Skip to main content

Full text of "Aquatic mammals; their adaptations to life in the water"

See other formats


i^<' 'jf;^* ii;;--^ 





I T 




.Aquatic Mammals 

Frontispikch. Convei\t?cnce of bt)(Jy form in three classes of vertcbiates: shark, 
a fish, above; ichthyosaur, a reptile; and porpoise, a mammal, below. 



Their Adaptations to Life in the Water 


Lecturer in Comparative Anatomy 
Department of Anatomy 

The Job /IS Hopkins University 





With fifty-four illustrations 








Dedicated to those who 

are endeavoring to save from 

commercial extinction the great 

whales J the largest animals 

that have ever lived. 




WOODS HOLE, ivlAbS. \ 
W. H. 0. i. 


Jtor some years I have been investigating the functional anatomy of 
mammals speciaHzed for particular modes of progression and endeavor- 
ing to correlate the facts discovered with their morphology and habits. 
I have already published several papers of a technical nature on the 
anatomy of aquatic mammals and the present book is the result of these 
studies. My endeavor is to present the subject in an essentially non- 
technical manner, for it is believed that it is of sufficient interest to 
appeal to a reading public which is larger than that contained in the 
anatomical or even the purely mammalian field. Partly for this reason 
lengthy and tedious descriptions and impressive array of data have 
been omitted. 

The subject of aquatic adaptation in mammals is somewhat peculiar 
in that there are numerous points which are difficult or impossible of 
thoroughly conclusive investigation. Habits are at times extremely 
hard to study and many of the specializations are of such a character 
that there are no intermediate stages with which they can properly be 
compared. As a result there has been advanced a truly appalling list 
of improperly founded theories in attempted explanation of these points. 
This, however, has perhaps been justified for the reason that only by 
this means can much progress be made. Many of the specializations 
are so unusual that in order finally to gain an understanding of them 
it is first necessary to have a variety of plausible theories from which to 

The present contribution is far from the final word on the subject. 
I believe that I have herewith shown that some of the theories so far 
tentatively accepted are really untenable, and I have advanced others 
that to me seem better founded, but some of my readers will justifiably 
hold contrary opinions, for in morphological interpretations there is 
invariably produced some chaff with the wheat and further discoveries 
will surely oblige some revision. In fact I am constantly revising my 
own beliefs as my knowledge of conditions increases. I have endeavored 
to evaluate every hypothesis that seemed to hold merit. In attempting 
to arrive at the proper explanation of some detail I have frequently ad- 
vanced two or more essentially diverse theories, possibly without in my 
own mind accepting either of them, but discussing them pro and con 


in order that the reader may understand the various angles of the case. 

It is thus desired to emphasize the fact that the present report is 
more purely philosophical in character than would be justifiable in 
almost any other branch of mammalogy. We really know so little re- 
garding the multitude of unique aquatic specializations that were state- 
ments confined to proven facts the most interesting parts of this fascinat- 
ing topic would be omitted. 

In formulating theories and endeavoring to interpret conditions I 
have taken frequent opportunity to discuss problems with my colleagues. 
Often have they made remarks and offered suggestions that have given 
me an entirely new angle from which to approach some particular de- 
tail. For this I am appropriately grateful, and although I have tried 
always to make acknowledgement at the proper place, still I have found 
it difficult to give stated credit where credit may really be due. I can 
and do, however, give full measure of appreciation to Remington Kel- 
logg, whose unflagging interest in my undertaking and broad knowledge 
of whales, unstintedly shared, have proven quite indispensable. 










The mechanics of swin-isnin^ 




Aquatic mammals . 




External features . 




The Senses 
Olfactory . 





Mouth and nose 
Adipose cushion . 




The skull . 




The neck 


Chapter VIII 

The trunk 




The tail 




The pectoral limb . 




The pelvic limb 




Other soft parts and physiologic 

al features 


Literature cited .... 





Convergence of form in shark, ichthyosaur, and porpoise 

1 Swimming principles of vertebrates 

2 Swimming movements of eel and fusiform fisti 

3 Some of the smaller aquatic mammals 

4 The insectivore otter Potomogale 

5 Pinniped postures . 

6 Sirenians .... 

7 Body postures of the manati 

8 Facial postures of the manati 

9 Types of the larger whales 

10 Heads of porpoises . 

11 Some extinct, aquatic reptiles 

12 Pharyngeal region of Kogia 

13 Blow-holes of whalebone and toothed whales 

14 Skeletons of pinnipeds . 

15 Blow-hole action in a porpoise 

16 Skulls of sirenians . 

17 Skull of a toothed whale . 

18 Skull of the sperm whale 

19 Skull of a whalebone whale 

20 Skull of the right whale . 

21 Extremes of cetacean neck vertebrae 

22 Dorsal musculature of a sea-lion . 

23 Dorsal musculature of a seal . 

24 Ventral musculature of a sea-lion 

25 Ventral musculature of a seal . 

26 Panniculus carnosus musculature of aquatic mammals 

27 Lateral musculature of the narwhal 

28 Skeletons of toothed whales 

29 Skeletons of zeuglodonts . 

30 Vertebrae of a porpoise . 

31 Sterna of whales and sirenians 

32 Development of the cetacean flukes 

33 Caudal restorations of ichthyosaurs 

34 Pectoral limb bones of some aquatic reptiles . 

35 Pectoral limb bones of pinnipeds and sirenians 






































36 Pectoral limb bones of some cetaceans 

37 Muscle attachments of the cetacean pectoral limb 

38 Medial musculature of the sea-lion pectoral limb 

39 Medial musculature of the seal pectoral limb 

40 Forearm bones of the Steller sea cow 

41 Outlines of cetacean and pinniped flippers 

42 Fore feet of some amphibious mammals . 

43 Hind feet of some amphibious mammals 

44 Hind feet of pinnipeds .... 

45 Hind limb of seal and sea-lion 

46 Hind limb musculature of the sea otter . 

47 Hind limb musculature of Potomogale 

48 Hind limb musculature of Desmana 

49 Pelvic bones of seal and sea-lion 

50 Pelvic bones of sirenians . 

51 Pelvic bones of cetaceans 

52 Femora of sea-lion and seal 

53 Foot bones of sea-lion and seal 



In such an investigation as the following the writer is forcibly im- 
pressed with the soundness of that tenet expressed by Jennings (1924) 
when he said that "under the same conditions objects of different ma- 
terial behave diversely; under diverse conditions objects of the same 
material behave diversely", and that "neither the material constitution 
alone, nor the conditions alone, will account for any event whatever; 
it is always the combination that has to be considered". 

In considering such a phenomenally specialized organism as the 
whale and studying it part by part one is also forcibly impressed by the 
staggering length of time that it must have taken to produce such a 
creature, and the intricacy of the evolutional processes that have shaped 
countless anatomical details to effect proper interaction and interrelation- 
ship of parts to a single harmonious end. 

The evidences of evolution are most abundant upon every hand, and 
still we know so very little about them. Tentative theories are advanced 
and discussed, but the ways of evolution are so exceedingly deliberate 
that the lifetime of a single investigator is not sufficiently long for him 
to observe its natural processes. O. P. Hay (1928) has stated "A learned 
writer on mammals tells me he doubts that a single new species has de- 
veloped since the first interglacial stage", which was "perhaps 400,000 
years" ago. If this be the case I do not see how the whales have evolved 
from a terrestrial ancestor in less than one hundred million years, and 
two hundred million might be all too short. How, then, can we expect 
successfully to investigate the processes of evolution in the laboratory.^ 
This element of vast time should be stressed, and also, although we 
know little about the ways of evolution, we cannot help formulating 
tentative hypotheses of explanation. 

The writer is strongly convinced that we cannot look to any one 
theory to explain evolution. The whale is very different indeed from 
the generalized type of mammal in a great many major details, and hun- 
dreds and even thousands of minor details, involving billions of cells. 
These items have been changed doubtless by numberless evolutional 
stimuli, some simple and others complex, involving the inherent tenden- 
cies and limitations of the organism, antagonistic influences, and all 
the intricacies of the relationship of an animal to its environment. 



The reader will doubtless discover in die following pages that the 
writer subscribes to a modification of the Lamarckian theory, or theory 
of the inheritance of acquired characters, perhaps better expressed as 
inheritance based upon the use or disuse of parts. This theory has 
fallen into disfavor for the reason that no one has been able to demon- 
strate it, but that is no reason for ignoring its possibility. The failure 
to prove it doubtless lies in the probability that it operates too slowly 
to be discernable in the lifetime of one man. There have been futile 
experiments such as the one in which the tails of white mice were re- 
moved throughout several generations in the expectation of producing 
a race of bobtails. Naturally this was foredoomed to failure under 
any circumstances for account was not taken of the fact that amputation 
of the tail did not remove the mouse's need for a tail as a balancing 
organ; or that the germ plasm remained unaltered. But if we could 
cause white mice to dig industriously in the ground for several hours 
each and every day, and oblige all their descendants to do likewise with 
conscientiousness, then would there be produced, according to my 
belief, a race of mice with feet much more specialized for digging: not, 
emphatically in twenty or a hundred generations, but in twenty or a 
hundred thousand generations. 

In a study of this kind one frequently encounters the puzzling condi- 
tion illustrated by the following situation. In the rorquals the narrow 
part of the tail — the peduncle — has built up a sharply ridged keel, both 
above and below, of fibrous tissue which acts to reduce water resistance 
during swimming. The broad part of the tail — the flukes — in this 
animal has built up precisely the same sort of fibrous expansions, but 
for the purpose of increasing water resistance. Thus the same sort of 
structure has been developed in response to stimuli which appear dia- 
metrically opposite. This is entirely beyond our comprehension, and if 
the condition is referred to at all it is usually in the most cautious, 
evasive terms. All that can be told is that often a structure will see^n 
to develop where there has clearly become a need for it without our 
being able to distinguish any really directive stimulus at all, in spite 
of the fact that logic dictates that a new modification cannot anticipate 
a new function. 

In considering the specialization exhibited by aquatic mammals one 
must constantly bear in mind that the results seen depend upon the 
strength of the stimuli involved, the capacity of the animal to respond 
to them, the strength of possible antagonistic stimuli, and the length 
of time involved. The aquatic evolution of one particular type of 



mammal may be in a straight line with slight deviation, while in the 
case of another, perhaps handicapped by some bodily equipment less 
readily adaptable to aquatic requirements, there may be much trial and 
error, with the employment of temporary makeshifts. Thus one pair 
of limbs may become quite highly modified for propulsion, but yet 
incapable, because of some mechanical defect in the method of employ- 
ment, of developing as high speed as can the altered tail, in which case 
the latter is apt to take over the duty of locomotion and the former fall 
into disuse. But this is so only up to a certain point, and the "law 
of irreversible evolution" is believed to prevent the redevelopment of 
a part once atrophy is very definitely under way. The sea-lion might 
develop its muscles so that the hind limbs would become the chief or- 
gan of propulsion, but whales have lost their hind limbs forever. 

Thorough modification for an aquatic life by a mammal involves 
not only more numerous but more profound anatomical and physiological 
alteration than for any other sort of existence. The volant specialization 
of the bat has resulted in considerable change; but this has concerned 
only lengthening or rotation of bones, development of membranes, and 
not particularly profound muscular modification. In the whale, how- 
ever, the evolution of bodily details has been truly astonishing, and to 
only a slightly lesser extent in sirenians. Thus in mammals the aquatic 
life is the only one that has resulted in the complete loss of the external 
portions of the hind limb, acquisition of supernumerary phalanges to 
the digits, migration of the nostrils to the top of the head, and entire 
inability to travel for at least a few yards when placed on land. The 
entire specialized development has been confined to the ends of securing 
food and of locomotion involving but a single pair of major movements. 
In some sorts the former specialization docs not have much eff^ect upon 
body form, so that as far as concerns external conformation the only 
major stimulus has been a simplified propulsive one. This being the 
case it should not cause undue astonishment that the most perfect degree 
of highly specialized convergence in bodily form to be found between 
different vertebrate classes is furnished by some species of shark, ichthyo- 
saur, and whale (frontispiece) . 

But while the external form, and even some of the internal anatomy, 
of a whale may be said to be simplified, internal alteration from the 
normal mammalian sort is often so profound in both quantity and 
quality that one may be at a complete loss to interpret it correctly. And 
yet all organs of the terrestrial mammal are present, even though they 
may be vestigeal. The change from what we consider as the mammalian 



norm has been insensibly deliberate, and if we could but have the com- 
plete picture before us there would be no difficulty in its correct in- 
terpretation. The change which the abandonment of the atmospheric 
for the completely aquatic environment necessitated was prodigeous, 
but nevertheless it is probable that with few exceptions (as the sperma- 
ceti organ of the cachalot) there have been no fundamentally new de- 
tails developed, but only the gradual change and specialization of old 

In discussing aquatic modification, and most other things, the majority 
of writers are prone to consider that there is one, or at most a few, 
stimuli operating to accomplish precisely the same end. In other words 
it seems to be a common belief that when mammals take to the water 
there will be a tendency for them all to develop along precisely the 
same lines to exactly the same end. This is true to only a definitely 
limited extent. A dozen, say, types of terrestrial mammals taking 
to the water may have twelve individually peculiar organizations, differ- 
ing in their capabilities for variation, amenable to diverse stimuli of 
dissimilar strength, with twelve combinations of interdependence of 
parts. The convergence which they will ultimately show will be in 
two ideal types of body form and toward two (presumably) final 
methods of swimming. But these ends may be attained in considerably 
different manners, involving entirely different specialization of details, 
and one must be cautious in stating that the aquatic life will result in 
such an item as a shortened neck without a very careful analysis of the 
factors involved. 

Almost all mammals can swim after a fashion, and even some of 
those which one would naturally expect to be the least capable of 
doing so, such as the sloths {Bradypus and Choloepus) can cross 
broad rivers. To merit the term aquatic, however, a vertebrate must 
be sufficiently at home in the water so that when near this element it 
will instinctively seek it for concealment or escape. It must not only 
be able to swim with adequate speed but must be able to dive with 
facility and remain submerged. In almost all cases there is necessary 
the further corollary that it habitually seeks its food in the water. The 
ability to propel itself through the water does not mean that it has 
necessarily developed organs that are appreciably more highly special- 
ized for such use than are found' in its nearest, wholly terrestrial rela- 
tive, but it does mean that such a vertebrate must have evolved some 
sort of valvular mechanism for closing both nostrils and ears while 
submerged. In other words an aquatic vertebrate must have the ability 
to close all orifices leading within the body. 



All aquatic mammals have, of course, evolved from terrestrial ones, 
although paleontology does not furnish us with evidence regarding 
the early stages of any aquatic genus, but many of the steps may be 
reconstructed with a feeling of considerable assurance. 

The surf is usually too boisterous to appeal to a small or medium- 
sized mammal with a desire for an occasional swim, and shore waters 
offer too few places for concealment. Large carnivores might take 
directly to the sea, as the polar bear is doing, but it is probable that 
most aquatic mammals first ventured into fresh water, the smaller sorts 
into bogs and small streams, and the larger into swamps or rivers. 
Those of carnivorous (including insectivorous) propensities took this 
step almost invariably in search of prey which seemed to them easier 
of capture than terrestrial food, or else more palatable. At this stage 
carnivores probably would not seek the water as a means of escape from 
their enemies and their swimming must have been only superficial, 
consisting of a plunge after a fish or a short sally across a stream. 
An aquatic herbivore probably would start its career either by seeking 
the shelter of densely grown swamps as a diurnal refuge from its 
enemies, to secure aquatic plants which attracted it, or to escape the 
torment of insects by standing in the water. It would not take long, 
geologically speaking, for either the carnivore or herbivore to discover 
that the water offered a safe refuge from most of its enemies. But this 
must also be taken into consideration: an abundance of aquatic enemies, 
such as crocodiles, would tend to discourage the adoption of an aquatic 
existence and just this factor may have prevented the aquatic develop- 
ment of many a promising mammalian stock. 

A mammal with somewhat palustrine tendencies might, therefore, 
likely have been forced into the water either by similar forms that 
competed with it for terrestrial food, or by enemies which pursued it — 
perhaps both. If in the water it encountered an abundance of enemies 
m the shape of large fish or hungry reptiles it might either be exter- 
minated, forced back to face terrestrial enemies as being the lesser 
of two evils, or obliged to content itself with a borderline existence, 
foraging in shallow water amid protective vegetation. This latter 
has constituted the fate of most small, semi-aquatic insectivores and 
rodents. Some of them may eventually encounter an environment 
sufficiently favorable so that they can relinquish such a game of hide- 
and-seek and boldly take to the open water, but most of them will 
not, and the latter can never develop very highly specialized aquatic 



It is probable that the original adoption of a habitat that was largely 
aquatic usually, if not almost invariably, meant for a terrestrial mammal 
a diminution in the severity of the competition to which it was ac- 
customed, else it would not have taken to the water. And it is well 
known that a life wherein competition is reduced is not conducive to 
the rapid modification of bodily form. On the other hand, any such 
drastic change in the functions of the body as is experienced by a terres- 
trial mammal when it takes definitely to the water is highly conducive 
to evolutionary changes. To what effect these diametrically opposed 
tendencies have acted in the past upon the aquatic mammals with which 
we are now acquainted cannot be conjectured. But we do know, by 
such paleontological evidence as is at hand, that the course of develop- 
ment and change in our aquatic mammals has been an extremely slow 
and lengthy process, probably lasting, in those groups most highly 
adapted, throughout tens if not scores of millions of years. 

Let us take the hypothetical case of a carnivore that eventually be- 
comes exclusively marine. For our purpose this must be of an active 
type that prefers live food, and not such an omnivorous and slightly 
sluggish animal as the racoon (Procyon), which is slightly aquatic in 
its habits. Such an active carnivore likes fish and finds that it is easier 
to catch them in the shallows than to compete with other sorts of preda- 
tors catching mice in the woods. He will spend an increasing amount 
of time in the water and throughout successive generations gain increas- 
ing confidence in swimming considerable distances. During this time 
he has naturally discovered that few of his terrestrial enemies will fol- 
low him into the water, and he therefore instinctively seeks this element 
when startled. But he still journeys overland from one stream to an- 
other, may hunt on shore occasionally when fish are temporarily shy, and 
sleeps and raises his family in a hole beneath a bank. While he is 
content to linger in a small river he can do little else and his bodily 
modifications will not only be correspondingly circumscribed, but the 
change throughout long ages will be slow. 

Perhaps in a few million years this carnivore has gained such a facility 
in swimming that he finds narrow quarters irksome, and seeks the greater 
freedom of large rivers. He is now able to catch fish in fair chase, 
and can also escape from his enemies by his speed. The water is his 
home but although he frequently takes short naps while floating on 
the surface, he still seeks a hole in the bank, or if large and bulky, a 
sandbar, for a sound sleep, and his mate must seek the land to raise a 

family. At this time it makes little difference in the degree of his 



aquatic adaptation whether he inhabit large rivers or coastal waters. 
And heretofore all steps in his development have presumably been by 
slow stages. He himself soon has no further need for the land, for 
although he still likes to bask on a rock in the sun, it is not long before 
he can forego this luxury, does such a course seem expedient. But 
his young must pass their early life upon the land and if this necessity 
persists he can never become exclusively aquatic. 

The latter is a fundamental factor in the evolution of a highly de- 
veloped aquatic mammal — the requirements of its young. The newborn 
seal and sea-lion will drown if forced into the water. If the terrestrial 
or amphibious enemies of pinnipeds should multiply sufficiently to 
destroy more than the critical number of young needed to replace the 
breeding stock this order would become extinct, for they are not yet 
ready to desert the land entirely. In a few million years they may be 
ready to do so. At that time, if driven from their rookeries, they may 
have become sufficiently modified so that enough young might survive 
an aquatic birth for the perpetuation of the race. The whales and 
sirenians have successfully taken this step and forever severed their 
slightest connection with dry land. 

The development of existing aquatic herbivores was probably con- 
siderably different, for this sort of mammal is under no necessity for 
rapid movement in order to secure sustenance. From a palustral habi- 
tat the ancestors of the sirenians doubtless subsided sluggishly into deep 
water that was free from large enemies, chiefly for the purpose of es- 
caping troublesome terrestrial carnivores, and have since been under the 
necessity of doing little but move from one submerged pasture to an- 

All the above resolves into the simple statement, first advanced by 
Kukenthal (1890) that the degree of aquatic specialization in mammals 
is corollary to the amount of connection retained with the land. 


Chapter One 

The Mechanics of Swimming 

In considering the characteristic modes of locomotion of any organ- 
ism there are other factors besides simple progression to be heeded. 
Perhaps the most important of these is posture, which in turn depends 
largely upon bodily form as well as many environmental conditions. 
The posture of an elephant and a mouse must be different, necessitating 
differences in the skeletal framework and consequently in the muscular 
equipment. Dissimilarity of muscular equipment involves correspond- 
ing diversification of the controlling nervous mechanism, according as 
a mammal may habitually trot, gallop, pace or hop. Muscular action 
that is certainly reflex, if not actually involuntary in such forms as can 
sleep while standing, is a necessary agent in maintaining posture in a 
terrestrial mammal, while in a thoroughly aquatic one, such as a whale, 
posture need not involve action of the muscles, and balancing actions 
while moving are entirely different from those in which the non- 
aquatic sort must indulge. Similarly there must be important modifi- 
cations in the nervous equipment for determining the swimming ac- 
tions of such a mammal as the seal as contrasted with that of the 

The laws underlying the mechanics of swimming as employed by 
vertebrates are of almost hopeless complexity for the reason that the 
body is not rigid and the force is not applied at any one point, but 
more or less continuously over a greater or lesser area. It has taken 
many years to calculate the factors encountered by a rigid ship moving 
through the water at a given speed with all the driving force applied 
at one point. It has taken years for a multitude of highly trained tech- 
nicians to discover the proper lines for a rigid aeroplane fuselage and 
wings. If the latter were propelled to the accompaniment of con- 
tortive wrigglings by the hinder end, or convulsive gyrations of the 
wings by means of a multitude of small engines (i.e. muscles) of 
unknowable horse-power delivering their power at vague points, one 
may visualize how little would be known about aerodynamics at the 
present time. Precisely this situation is encountered by one who would 
investigate the principles underlying the swimming of mammals, and 

' [9] 


the higher the attainments of any quahfied physicist with whom the 
questions are discussed the more emphatic is he in his indication that 
he will have nothing to do with it. It therefore seems entirely im- 
possible to arrive at an exact conception of the mechanics of swimming 
by mammals, but certain generalities we know to be facts, and the evi- 
dence for certain others appears to be sufficiently strong to justify us 
in tentatively offering hypotheses. 

Stream-line, as I understand it, is that indefinite term by which we 
designate the precise shape of a given mass in order that it will slip 
through the water (or air) with less friction or resistance than any 
other shape of equal mass. But it is perhaps requisite that this body 
be towed. If it be self-propelled then the propeller will introduce a 
complication. It therefore is self-evident that no vertebrate can be 
perfectly stream-line in form. Not only must it have appendageous 
means of propulsion, but it usually has a steering or equilibrating ap- 
paratus. Because of feeding or other requirements its head may not 
be of a shape best adapted for speedy passage through the water. Either 
to compensate for such external irregularities or because obligated by 
muscular or visceral requirements, the cross section of the body may 
depart in one direction or the other from the circle that is ideal in a 
stream-line form. This being the case, two fish may be equally efficient 
in body form and yet have a considerably different appearance. Both 
may be 90 per cent efficient, and yet those details which from one 
detract 10 per cent may be entirely different from those amounting 
to the 10 per cent in the other. Or another viewpoint may be accorded 
this fact by the statement that what proves stream-line for one mode of 
propulsion does not prove so for some other method of swimming. Not 
only that but if one animal swims with its entire caudal and lumbar 
regions involved, the degree of its departure from an ideal stream-line 
form will not be the same as if it swims by vibrations of the tip of 
the tail only. 

Thus, in effect, it is found that when we speak of a body as stream- 
line we mean merely that it tapers gradually, without sharp angles or 
excrescences that would offer resistance as it passes through the water. 
But even though our knowledge of what is, or might be, the ideal of 
stream-line form in any particular case is so vague, the evidence fur- 
nished by the tendency of every essentially active form of aquatic life 
to assume certain definite body-shapes is so overwhelming that we know 
beyond question that this is one of the most fundamental of aquatic 



Terrestrial vertebrates may swim with much lost motion, thrashing 
about and straining with most of the muscles of the body. As soon 
as one of these has become even slightly aquatic, however, movement 
through the water is accomplished more easily and effectively, as it 
must be in order not to exhaust the swimmer. 

By such of the vertebrates as spend much of their time in the water, 
aquatic propulsion would seem to be accomplished by three methods: 

(1) Rhythmic undulatory or oscillating movements of the body 

(2) Rhythmic movements of the appendages. 

(3) Expulsion of jets of water. 

As far as I am aware, no vertebrate swims exclusively by expelling 
jets of water, but Breder (1926) and others have shown that many fish 
habitually employ the force of the water expelled through their gills 
as a material though secondary aid to some other primary means of 
propulsion. For our present purpose this method of swimming may 
be dismissed with no more than this brief mention. 

Swimming by means of rhythmic movements of the body include 
oscillations either in a vertical or horizontal plane of the body proper 
and these movements are in consequence always transmitted to the tail, 
which, if the latter be of sufficient size, then acts as a primary means 
of locomotion. This includes all vertebrates which swim chiefly by 
means of the tail except such fishes as hold the body rigid while rapidly 
oscillating the tail tip. 

Swimming by means of rhythmic movements of the appendages is 
employed by all vertebrates which propel themselves through the water 
by movements of their limbs, even though their bodies be wriggled dur- 
ing the process, by undulations of long dorsal. or ventral fins, or of a 
vibratory tail tip. 

In the case of such forms as are but slightly modified for an aquatic 
life it is frequently difficult to decide which is the primary means of 
propulsion through the water, as body, tail and limbs may all be used 
to some extent, but such doubtful cases are always mentioned in the 
text. Another difficulty in properly classifying swimming methods is 
introduced by the fact that some particular vertebrate may have experi- 
enced a sequence in its propulsive mechanism during its evolutionary 
modification, a more inefficient method being temporarily employed 
pending development of the final primary propulsive organ. 



The means of aquatic locomotion mentioned above in (1) and (2), 
together with representative vertebrates employing them, may be ar- 
ranged as follows. 


(1) Propulsion chiefly by means of oscillations of the body and 
base of tail. 

(a) Body fusiform 

Pisces (the majority of fish with body form as in 
mackerel, shark etc.) 


Caudata (the majority of salamanders) 
Crocodilia (crocodiles and alligators; chiefly) 
Ichthyosauria (of the better known, short-tailed sorts) 


Insectivora (probably only Potomogale and Limnogale) 

Carnivora (Lutrinae — the river otters, chiefly) 


Cetacea (exclusive of zeuglodonts of Basilosaunis type) 

{b) Body not fusiform but largely anguilliform or eel-like 
Pisces (eels and fish of this form) 

Serpentes (all swimming snakes) 

Amblyrhynchus (marine iguanas) 


Cetacea (only zeuglodonts of Basilosaurus type) 

(2) Propulsion by means of appendageous movements. 
{a) Propulsion by both fore and hind limbs 


Chelonia (turtles of the mud-turtle type) 

Plesiosdun/s and doubtless others 

Carnivora {Thalarctos, the polar bear) 

Pinnipedia (Odobenidae) 

Ungulata (Hippopotamidae) 

Rodentia (Hydrochoeridae) 

[12 1 


{b) Propulsion by means of pectoral appendages 

Pisces (many fish use the pectoral fins for slow propul- 
sion; the skates and rays (Batoidea) exclusively) 


Chelonia (the marine turtles) 

Aves (penguins and such birds as "fly" under water) 

Pinnipedia (Otariidae) 

{c) Propulsion by pelvic appendages 

Reptilia (Salientia — frogs and toads) 
Aves (all birds that swim with the feet) 

Marsupialia {Chironectes) 

Insectivora (practically all aquatic forms save Potomo- 

gale and Limnogale) 
Carnivora (Enhydrinae) 
Pinnipedia (Phocidae) 

Rodentia (all aquatic forms chiefly, most of them ex- 

(^) Propulsion by undulations of a longitudinal fin 

Pisces (an important, although possibly not the most 
eff^ective, means of locomotion of such forms as Amia- 
tus and Gymnotus, and exclusively of the Hippocampi- 
dae or sea horses) 

{e) Propulsion by vibration or undulation of the tail tip 

Pisces (used by many fish at slow speeds, and by such 
as are incased in an unyielding body covering, as the 

In some cases a certain vertebrate may swim by a combination of the 
above methods, and at times, especially in the case of certain extinct rep- 
tiles, it is extremely difficult to determine the precise method of swim- 

Although beset with several difficulties which are at present insur- 
mountable, some of the principles underlying the aquatic progression 
of mammals may be discussed with confidence. 



If one drop a cutworm into a dish of water the worm is unable to 
progress, or to do aught but bend first to one side and then to the 
other. The reason for this is that the center of gravity and center of 

Figure 1. Diagrams illustrative of some swimming principles of vertebrates: 
(i) a cut worm or similar form; (2) mud turtle swimming freely; 
(3) mud turtle with fore legs bound down by adhesive tape; (4) fish with 
its snout held rigid; (J and 6) fish swimming free; (7) the sea-lion prin- 
ciple: («) pivot of movement; {b) fulcrum; (r) power arm; {d) weight 
to be moved (i.e. the water) ; {e) lever. 

motion are at the middle of the body, and both the form and mass of 
the anterior and of the posterior parts are equal, and hence, any motion 
by one end is equalized by a corresponding motion of the other. Bend- 
ing of the body in a curve hitches it first an infinitesimal distance to 



one side, and antagonistic bending by the body then hitches it to the 
other side, but all actions and all motions are equalized. i 

As our cutworm cannot swim because there are two antagonistic pro- 
pulsive strokes of equal force and two antagonistic body divisions of 
equal mass and shape, it is evident that in order to make a swimming 
organism out of it we would have to make certain interdependent altera- 
tions as follows: There would have to be provision for holding the 
forward part of the body relatively inert to furnish a base from which 
propulsive movements could be initiated. This could be accomplished 
by increasing the mass of the anterior part of the animal and altering 
its shape so that it offered more resistance to torsional stress. This 
would have the effect of displacing the center of mass of the entire ani- 
mal forward of the middle. A relative increase in the mass of the an- 
terior part would cause a relative decrease in that of the posterior por- 
tion, and the latter should also experience some flattening and broaden- 
ing, to provide a greater area for operating against the water. The result 
may be likened to a man that is propelling a skiff by a single sculling- 
oar from the stern. But this sculling-oar may be very short and light, 
or very long and heavy. If the former, conditions are comparable to 
those encountered in fishes of the Ostraciidae type, in which the entire 
body is incased in an unyielding covering and locomotion is accom- 
plished by vibration of the tail tip. Or if the sculling-oar be very long 
and heavy, conditions may better be compared to the mackerel, whale 
and similar marine types which propel themselves by oscillations of the 
entire body. Both these groups might conceivably have body forms of 
essentially the same degree of stream-line perfection, and the center of 
gravity would therefore tend to be located at the same point in both, 
although varying somewhat with the speed of movement. But the 
actual center or pivot of motion would be very different indeed. In 
the ostraciiform fish it would be located in the tail proper and far from 
the center of mass, while in the other sort it would be much farther for- 
ward and nearer the center of mass. In effect, a fish can hold its body 
rigid and move only the extreme tip of the tail in order to glide forward, 
or it can move the entire body; and this is what introduces difficult 
physics. Not only is the amount of force applied in swimming inde- 
terminate, but the location of the exact area over which it operates is 

^ It must be understood in this connection that there is no assertion made of 
the inability of a short, worm-like body to swim through the water. Some 
larval forms can do so very well, but this ability depends upon some particular 
specialization which they have been able to develop. 



unknown, and can never be known with certainty. Thus in some sorts 
of whales it seems that the entire body takes part in swimming move- 
ments, and that the center of motion accordingly migrates forward to the 
center of mass, while it has been observed that when some sorts of por- 
poises are swimming at speed the oscillations of the tail are through such 
a short arc and are so rapid that the animal appears to be making no 
movement at all. Which type of action is the more efficient cannot be 

If one should hold a fish by the snout and allow it to wriggle in the 
water from side to side its motions might in some respects be compared 
to the principles of a lever of the third order. The snout held immov- 
able may be called the fulcrum while the water to be moved at either 
side of the tail is the weight (fig. 1) The entire posterior half, or even 
more, of the fish is the lever, and the muscles concerned in pulling the 
tail first to one side and then to the other constitute a double, compound 
'power arm. When the fish is released and it darts away, the circum- 
stances will have been altered as follows. The fulcrum shifts backward 
and becomes either a center of mass or a pivot of motion, according to 
conditions, and to the pair of caudal power arms is added a pair of an- 
terior power arms (fig. 1), and consequently, a second lever. In fish of 
the usual type the anterior lever will always be much shorter than the pos- 
terior. The posterior one may also be so short that as a result the length of 
the anterior lever will be negligible (as in fish swimming by vibration of 
the tail tip), or the posterior may be of sufficient length so that it is 
continuous with the anterior lever — which is another way of saying that 
swimming will then be accomplished by throwing the entire body into 
a continuous curve. 

For an interpretation of the motions followed by a fusiform fish 
(Breder's carangiform, and Abel's "torpedo principle") I follow Breder 
(1926), who has discussed the aquatic movements of this class with 
great thoroughness, and those who are interested in following the de- 
tails at greater length may refer to him. If we take a fish of the form 
of a mackerel the swimming motions will theoretically be as follows: 
The contraction of the muscles of one side will throw the body into a 
curve, but because of the resistance which the flattened posterior part 
offers to the water this results in the head being thrown farther to one 
side than the tail. With the momentum thus acquired by the head and 
consequent inertia of the more massive anterior part of the body, the fish 
is enabled to swing the tail to the opposite side with considerable force 
(fig. 2). The pressure of the curved tail against the water initiates 

[16} ^ . 


forward motion and this is maintained by alternations of the movements 
from side to side. The important points to note are that the body and 
tail do not oscillate about the snout, but that the anterior and posterior 
parts of the animal oscillate from a point in the body anterior to the 
middle. Whether this point is actually the center of gravity I do not 
know, for an aquatic vertebrate of this sort really has no center of 
gravity while it is in the water, so I prefer to employ the term "center 
of equilibrium" to designate this point. If the animal throw itself 

Figure 2. Diagrams illustrative of the swimming principles of an anguilli- 
form or eel-like body above, and of a fusiform body below (redrawn from 

into an uninterrupted curve while swimming, then the center of equi- 
librium will constitute a pivot of motion, from which both ends are 
curved. If only the tail proper be involved in swimming motions (as 
in ostraciiform fish), then this curve of the caudal amphikinetic part 
(as of Breder) will begin at a point considerably posterior to the center 
of equilibrium, and as an accompaniment the curve of the cephalic 
amphikinetic part (always present in theory although at times reduced 
to an inappreciable amount) must begin at a point correspondingly an- 
terior to the center of equilibrium. The difficulty lies in defining the 



significance of the word "correspondingly" in the last sentence. The 
principle is clear but the formula cannot be worked out. 

The above mode of swimming, termed the fusiform type because 
a spindle-shaped body is the best fitted to employ it, is followed by 
most fish, Potomogale, Sirenia, Phocidae (in seals the hind feet are 
employed the same as a tail), and Cetacea. Almost always there are, 
or finally will be, at least one rudder or pair of rudders or stabilators 
disassociated from the primary organ of propulsion, and not infrequently 
there is one or more accessory equilibrators (as a dorsal fin) . 

If the body and tail become disproportionately long for their diameter 
this mode of swimming will change to the anguilliform or eel-like 
type. The length, relative to diameter, must be sufficiently great for 
the animal to assume an S-shaped posture, when it will then swim by 
what constitutes, in reality, a duplication of the fusiform method of 
swimming, slipping through the water as a snake slips through the grass. 
This is of interest in the present connection chiefly for the reason that 
this was undoubtedly the method of swimming employed by long- 
tailed 2euglodonts of the BasHosaums type. But there may be body 
shapes and swimming methods intermediate between the anguilliform 
and fusiform types. Some of the longer bodied, modern whales may 
be partially anguilliform in their motions, and the Weddell seal {Lep- 
tonychotes) certainly is remarkably serpentine in its terrestrial move- 
ments. Then too there were short-bodied and long, anguilliform-bodied 
2euglodonts, and ichthyosaurs of similarly different conformation. Pre- 
sumably the tendency should be for a shortening of an anguilliform ver- 
tebrate because the fusiform method of swimming is capable of higher 
speed, and the shorter body encounters less water friction. 

Propulsion by the fore limbs, and by the hind limbs in other manner 
from that in which the hair seals (Phocidae) use these members, in- 
volves different principles. The most specialized instance of propulsion 
by the fore limbs among Mammalia is encountered in the sea-lions and 
fur seals (Otariidae). The principle is much the same as that of a 
row-boat. The latter should be fairly long for its width, for if it be 
short and tubby it will tend to wabble rather than progress in a straight 
line. It is evident also that the oars should operate from near the center 
of the craft, and that the latter should taper from the middle toward 
either end. This description fits the sea-lion, which apparently can 
steer with ease by means of either the head and neck or the hinder end. 
The description does not fit the platypus, however, which also swims 
chiefly by means of the pectoral limbs. In the latter the mass anterior 



to these limbs is not approximately equal to the mass posterior thereto, 
but very much less, and there must therefore be specialized apparatus 
for equilibration, which exists in both the hind feet and the tail. 

But there are different methods of swimming by the anterior limbs. 
Thus the platypus apparently pushes the water directly backward with 
the palm and partly clenches the manus during recovery. The marine 
turtle holds the axis of the manus almost parallel with the body axis 
(at least in the more specialized sorts) and pushes the water to the rear 
by a vertical motion of the arm. The sea-lion also employs an oblique 
movement of the flipper, but by pushing the water from the ulnar 
border of the manus by adduction of this member in the transverse plane. 

There are two features of interest illustrated by mammals which swim 
by kicking the hind feet in alternation. In one, encountered in almost 
all aquatic rodents and insectivores, as the hind feet are kicked the tail 
is involuntarily moved from side to side by the wriggling motion of the 
hinder end of the body proper. If the tail be of considerable length 
this member will at first effect all the equilibration that is necessary, 
and later, as the tail becomes more speciaHzed, it will take over a sub- 
stantial part, if not the whole, of the function of propulsion. If the 
hind limbs be relatively small their action in swimming may be assisted 
by the fore limbs, or if the tail be too short to act efficiently as an 
equilibrator the pectoral limbs may be kicked in alternation chiefly as 
balancing agents, just as a man swings his arms when walking. 

It is easily seen that if an animal swim by alternate strokes of the 
hind limbs alone there will be no serious disturbance of equilibrium 
providing the legs be close together. If the animal be of a tubby shape, 
However, with hind legs far apart, each kick will throw the body to the 
side unless there be some separate provision for maintaining equilibra- 
tion. This may be nicely illustrated by the common mud turtle. If 
one bind down the fore limbs of one of these animals by means of ad- 
hesive tape and then place it in a pool of water it is found that the 
turtle is very greatly handicapped thereby and will progress by a series 
of erratic jerks, first to one side and then the other (fig. 1) . The reason 
for this is that in vertical dimension the body is very thin, and in trans- 
verse very broad, so that the four legs appear almost as though attached 
to the corners of a rectangle. If one foot only is kicked it will turn the 
rectangle, and to prevent such turning there would either have to be a 
heavy tail, which the animal lacks, or a compensating kick by the oppo- 
site limb of the other pair. Thus the mud turtle in reality trots through 
the water. As specialization advances and the feet become paddles in 



the marine turtles the tendency is to use the hind hmbs as equihbrators 
and the anterior pair as propellers, these being operated in unison with 
a sort of flying motion. 

The same principle as forces the mud turtle to trot through the water 
operates to oblige all slightly specialized aquatic mammals with insig- 
nificant tails to employ all four feet in swimming, especially if the body 
be given to corpulency. So apparently swims the hippopotamus, whose 
legs are far apart, and the capybara, which has no tail; also the polar 
bear. But I am inclined to think that the culminating development in 
such mammals would not be the acquisition of four paddles of almost 
equal efficacy, as marine reptiles so often show, but rather the ascendency 
of one pair of limbs over the other, such as is now encountered in the 
Otariidae or Phocidae. 

In this connection it may be of interest to mention that aquatic rep- 
tiles seem to have had modifications which it is probable that no mammal 
could develop. Thus reptiles can increase the number of cervical verte- 
brae to a phenomenal degree — a procedure which a mammal cannot 
effect. The tendency in marine reptiles always was to develop four 
paddles, while in mammals but one pair usually becomes very highly 
modified. Similarly there must be many other reptilian body details 
which respond to definite stimuli in an entirely different manner than 
do the same details in mammals. I therefore regard it as very unsafe 
to draw conclusions anent the Mammalia from data provided by the 

The extinct plesiosaurs followed the turtle path for a considerable 
distance. Their bodies were heavy and broad, as well as somewhat 
flattened, and it is not improbable that this shape was accentuated in 
their ancestors. They were equipped with four long paddles, and as 
they employed a four-cornered method of propulsion they steered with 
the feet, kicking according as they wished to turn to one side or the 
other, and neither the tail nor the neck needed to function as a rudder. 
These two extremities could therefore develop in response to other 
stimuli without much regard to any duty in swimming. The result was 
that some sorts were short tailed and short necked, while others had 
a long tail and a neck that was relatively much longer than any other 
vertebrate has ever had, so far as we know. They were rather sluggish 
beasts, evidently, and the long neck was doubtless developed so that 
the head might dart here and there in pursuit of active prey in spite of 
the clumsy body. No other instance is known either in reptiles (ap- 
parently) or mammals in which a highly specialized aquatic form had 



both a long tail and a long neck, and presumably this development could 
only take place in connection with a body in which each motion during 
swimming was counterpoised by a corresponding motion on the oppo- 
site side. 

The Crocodilia, mesosaurs, mosasaurs and longer-bodied ichthyosaurs 
really represent different steps toward the same goal in the development 
of their limbs, and in all these there is very little difference in size be- 
tween the fore and hind feet. The longer sorts were partly anguilliform 
in movement and the shorter, fusiform. The Crocodilia use the tail 
exclusively for speedy swimming, and the hind limbs to a considerable 
extent during sluggish progression. So probably did the mesosaurs, 
and the mosasaurs largely so. In the latter the limbs had become true 
paddles, the fore limbs differing but slightly in size from the hinder 
ones, but the powerful, laterally flattened tail is clearly indicated as the 
more efficient propulsive organ and the flippers had probably begun 
to be used merely as equilibrators and accessory organs, as they finally 
became in the later, shorter-bodied ichthyosaurs. In the latter, with 
their very powerful tails, the hind limbs were considerably smaller than 
the fore limbs, and hence it seems evident that they had long since fallen 
into relative disuse, but not at a sufficiently remote period to have disap- 
peared altogether. 

From a consideration of aquatic mammals it seems that there are 
certain mechanical conditions which the more specialized sorts are 
constrained to endeavor to fulfill during the course of their development. 
There seems to be a decided tendency for aquatic mammals to develop 
as the primary means for locomotion a single organ or pair of organs, 
and if the original conformation of the animal allow and nothing divert 
it from its goal, the situation of this propelling apparatus will most likely 
be at the hinder end and medially situated. In other words, the stimulus 
is for the acquisition of the fish-tail type of propulsion, although if we 
leave out of consideration the question of comparison of the muscles 
involved, it makes no difference whether this equipment be in the form 
of a true tail expanded vertically or horizontally, or of the hind feet 
held with soles adpressed (Phocidae) . Pending the development of a 
tail fitted for propulsion the chief means by which swimming is ac- 
complished will probably, in the majority of cases, be the hind feet. 
Unless the animal can use the hind feet in swimming after the manner 
of the Phocidae (and Enhydrinae?), however, these members will 
probably never develop into the final, primary propulsive organ, one 
reason possibly being that the mammalian foot is likely incapable of 


duplicating the mechanical conditions embodied in the foot of a grebe 
or duck, which is not the most efficient method of swimming anyway. 
Swimming by oscillations of the hinder end of the body (by feet or 
tail) is the most economical method partly for the reason that all motions 
are propulsive ones, without obligation of any braking action by recovery 
motions, and because each stroke of the rhythmic cycle is of approxi- 
mately equal power. 

If the tail be too small to be adapted for a propulsive organ it is 
then likely that the fore feet will finally be modified for taking over 
this function exclusively (as in the Otariidae) . Theoretically it is not 
quite so efficient a method of swimming, for it can never be brought 
to as high a state of perfection as can that involving the transformation 
of the tail into flukes, but when a stage will have been reached compar- 
able to that at which the sea-lions now are, the fore limbs will constitute 
very eflFective swimming organs indeed, any braking action by recovery 
motions being almost entirely overcome, as discussed in a later chapter. 
Whether a seal can actually swim faster than a sea-lion is unknown, for 
there are no trustworthy figures on the subject. Both are capable of 
high speed ; not to the extent of 60 miles an hour, of course, as I have 
heard claimed by some illiterate fisherman, nor is it likely that they 
could outstrip some of the speedier porpoises. On theoretical grounds 
the seal should prove the faster. If it is not it may be because it does 
not need to be. 

Before closing the present consideration of aquatic progression it will 
be well to emphasize the fact that there will usually be a regular sequence 
of swimming methods employed during the evolution of any aquatic 
vertebrate. This is conveniently illustrated by the alligator. This rep- 
tile may "walk" slowly through the water with all four feet, or may 
progress by means of the pedes alone, while during more rapid progres- 
sion the appendages are pressed to the body while the tail is lashed from 
side to side. This is entirely illustrative of a sequence in methods of 
progression according as one part after another of a vertebrate becomes 
difl^erentiated for special use. It is also illustrative of another and 
fundamentally important consideration with which every aquatic verte- 
brate is obliged to contend at some point in its evolution, and which 
the reader will readily appreciate after having read this chapter. No 
method of swimming involving forthright pushing backward of the 
water by any pair of appendages is ever truly efficient, for the speed 
thereby attained is limited by the inertia of the limbs and their inability 
to repeat the rhythmic motions sufficiently fast. For really high speed 




efficient pair of anterior (sea-lion) or posterior (seal) flippers, or 
a properly formed tail (whale) must act obliquely against the water, 
for the same reason that an ice boat can travel much faster quartering 
the wind than running directly before it. This principle has governed 
the whole course of development of all the more highly adapted aquatic 


Chapter Two 

JLquatic zMammals 

T HE FOLLOWING members of the class Mammalia are definitely, al- 
though not necessarily exclusively, aquatic in their predilections: 





Mustela (the mink and sump 
f otter only) 
Lutrinae (all genera) 
Enhydrinae (the single genus) 

My castor 


Microtus (partly) 
Castoridae (all) 


Sylvilagus palustris 
Sylvilagus aquaticus 

Hippopotamidae (all) 
Tapiridae (all) 


SiRENiA (all) Cetacea (all) 

Trichechidae Archaeoceti (extinct) 

Halicoridae Odontoceti 

Hydrodamalidae (extinct) Mysticeti 

Perhaps the tapir should not be included in the above list, while 
there are doubtless those who will consider that in it should be placed 
other mammals, such as other swamp-loving ungulates. 

No attempt will be made to present a complete diagnosis of the 
families and genera included, but only a brief characterization and the 
salient points which are likely of significance from the viewpoint of 
their aquatic specialization. Preceding these, in order to give the general 
reader a better understanding of the chapters to follow, will be a general 
consideration of the mammals and their habits. The precise derivation 
of different aquatic mammals is not here of particular concern, but it 
seems advisable to touch briefly, without any great array of supporting 
evidence, upon the probable stock from which the more specialized of 
aquatic mammals were derived, in order that we may more intelligently 
follow the probable course of their development. 

It is virtually certain that even in those mammals only slightly aquatic 
there are present valvular mechanisms for the closure of the nostrils 
and external ears, and these two points will receive no further mention in 
the present chapter. 


Ormthorhynchus — platypus or duck-bill: an inhabitant of quiet 
streams and rivers of Australia and Tasmania. It feeds upon worms 
and similar food upon the stream bottoms. It constructs burrows for 
resting and for raising its young, and is capable of traveling over land 
somewhat clumsily but at considerable speed if its stream should go dry. 
The premaxillaries and maxillaries are expanded anteriorly and support 
a naked beak which superficially resembles that of a duck, and which 
is covered with a soft and very sensitive membrane. The body is 
flattened horizontally and covered with a fine, dense underfur, beyond 
which projects a coat of coarser hairs. The tail is rather short, very 
broad and compressed. The eyes are very small and there is no ex- 
ternal pinna to the ear. The legs are short and well modified for 
swimming, the toes being webbed and with large claws. The webbing 
of the forefoot extends well beyond the claws, and this anterior part 
is folded back when the animal progresses upon land, is digging, or 
combing its fur. In swimming the forelimbs are used almost exclu- 



sively, apparently alternately, while the hind feet are extended laterally 
and used, with the tail, as equilibrating apparatus. 

We cannot rely too greatly upon the internal anatomy of the platy- 
pus in judging its changes for a life in the water; and we cannot even 
be sure that the flatness of the body was not inherited from terrestrial 
ancestors. The form of its feet, tail, "bill" and external ear are the re- 
sult of this life, indubitably, but we cannot be sure of its muscles, bones, 
and certain details of its internal organs. It and the echidna are the 
sole representatives of an exceedingly primitive group of egg-laying 
mammals, and we have no suitable material, either living or fossil, with 
which to compare it. As these are the most primitive of mammals the 
stock must be of tremendous age, and it is extremely likely that the 
platypus has been aquatic for a great many millions of years. It would 
doubtless be far more specialized in this direction but for the fact that 
it is most at home along streams of but moderate size, has few if any 
aquatic competitiors, and feeds chiefly upon food that is relatively in- 
active. It is not impossible that its aquatic preferences began even be- 
fore the placental mammals became differentiated. For the above rea- 
sons the internal anatomy of the platypus will either be entirely omitted 
or discussed with great circumspection. 


Ch'tYonectes — water-opossum or yapok: of Central America and 
northern South America. It subsists chiefly upon small fish and crus- 
taceans, but is attracted by almost anything edible. Its pelage is short 
and dense and the external ears are well developed. The tail is round 
but the hind feet are large and very broadly webbed, in some toes to 
the tip and in others to the last phalanx. 

But few details regarding the habits of the yapock — the only aquatic 
representative of the Marsupialia — are known, but its dependence upon 
the land appears to be great. The development of the hind feet seems 
to be its only change for swimming, if we except the character of the 
pelage, and it is remarkable that the feet should be so highly specialized 
while the tail is perfectly terete. The base of the latter tapers gradually 
from the body, as in many other marsupials. Although it is known that 
small marsupial pouch-young can suspend breathing for many minutes 
without harm and Carl Hartman has told me of a swimming opossum 
(Didelphis) with live pouch-young having been captured, it would seem 
that in Chironectes the female must curtail aquatic activity while she 
raises a family. Without doubt this factor has been of exceeding im- 



portance in limiting the degree of aquatic specialization of this animal, 
and it has probably contributed very largely to the fact that the Aus- 
tralian marsupials, so able in their plasticity to fill all vacant ecologic 
niches, have never developed an aquatic form, for such bold individuals 
as dared venture freely into the water would be likely to drown their 
young and die without issue. At the same time it is possible that dur- 
ing submergence the pouch might retain sufficient air for the needs of 
the young. 


Figure 3. Some aquatic mammals. Platypus {Ornithorhynchus) (redrawn 
from Lewin), beaver {Castor), desman (Desmana) (redrawn from Flower 
and Lydekker), and capybara {Hydrochoerus) (redrawn from Scott). 


Aquatic insectivores are all small and few of them can venture into 
deep water for fear of large fish. Hence they must retain in large de- 
gree their dependence upon the land, for throughout their lives along the 
rills which they inhabit escape by a scamper through pebbly shallows may 
be just as necessary as by diving. Aquatic forms have the fur even denser 
than usual: the tendency is for the hind feet to become enlarged and 
fringed with bristles as well as to be slightly pronated and for the tail 
to become keeled with hairs, or (and) flattened horizontally. Rarely 
have the external ears disappeared. Swimming, except in the aquatic 
Tenrecidae, is evidently always by alternate strokes of the hind feet as- 
sisted to some extent by accompanying motions of the tail. 


Desmana — the desman, occurs in southeastern Russia. The fur is 
especially dense and fine, the snout is long and flattened vertically, and 



there are no external ears. Although partly webbed the forefeet are 
small, while the hind feet are enlarged, webbed to the base of the nails, 
and heavily fringed by bristly hairs along the outer border. The tail 
is robust and laterally compressed. This genus is more modified for 
an aquatic life than any of the Soricidae. 

Gaiemys occurs in parts of Spain and adjoining France. It is similar 
to Desmana but smaller, snout somewhat longer, tail longer and later- 
ally compressed only near the tip, but its hairs form a faintly defined 
ventral keel. It is not quite so highly modified aquatically as the des- 


Neosorex and Atophyrax are two groups of water shrews inhabiting 
parts of the United States and Canada. Jackson (1928) has recently 
placed them with the genus Sorex but they have usually been accorded 
generic (or subgeneric) rank and may be so treated here for convenience. 
Their habitat is along the borders of little streams and in bogs and 
marshes, where they secure water insects and small fish. Appreciable 
modifications for an aquatic life consist only of short fringes of stiffened 
hairs upon the sides of the hind feet and toes, and slight pronation of 
these members. 

Neomys — the European water shrew, is found in Europe and Euro- 
pean Russia. Its aquatic modifications also consist of fringes to the 
hind feet and toes, but in addition one species (N. foidens) has a 
median keel of stiffened hairs extending the entire length of the under 
surface of the tail. 

Chimarrogale of the Himalayas and parts of China and Japan has the 
hind feet and toes fringed and the hands slightly so, while the tail has 
a well marked inferior fringe. 

Crossogale of the Malay Archipelago is said to be essentially like 

Nectogale of Tibet and adjacent China is more specialized. In gen- 
eral it too resembles Chimarrogale but there are said to be no external 
ears, the hind feet are webbed in addition to having slight fringes, and 
the tail is fringed below. 


Limnogale of Madagascar has a decidedly flattened muzzle and 
webbed feet. The whole tail is very powerful but only the distal part 
is laterally compressed. Presumably both hind feet and tail are used 
in swimming but probably the latter is the chief organ of propulsion. 



Potomogale is an inhabitant of the streams of equitorial Africa. Ex- 
ternal ears are present. The muzzle is flattened and the body is cy- 
hndrical, continuing uninterruptedly into the thick, powerful tail, which 
is excessively compressed laterally. The legs are short, rather weak, 
and the toes are not webbed, but the lateral border of the hind foot is 
broadened to a thin edge so that it may be folded back more smoothly 
against the tail. The latter is evidently the exclusive means of aquatic 
propulsion, the feet being then pressed against the body and base of 
the tail. 

The aquatic members of the Talpidae and Soricidae may be consid- 
ered as a unit. The insectivores are among the most primitive of 
placental mammals and reasonably close to the original prototype. They 
are notoriously conservative, some of them, we know, having remained 
virtually unchanged since Cretaceous times. Remains of desmans which 

Figure 4. The African insectivore otter Potomogale, from a mounted specimen 
in the National Museum. 

are considered to be congeneric with the living animal are known from 
the Middle and Lower Miocene of Europe. So we can rest assured 
that the aquatic insectivores are no very recent development but that 
their modifications have been brought about throughout very long 
geologic periods. That such specialization is not now more marked 
than merely in the form of the hind feet and at times, slight changes 
in the tail and ear, may be attributed to the conservativeness of the 
insectivore phylum and the fact that they are so largely dependent upon 
the land. 

