fflBBBBBBBBBBBQBB^l

S B3

Marine Biological Laboratory Library

3 Woods Hole, Mass. ffl

1 id

II ID

ii .~v. m

3 ID

B ID

U Presented by II]

D Ifl

II . ID

J Association of American

J University Presses [j]

J Aug. 24, 1963 I

D ID

D ID

E ffl

83 E3

3E3E3E3E3B3E3E3E3E3E3E3QE3E383E

THE LIFE OF

VERTEBRATES

THE LIFE OF

VERTEBRATES

BY

J. Z. YOUNG

M.A., F.R.S.

PROFESSOR OF ANATOMY AT
UNIVERSITY COLLEGE, LONDON

SECOND EDITION

OXFORD UNIVERSITY PRESS

NEW YORK & OXFORD

1962

© Oxford University Press ig^ 2
Library of Congress Catalogue Card Number: 62-21012

First Edition 1950
Second Edition 1962

Printed in the United States of America

ACKNOWLEDGEMENTS

My grateful thanks are due to the following, who spent much time reading and
criticizing various parts of the manuscripts and proofs:

E. H. Ashton, A. d'A. Bellairs, Q. Bone, B. B. Boycott, P. M. Butler, S. Crowell,

D. H. Cushing, F. C. Fraser, R. B. Freeman, H. Greenwood, I. Griffiths, R. J.
Harrison, R. A. Hinde, K. A. Kermack, J. Lever, N. B. Marshall, D. R. Newth,
C. Nicol, F. R. Parrington, P. Robinson, A. J. Sutcliffe, H. G. Vevers, E. I.
White, M. Whitear.

Thanks for permission for reproductions of illustrations are due to:

G. C. Aymar, E. J. W. Barrington, J. Berrill, T. H. Bullock, A. J. E. Cave,
W. E. Le Gros Clark, E. Crosby, F. G. Evans, Helen Goodrich, A. Gorbman,
J. Gray, W. K. Gregory, W. J. Hamilton, J. E. Harris, L. Hogben, A. Holmes,

E. Hoskings, W. W. Howells, J. S. Huxley, F. Knowles, D. Lack, N. A. Mackin-
tosh, G. H. Parker, A. T. Phillipson, R. J. Pumphrey, E. C. R. Reeve, A. S.
Romer, E. S. Russell, F. K. Sanders, A. H. Schultz, G. G. Simpson, N. Tinber-
gen, G. L. Walls, L. Waring, T. S. Westoll, E. I. White, F. F. Zeuner,

and to the following publishers and other bodies:

The American Museum of Natural History; Bailliere, Tindall and Cox; E.
Benn, Ltd. ; Biological Reviezvs ; the British Medical Journal; the British Museum ;
the Cambridge University Press; the Company of Biologists, Ltd.; Dodd, Mead
& Co.; Editors of the Ibis; Longmans, Green & Co., Ltd.; the Physiological
Society; Putnam, Ltd.; the Scientific Monthly; the Wilson Bulletin; the Wistar
Institute; Zoological Society of London.

PREFACE TO
THE SECOND EDITION

For this edition every part of the book has been revised and corrected,
but the basic plan and balance of interests have not been altered.
Changes of arrangement and emphasis might have suited some types
of reader but I have thought it better that the book should continue
to show the idiosyncracies and interests of the author. One of the
dangers of a textbook is, surely, that unsophisticated readers may sup-
pose that they are getting the authentic and complete treatment of the
subject. Some obvious imbalances may, therefore, even be an advan-
tage as reminders of the relativity of all statements.

Nevertheless, I have attempted to make the treatment rather more
complete and systematic than before. For example, the descriptions
of the parts of the body are now arranged more nearly similarly for all
groups. With the help of many friends mistakes have been removed
and accounts of recent work added. The anatomy of Mammals is not
dealt with in the same detail as that of other groups, being covered
separately in The Life of Mammals (Clarendon Press, 1957), where
also there is a fuller account of the comparative embryology of
vertebrates.

During the revision I have become even more conscious of the
defects of the work, both in general form and detailed treatment. It is
still not possible to see more than the vaguest outlines of a proper
science of comparative biology. We are faced with a great series of
wonderful systems, differing slightly from each other and maintaining
themselves in slightly different surroundings. But we have no proper
scientific words with which to talk about them. For example, it is
absurd that this book contains so little reference to genetics, bio-
chemistry, or control theory. No doubt this is partly my fault, but the
fact is that these more exact sciences have yet to show us how to treat
the organization of a whole creature.

Fortunately, the animals remain as fascinating as ever, indeed the
search for exact ways of describing makes them even more so. Those
of us who have revised the book will be well rewarded for our trouble
if others arc helped to look and think for themselves. If they do they
will find a really astonishing array of experiments made by natural
selection with every part of the vertebrate organization. To take one
example, we are offered the opportunity to learn how the endocrine

vi PREFACE TO SECOND EDITION

control system works by examining hundreds of different variants of
it. The more one thinks of it the more surprising it is that biology
has made so little use of the experiments that have been done for us
by nature. Surely soon someone will come along with sufficient know-
ledge and logical and mathematical ingenuity to show us how to study
vertebrate organization.

Besides those mentioned below who have assisted in the revision of
particular sections, I should like to thank the many people, including
teachers and students, who have written about particular points, and
especially Professor J. Lever of Amsterdam for his many detailed
comments. My grateful thanks are also due to Mrs. J. Astafiev, who
has redrawn many of the figures, Mr. C. Marmoy for assistance with
the Bibliography, Mr. P. N. Dilly, who has helped throughout, also
my secretaries and especially Miss S. Thistleton and Miss J. Everard,
for continuous help with the manuscript. It is also a pleasure to thank
the members of the Clarendon Press and in particular Miss M. Gregory
for the help with the revision.

J. Z. Y.

February 1962

PREFACE TO
THE FIRST EDITION

The history of textbooks is often dismissed by the contemptuous
assertion that they all copy each other — and especially each other's
mistakes. Inspection of this book will quickly confirm that this is true,
but there is nevertheless an interest to be obtained from such a study,
because textbooks embody an attitude of mind; they show what sort
of knowledge the writer thinks can be conveyed about the subject-
matter. It may be that they are more important than at first appears
in furthering or preventing the change of ideas on any theme.

The results of the studies of scholars on the subject of vertebrates
have been summarized in a series of comprehensive textbooks during
the past hundred years. Most of these works are planned on the lines
laid down by the books of Gegenbaur (1859), Owen (1866), and
Wiedersheim (1883), lines that derive from a pre-evolutionary tradi-
tion. This partly explains the curiosity that in spite of the great impor-
tance of evolutionary doctrine for vertebrate studies, and vice versa,
vertebrate textbooks often do not deal directly with evolution. They
derive their order from something even more fundamental than the
evolutionary principle. The essential of any good textbook is that it
should be both accurate and general. As Owen puts it in his Preface:
Tn the choice of facts I have been guided by their authenticity and
their applicability to general principles.' The chief of the principles
he adopted was 'to guide or help in the power of apprehending the
unity which underlies the diversity of animal structures, to show in
these structures the evidence of a predetermining Will, producing
them in reference to a final purpose, and to indicate the direction and
degrees in which organisation, in subserving such Will, rises from the
general to the particular'. He confessed 'ignorance of the mode of
operation of the natural law of their succession on the earth. But that
it is an "orderly succession" — and also "progressive" — is evident
from actual knowledge of extinct species.'

These principles were essentially sound, and Owen's treatment was
to a large extent the basis of the work that appeared after the Dar-
winian revolution. In English, following the translation of Wieder-
sheim's book by W. N. Parker (1886) we have H. J. Parker and
Haswell's work, now in its 6th edition. The books of Kingsley and
Neal and Rand are in essentially the same tradition, though they

viii PREFACE TO FIRST EDITION

incorporate much new work, especially from the neurological studies
of Johnston and Herrick. Further exact studies on these same general
morphological lines made possible the books of Goodrich (1930) and
de Beer (1935), which have provided the morphological background
for the present work. Throughout these works on Comparative Ana-
tomy the emphasis is on the evolution of the form of each organ
system rather than on the change of the organization of the life of
the animal as a whole.

Meanwhile many other treatises appeared dealing with the life and
habits of the animals, rather than with morphological principles.
Among these we may mention Bronn's Tierreich (1859 onwards), the
Cambridge Natural History, and many works dealing with particular
groups of vertebrates. The palaeontologists produced their own series
of textbooks, mainly descriptive, such as those of Zittel and Smith
Woodward, culminating in Romer's admirably detailed and concise
book, to which the present work owes very much. The results of
embryological work have been summarized by Graham Kerr (191 9),
Korscheldt and Heider (1931), Brachet (1935), Huxley and de Beer
(1934), and Weiss (1939), among others. Unfortunately there has been
little summarizing of what is commonly called the comparative physio-
logy of vertebrates. Winterstein's great Handbuch der vergleichenden
Physiologie (191 2) covers much detailed evidence, but comes no nearer
than do the comparative anatomists to giving us a picture of the
evolution of the life of the whole organism.

All of these books deal in some way with the evolution of vertebrates,
and vet curiously enough they speak of it very little. It is hardly an
exaggeration to say that they leave the student to decide for himself
what has been demonstrated by their studies. Huxley's Anatomy of
Vertebrated Animals (1871) is an exception in that it deals with the
animals rather than their parts, and at a more popular level. Brehm's
Thierleben (1876) gives a picture of the life of the animals, though in
this case not of their underlying organization. Kukenthal's great
Handbuch der Zoologie has the aim of synthesizing a variety of know-
ledge about each animal-group, and some of the volumes dealing with
vertebrates make fascinating reading- — notably that of Streseman on
birds. But the size of the work and the multiplicity of authors make
it impossible for any general picture of vertebrate life to appear from
the mass of details.

The position is, then, that we have good descriptions of the struc-
ture, physiology, and development of vertebrates, of the discoveries
of the palaeontologists and accounts of vertebrate natural history, but

PREFACE TO FIRST EDITION ix

that there is no work that attempts to define the organization of the
whole life and its evolution in all its aspects. Indeed, none of these
works defines what is being studied or tries to alter the direction of
investigation — all authors seem prepared to agree that biological study
is adequately expressed through the familiar disciplines of anatomy,
physiology, palaeontology, embryology, or natural history. In passing,
we may note the extraordinary fact that there are no detailed works on
the comparative histology or biochemistry of vertebrates — surely most
fascinating fields for the future, as is, indeed, hinted by the attempts
that have been made in older works, such as that of Ranvier (1878),
and the newer ones of Baldwin (1937 and 1945).

The present book has gradually grown into an attempt to define
what is meant by the life of vertebrates and by the evolution of that
life. Put in a more old-fashioned way, this represents an attempt to
give a combined account of the embryology, anatomy, physiology,
biochemistry, palaeontology, and ecology of all vertebrates. One of
the results of the work has been to convince me more than ever that
these divisions are not acceptable. All of their separate studies are
concerned with the central fact of biology, that life goes on, and I
have tried to combine their results into a single work on the way in
which this continuity is maintained.

A glance through the book will show that I have not been successful
in producing anything very novel — others will certainly be able to
go much farther, and in particular to introduce to a greater extent facts
about the evolution of the chemical and energy interchanges of verte-
brates, here almost omitted! However, I have very much enjoyed the
attempt, which has provided the stimulus to try to find out many
things that I have always wanted to know.

For any one person to cover such a wide field is bound to lead to
inexactness and error in many places. I have tried to verify from
nature as often as possible, but a large amount has been copied, no
doubt often wrongly. Throughout, the aim has been to provide
wherever possible an idea of the actual observations that have been
made, as well as the interpretations placed upon them. A proper
appraisal of general theories can only be reached if there is first a
knowledge of the actual materials, which is the characteristic feature
of scientific observation. A book such as the present has value only
in so far as it leads the reader to make his own observations and helps
him to know the world for himself.

Mammalian organization requires more detailed treatment than
that of other groups, and in providing this the work grew to beyond

x PREFACE TO FIRST EDITION

the length of a single book. Mammalian structure, function, and
development will therefore be dealt with in a separate volume, which
will also include a survey of comparative embryology.

The original plan was that the palaeontological parts of the book
would be written by J. A. Moy-Thomas. Had he lived this aspect of
the work would have been very much better, and his common sense
and laughter would have lightened the whole. I have tried to give
some compensation at least by the speculation that is possible from a
single point of view. To protect the reader against the limitations of
my ignorance I have consulted specialists on every part of the work,
and my deepest thanks are due to those who have helped in this way.
They have done wonders in correcting mistakes, but, of course, are
not responsible for any that remain, or for views expressed. Among
those who have helped in this way with particular parts are Professor
G. R. de Beer, Mr. R. B. Freeman, the late Professor W. Garstang,
Dr. A. Graham, Professor J. B. S. Haldane, Professor W. Holling-
worth, Dr. W. Holmes, Dr. J. S. Huxley, Dr. D. Lack, Mr. Maynard
Smith, Dr. F. S. Russell, Dr. Tyndell Hopwood, Mr. H. G. Vevers,
Professor D. M. S. Watson, and Professor S. Westoll. They have been
patient and severe critics, and the reader and I owe them very much.

One of the main problems of such a work is its illustration, and here
I have been extraordinarily fortunate in having the help of Miss E. R.
Turlington, who has not only provided brilliantly clear and beautiful
pictures, but has taken extremes of care to ensure their accuracy by
drawing from live animals, from dissections, and from skeletons, as
well as by research into the illustrations of others. Miss J. de Vere
has also given much help with drawing. We have borrowed good
pictures unhesitatingly and should like to thank those who have given
permission for their reproduction.

I should also like to thank particularly my secretary, Miss P. Codlin,
who has played a large part in making the book possible, and my
daughter Cordelia for help with the index.

Finally, I have to thank the Secretary and members of Oxford
University Press for the care with which the book has been produced,
and for their friendly co-operation, which has made the work a
pleasure.

J. Z. Y.

1950

CONTENTS

I. EVOLUTION OF LIFE IN RELATION TO CLIMATIC AND GEO-

LOGICAL CHANGE

i. The need for generality in zoology, i ; 2. What do we mean by the life of an
animal ? 2; 3. Li ving things tend to preserve themselves, 3; 4. What do we mean by
awareness of life ? 5 ; 5. The influence of environment on life, 7 ; 6. What is it that
heredity transmits ? 8 ; 7. The increasing complexity of life, 9 ; 8. The progression
of life from the water to more difficult environments, 9; 9. Changes of climate and
geological periods — (1) Changes of level of the continents, 11; (2) Changes of
climate, 13; (3) Geological time, 16; (4) Classification of geological history, 18;
10. Summary, 21.

II. THE GENERAL PLAN OF CHORDATE ORGANIZATION: AMPHI-

OXUS

1. The variety of chordate life, 23; 2. Classification of chordates, 24; 3. Amphi-
oxus, a generalized chordate, 24; 4. Movement of amphioxus, 26; 5. Skeletal
structures of amphioxus, 29; 6. Skin of amphioxus, 29; 7. Mouth and pharynx
and the control of feeding, 30; 8. Circulation, 33; 9. Excretory system of
amphioxus, 35; 10. Nervous system, 36; 11. Gonads and development of
amphioxus, 41; 12. Amphioxus as a generalized chordate, 46.

III. THE ORIGIN OF CHORDATES FROM FILTER FEEDING ANIMALS
1. Invertebrate relatives of the chordates, 47; 2. Subphylum Hemichordata
(= Stomochordata), 50; 3. Class Pterobranchia, 58; 4. Subphylum Tunicata.
Sea squirts, 60; 5. Development of ascidians, 66; 6. Various forms of tunicate,
69; 7. Class Ascidiacea. 70; 8. Class Thaliacea, 70; 9. Class Larvacea, 72;

10. The formation of the chordates, 74.

IV. THE VERTEBRATES WITHOUT JAWS. LAMPREYS

1. Classification, 81; 2. General features of vertebrates, 81; 3. Agnatha, 83;
4. Lampreys, 83; 5. Skeleton of lampreys, 85; 6. Alimentary canal of lampreys,
88; 7. Blood system of lampreys, 91; 8. Urinogenital system of lampreys, 93;
9. Nervous system of lampreys, 97; 10. The pineal eyes, 103; 1 1. Pituitary body
and hypophyseal sac, 106; 12. Lateral line organs of lampreys, 108; 13. Vesti-
bular organs of lampreys, 109; 14. Paired eyes of lampreys, no; 15. Skin photo-
receptors, in; 16. Habits and life-history of lampreys, 112; 17. The ammocoete
larva, 114; 18 Races of lampreys, a problem in systematics, 119; 19. Hag-fishes,
order Myxinoidea, 122; 20. Fossil Agnatha, the earliest-known vertebrates, 125.

V. THE APPEARANCE OF JAWS. THE ORGANIZATION OF THE HEAD
1. The elasmobranchs : introduction, 131; 2. The swimming of fishes, 133;
3. Equilibrium of fishes in water; the functions of the fins, 136; 4. Skin of
elasmobranchs, 141; 5. The skull and branchial arches, 142; 6. The jaws, 145;
7. Segmentation of the vertebrate head, 148; 8. The pro-otic somites and eye-
muscles, 149; 9. The cranial nerves of elasmobranchs, 152; 10. Respiration, 157;

11. The gut of elasmobranchs, 158; 12. The circulatory system, 159; 13. Urino-
genital system, 162; 14. Endocrine glands of elasmobranchs, 164; 15. Nervous
system, 167; 16. Receptor-organs of elasmobranchs, 170; 17. Autonomic nervous
system, 173.

81659

xii CONTENTS

VI. EVOLUTION AND ADAPTIVE RADIATION OF ELASMOBRANCHS
i. Characteristics of elasmobranchs, 175; 2. Classification, 175; 3. Palaeozoic
elasmobranchs, 176; 4. Mesozoic sharks, 180; 5. Modern sharks, 180; 6. Skates
and rays, 182; 7. Chimaera and the bradyodonts, 184; 8. Tendencies in elasmo-
branch evolution, 185; 9. The earliest Gnathostomes, Placoderms, 186.

VII. THE MASTERY OF THE WATER. BONY FISHES

1. Introduction: the success of the bony fishes, 190; 2. The trout, 191; 3. The
skull of bony fishes, 193; 4. Respiration, 196; 5. Vertebral column and fins
of bony fishes, 199; 6. Alimentary canal, 201 ; 7. Air-bladder, 201 ; 8. Circulatory
system, 201 ; 9. Urinogenital system and osmoregulation, 202; 10. Races of trout
and salmon and their breeding habits, 204; 1 1. Endocrine glands of bony fishes,
206; 12. Brain of bony fishes, 209; 13. Receptors for life in the water, 212;
14. Eyes, 212; 15. Ear and hearing of fishes, 216; 16. Sound production in
fishes, 218; 17. The lateral line organs of fishes, 218; 18. Chemoreceptors.
Taste and smell, 220; 19. Touch, 222; 20. Autonomic nervous system, 222;
21. Behaviour patterns of fishes, 225.

VIII. THE EVOLUTION OF BONY FISHES

1. Classification, 228; 2. Order 1. Palaeoniscoidei, 228; 3. Order 2. Acipen-
seroidei, 234; 4. Superorder 2. Holostei, 234; 5. Superorder 3. Teleostei, 236;

6. Analysis of evolution of the Actinopterygii, 237.

IX. THE ADAPTIVE RADIATION OF BONY FISHES

I. Swimming and locomotion, 244; 2. Various body forms and swimming habits
in teleosts, 248; 3. Structure of mouth and feeding-habits of bony fishes, 251;
4. Protective mechanisms of bony fishes, 252; 5. Scales and other surface armour,
252; 6. Spines and poison glands, 253; 7. Electric organs, 253; 8. Luminous
organs, 254; 9. Colours of fishes, 255; 10. Colour change in teleosts, 258;

II. Aerial respiration and the air-bladder, 261; 12. Special reproductive
mechanisms in teleosts, 265.

X. LUNG-FISHES

1. Classification, 268; 2. Crossopterygians, 268; 3. Osteolepids, 268; 4. Coela-
canths, 271; 5. Fossil Dipnoi, 273; 6. Modern lung-fishes, 275.

XI. FISHES AND MAN, 280

XII. TERRESTRIAL VERTEBRATES: AMPHIBIA

1. Classification, 296; 2. Amphibia, 296; 3. The frogs, 298; 4. Skin of Amphibia,
298; 5. Colours of Amphibia, 299; 6. Vertebral column of Amphibia, 303;

7. Evolution and plan of the limbs of Amphibia, 307; 8. Shoulder girdle of
Amphibia, 309; 9. Pelvic girdle of Amphibia, 312; 10. The limbs of Amphibia,
313; 1 1. The back and belly muscles of Amphibia, 318; 12. The limb muscles of
Amphibia, 322; 13. The skull of Stegocephalia, 325; 14. The skull of modern
Amphibia, 328; 15. Respiration in Amphibia, 332; 16. Respiration in the frog,
333; I 7- Respiratory adaptations in various amphibians, 334; 18. Vocal appara-
tus, 334; 19. Circulatory system of Amphibia, 335; 20. Lymphatic system of
Amphibia, 338; 21. The blood of Amphibia, 339; 22. Urinogenital system of
Amphibia, 340; 23. Digestive system of Amphibia, 342; 24. Nervous system
of Amphibia, 344; 25. Skin receptors, 349; 26. The eyes of Amphibia, 350; 27.
The ear of Amphibia, 353; 28. Behaviour of Amphibia, 354.

CONTENTS xiii

XIII. EVOLUTION AND ADAPTIVE RADIATION OF AMPHIBIA

i. The earliest Amphibia, 356; 2. Terrestrial Palaeozoic Amphibia. Embolomeri
and Rhachitomi, 357; 3. Aquatic Amphibia of the later Palaeozoic, 359;
4. Tendencies in the evolution of fossil Amphibia, 362; 5. Newts and Salaman-
ders. Subclass Urodela, 364; 6. Frogs and Toads. Subclass Anura, 365; 7. Sub-
class Apoda (— Gymnophiona = Caecilia), 366; 8. Adaptive radiation and
parallel evolution in modern Amphibia, 366; 9. Can Amphibia be said to be
higher animals than fishes ? 367.

XIV. LIFE ON LAND: THE REPTILES

1. Classification, 369; 2. Reptilia, 371; 3. The organization of reptiles, 372;
4. Skin of reptiles, 373; 5. Posture, locomotion, and skeleton, 373; 6. Feeding
and digestion, 378; 7. Respiration, circulation, and excretion, 378; 8. Reproduc-
tion of reptiles, 380; 9. Nervous system and receptors of reptiles, 383.

XV. EVOLUTION OF THE REPTILES

1. The earliest reptile populations, Anapsida, 386; 2. Classification of reptiles,
391; 3. Order 1. Chelonia, 392; 4. Subclass *Synaptosauria, 399; 5. Order
*Ichthyopterygia, 401 ; 6. Subclass Lepidosauria, 401 ; 7. Order Rhynchocephalia,
402; 8. Order Squamata, 404; 9. Suborder Lacertilia, 407; 10. Suborder
Ophidia, 411; 11. Superorder Archosauria, 416; 12. Order *Pseudosuchia, 417;
13. Order *Phytosauria, 417; 14. Order Crocodilia, 418; 15. The 'Terrible
Lizards', Dinosaurs, 421; 16. Order *Saurischia, 422; 17. Order *Ornithischia,
424; 18. Order *Pterosauria, 426; 19. Conclusions from study of evolution of the
reptiles, 429.

XVI. LIFE IN THE AIR: THE BIRDS

1. Features of bird life, 431 ; 2. Bird numbers and variety, 431 ; 3. The skin and
feathers, 432; 4. Colours of birds, 436; 5. The skeleton of the bird. Sacral and
sternal girders, 437; 6. The sacral girder and legs, 440; 7. Skeleton of the wings,
447; 8. Wing muscles, 449; 9. Principles of bird flight, 450; 10. Wing shape,
452; 1 1. Wing area and loading, 452; 12. Aspect ratio, 453; 13. Wing tips, slots,
and camber, 453; 14. Flapping flight, 455; 15. Soaring flight, 458; 16. Soaring
on up-currents, 458; 17. Use of vertical wind variations, 460; 18. Speed of
flight, 461; 19. Take-off and landing, 462; 20. The skull in birds, 464; 21. The
jaws, beak, and feeding mechanisms, 464; 22. Digestive system of birds, 468;
23. Circulatory system, 470; 24. Respiration, 471; 25. Excretory system, 474;
26. Reproductive system, 475; 27. The brain of birds, 477; 28. Functioning of the
brain in birds, 479; 29. The eyes of birds, 482; 30. The ear of birds, 488; 31.
Other receptors, 490.

XVII. BIRD BEHAVIOUR

I. Habitat selection, 491; 2. Food selection, 491; 3. Recognition and social
behaviour, 492; 4. Bird migration and homing, 493; 5. The stimulus to migra-
tion, 495; 6. The breeding-habits of birds, 496; 7. Courtship and display, 497;
8. Bird territory, 503; 9. Mutual courtship, 504; 10. Nest-building, 505;

I I. Shape and colour of the eggs, 507; 12. Brooding and care of the young, 507.

XVIII. THE ORIGIN AND EVOLUTION OF BIRDS

1. Classification, 509; 2. Origin of the birds, 510; 3. Jurassic birds and the
origin of flight, 510; 4. Cretaceous birds. Superorder Odontognathae, 513;

xiv CONTENTS

5. Flightless birds. Superorder Palaeognathae, 514; 6. Penguins. Superorder
Impennae, 515; 7. Modern birds. Superorder Neognathae, 516; 8. Tendencies
in the evolution of birds, 522 ; 9. Darwin's finches, 524 ; 1 o. Birds on other oceanic
islands, 530; 11. The development of variety of bird life, 532.

XIX. THE ORIGIN OF MAMMALS

1. Classification, 533; 2. The characteristics of mammals, 534; 3. Mammals of
the Mesozoic, 536; 4. Mammal-like reptiles, Synapsida, 539; 5. Order *Pely-
cosauria (= Theromorpha), 540; 6. Order *Therapsida, 541; 7. Mammals from
the Trias to the Cretaceous, 545; 8. Original cusp-pattern of teeth of mammals,
548; 9. Egg-laying mammals. Subclass Prototheria (Monotremata), 549.

XX. MARSUPIALS

1. Marsupial characteristics, 557; 2. Classification of marsupials, 562; 3. Opos-
sums, 563; 4. Carnivorous marsupials, 565; 5. Marsupial ant-eaters and other
types, 566; 6. Phalangers, wallabies, and kangaroos, 566; 7. Significance of
marsupial isolation, 568.

XXI. EVOLUTION OF PLACENTAL MAMMALS AND ITS RELATION
TO THE CLIMATIC AND GEOGRAPHICAL HISTORY OF THE
CENOZOIC

1. Eutherians at the end of the Mesozoic, 569; 2. The end of the Mesozoic, 569;
3. Divisions and climates of the Tertiary Period, 571; 4. Geographical regions,
572; 5. The earliest eutherians, 574; 6. Definition of a eutherian (placental) mam-
mal) 575; 7- Evolutionary trends of eutherians, 575; 8. Conservative eutherians,
577; 9. Divisions and classification of Eutheria, 577.

XXII. INSECTIVORES, BATS, AND EDENTATES

1. Order 1. Insectivora, 581; 2. Order Chiroptera. Bats, 585; 3. Order Dermo-
ptera, 592; 4. Order Edentata, 592; 5. Armadillos, 595; 6. Ant-eaters and sloths,
597; 7. Order Pholidota: pangolins, 601.

XXIII. PRIMATES

1. Classification, 602; 2. Characters of primates, 603; 3. Divisions of the pri-
mates, 607; 4. Lemurs and lorises, 609; 5. Fossil Prosimians, 613; 6. Tarsiers,
614; 7. Characteristics of Anthropoidea, 617; 8. New World monkeys, Ceboidea,
620.

XXIV. MONKEYS, APES, AND MEN

1. Common origin of Old World monkeys, apes, and men, 623; 2. Old World
monkeys, Cercopithecoidea, 623; 3. The great apes: Pongidae, 626; 4. The
ancestry of man, 633; 5. Brain of apes and man, 633; 6. The posture and gait of
man, 634; 7. The limbs of man, 635; 8. The skull and jaws of man, 637; 9. Rate
of development of man, 640 ; 10. Growth of human populations, 641 ; 1 1 . Time of
development of the Hominidae, 641; 12. The Australopithecinae, 643; 13. Early
Hominids, *Pithecanthropus, 645; 14. Man, 646; 15. Human cultures, 648.

XXV. RODENTS AND RABBITS

1. Characteristics of rodent life, 652; 2. Classification, 653; 3. Order Rodentia,
654; 4. Order Lagomorpha, 660; 5. Fluctuations in numbers of mammals, 663.

CONTENTS xv

XXVI. WHALES, 666.

XXVII. CARNIVORES

i. Affinities of carnivores and ungulates: Cohort Ferungulata, 677; 2. Classifica-
tion, 679; 3. Order Carnivora, 680; 4. The Cats, 680; 5. *Suborder Creodonta,
683; 6. Suborder Fissipeda, 684; 7. Suborder Pinnepedia, 691.

XXVIII. PROTOUNGULATES

1. Origin of the ungulates, 694; 2. Ungulate characters, 695; 3. Classification, 699;

4. Superorder Protoungulata, 700; 5. South American ungulates. *Order Notoun-
gulata, 701; 6. *Order Litopterna, 703; 7. *Order Astrapotheria, 703; 8. Order
Tubulidentata, 704.

XXIX. ELEPHANTS AND RELATED FORMS

1. 'Near-ungulates', superorder Paenungulata, 706; 2. Classification, 706;
3. Order Hyracoidea, 707; 4. Elephants. Order Proboscidea, 709; 5. *Order
Pantodonta (Amblypoda), 717; 6. *Order Dinocerata, 718; 7. *Order Pyro-
theria, 718; 8. *Order Embrithopoda, 718; 9. Order Sirenia, 720.

XXX. PERISSODACTYLS

1. Perissodactyl characteristics, 722; 2. Classification, 723; 3. Perissodactyl
radiation, 724; 4. Suborder Ceratomorpha, tapirs and rhinoceroses, 727;

5. Rhinoceroses, 728; 6. *Brontotheres (*Titanotheres), 730; 7. *Chalicotheres
(= *Ancylopoda), 731; 8. Palaeotheres, 732; 9. Horses, 732; 10. Allometry in
the evolution of horses, 737; 11. Rate of evolution of horses, 738; 12. Conclu-
sions from the study of the evolution of horses, 739.

XXXI. ARTIODACTYLS

1. Characteristics of artiodactyls, 741; 2. Classification, 745; 3. The evolution
of artiodactyls, 746; 4. Pigs and hippopotamuses, 748; 5. *Oreodonts, 750;

6. Camels, 751; 7. Ruminants, 753; 8. Chevrotains, 754; 9. Pecora, 755;
10. Cervidae, 755; II. Giraffidae, 757; 12. Antilocapridae and Bovidae, 760.

XXXII. CONCLUSION. EVOLUTIONARY CHANGES OF THE LIFE OF
VERTEBRATES

1. The life of the earliest chordates, 765; 2. Comparison of the life of early
chordates with that of mammals, 767; 3. The increasing complexity and variety
of vertebrates, 768; 4. The variety of evidence of evolutionary change, 769;
5. Rate of evolutionary change, 770; 6. Vertebrates that have evolved slowly,
771; 7. Varying rates of evolutionary changes, 774; 8. Vertebrates that have
disappeared, 774; 9. Successive replacement among aquatic vertebrates, 775;
10. Successive replacement among land vertebrates, 776; 11. Is successive re-
placement due to climatic change?, 776; 12. Convergent and parallel evolution,
777'. ! 3- Some tendencies in vertebrate evolution, 779; 14. Evolution of the
whole organization, 780; 15. Summary of evidence about evolution of verte-
brates, 781; 16. Conservative and radical influences in evolution, 783; 17. The
direction of evolutionary change, 784; 18. The influences controlling evolutionary
progress, 785.

REFERENCES 787

INDEX 797

THE LIFE OF

VERTEBRATES

I

EVOLUTION OF LIFE IN RELATION TO CLIMATIC
AND GEOLOGICAL CHANGE

1 . The need for generality in zoology

The aim of any zoological study is to know about the life of the ani-
mals concerned. Our object in this book is, therefore, to help the
reader to learn as much as possible about all the vertebrate animal life
that has ever been. Thinking of the great numbers of types that have
existed since the first fishes swam in the Palaeozoic seas, one might
well be appalled by such a task: to describe all these populations in
detail would indeed demand a huge treatise. However, in a well-
developed science it should be possible to reduce the varied subject-
matter to order, to show that all differences can be understood to have
arisen by the influence of specified factors operating to modify an
original scheme. Animal and plant life is so varied that it has not yet
proved possible to systematize our knowledge of it as thoroughly as
we should wish. Thinking, again, of the variety of vertebrate lives,
it may seem impossible to imagine any general scheme and simple set
of factors that would include so many special circumstances. Yet
nothing less should be the aim of a true science of zoology. Too often
in the past we have been content to accumulate unrelated facts. It is
splendid to be aware of many details, but only by the synthesis of
these can we obtain either adequate means for handling so many
data or knowledge of the natures we are studying. In order to know
life — what it is, what it has been, and what it will be — we must
look beyond the details of individual lives and try to find rules govern-
ing all. Perhaps we may find the task less difficult than expected.
Even an elementary anatomical and physiological study shows that
all vertebrates are built upon a common plan and have certain simi-
larities of behaviour. Our object will be to come to know the nature
of this plan of life, of structure, and action, to show how it is modified
in special cases and how each special case is also an example of a
general type of modification.

Since the problem arises from the variety of animals that have
lived and live today, our central task is obviously to inquire into the
reason for the existence of so much difference. If vertebrate life began
as one single fish-like type, why has it not continued as such until
now? Why, instead of numerous identical fishes, are there countless

2 EVOLUTION OF LIFE i. i-

different kinds, while descendants of most unfish-like form are found
living out of the water and even in the air and under the ground ?

To put it in a way more familiar, though perhaps less clear: what
are the forces that have produced the changes of animal form ? Know-
ing these forces, and the original type, it would be possible to con-
struct a truly general science of zoology, with sure premisses and
deductions. Even if we cannot reach this end, we should at least try,
hoping that after investigation of the biology of vertebrates it will be
possible to retain something more than a mass of detailed information.
At the end of such a study, if we deal with the subject right, we should
surely be better able to answer some of the fundamental biological
questions. We should be able to say something about the nature of
evolution and of the differences between types, to know whether there
have been rhythms of change at work to produce these differences,
and also — the acid test of any true science — to forecast how these
changes are likely to proceed in the future.

2. What do we mean by the life of an animal?

In biology we make much use of analogies, attempting to grasp the
nature of the processes at work by comparison with man-made
machines. We have a science of anatomy, which we are told is con-
cerned with the 'structure' of animals, and we feel that we understand
what 'structure' means. Physiology is the study of 'function', and
this, too, we seem to understand. We take the analogies from our
machines, which have what we call 'structure' and 'function'. How-
ever, the difficulty at once arises that the living things make and control
themselves. The whole scheme fails us when we ask what is it, then,
that we call the 'life' of the animal, and what is it that is passed on
from generation to generation, and that changes through the ages by
the process we call evolution ? It has gradually become apparent that
the body is not a fixed, definite 'structure' as it appears to casual
observation or when dissected. In life there is ceaseless activity and
change going on within the apparently constant framework of the body.
The movement of the blood is one sign of this activity, and since
Harvey's discovery of the circulation ( 1 628) we have learnt of innumer-
able others. Everyone knows that the skin is continually being renewed
by growth from below, and many other types of cell are similarly
replaced; for instance, red blood-cells last only for a few weeks in
man. Even in the cells that are not completely destroyed and replaced,
such as the nerve-cells, there is continual change of the molecules that
make up their substance. The full extent of this exchange has been

i. 3 DEFINITION OF LIFE 3

shown by using isotopes to discover for how long individual atoms
remain in the body; the work of Schoenheimer (1942), which by this
means first clearly established the rapidity of the turnover, is a classic
of modern biology.

There are no man-made machines that replace themselves in this
way, but in recent years there has been much study of machines that
control their own operations. Such work provides us with new analo-
gies and new mathematical techniques with which we can analyse
the control of living systems (see Yockey and Quastler, 1958). As yet
we have no means of grasping the enormously complicated network
of activities that constitutes a single life. Throughout this book, how-
ever, an attempt will be made to approach that end by use of certain
clues to help us to concentrate on significant features, to see the
rhythms or patterns common to the lives of the animals, and thus to
carry in mind many details. It is possible in this way to bring together
information collected by morphologists, geneticists, embryologists,
physiologists, biophysicists, and biochemists to give a single view of
the life of the organisms concerned. The task is admittedly a hard one
and the success achieved only partial. Continually one slips into the
discussion of particular structures, substances, or processes, forgetting
the whole life. A detail of form or of chemical composition attracts,
and thus distracts, attention; perhaps it can hardly be otherwise if we
are to describe exactly. But it is surprising how practice improves the
powers of selecting and emphasizing those patterns or details of
knowledge that are significant for the study of each life as a whole.

The first difficulty is to force oneself to remember all the time that
a living animal or plant system is in a continual state of change. When
making any observations, whether by dissection or with the micro-
scope, with a test-tube, microelectrode, or respirometer, it is necessary
continually to think back to the time when the tissue was active in the
living body, and to frame the observation so that it shall reveal some-
thing significant of that activity. This means that every biologist
must know as much as possible of the life of the whole organism with
which he deals ; indeed, something of the whole population from which
the specimen was drawn.

3. Living things tend to preserve themselves

The clue by which we recognize significant features during any
biological study is that living activity tends to ensure the continuance
of its own pattern. The processes of life draw materials into the
system, organize them there, and then send them out again, all in such

4 EVOLUTION OF LIFE 1.3-

a way that the arrangement or pattern of the processes remains almost
unchanged as the molecules pass through it. We see analogies in the
way that a waterfall or a human institution such as the Catholic
Church remains the same, though its components change. Our
business is to try to describe this arrangement or pattern of processes
that is preserved. It is this pattern that we call the life of the species.
The activities that go to make up one sort of life are not necessarily
all to be found in any one individual, still less in any part of an in-
dividual. The pattern is not to be seen in any single creature or
part. Though we speak of 'individuals' they are no more the final
units than are the cells, the heart, or the brain, the bones, hair, or
nails. A whole interbreeding population is the unit of life that tends
to preserve the type, assisted, in social species, by individuals that
play a part in the life without participation in reproduction, such as
worker-bees.

A wide range of activities, therefore, goes to make up any one type
of life, and we shall only appreciate these activities properly if we study
that whole life as it is normally lived in its proper environment. The
way to study animals or men is, first and foremost, to examine them
whole, to see how their actions serve to meet the conditions of the
environment and to allow preservation of the life of the individual and
the race. Then, with this knowledge of how the animal 'uses' its parts
we may be able to make more detailed studies, down to the molecular
level, and show how together the activities form a single scheme of
action.

A living animal is continually doing things. Even when it is asleep
it is breathing, its heart beating and brain pulsing, while countless
chemical changes go on throughout its tissues. The waking life, of
course, shows this restless action even more clearly. Animals may
indeed sometimes be still, but they are never wholly inactive. It is not
difficult to see startling glimpses of this activity if we watch animals
alive, especially when they are in groups. A hawk wheeling, a pond
full of tadpoles, or a crowd of people moving on a city street will
remind us that if we are to see the interesting side of life we have to
study activity and not, as is more easy to do and so often done, to
spend all our time examining the 'structure' or 'chemistry' of the dead.

The peculiarity of this activity of animals is that much, perhaps
most, of it has the effect of maintaining the integrity of the body and,
indeed, even of increasing the bulk, or of reproducing more bodies
like the first (homeostasis). The search for food provides raw materials
giving to the muscles energy for further search. If the situation

i. 4 ACTIVITIES OF LIVING THINGS 5

demands still greater efforts these efforts will themselves lead to
'hypertrophy', or increase in the muscle substance and power. Simi-
larly, the muscular movements of respiration provide the oxygen by
which these same movements and others are made possible.

We could go on indefinitely describing how the activity of each part
of the body tends, with some exceptions, to ensure the continuation
of the whole. The mere statement of the existence of this tendency to
self-maintenance does not, perhaps, sufficiently emphasize the power
that it represents. It is one of the great 'forces' that control the
matter of the earth. It causes huge masses of material to be moved
annually to the tops of high trees and millions of wonderfully built
animals to roam daily to find and consume uncounted tons of food
or, not finding it, to search on and maintain their activity while any
calorie remains available. The power of life is sufficient to bring about
the incorporation of an appreciable part of the matter of the earth's
surface into living things. Within the appropriate range of conditions,
found chiefly near the surface of the sea and on the damper parts of
the earth, life dominates the lifeless and provides a main influence on
the matter present.

Animals and plants are able to take these actions that tend to their
own preservation because they contain stores of information about
the conditions that are likely to be met with and the means by which
adverse changes may be prevented. A fish is born with a body so
shaped that it may swim, a gull can soar on air currents, and a monkey
leap from branch to branch. Every type may thus be said to represent
the environment in which it lives, that is to say, it has a hereditary
store of information about it. Moreover, this hereditary store provides
it with receptor organs and brain with which it can acquire further
information during its lifetime. The study by engineers of the means
by which information may be coded, transmitted, and stored has
provided biologists with further means for study of the living memory
stores, which are comparable in some ways with those of machines.

4. What do we mean by awareness of life ?

A man states that he is aware that he is alive. He says that he knows
his needs and that he feels satisfaction when they are fulfilled. One
of the most difficult problems of biology is to decide how to relate
such statements about 'subjective experience' to what may be called
the 'objective' descriptions of science. This is clearly a philosophical
problem too large and important to be discussed properly here but
it must be approached. Perhaps it begins to find a solution when we

6 EVOLUTION OF LIFE 1.4-

remember that in speaking of all these matters we are using the words
of a conventional code, trying with them to convey information to
our fellows or somehow to influence them. Then we shall stop
asking such questions as 'what is consciousness ?' substituting 'what
sort of information does he transmit when he says "I am conscious" ?'

This will help with the particular aspect of the problem that con-
cerns us here. In trying to define what we mean by the 'life' of an
animal should we assume that, in addition to the actions of its body,
which we describe, there are also actions of some other entity, its
'mind' ? It is true that we should feel that any description of our own
lives that left out 'awareness' was ludicrously incomplete. Since we
have evolved from animals, so the argument runs, why should we
deny that they have some form of 'consciousness' ? This seems logical
but overlooks that the essential feature of statements such as T am
aware that I am alive' or T feel pain' is that they are part of the
means by which man, the communicating animal, controls and
influences his environment. Statements are part of human life, just
as swinging from branches is a feature of the life of monkeys and
flying of birds. It will at once be objected that these animals also
communicate, but the point is that communication must be considered
as a part of the life system of each animal, like respiration, locomotion,
or reproduction.

This leaves us with the baffling problem of finding words with
which to describe the describing system itself. Where indeed can we
find a sure basis from which to start? Here, I think, we can only
proceed by humbly admitting both ignorance and inadequacy. In
our language we have a communication system with which we can
convey to each other incomparably more information than passes
between other animals. Our system is improving every year, but it is
still grossly inadequate to describe the more subtle features of the
world, and especially of living things. We may show the greatest
respect for the depth of these mysteries by recognizing that they are
still too great for us to describe in our simple language. To provide a
good description of all the marvellous features of the life of a man or
an animal requires a complicated and subtle terminology, for which
we are striving. In pre-scientific language all such problems are
simplified by supposing that the actions of any system are produced
by some agent rather like a human being that resides within it. Thus
a child says that the clouds move 'because they want to'. So we are
accustomed to say that the body moves because it is guided by 'the
mind'. This may indeed be the best way of speaking for some

i. 5 INFLUENCE OF ENVIRONMENT 7

occasions, with our imperfect language, but it is a feeble descriptive
technique. I do not believe that it is satisfactory for biology and
especially not for zoology. By the life of an animal we mean all those
activities that make a certain pattern and serve to maintain that
pattern. In so far as we can describe this as a whole it is by comparing
it with other self-maintaining systems and particularly with those
self-controlling machines that we have made for ourselves. Biology
today has a great opportunity to explore the means by which animals
remain alive, using many sorts of descriptive technique, chemical,
electrical and, not least, the means by which mathematicians and
engineers describe whole complicated self-maintaining systems. It
is in such language that a fuller and richer account of living things
can be given. It is curious that objections to the use of scientific
terminology often claim that it somehow 'reduces' or 'restricts' our
view of life. Exactly the reverse is the case. Explaining human or
animal life in terms of 'spirits', good or bad, is only describing them
by comparison with themselves. Scientific description allows us to
break out of our narrow prison and to show how each of the many
aspects of life can be measured and compared with the forces that can
be detected throughout the universe.

5. The influence of environment on life

Growth is the addition of material to that which is already organized
into a living pattern. But the pattern is not fixed and invariate, even
throughout any one life. Each individual changes through its lifetime,
develops, as we say, and moreover is modified by the action upon it
of its surroundings. Those parts that are exercised by the interaction
of the animal's tendencies and the surrounding circumstances increase
in amount (hypertrophy), while any disused parts undergo atrophy or
reduction. The pattern is thus able to conform to a considerable
extent to the exigencies of change in the external world. It could be
imagined that a sufficiently plastic animal organization would be able
in this way, if its tendencies to survival were strong enough, to mould
itself to all the changes of climate through the millennia, so that a
great variety of animal tvpes would arise by use and disuse alone.
Only a limited degree of change is possible in this way, however, and
it is not such changes either of development or by the direct influence
of environment that we call 'evolution'. There is abundant evidence
that the result of such interaction between organism and environment
is not handed on in the genetic code. Acquired characters are not
inherited.

8 EVOLUTION OF LIFE 1.6-

6. What is it that heredity transmit?

What is passed on is a coded pattern or plan controlling the
organization of the life processes of the next generation. The plan
takes the physical form of a series of molecules of deoxyribose
nucleotides (DNA) in the chromosomes. These by the specific
arrangements of the four types of base that they contain somehow
organize the proper linear sequences of the twenty or so amino-acids
that make up the proteins of each species. By the emergence at the
proper time during development of the appropriate proteins, enzyme
systems are produced that ensure the development and functioning of
the embryo and later the adult. We cannot fully understand how all
these processes are regulated but we see in outline how it all follows
if the DNA molecules provide a code from which natural selection
has chosen in the past those items that are suitable to provide viable
organisms for a particular environment.

The organization of life is very rarely identical in any two indi-
viduals; there is, therefore, a considerable range of potential patterns
resident in all those animals of a population that are capable of mating
together. The sum of those variants of the hereditary materials con-
stitutes the pattern or mould, as it were, of the life of the whole species.
Evolution consists in a change in this hereditary genotype, producing,
of course, a new set of adults. The genotype probably rarely stays for
long quite the same. Even in species that do not seem to be changing
rapidly there are continual adjustments, for example in the power to
produce antibodies or to manufacture enzymes. Evolution, proceeding
by mutation, recombination, and selection, is not some remote or
rare thing occurring only sporadically. It is a 'physiological' process as
much as is a change in respiration rate or in number of red cells, but
it has a longer time course than these. Evolution is the process by
which the whole population adjusts its control system to meet chang-
ing needs. Over long periods of years these adjustments produce the
new forms of life that appear as, say, the first fishes, or land animals
or mammals. Our aim is to try to discover the conditions under
which each new main group of vertebrates arose and so to understand
the processes that have been at work, modifying the basic organization.

We must therefore direct our studies continually to populations,
rather than to single individuals, thinking of all the creatures of a
kind, spread out wherever a suitable habitat for them occurs. They
will not all be alike genetically, and the circumstances of the lives of
some members of the group may become sufficiently dissimilar to

i.8 INCREASING COMPLEXITY OF LIFE 9

produce further divergences by use and disuse. Limitations of inter-
mating may occur on account of limitation of movement, accentuated
by partial and, perhaps, eventually complete geographical barriers.
Such variations in external circumstance become matched by diver-
gences in type, until two new races are produced, at first relatively
and then absolutely infertile, so that there are then two separate
populations or species instead of one.

7. The increasing complexity of life

The acquisition of new matter, and hence growth and reproduction,
occurred in the earliest animals by relatively simple means, as it still
does today in the bacteria, lower plants, and some protozoa. It is not
easy to provide rigid criteria for the definition of 'simple'; perhaps
some of the chemical changes involved may be quite complex, but the
whole system can, with meaning, be said to be simple. The number
of parts that it contains is relatively small and the number of 'adap-
tive' actions that it can take is limited. A population of bacteria in a
suitable culture medium obtains its raw materials by diffusion; the
chief device that it uses to secure these materials is to provide a large
number of spores, so that some may come to rest in suitable sur-
roundings. Such a life can be said to be more simple than that of a
vertebrate, whose system includes many special devices for obtaining
access to the raw materials that it needs. We can say that a species of
bacteria transmits less information than a vertebrate. LInfortunately
there are no satisfactory counts of the number of genes available; the
amount of DNA in bacteria is said to be about 0-05 mgm per gm
and in rat liver 2 mgm per gm. Bacteria of any one species are able to
alter their enzymes to suit the substrates available, but their life does
not depend upon the differentiation into numerous cell types each
with its special functions. The variety of information available in the
'higher' genotypes enables them to take actions that ensure survival
under conditions where the 'lower' organisms would die. Of course,
each type has its own special 'niche' and the comparison of higher and
lower is only possible if we can show exact quantitative differences.

8. The progression of life from the water to more difficult environ-
ments

In general, the new environments colonized have involved ever
wider departures from that watery one in which life first arose. This
is shown most strikingly if we contrast the simple way in which the
means of life are obtained by a marine bacterium with the complicated

io EVOLUTION OF LIFE in-

activities that go to maintain a man alive in a city. Yet all living
systems, even those that have changed most markedly since their first
origin, are still watery, and must have salt and nitrogenous com-
pounds with which to make proteins and so on. Perhaps, indeed, the
basic plan of the living activity differs less in the various types than
one might suppose. 'Protoplasm' is certainly not identical in all
creatures, but it may be that it differs less than do the outward forms
that support it.

In order to provide the conditions necessary for the maintenance of
such a watery system, in very different environments, many auxiliary
activities have been developed. It is these that give added complexity
to the higher animals and plants, enabling them to undertake what
can be called more difficult ways of life. In order to do this their
activity must also be physically greater than is necessary in more
lowly types. It may be presumed that more energy is transferred to
maintain a given mass of living matter in the less 'easy' environments,
and in this sense the higher animals are less efficient than the lower,
by a very crude criterion of efficiency.

According to this conception, then, evolution has involved a change
in the relationship between organism and environment. Life has
come to occupy places in which it did not exist before. Perhaps the
total mass of living matter has thus been greatly increased. It must
not, of course, be supposed that every evolutionary change has pro-
duced an increase in complexity in this way; examples of 'degenera-
tion' are too well known to need quoting. We have, however, a clear
impression that through the years there has been, in general, some
change in animals and plants and that in a sense some of the later
organisms are 'higher' than the earlier. It is hardly possible to deny
that there is some meaning in the assertion that man is a higher
animal than amoeba. Our thesis attempts to specify more clearly what
we can know about this evolutionary change, by saying that it con-
sists of a colonization by life of environments more and more different
from that in which life arose. This colonization was made possible
by the gradual acquisition of a store of instructions enabling adjust-
ments to be made by which life could be maintained in conditions not
tolerable before.

It is not easy to enumerate the complexity of any animal or to
define quantitatively the nature of its relations with its environment,
and for this reason it is difficult to prove our thesis rigorously. This
book nevertheless makes an attempt to show how the organization of
vertebrate life has become more complex since it first appeared, and

1.9 INVASION OF NEW ENVIRONMENTS n

that the increasing complexity is related to the adoption of modes of
life continually more remote from the simple diffusion of substances
from the sea. Of course, even the earliest vertebrates had already
departed a long way from the first conditions of life and were quite
complex organisms. However, in the history of their life through
nearly 500 million years since the Ordovician period we can trace con-
siderable further changes in complexity. During this time vertebrate
life has left the sea to live in fresh water, on swampy land, and finally
on dry land and in the air. It has produced special types able to sup-
port life by such an astonishing variety of devices that we cannot
possibly specify them all. We shall only direct attention to a few, and
thus attempt to obtain an impression of the scheme of life of the vast
hordes of vertebrate animals, which, in one shape or another, have
swarmed and still swarm in the waters and over the earth. We shall
try to discern whether there is reason to suppose that all this variety
is related in some way to changes in the surrounding world and we
may therefore finish this introduction by a brief survey of the evi-
dences for climatic and geographical changes such as may have been
responsible for the changes in organic life.

9. Changes of climate and geological periods

9.1. Changes of level of the continents

Changes of geography are mostly so slow that they cannot in them-
selves influence individual lives. On the other hand, nearly all living
things must be suited to daily and annual cyclical changes, unless they
live where no light enters. There is indirect evidence of further
changes in climate and geography, occurring with such long periods
that they are without appreciable effect on individual organisms, but
may greatly affect the history of the race.

The idea of geographical change is made familiar by the fact that
coast-lines and river-courses have changed appreciably in historical
times. We are familiar with stories of destruction of some houses or
of a village by the sea, though it may come as a shock to learn that the
sea-level has changed so much that England and France were con-
nected by land 8,000 years ago, and that man-made instruments fished
up from the Dogger Bank show that it was an inhabited peat bog
6,000 years B.C. These changes in height of the land are signs of the
'diastrophic movements', which are major features of long-period
geological evolution. The earth forces that produce these movements
are still obscure but they lead to repeated elevation and sinking of the
land masses. The action of frost, wind, and rain continually breaks up

12 EVOLUTION OF LIFE 1.9

and carries away the surface of the land, at a rate of the order of 1 ft
per 4,000 years, the processes known as weathering and denudation.
The material carried away is deposited in the river-beds and in the
lakes and shallow seas around the river mouths (sedimentation) (Fig.
1). Here it builds the sedimentary rocks, which may be many thou-
sands of feet in thickness, the whole continental platform continuing
to sink for long periods, perhaps with intervals during which it be-
comes raised above the water. Fossil remains are usually the result of

km
8

6-
4-

8,840m.

High mountain
ranges

Average height ^
r of land tr X$

\0km

Fig. i. Curve showing the areas of the earth's solid surface in relation to the
sea level. (From Holmes.)

the preservation of the harder parts of animals in sedimentary deposits,
and the most complete series of fossils are likely to be those of animals
living in the seas.

The surface crust of the earth is not a layer of uniform thickness and
density but consists of irregular masses of lighter material, rich in
silicon and aluminium (sial), forming the continents, and heavier
material, rich in magnesium (sima), under the ocean beds. The reason
for this non-uniform distribution is obscure, but it has the effect of
making the continents stand higher, floating on the plastic denser
medium beneath the crust. When material is removed from the con-
tinents by denudation they rise; conversely the addition of millions
of tons of ice will depress them. The continents are thus said to rest
in isostatic equilibrium, and following the small changes in level the
sea leaves more or less of the continental shelf uncovered. Such
upward and downward movements profoundly influence the climate.
Oceanic climatic influences tend to produce a damp, equable climate,
with large areas of marsh and forest. When the land stands higher

i. 9 GEOLOGICAL CHANGES 13

extremes of climate develop, some parts being cold, others forming
large, dry interior plains.

Besides changes in the balance produced by denudation and the
advance of ice-caps there are also from time to time marked move-
ments of uplifting or lowering of the land, which may be called
independent earth movements. Such vertical movements of the con-
tinental masses are produced by internal forces of unknown origin.
They are doubtless related to a second series of major movements of
crustal deformation that are due to tangential forces and lead to the
formation of new mountain ranges (orogenesis) by compression, or to
fracturing by tension. The upwelling of lava from the inside of the
earth at these times makes the igneous rocks, usually devoid of fossils.