Potoj)iogale is a paradox, for it is extremely unlikely that any land 
mammal would take to the water and from the very start use its tail 
as the sole means of propulsion. In other cases a flattening of the tail 
follows acquisition of webbing by, and at times an increase in the size 
of, the hind feet. Yet the feet of this genus are unwebbed and re- 


markably weak. It too belongs to a primitive family of insectivores 
and its modifications probably began at a very remote time. 



Thalarctos — the polar bear, is an inhabitant of the Arctic regions. It 
spends so much of its time in the water that it must needs be included 
in any consideration of aquatic mammals, yet the only way in which this 
bear differs from others is in the increased hariness of the soles of the 
feet, which was probably brought about by the coldness of the ice on 
which it walks rather than any influence of the water. It also has a 
build somewhat more rangy and slender — in other words more stream- 
line — than is usual in this family. All four feet are employed in 


Mustela. In this genus only the American mink and European sump- 
fotter may be said to be aquatic, and yet there are really no discernible 
external indications of the fact. The toes are partially webbed but so 
are those of many strictly terrestrial species of this genus. 

Lutrinae. AH river otters are very largely aquatic. Of this sub- 
family the following genera are now recognized by those who have re- 
cently worked with them: Aiicroaonyx, Paraonyx, Aonyx, Amblonyx, 
Hydrictis, Pteronura, Lutra, and Lutrogale. They occur almost through- 
out the world on the larger and many of the smaller land masses. 
The legs are rather short, the body sinuous and cylindrical, and the tail 
is very stout, especially at base, tapering gradually from the body to the 
tip. In Pteronura there is a lateral flange or keel upon each side of the 
tail. The webbing of the feet diflFers somewhat in the different gen- 
era, and this is discussed fully in the case of the Old World genera 
by Pocock (1921) . Thus in Aonyx and Paraonyx of Africa the forefeet 
are unwebbed, and even in the common otter they are no more webbed 
than in many of the strictly terrestrial mustelids, although in at least 
one species of this genus (Lutra macuUcolUs) they are. In Paraonyx, 
Mkroaonynx and Aonyx, the so-called clawless otters of Africa, the 
hind feet are practically unwebbed; but in most genera the webbing 
extends almost or quite to the tips of the toes, although the feet are 
little if any enlarged. The base of the tail is more robust than in ter- 
restrial mustelids. 



The remains of forms supposed to be intermediate between otters 
and true mustelids are known from the Tertiary, but these are frag- 
mentary, and nothing, of course, can be told regarding any aquatic 
adaptations which they may have shown. 

Enhydrinae — the sea otter, of a single genus and species, with dis- 
tribution in the north Pacific and south along the American coast into 
the waters of Lower California. Apparently this animal spends its 
entire time in the sea save occasionally when it may haul out on the 
rocks to rest. It is said to give birth to the young upon floating patches 
of kelp. External modifications for an aquatic life consist chiefly of a 
reduction in the size of the external ear and of the fore feet, increase in 
the size of the hind feet, which are fully webbed, with hairy sole and 
with the fifth toe the longest, in the slight vertical flattening of the 
moderate sized tail, and the excessive softness and denseness of the 
pelage. Significant points in its osteology consist of reported increase 
in the flexibility of the vertebral column, slight shortening of the neck, 
flaring of the anterior element of the pelvis, slight shortening of the 
fore limbs and of the thigh and shank as well, possible slight increase 
in stoutness of the hind limb and enlargement of the pes. 

In spite of my best eff^orts and numerous letters to those who should 
know, I have been entirely unsuccessful in ascertaining the precise move- 
ments of the hind feet whereby this animal accomplishes swimming. 
Possibilities will, however, be discussed in chapter eleven. 


All the pinnipeds are highly aquatic, to only a lesser degree than the 
whales and sirenians, and are chiefly pelagic. As in the case of the 
Cetacea there are many sorts of pinnipeds, especially phocids or true 
seals, which diflfer from one another in respects which for our present 
purposes are of a rather minor character. The order is separable into 
three well defined families, as discussed below. In addition there are 
several curious fossil pinnipeds whose allocation may prove puzzling. 
Because this order is of the utmost importance in a consideration of 
aquatic modifications its characteristics will be discussed in considerable 

As an order the Pinnipedia are mammals of markedly stream-line 
shape and with very short tails. TJie elbow and knee are always situa- 
ted well within the body contour and the crotch is located at or slightly 
above the level of the heel. The feet are webbed and in some are 
paddle-like. The fifth toe of the hind foot is approximately as long 



as the first and both are longer than the three middle toes. The pre- 
acetabular length of the pelvic bones is always less than the post-acetabu- 
lar. The humerus is relatively massive and the bones of the forearm 
very broad at one end, while the femur is much reduced in length and 
is flattened. There is no clavicle. In adults there is probably always 
a large hepatic sinus of the vena cava. Nictitating membranes are 
said to be present, and retia mirabilia occur to an unknown extent. 


These are the sea-lions and fur seals or sea bears. The pelage is 
short, either entirely hairy or with a dense and fine underfur, and the 
young do not have a coat that is exclusively woolly. The external ear 
is small and narrow. Both fore and hind feet are used in limited ter- 
restrial locomotion but the former exclusively in swimming. The area 
of the fore foot is great and the axilla is situated at about the middle 
of the forearm. The hind foot assumes a plantigrade posture during 
terrestrial locomotion and the astragalus is without a posterior exten- 
sion. Each digit of both fore and hind foot has a cartilaginous ex- 
tension and the palms and soles are naked. The nails of the fore foot 
and those of the first and fifth digits of the hind foot are vestigeal. 
The testes are scrotal. The canines are not unusually developed. The 
vertebral spines of the anterior thorax are well developed. The greater 
tuberosity of the humerus is higher than the lesser. The ilium is 
slightly curved — not markedly and abruptly bent laterad — and the femur 
has a lesser trochanter. 


This is the walrus, of the north Atlantic and north Pacific. 
This family, containing a single genus, is in most respects nothing 
but a specialized otariid, although in some points it is fairly inter- 
mediate between that group and the phocids. The body is almost 
hairless. There is no external ear. Both fore and hind feet are used 
in terrestrial locomotion, which is more limited than in otariids, but 
the hind feet are used in swimming by being moved from side to side, 
while the fore feet are also used, but alternately. In major details the 
feet are most like those of an otariid but the cartilaginous extensions 
are not as long, the fore foot is relatively smaller and the hind foot 
broader and more comparable in shape to that of a phocid. The hind 
foot assumes a plantigrade position during terrestrial locomotion and 
there is a slight posterior extension to the astragalus. The testes are 

[ 32 ] 


scrotal. The canines are enormously and phenomenally enlarged, 
especially in males. The greater tuberosity of the humerus is higher 
than the lesser. The ilium is moderately curved ourward, to a greater 
extent than in the otariids and less than in phocids, and the femur is 
without a lesser trochanter. 


These are the true seals or earless seals, of which there are numer- 
ous genera, are found locally almost throughout the world, even in 
some large lakes. The pelage is hairy and never with a furry under- 
coat, but the pelage of the young, at least in a number of genera, is 
long and woolly, this being shed either shortly after birth or m utero. 
The area of the fore foot is reduced, the axilla falls opposite the wrist, 
and in most sorts these members are of use only for equilibration, on 
land being employed only for surmounting obstacles; although in the 
elephant seal {Mboiinga) at least, they are usually placed fairly flat 
upon the ground. The area of the hind foot is somewhat increased, 
relative to length of limb, these appendages being plainly indicated 
as the means for aquatic propulsion, which is accomplished by rhyth- 
mic lateral movements as by the tail of most fishes. In spite of the 
apparent inefficiency of terrestrial locomotion, which is accomplished 
by a caterpillar-like wriggling of the body, with the hind feet elevated, 
seals have been known to travel overland for many miles, and even 
through mountainous country, when the freezing of coastal waters 
has made it advisable for them to seek new aquatic territory. 

The astragalus has a posterior projection as long as that of the cal- 
caneum and the foot is prevented from assuming a plantigrade posture 
by the unusual tension of the flexor hallucis longus tendon. There 
are no cartilaginous prolongations of the digits and the palm and sole 
are as well haired as the dorsum of the feet. The nails are all well 
formed, except that they are absent from the pedes of M'nounga. The 
testes are abdominal. The canines are but moderately developed. The 
vertebral spines of the anterior thorax are poorly developed. The lesser 
tuberosity of the humerus is usually higher than the greater, the ilium is 
markedly and abruptly bent outward, and there is no femoral lesser 

I have already discussed the pinnipeds at considerable length 
(Howell, 1929). Their derivation is a matter of controversy, for al- 
though both otariids and phocids are known from the Tertiary, they 
were even then true pinnipeds. It is a commonly accepted belief that 



they are biserial, and that the walrus is but an aberrant and modified 
otariid, but as in the case of the whales much of the discussion has 
been of a profitless nature. The meaning of the term biserial or 
diphyletic is loose and varies with the investigator. It is presumed that 
all placentals were derived from a common ancestor, and therefore all 
phyla eventually go back to this point. It is commonly believed that 
the pinnipeds were derived from the adaptive creodonts, which in turn 
were probably derived from a single ancestor. It may be pointed out, 
however, that because the pinnipeds have so far been unable to sever 
their connection with the land at the time of the birth of the young 

Figure 5. Pinniped postures. Seal {Phocidae) (above), sea-lion {Otariidae) , 
and walrus {Odobenidae) . 

their latter day (geologically speaking) evolutional velocity may have 
been at a slow rate, and they may well be nearly as old, phylogene- 
tically, as the Cetacea. 

Whether or not the otariids and phocids became individually dif- 
ferentiated before or after their ancestors first took to the water is, 
with the present paucity of fossil material, a matter that cannot be 
settled and over which one may waste much valuable time and ink. 
It has been frequently claimed — first by Mivart, I believe — that cer- 
tain anatomical resemblances which it is not necessary here to repeat 
at length indicate that the otariids have descended from the bears and 
the phocids from the otters. In the first place the otariid stock is al- 



most certainly older than the ursine line, and the phocid older than that 
of the otters. The resemblances are incontrovertible and undoubtedly 
of real phylogenetic significance, but probably have no greater value 
than indicating that the primitive ancestors of both otariids and bears, 
although entirely distinct, had certain characters in common, as now 
have any number of very diverse carnivores; and similarly with the 
prototypes of the phocids and otters. 

Be that as it may these two groups of pinnipeds are now fundamentally 
very different and have been distinct for a very long geologic time. 
Their true phylogenetic dissimilarities are further accentuated by the 
differences developed by their diametrically opposed methods of aquatic 
progression. The walrus seems certainly a branch of the otariid stock, 
modified by its enormous tusks, its feeding upon inactive prey, and 
its bulk. The way in which it swims both with its fore feet like the 
otariid and its hind feet like the phocid is extremely interesting and 


No rodent is very greatly modified for an aquatic life, probably be- 
cause they are either dependent to great degree upon the banks of the 
streams which they prefer, or feed upon plants which are terrestrial or 
of the shallow water. In addition many of them are small enough to 
be eaten by large fish if they venture into deep water. Besides web- 
bing or bristly fringing of the hind feet in the more specialized sorts, 
these members may be enlarged, the tail may be flattened, the fur 
dense and soft, and the dorsal outline of the skull in some sorts is 
less convex than usual, especially in the rostral region. 


Aiyocasfor — the coypu of eastern South America. This large rodent 
has beneath its outer coat a suit of fine fur that is not so dense as 
that of the muskrat but still heavier than one would expect to find 
in any terrestrial rodent of the tropics. The hind feet are quite large 
but the webbing of the toes does not extend to the tips. The tail is 
perfectly round and the external ears are rather small. The mammae 
are located dorsad. 


Hydrochoeriis — the capybara of eastern South America is the larg- 
est living rodent, attaining a body length of about four feet. The 
nostrils are situated rather high, the toes are webbed only at their 



bases, and the pelage is coarse. Like many of its relatives it is without a 
tail, differing in this respect from all other aquatic rodents, and it is 
by no means unlikely that this lack accounts for the fact that (from 
all accounts) it uses all four feet in swimming. It is partial to beds 
of reeds but seeks the water when disturbed, swimming and diving with 


Crossomys — one of the Australian water rats, in all three genera of 
which the fur is dense and soft, and in all the external ears are small 
but well formed. In this genus the toes are said to be webbed and the 
tail, which tapers gradually from the body, has a strongly marked swim- 
ming fringe below. In these water rats swimming is undoubtedly by al- 
ternate movements of the hind feet, assisted by the tail. 

Hydromys — a second genus of Australian water rat and the only 
one which I have examined. The feet are partially webbed, and the 
dorsal outline of the skull straight (as may be the case in the other 

Parahydromys — a third genus of Australian water rat, in which the 
aquatic specialization is said to be somewhat less than in Hydromys. 
The first and last hind toes are practically unwebbed. 

Dasmyms — a genus of African rodent which is said to be somewhat 
aquatic, but the only indication of this which is shown by specimens 
consists of the pelage, which is short, thick and soft. 

Nilopegamys is another genus of aquatic rodent from Africa, known 
from a single specimen and recently named by Osgood. Its pelage 
is also soft and dense, the hind feet are slightly enlarged, and the dor- 
sal outline of the skull has a concavity near the interorbital region. 

Arvicola and Microtus. Several members of the subfamily Micro- 
tinae (meadow mice or voles) live in bogs or along the banks of 
streams and swim freely. Perhaps the most aquatic of these is the 
European water rat (Arvicola), but it has no discernible modifica- 
tions that are considered really aquatic. 

Ondatra — the muskrat, of the United States and Canada, is the most 
aquatically specialized of all the Muridae. Its underfur is extremely 
soft and dense, the hind feet are enlarged and although not com- 
pletely webbed, the sides of the feet and all the toes are bordered by 
very heavy fringes of stiff hairs, and the tail is laterally much flattened. 
Swimming is accomplished by means of alternate thrusts of the hind 
feet and accompanying horizontal movements of the tail; and I have 



been led to believe that occasionally the tail is thus used alone, while 
the feet are trailed. 

Neofjber — the Florida water rat, is a smaller edition of the musk- 
rat, but the tail is round, the feet are but slightly webbed, and hairy 
fringes are lacking. 

Caslor — the beaver, of North America and locally in northern 
Europe. The under fur is famous for its softness and density. The 
nostrils are situated slightly higher than in most rodents, the external 
ear is reduced in size, the hind feet are large and very fully webbed, 
and the tail is naked, flat, and relatively broader than in any other 
partially terrestrial mammal. Swimming is usually accomplished by al- 
ternate movements of the hind feet, although occasionally I have seen 
the tail alone used as a sculling oar. The tail is also used to facilitate 
quick submergence, being then slapped loudly on the water as a warn- 
ing signal: but it is not used as a trowel, as has frequently been 
claimed in the past. 


There are three closely related genera of this family that are par- 
tially aquatic. They occur in central and South America and have 
short, dense fur. 

Ichthyomys. The muzzle is slightly flattened and the eyes and ears 
unusually small. The upper incisors are prolonged into sharp lateral 
points, which should facilitate the holding of slippery prey, and the 
dorsal outline of the skull is definitely concave. The hind feet are 
broad and fringed, but very incompletely webbed. The hairs on the 
under side of the long tail are slightly lengthened and are dense, form- 
ing a slight keel. 

Rheomys is very similar to the last genus but hairy fringes are 
limited to the sides of the feet and outer toes only, and the tail is not 
noticeably fringed. It is said to feed upon aquatic snails. 

Anotomys is said to be similar to Ichthyomys also, but the ear open- 
ing is reported to be a mere slit. 

As with the insectivores the line of aquatic modification which ro- 
dents usually follow is in the webbing and fringing of the hind feet, 
followed by enlargement of these members and a lateral flattening 
or ventral fringing of the tail. External ears are said to be absent 
in Anotomys but are present, although usually reduced, in all others. 



As previously mentioned, it is significant that the capybara is the 
only aquatic rodent without a tail and the only one in which the hind 
feet are not indicated as the primary means of propulsion. 



Sylvilagus palustris and S. aquatkus. The marsh and swamp rabbits 
of the southern United States are quite aquatic, readily taking to the 
water when dismrbed and swimming freely. They have no discernible 
aquatic modifications, however, and the hind feet are even smaller 
and narrower than in the majority of rabbits. It is not meant to im- 
ply that their swimming propensities have resulted in a reduction of 
foot size, but it is not unlikely that they are descended from some 
stock of small-footed running rabbits (as is Romerolagus) rather than 
of hopping rabbits, like most other sorts. 


Artiodactyla: Hippopotamidae 

This family is the most aquatic of any of the existing ungulates. 
The animals usually spend the day either in the water or basking upon 
some segregated sandbar, repairing to the land at night to feed along 
the banks, and sometimes foraging to considerable distances or travel- 
ing overland from one river to another. 

Hippopotamus — the common hippopotamus of Africa is highly 
aquatic in habits. It evidently swims with all four feet but these are 
not especially modified for the purpose. External specialization for an 
aquatic life consists of the elevation of the nostrils and the peculiar 
orbits, which allow the eyes as well as the nares to be kept above 
water while the remainder of the animal is submerged. The shortness 
of the legs may, to some extent, also be due to this influence. The 
short tail is exceedingly flattened laterally, but it is too small to as- 
sist in swimming and the reason for its shape is not clear. 

Choeropsis — the pigmy hippopotamus of parts of west Africa is much 
smaller than its cousin and its head is much less modified in the direc- 
tion which the other genus has taken, the orbits not being excessively 

The hippopotamus is related to the pigs and remains that are hardly 
distinguishable from the existing species have been found in the Euro- 
pean Pliocene. The wonder is that this animal is not considerably more 



modified aquatically than we find it. However, its critical activities are 
more terrestrial than aquatic, for it must seek the land for its food, of 
which it requires a prodigious amount, and this condition is unique 
among aquatic mammals, for no other spends the whole day in the 
water and feeds exclusively on land. This, then, is its primary require- 
ment — that it may walk with facility upon the land. At the present 
time it can have no dangerous terrestrial enemies save the lion and man, 
and there is really little reason why it should spend the most of its 
time in the water save that it is probably more comfortable there. And 

Figure 6. Dorsal representation of Sirenia: (a) manati {Trkhechus) from 

British Guiana (redrawn from Murie, 1880) ; {b) Trichechus americanus 

(from Murie, 1872); (f) dugong (Halicore) (from a mounted specimen 
in the National Museum). 

as it should be perfectly safe when in the water from everything except 
a large crocodile, there seems little reason for the high specialization of 
its nostrils and especially its eyes, which enable it to breathe and see 
when every other vestige of its body is submerged. 


The sirenians or sea cows are exclusively aquatic and inhabit rivers 
and coastal waters where aquatic plants and algae abound. The fore 
limbs are paddle shaped, there are no external hind limbs, and the tail 
is excessively flattened vertically and correspondingly expanded hori- 
zontally. The nostrils are considerably elevated, the eyes are small and 
are furnished with nictitating membranes, and there are no external 
ears. The skin is rugose and almost hairless. The mammae are lateral 
and postaxillary and the testes are abdominal. All the bones are ex- 
ceedingly dense and heavy, the external nares are high upon the rostrum 
and the nasal bones are usually absent (present in the manati). The 
neck is shortened, there is no clavicle or sacrum and the pelvis is ex- 
tremely vestigeal. The phalanges are moderately flattened. The lungs 
are long and shallow (vertically) and the diaphragm is almost hori- 
zontal and extends far back. There are very extensive retia mirabilia. 
Swimming is accomplished exclusively by the tail and individuals (at 
least of the manati) habitually rest with the back excessively curved 
and the tail tip touching the bottom (see figure 7). The pectoral 
limbs are used in feeding and are more mobile than in the Cetacea. 


Trichechus — the manati inhabits rivers in parts of eastern tropical 
America and locally in tropical Africa. Nasal bones are present and 
there are but six cervical vertebrae. The nostrils are not quite so ele- 
vated as in the Halicoridae, the tail is shovel-shaped, and vestigeal 
nails upon the flippers are usually present (absent in the form inun- 
guis) . 


Halicore — the dugong is found in shallow bays and rivers locally 
from the Red Sea to Australia, and feeds chiefly upon marine algae. 
The nares, both externally and in the skull, are more elevated than in 
the manati, and there are no nasal bones. There are seven cervical 
vertebrae. The tail is notched centrally and the lateral tips are quite 
pointed rather than rounded. Nails are absent. 


Hydrodmualis — Steller's sea cow inhabited the vicinity of Bering and 
Copper islands in the North Pacific, where they fed upon marine algae, 
but they were exterminated more than one hundred and fifty years ago 


and very little is known about them. The tail took the form of two lat- 
eral pointed lobes and the flippers were relatively small and said to have 
been covered with short, coarse hairs, although the body was naked. The 
cervical vertebrae numbered seven, and they probably lacked nasal bones. 
There were no teeth in the adult, their places being taken by horny oral 

The sirenians are commonly believed to have been derived from the 
progenitors of the proboscidian stock. Known remains date back to 
the Eocene, and even then they were highly modified for an aquatic 
life. They are not as progressively specialized as the whales but their 
sedentary existence with no need for seeking active food is not as stimu- 
lating for rapid evolutionary change and their aquatic habits are un- 
doubtedly of great antiquity. 


Archaecoceti or zeuglodonts were essentially whale-like in body form 
and general details, but were less specialized in some respects. Thus 
the anterior limb was less modified and the skull had few of the char- 
acteristics which we associate with modern whales. Externally the hind 
limbs had disappeared, however. According to Kellogg (1928) zeuglo- 
donts are first known from the Eocene, but were evidently a rather un- 
successful experiment for they suddenly disappeared in the Upper Oli- 
gocene. Although the narial aperture of the skull shows considerable 
retrogression, the bones exhibit no indication of the telescoping that is 
so characteristic of the modern whales. The dentition of the more 
primitive sorts resembles to a considerable extent that of certain creo- 
donts, but this is believed to be only superficial, and the skull is re- 
markably suggestive in general conformation of some Cretaceous insec- 
tivores, and indeed modern ones as well. The majority of paleontolo- 
gists, I believe, consider the zeuglodonts and modern whales to have had 
a common ancestry, but that it is impossible that the latter could have 
descended from any sort of zeuglodont now known. Most likely this an- 
cestor was very different from either, the modern whales diverging in one 
direction and the zeuglodonts branching at another tangent in a direc- 
tion that proved unsuccessful. It is interesting to note, as Kellogg says, 
that ""periotic bones of both the whalebone and toothed whale type have 
been found attached to skulls of zeuglodonts," and, I may add, no other 
mammal known, either living or extinct. 

Of those zeuglodonts so far known there are two extremes of body 
type. One of these is represented by BasHosaurus, with very long tail 


which unquestionably was utilized for anguilliform propulsion, and the 
other by Zeuglodojz osiris, with a rather short tail suitable for fusiform 
propulsion. Just as these two swimming types grade one into another 
in the case of fish, according to body length and slenderness, so may 
we presume that there were zeuglodonts at some period that were inter- 
mediate between the two types mentioned. Basilosaurus, however, certain- 
ly could not have had flukes like modern cetaceans, for it had not the 
musculature to control such equipment, as indicated by the low spinous 
processes. But the tail was very mobile, because the zygopophyses did 
not articulate, and it must have had some sort of caudal expansion, but 
of great lineal rather than lateral extent. The same applies in some 
measure to Zeuglodon osiris, except that having a short tail its caudal 
expansions must have been lineally shorter and possibly somewhat sug- 
gestive of the manati. 

Modern whales are exclusively aquatic. The body is fusiform, with 
no real constriction at the greatly shortened neck, and tapers gradually 
from the thorax to the base of the tail. The latter terminates in a pair 
of flukes, vertically flattened and laterally expanded into two pointed 
lobes with a notch between them. There is no vestige of external hind 
limbs and the anterior limbs are paddle-shaped, the integument being 
continuous over the digits, which have no sign of nails, and there is no 
external division of the limb into segments. The skin is smooth and 
hairless except that there are occasionally a very few bristles about the 
head. There are no sebaceous glands and the suderiferous ones are ab- 
sent or much reduced. There is a specialized blubber layer beneath the 
skin. The eyes, which are specialized to function in salt water, have 
no nictitating membrane nor lachrymal duct. There is no pinna to the 
ear and the auditory meatus is a minute aperture. The nostrils open 
not near the snout (save in sperm whales) but near the vertex of the 
head, with corresponding alteration in the bones of the skull. The lat- 
ter is very remarkable in that some of the bones have slid partly over 
others, resulting in the condition referred to as telescoping. The bones 
are spongy, the cavities being filled with oil. The cervical vertebrae 
are often phenomenally shortened and in some species most or all of 
them are fused. The articular connections of the body vertebrae are 
much reduced and are separated by elastic intervertebral discs affording 
great freedom of movement. The spinous processes are always well de- 
veloped. There is no sacrum and the body and caudal vertebrae are dif- 
ferentiated only by the presence upon the latter of chevrons. The ru- 
dimentary pelvis has no contact with the vertebral column. There are 



no clavicles, and all articulations of the forelimb except that of the 
shoulder are fibrous and virtually immovable. The radius, ulna and 
digital elements are greatly flattened, and the phalanges of the second 
and third digit always exceed the normal number (hyperphalangism) . 
The phalanges have epiphyses at both ends. The salivary glands are 
rudimentary or absent. The diaphragm slopes greatly and extends far 
backward. The vascular system is characterized by extensive retia mira- 
bilia. The testes are abdominal. The internal ears are highly special- 

Odontoceti — the modern toothed whales, including the cachalot or 
sperm whale. The integument of the throat never has numerous 
grooves. The nostrils unite to form a single external orifice, within which 
are a number of diverticula. The epiglottis is elongated and tubular. 
Teeth are present and there is no baleen. The olfactory apparatus is ab- 
sent or extremely rudimentary. The skull is asymmetrical, the right 
side being more developed than the left, the nasal bones are rudimentary 
and the telescoping of the skull is mainly from before posteriorward. 
Mediad of the angle of the mandible there opens a large space into 
which fits a specialization consisting of fatty tissue and trabeculated air 
passages opening from near the Eustachean tube (this may not be pres- 
ent in some sorts) . More than one pair of ribs articulate with the ster- 
num, which is composed of more than a single bony element (although 
these may fuse in adults) . The muscles of the lower arm are very ru- 
dimentary and the digits always number five. There is no vestige of 
the femur present except in the sperm whale. There is no caecum ex- 
cept in the Platanistidae. 

Mysticeti — the whalebone or baleen whales. The integument of the 
throat has deep longitudinal grooves in some genera, but not in others. 
The nostrils have two distinct openings. The epiglottis is not espe- 
cially elongated. Teeth are present only in the fetal state, adults hav- 
ing the roof of the mouth furnished with plates of baleen or "whale- 
bone." The olfactory organ is much reduced but is present. The skull 
is symmetrical, its chief peculiarities for the present purpose being the 
development of the rostrum for the accommodation of the baleen, the 
elevated position of the nares, and the marked anterior inclination of 
the occipital shield, the telescoping being mainly from behind anterior- 
ward. The nasals are much reduced but are not strictly rudimentary. 
But one pair of ribs articulates with the sternum, which is composed of 
only the manubrium. The muscles of the lower arm are not as much re- 
duced as in the Odontoceti and the digits number either four or five. 



There is a vestige of the femur present and, at times, of the proximal 
tibia. There is a small caecum. 

Modern whales are the most highly modified for an exclusively 
aquatic existence of any mammals that have ever lived. Externally some 
of the porpoises are, save in minor details, to all intents replicas of 
ichthyosaurs and some sharks, furnishing perhaps the most spectacular 
instance of convergence in the entire class of mammals. This fact is 
proof positive that the whales of the more typical shape are close to the 
ideal in body form for their habitat. They are excessively specialized 
— in fact more so than any mammal, living or fossil, of which we know 
— and their survival bespeaks eloquently of the lack of ecologic com- 
petition which they have encountered. 

The published discussions regarding the derivation of the Cetacea 
would fill many volumes, but most of this material is open to the same 
objection that has been presented in the case of the pinnipeds. Per- 
haps the majority still believes that they are descended from a hyaeno- 
dont stock, more particularly from the adaptive creodonts (a conten- 
tion refuted by Matthew, Gregory, Kellogg) ; but there exist those who 
argue from numerous other angles, including their descent directly from 
the Pro-Mammalia (Albrecht), the ungulates through the sirenians, 
that their closest aifinities are with the pinnipeds, or the edentates, and 
even that their ancestors never were land mammals but aquatic verte- 
brates. Some of the most interesting evidence in argument against the 
carnivorous ancestry of the whales has been presented by Anthony 
(1926), comprising anatomical details that are notoriously conservative. 
He stated that the following points are common to some of the whales 
and certain ungulates: Lateral nasal diverticula; diverticula of the Eus- 
tachean tubes ; communication of the pleura through the anterior medias- 
tinum; tapetum choroideum of the eye composed of fibrous lamellae; 
absence of os penis (some whales are said to have this) ; ovaries re- 
sembling Hyrax; frequent persistence of Mullet's canals in males; 
placenta adeciduate and diffused; hippomanes body present in fetal 
membrane ; usually but one young at birth ; long period of gestation ; and 
young large and well developed. 

Flower and Kellogg at least have gone on record in their belief that 
the terrestrial ancestors of the whales must be sought among Cretaceous 
remains, and this is a most plausible contention. It seems not improb- 
able that the cetacean stock is older than that of the true carnivores or 
of the ungulates, and therefore this ancestor was presumably a mem- 
ber of the carnivore-insectivore stem that is believed to have been ances- 



Figure 8. Some of the facial and nasal' postuies assumed by the manati, re- 
drawn from life, after Murie. 



tral to a large proportion of present-day mammals. If this were the 
case, then naturally there would be significant resemblances of the whales 
to the ungulates as well as to the carnivores. But far more and far older 
fossil material must be found before anything at all in this direction can 
be proved. 

Modern whales are biserial as are the pinnipeds, and as in the case of 
the latter it seems equally futile to argue as to whether the two groups 
separated before or after they took to the water. Those who are but 
casually acquainted with whales seldom realizq the fundamental dis- 
tinctiveness of the baleen whales or Mystaceti and the toothed whales or 
Odontoceti. The former, usually, but perhaps erroneously, believed to 
be the most recently developed, are known no farther back than the 
Ohgocene, while a few remains of the latter have been detected in the 
Upper Eocene. The differences that occur in the two groups may, how- 
ever, mean nothing more than that their habits or environment differed 
considerably at an early stage of their aquatic career, thus affecting their 
rates of differentiation. For instance the ancestor of the baleen whales 
may have lingered in rivers and shallow lakes, retaining considerable de- 
pendence upon the land, while the toothed whale ancestor may at the 
same time have taken to the sea and severed connection with the land at 
an earlier time ; or vice versa. 

Gigantism in whales is an indication of overspecialization and could 
be developed to the degree in which it now at times occurs only in an 
aquatic mammal completely emancipated from the land. The elephant 
seal, which may exceed 20 feet in length, seems now at the critical point 
in this regard. Some increase in size or a slight increase in its aquatic 
specialization would likely render it unable to leave the water, and we 
do not know whether it would perish or survive as a result. 

Some whales descend to great depths, in the case of the cachalot pre- 
sumably as far as a mile, where the pressure would be more than a ton 
to the square inch, and it is purported to be able to submerge for con- 
siderably more than an hour. These facts bring up a host of physiologic 
questions which are extremely difficult of solution. 


Chapter Three 

External Features 

JroR GREATEST efficiency, which simply means the attainment of con- 
siderable speed with the least expenditure of effort, certain requisites of a 
self-evident nature are necessary to an aquatic mammal of high specializa- 
tion. The body must be of streamline form without excrescences that 
would offer resistance to the water. The propulsive force should be 
from the rear, as has been learned from a study of shipbuilding, and 
there must be some sort of apparatus for steering and equilibration. The 
limitations of mechanical adaptation in a vertebrate necessitates that 
separate means for steering be situated where this will receive no great 
disturbance from motions of the body concomitant to propulsion. Hence 
the rudder must be elsewhere than near the activators for progression. 
But there may be accessory equilibrators, as a dorsal fin. An additional 
necessity is that in an active mammal that is thoroughly aquatic there 
must be an alteration by means of which respiration may be carried on 
while the animal is traveling at full speed. The above qualifications 
are mechanically essential to ideal efficiency, a thesis that may be accepted 
without argument, and is a goal toward which every mammal heads as 
soon as it takes to the water. Whether every such mammal will attain this 
goal eventually is a different matter, and is dependent upon its inherent 
capabilities and inhibitions for such evolution, and environmental factors, 
which at times deflect it at a tangent, so that later it may be incapable of 
regaining the straight course leading to ideal aquatic specialization. 

A streamline body form is not a fixed shape that must fit into a parti- 
cular mold, but it can be long or short, narrow or relatively broad, shal- 
low or deep. It must taper at both ends, after a fashion, however. Some 
terrestrial mammals already have a streamline form, while to attain it 
others must pass through successive steps. As a standard of perfection 
in this regard it is safe to take the swiftest of the cetaceans, and we can 
rest assured that this form is close to the ideal, for it is approached to al- 
most identical degree, with but minor variations, by three diverse phyla 
of vertebrates — the porpoise, ichthyosaur, and certain sharks — as already 
mentioned. Other aquatic mammals that differ from this body form 
either have had insufficient time to attain it, although headed in this 



direction, or else by inherent limitations or influence of environment, 
have been or are being prevented from doing so. But all existing aqua- 
tic mammals, excepting, if one prefer, the hippopotamus, are of stream- 
line form to a greater or lesser degree dependent upon the stage of aqua- 
tic perfection which they approach in other respects. The corpulent 
bulls of the Steller sea-lion and the walrus are more ponderous than we 
should judge best for their aquatic efficiency, but the bulk of the bulls 
of these two animals is doubtless a secondary sexual character the de- 
velopment of which evidently proved to be no critical handicap. Simi- 
larly with the female walrus, only slightly less corpulent. The latter 
pinniped feeds upon inactive prey such as clams. Therefore speed, and 
the more slender proportions which are requisite to its attainment, is not 
a necessity in the securing of food, and it is evident that there was not a 
strong stimulus for it to develop great speed in order to escape from its 
enemies, else it would have become extinct or more speedy long since. 
This shows that the adoption of an aquatic habitat will not develop a 
slender body form, but that the stimulus for this is the necessity to travel 
through the water speedily either for the purpose of securing fast-mov- 
ing food or to escape from rapid enemies. 

Although it has evidently been aquatic for a very long geologic time 
the form of the hippopotamus is far from suited to traveling with rapid- 
ity through the water, although the celerity of its movements is said to 
be quite surprising. It has, however, been under the necessity of lit- 
tle besides sinking sluggishly beneath the surface of its pool or moving 
languidly about. A hippopotamus that could dash about a small river 
at the rate of thirty miles an hour would in all probability promptly 
flatten itself against a rock. In other words, there has been no stimulus 
for this mammal to develop a more perfectly streamline form, or if 
some slight influence of this sort has been experienced, it would doubt- 
less overborne by the greater importance of its ability to seek enormous 
quantities of green fodder on shore. 

The long-tailed zeuglodonts did not conform to the aquatic ideal of 
bodily shape as they were anguilhform, and this relative inefficiency may 
have contributed to their disappearance ; but the short-tailed genera also 
became extinct and the critical factor doubtless lay elsewhere. The 
shape of modern whales is quite variable. Some of the small porpoises 
are very speedy and with ease encircle a fast boat, the inference being 
that they can travel considerably in excess of thirty miles per hour. The 
larger, pelagic sorts of toothed whales are slower, however. The sperm 
whale is decelerated by its enormous head and the mass of its sperma- 


Figure 9. Some of the larger whales, to illustrate type (compiled from various 
sources): (a) Physeter, the sperm whale or cachalot (an odontocete) ; 
{b) Eubalaena glacialis, the right whale; {c) Rhachianectes, the gray whale; 
{d) Megaptera, the humpback; and {e) Balenoptera physalus, the finback. 
The last four are mysticetes or baleen whales. 



ceti organ, but its abundance until harried by man proves that speed was 
not requisite to the filHng of its stomach and its lack of capacity for swift- 
ness, as well, indeed, as in the case of other large, slow whales, is doubt- 
less a secondary acquisition. At least most of the larger, pelagic Odon- 
toceti feed largely on squids and other cephalopods, which are among 
the most agile of sea forms, and how some of the slowest swimming of 
whales manage to secure the vast quantities which they require is a 

Among the baleen whales or Mysticeti, the speed of the more slender 
sorts, the rorquals, is astonishing for a body that sometimes approaches 
a length of 100 feet. By a possible speed estimated to be above thirty 
miles per hour they can often escape the killer whale, which is one of the 
most relentless enemies of the slower genera, and in addition, such speed 
probably secures for them a relative freedom from ectoparasites. At least 
it is a fact that these pests are rarely found upon the fast rorquals, while 
they usually abound upon the slower mysticetes. The baleen whales do 
not secure their food by direct pursuit but by swimming leisurely with 
mouth partly open through hordes of crestaceans and small fish. Hence 
the speed of the rorquals has been developed either for the purpose of 
escaping from thir enemies, or very likely partly to enable them to travel 
rapidly from one feeding ground to another; for the half-inch crusta- 
ceans which constitute their favorite food are local, and may not be easily 
found in the quantities necessary for such vast appetites, so that fre- 
quently it should be necessary for the whales to cover great distances 
between meals. 

The very appearance of these vast cetaceans bespeaks of speed, with 
their relatively slender bodies and great bulge of muscles and tendons 
above and below the peduncle. And equally eloquent are the shorter, 
more ponderous bulks of the gray, humpback and balaenid whales. They 
are much slower and the grays, at least, often succumb to onslaughts of 
the killer whale; but in spite of this they occurred within their rather 
restricted range in prodigious numbers until the depredations of the 
whalers decimated their ranks so sadly. 

When a mammal reaches the stage when it spends most of its time 
in the water it must either have developed a quality of pelage that will 
retain the air that fills the interspaces or have modified its skin to with- 
stand any detrimental action of the water. Especially if it be of boreal 
or even temperate habitat will the tendency rather be for the underfur 
to take on a fine and particularly dense texture, and accompanying this 
will usually if not always be fine adjustments in the functioning of the 



glands of the skin whereby there will be secreted just the correct amount 
of suitable substances to hinder the pelage in becoming water-logged but 
yet insufficient for the matting of the coat. All aquatic insectivores have 
this type of pelage, most rodents have it or seem to be in process of ac- 
quiring it, and it is a character of some of the aquatic carnivores. It is 
not to keep it warm that the Potomogale of tropical Africa has an un- 
dercoat almost as fine and heavy as the Siberian desman, but for the 
purpose of protecting its skin from detrimental action of the water dur- 
ing prolonged submersion, or to accomplish flotation with the aid of 
the air imprisoned in the fur. Among aquatic rodents the most notable, 
and indeed only, example of the lack of an undercoat of fine fur is the 
capybara, whose pelage is markedly coarse. Like all rodents of Cavidae 
affinity its coat was probably coarse to begin with and the easier line to 
follow was to modify its skin rather than its pelage. 

But there are other considerations concerned with the pelage of an 
aquatic mammal. In the case of a mammal of small or even moderate 
size the presence of a coat of fine fur whose surface is plastered smooth 
by the action of the water is doubtless not an appreciable reducer of 
speed, but although perhaps not impossible it is at least unlikely that a 
large whale could ever have had the velvety covering of a sea otter, and 
were a hairy covering present it would doubtless be of wiry texture, 
which would act as a definite retardant of progression comparable to a 
weedy growth upon the hull of a vessel. 

Yet another possibility must receive consideration. The Cetacea and 
Sirenia are the most completely hairless of marine mammals and they 
are the only ones which can not come ashore to bask and dry their hides. 
If they had coats of coarse hair it seems almost a certainty that during 
continuous immersion they would than accumulate such a crop of para- 
sites and sessile marine growth that life would be unbearable for the 
wretched creatures ; and resulting scaly and scabby condition of the hide 
would then inevitably eliminate the hair. The pinnipeds can occasionally 
haul out upon a rock or the ice, thoroughly dry the hide and remove at 
least such unwelcome attendants as need continued immersion. The 
northern fur seal, however, spends many months on its annual migra- 
tion far from land, and the sea otter very seldom leaves the water, so 
that the beautiful, soft pelage of these two mammals refutes any im- 
plication that frequent drying is requisite to the retention of hair in all 
aquatic mammals. 

Among the pinnipeds there are three types of body covering, repre- 
sented by the walrus, which is almost hairless, the majority of earless 



seals or phocids and most of the eared seals or otariids, in which the 
pelage is of course hair only, and the fur seals, famous for the softness 
and denseness of their undercoat. Nothing, of course, is known regard- 
ing the pelage of the pinniped ancestors and it is useless to speculate. 
Undoubtedly the progenitors of the walrus had a coat of some sort, 
which was largely lost for reasons unknown. That the earless seals are 
derived from an ancestor with a more luxuriant and softer pelage is in- 
dicated by the fact that at birth some phocids are covered with an ex- 
cessively long, woolly coat which in some is shed at the age of about 
one month (and until this time they cannot swim) , while in other gen- 
era this coat is shed in utero. And this is an important reason for con- 
sidering the phocids to be more primitive phylogenetically than the 
otariids. Probably contributory to these three conditions of pelage in 
the Pinnipedia are the facts that the walrus is reputed normally to 
have the thickest coat of fat and so has less need for hair; but I cannot 
see that the need of the fur seal for a warm pelage could be greater than 
that of the Steller sea-lion, and it is probable that there is an additional, 
obscure reason for the presence in the former animal of this type of coat. 

The only aquatic mammals with practically naked and rugose hides 
are the walrus and the sirenians. The former was probably derived 
from a hairy ancestor and the reason for its present naked condition 
is obscure. Embryos and young of the latter are said to be more hairy 
than the adult, which would indicate that the ancestral form was also 
more hairy. But the sirenians may never have had a thick coat, for this 
order is commonly believed to have been of the same stock as the 
proboscidians, and elephants have been prone to a hairless condition. 
And the latter theory may well account for the rugosity of the hide of 
these animals. Feeding upon inactive prey and having no great need of 
speed, a rugose hide and its consequent retardation of speedy passage 
through the water would doubtless prove of slight detriment. But 
theoretically no aquatic mammal should attain to high speed and at the 
same time retain a rugose hide. 

The hippopotamus might well be expected to have a rugose hide but 
on the contrary it is remarkably smooth for an animal of this size, and 
its present hairlessness, in a slow mammal that spends much of its time 
on shore, would indicate a hairless ancestor. Its hide appears to be 
rather tender in spite of its great thickness, and it is not surprising that 
some integumental provision has been made for the alternation of a 
whole day spent in the water followed by long terrestrial excursions at 
night in a hot and at times a dry climate. The glands of the skin ex- 



Crete a thick, apparently sticky, pinkish substance which formerly gave 
rise to the belief in the blood-sweating proclivities of this behemoth of 
holy writ. This is presumably accomplished for the purpose of furnish- 
ing a protective covering to guard against dessication of the hide, but 
for all we know to the contrary it may also function in fitting the skin 
for lengthy submergence. 

It was at one time argued by several investigators (as Abel) that a 
number of bony plates discovered associated with the remains of a 
zeuglodont constituted a part of a dermal armor, but it was later shown 
that these plates belonged to a turtle. Several other investigators, chiefly 
Kiikenthal, have vigorously championed the theory that the series of 
dermal dots upon the middorsum of Neomeris, and to a lesser extent 
one or two other genera of porpoises, is a remnant of a dermal arma- 
ment. Kiikenthal also found what he interpreted as dermal ossicles 
upon the anterior margin of the flipper, and scattered over the body of 
the common porpoise. His most significant evidence was that the dor- 
sal rows of dots were much better defined in an embryo of Neomeris 
than in adults; but this character is individually variable, for I have 
found the dots exceptionally sharp in a young specimen and not dis- 
cernible in another of equally tender age (Howell, 1927), nor in an 
adult female. In none of the animals which I have examined did these 
dots assume a squarish shape as claimed by Kiikenthal. Furthermore, in 
preserved specimens of porpoises there is often a roughness of the skin 
upon the anterior border of the flipper caused by the shrinking and 
cracking of the epidermis, and it is by no means unlikely that Kiikenthal 
was misled by his enthusiasm for his theory of dermal armature into 
mistaking such roughness for definite ossicles. A histological study 
showed me that the dots upon the dorsum of Neomeris are formed 
merely by a slight thickening and local cornification of the epidermis, 
and that it is more logical to consider them as the beginning of some 
integumental specialization rather than as the remnant of dermal plates 
— an opinion shared by Winge (1921). 

One would naturally presume that to a mammal highly modified for 
best efficiency in traveling through the water the shape of the head, as 
the part of the body which chiefly has the function of cleaving this ele- 
ment, would be of paramount importance. In our inquiry regarding 
what shape of head is best fitted for this use one naturally turns to the 
Cetacea as being the mammals most highly modified for an aquatic life, 
but from them we can learn but little. It is, of course, a self-evident 
fact that an animal with a pointed snout which tapers gradually to a 


smooth head can cleave the water with least resistance, and conversely 
that no whale with a broad, blunt forehead which piles up the water 
before it can attain to highest speed without expending a disproportion- 
ate amount of energy. But whales' heads are of almost every conveiv- 
able shape, varying all the way from those with an excessively long, 
tweezer-like beak and small head, to the cachalot with its amazing bulk 
of rostral tissue. In consequence we are forced to believe that in the 
Cetacea, no matter how strong the stimulus for a stream-line snout and 
head, the opposing stimulus for the development of an unwieldly frontal 

Figure 10. Heads of porpoises, illustrating frontal prominences and length 
of rostra: (a) Globiucephala; (b) Phocaena; {c) Tursiops; {d) Delphinus; 
{e) the extinct Zarhachis (restoration after Kellogg); and (/) Monodon. 

fat organ, or else for a huge and rather blunt snout (as in the Balaeni- 
dae) has at times proven the more powerful. The inference is there- 
fore drawn that moderately rapid propulsion through the water for many 
millions of years need not necessarily bring about a good stream-line 
form to the head. But it is logical to accept the thesis that the head is 
just as amenable to stream-line influences as is the remainder of the body, 
so it may therefore be accepted as a fact that the stimulus that has at 
times resulted in a cetacean head that offers more resistance to the water 
than all but a very few of the terrestrial Mammalia, is extremely strong. 



In other words, such broad frontal prominences as occur have developed 
not haphazardly but for a purpose that is of the utmost importance to 
those species having them. 

Broadly speaking the shape of the cetacean head is of two sorts. In 
the Odontoceti the throat contour passes almost straight back, while that 
portion of the head above the rostrum offers by far the greater part of 
the cephalic resistance to the water. In the Mysticeti the rule is for 
the dorsal outline of the head to extend straight back from the rostrum 
virtually parallel with the back, so that this part of the head does not 
offer aquatic resistance, while such resistance as is offered is experienced 
by the throat or the lateral dilation of the head. An exception to this 
rule is found in the Balaenidae, with their rostra excessively curved — a 
phenomenon which is a more recent specialization. 

The rostrum of the toothed whales is exceedingly variable. It may 
be so short that the fatty prominence of the forehead projects beyond it 
(Globiocephala) or it may extend slenderly for almost four feet (Zar- 
hachis). Presumably those with no beak to speak of should feed on prey 
that is not particularly agile, while a moderately projecting beak is of 
advantage in snatching fish that move at considerable speed. But theo- 
retically an excessively long beak such as that of the Miocene Zarhachis 
would be more of a disadvantage than otherwise. With a mobile neck 
a beak of this sort could be thrust quickly in all directions largely inde- 
pendently of the position of the body, but in the Cetacea, with their 
much shortened neck, the beak would have but slight directional mo- 
bility, so that to effect a decided alteration in the direction of the beak 
thrust the whole body would have to be moved. Evidently the ances- 
tor of these long-beaked porpoises responded first moderately to some 
stimulus for a lengthening of the rostrum but it seems very likely that 
this modification attained undue evolutional velocity, developed first be- 
yond the needs of the animals and then became of positive disadvan- 
tage, so that it seems likely that the handicap of such an excessively long 
rostrum may have contributed materially to their final extinction. 

The external form of the rostrum in the Mysticeti — at least as we now 
find it — is probably largely attributable to mechanical needs. We can 
tell little about it save that presumably the rostrum has developed in a 
way most suitable to act as a support for the baleen and in response to 
the need for a large mouth. The external grooving upon the throat 
and chest of some of the whalebone whales has evidently been brought 
about in the same way, and will be discussed later. 


Presumably the ancestors of the Cetacea were hairy, as seems clearly 
indicated by the fact that in the fetal state bristles are more numerous 
about the head than in adults, but the hair of the body has long since 
disappeared. Adult mysticetes may have a few bristles about the head 
as well as scattered cutaneous pits, evidently constituting the relics of 
hair follicles. These are most noticeable, at least in Balaenoptera, be- 
neath the chin tip, where they are gathered in a sharply defined cluster, 
which often shows to good advantage in a photograph. In fetuses there 
may be a row of widely spaced bristles along the rostrum, and scat- 
tered at other points upon the head. Among the odontocetes, Stenodel- 
phis is the only one that retains a few bristles in adult life, while Mono- 
don and Delphinapterus (the white whale or beluga) never have them 
at any stage in their development. In Tursiops, at least, among the por- 
poises there are scattered pits about the lips which may have follicular 

The specialization in this direction has progressed farther than in the 
loss of hair, however. The Cetacea have lost sudoriferous glands and 
possibly the sebaceous as well, although authors seem to be at variance 
on the latter point. For instance Beddard (1900) stated that the seba- 
ceous glands have begun to vanish. Certainly they are absent from most 
sections of the skin that have been examined, but perhaps they still 
persist in certain circumscribed areas. 

In a section of the skin of Tursiops before me, prepared by G. B. 
Wislocki, the corneum is tissue thin, and associated with it seem to be 
even thinner elements of a stratum lucidum. Next comes an excessively 
thickened stratum germinativum of very homogeneous character. The 
papillae of the corium are unusually slender and long (for their thick- 
ness), while the coreum insensibly merges with the tela subcutanae— 
the true blubber layer. In reality the blubber layer should probably 
be considered as a definite component of the coreum, for It is very 
tough and fibrous, whether relatively collapsed (in an emaciated ani- 
mal) , or gorged with fat cells, and totally different from the simple layer 
of soft, subcutaneous fat of the seals. 

In some ways the whole layer of skin and blubber is remarkably 
tender. In a freshly caught Balaenoptera the corneum is relatively even 
thinner and more tender than in a porpoise, and may be rubbed oflf in 
great patches by the palm of one's hand, the resulting sheet resembling 
tissue paper. I have easily scored the surface with a finger nail, and af- 
ter inserting a pocket knife for the full length of the blade, one may cut 
as easily as through so much cheese. One would suppose that the first 



scrape of so ponderous a bulk against a sharp rock would lay the ani- 
mal completely open. And yet many hundreds of pounds may be ap- 
plied to a cable passed through an incision in this whole layer without 
it tearing away. 

The constitution of the blubber layer in the Cetacea certainly appears 
unusual, but whether its precise histological structure is or is not unique 
among the Mammalia has not been determined. At any rate the mere 
presence of a considerable layer of fatty tissue beneath the integument of 
aquatic mammals is not surprising. Many land mammals store up just 
such a supply of extra fuel both for warmth and to tide them over periods 
of food shortage. Any carnivore must be prepared to live through times 
of hunger and this should be true in the case of large whales which re- 
quire such a prodigeous quantity of food, often of small size, and which 
travel so widely. Any sort of mammal which experiences times of 
plenty alternating with periods of hunger will quickly acquire the abil- 
ity to accumulate a reserve of fat. Some sort of insulation of the body 
is probably of critical necessity to an arctic marine mammal, for although 
it never experiences an aquatic temperature lower than a few degrees 
below freezing, the conductivity of the water is 27 times that of air. 
But there are probably other advantages derivable from the presence of 
a fat layer in such mammals. The tropical sirenians surely have little 
need for insulation and yet they are not only abundantly supplied with 
fat, but this seems singularly inefficient in insulating the body, for the 
manatis of Florida are known quickly to succumb to any unusual low- 
ering of the air temperature. Furthermore, we cannot know just how 
advantageous it is to an aquatic mammal to have the added buoyancy af- 
forded by an extensive deposit of fat. The blubber layer itself may be 
significant of nothing more than stated above, but the excessive oiliness 
of other parts of the body in marine mammals and of the proneness of 
the Odontoceti to develop fat organs of some sort seems to point to the 
probability of all this fat and oil serving some other and very important 
physiological purpose of which we do not know. This question will re- 
ceive further consideration elsewhere. 

It is perhaps proper here to discuss the question of a dorsal fin, al- 
though little can be stated in this connection. Some whales are without 
it (Balaenidae, Rhachianectes, Physeteridae, Delphinapterus, Monodon, 
Phocaena) ; in some it is barely indicated and too small for function, 
while in most odontocetes it is quite large, culminating in the killer 
whales, in which it may attain a length of several feet. All that it is 
now safe to say is that all aquatic vertebrates are prone to develop such 



a fin, almost certainly as an aid to equilibration. It should be men- 
tioned, however, that a dorsal fin can hardly act very efficiently as an 
equilibrator in an animal swimming by a vertical motion of the tail (as 
whales) as it can in one employing a horizontal motion (as most fish), 
for equilibrators should be situated at a right angle to the direction of the 
swimming force. In the Cetacea, therefore, the situation of the flippers 

Figure 11. Some of the extinct aquatic reptiles: (a) Mesosaurus ; (b) Elasmo- 
saurus (plesiosaur) ; (r) Clidastes (mosasaur) ; {d) Ichthyosaurus; 
(e) Trimacromerum (plesiosaur): (redrawn from Williston). 

renders these appendages by all odds the most efficient organs of equili- 
bration and it seems that a dorsal fin would not be of sufficient practical 
importance for all sorts of whales invariably to have developed it. It 
may be mentioned that theoretically a dorsal fin would be of greater use 
to the Phocidae in pure equilibration than the fore limbs, but these mam- 
mals are not yet sufficiently specialized to have developed the former. 



Other external features, such as position of the narial openings, length 
of neck, tail, and form of appendages will be discussed in detail in other 
chapters. Suffice it to say here that in the highly modified aquatic mam- 
mal the tendency always has been for the elimination of any unevenness 
of body contour that could offer resistance during progression through 
the water. The pinna of the ear disappears, the scrotum is eliminated, 
and the appendages that are not used definitely either for propulsion or 
in equilibration will first atrophy and then sink beneath the surface of 
the body. 


Chapter Four 

The Senses 


C^ERTAIN adjustments in the optic equipment of mammals that habitu- 
ally seek their food beneath the surface of the water is to be expected. 
The platypus, most of the insectivores and some of the smaller aquatic 
rodents secure their food largely through the sense of touch. It seems 
that they must actually keep their eyes closed for a large part of the 
time which they spend submerged, and the result is the same as is found 
in fossorial mammals, namely, a tendency toward reduction in the size 
of the eye. Among Cetacea it is interesting to note that Platanista 
at least has followed this same course, for it has become virtually blind 
after having lived for a great length of time in the waters of muddy 

Other sorts of mammals may have experienced a pressing need to 
watch for enemies above the water while keeping a maximum amount 
of head and body hidden from view. The result of this is best exem- 
plified among mammals by the common hippopotamus, with its dor- 
sally protruding orbits. The pigmy hippopotamus (Choeiopsis), how- 
ever, does not have markedly protruding orbits but appears more like 
a young individual of its larger cousin, indicating that the latter is prob- 
ably derived from a more generalized ancestry — not vice versa. The 
hair seals (Phocidae) show a tendency to acquire dorsal direction of 
vision to a considerable extent and where this is pronounced there is 
osteological indication of it in extreme interorbital constriction. The 
upward pointing of the axis of the eye is to be noted in the sea-lion 
(Otariidae) and walrus (Odobenidae) also but to a lesser degree. The 
eye in at least the majority of the Phocidae is relatively larger than in 
other pinnipeds, possibly because these seals are more in the habit of 
seeking smaller prey in the subdued light of deep waters, although it 
must be admitted that the dorsal direction of the eyes would hinder such 
action. Incidentally I have found scores of half inch shrimps in the 
stomach of a Phoca his[>ida, which fare is probably too insignificant to 
interest an otariid. 



In the Sirenia the dorsally directed axis of the eye is also quite ap- 
parent, but the size of the eyeball is very much reduced and in conse- 
quence it is inferred that visual efficiency has become much impaired. 

The beaver probably has a greater dorsal inclination of the eye than 
any other rodent, as might be expected, and the muskrat, capybara and 
other aquatic sorts exhibit this character to some degree, but in rodents 
this may be without much significance from the present standpoint, for 
some species that are strictly terrestrial have this character considerably 
developed. Among these are the Microtinae or meadow mice, most 
of which follow runways through grass and must watch above for ap- 
proaching danger. 

There may be marked visual differences, however, in eyes that have 
moved dorsad. One extreme is represented by Hippopotamus in which 
the protruding bony orbits are directed chiefly laterally with a slight for- 
ward trend, thus indicating vision that is largely monocular. The well 
developed supraorbital prominences of the platypus also obliges vision 
in this animal to be completely monocular. The tendency of some ro- 
dents and the hair seals, however, seems to be for the acquirement of 
dorsal binocular vision, although it is probable that none of these 
mammals has the power of true stereoscopic vision, in which the image 
of one eye is exactly superimposed upon that of the other and the two 
function together perfectly as a unit. 

Such binocular visual powers as the seal may have are doubtless of 
a character somewhat intermediate between monocular and stereoscopic 
vision. In order that the visual sense shall function perfectly in a reflex 
manner the physiology must be quite complicated, and our knowledge of 
the mechanics of the sort of monocular vision which must be employed 
by an animal having the eyes upon opposite sides of the head is still in- 
complete. Thus it is not certain whether the animal must first give 
its attention to an object upon the right and then to another upon the 
left, or whether both eyes can function as separate units with equal effi- 
ciency simultaneously, in which case the optic colliculus of the brain 
would have to function in a manner differing in some unknown degree 
from that of man. Also it is unknown whether the apparatus for ac- 
commodation of each eye can operate entirely independently of the 
other, so that the animal can simultaneously focus one eye on a near, 
and the other on a distant, object with an equal degree of perfection. 

It may be presumed that a tendency for the elevation of the optic 
axis of the character encountered in the seal is useful for the purpose 
of detecting enemies that are prone to pounce down from above, while 



the hippopotamus has had need to watch for foes approaching hori- 
zontally along the river banks. At any rate it seems likely that only 
those aquatic mammals which have experienced a definite need for 
eyes directed more dorsally will have acquired this characteristic, and 
one would not expect it to develop to as high a degree in the great 
majority of cases as it now occurs in the anomalous hippopotamus. 
For one thing, unless the need for it was extraordinarily strong, by the 
time that dorsal vision had become moderately efficient it is probable 
that aquatic specialization in other respects would usually be so far 
advanced that the animal would have but little reason to fear enemies 
approaching above the surface of the water. 

As soon as the last mentioned stage in aquatic specialization has been 
reached the occular stimulus would automatically change, for the chief 
need, and probably the only one, would then be for the discernment 
of submarine food and enemies. In the case of a mammal whose eyes 
had already turned upward to a greater or lesser amount, there would 
then be a secondary migration downward of the direction of vision, 
so as better to detect food which it was approaching. For all we know 
the progenitors of the whales may have passed through just this visual 
cycle, in which case there would have been left some complication in 
skull development which one cannot hope to decipher by means of the 
fossils now available. 

Horizontal direction of vision may be either forward or lateral, or 
even backward. If the former then it must presumably be binocular, 
and if lateral, monocular. We have no means of knowing which of 
these sorts of sight might prove most useful to an aquatic mammal but 
it seems likely that binocular vision would never be developed by any 
reasonably active mammal of high aquatic specialization. This would 
entail a forward direction of the orbits where the eyes would receive 
the full force of water pressure as the animal progressed — a position 
which obviously would prove of considerable, and perhaps critical, 
detriment. As a matter of fact the eyes of the Cetacea are directed at 
practically a perfect right angle to the body axis, where they receive 
the minimum of friction and irritative interference by the water. 

Eyelids that are at least partly functional are retained by aquatic 
mammals. In pinnipeds, serenians (apparently) and mysticetes these 
are largely comparable to what are found in terrestrial mammals, with 
lids moderately wrinkled, indicating that they may easily be closed 
and held in that position without strain as long as the animal wishes. 
This statement should perhaps be qualified as regards the whalebone 



whales. From examinations of many freshly killed specimens of the 
latter I am convinced that it is so, for the lids may be moved with the 
fingers very easily; but Putter has stated that they are immovable, as 
is also the case in odontocetes. It is true that in at least the majority 
of toothed whales the lids are unwrinkled when the eye is open. 
It therefore seems likely that their mobility is very much reduced. 
Complete closure of the eye is perhaps impossible, but I would be 
loath to believe that the lids are without power of movement entirely. 
The narwhal ( Monodon) , and perhaps some others, has a very peculiar 
modification of the eyelids, which will be described in detail by Ernst 

An optic tendency resulting from a life in the water is for the elim- 
ination of eyelashes, but their disappearance is slow. Lashes are en- 
tirely lacking in the Cetacea and I understand that they are practically 
so in the Sirenia. All other aquatic mammals have them in various 
degrees. One might expect that there would be a great advantage in 
the acquirement of a nictitating membrane, but the only sorts which 
have them are the Sirenia and Pinnipedia, so far as I have been able 
to learn. There should hardly be any use for a functional lachrymal 
duct in aquatic mammals and this shows a tendency to disappear (Ceta- 
cea, Sirenia and Pinnipedia) . 

Frequent or constant submergence will probably bring about some 
sort of alteration in composition of the lachrymal fluid in order that 
this will not so readily be washed from the eye, and very possibly in 
order to counteract any irritative influence of the sea water; and Putter 
has determined that this is actually the case. It should be noted in 
this connection that cetaceans are perhaps the only mammals which 
cannot rub the eye against some part of the body in order to free it of 
foreign particles, including parasites. 

Very probably there need be no optic stimulus in addition to the 
above for aquatic mammals inhabiting fresh water. The refractive 
index of the latter is not so different from that of air but that the 
normal eye can see well beneath the surface and there is no need for 
it to withstand any considerable pressure. In a marine mammal such 
as the whale, however, matters are entirely different. The refractive 
index of sea water differs so much from that of fresh or of air that 
when the normal eye is submerged in it very little can be done but to 
distinguish objects by their high lights and shadows, because of the 
fact that the point of focus for the image then lies far behind the 
retina. Hence, for a whale to see properly when submerged means 



that there will have been quite profound changes in its visual apparatus ; 
and in addition there must be alterations to enable the eye to with- 
stand the enormous pressure to which it is at times subjected. 

Putter (1902) has investigated this subject more thoroughly than 
anyone else and it is upon him that we must rely for our facts. He 
found that pinnipeds and sirenians also have the eyes importantly modi- 
fied, but his material representing the latter group was not altogether 
satisfactory. The changes which he found to have taken place in the 
eyes of pinnipeds, sirenians, mysticetes and odontocetes he has sum- 
marized as follows: 

Optic adaptations 

1. Lens almost spherical. 

2. Refractive index higher than in any terrestrial mammal; almost 
as great as in fish. 

3. Relationship of the neural elements of the retina; unusual num- 
ber of rods connected with a single ganglion cell. 

4. Superfluous ganglion cells in outer granular layer of retina. 

5. Extensive tapetum lucidum. 

6. Enlargement of fundus at the expense of the pre-equitorial zone. 
Thermal adaptations 

1. Diminution of the cornea in proportion to size of the bulbus. 

2. Form and number of lymphatics in cornea propria; large and 
relatively few in number. 

3. Unusual development of choroid and of the perichoroidal lymph 

4. Form of optic orifice: is so much reduced that only the cornea 
is visible. 

5. Tremendous development of musculature, with immovable bul- 

Hydrostatic adaptations 

1. Curvature of the cornea so as to receive support laterally. 

2. Lateral thickening of the cornea. 

3. Epitheleal cornification of cornea; cornified substance unites di- 
rectly with elastica anterior. 

4. Thickening of sclera; tremendous at equator and above fundus, 
slight at corneal sulcus. 

5. Thick optic sheath; supporting bulbus like a column. 
Arterial and venous network of the ciliary blood vessels (retia 
mirabilia) . 




7. Location of the bulbus away from the bony walls and embedded 
in muscular, fatty and glandular tissue. 

8. Acquisition of a special hydrostatic sensory organ in the Odon- 

Chemical adaptations 

1 . Development of glands to produce a fatty, oily secretion. 

2. Increase in size of Harter's and lachrymal glands, and develop- 
ment of a subconjunctival stratum of glands. 

Putter suggested that as the eyes are immovable, the very large re- 
tractor muscles may have developed a specialized function and that it 
is possible that they may now act to develop heat to warm the eye. 
But there is apparently not the slightest basis for this theory. Even if 
these muscles were continually contracted and relaxed the heat furnished 
could hardly be a fraction of that supplied to the eye by the blood 
through the retia mirabilia. Similarly with the structure in the eyes 
of odontocetes which he interpreted as a hydrostatic organ. This was 
a bit of neural epithelium in the connective tissue of the sclera at the 
angle of the iris and isolated from the retina so that it could receive 
no stimulus by means of light. I can see no reason for considering 
this to have any hydrostatic function. 

In summarizing the aquatic specialization which the whale's eye. has 
undergone Kellogg (1928) has stated that "in its gross features the 
whale eye differs from that of a land mammal in having an eyeball 
immovable, eyelids without eyelashes, no tarsus or supporting cartilage 
in the eyelid, no Meibomian glands, and a downward direction of the 
eye axis. As the result of an aquatic mode of life whales have acquired 
a more spherical lens and a greatly thickened sclera. The ciliary pro- 
cesses and their muscles are reduced in size and have lost their original 
function of controlling the shape of the lens. The tension of the 
suspensory membrane (the zonula Zinii) is not great enough to flatten 
the anterior surface of the lens, and as a general rule the latter retains 
a more or less spherical shape. Whales thus lack the power of ac- 

The more spherical lens projects the image upon a retina that is rela- 
tively much nearer the lens than in the normal eye. This is accom- 
panied by an enormously thickened sclera, and it is this which stands 
the optic stress experienced during deep diving. 

The above lack of accommodation is not unique, for it is a character 
shared by a number of land mammals. 


The manner in which Putter found the eyes of pinnipeds to differ 
from conditions in the Cetacea may be summarized as follows: The 
pre-equitorial segment is thick and the equitorial segment thin, while 
the latter is thick in whales; the fundus segment is thick in all. The 
choroid is thin in pinnipeds and mysticetes, but thick in odontocetes. 
The tapetum lucidum is poorly developed. The ciliary muscle is feeble 
in pinnipeds, absent in mysticetes, and represented by a few fibers in 
odontocetes. The ciliary processes are moderately long in pinnipeds but 
short in cetaceans. The lens is larger than in whales and the rods of 
the retina very long. Although the sheath of the optic nerve is not 
so robust as in whales, it is much thicker than normal. The tarsus of 
the lid, while poorly developed, is more so than in whales. Meibomian 
glands are absent in all three. Nictitating membranes are present in 
pinnipeds (and sirenians) but absent in whales. The eyeball in pinni- 
peds is slightly movable, which is not the case in cetaceans. In the 
former the axis of the eye is horizontal or directed somewhat dorsad, 
while in the latter it is horizontal or directed somewhat ventrad. It is 
further mentioned that of all mammals sirenians may be said to have 
the cornea least developed. 

As Kellogg has suggested, adaptive visual changes, when the need 
for them has arisen, are of critical importance to a marine mammal. 
He pointed out that the atrophy of the optic chiasma in the zeuglodont 
brain indicates a failure in their visual apparatus very possibly because 
of an inability to adjust their vision to salt water requirements, and 
this may very well have been the deciding factor in their eventual 

Kellogg's investigations of cetacean eyes were based on mysticete 
material and he informs me that there is not the slightest doubt but 
that the eyes of this group are well nigh useless for seeing above the 
surface of the water. Putter stated that the eyes of odontocetes and 
mysticetes differ in many important respects and to a degree that indi- 
cates that in each the eye has experienced its own particular type of 
specialization from a more generalized optic equipment. Having never 
worked on this question myself I am in no position to point out any 
erroneous conclusions which this author may have reached, but it 
seems that he was mistaken in ascribing to mysticetes, odontocetes and 
pinnipeds a common inability to see effectively through an atmospheric 
medium, very likely for the reason that apparently he did not determine 
mathematically the refractive powers of their visual equipment. At 
any rate it is well known that when a watchful seal is basking it is 


quick to discover the approach of a polar bear or other enemy, and 
that not only the killer whale but the cachalot is in the habit of thrusting 
the head above the water for the purpose of ascertaining what may be 
of interest above the surface. Hence it is evident that there yet remain 
for settlement several fundamental questions regarding the optics of 
aquatic mammals. 


There is a tendency for the elimination of the external ear in aquatic 
mammals, partly following the law that ultimately the aquatic life will 
eliminate superfluous prominences upon the body, and partly because 
of the ultimate disuse of the pinna as an accumulator of atmospheric 
sound waves. In the case of those sorts which inhabit streams this re- 
duction of the pinna may be barely or not at all appreciable, partly be- 
cause they are not very highly modified aquatically if they are still 
stream dwellers, and partly for the reason that although hearing has 
ceased to be an aid in capturing aquatic prey, they must still be on the 
alert for terrestrial foes. Also it must not be forgotten that mammals 
of this category are almost always densely furred, and that in these the 
ears are usually hidden in the pelage, constituting elimination of pinna 
as far as concerns the external contour. But this need not be of aquatic 
significance since many terrestrial representatives of these mammals 
(among Insectivora and Rodentia) exhibit the same condition. 

The platypus, anomalous in so many ways, has rather peculiar ears. 
Although it lacks the pinna, the musculature enables it to "cock" the 
orifice forward, as is shown to excellent advantage in plate 6 of Bur- 
rell's (1927) book. This authority also states that the auditory "ori- 
fice lies at the posterior end of a facial furrow, the eye lying at the an- 
terior end, while the furrow is incompletely divided into two by an 
oblique fold of skin. The edges of this furrow act as a long pair of 
lids, by means of which both eye and ear may be tightly closed at the 
will of the animal. The aural aperture may also be dilated and con- 
tracted while the eyes are open." 

The aquatic Talpidae — Desmana and Galemys — also lack the pinna, 
which has no aquatic import, for they belong to the earless family of 
moles. Nectogale is the only other insectivore which is said to lack a 
pinna. And Anotomys is the only such rodent, but I question this 
without verification (I have not seen this genus) . Among pelagic 
mammals the Cetacea, Sirenia, Phocidae, and Odobenidae have no ex- 
ternal ears. 



It is thus seen that mammals which inhabit fresh water usually re- 
tain the pinna, while pelagic mammals have mostly lost it, both be- 
cause they are more highly aquatic and because they have less need for 
it. It is retained by the sea otter and Otariidae, but in the latter 
the pinna is but a remnant and too slender to act as an acoustic aid. 

If an aquatic mammal show by the position of its eyes that it has 
experienced a need for peering above the surface of the water with the 
minimum of its body exposed to view — in other words, if its eyes have 
assumed a somewhat dorsal direction, then has it similarly experienced 
a need for hearing in this position, and the external orifices of its ears 
will also be situated more dorsal than usual. This is not apparent in 
any insectivore and is at all marked among rodents only in those aqua- 
tic forms having a rather flat skull, in which the external auditory meatus 
is relatively close to the dorsal profile (as Castor), and in such this de- 
velopment may have nothing to do with an aquatic modification, as 
many terrestrial rodents exhibit the same character. The well formed 
ears of the hippopotamus are markedly dorsal in position. In all Pin- 
nipedia the auditory lumen does not extend directly laterad from the 
meatus but turns quite sharply upward, reaching the body surface con- 
siderably dorsad of the meatus. This is most marked in the Phocidae 
and least so in the Otariidae, and is an expected condition. I strongly 
suspect that this is also the case in the Sirenia. In the Mysticeti (at 
least in Balaenoptera) the auditory tube extends practically in a direct 
lateral line to the surface of the body. At least in Neomeris, among the 
Odontoceti, there is a sharp bend of the lumen upward for about an 
inch. If this should prove to be a uniform character among the toothed 
whales it should be of considerable phylogenetic importance. 

As already mentioned the ability to close the ear so as to exclude 
water is one of the first acquirements of an aquatic mammal and it is 
confidently believed that all mammals that may be so classed have it. 
It is one which is easy of accomplishment. Most terrestrial mammals 
which burrow also have it, for the purpose of keeping loose soil out of 
the ear. Perhaps most often this closure is accomplished by the pulling 
into the orifice of a valvular plug, probably homologous with the anti- 
tragus, specialized for this purpose. This may be seen to excellent ad- 
vantage in the case of the seal (Phoca) but the precise functioning of 
the mechanism is not uniform and the result may be attained in different 
genera, and even in well differentiated species, by complex variations 
in the actions of the small auricular muscles, which are relatively very 
well developed. A somewhat difl^erent method for closure is employed 



by the hippopotamus, for it may be seen that directly preceding sub- 
mergence this animal contracts the entire base of the ear, which evi- 
dently compresses and shortens the lumen. The extremely vigorous 
and repeated twitching of the ears upon its reappearance is perhaps in- 
dicative that the closure is not directly at the orifice, but slightly deeper. 
so that some effort is necessary to dispose of lodged water. 

Still another way of closure is employed by the sea-lions. The car- 
tilage is very slender and somewhat furled, and muscular action results 
in further and tighter furling, and perhaps some longitudinal contrac- 
tion of the lumen. For this method of excluding the water at least 
some remnant of the external pinna would seem to be of distinct ad- 
vantage, and it is not unlikely that this is the sole reason for the reten- 
tion of a part of the external ear by otariids, while the lack of such 
need has hastened its disappearance in phocids and the walrus. 

Among Sirenia I have had an opportunity for examining only a 
hardened dugong. The auditory aperture in this was relatively smaller 
than in a seal, but larger than in the Cetacea, and appeared to be fairly 
intermediate in other respects. 

At first, in the early history of a hypothetical aquatic phylum, the 
closed position of the ear would be retained with some muscular exer- 
tion, and later with ease. A point would inevitably be reached, at a 
time when the mammal had ceased to leave the water for more than 
short periods, spending the greater portion of its time with the ears be- 
neath the surface, when the action of the auricular musculature had so 
changed that the closed position would be the one maintained involun- 
tarily, while muscular effort would be required to open the ear. Whether 
or not any pinniped has yet attained this stage is unknown. 

Thence it would be but a relatively short time until, through disuse 
of the musculature involved, the ear could not be opened at all, the au- 
ditory lumen would first remain permanently water-tight, and because 
its apparatus for closure had throughout a long period been operating 
with but little or no opposition from the opening musculature, its cali- 
ber would be reduced. This in effect, is just what is now to be found in 
the Cetacea. I have introduced a match stick into the auditory tube of a 
Balaenoptera 75 feet long, but it would not accommodate anything 
larger; and in some odontocetes it is even smaller. In the latter, as 
would be expected because of the continued action here in the past of 
the musculature for closure, the lumen is smallest at the distal end, ex- 
panding to a slightly greater diameter proximad. In the Mysticeti the 
external orifice is also small, the lumen thence expanding for a short 



distance until it is sealed entirely, once more becoming open and again 
expanding farther proximad. 

It is usually stated in the literature that auricular musculature has 
disappeared entirely in the Odontoceti, and is vestigeally represented by 
one or two remnants in the Mysticeti, but few investigators have spec- 
ialized sufficiently in the facial musculature to do a thoroughly satis- 
factory dissection of this portion of the Cetacea. A slip, believed to be 
a vestige of the ear muscles, was found by me in Neomeris but because 
of the bad condition of the specimen I could not be positive. Ernst 
Huber has found remnants of several auricular muscles in the narwhal, 
and I confidently expect that there are actually more than the two usu- 
ally stated to occur in mysticetes. A vestige of the auricular cartilage 
has also been found in some whales. 

It is extremely doubtful if any aquatic mammal swims with the audi- 
tory tubes open. Hence, as the tube is closed when beneath the water 
a mammal can use its ear in normal manner only for air-borne sounds. 
Thus the ears of a mink or beaver are probably largely inoperative un- 
der water save as sound waves may be transmitted by resonance through 
parts of the head. When, in an aquatic mammal, the opening me- 
chanism of the ear has ceased to function and the lumen remains closed, 
normal use of the ears will cease forever, for then neither air nor water 
can transmit extraneous sounds directly to the ear drum. The latter 
situation now obtains in the Cetacea. But there must be a gradual ac- 
commodation to this change — a gradual increase in ability to receive 
water borne vibrations and a gradual decrease in the power to receive 
those transmitted by air. And there must be a change in the quality of 
reception also, for it is unthinkable that during the thousands of years 
since the abandonment of atmospheric hearing in the Cetacea they re- 
ceive sounds under water only after the same fashion as do we when the 
head is submerged. It is not unlikely that the Pinnipedia are now un- 
dergoing this auditory alteration, and that the change is more marked in 
the Sirenia. 

There is abundant evidence that whales are sensitive to certain water 
borne vibrations which cannot possibly be transmitted through their 
auditory or Eustachian tubes. Hence these waves must be transmitted 
through some solid part of the head, but we have no means of ascer- 
taining which part is most resonant.^ There is further proof of ceta- 
cean keenness of hearing in the high development of the internal ear, 
and in the size and character of the acoustic colliculus, which in a por- 



poise G. L. Streeter (in Kellogg, 1928) found to be four times the size 
of the optic colliculus. 

The middle and inner ear of the Cetacea are sufficiently modified for 
us to be sure that the function is altered. Presumably the ear of the 
Sirenia has also undergone modification, but little or nothing is known 
about it. It is also permissible to surmise that in the Pinnipedia some 
change is going on, perhaps of two sorts, for the auditory bulla of most 
genera of seals is considerably inflated, while that of the sea-lion is 
not, appearing shrunken and rugose. Gray (1905) has reported on the 
former. He stated that the inner ear of this mammal is larger than in 
any other except the walrus. He found that two otoliths of remarkable 
size were present in the vestibule and as these were unlike those of any 
mammal yet recorded he surmised that they must have some particular 
physiological function. It is unknown whether otoliths are present in 
the Cetacea. That the large vestibule of the Phocidae is not a simple 
corollary of the aquatic life is indicated by the fact that in the Cetacea 
this is particularly small. 

The ear bones of the Cetacea are characterized by extreme hardness. 
Kellogg (1928) said "the tympanic bulla is the relatively dense and 
heavy sounding box fastened to the periotic by two thin pedicles, which 
can be set in vibration. Vibrations set up in these pedicles produce a 
corresponding amount of motion in the malleolus, whose anterior pro- 
cess is likewise fused with the bulla between these pedicles, and it in 
turn transmits these vibrations to the incus and stapes." The mal- 
leus is rigid and the stapes is immovable in the vestibule. Kellogg fur- 
ther said that "water-borne vibrations transmitted to the air contained in 
the tympanic bulla cause it to function as a sounding box, and its vi- 
brations reach the cochlea by way of the ossicular chain and vestibule." 
Kernan and Schulte (1918) stated that Denker (1902) thoroughly dem- 
onstrated vibration of the ossicular chain to be impossible, but the latter 
merely claimed that vibration cannot be activated by the tympanic mem- 
brane, for this is too lax for any such function. Kernan and Schulte 
also mentioned what is well known clinically, that an increase in the con- 
duction of sound through the bones of the head accompanies dimin- 
ished function of the middle ear (this is the condition in the author) . 

The ear bones of the Mysticeti and Odontoceti are of two distinct 
shapes and this is accompanied by other differences. In the former the 
ear drum looks like the finger of a gray leather glove, moderately pliable 
when fresh but inelastic, the part representing the finger tip extending 
distad into the auditory tube. In the Odontoceti the tympanic mem- 



brane is not at all finger-shaped but is gently bowed and often partly 
calcified. In the whalebone whales (at least in Balaenoptera) , the an- 
terior part of the bare bulla, without membranous covering, projects 
into a fossa the size of one's two fists, and in freshly killed specimens 
this is entirely filled with a coarse foam of albuminous, rather than 
greasy, texture. Whether this is so in living specimens cannot be dem- 
onstrated, but presumably it is, and the foam may have some function in 
determining the quality of sound reception. There is free communica- 
tion between this fossa and the choanae. In the odontocetes there is a 
different but analogous system of air sinuses adjoining the inner ear and 
connecting with an intricate labyrinth of ducts. Authors have been very 
vague and cautious about describing these ducts, and with good reason, 
for without the injection of a suitable colored mass into this part of a 
freshly killed specimen their proper definition is uttterly impossible, as 
their finer ramifications are otherwise not to be distinguished from ad- 
joining blood vessels and oil ducts. It must therefore suffice to say that 
this system of air sinuses communicate with the choanae and apparently 
send trabeculated branches ramifying through the peculiar fatty tissue 
that occurs in the odontocetes within the angle of the lower jaw. 

As already stated it is well known that cetaceans are sensitive to cer- 
tain sorts of water borne vibrations. It has been reported that por- 
poises are peculiarly sensitive to the waves that are transmitted by the 
sonic depth finder and will disappear in great haste and apparent dis- 
comfort form the vicinity of a vessel when one of these contrivances is 
put in operation. This fact suggests that the Cetacea may be sensi- 
tive to water-borne sound waves of a character and after a fashion that 
we do not yet understand. The transmission of the sounds that reach 
them and the ears themselves are so different from anything connected 
with our own acoustic apparatus that my personal opinion is to the ef- 
fect that we know nothing whatever about the matter. 

The function of a hydrostatic organ or depth gauge has been as- 
signed to the air sinuses and passages about the cetacean ear, but it has 
become the practice to assign this function to any part of cetacean an- 
atomy which seems in any way unique or peculiar. I cannot see that 
any sort of a hydrostatic organ would be necessary in this order, for they 
can always tell when water pressure upon the body becomes too great 
for comfort or safety. 

Numerous writers have concerned themselves with the question of 
how the air within the inner ear is equalized to correspond with the 
great external pressure experienced during deep diving. I cannot see 



that any equalization is necessary. There is no tension of the cetacean 
tympanic membrane and hence no need for nice adjustments of pres- 
sure within the ear to that of the auditory tube, such as is often so 
annoying to us during speedy ascent of a mountain. No water, and 
hence no pressure save that appHed to the body surface can reach the 
ear through the auditory tube. Judging by the powerful rush of the 
air which leaves the cetacean blow hole as soon as the nasal orifice is 
opened, the air is evidently retained between breaths at considerable 
pressure — greater than the external pressure when the body is at the 
surface, and it seems that this air pressure must reach the inner ear 
through the Eustachian tube, at least during the beginning of expira- 
tion. It is therefore logical to assume that the inner ear has been simply 
adjusted to withstand any pressure experienced without the necessity for 
nice muscular adjustments for the equalization of pressure. 


It is strikingly apparent that a mammal which seeks its food exclu- 
sively beneath the surface of the water and comes to land only for brief 
periods of basking or to have its young on some safe island retreat, from 
whence it can dive into the water without an instant's delay, has vir- 
tually no need for a sense of smell. Hence it is rather remarkable that 
the Pinnipedia are so well equipped with olfactory apparatus. The ol- 
factory bulbs are not as well developed as in most terrestrial carnivores, 
according to O. R. Langworthy (MS), but the sense of smell is by no 
means vestigeal in this order and turbinal bones of considerable com- 
plexity are retained. It accordingly seems likely that a fairly well de- 
veloped olfactory apparatus is a character which is not readily relin- 
quished, at least by mammals of this sort, and that it will be retained 
considerably after any great need for it has passed. 

Probably no mammal less specialized for an aquatic life than the 
pinniped has had the olfactory apparatus appreciably reduced. In the 
Sirenia it is considerably more reduced than in the Pinnipedia, and 
Owen (1868) mentioned that the olfactory nerves are fewer and the 
cribriform plates smaller in the dugong than the manati, as might be 
expected from the greater aquatic specialization of the former. But 
Murie (1872) stated that the size of the olfactory bulbs of the manati 
indicates that its sense of smell is "fairly well developed." It is in the 
Cetacea, however, than one finds the greatest olfactory alteration occur- 
ring among aquatic mammals. The whalebone whales retain an ol- 
factory apparatus, although it is vestigeal, but for what purpose it is 



difficult to understand, for ages ago they must have ceased to have the 
slightest need for it. The olfactory nerves are simple and after pierc- 
ing the cribriform plates, are distributed to the mucous membrane of 
the narial passages. Needless to say, water cannot come into contact 
with these nerve endings, so they could not have taken over any aber- 
rant taste function, and it is well nigh inconceivable that any message 
which they might receive by this means from the atmosphere would 
have any significance for them. Kellogg (1928), however, thinks that 
"the retention of the sense of smell (in the mysticetes) may be due in 
a larger measure to the actual mechanical construction of the skull than 
to the need of such organs." 

Olfactory apparatus has ben found in fetal Odontoceti but the adults 
have lost it and the cribriform plates are imperforate. Again Kellogg 
says that in modern toothed whales "mechanical changes in the re- 
lations of the component parts of the odontocete type of skull appear to 
have restricted at first and finally prevented the physiological function- 
ing of the olfactory apparatus." 

[75 1 

Chapter Five 

J\ioufh and Islost^ 


JiLxCEPT for the teeth where already noted the only aquatic mammals 
whose mouths are of noteworthy interest are Orn'ithorhynchus, the 
Cetacea, Hippopotamus, the walrus, and the Sirenia. 

Ornithorhynchus, or the platypus, has a mouth that is unique among 
Mammalia. Except for its extreme broadness it superfically resembles 
the bill of a duck, being used for the same purpose, but whereas the lat- 
ter is horny, the bill of the former, although firm, is covered with soft, 
moist, and extraordinarily sensitive skin. As the bones of both the man- 
dible and rostrum are bifurcated, some play is permitted and the ani- 
mal, according to Burrell (1927), can pucker the mouth sufficiently to 
form a central suction tube whereby small life of the stream bottoms 
may be acquired with the minimum of mud. Besides the serrated, 
horny part of the lower lips which develops after the early loss of the 
teeth, there is a pair of longitudinal structures in either jaw which 
Burrell designates secateuring ridges. Cheek pouches are also found, 
which the same authority considers are used mainly for holding fine 
gravel as an aid in masticating chitinous or other hard food. 

Never having seen a living sirenian my conclusions regarding their 
feeding actions have been derived from the literature. They are in the 
habit of using the fore limbs for pulling aquatic growth toward them, 
the action then being, according to reports and illustrations, for the 
herbage to be grasped and thrust into the mouth by a unique interaction 
of the lips and a pair of fleshy folds, one upon either side, which may 
be designated as the side lips. There is surely considerable difference 
in this labial equipment between the manati and dugong, but the various 
postures which may be assumed (see figure 8) so utterly and ridicu- 
lously change the expression of the creatures that we can judge little 
from the illustrations, for the posture shown by one animal may or may 
not be one that the other can adopt.- Certainly the labial musculature 
of the sirenians is very remarkable, and of it Murie (1872) wrote: 
"Were the trunk of an elephant cut short at the root, or better still, left 
entire, but contracted to a minimum of its long diameter, and with the 



terminal tactile appendage aborted, structurally the manatee's naso-labial 
organ would assimilate with it." Being derived from proboscidean stock 
(according to belief) it is perhaps not astonishing that sirenians should 
have such a highly developed and unique labial equipment. Neither 
is it surprising to find that in consequence (evidently) of this great 
mobility of the lips this order is incapable of protruding the tongue. 

In sirenians the ingestion of food is aided by bristles upon the inner 
lips, hairs and bristles upon the oral cheek surfaces, and fibrous papillae 
upon both the floor and roof of the oral cavity. These papillae are bet- 
ter developed in the dugong than the manati, and are said to have been 
best defined in the extinct Steller sea cow. In the dugong at least, a 
skinned head of which I examined in the National Museum, the papillae 
constituted rasp-like surfaces that must be extraordinarily efficient in 
tearing up vegetable food. This individual also had a most peculiar 
oral development consisting of a subglobular prominence the size of 
a small hen's egg upon the ventral part of the rostral tip. This was 
smooth, and hard and tough in the preserved specimen, but in life it was 
evidently soft and mobile, functioning as a sort of soft, accessory tongue 
tip. From Murie's figure the tongue proper of the manati is relatively 

The adult male walrus is so constructed that it cannot use the fore 
limbs as an aid to ingestion of food, and unless a morsel is small 
enough to slip readily between the bases of the huge tusks, difficulty is 
encountered. The midsection of the upper lip is relatively immobile, 
and the lower lip seems but slightly muscular and incapable of protru- 
sion. The difficulty is therefore partly overcome by the mystaceal pads. 
These are greatly developed, project considerably, are mobile, and fitted 
with short vibrissae of such large diameter that they are veritable 
spikes. By a medial contraction of these, food may be forced into the 
mouth. There is record of a captive walrus having killed and partly 
eaten a seal, using the bristly pads to tear off flesh — a feat which it 
would have been entirely incapable of accomplishing without their aid. 

The apex of the walrus tongue is rounded and entire, but cleft in all 
other pinnipeds, according to Sonntag (1923) . In comparing pinniped 
with fissiped tongues he stated that in the former the tongue is "shorter, 
wider, thicker. Apex cleft in all except Trichechus (Odobenus). Edges 
lobulated or have laterally-projecting papillae. Mucosa of pharyngeal 
part folded. Many glandular orifices present. Vallate papillae fre- 
quently absent. Lytta absent. Frenum slight. No trace of a spinous 
patch or papillae clavatae. Lateral organs variable." He also stated 



that Phoca has a fuller complement of gustatory organs than any other 
pinniped which he examined. 

The mouth of the hippopotamus is remarkable for its size. It has 
evidently been developed thus partly by the stimulus of excessive use, 
for it feeds upon succulent vegetation of very low nutritive value, and 
hence must consume an enormous quantity. Indicative of this is the 
fact that the stomach is upward of 15 feet in length, and even more 
complicated than is usual in Artiodactyla. But the huge mouth is not 
altogether an efficient apparatus, for mastication is a slow and laborious 
process, and the animal has much difficulty in properly handling the 
pendulous angles of the mouth. 

In the Cetacea the lips are always immobile, smooth and rubbery. In 
Odontoceti the width and size of the mouth is never excessive, but is 
quite variable, largely according to the size of the snout. Thus in Kogia 
the very short mandible makes the mouth relatively very broad, while 
in the long-snouted porpoises it is more beak-like. In the cachalot 
(Physeter) the enormous mass of the head dwarfs the mouth by com- 
parison, and yet the animal feeds upon the most bulky fare of any ceta- 
cean. The tongue of the toothed whales is variable, but may be said 
to be normal in function. It often has a fringe of papillae. Sonntag 
(1922) said that "the tongues of the Cetacea have their glandular or- 
gans better developed, but their gustatory and mobile functions are less, 
than in other Mammalia." The tongues of Mysticeti agree with those 
of Odontoceti in having filliform papillae scanty or absent, the mucosa 
more or less corrugated, no trace of foramen caecum, lytta, frenal la- 
mella, lateral organs, and apical gland of Nuhn. They differ as fol- 

Mysticeti Odontoceti 

tongue soft tongue firm and hard 

intermolar elevation present absent 

much oil in tongue absent 

apex massive not so 

marginal lobes absent present 

lateral borders ill-defined well-marked 

glands less numerous very numerous 

muscles slight or large well-developed 

mobility slight variable 

Salivary papillae and plicae fimbriatae are absent in both. In odonto- 
cetes the surface of the tongue is rather parchment-like. It usually fills 



the intermandibular space, but not always (Monodon). The esophagus 
is not noticeably reduced in diameter in this group, and the muscles for 
deglutition, although strong and compact, are not remarkable. 

In the baleen whales the oral condition is very different. In this group 
the mouth is of enormous size to allow for the unique feeding habits, 
but the details differ considerably. In all forms the lower jaw is very 
much broadened and bowed out, but the actual size of the mouth varies 
so as properly to accommodate the baleen equipment. The mouth is 
apparently larger, relatively speaking, in the humpback (Megaptera) 
than in any of the true balaenopterid whales, and of fairly prodigeous 
size in balaenids, for in the latter group it may be more capacious than 
the thoracic and abdominal cavities combined (Flower and Lydekker, 
1891). The reason for such oral disparity is evidently correlated with 
food habits, but it is difficult to observe precisely the finer points of such 
actions below the surface of the water, even upon the rare occasions 
when one can approach sufficiently close to a feeding whale. It is 
usually claimed that the mysticetes feed by swimming slowly along, the 
rostrum level with the surface and the mandible dependent, the mouth 
being wide open. It is not logical that the mouth should be held in this 
position for long or at speed. Except possibly in the balaenids, if these 
animals maintained the mouth ivide open while feeding upon small 
food, the mandible would fall below the ends of the baleen plates, 
so that all food would pass out of the mouth with the free flow of 
water. Even such an experienced observer as Andrews (1909) is 
noncommital in regard to the precise action of the mouth during feed- 
ing. He has stated, however, that when feeding on shrimp both Megap- 
tera and Balae)wptera turn upon the side, a posture which he has illus- 
trated by photographs, and that when the jaws are closed the water 
rushes out in streams. Regardless of the position of the body the pro- 
cedure of a balaenopterid when shrimp are being consumed is presum- 
ably for the animal to swim slowly through a swarm of the crustaceans 
with the mouth opened for but a foot or two. Then only will the ba- 
leen operate efficiently as a strainer, for the oral cavity will then act as 
a closed filter, the water flowing away laterally through the baleen 
fibers and leaving the shrimp in the mouth. The form of the mouth and 
length of baleen plates in the balaenids (see figure 9) allows the mouth 
to be opened much farther for best efficiency, though no one can say 
why this is the case. 

If balaenopterids cannot find shrimp, they will feed on anchovies, or 
if these are unavailable, then on smelts, or small mackerel. In other 



words, they prefer the smallest food they can get. I am unaware of the 
method they employ in securing fish as large as one-foot mackerel, but 
have been told of how the humback (Megaptera) obtains them. An ex- 
perienced and trustworthy whaler has assured me that off the coast of 
Lower California he has often watched these whales feeding on fish of 
this size. According to him, they "stand on their tails" in the water 
with snout protruding, presumably in a dense school of fish. With 
mouth opened at a right angle the water is fanned with the greatly 
elongated flippers and the fish (according to his theory) mistaking the 
dark cavern of the great mouth for a safe haven, rush thither. During 
this process he assurred me that the mandible was twisted from side to 
side in so outlandish a fashion that it was a wonder the animal could 
ever get it back into place. Then with a heave, the whole head was 
lifted above the surface, and as the jaws were closed, fish and water cas- 
caded from the mouth. I believe this statement, but give it for what it 
is worth. Certainly fish of this size must be secured after some other 
manner than that employed in catching shrimp. The bow-head whale 
is said to feed largely on jelly-fish of several sorts. This larger fare, 
being less crowded in the water, may have had something to do with 
the development of the enormous head, which may reach one-third of 
the total length, or it may be because of some other stimulus. 

In the Mysticeti the tongue is of two distinct sorts. In the balaenid 
whales and the gray whale it is of large size, apparently even higher 
than broad, firm and highly muscular. It is not a fortuitous accom- 
paniment that in these animals the external throat is either entirely 
without furrows or practically so {Rhachianectes has two or three short, 
shallow grooves, and it is of interest to note that among the toothed 
whales this is also the case in the Ziphiidae) . This indicates that the 
gular region is not extensible, and it does not need to be. When, while 
feeding with the mouth partially opened, it is wished to swallow the 
food that has been captured, all water is forced from the mouth by sim- 
ply elevating the tongue, just as we ourselves would do. 

This procedure in whales with grooved throats (at least in Balaenop- 
tera, and probably others) is entirely different. It has been frequently 
stated in the literature that the tongue of these whales is enormous. Of 
course it is, in an animal of this size, but not relatively so. On the 
contrary it is extremely small and so weak that it may really be con- 
sidered as nonfunctional. In an eleven-foot fetus of Balaenoptera the 
tongue was slightly muscular (considerably less so than in the normal 
mammal) — much more so than in the adult. It almost filled the mouth 



and was at least a quarter the size of the adult tongue, although the 
whole head was perhaps but one-hundredth of the bulk. Transection 
of the adult tongue shows but few muscle fibers and these are partly 
disassociated. Connective tissue is soft and flabby with much soft fat 
of oozy consistency, so that when a large piece of the tongue is turned 
over on the ground by means of an iron hook it is strongly reminiscent 
of a large bladder partly filled with mercury. In a freshly killed speci- 
men the tongue lays flabbily at the base of the oral cavity, appears 
shrunken and almost shapeless and, as said, is clearly almost nonfunc- 
tional. So there is ontogenetic reduction of the tongue in this group, 
indicating definite lessening of function. Incidentally it is hardly likely 
that just this change in the character of the tongue could take place 
without a corresponding reduction in the size of the esophagus, and 
hence, in the function of deglutition. 

Now the integument of the throat in these whales has an intricate 
system of longitudinal plicae, grooves, or folds, extending from the chin 
to the mid-ventral line for half its length or more, and from the angle 
of the mouth to the base of the fore limb (see figure 9) ■ Their pres- 
ence has given rise to fanciful theories, as that the grooves are highly 
vascular and so in this manner oxygen is secured from the sea water, 
but anatomical facts clearly illustrate the reason for their being. After 
death the gular musculature is often relaxed, allowing the throat to 
bulge, but occasionally it is more tensed, as in life. Unfortunately I did 
not dissect the fetus which I examined so my acquaintance with the 
gular musculature of this group is limited to what I could learn by hasty 
observation while numerous finbacks were being cut up. It is evident 
however, that neither Schulte (1916) nor Carte and MacAlister (1869) 
properly interpreted all of the conditions which they encountered in 
this group of whales. That Schulte's designation of his mylohyoid as 
a part of the three-sheeted gular musculature is erroneous is shown by 
its innervation as well as by its position in relation to adjacent muscles. 
Furthermore it seems that the muscle which he called a geniohyoglossus 
was in reality the mylohyoid, his hyoglossus the geniohyoid, and his 
sternomandibularis probably the specialized digastric — at least this 
seems, from its position, the most logical interpretation; but lack of 
precise statements as to the innervation of all muscles prevents incon- 
trovertible interpretations. At any rate it seems fairly certain that the 
three main sheets of the gular musculature comprise a specialization 
of at least two sheets of a primitive sphincter colli, now occurring as 
a sphincter colli superficialis, a sphincter colli profundus, and a re- 



markable third sheet — Schulte's longitudinal muscle of the ventral 
pouch — which may be either a superficial and extraordinary division of 
the mylohyoid, or an equally astonishing platyspia. The action of 
these three is undoubtedly assisted by what seems to be the digastric, 
specialized for this purpose and continuing caudad onto the sternum. 
It is interesting to note that the caudal limit of the long gular muscle 
(either mylohyoid or platysma) coincides exactly with the extent of 
the external plicae. 

The feeding actions, then, of those whales which have gular fur- 
rows is apparently as follows. When the mouth is opened for the 
purpose of securing food while the animal is slowly swimming, the 
inrush of water distends the gular region (as noted by Lillie, 1910), 
which is abundantly allowed for by the furrows and elasticity of the 
musculature involved. Then, when the whale wishes to swallow what 
it has captured, the weak tongue lies quiescent but the specialized gular 
musculature, including the digastric, contracts in the powerful way for 
which it is fitted, expelling the water from the mouth and restoring 
the throat and its plicae to their normal, static state of tone. The con- 
ditions occurring clearly indicate that this is fact — not theory. 

Of passing interest here is the presence upon the under side of the 
rostrum, near its tip, in the mysticetes of a pair of shallow pits which, 
in a specimen of Balaenoptera physalus, Lillie (1910) found led into 
two narrow tubes which ended blindly about two inches from their 
mouths. These have been termed naso-vomerine organs and are popu- 
larly believed to be the vestiges of Jacobson's organs, now separated 
from the nostrils by a distance as much as ten feet. 

The esophagus of the Odontoceti may be said to be of normal size, 
but that of the Mysticeti is remarkable for its small diameter. That of 
a finback of 70 feet was not larger than five inches at the most, and 
down its length was a close-ranked procession, in single file, of foot 
long mackeral, all headed to join the host of their fellows in the huge 
stomach. The latter is complex, and even more so in odontocetes. 

The teeth in aquatic mammals as in other sorts have developed to 
conform to the character of the food consumed and the manner of 
obtaining it, and as most aquatic mammals feed chiefly upon fish, the 
usual tendency is for the teeth to become simplified into sharp points 
for the grasping of slippery prey. 

[ 82 ] 


Among less modified sorts the rodent Ichthyomys has the lateral 
edge of the upper incisors prolonged into a sharp point. This un- 
doubtedly facilitates the capture of small fish, but the development may 
have been fortuitous, for the members of the terrestrial subgenus 
Mictomys exhibit the same character. It is caused by the very slender 
lower incisors working against the broader upper ones, leaving the 
lateral edges largely unworn. The upper incisors of the beaver are 
especially modified to act as chisels for cutting down trees — a character 
that really has no bearing on aquatics. 

Two aquatic mammals other than whales lack teeth in the adult 
state. Both the platypus and Steller sea cow (HydromadaUs) shed 
the molars at a very tender age, their places being taken by leathery 
or horny plates — a development accompanying the food habits of the 
former, feeding upon worms and similar soft foods, and of the latter 
in consuming soft marine algae. In the young Hdicore there are four 
pairs of teeth beneath the horny plate of the mandible but these are 
absorbed before maturity. In this genus a pair of the upper incisors 
have developed into small tusks in the male, but these are diminutive 
and non-protruding in the female. In the manati (Trichechus) in- 
cisors are said to be present only during the fetal stage. In this genus 
but in no other sirenian, either living or extinct, there is a progressive 
succession of the molariform teeth comparable to conditions in the ele- 
phant. This takes the form of a constant forward displacement and 
loss of the tooth from the front of the series as a new back one de- 
velops. After studying a considerable series of skulls of different ages 
Thomas and Lydekker (1897) came to the conclusion that during the 
span of a manati's lifetime there must be not less than 20 teeth in 
each series, which is the same as stating that this process of succession 
will continue ceaselessly until death, or at least until senescence. If 
this be the case the teeth can hardly be homologized with those of the 
normal mammalian complement, the whole process of succession being 

Another interesting alteration of the molariform teeth is that en- 
countered in the walrus, which feeds largely upon mollusks, and the 
back teeth have become very broad and flat for crushing these. Some- 
what similar in general respects are the teeth of the sea otter, which 
favors echinoderms or sea urchins, as an article of diet. The molars 
of seals and sea-lions are simpler than those of the average terrestrial 
carnivore, although whether they have always been so or have become 
secondarily simplified from a more complex pattern has not been estab- 



lished, but is a question of some controversy. Their precise pattern 
need not concern us here, but it may be mentioned that the dental arma- 
ture of the sea-Hon appears to be extraordinarily powerful for the use 
to which it is put in coping with fish, large ones being torn asunder 
by a powerful twist of the head rather than by shearing with the teeth 
proper, and small ones swallowed entire. The strong canines may be 
chiefly for the purpose of fighting among themselves. 

Among aquatic mammals there are several exceedingly interesting 
modifications of the front teeth. In the fetal narwhal there is found 
within a socket in either premaxillary a small; nonprojecting tusk. In 
males the right tusk does not ordinarily develop further, but the left 
one grows, projecting straight forward from and through the upper 
lip, and attains a length of eight feet or even more. It is twisted, with 
spiral grooves running in a sinistral direction. Occasionally the right 
tusk also develops and this too shows a sinistral twist. Females do not 
normally have a projecting tusk, though very rarely one develops to 
some extent. No definite use is known for the narwhal tusk. It has 
been claimed that the males sometimes fence with them in play or 
in battle, and the theory has been advanced that they are used for 
digging in the sea bottom, but they grow too long to be handily used 
for such purpose. The proper explanation probably is that the tusk 
at first developed to a moderate length of say, one foot, for the purpose 
of fulfilling a definite need, such as rooting in the mud or as an ice 
pick. But tusks, antlers, and appendages of this sort are exceedingly 
prone to acquire undue evolutional velocity and to develop beyond the 
point of real use, ultimately becoming a hindrance, and hastening ex- 
tinction through overspecialization (as some extinct elephants, the 
saber-toothed tiger, Irish stag, etc.). It therefore seems probable that 
the tusk of the narwhal is too long for any useful purpose and that it 
is more of a handicap than a help. The enormous tusks of the male 
Pacific walrus belong in the same category. Originally the canines 
were undoubtedly used for prying up clams and developed in response 
to this stimulus until they were ideally fitted for this purpose, with a 
length of several inches. But the development was not checked and 
ran wild, resulting in a length and bulk of tusk in the male that must 
surely prove most unwieldly, and if continued, threatens the extinction 
of this pinniped through constriction of the mouth to a degree which 
will hinder the ingress of food. The female walrus is provided with 
tusks also, but these are of but moderate length. 


It will be noted that as with terrestrial mammals with tusks, as well 
as in the case of the majority of mammals bearing antlers or horns, 
these develop to a considerably greater size in the male. The claim is 
usually made that this is for the reason that the males may more effec- 
tively do battle with one another for the females, but this natural 
selection resulting is likely only secondary. Rather does it seem that 
the male sex hormone almost invariably stimulates excessive growth of 
tusks, horns, or other excrescences of a secondary sexual character to a 
far greater degree than does that of the female, just as it frequently 
stimulates growth of body in the male. But this is an uncultivated 
field and I believe there are those who claim that there is no such 
stimulus in the male, but rather an inhibitional factor in the female. 

Of a somewhat different sort is the development of the large teeth 
of the hippopotamus, for these are equally developed in both sexes, 
but overspecialization is indicated here as well. The size of the teeth 
has evidently kept pace with size of jaw and the result is that the 
front teeth are so large and unhandily placed that they can serve but 
little useful purpose and seem detrimental. 

There is much dental variation among the toothed whales. The 
teeth are usually simply conical and attenuate, but the crowns may be 
chisel-shaped (Neomaris, Phocaena) and the tusks of the ziphioid 
whales are at times curiously flattened and twisted. There may be well 
over 200 teeth present (Euihitiodelphis, Stenodelphis), but two, situ- 
ated in the mandible (some ziphioids), or a single large tusk (Mo770- 
don). So far as known, however, nonfunctional back teeth are always 
present in the fetal state at least. The teeth of living odontocetes are 
always of a single or homodont pattern — never heterodont — and this, 
as well as the enormous increase in the number of the teeth of some forms 
over the normal mammalian complement has given rise to endless dis- 
cussion. To account for the latter condition there was first advanced 
the theory of the intercalation of milk teeth into the series, but this 
was abandoned when Kiikenthal discovered indications of tooth suc- 
cession in embryos; and at any rate this theory could not account for 
the great number of teeth occurring in some sorts. Abel has argued 
strongly that the original teeth were split up into numerous simpler 
units ; but this theory is unsupported by any good evidence and is viewed 
none too favorably by many. In fine, we know absolutely nothing about 
the matter. There is not always perfect alternation of upper and lower 
teeth in the closed jaws, and the smaller teeth near the jaw tips are often 
crowded and insecurely attached, indicating that if the jaws become 



further elongated, more teeth may be acquired at the end of the series 
to fill the space provided. But we do not know how or why. 

Tooth buds are present in fetuses of the Mysticeti or whalebone 
whales, but these are absorbed when the baleen starts to develop, or 
even before. Flower (1893) considered that the baleen plates devel- 
oped gradually over their entire present area from the oral ridges of 
the roof of the mouth, which are present to some extent in most mam- 
mals and are highly developed in ungulates. The baleen is analogous 
to the ridges but Flower was mistaken in his premise and they are not 
really homologous. I have examined a fresh eleven foot fetus of 
Balaeuoptera in which the start of the baleen growth was to be seen 
to good advantage. Slightly sunken within a groove which corre- 
sponded in position exactly with a maxillary dental arch was a soft, 
whitish tissue abruptly differentiated from the normal oral epithelium. 
This was about 3 centimeters in width and continuous save for a brief 
interruption anteriorly at the midline of the rostrum. It was plain 
that as this strip grew to form the young baleen there would be an 
accompanying widening in a medial direction so as to cover most of 
the roof of the mouth, which at the stage examined was formed of 
normal mucous membrane. So Tullberg (1883) was correct in prin- 
ciple in stating that baleen develops from a growth of papillae along 
the outer margin of the upper jaw. 

Of extreme interest in the above connection is a condition in Phoco- 
enoides dalli to which Miller (1929) has recently called attention. In 
a preserved section of the upper jaw of the specimen which he had 
the tooth tips appeared as being sunk in small pits and were below 
the gum surface so as to be nonfunctional. In compensation the gum 
along the dental arch and immediately adjoining on either side had 
developed rows of cornified papillae which had obviously taken over 
the grasping function of the decadent teeth. Histologically the struc- 
ture of these papillae certainly seems to be homologous with baleen. 
So we appear to have an illustration in an odontocete of just how 
and why the baleen first began to appear in the mysticete ancestry. 

Wlialebone whales may have nearly 400 plates of baleen on either 
side. Flower and Lydekker (1891) stated that whalebone consists of 
modified papillae of the mucous membrane, with an excessive cornified 
epithelial development, there being at the base and between the blades 
an intermediate substance consisting of a softer epithelium. The latter 
is white and of a cheesey consistency, and projects into the base of each 
blade. When an animal is examined in the flesh the baleen appears 



as a series of rather triangular blades, set close together, the outer or 
labial borders being almost vertical, smooth and firm. The formation 
is of long fibers cemented together. From the ventral tip of each blade 
obliquely upward and inward to near the median line of the roof of the 
mouth the softer cementing substance wears or dissolves away, leaving 
the fibers to form a brush-like inner border to each plate. These diffuse, 
intermingle with those of adjoining plates, and form an effective ap- 
paratus for straining from the water and retaining within the mouth 
the half inch crustaceans which form the favorite item of food. The 
baleen of some whales, as the California gray (Rhachianectes) is short, 
coarse, and commercially valueless. In this whale, as well as probably 
all baleanopterids, the baleen equipment fits into the closed mouth 
without bending. In the balaenid (as the right and bowhead) whales 
the plates are of fine texture and may reach a length of more than 12 
feet (bowhead), the ends folding upon themselves at the bottom of the 
oral cavity, but because of their excessive elasticity, they at once spring 
straight when the mouth is opened and pressure upon the blade tips is 

We may, because of the present condition in Phocoenoides, follow 
the probable course of the development of the baleen with some little 
feeling of assurance. First starting as short epithelial papillae along 
the maxillary dental arch, they were used in place of the disappearing 
teeth to hold and retain small active prey. As they increased in length, 
water could be squirted out between the papillae so as to separate small 
fish and similar food held within the mouth, and the further transition 
of the baleen was but a matter of the ability properly to respond to the 
stimulus provided, and of sufficient time. The degree to which the 
baleen has developed in the large balaenid whales surely constitutes an 
overspecialization at the present time. 