Changes in geography are, then, mainly changes in the height of
the land and the amount of it that is above water. Where the con-
tinent is surrounded by a rather shallow continental shelf, this leads
to considerable changes in appearance of the land-masses. The general
opinion is that the main outline of the continental masses has remained
much as at present, at least since Cambrian times. However there have
probably been considerable movements of the land-masses in relation
to each other. Some hold that the continents of lighter material are
continually expanding, at least in certain directions, having grown
from small centres to their present size. According to the hypothesis
of Wegener, the continents have all been formed by the splitting up
of one or a few land-masses. There is indeed evidence from both
geophysics and biology that the continents have been drifting apart
(Bullard, 1959). The direction of magnetization of rocks, which is
determined at the time of their formation, shows that the land-masses
must have changed their positions greatly. For example, such data
show that during the Triassic period the British Isles lay in the tropics
and in confirmation of this we find that many salt deposits (formed
only in very warm climates) lie in the Triassic formation (Droitwich,
Bath, Nantwich, &c).

9.2. Changes of climate

Evidence of marked changes of climate is the finding in England
and other regions now temperate of animal and plant remains appro-
priate to warmer or colder conditions (corals and woolly rhinoceros,
for instance). There is thus every reason to think that there have been
great changes from hot to cold and wet to dry conditions, in conjunc-
tion with the changes in latitude and in level of the land.

These fluctuations in geography and climate are obviously of great

14 EVOLUTION OF LIFE 1.9

importance to the biologist. We can hardly expect to treat animals
and plants as stable systems if the environment around them is
changing. In order to be able to assess the influence of such changes
on life we must know more about the rates at which they occur, and
careful study shows that some of the climatic changes are rhythmic.
Rhythmic changes of climate are, of course, very familiar to us in the
cycles of days, months, and years, and the immense importance of
these short-period changes for animal and plant life must not be
forgotten.

Here we are more concerned with changes of longer periodicity, of
which the best known are fluctuations of the amount of solar radiation
received at any given part of the earth's surface. These are likely to
be especially important since plants, and hence ultimately animals,
depend for their energy on sunlight. The cycle of number of sun-
spots (n*4 years) involves a change in amount of radiation, and this is
associated with some biological cycles, for instance in the distribution
of the rings of growth made by trees. Longer-period fluctuations in
the amount of radiation received on any part of the earth's surface
depend on the perturbations of the earth's orbit, particularly on
changes in the obliquity of the ecliptic. The effect of these perturba-
tions can be calculated, and the results show that at any one place
there are rhythmical variations in the amount of radiation received,
and in its seasonal distribution. The periodicity of these calculated
changes is about 40,000 years, with considerable irregularities and
variations in the sizes of the maxima (Fig. 2).

During the last million years (the Pleistocene epoch) there has been
a series of waves of glaciation (ice ages); the ice-caps have several
times advanced towards the equator and then retreated again. These
changes are usually classified into four periods of glaciation, separated
by interglacial periods. However, the last (fourth Pleistocene) glacia-
tion, of which we know the most, certainly had three separate climaxes
of cold. The correspondence of these with especially marked minima
in the curve of solar radiation is not perfect (Fig. 2), but it suggests
that the basic periodicity may have been something like 40,000 years,
and that the division of the whole Pleistocene period into four periods
of glaciation obscures a change with much shorter periodicity. From
about 120,000 to 180,000 years b.p. (Before Present) there were no
marked minima in the solar radiation curve, and this agrees with other
evidence of a long interglacial period (third Pleistocene interglacial).
Two marked minima agree with the other signs of a penultimate (third
Pleistocene) glaciation, and this was preceded by a very long warmer

i.9 CLIMATIC CHANGES 15

period, the second inter-glacial. As we go farther back the study
becomes more and more difficult, but the available evidence suggests
that fluctuations of climate considerable enough to alter the entire
fauna and flora may have taken place at a periodicity of something
over 40,000 years. It is a measure of the difficulty of geological
science that we cannot yet give a systematic account of the chronology
or climatic changes even of the relatively recent Pleistocene period
(variously estimated at 600,000 to 1,800,000 years) during which these
glaciations occurred.

RM590 R.M550
EGI 1 EGI 2

EARLY GLACIATION

RM476 RM 435

ApGI I ApGI 2

ANFEPENULTIMATE GLACIATION

Fig. 2. Curve of solar radiation received at 65 ° N. lat. in the summer. The radiation is
expressed in 'canonic units' (related to the solar constant in calories). Time in thousands
of years. R.M. 25, &c, indicate the radiation minima. (From Zeuner, based on the tables

of Milankovitch.)

As we proceed to study times still more remote our vision becomes
increasingly blurred. We can now only rarely distinguish periodicities
as short as 40,000 years, though there is evidence that they existed, for
instance from varved Cretaceous sediments. All we can see in the
study of geological deposits are the very marked changes produced by
the major movements of orogenesis and by the isostatic readjust-
ments. The surprising thing is that these immensely slow changes
have been sufficiently regular to leave layered deposits, allowing the
development of a system of geological classification. The process of
sedimentation was interrupted by periods when the continental shelf
on which the rocks rest was raised above the water surface and under-
went denudation for a while, before being again lowered below the sea
and covered with a new deposit. During the interval, while the shelf
was raised above the water, the animals and plants in the sea became

16 EVOLUTION OF LIFE 1.9

changed; thus rather sharp breaks appear in the series of fossils. The
occurrence of these breaks has been used by geologists to define the
major geological periods, which thus correspond to cycles of elevation
and depression of the continents. By comparing the fossils contained
in the rocks major geological periods have been recognized in various
parts of the world. The times of submergence and emergence differ
from region to region, however, and no very close detailed comparison
is possible. It is easy to forget that climates and land levels do not
always change in the same direction in different parts of the world.

9.3. Geological time

Until recently most geologists assumed that there was a regular
cycle of raising and lowering (diastrophism) and that comparable
periods could be recognized everywhere. It is now widely doubted
whether there has been any such 'pulse of the earth'. The rock series
are not the same in all the continents. For example, in South Africa
three long series, known as Cape, Karoo, and Cretaceous formations,
occupy the time covered in Europe by the many elevations and depres-
sions between Silurian and Cretaceous times. Probably the conditions
under which rocks were formed have remained about the same
throughout geological time but have been interfered with by periods
of elevation, depression, and folding that are peculiar to each region.

The study of fossils often establishes the order in which the rocks
were laid down, but other methods have to be used to discover the
period of time covered by each stage. This is especially important to
the biologist, who wants to know the rate at which animals or plants
have evolved. Reliable knowledge of the ages of the rocks has only
begun to accumulate since the discovery of radioactivity. Uranium
and thorium disintegrate, producing lead, at rates that are unaffected
by any known conditions. The age of any rock since its deposition can
therefore be calculated if we can estimate the amount of breakdown
products of these elements present in it. The lead present in a rock is
often not all derived from the uranium and thorium there, but separa-
tion of the lead isotopes enables those of radioactive origin to be
estimated, and the age of the deposit can then be determined, assum-
ing that the breakdown of uranium to lead began when the rock was
crystallized in its present position. Other methods of estimating the
ages of rocks from isotope ratios have been developed. Especially
promising is the determination of the ages of the deposition of sedi-
mentary rocks from the ratio of A 40 /K 40 and Sr 87 /Rb 87 in deposits
formed by erosion of micas or granites.

1.9

GEOLOGICAL TIME

17

The time at which the crust of the earth assumed its present form
is now thought to have been 4,500 million years ago (Holmes, 1959)
but the rocks laid down during the greater part of this long period
contain no undoubted animal or plant remains. Cambrian rocks,
when fossils become readily discernible, were laid down about 600
million years ago.

25 60
11 40 70

MILLIONS OF YEARS
135 180 225 270 305 350 400 440

500

600

21
42

68

98
110

£ 161

u. 205

CEpliocene
sTl i i i
^j- MIOCENE

"IjJ-o'ugocene
~3JJeocene

o

235
254
2 74
300

O ,

38

372

412

452

PRECAMBRIAN

_i I I I I I I i l_

100

200 300 400

MILLIONS OF YEARS

500

■ ■ ' ■

600

i i i i .

50
100
150
200
250
300
350
400
450

700

Fig. 3 shows the maximum thickness of sediment in each period plotted against estimates
of the absolute date. The error attached to these determinations is shown by the
marginal lines. Apparently the rate of sedimentation has not been constant (modified

after Holmes).

Classical geology is based mainly on studies in Europe and North
America. Although a terminology based on absolute time is beginning
to emerge, it is still necessary to use that based mainly on stratigraphic
studies, begun by William Smith in the British Isles early in the nine-
teenth century. In this system, the time since the Cambrian is divided
into eleven major periods, but several of these were double or triple
periods of advance and retreat of the sea. Even the most carefully
compiled radioactivity data are not yet adequate to provide us with
definite estimates of the durations of the periods, though there is
agreement on a total period of about 600 million years since the
Cambrian. Fig. 3 shows the maximum thickness of sediment in each
period plotted against estimates of the absolute dates. The error

18 EVOLUTION OF LIFE 1.9

attached to these determinations is shown. Apparently the rate of
sedimentation has not been constant.

It is conventional to postulate a series of crustal revolutions. The
extent of the movements has not been equal throughout and some of
them, more marked than others, were times of building of great
mountain chains such as the Alps or Andes (Fig. 4). There were also
many lesser rises and falls and changes of climate with shorter periods,

Table I

MAXIMUM THICKNESSES AND REVISED TIME-SCALE

(ACCORDING TO HOLMES)

THICKNESSES IN THOUSANDS OF FEET TIME SCALE IN MILLIONS OF YEARS

WORLD
MAXIMA

6

CUMULATIVE

MAXIMA PERIODS

6 PLEISTOCENE

SINCE

BEGINNING

OF PERIOD

1

DURATION

OF

PERIOD

1

15
21

21
42

PLIOCENE
MIOCENE

11
25

10

14

26

68

OLICOCENE

40

15

30

98

EOCENE

60

20

12

110

PALEOCENE

70 ± 2

10

51

161

CRETACEOUS

35± 5

65

44

205

JURASSIC

80 ± 5

45

30

235

TRIASSIC

25± 5

45

19

20 1 4 *
26j 46

254

300

PERMIAN

upper")
CARBONIFEROUS

LOWER J

70 ± 5

• 350 ±10

45
80

38

338

DEVONIAN

400 ±10

50

34

40

372
412

SILURIAN
ORDOVICIAN

450+10
500+15

40
60

40

452

CAMBRIAN

600 ±20

100

such as those of about 40,000 years that we can detect in the later part
of the Pleistocene. Many modern geologists are sceptical about the
existence of any regularities or rhythms in these changes (see Herbert,
1952, and Gilluly, 1949). It is useful when trying to adjust the mind
to periods of 30 million years to remember the frequent changes of
level and climate that have occurred in the last 100,000 years. In
spite of all that we know about the history of the earth's surface, it is
necessary every time that we make statements about the influence of
presumed climatic changes on organic evolution to remember how
scanty our knowledge is.

9.4. Classification of geological history

The period isolated as 'Cambrian' by geologists lasted 100 million
years and almost certainly included several inundations, perhaps
three. The Ordovician lasted for 60 million years and included three

i.9 RHYTHM OF GEOLOGICAL CHANGES 19

floods in North America. There were powerful earth movements at
the end of this period, at any rate in North America, known as the
Taconian revolution. The Silurian, lasting for 40 million years,
apparently included a single main cycle of inundation, ending in an
elevation of the land, which though slight in America, was marked
in Europe as the Caledonian revolution, producing the range of
mountains stretching across Scandinavia to Scotland and Ireland.

GENERALIZED CONTINENTAL

SUBMERGENCES

PALEOZOI C

ME S Z O I C

CEN0ZO1C

CAMBRIAN 0RO0VICIAN SIL DEVONIAN MISS PENN PERMIAN

h-RiAssir iiiRA«if L0W U p PER
TRIASS'C JURASSIC CRET CRETACEOUS

Tai n 1- Acadian Appalachian Nevadian Laramidt Cascadian

Fig. 4. Diagrams of main changes of areas of land and water and in climatic conditions

since the Cambrian. The chief times of mountain-building (orogenesis) in America are

also shown. (Redrawn by permission from Textbook of General Zoology, 2nd ed. by

W. C. Curtis and M. J. Guthrie, John Wiley & Sons, Inc., 1933.)

Throughout these early Palaeozoic periods the fossils are entirely
those of aquatic animals, except for some traces of land plants and
arthropods at the end of the Silurian. The oldest remains of verte-
brates are fish-scales from the Ordovician (p. 125). Details of the
Palaeozoic climatic changes are not clear, but the fact that corals,
which can now live only in warm water, were alive over a considerable
part of the earth's surface suggests that conditions were warmer than
at present at least at some early Palaeozoic times.

The Devonian is considered by some to include a single main
period, about 50 million years long, with one flood at the middle and
more arid conditions at the end, but other authorities divide it into
several periods. The first forests appeared at this time, and here, also,
are found tti3 first signs of vertebrate terrestrial life, in the form of
fossil lung-fishes and amphibians (p. 296). The period recognized as
Carboniferous in Europe includes two major periods of about 40

2o EVOLUTION OF LIFE 1.9-

million years each in America, the Mississippian and Pennsylvanian.
Throughout this long time conditions varied widely in different parts
of the world. In the early Mississippian there were many swamps in
North America. In the northern hemisphere the Pennsylvanian was
probably a time of warm, moist conditions, with no cold winters,
but there are signs that for part of this time India and Africa were
covered with an ice-sheet. The coal measures show us the remains of
the forests of spore- and seed-bearing plants that were then pro-
duced, and the land conditions evidently favoured the life of the
Amphibia.

The Permian probably constitutes a single 45-million-year period,
with very active orogenesis, leading to a more arid climate, perhaps
showing large seasonal changes, with deserts in some parts of the
world and glaciation in others. These conditions continued into the
Triassic, when the continents lay high. The reptiles, first found in
the Permian, developed throughout the Triassic and flourished in the
succeeding Jurassic period, which probably lasted 45 million years.
The Cretaceous period, during which the thick chalk deposits were
laid down, probably lasted for rather more than 60 million years,
including two major cycles of inundation. The lower Cretaceous
certainly included extensive periods of flooding, when there were
large shallow seas. Then later, towards the end of the upper Creta-
ceous, there were extensive orogenic movements, the Laramide
revolution, producing the Rockies and the Andes. The temperature
was warm until near the end of the Cretaceous, and we do not know
what condition led to the break that is found between the animals of
the Cretaceous and Eocene. Some groups of dinosaurian reptiles seem
to have died out suddenly, but it is important to notice that not all
disappeared at the same time, for instance, the stegosaurs and
pterodactyls (p. 569) disappeared well before the end of the Cretaceous.
However, it is probable that great changes went on at the end of this
period, and we may guess that a factor leading to the development of
the birds and mammals was the great rise of the continents, perhaps
accompanied by a fall in temperature over wide areas that had enjoyed
warmer weather. As always, when we look closely at such problems,
we are appalled by the vast lengths of time involved and the scanty
nature of our clues about them. The land lay very high at this time,
and the apparent abruptness of the break between Cretaceous and
Eocene fauna may be an artifact due to the scarcity of fossils. In
North America there is evidence from terrestrial deposits of a long
Paleocene period between the Cretaceous and Eocene.

i. 10 GEOLOGICAL PERIODS 21

It is usual to divide the last main geological period, the Tertiary, 1
into epochs, Paleocene, Eocene, Oligocene, Miocene, Pliocene, and
Pleistocene, the names originally referring to the percentage of fossil
genera surviving to the present day (see p. 571). Probably the whole
time since the end of the Cretaceous has been about 70 million years.
During the early part of the Tertiary period the climate was cold, but
as erosion of the mountains that had been produced at the end of the
Cretaceous proceeded the conditions became warmer, and throughout
the Eocene and Oligocene there were large forests and humid con-
ditions. Then during the Miocene there were marked earth move-
ments, leading to elevation of the land and accompanied by more arid
conditions, with wide areas of prairie and the widespread appearance
of important new food plants — the grasses. The weather probably
became gradually colder through the Pliocene, no doubt with many
fluctuations, culminating in the ice ages of the Pleistocene. Here we
come back to the period of which we have more detailed knowledge,
and are reminded that the ice age was not continuous, but interrupted
by many w r armer periods.

This very brief survey of geological history in the northern hemi-
sphere can hardly do more than remind us of the depths of our ignor-
ance. We see enough to be sure that climatic conditions have varied
throughout the millions of years, but we cannot yet see sufficient
details to allow us to discover whether there is any rhythm of major
cycles. It is easy to talk glibly of 'Carboniferous forests' or 'arid con-
ditions of the Permian', forgetting that these periods lasted for a time
that we can only roughly record in numbers and not properly imagine
in terms of our experience, although we are among the longest lived
of animals. The evidence suggests that conditions did not remain
stable for such a vast length of time as a whole geological period, but
fluctuated markedly, either irregularly or with complicated rhythms
of greater and lesser magnitude. We must not forget that very pro-
found 'climatic' changes occur every day, others every year, and some
every eleven years. It is not impossible that these shorter-period
changes, necessitating continual readjustment of animal and plant life,
have been as important as the slower changes in producing evolution.

10. Summary

To reduce to order our knowledge of vertebrate life we shall try
to discover its general organization and then examine the factors that

1 This word is a survival from an old-fashioned classification of rocks, the Tertiary heing
the period since the Cretaceous.

22 EVOLUTION OF LIFE i. 10

have produced all the varied types. The pattern of organization we
have to study is that of the animal as an active system maintaining
itself in its environment. This tendency to maintenance and growth
is the central 'force' that produces the variety of life. The opportunity
for change is provided by the fact that reproduction seldom produces
an exact copy of the parent, and thus a range of types is provided.
The tendencies to grow and to vary lead animals to colonize new en-
vironments and produce the variety of life. As evolution has proceeded
animals have come to occupy environments differing ever more widely
from the sea in which life probably arose. Life in these more difficult
environments is made possible by the development of special devices,
making the later animals more complex than the earlier and in this
sense 'higher'. It remains uncertain what influences have been respon-
sible for producing the changes in organic form. Geological evidence
shows that there have been many changes in climate and geography,
some of them proceeding at very slow rates in comparison with the
rhythms of individual animal lives. It is uncertain whether evolution-
ary changes follow these slow geological changes, or are a result of
the instability imposed on living things by climatic rhythms with
shorter periods, such as those of days, years, and the sunspot cycles.

II

THE GENERAL PLAN OF CHORDATE
ORGANIZATION: AMPHIOXUS

1 . The variety of chordate life

The Chordata occupy a greater variety of habitats and show more
complicated mechanisms of self-maintenance than any other group in
the whole animal kingdom. They and the arthropods and the pulmon-
ate molluscs have fully solved the problem of life on the land — which
they now dominate. This domination is achieved by most delicate
mechanisms for resisting desiccation, for providing support, and for
conducting many operations that are harder in the air than in water.
By even more wonderful devices the body temperature is raised and
kept uniform and thus all reactions accelerated. Finally, use is made of
this high rate of living for the development of the nervous system into
a most delicate instrument, allowing the animal not only to change its
response to a given stimulus from moment to moment, but also to
store up and act upon the fruits of past experience.

Besides these more developed types of chordate that dominate the
land and air there are also great numbers of extremely successful
aquatic and amphibious types. The frog is often referred to as a some-
what lowly and unsuccessful animal, but frogs and toads are found all
over the world. The sharks and bony fishes share with the squids and
whales the culminating ecological position in the food chains of the
sea, while the bony fishes are the only animals that have achieved con-
siderable size and variety in fresh water. Among the still more lowly
chordates the sea-squirts take a very important, though not dominant,
position among the animal and plant communities that occupy the
sea bottom, but they have never entered fresh water.

One could continue indefinitely with particulars of the amazing
types produced by this most adaptable phylum. Yet through all their
variety of structure the chordates show a considerable uniformity of
general plan, and there can be no doubt that they have all evolved
from a common ancestor of what might be called a 'fish-like' habit.
In the very earliest stages only the larva was fish-like, and the life-
history probably also included a sessile adult stage, such as the tuni-
cates still show today (p. 66). This bottom-living phase was then
eliminated by paedomorphosis, the larvae becoming the adults. There-
fore the essential organization of a chordate is that of a long-bodied,

24 CHORDATE ORGANIZATION n. i-

free-swimming creature. All the other types can be derived from such
an ancestor, though in some cases only by what is often called
'degeneration'.

2. Classification of chordates

We may conveniently divide the Phylum Chordata into four
subphyla :

Subphylum i. Hemichordata

Balanoglossus; Cephalodiscus; Rhabdopleura
Subphylum 2. Cephalochordata (= Acrania)

Branchiostoma
Subphylum 3. Tunicata

Ciona, Sea-squirts
Subphylum 4. Vertebrata

The Vertebrata, the largest of these groups, may be subdivided:
Subphylum Vertebrata
Superclass 1. agnatha

Class 1. Cyclostomata. Lampreys and hag- fishes

Class 2. *Cephalaspidomorphi. *Cephalaspis

Class 3. # Pteraspidomorphi. *Pteraspis

Class 4. *Anaspida. *Birkenia, *Jamoytius

Superclass 2. gnathostomata

Class 1. *Placodermi. *Acanthodes

Class 2. Elasmobranchii. Dogfishes, skates, and rays

Class 3. Actinopterygii. Bony fishes

Class 4. Crossopterygii. Lung-fishes

Class 5. Amphibia

Class 6. Reptilia

Class 7. Aves

Class 8. Mammalia.

3. Amphioxus, a generalized chordate

It has long been realized that through their great variety all these
types show certain common features, often referred to as the typical
chordate characters. It is better to regard these not as a list of isolated
'characters' but as the signs of a certain pattern of organization that
is characteristic of the group. There is much reason to suppose that
this basic chordate organization was that of a free-swimming marine
animal, probably feeding by the collection of minute particles. We
are fortunate in having still alive a little animal, amphioxus, the

ii. 3 AMPHIOXUS 25

lancelet, which shows nearly all of these features in diagrammatic form.
Study of amphioxus will go a long way to show the basic plan on which
all later chordates are built, and, indeed, gives us a strong indication
of what the early chordates must have been like.

Though it can swim freely through the water, amphioxus is essenti-
ally a burrowing animal, and many of its special features are connected
with this habitat. It lives in the sand, at small depths, and has been
found all round the oceans of the world. Evidently, in spite of its
simplicity, it is a successful type. It is found on British coasts and,
indeed, the first individual described was sent (preserved) from Corn-
wall to the German zoologist Pallas, who supposed it to be a slug and
called it Limax lanceolatus (1774). It was first figured and given the
name Amphioxus lanceolatus by Yarrell in 1836. However, the name
Branchiostoma had been given in 1834 by Costa and by the rules of
priority this is the official name of the genus. We may keep amphioxus
as a common name. Some eight species of Branchiostoma are recog-
nized, and in addition there is a group of six species referred to the
genus Asymmetron. These resemble Branchiostoma in general organi-
zation, but they have gonads only on the right side.

The adult Branchiostoma lanceolatum is rather less than 2 in. long
and has the typical fish-like organization, whose main external
features are related to the methods of locomotion and feeding (Fig.
5). The body is elongated, and flattened from side to side. The skin
has no pigment, and the muscles can be easily seen as a series of
blocks, the myotomes, serving to bend the body into folds. As the
name implies, the body is pointed at both ends; there is no recogniz-
able head separated from the body. Indeed, there are no separate eyes,
nose, or ears, and no jaws, so that the fundamental plan of chordate
organization appears in almost its fullest simplicity from one end of
the body to the other. The front end is, however, marked by a series
of buccal cirri, which form a sieve around the opening of the oral hood
and are provided with receptor cells.

Although the animal is provided with a large number of gill-slits
these do not appear externally, being covered by lateral folds of the
body, which enclose a ventral space, the atrium, opening posteriorly
by an atriopore. The outside edges of the atrium project as a pair of
metapleural folds, giving the body a triangular shape in transverse
section. The alimentary canal opens posteriorly by an anus, in front
of the hind end of the body, thus leaving a definite tail — a region of
the body not containing any part of the alimentary canal.

The general arrangement of the organs can best be understood by

26 CHORDATE ORGANIZATION n. 3-

considering the body as consisting of two tubes, the outer skin
(ectoderm) and the inner alimentary canal (endoderm), with a space
between (the coelom) lined by a third layer (the mesoderm). This
arrangement is actually found during the course of the development
(Fig. 18). The mesoderm at first forms thin layers, the somatopleure
applied to the outer body wall and the splanchnopleure to the gut.
Very soon the inner layer becomes much thickened where it is applied
to the nerve-cord and notochord, and here it forms the myotomes, or
muscle-blocks. In this dorsal part of the mesoderm the coelom,
known here as the myocoele, soon becomes obliterated, leaving the
ventral splanchnocoele around the gut. Besides the muscle that forms
in the myotomes, non-myotomal muscles develop in the somatopleure
and splanchnopleure. These are not divided into segments and are
innervated by the dorsal nerve-roots, the ventral roots supplying only
the myotomes.

4. Movement of amphioxus

The adult myotomes are blocks of striated muscle-fibres, running
along the body, separated by sheets of connective tissue, the myo-
commas. This repetition or segmentation is characteristic of the
organization of all chordates. The myocommas do not run straight
down the body from dorsal to ventral side but are V-shaped (Fig. 5).
However, each muscle-fibre runs straight from before backwards, and
the contraction of the whole myotome therefore bends the body. A
full discussion of the means by which forward motion is achieved by
such a system will be given later (p. 133). Essentially, contraction of
the myotomes results in transverse motion of the body inclined at
varying angles in such a way as to result in forward propagation. Each
myotome must therefore contract after that in front of it — the effect
being to produce an S-bend that moves backwards through the water
as the fish moves forward.

For our present purpose the point is that the contraction is serial,
that is to say, it depends on the breaking up of the longitudinal muscle
into blocks. It was probably the need for division of the musculature
that led to the development of the segmentation, and this, affecting
primarily the muscles, has come to influence a great part of chordate
organization.

Contraction of the longitudinally arranged muscle-fibres will only
produce a sharp bending of the body if there is no possibility of short-
ening of the whole. To prevent telescoping, an incompressible and
elastic rod, the notochord, runs down the centre of the body. It

II. 4

LOCOMOTION OF AMPHIOXUS

27

is usually stated that this is a
'supporting structure', but, of
course, an animal such as a fish in
water needs no 'support'. Nor is
the notochord a lever to which
muscles are attached, as they are
to the bones of many higher forms.
No muscles pull on it directly,
though the myocommas are at-
tached to its sheath. Its function
is to prevent the shortening of the
body that would otherwise be the
result of contraction of longitudi-
nal muscles. In fact, it serves to
make that contraction efficient in
bending the body; its elasticity
may also play an important
part.

The notochord is composed of
a series of flattened plates sur-
rounded by a fibrous sheath. The
plates are arranged in a regular
manner with their flat surfaces in
the transverse plane of the body.
They are of two sorts, fibrous
and homogeneous, which alternate
with each other. Each plate de-
velops as a highly vacuolated cell,
the nuclei being later pushed aside
to the dorsal or ventral edge. This
structure is well suited by the
turgidity of its cells enclosed in
the sheath to resist forces tending
to shorten the body. The cord of
amphioxus is peculiar in that it
extends from the very tip of the
head to the end of the tail, pro-
jecting, that is to say, beyond the
level of the myotomes, a condition
presumably associated with the
burrowing habit.

%

<N<

(1

|\

^

c *

§ 8

5 «.
3 <j

.t: «

it

e« 3

3 C

to «

•S c 2

--' M 2
-. t- m

3
X

o

'J:
0,

JZ J-i u

■£
S -Sa

JZ <•>

"a)

l>

73 -*-l

73

^ -5

> c

M O

SB

^ 3
* u
— g

;» O

- -C

28

CHORDATE ORGANIZATION

ii. 4-

Amphioxus probably does not often swim free in the water and the
body is not adapted for fast movements. It has no elaborate fins such
as those of later fishes, which ensure static stability like the feathers
on an arrow, or are movable, to allow active control of the direction

Fig. 6. Transverse section through amphioxus in the region of the pharynx.

atr. atrium; d.a. dorsal aorta; d. coel. dorsal portion of coelom; div. intestinal diverticulum;
d.n. 1 and d.n. 2 , branches of the dorsal nerve-root; end. endostyle; ep.gr. epipharyngeal groove;
/. fin-ray box; g. gonad; I y.b. primary gill bar containing coelom; my. myotome; metapl.
nietapleural fold; n. notochord; n.c. nerve-cord; ph. pharynx; sub. end. coel. subendostylar
coelom; t.b. tongue bar; v. a. ventral aorta; v.n. ventral nerve-root. (After Krause.)

of swimming (p. 136). There is a low dorsal ridge, which continues
behind as a small caudal fin. There are no definite paired fins, but the
metapleural folds might perhaps be considered comparable to the
lateral fin folds from which all vertebrate limbs are probably derived.
They are distended with coelomic fluid and, with the dorsal ridge,

II. 6

SKELETAL STRUCTURES OF AMPHIOXUS

29

probably serve to protect the body during the rapid dives by means of
which the creature enters the sand. The habit of swimming with the
front end downwards suggests the presence of a gravitational receptor
mechanism. The larvae of lampreys swim in a similar way (p. 114).

scut

Fig. 7. Section of the skin of amphioxus.

b.v. blood-vessel; cut. cutis; ep. epidermis; n. nerves; s.cut. sub-cutis.
(After Krause.)

5. Skeletal structures of amphioxus

Around the notochordal sheath is a further layer of gelatinous
material containing fibres. There are no cells within this material
but it is secreted by cells around the outside, which retain the epithe-
lial arrangement of the mesoderm from which they were derived.
This connective tissue continues as a sheath around the nerve-cord
and above this into a series of structures known as fin-ray boxes,
which support the median ridge. These are more numerous than the
segments and each contains a more rigid material referred to as
'cartilage'. The relationship of these structures to the fin supports of
vertebrates is obscure. Other skeletal rods occur in the cirri around
the mouth and in the gill bars.

6. Skin of amphioxus

The epidermis differs from that of vertebrates in being very thin,
composed of a single layer of cells, ciliated in the young, and with the
outer border slightly cuticularized in the adult (Fig. 7). It is not

3 o CHORDATE ORGANIZATION n.6-

known whether this cuticle contains a substance similar to the keratin
produced by the many-layered skin of later forms. There are receptor
cells but no glands or chromatophores in the skin.

Below the epidermis is a fibrous cutis, and below this again a
gelatinous material containing fibres, the sub-cutis. Both these layers
are secreted by scattered cells having some similarity to the fibroblasts

Fig. 8. Anterior end of amphioxus, from a stained and cleared preparation
of a young animal.

b.c. buccal cirri;/, fin-ray box; II. p. Hatschek's pit; my. myotome; n. notochord;
n.c. nerve-cord; p. pigment spot; ph. pharynx; v. velar tentacles; zi.o. wheel organ.

of higher forms. They contain a system of cutaneous canals, with
endothelial lining (Fig. 14).

7. Mouth and pharynx and the control of feeding

Amphioxus obtains its food by extracting small particles from a
stream of water, which it draws in by means of cilia. In all animals
that use cilia for this purpose a very large surface is provided (e.g.
lamellibranchs, ascidians), and the pharynx and gill bars of amphioxus
occupy more than one-half of the whole surface area of the body.
Special arrangements are made for the support and protection of this
ciliated surface, the wall of the pharynx being so greatly subdivided
that it needs the protection of an outer layer, the atrium.

The mouth lies covered by an oral hood whose edges are drawn out
into buccal cirri, provided with sense-cells (Fig. 8). When feeding the
cirri are curved to form a funnel-like sieve preventing the entry of
large particles. Around the mouth itself there is a further ring of

ii. 7

FEEDING OF AMPHIOXUS

3i

sensory tentacles, the velum. The oral hood contains a complex set of
ciliated tracts, the 'wheel organ' of Miiller, and this plays a part in
sweeping the food particles into the mouth (Figs. 8 and 9). Near its
centre is a groove, Hatschek's pit, formed as an opening of the left
first coelomic sac to the exterior (p. 44).

The main operation of food collec-
tion is performed by the pharynx, a
large tube, flattened from side to side,
whose walls are perforated by nearly
200 oblique vertical slits, the number
increasing as the animal gets older.
The slits are separated by bars con-
taining skeletal rods and further sub-
division is provided by cross-bars
(synapticulae). Since the bars slope
diagonally many of them are cut in a
single transverse section, but it must
be remembered that they are essen-
tially the vertical portions of the main
walls of body and pharynx, where
these have not been perforated by a
gill-slit. Such a portion of the body
wall must contain a coelomic space
and this can in fact be seen in the
original or primary gill bars. How-
ever, an increase of the ciliary surface
is produced by downgrowth of secon-
dary or tongue bars from the upper
margin, dividing each primary slit;

these secondary bars contain no coelom. The coelomic spaces in the
primary bars, of course, communicate above and below with con-
tinuous longitudinal coelomic cavities (Fig. 6).

There are cilia on the sides and inner surfaces of the gill bars, the
lateral ones being mainly responsible for driving the water outwards
through the atrium and thereby drawing the feeding current of water
in at the mouth. In the floor of the pharynx lies the endostyle, con-
taining columns of ciliated cells, alternating with mucus-secreting
cells, which produce sticky threads in which food particles become
entangled. Various currents then draw the sticky material along until
it reaches the mid-gut. The frontal cilia of the gill bars produce an
upward current, driving the mucus from the endostyle into a median

Fig. 9. Transverse section through
front end of amphioxus.

b.c. buccal cirri; e. eyespot; H.p. Hats-
chek's pit; ?i. notochord; n.C. nerve-cord.

32 CHORDATE ORGANIZATION n. 7-

dorsal epipharyngeal groove, in which it is conducted backwards.
The cilia of the endostyle also move mucus along the peripharyngeal
ciliated tracts, behind the velum, to join the epipharyngeal groove.
Radioactive iodine is concentrated by one of the columns of the endo-
style and secreted with the mucus. Barrington (1958) suggests that
these may be regarded as the precursors of the thyroid cells, serving
to produce iodinated mucoproteins, which are then absorbed farther
down the gut (see p. 119).

The pharynx narrows at its hind end to open dorsally into a region
best known as the mid-gut, the name stomach being inappropriate.

£{■ nig. ant. fg.post (cft

/ Ag

T ^^^' |l ^ | i | Mii)iifmiTn^ r . y

Fig. 10. Currents in the mid-gut of amphioxus, showing the appearance when an
animal is placed in a medium containing carmine particles. Arrows show the chief

ciliary currents.

div. diverticulum ;f.c. food cord; h.g. hind-gut; i.c.r. ileo-colon ring; m.g.ant. and m.g.post.

anterior and posterior parts of mid-gut; oes. oesophagus. (After Barrington.)

A large mid-gut diverticulum reaches forward from this region on the
right-hand side of the pharynx. From its position this organ is often
called the liver, but Barrington has given reasons for supposing that
it is the seat of the production of digestive enzymes. Zymogen cells,
similar to those of the mid-gut, are found in its walls. Its strong dorsal
and ventral ciliation maintains in it a circulation of food materials and
secretion, and its cells are capable of phagocytosis as well as secretory
activity. Amphioxus thus combines intracellular with extracellular
digestion, doubtless in connexion with its microphagous habit.
Particles placed in the diverticulum are swept backwards and join the
main food cord that passes through the mid-gut (Fig. 10).

The hind end of the mid-gut is marked by a specially ciliated region,
the ileo-colon ring, whose cilia rotate the cord of mucus and food. The
movement is transmitted to the portion of the food cord in the mid-
gut and presumably assists in the taking up of the enzymes that
emerge from the diverticulum. Extracellular digestion takes place in
the mid-gut and the enzymes responsible have been studied by Bar-
rington. The pH of the contents varies from 67 to 7-1. An amylase is
present in extracts of the diverticulum, mid-gut, and hind-gut, but
not in those of the pharynx. Lipase and protease are present in the

ii. 8 FEEDING OF AMPHIOXUS 33

same regions, the latter having an optimum action at about pH 8-o,
being, that is to say, a tryptic type of enzyme. There is no sign of
any protease with an acid optimum, similar to the pepsin of higher
forms.

Behind the ileo-colon ring the intestine runs as a straight hind-gut
to the anus. Absorption of food takes place here, and perhaps also
in the mid-gut, apparently partly by intracellular digestion, since
ingested carmine particles are taken into the cells.

The feeding current is regulated by the rate of beat of the cilia and
the degree of contraction of the inhalent and exhalent apertures. The
walls of the atrium contain an elaborate system of afferent and
efferent nerve-fibres. The receptors include a set of large peripheral
nerve-cell bodies, lying beneath the atrial epithelium and sending
axons in by way of the dorsal roots. The motor fibres also pass through
the dorsal roots and run without synapse to the cross-striated fibres
of the pterygial muscle, which forms the floor of the atrium. The
stream flowing into the pharynx is tested by the receptors of the velum
and atrium, and if noxious material is present, the water is expelled by
closing the atriopore and contracting the pterygial muscle, producing
a 'cough'. The system can distinguish between suspensions of food
material and inorganic particles. When sufficient food has been taken,
collection is suspended until it has been digested (Bone, i960).

The atrial nervous system probably regulates spawning as well as
feeding. It has often been compared with the sympathetic system of
craniates but there are almost no close similarities. The nerve cells in
it are receptors and there is no sign of the peripheral synapse on
the efferent pathway that is so characteristic of the true autonomic
system. The atrial system is developed in relation to filter feeding
and has perhaps been completely lost in higher forms that feed by
other methods and have developed new methods to control them
(p. 117).

8. Circulation

The blood-vessels of amphioxus show in diagrammatic form the
fundamental plan on which the circulation of all chordates is based
(Fig. 11). Slow waves of contraction occur in various separate parts
in such a way as to drive the blood forwards in the ventral vessels,
backwards in the dorsal ones. Below the hind end of the pharynx
there is a large sac, the sinus venosus, into which blood from
all parts of the body is collected. From this there proceeds for-
wards a large endostylar artery (truncus arteriosus or ventral aorta)

34

CHORDATE ORGANIZATION

II. 8-

ant. card-

d. cuv.

>m

d ao:

post, card'

d.ao.

from which spring vessels carrying blood up the branchial arches.
At the base of each primary bar there is a little bulb, functioning as

a branchial heart. From the gill bars
blood is collected into paired dorsal
aortae, which join behind the pha-
rynx. From the paired and median
aortae blood is carried to the system
of lacunae that supplies the tissues.
There are no true capillaries. From
the lacunae blood is collected into
veins, the most important of which
are the caudals, cardinals, and a
plexus on the gut. The cardinals are
a pair of vessels in the dorsal wall of
the coelom, and they collect blood
from the muscles and body wall.
They lead to the sinus venosus by
a pair of vessels, ductus Cuvieri,
which pass ventrally and across the
coelom to join the sinus venosus on
the floor of the gut. The caudal
veins join the plexus on the gut,
from which blood is collected by
a large subintestinal vein running
on to the liver; from here another
plexus leads to the sinus venosus.

Contractions arise independently
in the sinus venosus, branchial
bulbs, subintestinal vein, and else-
where. The rhythms are very slow
(once in two minutes), irregular,
and apparently not coordinated by
any control system.

The blood is colourless and is not
known to contain any respiratory
pigment. It contains no cells. Pre-
sumably the tension of dissolved
oxygen acquired by simple solution
is sufficient for the small energy needs of the animal, w r hich spends
most of its life at rest. It is by no means certain that any oxygenation
of the blood takes place in the gills. Orton has suggested that since

'subuit

^

Fig.

i i. Diagram of the circulation of
amphioxus.

aff.d. afferent vessel of diverticulum; ant.
card, anterior cardinal vein; br.a. bran-
chial arch; d.ao. dorsal aorta; d.cuv. ductus
Cuvieri; eff.d. efferent vessel of diver-
ticulum; post. card, posterior cardinal
vein; sin. sinus venosus; subint. subin-
testinal vein; v. a. ventral aorta. (After
Grobben and Zarnik.)

II. Q

EXCRETION OF AMPHIOXUS

35

these, through their cilia, do much of the work of the body, the
blood actually leaves the gills less rich in oxygen that when it enters
them. Oxygenation probably takes place chiefly in the lacunae close
to the skin, perhaps especially those of the metapleural folds.

9. Excretory system of amphioxus

One of the most mysterious features about the organization of
amphioxus is that there are flame-cells, comparable with those found

Fig. 12. Solenocytes of amphioxus, showing the nuclei, long flagella, and the openings
into the main excretory canal, which leads to the atrium. (After Goodrich.)

in platyhelmia, molluscs, and annelids. The excretory organs, there-
fore, do not conform to the basic chordate plan, and are in fact very
different from those not only of all other chordates but also from any
found in the remote invertebrate allies of the chordates which, as we
shall presently see, include the echinoderms, brachiopods, and polyzoa.
The nephridia lie above the pharynx. To each primary gill bar
there corresponds a sac, opening by a pore to the atrium and studded
with numerous elongated flame-cells (solenocytes) (Fig. 12). These

36 CHORDATE ORGANIZATION n. 9-

flame-cells do not open internally, but are in close contact with special
blood-vessels (glomeruli) whose walls separate the flame-cells from
the coelomic epithelium. Assuming that there are 200 of these nephri-
dia, each with 500 solenocytes 50 \x long, Goodrich, who has provided
the most accurate information about these organs, shows that the total
length available for excretion is no less than 5 metres. It is assumed
that excretion takes place by diffusion through the flame-cell wall, the
liquid being driven down the tube by cilia. Coloured particles injected
into the blood-stream are not excreted by the nephridia.

In development these remarkable organs arise from groups of cells
close to the meeting-place of ectoderm and endoderm; almost cer-
tainly they are derived from the former. They have no relation what-
ever to the mesoderm and this fact alone sufficiently indicates that
they are in no way comparable to the pronephros of vertebrates, as is
sometimes stated. There is no organ in vertebrates with which they
can be compared, nor is there any trace in amphioxus of organs com-
parable to the vertebrate kidney system. In fact we have here a
remarkable case of an isolated feature; evidently separate items of the
genotype may vary independently, and the whole bodily organization
does not necessarily change together.

The brown funnels are blind sacs at the front of the atrium,
invaginating into the epibranchial coelom. They are probably receptor
organs. Some parts of the atrial wall may perform excretory functions.
Masses of cells in the atrial floor, the atrial glands, contain granules
that may be excretory but may have been taken up from the food
current.

In the gonads, especially the testes, there are large yellow masses,
containing uric acid, which are extruded with the gametes.

10. Nervous system

Amphioxus possesses a hollow dorsal nerve-cord similar to that of
vertebrates. Though this is somewhat modified at the front end, it is
not there enlarged into an elaborate brain. The nervous system is con-
nected with the periphery by a remarkably simple set of nerve-roots,
a dorsal and a ventral on each side in each segment. The roots do not
join (Fig. 13): the ventral roots lie opposite the myotomes, to which
they carry motor-fibres, and these end on the muscle-fibres with
motor end-plates. The dorsal root runs out between the myotomes
and carries all the afferent fibres of the segment and motor-fibres for
the non-myotomal muscles of the ventral part of the body. This is the
fundamental pattern of the roots in all vertebrates.

NERVOUS SYSTEM OF AMPHIOXUS

d.n.

37

v.n.

Fig. 13a. Nerve-cord of amphioxus.
d.n. dorsal nerve-root; g. 'giant' nerve-fibres; v.n. ventral nerve-root. (After Retzius.)

Ret

rec

v.m.c

s.m c

median giant fibre

Fig. 13ft. Stereogram illustrating the structure of the spinal cord in an adult amphioxus.

The receptor system is made up of a more or less continuous column of bipolar cells of
Retzius (Ret.), together with smaller cells of various types (rec). According to Johnston
these receptor cells (1,2 and 3) can be regarded as equivalent to the dorsal root ganglion
cells of vertebrates. The other type of receptor cell is the giant Rohde cell (Roh.), which has
a large axon and elaborate dendritic system. It is probable that at least some of these cells
possess a peripheral axon running in the dorsal root. I.e. longitudinal connective cell.

The visceral motor cells (v.m.c.) are arranged segmentally, one per segment.

The somatic motor cells (s.m.c.) lie at a different level in the cord from the ventral
roots.

Other cells in the cord are internuncials of various types. (After Bone.)

The fibres of the peripheral nerves differ from those of vertebrates
in that they have no thick myelin sheath that will blacken with
osmium tetroxide. The nerve trunks are surrounded by an epineurium
with connective tissue cells but there seem to be no Schwann cells
accompanying the nerve-fibres (Bone, 1958).

38

CHORDATE ORGANIZATION

The afferent fibres of the dorsal roots are unique among chordates
in that the cell bodies are not collected into spinal ganglia but mostly
lie within the central nervous system. At least three types of central
neuron send fibres that terminate as free nerve endings in the skin. In
addition, on the head and tail there are peripheral receptor cells,
sending fibres centrally, also complicated encapsulated organs in the
metapleural folds (Bone, i960). There are numerous large multipolar
nerve-cells, presumably afferent, just beneath the atrial epithelium.

5 cut.

bucc. ep

Fig. 14. Sagittal section through the front end of amphioxus.
bucc.ep. buccal epithelium; cer. cerebral vesicle, with large nerve-cells; ep. epidermis;
my. myotome; n. notochord; p. pigment spot; s.cut. subcutis; vent, ventricle of cerebral vesicle.

(After Krause.)

These cells have many branched dendrites and an axon that runs
through a dorsal root to the spinal cord. Their status is discussed on

P-33-

The spinal cord has only a narrow lumen and its elements are

arranged as in vertebrates, namely, ependyma close to the canal, cell
layer ('grey matter'), and outer fibrous layer ('white matter'). The cells
are not arranged clearly in horns as they are in vertebrates. The most
conspicuous cells are the giant cells, which lie dorsally in the anterior
and posterior parts but are absent from about the 13th to 39th seg-
ments. Each of these cells has many dendrites, branching in the region
of entry of the dorsal root fibres, and a single axon, which runs back-
wards in the front part of the body, forwards in the hind, passing in
each case for the whole length of the cord. A median giant fibre,
which runs ventrally for the length of the cord, lies close to the
viscero-motor cells that probably produce the 'coughing' movements
of the atrium (p. 33).

II. IO

RECEPTORS OF AMPHIOXUS

39

Ten Cate has investigated the movements of amphioxus and found
that it responds to all stimuli by movements of 'flight'. There are no
isolated or local movements; the effect of any stimulus such as touch
on the side of the body is to produce waves of myotomal contraction.
These may, however, vary from strong waves going the whole length

g-ep.

s.ep.

p.sp.

B

n.pr. o.c. t.po.h

Fig. 15. Diagram of (a) the anterior end of the nervous
system of amphioxus and (b) the brain of a fish
(Polypterus).

A. Amphioxus. g.ep. granulated ependyma in the wall of the 'dorsal
central canal'; i.o. infundibular organ; p.sp. pigment spot; r.f.
Reissner's fibre in the central canal; s.ep. sensory epithelium,
u. Polypterus, a.c. anterior commissure; aq.s. aqueductus Sylvii; cer.
cerebellum; ep. epiphysis; m. medulla spinalis; n.h. neurohypophysis;
n.pr. nucleus praeopticus; o.c. optic chiasma; r.f. Reissner's fibre in
fourth ventricle; s.c.o. subcommissural organ; s.d. saccus dorsalis;
s.v. saccus vasculosus with primary sense cells; t.po.h. tractus praeop-
tico-hypophyseus. (After Olson and Wingstrand.)

of the body to single rapid twitches. The giant cells participate in the
spread of these waves. It seems likely that the arrangement ensures
that touch on the anterior part of the body, normally exposed when
feeding, produces backward movement (i.e. withdrawal into the sand)
but touch on the hind part the reverse movement of emergence and
escape.

At the front end the central canal is enlarged to form a cerebral
vesicle (Fig. 14). The whole neural tube is hardly wider here than in
the region of the spinal cord and there is no thickening of the walls,

40 CHORDATE ORGANIZATION n. 10-

which are indeed mostly formed of a single layer of ciliated epithelial
cells (Fig. 15). This is a striking indication of the lack of cephalization
of the animal. From the region of the cerebral vesicle spring the first
two dorsal roots, to which there are no corresponding ventrals. These
roots carry impulses from the receptors of the oral hood and its
tentacles.

A B C

Ifls jT^o*~r$5fc

Fig. 16. Diagram to show the direction of the eye-spots of amphioxus.
A, anterior, B, middle, and c, posterior regions of the body. The eyes are shown as if seen from
behind. D shows the direction of spiralling of the animal when swimming — as seen from in front.

(After Franz.)

The infundibular organ (Fig. 15) is composed of tall cells with long
cilia, which beat in the opposite direction to those of the rest of the
vesicle. From them fibres run backwards down the cord. The organ
is also the site of origin of Reissner's fibre (Fig. 15). This is a thread
of non-cellular material, present in all vertebrates at the centre of the
neural canal. It is secreted at the front end and then passed backwards
and is often collected and absorbed in a sac at the hind end of the
spinal cord. In vertebrates it arises from secretory ependymal cells of
the subcommissural organ, lying dorsally in the diencephalon (Fig.
1 5). The infundibular organ of amphioxus is clearly not exactly similar,
yet the Reissner's fibres are clearly comparable ; an interesting problem
in homology.

A further complication is that the cells of the infundibular organ
contain material that stains with the Gomori method, and is similar to
the neurosecretory material found in the fibres of the hypophysial tract
(Fig. 15). The organ thus seems to occupy a central position in the
control system as a receptor, originator of nerve-fibres, and of two
sorts of secretion. There is clearly much to be learned from this about
the origin and significance of the control systems of the diencephalon.

In young stages the cerebral vesicle opens by an anterior neuro-
pore, and at the point where the closure takes place there develops a
depression of the skin, lined by special epithelium, and known as
Kolliker's pit. It is said to receive no special innervation. The cells
at the front end of the cerebral vesicle contain pigment and there have

ii. ii RECEPTORS OF AMPHIOXUS 41

been attempts to show that this represents an eye. More probably
it serves to prevent rather than to receive photic stimulation; there
are other cells lying in the spinal cord that are clearly photoreceptors
(Fig. 9). In the front part of the body these are unprotected by pig-
ment, whereas more posteriorly they are so pigmented as to be pro-
tected asymmetrically from the light (Fig. 16). This asymmetry may
be connected with the fact that when swimming free in the water
amphioxus moves spirally about its axis, turning clockwise as seen
from behind. It was established by Parker that a small beam of light
produces movements of amphioxus only when it is directed on to the
region of the body or tail, not when it shines on the head. Since the
animal normally lies with the head protruding we may suppose that
the pigment spot serves to prevent light that strikes down vertically
from stimulating the photoreceptors in the cord.

Amphioxus is therefore provided with receptor and motor systems
that serve to keep it in its sedentary position, able to collect food from
the current that it makes by the cilia (p. 33). There are mechanisms
that help it to make appropriate movements of escape when it is
touched or when the body (but not head) is illuminated. The touch
receptors of the buccal cirri produce rejection of large particles and
those of the velum are chemo-receptors. The infundibular organ may
be some form of gravity or pressure receptor. By means of these
receptor organs and its simple movements of swimming, burrowing,
and closing the oral hood, the animal is maintained, probably mainly
by trial and error (phobotactic) behaviour, in an environment suitable
for its life. There are none of those elaborate mechanisms that we find
in higher chordates for 'seeking' special environments or for so
'handling' or managing them that they may prove habitable by the
animal. Amphioxus must take and leave the world very much as it
finds it. The 'correct' environment is chosen for it by the selective
settling of the larvae.

1 1 . Gonads and development of amphioxus

The gonads of amphioxus are hollow segmental sacs with no com-
mon duct. Each sac develops from mesoderm cells, perhaps originally
from a single cell, at the base of the myotomes in the branchial region,
the genital cells themselves developing on the walls (Fig. 6). The sexes
are separate and the genital products are shed by dehiscence into the
atrium, the aperture by which they escape closing and the gonad
developing afresh.

Extrusion of the gametes occurs in spring, on warm evenings

42

CHORDATE ORGANIZATION

following stormy weather. Fertilization is external and development
then occurs free in the water. Numerous eggs are produced and they are
small but yolky. Complex flowing movements take place in them after
fertilization, and cleavage is then rapid and complete, producing a
blastula composed of a dome of somewhat smaller and a floor of rather
larger cells (Fig. 17). These latter then invaginate to make the archen-
teron, opening by a wide blastopore, which later becomes the anus.
At about this stage the gastrula becomes covered with flagella, by
which it rotates within the egg case.

Fig. 17. Three stages in the development of amphioxus as seen in stained

preparations.
a, the blastula; b, early, and C, later gastrula.

The creature now elongates and its dorsal side flattens and eventu-
ally sinks in to form the neural tube (Fig. 18 a). At about this time
the dorsal side of the inner layer begins to fold near to the front end,
in such a way as to make a pair of lateral pouches. The walls of these
pouches are the future mesoderm and the cavity is the coelom. As in
other early chordates, therefore, the coelom is continuous at first
with the archenteron. The roof of the archenteron also arches up
dorsally and forms the notochord, the gut wall being completed by the
approximation of the edges of the remaining portion of the inner
layer, which is now the definitive gut wall or endoderm.

The analysis of the processes of development now enables us to say
something of the forces by which these formative foldings and cell
movements are produced. The formation of the neural tube, meso-
derm and notochord and the completion of the gut roof all involve an
upward movement of cells towards the mid-dorsal line. This process
of 'convergence' is a very marked feature of the development of all
chordates (Young, 1957, p. 609).

As the animal elongates, further mesodermal pouches are produced,
each separating completely from the endoderm and from its neigh-

II. II

DEVELOPMENT OF AMPHIOXUS

43

bours. The cells of each pouch push down ventrally on either side of
the gut, the outer ones applying themselves to the body wall to form
the somatopleure, the inner to the gut wall as splanchnopleure (Fig.
1 8 d). The inner wall of the mesoderm on either side of the nerve-
cord thickens to form the myotome, and a tongue of cells growing up
between this and the nerve-cord forms the sheaths of the latter and

lOOfJL

my coel.

som.pl.

A B C D

Fig. i 8. Further stages in the development of amphioxus as seen in transverse

sections.
A, stage of three somites; B, six somites; c, nine somites; D, eleven somites, arch, archenteron;
coel. coelom; tries, mesoderm; my. myotome; my. coel. myocoele; n. notochord; n.c. nerve-cord;
som.pl. somatopleure; spl.pl. splanchnopleure; spl.coel. splanchnic coelom. (After Hatschek.)

neurp.

g. ioo u.

Fig. 19. Young amphioxus, soon after hatching.
g. gut; «. notochord; n.c. nerve-cord; nenr.c. neurenteric canal; neur.p. neuropore.

probably also the fin-ray boxes and other 'mesenchymal' tissues. The
upper part of the coelomic cavity, the myocoele, becomes separated
from the ventral splanchnocoele. Whereas the former becomes almost
completely obliterated, the latter expands to form the adult coelom,
the cavities between the adjacent sacs breaking down.

While this differentiation of the mesoderm has been proceeding the
animal has elongated into a definitely fish-like form. The neural tube
is a small dorsal canal, opening by an anterior neuropore and con-
tinuous behind through a neurenteric canal with the gut (Fig. 19).
The larva hatches when only two segments have been formed and
swims at the sea surface by means of its ciliated epidermis, turning on
its axis from right to left as it proceeds with the front end forwards.

The mouth now appears as a circular opening and then moves
over to the left side and becomes very large. From this time onward
the whole development is markedly asymmetrical, presumably in

44

CHORDATE ORGANIZATION

ii. ii

connexion with the spiral movement and method of feeding. The first
gill-slit also forms near the midline but moves up on to the right side
(Fig. 20). At about the same time the right side of the pharyngeal wall
develops into a V-shaped thickening, the endostyle. Behind this there
forms a tube, the club-shaped gland, joining the pharynx to the
outside and formed by the closure of a groove in the side of the
pharynx. The significance of this organ is still obscure ; it is presum-
ably connected with the feeding process, which begins at this stage.
It has been thought to represent a gill-slit.

Fig. 20. Larval amphioxus before metamorphosis.

an. anus; c.s.g. club-shaped gland; end. endostyle; g. gill-slits lying on right

side, which will later move over to the left. H.p. Hatschek's pit (left first

coelomic sac); m. lower edge of mouth, lying on left side; n. notochord;

p. pigment spot.

The first two coelomic pouches differentiate, asymmetrically, at
this time. That on the right becomes the coelomic cavity of the head
region, while the left one acquires an opening to the exterior and a
heavily ciliated surface. This is perhaps also connected with the feed-
ing-systems and becomes developed into Hatschek's pit of the adult.
Its interest to the morphologist lies in the fact that the first coelomic
cavity opens to the exterior in other early chordates and in some
vertebrates (p. 206). The pit has thus some claim to be considered the
equivalent of the hypophyseal portion of the pituitary gland.

Further gill-slits develop in the mid-ventral line and move over on
to the right side until fourteen have been so formed. Meanwhile, a
further row of eight slits appears above that already formed. These are
the definitive slits of the right side and presently the larva proceeds to
become symmetrical by movement of eight of the first row of slits
over to the left side, the remainder disappearing. At this 'critical

ii. ii LARVA OF AMPHIOXUS 45

stage' with eight pairs of slits the larva pauses for some time before
further changes. It is interesting that this is the time at which it most
nearly represents what might have been an ancestral craniate, with
eight branchial arches (p. 145). Further slits are then gradually added
in pairs on both sides. Each slit becomes subdivided, soon after its
formation, by the downgrowth of a tongue bar. The atrium is absent
from the early larva. Metapleural folds then appear on either side and
are united from behind forwards to form a tube below the pharynx.
During the later stages of development the larva sinks and finally rests
on the bottom while undergoing the migration of gill-slits that con-
stitutes its metamorphosis. In other species the larva remains longer
in the plankton, becoming large and even showing quite large gonad
rudiments. These were at first thought to be adults of a new genus
(Amphioxides).

The development of amphioxus, like its adult organization, shows
us many features of the plan that is typical of all chordates and was
presumably present in the earliest of them. Thus the cleavage, in-
vagination, and mesoderm formation recall those of echinoderms and
other forms similar to the ancestors of the chordates, and also show a
pattern from which all later chordate development can be derived.
Unfortunately we cannot pursue this study as far as we should like
because of the difficulty of investigating the development of am-
phioxus. Modern embryologists aim at tracing the morphogenetic
movements by which the organism is built, and ultimately at dis-
covering the forces responsible for these processes. We still remain
ignorant of the details of these morphogenetic movements, and can
only guess that the system of cell activities by which an amphioxus
is built represents quite closely the original set of morphogenetic
processes of vertebrates (Young, 1957, p. 633).

There are, of course, some special features connected with the
method of life of the larva, and especially with its asymmetry. The
strange sequence of gill formation, the immense left-sided larval
mouth, perhaps the club-shaped gland, and Miiller's organ, may show
considerable modifications of relatively recent date. However, the
earliest chordates probably fed by means of cilia and were planktonic,
so we must not too hastily assume that even these asymmetrical features
are novelties.

The division of the mesoderm of amphioxus into a series of sacs
presents an interesting problem. The segmentation of the mesoderm
of vertebrates is restricted to the dorsal region. In the lowest chord-
ates (see p. 51), as in their pre-chordate ancestors, there are three

46 CHORDATE ORGANIZATION n. 11-12

coelomic cavities, but it is probable that the many segments of verte-
brates arose in order to provide a set of muscles able to contract in a
serial manner for the purpose of swimming. Their segmentation would
thus be a relatively late development, not related to the segmentation
of annelids, which divides the whole body into rings. Accordingly the
ventral part of the vertebrate coelom usually remains unsegmented.
But in amphioxus (and in the lamprey) it is subdivided from its first
appearance and only becomes continuous later. The best interpreta-
tion of this condition is to suppose that in order to provide a series of
myotomes a rhythmic process subdividing the mesoderm was adopted.
In its earliest stages this affected the whole mesoderm, ventral as well
as dorsal, but later became restricted to the dorsal region. New
morphogenetic processes may often pass through stages of refinement
and simplification in such ways.

12. Amphioxus as a generalized chordate

Amphioxus provides us, then, with a valuable example of a chordate
that retains the habit of ciliary feeding, which was probably that of the
earliest ancestors of our phylum. No doubt in connexion with this,
and the bottom-living habit, there are many specializations; the
enormously developed pharynx with its atrium, the asymmetry, and
so on; but the general arrangement of the body is almost diagram-
matically simple, and it may well be that amphioxus shows us a stage
very like that through which the ancestors of the craniates evolved.
Perhaps next the larva remained longer in the plankton and became
mature there. The Amphioxides larvae show signs of such a change.

This might give rise to a suspicion that amphioxus is not an
ancestral type but a simplified derivative of the vertebrates, perhaps
a paedomorphic form. It possesses, however, sufficient peculiar
features to make this view unlikely. Neoteny might explain the
regular segmentation, separate dorsal and ventral roots, and other
features, but can hardly account for the method of obtaining food,
for the condition of the skin, or for the presence of nephridia. It may
be, therefore, that amphioxus shows us approximately the condition
of the early fish-like chordates, living in the Silurian some 400 million
years ago, and that it has undergone relatively little change in all the
time since.

Ill

THE ORIGIN OF CHORDATES FROM FILTER
FEEDING ANIMALS

1 . Invertebrate relatives of the chordates

We have seen in the organization of amphioxus the plan of chordate
structure as it may have existed in Palaeozoic times. Before proceeding
to discuss the later forms that evolved from animals of this sort we may
first look yet farther backwards to discuss the origin of the whole
chordate phylum from still earlier ancestors. The great difficulty of
such an inquiry is itself a stimulus and a challenge. Typical fish-like
chordates were undoubtedly established by the Ordovician period,
but we have no good fossil record of their earliest form and this must
therefore be deduced from study of amphioxus and later animals. No
fossils that suggest chordate affinities have been found in the still
earlier rocks. There are, however, certain strange animals alive today
which, though not of fish-like type, show undoubted relationship with
our group. These might, of course, be degenerate offshoots from later
periods, but careful comparison suggests that they have been separa-
ted for a very long time and provide us with relics of some of the early
stages of our history.

The first step in our inquiry, however, before discussing these
forms, should be to find out, if possible, which of the main lines of
invertebrate animals shows the closest affinity with the chordates.
Almost every phylum in the animal kingdom has been suggested,
including the nemertines. Many still suppose that the annelids and
arthropods, because of their metameric segmentation, are related to
the chordates, but closer examination shows that the similarities are
superficial. The segmentation of these annulate animals is an almost
complete division of the whole body into rings, and all the organ
systems are affected by it to some extent. In chordates only the dorsal
myotomal region is segmented; even the mesoderm is not divided in
its ventral region in most animals. Moreover, the whole orientation
of the body differs in the two groups. The vertebrate nerve-cord is
dorsal to the gut, in annulates the nerve-cord is below and the 'brain'
above. The blood circulates in opposite directions, the limbs are
based on quite different plans, and so on. Attempts have been made
to get over these difficulties by turning the invertebrate upside down!
Patten and Gaskell carried such theories to extremes and tried to

48 ORIGIN OF CHORDATES in. i

show a relationship of chordates with the eurypterids, heavily ar-
moured arachnids of the Cambrian and Silurian. These animals show
a certain superficial resemblance to some early fossil fishes, the cepha-
laspids of the Devonian (Fig. 83), and these workers, with great
ingenuity, claimed to find in them evidence of the presence of many
chordate organs.

The safest evidence of affinity is a similarity of developmental pro-
cesses : animals that develop very differently are unlikely to be closely
related. The development of modern annulates is utterly different
from that of chordates. The cleavage by which the fertilized egg is
divided into blastomeres follows in annulates a 'spiral' plan, in which
every blastomere arises in a regular way and the future fate of each
can be exactly stated. In later annulates, such as the arthropods, this
plan is complicated by the presence of much yolk, but even in these
animals the cleavage does not resemble that of chordates, which is
radial or 'irregular', the cells not forming any special pattern. This
characteristic has been used to divide the whole animal kingdom into
two major groups, Spiralia and Irregularia.

The next stage of development, gastrulation, by which the ball of
cells is converted into a two-layered creature, also occurs very differ-
ently in the two groups. Our knowledge of the mechanics of the pro-
cesses by which this change is produced is still imperfect, in spite of
recent advances, but in lower chordates it occurs by invagination, the
folding in of one side of the ball of cells to form an archenteric cavity
communicating with the exterior. In annulates this is never seen;
the cells that will go to form the gut migrate inwards either at one
pole or all round the sphere and only later form themselves into a
tube, which comes to open secondarily to the outside. It is probable
that when we know more of the forces by which the gastrulation is
produced the difference will appear even more marked than it does
from this crude and formal statement that gastrulation in chordates is
by invagination, in annulates by immigration.

The same applies to the method by which the mesoderm and coelom
are formed. In lower chordates the third layer is produced by separa-
tion from the endoderm, so that the coelom is continuous with the
archenteron and is said to be an enterocoele. In annulates cells
separate in various ways to form the mesoderm and a coelom then
arises within this solid mass as a schizocoele. It is true that in some,
indeed many, of the higher chordates the coelom is never continuous
with the archenteron, but its method of development shows it to be
a modified enterocoele.

in. i INVERTEBRATE RELATIVES OF CHORDATES 49

In all these points of development the chordates differ from the
annulates, but resemble the echinoderms and their allies. Further
features support this latter relationship. One of the most important
of these is that the echinoderm-like animals, and some of the early
chordates, have a larva with longitudinal ciliated bands, very different
from the trochophore larva, in which the bands run transversely
round the body, which is found in the other line of animals. The
nervous system of annulates consists of a set of ganglionated cords,
whereas in echinoderm-like animals it is a diffuse sheet of cells and
fibres below the epidermis. The nerve-cord of the chordates can be
derived from the latter but not easily from the former condition.
Many further points could be cited, for instance, the presence of a
mesodermal skeleton in both chordates and echinoderms, but not in
annulates. It may be that there are also fundamental biochemical
differences. Most of the spirally cleaving types of animal conduct their
energy transfers with arginine phosphate, whereas vertebrates,
amphioxus, ascidians, and ophiuroid echinoderms use creatine phos-
phate. Balanoglossus and echinoids have both.

In the study of evolution it is not sufficient merely to make formal
comparisons, we must try to find out and compare the plan of develop-
ment and structure common to all members of two groups, a technique
often requiring great knowledge and good sense. When this is done in
the present case it will be found that the essential plan of development
of annulates involves spiral cleavage, gastrulation by immigration,
and a coelom formed as a schizocoele, a trochophore-like larva, and
full segmentation of the mesoderm. It is exceedingly unlikely that
such animals have given rise to chordates with their very different
development, which we may crudely define as showing radial cleavage,
gastrulation by invagination, and larva of echinoderm type.

Extending this method we may divide the whole world of Metazoa
by similar criteria into Spiralia or Polymera and Irregularia or Oligo-
mera. The former include besides the annulates the molluscs and
platyhelmia, whereas the latter group contains, in addition to the
chordates, the echinoderms, brachiopods, polyzoa (ectoprocta), grap-
tolites, pogonophora, and Phoronis. The animals in this latter group
seem at first sight to be very different from the chordates in outward
form, but the farther we look into their fundamental organization, the
more we become convinced that the ancestors of the fish-like animals
are to be found here. By study of the relics of the early chordates it is
possible to trace the history of this strange change with some plausi-
bility, though its full details will probably never be known.

5Q

ORIGIN OF CHORDATES

III. 2

2. Subphylum Hemichordata (= Stomochordata)

Class i. Enteropneusta

Balanoglossus ; Glossobalanus ; Ptychodera; Saccoglossus
Class 2. Pterobranchia

Cephalodiscus ; Rhabdopleura

Fig. 21. Balanoglossus, removed from its tube and seen from the dorsal side.
abd. abdomen; atr. atrium; an. anus; c. collar; h.c. hepatic caeca; p. proboscis; ph. pharynx.

(From van der Horst.)

Fig. 22. Balanoglossus in its tube in the sand. (After Stiasny.)

In the Hemichordata are placed animals of two types, the worm-
like Balanoglossus and its allies (Enteropneusta) and two sedentary
animals, Cephalodiscus and Rhabdopleura (Pterobranchia). The Entero-
pneusta are mostly burrowing animals (Figs. 21 and 22) varying in
different species from 2 cm to over 2 metres long. Several genera are
recognized (e.g. Balanoglossus, Saccoglossus, Ptychodera) and they
occur in all seas. Saccoglossus occurs around the British coast. The
body is soft, without rigid skeletal structures, and divided into
proboscis, collar, and trunk. The animals are very fragile and it is
difficult to collect specimens in which the hind part of the trunk

III. 2

MOVEMENT OF BALANOGLOSSUS

5i

('abdomen') is intact. The proboscis, collar, and trunk each contain
a coelomic cavity, and the coeloms of the proboscis and collar are
distensible by intake of water through a single proboscis pore and
paired collar pores. The skin is richly ciliated all over the body. The
outer epithelium is thus unlike the squamous, layered skin of higher
forms (Fig. 23). It contains numerous gland-cells, whose secretion
is very copious, so that the animals are always covered with slime.

neur.s

Fig. 23. Section of the epidermis of an enteropneust.
h.m. basement membrane; ep. epidermal cell; gl. I and 2, different types of gland cell;
neur. neuron; neur.s. neuro-sensory cell; n.g.p. process of epidermal cell acting as neuro-
glia in the nerve net. (After Bullock, v. der Horst and Grasse.)

A characteristic feature is an unpleasant smell, resembling that of
iodoform, which possibly serves, like the mucus, as a protection.

Below the skin is a nerve plexus receiving the inner processes of
receptor cells and containing ganglion cells (Fig. 23). Deep to this are
muscles running in various directions. It is said that the animal moves
by first pushing the proboscis and collar forward through the sand
and then drawing the body after it. Protrusion of the proboscis can-
not, however, be very vigorous. It may perhaps be produced by ciliary
action distending the coelom as is usually stated — more probably by
circular muscles, but these are weak. Numerous longitudinal muscles
are present, however, in the proboscis and trunk and are partly
attached to a plate of skeletal tissue in the collar. This tissue is
attached to the ventral side of a forwardly directed diverticulum of
the pharynx. The wall of this is thick, composed of vacuolated cells,
and bears a certain resemblance to a notochord (Fig. 24). A notochord
extending throughout the length of the body would clearly be dis-
advantageous for an animal whose main movements are lengthening

52

ORIGIN OF CHORDATES

and shortening. It is possible that the diverticulum and plate found
in the collar represent the remains of a notochord, serving as a
fixed point by which the body is drawn forward on to the proboscis.
However, many prefer to call it a 'stomochord' to avoid too close a

card.

Fig. 24. Diagrammatic section of front end of Balanoglossus.
c. collar coelom ; card.s. sac around heart ; div. pharyngeal diverticulum ('stomochord') ; dn
dorsal nerve-root ; dv. dorsal vessel ; gl. glomerulus ; £s. gill-slit ; Im. longitudinal muscles of
proboscis ; n.c. nerve-cord ; p.p. proboscis pore ; sk. skeletal plate. (Modified after Spengel.)

comparison with the notochord. The external cilia probably play a
considerable part in locomotion; possibly they are the chief burrowing
organs, the muscles serving mainly to perform escape movements.

The mouth lies in a groove between the proboscis and collar (Fig.
25). The proboscis contains many mucus-secreting cells and the food
particles are captured on its surface and conveyed by ciliary currents
to the mouth. In the anterior part of the trunk there is a wide pharynx,
opening by a series of gill-slits (Figs. 24, 26). These resemble the
gills of amphioxus in the presence of a supporting skeleton in the gill
bars; there are also tongue bars dividing the slits from above, and

III. 2

FEEDING

53

Fig. 25. Feeding-currents on proboscis of Glossobalanus, shown by placing the animal

in water containing carmine particles. The particles (gra.) are either taken directly into

the mouth (?«.) as at w., or are caught up in strands of mucus (sec.) and passed backwards.

(From Barrington, Quart. J. Micr. Set. 82, by permission.)

Fig. 26. Transverse section of the pharynx of Glossobalanus.

cil. cilia of the gill bars; dc. dorsal chamber of pharynx; es. epibranchial strip; gp. gill pore;
VC. ventral chamber of pharynx. (From Barrington. With permission as for Fig. 25.)

horizontal synapticulae strengthening the gill arches. The slits open
in some species into an atrium formed by lateral folds, usually turned
upwards to leave a long mid-dorsal opening. In some species each
slit opens to a gill pouch. The whole branchial apparatus perhaps

54

ORIGIN OF CHORDATES

III. 2

assists in the process of feeding, probably by serving to filter off the
excess water from the material already collected on the proboscis,
which often consists of large amounts of sand or mud. Relative to the
size of the animal the pharynx is less extensive than in amphioxus,
presumably because ciliary surfaces are provided on the outside and
also large masses of sand are forced into the mouth during locomotion.
There is no endostylar apparatus, but the ventral part of the pharynx
is often partly separated from the rest (Fig. 26). Along this groove the
matter ingested is passed to a straight oesophagus and intestine open-
ing by a terminal anus. There is no true tail in the adult but a post-
anal region is present in some species during development. Numerous

v'.v. ph.

Fig. 27. Diagram of the blood system of Balanoglossus.

col. collar; d.v. dorsal vessel; glom. glomerulus; hp. hepatic caeca; m. mouth; not. 'notochord';
p. proboscis; ph. pharynx; v.v. ventral vessel. (After Bronn.)

hepatic caeca in the anterior part of the intestine can be seen from the
outside as folds of the body wall, often highly coloured.

The blood system consists of a complex set of haemocoelic spaces,
communicating with large dorsal and ventral vessels (Fig. 27).
The former enlarges into a sinus anteriorly and this is partly sur-
rounded by the wall of a pericardial cavity, which contains muscles
and may be said to be the heart, though clearly lying in a very different
position from that of other chordates. From the sinus, vessels proceed
to the proboscis and round the pharynx to the ventral vessel. The
blood is said to move forwards in the dorsal and backwards in the
ventral vessels. The front of the sinus forms a series of glomeruli,
covered by a region of the proboscis coelom specialized to form
excretory cells, the nephrocytes, some of which drop off into the
coelom. The blood is red in some species but usually colourless. It
contains a few amoebocytes.

The nervous system is one of the most interesting features of
Enteropneusta. It resembles that of echinoderms in consisting of a
sheet of nerve-fibres and cells lying beneath the epidermis all over the
body (Fig. 23). This sheet is thick in the mid-dorsal and mid-ventral

BEHAVIOUR OF BALANOGLOSSUS

55

lines, and in the dorsal part of the collar region it is rolled up as a
hollow neural tube, open at both ends (Fig. 24). These unmistakable
resemblances not only to the uncentralized sub-epithelial plexus of
echinoderms but also to the hollow dorsal nerve-cord of vertebrates
are most instructive, showing the affinity of the groups and the origin
of the general plan of the vertebrate nervous system. There are no
organs of special sense, unless this is
the function of a patch of special cili-
ated cells on the collar. Receptor cells
all over the body send their processes
into the nerve plexus (Fig. 23), on the
primitive plan of neurosensory cells
found elsewhere in vertebrates only in
the olfactory epithelium and the retina.
The plexus is remarkable in receiving
fibres from the outer ciliated epithelial
cells, which thus represent the epen-
dyma, the earliest form of neuroglia
(Fig. 23). Nothing is known of the
organization of pathways or of the
connexions with the muscles. The
collar nerve-cord contains giant nerve-
cells whose axons proceed backwards
to the trunk and forward to the pro-
boscis (Fig. 28). They are probably
responsible for rapid contractions
(Knight-Jones, 1951).

Bullock has investigated the beha-
viour of the animals and found only

one clear-cut reflex, namely, a contraction of the longitudinal muscles
in response to tactile stimulation. Isolated pieces of the body are able
to show reflex responses, moving away from light or tactile stimuli.
Such local actions are an interesting sign of the uncentralized nature
of the nervous system, and similar actions are found in echinoderms.
A further sign of lack of special conducting pathways is that stimula-
tion of flaps of body wall partly severed from the rest produces
generalized contraction, proving that conduction can occur in all
directions. The dorsal and ventral nerve-cords do, however, act as
quick conduction pathways, and contraction of the trunk following
stimulation of the proboscis is delayed or absent if one, and especially
if both, cords have been cut.

Fig. 28. Diagram of certain tracts in
the nervous system of Balanoglossus .

com. circular connective; col.coel. collar
coelom; col.n.c. collar nerve-cord; ep.pl.
nerve plexus in epidermis of trunk; gp.
gill pore; tr. coel. trunk coelom; tr.n.c.
trunk nerve-cord. (From Bullock, J.
Comp. Neurol., vol. 80, by permission.)

56 ORIGIN OF CHORDATES m. 2

Perhaps the most interesting behaviour observed was the activity
shown by an isolated proboscis, collar, trunk, or portion of trunk.
These organs may move around vigorously in an exploratory manner;
evidently the main nerve-cords are not necessary for the initiation of
action, as is the central nervous system of higher chordates.

There are nerve-fibres in the walls of the pharynx and oesophagus,
where peristaltic movements have been observed. Their relationship

an.

Fig. 29. Young tornaria larva, seen from the side.

an. anus; ap. apical organ; cb. longitudinal ciliated band; m. mouth;

pb. posterior ciliated band; pp. proboscis pore. (After Stiasny.)

to the rest of the nervous system is unknown. They may represent
the beginnings of an autonomic nervous system.

The sexes are separate in enteropneusts and the gonads resemble
those of amphioxus in being a series of sacs developing from cells just
outside the coelom. These proliferate and bulge into the coelom,
covered by the somatopleure. They acquire a cavity and each opens
by a narrow duct to the exterior, fertilization being external. The
development is remarkably like that of echinoderms. Cleavage is holo-
blastic and resembles that of amphioxus and ascidians, gastrulation is
by invagination, and the coelom is formed as an enterocoele, later
becoming subdivided into proboscis, collar, and trunk coeloms. Hatch-
ing occurs to produce a pelagic tornaria larva, with a ciliated band that
has exactly the relations found in the dipleurula larva of echinoderms.
The band passes in front of the mouth, down the sides of the body,
and in front of the anus (Fig. 29). It then divides into more dorsal

III. 2

DEVELOPMENT OF BALANOGLOSSUS

57

and ventral sections, exactly as in the production of the bipinnaria

larva of a starfish. This arrangement differs essentially from the rings

of cilia that pass round the body in the trochophore larva found in the

annelids and other spirally cleaving forms. In later tornaria larvae

there is, however, in addition to the longitudinal bands always a

posterior ring of stout cilia (telotroch), and in large oceanic forms

(which may reach 8 mm in length) the longitudinal band itself is

prolonged into prominent tentacle-like loops (Fig. 30). The cilia of

the posterior ring are purely locomotive, while those of the band set

up feeding-currents converging to the

mouth. As the larva becomes larger the

ciliary surface needed for locomotion

and feeding has to increase relatively

faster than the increasing mass of the

body, the latter following the cube but

the former only the square of the linear

dimensions. Accordingly the cilia of

the locomotive ring become broadened

and flame-like, while the convolutions

of the longitudinal (feeding) band reach

fantastic proportions. In some types,

however (Saccoglossus), the pelagic

phase is brief and the telotroch alone

is formed.

Finally the larva sinks, becomes con-
stricted into three parts, and undergoes metamorphosis into the worm-
like adult. This development is so like that of an echinoderm that it
would be necessary to consider the enteropneusts to be related to
that group even if no other clues existed. Such close similarity in
the fundamentals of development cannot be due to chance.

These animals thus provide a very remarkable and sure demonstra-
tion that the chordates are related to the echinoderms and similar
groups. The general arrangement of the nervous system as a sub-
epithelial plexus, as well as the whole course of the development, show
the affinity with the invertebrate groups, whereas the hollow dorsal
nerve-cord and the tongue-barred gill-slits are by themselves sufficient
to show affinity with the chordates, this affinity being also perhaps
suggested by other features, such as the 'notochord'. As we have seen
already, affinities are not to be determined by single 'characters' but
by the general pattern of organization of animals and especiallv that of
their development. The organization of the enteropneusts is certainly

Fig.

30. Older tornaria larva seen
from ventral surface.

Letters as Fig. 29; coel. proboscis
coelom. (After Stiasny.)

58 ORIGIN OF CHORDATES in. z-

highly specialized for their burrowing life, but showing through the
special features we can clearly see a plan that has similarity with both
the echinoderms and the chordates. The special value of study of these
animals is that it proves decisively that an affinity between these
groups exists. Exactly how they are all related is a more speculative
matter, which we shall deal with later (see p. 74).

3. Class Pterobranchia

These are small, colonial, marine, sedentary animals, which show
some signs of the general echinoderm-chordate plan of organization
we have been discussing. Cephalodiscus (Fig. 31) has been found on the
sea bottom at various depths, mainly in the southern hemisphere:
there are several species. The colony consists of a number of zooids
held together in a many-chambered gelatinous house. The zooids are
formed by a process of budding, but do not maintain continuity with
each other. Each zooid has a proboscis, collar, and trunk; there are
coeloms in each of these parts, and proboscis and collar pores. The
collar is prolonged into a number of ciliated arms, the lophophore, by
means of which the animal feeds. There is a large pharynx, opening
by a single pair of gill-slits, which serve as an outlet for the water
drawn in by the cilia of the tentacles for the purpose of bringing food.
The intestine is turned upon itself, so that the anus opens near the
mouth. A thickening in the roof of the pharynx corresponds exactly in
position with the stomochord and contains vacuolated cells. The blood
system consists of a series of spaces arranged on a plan similar to that
in Balanoglossus. There is a dorsal ganglion in the collar, but this is
not hollow. The gonads are simple sacs and development takes place
in the spaces of the gelatinous house. Gastrulation is by invagination
at least in some species and the coelom is formed as an enterocoele.
The larva somewhat resembles that of ectoproctous polyzoa, which
is not closely similar to the echinoderm larvae, but could be derived
from the same plan.

Rhabdopleura occurs in various parts of the world, including the
North Atlantic and northern part of the North Sea. The zooids are
connected together and have proboscis, collar, and trunk, ciliated
arms, coelomic spaces with pores (not 'nephridia' as is sometimes
stated) and stomochord, but no gill-slit. The development is not
known.

The Pterobranchia thus show undoubted signs of the enteropneust-
chordate plan of organization and provide also an interesting sug-
gestion of possible affinities with Polyzoa, Brachiopoda, and Phoronis.

in. 3 PTEROBRANCHS AND POLYZOANS 59

Like the Pterobranchia the Polyzoa Ectoprocta are sessile, with mouth
and anus pointing upwards. They feed by means of the cilia borne on
a horseshoe-ring of tentacles (the lophophore); but there is no division

Fig. 31. Longitudinal median section of Cephalodiscus.

a. anus; b.c. 1, 2, and 3 body cavities; int. intestine; lo. lophophore;

m. mouth; nch. 'notochord'; n.s. nervous system; oes. oesophagus;

op. operculum (collar); ov. ovary; ph. pharynx; pp. proboscis pore;

ps. proboscis; St. stomach; st.k. stalk.

(Modified after Harmer, Cambridge Natural History, Macmillan.)

into proboscis, collar, and trunk, and no tripartite coelom. The nervous
system is in the condition of a sub-epithelial plexus, which is folded,
around the base of the lophophore, to form a hollow tube — a remark-
able point of similarity to the chordates. Even though it is difficult to
compare this tube exactly with the nerve-cord of chordates, it is at
least evidence of the organization of the nervous system on a plan that
allows of such folding. It is probable that the modern pterobranchs
are the surviving members of the ancient group of graptolites, but

6o

ORIGIN OF CHORDATES

in. 3-

mu.

gen

these are known only from the skeleton. The Pogonophora may also
be distantly related, their larva can be regarded as of tornaria type,
the coelom develops as in enteropneusts and the larval body shows
three parts, as does that of the adult in some species.

Although it would be unwise to
suggest close relationship between the
polyzoans and the pterobranchs, the
similarities are sufficient to suggest that
the chordates arose from sedentary
creatures, feeding by means of ciliated
tentacles. The evidence is sufficiently
strong to encourage us to look for the
presence somewhere in the line of verte-
brate ancestry of an animal with this
habit. The difficulties of this view arise
when we come to consider how the fish-
like organization of a free-swimming
animal first appeared, a question better
dealt with after consideration of the
tunicates.

4. Subphylum Tunicata. Sea squirts

In the adult ascidians or sea squirts
there is no obvious trace of the fish-like
form at all. The majority of these
animals are sac-like creatures living on
the sea floor and obtaining their food
by ciliary action. Often the separate
individuals are grouped together to form
large colonies, but in Ciona intestinalis,
common in British waters, the indi-
viduals occur separately, and this is possibly the primitive con-
dition for the group. The whole of the outside of the body is covered
by a tunic, in which there are only two openings, a terminal mouth and
a more or less dorsal atriopore, both carried upon siphons (Fig. 32).
The tunic is made mainly of a carbohydrate, tunicin, closely related to
cellulose, with which is combined about 20 per cent, of glycoprotein.
It is secreted by the epidermis but contains special cells that have
arrived there by migration from the mesoderm. In some tunicates
calcareous secretions of various shapes are found in the tunic. The
mantle that lines the tunic is covered by a single-layered epidermis.

Fig. 32. Diagram of structure of
Ciona.

atr.p. atriopore ; e. endostyle ; gen.d.
genital duct; h. heart; int. intestine;
m. mouth; mu. muscle; oes. oeso-
phagus; ph. pharynx; st. stomach.
(After Berrill.)

in. 4

ORGANIZATION OF CIONA

61

Ascidians are often brightly coloured, the pigment being either in the
tunic or the underlying body, which shows through the transparent
tunic. The colour can change, at least over a period of some days. Little
is known about the origin of the pigment, but it is sometimes derived
from the blood-pigment and may lie in pigment cells.

The mantle is provided with muscle-fibres running in various
directions but mainly longitudinally, and serving to draw the animal
together, with the production of
the jet of water from which the
animals derive their common
English name.

The greater part of the body is
made up of an immense pharynx,
beginning below the mouth and
forming a sac reaching nearly to
the base (Fig. 32). The sac is
attached to the mantle along one
side (ventral) and is surrounded
dorsally and laterally by a cavity —
the atrium. This pharynx is, of
course, the food-collecting appa-
ratus ; its walls are pierced by rows
of stigmata (gill-slits) whose cilia
set up a food current entering at
the mouth and leaving from the
atriopore. The entrance to the
pharynx is guarded by a ring of
tentacles, which may be compared

with the velum of amphioxus. The stigmata are very numerous
vertical cracks, all formed by sub-division of three original gill-
slits. Tongue bars grow down to divide each slit and then from each
tongue bar grow horizontal synapticulae. This arrangement has clear
resemblance to that of amphioxus and results in the production
of a pharyngeal wall pierced by numerous holes. Immediately within
the stigmata there is a series of papillae, provided with muscles and
cilia. There is an endostyle, which has three rows of mucus cells
on each side, separated by rows of ciliated cells and with a single
median set of cells with very long cilia (Fig. 33). The mucus secreted
in the endostyle is caught up on the papillae, whose muscles move
them rhythmically, spreading a curtain of mucus over the inside of
the pharynx. Food particles are caught in the mucus, which moves

Fig. 33. Transverse section of the endo-
style of Ciona.

lat. cil. lateral cilia; med. cil. long median cilia;
mu. mucous cell. (After Sokoloska.)

6z ORIGIN OF CHORDATES in. 4

upwards and is then passed back to the oesophagus by the cilia of a
dorsal lamina or of a series of hook-like 'languets'. Autoradiographs
made from tunicates that have been provided with isotopes of iodine
show that iodination occurs in certain cells lying above the glandular
tracts of the endostyle. Iodine is also abundant in the tunic, as it is in
the exoskeletal structures of molluscs and insects. When it became
of metabolic value its production may have become concentrated in
the pharynx (see p. 118).

The extensive ciliated surface of the pharyngeal wall ensures the
passage of large volumes of water inwards at the mouth and out at the
atriopore. Rapid change of the water is also produced by periodic
muscular contractions (p. 65). The pressure of the exhalant current is
sufficient to drive the water that has been used well away from the
animal.

The oesophagus leads to a large 'stomach' with a folded wall con-
taining gland-cells, which produce digestive enzymes. These include
much amylase, invertase, small amounts of lipase, and a protease of
the tryptic type. The organ is therefore not to be compared with the
stomach of vertebrates. A branching 'pyloric gland' opens into the
lower end of the stomach. From the stomach a rather short intestine
leads upwards to open inside the atriopore; this is apparently the
absorptive region of the gut.

The heart lies below the pharynx and is a sac, surrounded by a
pericardium (see p. 63) and communicating with a system of blood
spaces derived from the blastocoele. The larger of these spaces have
an endothelial lining; the biggest is a hypobranchial vessel below the
endostyle, from which branches pass to the pharynx. From the oppo-
site end of the heart springs a large visceral vessel and others pass to
the dorsal side of the pharynx, tunic, body wall, &c. The heart is
peculiar in that the beat can proceed in either direction. After passing
blood into the hypobranchial vessel and gills for a few beats, its direc-
tion reverses, passing the blood to the viscera. This reversal is pro-
duced by the presence of two pacemaker centres, each capable of
initiating rhythmical contractions, one at either end of the heart.
Stimulation of these by warming and cooling allows control of the
reversal of the beat. There are no capillaries and the blood system is
a haemocoele. The blood-plasma is colourless but contains corpuscles,
some of which are phagocytes, while others contain orange, green, or
blue pigment (in different species). The green and other pigments are
remarkable in that they contain vanadium. In some ascidians (Molgula)
some individuals contain vanadium, others niobium (Carlisle, 1958).

III. 4

ORGANIZATION OF CIONA

63

The vanadocytes contain much sulphuric acid and the metal is
associated with a chain of pyrrol rings. This haemovanadin is able to
reduce cytochrome but it remains uncertain what part the pigment
plays in respiration. The blood turns blue in air but cannot take up
more oxygen than can sea water.

The blood is isotonic with sea water, and ascidians appear to have
little or no power of regulating their osmotic pressure; none of them
is found in fresh water. They are not even able to colonize brackish
waters or those of low salinity. For example, they are rare in the
Baltic Sea, from which only six
species have been reported. Only
one species, Molgula tubifera, has
been reported from the Zuider
Zee (salinity 8-4 per mille).

A possible reason for this in-
ability to regulate the internal
composition is perhaps the need
to expose a large surface to the
water. There are no tubular ex-
cretory organs such as could be
used to maintain an osmotic
gradient. Ninety-five per cent of
the nitrogen is excreted as am-
monia. Cells known as nephro-
cytes found in the blood and
elsewhere contain concretions
within the cytoplasm and these
may in some cases be stored in an excretory sac until the animal dies.

There has been much debate as to whether the tunicates possess
a coelomic cavity. The heart develops from a plate of cells arising
early from the mesoderm and lying between ectoderm and endoderm.
This becomes grooved and folded to make the heart itself and the
pericardium. The irregular system of haemocoelomic spaces around
the pharynx and elsewhere is usually said to consist of 'mesenchyme'
and to be derived from the blastocoel and therefore not coelomic, but
its walls are mesodermal. The situation is complicated by the presence
of a pair of outpushings from the pharynx, the epicardia, or perivis-
ceral sacs, which end blindly on either side of the heart (Fig. 34).
Berrill and others have suggested that these epicardia may be com-
pared with coelomic cavities. Their function in the open condition
in which they are found in Ciona is perhaps to allow sea water to

Fig. 34. Section through base of Ciona,

showing heart, fit., in pericardium, p.,

and the epicardia, e.p., opening into the

pharynx, b.s.

at. atrium; g. gonad; int. intestine.

64 ORIGIN OF CHORDATES in. 4

circulate about the heart and hence to help excretion (and respiration ?).
In other ascidians the epicardium loses its connexion with the pharynx.
The closed sac functions in some cases as an excretory organ, con-
taining concretions of uric acid, whereas in other animals it becomes
the main source of the cells that make the asexual buds.

The central nervous system consists of a round, solid ganglion
(Fig. 36), lying above the front end of the pharynx. The ganglion has
a layer of cells around the outside and a central mass of neuropil and
is therefore quite unlike the nerve-cord of a vertebrate. From the
ganglion nerves proceed to the siphons, other parts of the mantle,
muscles, and viscera. Receptor cells with nerve-fibres ending around
the base have been described, especially in the siphons. The gut is
said to contain a plexus of cells and fibres, whose relation to the
autonomic system of higher forms remains uncertain.

Movement consists mainly of contraction and closure of the aper-
tures. Light touching of either siphon causes closure proportional to
the strength of the stimulus. Stronger stimuli cause closure of both
siphons and if very strong there is contraction of the whole body and
ejection of the water in the pharynx and atrium. Stimulation just
inside either siphon produces closure of the other one and also, if
strong enough, contraction of the body, ensuring that a jet of water
sweeps out the aperture that received the stimulus. These crossed
reflexes depend upon the integrity of the ganglion.

The surface of the body is sensitive to changes in light intensity,
and these are followed by local or total contractions, according to their
extent. After removal of the ganglion the wider reflexes can no longer
be obtained but local responses continue, suggesting the presence of
nerve-cells in the body wall. Electrical stimulation also provides evi-
dence of this. One shock may produce only a small response but if a
second shock follows shortly afterwards there is marked facilitation
and a large contraction occurs. These responses are also seen after
removal of the ganglion. The various parts of the body are not all
equally sensitive to light, the highest sensitivity being in the region
of the ganglion. The 'ocelli' are cup-like collections of orange-
pigmented cells around the siphons ; according to Hecht they are not
photoreceptors.

The neuromuscular system thus appears to function mainly as a
reflex apparatus for producing protective movements in response to
certain stimuli. This is the role that might be expected of it in an
animal that remains fixed in one place. The 'initiative' for food-
gathering activities comes from the continuous action of the cilia of

in. 4 MOVEMENTS OF ASCIDIANS 65

the pharynx. The nervous system shows little sign of those continuous
activities that produce the varied and 'spontaneous' acts of behaviour
in higher forms. Nevertheless, it would be unwise to suppose that the
nerves are only activated by external stimuli. There are some sugges-
tions that even in these simple animals rhythmical activities are
initiated from within. The food-collecting operations of the pharyn-
geal wall involve rhythmical movement of the papillae by their
muscles. Further, in many species of ascidians there are regular
contractions of the siphons and body musculature in rotation, with

Fig. 35. Rhythmical 'spontaneous' contractions of Styela shown by attaching levers

to the two siphons. Branchial siphon above, atrial siphon below. The time-marker

shows intervals of 5 minutes. (From Yamaguchi.)

a frequency of 8-27 per hour (Fig. 35). These contractions are
especially marked when the animal is in filtered water and they may
be some form of 'hunger' contraction, directed towards the obtaining
of food. More water is moved by these contractions than by the ciliary
current. Their presence is a striking warning of the dangers of assum-
ing that even the simplest nervous system operates only when
stimulated from outside.

The neural gland is a sac lying beneath the ganglion and opening
by a ciliated funnel on the roof of the pharynx. It arises mainly from
the ectoderm of the larval nervous system, in part from the pharynx.
This double embryological origin, and its position, suggest that the
neural gland may be compared with the infundibulum and hypophysis
of vertebrates. There is an obvious similarity with Hatschek's pit of
amphioxus. Both seem to be receptor organs, testing the water stream
and also producing mucus. The subneural gland has also been held
to have a similarity to the pituitary in that it controls the release of
gametes. When eggs or sperms of the same species are present in the
water, signals from the neural gland apparently produce discharge
from the gonad. The pathway of the signals is said to be partly hor-
monal, partly nervous. Discharge is produced by injection of extract of

66

ORIGIN OF CHORDATES

in. 4-5

neural gland or of mammalian gonadotropin, but these act through the
ganglion, since they produce no effect if the nerves leading from this
(and the dorsal strand) are cut.

Further similarities with the pituitary have been claimed, such as
the presence of vasopressor and oxytocic substances in the subneural
gland. However, oxytocin is present elsewhere in the tunicate and in

Fig. 36. Longitudinal section of the ganglion (g.) and subneural gland
(s.n.g.) of an ascidian.

cil. ciliated funnel; d.s. dorsal strand; n.a. and n.p. anterior and posterior nerves;
ph. wall of pharynx. (After L. Bertin from Grasse.)

any case differs from that of vertebrates. It cannot be claimed that the
relationship with the pituitary is clear, but it seems likely that there
is some. As in the thyroid, a pharyngeal mucus-secreting organ
stimulated by the environment has evolved into a glycoprotein-
secreting endocrine organ, controlled by substances reaching it in the
blood. (Barrington, 1959, in Gorbman, Symposium on Comparative
Endocrinology.)

5. Development of ascidians

Tunicates are hermaphrodite, the ovary and testis being sacs lying
close to the intestine and opening by ducts near the atriopore. Fer-
tilization is external in the solitary forms but internal in those that
form colonies, the development in the latter taking place within the
parent. The details of cleavage and gastrulation show a remarkable
general similarity to those of amphioxus. Indeed, the whole develop-
ment is so strikingly like that of chordates that it establishes the
affinities of the tunicates far more clearly than the vague indications

(6?)

end:

Fig. 37. Ascidian tadpole of Clavelina.

air. atriopore; c. mantle; cer.v. cerebral vesicle; e. eye-spot; end. endostyle; ep. epicardium;
h. heart; m. mouth; mu. muscle-cells; n.c. nerve-cord; not. notochord; ot. otocyst; St. stomach;

sub.n. subneural gland.

r. C. b.

Fig. 37 A. T.S. ocellus of the free swimming tadpole stage of the sea squirt Ascidia nigra.
(Drawing from an electron micrograph.)

The ocellus is situated in the posterior wall of the cerebral vesicle. It consists of three parts, a lens
cell, a pigment cell, and a retina. The lens cell usually contains three lens vesicles, which are spheres
of cytoplasm bounded by mitochondria. The pigment cell contains granules of melanin, which
protect the photoreceptor from stray light. The retinal cells have processes that penetrate the
pigment cell. They are similar to vertebral rods, composed of a pile of membranes, closely applied

to the uaner edge of the lens cell.
a. p. attachment plaque, a membrane specialization thought to function as an anchor of the retinal
cell process to the pigment cell membrane ; b.m. basement membrane, the outer limit of the cerebral
vesicle; c.v. cavity of the cerebral vesicle; I.e. lens cell; l.v. lens vesicle; m. mitochondrion; p.c.
pigment cell; p.g. pigment granule; p.m. piled menbrane of photoreceptor part of the retinal cell;
r.c.b. retinal cell body; r.c.n. retinal cell nucleus; r.c.p. retinal cell process.
(From a preparation by N. Dilly.)

68 ORIGIN OF CHORDATES in. 5-

of a chordate plan of organization seen in the adult. The result of
development is to produce a fish-like creature, the ascidian tadpole,
which is immediately recognizable as a chordate (Fig. 37). The
cleavage is total and produces a blastula with few cells, whose future

mus.c.

atr.

mus. c.

Fig. 38. The ascidian tadpole (Ascidia or Ciona type). 1. Tadpole ready to hatch.
2. Tadpole. 3. Sensory vesicle. 4. Cross section of tail.

atr. atrium; end. endostyle; fol. follicle cells; mus.c. muscle cells; mus.f. muscle fibrils;

n.c. nerve-cord; n.m. nerve to tail muscles; not. notochord; oc. ocellus; ot. otolith; su. sticking

gland; ves. sensory vesicle. (After Berrill.)

potentialities are already determined. Gastrulation by invagination
follows and the creature then proceeds to elongate into the fish-like
larva. This possesses an oval 'head' and long tail, the latter supported
by a notochord formed by cells derived from the archenteric wall.
Forty of these cells make up the entire rod, becoming vacuolated and
elongated by swelling.

On either side of this notochord run three rows of muscle-cells,
eighteen on each side, derived from mesoderm that arises from yellow-
pigmented material already visible in the egg and later forming part
of the wall of the archenteron. Other cells of this tissue migrate
ventrally to make the pericardium, heart, and mesenchyme. The

in. 6 DEVELOPMENT OF ASCIDIANS 69

muscle-cells contain cross-striated myo-fibrils at the periphery, these
being continuous from cell to cell.

The nervous system is formed by folds essentially similar to those
of vertebrates, making a hollow, dorsal tube, extending into the tail
and enlarged in front into a cerebral vesicle, within which is an
ocellus and also a unicellular otolith (Fig. 37 a). Nerve-fibres proceed
only to the front end of the rows of muscles and the rest of the cord
contains no nerve-cells or fibres (Fig. 38).

The larva takes no food and the gut is not well developed. There is
a pharynx with usually a single pair of gill-slits opening into an
atrium, which develops as an ectodermal inpushing. Below or around
the mouth various forms of sucker are formed.

The whole process of development occupies only one or two days,
and the larva, in the species in which it is set free, is positively photo-
tropic and negatively geotropic and so proceeds to the sea surface.
But its life here is also limited. Within a day or two, depending on the
conditions, its tropisms reverse so that it passes to the bottom, turns
to any dark place and thus finds a suitable surface. It attaches by the
suckers, loses its tail, develops a large pharynx, and grows into an adult
ascidian. Presumably its short life in the chordate stage is sufficient
to ensure distribution, and the simple nervous system serves to find
a place in which to live.

In addition to the sexual reproduction, tunicates have great powers
of regeneration and also often multiply by budding. The bud consists
of an outer epicardial, mesenchymal, pharyngeal or atrial tissue. The
epidermis develops only more tissue like itself and all the other tissues
are formed from the inner mass. This occurs by a process of folding
to make a central cavity; the nervous system, intestine, and peri-
cardium are then formed by further foldings. The bud thus begins in
a condition comparable to a gastrula but develops directly into an
adult, without passing through the tadpole stages. The fact that a
complete new animal is thus formed from one or two layers shows
that the separation into three layers during development does not
involve any fundamental loss of potentialities, as would be required
if the 'germ layer' theory held rigorously. The germinal tissue of the
bud is not necessarily derived from that of the parent.

6. Various forms of tunicate

Besides some 2,000 species of sessile tunicates, about 100 species
have become secondarily modified for a pelagic life. These pelagic

70 ORIGIN OF CHORDATES m. 6-

animals are perhaps all related, but the whole subphylum is con-
veniently subdivided into three classes.
Class i. Ascidiacea.

Typical bottom-living forms such as Ciona (solitary), Botryllus
(colonial).
Class 2. Thaliacea.

Pelagic forms, simple or colonial, swimming by means of circular
muscle bands. Salpa, Doliolum, Pyrosoma.
Class 3. Larvacea.

Pelagic tunicata without metamorphosis; the adult has a tail and
resembles the tadpole of the other groups. Oikopleura.

7. Class Ascidiacea

The typical sessile ascidians are found in all seas. They may be
divided into those that live as single individuals (Ascidiae simplices)
and those forming colonies (Ascidiae compositae). Both types include
many different forms, however, and the division is not along phylo-
genetic lines. The colonial forms produced by budding may consist
simply of a number of neighbouring individuals {Clavelina) or of a
common gelatinous test in which the individuals are embedded
{Botryllus, Amaroucium). The form of the body is related to the type
of bottom upon which they are found; there has thus been an adap-
tive radiation within the group; a great variety of habitats is avail-
able for bottom living creatures, and the animals become adapted
accordingly.

Most of the species live in the littoral zone, but a few deep-sea forms
are known, such as Hypobythius calycodes, found below 5,000 metres.

Many ascidians probably live only for a short time, becoming
mature in their first year and dying thereafter. In some species the
animals live over a second winter, during which they become reduced
in size, growing and budding again in the following spring (Clavelina).

8. Class Thaliacea

These are pelagic tunicates living in warm water. They have
circular bands of muscle, enabling the animal to shoot through the
water by jet propulsion. In Doliolum and its allies the muscle-bands
pass right round the body (Cyclomyaria), whereas in Salpa the rings
are incomplete (Hemimyaria). The mouth and atriopore are at
opposite ends of the body. The tunic is thin and, like the rest of the
body, transparent.

The life-history of these forms involves a remarkable alternation of

PELAGIC TUN1CATES

7i

Fig. 39. Doliolum, gonozooid.

I, inhalent aperture; 2, ciliated pit; 3, ganglion and nerves; 4, pharynx; 5, mantle;

6, sense-cells, 7, exhalant aperture; 8, ovary; 9, intestine; 10, heart; 11, endostyle;

12, testis; 13, ciliated groove. (After Neumann.)

atr

muse

ats

an

br.s

Fig. 40. Cyclosalpa affinis, oozooid with chain of five wheels of blastozooids.
an. anus; atr. atrium; at.s. atrial siphon; hi. blastozooid with egg; br.s. branchial siphon; en. endo-
style; gn. ganglion; gr. gill ridge; ht. heart; muse, muscle ring; ph. pharynx; s. stomach.
( X £ modified. After Ritter and Johnson and Berrill.)

generations. In Doliolum the ascidian tadpole develops into a mother
or nurse zooid (oozooid). This by budding gives rise to a string of
daughter zooids, which it propels along by its muscles. The daughter
zooids are of three types: (i) sterile, nutritive, and respiratory indi-
viduals, the trophozooids, permanently sessile on the parent; (2)
sterile nurse forms, which are eventually set free (phorozooids); (3)
sexual forms (gonozooids, Fig. 39), nursed and carried by the phoro-
zooids until sexually mature, when they also break loose.

In Salpa the sexual form (blastozooid), produces only a single egg,

72 ORIGIN OF CHORDATES in. 8-

which develops within the mother without passing through a tadpole
stage, nourished by a diffusion placenta, whose cells also migrate into
the developing embryo. This becomes the asexual oozoid and pro-
duces a long chain of blastozooids, which it tows about until these
break away by sections (Fig. 40).

The pelagic colonial Pyrosoma of warm seas consists of a number of
individuals associated to form an elongated barrel-shaped colony.
The mouths open outwards and the atria inwards into a single
cavity with a terminal outlet from which a continuous jet emerges.

The mode of budding from the epicardium
and other features suggest an affinity with
Doliolum and Salpa, but Pyrosoma also
resembles the ascidians in that its zooids
are all sexual and capable of budding. The
yolky eggs develop within the parent,
without forming a larva. The outstanding
characteristic of the creatures is the
powerful light that they shine. This is
Fig. 41. Photogenic cell of produced in photogenic organs on each

Pyrosoma. (After Kukenthal.) r . r mi 1

side or the pharynx. 1 he photogenic cells
contain curved inclusions about 2^ in diameter (Fig. 41). These
are considered by some to be symbiotic luminescent bacteria, but
this is doubtful. The light is so powerful that when large masses of
Pyrosoma occur together the whole sea is illuminated sufficiently to
allow of reading a book. A remarkable feature of the phenomenon
is that the light is not produced continuously but only when the animal
is stimulated, as by the waves of a rough sea. If one individual is
stimulated others throughout the colony may show their lights, but
the mechanism of this effect is not known and the groups of cells that
form the luminescent organs receive no nerves. Other types of animal
with luminescent bacteria emit light continuously. The sudden flashes
of light probably serve as a dymantic reaction (p. 302), giving protec-
tion against enemies by producing a flight-reaction in the same way
as do sudden manifestations of colour or black spots by other animals.
It has been observed in the laboratory that colonies of Pyrosoma that
are dying and do not light up may be eaten by fishes, whereas any
that light up when seized may then be dropped.

9. Class Larvacea

The (Appendicularia) Larvacea (Figs. 42 and 43) are minute neo-
tenous tunicates that live in the plankton. Instead of the test, each

ill. 9

LARVACEA

73

Fig. 42. Oikopleura, one of the Larvacea, in its house, showing

the feeding-currents.

e. exhalant aperture; e.e. 'emergency exit'; f. p. filter pipes; f.tv. filter

window; g. gill-slit; m. mouth; r. trough; ta. tail. (After Garstang;

this and Figs. 44 and 45 by permission of the Editors of the Quarterly

Journal of Microscopical Science.)

individual builds a 'house', by secretion from a special part of the skin,
the 'oikoplastic epithelium'. The tail is a broad structure held at an
angle to the rest of the body; its
movement produces a current in
which the food is carried and caught
by a most elaborate filter arrangement
ixi the house (Fig. 42). Water enters
the house by a pair of posterior
'filtering windows' and is passed
through a system of filter pipes in the
part of the house in front of the
mouth. The very minute flagellates
of the nanonplankton are stopped by
these pipes and sucked back to the
mouth. The pharynx has two gill-slits,
also an endostyle and peripharyngeal
bands. The general organization is that
of a typical ascidian tadpole, and there
can be no doubt that these forms have
arisen from tunicates by the accelera-
tion of the rate of development of the
alimentary organs and gonads so that
the metamorphosis and normal adult
stage are eliminated. This may, of
course, have happened long ago, so
that the modern Larvacea are not FlG - «■ Appendicular* t seen from

the side and from below.

closely related to any living forms, (After Lehmann.)

74 ORIGIN OF CHORDATES in. 9-

but the fact that they differ in many ways from known ascidian
tadpoles does not invalidate the hypothesis; it would be expected
that many special features would be developed during evolution
after the paedomorphosis. Garstang, however, believed that there
is sufficient evidence to show that the Larvacea are related to the
Doliolidae and suggested an ingenious hypothesis by which the appendi-
cularian home could be derived from the doliolid test, the animal
itself remaining attached at the front end by gelatinous threads, which
came to make the filter tubes (Fig. 44).

Fig. 44. Sequence of stages by which the Larvacea may have
been evolved from a doliolid type.

A, Thaliaccan type of individual in its test (t). b, Paedomorphosis has
occurred so that a tailed creature is found in the test; g. gill-slit, c, The
tadpole has moved away from the inhalant aperture, leaving a series
of threads that become the filter pipes (/./>.), the inhalant aperture
becoming exhalant and vice versa. (After Garstang.)

The tail is a highly developed organ, serving for locomotion, nutri-
tion, and in the building of the house. It has a wide, continuous fin
and is supported by a notochord of 20 cells. Bands of 10 large striped
muscle-cells extend down each side, giving an appearance that has
been compared with metameric segmentation. The small number of
the cells makes any such comparison very difficult. Moreover, the
muscles are not developed from anything resembling myotomes. The
nerve-cord is a hollow tube with ganglionic thickenings, each con-
taining one to four nerve-cells. From these cells fibres proceed to
the muscles and to the skin in a series of roots that usually remain
separate, the motor being more dorsal.

10. The formation of the chordates

We can now recapitulate the points that we have established about
the origin of the chordates and attempt to piece together the evidence
to show the sequence of events that led to the production of a free-

in. 10 LARVAL ANCESTRY OF CHORDATES 75

swimming, fish-like animal. The chordates are related to the echino-
derms and their allies. This is established by the similarities in early
development (cleavage, gastrulation, mesoderm formation); by the
presence in early members of both groups of three separate coelomic
cavities, some with pores; by the similarity of the larva of entero-
pneusts to the dipleurula, and by other points of general morphological
and biochemical similarity between early chordates and echinoderms,
especially the arrangement of the nervous system and presence of a
mesodermal skeleton.