Probably the first modification which an aquatic mammal undergoes 
is the acquisition of the ability to close the nostrils, for any mammal 
would be decidedly handicapped during under-water activity by having 
to guard continually against the sudden entry of water into the open 
nose. As in the case of the ear, there will tend to be a progressive 
adaptation in this respect, consisting first of closure with difficulty and 
then with ease, followed by the time when the closed position is the 
involuntary one and opening is the voluntary. Cetaceans and probably 
the Sirenians belong in the last category, while pinnipeds, the otters, 


Potomogale and the Hippopotamidae have either attained it in some 
degree or are approaching it. 

The procedure of closure is accompHshed in a variety of ways, and 
probably no two sorts of different mammals have precisely the same 
mechanism for attaining this end. Unfortunately, observation of the 
live animal is not always illuminating in this respect, and the inter- 
action of the small nasal muscles is so nice that dissection does not al- 
ways show to our entire satisfaction the exact method followed. Broadly 
it may be stated that closure of the mammalian nose is effected by ac- 
tions of the Mm. naso-labialis and maxillo-naso-labialis (of Huber), 
complicated by sundry specializations of these and the help of pads, 
flaps or valves. In the pinnipeds, for instance, the naso-labialis arises 
from near the middorsal line above the orbit, and, diverging slightly 
fanwise, inserts into the mystaceal pad. Its contraction lifts the pad 
dorso-caudad and crowds it mediad. The maxillo-naso-labialis arises 
from the zygomatic process of the maxilla below the infraorbital fora- 
men, and inserts into the mystaceal pad deep to the naso-labialis. Its 
contraction pulls the pad latero-caudad and opens the nostril. Hence, 
mass of pad and contraction of the naso-labialis (or its normal tone 
when the specialization is greater) closes, and contraction of the maxillo- 
naso-labialis opens — really a very simple arrangement. 

In most insectivores and rodents practically nothing can be told about 
the mechanism for nasal closure — at least without a long period of very 
painstaking work. They are too small to watch properly in life and 
similar difficulties are met in their dissection; and many others are not 
available, either dead or alive. Closure mechanism may, however, be 
divided into three classes. In one it is effected by a flap or valve; in 
the second, by the crowding of a fibro-muscular pad ; and in the third, 
by fibrous or muscular tension from two or more sides. But action may 
partly combine two of these modes. 

I am unacquainted with the precise method of narial closure em- 
ployed by the platypus, but the external apertures remain open, so there 
must be a deeper valvular arrangement. 

There is a tendency in many aquatic mammals {Potomogale, fissipeds 
and pinnipeds) for the broadening of the muzzle. This is popularly 
supposed to be a specialization in the direction of perfection of aquatic 
bodily form, but it is doubtful whether there is any real logic in this. 
The reason for its acquirement may be partly tactile, because of an in- 
crease in the sensitivity of the vibrissae, and hence in the branches of 
the infraorbital nerve that extend to their bases. But the increase in 



size of these pads is chiefly for the purpose of maintaining the nares 
in a closed position. The large pads of Potomogale, otters and phocids, 
by their bulk and elasticity crowd the narial apertures so that in their 
fullest development little or no muscular effort is necessary for closure. 
If effort be necessary, then it can be supplied by the radiating fibers 
of the pad itself and by tension of the naso-labialis muscle. In the 
fissipeds and pinnipeds the pad pressure is largely in a medial direc- 

FiGURE 12. Pharyngeal region of the pigmy sperm whale Kogia. from above 
(redrawn from Kernan and Schulte) ; {a) tongue; {b) nasopharynx; 
(r) oropharynx; {d) larynx; {e) arco palato-pharyngeus ; and (/) esopha- 

tion, while in Potomogale it would seem to be chiefly anteriorly. The 
reason for the latter condition is that in this insectivore otter there is 
a bilobed rhinarial shield, smooth and evidently firm in life. The nasal 
passages are situated posterior to this while the openings are to the 
sides of the shield. Hence, pulling the very broad mystaceal pads to 
the rear opens the nostrils, and relaxation presses them firmly against 
the shield. 



I have for hours watched the breathing actions of various pinnipeds. 
By far the greater part of the time phocids inhale rapidly and then 
close the nostrils, although occasionally they are kept open for the dura- 
tion of several breaths, especially when the animal comes to the surface 
after a period of active swimming. And a definite impression is given 
that the closed position is the relaxed one. In the sea-lions the mecha- 
nism for involuntary closure seems to be less perfected, but the mystaceal 
pads are not so broad, the animal is more prone to maintain the nostrils 
open between breaths (and, incidentally, to breathe through its mouth), 
and one is unable to decide from observation whether the opened or 
closed position is the involuntary one. Similarly with the otter. Al- 
though the muzzle of the latter appears very broad this is partly at- 
tributable to the rostral breadth of the skull and the pads are really 
not as large as they seem. 

I have been unable to tell by observation of both American and 
Asiatic tapirs just how the nasal passages are closed. Apparently there 
is a contraction of the proboscis and consequent crowding of parts of 
the passages deep to the apertures. In this group of mammals there 
are diverticula of the nasal passages comparable to those existing in the 
Equidae, but no use is known for them. 

In the Hippopotamidae the nostrils take the form of a pair of slits, 
closure seems to be partly voluntary, and opening accomplished by 
muscular pull along both borders of the slits. 

I have been unable to acquaint myself with the exact processes of 
narial action in the Sirenia. The apertures are each somewhat crescentric 
and closure is evidently chiefly involuntary and largely of the valvular 
type, like a modification of the condition in the Odontoceti. The nostrils 
are located at the angle of the muzzle and are evidently farther back in 
the dugong than the manati. Their present position may be perfectly 
ideal for the needs of these sluggish creatures and it is by no means 
certain that there is any stimulus whatever for further migration of the 
nostrils toward the rear. The extensive rostral basining in this order 
indicates a high development of the naso-labial musculature, but from 
Murie's (1872) descriptions I judge that this concerns chiefly the lip 
movements. The muscular complexity which exists in this region 
may well be an inheritance from proboscidean ancestry more properly 
than a relatively recent specialization in response to generic needs. 

For a proper understanding of the narial equipment in the Cetacea it 
will be necessary to consider the lungs and progress therefrom anterior- 
ward. Wislocki (1929) has announced that in the porpoise Tursiops 



the bronchioles with a diameter of less than 0.5 mm. are provided with 
numerous muscular sphincter valves and that the terminal air sacs are 
guarded by the same means, as discussed more fully in the last chapter. 
Heretofore investigators have been obliged to consider the possible 
means whereby the nasal, or laryngeal, equipment of the Cetacea was 
enabled to prevent the escape of air from the lungs under the enormous 
pressure to which the animal is subjected during deep dives. Wislocki's 
discovery puts an entirely different aspect on the matter, however. 
Although the individual sphincters of the bronchioles may be assumed 
to be relatively feeble in action, the amount of air which each imprisons 
is so minute, and the valves occur in such prodigious numbers, that 
this equipment alone seems entirely adequate to prevent the escape of 
air even into the esophagus against the wishes of the animal. Pre- 
sumably these valves close at the end of inspiration and open at the 
initiation of expiration, and it is reasonable to assume that their presence 
is characteristic of all Cetacea. The simultaneous relaxation of these 
small sphincter muscles at the same time that the nostrils are opened 
would account for the way in which the air, imprisoned in the lungs 
under the pressure applied by the relaxed thorax plus the pressure of 
the surrounding water, rushes forth with a veritable pop. But there 
is no reason for believing that bronchial sphincter valves must remain 
closed between breaths, and the animals may at times inflate a part of 
the nasopharynx before the nose is actually opened, as I have been led 
to believe by watching the escape just preceding expiration of small 
bubbles of air around the external slit of the blowhole. 

In the Odontoceti the larynx is prolonged in a tube-like manner, 
projecting the epiglottis into the nasopharynx, and this may be closely, 
clasped by the soft palatopharyngeus muscle. This is quite remarkably 
developed in the Physeteridae and Kogidae especially, in which the naso- 
pharynx branches from the oropharynx upon the side and the larynx 
is correspondingly situated. Lillie (1910) stated that the latter was 
situated upon the left side in two specimens of cachalot. Raven (MS) 
makes the same statement, as well as Kernan and Schulte (1918) in 
the case of Kogia; and the dissimilarity in size of the choanae is un- 
doubtedly the precise reason for the laryngeal asymmetry. 

Kernan and Schulte have shown the pharyngeal region of Kogia 
(fig. 12) to good advantage. They said that the conditions are such 
as to prevent the entrance of water from the mouth into the nasal 
passage, and this may be a superficial function also, but the construc- 
tion of the arc of the palatopharyngeus is clearly such that the greater 



the air pressure from the lungs the more securely it should clasp the 
larynx. Thus in these sperm whales at least, the air under pressure 
may be forced from the lungs through the larynx and into the naso- 
pharynx, thus pressing backward against the palatopharyngeus, closing 
it the more tightly the greater the pressure, its suggested primary func- 
tion then being to prevent the escape of air from the nasopharynx into 
the mouth. Some such procedure may be necessary in these animals 
in connection with the remarkable specialization of the right nasal 

H. C. Raven's preparations of Monodon, however, suggest that in this 
animal conditions are different, for the palatopharyngeus is very heavy 
and different from what Kernan and Schulte showed for Kogia, which 
may be due to the fundamentally different narial equipment of these 
two genera. It has usually been assumed that odontocetes at least may 
breath while at the surface at the same time that they swallow, and that 
when submerged they could not expel air through the mouth even if 
they so wished. But in the Monodon and Physeter which H. C. Raven 
(MS) dissected the larynx had been withdrawn and the epiglottis was in 
the oropharynx. Accordingly this authority is of the opinion that the 
larynx is thrust upward only temporarily for breathing or other pur- 
pose. If this actually be the case then it is certain that the small sphincter 
valves of the lungs act alone in preventing the escape of air, for the 
formation of the epiglottis is such that it could offer but little hindrance 
to the involuntary escape of air from the trachea into the mouth. 

It may be mentioned that an elongated larynx for thrusting the epi- 
glottis into the nasopharynx is not in itself an aquatic adaptation, for it 
is a character that is developed to a considerable extent in some mar- 
supials and ungulates. The odontocete ancestry may have had this pro- 
vision in moderate form to begin with, or else have evolved it completely 
for their particular needs. 

The Mysticeti do not have this tubular type of larynx. Benham 
(1901) and others have described it but from these and existing sketches 
I am unable to envision the conditions with satisfactory clarity and I 
neglected to investigate the matter myself when I had the opportunity. 
Apparently the arytenoid body closes the epiglottis from above, the latter 
is triangular rather than tubular, and of very moderate length. 

In a consideration of the precerebral portion of the nasal apparatus 
of the Cetacea it must be remembered that in most mammals the muscles 
of the face proper have a variety of functions to perform, such as in- 
volved twitching of the nose, snarling of the lips, and a host of others 



■.',. ^:»^-.v 


Figure 14. Mounted skeletons of a seal (Phoc/dae) above and fur sea 
{OtiiriicJcie) below, in the U. S. National Museum. 



having to do with facial expression, while in the whale all these muscles 
are subservient to the single purpose of opening the blowhole. Such 
as cannot readily lend themselves to this use have become virtually non- 
functional. It is therefore of small wonder that we find the nasal 
mechanism as perfected as it now is in the odontocetes. Furthermore, 
in view of the bronchial conditions as reported by Wislocki it is neces- 
sary that we free our minds of the assumption that a special provision 
for the purpose of retaining air under great pressure is situated in the 
blowhole. Nor is a complicated contrivance for excluding water essential. 
The external area of the closed blowhole is small and it has been a simple 
process for this to acquire such a form that the greater the pressure the 
tighter is automatic closure. 

Precranial narial conditions in the Cetacea are really of three sorts, the 
most simple being in the Mysticeti, a more involved situation in most 
Odontoceti, and a remarkably complicated one in the Physeteridae and 
Kogidae. In mysticetes the nostrils are in the form of two slits running 
in a sagittal direction (see fig. 13) , but converging to form a rather steep- 
sided V with angle pointing forward. They do not actually join, how- 
ever. Along the margin of each slit is a raised area, the elevation being 
slight along the medial margin and considerably higher along the lateral. 
To the touch this region is, like the rest of the body, tough and elastic. 
I have run my arm to above the elbow down one nostril and found that 
there was decided, though not heavy, pressure upon my arm for per- 
haps one foot below the surface, and as my arm was withdrawn, the 
elasticity of the tissue closed the passage completely. This involuntary 
closure may likely be assisted, does the need arise, by voluntary tension 
at both ends of the nasal slit, and by a downward contraction from the 
surface. But the conditions are such that the greater the surface pressure 
the more securely do the nostrils close, and I regard it as unlikely that 
the need ever arises for voluntary closing action in order to exclude sea 
water. In mysticetes there are no widely divergent diverticula. There 
is well known to be a "subcircular diverticulum from the dorsal wall 
of the respiratory passage," as stated by Schulte (1916), marking the 
olfactory region, and a "spritzsac" along the anterior wall, but from 
published descriptions I am unable exactly to envision the conditions. 
By manual investigation within the passages of a fresh adult I could 
not discover any true diverticula but only a slight folding and wrinkling 
of the mucosa rostrad and to a lesser extent laterad. Schulte considered 
the arrangement of this to be such as to aid closure when pressure is 
supplied from without. 



The interaction of the speciahzed facial musculature in opening the 
nostrils of mysticetes has not yet been investigated with sufficient exacti- 
tude for us to be sure of all the actions involved. Of course the facial 
musculature is very highly modified, as it must be to control nostrils 
situated upon the top of the head, but the intricacy is not nearly so great 
as in that of the odontocetes. There are sufficient good photographs 
(see fig. 13) of breathing mysticetes to gain an understanding of the 
external results when the opening mechanism of the nostrils is in opera- 
tion. In these there is apparent a marked lateral dilation of the outer lips 
of the nares, and a surprisingly high elevation of the anterior margins — 
an arrangement which evidently operates as an efficient barrier to the 
entry of water within the respiratory tract while the animal is in loco- 
motion. There must be definite mechanical provision for this elevation, 
either in excessive elasticity of the tissue deep to the anterior margins 
or in wrinkling, and the latter is exactly what we find in the spritzsack, 
to which theory Schulte also subscribed. 

The latter authority, in his study of Balaenoptera borealis, indicated 
uncertainty regarding just how the elevation of the anterior narial lip 
was instigated. From what I could learn from adult finbacks there is 
no great difficulty in this. The closed nares are suggestive of a V with 
apex directed forward, and the open nares are each broadly oval, with 
the long axes almost parallel. It is clearly evident that the pull of 
superficial muscles converging to the apertures from rostrad and laterad 
open the nares. The elevation could easily be provided by superficial 
muscles pulling rostrad upon the anterior and lateral margins while at 
the same time antagonistic action is supplied by deeper muscles pulling 
largely caudad, and relatively nonyielding tissue anterior to the narial 
passages. This seems to be the general principle of the mechanism, 
although in reality the actions are likely very intricate. 

The narial conditions in the Odontoceti are very difi^erent. In all of 
them the external aperture is single, in all but the sperm whales medial, 
and almost always it is crescentric in shape, the concave aspect being di- 
rected anteriorly. The only known exception to the last detail is in 
Platanista, in which the orifice is said to be in the form of a longitudinal 
slit, but in exactly what manner this is opened is unknown. Excepting 
the latter for the time being, narial development of the toothed whales 
is of two definite sorts — that represented by the Kogidae and Physeteri- 
dae or sperm whales, and all other odontocetes. The latter, as the least 
intricate, will be discussed first. In perhaps no two distinct sorts is the 
facial musculature arranged exactly the same, for as the genera diverged 



from a common ancestry, more pronounced differences developed each 
in its own particular way as specialization increased. It is thus easily 
seen that the narial musculature of a beaked porpoise with rather low 
forehead, such as Tursiops, should be considerably different from that 
of a nonbeaked form with large frontal bulge, such as Globiocephala 
(see fig. 10). Of the first type I have examined only Tursiops, and of 
the second, Monodon and Neomeris, and the last was too hardened to 
be satisfactory. In this only did I attempt to dissect the facial muscu- 
lature, for in the others this was done by Ernst Huber. I will there- 
fore not go into precise details but will discuss only the generalities of 
this feature. 

In the majority of odontocetes, then, the external nares take the form 
of a single crescentric orifice. Within this, at a distance varying with the 
genus or species, the passage is divided by a membranous partition into 
two. In this, as well as in mysticetes, the supracranial part of the nasal 
passages is practically vertical or slopes gently forward as it approaches 
the surface. But between the skull and the orifice in odontocetes there 
is a system of diverging diverticula or membranous folds (see descrip- 
tion of Neomeris, Howell, 1927). These apparently vary with the spe- 
cies, or even with the individual, but as I understand their underlying 
principle they had better be described as two different parts of the nos- 
trils. Lying forward of the bony nares and in contact with the pre- 
maxillae in a slight hollow of these bones, the extent of which can be 
determined upon the skull, is a diverticulum of each nostril which may be 
designated the premaxillary diverticula. These and the supracranial part 
of the nostrils proper describe an angle of, say, 90 degrees or less. 
Within this angle the tissue is very soft because of the soft fat of 
which it is largely composed. At the apex of the angle this soft tissue 
fits over the bony narial openings, effectively sealing them, especially 
when there is any pressure from without. When the animal wishes to 
take a breath the soft tissue, acting as a plug, is drawn forward by mus- 
cular action in a way that is greatly facilitated by the premaxillary di- 
verticula. This action is strongly suggested by conditions in an adult 
narwhal (Monodon) the entire head of which H. C. Raven had sawn 
lengthwise in thin sections. When examining figure 15, however, it 
should be clearly understood that the sketch is only roughly diagrammatic 
and the degree of expansion of the deeper part of the passages and 
diverticula is unknown. Neither is it meant to imply by the above state- 
ments that these deep parts are always kept closed between breaths. Per- 
haps the complete sealing of the entire mechanism is indulged in only 
at some depth. 



The remainder of the nasal diverticula is more puzzling. They occur 
somewhat nearer the surface and may be either forward, backward or to 
the side of the main nasal passage. Histologically the mucosa of these 
diverticula consists of simple, stratified epithelium without speciali- 
zation. To them has at times been assigned the function of a hydro- 
static organ, which theory I consider to be untenable. They may orig- 
inally have been a phylogenetic relic, like the nasal diverticula of the 
Equidae are supposed to be, since become somewhat more complex but 
still without useful function; they might serve in some slight measure 
to divert water that might slop over into the open blowhole, which 
would hardly be of sufficient importance to account for their existence. 

Figure 15. Diagramatic representation of the probable blow-hole action, closed 
and open, of the porpoise Tursiops: (a) external orifice; (b) supracranial 
part of nostril; (c) premaxiilary diverticulum; (d) bony nares; arrow in- 
dicates direction of movement in opening, and broken line the outline of 
the skull. 

Winge (1921) considered that "the air, which is exposed to strongly 
varying pressure and temperature, has a tendency to provide itself with 
greater space by widening out the nasal passage and Eustachian tube 
wherever it meets with least resistance." For all we know this may 
be the case and no one can prove the contrary; but I believe there has 
been a more primary stimulus for the initial development, or at least 
retention from a primitive ancestor, of the system of nasal diverticula 
of the odontocetes. Let us liken the narial passage to the soft, elastic, 
rubber tube which it so greatly resembles. The intricate and laminated 
muscle layers which converge to operate its opening cannot function 
as dilators with precisely equal force at each and every point, however, 



but their pull must be very unequal. Some must pull more strongly than 
others, or act obliquely, and there would result a bulging stimulus at 
some points which, throughout the ages, might well have resulted in 
the formation of the diverticula as we now encounter them. It hence 
seems likely that they are now a part of the opening mechanism of 
the nose. 

The external aspect of the blowhole of, say, Tursiops, is crescentric 
and somewhat greater than the half of a circle, the concavity facing 
forward. Its outer margin is firm, elastic and of rubbery texture. 
Within this curve is a sort of valve that is much softer to the touch, 
intrinsically muscular and mobile, which is really a part of the deeper, 
soft, valvular plug, so that in the living animal while the blowhole 
is closed one or another part of its surface may move or "work" slightly. 
The "hinge" of this valve stretches transversely between the points 
of the arc, and when expiration is desired the posterior part is merely 
withdrawn within the spiracle and sinks into the highly elastic tissue 
anterior to the nostril. 

As far as concerns surface indications of muscular action during respir- 
ation, all we can tell is that the nasal valve is opened by contraction 
of certain of the rostral muscules, and that in closure there is tension 
from laterad of the blowhole, but whether the latter is voluntary or in- 
voluntary is not clear. Of course there is intricate action and interaction 
of the deeper nasal musculature, but discussion of this I shall leave in 
the capable hands of Ernst Huber. 

There still remains for discussion the astonishing nasal passages of 
the Kogidae and Physeteridae or sperm whales, reported on by Pouchet 
and Beauregard (1885), Le Danois (1910), and Kernan and Schulte 
(19I8) . The conditions are so excessively involved, however, that writ- 
ten descriptions can hardly be clear without extensive diagrams, which 
so far have not been published. For a proper understanding of these I 
am indebted to H. C. Raven, who has most generously placed at my 
disposal a description and diagrams of the head of a young cachalot 
which he recently dissected. He will himself report upon this in detail 
so it is wished here to describe briefly only such points as are necessary 
for a general discussion. In the cachalot, then, the left nasal aperture of 
the skull is huge and the right relatively minute. From the skull the 
lumen of the left passage extends for many feet along the left side of 
the spermaceti organ, opening upon the left side of the antero-superior 
part of the snout in a somewhat S-shaped orifice. The right nostril 
leaves its small cranial orifice and expands laterally to extraordinary 




proportions. It is situated fairly between the spermaceti organ and the 
adipose cushion below. It has a large diverticulum between the organ 
and the skull and another anterior of the organ, and joins the left nos- 
tril just within its orifice. The mucosa of the left nostril is plicated 
but that of the right is smooth in Physeter. In Kogm, whose spermaceti 
organ and nasal passages are relatively very much shorter indeed, the 
mucosa of the right passage is thrown into complex folds, evidently in 
compensation for its shortness. These folds, according to Kernan and 
Schulte, are highly vascular. 

It seems almost certain that originally the narial conditions of the 
sperm whales were very similar to those found in other odontocetes. 
At least we are justified in assuming so, for the conformation of the 
skull indicates that there was retrogression of the external nares to the 
top of the head in the sperms as in other toothed whales. What I be- 
lieve to have since taken place in the case of the latter is a secondary 
displacement forward of the blowhole proper and its controlling mus- 
culature by crowding of the spermaceti organ, although H. C. Raven 
is of the opinion that the organ began to develop in front of the nos- 
trils and intruded between them as it increased in size. Ordinarily ex- 
piration and inspiration in this group must be accomplished solely by 
means of the left nostril. At least it seems certain that because of 
the disparity in size between the right and left narial apertures of the 
skull an insignificant amount of air could pass through the former during 
the short time occupied by the act of breathing. But the enormous dila- 
tion possible in the case of the right passage can hardly be fortuitous and 
must have some function. If it be dilated with air, as seems certain, then 
this should occur chiefly between breaths. It might hold an accessory 
air supply, but hardly enough to account for the long submergence of 
these whales, which are purported to remain below in excess of an hour. 
It is not impossible that the right nostril may be emptied of air to facili- 
tate deep diving, and filled from the lungs so as the more easily to bring 
the animal straight up from the depths, as suggested by H. C. Raven 
(MS) . And one must not overlook the possibility that when beneath 
the surface for a considerable period a part of the air in the lungs might 
be forced into the right passage and again withdrawn for some particular 
physiological purpose which might or might not have something to do 
with the spermaceti organ. 

No discussion of the position of the external nares in aquatic mammals 
has yet been made in the present contribution. The position will cor- 
respond to the sort of stimulus experienced, the length of time that it has 



been in operation, and the capabilities of an animal for responding to it. 
Thus the whales are relatively speedy, compared to swimming land 
mammals ; most of them cover much territory ; and we can well see that 
they would be greatly handicapped if they could not renew their air 
supply at full speed in a choppy sea, but had to slow down and twist 
their nostrils above the surface at every breath. Hence the nostrils would 
be expected eventually to occupy the position upon the top of the head 
that they now do. Sea-lions and otters might be expected ultimately to 
follow the same development. On the contrary an animal like a sirenian 
or hippopotamus which does not seek active prey but only pokes the nose 
cautiously above the surface, usually while stationary and in a sheltered 
lagoon or river, might never attain a like nasal development. 

Nothing much in this respect can be told about the platypus. The 
nostrils are dorsally located upon the "bill" but we have no means of 
judging the condition in its terrestrial ancestor. No migration of the 
nostrils is apparent in any of the insectivores, and in few rodents. Thus 
the nostrils of the coypu, capybara, beaver and muskrat are considered to 
be located slightly more dorsal than usual, but the alteration is really very 
slight. In the hippopotamus the nostrals are definitely upon the dorsal 
side of the snout, so that when the remainder of the body is submerged 
the animal may breathe with only the nostrils, eyes and ears exposed; 
or when thoroughly frightened, the nostrils alone may be quietly thrust 
forth. If the latter were situated farther back on the head this could not 
be accomplished without showing the eyes also, and this may militate 
against any farther migration caudad of the nares. The same may be 
said of the Sirenia. The nostrils might conceivably fuse to form a 
single blowhole and this might protrude somewhat so that breathing 
could be accomplished with the very minimum of exposure, but unless 
there was a radical change in habits there could be little stimulus for a 
migration of the nostrils to the top of the head. At present they are 
not even as dorsally situated as in the hippopotamus, for the animal 
evidently does not make much use of its eyes for peering above the 
surface, and all it need do is extrude the muzzle, with the axis of the 
head at an angle (not parallel) with the surface. 

In a different class are otters, pinnipeds and whales. These either 
pursue active prey, travel at speed from place to place, or both: The 
otter is not as yet sufficiently modified for more than a slight change in 
position of the nostrils to be apparent. And unless there is a great in- 
crease in size throughout future geologic time the stimulus for posterior 
migration of the nares will doubtless be very slight. This animal can 



hold its nose above the surface with sHght effort because of the small 
size of its head. A shght tilt in the cranial axis is all that is necessary. 
If it had a blowhole upon the forehead it would be but two or three 
inches from the present position of the nostrils and I judge that the 
difficulty it might experience in keeping water from slopping into the 
nose would be even greater than at present. Therefore small size should 
be a retardant of nasal migration. On the other hand, a long-snouted, 
large-headed beast with a body of ten or twelve feet would encounter a 
different set of conditions. The snout would travel through a larger arc 
for breathing, nostrils upon the forehead would be of more definite 
advantage than in the otter for the reason that they could then be thrust 
higher out of water with but slight exertion. The critical factors here 
would be size of head and speed; but the forces at play are intricate. 
Thus, in a mammal propelling itself from the rear the head (if the neck 
be shortened as in the whales) is a critical part of the swimming ap- 
paratus. If the head be large it cannot be swung about to thrust the 
snout above the water without acting as an undesired rudder. Further, 
if the neck be shortened to facilitate speed of locomotion the effort 
to elevate the snout would be disproportionately great. To me these 
facts point to the probability that when the whales were undergoing 
the most marked migration of their nostrils they were of a size perhaps 
comparable to a sea elephant or larger, with relatively large heads and 
necks that had already become markedly shortened. The difficulty both 
muscular and mechanical, of raising the snout while swimming at speed 
obliged them to elevate and thus retract the nostrils to their full ability, 
and this stimulus was largely critical in the migration of the latter to their 
present position. 

The nostrils of the Phocidae, or those which I have observed, are 
slightly elevated, and judging from other aquatic specializations of this 
group they might be expected to show more marked modification. Ac- 
cording to my experience the true seals are not very prone to breathe 
while swimming at speed, but are more often in the habit of doing so 
at rest, pausing after swimming beneath the surface quietly to renew 
their air supply before starting off afresh. If this be really characteristic 
and is persisted in they may never undergo more extensive migration of 
the nares. 

The sea-lions (Otariidae) with which I am familiar habitually breathe 
while swimming rapidly at the surface, but the alteration in the position 
of the nares is even less marked than in the seals. For one thing there 
seems to be a different set of stimuli here. Swimming with the anterior 



limbs, the head and anterior part of the neck do not constitute such a 
critical part of the natatory equipment. Hence the head may be thrust 
moderately in any direction without disturbing equihbrium providing 
the hinder end acts in compensation. Furthermore the mouth is used 
to a considerable extent for breathing, and we cannot be sure that 
there is any really marked stimulus experienced for the retraction of 
the nares. 


Because of a complex of reasons, chief among which are doubtless 
the size and breadth of the head, and feeding habits, the speedy balaenop- 
terid whales carry the flattish dorsum of the head parallel with the body 
axis, so that there is the minimum of cephalic resistance offered to the 
water. The slower Rhachianectes and balaenid whales have a down- 
ward-curving rostrum, but the base of the rostrum and posterior part 
of the head are carried parallel with the body axis also. In these there 
is no adipose cushion in the frontal region, unless the "bonnet" of 
Efibalaena, consisting of a raised, warty area upon the anterior rostrum, 
could be considered as the beginning of such a structure. 

All odontocetes, without exception I believe, have some sort of recog- 
nizable adipose cushion equipment upon the forehead. In these whales 
the outline of the head from snout to vertex is never a straight line 
that can be carried parallel to the body axis so that the only resistance 
to the water is offered by the tip of the snout. The latter is always 
carried slightly depressed. In the beaked porpoises (as Tursiops and 
Delphinus, fig. 10), the dorsal outline is concave where the rostrum 
meets the cerebral part of the head, while in the short-snouted sorts the 
forehead is protuberant, but in all, the frontal region is anteriorly pre- 
sented so that it receives the impact of the water during swimming. 
This, being the front of the braincase, is a region of rather critical 
delicacy, and it might be expected to respond to this stimulus by building 
up some sort of protective buffer or shock-absorber: but the subject 
cannot be abandoned with any such simple statement. 

The adipose cushion in the majority of odontocetes is a thick pad of 
soft, elastic tissue upon the forehead in front of, and even partly sur- 
rounding, the blowhole. It need not be sharply defined, but the fibrous 
tissue and soft fat of which it is composed is surrounded and to some 
degree penetrated by the facial muscles which converge to or toward the 
blowhole. Directly anterior to the blowhole it is composed of pure, 
soft fat that oozes oil when transected. This fat is not the same as 



that of the blubber layer but, together with the fatty tissue within the 
angle of the jaws, is of different consistency, and hence specialized, the 
refined oil from these regions being the finest lubricant for precision 
instruments so far known, and correspondingly valuable. The develop- 
ment of this frontal fat body is variable in different sorts of toothed 
whales. In long-snouted porpoises such as Tuvsiops and Delphinus its 
development is but moderate, while in the adult Monodon it is large, 
and is relatively so great in Globiocephala (fig. 10) as to cause the 
forehead to bulge forward beyond the jaws. 

In the latter case the frontal bulk must act as a definite retardant 
to speedy locomotion. If the adipose cushion, or "melon" as it is 
called in the trade, had the function simply of a shock-absorber it would 
hardly develop to such proportions, nor would it have its present con- 
sistency, but should be more gristley and perhaps, as whales are prone 
to fatness, have some deposit among the fibers of the same sort of fat 
as constitutes the blubber layer; but hardly a particular grade of fine 
fat of the same composition as that found within the angle of the jaws. 

In explanation of the above state of affairs it seems to be most likely 
that originally the water pressure against the anterior braincase of the 
toothed whales stimulated the formation of a simple fibrous thickening in 
this area to act as a buffer. This development once having begun, the 
region should then have been in a state plastic for further adaptations, 
and I believe that its present condition indicates a specialized physio- 
logical function of the adipose cushion to which that of a shock-absorber 
is now of secondary importance. This is but a personal opinion, how- 
ever, and entirely unproven. 

Sharply marked off from the other toothed whales in details of the 
facial regions are the sperm whales (Kogidae and Physeteridae) . In 
the cachalot the fatty tissue of the head is of two sharply differentiated 
kinds. Occupying the entire bottom of the facial basining and project- 
ing beyond the bony snout is a prodigious mass of fibrous tissue so 
tough that it must be hewn with an axe. This holds fat as a sponge 
holds water, and from it may be secured about 10 or 12 barrels of oil. 
It is the adipose cushion, or "junk" of whalers and seems largely homolo- 
gous with the adipose cushion of other toothed whales. The sperma- 
ceti organ proper is a huge ovoid body occupying the upper half of the 
rostral basin and separated from the junk by the expanded right narial 
passage. After surrounding tissue has been removed it is said to be 
composed of an envelope of extremely strong, fibrous tissue of tendinous 
aspect. Within this is a zone of partly solid oil held by a spongy net- 



work, and in the center is liquid spermaceti oil. This spermaceti is 
very different from the blubber oil, just as is the case with respect to the 
adipose cushion of other odontocetes. The spermaceti organ is said to 
be a closed, ductless system, but it must be supplied by blood vessels. 
In a diagram with which H. C. Raven has kindly furnished me it is 
shown as resting for its entire length upon the right nasal passage, so 
phenomenally expanded in lateral direction. Diverticula of this passage 
are situated directly in front of and behind it, so when the passage is ex- 
panded with air, the spermaceti organ would seem to be suspended upon 
a pneumatic cushion upon three of its sides. This act is not as difficult 
as might be supposed ; for it is not improbable that the organ is lighter 
than the water which it displaces. Occasionally, because of poor con- 
dition of the animal or for some unknown reason, the spermaceti case 
is empty, but ordinarily it contains up to 15 barrels of oil. Shortly be- 
fore my visit in 1926 to the whaling station at Trinidad, California, a 
large sperm whale was captured which yielded 27 barrels of oil, pre- 
sumably from the case and the adipose cushion combined. The size 
of the spermaceti organ and junk should be emphasized. In a large 
male the head constitutes two-fifths of the total length, so the cephalic 
fatty equipment must be at least one-third the length of the entire animal, 
or say 20 feet. It may be noted in this connection that in drawings of 
this whale the end of the snout is usually represented as truncated and 
ending even with the tip of the lower jaws, whereas all photographs 
show that the snout is bluntly and evenly rounded, and that in large 
males it projects for several feet beyond the lower jaws. Conditions are 
evidently relatively the same in the pygmy sperm whale (Kogia) except 
that the perfectly formed spermaceti organ is very small, and the Ziphii- 
dae — notably Hyperoodori — are said to have some sort of fatty frontal 
organ, but precise descriptions are lacking. 

We know nothing about the manner in which this great cephalic 
oil equipment began to develop except by inference. In other odonto- 
cetes the adipose cushion is situated in front of the blowhole, and the 
junk of the sperm whale is probably homologous with this. The present 
conformation of the nasal passages in the cachalot is to me strongly sug- 
gestive of the theory that the spermaceti organ is not homologous with 
any part of the adipose cushion of other toothed whales, but is a distinct 
development, originally having had its inception back of the blowhole, 
which latter was forced farther and farther forward as the organ increased 
in size. This belief, however, is not shared by H. C. Raven (MS) who 
is of the opinion that the spermaceti organ is the part that is homologous 



with the adipose cushion of other odontocetes, and that as it developed 
in the more primitive Physeteridae it forced itself rearward hehveen the 
nasal passages. 

Be that as it may, the markedly posterior position of the bony nares 
in the cachalot is suggestive, to me at least, that the blowhole of early 
physeterids had receded almost as far to the rear as we now find it in 
other odontocetes before the spermaceti organ had attained even mod- 
erate size. In other words, in the sperm whale ancestors there was likely 
a primary recession of the nares as in other odontocetes, but the sperma- 
ceti organ forced a secondary migration of the spiracle in a forward di- 
rection. But no fossil sufficently primitive to settle this point is yet 

I think it is perfectly clear that this organ, developed to the pro- 
portions which It now assumes in the sperm whale, cannot be a for- 
tuitous modification. No animal would ever acquire such a cumbersome 
contrivance which would be so great a handicap in swimming unless 
there were definite physiological need for it. It may be, and probably 
is, an overspecialization at the present time, its evolutional velocity out 
of control, but it undoubtedly began in response to some definite need 
which it now fulfills. It seems extremely unlikely that it has any pri- 
mary function for the storage of a surplus of fat against periods of 
food scarcity, for the animal is abundantly supplied with blubber, or 
that it is used as an aid to the flotation of the head. But what the true 
function of this remarkable organ may be is entirely unknown. 


Chapter Six 

The SkuU 

OEVERAL volumes might be written descriptive of the skulls of the 
Cetacea alone, and indeed the literature on this subject is already so 
voluminous that it would be unjustifiable here to do more than mention 
the most salient points and discuss those details which are believed to 
have particular bearing on aquatic modifications. 

Skulls of the Monotremata are so unique that there are but a few 
details in that of the platypus that we can be sure are largely the result 
of its aquatic habits. These, briefly, are the dorsal position of the orbits, 
and the form of the broad, bifurcated rostrum and mandible as the 
framework of the spatulate bill. No insectivore has any well marked 
cranial modifications for an aquatic life, and almost the same can be 
said for rodents. In this order it is popularly presumed that there is 
some tendency for aquatic sorts to exhibit shortening of the nasals, to 
allow for a more dorsal position of the nostrils, and flattening of the 
dorsal side of the skull to permit of more dorsal vision ; but these char- 
acters should be evaluated with caution. Evidently a straight dorsum 
means little in itself. This part of the skull in such an essentially aquatic 
genus as the muskrat (Ondatra) is indubitably convex, but is straight in 
Hydroxys. In the terrestrial wood rat subgenus Teonoma the inter- 
orbital region is equally straight or gently concave, however. In Nilop- 
egamys it is also concave, and still more so in Ichthyomys, so it is prob- 
ably justifiable to accept this tentatively as an aquatic character in rodents. 
In Hippopotamus the nasals are somewhat shortened and there is marked 
elevation of the bony eye sockets. The peculiarities of the skull of the 
tapir probably have nothing to do with its slightly aquatic habits. 

The skulls of aquatic fissipeds and of pinnipeds show only slight 
aquatic adaptations in the details where one might expect them to be 
most marked. They have rostra that are perhaps shorter than in the 
majority of terrestrial fissipeds; but so have the cats and others. The 
same may be said about the recession of the anterior margin of the nasals. 
This is more pronounced in PhocaW\din Zalophus, but in relation to 
total length of skull it is no more marked than in the Canidae, for in- 
stance. The lachrymal bone has virtually disappeared in the pinnipeds, 

[ 107 ] 


although it is occasionally present in Otariidae and possibly more rarely 
in the Phocidae, and is imperforate. Aquatic modification seems appar- 
ent in the interorbital region of otters and pinnipeds. The interorbital 
septum is thinner in the former than in most fissipeds, is 11 per cent of 
the skull length in Zalophus and may be as little as 3 per cent in Phoca, 
in which it is so reduced that the ethmoturbinals have forced their way 
to the surface of the bone. This reduction in the interorbital width is 
ostensibly for the purpose of permitting more dorsal vision, best ac- 
complished by reducing this width so that the eyes may roll farther 
mediad and hence dorsad. The size of the orbits in the Phocidae are 
usually larger than in the Otariidae to accommodate the larger eyes of 
the former, and this has had a tendency to force the postorbital processes 
of the zygomatic arches farther caudad. The temporal fossae of the 
Otariidae indicate temporal muscles which are much more powerful 
than one would expect to find in a fish-eating mammal — an anatomical 
arrangement which is also the case with the masseter muscles — and these 
conditions are reflected in the rather robust mandible. The skull of 
the otter is also well muscled and that of some of the Pocidae, but 
the tendency in most of the latter is for a reduction in the size of the 
temporal fossae, which may not meet the sagittal line; and correspond- 
ing weakening of the mandible. 

No trustworthy conclusions may be drawn anent the palatal and 
pterygoid regions, nor regarding the bullae, for these may or may not re- 
flect conditions in terrestrial ancestors. Although no more marked than in 
many terrestrial forms, the occipital condyles in Zalophus, for instance, 
are rather narrow, for this animal has occasion to twist its head in all 
directions — movements which are facilitated by a circumscribed joint. 
In Phoca the condyles are somewhat wider, for the diflferent method of 
swimming which it employs militates again great mobility of the head 
during progression, and broader condyles make a stronger joint. 

It is in the occipital region that the pinniped skull shows the greatest 
results of the aquatic stimulus. In the usual fissiped, as in the otter, 
the occipital plane is either practically certical or else slopes rearward 
while in Zalophus the slope is definitely forward, and still more so in 
Phoca. The reason for this is discussed under the Cetacea. In Zalophus 
the lambdoidal crest bordering the occipital plane is strongly developed, 
for strong muscles (cephalohumeral, sternomastoid, trachelomastoid, 
rhomboideus anticus, splenius and semispinalis capitis) which, as a 
whole, extend uninterruptedly from the vertex to the mastoid process. 
In Phoca there is no crest formed in the lambdoid region, and the semi- 




spinalis capitis, rhomboideus amicus, humerotrapezius and cephaiohu- 
meral are confined to a narrow area near the vertex, while the sterno- 
mastoid, trachelomastoid, and splenius are confined narrowly to the 
mastoid process. Thus, in this animal, the most powerful muscles for 
motivation of the dorsal region are confined to two circumscribed areas, 
one for vertical movement of the occiput and the other for lateral move- 
ment, which is an excellent indication that Phoca has very much less 
need for such complex cranial actions as are indulged in by the sea-lions. 
As already indicated, because of the fundamentally different positions 
of the primary swimming apparatus in these two pinnipeds, the otariid 
can move its head in all manner of ways without disturbing equilibrium, 
or else this may be compensated for by movement of the hind feet ; and 
furthermore, the head and neck, by swaying motions, are of decided use 
in the terrestrial locomotion of this animal. On the other hand, the true 
seals travel on land by vertical undulations of the body, in which the head 
and neck do not play an important part, and because the swimming 
organs are at the rear, the neck, and consequently the head, can be moved 
only in moderate degree during swimming, as discussed in succeeding 

The only cranial modifications of the order Sirenia which may with 
certainty be ascribed to aquatic influences are the exceeding density of 
their bones, the recession of the nares, probably, and the slight forward 
tilting of the occipital plane, conforming to the usual posture of the head. 
In addition, adults of the extinct Steller sea cow (Hydrodamalis) were 
toothless, the place of the teeth being taken by horny oral ridges devel- 
oped for masticating soft marine algae. The phenomenally long, de- 
pressed snout and mandible of the dugong is doubtless for the same pur- 
pose, as these support horny rugosities of the membrane which should be 
equally efficient for this purpose. Why bones of the Sirenia should be 
denser, harder and heavier than in other aquatic mammals is unknown, 
but presumably they have become so in response to a long continued need. 

In Sirenia the recession of the bony nares is marked (fig. 16), and 
although a lengthy description is hardly necessary it may be mentioned 
that in this group the posterior narial border is well behind the orbits 
proper, and that anteriorly there is a great narial basin between the.pre- 
maxillaries. But the external position of the nostrils is near the anterior 
angle of the snout, to allow for breathing at the surface with no other 
part of the head exposed, a position that is definitely forward of any part 
of the narial basining of the skull. So it is seen that in this group the 
recession of the bony nares has been far posterior to the position indi- 


cated for them by the external nostrils. Conditions are not comparable 
to those in the seals, for instance, but more like those in such mammals 
as the tapir and moose (Alces), in which, in spite of the fact that the 
nostrils are situated upon a prolonged proboscis, the nasal bones have 
retreated far backward to allow for complicated musculature at the 
base of the proboscis to give it the mobility required. The narial basin- 
mg of sirenians — chiefly its broadening — is likely traceable to the same 
stimulus, so that the muscles embracing the nasal passages may have 
broader, firmer attachment upon its margins. To me these facts are 
suggestive of the possibility that originally, in the sirenian ancestors, 
external evidence of complicated nasal conditions may have been more 
marked and that possibly they were equipped with a well developed 
proboscis or comparable narial equipment. Strength is added to this 
postulation not only by the fact that in the Eocene sirenian Eosiren, with 
very long rostrum, the nasal basining extended to the posterior part of 
the orbits, but also by the present sirenian equipment consisting of 
highly specialized and mobile "lateral lips" and by the almost uni- 
versally accepted belief that they are of proboscidean ancestry. The 
conclusion suggested is that mammals of proboscidean derivation are 
unusually amenable to the specialization of the narial and other facial 
muscles in the production of probosceal or comparable equipment and 
that the present result of this in the case of Sirenia is found in peculiar 
labial specialization, and in the muscular conditions at the base of the 
narial passages as indicated by the recession of nares and narial basining, 
this being in part a relic suggestive of former and more specialized (in 
one sense) nasal musculature. 

According as the trend of specialization be in one way or another 
we might expect that eventually in Sirenia the nasal basining might either 
be reduced in size by the recession backward of the suture between the 
two premaxillae, and possibly by a narrowing as well, or else that it 
might become shallower by a reduction in height of its bony borders, 
and final broadening, the logical conclusion being a facial condition ap- 
proximating in some respects that now occurring in the Odontoceti. 

Small, thickened, ovoid nasal bones at the side of the middorsal line 
are present in Trkhechus\ there is no place for any such bones in Halt- 
cove, and Hydrodamalis seems to occupy a middle position in this respect. 
It is not known for a certainty that they were lacking in the latter genus, 
for they would naturally have fallen from the weathered skulls that are 
now available, and this cannot be settled by an examination of the speci- 
mens. Some have considered that nasal bones were present but this I 



doubt. On the whole it is difficult to judge in which of the three sirenian 
genera the posterior border of the nares Hes farthest caudad, as the con- 
formation of the details of this region differs considerably, but it seems 
to be most pronounced in Halkore. We know that in bodily form 
Hydrodajnalis and Halicore are more perfected for aquatic locomotion 
than Trichechus, and from the skull it seems that Halkore is slightly 
more modified aquatically than Hydrodamalis. In all three the orbit 
•proper is comparatively small, following reduction in the size of the eye- 
ball. The temporal fossae indicate that these muscles are quite well devel- 
oped. Deserving of passing mention is the peculiar hypertrophy in 
Trichechus of the zygomatic processes of the squamosals. 

Among the Cetacea the skull of the sperm whale departs most widely 
from that of the usual terrestrial type than that of any mammal living. 
But in some respects it occupies an intermediate position between that 
of the mysticetes and of other toothed whales. A better idea of the 
cetacean skull can be obtained from the accompanying illustrations than 
from a description and only some of the points of most importance in 
the present connection will be discussed. 

On the whole the bones of Cetacea are very spongy, and oily, and 
this is more marked in the Mysticeti. Perhaps the latter circumstances 
is in compensation for the fact that this group is not so prone to develop 
particular fatty areas upon the head, nor cavities and ducts for the 
accommodation of oil between certain muscles and beneath the blubber 
layer. Winge (1921) has stated that "the effect of water pressure is 
to develop unusual strength in those rostral bones which project farthest 
forward", and many others have been of the same opinion. I cannot 
subscribe to this belief. Of course the bones are heavy, as they must be 
in large mammals, but I cannot see that they are relatively as heavy, 
or rather that they form as strong a complex, as in land mammals. 

The rostrum of the usual land mammal is affected by several strong, 
fundamental stimuli ; as a support for a nose which must be well de- 
veloped as an apparatus for taking in air and odors, as a framework for 
mystaceal pads with vibrissae that may be highly tactile and for the an- 
terior facial musculature, as a housing for intricate turbinal bones, and as 
a support for dental armature. In addition certain mechanical elements 
are introduced following the fact that the lower rostrum supports the 
arch of the nasal bones covering the nasal passages. The Cetacea are 
the only mammals in which none of these stimuli is present, save very 
simple and usually weak teeth. Small wonder that in this order the 
rostrum departs widely in form from the average mammalian type. 



Unless some complication be introduced there is fundamentally little 
need for a rostrum of more mass than the mandible, and this is the case 
in Platanista. Unless there be some antagonistic influence in operation 
the tendency should be for the cetacean rostrum to diminish in width 
and perhaps increase in length to a reasonable extent, thus acting as a 
cut-water. This it has done in some instances, but whether for this 
reason or because of food habits is of course unknown. In Platanista 
the two sides of the maxillary dental arch are so close together that they 
partly blend. The extinct Zarhachis had a rostrum five times as long as 
the cranium proper, and in Eurhinodelphh bossi the rostrum constituted 
nine-elevenths of the total skull length (Kellogg, 1928), although in 
some specimens of the last the mandible was considerably shorter. Theo- 
retically a long, tweezer-like beak would be best for capturing small 
active fish which dart about in schools, but it is difficult to see how a 
beak of such length as that of Eurhinodelphis could be effectively used 
in conjunction with the limitations of a short cetacean neck; and it 
doubtless constituted an overspecialization. Because there have surely 
been many complications, however, we find rostra of all shapes, down to 
exceedingly short, broad ones such as in Globiocephala in which the 
stubby mandible, overhung by a rostral frontal bulge, appears ideal for 
scooping up relatively inactive, bottom-living food; and yet these whales 
are said to feed on cephalopods. 

The rostrum of a porpoise with moderate beak, such as Tursiops, com- 
posed of the maxillaries and premaxillaries, is relatively thin in vertical 
dimension, and certainly not as strong as the average land mammal with 
equal rostral length would require. The killer whales, with their carniv- 
orous habits, need a stronger rostrum and have it. But the rostrum of 
Physeter is certainly not as massive as one would expect to see as the 
support of its huge frontal fatty equipment, and the rostrum of the 
balaenid whales appears very fragile as a scaffold for their remarkable 
baleen armature. Mechanically the attachment of the Mysticeti rostrum 
to the rest of the skull is very weak, for the principal sutures are largely 
in one plane where they could all be acted upon by a single oblique force. 
But it has been entirely adequate for the needs of the animal and that is 
all that is requisite. On the other hand, the attachment of the base of 
the odontocete rostrum to the cranium is mechanically of exceeding effi- 
ciency. It is broadly distributed ijn laminated, squamous sutures in a 
manner that is stronger than could be accomplished by dentate sutures 
of the usual sort. And this is not surprising, for the stress which the 
cetacean rostrum must undergo is applied almost exclusively in a pos- 



terior direction. Hence the force of the water pushing against the tip 
is distributed over the anterior part of the skull segment next caudad. 
Even in the baleen whales is this largely true, for when the mouth 
is open and the baleen is subjected to water force the animal is presum- 
ably always swimming at low speed. 

If we examine the stimuli experienced by other parts of the head we 
find that the eyes are reduced and unusually placed, the function of the 
ear has altered, the masticatory musculature is reduced in odontocetes 
and altered in all, the cranium proper must, with the base of the rostrum, 
act as an anchor for the narial musculature, and the need for supporting 
the head has been simplified by flotation. Hence almost all of these are 
difi^erent from what the usual land mammal meets. 

In zeuglodonts the bones of the skull retain essentially the normal 
mammalian relationship, although by a recession of the nasals the bony 
nares have become elevated from one-half to two-thirds the distance to 
the orbit (Kellogg, 1928). But in some other respects, as reduction of 
the hind limb, these mammals at the time of their extinction were al- 
most as highly aquatic as most existing whales. Hence it is evident that 
they were either inherently lacking in the ability to respond to the stimuli 
that have resulted in the telescoping of the cetacean skull, which, under 
the circumstances, appears unlikely, or else that they did not experience 
those stimuli that were most critical in bringing about telescoping. What 
these may have been will be discussed in the case of existing Cetacea, 
but it may here be mentioned that the mere fact that the gradual recession 
in zeuglodonts of the anterior border of the nasals to a position from 
one-half to two-thirds the distance to the orbits does not necessarily 
mean that the external nostrils must have shifted to this position. This 
has been accepted without question, it seems, but the bony nares lie far 
to the rear of the nostrils in sirenians, and the same may have been true 
in zeuglodonts. Not only that but the recession of the nasals may, as 
in the moose, indicate that they had a small proboscis, and so the nostrils 
may actually have been situated anterior to the tip of the rostrum. If 
this were really the case, then the zeuglodent skull lacked all those 
stimuli which I now believe to have been at all important in bringing 
about the condition of telescoping, except that of backward water pres- 
sure against the head. 

As Miller (19'23) says, in "modern Cetacea the most conspicuous 
facts are these: (a) That the telescoping of the skull was far advanced 
in the earliest known extinct genera, and (b) that this process has de- 
veloped according to two difl^erent plans". Briefly, in odontocetes the 



telescoping, or the sliding of some bones over or under others so as to 
result in bony laminations and shortening of some of the cranial ele- 
ments, has clearly been from before in a backward direction against the 
cranium. The maxillaries and premaxillaries have spread backward and 
to the side so as to override the frontals, the parietals have been crowded 
far to the side and the interparietal eliminated. In some forms {Kogia) 
this is so marked that the maxillaries or premaxillaries may actually meet 
the supraoccipital. There is no tendency for the elimination of the 
lachrymal as in the pinnipeds, but like the Sirenia this bone becomes 
thickened and more massive. In practically all living forms the meseth- 
moids and ectethmoids fuse into a flat bony plate to form the posterior 
wall of the nares and the nasals are reduced to thickened ossicles above 
these. The rounded occipital plane has a marked forward inclination, but 
this is most moderate when compared to conditions in the Mysticeti. 
Other peculiarities often exist, as the fact that in Platanista the palatines 
meet the maxillaries and the latter have a remarkable development in 
broad extensions stretching upward, while in adult males of Hyperoodon 
there is a comparable development of the premaxillaries. It is not un- 
likely that these bony extensions are either cause or efl?ect of unusual 
specialization of the blowhole musculature attached to them. In life the 
bony passages are approximately vertical, although the angle, from below 
upward, which they describe with the long axis of the skull is slightly 
more than a right angle. Thus the nares have migrated as far to the 
rear as the braincase will allow, and farther, I may add, than would be 
possible without the recession of the olfactory lobes of the brain. One 
bony narial aperture is usually slightly larger than the other and this 
varies individually. In all odontocetes without exception, however, the 
right half of the facial region of the skull is larger than the left. In 
other words, the measurement from between the nares to the lateral 
border of the frontal is always larger upon the right side than the left. 
And the Odontoceti are the only mammals either living or extinct in 
which this cranial asymmetry is the normal condition. In some forms 
it is much less marked than in others, while in the Physeteridae, because 
of accompanying asymmetry of the bony nares, it is carried to an extreme. 
All manner of theories, most of them somewhat fanciful, have been ad- 
vanced to account for it. Abel claimed, without any convincing argu- 
ment, that it was due to the atrophy of the nasal bones and shortening 
of the braincase. Kiikenthal believed that a sculling motion of the tail 
tended to turn the animal to the left, resulting in a thickening of the 
cranial bones upon that side and a consequent broadening of the right 




side. Lillie (1910) inferred that the laryngeal asymmetry in the Physe- 
teridae caused the asymmetry of their skulls, but failed to indicate how. 
I attempted to discover whether there was not a twisting action of the 
facial muscles in closing the blowhole so that those of the right side de- 
veloped more strongly than those upon the left, but the results were 
entirely negative. Nor has Ernst Huber found any differences in the 
facial muscles of the two sides that might contribute to the condition. 
In short we know absolutely nothing regarding the cause of this unique 
cranial asymmetry and no logical theory for it has yet been advanced. 