The echinoderms we have to consider are not the modern star-
fishes and sea-urchins, which are relatively active animals, but their
sessile Palaeozoic ancestors. These were sedentary, often stalked
animals, the cystoids, blastoids, and crinoids, feeding by ciliary
action. Surviving animals of related phyla, such as Polyzoa Ecto-
procta and Phoronis suggest that the ancestor for which we are looking
may have possessed a ciliated lophophore for food-collecting. For
purposes of dispersal its life-history presumably included a larval
stage with a longitudinal ciliated band, similar in plan to that of the
auricularia.

One might well ask how such an animal could possibly become
converted into a motile, metameric fish, feeding with its pharynx.
Yet the evidence of the lower chordates is sufficient to establish that
this change has occurred, and even provides us with an outline of the
main stages in the process of the change. Cephalodiscas, which is in
some ways the most primitive of surviving chordates, with its lopho-
phore also possesses gill-slits. This suggests that the pharyngeal
mechanism was substituted for the lophophore as a means of feeding
in the adult stage. There are other possible interpretations. It has
been suggested that Cephalodiscus was derived from a larval entero-
pneust (Burden-Jones). However, it is possible that ciliary mechanisms
developed in the pharynx first to deal with food collected outside by
tentacles or proboscis. Later the pharynx became developed into a
self-contained feeding mechanism, making unnecessary the tentacles,
which provide a tempting morsel for predators. The adoral band of
cilia of the auricularia probably serves to carry food into the mouth,
and for this purpose it is actually turned in to the floor of the pharynx.
Garstang suggests that the endostyle has been derived from this loop
of the adoral band.

The pharyngeal method of food-collecting thus replaced the ten-
tacles in the adult and the whole apparatus of an endostyle and an
atrium to protect the gills became developed. We may notice here the

7 6

ORIGIN OF CHORDATES

III. 10

remarkable similarity of this arrangement of the pharynx in tunicates,
amphioxus, and cyclostome larvae, and the partial similarity in
Balanoglossus.

The tunicates show us a stage in which branchial feeding has fully
replaced tentacle feeding in a sessile adult. But they have a larva that
is beyond all question a fish-like chordate. If the adult tunicate has
evolved from a modified lophophore-feeding creature, how has the

Fig. 45. To show the method by which a protochordate animal might have been

derived from an echinoderm larva such as the auricularia.

a. Auricularia in side view; b. protochordate in side view; c. same, dorsal view. ad.b. adoral

band; an. anus; coel. coelom; end. endostyle; g. gill-slit; Lb. longitudinal ciliated band;

m. mouth; n.c. nerve-cord; not. notochord. (After Garstang.)

ascidian tadpole arisen from the auricularia larva ? Garstang's auri-
cularia theory, first propounded in 1894, provides a possible answer.
As a ciliated larva grows its means of locomotion becomes inadequate
because the ciliated surface increases only as the square of the linear
dimensions, the weight as the cube. Muscular locomotion is not sub-
ject to this difficulty, and some of the starfish larvae actually show
flapping of the elongated processes, movements that presumably
assist them to remain at the surface. Garstang suggests that the fish-
like form arose by development of muscles along the sides of the
elongated body, the ciliated bands being pushed upwards and even-
tually rolled up with their underlying sheets of nerve plexus to form
the neural tube. The adoral ciliated band might then well be the
endostyle (Fig. 45).

This theory may seem at first sight fantastic. It is necessarily
speculative, but it has certain strong marks of inherent probability. It

in. 10 LARVAL ANCESTRY OF CHORDATES 77

violates no established morphological principles and certainly enables
us to see how a ciliated auricularia-like larva could be converted by
progressive stages into a fish-like creature with muscular locomotion,
while the adults, at first sedentary, substituted gill-slits and endostyle
for the original lophophore. The alternative is to suppose that the
ascidian tadpole arose as a purely tunicate development, providing
sufficient receptor and muscular organs to allow for the finding of
suitable sites on the bottom (Berrill, 1955).

We may plausibly regard the adult tunicate organization as directly
derived from that of sessile lophophore-feeding creatures, and the
larval organization as descended from an echinoderm-like larva.
There is no need, on this view, to regard the sessile adult tunicate
as a 'degenerate' chordate. The problem that remains is in fact not
'How have sea-squirts been formed from vertebrates ?' but 'How have
vertebrates eliminated the sea-squirt stage from their life-history?'
It is wholly reasonable to consider that this has been accomplished by
paedomorphosis. Advance of the time of development of the gonads
relative to that of the soma is well known to occur in certain special
cases such as the axolotl. The example of the Appendicularia shows
that a similar process can happen among tunicates! Various workers
have stressed the differences between the ascidian tadpole and the
adult appendicularian, in attempts to show that the two are not
comparable. But the differences, though considerable, are superficial :
the similarity of organization is profound. Any sensible biologist with
an understanding of the way in which the characteristic forms of
animals arise by change in the rate and degree of development of
features can see how the Appendicularia may represent modified
ascidian larvae.

The appendicularians do, indeed, carry certain characters of the
'adult' sea-squirt, in particular they have gill-slits, though of simple
form. Nothing is more likely, however, than that some features of the
sessile adult would be adumbrated in its larva and capable of fuller
development therein if advantageous. Larva and adult, it must be
remembered, possess the same genotype; the remarkable feature in all
animals with metamorphosis is the difference between the two stages,
not the similarity. Any characteristic may appear at either larval or
adult stage or be transferred by evolutionary selection from one to the
other. There is no serious objection to the view that the early adult
free-swimming chordates arose by paedomorphosis of some tunicate-
like metamorphosing form. If the creatures abandoned the habit
of fixation it would be possible for characters previously present

78 ORIGIN OF CHORDATES in. 10

separately in larva and adult to become combined in a single stage.
This is indeed what has happened in the Appendicularia.

Strangely enough, one of the chief difficulties of this theory is to
find the position of the enteropneusts. Since the larva is still in the
ciliated-band stage there should be no sign of organs characteristic of
the muscle-swimming, fish-like pro-chordate. Yet such signs are
present in the adult Balanoglossiis; there is a hollow nerve-cord and
some sign of a notochord. These features almost compel us to suppose
that the group has at one time possessed a free-swimming, fish-like
stage. The only escape from this conclusion would be by supposing
the hollow nerve-tube to be a case of convergence, for which a parallel
might be cited in the hollow nervous system of Polyzoa. But there is
no clear reason why the nerve-cord should become rolled up in the
collar, and it is easier to suppose it a vestige. This imposes two further
hypotheses on us. First that a fish-like stage once followed an advanced
ciliated-band stage in ontogeny, and secondly that this fish-like stage
later became adapted to a burrowing life, in fact that Balanoglossus is
a 'degenerate' chordate. Neither of these propositions is impossible,
but it must be admitted that the position of the enteropneusts is not
clear. Showing a combination of ciliated larva and chordate characters
they provide a valuable proof of the affinity of chordates and echino-
derm-like creatures, but these very chordate characters become an
embarrassment when we try to explain in detail how they have arisen!

There is strong reason to suppose that what we may call the Bate-
son-Garstang theory of the origin of chordates is correct. There is little
doubt that chordates are related to the sessile lophophore-feeding
type of creature rather than to any annulate, and we can reconstruct
the course of events by which the lophophore-feeder may have come
to have a pharynx with gill-slits and its larva to have muscles, a noto-
chord, and a nerve-tube. Then by paedomorphosis the sessile stage
disappeared and the free chordates began their course of evolution.
There are some reasons for supposing that a type such as amphioxus
could have been derived from a creature not distantly related to the
simpler Appendicularia and this in turn from a neotenous doliolid or
some similar ancestral type.

We need not, however, follow the theory into its details, which are
speculative. The whole treatment provides a conspicuous example
of close morphological reasoning, allied with proper consideration of
general biological principles, and establishes with some probability
the main outlines of the origin of our great phylum of active creatures
from such humble sedentary beginnings.

80 ORIGIN OF CHORDATES in. 10

Can we see in the production of the first fish-like creatures clear
signs of an 'advance' in evolution ? In acquiring the power of active
muscular locomotion the animals became able to live and feed in a
variety of habitats, either at the sea surface or on the bottom. Forms
with a sedentary adult stage are limited by the necessity for the
presence of a sea bottom of suitable character. The larvae were evolved
to provide the information to make sure of reaching such conditions.
But whereas suitable situations on the bottom are not common, and
are liable to change, the sea surface provides a generalized habitat in
which there is always abundant food, though no doubt also strenuous
competition for it. Paedomorphosis in this case, as in others, allows
the race to eliminate from its life-history the stage passed in a 'special'
environment, which is difficult to find. Although the fish-form that
was thus produced proved to have great possibilities for further
evolution, the change was not at first a strikingly progressive one.
The surface of the sea is perhaps the most general of all environments ;
possibly it was the seat of the origin of life. Races that have devised
means of living on the sea bottom may therefore be said to have
advanced, because they have invaded a more difficult habitat. To
abandon the sedentary life might in this sense be regarded as a retro-
grade step. The peculiar feature of the early fishes, however, was that
they developed powers of active movement in a relatively large
organism provided with efficient receptors, and by making use of the
feeding mechanism developed at first by the bottom-living adult were
able to live successfully at the sea surface. They acquired their
dominance at this stage not by invading new habitats but by develop-
ing effective means of living in the richly populated plankton.

IV

THE VERTEBRATES WITHOUT JAWS. LAMPREYS

1. Classification

Phylum Chordata

Subphylum 4. Vertebrata (= Craniata)
Superclass 1. Agnatha
Class 1. Cyclostomata
Order 1. Petromyzontia

Petromyzon; Lampetra; Entosphenus; Geotria; Mordacia
Order 2. Myxinoidea
Myxine; Bdellostoma

Class 2. *Osteostraci. Silurian-Devonian
*Cephalaspis; * Tremataspis

Class 3. *Anaspida. Silurian-Devonian
*Birkenia; *Ja?noytius

Class 4. *Heterostraci. Ordovician-Devonian
*Astraspis; *Pteraspis; *Drepanaspis

Class 5. *Coelolepida. Silurian-Devonian.
*Thelodus; *Lanarkia

Superclass 2. Gnathostomata

2. General features of vertebrates

All the remaining chordates are alike in possessing some form of
cranium and some trace of vertebrae; they make up the great sub-
phylum Vertebrata, also called Craniata. The organization of a verte-
brate is similar to that of amphioxus, but with the addition of certain
special features. A few of these novelties may now be surveyed, with
emphasis on those that provide the basis for the capacity to live in
difficult environments that is so characteristic of the vertebrates.
Firstly the front end of the nervous system is differentiated into an
elaborate brain, associated with special receptors, the nose, eye,
and ear. Through these receptors the vertebrates are able to respond
to more varied aspects of the environment than are any other animals.
Some of them have the ability to discriminate between visual shapes
and colours, and in the auditory field between patterns of tones, also
between a host of chemical substances. The motor organization
allows the performance of delicate movements to suit the situations

82 VERTEBRATES WITHOUT JAWS iv. 2-

that the receptors reveal. The swimming process, by the passage
of waves down the body, is itself perfected by improvements in the
shape of the fish, allowing rapid movements and turns. Besides the
median fins there develop lateral paired ones, serving at first a
stabilizing and steering function and then converted, when the land
animals arose, into organs of locomotion on the ground or in the air
and finally, in the shape of the hands, into a means of altering the
environment to suit the individual.

The brain itself, at first mostly devoted to the details of sensory
and motor function, comes increasingly to preside, as it were, over
all the bodily functions, and to give to the vertebrates the 'drive' that
is one of their most characteristic features. The skull is developed as
a skeletal thickening around the brain, probably at first mainly for
protection, but later serving for the attachment of elaborate muscle
systems. The study of vertebrates is especially identified with study
of the skull, because in so many fossils this is the only organ preserved.

The food of the earliest vertebrates was collected by ciliary action,
but this habit has long been abandoned and only in rare cases today
does the food consist of minute organisms. The pharynx of most
vertebrates is small, there are relatively few gill-slits and these are
respiratory. In all except the most ancient forms the more anterior of
the arches between the gills became modified to form jaws, serving not
only to seize and hold the food but also to 'manipulate' the environ-
ment.

The blood system shows two of the most characteristic vertebrate
features, namely, the presence of a heart that has at least three
chambers and thus provides a rapid circulation, and of haemoglobin
within corpuscles, serving to carry large amounts of oxygen to the
tissues. The efficiency of this system must have been a major factor
in producing the dominance of the vertebrate animals. In the air-
breathing forms, and especially the warm-blooded birds and mam-
mals, the respiratory and circulatory systems allow the expenditure of
great amounts of energy per unit mass of animal, so that quite extra-
vagant devices can be used, allowing survival under conditions that
would otherwise not support life.

The excretory system is based on a plan quite different from that
of amphioxus. It consists of mesodermal funnels, leading primarily
from the coelom to the exterior. It may be that this type of kidney
arose in connexion with the abandoning of the sea for fresh water.
Probably all but the earliest vertebrates have passed through a fresh-
water stage, and it is significant that all except Myxine have less salt in

iv. 4 AGNATHA 83

their blood than there is in sea water. Elaborate devices for regulation
of osmotic pressure have been developed, and the mesodermal kidneys
play a large part in this regulation.

This outline only gives a few suggestive features of vertebrate
organization. The details differ bewilderingly in the different types
and it is our business now to survey them. In the earliest forms the
more special mechanisms are absent or at least function only crudely,
and passing through the vertebrate series we find more and more
devices adopted, along with more and more delicate co-ordination
between the various parts, culminating in the extremely highly
centralized control of almost every aspect of life that is exercised by
the mammalian cerebral cortex.

3. Agnatha

The earliest vertebrates, while showing most of the characteristic
features of the group, differ from the rest in the absence of jaws and
are therefore grouped together in a superclass Agnatha, distinguished
from the remaining vertebrates, which have jaws, and are therefore
called Gnathostomata. The only living agnathous animals are the
Cyclostomata, lampreys and hag-fishes, but the first vertebrates to
appear in the fossil series, mostly heavily armoured and hence known
as 'ostracoderms', found in Silurian and Devonian strata, also show
the agnathous condition, and have some other features in common
with the Cyclostomata. This group of agnathous vertebrates shows
some interesting experimentation in methods of feeding, before the
jaw-method became adopted. The modern cyclostomes are parasites
or scavengers, in the adult state, but as larvae the lampreys still feed
on microscopic material, using an endostyle resembling that of
amphioxus in many ways, but making use of muscular contraction
rather than ciliary action to produce a feeding current. The methods
of feeding of the Devonian forms are not known for certain, but
probably included shovelling detritus from the bottom.

The Cyclostomata are therefore worth special study as likely to
show us some of the characteristics possessed by the earliest vertebrate
populations.

4. Lampreys

The most familiar cyclostomes are the lampreys, of which there are
various sorts found in the temperate zones of both hemispheres. All
lampreys have a life-history that includes two distinct stages: the
ammocoete larva lives in fresh water, buried in the mud, and is

8 4 VERTEBRATES WITHOUT JAWS iv. 4-

microphagous: the adult lamprey has a sucking mouth, and usually
lives in the sea, where it feeds on other fishes. Lampetra the lamprey
(Fig. 47), is a typical example, common in Great Britain. The adult
is an eel-like animal about 30 cm long, black on the back, and white
below. The surface is smooth, with no scales. The skin is many-
layered (Fig. 48). The outermost cells have a striated cuticular border.
Mixed with these epithelial cells the lamprey, like most aquatic
vertebrates, has many gland-cells for producing slime. Below the
epidermis lies the dermis, a layer of bundles of collagen and elastin

Fig. 47. Brook lampreys, Lampetra planer i.

A, ripe female, with anal fin; B, ripe male; note shape of dorsal fin and presence of

copulatory papilla. (Curves due to fixation.)

fibres, running mostly in a circular direction. This tissue is sharply
marked off from a layer of subcutaneous tissue containing blood-
vessels and fat, as well as connective tissue. There are pigment cells
in the dermis and a thick layer of them at the boundary of dermis and
subcutaneous tissue. The chromatophores are star-shaped cells whose
pigment is able to migrate, making the animal dark or pale. This
change is especially marked in the larva and is produced by variation
in the amount of a pituitary secretion (p. 107).

The head of the lamprey bears a pair of eyes and a conspicuous
round sucker. On the dorsal side is a single nasal opening, and behind
this there is a gap in the pigment layers of the skin through which the
third or pineal eye can be seen as a yellow spot. There are seven pairs
of round gill openings, which, with the true eyes (and some miscount-
ing or perhaps inclusion of the nasal papilla), are responsible for the
familiar name 'nine eyes'. There is no trace of any paired fins, but the
tail bears a median fin, which is expanded in front as a dorsal fin.
There are sex differences in the shape of the dorsal fins of mature
individuals and the female has a considerable anal fin (Fig. 47).

iv. 5

SWIMMING OF LAMPREYS

85

The lamprey swims with an eel-like motion, using its myotomes in
the serial manner that has been mentioned in amphioxus and will be
discussed later (p. 133). The waves that pass down the body are of
short period relative to the length, so that the swimming is mechani-
cally inefficient; lampreys show great activity, but their progress is
not rapid. The animal often comes to rest, attaching itself with the
sucker to stones (hence the name, 'suck-stone') or to its prey. In this
position water cannot of course pass in through the mouth, but both

s. cut.

Fig. 48. Section of skin of lamprey.

c. club cells; der. dermis; ep. epidermis; gr. granular gland-cells; m. myotomal muscle;
pig. pigment cells; s.cut. subcutaneous connective tissue. (After Krause.)

enters and leaves by the gill openings. When swimming the backward
jet of water may assist in locomotion.

The trunk musculature consists of a series of myotomes separated
by myocommas. Each myotome has a W-shape, instead of the simple
V of amphioxus. The muscle-fibres run longitudinally and they are
striped, but of a somewhat peculiar fenestrated type.

5. Skeleton of lampreys

The skeleton of lampreys consists of the notochord and various
collections of cartilage. This latter is partly of the typical vertebrate
type, that is to say, consists of large cells in groups, separated by a
matrix of the protein chondrin, which they secrete. In other regions
a tissue containing more cells and less matrix is found, the so-called
fibro-cartilage, and this more nearly resembles fibrous connective
tissue and serves to emphasize that no sharp line can be drawn
between these tissues. There is also, in the larva, a tissue known as

86 VERTEBRATES WITHOUT JAWS iv. 5

muco-cartilage, which is an elastic material serving more as an
antagonist to the muscles than for their attachment.

The notochord remains well developed throughout life as a rod
below the nerve-cord. It consists of a mass of large vacuolated cells,

Fig. 49. Transverse section through notochord of lamprey.
c. cells; s. sheath. (After Krause.)

anc

Fig. 50. Lateral view of skeleton of head and branchial arches of Petromyzon.

ac. auditory capsule; adc. antero-dorsal cartilage; anc. annular cartilage; ha. branchial
arch; bd. basidorsal; hac. hyoid cartilage; hbc. hypo-branchial rod; lit. horizontal bar; n.
notochord; ?ic. nasal capsule; oc. orbital cartilage; per. pericardial cartilage; pdc. postero-
dorsal cartilage; pic. posterolateral cartilage; st. styliform cartilage; /. tendon of tongue;
II-X, cranial nerves. (After Parker.)

enclosed in a thick fibrous sheath (Fig. 49). The rigidity of the whole
rod depends on the turgor of the cells and it often collapses com-
pletely in fixed and dehydrated material (Fig. 59). No doubt in life it
serves, like the notochord of amphioxus, to prevent shortening of the
body when the myotomes contract.

The notochordal sheath is continuous with a layer of connective
tissue, which also surrounds the spinal cord and joins the myocom-

iv. s SKELETON OF LAMPREYS 87

mas and thus eventually the subcutaneous connective tissue. Within
this connective tissue there develop certain irregular cartilaginous
thickenings that are of special interest because they may be compared
with vertebrae, perhaps with the basi-dorsal element (p. 132). They
lie on either side of the spinal cord (Fig. 50), that is to say, above the
notochord, and consist either of one nodule on each side of the seg-
ment, through the middle of which the
ventral nerve-root emerges, or of two
separate nodules, with the nerve between
them. Rods of cartilage extend dorsally
and ventrally into the fins, but are not
attached to the 'vertebrae'.

The lamprey skull shows even in the
adult the basic arrangement found only
in the embryo of higher vertebrates. The
floor is formed of paired parachordals on
either side of the notochord and in front
of this paired trabeculae. Attached to this
base is a series of incomplete cartilagi-
nous boxes surrounding the brain and
organs of special sense (Fig. 51). To
this skull is attached the skeleton that
supports the sucker and gills. The
arrangement of the skull differs consi-
derably from that of later vertebrates.
The cranium has a floor around the end
of the notochord, and in front of this
there is a hole containing the pituitary
gland. The side walls are strong but the roof is composed only of
a tough membranous fibro-cartilage. The auditory capsules are
compact boxes surrounding the auditory organs at the sides. The
olfactory capsule, imperfectly paired, is also almost detached from
the cranium. Other ridges of cartilage lie below the eyes and there is
a complex support for the sucker.

The skeleton of the branchial region consists of a system of vertical
plates between the gill-slits, joined by horizontal bars above and
below them. This cartilage lies outside the muscles and nerves and is
therefore difficult to compare with the branchial skeleton of higher
fishes, which lies in the wall of the pharynx. The elastic action of the
cartilages produces the movement of inspiration. A backward exten-
sion of the branchial basket forms a box surrounding the heart.

Fig. 51. Dorsal view of skull of

Petromyzon.

Lettering as Fig. 50. /, olfactory

nerve; /. hole in roof of cranium.

(After Parker.)

88 VERTEBRATES WITHOUT JAWS iv. 6

6. Alimentary canal of lampreys

The sucker is bounded at the edges by a series of lips, which besides
being sensory serve also to make a tight attachment when the lamprey
sucks (Fig. 52). In the sucker are numerous teeth, whose arrangement
varies in the different types of lamprey. These teeth are horny epi-
dermal thickenings, supported by cartilaginous pads, and are there-
fore not comparable with the teeth of vertebrates, which are derived

Fig. 52. Sucker of Petromyzon show- Fig. 53. Section through tooth of lamprey

ing outer circular lip, teeth, and tongue, 1, horny cap; 2, stellate tissue; 3, cap to replace

with special teeth, at the centre. , ; 4> connective tissue; 5, epidermis of mouth;

(After Parker.) 6, cartilage; 7, proliferative layers of epidermis

that produce the horny cells.
(After Hansen, from Kukenthal.)

mainly from mesodermal tissues (Fig. 53). The sharper and larger
teeth are borne on a movable tongue, which is used as a rasp (Fig. 54).
An annular muscle runs round just above the lips of the sucker and
presumably serves to narrow the margin and hence to release the fish.
The remaining muscles are mostly attached to the tongue and base
of the sucker. The largest of these muscles, the m. cardioapicalis, is
attached posteriorly to the cartilage surrounding the heart and in front
is prolonged into a conspicuous lingual tendon, which is attached to
the tongue and serves to pull it backwards. Presumably the action of
this muscle deepens the oral cavity and is thus the main agent securing
attachment of the sucker. There is a collar of circular fibres around
the front end of the cardio-apical muscle, serving to lock the tendon
and maintain the suction. Dorsal and ventral to the main tendon are

iv. 6

FEEDING OF LAMPREYS

groups of muscles that rock the tongue up and down to produce a
rasping action. The muscles of the sucker are all derived from the
lateral plate and are innervated from the trigeminal nerve; their fibres
are striated.

The mouth is a small opening above the tongue and leads into a
large buccal cavity. At the hind end this divides into a dorsal passage,
the oesophagus, for the food, and a ventral respiratory tube, which
leads to the gill pouches but is closed behind. At the mouth of the

ann.

m. card.ap.

circ.m.

mus.2 mus.t

Fig. 54. Longitudinal section through head of lamprey.

ann. annular muscle of sucker; b. brain; circ.m. circular fibres of tongue-muscles; oes. oeso-
phagus; g. gill aperture; h.s. hypophysial sac; m.card.ap. cardio-apical muscle; mus. 1 and
2, muscles that rock the tongue; n. notochord; nas. nasal sac; nos. nostril; p. pineal; pit.
pituitary gland; t. tooth; tend, tendon of tongue, pulled back by m.card.ap.; to. tongue. (Partly

after Tretjakoff.)

respiratory tube is a series of velar tentacles, corresponding exactly in
position to those of amphioxus, and serving to separate the mouth
and oesophagus from the respiratory tube while the lamprey is feed-
ing. The seven branchial sacs are lined by a folded respiratory epi-
thelium and surrounded by muscles, and these, together with the
elastic cartilages and appropriate valves, ensure the pumping of the
water tidally, in and out of the external openings. In front of the first
sac is the remains of an eighth pouch, whose surface is not respiratory.
The 'salivary' glands are curious organs of which little is known.
They are a pair of pigmented sacs, embedded in the hypobranchial
muscles. Each has a folded wall, from which a duct proceeds forward
to open below the tongue. The salivary glands produce a secretion
that prevents coagulation of the blood of the fishes on which the lam-
prey feeds. The nature of this secretion is not known, but it rapidly
turns black on exposure to the air and the glands for this reason
appear to be pigmented. It has been observed that in lampreys taken
from fishes the intestine is filled with red corpuscles, and there is
therefore no doubt that they feed mainly on the blood of their prey.

90 VERTEBRATES WITHOUT JAWS iv. 6-

Little is known of the habits of lampreys in the sea, but in North
America there are races of lampreys that are land-locked and feed on
the fishes in the lakes, where they have recently become a most serious
pest (Fig. 55).

The oesophagus (fore-gut) leads directly into a straight intestine
(mid-gut); there is no true stomach in lampreys (Fig. 56). The surface
of the intestine is increased by a typhlosole, running a somewhat
spiral course. There is a liver, gall-bladder, and bile-duct of typical
vertebrate plan, but no separate pancreas. However, in the wall of the

Fig. 55. Lake lamprey attached to a bony fish, which also shows the scars of the
attacks of other lampreys. (After Gage.)

anterior part of the intestine there are large patches of cells that
resemble those of the acini of the pancreas of higher forms and con-
tain secretory granules. Barrington has shown that extracts of this
region have a high proteolytic power, the enzyme being of the tryptic
type, with its optimum between pH 7-5 and 7-8. Some of this tissue is
collected in the walls of short diverticula, reaching forwards. The
situation is therefore essentially similar to that found in amphioxus,
and we may regard these patches of zymogen cells, or the diverticula,
as the forerunners of the exocrine portions of the pancreas. In the
lampreys the endocrine portion, not yet identified in amphioxus, also
appears. Around the junction of the fore-gut and intestine are groups
of follicles that do not communicate with the lumen of the intestine.
These 'follicles of Langerhans' were, appropriately enough, first seen
by the discoverer of the islets in higher forms, and Barrington has
now shown that following destruction of this tissue by cautery there
is a rise in blood-sugar. Moreover, after injection of glucose, vacuola-
tion of the cells occurs. We may safely conclude that these cells are
involved in carbohydrate metabolism, but only one type of cell is
present.

iv. 7

(9i)

7. Blood system of lampreys

The blood vascular system is arranged on the same general plan
as in amphioxus but there is a well-developed heart. This lies behind
the gills and can be considered as a portion of the sub-intestinal vessel,
folded into an S-shape and divided into three chambers. The heart
is suspended in a special portion of the coelom, the pericardium,
whose walls are supported by cartilage. In the larva the heart first
appears as a straight tube and owing to an abnormality of development
it sometimes fails to develop its S-shape. Contractions can neverthe-

Fig. 56. Mid-gut of larval lamprey.

ai. anterior region of intestine; bd. bile-duct; ca. coeliac artery; gb. gall-bladder; hp. hepatic

portal vein; /. liver; oes. oesophagus; p. position of 'pancreas', containing islet tissue;

pi. posterior intestine; y. yellow area where wall of intestine contains zymogen cells.

(From Barrington.)

less be seen in these abnormal hearts, passing from behind forwards
along the straight tube. Similarly in the normal heart contraction
proceeds in the chambers from behind forwards. The most posterior
chamber is a thin-walled sinus venosus, into which the veins pour
blood. This leads to an auricle (atrium), also thin-walled, lying above
the sinus. The atrium passes blood into the ventricle below it, a
thick-walled chamber, providing the main force for sending the blood
round the body.

The heart receives nerve-fibres from the vagus nerve and contains
nerve-cells, some of which give a chromaffin reaction suggesting the
presence of adrenalin-like substances. Stimulation of the vagus nerve
produces acceleration of the heart-beat, followed by slowing. Acetyl
choline also accelerates the heart. In Myxine there are no nerves to
the heart or nerve-cells in it and acetyl choline has no effect. Both
hearts contain much adrenaline and similar substances but show little
change when adrenaline is added to a perfusate.

Blood leaves the ventricle by a large ventral aorta, running forwards

92 VERTEBRATES WITHOUT JAWS iv. 7-

between the gill pouches, to which it sends a series of eight afferent
branchial arteries. These break up into capillaries in the gills, and
efferent branchial arteries collect to a pair of dorsal aortae, running
backwards, which join and form the main dorsal aorta. This passes
down the trunk and carries blood to all the parts of the body by means
of series of segmental arteries and special vessels to the gut, gonads,
and excretory organs. A curious feature is that many of these arteries
are provided with valves at the point at which they leave the main

trunks (Fig. 57). It may be significant
that such valves are not found where
the efferent branchials join the dorsal
aorta, nor at the points of exit of the
renal arteries, so that perhaps the
valves serve to reduce the pressure in
the majority of the arteries, while
leaving it high in those to the kidneys.
The removal of large quantities of
water is an important problem in all
freshwater animals and is facilitated
by a high pressure in the kidneys.
This must be difficult to maintain in
an animal with a branchial circulation
and hence a double set of capillaries.
The venous system consists of a
network of sinuses, with contractile
venous hearts in various places.
There is a large caudal vein, dividing
where it enters the abdomen into two posterior cardinals. These
run forward in the dorsal wall of the coelom, collecting blood from
the kidneys, gonads, &c, and opening into the heart by a single
ductus Cuvieri on the right-hand side, this being the remains of a pair
found in the larva. Anterior cardinals collect blood from the front
part of the body, and there is also a conspicuous ventral jugular vein
draining venous blood from the muscles of the sucker and gill pouches.
Besides the veins proper there is a large system of venous sinuses,
especially in the head. Blood from the gut passes by a hepatic portal
vein through a contractile portal heart to the liver, from which hepatic
veins proceed to the heart.

The blood of lampreys, like that of all vertebrates, contains the
respiratory pigment haemoglobin, enclosed in corpuscles, here nucle-
ated. This arrangement immensely increases the oxygen-carrying

Fig. 57. Valves at the origin of
segmental arteries of a lamprey.

1, notochord; 2, segmental artery; 3,

aorta.

(From Kukenthal, after Keibal.)

iv. 8 EXCRETION IN LAMPREYS 93

power of the blood. Haemopoietic tissue occurs in the intestinal wall
of the larva and this has been regarded by some as representing the
spleen. In the adult the blood-forming tissue lies below the spinal
cord and in the kidney. White corpuscles resembling lymphocytes
and polymorphonuclear cells occur, produced by lymphoid tissue in
the kidneys and elsewhere. However, there is no distinct system of
lymphatic channels.

8. Urinogenital system of lampreys

The excretory and genital systems of vertebrates consist of a series
of tubes opening from the coelom to the exterior and serving to carry
away both excretory and genital products. This plan of organization is

Fie. 58. Diagram to show arrangement of the pronephros in a freshly hatched

lamprey.
g. gonad; pr. pronephros; prd. pronephric duct. (After Wheeler.)

quite different from that found in amphioxus and represents a new
acquisition by the vertebrates. It is not clear whether the excretory
or genital component of the complex is the primary one, nor indeed
why they are associated. The gonads develop from the walls of the
coelom in all animals possessing that cavity; some hold that the coelom
represents an enlargement of a sac that at first served purely as a
gonad. Genital ducts leading from the coelom to the exterior are
common in invertebrates, and we may guess that at their first appear-
ance the urinogenital tubules of vertebrates served only for genital
products.

The conversion of these tubules to excretory purposes may have
been a result of the adoption of the freshwater habit. The blood of
lampreys, when in fresh water, contains a higher concentration of salts
than the surrounding water. Little is known about the condition in
sea lampreys, where blood is probably hypotonic to the sea. When in
the river the animals must deal with the tendency for water to flow in.
This water must be removed without losing salt ; accordingly in most
freshwater animals, including vertebrates, we find some system by
which the separation can be achieved.

94 VERTEBRATES WITHOUT JAWS iv. 8

The region that gives rise to the kidney during development lies
between the dorsal scleromyotome and the more ventral lateral plate

card v

glom

on. t

oes.

pron. F

Fig. 59. Section through newly hatched larvae of Lampetia hehind the pharynx.

ao. aorta; card.v. cardinal vein; glom. glomerulus; fit. heart; my. myotome; n.c. nerve-cord;
not. notochord, which has collapsed because of lack of turbidity after fixation; oes. oeso-
phagus; pron.f. ciliated funnel of pronephros; pron.t. twisted pronephric tubule; sp. space

around nerve-cord.

Fig. 60. Kidney system of a 22-millimetre larva of Lampetra.

tnes. mesonephric tubules; mesg!. mesonephric glomeruli; pr. pronephric funnels;
prd. pronephric duct; prgl. pronephric glomeruli. (After Wheeler.)

mesoderm; it is known as the nephrotome. This tissue differentiates
during development from in front backwards, making a series of
segmental funnels, opening into a common archinephric duct (Fig.
58). The most anterior funnels open into the pericardium; usually
there are four of these in a freshly hatched larva, opening into a single

iv. 8

EXCRETION IN LAMPREYS

95

mid

duct, which reaches back to an aperture near the anus. Close to each
funnel there develops a tangle of blood-vessels, the glomerulus (Figs.
59 and 6o). Presumably the osmotic flow of water into the body is
relieved by the pressure of the heart-beat forcing water out from the
glomeruli into the coelomic fluid, whence it is removed by the funnels,
with the aid of their cilia. The tubules become longer and twisted
after hatching and may perhaps serve for salt-reabsorption.

These anterior funnels constitute
the pronephros. As the animal
grows they are replaced by a more
posteror set, the mesonephros.
There is, however, a gap of
several segments in which no
tubules appear (Fig. 6o), a strange
and unexplained discontinuity,
common to all vertebrates. The
pronephric tubules gradually dis-
appear and finally in the adult all
that remains of the organ is a mass
of lymphoid tissue. Meanwhile the
mesonephros develops as a much
larger fold, hanging into the coelom
and containing very extensive wind-
ing tubules. These do not open to
the coelom (at least in the adult) but
each to a small sac, the Malpighian
corpuscle, which contains a portion
of the coelom and the glomerulus.
This is obviously a more efficient method for allowing the heart to
pump excess water out of the blood and down the tubules. The latter
themselves have become greatly elongated and make up the main
bulk of the organ (Fig. 6i). The segmental arrangement is there-
fore much obscured and as extra glomeruli are added it disappears
completely. The mesonephros extends at its hind end as the animal
grows, until it forms the adult kidney, a continuous ridge of tissue
reaching back to the hind end of the coelom. Besides the excretory
apparatus the kidney also contains much lymphoid tissue and fat, and
it probably plays a part in the formation and destruction of red and
white corpuscles.

The gonads are unpaired ridges medial to the mesonephros. Pri-
mordial germ-cells, set aside very early in development, migrate into

Fig. 6i. Transverse section of kidney of
hampetra.

gl. glomerulus; mid. middle section of

tubule; pr. proximal region of tubule; term.

terminal region, opening into W.d. Wolffian

duct.

(After Krause.)

iv. 8-

96 VERTEBRATES WITHOUT JAWS

these ridges and develop into eggs or sperms. The differentiation of
the gonad occurs relatively late in lampreys, so that in young am-
mocoetes the organ is 'hermaphrodite', containing developing oocytes
and spermatocytes together. The ripe ovary consists of ova each sur-
rounded by single-layered follicular epithelium, which finally ruptures
and liberates the egg into the coelom, whence it escapes by pores to be
described presently. The testis consists of a number of follicles con-
taining sperms; it is unique among vertebrates in that the follicles

CL U 9 .

Fig. 62. Cloacal region of fully adult Lampetra.

C. coelom; CI. lips of cloaca; Ct. connective tissue; D. duct leading from coelom to the

mesonephric duct; Df. dorsal fin; M. muscle; Md. mesonephric ducts; N. notochord;

R. rectum; Ug. urinogenital papilla. (After Knowles.)

have no ducts; when ripe they rupture into the coelom, which becomes
filled with spermatozoa and these escape, like the ova, by pores.

These apertures by which the gametes escape are similar in the
two sexes and consist of short channels, one on each side, leading from
the coelom to the lower end of the kidney duct (Fig. 62). They nor-
mally become open only a few weeks before spawning, but Knowles
has shown that injections of oestrone or anterior pituitary extract will
cause perforations of the ducts in young lampreys, indeed even in the
ammocoete larve.

Fertilization is external, but there are modifications of the cloaca in
the two sexes to assist in ensuring fertilization and proper placing of
the eggs in the 'nest' (p. 113). The lips of the cloaca of the ripe male
are united to form a narrow penis-like tube. The cloacal lips of the

iv. 9 NERVOUS SYSTEM OF LAMPREYS 97

female are enlarged and often red; in addition she has an anal fin,
probably used, as in salmon and trout, to make a nest. These sex
differences, which develop shortly before spawning, can also be
initiated by injection of anterior pituitary extracts (p. 107).

9. Nervous system of lampreys

The nervous system of the cyclostomes is very much better
developed than that of amphioxus and shows the characteristic plan
that is present in all vertebrates. The essence of the vertebrate ner-
vous organization may be said to be that it consists of large amounts
of tissue and is highly centralized. The brains of vertebrates contain
much larger aggregates of nervous tissue than are to be found in any
other animals, and this tissue produces by its actions the most charac-
teristic features of vertebrate life. Vertebrates are active, exploratory
creatures, and their behaviour is much influenced by past experience.

We shall return later to detailed discussion of the organization of
the central nervous system; now we may look briefly at the plan
found in the lamprey, as an introduction to that of other vertebrates.
As compared with amphioxus there has been a very high degree of
cephalization. The front end of the spinal cord is enlarged into a
complicated brain, and the nerves connected with a number of the
more anterior segments have become modified to form special cranial
nerves.

The spinal nerves, however, still show the plan found in amphioxus
in that the dorsal and ventral roots do not join. In amphioxus the
ventral roots contain motor-fibres for the myotomes and some
proprioceptive fibres, while the dorsal roots contain sensory fibres and
motor-fibres for the lateral plate musculature (p. 36). The details of
the composition of the nerves of lampreys are still unknown, but there
are hints of considerable deviations from this plan. The ventral roots
contain many motor fibres passing to the myotomes. The dorsal roots
consist largely of sensory fibres with bipolar cell bodies collected into
dorsal root ganglia including proprioceptor fibres from the myotomes:
it is not known whether the dorsal roots also contain any efferent
fibres. In the young larva many of the afferent fibres are the processes
of cells lying in the spinal cord (Rohon-Beard cells), which are
typical of the early stage of many chordates. There are few types
of cells in the cord at this time, allowing for only the simplest
reflex arcs.

The autonomic nervous system shows some generalized and some
special features. The gut is mainly innervated by the vagus, which

98 VERTEBRATES WITHOUT JAWS iv. 9

extends far back along the intestine. There is little contribution of
fibres from the spinal nerves to the alimentary canal, since this has no
mesentery, being attaached only at its cranial and caudal ends. There
are, however, numerous fibres from the spinal nerves to the rectum,
ureters, and cloacal region, and numerous postganglionic neurons
are found here. Nerve-cells are also found in the intestinal plexuses.

The sympathetic system consists of isolated fibres running in both
dorsal and ventral roots. Many of these run directly to their endings,
for instance in the arteries, without interpolation of neurons. A few
postganglionic cells are present, however, but they are seldom collec-
ted into ganglia. The system is therefore even more scattered than
in elasmobranchs (p. 173). The 'adrenal' system is also diffuse. There
are scattered masses of interrenal (cortical) tissue and large groups
of suprarenal (medullary) cells, especially in the walls of the veins and
the heart. The suprarenal tissue receives 'preganglionic' fibres from
the spinal nerves. Its cells sometimes seem to be connected with each
other by fibres like those of neurons and they may operate a form of
control intermediate between nervous and hormonal (Johnels, 1956).

The nerve-fibres in the nervous system of cyclostomes are not pro-
vided with myelin sheaths ; in this they resemble the nerves of amphi-
oxus. Conduction is slow in such non-medullated fibres, the only case
actually investigated in cyclostomes being the lateral line nerve of
Bdellostoma, found by Carlson to conduct at the low rate of 5 metres
a second (frog about 50 m/sec, mammals up to 100 m/sec).

The spinal cord is of a uniform transparent grey colour and is
flattened dorso-ventrally, apparently to allow access of oxygen, and
metabolites, no blood-vessels being present within the cord. How-
ever, vessels are present in Myxine in which the cord is also flat. The
nerve-cell bodies lie, as in higher vertebrates, towards the centre, but
the synaptic contacts are not made in this 'grey' matter but at the
periphery, in what would correspond to the white matter of higher
forms. The outer part of the cord is thus made up of a neuropil or
nerve feltwork, formed of the terminations of the incoming sensory
fibres and the dendrites of the motor-cells. These cells (Fig. 63) lie
in the ventral part of the cord, their axons running out to make the
large fibres of the ventral roots and their dendrites passing to all parts
of the peripheral regions of both the same and the opposite sides of the
cord. They are thus presumably able to be stimulated directly by
impulses in the processes of the afferent fibres that end in these
regions.

Direct control of the spinal cord from the brain is obtained through

(99)

neur.p.

Fig. 63. Cells of the spinal cord of the larva of Lampetra.
A „nH R We motor-cells with dendrites reaching to the opposite side; C, small cells with
^p^SSno axon; «. axon; M/. Mailer's fibres; „*«,.>. neuropil at
P periphery of spinal cord. (After Tretjakoff.)

VERTEBRATES WITHOUT JAWS

opt.L.

IV. 9

cer.

hi/pot.

Fore.
cerh. h.

Mid- Hind- brain.

opt.L med. c ^ or

Fore. - Mid- Hind- brain.

Lam.i

Fore.

Mid.

Hind — brain.

Fig. 64. Brain of the lamprey.

A, side view; B, dorsal view with choroid plexus intact; c, after removal of choroid, cereb.
cerebellum; cer. h. cerebral hemisphere; chor. choroid plexus; hypot. hypothalamus; it. iter
between third and fourth ventricles; lam.t. lamina terminalis (thickened anterior wall of
third ventricle); med. medulla oblongata; opt.L optic lobe; pin. pineal eye; thai, thalamus;
3rd v., 4th v., third and fourth ventricles. (After Sterzi.)

a number of very large Miiller's fibres, originating from giant cells in
the reticular formation of the brain, whose large dendrites (Fig. 65)
receive fibres from several higher centres, providing an uncrossed final
common pathway to the spinal cord. There is some difference of
opinion as to whether any branches of these large fibres proceed

iv. 9 BRAIN OF LAMPREYS 101

directly into the ventral roots; probably they do not do so but the
dendrites of the motor-cells branch around them and thus receive
stimulation (Fig. 63). In the earliest larva co-ordination is by a pair of
giant Mauthner cells, with dendrites among the entering fibres of the
eighth nerve and an axon descending on the opposite side. Such cells
are present in the earliest stages of nearly all fishes and amphibians.

Other nerve-cells in the more dorsal parts of the cord have no long
axons and apparently serve to connect the neuropil of the various
regions. The afferent fibres reaching the cord in the dorsal roots give
off branches that ascend for a short distance and descend for long
distances. The pathways to the brain thus pass through multiple
relays.

The brain itself (Fig. 64) is built on the typical vertebrate plan, as
an enlargement of the front end of the spinal cord, with thickenings
and evaginations corresponding to the various organs of special sense.
Although we know little of its internal functional organization in
lampreys, it is probably not far wrong to regard it as chiefly consisting
of a series of hypertrophied special sensory centres; thus the forebrain
is connected with smell, midbrain with sight, hind-brain with
acoustico-lateral and taste-bud systems. The forebrain and olfactory
sense are moderately well developed in adult lampreys, as is the visual
sense, with its chief centre in the midbrain. The auditory and acoustico-
lateral systems are not very well marked, and the cerebellum is small.
Taste is also much less developed than in the higher fishes (p. 220).

Parts of the brain

Forebrain (prosencephalon) Cerebral hemispheres (telen-

cephalon)
Between-brain (diencephalon)
Midbrain (mesencephalon) Optic lobes

Hind-brain (rhombencephalon) Cerebellum (metencephalon)

Medulla oblongata (myelen-
cephalon)

The upper surface of the brain is covered by an extensive vascular
pad, the choroid plexus or tela choroidea (Fig. 64). This extends into
the ventricles of the brain at three points — into the third ventricle of
the diencephalon, into the iter (duct) leading through the midbrain
from third to fourth ventricles, and into the fourth ventricle itself.
The roof of the brain is thus non-nervous in these regions. In later
vertebrates the choroid extends only into the third and fourth

io2 VERTEBRATES WITHOUT JAWS iv. 9-

ventricles. Presumably the vascular membranes of the brain are highly
developed in lampreys because of the absence of cerebral blood
vessels.

From the lower part of the mid- and hind-brain arise all the cranial
nerves except the olfactory and optic. These nerves follow the same
plan as those of gnathostomes but they are difficult to make out by
dissection in the lamprey and will be left for consideration in con-
nexion with the dogfish, in which they can easily be dissected. The

beet. opt.
chor. pL3

chor.pt, 4.

otf.ep — IM-<^9 %

/nterped.

Fie. 65. Sagittal section through head of lamprey.

cereb. cerebellum; cer.h. cerebral hemisphere; chor.pl. 3 & 4, choroid plexuses of the 3rd and
4th ventricle, extending also into the midbrain; h.s. naso-hypophysial tube; hab. habenular
region; hyp. hypothalamus; interped. interpeduncular region; med. medulla oblongata; Mull.
M Oiler's cell; not. notochord; o. glandular organ of nasal sac; olf.ep. olfactory epithelium;
olf.n. olfactory nerve; p. ant., p. int., and p.nerv. partes anterior, intermedia, and nervosa of the
pituitary gland; parap. parapineal; pin. pineal; ted. opt. tectum opticum.

cranial nerves represent nerves similar to the dorsal and ventral nerve-
roots of the trunk, much modified as a result of the special develop-
ment of the head (p. 148). They carry afferent fibres from the skin of
the head and gills and motor-fibres for moving the eyes, sucker, and
branchial apparatus.

From the relative sizes of the parts of the brain it can be seen that
the various special sensory centres are still small. The largest part of
the brain is the medulla oblongata, which is well developed because

iv. io PINEAL EYES OF LAMPREYS 103

of the extensive sucking apparatus, innervated from the trigeminal
nerve.

The forebrain consists of a pair of large cerebral hemispheres and
these open by the foramina of Munro into a median third ventricle,
whose walls constitute the diencephalon or between-brain (Fig. 65).
This diencephalon, besides connecting the forebrain with the mid-
brain, includes the thalamus and serves important functions of its own.
Its ventral part, the hypothalamus, is well developed in all vertebrates
as a central organ controlling visceral activities and the internal life
of the organism. Nerve-fibres from the supraoptic nucleus of the
hypothalamus proceed to the pars nervosa of the pituitary and, as in
other vertebrates, are filled with granules of neurosecretory material,
which presumably controls pituitary action. A simple portal system
of blood-vessels connects the hypothalamus with the pituitary.

10. The pineal eyes

The diencephalon is also the region of the brain from which the
eyes are formed. In lampreys, besides the usual pair of eyes, there is
also, attached to the roof of the between-brain, the so-called third,
epiphysial, or median eye, better developed in these animals than in
any other living vertebrate except perhaps certain reptiles.

This organ is actually not median but consists of an unequally
developed pair of sacs, that on the right, the pineal, being larger and
placed dorsal to the morphologically left parapineal (Fig. 66). The
sacs form by evagination from the brain and remain connected with
the dorsal epithalamic or habenular region of the between-brain by
two stalks. The two organs are similar in structure, consisting of
irregular flattened sacs with a narrow lumen. Both upper and lower
walls of each organ contain receptor cells, with processes that project
into the lumen and nerve-fibres directed outwards. These fibres
apparently mostly end within the organ, in contact with ganglion
cells whose axons run to unequal right and left habenular ganglia.
In addition there are supporting and pigment cells in the retinas.
Knowles has shown that the retinal cells of the pineal make movements,
being arranged differently under conditions of illumination and dark-
ness. The significance of these photomechanical changes is unknown
but they demonstrate that the pineal cells are sensitive to light.

The structure of these pineal organs shows that they consist of
portions of the diencephalic wall where the ciliated cells of the epen-
dyma are specialized as photoreceptors. They show the same general
plan as the paired eyes, but with no differentiated dioptric apparatus.

IV. IO

104 VERTEBRATES WITHOUT JAWS

It has been possible to find out something of the part that these
organs play in the life of the lamprey. When a bright spot of light is
directed upon the pineal region of a stationary ammocoete larva move-
ment is usually initiated, but only after illumination for many seconds.

o.s.s.

Fig.

t.s.s., o.s.s

66. Pineal and parapineal organs of adult Lampetra fluviatilis
A. larva, B. adult. Sagittal section.

, inner and outer sensory cells; p. process; pin. pineal; p. pin. parapineal.
(After Tretjakoff.)

Moreover, these movements can be elicited even after the pineal
organs have been removed! In the larval lamprey the paired eyes are
deeply buried below pigmented skin, so the movement is not likely
to be due to them; indeed it continues when they too have been taken
out! Evidently there must be still other receptors, able to respond to
changes of light intensity in the wall of the diencephalon. This recalls
the fact that photoreceptors are found within the substance of the
nervous system of amphioxus. This power of response to changes of
illumination has been retained in the vertebrates, and persists in some

FUNCTION OF PINEAL EYES

105

as yet unknown cells in the brain, even after the paired and pineal
eyes have become specialized for light reception. The whole study is
of special interest as showing the stages by which the eyes may have
been evolved. Higher fishes also show the power of responding to

PINEALS REMOVED
1

rm — 1 1 1 1 1 1 1 1 1 1 M 15.

I I I i I I I I 1 ! ! I I I I II 1 I I II I I

\t up

hi

Y f

15 20 25 30 I S

I I M I I I I I I I I I I I I I I I I I I I I I I

DECEMBER JANUARY

2S26
I I

17

nil

10 15 ■
' '' 1 1 1 1 1 1 1 1

MARCH

Fig. 67. Colour-changes of larval lampreys, measured by the melanophore index (see
p. 300). Animals kept out of doors except as shown along the line AB, where rect-
angles above the line show illumination with electric light and below the line total
darkness. Normal animals show a regular daily rhythm, becoming pale at night.
Reversal of normal day and night illumination stops the change. On 19 December
the pineal eyes were removed from five out of the ten individuals and these there-
after remained dark (upper chart); the other five continued to show the normal
rhythm, until placed in total darkness. (After Young.)

changes of illumination after the paired eyes and epiphysis have been
removed (p. 210).

If the pineal eyes are not essential for the initiation of movement,
what is then their function ? In the ammocoete larva there is a daily
rhythm of change of colour, the animals becoming dark in the day-
time and pale at night. After removal of the pineal eyes this change no
longer occurs: the animals remain continually dark (Fig. 67). This
effect on the colour is produced by the action of influences from the
pineal, passing to the pituitary gland (see p. 103). It seems that
the pineal apparatus is an organ concerned with adjustment of the
internal activities of the animal to correspond to the changing con-
ditions of illumination. The control may be effected by impulses

io6

VERTEBRATES WITHOUT JAWS

carried in the large tract that proceeds from the habenular ganglion
to the hypothalamus, in the floor of the diencephalon. The latter is
known to be concerned, throughout the vertebrate series, with the
integration of the internal activities of the animal.

¥ d Y

pner.

p. int.

pant

Fig. 68. Sagittal section of the pituitary gland of lamprey.

dien. diencephalon; inf. infundibulum; p.ant., p.int., and p. tier, partes anterior, intermedia,

and nervosa, there are two types of cell in the pars anterior; 3rd V. third ventricle.

(After Stendell.)

3 4 5 6

Hours after operation

FlG. 69. Onset of pallor in a larval lamprey after removal of the pituitary gland,
as shown by the decline in the melanophore index. (From Young.)

1 1 . Pituitary body and hypophysial sac

The lower portion of the diencephalon, the hypothalamus, forms
a prominent pair of sacs, the lobi inferiores, which contain a partly
separated diverticulum of the third ventricle and end below in the
infundibulum (Fig. 65). The pituitary gland (hypophysis) is pressed
against the underside of the hypothalamus (Fig. 68). The lower wall
of the brain in this region consists not of nerve-cells but of a single

IV. II

PITUITARY OF LAMPREYS

107

-^fMvn/w^^

epithelial layer, corresponding to the pars nervosa of the pituitary of
higher forms. The major portion of the pituitary gland is a mass of
secreting cells in which two parts can be recognized, the partes
anterior and intermedia. After experimental removal of the inter-
mediate portion of the pituitary lam-
preys become permanently pale in
colour (Fig. 69), showing that, as in
other vertebrates (p. 299), a melano-
phore-expanding substance is liberated
into the blood by this gland, the secre-
tion being presumably under the control
of the pineal eyes (p. 105). The lamprey
pituitary has been shown to contain
oxytocic and 'water balance' hormones
as well as one producing melanophore
expansion. Moreover injections of
mammalian anterior pituitary extracts
induce appearance of the secondary
sexual characters of lampreys. Evi-
dently the functions of the pituitary
have remained essentially the same
through the whole chordate series.

The pituitary of lampreys is peculiar
because of great development of the
naso-hypophysial sac (Fig. 70). Charac-
teristically in vertebrates the pituitary
body develops by the formation of a
pocket of buccal ectoderm, whose walls
then become folded, so that the part
in front of the lumen becomes the pars
anterior, that behind the pars inter-
media. In nearly all vertebrates the
lumen then loses its connexion with the
exterior. In lampreys the hypophysial
rudiment is continuous with that of the
olfactory epithelium. The latter then
moves dorsally and the two remain

connected throughout larval life by a strand of cells. At metamorphosis
this acquires a lumen and forms a tube extending from the nostril
below the pituitary and brain. Because of its development this is
sometimes called the naso-hypophysial tube but others doubt that

Fig. 70. Dissection of lamprey from
the ventral surface after injection
of coloured gelatine to show the
outline of the naso-hypophysial sac
(.1) and its duct (d), which is shown
dotted where it runs upwards be-
tween the nasal sacs («). g, gill
pouches. Contraction of the bran-
chial apparatus squeezes the sac s,
so that water is drawn in at each
relaxation.

io8

VERTEBRATES WITHOUT JAWS

IV. II-

it represents the cavity of the hypophysis and prefer the name
nasopalatine canal.

Inside the single nostril, guarded by a valve, are openings into the
nasal sacs, which are cavities with folded walls. Some of the cells of
these walls are the olfactory receptors and give off the axons that make
up the olfactory nerves, entering the olfactory bulbs on the anterior
end of the hemisphere (Fig. 65). Behind the nasal sacs lie numerous
glandular follicles opening into the sac in the larva, but completely
closed in the adult (Fig. 65). They may be comparable to Jacobson's
organ (p. 405).

Fig. 71. Section of lateral line organ of tail of adult Lampetra.

p. pigmented cells around the pit; s.c. receptor cells
(not showing long hairs). (After Young.)

The naso-hypophysial tube proceeds back behind the pituitary to
a closed sac lying between the first pair of gill pouches (Fig. 70).
During the movements of respiration this sac is squeezed and water
is expelled with some force through the nostril. When the gills relax
water flows in at the nostril, and in this way the olfactory organ is
provided with samples. If the naso-hypophysial opening is closed with
a plug of plasticine the lamprey no longer reacts to solutions, for
instance of alcohol, to which it normally responds by freeing its
sucker and swimming away.

12. Lateral line organs of lampreys

The lateral line receptors, peculiar to fish-like vertebrates, are little
patches of sensory cells found along certain lines on the head and
trunk. They are all innervated by cranial nerves, those on the body
and tail being served by a special backward branch of the vagus nerve.
The receptor cells carry long hairs and are thus able to detect either
movement of the water relative to the fish or of the fish itself. Objects
moving nearby set up disturbances that may also be detected (p. 218).

v. i3

LABYRINTH OF LAMPREYS

109

In the lamprey the lateral line organs are very simple (Fig. 71), being
open to the exterior and not sunk in a canal as in higher forms. The
rows are somewhat irregular, especially those on the body.

13. Vestibular organs of lampreys

The labyrinth may be considered as a specialized portion of the
lateral line system, concerned with recording the position of the head

end.

mac.laq. \ mac.ut.
3 mac. sac.

smp.p.

amp a.

Fig. 72. Labyrinth of right side, seen in lateral view. A and B,

Lampetrci, c, Myxitie.

a.c. anterior canal; amp. a. and p. anterior and posterior ampullae; cil.

ciliated chamber; cr.a. and p. cristae; end. endolymphatic duct; lag.

lagena; mac. lag., neg., sacc, ut. maculae of the lagena, neglecta,

saccule, and utricle; p.c. posterior canal. (After de Burlet.)

and angular accelerations. There is no evidence to decide whether
lampreys can respond to sound. The labyrinth develops by an in-
pushing of the wall of the head, and this then becomes closed off from
the exterior. Internal foldings divide up the sac into a number of
chambers, which differ considerably from those of gnathostomes.
There is a large central vestibule, into which open below several
partially separate sacs, provided with patches of sensory hairs. These
correspond, from in front backwards, to the maculae of the utricle,
saccule, and lagena of higher forms (Fig. 72). The hairs of the maculae
are loaded with otoliths. There are only two broad semicircular
canals, corresponding to the anterior and posterior vertical canals
of other vertebrates, each with an ampulla, containing a receptor
ridge, the crista. Also opening to the vestibule are two large sacs,

no VERTEBRATES WITHOUT JAWS iv. 13-

covered with cilia (Fig. 72), whose beat produces complicated
counter currents in the dorso-ventral plane. It has been suggested
that these function as a gyroscope, compensating for the absence of
a horizontal canal.

In Myxine the condition is even simpler, there being only a single
vertical semicircular canal (Fig. 72). However, it is claimed that this

Fig. 73. Horizontal section of the eye of a lamprey.

er., ir., sr., external, internal, and superior rectus muscles; v.v. venous sinuses which cushion
the eye. (From Walls, The Vertebrate Eye, Cranbrook Institute of Science.)

has cristae at both ends. The macular system also does not show the
characteristic subdivisions but is a single macula communis.

14. Paired eyes of lampreys

The structure of the paired eyes is similar to that in other verte-
brates. They are formed, like the pineal eyes, by evaginations of the
wall of the diencephalon; the so-called optic nerve is therefore not
really a peripheral nerve but a portion of the brain; it should strictly
be called the optic tract. The eyes are moved by extrinsic muscles
arranged in a somewhat unusual manner. Accommodation is effected
by a process found in no other vertebrates. The cornea consists of two
distinct layers, separated by a gelatinous substance. Attached to the
outer (or dermoid) cornea is a cornealis muscle, apparently of myo-
tomal origin, which flattens the cornea and pushes the lens closer
to the retina (Fig. 73). There is an iris, outlining a round pupil, which
changes little, if at all, in diameter under different illuminations. Most
species of lampreys are diurnal animals. They are said to move towards

iv. is

PHOTORECEPTORS OF LAMPREYS

white objects and probably use both the eyes and the nose to find
their prey. In the ammocoete larva the paired eyes are buried below
the pigmented skin and the animal makes no movements when light
is shone on to this region.

The optic tracts of adult lampreys end in the roof of the midbrain
(tectum opticum) which is a highly differentiated, stratified region.
Besides the optic fibres it receives also impulses from fibres ascending
from the spinal cord and others from the auditory and lateral line

Fig. 74. Experiment to show behaviour of larval lampreys when illuminated. The
tank is left in total darkness and the larvae settle in all parts. When the light is
switched on those in the illuminated part begin to swim and continue to do so until
by chance they arrive in the darkened part, where they settle down. (From Young.)

centres. The midbrain is therefore undoubtedly one of the most
important parts of the brain in lampreys, though nothing is known in
detail of its functions. Its cells control movements of the animal, by
means of fibres that run to make connexion with the dendrites of the
large Midler's cells, whose axons pass down the spinal cord; other
fibres from the tectum opticum reach to various parts of the brain,
and it is probable that its activities are closely correlated with those of
many other regions.

15. Skin photoreceptors

Like many lower vertebrates the lamprey has light-sensitive cells in
the skin, as well as those in the eyes. These receptors are abundant in
the tail and if a light is shone on to this region the animal rapidly
moves away (Figs. 74 and 76). If the spinal cord is cut just behind the
head and a light then shone on to the tail, the head will be seen to
move. This suggests that the impulses are carried forwards by means
of the lateral line nerves, which is confirmed by the fact that if these
latter are sectioned, leaving the spinal cord intact, then no movements
follow when the tail is illuminated. This sensitivity of the lateral line

ii2 VERTEBRATES WITHOUT JAWS iv. 15-

organs to light is not found in other fish-like vertebrates. Indeed the
receptors are not strictly lateral line organs but pigmented epidermal
cells. The sensitivity curve shows a sharp peak at 530 m/x, this being
the region of the spectrum at which light penetrates farthest into sea
water. The pigment is probably a porphyropsin (Steven, 1950).

In hag-fishes (Myxine) the head and cloacal regions are more
sensitive to light than is the rest of the body. The impulses from the
skin are conducted through the spinal nerves in these animals, not
the lateral line nerves.

16. Habits and life-history of lampreys

We have very little information about the life of lampreys during
the time that they are in the sea. They are caught in considerable
number attached to other fishes. It is not known how many years a
lamprey spends in the sea, but it returns only once to the river for
spawning and dies after this act. The up-river migration of L. fluvi-
atilis occurs in the autumn, for instance large numbers come up the
River Severn and are caught in traps on the way, for use as food. The
spawning migrations of lampreys may take them for hundreds of
miles, for example, those of the eastern Pacific ascend to the head-
waters of the Columbia River. They are said to perform remarkable
feats of climbing, leaping from stone to stone and hanging on by their
suckers. During this period of migration some lampreys assume
brilliant orange and black colour patterns. On the other hand, lam-
preys land-locked in the lakes of New York (Petromyzon marinus
unicolor) feed in fresh water and ascend only a few miles up streams
to breed.

Once in the river the lampreys do not feed again but live over the
winter on the reserves accumulated in the form of fat, especially under
the skin and in the muscles. During the winter the gonads ripen pro-
gressively and the secondary sexual characters begin to become appar-
ent only in February. The females then develop a large anal fin, while
in the male a penis-like organ appears (Fig. 47) and the base of the
dorsal fin becomes thickened.

Spawning occurs in the spring and is preceded by a form of nest-
building. Numerous lampreys collect together, usually at a place
below a weir where the water is shallow and rather swift, and the
bottom both stony and sandy. Stones are then dragged by the mouth
in such a way as to make a small depression. Fertilization is secured
by a process of copulation in which the male fixes by the sucker on to
the fore-part of the female and the two then become intertwined and

iv. 1 6

REPRODUCTION OF LAMPREYS

"3

undergo rapid contortions, the eggs being squeezed into the water,
while sperms are ejected through the 'penis' (Fig. 75). Fertilization
is therefore external, but the sperms must be placed very close to the

Fig. 75. Spawning lampreys seen in their nest.
(After Gage.)

eggs, for they remain active only for about one minute after entering
the fresh water, which provides the stimulus that activates them. The
eggs and sperms are not all laid at once; mating is repeated several
times until all the products have been shed, after which the animals
are exhausted and soon die. The movements of the animals stir up
the sand in the nest (this is probably the function of the anal fin of
the female) ensuring that the eggs are covered up as they are carried
away by the current.

ii 4 VERTEBRATES WITHOUT JAWS iv. 17

17. The ammocoete larva

The eggs contain a considerable quantity of yolk, but their cleavage
is total and proceeds in a manner not unlike that of the frog. After
about three weeks the young hatches as the ammocoete larva, about
7 mm long. At first this is a tiny transparent creature, but its larval
life lasts for a long time, during which it grows into an opaque eel-like
fish, up to 170 mm long (Fig. 76).

Fig. 76. Ammocoete larva of Lampetra planeri, showing the effect of shining a narrow

beam of light on to various parts of the side of the body. Illumination of 1, 2, or 8 is

followed by movement after a few seconds, but no movement follows illumination at

points 3-7. (From Young.)

end.

Fig. 77. Young ammocoete larva of lamprey fixed while feeding on green flagellates
and detritus and then stained and cleared.

an. auditory sac; br. brain (covered by meninges); e. eye; end. endostyle \f.c. food cord in
pharynx; h. heart; /. liver; m. mid-gut; oes. oesophagus; v. velar fold.

This portion of life is spent buried in the mud, the animals emerg-
ing only occasionally to change their feeding-ground, presumably if
the mud is not sufficiently nutritious. There is no sucker, the mouth
being surrounded by an oral hood rather like that of amphioxus (Fig.
77). The paired eyes are covered by muscles and skin. The head at
this stage is little sensitive to light, but the animal quickly begins to
swim if the tail is illuminated. We have seen already (p. 108) that in
lampreys there are photoreceptors in the tail, connected with the
lateral line nerves. In the larva these are the main photoreceptors, and
they ensure that the animal lies completely buried.

If a number of larvae are left in a vessel with a layer of mud on the

iv. i7

AMMOCOETE LARVA
r 2 3

H5

Fig. 78. Development of the endostyle of the lamprey. Sagittal sections through

the head at three stages.

1, auditory sac; 2, medullary tube; 3, myotome; 4, conus arteriosus; 5, endostyle; 6, first

gill-slit; 7, first arterial arch; 8, notochord; 9, inpushings which cut off the endostyle from

the pharynx; 10, aorta; 11, stomodaeum. (After Dohrn, from Kukenthal.)

bottom they rapidly disappear and remain hidden indefinitely, the
heads perhaps just visible in small depressions made by the rhythmic
respiratory movements. When disturbed they always swim with the
head downwards and in contact when possible with the ground. This
habit leads them to burrow rapidly. It is not known whether they
have other receptors to guide them to mud rich in possible food
organisms. The nasal and hypophysial sacs are poorly developed in the
larva, and the sense of smell can hardly serve this purpose.

(n6)

t.m

bEd.

Fig. 79. A, Transverse section of endostyle of ammocoete larva.

cav. cavity of the gland; cil.g. ciliated groove in floor of pharynx; lam. lamellae of gills;
ph. cavity of pharynx; seer, mucus-secreting cells of the gland.

b, Transverse section of thyroid follicles of adult. (After Young and Bellerby.)

C, Cross section of endostyle of ammocoetes larva of Petromyzon tnarinus at level
where it is connected by a duct to the pharynx. Autoradiograph showing distribu-
tion of protein-bound I 131 . The radioactive clumps of cellular debris in the
glandular lumen and in the duct suggest that the material represents a holocrine
secretion, which will probably be absorbed in the intestine.

d. duct; hs. cellular debris; ph. pharynx; t II d. type II dorsal cells;
t III epithelial cells,

iv. 17 FEEDING OF AMMOCOETE LARVA 117

Feeding takes place by the intake of water through the mouth and
the separation of small food particles from it in the pharynx (Fig. 77).
For this purpose there is used a great quantity of mucus, which is
secreted by the endostyle and gathered into a strand by the cilia of
the pharynx. This endostyle is a most remarkable organ, forming early
in development as a sac below the pharynx (Fig. 78). It consists of a
pair of tubes, on the floor of which there are four rows of secretory
cells (Fig. 79). There is a single opening to the pharynx, by a slit
at about the middle of the length. As development proceeds the inner
rows of cells at the hind part of the organ become coiled upwards, and
at the end of larval life the endostyle therefore forms a very large mass
below the pharynx, composed of tubes lined partly by secretory and
partly by ciliated cells. Probably no enzymes are secreted by the
endostyle, its function being to produce mucus in which the food
particles become entangled. Although it resembles the endostyle of
amphioxus in the arrangement of the secretory columns, there is a
difference in that the organ in the ammocoete larva is not an open
groove. There is, however, a ciliated groove in the floor of the pharynx,
that is to say, on the roof of the endostyle (Fig. 79).

The details of the feeding-currents of the ammocoete larva are not
understood. An important difference from the arrangement in am-
phioxus is that the current is produced by muscular rather than ciliary
action. The velum, a pair of muscular flaps, provides the main current
when the animal is at rest. The branchial basket can also be expanded
and contracted by an elaborate system of muscles. It is not easy to
observe how the food particles are taken up from the current, but
apparently a strand of mucus shoots from the endostyle and occupies
the whole of the centre of the pharynx (Fig. 77). This strand probably
rotates and as it passes backwards into the eosophagus it catches the
particles.

Evidently the system enables the animals to feed efficiently on the
small unicellular algae and bacteria of the mud. In amphioxus the
ciliated pharynx, occupying a considerable proportion of the whole
surface, is only able to support a tiny creature, but the muscular
feeding-system of the ammocoete allows a relatively small pharynx to
feed a fish 170 mm long and weighing up to 10 grams. This use of
muscles for moving the gills was evidently an important step in
chordate evolution. It allowed the animals to escape from the limita-
tion of size imposed by the ciliary method of feeding. After the
development of jaws to form a still more efficient feeding mechanism
the rhythmic movement of the branchial apparatus persisted for the

n8 VERTEBRATES WITHOUT JAWS iv. 17-

purpose of respiration. We cannot be certain about changes which
occurred so long ago, but it seems likely that the respiratory move-
ments of a fish were first introduced to provide food rather than
oxygen.

The endostyle therefore shows the survival of the primitive feeding-
methods of chordates, but it also undergoes at metamorphosis an
astonishing change into a thyroid gland. The mucus-secreting columns
shrink and the whole organ becomes reduced to a row of closed sacs,
lying below the pharynx (Fig. 79 b). Each of these sacs is lined by an
epithelium, contains a structureless 'colloid' substance, and is there-
fore closely similar to a thyroid vesicle. Moreover, experiments have
shown that extracts of this organ contain iodine and exert an accelerat-
ing effect on the metamorphosis of frog tadpoles. Although nothing is
known of the part played by the secretion of this gland in the life of
the adult lamprey, we may safely conclude that we have here the
conversion of an externally secreting feeding-organ into a gland, of
internal secretion. The actual mucus-secreting cells are not trans-
formed into those of the thyroid follicles, these latter are derived from
epithelial cells in the wall of the larval organ. One cannot avoid specu-
lating on this extraordinary change of function. It may perhaps be
significant that the endocrine gland that regulates basal metabolism
(the thyroid) is derived from the part of the feeding-system that in
the earliest chordates was responsible for providing the raw materials
of metabolism. Experiments with radioactive iodine show that this
element is concentrated in certain cells of the larval endostyle (Fig.
79 c). Moreover, after addition of the anti-thyroid substance thiourea
to the water there are changes in the endostyle. Thyroxine has been
extracted from the gland and it probably has an endocrine function
as well as secreting mucus, though no one has ever produced any
changes in larval lampreys by administering thyroid hormones.
Lampreys thus show, as larvae, a stage in which the accumulation of
iodoproteins, previously widespread, becomes concentrated in the
pharynx. Perhaps at this site there were already cells specialized for
halide transport (cf. the chloride-secreting cells of teleosts, used for
osmogulation, p. 203). In adult lampreys and all higher chordates the
iodoprotein is secreted into the blood under the control of blood-
borne signals (Fig. 80). The change may well be related to develop-
ments in the regulation of metabolism, which, in the animals with a
fully endocrine thyroid becomes more nearly independent of varia-
tions in the external supply of iodine.

The great change in the endostyle is only part of the complete

RACES OF LAMPREYS

119

metamorphosis by which the ammocoete larva changes into an adult
lamprey. The mouth becomes rounded and its teeth, tongue, and
complex musculature develop. The paired eyes (previously buried)
appear; the olfactory organ becomes internally folded, and the olfac-
tory nerve and tracts much enlarged. The naso-hypophysial sac grows
backwards to the gills. In the pharynx the gills develop into sacs
opening to the branchial chamber. Changes also take place in the

B

Fig. 80. Diagram to show distribution of iodoproteins, at first in exoskelctal struc-
tures, as in many invertebrates and in tunicates (a). Some of this material is
concentrated in the pharynx. This tendency is exaggerated in amphioxus and the
ammocoete larva, and in the adult lamprey and later animals this pharyngeal
material forms the thyroid. A. Many invertebrates and tunicates ; n. Amphioxus ;
c. Ammocoetes; D. Metamorphosis of ammocoetes; e. General vertebrate type.
(After Gorbmann, A., in Comparative Endocrinology. Wiley, New York.)

intestine. The yellow-brown colour of the larva gives place to the
black with silver underside of the adult. The animal more and more
frequently leaves the mud and finally migrates to the sea to begin its
parasitic life.

18. Races of lampreys, a problem in systematics

Besides the river lampreys, such as L. fliwiatilis (Linn.), which
show this characteristic migratory life-history, there are also in various
parts of the northern hemisphere small brook lampreys ('prides'),
such as L. planeri Bloch, which remain throughout their life in fresh
water. These prides are very abundant in many English rivers and
streams, but since the greater part of their life is passed in the am-
mocoete stage they are not often seen. The larvae remain in the mud

izo VERTEBRATES WITHOUT JAWS iv. 18

probably for three years and undergo metamorphosis in late summer
and autumn. The characteristic of this type of lamprey is that the
adults never migrate and never feed. The gonads are already well
developed at metamorphosis and ripen during the winter. Spawning
takes place in March or April and the animals then die.

There has been much dispute about the status of these freshwater
races. In structure the adult L. planeri is nearly if not quite identical
with an adult L.fiuviatilis, except that the latter is much the larger and
has sharper teeth. Crossing of the two sorts could presumably never
take place in nature, on account of the size difference, but by artificial
stripping of the adults cross-fertilization in both directions can easily
be achieved. Unfortunately the hybrid larvae have never been reared
to maturity; we cannot therefore say whether the small size and
failure to migrate of the planeri forms are inherited characters or
are produced by the influence of the environment. The effect of the
non-migratory condition is to enable the lampreys to colonize very
fully rivers that, because of effluents, they would be unable to occupy
if a migration to the sea was necessary. By this process of acceleration
of the development of the gonads a dangerous stage in the life-history
has been avoided.

Similar pairs of migratory and non-migratory forms of lamprey are
found in Japan and in North America. Indeed, the condition appears
to be developing independently in several river systems in the United
States. Since it may be difficult for the brook lampreys to spread
from one river system to another it is possible that many of the
planeri forms have evolved separately, perhaps quite recently. If so,
this is a remarkable example of a similar response produced in different
parts of a population by a similar environmental stimulus, in this
case the effluents. This process of alteration in the relative times of
metamorphosis and sexual maturity (paedomorphosis) has occurred
also in certain amphibians (the axolotl) and in tunicates (Larvacea).
Similar changes in rates of development may have been essential
factors in the development of the whole chordate phylum (p. 77).

In one race, found in Italy, ammocoetes with mature gonads have
been reported. However, in most of these lampreys the paedomor-
phosis is only partial : metamorphosis does take place, but is immedi-
ately followed by maturity. Since in mammals injections of anterior
pituitary extracts accelerate development of the gonads, it was thought
possible that complete neoteny might be produced by making such
injections into larvae of L. planeri. No completely sexually mature
ammocoetes have yet been produced by this method, but following

iv. 18 NOMENCLATURE FOR LAMPREYS 121

the injections the larvae assumed the secondary sexual characters,
which are normally shown only at maturity, namely, swelling of the
cloaca, opening of the pore from coelom to exterior, and the changes
in body form. No signs of metamorphosis were produced by these
injections and we are left without information as to the cause of that
change in the lamprey. In Amphibia even very young larvae undergo
metamorphosis when treated with thyroid extracts, but similar treat-
ment of ammocoete larvae has failed to produce any change. Further
investigation of the problem should be very interesting, since it seems
likely that the differences between the fluviatilis and planeri forms are
the result of an endocrine factor accelerating the onset of sexual
maturity in the latter. The fact that the change is occurring in various
parts of the world adds further interest to this example of evolution
in progress.

Besides all these relatively small lampreys, there is a much larger
form, the sea lamprey, Petromyzon marinns Linn., reaching to over a
metre in length. This animal differs from Lampetra in body form,
structure of sucker, and other features, as well as in size. Like most
other groups of animals lampreys therefore present several problems
of nomenclature. Linnaeus included the three types that occur in
Europe in the one genus Petromyzon ; since they are all rather alike in
shape this is in some ways a reasonable procedure. But are we then
also to include in the same genus forms that differ more widely, such
as those occurring in the southern hemisphere ? As so often happens,
systematists have chosen the course of splitting up the Linnaean
genus, even though several of the resulting genera have only one
species. Thus Gray suggested the genus Lampetra for the brook and
river lampreys, keeping Petromyzon for the larger species of sea
lamprey. Other genera have been added, such as Entosphenus Gill for
some of the North American forms and Mordacia Gray and Geotria
Gray for the forms from the southern hemisphere (Chile, Australia,
and New Zealand). Such distinctions, though they may seem irritating
at first sight, are an advantage in that they call attention to the differ-
ences which exist. For instance, it is a striking fact that lampreys are
found in temperate waters of both hemispheres, but not in the tropics,
and it is interesting to learn that the forms from New Zealand,
Australia, and South America (there are none in South Africa) show
distinct peculiarities. Thus Geotria possesses a large sac behind the
sucker.

A special problem of nomenclature arises from the fact that the
river and brook lampreys are almost identical in structure and differ

i22 VERTEBRATES WITHOUT JAWS iv. 18-

mainly in size, time of sexual maturity, and habits. A further com-
plication is that the germ-cells of the two races allow cross-fertiliza-
tion, although this probably never occurs in nature! We may take
Dobhzansky's definition of species as 'groups of populations which
are reproductively isolated to the extent that the exchange of genes
between them is absent or so slow that the genetic differences are not

Fig. 8i. Myxine, partly dissected.

i, cloaca; 2, testis; 3 and 4, ovary with eggs; 5, liver; 6, branchial opening; 7, mouth;
8, nostril; 9 and 11, slime glands; 10, intestine. (After Retzius, from Kukenthal.)

diminished or swamped', and in this sense we may retain the specific
names L. fluviatilis and L. planeri for the two populations.

19. Hag-fishes, order Myxinoidea

The hag-fishes, Myxine and Bdellostoma (Fig. 8 1 ), are animals highly
modified for sucking. They live buried in mud or sand and probably eat
polychaetes and other invertebrates, as well as scavenging dead fishes.
The eyes are functionless rudiments, though the animals are sensitive
to changes of illumination, through skin receptors. There are sensory
tentacles around the mouth, and in both hag-fishes the teeth and
sucking apparatus are well developed. They burrow into the bodies
of dead or dying fishes. As many as 123 Myxine have been taken from
a single fish. Since the introduction of trawling they have become less
common in the North Sea, where they used to be a serious source of
loss to fishermen by their attacks on fishes caught in drift nets or on
lines. They seem to find fish when they are dying or just dead, and
entering by the mouth of their prey eat out the whole contents of the
body, leaving a sack of skin and bones. When they are themselves

IV. 19

HAG-FISHES

123

caught on lines (for instance, with a salted herring bait) the hook is
swallowed so deeply that it may be found near the anus!

The gills are modified into pouches (6-14 in Bdellostoma, 6 in

-5

Fig. 82. Arrangement of gills in Bdellostoma and Myxine.

1, tentacles; 2, wall of pharynx ; 3, branchial sac opened to show gill lamellae; 4, branchial

duct; 5, branchial sac; 6, mouth; 7, common branchial aperture in Myxine.

(From Kukenthal after Dean.)

Myxine), opening by tubes into the pharynx, and to the exterior (Fig.
82). In Myxine all the tubes are joined and open by a single posterior
aperture on each side. Water enters at the nostril and is pumped back-
wards by a muscular velum through the gill chambers and out behind.
There is also a single posterior oesophago-cutaneous duct on the left
side, which is probably closed during normal respiration but is
opened to allow expulsion of large particles. If the nostril is closed
experimentally with a plug no water enters by the mouth or posterior
apertures but the fish survives well, presumably respiring through the
skin.

124 VERTEBRATES WITHOUT JAWS iv. 19-

The thyroid gland consists of a long series of sacs formed by
evagination from the floor of the pharynx.

Down the sides of the body are pairs of slime glands, able to secrete
large amounts of mucus, which may be protective and is said also to
be produced under the operculum to hasten the end of a dying fish
that the hag has attacked.

A curious difference from the nervous system of lampreys is that
the dorsal and ventral roots join, though the details suggest that the
union is not similar to that found in gnathostome vertebrates. The
brain shows several features of reduction and simplification and no
pineal eyes are present. There is only one semicircular canal in the ear
(p. 109). The kidneys show a more generalized condition than in any
other vertebrate in that the pronephros persists in the adult and is
hardly marked off from the mesonephros, so that an almost continu-
ous series of funnels and glomeruli can be recognized. Moreover, there
is a regular series of mesonephric glomeruli, a pair in each segment.

The development is known only in Bdellostoma, where the egg is
yolky and cleavage partial, leading to the formation of an embryo
perched on a mass of yolk. It is often stated that Myxine is a protandric
hermaphrodite, because individuals are found in which the front end
of the gonad contains eggs, whereas the hind part is testis-like (Fig.
81). No ripe sperms have ever been found in this region, however,
and, moreover, individuals with fully testicular gonads do occur.
Since it is known that in other vertebrates (including the lampreys)
the gonads go through a hermaphrodite stage during development it
seems likely that Myxine is not a functional hermaphrodite but that
the double-sexed gonad shows a rather late persistence of the indeter-
minate stage.

The hag-fishes all live in the sea and their blood differs from that of
other chordates in that it is isosmotic with sea water. However, the
individual ions are regulated; sodium and phosphate exceed their
values in sea water, and the other ions are present in lower concentra-
tion. It is usually assumed that fishes, with their glomerular kidneys,
evolved in fresh water. However, the very earliest fragments of
armoured agnathans are from Ordovician deposits that may be littoral
or marine and it might be that the condition of the blood and kidney
of Myxine is that of the earliest agnathans and that the glomerulus was
not evolved as an adaptation to freshwater life, as is often supposed
(Robertson, 1954).

The organization of the lampreys and hag-fishes shows that they
preserve many characteristics from a very early stage of chordate

IV. 20

FOSSIL AGNATHANS

125

evolution, probably that of about the Silurian period. Their special
interest for us is in giving an insight into the organization possessed
by the vertebrates before jaws were evolved. However, no doubt many
changes have gone on during cyclostome evolution and we must not
suppose that all Silurian vertebrates were like lampreys. Indeed, we
may now complete our picture of this stage of evolution by examining
the fossil fishes known to have existed at that period. We shall find
them superficially so different from modern cyclostomes that only
careful morphological comparison reveals the similarities. The inquiry
will show us once again how a common plan of organization can be
found in animals of very different superficial form and habits.

Fig. 83. A ccphalaspid restored (Hemicyclaspis).

d. dorsal fin; bf. lateral field; pec. pectoral fin; p. pineal; sclr. sclerotic ring.
(From Stcnsio.)

20. Fossil Agnatha, the earliest-known vertebrates

The ostracoderms are fossil forms from freshwater Silurian and
Devonian deposits. They are therefore the oldest fossil vertebrates
known to us (except for a few Ordovician fragments), and this makes
it specially interesting that they show affinity with the cyclostomes.
These are fossils that are rarely found complete, particularly the
pteraspids, but a quarry in Herefordshire yielded numerous whole
specimens of Cephalaspis and Pteraspis of Old Red Sandstone age,
probably all from a single dried-up pool.

In the cephalaspids (Osteostraci) the head was flattened and com-
posed largely of a shield. The rest of the body was fish-like, with an
upturned tail (heterocercal, see p. 136) covered with heavy bony scales
(Fig. 83). A pair of flaps behind the gills may have functioned like
pectoral fins.

On the dorsal surface of the shield are two median holes, one
behind the other, which served a naso-hypophysial opening and a
pineal eye. The whole outline of the cranial cavity is preserved and
shows a brain remarkably like that of a lamprey, with a naso-hypo-
physial canal below it (Fig. 84). There were paired eyes and only two

(126)

eFf.pm

l.br.l

Fig. 84. Head shield of the ccphalaspid Kiaeraspis, see from below. 5 X natural size.

car. a. internal carotid; d.ao. dorsal aorta; eff.br. 4. 4th efferent branchial; eff.pm. efferent branchial

of 1st arch; i.br.i. 1st interbranchial ridge; oes. oesophagus; orb. depression made by orbit; vest.

depression made by vestibular apparatus. V1-X2. Cranial nerves.

(After Stensio.)

hup. foss

orb

■ C3p tat

Mf/m

Fig. 85. Cast of the endocranium and system of canals in the head shield of Kiaeraspis.
amp. ampulla of posterior semicirculr canal; can. canal leading to field; car. a. carotid artery;
hyp.foss. hypophysial fossa; /./. lateral 'electric' fields; med. medulla oblongata; nas.c. naso-
hypophysial canal; orb. orbit; v.cap.lat. vena capitis lateralis; vest, vestibular apparatus; III-X

cranial nerves. (After Stensio.)

iv. 2 o CEPHALASPIDS 127

semicircular canals. Long tubes leading through the shield contained
the cranial nerves, which can be reconstructed in detail (Fig. 85).
On the under side of the shield is a series of ridges, which outline a
set of ten pairs of branchial pouches. The first of these lies far forward
at the sides of the mouth and the ridge in front of it is probably the
premandibular arch; it carries the profundus nerve (p. 152), which
was large. The ventral surface of the head was flat and covered with
small scales. Probably the gills were pouches, as in lampreys. The
canals of the aorta, epibranchial arteries, and some features of the veins
and heart have been preserved.

The mouth was a slit at the extreme front end with which the
animals may have scooped decaying matter from the lake floor.
On the dorsal surface there are sunken areas, covered by small scales,
known as the median and lateral fields, and supposed by some to have
contained electric organs. They were apparently served by a very rich
blood-supply and a system of wide canals leads to the vestibular
region. These canals might have contained nerves, but Watson makes
the far more likely suggestion that they housed tubular extensions of
the labyrinth and served to carry pressure waves to the ear, perhaps
providing a substitute reinforcement for the defective lateral line
system.

We therefore know in some respects as much about these fossils as
of many living fishes. They show in the complete segmentation of the
head the most primitive condition known among craniates. Many of
their features are very like those of modern lampreys and there can be
little doubt that, as Stensio suggests, the latter represent their sur-
viving descendants, which have lost the bony shield.

The Anaspida (mostly Silurian) are placed by Stensio near the
Cephalaspids but they are less well known. They were small fishes
(up to 7 in. in length) covered with rows of bony scales (Fig. 87). The
tail shows a lower lobe larger than the upper ('hypocercal'). This
would presumably serve to drive the head end upwards perhaps to
compensate for the weight of its armour. The opposite ('heterocercal')
condition, found in cephalaspids and many modern fishes (for
instance, the dogfish), produces a tendency to negative pitch and is
associated with the presence of pectoral fins (p. 136). The anaspids
possessed a curious ventral or ventro-lateral fin fold (Fig. 87) or
perhaps a series of them. There were large paired eyes, median holes
presumed to be nasal and pineal and a series of up to fifteen small
round gill openings.

We may consider here the fossil Jamoytius from the Silurian. The

i 2 8 VERTEBRATES WITHOUT JAWS iv. 20

notochord was persistent and there was no calcined endoskeleton.
There were long continuous lateral fin folds and a hypocercal tail. A
series of transverse structures were at first interpreted as myotomes

Pterolepis

Rhyncholepis

sec/.

pec.

Rhyncholepis

Figs. 86 and 87. Anaspids seen in dorsal and lateral views.

an. anal fin; na. nasal aperture; orb. orbit; pec. pectoral spine; pi. pineal foramen;
ros. rostrum; sc.d. dorsal scales. (After Stensio and Kiaer and Grasse.)

but Stensio and Ritchie (i960) consider these to be scales and place
Jamoytins with the Anaspida. In either case, the form is of the greatest
interest, and represents as White says 'the most primitive of the
"vertebrate" series of which we have knowledge'. It is suggested that
it might be the ammocoete larva of an ostracoderm (Newth).

The Heterostraci are actually the oldest known craniates, since their
scales occur in the Ordovician. They were common in the Silurian

IV. 20

HETEROSTRACI

129

and lower Devonian. There were ventral as well as dorsal shields (Fig.
88), and a long series of gill pouches, but only a single pair of exhalent
branchial apertures, suggesting to Watson respiration by a moving
flap (velum). The shields were of cell-less bone (isopedin) covered
with dentine. The body was covered with scales of similar material.
The tail was hypocercal and there were lateral horizontal keels but
no fins. Theie were paired eyes, two semicircular canals and clearly

Fig. 88. Three views of a restoration of Pteraspis.
d.sp. dorsal spine; e. eye; m. mouth; r. rostrum. (From White.)

marked lateral line canals. There was a pineal opening, closed in the
adult, but no sign of the nostril, which may have opened into the
mouth. The latter was surrounded by long plates, suggesting that it
formed a protrusible apparatus, which could be pushed out to form
'a kind of scoop or shovel (Fig. 88) whereby mud and decaying refuse
could be taken off the bottom, for it seems likely that such were their
food and habit' (White).

The coelolepids or thelodonts are the least known group of agna-
thans. The outer surface was covered with fine, placoid-like scales or
hollow spines, which in isolation are often found in late Silurian and
Early Devonian rocks. The anterior end was usually flattened and
wide but the body behind was narrow, with a forked, probably
hypocercal tail. Structures that are probably eye-spots occurred
widely separated near the front margin. The mouth was ventral and
traces of seven branchial arches have been found. There were flap-like

i 3 o VERTEBRATES WITHOUT JAWS iv. 20

extensions on each side of the head but no paired fins. The only
median fin was the anal.

The affinities of these ostracoderm fossils with each other and with
the cyclostomes have been much disputed. Lankester claimed that
pteraspids were related to cephalaspids 'because they are found in the
same beds, because they have a large head shield and because there is
nothing else with which to associate them'. At the other extreme
Stensio holds that we have sufficient evidence to assert that the
pteraspids have given rise to the myxinoids, and the cephalaspids to
the lampreys. Except for the absence of jaws there is indeed little in
common among the fossil forms. The differences in the shape of the
tail are especially baffling. As White points out, an animal with a
heterocercal tail and pectoral fins can hardly have lost either of these
organs independently. He suggests that the earliest vertebrates pos-
sessed straight ('diphycercal') tails and that from these were evolved
on the one hand the pteraspids with hypocercal tails and on the other
the cephalaspids with upturned heterocercal tails. The modern cyclo-
stomes are perhaps derived from the latter, but which, if either, group
gave rise to the earliest gnathostomes is unknown.

The Agnatha were the first animals of the chordate type to become
large, and they apparently all did so by feeding on the detritus at the
bottom of rivers and lakes. They evolved into various types, mostly
rather heavily armoured and perhaps slow-moving forms. The lam-
preys and hag-fishes have been derived from early Agnatha by the
evolution of a sucking mouth, perhaps with loss of the bony skeleton
and paired limbs. However, it was the unknown forms that evolved
a biting mouth that made the next great advance in vertebrate
evolution.

V

THE APPEARANCE OF JAWS.
THE ORGANIZATION OF THE HEAD

1 . The elasmobranchs : introduction

In all parts of the sea there are to be found members of the class of
the elasmobranchs (literally 'plate-gilled' fishes), including sharks
ranging from monsters of 50 ft long to the common dogfish Scylio-
rhinus caniculus of 1-2 ft. Nearly all the fishes in the group are carni-
vorous or scavengers: the skates and rays are bottom-living relatives,
feeding mostly on invertebrates. Although they are not quite so fully
masters of the water as are the bony fishes, they are yet well enough
suited to that element to survive in great numbers in all oceans.
Perhaps the skill and cunnning of a shark is exaggerated by the
frightened boatman or bather, who is apt to mistake a keen nose and
the persistence of hunger for intelligence, especially when he is faced
at intervals with a well-armed mouth; but the sharks have a large
brain and their active, predacious habits enable many of them to
live by eating the more elaborately organized bony fishes.

Evidently such active creatures have changed considerably if they
have been evolved from the heavily armoured and probably slow-
moving agnathous vertebrates that shovelled up food from the bottom
of Palaeozoic seas. It used to be supposed that these elasmobranch or
cartilage fishes represent a very primitive stock, but we now realize
that there have been great changes since the biting mouth was first
evolved; we cannot be sure that any features we find in the elasmo-
branchs were possessed by the earliest gnathostomes.

The typical shark is a long-bodied fish, swimming by the passage of
waves of contraction along tne metamerically arranged muscles. As in
the lampreys and eels, the wave that passes down the body is of short
period, relative to the length of the fish, and is therefore evident as it
travels along. This is probably a less efficient system than is provided
by the longer period waves of the most highlv developed bonv fishes;
the sharks are good swimmers, but except for the mackerel sharks
(Isuridae) not among the swiftest. Stability and control of direction
are ensured by the upturned tail and the fins. The tail, with its dorsal
lobe larger than the ventral, is called heterocercal, and tends to drive
the head downwards. This is corrected by the flattened shape of the
head itself and by the pectoral fins, which act as 'aerofoils', allowing

i 3 2 THE APPEARANCE OF JAWS v. i-

steering in the horizontal plane (p. 140). There are two dorsal fins,
which secure stability against rolling, and also assist in making possible
the vertical turning movements.

The muscles for the production of these movements are a serial
metameric set, with longitudinal fibres, essentially like those of the
lamprey or amphioxus. The central axis is no longer simply a rod;
the notochord has become surrounded and partly replaced by a series

df f s P .by Ld

ac.

bw.

msv. pr b.

Fig. 89. Diagram of the organization of a vertebrate.

ac. wall of abdominal coelom ; b. body wall ; bd. basidorsal ; bv. basiventral ; bw. body wall ; dr. dorsal

rib; i. intestine; iv. interventral; m. myocomma; ms. mesentery; msd. median dorsal septum; msv.

ventral mesentery; nes. neural tube; ns. notochordal sheath; pr. ventral (pleural) rib; sp. neural

spine; ts. horizontal septum. (From Goodrich.)

of vertebrae (Fig. 89). These develop as two pairs of cartilaginous
nodules in each segment, the basidorsals and basiventrals behind, and
smaller elements, the interdorsals and interventrals, in the front. The
basiventrals, lying on either side of the notochord, form the centrum
of each vertebra, invading and almost interrupting the notochord,
which widens again, however, between the vertebrae. The vertebrae
are held together by ligaments, but are not articulated by complex
facets as they are in land animals. The basidorsals form neural arches
above the nerve-cord, and the interdorsals make intercalary arches.
The interventrals partly separate the centra. Attached to each basi-
ventral is a pair of transverse processes, which in the anterior region
bear short ribs and in the tail are fused in the midline to make the
haemal arches.

The median and paired fins are supported by cartilaginous rods,
the radials, and their edges are further strengthened by special horny

SWIMMING OF FISHES

i33

rays, the ceratotrichia. The radials of the paired fins form a series
attached to larger rods at the base. These more basal rods are attached
to a 'girdle' of cartilage embedded in the body wall. The pectoral

Fig. 90. Successive positions of a swimming dogfish at intervals of o-i sec. The
lines are 3 in. apart. The passage of a wave is marked by dots. (After Gray.)

Fig. 91. Successive positions of a swimming eel at intervals of 005 sec. Scale 3 in.
The wave-crests are marked. (After Gray.)

girdle is a hoop extending some way round the body, but the pelvic
girdle is simply a transverse rod in the abdominal wall. The origin of
these girdles and of the fins will be discussed later (p. 136).

2. The swimming of fishes

The propulsive forces that move a fish through the water are usually
produced by the longitudinal muscle-fibres of the myotomes, but in
some forms the propulsion is produced by movement of the fins,
whose function is usually rather to give the fish its stability, enabling it
to keep on a constant course, and also to change its course.

i 34 THE APPEARANCE OF JAWS v. 2

The myotomes consist of blocks of longitudinal muscle-fibres,
placed on either side of an incompressible central axis, the notochord

or vertebral column. The effect of
contraction of the muscle-fibres in any
myotome is therefore to bend the body.
In forward swimming the contraction
of each myotome takes place after that
in front of it. In this way waves of
curvature are passed down the body,
alternately on each side. This can be
illustrated by a series of photographs
of a fish such as the dogfish or eel in
which the amplitude of the waves is
large (Figs. 90 and 91).

In other fishes the waves are not so
immediately obvious, but serial photo-
graphs show that even in such forms as
the mackerel and whiting there is a
backward movement of waves. The
number of waves per minute in steady
swimming varies from 54 in the dog-
fish to 170 in the mackerel, the
corresponding velocities of the waves
being 55 and 77 and of the whole fish
29 and 42-5 cm /sec.

Gray has shown how the muscle
contractions produce movements of
the parts of the body, related to one
another in such a way as to transmit
a backward momentum to the water.
Fig. 92 shows superposed drawings of
an eel, made from successive photo-
graphs. The region marked XY is
moving from right to left and that
X 1 Y 1 from left to right and evidently,
as Gray puts it, 'all parts of the fish's
body which are in transverse motion have their leading surfaces directed
backwards and towards the direction of transverse movement, but
the angle of inclination is most pronounced when the segment is
crossing the axis of longitudinal motion, and at this point the segment
of the body is travelling at its maximum speed. Each point of the body

Fig. 92. Enlarged drawings of suc-
cessive photographs of a young eel
superimposed on each other so that
the tips of the head are on the same
transverse axis and the longitudinal
axes of motion (ab) are made to
coincide. As the wave passes the
section XY it first moves to the left
and is directed backwards and to
the left, whereas X x Y\ moves
in the opposite direction. The tip
of the tail follows a figure of 8.
(From Gray)

v. 2 SWIMMING OF FISHES 135

is travelling along a figure 8 curve relative to a transverse line which is
moving forward at the average forward velocity of the whole fish.
The track of any point on the body (relative to the earth) is a sinu-
soidal curve whose pitch or wave length is less than that of a curve
which defines the body of the fish. There is therefore a definite angle
between the surface of the fish and its path of motion.'

Each portion of the side of the fish can thus be considered as moving
like the blade of an oar used for sculling at the back of a boat. The
principle used, that of an inclined plane, is the same as in screw pro-
pulsion, the essential feature being that the moving surface is inclined
at an angle to its line of motion. The effect of the movement is greatly
increased by the fact that the amplitude of the oscillations grows
passing backwards, as is necessary to produce additive effects in any
coupled system of screws or turbines. The whole fish thus operates
as a single self-propelling system.

The magnitude of the forward thrust thus generated depends
among other things on (a) the angle that the surface of the fish makes
with its own path of motion, (b) the angle between the surface of
the fish and the axis of forward movement of the whole fish, and
(c) the velocity of transverse movement of the body (Gray). These are
evidently factors that will vary with the shape of the body and the
action of its muscles. The body form of the faster-moving types of
bony fishes provides substantial advantages for swimming over that
of the more elongated types. The essential differences are that the
bony fishes have (1) large caudal fins, (2) a much smaller length of the
body relative to its depth, (3) less flexibility.

The role of the large caudal fin is to resist transverse movements;
its effect is, again quoting Gray, 'to keep the leading surface of the
body directed obliquely backwards during both phases of its trans-
verse movements and thereby to exert a steady pressure on the water'.
Since, however, the tail does execute transverse movements, and at
the same time is being rotated towards and away from the axis of
motion, it exerts a very large propulsive effect, probably as much as
40 per cent, of the total thrust.

The effect of the caudal fin, combined with the shortness of body
and reduced flexibility, is that the front part of a bony fish makes only
small transverse movements; the track of the head is therefore nearly
straight and the whole front of the body presents a streamlined
surface with little resistance. Further, the muscles just in front of the
tail exert their tension with very little change in length.

No doubt the shape of the body also has an important influence on

136 THE APPEARANCE OF JAWS v. 2-

the effect of the fish on the water and hence on the turbulence in the
flow of water and the resistance that must be overcome. Gray has
shown that in a dolphin the resistance cannot be that of a rigid model
towed at the speed at which the animal moves, since this would require
that the muscles generate energy at a rate at least seven times greater
than is known in the muscles of other mammals. By watching the
flow of particles past the body of fish-like models he showed that
movements such as those produced in swimming accelerate the water
in the direction of the posterior end, and this would greatly reduce
the turbulence.

Something is known of the nervous mechanism responsible for the
production of the swimming waves. An eel can swim if its whole skin
has been removed. If a region of the body is immobilized by a clamp,
swimming waves can pass along. Therefore the rhythm is determined
by some intrinsic activity of the spinal cord and not by any mechanism
such as proprioceptor impulses arising in active muscles and causing
others to contract.

Experiments in which the spinal cord was cut across show that in
the eel the rhythm is only initiated when suitable impulses reach the
cord either from spinal afferents or from the brain. Thus the spinal
eel can be made to swim either by fixing a clip on to its caudal fin or by
electrical stimulation of the cut end of the spinal cord. Though the
cord requires such afferent stimuli for its functioning, they do not
determine the frequency of the rhythm, which bears no relationship
to that of the applied stimuli.

In the dogfish the isolated spinal cord is able to initiate rhythmic
swimming. After transection behind the brain the posterior portion
of the fish exhibits continuous swimming movements for many days.
Light touch on the sides of the body inhibits these movements, but some
sensory impulses are necessary for their initiation; after complete de-
afferentation, by section of all the dorsal roots, the movements cease.

The information available does not yet enable us to understand
fully how the swimming rhythm is initiated and maintained, nor how
it is influenced by the brain. It would be very interesting to have
further knowledge on these topics, especially because the locomotor
rhythms of land animals are probably based on the serial contractions
of their fish ancestors.

3. Equilibrium of fishes in water ; the functions of the fins

Making use of the methods of investigation of aeronautical engineers,
studies have been made of the forces that operate to keep a fish stable

v. 3 EQUILIBRIUM OF FISHES 137

as it moves through the water, or allow it to become temporarily un-
stable and hence to change direction. Instead of attempting to study a
living or dead fish moving in water, Harris made models and supported
them in a wind-tunnel in an apparatus suitable for measuring the
forces at work in the various directions. Such a method, in which no
compensating movements of the fins are allowed, makes it possible
to investigate the so-called 'static stability' of the fish, that is to say,

Fig. 93. Diagram of model of the dogfish Mustelus, showing the conventional terms
for describing deviations of motion. The longitudinal axis X is that of the wind
tunnel and Y (horizontal) and Z (vertical) are at right angles to it. The arrows show
the directions known as positive rolling, pitching, and yawing, which occur about the
X, Y, and Z axes respectively, a. is the angle of attack between the axis of the model
and the X axis. (From Harris, J. exp. Biol. 13.)

to see whether the body and fins are so shaped as to provide forces
that tend to bring the fish back into its previous line of movement after
it has deviated in any direction. Any body such as a fish or aeroplane
is said to be in stable motion if when it veers slightly from its line of
progress the new forces produced upon its planes tend to restore the
original direction of motion.

The forces acting on the fish are measured along three primary axes,
longitudinal, horizontal, and vertical. Deviation from the line of
motion about the longitudinal axis is known as rolling, about the
transverse axis as pitching, and about the vertical axis as yawing
(Fig. 93). The forces along these three axes are known as drag, lateral
force, and lift.

In order to discover the effect of the median fins and tail on the
stability, these fins were removed, the heterocercal tail being replaced
by a cone having the same taper as the actual caudal fin. The model
was then placed in the wind-tunnel with a wind at 40 m.p.h., which

138 THE APPEARANCE OF JAWS v. 3

corresponds to a motion of 3 m.p.h. in water. The lateral force was
measured when the body was made to yaw at various angles. The

+0-2
+ 0-1

-15 -ipl^^ f

-0

-0-2

+5° +10° +15°

D

+ 5" oX S+Yo° +iV

+ 6

y

+ 05

/

+0-4/

/

/

/

/o-z

+03
*0-1

+5° +10*

+15

-0-1

02

03

-04

-0-5

-0-5

Fig. 94. A. Results of yawing test on model of Mustelus without fins. The lateral force
is plotted as a light full line, drag force as a light broken line; yawing moment about
centre of gravity as full heavy line. Abscissae show the angle of attack in degrees,
ordinates the lateral force and drag in pounds weight, yawing moment in in. -lb. X ^j.
n. Yawing test similar to (a) but with the fins behind the centre of gravity in place.

C. Yawing test with all median fins in place.

D. Pitching test on model of Mustelus with all fins intact and pectoral fins set at an
angle of incidence of 8°. Lift force is shown as a light full line, drag force as a light
broken line, pitching moment about the centre of gravity as a heavy full line.

(From Harris.)

results showed that the equilibrium in this plane is quite unstable; a
slight turn off the direct course would produce a turning moment
tending to increase still further the deflection (Fig. 94). This is a
well-known property of all airship hulls, and is known as the 'unstable
moment' of the hull. It is corrected in the airship by the addition of

v. 3 FUNCTION OF FINS 139

suitable horizontal and vertical fin surfaces at the rear end, when the
airship becomes in effect a feathered arrow. The forces operating on
the fins tend to bring the body back into the original line of motion.
The fins of the fish operate in a similar manner. If the experiment
is performed with a model to which all the fins behind the centre of
gravity have been added, namely, the caudal, anal, and second dorsal
fins, it is found that the curve for the yawing moment now has a steep
negative slope (Fig. 94 b), that is to say, every deviation produces
forces that tend to give directional stability. With the first dorsal fin
also in position the model possesses a remarkable neutral equilibrium
(Fig. 94 c). Deviations by as much as io° produce no resultant yawing
moment about the centre of gravity. The form of the dorsal fins is
therefore definitely such as to maintain stable swimming and prevent
yawing.

Turning of a fish is produced either by the propagation of a wave-
down one side only of the body or by asymmetrical braking with the
pectoral fins (see below). The former type of turn has been investi-
gated by Gray in the whiting, where there is a large caudal fin. This
gives great lateral resistance, so that the first part of the turn is
executed by bending the front part of the fish on the tail as a fulcrum.
This enables the animal to turn through 180 within a circle of the
diameter of its own length. After removal of the caudal fin the turns
are much less effective.

In both elasmobranchs and teleosts the dorsal fins are well developed
in the active swimmers. In most elasmobranchs they are fixed, but in
many teleosts the dorsal fin can be folded up and down, and it is
observed that the fin is raised during turning. This would have the
effect of increasing the yawing moment produced by asymmetrical
action of the body muscles or by unilateral braking with the pectoral
fins.

Since the body is so markedly flexible in the lateral plane and there
are powerful muscles available for turning it in this direction,. the part
played by the fins in determining the stability is important mainly
when the body is held straight. The fish thus has the double advantage
of great stability (by keeping the body straight) and great control-
lability (by bending it). In a body unable to change its shape in this
way, stability and controllability would be inversely related. This is
the case for the stability of the fish in the vertical plane, in which the
body is little flexible. Fig. 94 n shows the positive slope of the curve
for the pitching moment and clearly the equilibrium in this plane is
quite unstable. The pectoral fins contribute more than any others to

Ho THE APPEARANCE OF JAWS v. 3-

movement in this plane, and since they lie in front of the centre of
gravity they greatly increase the instability. The fish must be able to
alter direction in the vertical plane, and it has apparently sacrificed
static stability for controllability. The equilibrium in this plane is a
dynamic one, controlled by the movable pectoral fins, and it is so
unstable that only a small movement of these fins is necessary to
produce a deflecting force that restores the original direction of
motion.

The pectoral fins, lying in front of the centre of gravity, tend to
produce a movement of positive pitch, that is to say, they force the
head upwards. This effect is normally compensated by a component
produced by the heterocercal tail. The upper lobe of this is rigid and
the lower more flexible, therefore the lateral motion given by the
swimming movements of the body produces a vertical lift force on the
tail, giving, of course, negative pitch. After amputation of the hypo-
caudal lobe and anal fin a dogfish swims continually along the bottom
of the tank : in order to compensate for the absence of negative pitch
the pectoral fins are held horizontally and hence there is no moment
to counteract the weight of the fish. If the pectoral fins are then also
removed the anterior end of the body is pointed upwards, often so
much so as to cause the fish to swim with its head out of the water.
This is the result of an over-strenuous attempt to compensate, by
raising the head, for the negative pitch produced by the tail. The
system is no longer suitable for making the continuous adjustments
necessary to ensure stability.

This analysis makes it clear why a heterocercal tail is found in
almost all the primitive swimming chordates; it is almost a necessity
for an animal with a specific gravity in excess of the medium and little
flexibility in the vertical plane. The component of positive pitch
could be provided by the flattened head or by continuous lateral fin
folds, such as may have been present in early fishes, and adjusted
by the limited flexibility possible in the fin. The development of
movable pectoral fins confers much greater control. Since the useful
portions of a fin fold for this purpose would be those well in front of
and behind the centre of gravity, we can perhaps see the reason why
the intervening portion has become lost. In the modern sharks the
pelvic fins have little influence on the stability and are perhaps retained
only for their modification as claspers.

It is not surprising that races of fishes with stability ensured by
systems of this sort should tend to adopt a bottom-living habit, with
dorso-ventral flattening, such as is found in the skates and rays.

v. 4

SKIN OF ELASMOBRANCHS

141

Expansion of the front end is developed at first to compensate the
effect of the tail, but the pectoral fin becomes expanded to allow ver-
tical adjustments and then reduction of the hypocaudal lobe of the tail
accompanies the adoption of life on the sea bottom. Eventually all

Fig. 95. Development of denticles in the dogfish.

A and b, first gathering of odontoblasts (sc.) below the basement membrane (brn.); ml. are the

epidermal cells that will become modified, c, first deposition of dentine (d.). In D there is

more dentine and a pulp cavity (p.) is seen. In E are shown stages in the formation of enamel

(e.) and of the basal plate (bp.) while the denticle cuts the epidermis (ep.).

(From Goodrich, Vertebrata, A. & C. Black, Ltd.)

locomotion is produced by undulatory movements of the fins, which
were at first used only to raise the fish off the bottom.

4. Skin of elasmobranchs

Being swift and predatory animals, more attackers than attacked,
the sharks do not possess a very heavy external armament. The skin
itself is tough, being covered by layers of epidermis. Beneath this is a
thick dermis of connective tissue with fibres arranged at right angles

142 ORGANIZATION OF THE HEAD v. 4-

as in a carpet, giving a tissue of great strength and flexibility, able to
maintain the shape of the body. Scattered over the skin are the charac-
teristic denticles or placoid scales (Fig. 95). Each of these consists of a
pulp cavity, around the edge of which lies a layer of odontoblasts
secreting the calcareous matter of the scale, known as dentine. This
has a characteristic structure resulting from the fact that the odonto-
blasts send fine processes throughout its substance. The outside of the
dentine is covered by a layer of enamel, secreted by the overlying
ectoderm. Usually the denticles pierce through the ectoderm, after
which no further enamel can be added to their surface. Obviously the
scales are similar to teeth, which are indeed to be considered as
specialized denticles developed on the skin of the jaws. It has often
been supposed that the denticle is the primitive type of fish scale,
from which others have been derived, but it now seems more likely
that the earliest covering was a continuous layer, later broken into
large scales, from which the denticle was ultimately derived (p. 269).
The skin also gives protection to the fish by its colour, produced
by a layer of chromatophores beneath the epidermis. Many sharks
have a spotted or wavy pattern, which breaks up their visible outline
as they move in the water, especially near the surface. They are able
to change their colour, though only slowly, becoming darker on a dark
background (see p. 164).