In regard to the Physeteridae or sperm whales, Kellogg (1928) says 
it is confidently believed that "generalized sperm whales had been dif- 
ferentiated from the main odontocete stock subsequent to the elimination 
of the postorbital constriction, but at a time long before the beginning 
of the Miocene". In the cachalot {Physeter) the telescoping has also 
been mainly from before backward, as in other odontocetes, but there 
has been a basining of the whole facial region, this extending, for the 
accommodation of the huge spermaceti organ, well back of the bony 
nares to the cranial crest, which is vertical to the condyles. The maxil- 
laries reach this at some points, over-riding the frontals, but while the 
left nasal passage is huge, the right is but a fraction of its size. In conse- 
quence, extraordinary osteological peculiarities have resulted. While the 
right premaxilla almost reaches the cranial crest, the left is much shorter, 
stopping at the blowhole (fig. 18). Upon the left, the flattened, ex- 
panded nasal takes the place, posterior to the nares, of the position oc- 
cupied upon the right by the backward extension of the premaxillary, 
and the right nasal bone has been eliminated. 

In the case of the odontocete mandible the condyle is reduced and the 
articulation peculiar in the spreading of the ligaments, but this is more 
marked in mysticetes. There is a broad, flaring aperture to the dental 
canal, opening to the rear and partly filled with fatty tissue. The an- 
terior mandible is usually relatively weak in accordance with reduction 
in tooth size, but in ziphioids having well developed tusks it may be 
larger and of peculiar form. There is a true symphysis menti, which in 
most forms is small in area and lacks strength, but especially in some 
of the long-beaked, extinct forms, was large. Thus in Argyrocetes, 
Kellogg (1928) said that the mandible had a length of about 35 inches, 
three-fifths of which was symphysis. Why the mandible of such forms 
as Eurhinodelphis stopped far short of the rostral tip is unknown. Ii 
some ziphioids and mysticetes the mandibular tip projects beyond the 



Cranial conditions in the Mysticeti are more difficult to describe and 
the reader should turn to the illustration of Balaenoptera (fig. 19)- It 
is seen in this skull that although the premaxillae and maxillae are long 
and of very specialized shape, they really do not extend farther backward, 
using the position of the eye as a criterion, than in the normal mammal. 
As in odontocetes it is the central elements of the skull — frontals and 
parietals — that have suffered crowding, but in a different manner. In 
both does this section appear to have been squeezed between the rostrum 
and occipital, but whereas in the odontocetes the occipital seems to have 
remained more or less stationary while the base of the rostrum has been 
forced caudad, in mysticetes the appearance is that the rostrum main- 
tained its position while ^the occipital did the pushing, over-riding the 
parietals and overhanging the inferior part of the frontals, which are 
expanded to form the floor of the temporal fossae in an extraordinary 
manner. Beneath this is situated the eye. The relative weakness of the 
rostrum should again be emphasized. Reference to the illustration (fig. 
19) shows that envisioning the mysticete skull as though it were but 
a few inches in length one would never suspect that it was anything more 
than a purely edentulous mammal whose rostrum was not subjected to 
many vicissitudes of strain — certainly not that it acted as a support for 
an enormous armature of baleen (especially is this the case in the balaenid 
whales). But like other parts of the animal the weight of the baleen 
is largely counteracted by the flotation of the water and the lack of much 
resistance applied to it, such as is experienced by the grinding of an upper 
dental armature against a lower, save in a fore-and-aft direction by water 
pressure, allows for the relatively moderate strength of the whole ros-' 

The nasal bones of mysticetes are much reduced but are otherwise 
approximately normal. As usual in terrestrial mammals they roof over 
the nasal cavities, which are more gently sloping, and not vertical as in 
odontocetes. As in odontocetes the interparietal has been eliminated in 
mysticete adults, although Ridewood (1922) found it in small fetal 
balaenopterids, but not in those of Sibbaldi/s or Megaptera. It was also 
present in the Balaenoptera borealis dissected by Schulte (1916). The 
palatine is excluded from the wall of the nasal cavity. The skull is 
symmetrical, although Kiikenthal, with characteristic persistence in his 
theories, attempted to prove asymmetry. He did show fractional differ- 
ences, naturally, for it is rarely that any mammal skull has the two sides 
precisely alike, but these were not sufficient to be designated as asym- 



The mysticete mandible is profoundly modified by the necessity for 
its enlargement and bowing to accommodate the baleen equipment. The 
posterior processes are much reduced and the articulation with the skull 
altered in that the articular ligaments are greatly spread so as to allow 
a phenomonal amount of movement. The latter is further facilitated 
by the lack of a true symphysis menti, the connection of the mandibular 
tips being highly elastic; so that the mandible is perhaps capable of a 
greater variety and degree of twisting, distortional movements than that 
of any other mammal. This is doubtless of economic value in the se- 
curing of food, and to facilitate the replacement of the baleen tips in 
the eventuality that these become extruded from the mouth in some man- 

The development of the muscles of mastication in the Cetacea de- 
serves brief mention. In the extinct squalodonts, with heavy teeth clearly 
fitted for actively predaceous uses, the temporal muscles were quite well 
developed, if we compare their fossae with the relative size of the brain- 
case. In most living odontocetes, however, these fossae are much smaller ; 
but it is interesting to note that they still occupy the area from zygomatic 
process of the squamosal to the frontal border, as is usual in terrestrial 
carnivores. In other words, the reduction in size of the temporal muscles 
has kept pace with lateral displacement of the parietals (in Tursiops for 
instance) . This may or may not have any real significance. It is well 
known (see, for instance, Anthony, 1903), however, that mammals of 
small brain capacity are very apt to have powerful temporals reaching 
the sagittal crest, which may be very high, and that as the brain becomes 
larger and more highly specialized, the tendency will be for a definite 
reduction in size of the temporal fossae. The latter condition is probably 
not primarily a result of the increase in brain size, but is a reflection of 
the fact that to all intents a mammal to which great power of the dental 
armature (to which condition large temporal muscles are necessary) 
and consequent development of the anterior face, is critically requisite, 
is hindered in the development at the same time of a large brain. Again, 
this may or may not be of significance in the case of odontocetes. 

Following the reduction in size and masticatory importance of the 
teeth the masseter muscles of at least some odontocetes are decadent. 
In fact they may not have any definite attachment to the skull at all 
{Neonieris) and but a very slight one to the mandible. But they were 
probably much better developed in the extinct porpoises with phe- 
nomenally long beaks. 



The huge mandibles of mysticetes, being supported to a considerable 
degree by flotation, do not need a great mass of muscle for simply opening 
and closing the mouth, but with mandibles of such mass there must be 
a large supply of reserve muscular power, else a chance encounter with a 
companion, backed by all the momentum of its huge body, would have 
disastrous results in the shape of dislocation or fracture. Hence we 
would expect to find the temporal muscles of considerable size. Of 
course they are not as powerful relatively as in a terrestrial carnivore 
which is used to rending sinew and crushing bone, but they are, neverthe- 
less, well developed. At present in balaenopterids they extend far for- 
ward of the level of the eye and quite to the base of the rostrum, this 
being permitted by the position of the eye beneath the intervening supra- 
orbital plate of the frontal, above which the temporal muscle lies. Kel- 
logg's illustrations (1928) indicate that there has been progressive for- 
ward extension of the temporal fossae since the Oligocene. In Patrio- 
cetus of this period the over-riding of the frontals by the temporal 
muscles is not apparent or has only just begun; in Miocene Cetotherium 
the fossae do not quite include the entire supraorbital part of the fron- 
tals ; while in living adult Balaenoptera they apparently extend beyond 
and onto the maxillae, although in the fetal state they are naturally more 

There now remains for discussion the question of the telescoping 
of the cetacean skull. This telescoping may be divided into two cate- 
gories: the sliding of one bone over another, and the crowding by the two 
terminal elements of the skull upon intervening elements so that the 
latter are reduced or eliminated. The latter may well be but a result of 
the former, for presumably the central elements have been mechanically 
the weaker and have been largely overcome by a stronger force applied 
from before or behind. 

Miller (1923) is the latest to have considered at length this process 
of telescoping. He unqualifiedly subscribes to the belief that it has 
resulted from the backward push of the water against the head as the 
animal moved forward, combined with the forward push supplied by 
the moving body. In other words, that the head has been squeezed 
between the compressed wall of water against the snout at one end and 
the neck at the other. To account for conditions in odontocetes he con- 
siders that these two stresses have been relatively uncomplicated and 
that the base of the rostrum has overspread the braincase for the reason 
that originally the sutures in this region were of the squamous type, 
such as now occur in the fox, and hence would more easily over-ride the 



frontals than could the occipital, this presumably having had a dentate 
type of suture. 

He considers that the simpler type of water-and-body force acting upon 
the odontocete skull was complicated in mysticetes by the downward pull 
exercised by the enlargement of the head for baleen armature, the down- 
ward pull offered by the latter when the mouth is wide open for feeding, 
and the upward pull at the back of the skull needed to counteract these 
forces. He considers that the telescoping in this group has taken the 
form largely of an overthmst of the occipital element from behind 
forward chiefly because, like the present sea-lion, the occipital suture 
of the mysticete stock was of the squamous type, over-riding the brain- 
case with more ease than could the maxillary-premaxillary element which 
presumably had a dentate type of suture with the frontal. 

In final analysis this theory of Miller's rests solely on the premise that 
the only stimulus for the telescoping of the skull in whales of all types 
has been the backward, or at times, oblique, pressure of the water through 
which the animal moves. He advances as reasons for lack of telescoping 
in the skulls of other aquatic mammals the facts that their heads are 
relatively smaller, offering less resistance to the water, they are not held 
so stiffly or uniformly pressed into the resisting element, and that they 
all swim at lower speed. In explanation of the importance of the last 
contention it may be mentioned that the resistance offered by water to 
a body moving through it increases as the square of the velocity. Hence 
the pressure against the head of a porpoise swimming at 30 miles per 
hour would be 36 times as great as against a sea cow (did it have a head 
of the same shape and size) moving at five miles per hour. 

There is certainly a definite stimulus supplied by strong, backward 
water pressure which we would expect to find reflected in cranial details. 
It is perhaps certain that no mammal that had not been accustomed to 
moving through the water with speed for long ages could exhibit telescop- 
ing of the skull in a form as perfected as we now find it in odontocetes. 
We may therefore put this down as an essential stimulus. But I am very 
loath to believe that this was of sole importance in producing telescoping. 
We will therefore proceed to hunt for stimuli in this direction which 
may also have been essential to the attainment of this condition, although 
perhaps in themselves of insufficient strength to have produced it without 
the aid of water resistance. 

It is clear that there is no need of great strength at the tip of the ceta- 
cean rostrum. At least they get along very well without it, and its at- 
tenuate form is permitted by the elimination of strong front teeth, the 



mystacial pad with vibrissae, functional facial musculature in this region, 
and the recession of the nares and narial musculature. We have no means 
of knowing just what effect these alterations alone would have upon the 
skull — what proportion of present conditions in Cetacea should be 
ascribed to them and what to water resistance — for in no other mammal 
have they all operated ; but it is evident that they would have some defi- 
nite effect. At least this has been followed by some atrophy of the 
rostral tip in vertical dimension, and it is not unreasonable to look for 
a compensational hypertrophy in these elements at the rostral base. In 
mysticetes this is found to some extent in the broadening of this part 
of the rostrum. But in Patriocetus of the Oligocene both the widening 
of the rostral base and the attenuation of its tip is more marked than in 
any living mysticete, showing that both of these details have experienced 
secondary modification through requirements of the baleen equipment as 
this developed. In odontocetes this rostral change has taken the form 
both of the broadening of the rostral base and the spreading of the 
maxillae over the frontals. This as a lone stimulus would probably not 
be of very great importance to telescoping, but it must be considered as 
contributing in some degree to the total of stimuli for skull change. 

The odontocete skull is so formed that both the rostral tip and the 
forehead receive the backward pressure of the water, while in ir^sticetes 
the forehead does not receive it. The concomitant condition of the 
telescoping of the rostral elements in the former group is not believed 
to be purely fortuitous. At least it is presumed that the water pressure 
at this critical point would be largely instrumental in strengthening the 
anterior wall of the braincase. Admittedly this might be accomplished 
as effectively by a mere thickening of the bones, but the fact remains that 
the end has been attained by a process of laminations. The theoretical 
effect in this connection of the presence of a frontal adipose cushion is 
of course unknown, but it would naturally have some sort of stimulus, 
probably in broadening the face. These stimuli are absent in mysticetes 
and there has been no lamination of rostral elements over the frontals. 

In skulls of toothed wales (excepting present conditions in the Physe- 
teridae and the Kogidae) the nasal passages have migrated just as far 
to the rear as the anterior wall of the braincase has allowed. With them 
has shifted a complicated muscular mechanism for the control of the 
blowhole. This is no system of small muscular slips, weakly attached 
to the bone, as in many mammals, but a robust complex, firmly anchored 
to the skull over the entire breadth of the forehead and converging to 
the blowhole and a part of the adipose cushion. These have dragged 


with them the main trunk of the facial nerve, which now turns abruptly 
backward over a groove in the maxilla and lachrymal at the base of the 
rostrum and so to the dorsal surface. A like condition obtains in the 
case of the usual anterior opening of the infraorbital nerve of the maxil- 
lary branch of the trigeminal, the infraorbital foramen having been 
forced first dorsad and then caudad, where the nerve now emerges in 
separate branches from several foramina, which may be located directly 
above the eye {Tursiops) and even within the bony narial orifice. 
Whether or not this actually be the case it appears superficially very 
much as though the posterior extension of the premaxillaries and the 
expansion caudad and laterad of the maxillae had been stimulated to 
this action by the migration in these directions of the facial muscles 
which are so firmly attached to these bones; — that the attachments of 
these muscles had not shifted from one bone to another but had dragged 
the bones with them, just as anterior migration of the occipital muscles 
alters conformation of the occiput. And that the more circumscribed 
area which these narial muscles needed in attachment to the frontals 
had permitted all but a narrow border of these to be covered by the 
maxillae. At any rate this excessive migration of the nostrils has had 
a profoundly stimulating effect upon osseous conformation. This is 
beyond question. 

The mysticetes indicate an entirely different state of affairs. There is 
no backward pressure of water against the anterior wall of the brain- 
case, save as applied through the rostral tip, and there has been no oc- 
casion to build up a thickened or laminated bony wall for its proper 
resistance, nor is there an adipose cushion in this region. Unlike odon- 
tocetes the baleen equipment has retained considerable growth-stimulus 
in the rostrum proper, and the nostrils and their musculature have not 
been forced as far backward against the braincase as they could go. On 
the contrary the bony nares are more nearly comparable to the usual 
mammalian condition. And most important of all, there has not been 
such a pronounced migration backward of the narial musculature, and 
the requirements in this respect are not such as to necessitate any great 
broadening out of bony elements to accommodate a pyramid of muscles as 
broad as any part of the skull, for in mysticetes the intertemporal region 
is quite narrow. 

Kellogg (1928) considers that the mysticete maxillae cannot over- 
spread the braincase because the posterior part straddles the supraorbital 
processes of the frontals, and he presumably accepts this as a fundamental 
reason why these whales do not exhibit more telescoping at the base of 


the rostrum. I doubt whether this is a definite hindrance. If the stimu- 
lus in other directions were of sufficient strength at least the superior 
part of the maxillary could extend farther to the rear. Nor do I consider 
that the possible presence of a dentate type of suture between the maxillae 
and the frontals of the mysticete ancestor would be a permanent inhibitor 
of marked telescoping in this region, as contended by Miller. Sutural 
changes cannot be instigated in the adult skull, in which the attachments 
of all bones are secure, but begin in the younger animal whose sutures 
are not securely locked; and the individual time of sutural closure is 
notoriously variable. As far as present knowledge goes the cranial 
sutures will change according to the needs of the animal and it seems 

Figure 18. Dorsal view of the skull of the cachalot or sperm whale, Physeter. 

unlikely that they could have been so conservative and tenacious of 
type in the cetacean ancestors as to predetermine the whole course of 
cranial development in the two groups. 

There is little or no sliding of one bone over the other in the posterior 
elements (occipital, parietal and squamosal bones) of the odontocete 
skull, and no details that may not be understood rather readily. In the 
existing sperm whale the occipital plane is vertical, which gives the ap- 
pearance that all force has been from before toward the rear, but this 
is probably a secondary result, for in the Miocene Diaphorocetus, and 
doubtless other early physeterids, the facial basin appears to have been 
not quite so well defined and there was a marked forward inclination 


of the occipital. Some of the more peculiar types of ziphioid whales 
also have an occipital that is almost vertical: but in odontocetes of the 
porpoise sort there is marked forward inclination of the occipital plane 
(discussed later). This is definitely rounded and the supero-medial 
part extends to meet the facial plane. This crowding has forced the 
parietals far laterad. The restricted temporal fossae project backward to 
indent the occipital expanse, and the lower part of the exoccipitals and 
lateral part of the basioccipital are projected into marked processes. The 
bony border resulting is not continuous but has a deep fissure for the 
jugular vein and nerves IX, X, and XI (glossopharyngeal, vagus and 
spinal accessory), with the condyloid foramen (for the hypoglossal 
nerve) upon its border. Upon this falcate border of the basi- and exoc- 
cipitals is inserted the great dorsal scalene muscle, which is probably for 
the purpose of providing static strength in the same way that the rectus 
capitis superior is so used upon the opposite side of the neck, but it 
doubtless also assists in swimming motions. At any rate this falcate 
border is not present in Mysticeti, the downward motions of whose heads 
are (presumably) largely governed by the specialized gular musculature, 
which has leverage clear from the tip of the mandible. 

The occipital condyles of the Cetacea are massive but are not any 
broader in relation to the width of the skull than in many terrestrial 

On the whole the cetacean brain is of a very high type — quite extra- 
ordinarily so when we consider that as far as can be judged members of 
this order have less need for such equipment than almost any other 
mammal. It is perfectly apparent upon inspection of skulls that the 
brain capacity of an ordinary porpoise such as Tursiops is relatively con- 
siderably greater than in a large baleen whale. But this may be merely 
in conformity with Haller's law, which is to the effect that in related 
animals the relative size of the brain is greater in the smaller sorts, 
and that endocranial capacity decreases as size increases. Dubois (1924) 
has stated that volume of brain in such comparable material ("species 
of equal organization") increases as the 5/9 power of the body weight, 
and that in mysticetes this cephalic index (not actual capacity of the 
brain case) is only one-third of what it is in porpoises. It is therefore 
difficult to determine whether there is actually a difference in the endo- 
cranial capacity between equal-sized representatives of the two cetacean 
groups, for the skulls of the larger toothed whales are curiously distorted 
in some respects and the smallest mysticetes are depauperate and not 
representative of the group. At any rate, regardless of this point, the 



brain capacity of a porpoise is relatively much larger than of a finback 
whale. In the former at least the brain shows evidence of having been 
compressed in an antero-posterior direction and is broader than long, 
this having been caused evidently by the recession of the anterior cranial 
elements against the posterior ones; and one result of this great brain 
size is the curvature of the occipital plane. In effect, then, the stimulus 
for a large brain in the porpoise has been greater than any muscular 
stimulus which might have been productive of a flat, ridged or crested 
occipital for the accommodation of unusually powerful muscles, and 
greater than this endocranial stimulus in large baleen whales. In the 
sperm whale, however, the osteological crowding that has resulted from 
the frontal basining for the accommodation of the spermaceti organ has 
forced the cerebral cavity to occupy a restricted space at the apex of the 
angle formed by the vertical part of the occipital elevation and the hori- 
zontal rostral portion. I judge that the size of the brain in an adult 
Physeter is not at the most over four or five times as great as in the 
porpoise Tnrsiops, and the foramen magnum passes downward (in an 
anterior direction) at an angle of more than 45 degrees. 

Weber (1904) has stated that the cetacean supraoccipital is paired, just 
as it is in Tatusia (Edentata) and Erinacetis (Insectivora) . 

The occipital conditions in the Mysticeti are very different indeed from 
those in the Odontoceti. Tilting of this element in the former group is 
excessive and the forward inclination of the occipital plane more accen- 
tuated than in any other mammal. In whales of the balaenopterid type 
this tendency has forced the parietals forward so that considerable por- 
tions of them surmount the frontals, and the occipital covers all but the 
most anterior part of the parietals. As is frequently remarked, in whales 
of this type (including Sibbaldus) parts of the nasals, premaxillae, maxil- 
lae, parietals and frontals occur in one transverse plane. Among living 
mysticetes the large balaenid whales (Balaem and Ei/balaeiia) usually 
are said to exhibit the most pronounced tilting of the occipital, but this 
statement needs qualification, as explained elsewhere. 

It is difficult to decide whether the occipital overthrust in mysticetes is 
now actually on the increase or wane. Kellogg (1928) has remarked 
that a greater overthrust occurred in one of the Pliocene balaenopterids 
than in any whale since. He is also of the opinion that in mysticetes 
the forward overthrust of the occipital elements has not had precedence 
over the backward interdigitation of median rostral elements, for in 
Miocene cetotheres there was less occipital tilting and more interdigita- 
tion of rostral with cranial elements. But these details hardly should 


be compared in this way for the stimuH involved were doubtless very 
different. Observation tells us that at least the stimulus for occipital 
tilting in modernized mysticetes has been one of the strongest experienced 
by the skull for the reason that it is more pronounced than in any other 

The difficulty experienced in explaining telescoping of the odontocete 
facial elements because no other mammal exhibits an intermediate con- 
dition in this respect is not so applicable in the case of the telescoping 
shown by the mysticete occipital. In the first place the latter is of a 
simpler, less astonishing sort, with the lamination of the bones not 
marked. In fact it is just about what one would expect to see result from 
an excessive foreward tipping of the occipital plane in any type of mam- 
mal. Several other mammals show this to a modified degree, and several 
exhibit incipient telescoping of the posterior cranial elements. Among 
terrestrial forms occipital tilting is perhaps most marked in the European 
blind rat {Spalax) , but in all rodents sufficiently specialized for a fossorial 
life as to be blind or nearly blind (Spalacidae, Rhizomyidae and Bathyer- 
gidae) this character is very pronounced, while in burrowing mammals 
not quite so strongly modified (fossorial octodonts, murids and geomyds) 
the occipital change is but moderate or slight. It is also a character of 
some few ungulates. 

That part of telescoping which is exemplified by the broadening of 
the surface of contact between two bones is shared by Spalax also, and 
by most if not all of the more specialized fossorial rodents, this con- 
sisting in a broadening of the squamate sutural contact of occipital with 
parietal. In varying extent it is found also in the pig tribe, which roots 
in the ground, in Arctonyx the hog badger, which does the same thing, 
and in the fur seal {Callorhinus) , but not the hair seal (Pboca) . In all 
of these as well as in the Cetacea the head frequently or almost continually 
experiences a posteriorly directed force applied against the muzzle. Water 
presses against the head of the aquatic sorts as they swim, and the re- 
mainder push against the ground, rooting, tamping the earth of exca- 
vated runways with the nose (Spalax) or using the incisors as a pick to 
loosen the earth of their burrows after the manner of many fossorial 
rodents. Mechanically it would be very easy for the forward tilting of 
the occipital to bring about a broadening of occipital-parietal suture. 
Alone it might not have this effect, but the condition may well have 
been helped by the backward pressure against the rostrum experienced 
by those mammals which exhibit this tilting to best advantage. That 
forward tilting of the occipital is not a requisite of marked broadening 




of the occipital-parietal contact is shown by the fact that the latter situa- 
tion obtains in pigs, in which the occipital plane is tilted quite far 
backivard. And that no stimulus of such sort is absolutely necessary is 
shown by the fact that broadening of sutural contracts between parietal 
and squamosal is well developed in at least some of the Kangaroos 
{Macropus) and very likely some other terrestrial mammals. The in- 
ference is therefore that such telescoping of the posterior cranial ele- 
ments as results purely in the extension of one bone over the other so as 
to broaden the line of contact may be either a character of no conse- 
quence (as far as we can see) other than possibly a phylogenetic one: 
may be largely caused by a forward tilting of the occipital, backward 
pressure frequently applied to the skull, or both. 

Invariably, I believe, it has been stated that the rostral tilting of the 
occipital plane of the mysticeti, and inferentially of fossorial rodents, 
is for the purpose of muscular strength, at which point the topic is 
abandoned without qualification or explanation regarding the quality 
of strength that is meant. Such strength is of five sorts, or is indicated 
in five separate ways, which fundamentally affect the conformation of the 
occipital shield by reason of the fact that the supraoccipital muscles seem 
utterly incapable of migrating from one bone to another, so that when 
they shift the supraoccipital must shift with them. In order properly 
to understand the cetacean conditions in this region of the skull it will 
be necessary to scrutinize the situation in certain pertinent terrestrial mam- 
mals, under five separate headings as follows: 

{A) There is the strength exemplified in the occipital conditions 
of the lion or pig. The former must have powerful neck muscles of 
a sort to enable the head to be strongly twisted in all directions in order 
that it may, for instance, break the neck of a zebra. The pig roots in 
ground that is often hard, can do this all day long, and can lift a truly 
astonishing weight with its snout. This sort of active muscular strength 
is accompanied by high and strongly-defined lambdoidal ridging along 
the supraoccipital border. 

(B) A different situation is encountered in mammals which we 
know must have great strength of occipital musculamre for the support 
of a heavy equipment of horns or tusks, but which have no particular 
need for cephalic activity, embodying strong twisting motions. Although 
not occurring in its perfected, simplest form, this sort of passive muscular 
strength is illustrated by the moose and others of the deer tribe with large 
antlers, large-horned bovines, and old male elephants. The bony indi- 
cations of this character of muscular strength do not take the form of 



high ridging, but on the contrary the occipital border is unridged and 
superficial examination would give the impression of definite muscular 
weakness in this area. The muscles need not be thick, or "powerful" in 
the usual sense, but they are either shorter or the muscular fibers less 
contractible, and the muscle must be tougher. This latter is best accom- 
plished by the presence of numerous tendinous fasciculi or bundles, and 
in the perfected state of this specialization, which no mammal has yet 
attained, these occipital muscles would be nothing but tendinous sheets, 
practically lacking muscle fibers, and resulting in the incapacity for 
muscular movement of the head in a dorsal direction. 

Certain modifications of either A or B, or a combination of both, may 
be effected mechanically by body conformation (a longer or shorter 
neck, or size of head), or by the provision, for instance, that has been 
developed by some ungulates in quite perfected form. This may be well 
demonstrated in a small antelope, which has need while grazing to lift the 
head perhaps every half minute to watch for enemies. The great muscu- 
lar energy that would otherwise be expended in this action during the 
course of a day is conserved by the presence of a highly elastic nuchal 
ligament extending middorsally to the head. If one props up a dead 
individual in standing attitude, presses the head to the ground and then 
releases it, the spring of the nuchal ligament will either automatically 
raise the head or almost do so. This provision thus economically takes 
the place of much more powerful cephalic muscles that would otherwise 
be wasteful: but no aquatic mammal has this equipment. 

(C) Length of leg in a land mammal largely influences length of 
neck (partially excepting browsers) . This, correlated with posture in- 
duced by life habits,^ to some extent, determines the normal position for 
carriage of the head. The giraffe, with its long neck, carries its head 
with axis at a right angle to the neck, while many bovines carry the 
cephalic axis practically in a line with those of the neck and body. This 
naturally has an effect upon the occipital musculature. The pig, while 
rooting, begins to apply the muscular force necessary when the cephalic 
axis is almost at a right angle to the neck. If the occipital plane were 
tilted forward or even vertical to the skull axis the occipital muscles 
would have no effective lever arm on which to work. Therefore it is 
mechanically necessary for the slope of the occcipital plane to be toward 
the rear, which it is to a very marked extent. The projection of the lamb- 
doidal crest but accentuates this feature. Surely the head and neck of a 
horse is not as strong as of a large bull, but the horse has an occipital 
plane caudally tilted and prolonged into a large crest, and so exhibits 


much greater "strength", from the usual but erroneous standpoint, than 
does the bull. The lambdoidal crest of the horse is largely due to the 
mobility of the head, and the backward inclination of the occiput to the 
fact that the head is usually held at a sharp angle with the neck. Even 
more strikingly illustrative of these points is the condition in man and 
other primates which hold the body erect for most of the time. Hence 
it is a law that in those mammals which habitually hold the axis of the 
head at near a right angle to that of the neck, the occipital plane is found 
to slope backward, to a greater or lesser degree according to the strength 
of possible antagonistic stimuli. 

(D) Among the Bovidae conditions opposite to those mentioned 
above are perhaps best exemplified in the water buffalo. In observing 
a living specimen one is immediately struck by the fact that it carries 
the axis of the head in almost a straight line with that of the neck, and 
that it does so almost continuously. We know that it has far more 
strength in this region than the horse, but there is no lambdoidal crest 
and the occiput shows some tendency to slope forward rather than to 
the rear. In the more strongly modified fossorial rodents this occipital 
tilting is strikingly marked. Those which can still see never have occa- 
sion to look in a ventral direction, but only above, to watch for an enemy 
while the head is level with the burrow entrance. The head is held 
on a line horizontal with the body, in rest, when tamping the burrow 
walls with the nose (as Spalax), and even above the body when the 
teeth are being used as a pick while extending the burrow system. Ac- 
cordingly the tilting of the occipital plane is most pronounced in those 
forms which have been subjected to fossorial conditions for the longest 
time, as indicated by their other adaptations in this direction. These 
rodents have great bulging occipital muscles but usually a very slight 
lambdoidal crest. The reason for the lack of the latter is probably partly 
because movements, although they may be of great strength, are limited 
in distance, that a good lever arm is secured without the necessity of a 
high crest, and, as seems probable, that a crest is not nearly so liable 
to develop upon the apex of an obtuse as of an acute bony angle. From 
contemplation of the above facts it would seem that when the head is 
normally carried parallel with the axis of the neck the occipital plane 
will be tilted in an anterior direction. 

(£) Certain developments that I have noted only in Cetacea would 
seem to point to the possibility of a fifth set of conditions that could 
be of import in influencing occipital conformation. In Balaenoptera 
horealis Schulte (1916) showed a single semispinalis capitis muscle as 



huge, occupying practically the whole of the supraoccipital region, while 
the insertion of the posterior rectus capitis was reduced and shifted 
largely lateral to the condyle so as to have almost no function in sagittal 
movements. In Neonieris I found the posterior rectus to be much larger, 
the attachment covering an area perhaps one-third of that occupied by 
the correspondingly reduced semispinalis, which was double; while in 
Monodon the rectus was enormous and the double attachment of the long 
back musculature very thin; and Murie (1873) showed that the same 
condition obtains in Globiocephala. Obviously the rectus aids more in 
static than in active strength, while the longer muscles to the occiput 
are of chief use in active movements. The former does not need an oc- 
cipital with a marked slope, while for proper efficiency in the case of 
a mammal holding its head parallel with the body axis, the latter does. 
This will be discussed further. 

It will thus be seen that there may have been, and doubtless were, more 
than one stimulus for a given development. These can be discussed, 
but we cannot tell which were the stronger or, indeed, if some have not 
been counteracted by antagonistic influences. 

We have no means of knowing the static or tensile strength of the 
cetacean occipital musculature. The term strength is but relative and 
indefinite at best. Certainly in this order the unbroken body contour 
in the cervical region means that the underlying muscles are very robust, 
but in Neomeris, for instance, although the complexus was extremely 
heavy at the base of the neck, it was very thin indeed at its broad attach- 
ment to the skull. Of course a porpoise with skull of moderate size may 
toss the head for a few inches with some show of force, and it is likely 
that relatively prodigious power would be needed in order to twist the 
creatures head far to the side in opposition to its wish, but nevertheless 
cetaceans with large heads cannot even lift these from the ground if they 
become stranded. The cetacean's head can be but little heavier than the 
water it displaces, so levitation is almost entirely by flotation. Further- 
more it is inconceivable that a whale while swimming does not adjust 
the position of the head so that it is in equilibrium, balanced so that 
there is no aquatic force tending to twist the head below or to one side. 
Hence, leaving out of consideration the fact that the side muscles are 
employed only occasionally while the dorsal ones are in use almost con- 
stantly, the supraoccipital muscles need not be any stronger than those 
of the side of the neck, unless complications be introduced by the flatten- 
ing of the whole head. 



In all cetaceans there has been a mechanical stimulus for the quantita- 
tive shortening of the neck, and this has been accompanied by a shorten- 
ing of the muscles extending from the thorax to the head. This acts 
in anchoring the latter more firmly, which is of decided advantage to 
the animal. In a massive body which moves through the water with 
the velocity characteristic of whales of less extreme specialization, any 
really marked lateral movements of the relatively large head might be 
as disastrous as is the abrupt turning of the front wheels of an automobile 
running at speed. It would enable the pressure of the water to snap the 
head sharply to the side and might result in a broken neck. A whale 
does not need any movement of the head in a sagittal plane for steering, 
because this is accomplished by tilting the flippers. It swims by inducing 
vertical curvature of the entire body, usually from a point in the anterior 
thorax as already explained, and this necessitates slight movements of 
the head in this plane. This is a constantly repeated action, whereas 
slight lateral movements of the head for steering in the horizontal plane 
(also facilitated by flipper action) are only occasional. As already stated, 
these lateral movements must be moderate in degree and should be, as 
they are, but little if any more decided than the possible curvature of 
the body proper. In a right turn, for instance, the muscles of the right 
side need apply practically no power. What is principally needed is a 
relaxation of the muscles of the left side of the neck, when the water 
pressure against the left side of the rostrum will force the head toward 
the right. Then must the left muscles have great static strength to keep 
the head from turning too far, and active strength through a short 
distance only- to overcome the water pressure and bring the head back 
into position when it is wished to progress once more in a straight line. 

In addition to essential but short sagittal movements of the head in 
swimming most odontocetes probably have need for tilting the head 
downward in securing food. The amount of this is unknown, but at 
least we know that in almost all sorts the rostrum is carried somewhat 
more depressed and is not directly on a line with the body axis. We 
would therefore not expect to find the occipital plane so sharply tilted 
forward as would be the case were the rostral axis parallel with the 
body. Furthermore, in most odontocetes the rostrum is relatively not 
nearly so large as in mysticetes and the occipital musculature does not 
need such leverage to control its movements. In considering this detail, 
however, there is encountered a great complexity of possible stimuli 
which it is difficult to explain with clarity, and almost impossible at the 
present time properly to evaluate. 



The Mysticeti never pursue an individual food item but whole schools 
of relatively small fish or shrimp. Hence there should be no need 
for any motion of the head connected solely with feeding save that of 
opening the mouth. There is thus need only for such slight movements 
as are involved in swimming and steering — very simple ones which have 
allowed the occipital musculature to assume a correspondingly simple 
form. They are unlike Odontoceti also in that the rostrum is not stati- 
cally held in a depressed posture. In balaenopterids the whole head 
anterior to the occipital plane presents dorsally a practically straight line. 
In other sorts the whole dorsum of the skull is curved, but still the "aver- 
age" rostral axis seems fairly on a line with that of the body. At any 
rate all mysticetes hold the head in a more elevated position than do 
generalized odontocetes. 

In connection with the occipital tilt of mysticetes Miller (1923) con- 
sidered as essential the presence of a downward pulling force applied 
to the head. But the downward force, furnished by water pressure, 
against the head of a baleen whale could hardly be relatively greater than 
that experienced by a pig while rooting. A force is present, certainly, 
but the critical factor is the position in which the head is normally held 
while the force is applied, rather than the force itself. If a whale habitu- 
ally held its head at a right angle to the neck it would not have a forward 
tilt to the occipital, but a backward one. A better understanding of the 
situation may be obtained by examining occipital details in relation with 
external details. It will be found that in Rhachianectes, the most primi- 
tive of living mysticetes, the occipital tilting, while perhaps more pro- 
nounced than in odontocetes, is very moderate: at the same time it has 
the longest neck. Tilting is most accentuated in the right whale (ba- 
laenid) type, and intermediate in other sorts {Sibbaldus, Balaenoptera, 
Megaptera) . Some qualification is necessary in the statement regarding 
balaenid whales, however. In these the occiput is most sharply tilted in 
relation to the cranium proper, but the entire skull is sharply curved and 
sickle-shaped (fig. 20), and the head so held in normal posture that 
the occipital tilt in relation to body axis is only moderate, and actually 
less than in balaenopterids. In reality occipital conditions in living mysti- 
cetes may be divided into two categories. In one, comprising Sibbaldus, 
Balaenoptera and Megaptera, the occipital tilt to body axis is extreme, 
and may be referred to as the balaenopterid type. In the other group, 
consisting of the gray and balaenid whales, it is more moderate, and may 
be called the balaenid type. The rostra of the latter are always down- 
wardly curved (see fig. 9) in varying degree, and the conformation of 



the head is such that backward pressure of the water is appHed both to 
the rostrum and the throat, so that when the animal is swimming the 
upward and downward force of the water could easily be equalized by 
slight adjustments in the posture of the head. And this position of 
stability must be regarded as the normal one. In the balaenopterid type 
the dorsum of the whole head is so flattened and so held horizontally 
that any downward pressure by water resistance would be relatively 
negligible. The gular region is more curved, however, off^ering more 
resistance to the water when the mouth is closed and the throat expanded 
slightly by force of the baleen plates. Hence in these whales the only 
action of water resistance when the animal is swimming at speed with 

Figure 20. Left lateral view of the skull of the right whale, Eubalaena glacialis, 
redrawn from Holder. 

the mouth closed is to push upward and backward against the throat, 
forcing the head up and supporting it without the aid of muscular eff^ort. 
If one subscribe to the belief, which I do not, that these whales ever 
move to any extent while feeding with the mouth wide open, the chief 
water force would be applied through the condyles of the mandible. 
And water force applied to the baleen plates while the mouth is open 
could be obviated by a slight tilting upward of the snout. In short, all 
of the above mechanical forces save purely backward pressure at the 
muzzle may be simply overcome by a slight tilting of the head in one 
direction or another. The result is th^t the balaenopterid type of whale 
is enabled consistently to hold its head in a more elevated position than 
is practicable for the balaenid type. In fact it must do so. In the balaenid 



type the rostrum is rather evenly rounded from side to side, while in 
balaenopterids it is flattened and expanded, so that in the latter the 
rostrum must be kept safely elevated to obviate the danger that when 
moving at thirty miles an hour the rostrum might become sufficiently 
depressed for the water to take hold of the broad dorsal plane and force 
the head sharply downward, with results that might be highly uncom- 

In addition to the fact that the occipital should slope more in mysticetes 
than odontocetes because the former hold the head more elevated, there 
is the thesis, explained in the next chapter, that because of the larger 
head the pivot of motion for swimming, normally situated in the anterior 
thorax, has been shifted somewhat farther forward in mysticetes, which 
might very likely have a tendency to force the occipital musculature more 
decidedly rostrad. There is the further theory that the larger, flatter head 
of the balaenopterids requires a longer power arm for the occipital 
muscles than the short neck alone can provide ; and the possible influence 
for the same effect suggested by the fact that in the balaenopterid the 
semispinalis capitis is the one developed for chief control in elevating 
the head, while in at least some odontocetes the less potentially active 
posterior rectus is fully as important and may be far more so. In addi- 
tion the rounded braincase of most porpoises shows that there has been 
an endocranial opposing force which the occipital has had less success 
in overcoming than in the much larger mysticetes. Finally, facial mi- 
gration toward the rear is not so pronounced in mysticetes, and this sug- 
gests that in odontocetes this stimulus for facial recession has proved 
too strong for the occipital tilt to overcome so readily, and that its for- 
ward inclination has at least been retarded by this influence. That this 
latter is not a mere fancy is suggested by the situation in the sperm whale. 
Although fossils of primitive type {Diaphorocetus) show a marked oc- 
cipital tilt, backward pressure supplied by the developing spermaceti or- 
gan has proved the stronger stimulus and overcome it, until now the oc- 
cipital plane of the living cachalot is practically vertical. 

The above are at least some of the reasons accounting for the fact 
that in Cetacea the forward tilting of the occipital plane is extreme, that 
it is more pronounced in mysticetes than odontocetes, and more marked, 
in relation to body axis, in balaenopterid than in balaenid whales. 

There are many having their own particular theories who will not 
agree with me, but to me the evidence as marshalled above is sufficiently 
strong for the acceptance, at least until many more data are available, 
of the following thesis. That flotation and water pressure has had an 



important influence in modelling the external shape of the cetacean head. 
That backward force offered by the water as the animal moved forward 
has been requisite to the extreme condition of the telescoping of the 
odontocete rostral elements, but it has been only a secondary factor, the 
primary one consisting of muscular influence accompanying the extreme 
recession of the odontocete type of narial musculature. That the moder- 
ate forward tilt of the odontocete occipital plane was caused by muscular 
conditions induced solely by the elevated and static position of the head 
unaccompanied by the need for cephalic agility, and that as far as the 
evidence points this region has been totally unaffected by backward pres- 
sure of the water. That the elimination or lateral displacement and re- 
duction of the central cranial elements of the Odontoceti have been 
caused solely by backward crowding of the face and forward crowding of 
the occiput, and hence by forces that were exclusively or chiefly muscu- 
lar. That the telescoping and more excessive forward tilting exhibited 
by the occipital region of mysticeti is due solely to muscle migration 
finally caused by the static position of the head with need for the mini- 
mum of cephalic movement. 


Chapter Seven 

The l^eck 

In most mammals the head and neck can develop partly independent 
of the trunk and the two together may in some respects be compared to 
an appendage such as the arm or leg. The neck may, and usually does, 
become long or short according to leg length, so that the mouth may 
reach the ground ; and it almost invariably shortens in mammals which 
can employ the fore foot as a hand for conveying food to the mouth. 
The neck must have strength according to size of head and the sort of 
work that the latter performs, and agility according to the needs of the 
animal. In the case of a mammal the size of a rat no more need be 
said, for anything of this size and such light weight could doubtless run 
as fast and easily were its neck twice as long. But in a large body weigh- 
ing hundreds of pounds another and vital element is added, constituting 
adequate equilibration according to the manner of locomotion habitually 
employed, and this introduces many complex factors which we can only 
partially understand. In the case of such an amphibious mammal as the 
sea-lion, the head and neck must act as a balancer, in a gravitational 
sense, during terrestrial progression. This no longer figures in a mam- 
mal that is exclusively aquatic. In such the head and neck usually must 
be of some particular mass and length either so as efficiently to play an 
important part in swimming, or more rarely the method of progression 
is such that the anterior end of the animal plays no important part in 
it, and this portion of the body can then develop in response to some 
other stimulus. 

The conditions which aquatic reptiles seem to have encountered are 
so at variance with those that have applied to mammals that separate 
though brief mention should be made of them. Most of the large, 
extinct, aquatic reptiles seem to have progressed for ages after taking 
to the water by a rhythmic, diagonal movement of all four limbs that 
may be compared to the trot, so that the head and neck had no need to 
act in maintaining equilibrium, or as a rudder. Hence, while this mode 
of swimming is employed the neck may lengthen or shorten in response 
to other stimuli. Further, the neck in the two classes can hardly be 
compared, for mammals exhibit a truly extraordinary conservatism in 


the number of cervical vertebrae, which hmits agihty, while the vertebrae 
of this region in reptiles are as remarkably plastic, and may multiply in 
some manner or other to a phenomenal number, allowing a length and 
mobility of neck utterly unattainable in mammals. So an aquatic reptile 
with a figure suggestive of a wash-tub may have a neck like the half of 
a huge snake, with as many as 76 cervicals in some plesiosaurs, enabling 
it to strike at and seize a fish which the body would be incapable of over- 

There are two body forms that an aquatic mammal might assume al- 
lowing either a lengthening or shortening of the neck according as 
there are present or absent certain stimuli connected with the acquisition 
of food. In other words, there are two possible forms in which the 
proportions of the neck would not vitally aflfect locomotion. One is 
the anguilliform or eel-like type, which might be attained by elongation 
of the body and tail without the length of neck being aflfected. No 
living mammal swims by this method. The long-tailed zeuglodonts 
were definitely modified in this direction and the neck was not elongated. 
The other type of body referred to above is that in which the legs of one 
side are separated from those of the opposite side by a relatively wide 
interval. If this were the original body form, as it likely was in the 
hippopotamus, we would expect all four legs to be employed in rapid 
swimming, as is the case in the mud turtle, the ultimate probability being 
that four paddles of approximately equal size would be developed, and 
these could perform all functions of swimming and steering, permitting, 
as far as locomotive factors are concerned, either the lengthening or 
shortening of the neck. The size of the head in Hippopotamus might 
be expected to inhibit the development of a long neck. The only other 
mammalian instance for discussion under this head is the platypus. Its 
body is now somewhat turtle-shaped, and may always have been so, for 
all four feet are well developed for swimming. The fore feet are now 
chiefly used for this purpose, which was permitted by the development 
of an excellent equilibrating organ from the tail. Theoretically it could 
now undergo a lengthening of the neck, which at present is short; but 
it would seem that a short neck is advantageous to its habits of feeding. 
It will occur to the reader that the duck employs somewhat similar 
methods of feeding and it has a long neck, but this is a necessity for 
proper equilibration in flight. If it were not then it would not be held 
in an extended position, but curved as in the heron. 

Most of the aquatic mammals listed are not sufficiently specialized for 
us to tell whether the length of neck is different from that of their an- 



cestors or not. Furthermore, we lack a proper yardstick in that there is 
no means of teUing in a particular case whether an apparent slight 
shortening of the neck is real or due entirely to a relative elongation of 
the trunk. Of one thing, however, we can be reasonably sure in most 
instances. In perhaps a majority of large terrestrial mammals a primary 
regulator of neck length is limb length, for the neck must usually be long 
enough to reach the ground. Aquatic mammals lack this stimulus and 
we can feel reasonably sure that this provides one factor for neck shorten- 
ing. It seems that the only factor for decided elongation of the neck 
that could possibly occur for this group would be in the case of one with 
a slow-moving body which found it of preponderant advantage to secure 
very active prey by darting movements of a mobile neck. Unless its prey 
be of marked agility then the aquatic mammal finds it as easy to turn 
the head by a shift of the entire body as to bend the neck. This is 
certainly so in the case of Sirenia, and it is a stimulus for cervical shorten- 
ing, probably to a greater degree than in most aquatic mammals, for these 
beasts are so sluggish that feeding habits are likely a stronger stimulus 
for neck length than those arising from locomotion. 

The above remarks suggest that the sirenian neck should be definitely 
but not excessively shortened, and this it is. Additional evidence is 
furnished by the fact that occasionally in the manati two or three of the 
cervical vertebrae are fused, the second, third, and fourth then being the 
ones involved. This is one of the only two living mammals having but 
six vertebrae in this series. Murie (1872) considered that it is the third 
cervical that is missing because of the conformation of a slip of the 
scalenus muscle, while several others have considered that it is the sev- 
enth which is lacking. Considering the phenomenal regularity with 
which just seven cervical somites are laid down in the case of mammals 
I would be extremely loath to believe, without extremely strong evidence 
furnished by the conformation of the cervical nerves, that Trichechus 
constitutes an exception to the rule. Rather do I prefer for the present 
to assume that this genus has but six cervicals for the reason that the 
thorax long since shifted forward and took unto itself the seventh cervi- 
cal. In fact it seems that the dugong may even now be undergoing this 
process, for occasionally if not invariably there is a pair of rudimentary 
ribs attached to the seventh cervical. This brings to mind the possibility 
that the stimuli for a short neck in the Sirenia may have been stronger 
than all facts now indicate, but that this order has inherently lacked 
the ability to respond to them as readily as have the Cetacea. According 
to the literature there is some question whether Hydrodanialis had six or 
seven cervicals. 



Taylor (1914) stated that the neck of the sea otter is two-tenths of 
the trunk length, while it is three-tenths in the river otter. With this 
exception it seems that all aquatic mammals other than the Cetacea, 
Sirenia and Pinnipedia are either insufficiently specialized to exhibit 
alteration in the length of the neck, or else they have no terrestrial rela- 
tives close enough for significant comparison. 

It is from the Pinnipedia and Cetacea that we can tell most about the 
effect which speedy aquatic locomotion has had upon the neck, and they 
will be discussed at considerable length. They are amenable to two 
sets of fundamentally different influences imposed by widely diverse 
swimming methods. The Phocidae or true seals together with the Ceta- 
cea comprise one group, and in this really belong the Sirenia as well. 
The other group consists of the Otariidae of fur seals and sea-lions. The 
walrus {Odobenus) occupies an intermediate position in some respects. 
The differences in these two methods of swimming which are of concern 
to us in the present instance is that in the first group the locomotive im- 
pulse is purely from the rear, while in the second it is from the anterior 
thoracic region. 

The cervical complex of the Otariidae is long and sinuous, and the 
function which it, with the head, plays is two-fold. The animal when on 
land travels by a shuffling gallop, a gait made necessary by the shortness 
of its limbs. This is accomplished only by the expenditure of much 
effort and is greatly facilitated by the violent swinging backward and for- 
ward of the neck. Especially is this so in the case of large bulls with 
huge mass of cervical tissue. This swinging of the neck is an absolute 
requisite for their terrestrial mode of locomotion. If the neck were but 
half its length it is likely that travel upon the land would be so difficult 
that they could do little but wriggle onto a rocky ledge and roll off 

In swimming, the impulses are from the side in the case of Otariidae, 
with the center of motion presumably between the two organs for pro- 
pulsion and near the midthorax. Conditions are thus very similar to 
the case of a rowboat. A short, tubby skiff is difficult to handle, and 
one of usual length is best managed when the rower is near the center. 
Thus, for mechanical reasons, the sea-lion should have considerable 
mass of body both before and behind the anterior flippers, and this has 
undoubtedly constituted the chief stimulus for length of neck in this 
animal. Whether this has been of any influence in the acquisition by 
adult bulls of their enormous mass of cervical tissue seems doubtful, 
but must be considered as a possibility, in which case there would have 



to be present in the female a sex hormone with action of inhibiting in 
this sex the developing of extraordinary neck size ; or vice versa. 

This "bull-neck" character in male otariids increases to a quite phe- 
nomenal degree at the beginning of the breeding season, when they spend 
more time on land. In fact, bull fur seals and Steller sea-lions, at least, 
then spend many weeks on land without feeding or entering the water. 
Unfortunately those in a position to do so have never reported the con- 
ditions involved. Inferentially there is much fat deposited in the cervi- 
cal region, this being increased as a reserve supply when the breeding 
period approaches. O. J. Murie has told me that at the time of rut in 
caribou (Rangifer) there is a definite thickening of certain of the cervical 
muscles, which would be useful in the battles between males. This is 
probably caused by a hormone released by the awakening sex-glands 
and is entirely comparable with the lengthening of certain perineal 
muscle fibers in females at the imminent approach of parturition. I deem 
it likely that a similar thickening of cervical muscles of bull sea-lions 
takes place at the beginning of the mating season. 

In young males and females the neck is very sinuous with remarkable 
precision of movement, as all who have watched "trained seals" (usually 
Zalophus calijornJaniis) will agree. This is inherent and assuredly of 
fundamental importance to the economy of these mammals. It is not 
merely developed by training, as indicated by the statement of Rowley 
(1929) that when a stone is thrown at a sea-lion cow, "no matter how 
violently nor how short the range, she will catch the stone with marvelous 
accuracy in her mouth, often at the expense of breaking off her teeth". 
Because of the shape and the fact that the propulsive mechanism is situ- 
ated near the center of the mass the neck may act importantly as a rudder, 
and yet it may be thrust in all directions for the capture of food without 
disturbing equilibrium provided that at the same time the hinder end 
is moved in proper compensation. If the aquatic stimulus were for a 
shorter neck, however, terrestrial activity would be curtailed, and this 
would result in still greater independence of the land. So the fact that 
the long neck which is advantageous for swimming happens also to 
help terrestrial locomotion should result in a tendency to slow up the 
rate of aquatic specialization in other directions. 

As compared with the seal the cervical transverse processes of the sea- 
lion are broader, following greater complexity of the longus colli. The 
spinous processes are also better developed. The most striking additional 
muscular differences are that the cephalohumeral and humerotrapezius 
are better situated for twisting lateral movements of the neck and the 



occipital muscles are evenly distributed along the lambdoidal area to 
facilitate diversity of head movements. 

In the case of an aquatic mammal as highly specialized as the seal or 
whale, in which the propelling mechanism is situated at the hinder end, 
an entirely different set of physical laws is introduced. For proper effi- 
ciency it is obliged to be of fusiform shape, just as must an airship of 
the Zeppelin type, the fuselage of an airplane, or (with proper camber) 
the cross section of its wing. We are entirely justified in accepting it 
as an incontrovertible fact that with this type of body and of propulsion 
a long, mobile neck would be impracticable, and that the neck form or 
its musculature or both must be such that when swimming at speed, 
this part of the animal is included within the uninterrupted contour of the 
fusiform body. There must be a point constituting the center of equi- 
librium, which may also be the pivot of motion, situated anterior to the 
middle of the body and theoretically this should be at that portion of 
the thorax having the greatest circumference, which will usually fall 
at the shoulders. In the Phocidae, especially, the pivotal point of swim- 
ming motion is fixed in this region by the fact that here are anchored 
the lateral muscles comprising the power arm of the tail in one direction, 
and of the neck and head in the other. Now for most effective results 
for the muscular power expended it is absolutely essential that there be 
sufficient mass to the prethoracic part, while if the mass be too great, 
then efficiency is reduced. 

Breder (1926) considered that in most fish the pivotal point lies 
through the atlas, and this may very well be correct. As discussed in 
a previous chapter it is deemed that there may be a single pivot of motion 
coinciding with the center of equilibrium, or in an animal in whose swim- 
ming less of the tail is involved there may be two pivots of motion, one 
posterior and one anterior to the center of equilibrium. In the seal, 
however, with its relatively small head, it seems that the pivot of motion 
(or pair of pivots) is situated in the region from which the lateral 
lumbo-caudal and cervico-cephalic impulses arise. Some of these muscles 
may shift their points of anchorage to the thorax over the distance of 
one, two, or three intercostal spaces, or those to the arm may wander 
slightly, but on the whole their possible migration is very limited. 

In general principles the body of the seal (Phocidae) is subjected 
while swimming to the same laws encountered by whales. We might 
therefore expect to find in the former the same marked tendency toward 
a shortening of the cervical series as characterizes the latter group. On 
the contrary, however, it may seem somewhat surprising to find that in 

[ 143 ] 


bony details the neck of the seal is little or no shorter than that of the sea- 
lion, which as far as we can see has no stimulus whatever for a shortened 
neck. In spite of the latter feature the seal is apparently enabled to 
meet the condition imposed by its method of swimming that the mass 
anterior to the thorax be not too great in volume (for proper efficiency) 
by the fact that its head is relatively small. We cannot be sure, of course, 
that the length of its neck is absolutely ideal for swimming. In fact 
it seems likely that there is present some indeterminate amount of stimu- 
lus for a shorter neck because in swimming or in resting posture on land 
the neck is retracted to a marked extent. That this is so is indicated by 
the fact that if a fish be held above the head of a seal in captivity it will 
stretch the neck to a phenomenal degree, when it appears fully as long 
as in the sea-lion. It thus seems probable that in antagonism to the 
stimulus for a somewhat shorter neck during swimming, there may also 
be some stimulus, connected with the acquisition of food, for the reten- 
tion of a moderately long neck. Incidentally the Phocidae may be as 
yet insufficiently specialized for either one of these to have gained de- 
cided ascendency over the other. 

Examination of an embalmed specimen does not throw any light on 
the manner in which this extensibility of the neck is made mechanically 
possible, but it is doubtless due in part to unusual elasticity of the in- 
tervertebral cartilages. The apparent retraction of the neck while swim- 
ming is partly real (to as great an extent as the vertebrae will allow) and 
partly illusory, both because of the unusual breadth through the base 
of the neck and probably because tension of the lateral neck muscles 
tends to draw the shoulders forward. 

It has been noted in the previous chapter that the muscles of the phocid 
occiput are so distributed as to facilitate movements in the sagittal and 
in the horizontal plane. Other muscles of the neck follow the same 
plan, and are distributed so as theoretically to pull the head up or to 
the side with less effort than in the sea-lion, and with a minimum of 
diagonal twisting. If we analyze the swimming movements of the seal 
it will be seen that practically the entire musculature of either side is 
concerned. Neither the forward nor the hinder end can be thrown 
to the side without curving the entire body ; hence the muscles concerned 
in both head-swing and tail-, or pedal-swing really constitute one single 
group, all the components of which have been specialized toward the 
single end of efficient aquatic propulsion. Some operate from the an- 
terior thorax, while others — especially the deep and abdominal pectorals 
on the one hand and the inferior atlantoscapular, cephalohumeral and hu- 



merotrapezoid on the other — work in either direction from the anterior 
hmb, so that in such motions the arm acts as a sort of raphe between 
these two groups of muscles. These muscles concerned with lateral 
neck movements are very robust and diverge from the head to the sides 
of the broadened thoracic region in a manner to give them very powerful 
leverage. They must certainly have an extraordinary amount of normal 
tone, so that when those of one side contract the muscles upon the op- 
posite side will not relax to a greater extent than is proper. And yet, 
when the animal wishes to stretch forth its head, the tone is removed 
and the muscles are relaxed to an abnormal degree. The same may be 
said of the muscles chiefly concerned with movement of the head in 
the sagittal plane — the anterior rhomboid and semispinalis for raising, 
and sternohyoid and tracheovertebral muscles for lowering it. They 
are powerful and well situated for performing the work which they 
have to do, which evidently consists chiefly in strong but short depression 
or elevation of the head to facilitate steering. 

As before mentioned it seems certain that because the head of the seal 
is rather small, instead of relatively large, for the size of the body, the 
lateral movements of the neck which accompany the act of swimming 
are distributed over the entire cervical series of vertebrae, and they ac- 
cordingly would not be expected to depart widely from the usual type. 
Really the only noteworthy detail is the fact that the spinous processes 
are reduced and very much smaller than in the sea-lion, which seems an 
indication that active strength (as contrasted with passive strength) in 
elevating the head is not of great importance to the animal. 

The only ways in which the principle of aquatic locomotion of whales 
differs from that in seals is that in the former the flattened tail consti- 
tutes the primary organ for propulsion instead of the hind feet, and that 
its plane of movement is vertical instead of horizontal. This means 
that in the Cetacea the dorsal and ventral neck muscles are required to 
perfom those acts during swimming which constitute the chief function 
of the lateral neck muscles of the seal. But the lateral neck muscles are 
attached to the thorax or arm and belong in entirely diiferent groups from 
those which swing the tail laterally, while the dorsal neck muscles of chief 
use are an integral part of the spinal musculature of the entire animal. 
This permits the fundamental differences that in the seal the center of 
swimming motion must be in the anterior thorax, from which the neck 
muscles arise, while in the whale there is no muscular inhibition to its 
shifting either forward or backward, according to requirements intro- 
duced by external body form. 

[ 145} 


The whale has no external neck, or at least there is no constriction at 
this point, save to a barely discernible extent in Platanista and Delphinap- 
terus, in both of which the series of cervical vertebrae is relatively longer 
than usual in this order. This is for the reason that the cervicals have 
become much shortened in all Cetacea, but in some more than others, and 
there is variation in other respects, this being best illustrated perhaps in 
the Odontoceti. The individual vertebrae may be distinct with the cen- 
tra either fairly thick or wafer thin, all may be fused into one bony com- 
plex, or intermediate conditions may obtain, save that invariably, with 
the exception of the cachalot, any fusion which occurs is first manifested 
anteriorly. It is probable that this series is shortest in those sorts in which 
all the elements are fused. Among odontocetes it is likely that Pla- 
tanista has the longest neck of any genus now living, but the skeletons 
available are disarticulated and unsatisfactory for determining this point. 
The neck of Delphinapterus is also relatively long, as mentioned. Among 
mysticetes the longest neck occurs in the gray whale (Rhachianectes), 
which is famous for having this member unusually mobile, and thereby 
hoisting whaleboats several feet above the water by a thrust of the head. 

It is difficult to determine the exact percentage of the cervical series 
to body or even total skeletal length. The cervical length may be taken 
with ease for the elements are either fused or were in hfe separated by an 
insignificant amount of intervertebral substance; but this dimension is 
almost invariably omitted from existing osteological reports, or else one 
does not know if the intervertebral disks of the remainder of the skele- 
ton were allowed for. In Neomeris I found that the cervical length 
constituted about four and one-half per cent of the skeletal length, but 
the specimens were disarticulated so that this percentage is greater than 
is actually the case in life. From rather rough measurements of the 
mounted baleen whales in the National Museum, having artificial in- 
tervertebral disks of an unknown degree of error, I found that the per- 
centage of cervical length to skeletal length was in Rhachianectes, 4.2 ; 
Sibbaldus musculus, 3.7; Megaptera nodosa. 3.6; Balaenoptera physalus, 
3.4; and in Eubalaena, with its fused cervicals, but 2.4. The latter figure 
also obtains in Physeter, according to measurements furnished by Doctor 
Stone of the specimen in the Philadelphia Academy of Sciences. 

Shortening of the cetacean neck has always awakened the liveliest in- 
terest and has caused much speculation. The most popular belief may be 
summed up in the words of Winge (1921) : "The head, during swim- 
ming is held directed as firmly as possible forward. The neck is not 
moved, and for this reason it becomes short and stiff. During motion 



through the water the head is pressed from the front ; it is forced back- 
ward against the cervical vertebrae, which are thereby squeezed exces- 
sively together and pressed back against the anterior dorsal vertebrae." 

The thesis that water pressure against the head has reduced the length 
of the neck by squeezing it between the head and the thorax hardly seems 
sound. In the first place it is not alone the neck that has experienced the 
sort of squeezing meant by Winge, but all vertebrae anterior to the flukes, 
and the skull as well. In the second, if such a mechanical squeezing 
would have had any effect this should be just the reverse of a shortening 
of the cervical complex. According to modern concepts of morphology 
this should rather have consisted of a narrowing of the skull and elonga- 
tion of the neck. The whale has of necessity assumed a short neck in 
order to fulfill certain mechanical requirements contributing to effi- 
ciency in swimming and I seriously doubt whether water pressure, in the 
sense referred to above, has entered into the question at all. 

Almost if not quite invariably in cetacean literature the shortness of 
the neck is discussed as though this detail had in itself suffered reduc- 
tion without reference to the thoracic elements, and this has resulted in 
much misunderstanding. The cervical vertebrae have not alone been re- 
duced, but all the vertebrae anterior to just forward of the center of the 
thorax. In many mammals, including man and cetaceans, the lumbar 
vertebrae are the most robust of the column. In man the centra of the 
anterior thoracic vertebrae are progressively thinner and this, to an al- 
most insensible degree, may be said to continue to, and culminate in, 
the second cervical. In whales the reduction in centra length is more 
marked anteriorly. The exact number of vertebrae involved is variable, 
but perhaps in most sorts the fifth thoracic is the first to show really ap- 
preciable shortening. The fourth is slightly shorter than the fifth, the 
third than the fourth, and so on. By the same slight degree the seventh 
cervical is slightly shorter than the first thoracic and this shortening may 
culminate in the third or fourth cervical in such whales as have these 
vertebrae unfused. This is but a general statement and details fre- 
quently differ, however. In the zeuglodont Basilosaurus the neck is 
clearly shortened, but the transition between the cervical and thoracic 
series is not gradual. On the contrary, the last cervicals are somewhat 
abruptly shorter than the first thoracics, and this might be an indication 
that the cervicals are more responsive to the stimulus for shortening than 
the thoracic vertebrae; but I believe that rather it is the result of its 
anguilliform conformation. In some porpoises, such as Tursiops, al- 
though the anterior thoracic vertebrae are considerably shorter than the 



middle or posterior ones, they are quite a bit thicker than the last cervical. 
This may be explained, I believe, by the probability that the posterior 
cervicals have become as thin as possible without fusion, and thinner 
than any rib-bearing vertebra could become. 

Not only is the skull relatively larger in all whales than in the seal, 
but almost invariably that of the former has accessory equipment in the 
way of special fatty deposits or baleen armature, which makes the head 
still larger in comparison to body mass. The head alone therefore fur- 
nishes almost or quite the prethoracic mass and weight requisite to high 
swimming efficiency. In other words the head alone is so large that if 
there were a neck of respectable length the part of the animal anterior 
to the thoracic "pivot" of motion in swimming would either be too mas- 
sive for the part posterior thereto, or else, if the pivot could be shifted 
forward, and therefore from the thorax to the skull, the neck would 
really constitute a mechanical part of the body. 

In scrutinizing various cranial and cervical conditions there seem to be 
recognizable two possible factors which might show a tendency for re- 
tarding maximum shortening of the neck and fusion of its elements. One 
of these is the condition imposed by the presence of a tweezer-like beak, 
or of the comparable tusk of the narwhal. It is readily seen that this 
equipment would necessitate at least moderate shifting about of the head 
with a consequent amount of flexibility of the neck, and this should delay 
fusion of the vertebrae. At any rate the narwhal has all the cervical ver- 
tebrae free, as have such long-beaked forms as Platanista and the extinct 
Eurhinodelphis. The other is the fact that in those forms with long or 
slender beaks the cranial part of the skull is relatively small in relation to 
body size. Hence the head has less mass and this might be compensated 
for by a slightly longer neck. But non-beaked odontocetes such as the 
beluga (Delphwapterus) may also have the cervicals free, and this fact 
introduces an element of serious doubt. 

At the other extreme are such odontocetes with head of moderate ceta- 
cean size as Hyperoodon, Ziphius, Mesoplodon. Grampus, and at times 
Fseudorca and Phocaena, in which all the cervicals are fused into one 
solid bone. Still other sorts exhibit every intermediate condition between 
complete fusion and complete freedom of the cervical elements, but in 
all existing Delphinidae excepting the narwhal and beluga at least the 
atlas and axis are fused. At one time I endeavored to reach some con- 
clusion regarding this point by making a list of cervical conditions in all 
species of whales which I could examine, and attempting to correlate 
them with any one of numerous other osteological features ; but with re- 


suits that were utterly discouraging. It may be put down partly to the 
variation in the cervical needs of different sorts, and another possible ex- 
planation is furnished by the fact that a very few species (as Neomeris) 
develop a slender, short pair of accessory ribs attached to the seventh 
cervical vertebra. It thus seems possible that this porpoise has found 
it less difficult to begin a shifting forward of the thorax than further to 
reduce the length of the cervical series as a whole. 

Certain additional reasons for cervical variation in the larger whales 
may also be considered, bearing in mind the thesis already explained 
that in whales with heads of even moderate size the head alone furnishes 
practically all of the prethoracic mass necessary for efficient locomotion ; 

Figure 21. Extremes of cervical vertebrae among toothed whales, illustrating 
a series in which the seven components are separate, and one in which they 
are all fused. Delphinapterus on left, Grampus on right. 

and this matter of head size is really more pertinent to the present chap- 
ter than the last. 

One is so used to reading of the prodigeous size of the head in balaenid 
whales that he is apt to take for granted that their heads are also of unu- 
sual length. Because of the question of the intervertebral disks it is 
difficult to find the precise proportions, but I have secured what measure- 
ments I could from the literature, and these indicate not only that the 
skull of Megaptera (25 to 31.4 per cent of the total length) may average 
longer than in the balaenid whales (26.8 to 28.8 per cent) , but that there 
is more individual than generic difference in Mysticeti. Rhachianectes 
(22) and Balaenoptera acuto-fostrdta (22.6 per cent) both have rather 
short skulls and the others are intermediate. But there are important 
differences elsewhere than in length of skull. In Balaenoptera, Sibbal- 



dus, Megaptera, and Rhachiauectes the whole head is either flattened 
dorso-ventrally or moderately in transverse direction (Rhachiauectes). 
In these the vertebrae are either all free, or, according to Andrews 
(1916) "two or more of the cervical vertebrae usually become ankylo- 
sed," although I personally have seldom noted this condition. This 
lack of complete ankylosis in mysticetes whose heads average larger than 
in odontocetes may be attributed to the common belief that the former 
group is not as far advanced along the evolutional road or as "special- 
ized" (ambiguous term!) as the latter, and consequently that the stimu- 
lus for fusion has not been operative for as long a time. In balaenid 
whales (Balaena and Ejibalaena), however, although the skull is no 
longer, the head is of prodigeous depth. This increase in total size and 
mass of head, necessitated by the hypertrophical development of the 
baleen equipment, has apparently obliged these whales to adopt a make- 
shift for the accomplishment of swimming. This is even more the case 
in Physeter, whose great spermaceti organ has enlarged the head to a 
degree where the snout projects several feet beyond the rostrum. In these 
it would seem that if the pivot of motion occupied its more natural posi- 
tion in the anterior thorax, the increase in the relative size of the head 
would tend to place this too far to the rear — or too near to the lineal cen- 
ter of the animal — for best efficiency. Whereas in the usual fusiform 
method of swimming at high speed both the tail and head are curved 
from some point in the anterior thorax, a disproportionately large head 
would disturb the proper balance, and it is suggested, partly by inference 
and partly by the shift forward of the greatest circumference of the body, 
that in such whales the center of motion has, to as large an extent as it 
was able, migrated from the thorax to the posterior part of the skull. 
Theoretically this is rendered possible by the increase in the inertia of 
the head and the essentially homogeneous nature of the dorsal cervical 
and the dorsal thoracic musculature. But this condition is in principle 
less eflficient and ha§ resulted in the definite reduction of speed. The 
stimulus for excessive head size has simply proved stronger than any 
forces acting in antagonism and under the conditions which the animals 
have encountered it has not been too great a handicap for continued exis- 

Thus the neck, body and tail of these whales seem to all intents to 
comprise the lever arm for swimming and the head can more properly be 
compared to a fulcrum (as in 4, figure 1). It might thus be said that 
while in porpoises the body wags the head, in the sperm and right whales 
the tendency is for the head to wag the body. The neck is functionally 


a part of the body and as a fused complex virtually constitutes a thoracic 
vertebra without ribs. The series has probably become as short as it can 
while at the same time accommodating essential muscles which are at- 
tached at this point, as well as allowing for the emergence through fora- 
mina of the cervical nerves. The stimulus has been for the thorax to en- 
gulf the neck and to this extent the latter has experienced squeezing, but 
from a morphological or evolutional rather than a mechanical sense. Al- 
though there is no reason for considering the balaenid as any older than 
the balaenopterid stock the cervicals in the former group have fused more 
completely probably because the larger mass of the head has increased 
the strength of the stimulus for fusion. Normally in the former they 
are completely fused but Flower (1876) has stated that occasionally the 
seventh is free. The cervicals of the cachalot (Physeter) are also usu- 
ally fused, I believe, but this is not established for the reason that in 
some of the few specimens known the atlas is free ; and this is a unique 
situation among Cetacea, for in all other whales fusion of the cervical 
elements first takes place anteriorly and then progresses toward the rear. 
In those specimens of the manati which show fusion of two or three 
cervicals the atlas is said always to remain free, however. 

In the above discussion account has not been taken of the pygmy 
right whale (Neobalaena) or pigmy sperm whale (Kogia) for the rea- 
son that these are depauperate, freakish forms. The former is anomalous 
in that the first thoracic vertebra lacks ribs. In other words, it has eight 
cervicals, which indicates an exactly opposite trend from that exhibited 
by all other whales. This circumstance is very puzzling, but it is felt that 
one exception does not overthrow the arguments that have been ad- 

Before closing this chapter brief mention should be made of other cer- 
vical muscles not yet discussed. During swimming the majority of the 
subcervical muscles of whales act in antagonism to the supracervical ones, 
but many of them have other functions not connected with locomotion. 
Thus the hyoid complex is very powerful in odontocetes and these doubt- 
less assist in depressing the head, but whether this is now their primary 
stimulus cannot be told. They must also be used in the deglutition of 
bulky food, but this is hardly to be reckoned with in Mysticeti, especially 
the balaenopterids, because their esophagus is relatively so small, but in 
this group the depression of the head can probably be largely brought 
about by flexion of the superficial gular musculature (previously dis- 
cussed) operating upon the mandibular tip. The hyoideal as well as 
other small muscles of the prevertebral and lateral parts of the neck are 



more diverse in their affinities than are the cervical elements of the spinal 
system and they could never become as simplified as the latter. 

To recapitulate briefly; it is believed that the long, flexible neck of 
the otariid is essential to its special method of swimming. The dif- 
ferent method of swimming employed by the phocid and cetaceans ne- 
cessitates a different cervical conformation. Although feeding habits of 
the seal may require a neck of considerable length, its effective length 
while swimming is reduced by the small size of the head and great 
breadth of the neck base, as well as by as much contraction as possible. 
The larger size of the head in whales requires marked reduction in neck 
length, in the porpoise so as to bring the head nearer the thorax, and pos- 
sibly in balaenid and sperm whales so as to bring the thorax nearer the 
head and mechanically to make of the cervical series a single element 
comparable to a thoracic vertebra without ribs. 


Chapter Eight 

The Trunk 

There are numerous factors affecting the precise conformation of the 
mammahan thorax, but it is invariably of such shape that only very minor 
adjustments are necessary in order to bring it to an ideal stream-line form. 
Perhaps in the majority of quadrupeds the broadest or deepest part of 
the body contour is just back of the anterior limbs, although this point 
not infrequently is situated in the posterior thorax, or even through the 
abdomen, in the more paunchy sorts. But no mammalian thorax is so 
unsuitable in form that it will offer any real resistance in swimming and 
in consequence there is no very strong stimulus for a definite change in 
its shape even when sufficient time has elapsed for there to be very ad- 
vanced aquatic specialization in other respects. I deem the Pinnipedia, 
Sirenia and Cetacea to be the only aquatic mammals to have reached this 
point. It is commonly accepted as fact that the great thoracic diameter 
of Hippopotamus is a result of aquatic habits. I doubt this and although 
I may be entirely mistaken I believe that the primary reason that the 
thorax is larger in this animal than the rhinoceros, for instance, is that 
the more succulent food of the former necessitates an alimentary tract of 
prodigious capacity, with stomach more than a dozen feet in length. 

We can be sure, however, that in pinnipeds, sirenians, and cetaceans, 
the aquatic life has been lived for sufficient time for there to have been 
modifications in the external form of the trunk — especially in the first 
and last orders, which are more speedy. The requirements for external 
conformation are simple and necessitate merely that the greatest girth be 
approximately at the pivot of motion for swimming. In pinnipeds, 
sirenians and most cetaceans this is in the region of the anterior thorax, 
a possible exception being in the case of adult bull sea-lions, where the 
greatest girth may be at the base of the neck. In balaenid and sperm 
whales, where it has been argued that because of the huge size of the head 
the pivot of swimming motion has tended to shift forward, the greatest 
girth seems to be through the posterior part of the head, as would be 
expected. For practical purposes, apparently, this is all that is neces- 
sary, for it would make very little difference whether the cross section 
of such a body be slightly flattened in the vertical or horizontal plane. 
We can be sure, however, that if the aquatic influence were uncompli- 



cated by any muscle pull or lifting action of the lungs the tendency would 
be for the cross section of the anterior trunk eventually to assume a 
shape that was exactly circular, for this is the ideal both for locomotion 
and for such purposes as retention of body heat, as well as that external 
pressure may be distributed evenly to all the internal organs. 

The point of greatest bodily circumference may be, and is, shifted ac- 
cording to aquatic requirements without especial reference to the form 
of the thoracic cavity itself, as fleshy or fatty tissue may easily be de- 
posited in the required region. The actual external shape of the an- 
terior trunk may be influenced by deposits of fat, by the conformation of 
the shoulders and their muscles, by the height of the vertebral spines and 
consequent thickness of the spinal musculature, and by the form of the 
thoracic cavity. 

In the fetal state the shape of the cross section of the thoracic cavity, 
in all mammals without exception, I believe, is slightly broader trans- 
versely than in a sagittal direction. In terrestrial sorts its shape in after 
life will alter according to posture, the fundamental forces concerned 
being gravity and muscle pull. In a strictly quadrupedal form, and even 
such a primate as the baboon, gravity will tend to pull down the ster- 
num. This is doubtless assisted in some obscure way by muscular stress, 
and by the requirement in swifter, larger mammals that the anterior 
limbs be as near together as practicable to reduce any propensity to wad- 
dle. The result will be a thorax narrow transversely. That muscle pull 
need not be of great influence in the attainment of this result is indica- 
ted by the situation in the existing sloths, which spend their lives either 
hanging by their limbs or curled up in sleep. We might expect to find 
that the constriction of the chest between the pendent arms had resulted 
in its becoming narrower but on the contrary it is much broader than 
thick, indicating that muscle pull has not been an important factor in 
the expected direction, but that the chief influence was the removal of 
ventral, gravitational pull, which has permitted the ribs to spring out- 
ward. In mammals of erect posture, as man, there is also no gravita- 
tional pull in a ventral direction upon the sternum, but here the throw- 
ing back of the shoulders (ostensibly) has permitted a broadening, usu- 
ally, of the thorax chiefly by tension of the muscles extending from the 
arm to the midline, both dorsally and ventrally. Occasionally, in man, 
a "pigeon-breasted" individual is encountered, in which the thorax is al- 
most circular, so other and obscure factors are indicated. Or there may 
be some such unusual situation as is found in bats, which require a broad 
chest to accommodate a huge mass of pectoral musculature. 


In the case of mammals which are exclusively aquatic gravitation does 
not enter the question. It is true that at least most whales are slightly 
heavier than the water which they displace and that for equalized flota- 
tion they need a lung-full of air. Hence there is some slight force of 
gravity experienced ; but when distributed over the entire animal it is 
absolutely negligible. In consequence the conformational stimuli ex- 
perienced by the thorax are those supplied by muscle pull, levitation of 
the lungs within the cavity, such amount of tendency for a circular trunk 
as progression through the water may determine, the influence which the 
location of heavier bones may introduce, and at times a possible fifth 
factor. In regard to the latter, certain odontocetes indulge in a rocking 
motion during progression. The killer whale especially is in the habit 
of swimming slowly and taking a fresh breath every few seconds. Ac- 
cordingly the head is first elevated above and then depressed below the 
surface, and the back rocks to and fro in the sagittal plane. Undoubtedly 
this has had some eff^ect, however slight, upon the conformiation of the 
entire animal. But it is impossible to know the relative importance of 
any of these influences. 

It is popularly believed that aquatic life augments the lung capacity 
and in consequence that there is an increase in the diameter of the thorax. 
This is entirely logical and may be conceded. Certainly the thorax of the 
sea otter seems to be considerably more capacious than in its river cousin. 
But usually there is no sure yardstick whereby this can be measured and 
it must be inferred. For one thing there are no thoroughly aquatic 
mammals sufficiently close to terrestrial forms for adequate comparison; 
and an expansion of the thorax may merely indicate that there has been 
a shortening of the abdomen, or an increase in the size and capacity of 
the alimentary tract may have crowded the thoracic cavity. 

Unfortunately I have neglected to examine the cross section of the 
chest of any fresh pinniped or cetacean, and a preserved specimen soon 
becomes so distorted by its own weight it is valueless for this pur 
pose. Nor can great reliance be placed in a mounted specimen. 

Among the smaller aquatic mammals the only one (so far as I know) 
having details of the thorax that merit attention in the present connection 
is the insectivore otter (Potomogale). Dobson (1882) showed that in 
this animal the pectoral muscles are surprizingly unlike the general pat- 
tern that is so characteristic of insectivores. Rather is it suggestive of 
conditions in the sea-lion. I believe, however, that this resemblance is 
fortuitous and that there is shown merely an intermediate stage in even- 
tual convergence toward a similarity in appearance to pectoral conditions 



in the seals (Phocidae) . In certain respects the swimming of Potomogale 
and the Phocidae is similar. In neither are the fore limbs used in swim- 
ming, and in both the hinder end is swung from side to side. Therefore, 
as discussed more fully in the case of the seal, there is a tendency for 
the posterior part of the pectoralis and the latissimus dorsi to develop 
toward the end that they may assist in pulling the posterior part to the 
side. In Potomogale there is a clear tendency in this direction, and the 
two muscles are continuous along their borders. The anterior part of 
the pectoralis unites over the arm with the inferior margin of the trape- 
zius, insertion being upon the humerus distad of the greater tuberosity. 
By means of these two muscles, at least, the arm may easily be held sta- 
tionery, acting as a base from which other muscles may operate. De- 
serving of mention in this connection is also the sternocleidomastoid, 
which is apparently of much importance in lateral movements of the head 
that are requisite to swimming. This muscle arises from the midventral 
line for a surprising distance beneath the pectoral mass. 

In mounted skeletons of seals the anterior thorax is usually slightly 
broader than high, or it may be markedly broader (P. groenlandica) ; in 
sea-lions it is definitely narrowed ; and in the walrus fairly intermediate 
between the two. These facts need not be of great significance in the 
present connection, however, for there is equal variation in terrestrial 
carnivores. At any rate this difference in pinnipeds is accentuated, in 
the case of preserved specimens, to the point where a seal can hardly 
be balanced upon its side, while one has difficulty in so balancing a sea- 
lion upon its sternum. The sea-lion spends considerable time on land, 
supported partly by the anterior limbs. The thorax thus experiences a 
ventral gravitational pull for at least some of the time, and during this 
time there is the same static muscle stress as acts upon the usual terres- 
trial quadruped. High spinous processes and accompanying muscula- 
ture in this region further accentuate the depth of the trunk as compared 
to its width. What effect the levitating influence of the inflated lungs or 
of the swimming actions of the anterior limbs have had while the animal 
is in the water we do not know, although it is not improbable that the 
constant adductive movements of the flippers have tended to make the 
chest narrower. 

At least most seals of the genus Phoca never use the anterior limb 
on land for supporting the body save for the briefest periods, although 
some others of the Phocidae, as Mirounga, habitually do so. Further- 
more, gravitational force can very rarely act upon the sternum. On the 
contrary the sternum presses upon the ground both when the animal is 


resting, and wriggling along the surface, so there is at such times actually 
as much force pressing iipivard against the strenum as downward. Pre- 
sumably these factors have been of importance in shaping the thoracic 
cavity. There are other elements, however, which further modify the 
shape of the trunk. Because the spinal musculature has become greatly 
broadened for the purpose of lateral movements of the hinder end and 
because there is no need for much movement in the sagittal plane, the 
spinous processes of the anterior thorax are very low and the musculature 
concerned very thin. This further reduces the sagittal dimension of the 
trunk. The transverse dimension is increased by the fact that the lateral 
muscles of both trunk and neck need as great a lever arm as possible 
more effectively to accomplish the lateral movements used in swimming. 
The muscles not only are very robust but they are overlain by a thick layer 
of fat, and between them and in all interstices there is much connective 
tissue surrounding networks of blood vessels. The broad thorax is of 
distinct advantage to the animal in swimming. Whether flattening in 
the sagittal plane actually increases the swimming ability of the seal or 
whether, not of advantage, it is yet of insufficient disadvantage to be of 
consequence, is unknown. 

In comparison with a typical terrestrial carnivore the chief character- 
istics of the vertebral column of the Pinnipedia are the looseness of the 
articulations, the elasticity of the intervertebral disks, the latter being 
difficult to investigate after death, and the fact that the spinous processes 
exhibit no definite change in slope, or anticline. The latter is a character 
shared by all essentially aquatic mammals. The usual quadruped has a 
definite center of motion in the vertebral column, which it would be more 
logical to call the center of suspension for the reason that the back-bone 
of a quadruped may with propriety be compared to a double-pier, canti- 
lever bridge. From the above center the anterior spinal and some of the 
shoulder muscles operate in one direction and the posterior spinal and 
some of the pelvic muscles in the other, resulting in a backward slope 
of the spines in the anterior thorax and their forward inclination in the 
extreme posterior thorax and lumbar region. And there is usually a 
rather abrupt alteration in the character of the spines where the slope 
changes. The position of this center depends upon the stress encoun- 
tered, as pointed out by D'Arcy Thompson (1917) . If the chief weight 
is borne by the fore legs the center will be farther forward, and if the 
hind limbs bear all the burden, as in the kangaroo, then the center will 
be shifted far to the rear. In a completely aquatic mammal the skeleton 
has no resemblance to a double-pier, cantilever bridge, but may better be 




LATIS. D0T\5\ 

Figure 22. Dorsal musculature of a sea-lion {Zalophus): superficial layer upon 
the left, and much of the next deeper layer to the right of the medial line. 





^TL^^lToseKP. \Hft-TV. 

TR\CE.Ti LOUa.- 
Vhr\S DO"RS\ 1- 

v.KT\^ oo'v\s\ 2/^'; 

LONQr\?.. DC)-K5\ 


"RtCT. TEr'\ 



3LfA\TE.rAD 2,^ 

J >S\ .BICEPS FOA. 2. 
SErA\TtND. 1 

Figure 23. Dorsal musculature of a seal {Phoca): superficial layer upon the 

left, and much of the next deeper layer to the right of the medial line. 



likened to a pontoon bridge, for the body is supported in the water 
throughout its entire length, and in theory each individual vertebra bears 
its proportionate share of the load, although muscle stress introduces a 
disturbing element. The curve of the vertebral column is always rather 
gradual, but there must be some one point upon the arc from which the 
column curves in either direction. In the anterior thorax as well as the 
posterior neck the spinous processes of otariids are moderately high, 
which is chiefly for supplying an efficient lever arm for the muscles rais- 
ing the head and neck, but spinous height rapidly decreases toward the 
posterior thorax and in the lumbar region, indicating that the lumbar 
spinal musculature need not be very strong for such actions as the ani- 
mal finds necessary. In phocids, on the other hand, the thoracic spines 
are very low, with no better definition than the lateral processes. Their 
height is slightly greater in the lumbar series, however, and here the cen- 
tra are also more massive. 

As in all aquatic mammals the articulations of the pinniped vertebrae 
are reduced, for as stresses are applied while swimming to a large ex- 
tent throughout the columnar length rather than at particular points there 
is little need for local strength, while there is increased need for flexibility 
of the column. This is attained (especially in the Otariidae) by virtual 
abandonment of an interlocking type of zygapophyses, and at the same 
time reducing all other processes which might offer mechanical restric- 
tion to limberness. Anapophyses are absent as such in the lumbar series. 
In otariids (at least in Zalophus) the zygapophyses of one side are very 
close to those of the other and in the lumbar region the articulations are 
such as apparently to prevent any marked concavity in the outline of the 
dorsal surface, but permitting an unusual amount of convexity. In a 
seal (Phoca), relative to size, the zygapophyses of opposite sides are about 
three times as far apart, and in the lumbar region the articulations are 
such as to allow very definite concavity in the outline of the dorsal sur- 
face. Especially is the latter the case between the lumbar and sacral ser- 
ies, to the degree where the sacrum may be elevated to a quite surprising 
extent. And although the cleaned skeleton does not show it there is pos- 
sible a very remarkable amount of concavity in the posterior thoracic 
and anterior lumbar regions of at least some phocids, for photographs of 
Mirounga show that this animal can. bend the column to quite a right 
angle at this point. Convexity of the lumbar series cannot be determined 
from osteological examination, for this would depend mainly upon the 
amount of play allowed by the zygapophyseal articulations. Presumably 
the possible amount of convexity is not very great and somewhat less than 


in otariids, but intervertebral flexibility would render it easy for the seal 
to develop this ability did it have occasion for doing so. 

The number of ribs in the most generalized mammals is thirteen pairs, 
while in the Pinnipedia they number fifteen, except in the walrus which 
has fourteen pairs. This is accompanied in the former case by five, and 
in the latter, by six lumbar vertebrae. So it is seen that the members 
of this order have responded to a stimulus for a longer thorax, although 
it is conceivable that this result was attained before aquatic habits were 
adopted. Ten (usually) pairs of ribs are attached to the vertebrae by 
both capitulum and tuberculum, and this, I believe, is the situation in at 
least the majority of terrestrial carnivores. So in the latter detail there 
has apparently been no alteration. The ribs are not otherwise note 

In pinnipeds the sternum is usually composed of six or seven bony 
elements, while in terrestrial carnivores it averages eight or nine. So the 
sternum in the former group has suffered reduction, this being more ac- 
centuated in the Phocidae than the Otariidae because in the latter the in- 
dividual elements have a greater length. In all pinnipeds there is a pres- 
ternal extension of the manubrium, partly cartilaginous, which is mark- 
edly well developed, especially in the Phocidae, and this is of very definite 
importance to the animal. 

With all of the above bony details of the pinniped thorax in mind we 
can proceed to scrutinize some of the muscular stimuli involved. In 
otariids the cervical and anterior thoracic muscles are developed not only 
for agility of neck muscles for enabling the head and neck to act as a 
balancer both on land and in the water, but to assist in movements of 
the anterior limb in both situations. The spinal musculature retains to a 
large extent its regular function of bending the vertebral column in the 
sagittal plane. The lateral and ventral thoracic and the abdominal mus- 
cles are mostly subservient to the functions of pulling the anterior limb 
posteriorly and medially, and pulling the sacral region down and for- 
ward during terrestrial progression. The musculature controlling the an- 
terior limbs is discussed more fully under the latter heading, but it may 
here be noted that the absence of a clavicle in the Pinnipedia (as is usual 
in Carnivora), allows much freedom in movement of the shoulder, and 
the specialized functions involved may or may not have had some influ- 
ence upon the conformation of the anterior thorax. In both the sea-lions 
and seals the noteworthy forward extension of the manubrium, partly 
bony and partly cartilaginous, has undoubtedly been brought about in 
response to the need for more powerful action of the pectoral muscles 


TRKCnE£i.-\lE^T fAUSC 





Figure 24. Ventral musculature of a sea-lion {Zalophus): superficial layer upon 
the right, and much of the next deeper layer to the left of the medial line. 




SE-R-RAT. rAA&y 
RECTUS f\-BIiO(«\ 




Figure 25. Ventral musculature of a seal (Phoca): superficial layer upon the 
right and much of the next deeper layer to the left of the medial line. 


during adduction combined with extension forward of the limb. In the 
sea-lion this correspondingly affects the cephalohumeral, part of which 
arises from the anterior border of the pectoral. 

In the seal the panniculus carnosus muscle does not converge mark- 
edly to the arm pit, but the fibers have an even cranio-ventral inclination. 
This condition may have been purely an ancestral inheritance or may be 
useful in the wriggling motions accompanying terrestrial locomotion, 
but could hardly be of any help in swimming. Contrasted with this is 
the situation in the sea-lion, in which fibers of the postbrachial part of the 
panniculus all converge strongly to the arm pit. In this animal the sheet 
of muscle covers the knee and extends quite to the base of the tail, and 
as a result it is of great help in movement upon the land. Contraction 
of this panniculus assists in flexing the lumbar region so that the hind 
feet may be placed flat on the ground, and in galloping, contraction of 
the panniculus after extension of the anterior limbs helps in pulling for- 
ward the entire hinder end of the animal. In swimming the panniculus 
can act from the other end and lend power to backward thrusts of the 
fore flippers. This, combined with adductive motion, is also the func- 
tion of the posterior part of the pectoralis. 

In the seal lateral movements of the hinder end are prerequisite to 
swimming, and one would imagine that for it a panniculus of the sea- 
lion type would be very useful, but as already said this has not been de- 
veloped, possibly because any purely brachial stimulus for it has been 
lacking. It is logical however, to expect lateral movements, accom- 
plished mainly by the spinal musculature, to be markedly assisted by 
ventral muscles, and this is brought about by an extraordinary develop- 
ment of the pectoral muscles. Midventrally the latter extend from the 
sternum (deep pectoral part) practically to the pelvis, although muscle 
fibers do not occur quite so far caudad, while more laterally the abdom- 
inal pectoral virtually reaches the knee. This abdominal division is ex- 
tremely heavy and thick in its anterior portion. The whole postbrachial 
part of the pectoral thus can operate to pull the hinder end sidewise while 
the dorsal musculature counteracts a downward pull by the posterior pec- 
toral. But such flexion tends to pull the arm to the rear and this must 
be counteracted by antagonistic action of those lateral cervical muscles 
that are attached to the arm. Thus the arm operates mechanically as a 
sort of raphe, from which cervical muscles act in pulling the head to the 
side while at the same time the posterior pectoral is doing the same for 
the hinder end. 



As the above is the case one might expect to find a somewhat similar 
situation in the dorsal muscles, and this is so to a modified degree. In 
the seal the spinal musculature is the chief agent in lateral movements 
of the posterior end. This has phenomenal breadth, as well as consid- 
erable thickness, in the lumbar region. The sublumbar musculature is 
well, although not remarkably, developed, but the iliocostalis and longis- 
simus have expanded and where the latter is attached to the ilium this 
bone has turned laterad to provide greater accommodation for the mus- 
cular connection. As the iliocostal is the most lateral division of the 
erector spinae it naturally is of most consequence in the lateral move- 
ments employed for swimming. It might very well be expected to main- 
tain its robustness as far as the occiput, but this it does not do. As it con- 
tinues onto the thorax from the lumbar region it gradually thins and vir- 
tually disappears over the anterior thorax. The function of the iliocos- 
talis is therefore almost exclusively for operating the hinder end in swim- 
ming. The forward end of the animal must consequently be controlled 
by a different group of dorsal muscles. As with the more ventral an- 
terior muscles concerned in swimming, this is accomplished partly by 
muscles extending from the arm to the head or neck. The splenius, 
humerotrapezoid, cephalohumeral and both atlantoscapulars are better 
situated in the seal than the sea-lion for purely lateral motion of the 
head, and for antagonism the spinotrapezoid projects farther back in the 
former, while the latissimus is more extensive and is double. The con- 
formation of the latter indicates that it might be of distinct aid in side- 
wise curving of the hinder end, but it seems probable that the chief 
stimulus for its development was to act in antagonism to prebrachial mus- 

By the above statements it is not meant to imply that only those muscles 
mentioned are used by the seal in swimming. There are present ex- 
tremely heavy subvertebral cervical muscles which I suspect are used in 
antagonism to translate into lateral movement the action of the semi- 
spinalis capitis, which otherwise would largely result in raising the head. 
Similarly almost every muscle of the body and neck should have some use 
in swimming, but the actions of many of them in this connection are ob- 

The sirenian skeleton is remarkably heavy and dense, especially in the 
manati. This is popularly believed to be for the purpose of enabling 
these animals more readily to sink from the surface to their pastures upon 
the bottom. This may be the proper explanation but it should not be ac- 
cepted without considerable reservation. Sirenians descend to very mod- 



erate depths. Why should they be better fitted for doing so than whales, 
which descend to great depths? Also it would seem of at least equal 
importance that they should be able to rise to the surface with celerity. 
As a matter of fact they have doubtless adopted a middle course and 
their specific gravity is probably almost the same as that of the water 
which they displace. If this be the case then their bones are particularly 
dense to compensate for the unusual lightness of the rest of the body. 

As sirenians never leave the water the thoracic stimulus of fore leg 
support usually present in the terrestrial mammal is entirely lacking, as 
it is in cetaceans, with this difference, that in sirenians the anterior limb 
is much more mobile. Appreciable gravitational pull upon the thorax 
is also lacking, and there remain as discernible influences only the 
stimulus for an aquatic mammal to assume a circular thoracic cavity, and 
the effect that levitation by the inflated lungs may have. As a matter of 
fact in cross section the chest seems to be definitely broader than high. 
Levitation by the lungs might be expected to raise the curve of the ribs 
well above the vertebral column, but this is no more marked than the con- 
dition encountered in many terrestrial mammals, both quadrupedal and 

Sirenians have experienced quite a remarkable lengthening of the 
thorax which has operated to shorten the lumbar region. There seems to 
be considerable variation in the number of ribs. Stannius recorded a 
manati with 15 ribs while Murie encountered individuals with 16, 17 
and 18 pairs. Halicore may have 18 or 19- In the latter genus the con- 
formation of the ribs is not unusual, but in Trichechus they are very re- 
markable. In this animal the individual thoracic vertebrae are much 
longer and therefore the rib centers are farther apart. Presumably in 
compensation the ribs are phenomenally broadened so that the intercostal 
spaces are not particularly wide. Furthermore in this genus (at least 
in the species latirostris) the distal ends of the first twelve ribs are on a 
line virtually parallel with the vertebral column, the remainder of the 
series becoming successively shorter. It is said that always in this order 
all ribs have both capitular and tubercular attachment to the vertebrae. 
This character increases the rigidity of the thorax and might be expected 
to accompany a long sternum with strong and well calcified sternal ribs. 
On the contrary the sternum is much reduced and in the manati espe- 
cially about two-thirds of the costal series apparently have no costal car- 
tilages at all, their atrophy leaving a relic in the shape of nodular bony 
growths upon the distal extremities of the true costae. Flower has il- 
lustrated the sternum of a young Halicore in which four pairs of costal 


cartilages join the manubrial-xiphoid interval (figure 3) . In a mounted 
skeleton of this genus in the National Museum, with sternum completely 

Figure 26. Panniculus carnosus and cranial continuation of (P) seal (Phoca), 
(Z) sea-lion (Zalophus). (T) manati (Trichechus) redrawn from 
Murie), and (N) porpoise (Neomeris). 

ossified, there are but three pairs of ribs extending to this area, and also 
three pairs in a Trichechus, the first, however, apparently attached to the 



posterior part of the manubrium. In the Sirenia, therefore, there is en- 
countered the condition of a thorax extremely strong above and extremely 
weak below. It accordingly seems likely that as the Sirenia do not des- 
cend very deeply the aquatic pressure upon the thorax is relatively uni- 
form. The thorax does not need to adapt itself for fluctuating compres- 
sibility and the attachment of the ribs to the vertebrae is therefore less 
elastic. A necessity in this connection would seem to be that the animal 
breathe almost entirely by the diaphragm. 

In Sirenia the'articulations of the vertebrae in the lumbo-caudal series 
are definitely reduced, this being most marked in the manati. Anapophy- 
ses are lacking as are also well defined postzygapophyses from this series. 
Metapophyses are present in the manati although they are fairly fused 
with the prezygapophyses. In the dugong there are no metapophyses 
occurring as processes in the thoracico-lumbar series and the zygapophyses 
do not project laterad but are flush with the bony laminae of the spinous 
processes. In the manati the spines and zygapophyses are rugose with 
numerous small but well defined prominences. In the dugong the spin- 
ous processes are of moderate and approximately equal height in the 
thoracico-lumbar series. In the manati the spines are not quite so long. 
Murie (1872) has stated that in the manati the intervertebral substance 
is very limited in amount, being no more in thickness than a tenth of an 
inch, and this seems quite surprising in view of the fact that in other 
aquatic mammals this substance is unusually generous in amount, and in 
consideration of the extent to which this animal can curve its body and 
tail. The latter is often used as a prop (see figure 7) to keep the body 
from resting on the ground while feeding, and the curvature of the 
lumbo-caudal region is then excessive. 

From Murie's excellent illustrations it is seen that the panniculus 
carnosus of the manati is a most extraordinary muscle. Although some 
of its fibers stretch to the fore limb, the main body of the muscle, which 
may be as much as an inch and a half in thickness, extends from the pel- 
vic region to below the eye, in a broad, powerful sheet (figure 26) . It 
reaches quite to the midventral line, and passes both laterad and mediad 
of the arm, the sphincter colli then becoming more complex over the 
neck. Clearly this great muscle should have two functions; one as an 
accessory belly strengthener, as pointed out by Murie, to support the 
viscera and partly to make up for the absence of costal cartilages of the 
usual sort. The other function is to help bend the body in swimming. 
When one side of the panniculus is contracted there evidently follows a 
twisting, more or less lateral motion, but when both are flexed together, 




curvature is in the vertical plane, the result being that the tail is de- 

The manati thus effects swimming by flexion of the hypaxial caudal 
mass, and of the panniculus and rectus abdominis to depress the rear 
half of the animal, and this may be assisted by some of the neck muscles 
acting to depress the head. The hypaxial muscle is less extensive than in 
the Cetacea because the lumbar region is shorter. It is made up of a 
massive superficial and a robust deeper division, both caudad of the last 
rib, and in addition, a smaller muscle which reaches within the thorax 
and which probably represents a quadratus lumborum. The antagonist 
of this ventral group is, of course, the erector spinae. The divisions of 
this seem to be well fused, as one would expect. A spinalis dorsi con- 
tinuous with a levator caudae internus was said to have been distinct, 
while a broad sacro-lumbalis was confined to the thorax. 

As previously indicated one cannot be sure whether the theoretical 
mechanical stimulus encountered by aquatic mammals is for an exactly 
round thoracic cavity, or a round trunk as a whole. Personally I con- 
sider it likely that this stimulus alone is rather feeble, or else that it is 
overcome by much stronger ones. Just as a combination of factors has 
resulted in the assumption of a trunk laterally flattened by speedy pelagic 
fishes such as the mackerel, so might one expect to find that the swiftest 
whales are so flattened but in the sagittal plane, to correspond with the 
different plane of tail motion. As a matter of fact these two conditions 
are complicated by at least two factors that should influence tail shape, 
as discussed in the next chapter. 

The thoracic shape in whales is so variable that all our arguments fail 
of application. It is impossible to ascertain the precise thoracic cross 
section of a living, adult mysticete. Mounted skeletons indicate that 
the cavity is either slightly broader than high, or else is approximately 
circular. Allowing for the spinal musculature this would make the en- 
tire trunk either approximately circular or else higher than broad. And 
the latter seems to be the situation in at least the majority of porpoises, 
to a really quite marked extent in some. Levitation by the lungs may 
have been of influence in shaping those with broadened thorax, but this 
argument takes us no farther. We might expect to find that the very 
strong levitation of the lungs during deep diving had elevated the dor- 
sal curve of the ribs well above the vertebral centra, but this is no more 
marked than in many terrestrial mammals. 

As with the shape of its cross section there is really extreme variation 
in the length of the cetacean thorax, it having become elongated in some 




sorts and shortened in others. Neobalaena has no less than 17 pairs of 
ribs, while Hyperoodon has 9, which is the least number in any living 
mammal. Other sorts vary between these two extremes. There is in- 
dicated some instability in the anterior thorax by the fact that a few 
cetaceans have a pair of rudimentary cervical ribs, and that occasionally 
in mysticetes the first rib is bifurcated, one head going to the seventh 
cervical and the other to the first thoracic vertebra. It is perhaps usually 
stated that some whales exhibit this condition, which gives the impres- 
sion that it is a character of certain species. As far as I can ascertain, 
however, it is an individual though perhaps fairly common develop- 
ment. Thus no mounted mysticete skeleton in the National collection 
shows any bifurcation of the first rib, while the Sei whale obtained in 
Japan by R. C. Andrews (1916) does show it. In view of the fact that 
the more primitive fossil whales all had seven cervical vertebrae, so far 
as known, the two details mentioned above might well be interpreted as 
an indication of morphological effort to attain, in a somewhat different 
manner, a forward shift of the thorax. But against this reasoning there 
is the totally antagonistic fact that Neobalaena, anomalous in so many 
ways, has the first eight vertebrae without ribs, so that in this genus there 
seems to have been a shift of the thorax toward the rear. 

The thorax of whales varies in the strength with which the ribs are 
attached to the vertebral column. Ribs with both tubercular and capitu- 
lar attachment vary in number among toothed whales from at least eight 
(and possibly more in some sorts not examined) in Berardius, Kogia, 
Monodon, and Phocaena to as few as four in Stenodelphis and some in- 
dividuals of Tursiops. It has been stated in the literature that the ribs 
of existing mysticetes are all single-headed, but this is erroneous. Condi- 
tions are variable and the articulations of the ribs are certainly reduced, 
but all individuals of this group which I have examined show at least a 
few ribs that cannot be called single-headed, possibly with the exception 
of Megaptera. It seems normal for the first rib to have no capitular pro- 
jection toward the centrum. In Eiibalaena this is also lacking from the 
second rib, it is slight upon the third, and upon the fourth reaches half 
way from the tuberculum to the centrum, thence gradually shortening in 
caudal sequence. In Rhach'mnectes the second to sixth ribs inclusive have 
capitular projections that almost reach the centra. In Sibbaldus the sec- 
ond, third and fourth ribs show this character, while in Megaptera there 
is some capitular projection upon the third rib, which, however, falls 
considerably short of the centrum, and this is still less defined in the more 
posterior ribs. Thus, although there may not be bony connection between 



capitulum and centrum, a process representing the former undoubtedly 
indicates that there is strong hgamentous connection. 

Accompanying the increased elasticity of the dorsal thorax is a com- 
parable development of the ventral thorax brought about by a reduction in 
the costal attachments to the sternum. This is attained partially by an 
increase in the number of the so-called floating ribs through the elimina- 
tion of the cartilages which more usually are attached to the series of 
costae often but unfortunately termed false ribs, and by a reduction in 
number of the true sternal or cartilaginous ribs, which in odontocetes 
seem invariably to become calcified with age. Reduction of the sternal 
ribs is inevitably accompanied by a corresponding reduction in the length 
of the sternum. Atrophy of the latter complex does not follow any one 
rule but may be attained by elimination of any of the posterior elements. 
The xiphoid and posterior sternebers may disappear, but whales ap- 
parently never show the condition encountered in Halicore, in which the 
sternebral series has become so reduced that this has no individual cen- 
ters of ossification (apparently), several costal cartilages thus having at- 
tachment to the xiphomanubrial interval, although this later experiences 
general calcification. 

That a reduction of the cetacean sternum is now in progress is indicated 
by the statement of Beddard (1900) that in Phocaeua it shrinks progres- 
sively from the young to the adult state. Among Odontoceti there may 
be seven pairs of ribs with sternal attachment (Betaidius), six pairs 
( Monodon), or the sternum may be much more reduced, the elements 
finally fusing in the adult, and resulting in a single bone that may be 
broader than long (as Neomeris). Even when the sternum consists of 
several elements complete ossification of the whole often occurs in old 
age. In all porpoises without exception, I believe, the manubrium is 
broader than any other sternal element, is indented antero-medially, and 
produced in a process antero-laterally (chiefly for the sternohyoid mus- 
cle) , with a lateral expansion for the pectoraHs minor attachment. The 
sternum of Physeter is of peculiar form and Flower (1876) considered 
that it consists of a manubrium and two sternebrae ; but Owen stated that 
it has four elements. In all odontocetes at least the ossification of the 
twe centers of each sterneber is delayed. Occasionally (as in Monodon, 
figure 31) two pairs of ribs have attachments well forward upon the first 
sternal element and there thus seems to be some question regarding the 
homology of the latter. Either all or most of the ribs have shifted for- 
ward, thus placing the second pair upon the manubrium, the latter and 
first sternebra have experienced first crowding and then fusion of their 



ossification centers, or, more likely, the two centers of the manubrium 
were first separated by an antero-medial indentation, then becoming atro- 
phied, and the first sterneber finally became hypertrophied to take the 
place of the true manubrium. This point should be more thoroughly 
investigated when adequate material is available. 

In mysticetes the sternum is still further reduced to a single bone, evi- 
dently the manubrium, to which is attached but the first pair of ribs. It 
is frequently stated that this bone is of some particular shape in one spe- 
cies and of another shape in a second, but as a matter of fact the shape 
appears to be almost as variable individually as specifically. It is prob- 
ably always more or less roughly triangular or heart-shaped in the ba- 
laenid whales, while in the others it may be cither U-shaped, T-shaped, 
cross-shaped or rather irregular. Usually, however, there is a postero- 
medial projection forming with the lateral process a concave border into 
which fits the end of the first rib. Certainly in these whales the sternum 
has suffered as much reduction as is possible without a change in the at- 
tachments of (chiefly) the smaller, more fleshy pectoral (when there are 
two divisions) and of the sterno-mastoid. 

Winge (1921) stated that when there is a tendency for the costal car- 
tilages to disappear the sternum loses an essential stimulus and becomes 
reduced. This is self-evident, whether it be considered that the sternal 
elements are morphologically derived from the costae, or (according to 
a later thesis) , that the derivation of this complex was originally from 
the medioventral part of the shoulder girdle. Schulte (1916) ascribed 
to J. C. Vaughan the verbal opinion that in descending to great depths 
the pressure of the water on the abdomen would force rostrad the dia- 
phragm, which would in turn force the ribs out, accounting for the re- 
duction of rib connections and of the sternum. Muller (1898) believed 
that the reason the mysticcte sternum is more reduced than in odontocetes 
is that the diaphragm of the former is less muscular, indicating a greater 
use of the thorax in breathing. 

The sternum as a whole can hardly suflFer reduction in the number of 
its elements after the fashion characteristic in Cetacea without corres- 
ponding reduction of the costal cartilages. That the latter is not neces- 
sarily dependent upon decrease in strength of the costal-vertebral arti- 
culations is suggested by the fact that in the manati these articulations 
arc strengthened, while the sternal ribs are considerably reduced. While 
not absolutely dependent one on the other it is but logical to infer that 
increase in the elasticity of the dorsal thorax has usually gone hand in 
hand with the same adaptation ventrad. Then there is the matter of the 





function of the ventral thorax as a scaffold for muscular attachments. 
The costal cartilages cannot be reduced unless conditions permit of al- 
teration in the costal attachments of the transversalis abdominis, and in 
the more ventral of the intercostal muscles. The sternum cannot be re- 
duced without alteration in certain aspects of the rectus abdominis at- 
tachments and reduction in origin of the pectoralis. Chief of these are 
the latter two. The major sternal attachment of the rectus may easily 
shift forward anyway, and an extensive origin of the ectal pectoral may 
not greatly involve the sternum proper but only the linea alba. Where 
there is a heavy anterior pectoral mass attached broadly to the bone, 
which is the usual character of the minor division when this is present, 
or of the major if this be greatly reduced, this affects the sternum mainly 
in a broadening of the manubrial part. Hence it appears likely that mus- 
cular adjustments permitting the shortening of the sternum are easily 
made. In this connection there should also be mentioned the possibility 
that the extraordinarily developed gular musculature of mysticetes may 
well have been instrumental in reducing the sternum in this group. 

In fine, all we can logically infer regarding the reduction of the ster- 
num in the Cetacea may be summed up in a single paragraph. There is 
lacking the brachial and muscular stimulus for a sternum of moderate 
size. Vaughan's opinion that water pressure at great depths would force 
the diaphragm forward hardly seems well taken. Such pressure is na- 
turally the same over all parts of the animal at any given depth and we 
know that at great depths external and internal pressure must be equal- 
ized, for no thorax could otherwise withstand a pressure in excess of a 
ton to the square inch. It therefore seems logical to infer that the in- 
creased elasticity dorsad and the reduction of the sternum in this order is 
to allow for the amount of compressibility of the thorax needed to pre- 
vent water pressure from cracking the ribs. 