5. The skull and branchial arches

In general organization a dogfish follows closely the fish plan,
which we have already considered. Most of its special new features
are in the head, and we may now turn to a consideration of the
organization of the head and jaws of a gnathostome vertebrate. The
jawless vertebrates of the Silurian and Devonian included fresh-
water animals of various sorts, but the vertebrate type began to
flourish and increase more abundantly with the appearance of creatures
with jaws in the late Silurian. From this stage onwards we have to
follow the parallel history of numerous orders and families, as the
vertebrate plan of structure became adapted for various habitats. It
seems likely that the development of a biting mouth greatly increased
the range of possibilities of vertebrate life. The most obvious use of
a mouth is for attacking other animals, but it may also have been used
to collect plant food from all sorts of situations where it would not
be available to the microphagous or shovelling Agnatha. Probably the
mouth was also early used for defence, and in this way influenced the
whole bodily organization, making unnecessary the heavy armature

v. 5 DEVELOPMENT OF THE JAWS 143

that is so characteristic of many early vertebrates. Modern research
has shown that the armour has become progressively reduced along
various lines of iish evolution. Older ideas of comparative anatomy
regarded the 'cartilage fishes' as showing a primitive stage, preceding
the appearance of bone. We now realize that this is the opposite of the
truth and that the dogfish and its relatives represent a higher type,

rosC

Fig. 96. Skull and branchial arches of the dogfish (Scyliorliimts).
au.c. auditory capsule; b.b. basibranchial; b.li. basihyal; c. centrum; cer.b. ceratobranchials;
cer.h. ceratohyal; d.r. foramen for dorsal root; e.b. extrabranchials; e.c.f. external carotid
foramen; e.l. ethmoid ligament; ep.b. epibranchials; gr. groove for anterior cardinal sinus;
g.r. gill rays; hy.a. foramen for hyoid artery; hymd. hyomandibula; i.d. interdorsal; io.c.
interorbital canal; I.e. labial cartilages; M.c. Meckel's cartilage; na. neural arch; nas.c. nasal
capsule; o.n.f. orbito-nasal foramen; op. foramen for ophthalmic nerve; op.g. groove for
op.V; op.V, op. VII ophthalmic branches of V and VII; orb. orbit; ph.b. pharyngobranchials ;
p.sp.l. prespiracular ligament; r. rib; rost. rostral cartilages; spd. supradorsals ; tr. transverse
process; vr. foramen for ventral root; II-IX, foramina for cranial nerves. (After Borradaile.)

able to defend themselves by mobility, by biting, and by efficient
sensory and nervous organization. Heavy defensive armour is a
primitive form of protection for animals, as for man.

Besides its use in feeding and defence, the mouth can also be used as
a means of 'handling' the environment, for instance in the nest-build-
ing activities of many fishes. Indeed, it is difficult for us to realize the
utility of the jaws for an animal not provided with any other means of
seizing hold of objects.

The development of the mouth to a point at which it could be used
in these varied ways was, therefore, a very important stage in evolution.
Recognition of the Gnathostomata as a separate group of animals is
far more than a matter of classificatory convenience, it marks the
achievement of the possibility of life in a greatly increased range of
environments.

144 ORGANIZATION OF THE HEAD v. 5-

Morphological analysis enables us to see how this biting mouth
was produced, by modification of one or more of the gill-slits. The
main differences that separate the gnathostome from cyclostome ver-
tebrates are therefore in the head and its skeleton. Although the
modern elasmobranchs show the skull and jaws in a modified and
reduced condition, they provide by their simplicity a good starting-
point for discussion. The 'skull' of a dogfish consists of a series of

ac.

pan

--dao

Fig. 97. Diagram of skull of selachian embryo before fusion of the main cartilages;
cranial nerves black, numbered; arteries cross-lined.

ac. auditory capsule; bra. epibranchial; dao. dorsal aorta; eps. efferent pseudobranchial;

ha. efferent hyoid; hv. hypophysial vein; i.e. internal carotid; nc. nasal cartilage; oc. orbital

cartilage; oca. occipital arch; op. optic; oph. ophthalmic; or. orbital; pan. pila antotica; pf.

profundus nerve; pch. parachordal; poc. polar cartilage; tr. trabecula. (From Goodrich.)

cartilaginous boxes surrounding the brain and receptor organs (Fig.
96). The nasal capsules, orbital ridges, and auditory capsules are
largely fused with the main cranium, producing a single continuous
structure, the chondrocranium. It is interesting to consider how this
structure has arisen during the process of cephalization. Presumably
parts of it represent the modified sclerotomes of trunk regions. We
shall see presently that there is strong evidence that the head has arisen
by modification of a segmental arrangement such as is seen in the
trunk; the morphogenetic processes that build the skull must there-
fore be related in some way to those of the vertebrae. The first rudi-
ment of the skull in the embryo consists of two pairs of cartilaginous
rods, the parachordals and trabeculae (Fig. 97). The former lie on
either side of the notochord, the trabeculae in front of the notochord.
These first rods fuse up to make a continuous plate; from this grow
sides and roof, completing the cartilaginous neuro-cranium around the
brain. Meanwhile cartilaginous capsules form around the nose, eyes,
and ears, and become joined to the neuro-cranium. Posteriorly, behind

v. 6 BRANCHIAL ARCHES 145

the auditory capsules, the cranium is completed by the addition of a
number of segmented elements, evidently modified vertebrae.

The problem is, therefore, to determine the nature of the pro-otic
part of the skull. Before we can settle this we must consider the
visceral or branchial arches. These are pairs of rods of cartilage
developed in the walls of the mouth and pharynx, between the gill-
slits. In the dogfish each typical branchial arch (Fig. 96) consists of a
series of four pieces, the pharyngo-, epi-, cerato-, and hypo-branchials.
Ventrally some of the arches join a median basibranchial plate. These
rods lie in the pharynx wall and on their outer sides carry a series of
projecting rods, the branchial rays and extrabranchial cartilages,
whose function is to support the lamellae of the gills.

There are five such branchial arches, differing only slightly from
each other. In front of these lie two arches, the hyoid and mandibular,
which, though modified, are obviously of the same series. The hyoid
the more nearly resembles a typical branchial arch. Its most dorsal
element, the hyomandibular cartilage, is a thick rod attached dorsally
to the skull by ligaments and at its lower end forming the support
for the hind end of the jaw. It apparently corresponds to the epi-
branchials. The more ventral elements, cerato- and basihyal, resemble
the corresponding members of more posterior arches. The jaws them-
selves (mandibular arches) depart more widely from the form of a
typical branchial arch, but the two thick rods of which each is com-
posed, the upper palato-pterygo-quadrate bar and the lower Meckel's
cartilage, are recognizably members of the branchial series. Looking
at the whole apparatus with a thought to the embryological processes
that have produced it, with as it were a manufacturer's eye, we can
see at once that the jaws and hyoid arch have been produced by a
modification of the processes that make the branchial arches.

6. The jaws

Study of the serial relationship of the jaws and branchial arches
gives us an understanding of the course of evolution of the mouth.
We may suppose that the ancestors of the gnathostomes possessed a
nearly terminal mouth, either on the front end of the body or on the
ventral surface. The pharynx was pierced by a series of gill pouches,
beginning shortly behind the mouth and separated by arches, each
containing a set of cartilaginous bars (Fig. 96). There is some evidence
that this condition persisted in the cephalaspids (p. 125), where there
is found to be a series of ten pairs of gill-slits, beginning far forward
on either side of the mouth. The muscles moving the more anterior

146 ORGANIZATION OF THE HEAD v. 6

parts of the pharynx wall and the anterior arches could be called into
play to help in the collection of food. In this way the mouth came to
be used for prehension, and the grasping jaws of the gnathostomes
appeared as the more anterior arches became modified to allow more
efficient seizing, and the skin over them was modified to form the
teeth. The mouth probably shifted backwards during this process and
its lateral edges joined the first gill-slit. The rods supporting the
posterior wall of that slit thus became bent over into the characteristic
position of the vertebrate jaws.

There is some uncertainty as to the means of support of the jaws
in the earlier stages of their evolution. The front end of the palato-
pterygo-quadrate bar is attached to the cranium in the dogfish by the
ethmo-palatine iigament'. In most elasmobranchs the hind end of the
upper jaw is not fixed to the cranium but is slung from the latter by
the hyomandibula and by a prespiracular ligament. This means of
support, known as hyostylic, was for long supposed to have been the
original one. But the earliest gnathostomes (the acanthodians) do not
have this arrangement (p. 187), indeed, their hyoid arch is an almost
typical branchial arch, not modified to support the jaw. In the primi-
tive condition one would not expect the hyoid arch to have any con-
nexion with the mandibular. In the acanthodians the jaw is supported
by direct attachments to the cranium at its hind as well as front end,
a condition known as autodiastylic.

The early elasmobranchs themselves do not have a hyostylic jaw
support, but an arrangement in which the upper jaw is both attached
to the cranium and also supported by the hyomandibula. This
amphistylic condition persists to-day in the primitive shark Hexan-
chus. Apparently the jaws, which at first swung from the skull, later
became fixed at the hind end to the hyoid, and this finally became the
only means of support posteriorly. The advantage of this last arrange-
ment is presumably that it allows a wide gape for swallowing, the prey
whole. As the sharks sought to eat larger and larger fishes, those in
which the hind end of the upper jaw was less firmly fixed to the skull
were the more successful and so the hyostylic condition was achieved.

If this theory of the origin of the jaws is correct we may expect to
find some trace of a cartilaginous support for the side wall of the
pharynx in front of the original first gill-slit, a premandibular visceral
arch. Many sharks have two pairs of labial cartilages in this position,
which have been held to represent arches. However, there are strong
grounds for believing that this is represented by the trabeculae cranii,
the rods lying on each side in front of the parachordals and contribut-

v. 6 ORIGIN OF THE JAWS 147

ing to the floor of the skull (Fig. 98). Many points indicate that these
rods are not part of the axial skeleton. The main axis of the body
presumably ends at the front end of the notochord, that is to say, at
the level of the front ends of the parachordals. Indeed there is much
confirmatory evidence to show that this level represents the end of the

^ 12.2m R £

gsX rn.a.'ka'.'*br.a.l" 9 s - 9
Yt ¥2,3.

gut

FlG. uS. Diagrams to show the condition of the visceral arches and
jaws in early vertebrates.
A. cephalaspid; B. acanthodian; C. elasmobranch. (i.e. auditory capsule;
"br. a. 1" first branchial arch; c.n.c. nerve-cord; e. eye; gs. I. first gill-slit;
//. hypophysis. /i.a. hyomandibular arch; m. a. mandibular arch; m. mouth; not.
notochord; p.m.a. premandibular arch (trabecula); sp. spiracle; Vi, pro-
fundus nerve; V2, 3, trigeminal nerve; VII, facial; IX, glossopharyngeal;
X, vagus. (Modified after Westoll.)

segmented part of the body, everything in front of this level being as
it were pushed forward from above or below. The trabeculae have
exactly the relations to the most anterior nerves and blood-vessels
that would be expected of visceral arches. Confirmation of the theory
comes from the discovery that the cartilage of the front part of the
trabeculae, like that of the visceral arches, is formed by material
streaming down from the neural crest, that is to say, from ectoderm.
The branchial arches, hyoid, jaws, and trabeculae thus all constitute
a single series, the result of the working of a repetitive or rhythmic
process, appropriately modified at each level.

148 ORGANIZATION OF THE HEAD v. 7-

7. Segmentation of the vertebrate head

The rhythmicity or metamerism seen in the cartilages can be traced
throughout the structure of the head. Although in higher vertebrates
the head appears as a distinct structure, separated from the body by
a neck, yet there is every reason to think that it has arrived at that
state by gradual modification of the anterior members of an originally
complete metameric series. The jaws, the receptor-organs, and the
brain have become developed at the front end of the body, producing
what zoologists conveniently if pretentiously call cephalization.

The fundamental segmentation of the head is not very easily appar-
ent to superficial observation; the working out of its details is an
excellent exercise in morphological understanding. Recognition of the
segmental value of the various structures also makes them the more
easily remembered. For instance, the nerves found in the head have
been named and numbered for centuries by anatomists in an arbitrary
series:

I. Olfactorius
II. Opticus

III. Oculomotorius

IV. Trochlearis (patheticus)
V. Trigeminus

VI. Abducens
VII. Facialis
VIII. Acousticus
IX. Glossopharyngeus
X. Vagus
XI. Accessorius
XII. Hypoglossus

Morphological study has shown that these nerves are not isolated
structures, each developed independently, but that they represent a
regular series of segmental dorsal and ventral roots of the head
somites. The satisfaction and simplification given by this generaliza-
tion is one of the clearest advantages of morphological insight. More
important still, such understanding of the morphology of a structure
shows us how to look for the morphogenetic processes that produce it;
such knowledge of how organs are made is an essential step in mending
or remaking them.

The idea of the essential similarity of structure of the head and
trunk was early developed by Goethe, who tried to show that the
mammalian skull is a series of modified vertebrae. Unfortunately this

v. 8

PRO-OTIC SOMITES

149

view cannot be maintained in detail and the theory was brought to
ridicule by T. H. Huxley and others. The segmental value of the skull
floor and sides is not at all easy to determine; the parachordals arise
as a pair of unsegmented rods on either side of the notochord.

ma: ha

Fig. 99. Diagram of the segmentation of the head of a dogfish.

cr. limit of neurocranium; vr. limit of visceral arch skeleton; a. auditory nerve; aa.i, pre-
occipital arch; aa.2, occipital arch; ab. abducens nerve; ac. auditory capsule; ah. anterior
head cavity; c. coelom;/. facial nerve; gl. glossopharyngeal nerve; ha. hyoid arch; hm. hypo-
glossal muscles; hy. hypoglossal nerve; la. pila antotica; m. mouth; m.2-6, myomere 2-6;
ma. mandibular arch; tnb. muscle-bud; nc. nasal capsule; om. oculomotor nerve; prf. pro-
fundus nerve; scl. sclerotome of segment 10; sp.1-2, ganglion of spiral nerve 1-2; t. trochlear
nerve; tr. trigeminal nerve; v. vagus nerve; vgl. vestigial ganglion of segment 7; vc. ventral
coelom; vr. ventral root of segment 6. (From Goodrich.)

8. The pro-otic somites and eye-muscles

Ideas about the segmentation of the head were first correctly for-
mulated by F. Balfour. In his studies of the development of elasmo-
branchs (1875) he showed that three myotomes, the pro-otic somites,
can be recognized during development in front of the auditory capsule
(Fig. 99). The auditory sac, pushing inwards and becoming sur-
rounded by cartilage, then breaks the series of myotomes, so that
several are missing in the adult, though the series is complete in the
embryo.

If this analysis is correct we should be able to recognize that the
nerves of the head belong to a series of dorsal and ventral roots,
similar to that in the trunk, the ventral roots being those for the
myotomes and the dorsal roots, running between the myotomes,
carrying sensory fibres for the segment and motor-fibres for any non-
myotomal musculature present (p. 36). In the spinal region the

i5o ORGANIZATION OF THE HEAD v. 8

dorsal and ventral roots join, but this is not the primitive condition
(witness amphioxus and the lampreys), and in the head region the
earlier state of affairs is retained, the dorsal and ventral roots remain
separate. Presumably the arrangement we find in the head today was
laid down in very early times, in the Silurian period or earlier, when
the dorsal and ventral roots were still separate. The head, in spite of
its specializations, preserves for us a relic of that ancient condition.

The branchial nerves, such as the glossopharyngeal, show clear
signs of this condition. Each has a small pre-trematic branch in front
of the slit, a larger post-trematic branch behind it, and a pharyngeal
branch to the wall of the pharynx. The pre-trematic branch usually
contains mostly sensory fibres from the skin, the pharyngeal branch
visceral sensory fibres, including those from taste buds. The post-
trematic branch contains both motor and sensory fibres. In addition
to these more ventral branches the branchial nerves also usually
provide dorsal rami to the skin of the back.

The three pro-otic somites become completely taken up in the
formation of the six extrinsic muscles of the eye, arranged similarly
in all gnathostome vertebrates. The four recti roll the eye straight
upwards, downwards, forwards, or backwards, and the two obliques,
lying farther forward, turn it, as their name suggests, upward or
downward and forward (Fig. ioo). Of these muscles the superior,
anterior, and inferior rectus and inferior oblique are all derived from
the first myotome and are innervated by the oculomotor (third cranial)
nerve. The superior oblique, innervated by the trochlear nerve (fourth
cranial), is the derivative of the second and the posterior rectus
(external rectus of man), innervated by the abducens (sixth cranial),
of the third somite. These three nerves are evidently the ventral roots
of the three pro-otic somites. At some early stage of vertebrate
evolution all the myotomal musculature of the front part of the head
became devoted to the movement of the eyes. The muscles originally
forming part of the swimming series became attached to a cup-like
outgrowth from the brain.

Most of the rest of the musculature of the head, including that of
jaws and branchial arches, is derived from the somatopleure wall of the
coelom and is therefore lateral plate or visceral musculature. This
lateral plate muscle is indeed better developed in the head than in
the trunk, where all the muscles, even of the more ventral parts of the
body, are formed by downward tongues from the myotomes. The
lateral plate origin of the jaw-muscles at once gives us the clue to
the nature of some more of the cranial nerves, the fifth, seventh, ninth,

V.

THE ORBIT

151

and tenth. These nerves all carry ganglia containing the cell bodies
of sensory fibres and these are comparable to the spinal dorsal root

' op prof

olf.b.

cbl sip.

c'd.a.

Fig. 100. Orbit of the dogfish.

ant. reel, anterior rectus; cereb. cerebellum; cil.a., cil.p. anterior and posterior ciliary nerves;
epa. anterior carotid artery; ep. epiphysis; g.cil. ciliary ganglion; inf. obi. inferior oblique;
lam. lamina terminalis of cerebrum; O.V. and VII, superficial ophthalmic branch of trigeminal
and facial; obi. sup. superior oblique; olf.b. olfactory bulb; opt. I. optic lobe; post.rect. posterior
rectus; r. op. prof, ramus ophthalmicus profundus of trigeminal; rs. sensory root of ciliary
ganglion; sup.rect. superior rectus; thai, thalamus; // to A', cranial nerves.
(After Young, Quart. J. Micr. Sci. 75.)

ganglia. But the nerves also transmit motor-fibres to the muscles of
the jaws and branchial arches. They are in fact mixed roots, just as
we have seen that the primitive dorsal roots should be, carrying the
sensory fibres for the segment and motor-fibres for the non-myotomal
muscles (p. 36).

152 ORGANIZATION OF THE HEAD v. 8-

9. The cranial nerves of elasmobranchs

These nerves are more easily studied in elasmobranchs than in any
other vertebrates, because of the relatively soft and transparent
cartilage through which they run. We may therefore take this oppor-
tunity to examine the whole series of cranial nerves in some detail in
the dogfish, beginning with the oculomotor nerve, the first ventral
root. Examination after removal of the brain will show clearly in any
vertebrate, including man, that this nerve arises from the ventral
surface, at the level of the hind end of the midbrain (optic lobes).
This is not true of the trochlear or pathetic nerve, which emerges
from the dorso-lateral surface of the brain but nevertheless is the
ventral root of the second segment. Its cells of origin lie close behind
those of the oculomotor nerve, in the ventral part of the brain. The
reason for the dorsal emergence is that the muscle lies dorsally and the
nerve has been modified so as to reach its muscle by running partly
within the tissues of the brain. The third ventral root (abducens),
which is very short, is clearly ventral.

In looking for the dorsal roots that correspond to these three
segments we have to examine the trigeminal, facial, and auditory
nerves. The trigeminal of the dogfish, like that of man, has ophthalmic,
maxillary, and mandibular branches (Fig. ioo), but can be shown to
represent the dorsal roots of the two first segments. The ophthalmic
branch is a sensory nerve carrying fibres for skin sensation from the
snout. The maxillary branch supplies sensory fibres to the upper jaw,
whereas the mandibular is a mixed nerve to the skin and muscles of
the lower jaw. Besides these main branches there is also a small but
important sensory branch from the trigeminal to the eyeball (Fig.
ioo). This joins a motor root from the oculomotor nerve where the
latter swells slightly to form a ciliary ganglion. Two ciliary nerves
then carry motor and sensory fibres to the eyeball. In some specimens
a branch of the more anterior ciliary nerve leaves the eyeball anteriorly,
runs between the oblique muscles, and out of the orbit again to end
in the skin of the snout (Fig. ioo). Though this branch is small and
inconstant in the dogfish, its course corresponds exactly with that
of a much larger nerve in the related shark Mustelus and in skates and
rays. In these animals there are two ophthalmic branches of the
trigeminal nerve; one, having a course similar to that of the main nerve
in the dogfish, is the ramus ophthalmicus superficialis; the second,
the ramus ophthalmicus profundus, runs across within the orbit, gives
off the long ciliary nerve to the eyeball, passes between the oblique

v. 9 PARTS OF THE TRIGEMINAL 153

muscles, and leaves the orbit for the skin of the snout. In higher
vertebrates the nasociliary nerve and the long ciliary, innervating the
eyeball, represent the profundus, while the rest of the ophthalmic
nerve of mammals corresponds to the superficial ophthalmic of
elasmobranchs.

The relations of these nerves to other structures shows that the
so-called trigeminal nerve really includes the dorsal roots of two
segments combined. The profundus can be traced back in develop-
ment to a nerve that is obviously the dorsal root of the first somite,
of which the oculomotor nerve is the ventral root. Indeed it may be
noticed that the profundus and oculomotor partly join, at the ciliary
ganglion. The ramus ophthalmicus profundus and the oculomotor
nerve thus constitute the dorsal and ventral roots of the first or pre-
mandibular somite, whose corresponding branchial arch is presumably
the trabecula cranii (p. 147). The dorsal root does not show the full
structure of a branchial nerve, presumably because there is no gill-slit.
The profundus represents only the dorsal branch of a typical branchial
nerve, innervating the skin.

The ramus ophthalmicus superficialis, and the maxillary and mandi-
bular branches together constitute the dorsal root of the second pro-
otic somite, whose ventral root is the trochlear nerve. The correspond-
ing gill arch is the mandibular (palato-pterygo-quadrate bar and
Meckel's cartilage), whose gill-slit we have suggested has been in-
corporated with the edge of the mouth. The trigeminal nerve shows
considerable similarity to a branchial nerve, its maxillary branch repre-
sents the pre-trematic and the mandibular the post-trematic ramus,
while the ophthalmicus superficialis is the dorsal branch to the skin.
There is no pharyngeal branch. An anomalous feature of the trige-
minal is that it contains sensory fibres whose cells of origin lie within
the brain (mesencephalic root). These fibres are probably proprio-
ceptors from the masticatory muscles and eye-muscles. The latter run
from the eye-muscle nerves to join the trigeminal.

The dorsal root of the third segment, whose ventral root is the
abducens, includes the whole of the facial and also the auditory nerve.
The facial is a large mixed nerve in the dogfish. Its ophthalmic branch
runs to the snout, carrying mainly fibres for the organs of the lateral
line system that lie there. A large buccal branch supplies sensory
fibres to the mouth and a palatine branch joins the trigeminal. A small
prespiracular branch carries sensory fibres from the skin in front of the
spiracle, and the main portion of the nerve continues behind the
spiracle as the hyomandibular nerve, dividing up into motor branches

i 5 4 ORGANIZATION OF THE HEAD v. 9

for muscles of the hyoid arch and sensory ones for the skin of that
region.

This nerve is obviously the branchial nerve to the spiracle; we can
safely say that the facial and abducens are the dorsal and ventral roots
of the third or hyoid segment. The auditory nerve is included as part
of the dorsal root of the third somite because the auditory sac is
formed by sinking in of a portion of the ectoderm within the territory
of the facial nerve. The labyrinth still communicates with the surface
of the head in the adult dogfish by a canal, the aquaeductus vestibuli.
The nerve that innervates the auditory sac, whatever complexities it
may acquire, is to be regarded morphologically as a portion of the
dorsal root of the hyoid segment.

The segmental nature of the structures in the pro-otic region can
therefore be made out without serious difficulty. The disturbance
introduced by the auditory capsule makes the segmental arrangement
of the more posterior region of the head somewhat confused. The
series of dorsal roots is uninterrupted; the ninth (glossopharyngeal)
nerve is the dorsal root of the fourth segment of the series and runs
out through the cartilage of the auditory capsule. The dorsal roots of
the succeeding segments are then fused to form that very puzzling
nerve the vagus. The branches it sends to the gills are clearly typical
branchial nerves, but why should they all come off together from the
medulla oblongata, and if there is any advantage in this union, why
is the ninth nerve not also so incorporated? Above all, why does
the vagus send two branches far outside the segments of its origin, the
lateral line branch carrying fibres to the organs right to the tip of the
tail and the visceral branch fibres to the heart, stomach, and probably
small intestine?

Evidently these 'wanderings', from which the vagus gets its name,
began very long ago. The nerve reaches as far back in cyclostomes as
in any other vertebrates. It is easy to understand that if visceral
functions are to be directed from the medulla oblongata there is an
advantage in having sensory impulses sent direct to that region of the
brain and motor impulses sent out direct to the viscera. It may be
that these advantages allowed the centralization of these visceral
functions, while the need for serial contraction of the swimming
muscles led to the retention of the segmental arrangement of the
spinal cord. It is an interesting thought that but for the swimming
habits of our ancestors our nervous system might by now consist of a
central ganglion with nerves passing from it direct to all the organs.
Indeed we are tending in that direction, as the spinal cord shortens

v. 9 VAGUS NERVE 155

and becomes more and more nearly a simple pathway between the
brain and the periphery.

However this may be, the vagus is certainly a nerve compounded
of the dorsal roots of several segments and it is a mixed nerve, con-
taining both receptor and motor fibres. Some of the more posterior
rootlets of this series are separated off in higher animals (not the dog-
fish) to form the eleventh cranial nerve, the accessorius or spinal
accessory, which in mammals sends motor-fibres to certain muscles
of the neck, the sternomastoid, and part of the trapezius. Its motor
nature has led some to suppose that this nerve is a ventral root, but
these muscles are derived from lateral plate musculature and the
accessorius represents the motor portion of the hinder dorsal roots
of the vagus series.

The ventral roots of this post-otic region have become much
reduced. Several myotomes are always missing completely, so that
there are no ventral roots corresponding to the glossopharyngeal and
first three or four vagal segments. The more anterior of the surviving
post-otic somites are to be found not in the dorsal region but ventrally,
as the hypoglossal musculature of the tongue. The muscle-buds have
grown round into this portion behind the gill-slits, and the nerve
(hypoglossal) that innervates them represents the ventral roots of the
more posterior segments of the vagus-accessorius series (Fig. 104).
The origin of this nerve from the floor of the medulla is a clear sign
that it is a ventral root.

Thus the entire series of cranial nerves is :

Segment

Arch

Dorsal root

Ventral root

Pre-mandibular

Trabecula

R. op. profundus V

Oculomotorius III

Mandibular

Palato - pterygo - quad-

Rr. op. superficialis,

Trochlearis IV

rate bar and Meckel's

maxillaris, and man-

cartilage

dibularis V

Hyoid

Hyoid

Facialis VII
Acousticus VIII

Abduccns VI

1st Branchial

1 st Branchial

Glossopharyngeus IX

(absent)

2nd Branchial

2nd Branchial^

Vagus X

3rd Branchial

3rd Branchial 1

+

Hypoglossals XII

4th Branchial

4th Branchial j

Accessorius XI

5th Branchial

5th Branchial J

Two cranial nerves have not yet been considered, the first, olfactory,
and second, optic. Our thesis is that all connexions between centre
and periphery are made by means of a segmental series of dorsal and
ventral roots and therefore these nerves, too, should be fitted into the
series. No embryological or other studies have enabled this to be done

156 ORGANIZATION OF THE HEAD v. 9-

and the reason in the case of the optic nerve is quite clear. It is not
morphologically a peripheral nerve at all. The eye is formed as a
vesicle attached to the brain ; the optic 'nerve' therefore develops as a
bundle of fibres joining two portions of the central nervous system; in
fact it is now usually called the optic tract, not the optic nerve.

This reasoning will not apply to that very peculiar and interesting
structure the olfactory nerve. This is unique among all craniate
nerves in consisting of bundles of fibres whose cell bodies lie at the
periphery. The cells of the olfactory epithelium, like the sensory cells
in invertebrates and some of those of amphioxus, are neurosensory
cells, that is to say, their inner ends are prolonged to make the actual
nerve-fibres that pass into the brain. This fact does not by itself solve
the problem of fitting the nerve into the series of dorsal and ventral
roots, but it reminds us that the nerve is very ancient, and suggests
that it does not fall into the rhythm of the rest of the series because it
precedes the other cranial nerves either in time or space, or perhaps
even both. The olfactory nerve may have existed before any seg-
mental structure appeared, possibly as the nerve of sense-organs on
the front end of the ciliated larva which we suppose gave rise to our
stock (p. 76). Alternatively we can say that the olfactory nerve is as
it is because it lies in front of the region over which the segmenting
process operates; it is, as it were, 'prostomial'. If we wish we can hold
both these views together.

There are one or two other exceptions to the rhythmic arrangement
of nerves, perhaps more difficult to account for than the first and
second cranial nerves. If all connexions between centre and periphery
are made by dorsal and ventral roots what is the status of the fibres
that run down the infundibular stalk to reach the cells of the pituitary
body? This glandular tissue, derived from the epithelium of the
hypophysial folding of the roof of the mouth, is undoubtedly a peri-
pheral organ. Does it receive its nerve-fibres direct from the brain ?
If so presumably we must say that the pituitary, like' the nose,
is prostomial, lying in front of the segmental region, and this is
reasonable enough from its position. There is good reason to believe
that it is an extremely ancient organ, already present in the earliest
chordates.

A still more puzzling exception is the nervus terminalis. This is
a small bundle leaving the brain ventrally behind and below the olfac-
tory nerves and running to the olfactory mucosa or to the accessory
olfactory organ of Jacobson, where this is present (p. 350). In some
vertebrates it carries a small ganglion. The fibres are probably

v. io RESPIRATION 157

afferents and they run backwards through the brain tissue to the
pre-optic nucleus of the hypothalamus. A possible clue to its origin is
that this is the region of the brain where the morphologically ventral
region of the neuraxis ends (p. 147). The nervus terminalis may repre-
sent the ventral olfactory nerve, the much larger main nerve being
morphologically dorsal.

A further puzzle of some importance which may be mentioned here
is the course of the proprioceptor fibres for those muscles that are
supplied purely by ventral roots. The eye-muscles contain proprio-
ceptor organs and Sherrington and others have shown that the affer-
ent fibres connected with these run to the brain through the third,
fourth, and sixth nerves, that is to say, through ventral roots. Simi-
larly, it has been shown that there are afferent fibres in the hypo-
glossal nerves in mammals. Conversely it is now known that there are
efferent fibres running from the brain to many receptor organs. For
example, such fibres run in the auditory nerve. To pursue these ques
tions farther would lead us into discussion of the factors that control
the making of connexions within the nervous system Here we are
concerned only with analysis of the plan that produces the main out-
lines of the structures in the head, a plan which, with all its modifica-
tions, is essentially segmental.

10. Respiration

The function of the branchial arches is not merely to support the
gills but to allow the movements of the pharynx wall by which the
respiratory current of water is produced. It is for this reason that
the jointed system of rods is present. The respiratory movements con-
sist in a lowering of the floor of the mouth by means of the hypo-
glossal muscles, with at the same time an expansion of the walls of
the pharynx. This causes an inrush of water through the mouth,
which is then closed and the floor raised, forcing the water out through
the gill-slits. The whole movement is worked by the 'visceral' (lateral
plate) muscles of the pharynx wall, innervated by the trigeminal,
facial, glossopharyngeal, and vagus nerves, in co-operation with the
myotomal hypoglossal muscles, innervated by the hypoglossal nerve.

The gill filaments bear lamellae that meet at the tips, leaving minute
channels for the water. The blood flows through the lamellae in the
opposite direction to the water so that just before leaving the gills
the blood meets the highest concentration of oxygen and lowest of
carbonic acid.

158

ORGANIZATION OF THE HEAD

1 1 . The gut of elasmobranchs

The digestive system of sharks shows several changes from the plan
found in lampreys, especially the presence of a true stomach, charac-
teristic of all gnathostomes. Apparently little or no digestion goes on
in the mouth and pharynx. The teeth consist of rows of backwardly
directed denticles (Fig. ioi). They are carried on special folds of skin
lining the jaws and are continually replaced as they are worn away on

Fig. ioi. Sections through the jaws of a, dogfish (Scyliorhinus), and n, sand-shark

(Odontaspis), showing the transitions between dermal denticles (d) and teeth (t).

(From Norman, partly after Gegenbaur.)

the edge. The replacement of milk by permanent teeth in mammals
is a relic of such serial replacement in a fish. The 'gill rakers' arc
rods attached to the branchial cartilages and serving to prevent
the escape of prey. The basihyal supports a short non-protrusible
tongue.

The wall of the pharynx is lined by a stratified epithelium on to
which open numerous mucous glands, sometimes complex. The mucus
serves to assist the passage of the food, but probably has no strictly
digestive function, though the salivary glands of higher vertebrates
no doubt originate from a modification of these mucous glands.

The pharynx narrows to an oesophagus with thick muscular walls,
leading without sharp transition to the stomach. We have seen that in
cyclostomes the oesophagus opens directly into the region of gut that
receives the bile and pancreatic secretion. The stomach, which we

v. 12 DIGESTION IN ELASMOBRANCHS 159

now meet for the first time, has probably been formed as a special
portion of the oesophagus. Barrington suggests that it evolved with the
jaws, serving originally as a receptacle for the large pieces of food, or
even whole fishes, which could now be swallowed. The mucous glands
became modified to produce acid, since this prevents bacterial decay.
Finally, an enzyme, pepsin, was evolved able to digest proteins in
acid solution. In the dogfish this condition has been fully established
and the stomach has essentially the structure and functions found in
all higher vertebrates. However, in the gastric glands only one type
of cell is recognizable, there are no separate pepsin-secreting and
acid-producing cells. Nevertheless, there is a pepsin-like enzyme
present and the contents are acid. The stomach is divided into two
parts, a descending cardiac and ascending pyloric limb, the signifi-
cance of the divisions being unknown.

The region where the stomach joins the intestine is guarded by a
powerful pyloric sphincter, immediately beyond which open the bile
and pancreatic ducts. The liver is a large two-lobed organ, receiving
the hepatic portal blood from the gut. It serves as a storage organ
containing much glycogen and fat and sometimes the hydrocarbon
squalene. It probably also plays a part in the destruction of red blood
corpuscles. Bile is carried away to a gall-bladder, from which a bile-
duct leads to open at the front end of the spiral intestine.

The pancreas, hardly recognizable as a distinct organ in the lamprey,
forms in the dogfish an elongated body between the stomach and
intestine. It contains both exocrine and endocrine cells and its duct
enters the intestine shortly below the pylorus. The 'small' intestine
of elasmobranchs is of a peculiar form, being short but with its
surface greatly increased by the presence of a spiral ridge or 'valve'.
The intestinal contents are alkaline and contain trypsin, amylase,
and lipase. There is no constant fauna of commensal bacteria. Ab-
sorption presumably takes place wholly in this organ, for the remain-
ing length of gut consists only of a short rectum, to which is attached
an organ of unknown function, the rectal gland, containing branched
glands and much lymphoid tissue.

12. The circulatory system

The heart develops as a specialization of the subintestinal vessel
between the place where it receives the veins from the liver and the
body wall and the gills, which are to be supplied under high pressure.
It consists of a single series of three main chambers, sinus venosus,

i6o

ORGANIZATION OF THE HEAD

atrium, and ventricle, all of which are muscular, and there is also a
muscular base to the ventral aorta, the conus arteriosus, provided with
valves (Fig. 102).

The five afferent branchial arteries carry blood to the gill lamellae,
whence it is collected by a system of four efferents and connecting
vessels into a median dorsal aorta, carrying blood to all parts of the
body. Oxygenated blood is supplied to the head from three sources.
(1) From the top of the first gill a carotid artery leaves the efferent
branchial and runs forwards and towards the midline: it then divides

f. eF pef

Fig. 102. Diagram of the branchial circulation of an elasmobranch fish.

aa. median anterior prolongation of aorta; ac. anterior carotid; afa. afferent vessel of spiracu-
lar gill; aef. afferent vessel from last hemibranch; af. anterior efferent vessel; a/. 2-6 , five
afferent vessels from ventral aorta; afa. afferent artery of spiracular gill; c. conus leading to
ventral aorta; cl. coeliac artery; d. ductus Cuvieri; da. dorsal aorta; ef. epibranchial artery;
h. a. hyoid afferent vessel; hp. hepatic veins; ht. heart; pc. posterior carotid; pef. posterior

efferent vessel; s. spiracle; va. ventral artery; I-V, branchial slits.

(From Goodrich, Vertebrata, A. & C. Black, Ltd., after Parker.)

into an external carotid to the upper jaw and internal carotid to the
brain. (2) The dorsal aorta divides at its front end into branches,
which join the carotids before their division. (3) From the vessel
that collects blood from the first gill arises a hyoidean artery, carry-
ing oxygenated blood to the spiracle. From here the hyoidean artery
runs on as the anterior carotid (Fig. 102) across the floor of the orbit
to join the internal carotid within the brain-case.

The heart is supplied by a cardiac artery arising from the dorsal
aorta behind the gills. The blood-pressure in the ventral aorta is 30-40
mm Hg and there is a drop of 10-20 mm Hg across the gills. The
circulation is slow, with a mean circulation time as low as 2 minutes.
The venous return of the fishes is ensured by a system of very large
sinuses. The pericardium is almost completely enclosed in a carti-
laginous framework by the basibranchial plate above and pectoral
girdle below it. It may be that this produces a negative pressure in

v. iz VASCULAR SYSTEM OF ELASMOBRANCHS 161

the veins. There is a passage of unknown function, the pericardio-
peritoneal canal, leading from the pericardium to the abdominal
coelom and the hinder end of this is very narrow.

A caudal sinus from the tail opens into a renal portal system above
the kidneys. From the latter, and from the muscles of the back, blood is
collected into the pair of very large posterior cardinal sinuses, lying on
the dorsal wall of the coelom. Above the heart these receive the open-
ings of other large sinuses, such as the anterior cardinal sinus, running
above the gills and collecting blood from the head by way of an orbital
sinus, and the jugular, lateral cardinal, subclavian, and other sinuses from
the body wall. Blood then passes round the oesophagus in the two
ductus Cuvieri into the sinus venosus, where hepatic sinuses also open.

The resistance offered by a vessel to flow within it decreases with
approximately the fourth power of the diameter, therefore the large
size of these vessels substantially assists in allowing return to the
heart. The heart-muscles, like any others, require antagonists; they
can contract in one direction only, and each chamber therefore needs
to be actively dilated. It will be noted that the fish heart consists of
a series of three muscular chambers, presumably because the low
venous pressure is able to dilate only a chamber with very thin walls,
such as the sinus venosus. Contraction of the sinus then inflates the
auricle, and the auricle inflates the ventricle, which thus constitutes
the third step in this serial pressure-raising system. In the land animals,
where most of the blood only passes through a single set of capillaries,
a two-step system (auricle and ventricle) is sufficient for each part of
the circulation.

Little is known of the control of the circulation but it is probably
less effective than in higher animals. There is a cardiac branch from
the vagus ending in an elaborate plexus in the sinus venosus (Fig. 104).
Stimulation of this nerve slows the heart. There is no anatomical or
physiological evidence of a sympathetic nerve to the heart, but
abundant sympathetic fibres run to the arteries. Small doses of adren-
aline cause prolonged rise of blood-pressure.

There are receptors in the efferent branchial vessels and in the post-
branchial plexus above the cardinal veins (p. 175). Nerve impulses
from these receptors can be recorded in the vagus at each systole and
are increased by raising the blood-pressure. Their reflex effects are
to slow the heart and respiration and decrease the blood-pressure,
perhaps for protection of the gill capillaries. These reflexes are pre-
sumably the ancestors of the carotid sinus and similar reflexes of land
vertebrates.

162 ORGANIZATION OF THE HEAD v. 13

13. Urinogenital system

The blood of the elasmobranchs differs from that of all other verte-
brates in its very high content of urea. As measured by the depression
of the freezing-point the blood is isotonic with the surrounding sea
water (say, 3-5 per cent. NaCl); it may even be slightly hypertonic.
But there is far less salt in the blood than in the sea, in fact only
about 1-7 per cent. NaCl. Although the blood is nearly isotonic with
the sea its composition is therefore regulated (homeosmotic). This
arrangement is apparently a legacy of the fact that the ancestors of the
elasmobranchs were originally fresh- water animals (p. 187). The
return passage to the sea has been accomplished by the elasmobranchs
through the device of urea retention. The gill surfaces, in which alone
the blood comes into close contact with sea water, are not permeable
to urea, but this substance penetrates freely into the tissues, as it does
in other animals. Elasmobranch tissues if placed in sea water are
therefore in contact with a strongly hypertonic medium. They are so
habituated to the presence of urea that they are unable to function
unless it is present in a concentration that would be toxic to most
animals.

This arrangement has presumably been responsible for the fact
that few of the elasmobranchs have returned to fresh water. In the
case of the saw-fish Pristis, which lives some hundreds of miles
up the Mississippi and various rivers in China, Smith found that a
considerable concentration of urea is still maintained in the blood,
thus further increasing the work that these fishes, like any fresh-water
animal, must do in order to maintain an osmotic concentration above
that of the surrounding water. One shark (Carcharhinus nicaraguensis)
and some rays, Trygon, also live in fresh water.

In the ordinary marine elasmobranchs the high urea concentration
is maintained by the presence of a special urea-absorbing section of
the kidney tubules. The urinary apparatus is a mesonephros and
these fishes show a considerable specialization in that the urinary
functions of this organ are separated from its generative ones in the
male. The hinder part of the kidney (sometimes called opisthone-
phros, the term metanephros should be used only for the definitive
kidney of amniotes, which has a different method of development)
consists of a mass of tubules ending in very large glomeruli, and a
section of each tubule has the function of urea absorption. All the
tubules join to form a series of five urinary ducts and these enter a
urinary sinus, opening to the cloaca. The sinus can be compared

v. 13 URINOGENITALS OF ELASMOBRANCHS 163

functionally with a bladder, but it is a mesodermal structure, derived
from the main kidney duct, and not strictly comparable to the endo-
dermal bladder of tetrapods. The urinary sinus is a small organ and
the volume of liquid excreted is small.

Fie. 103. A. Urinogenital system of the female, B, of the male dogfish.
ah.p. abdominal pores; cl. cloaca; cp. claspers of the male; F. rudiment of the oviducal opening
in the male; Md. urinary ducts; mtn. hinder (excretory) part of mesonephros; od. oviduct;
oe. cut end of oesophagus; of. oviducal funnel; og. oviducal gland; ov. ovary; P.f. pelvic fins;
R. rectum; s.s. sperm-sacs; T. testis; up. urinary papilla in the female; ugp. urogenital papilla
in the male; vs. urinary sinus; vc. vasa efferentia; vs. vesicula seminalis; WD. Wolffian duct;
Wg. Wolffian gland or mesonephros. (After Bourne, from Goodrich, Vertebrata, A. & C.

Black, Ltd.)

The genital system is highly specialized to allow internal fertiliza-
tion and the production of a few very yolkv and well-protected eggs.
There is a single large ovary, from which the eggs are carried by the
cilia of the peritoneum to a pair of funnels lying on either side of the
liver behind the heart (Fig. 103). These are apparently formed from
proncphric funnels and the Mullerian duct (oviduct) separates from
the original nephric duct. In the adult it becomes a thick-walled
muscular tube, bearing a swelling, the nidamental gland, the upper
part of which produces albumen, the lower the horny egg case.

1 64 ORGANIZATION OF THE HEAD v. 13-

The testes are paired and sperms are collected at their front ends
by vasa efferentia leading into the anterior or reproductive portion of
the mesonephros. This consists of a much coiled, thick-walled, vas
deferens, whose glands produce material that aggregates the sperms
into spermatophores. The vas expands into a broader ampulla
(seminal vesicle), which at its lower end gives off a forwardly directed
blind diverticulum, the sperm sac, developmentally the lower end of
the Miillerian duct, reduced of course in the male; small funnels are
still visible at the upper end.

Transmission of the sperms is produced by a large and complicated
pair of claspers. These are modified parts of the pelvic fins of the
male, developed into scroll-like organs and containing a pumping
mechanism and erectile tissue; they are inserted into the female
cloaca. The mechanism of erection is operated by nerves and may
involve the liberation of adrenaline; experimental injection of that
substance will produce erection, and it is perhaps significant that the
male possesses a reserve of adrenaline-producing tissue (see p. 167).

Development of elasmobranchs is by partial cleavage, producing a
blastoderm, perched on the top of a large mass of yolk. The egg is
protected by an elaborate egg-case, the 'mermaid's purse', within
which development proceeds until the yolk has been used up.

In several elasmobranchs development is viviparous, the oviduct
forming a 'uterus'. In Mustelus there is a yolk-sac placenta, but in
Trygon 'uterine milk' is secreted into the embryo by villi (trophone-
mata) inserted through the spiracles.

14. Endocrine glands of elasmobranchs

Elasmobranch fishes possess the full complement of endocrine
glands, but these show some interesting differences from those of
higher vertebrates. The pituitary body lies in the usual place below
the diencephalon and anterior, intermediate, tuberal, and neural
divisions can be recognized. Little is known of the functions of the
gland. The gonads of the dogfish retrogress after removal of the pitui-
tary. Little or no vasopressin or oxytocin is present. The neuro-
intermediate lobe contains a substance that produces the expansion
of melanophores. Hogben and Waring have also produced evidence
that the pars anterior produces a substance causing contraction of the
melanophores, but this has not yet been isolated and the evidence for
its existence is indirect.

The thyroid is formed by a downgrowth from the floor of the
pharynx, to which it often remains attached by a narrow stalk con-

v. i 4

ADRENAL TISSUE

165

sp n

Fig. 104. Dissection of suprarenal bodies and sympathetic nervous system of the

dogfish.
b.a. brachial artery; c.a. coeliac artery; d.a. dorsal aorta; g. first sympathetic ganglion; hyp.
hypoglossal nerve; I.e. longitudinal sympathetic 'chain'; n.card. cardiac branch of vagus;
p.b.p. post-branchial plexus; r.c. ramus communicans; s. sensory fibre; s.a. segmental artery;
sg. sympathetic ganglion; sp.a. anterior splanchnic nerve; sp.m. middle splanchnic nerve;
spa. spinal nerve; sup. suprarenal body; v. ventricle; v.d. vas deferens; vs. vago-sympathetic
anastomosis; X. vagus; X.br. branchial branch of vagus; X.visc. visceral branch of vagus.
(After Young, Quart. J. Micr. Sci. 75.)

taining a small ciliated pit, a reminder that the organ was once a
ciliated mucus-secreting gland.

The adrenal tissue is especially interesting because the two parts,
so closely associated in mammals, are here found widely separated. A
segmental series of glands, the suprarenals, are rich in noradrenaline.
They project into the dorsal wall of the posterior cardinal sinus and can
be seen when it is opened (Fig. 104). The more anterior ones are
fused to form an elongated structure on either side of the oesophagus.

(i 66)

n v

Fig. 105. Diagram of transverse section through hind region of mesonephros of

dogfish.

cv. cardinal vein; da. dorsal aorta; int. interrenal body; k. mesonephros; I c. longitudinal

sympathetic connective; n c. sympathetic nerve-cells scattered in suprarenal body; nv.

sympathetic nerves; re. ramus communicans; sg. sympathetic ganglion; sv. subcardinal vein;

sup. suprarenal body. (From Young, Quart. J. Micr. Sci. 75.)

Fig. 106. Diagram of arrangement in hinder mesonephric region of dogfish.
Lettering as Fig. 105. tr. transverse sympathetic nerve. (From Young, Quart. J. Micr. Sci. 75.)

v. i 4 -i5 BRAIN OF ELASMOBRANCHS 167

The sympathetic ganglia are closely associated with these suprarenal
bodies, as would be expected from their common origin from cells
of the neural crest. The segmental series continues along the whole
length of the abdomen, the more posterior members being embedded
in the kidney tissue (Fig. 105). These posterior suprarenal bodies are
larger in the male than in the female, but only the central part of the
male glands shows the reaction with chrome salts that indicates the
presence of adrenaline. The peripheral portion of each gland appears
to consist of non-functioning cells, possibly a reserve used only during
reproduction (see p. 164).

The part of the adrenal corresponding to the cortex of mammals is
represented in elasmobranchs by the interrenal bodies, lying medially
in some species, paired in others, in the kidney region (Figs. 105 and
106). The cells of these organs resemble cortical adrenal cells. Since
they are not in contact with the suprarenals at any point, it would seem
that the association of the two parts is not necessary for their function-
ing, at least in these animals. Removal of the interrenal is always fatal.
The gland is stimulated by 'stress' or by mammalian ACTH. Extracts
of it prolong the life of adrenalectomized rats. There is evidence that
it influences carbohydrate metabolism and activity of the gonads but
not electrolyte balance.

The islets of Langerhans contain two cell types as in mammals.
The pineal body is small and without any trace of eye-like structure.

The gonads contain endocrine organs, producing steroid hormones.
These are formed by interstitial cells in the testes. Oestrogens probably
come from the outer (theca) cells of the follicles that surround the
eggs. The inner (granulosa) cells of the capsule assist in yolk produc-
tion but may also produce progesterone and in viviparous species
they develop into a distinct corpus luteum after ovulation.

15. Nervous system

The brain is large and well developed in elasmobranchs, having a
structure characteristically different from that of both the cyclostomes
and bony fishes (Fig. 100). The forebrain is large and has cerebral
hemispheres thickened both in floor and roof, whereas in teleosts the
roof is thin. The hemispheres are wide relative to their length and the
end of the unpaired portion of the forebrain between the hemispheres,
the lamina terminalis, is also much thickened. Attached to the ends
of the cerebral hemispheres are large olfactory bulbs and there are
also large nasal sacs. Evidently the olfactory sense is well developed
in these animals and they depend greatly on it for hunting.

1 68

ORGANIZATION OF THE HEAD

v. 15

All parts of the cerbral hemispheres receive fibres from the olfac-
tory bulbs and the forebrain serves mainly for analysing the olfactory
impulses. However, it is stated that there are fibres reaching forward
to one area at the back of the roof of the hemispheres from other
centres. Johnston therefore called this region the 'general somatic
area' and suggested that it represents the beginnings of that develop-
ment so characteristic of mammals by which all the senses are centred
on the cerebral hemispheres. Further work is needed to confirm the

Dec interhemisph

., Tr medianus

Ventr o/f

Tr olFacb.

Tr olFacb. epistr cruc.

Fiss Urn tel.

Ventrical, /at.

Fig. 107. A cross-section through the forebrain of a shark.

Dec. interhemisph. decussatio interhemispherica; Fiss.lim.tel. fissura limitans telencephali;

Prim. hip. primordium hippocampi; 5. septum; Striat. striatum; Tr. medianus. tractus

medianus; Tr.oljact. tractus olfactorius; Tr.olf act. epistr. cruc. tractus olfacto-epistriaticus

cruciatus; Ventricul.lat. ventriculus lateralis; Ventr. o.f. ventriculus olfactorius.

(From Kappers, Huber, and Crosby.)

existence of this pathway, and even if present its significance must
not be exaggerated. There is of course no cortical arrangement of
tissue in the hemispheres. The cells form thick masses around the
ventricle (Fig. 107). The roof is quite thick and contains decussating
fibres in the midline. The sides and floor make up the main bulk of the
organ, the lateral wall being known as the striatum, its upper part the
epistriatum. The medial wall is known as the septum and its upper
portion is often referred to as the primordium hippocampi, having a
position similar to that of the hippocampus of mammals. The main
efferent pathways are tracts leading to the hypothalamus and to the
optic lobes. After removal of the forebrain the sense of smell is lost
but the fish shows no obvious disturbance of posture, locomotion, or
behaviour.

The diencephalon is a narrow band of tissue, there are no extensive
tracts leading forward through it, and the optic and other pathways
do not end here as they do in higher animals. The lower part of the

v. 15 BRAIN OF ELASMOBRANCHS 169

between-brain, the hypothalamus, is, however, as well developed
(relatively) in these animals as in mammals. Its hind part (inferior
lobes) receives olfactory impulses via the forebrain (the 'fornix' of
higher vertebrates) and gustatory pathways from the medulla. Its
efferent fibres run to reticular centres. The more anterior part of the
hypothalamus lies above the pituitary and contains the supraoptic
nucleus, whose axons form the hypophysial tract, ending in the inter-
mediate lobe. The supraoptic cells of all vertebrates are large and
contain granules of neurosecretory material that is probably passed
down the axons and liberated in the pituitary. The anterior hypo-
thalamus is a higher centre for visceral control, regulating, for
example, circulation, respiration, and many metabolic activities.
Attached to the hind end of the hypothalamus of fishes is a peculiar
organ, the saccus vasculosus, with folded, pigmented walls. It has
been suggested that this acts as a pressure receptor, since it is well
developed in deep-sea fishes. It is one of the characteristic features
that the sharks and bony fishes have in common.

The midbrain, as in cyclostomes and teleosteans, is very large and is
perhaps the dominant centre of the brain. The optic tracts end in its
roof (tectum opticum) after complete decussation below the brain.
The cells of the tectum are arranged in a complicated pattern of
layers. Other sensory centres that send tracts to the optic lobes are
the olfactory (cerebral hemispheres), acustico-lateral, cerebellar,
gustatory, and probably also the general cutaneous centres of the
spinal cord. Efferent tracts leave the midbrain roof to the base of
the midbrain and extend backwards into the medulla, perhaps into the
spinal cord. The efferent midbrain fibres have direct influence on
the spinal cord, and electrical stimulation of points on the tectum
opticum produces various movements of the fins, suggesting a system
of control similar to that exercised over spinal centres by the cerebral
cortex of mammals through the pyramidal tract. Various forced
movements follow injury to the midbrain.

The cerebellum is a very large organ in clasmobranchs, as in all
animals that move freely in space. Its main source of sensory fibres is
from the ear and from the organs of the lateral line system, whose
afferent fibres enter through the seventh, ninth, and tenth cranial
nerves. The internal structure of the cerebellum is very uniform and
essentially similar in all vertebrates. Removal of portions of it from
dogfishes produces aberrations of swimming.

The medulla oblongata is the region from which most of the cranial
nerves spring and especially those that regulate the respiration and

i7o ORGANIZATION OF THE HEAD v. 15-

visceral functions. In mammals this control is indirect, but in fishes
the nerves that spring from the medulla directly innervate the
respiratory muscles of the gills and floor of the mouth. It is no doubt
for this reason that the centre for the initiation of the respiratory
rhythm developed in the medulla.

16. Receptor-organs of elasmobranchs

The paired nasal sacs have much-folded walls. Water enters by a
single opening but this may be partly divided by a fold, making a
groove, which may open to the mouth. There are taste-buds scattered
over the wall of the pharynx. It has been shown experimentally that,
as in higher animals, these are receptors for sampling the food after
it has been brought close to the animal, whereas the nose acts as a
distance receptor. Smell and taste are therefore different senses for a
dogfish, as for us. By training fishes to discriminate between various
substances it can be shown that those that we should smell are
detected by the nose in the dogfish, but its organs of taste, like ours,
can discriminate only between a few qualities, including salt, sour,
and bitter.

The eyes are well developed in sharks and no doubt serve as an
important means of finding the prey and avoiding enemies. However,
the retina usually contains only rods, and visual discrimination is
probably poor, but there are cones in Mustelus and Myliobatis. Unfor-
tunately details as to the functional performance of the eyes, ability to
discriminate shapes, &c, are scanty. Behind the retina there is often a
reflecting layer, the tapetum lucidum. This may be provided with
pigment cells, which expand in the light but contract in darkness,
allowing the underlying guanophores to reflect, thus increasing
sensitivity. The lens is spherical and very hard, as in all fishes, since
it must perform the whole work of refraction. It is provided with a
protractor-lentis muscle, presumed to produce active accommodation
for near vision by swinging the lens forward. The iris is peculiar in
those elasmobranchs that hunt by day; when it narrows it divides
the pupil into two slits by the descent of an upper flap or operculum.
The muscles of the iris are better developed in elasmobranchs than
in most bony fishes and the pupil makes wide excursions. The
sphincter iridis muscle, which narrows the pupil, works as an inde-
pendent effector. It is stimulated to contract by light, but its move-
ments are not controlled by any nervous mechanism. The radial
dilatator fibres, which open the pupil, receive motor-fibres from the
oculomotor nerve. The closure of the iris when illuminated is

v. i6 RECEPTORS OF ELASMOBRANCHS 171

relatively slow. If the whole eye is cut out from the head, in the dark,
the sphincter, being an independent effector, still closes when illumi-
nated. The muscle, being without nerves, is not affected by any of the
usual drugs that mimic action of the autonomic nervous system,
though some of these affect the innervated dilatator muscle. We have
therefore the curious situation that no 'autonomic' drugs applied to
the isolated dark adapted eye cause closure of the pupil ; this can only
be produced by illumination (Fig. 108).

The ear of elasmobranchs contains receptors concerned (1) with
maintenance of muscle tone, (2) with angular accelerations, (3) with

Red

Red

White

Adrenaline j//00,000

Minutes

Fig. 108. Movements of margin of pupil of an isolated iris of the shark Mustelus,
followed by plotting with a camera lucida and here shown magnified 53 X . Addition
of adrenaline causes slight dilation of the already dilated pupil and illumination then
causes closure. Acetyl choline even in concentrations of 1 in 10,000 has a similar
dilatory effect. (From Young, Proc. Roy. Soc. B. 112.)

gravity, (4) perhaps with hearing. There are three pairs of semi-
circular canals, each with an ampulla containing receptor cells, whose
hairs are embedded in a gelatinous cupula. This behaves as a highly
damped torsion pendulum, swinging with movement of the fluid.
These receptors discharge impulses continuously and during angular
rotations the frequency is either increased or decreased in the appro-
priate ampullae, initiating compensatory movements of the eyes and
fins.

The otolith organs include three patches of receptor cells in par-
tially distinct sacs, the utricle, saccule, and lagena. The endolym-
phatic duct is an open canal and in some species serves to admit sand
grains, which are attached to the maculae as gravity receptors. The
utricle seems to be the main receptor producing appropriate postures
in relation to gravity. The lagena shows a maximum discharge rate
near the normal position of the head and thus serves as an 'into level'
receptor. The areas of these maculae that carry otoliths do not
respond to vibrational stimuli but carry only gravitational receptors.
Vibration responses in the form of nerve impulses have been seen
in rays but only up to 120 c/sec, although vestibular microphonics
up to 750 c/sec occur. At high intensity there is much synchronization

172 ORGANIZATION OF THE HEAD v. 16-

of units. These results suggest that the ear may function as a vibration
receptor, but there are no conditioning experiments to show whether
these fishes can hear.

There is a well-developed system of lateral line organs, whose
function is considered later (p. 218). The organs of this system on the
head are highly modified in elasmobranchs to form the ampullae of
Lorenzini, long canals filled with mucus. Sand showed that these

Fig. 109. Drawing of a sympathetic ganglion and related structures in a dogfish.
Lettering as in Figs. 104 and 105. s.o. sense-organs. (After Young, Quart. J. Alter. Sci. 75.)

organs increase their discharge of nerve impulses with very slight
falls of temperature, and he suggested that their function is to detect
such changes. They are also sensitive to weak tactile stimulation and to
small voltage gradients in the water. Their function therefore remains
uncertain. It may be related to determining changes of hydrodynamic
pressure distribution over the surface of the aerofoil-like body,
especially in skates and rays. They may thus act as mechano-receptors
detecting local changes of pressure near the body surface.

No doubt elasmobranchs, like other animals, have many senses
referred to the skin, such as we call touch, pain, and the like, but few
studies of these exist. Sand has shown the presence of volleys of
impulses in the nerves connected with muscles when the latter are
stretched. Proprioceptors have been demonstrated histologically in
the muscles of Raja. This agrees with the fact that after severance of
the spinal cord the swimming rhythm only continues if some afferent
nerves are intact.

v. 17

SYMPATHETIC OF ELASMOBRANCHS

173

17. Autonomic nervous system

The sympathetic system of elasmobranchs consists of an irregular
series of ganglia, approximately segmental, lying dorsal to the pos-
terior cardinal sinus and ex-
tending back above the kidneys.
These ganglia contain motor
nerve-cells (post - ganglionic
cells) whose ascons end in the
smooth muscles either of the
arterial walls or of the viscera.
The cells themselves are con-
trolled by pre-ganglionic nerve-
fibres whose cell bodies lie in
the spinal cord and whose pro-
cesses run out in the ventral
spinal roots and rami com-
municantes (Fig. 109). In
higher animals the sympa-
thetic ganglia send postgangli-
onic fibres back to the spinal
nerves for distribution to the
skin ('grey rami communi-
cantes') but these are absent in
elasmobranchs and correspon-
dingly there is no evidence of
sympathetic control of skin
functions (e.g. chromato-
phores); a very different con-
dition is found in bony fishes
(p. 222). Another peculiarity
of the sympathetic system of
elasmobranchs is that it does
not extend into the head. This
condition is unique among
vertebrates, but it is not clear
whether it is primary or the
result of a secondary loss.

In mammals it is usual to
recognize a parasympathetic
system acting in antagonism to

symp

Fig. i 10. Diagram of the autonomic nervous

system of the dogfish.
art. artery; card.n. cardiac nerve; cil.g. ciliary gang-
lion; ft. heart; in. intestine; k. mesonephros; ov.
oviduct; ph. pharynx; pr. profundus nerve; py.
pylorus; st. stomach; symp. sympathetic ganglion
(with suprarenal near it); u.s. urinogenital sinus;
III, V, VII, IX, X, cranial nerves. (From Young,

Quart. J. Micr. Sci. 75.)

i 7 4 ORGANIZATION OF THE HEAD v. 17

the sympathetic, but this is not easy to define in the elasmobranchs
(Fig. no). The vagus, it is true, is well developed, with branches
to the heart and gut, but little is known of autonomic fibres in the
other cranial nerves, or of a special 'sacral' parasympathetic system.
Stimulation of either the vagus or the sympathetic nerves causes
contraction of the stomach. A ciliary ganglion connected with the
oculomotor nerve is present as in other animals, but there is no
sense in which it can be called antagonistic to the sympathetic
system, since the latter does not extend into the head. The post-
branchial plexus is a network of fibres and cells connected with the
vagus but stretching back above the posterior cardinal sinus
(Fig. 104). Receptors in this plexus and in the afferent branchials
(Fig. 109) may be concerned with vascular reflexes (p. 161).

VI

EVOLUTION AND ADAPTIVE RADIATION OF
ELASMOBRANCHS

1 . Characteristics of elasmobranchs

The organization of a shark used to be considered to show the earlier
stages of fish evolution, but we have seen evidence that this is a mis-
take (p. 131). The sharks and skates and rays are highly developed
creatures; in particular, the absence of bone is a secondary feature;
they have been able to give up their defensive armour because of the
development of other means of protection, swift swimming, good
sense-organs and brain, and powerful jaws. We can now examine the
history of these changes and study the varied creatures that can be
classified as elasmobranchs. As usual in examining such histories we
must try to discover evidence about the forces that have operated to
produce the changes of type, and look for signs of any consistent
trends, persisting for long periods of years.

2. Classification

Superclass Gnathostomata
Class Elasmobranchii ( = Chondrichthyes)
Subclass 1. Selachii
*Order 1. Cladoselachii. Devonian-Permian

*CIadoselache; *Goodrichia
*Order 2. Pleuracanthodii. Devonian-Trias

*Pleur acanthus
Order 3. Protoselachii. Devonian-Recent

*Hybodiis; Hetcrodontus
Order 4. Euselachii. Jurassic-Recent
Suborder 1. Pleurotremata. Jurassic-Recent
Division 1. Notidanoidea. Jurassic-Recent
Hexanchus ; Clilamydoselache
Division 2. Galeoidea. Jurassic-Recent

Scyliorhinus; Mustelus; Cetorhinus; Carcharodon
Division 3. Squaloidea. Jurassic-Recent
Squalus; Sqaatina; Pristiophorus ; Alopias
Suborder 2. Hypotremata. Jurassic-Recent
Raja; Rhinobatis; Pristis; Torpedo; Trygon

176 EVOLUTION OF ELASMOBRANCHS vi. 2-3

Superclass Gnathostomata (cont.)

Subclass 2. Bradyodonti. Devonian-Recent
*Order 1. Eubradyodonti. Devonian-Permian

*HeIodus
Order 2. Holocephali. Jurassic-Recent

Chimaera

The elasmobranchs form a very compact group of fishes, nearly
always marine and of predaceous habit, having a great quantity of
urea in the blood, with no bone in the skeleton, no operculum over
the gills, and no air-bladder. The tail is usually heterocercal. The
pectoral fin is anterior to the pelvic and the latter is usually provided
with claspers, fertilization being internal. The body is more or less
completely covered with placoid scales (denticles) and these are
specialized in the mouth to form rows of teeth. The intestine is short
and provided with a spiral valve. The typical cartilage-fishes with
these characters may be placed in the subclass Selachii, to distinguish
them from an early aberrant offshoot the Bradyodonti, represented
today by the peculiar creature Chimaera (p. 184).

3. Palaeozoic elasmobranchs

The selachians are among the most numerous of the various pre-
datory animals in the sea. There have, however, been many side-
branches of the main shark line and we may now survey the history
of the group from its first appearance. The characters we have used
in our definition mark the elasmobranchs off from the earliest-known
gnathostomes, the acanthodians and other placoderm types (Fig. 1 1 1),
which we shall consider later (p. 186). Presumably the elasmobranchs
were derived from some placoderm, but the earliest evidence of the
existence of true sharks is in the form of isolated teeth and scales from
Middle Devonian deposits, and the earliest type about which full
information exists is *Diade?nodus from the Upper Devonian, 'an
early and not distant offshoot from the primitive Chondrichthyan
stock, the main line of which led through *Ctenacanthus and the
hybodonts to the modern elasmobranchs ; *Cladoselache is a specialized
side-line of this main stock and is not an appropriate ancestral type
for the Chondrichthyes' (Harris). The teeth of *Diademodus are
many-cusped and resemble the scales more closely in sculpturing
than in other primitive sharks. The jaw suspension was amphistylic
and the notochord unconstricted. The pectoral fin was continuous
posteriorly with the body wall and there was no well-developed

(i77)

Cenozoic

Cretaceous

100-

Jurassic

^5

Triassic

Perm/an
200

Carbonifcrou.

, t ha

,. .„, Torpedo-

II. HIP

AcMtnedu ^V^vxi L I f)

Silurian

S. *

DieiroleP'i

Ordoi/iaan

400-

Cambrian

FlG. hi. The early evolution of vertebrates.

i 7 8

EVOLUTION OF ELASMOBRANCHS

VI. 3

pectoral girdle. The tail was heterocercal and there are no signs of skele-
tal support for lateral keels. All of these Harris regards as primitive
features; *Diademodus was specialized in having no spines in front
of the dorsal fin and no clasper on the head. Both of these features are
frequent in hybodonts and in *Cladoselache there is a large spine

zd nsi

Fig. 112. Development of the fins of the dogfish, i, Adult showing the nerve-supply
of the fins ; 2, adult with the fins shown expanded and their nerves and muscles shown
as if concentration had not taken place; 3, a 19-mm. embryo, showing the actual

condition.

a. anal fin; ac. anterior collector nerve of first dorsal fin; cr. (black) cartilaginous radial

partially hidden by the radial muscle; n. 1-57, spinal nerves and ganglia; pc. collector

nerve of second dorsal fin; pi. pelvic fin; pt. pectoral fin; rm. radial muscle; id. and 2d.

first and second dorsal fins. (From Goodrich, Vertebrata, A. & C. Black, Ltd.)

in front of the first dorsal fin. These fishes were thus like modern
sharks in their general form, but the fins were remarkable in having
a broad base, not sharply marked off from the body- wall. It has been
suggested by Goodrich and others that this was the earliest condition
of the pectoral fin, perhaps showing its derivation from a continuous
or extended fin-fold (Fig. 113). This theory has the advantage that it
agrees with the embryological development of the fin by concentration
of a series of segments (Fig. 112). It also seems likely that anterior
and posterior fins expanded in the horizontal plane would be neces-
sary for stabilization (p. 136). Moreover, this theory of the origin of
paired fins has the great advantage that it compares them with the
median fins, which are also continuous folds. It has been argued,

vi. 3 PALAEOZOIC ELASMOBRANCHS 179

however, that the cladoselachians are very far from the earliest known
fishes and that in both ostracoderms (p. 125) and placoderms (p. 186)
fins are known that have a narrowly constricted base. We cannot
yet say for certain what has been the course of evolution of the paired
fins, but the fin-fold theory has much plausibility, in spite of the
difficulties raised by palaeontologists.

Fig. 113. Pectoral fins of various fishes.
a, *Cladoselache\ b, *Pleur acanthus; c, Ncoccratodus; d, Gadits. (From Norman.)

The cladoselachians represent the ancestral Devonian sharks, from
which all later forms have been derived. Animals of similar type
were fairly common in late Devonian and Carboniferous seas. The
ctenacanths, such as *Goodrichia, reached a length of 8 ft. Later
radiation of the selachians took place along three different lines,
represented by the three remaining orders shown in the classification.
The pleuracanthodians (*Pleuracanthus) were a specialized group of
freshwater carnivores. The tail was straight (diphycercal) and the
paired fins had become modified accordingly (see p. 137). The axis
was completelv freed from the body wall to give a paddle-like fin,
with pre- and post-axial rays, a type known as archipterygial (Fig.
113), because it was once supposed to be ancestral to all others. A
large spine on the head gives the group its name. Claspers were pre-
sent. These animals were common in the Carboniferous and Lower

180 EVOLUTION OF ELASMOBRANCHS vi. 3-

Permian, but in subsequent times they disappeared without leaving
descendants.

4. Mesozoic sharks

After flourishing in Palaeozoic seas the shark line seems to have
become nearly extinct during the Permian and Trias. During this
period there was probably little fish life in the sea and the stock
seems only to have survived by adopting a varied diet, including
invertebrate food. The protoselachian or heterodont sharks of this
period had two types of tooth, pointed ones in front and flattened
ones, for crushing molluscs, behind. Heterodontus, the Port Jackson
shark of the Pacific, is a surviving form having a dentition of this
type.

There is total cleavage of the yolk of the egg. The meroblastic form
typical of modern elasmobranchs and teleosts was therefore a rela-
tively late development and other survivors of the mesozoic period
besides Heterodontus also show holoblastic cleavage (pp. 184-236). In
later Triassic times sharks again became more abundant, and this
agrees with the presence of numerous bony fish types, on which they
presumably fed. Some of the Triassic sharks still possessed a hetero-
dont dentition (*Hybodus), though otherwise much like the modern
forms.

In Jurassic times or earlier, however, the sharks divided into the
main lines that exist today. In the suborder Pleurotremata or true
sharks the teeth all became sharp and the animals swift swimmers. In
the suborder Hypotremata, on the other hand, the teeth remained
flattened and sometimes became highly specialized for a mollusc-
eating diet (Fig. 1 14), producing the flattened bottom-living creatures,
the skates and rays. The stages of this transition can be followed, and
some of the intermediate types still exist. Thus in Rhifiobatis, the
banjo-ray (Fig. in), the pectoral fins are enlarged but still distinct
from the body. Almost identical creatures have been found in Jurassic
rocks. It is probable that several separate lines showed this flattening
of the body.

5. Modern sharks

The Pleurotremata may be divided into three divisions all dating
from the Jurassic. The Notidanoidea show many primitive features,
such as an amphistylic jaw, the presence of six or seven gill-slits, and
an unconstricted notochord. Hexanckus and Heptranchias, are long-
bodied, slow-moving sharks from warm waters. They are viviparous

vi. 5

SHARKS

181

but without placentae. Chlamydoselache, the frilled shark, lives in
deep water and feeds on cephalopods. The division Galeoidea is much
larger and includes the sharks with two dorsal fins, not supported by
spines. Here belong the dogfishes Scyliorhinus and Miistelus, both
mainly bottom-living animals feeding on a mixed diet, including

Fig. 114. Teeth of various elasmobranch fishes.

1, Man-eater (Carcharodon); 2, tiger shark (GalaeocerJo); 3, comb-toothed shark (Hexanchus);

4, sand-shark (Odontaspis); 5, blue shark (Carcharinus); 6, nurse shark (Ginglymostoma);

7, guitar rish (Rhina), 8, eagle-ray {Myliobatis), (After Norman.)

crustaceans and molluscs. In Cetorhimis, the basking shark, the pre-
daceous habit of the group has been abandoned in favour of straining
small food directly from the plankton by means of special combs on
the gills (gill rakers), an arrangement recalling that of the whalebone
whales. The great effectiveness of this method of feeding may be seen
in the length of 35 ft or more attained by some of these sharks. Bask-
ing sharks produce very numerous small eggs, which develop within
an 'uterus', but without placentae. Rhineodon, the whale shark, is also

1 82 EVOLUTION OF ELASMOBRANCHS vi. 5-6

a plankton feeder and becomes very large. It is not closely related to
the basking sharks. It moves up and down vertically, the mouth open,
sucking in plankton. In this group there are also many of the fiercest
man-eating sharks, such as Carcharodon, often 30 ft long, found in
many seas. Some fossil forms of this genus are estimated to have
reached a much greater length, possibly of 90 ft.

The division Squaloidea includes those sharks in which there is
a spine in front of each dorsal fin. They are not, however, otherwise
different in habits from the other sharks. The spiny dogfish (Squalus)
is a well-known type and here belong also the saw-sharks (Pristio-
phorus) and a group of bottom-living forms, the angel-fishes or monks
[Squatina), which acquire a superficial similarity to the skates and
rays. Alopias, the thresher, is said to differ from most sharks in that
instead of seizing the prey as it is presented, it hunts systematically,
several sharks working together and using their whip-like tails to
drive smaller fishes such as mackerel into shoals, where they are then
seized.

6. Skates and rays

The Hypotremata, skates and rays, have become specialized for
life on the bottom of the ocean in shallow waters, feeding mainly on
invertebrates, and usually having blunt teeth (Fig. 114). Locomotion
is no longer by transverse movements of the body but by waves that
pass backwards along the fins. In the earlier stages, such as Rhino-
bath, the banjo-ray, which has existed from the Jurassic period to the
present, the edges of the fins are still free and the tail is well developed.
In Pristis, another saw-fish type, outwardly similar to Pristiophorus
and known since the Cretaceous, the head is drawn out into a long
rostrum armed with denticles. Its use is uncertain but the head strikes
from side to side among shoals of fishes. There are species in India,
China, and the Gulf of Mexico that live in fresh water. In Raja, first
found in the Cretaceous, the pectoral fins are attached to the sides of
the body and the median fins are very small, whereas in the more
recent Trygon and other sting-rays the tail is reduced to a defensive
lash, the dorsal fin persisting as a poison spine. In the eagle-rays
(Myliobatis) the teeth are flattened to form a mill able to grind mollusc
shells (Fig. 114). The sea-devils (Mobula) have expansions of the fins
at the front of the head, which they use to chase fishes to the mouth,
hunting in packs. In Torpedo, the electric ray, the fins extend so far
forward that the front of the animal presents a rounded outline. The
animal is protected by a powerful electric organ, formed by modified

(i8 3 )

Torpedo

Raja

FIG. 115. Various elasmobranch. fishes.

Chimaera ^ T

184 EVOLUTION OF ELASMOBRANCHS vi. 6-

latcral plate muscle, innervated by cranial nerves. Several species of
Raja have weak electric organs perhaps used for guidauce (p. 253).

Life on the bottom has produced many further modifications in the
skates and rays. In those that live in shallow and hence well-illuminated
waters the colour of the upper surface is often elaborate, the under
side being white. In certain species of Raja, for example, there is a
pattern of black and white marks, which probably serves to break up
the outline of the fish.

The eyes of the skates and rays have moved on to the upper surface
of the head and are protected by well-developed lids. In most forms
the pupil is able to vary widely in diameter and often has an operculum
by which the aperture can be reduced to two small slits.

There is a special modification of the respiratory system so that
water is drawn in not through the mouth but by the spiracle, which is
provided with a special valve that shuts at expiration, as the water is
forced out over the gills. The Hypotremata have therefore developed
many special features for their bottom-living habits and have diverged
among themselves into many varied lines. They have been very
successful and are among the commonest fishes in the sea.

7. Chimaera and the bradyodonts

Finally we must consider an aberrant group, the bradyodonts, which
diverged from the main stock at least as early as the Carboniferous
and preserves for us today some features of elasmobranch life at that
time as the strange Chimaera, the rat-fish of deep seas (Fig. 115).
Instead of the usual large, toothed mouth these Holocephali have a
small aperture surrounded by lips, giving the head a parrot-like
appearance. The teeth are large plates firmly attached to jaws, and
the upper jaw is remarkable in being fused to the skull ('holostylic'),
the hyoid arch being free. There is no stomach or spiral intestine.
These peculiarities are probably associated with a capacity to eat
small pieces of animal food. The Holocephali differ further from the
Selachii in the presence of an opercular flap attached to the hyoid
arch. There are also extra claspers in front of the usual pelvic ones
and an organ on the head of the male known as the cephalic clasper,
whose function is obscure. The notochord is unconstricted and the
vertebrae reduced to separate nodules. The cleavage is holoblastic,
as in other fishes with features of mesozoic type (p. 180).

Many of the internal features resemble those of selachians, for
instance the conus arteriosus, and urinogenitals in which there are
separate urinary and spermatic ducts. The brain has a peculiar shape

vi. 8 TENDENCIES IN EVOLUTION 185

on account of the large size of the eyes, which almost meet above the
brain, so that the diencephalon is long and thin.

These strange creatures appear in the Jurassic, apparently de-
scended from the somewhat similar bradyodonts (such as *Helodus),
which were common in the Carboniferous and Permian. They pre-
serve some primitive features (vertebrae, jaw support, open lateral
line canals, cleavage) but have developed many specializations in the
teeth, operculum, fins, and brain, probably in connexion with life on
the bottom of deep seas.

8. Tendencies in elasmobranch evolution

The elasmobranchs have been in existence ever since the Devonian,
and for much of this long period of nearly 400 million years we can fol-
low their changes with some accuracy. This type of fish was first
formed by loss of the heavy bony armour of the earliest gnatho-
stomes, associated with the adoption of a rapidly moving and car-
nivorous habit. The resulting shark-like form has remained with
relatively little change through the whole history of the group; clado-
selachians from the Devonian are remarkably like modern sharks, and
it would be difficult to assert that the latter show clear signs of being
in any way of a 'higher' type. Both are in fact suited to the same mode
of life.

If our interpretation of the evidence is right, however, the modern
shark type has been evolved from the Devonian type through a hetero-
dont stage. During the late Permian and Trias there was little fish
food for the sharks and they appear to have taken to living on inverte-
brates. Eating this diet was presumably easier for animals possessing
the two types of teeth described on p. 180, and the animals also
became rather flattened with their life on the bottom. On the
reappearance of numerous fishes in the sea, in the Jurassic, some of
these heterodonts resumed the shark-like habit, lost the crushing
teeth, and developed into the varied fish-eating types alive today.
Others of the heterodonts, however, became still more specialized for
bottom life, as the modern skates and rays.

It is difficult to see any persistent tendency in all this, except to eat
other animals of some sort. When fishes are available sharks will eat
them, and the bodily organization for doing so seems to have been
evolved at least twice. Similarly other members of the same stock ate
molluscs and Crustacea and became modified for this. The tendency
is for survival or continuance of the animals and this leads them
to adopt whatever habits are possible given their surroundings. In

i86 EVOLUTION OF ELASMOBRANCHS vi. 8-

meeting the circumstances certain types will be suitable at one time,
others at another. We know that genetic variations will produce fluctua-
tions of type — at a time when circumstances force the animals to strive
in one direction those with a particular bodily type, say, broad, 'hetero-
dont' teeth will be selected. When fish food again becomes available
those animals born with quicker habits and sharper teeth will be able
to eat the fish and the shark type returns.

The method of ensuring stability in the pitching plane adopted by
elasmobranchs (p. 136) necessitates a certain flattening of the front
end of the animal. It is not therefore surprising that this tendency is
often exaggerated and has several times produced flattened bottom-
living creatures, such as the skates and rays. The Actinopterygii
show the opposite tendency, to lateral flattening (p. 248). We might
imagine that most of the modern skates and rays had become so
modified in structure that only life on the bottom is possible for them
and that there could be no return to a free-swimming, fish-eating
habit, but it would not be true to say that this is certain or that the
past history of the group shows undoubted evidence of such irrever-
sible specialization.

The only general conclusion from our study of elasmobranchs since
the Devonian, then, is that they have tended to keep alive by eating
fish or invertebrates, that some have changed little during this time
but that, judging especially from the modern forms, the group tends
to produce varied types at any one time, each able to find its food in a
special manner. It is not clear that the group has advanced, in any
sense, since the Devonian. The type has always been a successful one,
able to produce specialized carnivores. We do not know enough to be
sure whether the number of creatures with this organization has
changed greatly, but it seems that, except for a reduction in numbers
in the Triassic, they have always been moderately abundant and are
perhaps at present on the increase.

9. The earliest Gnathostomes, Placoderms

*Class Placodermi (= Aphetohyoidea)

*Order 1. Acanthodii. Silurian-Permian (*Climatius)
*Order 2. Arthrodira. Silurian-Devonian (*Coccosteus)
*Order 3. Macropetalichthyida. Devonian (*Lunaspis)
*Order 4. Antiarchi (= Pterichthyomorphi). Devonian (*Bothri-

olepis)
*Order 5. Stegoselachii. Devonian-Carboniferous (*Gemundina)
*Order 6. Palaeospondyli. Devonian {* Palaeospondylus)

vi. 9 PLACODERMS 187

It has already been mentioned that the earliest gnathostome verte-
brates found in the rocks do not have the shark-like form, and present
so many peculiarities that they are placed in a distinct class. In the
past the fossils included here have been referred to various groups,
usually either to the agnatha or the elasmobranchs, and there is still
some doubt as to their position. In many respects they are highly
specialized, but they all have one feature that may be presumed to
have existed in the ancestral gnathostome, namely, that the hyoid arch
played no part in the support of the jaws and the spiracle was there-
fore a typical gill-slit. For this reason they are often given the name
Aphetohyoidea, but we shall prefer to call them Placodermi, to em-
phasize that they all have a heavy armour of bone-like material. The
class contains several orders, not obviously very closely related to
each other; all are fossil forms, none of which is known to have sur-
vived the Permian.

The best-known, earliest, and perhaps most interesting group is
the acanthodians, found in freshwater deposits extending from the
Upper Silurian to the Permian but chiefly in the Devonian. These
were small fishes with a fusiform body (Fig. 115), with heterocercal
tail and two, or later one, dorsal fins. The lateral fins consisted of a
series of pairs, often as many as seven in all, down the sides of the
body. The effect of these in stabilizing the fish would presumably be
different from that of a continuous fold, and the problem of the form
and function of the earliest paired fins remains obscure. The fins were
all supported by the large spines from which the group derives its
name.

The whole surface of the body was covered with a layer of small
rhomboidal scales, composed of layers of material ressembling bone,
covered with a shiny material similar to the ganoin of early Actino-
pterygii. On the head these scales were enlarged to make a definite
pattern of dermal bones, numerous at first but fewer in the later forms.
The pattern of the bones has no close similarity to that of later fishes.
The reduced bones of the later acanthodians are related to the lateral
line canals, which have an arrangement similar to that in other fishes,
but run between and not through the scales and bones of the head.
The teeth are formed as a series of modified scales. The skull is partly
ossified — important evidence that the boneless condition of elasmo-
branchs was not typical of all early gnathostomes.

The jaws of acanthodians were attached by their own processes to
the skull (autodiastyly) and are remarkable in that four separate
ossifications take place in them (two in the upper and two in the lower

1 88 EVOLUTION OF ELASMOBRANCHS vi. 9

jaw), making a series of elements similar to that found in the typical
branchial arches. The hyoid was an unmodified branchial arch. At
first the mandibular, hyoid, and each of the branchial arches were pro-
vided with small flap-like opercula, but in later forms the mandibular
operculum became especially developed and covered all the gills.

These animals might well represent the ancestors of many if not all
other groups of gnathostomes. They have not the peculiar features
that we characterize as shark-like, and though they may well have
been carnivorous they are not very highly specialized for that mode of
life. Whether or not the known acanthodians represent the actual
ancestors of the other gnasthostome groups, it is clear that knowledge
of their anatomy forces us to discard two conclusions which have
often been accepted in the past, namely, that lack of bone and an
amphistylic jaw support are primitive gnathostome features. Here
already in the Silurian we find animals that possessed both endo-
chondral bone and scales composed of bony substance. Moreover,
some of them have no trace of denticles and we must therefore regard
with suspicion any theory that considers the placoid scale as the
original type of all scales. It is at least as likely that scales composed of
simple layers of bone in the dermis were the ancestral type and that
placoid forms with a pulp cavity were a later specialization.

Several other types of placoderm fish are known, mostly from the
Devonian strata. The Arthrodira, Macropetalichthyida, and Anti-
archi (Fig. in) were mostly heavily armoured fishes with dermal
bones on the head and often a large shield over the body. There was
usually a heterocercal tail and a covering of scales. The earlier fishes
were mostly from fresh water, the later from the sea. Many were
rather flattened, probably bottom-living and invertebrate-eating
forms. The bony plates on the head were often arranged in charac-
teristic patterns, none of which, however, shows close similarity to
the pattern of bones on the head of bony fishes or tetrapods. Lateral
line canals of typical arrangement were present and the 'bones' follow
these to some extent.

*Gemundina was a flattened animal, superficially similar to a skate,
from marine Lower Devonian deposits. The skin was covered with
denticles, but under these were large plates, apparently of bone. This
fish is placed in a special order Stegoselachii and its affinities are
unknown, but it shows again that the tendency to develop a flattened
form has been present from the earliest appearance of fishes. *Palaeo-
spondylus from the Devonian is another isolated form, in the past often
classed with the cyclostomes. Moy-Thomas showed, however, that

vi. 9 PLACODERMS 189

jaws were present and that probably the hyomandibula was not sus-
pensory. He therefore classed the fish with the placoderms, in spite
of the absence of any dermal skeleton. So far as can be discovered,
all these placoderm fishes except the acanthodians were specialized
types and have not left any later descendants. Indeed it may well be
that they have been preserved only because of the great extent of their
armour; less heavily protected relatives may have existed but have
not survived as fossils. The remains that are known are sufficient to
establish the fact that there were, in the Devonian period, numerous
types of fish possessing a bony skeleton.

VII

THE MASTERY OF THE WATER. BONY FISHES

1 . Introduction : the success of the bony fishes

The acanthodians and some other of the late Silurian and Devonian
gnathostome fishes possessed bony skeletons; from these, or some
placoderm animals like them, may have been derived not only the
elasmobranchs but also the bony fishes and the lung-fishes, which
gave rise to the land animals. These presumed descendants of the
placoderms can be divided into three groups : first the elasmobranchs,
secondly, the crossopterygians, the lobed-fin or lung-fishes, including
the Devonian forms that led to the amphibia, and thirdly, the actino-
pterygian or rayed-fin fishes, culminating in the modern bony fishes.
In Devonian times the Crossopterygii and Actinopterygii were very
alike and both, like the placoderms, contained bone. The term bony
fishes or Osteichthyes is often applied to these two groups together,
since they have some features in common and distinct from the
elasmobranchs.

The great group of Actinopterygii, which, for all the importance of
the elasmobranchs, must be reckoned as the dominant fish type at the
present time, includes most of our familiar fishes, perch, pike, trout,
herring, and many other types of 'modern' fish. In addition there are
placed here some surviving relics of the stages that have been passed
before reaching this condition, such as the bichir, sturgeons, bow-fin,
as well as related fossil forms.

Many groups of animals have been successful in the water; Crus-
tacea, for instance, are very numerous and so are cephalopod molluscs
and echinoderms, but the success of the bony fishes surpasses that of
all others. From a roach or perch in a stream, to a huge tunny or a vast
shoal of herrings in the sea, they all have the marks of mastery of the
water. They can stay almost still, as if suspended, dart suddenly at
their prey or away from danger. They can avoid their enemies by
quick and subtle changes of colour. Elaborate eyes, ears, and chemical
receptors give news of the surrounding world and complex be-
haviour has been evolved to meet many emergencies. Reproductive
mechanisms may be very complex, involving elaborate nest-building
and care of the young; social behaviour is shown in swarming move-
ments, which may be accompanied by interchange of sounds (p. 217).

vii. 1-2 THE TROUT 191

Bony fishes abound not only in the sea but also in fresh water,
which has never been effectively colonized by cephalopods or elasmo-
branchs. They can exist under all sorts of unfavourable or foul-
water conditions and a considerable number of them breathe air and
live for a time on land. Perhaps the majority are carnivorous, but
others feed on every type of food, from plankton to seaweeds.

To whatever feature of fish life we turn we find that the bony fishes
excel in it in several different ways in different species. It is small
wonder that with all these advantages they are excessively numerous.
There are some 3,000 species of living elasmobranchs, but more than
20,000 species of bony fish have been described.

The number of individuals of some of the species must be really
astronomical. For instance, at least 3,000 million herrings are caught
in the Atlantic Ocean each year, so that the whole population there
can hardly be less than a million million. Again, it is estimated that a
thousand million blue-fish collect every summer off the Atlantic coast
of the United States and, being very voracious carnivores, they con-
sume at least a thousand million million of other fishes during the
season of four months. This gives some idea of the tremendous
productivity of the sea, and of the way the bony fishes have made
use of it. Needless to say, man has also made considerable use of the
bony fishes, which indeed provide, with the elasmobranchs, a not
inconsiderable portion of the total of human food.

2. The trout

Salmo trutta, the brown trout, may be taken as an example of a bony
fish; we shall also refer at intervals to conditions in other common
freshwater fish such as the dace, Leuciscus, and perch (Perca fluvi-
atilis). There is considerable confusion about the various types of
trout and their close relatives the salmon. The brown trout is abundant
in rivers and streams throughout Europe and is commonly about
20 cm long at maturity, though it may grow larger. It is grey above
and yellowish below, with a number of dark spots scattered down
the sides of the body (Fig. 116).

The body form is typical of that of teleostean fishes in being short,
narrow in the lateral plane but deep dorso-ventrally, in fact more ob-
viously streamlined than the shape of elasmobranchs. The movements
of a trout do not at first sight obviously involve the bending of the body
into an S; nevertheless, the method of swimming is essentially by the
propagation of waves along the body by the serial contraction of the
longitudinally directed fibres of the myotomes (p. 133).

192

BONY FISHES

VII. 2-

The tail differs from that of elasmobranchs in being outwardly
symmetrical, though internally there are still traces of the upturned
tip of the vertebral column (Fig. 118). Besides the typical caudal
'fish-tail', supported by bony rays, there are two dorsal fins and a
ventral fin, but the hinder dorsal fin differs from the others in having
no rays to support it and is called an adipose fin, because of its flabby
structure. The paired fins are rather small and it is from their struc-
ture that the whole group derives the name Actinopterygii or rayed-fin
fishes. There is no lobe projecting from the body and containing

Fig. i i 6. Male and female brown trout (Salmo trutta) spawning. The male is quivering — a

short sequence of rapid shudders of whole body which excites the female.

(After J. W. Jones, The Salmon.)

basal fin supports, as there is in the fin of lung-fishes. All the basal
apparatus of the fin is contained within the body wall and only the fin
rays project outwards, as a fan. The pelvic fin of bony fishes often
lies relatively far forward; in the trout, however, it is unusually far
back, just in front of the anus; in other types it may be level with the
pectoral fin, or even anterior to it (Fig. 118). The significance of the
shape of the body and fins in swimming will be discussed later (p. 244).
The skin consists of a thin epidermis and thicker dermis, the former
has stratified squamous layers but contains no keratin (Burgess, 1958).
It contains mucous glands. The mucus of some eels and other fishes
has remarkable powers of precipitating mud from turbid water. The
mesodermal dermis provides an elaborate web of connective tissue
fibres. It also contains smooth muscle, nerves, chromatophores, and
scales. The latter are thin overlapping bony plates, covered by skin,
that is to say, they do not 'cut the gum' as do placoid denticles. The
exposed part of each scale bears the pigment cells, which control the
colour of the animal, in a manner presently to be described. The bone
of the scales is absorbed at intervals by scleroclasts, making a series
of rings, which, like the growth-rings on a tree, are due to the fact
that growth is not constant but occurs fast in the spring and summer

VII. 3

FISH SCALES

193

and hardly at all in the winter. The age of the fish can therefore be
determined from these rings (Fig. 117), or from the similar markings
on the ear stones (p. 216). While an adult salmon is in fresh water no
growth occurs, leaving a spawning mark on the scale.

The head of the trout shows some of the most specialized and typical
teleostean features (Fig. 119). There are two nostrils on each side, but
no external sign of ears. The mouth is very large and its edges are
supported by movable bones, to be described below. The maxillary
and mandibular valves are folds
of the buccal mucosa, serving to
prevent the exit of water during
respiration. The tongue, as in
Selachians, has no muscles, but
may carry teeth and taste-buds.
Behind the edge of the jaw is the
operculum, a flap covering the
gills and also supported by bony
plates. In connexion with these
special developments of jaws and
gills the skull has become much
modified and has developed com-
plex and characteristic features
(Fig. 118).

Fig. 117. Spawning mark (sp. ?nk.), the
result of erosion or absorption of the scale
margin due to a calcium deficiency fasting
period. (After J. W. Jones, The Salmon.)

1st river winter. 2. 2nd river winter.
3. 1 st sea winter.

3. The skull of bony fishes

The main basis for the skull
is a chondrocranium and set of

branchial arches, exactly comparable to those of the elasmobranchs.
In the early stages of development there is a set of cartilaginous boxes
around the nasal and auditory capsules, brain and eyes, and a series
of cartilaginous rods in the gill arches. Bones are then added in two
ways: either (i) as cartilage bones (endochondral bones) by the re-
placement of some parts of the original chondrocranium, or (2) as
membrane or dermal bones, laid down as more superficial coverings
and considered to be derived from a layer of scales in the skin. This
outer position of the bones can be clearly seen in many cases by the
readiness with which the membrane bones can be pulled away from
the rest of the skull.

The skull bones are arranged in a regular pattern, whose broad
outlines can be seen in all fishes and in their tetrapod descendants.
However, there are many confusing variations and the naming of

(i94)

« in-.

j? e a a x -9

u 2. & 2 «"<=«'

s3 •'2 • -5

n ii ? 2 ° -2

F P 0« 3 o} -;

"^ O >. ~ .a „-

C • 2 o c '

B 05 "?

* H

o %

£ T3 ™
° 2 tl
w c c

° c

&

1)
J5

3 E'CsS o

A A . U . .

« 5? ^ a o P o

S*-. SE ... «s <n ■ n
,_ c <u --^3
„ ^.* c o

re — . *-■ en

i; -c

2 2

tj c c

dt tu c

a -'t.

2 E

a- 2.

"S

^4

bs. bra
al (hyp
lum ; f
t". ecti
"; oute
ring; s

A

WD

E-

od

r^i

"a

d

ages;
ssohy
percu
and F

G. SG

bital

1

£

basal and radial cartil;
ic; FR. frontal; GH. glo

NS. neural spine; o. o
cesses of ribs; pt, pt',
g'. dorsal process of s
a-occipital; sor. subor

• <-. • - o <o u -

« o "0 u ft

+ a o a « g

. <J C iL Vl - 1

: - o o •£• ■- o

{3 a, £ t: m ^

— w 2 o J, «

3 c a o n

ang
oid;

NC.

sup
uld

. cr

• P • 2 <->

< ■£ £? ^ M w

. . u ja »« ■ • -

,£ a ~ E »i E

<G W X 3 3

a. anal
of fins
MX. m
opercu
R. ribs
opercu

vii. 3 SKULL OF TELEOSTS 195

bones has given much controversy. No generally acceptable theory
has yet appeared, perhaps because we know little of the factors that
cause separate bony elements to develop. There is some evidence of
relations between bones and teeth and bones and lateral line organs
and the pattern of the latter may play a large part in determining the
plan of the skull (p. 325). Provisionally we may recognize four classes
of dermal bones (1) canal bones, (2) tooth bones, (3) 'ordinary' bones,
whose determination is unknown, (4) extra bones, filling special areas
(Wormian bones).

The arrangement of the numerous bones is made less difficult to
understand and remember if they are considered in the following
order. First the endochondral ossifications in the original neuro-
cranium, then the dermal bones that cover this above and below; next
the endochondral bones formed within the original cartilaginous jaws,
then the dermal bones that cover the edges of the jaws, and finally the
ossifications in the branchial arches and pectoral girdle, which latter
is in bony fishes attached to the skull.

The endochondral ossifications may be considered by beginning at
the hind end of the skull: here the floor ossifies as the basioccipital,
the sides as the exoccipitals, and the roof, over the spinal cord, as the
supra-occipital bone; these posterior bones are not well marked off
from each other in the adult skull. In the auditory capsule are five
separate otic bones, of these the epiotic and pterotic can be seen
externally (Fig. 118).

The floor in front of the basioccipital is occupied by a basisphenoid
bone and the walls above this by alisphenoids. The eyes nearly meet in
the midline and the orbits are here separated only by a thin orbito-
sphenoid. The only more anterior part of the chondrocranium to
ossify is the region between the nasal capsule and the orbit, forming
the ectethmoid.

The dermal bones that cover this partly ossified neurocranium may
be identified as on top and in front a pair of frontals and a median
supraethmoid, behind which are large paired parietals and small
paired post-parietals. These names have been inferred from study of
crossopterygians and early amphibians (p. 325), which showed that
the homologies earlier accepted were wrong. Fig. 118 carries the old
nomenclature in which the large paired bones were called frontals.
Around the eyes is a ring of circum-orbitals, and on the floor of the
skull two median bones, the parasphenoid and vomer.

The jaw bones are numerous, including both endochondral and
dermal elements, and the relation of the method of support to that

196 BONY FISHES vn. 3-

found in other animals is not clear. In the embryo palato-pterygo-
quadrate bars and Meckel's cartilages are seen. The upper jaw bears
inward projections, which extend towards the chondro-cranium and
probably represent the traces of an autostylic means of support (see
p. 187). But the effective support in the adult is achieved by the ossi-
fied hyomandibular cartilage. The palato-pterygo-quadrate bar ossifies
in several parts and palatine, pterygoid, mesopterygoid, metaptery-
goid, and quadrate bones appear, some of them partly formed in mem-
brane. The only part of Meckel's cartilage to ossify is the articular
bone, at the hind end. The actual edges of the jaws are supported by
membrane bones, the premaxilla, maxilla, and jugal, covering the
upper jaw. The dentary covers most of the lower jaw, except for a
small bone, the angular, that lies on the inner side at the posterior end.

The hyomandibular bone runs from an articulation with the otic
capsule to the upper end of the quadrate. The symplectic is a small
separate ossification at the lower end of the hyomandibula. The rest
of the hyoid arch is present as epi-, cerato-, and hypohyals, which
support a large toothed tongue. Bony fishes only rarely possess an
open spiracle and immediately behind the hyoid are attached the
bones supporting the operculum that covers the gills. The branchial
arches are formed of several pieces, as in elasmobranchs, each being
ossified separately.

The effect of this complicated set of bones is to provide an efficient
apparatus for the protection of the brain and sense-organs, support
of the jaws and teeth and of the respiratory apparatus. Teeth are found
on the vomers, palatines, premaxillae, maxillae, dentary, and on the
tongue. Covering the typical dentine (orthodentine) is a layer of
harder vitrodentine, poor in organic matter and perhaps derived
partly from ectodermal ameloblasts. The teeth are usually spikes
pointing in a backwards direction, used to prevent the escape of the
food and not usually for biting or crushing. They may, however, form
plates or be firmly attached to the bones. Folds of the mucous mem-
brane, supported by cartilage and carrying gill-rakers, are found in
species that feed on small prey.

4. Respiration

Limitations are imposed on the respiration of fishes by the facts
that water is 800 times more dense than air and the dissolved oxygen
is 30 times more dilute. The cost of respiration is therefore high. There
is a 70 per cent, increase in metabolism when a trout increases its
ventilation volume four times in water poor in oxygen.

vir. 4

RESPIRATION

197

Respiration is produced by a current passing in a single direction,
as in elasmobranchs, namely, in at the mouth and out over the gill
lamellae, but the mechanism by which the current is produced is some-
what different. The pumping action is produced by a buccal pressure
pump and opercular suction pumps, resulting from sideways move-

A B

Fig. 119. Arrangement of organs of head and gills in A, a shark, and,

B, a teleostean fish.

gb. gill bar; GC. outer opening of gill cleft; GF. gill filament; gr. gill-rakers; cv. gill vessels;

J, ]'. upper and lower jaw; M. mouth; N, n'. openings of nasal chamber; op. operculum;

sp. spiracle; ST. septum between gill filaments. (From Dean, Fishes, Living & Fossil,

The Macmillan Co.)

ments of the operculum, enlarging the branchial cavity. The branchio-
stegal folds, below the operculum, prevent inflow of water from
behind. When the operculum moves inward dorsal and ventral flaps
in the throat prevent the exit of water forwards.

The gill lamellae differ from those of elasmobranchs in the great
reduction of the septum between the respiratory surfaces (Fig. 119).
This has the effect of leaving the lamellae as free flaps, increasing the
surface available for respiration.

The area of the gills varies greatly, being relatively larger in more
active species. The rate of respiration is controlled by a medullary
centre but the most active rate of respiratory exchange is only some
four times the standard rate (as against twenty times in man). The

(198)

'•5 E

Q

«

ft

</>

a

i>

C3

(A '"7'

C

.- O

a

■JH

«U

"3

T3

C

c

ft .2

T3

ft

ftr=

U E

... <u
r w
«5 *

> "O — -

"a K

° a a

3 S

.2 ^3 k,

^ M

ji

?.'£$

<£

•^ o

W3 >■-

U-

>

u ° £
is &.-:•

c 3 rT

• - ft ^ c

pa 3 d °

• 3 b be t -
•- c w cfc

w « l- .3 w

3 n O fe

o "' c s a

2* g ,-s

c

bo

h

3= C
3 C
J3 JO

«

ft 2 **

_ jg c/) CO ■«

• - S -2 -5 ?

^3

6S

? E

? ° £

* >•-£ Si

Ch ta •- ft

< a u .

C H

M

3-8 2!:

C y J « .S

M

B-a

* " >

J OS

I ft .

.3 Oh J

ra 3 ._

< ^ u

.- E £ 13

"eg*

vii. 4-5 BACKBONE OF FISHES 199

area of the respiratory surface is thus an important limiting factor in
the movement and growth of fishes. During activity of a fish lactic
acid accumulates in the blood and the pH falls. The fish is thus able
to display a considerable burst of activity and then to repay the
oxygen debt over a long subsequent period.

5. Vertebral column and fins of bony fishes

The vertebral column of bony fishes performs the same function as
in other fishes, namely, to prevent shortening of the body when the
longitudinal muscles contract. It has, however, become very compli-
cated and with the ribs and neural and haemal arches forms an
elaborate system serving to maintain the body form under the stresses
of fast swimming. Like other parts of the skeleton it is extensively
ossified, and the necessary lateral flexion is obtained by division of
the column into a series of sections joined together. Typically there
is one such section (vertebra) corresponding to each segment, but in
the tail region of Amia there are twice as many vertebrae as segments.

Each vertebra consists of a centrum, neural arch and neural spine,
and in the tail region, in addition, haemal arch and haemal spine.
These parts are formed partly by ossification of cartilaginous masses,
the basidorsal and basiventral, interdorsal and interventral, such as
we saw in elasmobranchs, and partly by extra ossification in the sclero-
genous tissue around the notochord and nerve-cord and between the
muscles. The vertebrae are inter-segmental, the middle of each lying
opposite the myocomma that separates two muscle segments.

The centra are concave both in front and behind (amphicoelous),
and in the hollows between them are pads made of the remains of the
notochord, an arrangement that allows the column to resist longi-
tudinal compression and yet remain flexible; similar flat or concave
articulations of the centra are found in other aquatic vertebrates from
the elasmobranchs to the whales. Extra processes on the front and
back of the vertebrae ensure the articulation and are comparable to the
zygapophyses found in tetrapods. The ribs, which are so prominent
in the backbone of many fishes, are of two sorts; pleural ribs between
the muscles and the lining of the abdominal cavity, and more dorsal
intramuscular ribs. Both sorts are attached to the centrum. The bony
rods attached above the neural and below the haemal arches are often
called neural and haemal spines, though it is doubtful whether they
correspond to the neural spines of land vertebrates. They form the
supporting rods or radials of the median fins and are usually divided
into two or three separate bones in each segment. In addition to these

200

BONY FISHES

vii. 5-

radials the fins are also supported by a more superficial set of bony-
rods, the dermal fin rays (dermotrichia or lepidotrichia), which may
be considered as modified scales and accordingly lie superficial to the
radials. These dermal fin rays are usually forked at their tips. They
make an extra support for the fin margin and to them are attached the
muscles that serve to throw the fin into folds.

In the tail region the internal skeleton is not quite symmetrical and
shows signs of origin from an animal with a heterocercal tail. The

<ev.msx.

3<j-mand.

Fig. 121. Muscles of a teleostean fish, mainly based on Mullus.
ad-mand. adductor mandibulae; ep. epaxonic muscles; h. horizontal myoseptum; hy. hypo-
branchial muscles; hyp. hypaxonic muscles; lev. max. levator maxillae; my. myocomma;
op. operculum. (From Ihle.)

notochord turns up sharply at the tip, so that the neural spines are
very much shorter than the haemal spines, known here as hippural
bones. The final portion of the notochord is often surrounded by a
single ossification, the urostyle, and the whole makes a rigid support
for the dermotrichia of the tail. Such a tail with internal asymmetry
but external symmetry is said to be homocercal.

The myotomes are arranged in a complicated pattern having the
effect that contraction of each affects a considerable section of the
body (Fig. 121); in fast swimmers such as the tunny each myotome
may overlap as many as nineteen vertebrae. Between the lateral and
ventral muscle masses there is in many fishes a layer of red muscle
and this is especially well developed in the tunnies and bonitos.

The paired fins are similarly supported by ossified radials, covered
by dermal fin rays. At the base the radials are connected with 'girdles'
lying in the body wall. The pectoral girdle (Fig. 118) consists of a
cartilaginous endo-skeletal portion in which ossify the scapula, cora-
coid, and sometimes mesocoracoid, while dermal bones, large cleith-
rum, and one or more small clavicles, become attached superficially.

vii. 8 ALIMENTARY CANAL 201

Above these a further series of dermal bones, the supra-clavicle and
post-temporal, attach the pectoral girdle to the otic region of the skull.
The pelvic girdle is very simple, consisting only of a single bone,
the basipterygium.

6. Alimentary canal

The food of the trout consists mainly of small invertebrates such as
Gammarus, Cyclops, and other crustaceans, and aquatic insects and
their larvae, together with the fry of other fishes and perhaps some-
times larger pieces of 'meat'. The food is mostly swallowed whole,
being helped down the pharynx by the mucous secretions, but these,
as in elasmobranchs, contain no enzymes. The entrance to the stomach
is guarded by a powerful oesophageal sphincter, no doubt serving to
prevent the entry of the water of the respiratory stream. The stomach
is divided into cardiac and pyloric portions, though the distinction is
less clear than in elasmobranchs. The duodenum is beset by a number
of wide-mouthed pyloric caeca, serving to increase the intestinal
surface (Fig. 120). The intestine and caeca are lined throughout by a
simple columnar epithelium and there are no specialized multicellular
glands such as the Brunners glands or crypts of Lieberkiihn of mam-
mals. The exocrine pancreas consists of numerous diffuse glands in
the mesentery. The endocrine portion, however, forms a compact
mass of tissue. This is very rich in insulin and after its removal a fish
shows hyperglycemia and glycosuria. The intestine is relatively longer
than in elasmobranchs and often coiled; its internal surface may be
increased by folds, but there is no true spiral valve, though this was
present in the ancestors of the Teleostei (p. 233). There is no gland
attached to the rectum.

7. Air-bladder

Dorsal to the gut is a very large sac with shiny, whitish walls, the
air-bladder, filled with oxygen. A narrow pneumatic duct connects
this with the pharynx in the more primitive forms. The origin and
functions of the air-bladder will be discussed below (p. 261); it serves
as a hydrostatic organ, enabling the animal to remain suspended in
the water at any depth.

8. Circulatory system

The general plan of the circulation is similar to that of an elasmo-
branch (Fig. 122), that is to say, there is a single circuit and all
the blood passes through at least two sets of capillaries. The heart

202 BONY FISHES vu. 8-

contains a series of three chambers, sinus, auricle, and ventricle, but the
muscular conus arteriosus is absent, there being only a thin-walled
bulbus arteriosus at the base of the ventral aorta. The walls of the
bulbus are elastic but not muscular, and study of its action by means
of X-rays shows that it is dilated by the ventricular beat and then
contracts, thus maintaining the pressure against the capillaries of the
gills. The ventral aorta is short, but the arrangement of the afferent
and efferent branchial vessels is essentially as in elasmobranchs.

The blood-pressure in the ventral aorta is less than 40 mm Hg in
most fishes at rest, and in the dorsal aorta about half this. The venous
pressures are around zero, the pericardium being fibrous but not
rigid as it is in elasmobranchs (p. 160). There is no communication
between the pericardial and peritoneal chambers. There is a vagal
cardiac depressor nerve, but no sympathetic nerve to the heart.

There is a well-developed lymphatic system beneath the skin and
in the muscles and viscera. Lymphoid tissue is abundant in various
organs but there are no lymph-nodes along the vessels. There is a
large spleen concerned with haemopoiesis, which also proceeds in the
kidneys. The red cells are smaller in bony fishes (8-10 ix) than in
elasmobranchs (up to 20 /z). A continuous series of white cells is
present and acidic and basic granules may occur in the same cell.

9. Urinogenital system and osmoregulation

The kidneys are mesonephric in the adult and consist of an elon-
gated brown mass above the air-bladder. The ducts of the two kidneys
join posteriorly and are swollen to form a bladder which, being meso-
dermal, must be distinguished from the endodermal cloacal bladder
of tetrapods. The urinary duct opens separately behind the anus,
there being no common cloaca.

Nitrogenous elimination is a function mainly of the gills, which
excrete as ammonia and urea more than six times as much nitrogen as
the kidneys. The latter excrete creatine, uric acid, and the weak base
trimethylamine oxide, which is present in large amounts in the blood
of marine teleosts.