The vertebral column of the Cetacea is noteworthy for the very marked 
reduction (least pronounced in Platanista), which really amounts to en- 
tire abandonment, of interlocking or articulation of the vertebrae, and 
increase in size of the intervertebral disks, both of which are modifica- 
tions to increase the uniformity with which the column may be curved 
without tendency to bend at one or more particular points. It is per- 
mitted by the fact that the body has been supported by flotation only for 
a very long period of time and, incidentally, is a step in retroversion to- 
ward the primitive chordate condition. As support is by flotation acting 
upon all parts of the body it might be expected that the static position as- 
sumed by the vertebral column would be a straight line. The column, 



however, is affected by the fact that it is located above the longitudinal 
body axis and by tension of the muscles that extend toward the head on 
one hand and the tail on the other, and in life is slightly curved to a 
greater or lesser degree depending on muscular differences and body 

The terminal epiphyses of the cetacean vertebrae are very distinct and 
become thoroughly ossified into thin disks which fuse with the bodies 
very late, especially in balaenid whales. Flower (1876) has stated that 
whales appear to differ from all other mammals inasmuch as the neuro- 
central suture is always placed a little above the junction of the arch 
with the body. Anapophyses and postzygapophyses are absent as true 
processes. Prezygapophyses may be well developed and present upon 
all vertebrae, as in Mysticeti, Stenodelphis, Monodon, Mesoplodon, etc., 
indicating that there is definite zygapophyseal articulation (at least by 
ligaments) throughout the entire column ; or on the other hand they may 
be totally absent from the central lumbar series, becoming gradually 
differentiated craniad upon the first few lumbar and caudad on the last 
few, or even entirely eliminated {Grampus griseus, figure 28). This 
shows that in the area where they are not well defined zygapophyseal ar- 
ticulations do not exist, which increases the limberness of the column by 
just this much. Metapophyses are developed and in caudal sequence 
gradually arise from low upon the neural arch to the base of the spinous 
process. In the anterior thoracic series the di- and metapophyses usu- 
ally occur unseparated, I believe, the diapophyses in most sorts of whales 
gradually separating and descending from high on the arch to the cen- 
trum, thus becoming parapophyses. In the Physeteridae, Kogidae, and 
the Ziphiidae, however, there is a different condition, for the diapophyses 
do not gradually change to parapophyses. On the contrary the former 
diminish and disappear, while at the same time the latter become defined 
well ventrad of the diapophyses and gradually increase in size. 

As previously remarked most cetologists refer to the phenomenal short- 
ness of the whale's neck as though this character were entirely disasso- 
ciated from thoracic conditions. As a matter of fact this shortening 
process merely culminates in the neck, but also involves the anterior third 
of the thorax. In mysticetes there is an almost insensible, gradually in- 
creasing shortening of the centra from about the fifth thoracic to the 
third cervical (in those sorts with fjee neck vertebrae) . In most por- 
poises which have the posterior cervicals unfused these may be relatively 
thinner, to a considerably greater extent than would be possible for any 
thoracic vertebra bearing a rib. Consequently there is usually a more 



abrupt transition in the thickness of the centra between the thoracic and 
cervical series. The proper interpretation of these facts seems to be not 
merely that the cetacean neck has become markedly shortened, but that all 
vertebrae anterior to the thoracic pivot of motion have experienced strong 
stimuli for shortening, to which they have responded to various degrees 
according to their capabilities, this shortening having been necessary in 
order that the animal might assume the fusiform shape best fitted for 

Winge (1921) has stated that the abandonment by the fore limbs of 
the function of supporting the body results in the reduction of the height 
of the anterior thoracic spines. This character, however, is only secondar- 
ily dependent upon function of the anterior limb. If in quadrupeds the 
fore limb be of greater importance during locomotion than the hind, then 
not only must the spinal musculature have corresponding strength in this 
region but the spines must be higher to give better leverage for muscles 
that elevate a heavier head and neck (as in Bison) (Thompson, 1917). 
When the fore limb is no longer used for support the anterior part of the 
erector spinae usually becomes weaker (phocids, bats, etc.) , or it may un- 
dergo an independent modification for increased strength, resulting in 
spinous processes that are phenomenally developed, as in most Cetacea. 
The height, and to a lesser extent the character, of these varies consider- 
ably, however, reflecting corresponding variation in the spinal muscula- 
ture. In Aiesoplodon the height of the spines of the entire column is quite 
phenomenal. In mounted skeletons of Delphinus and Lagenorhynchus 
the spines are relatively broad sagittally and so close together that it is 
difficult to understand how the requisite amount of movement would be 
possible; but perhaps in life the intervertebral disks are unusually thick. 
In Globwcephala, on the other hand, there is sufficient distance between 
the spines for great mobility of the column. Winge (1921) has said 
that "bending of the column in the vertical plane — is reduced or aban- 
doned." This is an extraordinary statement in view of the fact that all 
the modifications of the vertebrae with the possible exception of spinous 
height are for increase of mobility and all locomotor movements in whales 
are instigated by motion of the column in this plane. 

In some respects the lumbar region of Cetacea really constitutes a part 
of the tail, from a functional standpoint, but certain of its characteristics 
may be discussed in the present chapter. Flower and Lydekker (1891) 
have stated that Neobalaena and Inia have but three lumbar vertebrae. At 
the other extreme Grampus may have 21, Lagenorhynchus, 23, and Del- 
phinus, 24 (21 in another skeleton). This of course, is figuring on the 



basis that there are no sacrals present, the length of the lumbar series be- 
ing determined by the position of the first chevron bone of the tail. As 
a matter of fact we have no assurance that in such a mammal without a 
sacrum there can not have been great alteration in the chevron compli- 
ment so that the first of these bones might now occur anterior or far 
posterior to its ancestral position, meaning, in the latter case, that the 
first few of the series have been eliminated. Either this has been the case, 
lengthening of the lumbar region having been brought about by absorp- 
tion of the anterior caudal elements, or else this has been due to the lay- 
ing down in embryo of accessory lumbar vertebrae anlage. It is, how- 

FiGURE 30. Vertebrae of the porpoise Phocaena, illustrating differences in the 
positions of the various details: {a and b) fourth thoracic; {c) eighth 
thoracic; and {d and e) lumbar vertebrae. 

ever, probably not astonishing that there should be such lumbar varia- 
tion in a mammal in which there is no attachment of the pelvis to the 
vertebrae and in which the apaxial and hypaxial musculature have each 
experienced such a complete degree of fusion. It seems that this would 
involve merely a shifting forward or backward of the pelvic region by 
shortening or lengthening of certain muscles with little regard for the 
bones, and this should not entail any great difficulty. Presuming that 
this has been the case, I have no iciea regarding the stimuli involved or 
the advantages gained. There is just one point that might throw some 
light on the question. In at least most cetaceans having short or fairly 
long (as many as about a dozen vertebrae) lumbar regions, prezygapophy- 



ses are continuously and uniformly present as far as the peduncle and in 
these, as far as can be told from bony details, the post-thoracic apaxial 
musculature is relatively homogeneous, so that the lumbar region seems 
to constitute the base of the tail. In porpoises with an excessive num- 
ber of lumbar vertebrae it seems that all spines except those near either 
end lack zygapophyses, so that in these sorts there are involved two re- 
gions, in some manner separately specialized, of the lumbo-caudal part 
of the erector spinae, one in the lumbar and the other in the caudal area. 
Conditions suggest that a need experienced for longer lumbar muscles 
may have been instrumental in lengthening this region in certain por- 
poises. This in turn suggests that there may be some difference in the 
exact muscular action by which these two sorts of cetaceans accomplish 

In Cetacea the spinous processes and parapophyses, constituting simple 
transverse processes, of the lumbar series are greatly developed, while 
other bony protuberances are either much reduced or entirely absent. This 
corresponds to the simplification of the spinal musculature. The whale 
swims by movements in the vertical plane of the caudal appendage or 
flukes, and the spinal muscles are called upon for little else. Naturally, 
with almost perpetual use these have acquired phenomenal thickness. In 
the lumbar area the dorsal muscles are imperfectly divisible into two ser- 
ies (iliocostalis and longissimus) while in the posterior thorax these may 
be even more homogeneous, and continue onto the head with little 
change either in mass or character. The muscular action concomitant to 
swimming has already been discussed to a considerable extent so that here 
little need be said, save to repeat that in whales, as contrasted to seals, 
the back muscles can act more uniformly throughout the entire vertebral 
length, permitting more latitude in possible shifting of the pivot of mo- 

The erector spinae operates to raise the tail. Consequently in whales 
there must be an antagonist to depress the tail with potential force that 
is approximately equal. This is provided by the extraordinary modifica- 
tion of the infravertebral or hypaxial musculature of the tail base and 
lumbar region. In the latter situation there are presumably psoas and 
quadratus lumborum elements, but it is entirely out of the question to de- 
cide whether all of these have become hypertrophied or some have in- 
creased at the expense of others, for anteriorly they are so simplified that 
only one superficial separation is possible. In what may be termed the 
posterior lumbar area the hypaxial and apaxial musculature are of ap- 
proximately equal mass, but whether they are each capable of exerting 



precisely the same force is unknown. If there be some inequality in- 
volved, then this must be compensated for by tilting of the flippers. Be- 
yond question, however, there is a tendency for perfect equalization, 

Figure 31. Sterna of whales and sirenians: (a) Balaena and (f) young dugong 
(Halicore) (both redrawn from Flower) ; (b) Sibbaldus and {d) Balaenop- 
tera borealis (both from specimens in the U. S. National Museum) ; and 
{e) young narwhal {Monodon) (from a photograph in the American 
Museum of Natural History). 

which presumably has been accomplished as fully as the muscular con- 
formation will allow. In respect to the latter point, the hypaxial muscles 
continue robustly only to the last rib, and thence disappear within two or 



three costal spaces. Hence, while the apaxial muscles operate through- 
out the entire length of the vertebral column, the hypaxials can bend 
only the lumbo-caudal series. Anteriorly what downward bending of the 
posterior thorax and of the head is necessary in swimming must be taken 
over chiefly by the powerful rectus abdominis on the one hand, and the 
ventral neck muscles on the other. In this connection a point should be 
mentioned that may already have occurred to the reader. In theory the 
swimming motion of whales has been discussed as though the curve as- 
sumed by the body were a perfect arc. In practice, however, this may not 
be so. Not only may the costal equipment prevent as much possible cur- 
vature of the thoracic vertebrae as of the lumbo-caudal series, but this 
may further be reduced by the greater mass in the former region. It 
might therefore be more correct to consider that the neck and tail each 
bend from the thorax to a considerable degree, while the curve of the 
thoracic vertebrae is more moderate. This, at the present time, is almost 
impossible of determination. 

Theoretically lateral motions by the tail are unimportant to the Cetacea, 
■ but actually there must be considerable strength in this plane for thrash- 
ing about when the need arises. Such movement may be accomplished 
chiefly by the flexion of the apaxial and hypaxial muscles of a single side, 
and also by the intertransversarii, which occupy the space between the 
transverse processes of the lumbo-caudal vertebrae. In cetaceans these 
are unusually modified and upon the thorax, of at least most odontocetes, 
spread out into a thin sheet covering a large part of the ribs deep to the 
latissimus dorsi layer. In Monodon there is a convergence of the fibers 
from each direction to (about) the sixteenth lumbo-caudal vertebra, in- 
dicating a center in this region for lateral movement. It is likely that the 
intertransversarii act fully as much in preventing too much curvature as in 
instigating it. In other words the function is probably as much static as 

[182 1 

Chapter Nine 

The Tail 

1 HE QUESTION of the physiological development of the hinder end of 
an aquatic mammal for the function of primary propulsion is an in- 
volved one. Of certain points one can be sure, while regarding others 
there is some uncertainty and only probabilities may be advanced, this for 
the reason that there are indications that in particular cases the develop- 
ment has not been in a straight line but has been somewhat by trial and 
error. As a fundamental concept, however, I have no hesitation in mak- 
ing the unequivocal statement that the evolutional tendency when a mam- 
mal takes to a free-swimming type of water habitat is always for it even- 
tually to develop the rear end into the primary, oscillating organ for its 
propulsion through the water. If it does not do so it is a sign that it 
originally made a wrong start or that it has encountered antagonistic 
stimuli of such strength that it was diverted from the most efficient evo- 
lutionary development. In theory it makes no difference whether this 
ideal propulsive force is furnished by the flattened tail as in the whale or 
by the adpressed hind feet of the seal. In practice, however, it is prob- 
able that the muscles of the seal that are involved in swimming can never 
become as homogeneously specialized for a single function as are those 
of the whale's tail. I do not know and consider that speculation on this 
point would be well nigh useless. 

The important point is that although swimming by the whale and the 
seal entails widely different muscular action, the principle is in theory 
the same, and this principle is the only one that a mammal can adopt 
which is thoroughly economical in practice, for it is the only one by which 
there is no lost motion or energy. In the case of other sorts of swimming 
motion by the hind feet, or by the fore feet, either recovery motions are 
necessary or a part of the flippers must overcome water resistance — al- 
ternatives which detract by just so much from the propulsive strokes. 

For a proper consideration of the situation some brief recapitula- 
tion is necessary. When a mammal first takes to the water it has a caudal 
equipment that may be divided into three categories. It is either without 
a functional tail, like the bears for instance, it has a long tail, or this 
member is of intermediate length. If the tail originally was short it is 



likely to remain so and never be of economic value to the animal, in 
w^hich case the limbs will be the members developed for propulsion. 
If the tail be long it is extremely likely that it will eventually become the 
prime means of swimming, unless it be diverted at a rather early stage 
in its development by the necessity for fulfilling some other function. If 
it be of intermediate length one of two things may happen, depending 
upon its exact size originally and the other conformational features of the 
mammal. In this must be included inherent capacity for change in the 
desired direction. It may become shorter and cease to be of consequence 
or it may become longer and of greater importance. In illustration of 
the possibilities in this line it may be mentioned that in the case of a 
mammal with moderate length of tail we can never predict what the de- 
velopment will be because the antagonistic stimuli may be too nearly 
equal. Although the modification is entirely according to law it appears 
to us as fortuitous whether the tail or the hind limbs gains the initial 
ascendency and implied evolutional velocity of the chief swimming or- 

In approaching this subject of caudal evolution one must clear his 
mind of all idea that this member changed to the form in which it may 
now occur in the most specialized of aquatic mammals for some single 
fundamental reason. True, its development has followed definite laws 
and in most cases it is predictable what form the tail will finally assume 
merely by observing the methods of swimming which any mammal now 
employs, but the development is, nevertheless, step by step. Thus the 
current belief that the cetacean tail is flattened vertically so that the ani- 
mal may more readily ascend to the surface for breathing is not only 
erroneous in practice, but involves an improper mental approach to the 
entire subject of aquatic specialization. Not only can an animal with 
horizontally flattened tail ascend as easily for the reason that elevation of 
the body is accomplished by the equilibrating rather than the propelling 
mechanism, but no animal could start on its evolutional career with any 
such particular end in view. The direction in which the tail will finally 
be flattened is dependent not upon the tail itself but upon the direction in 
which it is involuntarily (usually) moved during swimming movements, 
of the unspecialized ancestral form. 

For ease in discussion the tails of aquatic mammals may be divided 
into two classes, which involve two fundamentally different principles: 
Those which are narrow in the horizontal plane or which will eventually 
become so; and those which are flattened in the vertical plane or will 
eventually become so. In these two classes are included even such tails 



as are too short to be of any possible use for propulsion, and they thus 
comprise all aquatic mammals except the capybara, which is entirely de- 
void of a tail. 

1. Tails narrow in the horizontal and broadened in the vertical plane. 
In the first group below are included all those genera in which the tail 
has been flattened in the horizontal plane, or else is provided with a 
ventral keel of stiff hairs. In the second table are those genera in which 
the tail is still terete but which may be expected eventually to develop a 
narrow tail rather than a flat one. There is some doubt in this regard re- 
specting Chironectes, as explained later. 

Tail narrow Tail round 

Desmana Chironectes 

Galemys Neosorex 

Neomys Atophyrax 

Chitnarrogale Myocastor 

Crossogale Dasymys 

Nectogale Nilopegatnys 

Limnogale Arvkola 

Potomogale Neofiber 

Crossomys Rheomys 

Hydrof7iys Anotomys 

It will be noted that an the above mammals having horizontally flat- 
tened tails are either insectivores or rodents with the single exception of 
the hippopotamus, and in none except Potomogale, and possibly Limno- 
gale, does the tail constitute the principal organ of propulsion. Also that 
none, except Potomogale perhaps, is really very highly specialized in an 
aquatic direction. 

From a study of the question I am led to believe that in order that 
an aquatic mammal shall finally acquire a tail that is horizontally flat- 
tened its terrestrial ancestor must have had the following characteristics: 
a normally cylindrical body not particularly elongated ; a tail, preferably 
of considerable length, which was not much enlarged at the base, thus 
showing an abrupt transition in size between the hinder end of the 
body proper and the base of the tail (of the character occurring in 
typical rodents, as the rat) ; feet of the normal rodent or insectivore 



character, in which the hind feet are sufficiently larger than the fore 
feet so that there will be no doubt but that the former will be of greater 
importance to locomotion while the animal is swimming "dog-fashion". 

Given the above characters the course of aquatic development wiH 
ordinarily be as follows: When the animal first takes to the water, swim- 
ming will normally be accomplished by the movement of all four feet. 
The hind feet, however, being larger than the fore feet will be de- 
pended upon more and will gradually become larger, acquiring web- 
bing or a fringe of bristles. At this stage the fore feet are only an 
incidental aid to locomotion. Swimming will normally be by alternate 
strokes of the hind feet. I accept this as the most efficient method 
that will almost invariably be employed merely because it is the rule, 
in birds as in mammals. Swimming by alternate strokes of the hind 
feet involves wriggling the hinder end of the body, which will cause 
a sinuous motion of the tail. This latter will aid forward locomotion 
in degree according to the area of the lateral tail surface. I have no 
hesitation whatsoever in stating that this horizontal lashing of the tail, 
so well illustrated in the case of the muskrat {Ondatra) constitutes a 
strong stimulus for lateral flattening of this member. Why this is so 
no one knows, but the evidence is overwhelming that just such stimula- 
tion will initiate development in a useful direction. It may be purely 
natural selection, it may be chiefly because the friction of the water 
against the upper and under sides of the tail tends to develop a dorsal 
and ventral thickening, or a complex of unknown factors may be in- 
volved. At any rate, as the flattening of the tail progresses it will be of 
greater and greater proportional importance in swimming, for it is 
theoretically much more efficient than can be the alternate kicking of 
the hind feet because the latter necessitates recovery motions. Finally 
this specialization of the tail will increase possible speed to the point 
at which movement of the hind feet would be more of a hindrance than 
a help and use of the feet as a primary, or even secondary, means of 
speedy aquatic locomotion will be abandoned. The final step in this 
direction would be the assumption of a fish-like tail comparable to that 
of the whale, but with flukes vertical instead of horizontal, and pre- 
sumably the elimination of the hind feet. No mammal is thus devel- 
oped, for none with horizontal flattening of the tail is independent of 
the land. All are of rather small size, are inhabitants of streams more 
properly than of large rivers, and are not yet very highly specialized in 
an aquatic direction, Potomogale being the most modified of the lot. 

The above thesis is perfectly consistent with the facts except in two 



instances, one of these being Hippopotomus and the other Potoniogale. 
The size of the tail in the former animal is entirely too small to be of 
the shghtest use either in swimming or in steering, and yet it is laterally 
flattened to a phenomenal extent. The only explanation for this condi- 
tion would seem to be that in spite of its small size it has been wriggled 
from side to side by the alternate strokes of the hind limbs for a suffi- 
ciently long time for it to have responded to the same stimulus that 
would have aflPected it were it sufficiently large to be an aid in pro- 
pulsion. The case of Potoniogale is more difficult. Its tail is remarkably 
efficient as a flattened propulsive organ and yet the feet are small and 
so unmodified that we are justified in assuming that they are not used 
while the animal is in the water. It seems that the only logical theory 
is that originally the terrestrial ancestor of this mammal was equipped 
with small feet while its tail, although round, was unusually heavy, 
especially at the base, so that in connection with a slim body, it was 
even then a more efficient swimming organ than the small feet. If it 
came more natural for this terrestrial ancestor to move the tail from side 
to side, as a dog does, rather than in the vertical plane, flattening in 
the horizontal plane would be the inevitable result. 

The insectivores and rodents listed above as having terete or round 
tails can be dismissed with the statement that they are either not suffi- 
ciently specialized for there to have been caudal change, or else this 
member has been unusually conservative. The case of the water opos- 
sum (Chivonectes) is somewhat diff^erent. The webbing of its hind 
feet is quite extraordinary in degree, and it is difficult to understand 
how this was brought about so completely without some flattening of 
the tail, providing that the animal swims by alternate kicks of the feet- 
movements which would of necessity involve the wriggling of the tail 
from side to side and this should be easy of accomplishment because of 
the robustness of the tail base, as in so many marsupials. There is no 
authentic statement, I believe, of the exact method of swimming em- 
ployed by this animal and it does not seem impossible, although it is 
improbable, that it might propel itself by means of kicking the feet in 
unison, and if this should ultimately prove to be the case, then the round- 
ness of its tail would be expected. 

It seems that at least in insectivores and rodents the first change in 
the tail brought about by their aquatic life is usually the acquisition of 
a border of hairs, above or below or both, that are longer and stiflfer 
than those upon other parts of the tail. // there be anything to the 
theory of the eventual inheritance of acquired characters throughout 



lengthy periods of time, then it seems that a hairy border was in some 
way stimulated to growth by the endlessly repeated lashing of the tail 
from side to side during propulsion through the water. This motion 
results in considerable friction by the water along both the upper and 
under side" of the tail and this may have been the activating agent, 
throughout a very long period of time in order to have final effect. 

A similar possibility may explain the development in insectivores 
and rodents of a tail flattened transversely and expanded in the vertical 
plane. Any stimulus that increased growth of hair upon the dorsal 
and ventral borders of the tail would presumably result, at a later stage 
of aquatic specialization, in the deposition of more subcutaneous tissue 
along these borders. And this is mainly fatty or even partly glandular, 
rather than muscular or bony. But this is pure speculation, for as we 
cannot even find out what really causes baldness in man one can hardly 
speak authoritatively on the reasons governing hair growth upon the tail 
of a shrew. A rather baffling coincidence in the tails just considered is 
the fact that although hairy caudal keels evidently precede (usually) 
flattening of the member itself, in those forms in which the latter is the 
case the tail is no hairier above or below than upon the sides. 

Of those insectivores and rodents whose tails are flattened in the 
horizontal plane Desmana, Potomogale and Ondatra are the only ones 
in which this character is really pronounced, although Galemys and 
Limnogale approach the conditions existing in the first two genera re- 
spectively. The tails of Desmana and Galemys are somewhat swollen 
posterior to the base, as is usual in the family Talpidae, this being due 
to the presence of bodies that are largely fatty but which may also be 
glandular in character. In the muskrat {Ondatra), however, the tail 
is unusually slender, the skin not only being rather thin but without 
any sign of subcutaneous fat and so closely adherent to the muscular 
and tendinous tissue that this member is quite difficult to skin. In these 
aquatic moles and in the muskrat the base of the tail is not enlarged 
but is abruptly smaller than the hind quarters, approximately as in the 
common brown rat. In the muskrat at least the tail describes sharp 
sinuous movements throughout a considerable arc during aquatic pro- 
pulsion. This motion follows oscillation of the hind quarters while the 
hind feet are kicked alternately and is, undoubtedly, chiefly involuntary. 
The flattening of the tail, however, has progressed to the point where 
this member is a definite, although still a secondary, aid to swimming, 
and but a slightly greater specialization in this direction would enable 
it to constitute a propulsive organ equal in importance to the hind feet. 


In fact I feel sure that I have seen a muskrat swimming slowly by means 
of the tail alone, while the feet were trailed, but at speed the feet must 
still be employed. 

In the above mammals the musculature controlling lateral movements 
of the tail is not as yet appreciably specialized, although it must be in- 
appreciably so. In other words the muscles must be considerably more 
fitted for this work than is the case in the nearest terrestrial relatives, 
but not to the extent that this has as yet had much effect upon gross 
conformation. All that can now be said is that as this specialization 
becomes more marked, the muscles controlling lateral motions of the 
tail will become larger, broader, and there will develop a tendency for 
muscular simplification and partial fusion. 

I regard it as doubtful if the tail of Potomogale ever presented quite 
the appearance that we find in most terrestrial insectivores. As al- 
ready mentioned the case of this genus is somewhat puzzling, but it 
seems to me probable that its tail was already quite robust at its base 
when it first took to the water. The tail of this genus is now as perfected 
a swimming organ as it can probably ever be unless the animal become 
less dependent upon the land. Swimming is evidently accomplished 
not by a violent wriggling of the tail, as in the muskrat, but by more 
circumscribed, although stronger, motions, enabling the animal to slip 
through the water with remarkable ease and speed. I have never had 
a specimen for dissection but Dobson (1882) has illustrated its muscu- 
lature to good advantage. The surprising detail shown is the way in 
which the gluteal complex has become specialized. One division of the 
latter has expanded so as to stretch from the knee posteriorly and far 
over the base of the tail. Clearly this gives the latter added lateral 
leverage, so that a number of the muscles of the posterior limb can 
assist oscillations of the tail, either directly, or indirectly as antagonists. 
The muscles from the pelvis, consisting of Dobson's ischio- and ileo- 
caudalis, are both large and long, and indeed there is marked increase 
in the robustness of all the muscles of the tail proper. Unlike most mam- 
mals of this size the muscles of the base of the tail are entirely fleshy, 
rather than chiefly tendinous, and this is the reason for the gradual ex- 
tension onto the tail of the body contour. 

Before abandoning the subject of mammalian tails that have become 
flattened transversely it is justifiable to discuss future potentialities and 
successive steps in their eventual development which seem most likely. 
This, of course, is pure speculation, but is of great interest, and possi- 
bilities can at least be mentioned. 



The probabilities certainly are that the tail of the muskrat, for in- 
stance, will continue to expand vertically for some time to come, for 
we have every reason to believe that the stimulus that started it on its 
present course of development will continue. It is therefore improbable 
that the present direction of flattening would change through an angle 
of 90 degrees, so that eventually its tail would ever resemble that of the 
beaver. Evidently the stimuli for the two sorts of caudal flattening 
exhibited by these two mammals were always different. At least this 
thesis can be accepted as a likely one, and also that the tail of the muskrat 
will continue throughout the ages, to grow higher. 

Bilateral symmetry among vertebrates is almost invariable, while 
dorso-ventral symmetry, in its exact meaning, is very rare, but still, 
after a general fashion, is sometimes encountered. Thus, the dorsal 
and ventral halves of some sorts of fish are often exactly alike in shape. 
The difficulties arising from the situation were the tail of a whale longer 
and wider on one side than the other are too apparent to need discus- 
sion. But asymmetry in the vertically expanded tail of some vertebrate 
is a different matter. In a lop-sided whale of the former character all 
manner of gyrations would be performed because the asymmetry of 
the tail would throw the body to the side and there is present no pair 
of vertical rudders, one above and the other below the body, to counter- 
act the uneven influence. But were the tail expanded vertically instead 
of horizontally asymmetry need have no disturbing influence that could 
not easily (theoretically) be overcome by compensating tilting of the 
fore limb rudders, or proper curvature of the body. Tails dorso-ventrally 
asymmetrical with which their owners get along very well indeed are 
seen in certain newts and sharks. 

But asymmetry of the sort exhibited by the newt's tail might be pas- 
sively developed, as far as concerns swimming. Could there be any 
active stimulus connected with swimming that could cause asymmetry 
in a tail that is vertically expanded.'' Undoubtedly yes. There might 
be sufficient dorso-ventral asymmetry of the body, advantageous for 
feeding or other habits, so that there would eventually and naturally 
follow a compensating asymmetry of the tail to counteract it: or there 
might be asymmetrical influences introduced by an organism swimming 
habitually near the surface or near the bottom. Obviously, if a shark 
swam at the surface it would be disadvantageous to have one lobe of 
the tail projecting into the air, or conversely, a dependent tail-fork drag- 
ging over the bottom would be a handicap. From another aspect, swim- 
ming near the surface or bottom respectively introduces uneven degrees 
of water friction operating upon the upper and lower parts of the body. 


Although it is not denied that the vertically expanded tail of an 
aquatic mammal might eventually develop into perfect vertical flukes 
without ever showing any definite asymmetry, it is deemed more likely 
that there will be marked asymmetry exhibited during some stage of the 
process. If this be so the form of the tail might be either epibatic (the 
upper lobe longer and larger than the lower) as in most sharks, or 
hypobatic (the lower lobe longer than the upper) as in some of the 
extinct marine reptiles. The latter condition would probably be the 
more likely, for I believe an epibatic condition of the tail in any air- 
breathing vertebrate is unknown. 

Before abandoning the subject of tails that have expanded in the 
vertical plane it may be well briefly to discuss conditions in some of 
the extinct aquatic reptiles. Always in the flukes of the Cetacea and 
Sirenia the caudal vertebrae extend straight toward the rear and pass 
through the center of the tail to the vicinity of its medial notch. Thus 
the caudal tendons are enabled to operate practically from the posterior 
border of the tail and the force exerted upon each of the two lobes is 
symmetrical. Always in reptiles having a bilobed tail, however, the ver- 
tebrae followed the lower lobe to the tip, the caudal axis bending 
sharply ventrad at the tail base (peduncle). This clearly indicates a 
fundamental difference that from the very beginning has underlain the 
evolution of these two sorts of tails. Fraas has offered a reconstruction 
of what he considered to have been four of the stages in the attainment 
of this reptilian development. The precise shape of each tail is largely 
speculative, of course, and it is questionable whether in Mixosaurus the 
upper lobe should not be placed farther back (fig. 33), because the 
greater height of the spines near the tail base may really indicate that 
an augmented muscle mass existed at this point ; but the principle seems 
sound, for only this sort of gradual, asymmetrical development of the 
tail could account for the situation of the caudal vertebrae within the 
lower lobe. 

Obviously in the case of sharks the development of the epibatic tail has 
been the most favorable for bottom feeding habits, so that the shorter 
lower lobe would not drag upon the bottom. Equally obvious is the ad- 
vantage of a markedly asymmetrical hypobatic tail for swimming near 
the surface, so that no high upper tail lobe will project above the water. 
As deeper swimming was habitually indulged in a higher upper lobe 
could develop, this ultimately attaining the size of the lower lobe. 

Von Huene (1922) mentioned that in the latipinnate ichthyosaurs 
(as Mixosaurus) the tail seems to have been a very poor propeller and 



that locomotion was probably accomplished chiefly by the paddles. In 
longipinnate sorts such as Ichthyosaurus, however, he claimed that the 
upper lobe of the tail was immovable and together with the dorsal fin 
functioned as a rudder. It is impossible that in such a highly specialized 
form one-half of the tail could have acted efficiently as a rudder and 
the other half as a primary propeller. Undoubtedly as the upper half 
of the tail approached the lower half in size and form there was a 
stiffening of the caudal tissue, just as in whales, toward the ultimate 
goal that the caudal tendons operated from the base of the tail rather 
than its lower tip, enabling the upper lobe to impart almost, if not 
quite, as much propulsive force as the lower half. Any asymmetry in 
this force could then have been equalized by tilting of the flippers. 

The second class into which I have divided the tails of aquatic mam- 
mals is as follows: 

II. Tails flattened in the vertical and broadened in the horizontal 
plane. In the first group below are included all those genera in which 
the tail has been flattened vertically, and in the second, those in which 
the tail is still terete but which may be expected eventually to develop 
a vertically flat tail. 

Tail flattened Tail round 

Ornithoyhynchi/s Mustek (the aquatic forms) 

Lutrinae (part) Lutrinae (part) 


There are two members of the above groups which I am unable to 
discuss with any great feeling of certainty. These are Ornithovhynchus 
and Castor. It is not difficult to determine the economic value to them 
of their caudal equipment, but it is very puzzling to envision the process 
by which the specialization was initiated. The case of the platypus is 
of lesser moment, perhaps, It is such an anomalous beast in so many 
ways; we are entirely ignorant of the ancestral type, and it is the only 
aquatic mammal with a sizable tail that swims chiefly by means of the 
fore feet. Hence we must be satisfied with assigning the proper func- 
tion to its tail without attempting to visualize the successive steps through 
which it originally passed. According to all reports the tail is now 
used as a rudder, and for the purpose of keeping the head depressed 



when the animal is nuzzhng about in the mud for its food. The latter 
function would seem to be the critical one. 

Externally the tail of the beaver is phenomenally broadened and the 
integument is scaly in appearance, although cornification is not marked. 
Subcutaneously the tissue is fatty, for the most part, but the lateral 
muscles at the base of the member are very broad and powerful. The 
caudal vertebrae are unusual in that they exhibit marked broadening, 
especially those proximad, and this character is shared by the sacral 
series. In the caudal compliment there are about 23 to 25 vertebrae, 
which are short, wide and depressed, with very wide transverse processes 
which become double at the middle. 

The reason for the sort of flattening shown by the tail of the beaver 
is puzzling. The function, once popularly believed, as a vehicle for 
carrying mud, was long since proved to be erroneous. Many people 
still consider that the flat tail was developed so the animal might slap 
it on the surface of the water and thus more quickly submerge, but 
this hardly seems to be sound logic, for the muskrat can disappear with 
an abruptness that is equally startling. Its present function as an organ 
for giving warning signals (by surface slappings) is undoubtedly in- 
cidental and secondary, as is any use to which the tail may occasionally 
be put for tamping mud while building dams. It is obvious, however, 
that without a tail flattened in some manner the beaver would progress 
in circles while towing logs and sticks. This, then, must be listed as 
a primary need for a very broad tail, but one flattened in the horizontal 
plane should be much more effective for this purpose, hence, while the 
rudder function may have very materially assisted in the expansion 
process, it could hardly have initiated the present direction of flattening. 
Similarly with the fact that the tail is occasionally used as a scull. I 
have watched a beaver swimming slowly by the tail alone, pulling the 
appendage latero-ventrad first to one side and then the other. In this 
also a laterally flattened tail would be more effective. As the beaver 
swims by alternate strokes of the hind feet it seems that this almost 
certainly introduced some original stimulus for flattening the tail in 
the transverse plane. Hence there appear to be at least three factors 
for which a narrowed tail would be more favorable than a broadened 
one, and it accordingly appears likely that the broadening stimulus 
would have to be an unusually strong one to overcome them. 

The beaver often walks erect witfi an armful of mud, and also uses 
the tail as a prop while cutting trees (Bailey, 1923), or it may make 
more use of this member for keeping near the stream bottoms than is 



suspected. Of these three circumstances it seems that the second would 
be more critical in possibly causing a flat tail, but I do not feel assured 
that this is the case and I prefer to make it plain that I have no conviction 
on the subject. 

The Pinnipedia may be said to lack a functional tail and it seems 
highly probable that they have descended from an ancestry with tail 
much as in the living bears. In the walrus the "crotch" stretches un- 
interruptedly between the heels caudad of the bony tip of the tail. In 
a sea-lion {Zalophus) the external tail constituted but six per cent 
(60 mm.) of the total length, while in a seal (Pboca) this percentage 
was seven (72 mm.). The tail can have not the slightest use as an 
active aid either for propulsion or steering, but in the sea-lion the tail 
was slightly thicker in horizontal than in vertical dimension, while in 
the seal this was very much more pronounced. In the latter at least 
it was plainly to be seen that the shape of the tail permitted this member 
to fit perfectly into the cleft between the heels so as to effect an uninter- 
rupted body contour. The stimulus for this specialization is unknown. 
It was probably of an entirely different nature from that which has 
caused a broadening of the tail in aquatic forms in which this member 
is the primary propulsive organ, and for the present all that can be said 
is that it constitutes another instance of the fact that when an animal 
experiences the need for a modification, the latter will often appear even- 

The tail of the otter is of particular interest in the present connection 
for the reason that it seems to me likely that the original ancestors 
of the whales had a largely similar conformation of the body. The tail 
of this animal is exceedingly thick at base, so that there is a gradual 
taper from the hindquarters to the tail tip. At the same time the body 
is rather long and sinuous and the legs are short. The feet are often 
used in swimming, but as propulsive organs they seem to be of decidedly 
secondary importance. Swimming is mainly accomplished by curvature of 
both the lumbar and caudal regions in the vertical plane. Extension of 
the vertebrae, after flexion, is often accompanied by a kick of both hind 
legs, not in unison nor yet in alternation, but with a sort of galloping 
action, but the animal appears able to swim with equal ease and speed 
when the feet are folded against the body. The shape of both body 
and tail base, together with great muscular power in the latter region, 
is largely accountable for this, and as the feet are relatively small and 
little changed for swimming (in the common Lutra), it seems either 
that the feet of mustelids have been unusually resistent to modification 



or else, more probably, that the tail has always been of such character 
as to lend itself readily to swimming. This is also indicated by the fact 
that the feet of the mink and sumpfotter are not in the slightest modified 
for swimming and the tail, although much less specialized in this direc- 
tion than that of the otter, plays a very important part in propulsion. In 
this connection the feet of the otter will be more fully discussed in 
a future chapter. 

In but one genus of river otter is the tail other than round, at least 
to an extent that has proven noticeable. In Pteronura this member has 
a sort of fleshy keel upon either side, and hence, is flattened vertically 
and expanded horizontally. This is a development which other otters 
may be expected to follow. 

In the sea otter {Enhydra) the tail is also expanded laterally, evidently 
to' a more marked degree than in any river otter, but it is relatively 
shorter and does not reach beyond the tips of the toes when the hind 
limbs are extended backward. But one cannot discuss the development 
of the tail of this animal with any degree of assurance until more is 
known about its swimming habits. Certainly the hind feet of the sea 
otter are of greater importance in swimming than is the case in the 
river otter, and therefore, by analogy, the tail is of less importance. But 
the tail is more specialized, and it is therefore likely that it has experi- 
enced aquatic influences for a longer period of time. It may once 
have been relatively shorter and have experienced a stimulus for elonga- 
tion, or it may always have been approximately of its present length but 
the use to which the hind feet were put in swimming enabled these 
members to gain the evolutional ascendency. It appears very likely, 
however, that the tail of the ancestor of the sea otter when it first took 
to the water was not equally as long and robust as of the river otter at 
the same stage of its history, else the tail, being theoretically a more 
ideal organ for propulsion, would have gained the evolutional ascend- 
ency, as it has in the Lutrinae, before the hind limbs had gotten well 

The future course of enhydrine development is uncertain. If the 
tail were longer I should predict as a matter of course that this appen- 
dage would increase in importance and size, gradually supplanting the 
hind feet as the chief organ of propulsion, and that the latter would 
dwindle in size. Actually the tail now seems to be at a critical stage. 
The hind feet are so large and (apparently) efficient that the animal 
may be unable to alter its present course of development, or the tail may 
be unable to catch up in the race of aquatic adaption and it may well be 



that it will henceforth play a role of increasing unimportance. Inci- 
dentally any function of steering which it may now have would be insuf- 
ficent to save it from atrophy, in the latter course of events, for it is situ- 
ated too close to the primary propulsive organ to act efficiently as an 

The tail in the Sirenia varies. In Halicore and Hydrodauiaiis it is 
quite whale-like, there being two pointed flukes with a partly defined 
medial notch between, and a constricted peduncle. Little more than 
this can be said, for the muscular anatomy of neither has been inves- 
tigated; but in main features, consisting chiefly of simplification of the 
apaxial, and fusion and hypertrophy of hypaxial elements, it undoubt- 
edly resembles the Cetacea. In Trichechus the tail is usually stated as 
shovel-shaped, but in its exact conformation there is probably specific 
variation. Thus Murie (1872) figured a specimen (figure 6) of what 
he calls Aianatus americaniis { = T. latirostris) from the West Indies 
in which there is a medial indentation of the posterior tail, while else- 
where (1880) he showed an individual from British Guiana with tail 
tip somewhat pointed (figure 6) . But even yet it is not known whether 
the animals from these two regions really represent two species. The 
illustrations will give a better idea of the sirenian tail than can a de- 
scription. It will be noted that in this respect the manati is much less 
specialized than the dugong, but that the former represents a stage 
through which it is not improbable that the latter at one time passed. 
In Trichechus there is no well defined peduncle, but merely a slight 
taper of the posterior part before the lateral expansions of the tail be- 
gin. The latter is readily seen to be fairly intermediate between a stage 
on the one hand wherein there was either no lateral broadening of the 
tail, or one comparable in degree to that existing in Enhydra, and on the 
other, conditions as now to be found in Halicore. In a mounted skele- 
ton of Trichechus in the National Museum there are 22 lumbo-caudal 
vertebrae, while in a specimen of Halicore this series numbers 26. In 
the latter the transverse processes gradually diminish in width from the 
thorax to the region of the peduncle, thence widening once more until 
near the caudal tip. This is an interesting occurrence for the reason that 
it is a character which the Cetacea do not show to the slightest degree. 
It seems to be an additional instance of the fact that two animals fre- 
quently respond differently to the same stimulus. 

The caudal conditions in the Sirenia are not readily analyzed. It is 
currently believed that the sirenian ancestor was of proboscidean stock 
and we would therefore picture the caudal stimuli to have been some- 



what similar to those that the hippopotamus has experienced. One would 
naturally envision this ancestral type as swimming by alternate strokes 
of at least the posterior, if not all four, limbs, while an inadequate tail 
trailed behind. If this had been the case the stimulus would have been 
chiefly for a transversely flattened tail. In order for the tail to have at- 
tained the shape that we now find it, it is probable that either one of two 
things was an original requisite: (d)That the terrestrial ancestor of the 
sirenians had a sinuous body with long tail that was especially robust at 
base, and was of a rather active disposition. These premises seem neces- 
sary in order that it could swim with the sinous movements in the ver- 
tical plane that the otter now employs in swimming. But we surely 
have no justification for considering that any member of the probosci- 
dean stock was ever this sort of mammal, (b) The other alternative is 
that the sirenian tail developed along the lines that it followed (verti- 
cal flattening) for the same reason that the tails of the platypus and 
beaver did. Of these I regard the platypus' tail as the more significant, 
for this animal and the sirenians (at least the manati) utilize the tail 
extensively for keeping near the bottom. Murie's illustrations, repro- 
duced in figure 7, were drawn from life and show this in an interesting 

The subject of the sirenian tail should not be abandoned without call- 
ing attention to another possibility. As with most highly specialized 
aquatic mammals it is likely that the more distinct genera are of great 
antiquity. It is very likely that the manatis on the one hand, and the 
dugongs (and possibly Hydrodamalis) on the other began to diverge as 
separate groups soon after the sirenian ancestor took to the water, if not 
before. It is not improbable that this took place before the tail had ex- 
perienced any aquatic modification at all, in which event the tail of the 
manati and of the dugong has each followed, an entirely independent 
course of development, diverging in details to a greater or lesser extent 
according to varying conditions which each has experienced. Presum- 
ably the tail of the dugong is of the higher aquatic type, but this does 
not necessarily mean that it ever passed through the exact stage which 
is now exemplified by the manati. Each type of tail may now be per- 
fectly fitted to the needs of the animal and the manati may never de- 
velop the more specialized tail of the dugong. There is the possibility 
that the flukes of the latter have developed from skin folds of rather 
limited extent near the tail tip, while the tail of the manati may have 
been evolved from an appendage having lateral keels along its entire 
length. There is little evidence either for or against this possibility. 



Caudal conditions in living and known fossil Cetacea seem to be of 
three sorts ; one represented by the zeuglodont Basilosaurus, the second by 
Zeuglodon os'tr'ts, and the third by all odontocetes and mysticetes. Basi- 
losaurus and its ilk had an anguilliform type of body, the tail being ex- 
cessively long. Posterior to the thorax the zygapophyses did not articu- 
late and the transverse and spinous processes were short, indicating that 
the spinal musculature was not developed to a degree where it could 
handle flukes of the modern cetacean sort. These vertebral details show 
that there was great mobility of the tail to allow for serpentine motions 
and it seems certain that the propulsive mechanism must have been in 
the nature of some sort of continuous fin fold running in a fore and aft 
direction, although this may not have been of entirely uniform width. 
Presumably there was a symmetrically placed pair of these extending for 
most of the length of the tail. As the caudal expansion of living ceta- 
ceans is horizontal it is certainly reasonable to assume that this was also 

Figure 32. Six stages in the ontogenetic development of the cetacean flukes, 
after Ryder. 

the case in zeuglodonts, and that it progressed by undulations in the 
vertical plane after the manner of the traditional sea serpent. 

The second sort of cetacean caudal condition is represented by Zeuglo- 
don osnis. In this the tail length was comparable to modern porpoises, 
and although the high spinous processes of the thorax and anterior lum- 
bar region indicate that there was powerful musculature in these areas 
for control of the tail, the spines of the posterior lumbar region and 
tail were remarkably weak, showing that here there were no heavy mus- 
cles. It is therefore doubtful if this animal had abruptly expanded 
flukes ; and the tail was too short to have anguilliform keels upon either 
side. It therefore appears not unlikely that it was provided with a cau- 
dal equipment more nearly resembling the manati than a living por- 
poise. Incidentally there were likely at one time or another zeuglodonts 
which had a caudal propulsive apparatus fairly intermediate in character 
between that of Basilosaurus and Zeuglodon osiris. 



As far as I know there is no reason for believing that any odontocete 
or mysticete whose remains have yet been discovered had flukes very 
different from what we are accustomed to consider as characteristic of 
this order, although it is true that I have not examined the details of the 
caudal vertebrae of a great many fossils. 

The caudal equipment of the whale has been a subject for the live- 
liest controversy, as has been the case with so many cetacean details. 
Thus Gray held the belief that the flukes were derived from the entire 
hind limbs, while Ryder (1885) considered that they represent the pedes 
only. The latter's lengthly defense of this belief is a curiously artless 
combination of established fact and fancy. He argued that the original 
development in the Cetacea was along somewhat the same lines as that 
followed by the seals, and that the hind limbs were at one time used in 
oscillating movements for propulsion. He believed that the feet and tail 
were later inclosed in a single fold of integument and that finally the 
bony and muscular part of the hind limbs became atrophied and shrank 
toward the pelvis, leaving the integumentary part of the pedal expansions 
attached to the side of the tail. There were some converts to this view, 
chiefly among those who were unwilling to relinquish the thesis that 
pinnipeds, sirenians and cetaceans all represent different stages of de- 
velopment from a common derivative. It seems hardly necessary to 
point out that if this were the actual situation there would not only be 
clear and incontrovertible proof of it in cetacean embryos, but if the 
flukes had been derived from anything except fibrous, dermal dilations 
of the lateral tail the adult would necessarily exhibit some cartilaginous 
or muscular relic of the fact in this region. On the contrary I have no 
fear of contradiction when I say that in the light of present knowledge 
the evidence is conclusive that the posterior termination of the whale 
is composed of caudal elements only. 

In attempting to explain the asymmetry of the odontocete skull Stein- 
mann (1912) made the claim that the Cetacea originated from the ich- 
thyosaurs, and hence that the whale's flukes have become horizontal from 
an originally vertical position. This theory is also untenable. Had this 
been the case the whale would have experienced a period of (say) from 
one to several millions of years during which its flukes would have been 
at an angle of 45 degrees (more or less) to both the vertical and hori- 
zontal. This would have obliged one-half of the erector spinae muscu- 
lature to have become atrophied and the other half hypertrophied, while 
a similar fate, but in reverse order, would have overtaken the hypaxial 
muscles. In truth we would then have a cetacean of astounding asym- W 

metry ^tO fS^^^ 






Also in trying to explain the asymmetry of the skull in toothed whales 
Kiikenthal (1908) made the claim that the flukes are asymmetrical. He 
was led to this belief because in many of the embryos which he examined 
the caudal lobes were set somewhat awry, and from this he argued that 
a similar condition in the tails of baleen whales was accompanied by 
slight though recognizable asymmetry in the mysticete skull. His thesis 
was that an asymmetrical tail, used in sculling movements, tends to turn 
the animal to the left, when the unequal pressure of the water upon the 
two sides of the head will have resulted, throughout long ages, in an 
asymmetrical skull. At the present time this theory is given very slight 
credence. It is true that Kiikenthal figured the transected tail of a ror- 
qual which showed the flukes in somewhat oblique relation to the verte- 
brae, but even if this accurately depicted the conditions in life this in- 
dividual may have been pathological. The caudal conditions in pre- 
served embryos are entirely without significance, for the preservative act- 
ing upon soft tissue in a cramped position invariably distorts the flukes, 
and almost always they will remain fixed in the shape of an S. I have ex- 
amined numerous cetaceans, both odontocetes and mysticetes, with pos- 
sible asymmetry of the tail in mind and have never found the slightest 
indication of such being the case. 

Ray seems to have been the originator of the belief that the horizontal 
direction of the whale's flukes was attained because the animal is an air- 
breather, and such provision enables it more easily to seek the surface 
for a fresh breath. In the literature this is often repeated without com- 
ment. A little reflection, however, will convince one that there can be 
nothing to this reasoning, as Beddard (1900) and a few others seem to 
have concluded. In the first place no such function as ease of ascent 
from great depths could have had the slightest influence upon the ini- 
tial stages of tail change in the cetacean ancestor. Caudal development 
must have started in conformity with the manner of swimmng then em- 
ployed without regard to any final use to which the tail would be put. 
In the second place, when the tail is used as the primary means of pro- 
pulsion it will become specialized so as to drive the animal forward in 
a straight line, regardless of the direction of flattening, while other de- 
tails of the body will take over the function of steering. At speed the 
whale evidently elevates or depresses its line of progress by the flippers 
alone, save in the case of an abrupt turn, and it could with the greatest 
ease swim either straight up or straight down were the flukes vertical 
instead of horizontal. 



As already mentioned it seems to me highly probable that whales have 
descended from an essentially active ancestry, and that some method of 
swimming has always been employed which involved movement of the 
tail in the vertical plane, originally somewhat after the style now to be 
seen in the river otters. Successive steps in caudal development then 
would be a lateral flange upon either side, next the broadening of the 
terminal part of this flange without a corresponding increase in the 
width of the proximal portion. Following an increased tendency to- 
ward segregation in the extreme distal portion of the tail of the lateral 
expansions, there would finally result the graceful, bilobed flukes of the 
rorquals, which are probably more specialized than those of any other 

That an immense length of time has elapsed since the first step in the 
development of the cetacean flukes is indicated by embryological evi- 
dence. Ryder showed that in a Delphinapterus fetus of about an inch 
and a half in length this expansion is apparent, and of spear-shaped 
form. Indeed, the literature is replete with such evidence. Contrary to 
what one might expect this caudal expansion does not occur upon the 
peduncle proper but apparently is confined to the area of the tail tip 
which will later support the fully-grown flukes (figure 32). This fact 
may contribute some evidence to the theory that the cetacean tail was not 
comparable to that of the present manati during any stage of its evolu- 
tion, but that the lateral expansion had always been confined to the po- 
sition which it now occupies. 

Naturalists are not entirely in accord regarding the precise manner in 
which the flukes of whales are used, for opportunities for observation 
are infrequent and porpoises may move the flukes too rapidly for the 
human eye to follow. A number of observers have marvelled that when 
looking down upon a porpoise swimming at speed, as just in front of 
the bow of a boat, no movement was appreciable, and yet the animal 
not only maintained its position but easily darted ahead when it so 
wished. From this it must be inferred that the fluke movement of at 
least some porpoises is through a short arc, is rapid, and correspondingly 
very powerful. It has been claimed from time to time that the move- 
ments of the cetacean flukes are not strictly in the vertical plane but are 
slightly oblique, first to one side and then the other (scull-like) or even 
somewhat twisting (partly screw-like) . It is possible that both are used 
at times, in addition to strictly vertical thrusts, this depending upon the 
speed, the sort of whale concerned, and the conformation of the flukes. 
I think it more likely that a cetacean with relatively narrow flukes would 



employ a sculling motion, so as to reach out first on one side and then on 
the other for undisturbed water, more readily than one with very broad 

Breder (1926) has stated that in twelve high speed fish the width (or 
height) of the tail averaged 21 per cent of the total length. The ror- 
quals are the speediest of the large whales, and one which I measured 
had a length of 63 feet, with flukes 15 feet broad, so this same propor- 
tion in this animal was almost 24 per cent. The tips of the flukes were 
therefore of suflftcient length to reach well laterad into undisturbed water 
as the animal swam, and I have been assured by observers, including 
trained naturalists, that in this sort of whale there is no lateral or scull- 
ing motion to be noted when the animal is swimming. 

The shape of the flukes differs in various types of whales. Especially 
in less speedy sorts the posterior border may be straight, or this edge of 
either lobe convex in outline and with a median notch. In the rorquals, 
however, the flukes are more falcate in shape, the tips extending farther 
back and the posterior border being more suggestive of an S. Hence, 
the tail is thus inclined to be forked. Apropos of this, Breder (1926) 
has stated that fishes with squarish or spatulate tails are comparatively 
slow but capable of extremely sudden short bursts of speed, while those 
with deeply forked or lunate tails are capable of long continued swim- 
ming at high velocity, the more lunate the tail the faster being the fish. 
In discussing this fact Nichols (1915) has pointed out that during 
speedy locomotion the water displaced by either side of the body of a 
fish should, directly the fish has passed, meet again with a minimum of 
disturbance by the median part of the vibrating tail. Breder considers 
that in the main this is true, but that there are other factors involved 
is shown by the fact that when he cut a prominent fork in the tail of a 
fish normally having this member of spatulate shape, no greater speed 
was attained nor was this reduced, but the motions of the body were dif- 
ferent. It is therefore justifiable to presume that in the faster sorts of 
whales such as the rorqual, the broad, falcate, slightly forked flukes are 
of prime importance in attaining and maintaining speed, but that there 
are also other factors involved, perhaps of equal import, including, at 
least, shape of body and muscular conformation, In other words if a 
gray whale could temporarily be equipped with the tail of a rorqual there 
is no reason to believe that it could swim any better for the reason that 
it probably has not the equipment to operate such a tail in the most ad- 
vantageous manner. 

The structure of the whale's flukes is truly remarkable. The smooth 



slope of their contour and nice variation in thickness, as the situations of 
the various stresses dictate, are very suggestive of their evident fitness 
for function. The epidermis is, of course, very thin as it is over the re- 
mainder of the body, and beneath this is a pure white, fibrous tissue, 
somewhat fatty and very elastic in texture, which has phenomenal tough- 
ness and yet may be shced with a knife without difficulty. The strength 
of this tissue is forcibly impressed upon an observer as he watches a 
whale of 75 tons being drawn up the slip tail first. If the flukes catch 
beneath some obstruction either the latter is torn rrom its moorings or 
the flukes snap into place with an abruptness and strength that shakes 
the entire animal, but very seldom is any serious damage suflfered by the 

_ .^^ X 'rmsismS^^^r- 

Figure 33. Suggested restoration of caudal outlines, illustrating progressive de- 
velopment, of the Ichthyosauridae: (a) Mixosaurus ; (b) young and (c) 
adult stages of Stenopterygius ; and {d) Ichthyosaurus (after Fraas). 

flukes themselves. It is difficult to understand why, during the time 
that it took for cetacean flukes to reach their present stage of perfection, 
the caudal vertebrae themselves did not experience profound alterations 
in the way of broadening, and as a result we must believe that any un- 
yielding stiffener of such a sort would not prove advantageous. Of 
course there may ultimately be acquired within the caudal lobes an equip- 
ment of supporting cartilages, in which ossification centers might ap- 
pear, but as there is as yet no sign of any such development there is no 
reason to think that there ever will be. 