One of the most striking features of the life of bony fishes is that
they occur both in fresh water and the sea, and many, such as the
trout itself, can move from one to the other. It is supposed by some
that the earliest gnathostomes were freshwater animals (p. 187), and
the bony fishes might be said to show evidence of this in that the
concentration of salt in their blood is always less than in the sea, in the
neighbourhood of 1-4 per cent. NaCl against 3-5 per cent, outside.

vii. 9 BLOOD OF TELEOSTS 203

In fishes in fresh water the blood is more dilute, about o-6 per cent.
NaCl, but is, of course, more concentrated than the surrounding
medium, which contains only traces of inorganic ions. Freshwater
fishes are able to take up salts from the water through the gill surfaces.
The kidney apparatus, with its filtration system of glomeruli and
tubules for salt reabsorption, was probably developed for life in fresh
water and still serves in this way in the freshwater forms. Various
special devices are adopted in fresh water for minimizing the tendency

Fig. 122. Diagram of the branchial circulation of a teleostean fish.
ab. artery to air-bladder; a/ 3-6 , four afferent vessels from ventral aorta; ca. carotid artery;
cc. circulus cephalicus; cl. coeliac artery; d. ductus Cuvieri; da. dorsal aorta; ef 3 . efferent
vessel of first branchial arch; ep. epibranchial artery; ha. hyoidean artery (afferent vessel of
pseudobranch); hp. hepatic vein; ht. heart; mis. mesenteric artery; oa. ophthalmic artery
(efferent vessel of pseudobranch); ps. pseudobranch (hyoidean gill, possibly with spiracular
gill); s. position of spiracle (closed); va. ventral aorta; I-V. five branchial slits. (From
Goodrich, Vertebrata, A. & C. Black, Ltd., after Parker.)

to gain water and lose salt. The skin is little vascularized and probably
makes an almost waterproof layer. The production of mucus assists
in this waterproofing, and abundant mucus is secreted when an eel is
transferred from salt to fresh water: the full change cannot be made
suddenly without killing the fish.

In marine teleosts the problem is the opposite one of keeping water
in, or keeping out salt. The usual kidney mechanism is clearly ill
suited for this and it is found that the glomeruli are few, or often
completely absent from the kidneys. This no doubt reduces the loss
of water, but is not enough by itself to solve the problem, which is
met by taking in water and salts and excreting the salts. For this
purpose special chloride-secreting cells are present in the gills and it
has been shown that the amount of oxygen they use, and hence the
work they do in diluting the blood, is proportional to the difference of
concentration between the inside and the outside. A marine fish is
able to drink and absorb sea water in spite of the fact that this is more

204

BONY FISHES

vir. 9-

concentrated than its blood. The chloride-secreting cells dispose of
the excess salts. It remains to explain the means by which a solution
passes against the osmotic gradient from the cavity of the gut to the
blood; the membranes here must have some special properties of
which we are ignorant. Sodium and chloride enter with the water but
magnesium is excluded and may be precipitated in the intestine.

The genital system is nearly completely separated from the excre-
tory in both sexes. The testes (soft roes) are a large pair of sacs opening

into the base of the urinary
ducts. The ovaries (hard roes)
are also elongated in the trout
and the eggs are shed free into
the coelom (Fig. 123) and
passed to the exterior by
abdominal pores. This condi-
tion is unusual among teleosts,
in most the ovaries are closed
sacs, continuous with the ovi-
ducts.

Fertilization is external and
the eggs of the trout are shed
in small pits or depressions in
the sand; being sticky, they
become attached to small
stones. The eggs are very
yolky and cleavage is therefore only partial, forming a cap of cells,
the blastoderm, which eventually differentiates into the embryo.
After hatching, the young fish may still carry the yolk sac and
obtain food from it for some time, while beginning to eat the small
crustaceans and other animals that are its first food.

10. Races of trout and salmon and their breeding habits

There is considerable confusion about the various races of trout
and their allies the salmon. In both trout and salmon the adult
originally spent the great part of its life in the sea but returned to the
rivers to breed. Trout and salmon that do this are still abundant on
the West Atlantic coasts and ascend all suitable rivers to breed, the
process being known as the 'run'. But the trout has produced many
races of purely fluviatile animals, living either in lakes (where they often
become very large) or rivers, and never returning to the sea to breed.
These freshwater races differ in small points from each other and are

Fig. 123. Ovaries and kidneys of A, typical

teleostean fish, B, trout and some other fishes

where the eggs are shed into the coelom.

(From Norman, after Rey.)

VII. IO

LIFE CYCLE OF SALMO

205

given various common names (phinock, Severn, Loch Leven trout,
brook trout, &c). There can be no doubt that interesting genetic
differences between these forms exist, but they have not yet been
fully studied. The salmon are much less prone to form purely fresh-
water races, though such are known.

During the breeding-season characteristic changes take place in the
fishes and differences between the sexes appear. In the salmon the
jaws become long, thin, and hooked, especially in the male. The
animals make pairs and the males fight with others that approach the
female. As the gonads ripen, the other
parts of the fish, which were well supplied
with fat at the beginning of the run,
become progressively more watery.
Finally, spawning takes place, the female
laying the eggs in a shallow trough (redd),
which she has 'cut' in the gravel by move-
ments of her tail, while the male sheds
sperms over them. She then covers the
eggs with gravel by further cutting move-
ments. The young male salmon (parr),
which have not yet been to the sea, may
become sexually mature. They accompany
the fully grown fish, hanging around the
cloacal region and shedding their sperms
at the same time as the large male. It is possible that this develop-
ment of a kind of third sex serves to increase the variability of the
population. The spent parr eat some of the eggs and they then proceed
to grow, migrate to the sea, and return later.

Male trout will follow a spawning salmon and fertilize her eggs if
her own male is not looking. Hybrids formed in this way can develop,
but are said to be less fertile than the normal types; indeed, the males
are wholly sterile.

After fertilization the salmon are very exhausted (known as kelts) ;
the males seldom return to the sea. The females, however, may recover
and after a period in the sea return to breed again, and this process
may be repeated several times.

Very young trout or salmon are known as alevins or fry and remain
mostly among the stones (Fig. 124). When they emerge they are called
parr and have a number of characteristic parr-marks along their sides.
After two to four years spent as parr in fresh water salmon acquire a
silver colour and pass to the sea as smolts. Young salmon returning for

Fig. 124. Three stages in the

development of the salmon.

I and II are alevins; III, parr.

(From Norman.)

206

BONY FISHES

the first time to breed are called maidens. If they have spent only one
and a half years in the sea they are called grilse and may then return
to the sea as kelts. Others ascend for the first time after three years
or more at sea.

It is well established that salmon nearly always return to breed in
the river in which they were born, and it is certain that they may
journey for considerable distances in the sea. The mechanisms by
which these migrations are initiated and guided are only partly known.

Fig. 125. Pituitary gland of the primitive teleost Elops, showing the persistent
Rathke's pouch in the form of a hollow bucco-hypophysial canal, piercing the

parasphenoid bone.

ah. adenohypophysis ; b.h.c. bucco-hypophysial canal; h.v. blood-vessel; c. continuation with

pharynx; i. infundibulum; n.h. neurohypophysis;/)^, parasphenoid.

(After Olsson.)

They probably involve endocrine changes, for example the thyroid
is very active in the smolt as they begin to migrate. The return to the
home river may be a result of olfactory conditioning (see p. 221).

11. Endocrine glands of bony fishes

The pituitary gland occupies the same central part in the endocrine
signalling system that it has in mammals. Neural and glandular
regions are present and the adenohypophysis has three parts, the
two more posterior corresponding to the mammalian intermediate and
anterior lobes. The most anterior glandular region may be comparable
to the pars tuberalis. Experiments by removal and injection have
shown that the middle portion produces hormones that stimulate
growth, the gonads, the thyroid, adrenal, and probably the pancreas.
The posterior lobe produces a melanophore-dispersing hormone

ENDOCRINE GLANDS

207

(p. 260) and there may be a melanophore-concentrating one in the
anterior lobe. Oxytocin and vasopressin are present but there is no
evidence that excretion is controlled by the pituitary.

The thyroid tissue is not aggregated into a compact gland but forms
scattered masses along the ventral aorta. Its hormones appear to be

ns.a.

Fig. 126. Urohypophysis of A, eel; b, loach (Misgurnus); c, the same in a loach

after sectioning the spinal cord and injecting hypertonic saline.

b. blood-vessels; ep. ependyma;//. filum terminate; /;/. lumps of secretion (? 'Herring bodies');

tie. nerve endings; ns.a. neurosecretory axon; ns.c. neurosecretory cells; s.d. storage depot of

neurosecretion. (After F.nami, N., in Symposium on Comparative Endocrinology. Wiley, New

York.)

identical with those of mammals, including mono- and di-iodotyrosine
and thyroxin. Thyroid follicles are often found in the kidneys, heart,
eye, and elsewhere in the body of fishes, especially those deprived of
iodine.

The suprarenal and interrenal tissues are partly associated in masses
around the thickened walls of the posterior cardinal veins. Because of
the difficulty of isolating these tissues there is little information as to
their function. The corpuscles of Stannius are groups of gland cells
dorsal to the kidnevs, they have been held to be related to the adrenals,
but their nature is still uncertain.

208

BONY FISHES

The ultimobranchial gland is a mass of cells developed from the
last branchial pouch and perhaps related to the parathyroids.

The hormone rennin, which raises the blood pressure, is said to be
present in freshwater teleosts, with their high glomerular filtration,
but not in marine ones.

The gonads produce steroid hormones as in elasmobranchs (p.
167). The secondary sex characters depend upon their presence, thus

ns. c.

n.e

Fig. 127. Comparison of pattern of organization of the caudal neurosecretory
system (b) with the hypothalamo-hypophysial system (a).

ah. adenohypophysis; b.v. blood-vessels; c.c. central canal; /.i. filum terminale; h. 'Herring
bodies', neurosecretory products; n. neurohypophysis; n.e. nerve endings; ns.c. neurosecre-
tory cells; o.c. optic chiasma; p. pituicytes; p.v. hypophysial portal vessels; r.f. Reissner's fibre;
11. urohypophysis; v. vtntricle. (After Enami, N., in Symposium on Comparative Endocrinology.

Wiley, New York.)

the gonopodium of the male of viviparous fishes (p. 267) is developed
if sex hormone is added to the water.

At the hind end of the spinal cord of fishes is a small lump consisting
of masses of secretion produced by neurosecretory cells of the spinal
cord and hence called the urohypophysis (Figs. 126 and 127). In
function it appears to be connected with salt regulation; injection of
hypertonic NaCl produces hypersecretion, the products accumulating
at the cut surface if the cord has been severed. Injection of extracts
produces changes in the sodium content of the fish and also changes

VII. 12

FOREBRAIN

209

in buoyancy, perhaps due to an influence on the carbonic anhydrase
of the gas bladder.

12. Brain of bony fishes

The brain of bony fishes is built on the same general functional and
structural plan as that of elasmobranchs, namely, the development of a
number of separate centres, one concerned with each of the main
receptor systems. The forebrain is often large, but it is characterized

chiasma
opticum

rec. pr.

Fig. 127 {a). Cross-section of the forebrain of the cod.
lat.tr. and tn.tr. lateral and medial tracts between olfactory region and hypothalamus; hyp.
hypothalamus; m. membranous roof of forebrain; n.mag. nucleus magnocellularis preopticus;
n.opt. optic tracts (that on the right has atrophied in this specimen) ; rec. pr. preoptic recess ;
str. hind end of striatum; thai, thalamus; 3rd v. third ventricle. (From Kappers, Huber,

and Crosby.)

by great development of its ventral regions (the 'corpus striatum'), the
roof being wholly membranous (Figs. 127 (a), 128). This condition is
known as 'eversion' and is the very opposite to the inverted or thick-
roofed forebrains that are found in the lung-fishes, close to the line
of tetrapod descent (p. 278). The whole of this forebrain is reached by
olfactory fibres, and there is little evidence that fibres from other
receptor centres reach forward to it; it is mostly a smell brain.
Extirpation of the telencephalon from various teleosts has not
produced changes in locomotion, balance, or vision; there may be
slight changes in general activity and social behaviour. No movements
have been seen following electrical stimulation of it.

The diencephalon is not large, since most of the optic fibres end
not here, but in the midbrain. The roof is everted to form a pineal
body, and this and other parts of the diencephalon may contain

210 BONY FISHES vn. 12

receptors sensitive to light. The minnow Phoxinus has a transparent
patch on the head in this region, and it has been found possible to
train the fish to give appropriate responses to changes of illumination
even after removal of the paired eyes and the pineal body. Evidently
there are light-sensitive cells in other parts of the walls of the dien-
cephalon, besides those that become evaginated to form the eyes.
Experiments on lampreys also showed the presence of such cells
(p. 105). The hypothalamus is well developed and receives large tracts
from the forebrain. Below and behind it is a large saccus vasculosus
in some forms (p. 169).

The midbrain is often the largest part of the brain. The cells spread
out over its roof (tectum opticum) are not all collected round the
ventricle but have migrated away to make an elaborately layered
system. Into this midbrain cortex there pass not only the great optic
tracts but also ascending tracts from the sensory regions of the spinal
cord, lateral line system, gustatory systems, and cerebellum. Large
motor tracts pass back towards the spinal cord; the details of their
endings have not been traced, but they certainly exercise control over
motor functions. Electrical stimulation of the optic lobes produces
well-coordinated movements of local groups of muscles, for instance
those of the eyes or fins. It can hardly be doubted that this well-
developed midbrain apparatus thus controls much of the behaviour of
the fish and is able to mediate quite elaborate acts of learning and
other forms of more complex behaviour. After removal of the tectum
of one side a minnow is blind in the opposite eye. Each part of the
retina is mapped on to a distinct area of the tectum and if the optic
tract is cut and allowed to regenerate this projection is exactly replaced.
When a goldfish is trained to respond to some visual stimulus the
learning process occurs in the midbrain and continues unaffected after
removal of the forebrain. Conversely, olfactory learning takes place
in the latter and is undisturbed by injury to the tectum opticum.

The base of the midbrain (tegmentum) contains motor centres.
Electrical stimulation here produces abrupt and massive responses of
the locomotor apparatus, very different from the sequences of co-
ordinated movements that appear after stimulation of the roof of the
tectum.

The cerebellum is very large in teleosts, especially in the more
active swimmers, and a forwardly directed lobe of it, the valvula
cerebelli, extends under the midbrain. Various disorders of movement
have been reported after removal of the cerebellum, such as swaying
when moving quickly. Presumably it plays an important part, as in

HKA1N OF TELEOSTS

Fie. 128. Transverse sections of forcbrain in various vertebrates to show the condition

of inversion (thick roof) in A, n, and c and eversion (thin roof) in D, E, and F.

A, lamprey; B, frog; C, chelonian; D, chimaera; E. sturgeon; F, teleostean; pall, pallium;

sep. septum; str. striatum. (From Kappers, Huber, and Crosby.)

m.

Fig. 129. Sagittal section of brain of the gurnard, showing the swellings in the

spinal cord at the point of entry of the nerves from the fin.

C. cerebral hemisphere; ce. cerebellum; hy. hypothalamus; in. midbrain; no. swellings of

spinal cord; v. valvula. (From Scharrer, Z. verg. Physiol. 22.)

other vertebrates, in producing precise and correctly timed movements.
It is enormous in the Mormyridae, where it may assist in direction-
finding by electrical pulses (p. 253), perhaps acting as a timing device.
The medulla oblongata is also well developed, having special lobes
connected with the entry of the lateral line nerves and gustatory fibres
of the cranial nerves. In the gurnard, Trigla, there are chemical
receptors in the elongated fins. These are innervated from spinal
nerves, and there are swellings of the dorsal part of the spinal cord at
the points where these nerves enter (Fig. 129).

212 BONY FISHES vn. 13-

13. Receptors for life in the water

The features of the environment that are relevant for life are very
different in air and water. Man is so well used to the air that it is not
easy to appreciate fully the conditions underwater, where changes of
illumination, though obviously important, provide less detailed evid-
ence of the sequence of distant events than they do in air. We can
say that light carries less information for a fish; to put it in another
way, fewer distinct choices between alternative behaviour pathways
are made on the basis of visual clues by a fish than by a man.

On the other hand, the water around the fish provides mechanical
stimuli both at low and high frequency that are more closely related
to distant events than is generally true in air. Both hearing and touch
are of great importance in the water and the lateral line system pro-
vides a system of 'distant touch' that is perhaps wholly outside our
experience. Localization of distant objects by such a sense, perhaps
assisted by echo-location by water movements, provides the fish with
many relevant clues. It is interesting that these receptors are connected
with a very large cerebellar system, perhaps concerned with measur-
ing time differences.

Chemical changes in the water also provide much information and
both taste and smell are well developed. That smell is analysed by a
distinct system in the forebrain, not directly related to the cerebellar
system, is one of the fundamental principles of control of vertebrate
behaviour. Distant chemical changes provide the first clue to the
presence of food, a mate, or an enemy, whereas the detailed finding
of these involves eyes, ears, touch, and an accurate timing system.
There thus arises the distinction between the systems for initiation of
action in the forebrain ('emotive') and for its fulfilment (executive)
by centres farther back.

14. Eyes

An animal provided with suitable receptors can obtain much
information about the environment from the changes in illumination.
Control of the whole physiology to follow the rhythm of day and night
may have been the original reason for the development of photo-
sensitivity in the diencephalon (see p. 105). At the stage of evolution
reached by teleosts information is gained from the fact that light
varies in frequency (colour) and intensity (brightness) and that it is
reflected from many substances, revealing their movement and shape.
The greatest sensitivity of the fish eye is in the yellow-green, which

vii. 14 EYE OF TELEOSTS 213

is the wavelength that penetrates farthest into the water. In order to
extract the maximum of information at high as well as low intensities
it is necessary to adjust the sensitivity, and hence the signal/noise
ratio. For this purpose teleosts have developed retinas with distinct
rods, cones, and twin-cones and in some there is a fovea composed of
numerous thin cones (e.g. in Blennius). The pupil usually varies little
in diameter, and adjustment of sensitivity is by migration of pigment

scleral cartilage

epichorioKJal lymph space

choriad
'glond

Fig. 130. Diagrammatic vertical section of a typical teleostean eye. Not all the structures
here shown are found in all species. (From Walls, The Vertebrate Eye.)

between the receptors and contraction of a 'myoid' segment of the
latter. In bright light the pigment expands, the cones contract for-
ward, towards the light and the rods contract back, beneath the pig-
ment. These photo-mechanical changes thus serve the same end as
changes of pupil diameter in other vertebrates.

The photochemical change in the rods of marine fishes is the same
as that of land vertebrates, namely the breakdown of the rose-
coloured 'visual purple' (rhodopsin) first to the yellow retinene and
then to colourless vitamin A v In freshwater fishes there is a different
pigment porphyropsin, or visual violet, which breaks down to vita-
min A 2 . Intermediates between these may be found.

In all fishes there is a very large, dense, spherical lens, to which is
attached a retractor muscle (campanula Halleri) inserted on to a fal-
ciform ligament, which occupies the persistent choroidal fissure in the
retina (Fig. 130). The eye is usually said to be myopic at rest and to be
accommodated for distant vision by pulling the lens nearer to the

214 BONY FISHES vn. 14

retina. However, this has recently been disputed by Verrier, who
denies that the campanula is muscular and believes the eye to be
hypermetropic at rest and accommodated, if at all, by fibres of the
ciliary body, as in other vertebrates. It may be that the fixed focus is
already sufficiently deep and the campanula perhaps serves mainly
to steady the lens.

It is unwise, however, to generalize about teleostean eyes, for they
are very varied. Whereas the trout, like most, has a round pupil,
which varies little if at all in size, other fishes, whose eyes are more
exposed to light from above, have a more mobile iris. In flat-fishes
and the angler-fishes, such as Lophius, and the Mediterranean
Uranoscopus, the star gazer, the iris has an 'operculum' and is very
muscular; its movements are controlled by nerves and not, as in
selachians, by the direct effect of light. The sympathetic system sends
branches into the head in these animals (Fig. 138) and its fibres cause
contraction of the sphincter of the iris, whereas fibres in the oculo-
motor nerve cause contraction of the dilatator, the opposite arrange-
ment to that in mammals. In the eel the pupil is also capable of wide
changes of diameter, but here the control is mainly by the direct
response of the circular sphincter iridis muscle to light incident upon
it. The pupil of the isolated eye of an eel closes when illuminated and
reopens again in darkness (Fig. 131). Presumably because of its lack
of nervous control this iris is not affected by many of the usual
'autonomic' drugs. For instance, closure will occur in the presence of
atropine and the dark-adapted pupil remains unchanged when placed
in a solution as strong as 1 per cent, pilocarpine, but then closes
immediately on illumination (Fig. 132). The isolated pupil of Urano-
scopus, however, closes when pilocarpine is applied (Fig. 133), and
in this case the sphincter muscle is innervated by sympathetic nerve-
fibres. Adrenaline also causes the sphincter to contract and acetyl
choline in moderate concentrations causes dilatation

The eyes may be small or absent in fishes living in caves, muddy
waters, or the deep sea. In this last habitat, however, many have ex-
ceptionally large eyes, with, apparently, high acuity as well as sensi-
tivity. They may be elongated ('telescopic') and with large binocular
fields and a fovea of 'rods'. In Bathylagus the rods reach a density
of 800,000 mm 2 and are arranged in six superimposed layers, which
presumably come into action successively as an object approaches.
The cells of the deeper layers are less closely packed.

The tropical fish Anableps lives with the head half out of water and
the eyes are adapted for use in both media. The upper part of the

(215)

Red

White light

Red light

8
7

x5

i

14

^3

2
1

i

i

i i i

10

Time,

15

minutes

20

25

Fig. 131. Closure of the pupil of the isolated eel's eye, followed by plotting the move-
ment of its margin with a camera lucida. The movements are shown magnified 54 X ,

Time in minutes.

U-

2-35

Ringer

2;Z0

-2-05
-1-90
1-75
1-60
145
1-30
1-15

Acetul choline . . , L ■■
Vmnnnnn Acetylcholine

1/100.000 Atropine 7100,000

Vl .000.000

Red light

White

Red

15

20 11

35

40

i

45

50

55

Time, minutes

Fig. 132. Changes of diameter of isolated eel's iris in Ringer's solution with the

addition of various drugs. Acetyl choline produces some closure, atropine some

opening. Light still produces closure after application of atropine.

Pilocarpine

Fig. 133. Movements of margin of pupil in isolated iris of Uranoscopus in isotonic

solution. Pilocarpine produces closure and atropine opening of this pupil whose

sphincter muscle is innervated by sympathetic nerve-fibres.

(From Young, Proc. Roy. Soc. B. 107.)

2l6

BONY FISHES

vii. 14-

cornea is thickened, the iris provides two pupils, the lens is pear-
shaped, and there are two retinas in each eye.

15. Ear and hearing of fishes

The ear provides receptors that ensure the maintenance of a correct
position of the fish in relation to gravity and to angular accelerations.
In addition, in many species it serves for hearing. The inner ear is
completely enclosed in the otic bones. There is a perilymphatic space
only in those species that hear well.

ant.

Ns Sag. S. L.

Fig. 134. Diagram of ear of the minnow Phoxinus.

Ast. asteriscus; L. lagena; N.I., N.s. nerves of lagena and saccule;

S. saccule; Sag. sagitta; U. utricle. (From V. Frisch, Z. vergl.

Physiol. 25.)

Each ear sac is subdivided into three semicircular canals and three
other chambers, the utriculus, sacculus, and lagena (Fig. 134). In
each chamber is carried an ear stone (otolith) and these are given
special names, the lapillus, sagitta, and asteriscus, occupying the above
three chambers respectively.

The sensitive macula of the utricle lies horizontally, with the
lapillus resting upon it, whereas the maculae of the saccule and lagena
are vertical. These receptors with otoliths have double or triple
functions. At rest they act as static receptors, signalling the position
of the fish in relation to gravity and setting the fins and eyes in appro-
priate positions. In movement, together with the semicircular canals,
they signal angular accelerations, initiating compensatory movements.
Thirdly, some of the otolith organs respond to sonic vibrations.

In the fishes that hear well there is a connexion between the air
bladder and the ear. This may be either direct, by means of a sac
extending forwards (in Clupeidae and others) or indirectly by a chain
of modified vertebrae, the Weberian ossicles (Fig. 135). This latter
arrangement is found in the freshwater Ostariophysi, which hear

vii. 15 HEARING OF FISHES 217

particularly well. In these fishes the receptors are in the inferior part
of the ear (saccule and lagena, Fig. 134). The sagitta carries a special
wing projecting into the cavity and is so suspended as to serve to
amplify vibrations. Near to it is
a thin portion of the wall of the
sac, which would favour the
passage of variations of pressure
transmitted to the endolymph by
the ossicles.

These Ostariophysi respond to
sounds between about 60-6,000
vibrations/sec. After removal of
the pars inferior responses con-
tinue only up to 120/sec. If the
air bladder is punctured a min-
now can still respond, but only
up to 3,000/sec and with a sen-
sitivity diminished by more than
fifty times.

Minnows can be trained to
discriminate between warbled
notes separated by I tone. Non-
ostariophyse fishes have mostly
a much lower upper limit of
hearing and lower capacity for
discrimination. The Mormyridae
(p. 254), however, approach the
minnows in this respect and here
there is a special isolated portion
of the air bladder within the otic
bone.

In the best cases the sense of
hearing of fishes thus approaches
that of man, in spite of the
absence of a coiled cochlea and

basilar membrane with fibres of different lengths. Clearly the discrimi-
nation of tones cannot here depend upon differential resonance as the
theory of Helmholtz requires. In spite of the considerable powers
of pitch discrimination there is little evidence of capacity to localize
sounds except when they are loud and near.

Sch.

Fig. 135. Position of ear in Ostariophysi

and its relation to the Weberian ossicles,

which are shown in black.

C.tr. transverse canal between the two sacculi;
//. brain; /. 'incus'; L. lagena; M. 'malleus';
S. sacculus; Sch. swim-bladder; S.i. sinus impar
(perilymphatic space); St. 'stapes'; U. utriculus.
(From V. Frisch.)

218 BONY FISHES vii. 16-

16. Sound production in fishes

A surprisingly large number of fishes can produce sounds audible
to ourselves, and these noises are used by the fishes either for shoaling,
or to bring the sexes together, or to warn or startle enemies. Some fish
may use the sounds they produce for echo-location. Among the loudest
of the sounds is that produced by the drum-fish (Pogonias) of the
Eastern Atlantic. The 'whistling' and other noises of the 'maigre'
(Sciaend) are supposed to be the origin of the song of the Sirens, since
they can easily be heard above the water. In both these fishes the
sounds are made mostly if not wholly in the breeding season. In
others, such as siluroids and Diodon, the noise is associated with the
presence of spines and may be a warning. In Congiopodas the nerves
that innervate the muscles of sound production also supply muscles
that raise the spines (Packard).

The mechanism for sound production is very varied, involving
either stridulation by the vertebrae (some siluroids), operculum
(Cottus, the bull-head), pectoral girdle (trigger-fishes), teeth (some
mackerel and sun-fish), or phonation by the air bladder. The latter
may be involved either by its use for 'breathing' sounds in physosto-
matous forms (p. 261) or as a resonator. Noise production is common
in some families (Triglidae, Sciaenidae, Siluridae) but almost absent
from others. The advantages to be obtained from sound production
underwater have led to parallel evolution of similar mechanisms in
several different groups.

17. The lateral line organs of fishes

The lateral line organs occur partly as rows of distinct pits, partly
in canals that communicate with the surface through pores in the
scales. Besides the main canal running down the body and served
by the lateral line branch of the tenth cranial nerve, there are also
lines following a definite pattern on the head, namely, supra- and
sub-orbital lines, a line on the lower jaw, and a temporal line across
the back of the skull. The canals on the head are innervated mainly
from the seventh, partly from the ninth cranial nerve. The nerve-
fibres enter the very large acoustico-lateral centres of the medulla and
valvula cerebelli.

Fishes possess the capacity to react to an object moving some dis-
tance away in the water ('distant touch sense') and this is reduced or
absent after section of the lateral line nerve. Presumably the moving
object sets up currents in the water, which move the fluid (or mucus)
in the canals. It has also been suggested that the canals serve to record

vii. 17

FUNCTION OF LATERAL LINE

219

displacements produced by the swimming movements of the fish
itself, but this has not been proved. Fishes deprived of the lateral
line show no muscular incoordination, although if blind they collide
frequently with solid objects. It has often been suggested that these
organs serve for hearing, perhaps at low frequencies, but this is
probably not so.

Fig. 136. Responses of a single end organ in a lateral canal of a ray, shown with an
oscillograph after amplification. Time signal 10 sec. intervals. The movements of
the continuous white line show A, the beginning of a headward flow, increasing the
frequency of discharge; B, the end of this flow; c, return of spontaneous discharge
after an interval of 28 sec; D, spontaneous discharge 60 sec. later; E, beginning of
a tailward perfusion, inhibiting the discharge; F, the end of this perfusion; G, the
spontaneous discharge 10 sec. later. (From Sand, Proc. Roy. Soc. B. 123.)

Study of the electrical activities of these organs in rays has shown
that many of them discharge impulses all the time, even when not
under the influence of any external stimulation (Fig. 136). By passing
currents of water along the tubes Sand showed that a tailward flow
checks and a headward flow accelerates this 'spontaneous' discharge of
impulses. Such changes in the streams of impulses arriving at the
brain could, no doubt, form the basis for initiation of movements of

220

BONY FISHES

vii. 17-

the fish. This, however, still leaves open the question of what agency
initiates movement of the fluid in the canals during life. It has been
shown that when small streams of water are directed against the side
of the tail some fish make escaping movements, but that these no
longer appear when the lateral line nerve has been cut. The lateral
line organs thus provide signals when agitation of the water causes
pressure changes. In fact they provide the animal with a kind of
water touch, though it is not certain whether this is their only func-

/^v Lat.d.
gust.

max

Lat.v.
pelv. pect.

Fig. 137. Dissection of whiting to show the cranial nerves, and especially the
nerves for the taste-buds.

an. anal fin; gust, gustatory branch of facial; lat.d. and lat.v. dorsal and ventral lateral line

nerves of vagus; mand. and max. mandibular and maxillary divisions of trigeminal; op.

ophthalmic; pect. pectoral fin; pelv. pelvic fin; VII, hyomandibular branch of facial; IX,

glossopharyngeal; X.visc. visceral branch of vagus.

tion. Why this type of receptor should need such a peculiar apparatus
of canals, rather than a system of nerve-fibres in the skin, innervated
by the spinal dorsal roots, is not clear, nor do we know the significance
of the pattern of lines on the head. The lateral line system must cer-
tainly be of great importance in aquatic life, for it is found in all types
of fishes and also in the early Amphibia and in the aquatic larvae of
modern members of that group. The distant touch receptors could
obviously be used in many ways, not only to locate moving objects
and water currents but to serve for echo-location, by computing the
time relation of reflected waves set up by the fish itself.

18. Chemoreceptors. Taste and smell

As in all vertebrates, there are two separate chemical senses, taste
and smell. The former serves mainly to produce appropriate reactions
to food near the body, such as snapping, swallowing, or movements
of rejection. Smell, on the other hand, is a 'distance sense', by which
the whole animal is steered. The distinction between the two types
of receptor is somewhat obscured in bony fishes by the fact that taste-
buds are not restricted as they are in mammals to the tongue and

vii. i8 TASTE AND SMELL 221

pharynx but may occur on the whiskers and all over the body. They
are innervated by branches of the seventh, ninth, and tenth cranial
nerves, which may reach far backwards (Fig. 137). In some species it
has been shown that the fish is able to turn and snap at a piece of food
placed near the tail. This power is lost if the branches from the cranial
nerves are cut. In mammals taste-buds serve to discriminate only four
qualities (salt, sour, bitter, and sweet), most of our so-called 'tasting'
being in reality the smelling of the food in the mouth. In fishes, also,
the four taste qualities are discriminated by the taste-bud system, and
it has been shown that the minnow (Phoxinus) continues to make such
discriminations after the forebrain has been removed. Other chemical
discriminations are made by the nose, however, and can only be
performed with an intact forebrain. Thus Phoxinus tastes and smells
the same classes of substances as man does. The taste-buds are
exceedingly sensitive, the threshold for sweet substances being 500
times and for salt 200 times lower than in man. On the other hand,
some substances that are very bitter for us produce little reaction in
Phoxinus.

In many fishes the nose is one of the chief receptors (macrosmatic).
There are two nostrils on each side, allowing for the sampling of a
stream of water (Fig. 119). The nose does not communicate with the
mouth, except in a few fishes that live buried in the sand (Astroscopus).
The sense of smell is used to find food and for recognition of the
sex of members of the same species. Minnows can be trained to give
distinct reactions to extracts made from the skin of other species of
fish living in fresh water. In the presence of 'alarm substances' pro-
duced by damaged skin of a member of the same species, minnows
(and other fishes) show a 'fright reaction', scattering and refusing food.
The state of development of the nose is very varied. It is large in
macrosmatic solitary predators such as Anguilla and in many schooling
species that also have well-developed eyes {Phoxinus, Gobio). Daylight
predators, on the other hand, are microsmatic (Esox, Gasterosteus).
Other evidence shows that fishes can discriminate between the smells
of water plants and between the waters of different streams. It is
likely that this provides part of the mechanism by which salmon
return to the stream in which they were born, having been conditioned
as fry to the smell of its water. It has been suggested that they might
be decoyed to return to a stream other than that where they were
hatched by conditioning them as fry to a substance such as morpholene
to which they have a high sensitivity although it is neither an attract-
ant nor repellant.

222

BONY FISHES

vii. 19-

Fig. 138. Diagram of ventral view of the
sympathetic system of the front part of the
body of Uranoscopus, showing the fibres
in the sympathetic and oculomotor that are
responsible for the light reflex.
(From Young.)

cil.brev. cil.long., short and long ciliary nerves;
cil.gn., ciliary ganglion; dors.r. dorsal root; dil.
dilatator muscle; hypo, hypoglossal; n.splanch.
splanchnic nerve; opt. optic nerve; p. pupil;
pal. VII, palatine branch of facial ; prof, profun-
dus; r.b. short root of ciliary ganglion; r.comm.
ramus communicans; r.l. long root of ciliary
ganglion; sph. sphincter muscle; ventr.r. ven-
tral root; 1II-X, cranial nerves with their
sympathetic ganglia (V.symp. &c.); VII hyo.
hyomandibular branch of facial; r— 4 sp.symp.
spinal sympathetic ganglia.

19. Touch

Touch is, of course, well
developed in fishes, and in many
species there are special sensory
filaments, which presumably serve
this sense. They are usually de-
veloped around the mouth, as in
the catfish; in other fishes they
are modifications of the fins, for
instance, the pectoral fins of
gurnards, which also contain
chemoreceptors.

20. Autonomic nervous system

The autonomic nervous system
of bony fishes is organized on a
plan rather different from that
both of elasmobranchs and of
land animals. There is a chain of
sympathetic ganglia, extending
from the level of the trigeminal
nerve backwards, a ganglion
being found in connexion with
each of the cranial dorsal roots
(Figs. 138 and 139). These gan-
glia do not receive pre-ganglionic
fibres from the segments in which
they lie, but by fibres that run out
in the ventral roots of the trunk
region and thence forwards in the
sympathetic chain. This emer-
gence of the pre-ganglionic fibres
for the head in the trunk region
recalls the arrangement in land
animals.

Each trunk sympathetic gan-
glion, besides receiving a white
ramus communicans of pre-gan-
glionic fibres from its spinal nerve,
also sends a grey ramus back to

VII. 20

AUTONOMIC NERVES OF TELEOSTS

223

that nerve, this ramus carrying post-ganglionic fibres to the skin.
Some of these fibres control the melanophores, causing them to
contract (p. 259). In elasmobranchs there are no grey rami communi-
cantes and no sympathetic system in the head (p. 173); the differences
between the two groups are therefore very striking.

HI prof Viymp

Fig. 139. Diagram of the autonomic nervous system of Uranoscopus seen from the side.

bl. mesonephric bladder; cil.gn. ciliary ganglion; dors, dorsal root; n.sph. nerve to anal
sphincter; n.spl. splanchnic nerve; prof, nervus ophthalmicus profundus; rad. brev. short
root of ciliary ganglion; r.comm. ramus communicans (including both white and grey fibres);
stan. 'corpuscle of Stannius' (adrenal cortical tissue?); ventr. ventral root; ///, oculomotor
nerve; V— X symp. sympathetic ganglia associated with the cranial nerves. (From Young,

Quart. J. Micr. Sci. 75.)

Fig. 140. Tracing of the contractions of a strip of the stomach muscle of the angler-
fish, Lophins, attached to a lever. Time in minutes. At A, faradic stimulation of the
vagus nerve. Drugs then added to the solution to make, at B acetyl choline 1/1,000,000;
at c acetyl choline 1/100,000; at D adrenaline 1/100,000.
(From Young, Proc. Roy. Soc. 120.)

Little is known about the parasympathetic system of bony fishes.
The oculomotor nerve carries fibres to the iris, which work in the
opposite direction to fibres from the sympathetic (p. 214). There is
also a well-developed vagal system, but so far as is known no para-
sympathetic fibres in other cranial nerves and probably no sacral
parasympathetic system. Electrical stimulation of the vagus nerve

224 BONY FISHES vn. 20-

produces movements of the stomach but not of the intestine; the
latter, however, shows movements when the splanchnic nerve is
stimulated. In most of the viscera acetyl choline causes initiation of

nnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnr

11111111111J1I.

Fig. 141. Tracing of contractions of the muscle of the urinary bladder of Lopliius,

attached to a lever. At A, D, and F faradic stimulation of vesicular nerve. Drugs added

to make, at b acetyl choline 1/2,000,000; at c adrenaline 1/500,000; at E ergotoxine

1/50,000. Time, minutes. (From Young.)

Fig. 142. Tracing to show effect of atropine 1/50,000 added at A, on the contractions

of the bladder of Lophius produced by faradic stimulation of the vesicular sympathetic

nerve. Time, minutes. (From Young.)

rhythmic contractions and these are inhibited by adrenaline (Figs.
140 and 141). In Lophius this is true of the stomach, with motor-fibres
from the vagus, intestine, with sympathetic motor-fibres, and of the
muscles of the bladder, which contract on stimulation of the hinder
sympathetic ganglia. However, in the trout adrenaline causes contrac-
tion of the stomach (Burnstock, 1958). The effect of the nerves to the

vii. 21 BEHAVIOUR OF FISHES 225

bladder is prevented by atropine (Fig. 142) but not by ergotoxine
(Fig. 141), though the latter is the drug that in mammals often inhibits
sympathetic motor-fibres. In these fishes, therefore, it is not possible
to divide up the autonomic nervous system into sympathetic and
parasympathetic divisions by either anatomical, physiological, or
pharmacological criteria. Presumably the two 'antagonistic' systems
found in mammals are a late development, allowing for a delicate
balancing of activities for the maintenance of homeostasis.

21. Behaviour patterns of fishes

The well-developed receptors and brain of the teleostean fishes
constitute perhaps the most important of all factors in giving them

Fig. 143. The red belly of the stickleback releases attacking behaviour in other males

and following by females. Of the above models only the two on the left acted as

releasers. (From Tinbergen, Wilson Bulletin, 1948.)

their great success. Varied habits and quick actions enable the fish to
make full use of the possibilities provided by the special features of
their structure — the air bladder, mouth, and so on. The receptors and
brain make it possible for the fish to learn to react appropriately to
many features of its surroundings. Thus the eyes besides orientating
the fish to movements in the visual field allow the discrimination of
wavelengths and distinct reactions to differing shapes (see Bull, 1957).

The social behaviour of many species includes the development of
special 'releasers', shapes, colours, or postures that are displayed by
one individual and elicit specific reactions in another (Fig. 143).

There is no doubt that fishes possess great powers of learning.
They can form conditioned reflexes involving discrimination of tones,
also second-order conditioned reflexes, in which after the animal has
learnt to give a certain behaviour in response to a visual stimulus it is
then taught to associate the latter with an olfactory stimulus. There
are many other examples of such powers, but unfortunately we have
as yet little information as to the way in which they are brought about
by the brain. Nor have the naturalists provided us with very clear
examples of the use of these powers by fishes in nature. There are

226

BONY FISHES

VII. 21

many tales of carp coming to be fed at the ringing of a bell, and similar
powers of association must play a part in the life of fishes in more
natural situations. Bull has shown that fishes can be trained to dis-
criminate between very small differences of water flow, temperature,
salinity or pH, and no doubt it is by means of such powers that they
normally find a suitable habitat.

Fig. 144. Migrations of the eels. The European species (A. anguilld) occurs along the
coasts outlined with lines, the American species {A. rostrata) where there are dots.
The curved lines show where larvae of the lengths indicated (in millimetres) are taken.

(After Norman.)

The migrations of fishes have attracted much attention, but are
still imperfectly understood. They vary from the 'catadromous' down-
ward migration of young animals to the sea and the reverse 'anadro-
mous' movement to breed, to the astounding journeys of the eels,
3,000 miles westwards from Europe or eastwards from America to
their breeding-place in the Sargasso Sea (off Bermuda) and the return
of the elvers to the homes of their parents (Fig. 144). No one has yet
discovered the factors that direct these movements, currents may play
a part, but can hardly be the only influence. Indeed it has been
suggested that European eels never complete the journey but die in

vii. 21 MIGRATION OF EELS 227

their own continental waters. The populations of so-called European
eels {Anguilla anguilld) would then be maintained by reinforcements
of larvae of the American A. rostrata, the differences between the
two being due to temperature and other factors (Tucker, 1959).

Social behaviour is marked in many species and shoals of some
fishes may contain many thousands of individuals. The animals are
presumably kept together in most species by visual stimuli, though
sounds may play a part. Shoaling gives protection to small fishes, and
in some species the animals come together in shoals to breed (her-
rings). There may also be some advantage for the finding of suitable
feeding conditions, but on all these points we can do little more than
speculate and hope for further information.

VIII

THE EVOLUTION OF BONY FISHES

1 . Classification

Class Actinopterygii
Superorder i. Chondrostei

Order i. Palaeoniscoidei. Devonian-Recent

*Cheirolepis; *Palaeoniscus; * Amphicentrum; *Platysomus,
*Dorypterus ; *Cleithrolepis; *Tarrasius; Polypterus, bichir
Order 2. Acipenseroidei. Jurassic-Recent

*Chondrosteus; Acipenser, sturgeon; Polyodon, paddle-fish
Order 3. Subholostei. Triassic-Jurassic
*Ptyeholepis
Superorder 2. Holostei. Triassic-Recent

*Acentrophorus; *Lepidotes; *Dapcdius; *Microdon; Amia,
bowfin; Lepisosteus, gar-pike
Superorder 3. Teleostei. Jurassic-Recent
Order 1. Isospondyli

*LeptoIepis; *Portheus; Clupea, herring; Salmo, trout
Order 2. Ostariophysi

Cyprinus, carp; Tinea, tench; Silurus, catfish
Order 3. Apodes

Angnilla, eel; Conger, conger eel
Order 4. Mesichthyes

Esox, pike ; Belone; Exoeoetus, flying fish ; Gasterostens, stickle-
back; Syngnathns, pipe-fish; Hippoeampus, seahorse
Order 5. Acanthopterygii

*HopIopteryx; Zens, John Dory; Perca, perch; Labrus, wrasse;
Uranoscopns, star gazer; Blennins, blenny; Gadus, whiting;
Pleuronectes, plaice; Solea, sole; Lophius, angler-fish

2. Order 1. Palaeoniscoidei

The actinopterygian stock has been distinct since Devonian times.
The early representatives lacked many of the specializations that we
find in the successful bony fishes today and showed features of simi-
larity to the Crossopterygii. These Devonian Actinopterygii had not
yet acquired the striking signs of full mastery of the waters, which are
so characteristic of the group today. They resembled their ancestors
the placoderms and their cousins the crossopterygians in being rather

VIII. 2

PALAEONISCIDS 229

clumsy, heavily armoured creatures. From this early type many lines
have been derived and can be followed with some completeness to
their extinction or modern descendants. Various classifications have
been suggested. The one used here is simple but for that very reason
obscures the multiplicity of parallel lines. A recent classification
recognizes fifty-two orders of Actinopterygii (Grasse).

d.

Fig. 145. Scales of some early fishes.
A, hypothetical condition with denticle-like substance (d.) attached to a basal bony plate lying
in the connective tissue (ct:)\ B, 'cosmoid' scale of early crossopterygians, showing the cosmine
layer (co.) ; epidermis (ep.) ; vascular canals (hv.) and underlying 'isopedin' {is.); C, palaeoniscoid
scale with layers of 'ganoin' (ga.); d, lepidosteoid scale of the gar-fish with tubules (/.). (From
Goodrich, Vertebrata, A. & C. Black, Ltd.)

The Devonian and Carboniferous forms are grouped together in
the order Palaeoniscoidei, and animals of similar type survive today
as Polypterus, the bichir of African rivers, which though showing some
specializations remains in its general organization near the palaeoniscid
level.

A typical Palaeozoic palaeoniscid such as *Cheirolepis was a long-
bodied creature (Figs. 146 and 147) with a heterocercal tail, single
dorsal fin, and pelvic fins placed far back on the body. The pectoral
and pelvic fins had broad bases and the radials fanned out from a small
muscular lobe, present in all early actinopterygians but lost in later
forms. The body was covered with thick rhomboidal scales very
similar to those of acanthodians. They articulated by peg and socket
joints and have a structure known as palaeoniscoid (Fig. 145). The
scale is deeply embedded and grows by addition both to the bony or
isopedin portion and to the shiny surface-layer, the ganoin, which
thus becomes very thick. There is a middle layer of 'pulp' correspond-
ing to the cosmine layer of the cosmoid scale of Crossopterygii

230 BONY FISHES vm. a

(p. 269) and the two types have obvious similarities, though it is not
clear how they are related.

The skull was built on a distinctly different plan from that of
Crossopterygii, in that there was no joint such as was present in those
fishes to allow the front part to flex on the hind. The jaw support was
amphistylic, in the sense that the palatoquadrate was attached to the
neurocranium by a basal process, but the otic process did not reach
the skull and the hind end of the jaw was supported by the hyomandi-
bula. There were even more dermal bones than are found in modern
Actinopterygii, arranged so as to form a complete covering for the
chondrocranium and jaws. These bones were derived from the original
scaly covering of the head and the naming and comparing them with
the bones of other forms is a matter of some difficulty. Some of the
main bones resemble in appearance and shape those found in tetra-
pods, but there are others for which no such homologues can be
found, and sometimes there is considerable difficulty in recognizing
even the main outlines of the pattern. The problem is that we have
no rigid criterion by which to set about giving names to the skull
bones. No system yet discovered is wholly satisfactory, and we must
admit to insufficient knowledge of the factors that determine that
bone shall be laid down in certain areas and that sutures shall separate
these from each other. However, some of the dermal bones lie in
relation to the lateral line canals (or rows of neuromasts), which latter
may provide the stimulus to bone formation. The lines are remarkably
constant, perhaps because of their function in detecting water move-
ments in relation to swimming, and this is the factor that determines
the position of many of the bones. Others fill in the spaces between
(anamesic bones). Yet others may be differentiated in relation to the
teeth. However, the number of bones along any one line may vary
greatly even in one species (e.g. in Amid). The whole pattern is more
variable in fishes than in higher vertebrates, but it is usual to consider
that the bones of early Actinopterygii resemble those of Crossopterygii
and of the early amphibians (Fig. 194).

The roof of the skull usually shows a large pair of parietals between
the eyes, and post-parietals behind these. Between the parietals and
the nostrils there are frontal bones and the front of the head usually
also carries a number of rostral bones, not found in higher forms.
Behind the post-parietals in the midline is a series of extrascapular
bones.

The side of the skull of palaeoniscids is covered by numerous bones,
including a series of pre- and post-frontals, post-orbitals, and jugals

VIII. 2 PALAEONISCIDS 231

around the eyes. The outer margin of the upper jaw is covered by
premaxillae and maxillae, which are the main tooth-bearing bones.
Behind the orbital series of bones the cheek is very variable. Sometimes
there is a large bone identifiable as a pre-opercular, with a series of
opercular bones behind it. The lower portion of the throat was covered
by a series of gular plates. The spiracle in these early forms opened
above the opercular bones. The pectoral girdle was attached to the
back of the skull by a supracleithrum, below which a cleithrum and
clavicle made a series of dermal bones behind the gills, covering the
cartilaginous girdle. The roof of the mouth contained a median para-
sphenoid, with paired prevomers in front of it, and a series of
pterygoid bones occupied the space between it and the edge of the
jaws, the palatine, ectopterygoid, pterygoid, and sometimes others.
Finally the lower jaw, besides the main dentary carrying the teeth,
shows many small bones such as the pre-articular and coronoid on
the inner surface; splenial, angular, and surangular on the outside.

It will be clear that this skull of *Cheirolepis may be closely com-
pared with the skull of a crossopterygian or a modern teleostean.
The general plan is related to that of the lateral line organs arranged
along occipital, supratemporal, and infra-orbital lines. The numerous
small bones are evidently similar in the different groups, though it is
not easy to assign a suitable name to every one of the more numerous
bones of the earlier forms.

These palaeoniscids from the Middle Devonian were rather rare
freshwater fishes ; they had sharp teeth and probably lived on inverte-
brates. We have no information about their internal anatomy, but it
seems not unlikely that the air-bladder possessed a wide opening to
the pharynx (as it still does in Polypterns, descended from this stock)
and that they breathed air, as did other Devonian fishes. However,
they did not have internal nostrils, which are found in the old crosso-
pterygians.

During the Carboniferous and Permian the palaeoniscids were
numerous, mostly as small, sharp-toothed fishes. Several distinct lines
became laterally flattened and acquired an outwardly symmetrical tail
and blunt crushing teeth (Fig. 147). These characteristics probably
indicate a habit of feeding in calm waters, perhaps mainly on corals,
and they have appeared several times in the actinopterygian stock
(p. 241). Palaeoniscids of this type were formerly placed together in
a family Platysomidae, but it is now considered probable that the
type arose independently several times; thus * Amphicentrum is found
in the Carboniferous, *PIatyso?nus and *Dorypterus in the Permian,

(232)

Solea

Uranoscopus

Fig. 146. Various actinopterygians.

viii. 2 PALAEONISCIDS 233

*Cleithrulepis in the Triassic. Similar forms arose again later among
the holosteans and teleosteans and we have therefore evidence that
this type of animal organization tends to evolve into deep-bodied
creatures. *Dorypterus further resembles modern teleosteans in a great
reduction of its scales and in the forward movement of the pelvic fins.

Towards the end of the Triassic animals of typical palaeoniscid
type became rare; they were replaced by their more active and speedy,
mainly marine descendants, the Holostei (p. 234). Certain of the lines
that branched off in the Palaeozoic have, however, survived to the
present time, and in spite of subsequent specializations they give us
some idea of the characteristics of these early Actinopterygii. Perhaps
the most interesting of these survivals are Polypterus, the bichir, and
the related Calamoichthys, both inhabiting rivers in Africa. The air-
bladder shows some similarity to a lung. It forms a pair of sacs lying
ventrally below the intestine and opening to the pharynx by a median
ventral 'glottis' (Fig. 157). This is the arrangement found in lung-
fishes (except Ceratodus) and in tetrapods, and it seems reasonable to
suppose that it has survived in Polypterus from Palaeozoic times.
However, it is not certain to what extent the air-bladder is still used
as a lung, for Polypterus cannot survive out of the water.

This fish shows many other ancient characteristics. The covering
of thick rhomboidal scales, hardly overlapping, gives the animal an
archaic appearance; the structure of the scales is 'palaeoniscoid'. In
the skin there is a layer of denticles outside the scales. The presence
of a spiracle, the arrangement of the skull bones, and many other
features suggest that Polypterus is essentially a palaeoniscid surviving
to the present day. In the intestine there is a spiral valve, which
appears to have been present in the early Crossopterygii and Actino-
pterygii (as judged from fossilized 'coprolites') and occurs today not
only in the Dipnoi but also in sturgeons and, though much reduced, in
Lepisosteus and Amia. There is a single pyloric caecum in Polypterus
(the caeca are well developed in sturgeons, Lepisosteus, and Amid).
The tail of Polypterus is no longer markedly heterocercal, but shows
distinct signs of that condition. We can even find a parallel among
Carboniferous palaeoniscids for some of the special features of Poly-
pterus. The long body and dorsal fin are found in the fossil *Tarrasius,
which may have been close to the ancestry of Polypterus, though it
lacks the covering of scales. The pectoral fin in *Tarrasius, as in
Polypterus, has a peculiar lobed form, which has been compared with
the 'archipterygial' pattern (p. 269) and hence held to show that
these animals are related to the Crossoptergyii. The resemblance is,

234 BONY FISHES vm. 2-

however, only superficial and the plan of the fin is essentially actino-
pterygian. In the brain there is a thin pallium, thick corpus striatum,
and a valvula cerebelli. The pituitary is remarkable in that the hypo-
physial sac remains open to the mouth. In this and other features
(persistent pronephros) there are signs of neoteny.

3. Order 2. Acipenseroidei

The sturgeons are a rather isolated line descending from the palaeo-
niscids and characterized by reduction of bone. This was already
apparent in the Jurassic *Chondrosteus. Acipenser and other modern
sturgeons live in the sea but migrate up the river to breed. They may
reach a very large size (1,000 kg) and since a tenth of this is caviar
they are exceedingly valuable. They feed on invertebrates, which they
collect from mud stirred up from the bottom by a long snout. This is
flattened into a pear-shaped structure in Polyodon, the purely fresh-
water paddle fish of the Mississippi and in Psephurus in China. The
mouth of all sturgeons is small and the jaws weak and without teeth.
In Polyodon there is a filtering arrangement of gill-rakers in the
pharynx. The jaws of sturgeons hang free from the hyomandibular
and symplectic, and can be swung downward and forward during
feeding. The skull and skeleton is almost wholly cartilaginous and the
dermal skeleton much reduced. The tail is covered with rhomboidal
scales, but on the front of the body there are five lines of bony plates
bearing spines, with the skin in between carrying structures similar to
denticles. There is an open spiracle. The internal anatomy of the
sturgeons shows various features that have been held to show affinity
with the elasmobranchs ; for instance, besides the spiral valve there is
a conus arteriosus in the heart and a single pericardio-peritoneal canal.
However, there can be no doubt that they are descended from an
early offshoot from the actinopterygian line. They retain some fea-
tures lost by most members of the line, but resemble the Teleostei in
other characters, for instance a thin roof to the cerebral hemispheres.

The palaeoniscids and sturgeons may be grouped together in a
Superorder Chondrostci and placed with them is a third Order Sub-
holostei, probably a mixed group, including forms that resemble
palaeoniscids, but show various trends towards the holostean grade
of organization i*Ptycholepis).

4. Superorder 2. Holostei

During the later Permian period the palaeoniscids gave rise to
fishes of a different type, which replaced their ancestors almost com-

viii. 4 HOLOSTEANS 235

pletely during the Triassic and flourished greatly in the Jurassic. We
may group together the fishes of this type as Holostei but the term
is used variously by different authors and includes several lines, whose
relationships are not clear. The earliest holostean, *Acentrophorus
from the upper Permian, is much like a paleoniscid but with a small
mouth, shorter, deeper body and slightly upturned tail. This 'abbrevi-
ated heterocercal' tail was presumably made possible by the changed
swimming habits resulting from the use of the air-bladder as a hydro-
static organ. If the fish floats passively there is no need for a hetero-
cercal tail to direct the head downwards (p. 140). Similarly, the head
does not need to be flattened to produce an upward lift. The develop-
ment of the air-bladder has thus made possible the lateral flattening
and shortening of the body so characteristic of later Actinopterygii.
The body of holosteans was