In all the Cetacea the transition from peduncle to flukes is very 
abrupt, the latter always diverging at almost a right angle. The cross 



section of the peduncle varies greatly according to the habits — to a large 
extent the speed — of the animal. In some of the slower sorts of whales 
(balaenids, gray whale, etc.) this may be almost circular and fairly el- 
liptic. I believe that the height is invariably greater than the width, 
however. At the other extreme in peduncular shape are the rorquals. 
In a finback (Balaenoptera physah/s) of 65 feet the peduncle directly 
anterior to the flukes measured approximately one foot in width and four 
feet in height. Instead of being purely elliptic the peduncle was quite 
sharply keeled above and below. It is obvious that such a shape would 
prove of very great economic benefit in reducing water resistance to the 
minimum as the animal elevates and depresses the tail while swimming. 
Sections cut from the peduncle show that in a specimen of this size the 
dorsal and ventral 8 inches of the respective keels are composed of the 
same sort of fibrous tissue as the flukes. So it is obvious, I think, that 
these narrow, angular, peduncular keels have been built up above and 
below the great spinal tendons for the sole function of reducing water 
resistance during swimming, which is just what the flukes have done to 
increase resistance. But such contradictory situations are frequently en- 
countered in any study of specialized organisms. 

Save for the fibrous keels as described above, the shape of the pe- 
duncle of the Cetacea is dependent upon details of the vertebrae of this 
region, and therefore upon the character of the musculature concerned. 
It is clear that in those sorts of whales with relatively broad peduncles lo- 
comotion is retarded by just so much, and a broad peduncle can be of 
use only to give the caudal muscles greater leverage in lashing the tail 
laterally — a movement that cannot be of very great importance to the 
animal. With a peduncle that is relatively very high or deep a whale is 
equipped not only to elevate and depress the peduncle with the minimum 
of water resistance, but also to secure, by means of long spinous proces- 
ses and chevron bones, increased leverage by which the tail may be 
raised and lowered with greater power or greater ease. 

The caudal vertebrae of living Cetacea vary in number from 16 in 
Neobalaena to 32 (or possibly more) in Steno and Lagenorhyiichus, and 
the variation in height of spines and width of transverse processes is 
great. Prezygapophyses are present in mysticetes, and in most odon- 
tocetes, but in some forms (as Grampus) of the latter these are sup- 
pressed. Chevron bones are present, taking the place below that is 
filled above the column by the spinous processes. Their development 
corresponds to that of the spines, the latter invariably being slightly 
longer. As Flower has said regarding the caudal vertebrae, in passing 



backward the arches and processes gradually disappear, and the bodies 
become compressed and elevated vertically. At a point corresponding 
to the posterior part of the peduncle there is an abrupt change in the 
character of the vertebrae, and thereafter those which in life were sit- 
uated within the confines of the flukes become smaller and broader, the 
column ending in a series of vertebrae that are no more than bony but- 
tons. To the vertebrae of the flukes pass bundles of large tendons, both 
from the apaxial mass above, and the hypaxial musculature below. 


Chapter Ten 

The "Peroral Limb 

It is a well established belief that all terrestrial vertebrates were orig- 
inally derived from a fish-like aquatic ancestry. The anterior limb, or 
ichthyopterygium, of fish has certain well defined characteristics, among 
which is the lack of clear distinction between the proximal elements 
("brachium" and "antibrachium"), the ocurrence usually of more than 
five series of distal components ("digits"), and the indeterminate num- 
ber of elements ("phalanges") of which each of these are composed, ac- 
cording to the exigencies of individual cases. In mammals the anterior 
limb, or cheiropterygium, is composed of distinct brachial and anti- 
brachial segments, the digits never number more than five, nor the 
phalangeal ones (including metatarsi) more than four except in whales. 
What is more logical than to presume that some of the steps taken by 
the mammalian stock as it arose, by whatever process and by whatever 
path, from a fish-like ancestry, should eventually be retraced in more or 
less complete degree as some of its representatives once more become 
completely fitted for an aquatic existence ? Working with such a thesis 
as a tentative basis certain of the trends which are found to be exhibited 
take on added significance. 

When a mammal first takes to the water the fore limb is usually, if 
not always, used as an aid to locomotion, and there will be a lengthy pe- 
riod during which its function is very inefficient for the reason that it 
is ill fitted for the part that it plays. The final fate of the fore limb 
in respect to aquatic modifications undoubtedly depends upon a great 
number of factors, but it seems that the chief determinant rests upon 
the question of whether or not the hind limbs or (and) the tail gain 
evolutional ascendancy over the pectoral appendages. The chances of 
the hind limbs gaining the lead are much more than even. In the first 
place, with very few exceptions (some sorts of bovines, as Bison, the 
hyenas, etc.) the hind limbs of terrestrial mammals are larger, more pow- 
erful, or both, than the complementary pair. They would then be more 
vigorously kicked, and would gradually play an increasingly important 
part in aquatic locomotion. Second, the fore limbs may be used for 
other purposes besides propulsion when the animal is in the water, as 


in helping to introduce food into the mouth or in prying up stones upon 
the bottom. Third, the most efficient method of applying the propulsive 
force in swimming is from the hinder end, and it is likely that the ani- 
mal very quickly discovers this fact. Especially if there be present a 
tail of respectable proportions and length I regard it almost as a cer- 
tainty that in mammals the fore limbs will not constitute the primary 
means of propulsion. It seems likely, then, that unless there be some 
special feature of body conformation (as a short tail) or special feeding 
habits that might introduce disturbing elements, the final, primary, pro- 
pulsive organ of an aquatic mammal will not be the pair of pectoral 

If a highly specialized aquatic mammal swim by means of oscilla- 
tions of its hinder end — either the tail or hind feet — it should have 
some anterior provision for steering and equilibration. A slow baleen 
whale, with its enormous head, might possibly get along very well with- 
out a rudder, but for an odontocete or pinniped, pursuing individual 
food items, inability to steer (i.e. make sharp turns) would spell speedy 
death by starvation. If the head be of sufficiently small size so that a 
neck of considerable length be possible (see chapter 7), then this part 
of the animal can take over most of the function of rudder. It can 
swim along in a straight line with neck retracted until such time as it 
desires to turn abruptly, when the head can be thrust sharply to the 
side and the body will follow. But at times the head has other duties 
to perform. Perhaps a speedy fish is being pursued and the endeavor is 
to seize it as quickly as possible. Then is it advisable that there be a 
separate apparatus for steering, and this is supplied by the fore feet. 
Thus in the true seals (Phocidae) there is such nice interaction between 
the head and the neck on the one hand, and the anterior limbs on the oth- 
er, that it is impossible to tell which is of the most importance in steering. 
But it is important to note that in this animal it is likely that if the fore 
limbs were held immovable against the sides steering would be accom- 
plished with just as much effectiveness, at least for a short while, al- 
though it is probable that exhaustion would follow more speedily. 

In the Cetacea matters are somewhat different. The whole body may 
be curved moderately in any direction and thus effect turning, but the 
neck is so short that it is capable of little more curvature than the thorax 
and it is likely that when the animal is progressing at speed the entire 
vertebral column is so occupied with^oscillations concomitant to swim- 
ming that other motions of the body concerned purely with vertical steer- 
ing are not often attempted, this preferably being accomplished by sim- 



pie tilting of the flippers. Being of importance in this regard to a mam- 
mal capable of high aquatic speed it is to be expected that they would 
readily become specialized in a variety of ways according to the needs 
of the particular species or genus concerned. 

There here may arise the question of whether the pectoral limbs of 
the Cetacea ever constituted a primary swimming organ. Beddard 
(1900) considered this to be quite likely and others have entertained 
the notion. No one can deny this conclusively but in view of what evi- 
dence there is it is not at all likely. There is nothing in either the oste- 
ology or myology of the Cetacea to support such a theory, purely me- 
chanical reasons would prevent it from being a likelihood, and it is not 
probable that any aquatic mammal capable of developing such a per- 
fected propeller as the cetacean flukes would have so used the flipper to 
any marked extent. On the other hand, the osteology of the dugong does 
render it possible that at some comparatively recent time its flippers could 
have been used for propulsion in quite efficient fashion; but it seems 
more probable that osteological development has been merely conver- 
gent and that its brachial musculature has been put to some other use 
than as even a moderately important aid to swimming. 

It is seen from the above that in a mammal that is highly specialized 
for an aquatic life the anterior limb should normally have either one of 
two functions; that of steering and equilibration more often, their im- 
portance in this function depending upon other bodily details as well as 
habits, and that of primary organs for propulsion more rarely. 

If the fore limb be not of great importance in steering (or swimming) , 
it may never become essentially fin-like. If it is constantly used for this 
purpose, or to apply the propulsive force in swimming, it will finally 
assume the characters of an efficient paddle or fin. In either case effi- 
ciency demands that there should be a surface of broad area which may 
be brought into action against the water, and a relatively thin border, 
which may cut through the water with the least amount of resistance that 
is practicable in connection with requisite strength. As an ideal, how- 
ever, the cross section of this paddle or flipper will not show two lines 
that are perfectly parallel, but rather will such a cross section be of a 
fusiform shape, as is the wing of a bird or of an aeroplane. For one 
thing the anterior border can not be too thin because of needed strength. 
The posterior not only can be but should be thinner to allow for greater 
limberness, and to reduce suction or partial vacuum as the flipper passes 
through the water. 



Whether for propulsion or as a speciahzed rudder the anterior hmb 
of an exclusively marine mammal may be expected to have become rela- 
tively stiff, although elastic. In effect it becomes either a horizontal rud- 
der or an oar. As such it should have a single joint, analogous to a 
row-lock, and this should be situated just within the body contour. There 
should not be additional joints and those already existing will tend to be- 
come immobile, unless there be need for them because of feeding habits. 

In attempting to determine the ideal position for the flipper of a ma- 
rine mammal that uses this member for steering there are encountered 
a number of interacting factors which render the question a difficult one. 
It is apparent that in the whale the most effective position for simple 
steering fins would be in the vertical plane, one above and another be- 
low, so that by slight tilting the body would be thrown to the right or 
left: but such an alteration in fore limb posture would be impossible. 
The same act could be accomplished by carrying the limb horizontal but 
tilting it so as to offer as much resistance to the water as possible and 
at the same time pressing the opposite limb against the body. This 
concerns only rudder action. But some equipment for pure equilibration 
should also be advantageous. If a whale swim but languidly with only 
the tail proper involved then this may not be necessary, but if it swim so 
violently that there is no real interval between the cephalic and caudal 
amphikinetic parts, there will be a greater or lesser tendency for the 
thorax to move upward or downward at each stroke of the flukes. In 
order to reduce this lost motion to a minimum, large flippers widely ex- 
tended would be of the greatest aid. In addition, an animal provided 
with flippers of this character so held would find it easiest to depress or 
elevate the head by slight tilts of these members. Thus it is seen that 
in an animal such as a whale a pair of horizontal flippers should best ac- 
complish elevation and depression of the body (equilibration proper) 
and by operating one at a time these might be just as efficient in steer- 
ing from side to side as if they were vertically placed. 

The above argument sounds logical but it may be complicated to a 
large degree by yet another factor. Back in the early history of the 
world when the first vertebrates were being evolved these developed an 
equipment of primitive fin folds. Opinion differs as to the ideal arrange- 
ment of these and there are proponents of two different theories, but all 
are agreed that there was a pair situated latero-ventrally from which 
were derived all four vertebrate limbs. The point which is of interest 
in the present connection is that there was evidently an elemental stimu- 
lus connected with aquatic requirements which operated to develop pec- 



toral fins at a certain angle to the body axis. It is by no means unlikely 
that there is now the same stimulus, but in reverse order, operating in the 
case of aquatic mammals, to place the axis of the fore limbs at some par- 
ticular angle. Such a stimulus is impossible of analysis, and the question 
seems too speculative to follow further. 

In the case of a mammal such as the sea-lion which swims exclusively 
by means of the anterior limbs, there are certain principles of efficiency 
which must govern the motions employed. This does not concern low 
speed, during which several sorts of makeshifts may be employed, some 
for relaxation and some in pure fun. During leisurely progress the 
flippers may be advanced and then brought rearward with a broad sweep, 
this corresponding to the overhand stroke employed by human swim- 
mers. It is true that during action of this sort there is an active recov- 
ery motion which, during the forward stroke, must substantially increase 
water resistance, but what resistance is added at this time is subtracted 
during the backward stroke and after much thought I have come to the 
conclusion that as far as concerns only resistance it makes no practical 
difference whether the flippers, during slow progression, are held with 
their forward borders in the same transverse plane, or whether they are 
alternately advanced and retarded. 

The above "overhand" movement is not employed by sea-lions for 
speedy progression, so far as my experience shows. The reason for this 
seems to be that when a certain degree of speed has been passed, the 
animal is incapable of operating the long flippers on the backward stroke 
sufficiently fast to supply any propulsive force. The principle is some- 
what the same as that which correlates running speed with the celerity 
with which the legs may be moved. Another deterent is the fact that 
during such a backward stroke the broad surface of the flippers would 
be presented in an antero-posterior direction. Hence, even though the 
tip of the flipper could be retarded sufficiently quickly to supply some 
propulsion, the broad part of the wrist would be so situated as to act as 
a strong brake. By this method of overhand swimming, therefore, 
there appears to be a certain unknown speed limit which sea-lions are 
incapable of exceeding. So something more efficient must be evolved — 
some way of utilizing an oblique, rather than a direct, thrust against the 

As a result the Otariidae now swim at high speed not by flexing and 

extending the arm, but by adducting and abducting it, with hardly any 

antero-posterior action at all so far as I can determine. As the flipper is 

adducted it is turned obliquely or "feathered," so that the thicker an- 


if ' 

o Q □ a ta ca C3 a a CSlCSi 
.= racscsi ra C3 O C3 Cr3K)' 


Figure 34. Left pectoral limbs of some aquatic reptiles: {a) leatherback turtle 
(Dermachelys) ; (b) Geosaurus ; (c) Clidastes (a moasaur) ; {d) Op- 
thalmosaurus (an ichthyosaur) ; and {e) Elasfnosaurus (a plesiosaur). The 
last four are redrawn from Williston. 



terior border really supplies the force, and precedes the posterior bor- 
der, the motion of which is more passive. By this method of swimming 
the possible speed is very great indeed, and in practice depends directly 
only upon the power with which the flipper is adducted. Another ad- 
vantage is that this need not entail any lost motion, for abduction, al- 
though necessarily much weaker because of muscle conformation, con- 
sists of the same motions as adduction but in the opposite direction, and 
can contribute at least some propulsive force. Thus in swimming by 
adduction and abduction in the transverse or largely vertical plane all 
movements can be utilized for forward propulsion and the only resis- 
tance is offered by the anterior borders of the flippers. It is undoubt- 
edly by just this method, or one substantially the same, that aquatic 
birds, including penguins, which habitually pursue speedy prey by "fly- 
ing" under water utilize their wings; but the case of the marine turtles 
is different. 

I have had but limited opportunity for observing the actions of pen- 
guins, but so far as my experience goes the motions of their pectoral ap- 
pendages differ in no important respect from those of sea-lions. Move- 
ment is almost entirely in the abductive-adductive (transverse) plane, the 
wings are "feathered" during the stroke so that the force applied is ob- 
lique, and although the abductive or upward stroke is too fast to follow 
satisfactorily it seems likely that it is so performed that it furnishes at 
least a slight amount of forward propulsion. There is the difference, 
however, that in the penguin the static posture of the paddles is al- 
most horizontal, while in the sea-lion it is more adducted, and the arc 
of movement varies accordingly. 

The swimming movements of the marine turtles are rather hard to 
describe. The elbow protrudes from the body contour and the enormous 
humerus is worked chiefly up and down. The forearm segment, how- 
ever, is to all intents a mechanical part of the paddle and the whole is 
extended (see figure 34) with respect to the humerus in a manner never 
encountered in the Mammalia. By virtue of this alteration at the elbow 
the turtle's flipper is given a definitely caudal inclination, so that al- 
though the humerus is chiefly worked dorsad and ventrad from a trans- 
verse position, the flipper, with its axis almost parallel to that of the 
body, operates by thrusting the water chiefly backward by means of the 
palm. Perhaps this situation has some bearing upon the fact that hyper- 
phalangy is not met with among the turtles. 

By the above it is not meant to imply that no other swimming motions 
are indulged in by either penguins or turtles. Those described are 



merely believed to be the most efficient ones by means of which their best 
speed is attained. 

This outhne of the known factors responsible for fore limb conforma- 
tion and action in highly specialized aquatic mammals has been given 
first in order that the reader may have a better understanding of what 
follows. The details to be discussed in the present chapter have such 
great interdependence that it is difficult to arrange the subjects properly 
and to avoid some repetition. 

Clavicular conditions vary greatly among all sorts of mammals and it 
is not easy to distinguish just the critical factor that determines the 
presence or absence of this bone. In general it may be said to be lack- 
ing in those mammals which use the pectoral limb for support only, 
and present in those which are in the habit of using the hands for grasp- 
ing; but there are many exceptions, to the latter statement especially. The 
clavicle should be considered not in the nature of a strengthening mem- 
ber, but rather as a strut to prevent an undesirable degree of adduction of 
the shoulder. In heavy mammals that bound about, landing solidly upon 
the fore legs, it would be in danger of breakage, and is accordingly ab- 
sent; nor does there seem to be much need for it in the case of the more 
narrow-chested mammals. 

The clavicle is lacking from all the more highly specialized aquatic 
mammals — pinnipeds, sirenians and cetaceans — it is functionally ab- 
sent in the carnivores, usually present in rodents but absent in a number 
of heterogeneous sorts (including the capybara) , and present in all in- 
sectivores except Potomogale, and presumably Limnogale. The latter 
is really the only significant fact. For all we know the terrestrial an- 
cestors of pinnipeds, sirenians and cetaceans may have lacked a clavicle. 
It seems reasonable to suppose, however, that the way in which whales 
use the flippers would introduce a stimulus for the elimination of the 
clavicle were one present. On the other hand, one would surmise that 
the otariids might find a clavicle of advantage to the way in which they 
constantly adduct the flippers during swimming. 

It is perhaps unsafe to attempt to analyze the muscular factors under- 
lying the attachment of the shoulder to the body of aquatic mammals, 
but existing conditions may be mentioned and a few possibilities ad- 

A trapezius is lacking in whales but is present in both pinnipeds and 

sirenians, and it seems fully as likely that this muscle has always been 

absent from the cetacean stock as that it has been eliminated by aquatic 

habits. It is not known what scapular motions a whale finds of advan- 



tage to its well-being, but there appears to be a rather uniform plan of 
scapular suspension employed in this order and it is likely that very 
few, or very circumscribed, movements of the shoulder are indulged in. 
There is a tendency apparent for reduction of the width of the shoulder 
attachments, which are more segregated into three areas than is the usual 
case: below and slightly to the rear there is a narrowed pectoralis (either 
single or double) ; above and slightly to the rear a broad rhomboid, 
which may be continuous along its border with a narrowed latissimus 
dorsi ; and anteriorly there is anchorage to the atlas and mastoid region 
by two or three narrow muscles (atlanto- and mastoscapular, and masto- 
humeral). The serratus magnus may be either very narrow, or very 
broad as as to function as a very efficient antagonist to the extensive 

In pinnipeds the suspension of the shoulder is according to a different 
principle. The dorsal, ventral and cranial anchors are all spread to a re-, 
markable degree, allowing powerful movement in any and all directions. 
There are three widely-spreading trapezius divisions, the anterior rhom- 
boid reaches the head, and other shoulder muscles are specialized ac- 

There is no movement of the shoulder involved in the act of swim- 
ming by the true seals, and yet it is inconceivable that such specialized 
shoulder musculature could constitute merely a phylogenetic inheritance 
from a more specialized ancestor. If we look upon it as having de- 
veloped for the purpose of lending all possible assistance to the act of 
swimming, then is it more understandable. The action is entirely too 
complex for simple analysis, but after much study of the question I must 
believe that the end toward which the shoulder muscles of these animals 
have striven is for the purpose of accentuating the lateral movements 
of the hinder end of the body in one direction, and the forward end in 
the other. The muscles anterior to the shoulder would in this case act 
largely as antagonists to those posterior thereto, and the shoulder is thus 
comparable functionally to a sort of raphe between the two groups. The 
chief muscles concerned are the phenomenal pectoralis and latissimus 
dorsi on the one hand, and the cephalohumeral and humerotrapezius on 
the other. Movement of the shoulder muscles other than those of im- 
portance to swimming may be relatively inddental in this animal. 

In the Otariidae or sea-lions the functions of the shoulder muscles are 

very different from those in the seal. The shoulder proper does not play 

a largely passive part during swimming, but an essentially active one. 

Some of the extrinsic muscles (chiefly the pectoralis, but the cephalohum- 



eral and latissimus are also of major import) are directly concerned with 
flipper movement, while others have the function of adjusting the po- 
sition of the shoulder itself. It is certain that these latter adjustments 
are extremely frequent and extensive. The part which they play in actual 
swimming is unknown, but during terrestrial activity the scapula slides 
about beneath the skin in a quite surprizing fashion. The vertebral bor- 
ders of the scapulae may appear to meet well above the back bone, or 
they may be slid far ventrad. The pectoralis is of course largely re- 
sponsible for the latter act, and the humerotrapezius is peculiarly fitted 
for elevation of the scapula, inserting upon almost the entire length of 
the humerus and with only incidental attachment to the spine of the 

It is only from the pinnipeds, sirenians and cetaceans that we can hope 
to le'krn anything regarding scapular tendencies in aquatic mammals. 
Beyond any question this bone has undergone some degree of broaden- 
ing in both sea-lions and whales. There may be considered to have been 
some tendency in this direction in the manati, and in some seals, while 
in the dugong and the majority of seals the scapula assumes a more falci- 
form shape. It may safely be assumed that the variation in the general 
shape of the scapula, including the position of the spine, is due entirely 
to muscle stress. In Cetacea this bone is usually quite broadly fan- 
shaped, more so in some and less in Physeter, in which the flipper may 
be presumed to be less efficient as a rudder because of the unwieldly size 
of the head. Invariably the infraspinous space occupies practically the 
entire lateral aspect of the bone, while the bony area of the supraspinous 
fossa is insignificant and perhaps but one-hundredth as large, or occa- 
sionally it actually does not exist, as in Platanista (figure 36) and Me gap - 
teva. The supraspinous muscle is, then, of decreased importance. But 
the infraspinatus is not relatively larger so as to fill the infrapinous space. 
On the contrary this muscle covers but a half or two thirds of the lat- 
ter; but its details are variable. We must therefore seek other muscles 
that have had need of a greater angle of leverage and have accordingly 
stimulated the broadening of the scapula. Judging by conditions in 
Monodon and Neomeris it is likely that in toothed whales this stimulus 
has been supplied either by the subscapularis, which covers the entire 
medial surface of the bone, or (and) the deltoid, which in the latter 
genus arises from the entire, and iii the former, almost the whole, verte- 
bral border. Contributing to the situation may also have been a stimu- 
lus for extension of the glenovertebral angle by the teres major and (or) 
serratus magnus, which latter in Monodon is especially extensive. In at 



least some of the Mysticeti the stimulus for broadening of the scapula 
has resulted in an extreme development at the glenovertebral, and to a 
lesser extent at the coracovertebral, angle of the suprascapular cartilage 
(figure 36) . The chief reason for this extreme cartilaginous extension 
in a posterior direction is seen in figures given by Schulte (1916) to be 
probably the serratus magnus muscle, attached to the posterior part. This 
muscle is thereby given an especially efficient lever arm for operating as a 
depresser anguli scapulae, ostensibly of great use in tilting the flipper for 
equilibration. It should also be mentioned that in Mysticeti the deltoid 
has become differentiated both in origin and insertion for assuming in 
even fuller degree the function normally held by the supraspinatus. 

In the Cetacea the condition of the acromion is quite curious and there 
is apparently no good reason for its existence. It does not increase the 
leverage of any muscle, for nothing is attached to it save inciderftally. 
The same may be said of the coracoid process. Both may be developed 
to an extraordinary degree, as in Sibbaldiis for instance, or both may be 
entirely obliterated, as in Megaptera (a southern species of this genus 
is said to have a short acromion), apparently with an equal lack of 
reason for both conditions. 

It is interesting to note that the scapula in the zeuglodont Basilosaurus 
is exactly what one would expect it to be supposing that it represents a 
stage through which modern whales have passed. In shape it is entirely 
whale-like. The supraspinous fossa is relatively smaller than usual but 
still much larger than in any whale, being about a third or two-fifths 
of the infraspinous fossa. Along the glenoid border there is also a 
fossa for a strong teres muscle. The acromion is greatly developed. 

To a large extent scapular conditions in the sirenians resemble those 
in the pinnipeds, but Murie did not give sufficient detail regarding this 
part of the manati to make it advisable for me to pursue the subject fur- 
ther. Noteworthy in the manati is the irregularity of the spine, appar- 
ently attributable to some detail of the deltoid, and the pointed acromion. 
In the dugong the acromion is strongly distinct but it projects at a right 
angle to the scapular plane. 

Ostensibly the medial muscles of the sea-lion scapula must be those 
which have been instrumental in causing the broadening of this bone, 
for the lateral ones are very different from those in whales. The supras- 
pinous fossa is considerably larger than the infraspinous space, while 
in the Phocidae it is smaller. In both groups the origin of the infraspin- 
ous muscle occupies about half the space posterior to the spine, but the 
muscular conditions over the remainder of this area are so involved and 




diverse that but little more can be deciphered. There seems to be but 
Httle doubt, however, that the more falcate shape of the scapula in Phoca 
signifies a more caudal extension of the glenovertebral angle for the chief 
purpose of furnishing increased leverage for the enormous triceps, rather 
than to any extrinsic limb muscle. 

In all aquatic mammals the ball-and-socket character of the shoulder 
joint is retained. In the Pinnipedia, Sirenia and the Zeuglodonts (at 
least in some of them) the synovial character of the elbow joint is also 
retained, and this is true also of the carpal bones of pinnipeds, save that 
in the Otariidae the mobility of the latter region is reduced. In modern 
whales, however, the joints distad of the shoulder have all lost their 
synovial character and instead are entirely fibrous. That part of the limb 
that projects beyond the body contour of whales is thus incapable of 
bending movement save that the fibrous interosseous tissue gives con- 
siderable elasticity to the flipper. This character of resilience is appar- 
ently of much value and it is to be expected that it will increase to a 
certain ideal optimum. One way in which this might be accomplished 
is in the reduction in length of the interfibrous (i.e osseous) elements. 

The advantage of an elastic but non-jointed paddle to a whale is too 
obvious to need discussion. Apparently it might be of equal desirability 
to a swimming sea-lion, but this animal is obliged to have adequately 
bendable joints in the arm if it is to continue movement on land. Al- 
though the latter function is of no use to a sirenian, it does need to bend 
the elbow during feeding, for the flipper is then employed for drawing 
herbage toward the mouth, and Murie mentioned that in the manati the 
joints are very lax and their ligaments simple. 

But a fore limb in the shape of a true paddle is needed by whale, 
sirenian and sea-lion — not a spatulate enlargement of the manus upon 
the end of a long arm. I think this is obvious and that it may be accepted 
without argument that the normal tendency in such aquatic mammals as 
whales and sea-lions is for increase in relative size of the paddle part 
(the manus) and decrease in relative size of the non-paddle part (the 
antibrachium and humerus) . There undoubtedly are several different 
factors that may help this development, and likely some unknown ones 
that tend to hinder it. 

In such a non-jointed paddle aswe may presume to be ideal for pro- 
pulsion or equilibration in the water the only essential limb muscles are 
those which operate the arm as a whole, bending it upon the shoulder 
joint. Not only must there be adequate provision for flexion and ex- 
tension, abduction and adduction, but also for tilting or rotation, in or- 



der that a whale may elevate or depress its line of progress. It is clear 
that small muscles inserted near the head of the humerus would be cap- 
able of but feebly waving about a long arm with broad paddle upon its 
end. For efficiency the arm should be shortened, as already argued, the 
critical muscles should be strengthened, and their effective leverage in- 
creased by a migration of their insertions distad from the head of the 
humerus. Such alteration in muscle attachment will effect alteration in 
the bones, and this in turn may greatly change the functions of the mus- 
cles involved. 

As already discussed there is no longer any reason for aquatic forms to 
hold the limbs in vertically dependent posture and the tendency, if un- 
complicated, is probably for these members to be held at an angle of 
about 45 degrees. We have no means of knowing the exact angle fav- 
ored by various sorts of cetaceans, but the osteological evidence would 
indicate either that in most porpoises the habitual posture of the flippers 
is more abducted than in Mysticeti — which seems unlikely — or that the 
chief u'ork performed is instigated from a position with the flippers 
more elevated or abducted in the former cetaceans. Theoretically this 
chief work should consist of strong downward movements of the flippers 
after they have first been elevated, ostensibly for elevating the anterior 
end of the animal. 

Incidentally it should be mentioned that at least most cetaceans prob- 
ably cannot extend the flipper forward to an angle greater than 90 de- 
grees to the body axis, if indeed even this much extension be possible. 
The point at issue, however, is that whales abduct (elevate) the limb 
at the shoulder joint to a considerably greater amount than does the aver- 
age terrestrial mammal, the degree depending upon the sort of whale. 
Thus, whereas the limb movement in most mammals is fore and aft, or 
by extension and flexion, in whales it is rather in the transverse plane, 
involving abduction and adduction. This change is accompanied by cer- 
tain definite alterations in the shoulder, which are reflected in the hu- 

It is frequently stated that there has been rotation of the cetacean hu- 
merus. I do not altogether approve of this term as it is somewhat mis- 
leading, and the process has evidently been entirely different from that 
experienced by man, in which the upper arm has been rotated by a shift 
in the usual position of the elbow joint. In whales the elbow is in the 
plane usual in Mammalia. In the" Mysticeti the elements of proximal 
humerus have not undergone much alteration in position, but in the 
Odontoceti they have, apparently not by any twisting of the bone but 



by individual shifting of the items. In the whalebone whales the hu- 
meral head may be said to occupy a caudo-lateral position in respect to 
the axis of the shaft indicating that there is not such definite or else 
such habitual flexion of this segment as in most mammals, and that it is 
held somewhat more abducted. The tuberosities have not altered their 
positions. The lesser, situated mediad, is practically undeveloped as a 
process and is indicated merely by a pronounced rugosity and slight 
elevation of the bone. The determinant in developing this region into 
a true process is the subscapularis, although for Balaeiioptera borealis 
Schulte (1916) showed the coracobrachialis and mastohumeralis as also 
attached upon this area. This lack of development of the lesser tuber- 
osity indicates that either the subscapularis is unusually weak, which its 
extent belies, that it normally operates when the arm is much adducted, 
or else that it has a somewhat altered function, for instance to effect ro- 
tation of the humerus when the arm is considerably flexed. I regard 
the latter as the most likely of these three possibilities, although there is 
nothing else to indicate that this is the case. 

In the Mysticeti the greater tuberosity is fairly well developed and 
may be practically as high as the head. According to Schulte's figures 
the reason for this is partly, though perhaps unimportantly, the infras- 
pinatus attachment, and (chiefly) the deltoid, which here inserts. We 
must presume for the present that this muscular condition is also found 
in other mysticetes. It will thus be seen that although this process ap- 
pears from an exclusively osteological viewpoint to be homologous with 
a greater tuberosity it is hardly so except in position, and is actually a 
deltoid process. The inference then, according to Schulte's figures, is 
that in at least one mysticete the deltoid has gradually been altered so 
that its origin occupies the scapular area normally held by the supraspina- 
tus, and that its insertion enables it to function in the same way, except 
that the latter also stretches far distad and onto the forearm in a way 
that no supraspinatus ever does, thus not only effecting moderate exten- 
sion from a more flexed posture of the arm (shown by the height of the 
process) but also effectively aiding to maintain with a minimum of effort 
such moderate extension as far as an angle of 90 degrees with the body. 

In odontocetes the details of the proximal humerus are quite dif- 
ferent. Instead of the head being. located somewhat toward the rear of 
the shaft it is usually situated to the side, and the lateral side at that. 
Conformation of the tuberosities is essentially variable, undoubtedly re- 
flecting important differences in the muscular equipment. On the whole 
the lesser tuberosity may be said to occupy its normal position, at least 



in those sorts examined, but because of the shift of the head this process 
is so located as to allow the subscapularis, which inserts upon it, to act 
even more directly as an adductor than usual. If the habitual posture 
of a humerus be vertical (to the body axis) then the subscapularis can 
operate efficiently upon a lesser tuberosity that is either low or poorly de- 
fined. If the humerus be held markedly abducted then equal efficiency 
will demand a lesser tuberosity of great height and projection. The lat- 
ter is the case in such an odontocete as Tursiops, in which the tuber- 
osity is just about as large as, and higher than, the head, indicating in 
connection with the lateral situation of the latter that the normal posture 
of the flipper may be a pronouncedly abducted one. But this process 
is not so high in a number of other toothed whales, and there is even 
some variation in its precise situation. 

As there is no occasion to extend the arm beyond an angle of 90 de- 
grees with the body axis there is not only no muscle corresponding in 
function to a clavoacromiodeltoid, save a weak mastohum.eralis poorly 
placed for this purpose, but the supraspinatus, normally an extensor of 
the humerus, has not only suffered enormous reduction but its function 
has changed. In those sorts dissected its insertion has shifted mediad to 
the anterior border of the lesser tuberosity and it accordingly acts as a 
rotator to elevate the anterior border of the humerus. The infraspinatus, 
also normally inserting upon the greater tuberosity, has shifted its at- 
tachment distad and slightly laterad and now is inserted chiefly into a 
fossa, very characteristic of most odontocetes, situated upon the hum- 
eral shaft, so that it acts not only as an upward rotator of the anterior 
border but also effectively as an abductor of the arm. Usually in mam- 
mals the greater tuberosity has developed into an eminence because of 
the stimulus supplied by the insertions upon it of the supraspinatus, in- 
fraspinatus, and teres minor, but in the toothed whales the first two mus- 
cles have shifted elsewhere and the last does not occur as a distinct di- 
vision, so the greater tuberosity has ceased to exist as a process strictly 
homologous with the eminence to which this term is applied in most 

In a number of toothed whales which I have examined there is some 
variation in the conditions of the proximal humerus as recounted above. 
Slight eminences, situated in this or that direction from their situation in 
the dissected specimens bespeak corresponding muscular variation, and 
there may actually be a prominence which might be mistaken for a 
greater tuberosity of slight definition, but it is believed that this is only 
analogous, rather than homologous, as discussed below. 



In those odontocetes dissected the extremely robust deltoid inserts 
upon the whole lateral face of the distal humerus, acting as a powerful 
abductor of the arm, but this insertion is thicker and stronger craniad and 
it likely has some rotating action as well. The latter fact is not shown 
osteologically in all toothed whales but in fully adult Tursiops, for in- 
stance, it seems clearly indicated by a very pronounced process upon the 
distal third of the anterior aspect of the humerus. This strongly im- 
presses one with the fact that if the latter protuberance were shifted 
more to the proximal end of the bone, as is largely the case in some in- 
dividuals of Kogia, it would form precisely the same sort of "greater 
tuberosity" as appears to have been caused in the Mysticeti by the deltoid. 
There is a similar process, evidently attributable to the deltoid, in zeu- 
glodonts, but located just distad to the middle of the shaft. 

The above muscles taken together seem to show that in the Cetacea 
there is a reduced power of extension of the arm, and increased power 
of abduction, and perhaps a slightly greater power of both adduction and 
rotation upward of the anterior border. The latter act is also markedly 
assisted in mysticetes by flexion of the serratus magnus. There is not 
apparent any special provision for depression of the anterior border of 
the arm. The general conformation perhaps makes this motion not so 
necessary, or it may be effected by special action of the latissimus dorsi 
or other muscles. 

The distal humerus of the Cetacea lacks the stimulus supplied by 
functional antibrachial muscles and a synovial joint. Accordingly the 
lateral and medial elevations of the condyles have atrophied. Even were 
there absent some undefinable stimulus for a broadening in an antero- 
posterior direction of all the arm bones, which indubitably exists, this en- 
tire lack of reason for condyle definition would likely be sufficient to al- 
low the distal humerus to broaden out in an antero-posterior direction to 
conform to the extent in this plane of the antibrachial bones. 

It is unfortunate but nevertheless true that when I was preparing a 
previous paper on the pinnipeds (Howell, 19'29) I was unable clearly 
to observe all the motions followed by a sea-lion when swimming at 
speed. I gained a satisfactory understanding of movements during more 
leisurely progress and made the mistake of supposing that these were 
used at a faster gait, for I had not then encountered the correct condi- 
tions of light and clarity of the water to see them dart at speed deep 
below the surface. This course seemed entirely justified for the reason 
that the posterior part of the pectoralis and the latissimus dorsi are so 
well developed for executing powerful backward thrusts. The statements 




which I then made were that the sea-hon swims by advancing the flippers, 
not in unison nor yet in alternation but with a sort of galloping move- 
ment, and then by partial rotation so as to present the broad aspect of 
the flipper, progressed by means of strong backward sweeps of these 
members. More recently, however, I have succeeded in obtaining clear 
views of the whole process. Undoubtedly the muscles mentioned ivere 
developed for backward thrusts at a time when the animals were less 
specialized and as yet incapable of the speed that they now attain. At 
present, however, they are too speedy for just this method of propulsion 
to be efficient, as already discussed at the beginning of this chapter, and 
for mechanical reasons rather than purely muscular ones, they are 
obliged when progressing as fast as possible to do so by means of adduc- 
tive and abductive thrusts of the flippers, with the minimum of ex- 
tension or flexion. The anterior border rather than the posterior is where 
the force is applied, while the latter is allowed passively to follow 
through, so that there is an oblique thrust and both adduction and ab- 
duction are utilized for propulsion, although the latter undoubtedly is 
productive of much less power, possibly an insignificant amount. 

Unlike the case of the whales the flipper of a sea-lion is not stiflfened 
by fibrous joints. The elbow does not cut much of a figure in the bend- 
ing of the external arm for it is close to the body contour, but although 
possible flexion at the wrist is not great, extension at this point is through 
an angle of 90 degrees to allow for terrestrial progression, and this 
must be overcome by mechanical means during swimming so that the 
wrist will not bend backward during the adductive stroke, and so that 
muscles will not be wearied by continual effort to prevent such bending. 
Extension of the arm is no more needed than in whales, the degree of 
abduction need not be greater, but adduction and flexion must, or at least 
should be, greater. In relaxed posture the radial or anterior border 
should be presented straight, forward, for this position is the mean from 
which pronation and supination are instigated. Accordingly the normal 
terrestrial position is for the manus to extend from the wrist in an exactly 
lateral direction rather than somewhat craniad as usually shown in 
mounted skeletons. ^ 

In the seals (Phocidae) the pectoral limb is not used in swimming, as 
already discussed, save during such acts as turning. It is supposed to be 
employed for such purposes as scratching holes in the ice, and its strictly 

^This inaccuracy will be noted in the photograph (fig. 14) of the fur seal 
herewith depicted, which is the same as that shown in my previous paper re- 
ferred to. 



terrestrial uses are very incidental, save in the elephant seal ( Aiiro/niga), 
in which the manus helps support the weight. It is usually kept folded 
back against the body with the segments markedly flexed so that all of 
it save the manus is contained within the body contour. Extension is 
possible, however, and I have seen an animal on the steep margin of its 
pool stretch forth the manus so that as well as I could judge the arm was 
extruded from the body contour as far as the elbow. 

The normal or static position of the humerus in respect to the scapula, 
determined to the best of my ability during dissection, differs only 
slightly in sea-lions and seals. The angle formed by the humeral axis 
with the spine seems to be slightly less in the latter, showing a greater 
degree of flexure, but the difference is not sufficient to be of much signifi- 
cance. The position' of the humeral head in relation to the shaft is 
slightly more mediad in the sea-lion. This, I judge, is less marked than 
one might expect, in consideration of the fact that the most efficient 
method of swimming is by abduction-adduction movements, but this 
development may be so recent that it has not had sufficient time to have 
had marked osteological effect, and after all many other motions of the 
humerus are of great importance. In phocids the head is more posterior 
to the axis of the shaft, showing that extension-flexion motions are of 
more pronounced import, and its conformation is such as to indicate 
that possible flexion of the humerus in relation to the scapula is more 
extreme in the seal. 

In the sea-lion the greater tuberosity is markedly higher than the head 
and than the lesser, while in the seal it is much lower, the lesser tuber- 
osity having greater elevation. The reason for the height of the greater 
tuberosity in the sea-lion seems solely attributable to the supraspinatus 
muscle, acting chiefly as an extensor, and the strength of this muscle 
is shown in the scapula also by the great extent of the supraspinous 
fossa. The infraspinatus, chiefly a rotator, also inserts upon this process, 
but more proximad, and its position in Phoca is quite comparable, but in 
this animal the supraspinatus is evidently not required to act so strongly 
as an extensor, for not only is the supraspinous fossa of the scapula much 
smaller, but the greater tuberosity, upon which it inserts, is hardly higher 
than the head and very much lower than in the sea-lion, giving this mus- 
cle reduced leverage. 

The conformation of the lesser tuberosity in the two pinnipeds con- 
sidered is not altogether what one" would expect. In the sea-lion, al- 
though robust and subtended by a heavy ridge, it is rather low and not 
nearly as high as the head; consequently much lower than the greater 



tuberosity. Its height proximad has nothing to do with the strength of 
the attached muscles, of course, for this detail depends upon the opti- 
mum angle of leverage. Its relatively low height, then, would indicate 
that the muscles operating from it are chiefly flexed when the arm is in 
a rather adducted posture. Not only is the subscapularis well placed in 
respect to the head for leverage in adduction, but insertion of the 
episubscapularis (absent in seals) is located still farther distad along the 
ridge, giving added power. What one might not expect in this animal 
is that the lesser tuberosity is not situated directly mediad of the head 
(to be most efficient in adduction) but medio-craniad. This situation 
might easily translate movement that might otherwise be adduction 
to one largely of rotation. This, in effect, may be actually what occurs 
at times; or flexion of the same muscles may be productive of pure ad- 
duction when other muscles are used as antagonists. It seems that chief 
among the latter might be the teres major, whose insertion is broadly 
along the medial shaft of the humerus distad of the middle of the bone. 

In the seal these medial insertions are disposed upon the humerus for 
different action. The elevation of the lesser tuberosity is quite surprising, 
it being much higher than either the head or the "greater" tuberosity. 
The attachment of the cephalohumeral thereto can hardly account for 
its height, and the only reason apparent is that the chief work of the 
subscapularis, and attendant subscapulo-capsularis (absent in sea-lions), 
is performed when the arm is pronouncedly abducted or elevated. This 
might well take the form of a strong downward heave of the manus 
while swimming for the purpose of quickly elevating the body, or while 
scratching holes in the ice. Other muscles of this region are less well 
situated for adduction of the humerus than in the sea-lion. There is no 
episubscapularis and the insertion of the teres major is farther craniad 
in respect to the head, theoretically giving a greater rotating action to 
this muscle. 

The deltoid ridge, comprising a proximal continuation of the greater 
tuberosity, is phenomenally developed in the pinnipeds, and a complex 
of muscular stimuli is brought to bear upon it. In the seal this ridge 
with tuberosity is but little more than half the length of the bone, while 
in the sea-lion it is almost two-thirds, and in the latter especially the 
concerned muscles accordingly have a phenomenal leverage. While in 
the seal the humeral attachment of the cephalohumeral is confined to 
the greater tuberosity, in the sea-lion it extends over the entire length 
of the deltoid crest. In the seal this is the case with the humerotrapezius, 
while in the sea-lion the insertion of this muscle is for practically the 



entire length of the humerus. In both animals the pectoral is double, 
one division being confined to the deltoid ridge and the other being al- 
most as long as the whole humerus. In the seal the atlantoscapularis 
inferior inserts upon most of the deltoid ridge, while in the sea-lion this 
muscle does not reach the humerus but inserts upon the spine of the 
scapula. Most of these insertions are by fascia and although it would 
be utterly unjustifiable to say that the muscles mentioned have been only 
of secondary importance in the development of such a high deltoid crest, 
still it may be stated that all of them would appear to be equally effective 
if operating upon a humerus without a crest. The large deltoid, how- 
ever, which operates chiefly as a flexor by virtue of the shortness of the 
humerus, is given the function also of a powerful rotator by the lateral 
definition of this crest. 

Little more need be said regarding the pinniped pectoralis, for it has 
already been considered in preceding chapters. In Phoca the main de- 
velopment of this complex is posteriorly, for swinging the hinder end 
sidewise, but the part medial to the arm is also very powerful, either for 
quickly adducting the arm for steering, holding it firmly adducted while 
the muscles anterior and posterior from the arm are used in swimming, 
or probably both. In the sea-lion although the abdominal pectoral is 
powerful it is far less so. It can be used to help control flexion of the 
arm during terrestrial progression or backward sweeps of the flipper 
when the need arises. In the seal the insertion of the pectorals upon 
the humerus gives unusually efficient leverage, and in the sea-lion this 
is even better developed. Not only is there double insertion over the 
length of the humerus but extension of a superficial sheet onto the 
forearm, giving the greatest possible leverage for strong and constantly 
repeated adduction of the appendage. Flexion of the part inserting 
over the deltoid crest also greatly aids the rotation of the anterior 
flipper border that is so necessary during swimming. 

The significance of most of these extrinsic muscles of the humerus 
has already been discussed in relation to the neck and other parts of 
the body, but they must receive further consideration in the present 
connection. Although these muscles have many significant similarities 
in seals and sea-lions, yet their functions appear to be very different. 
It seems almost certain that in the sea-lion the muscles extending from 
the humerus to the head and neck are used chiefly for extending the 
arm craniad, both when the need arises while in the water and during 
progression by a lunging gallop while on land. And similarly that 
the abdominal pectoral, latissimus, and panniculus may be likewise used 



for flexing the arm. It appears equally likely that the corresponding 
muscle groups in the seal have their chief use in pulling the head and 
the hinder end respectively in a lateral direction while swimming. To 
me it seems beyond question that if these highly specialized muscles 
extending from the phocid humerus were actually a relic from a time 
when seals might have used the arm as a primary means of propulsion 
the external form of the manus would now show far more indication 
of this fact. 

In the sea-lion the lateral epicondyle of the distal humerus projects 
scarcely laterad of the trochlea, but is much better defined in the seal, 
while in the former the medial epicondyle is greatly developed, but 
slightly so in the seal. This is an indication that in the sea-lion the 
flexor muscles of the fore arm have more leverage and hence are pre- 
sumably more efficient than the extensors, and that in the seal the re- 
verse is the case. These muscles will be more fully discussed later. 

The musculature of the dugong is almost unknown but it is likely 
that in brachial details it is considerably different from the manati be- 
cause there are very important differences in the form of the humerus. 
In addition, Murie failed to figure many important items of the manati, 
and in consequence my discussion of the sirenian brachium should be 
considered as only, tentative and subject to amendment. 

The humerus in the dugong and the manati show two very different 
trends, that of the former being very pinniped-like in some general 
respects, while in the latter this is not the case. The head in both is 
situated fairly posterior to the shaft axis. This is a rather trustworthy 
indication that the chief direction of movement is in the sagittal plane, 
or at least that any other primarily important movement that might have 
been recently adopted has not yet had time to cause appreciable alter- 
ation osteologically. This is also borne out by the fact that the lesser 
tuberosity is not medially situated. Rather is it continuous with the 
greater, the two conjoined forming a broad, transverse ridge fairly 
anterior to the head. The medial part of this ridge, homologous to the 
lesser tuberosity, is much higher than the latter, indicating, I should 
think, that the humerus is normally held somewhat abducted from the 
scapular plane. There is no teres minor in this genus, so the scapular 
muscles inserting upon this ridge, or broadened tuberosity, are the sub- 
scapularis upon the medial part, the infraspinatus upon the outer, and 
the supraspinatus between them. The first, operating alone, should ac- 
complish some inner rotation with adduction; the second, slight ab- 
duction with outer rotation, or flexion when the humerus is also strongly 



flexed; and the third extension: or by operating the first in antagonism 
to the second, all three can act in extension. And extension of the flipper 
is very important in the act of drawing herbage toward the mouth. In 
the manati the deltoid is inserted not upon a ridge but on a rugosity 
upon the middle of the shaft. It accomplishes chiefly flexion, this 
being strongly aided by action of the teres major, which with the latis- 
simus inserts upon a similar but larger and more distal rugosity upon 
the opposite (medial) side of the shaft. 

In the dugong the humeral head is also posterior to the shaft axis, 
indicating that flexion and extension is the chief movement, the greater 
tuberosity is higher than the lesser, the former is located craniad of 
the head as usual, and both are individually distinct instead of continu- 
ous as in the manati. In addition there is a heavy and high deltoid 
crest continuous distad with the greater tuberosity. Humeral condi- 
tions are essentially similar in the Steller sea cow. 

In the dugong osteological details of the brachium so greatly re- 
semble those in the sea-lion that I have no choice but render the opinion 
that the musculature of this region must have many points of similarity. 
The supraspinous fossa is very much smaller in the dugong, but the 
muscle should have a comparable action and insert with the infra- 
spinatus upon the greater tuberosity. Origin 9f the teres minor cannot 
be as extensive and this muscle must be reduced — perhaps absent. The 
subscapularis is likely used separately for adduction and there should be 
a strong deltoid inserting upon the deltoid crest for use in rotation 
of the humerus. This is all that can be said. In view of the fact that 
we know so little of the way in which the limb of this animal is mostly 
used and consequently are unable intelligently to compare its real' 
function with that of the manati limb, it would hardly be justifiable 
to advance further possibilities. It may be stated, however, that the 
anatomical details of the manati indicate that this animal has never 
used the pectoral limb as a means of propulsion through the water since 
this member became specialized, while the osteology of the dugong 
does show that the limb might once have been used either for this func- 
tion or some other that necessitated muscle action of much the same sort. 

After having discussed the brachium and before taking up details 
of the antibrachium it is proper to discuss alterations in the length of 
both of these segments, and consequently in the whole arm, which 
aquatic mammals have experienced. For the reason that different sorts 
of vertebrates may respond in such diverse ways to the same stimuli 



we may not derive much help in this from a scrutiny of brachial condi- 
tions in aquatic reptiles, but to do so will at least be interesting. 

In discussing the modification which the pectoral limb has undergone 
in both mammals and reptiles of aquatic habits Williston (1914) stated 
that the humerus has become greatly shortened in aquatic types having 
a tail fitted for primary propulsion and even in some having short tails, 
as seals, and to a lesser degree, sea otters. In those which use the legs 
for direct propulsion, as plesiosaurs and marine turtles, the humerus 
is elongated. In all save seals and otters whose limbs are used rather 
as sculls than oars the lower limb bones are always shortened. These 
statements are not entirely accurate. In all highly specialized aquatic 
mammals now living, regardless of the method of propulsion em- 
ployed, the humerus has become shortened. On the other hand, so 
far as I am acquainted with the facts, in all aquatic reptile the sequence 
seems to be for the antibrachial segment first to experience shortening, 
followed later by a shortening of the humerus. At times, as in the 
leatherback turtle (Dermochelys, fig. 34) , the disparity in size of 
these segments is still very pronounced, the humerus being huge and 
the antibrachium really of insignificant size. Apparently this is funda- 
mentally characteristic of the order and different from what seems to be 
the usual sequence in aquatic mammals. The likelihood is that this 
difference is chiefly attributable to the dissimilarity in the osteological 
plan and basic muscular equipment of the two orders. For the present 
it is hardly profitable to pursue the subject further. 

In spite of the elongation of the sea-lion manus the total bony arm 
length from the shoulder joint is, relative to length of body vertebrae, 
very much shorter than in such a terrestrial carnivore as a cat, while 
in the seal it is 20 per cent less. This may be said to be due entirely 
to a diminution in the length of the brachial and antibrachial segments. 
Relative to length of body the humerus in the sea-lion is but 65, and 
in the seal but 45 per cent of the length of this bone in the cat. The 
radius, as representing the antibrachial length, is found to be 105 in 
the sea-lion and 96 per cent in the seal of their respective humeral 
lengths, while in the cat this percentage is about 104. It is somewhat 
surprising to find the proportions of these two segments so uniform 
in these three mammals, and it is an indication that the stimulus for re- 
duction in the arm length of these pinnipeds has apparently been quite 
uniform, or at least has had a uniform result, in both segments between 
the scapula and manus. In the walrus, however, the radius is only about 
80 per cent of the humeral length. 



That there has been a pronounced reduction in brachial and anti- 
brachial length in sirenians is apparent and yet it is difficult to compare 
its degree with this detail of the Pinnipedia. It is clear, however, that 
in relation to arm length there has not been as great reduction in this 
dimension of the sirenian humerus, while there has been a greater 
amount in the radius and ulna. But it is unexpected to discover that 
in both genera considered the radius is just 68 per cent of the humeral 

In living cetaceans there is invariably a phenomenal shortening of 
the humerus to the extent where this bone is at times two thirds as 
broad as long, but its proportion to the radial length is extremely vari- 
able. In almost all sorts it is definitely shorter than the radius, but in 
Physeter, Kogia, Stenodelphis and Platanista it is longer — in the latter 
over 200 per cent of the radial length. In other odontocetes it is slightly 
shorter, while in mysticetes this is considerably more pronounced. Thus 
in Rhachianectes the humerus is about 68 per cent of the radial length, 
and 60 per cent in Sibbaldus. Speculation of the above cetacean facts 
is, however, sadly complicated by the circumstance that the arm of 
Basiolosaurus, more primitive than any recent whale, shows a much 
reduced antibrachium and a rather long humerus. 

As with the carpus and digits of the Cetacea the elbow joint in this 
order is entirely fibrous in character. Apparently this is due to the non- 
development of the articular structures characteristic of a synovial joint, 
rather than to the alteration of these. It seems that the fibers of the 
existing joint develop directly from the perichondrium or periosteum 
to the degree where they supply the amount of stiffening that the animal 
finds essential. 

Whereas the external axilla of the sea-lion falls midway of the anti- 
brachium and of the seal opposite the wrist, it. is apparently situated 
just proximad of the elbow in the Sirenia, at about the proximal third 
of the antibrachium in at least some odontocetes (porpoises), and at 
about the same point in Mysticeti. The optimum position of the elbow 
in respect to the body contour in aquatic mammals thus seems to be 
somewhat variable, depending upon individual requirements, and of 
course upon the amount of specialization. 

In no other mammals than those discussed above does it seem safe to 
say that there has been alteration in the osteology of the fore limb in- 
duced by life in the water. Taylor (1914) has made the observation 
that in the sea otter this member is relatively smaller than in the river 
otter, and the short legs of the hippopotamus may be considered either 



in this light or as having been caused by decreased terrestrial activity. 
But such alteration has been too sHght in degree to make discussion 

As the anterior limb of aquatic mammals has always been a subject 
for the liveliest speculation there will undoubtedly be expected of me 
some statement regarding my conviction on this question. Unfortunately 
the facts do not warrant any very strong convictions, for each case is 
different and seems to constitute a law unto itself. In the first place it 
is likely that the paths followed by aquatic reptiles are throughout most 
of their course so different, because of a different equipment to begin 
with, that they can certainly not be compared with any intelligence, at 
least during their earlier stages. At long last, after a staggering length 
of time, the paddle of an ichthyosaur and that of a cetacean, if used 
for the same purpose throughout ages, may show a convergence of char- 
acters to the point where they have essentially similar details. The 
Cetacea already show promise of this, but no other aquatic mammal is 
sufficiently specialized for this to be apparent. Diverse sorts pursue in- 
dividual paths and even though these trend in the same general direc- 
tion they are often far apart and have become deflected from the straight 
line by numerous obstacles. All that I feel convinced of at present is 
that in aquatic mammals there is a stimulus for the shortening of the 
length of the part of the arm situated between the scapula and the manus. 
It further seems likely that the control of the arm as a swimming paddle 
or equilibrator is facilitated by a humerus shortened to some degree and 
with the insertions of some of the critical muscles shifted farther distad. 
In addition it appears logical that after the disappearance of a synovial 
elbow joint and marked atrophy of the musculature of the lower arm 
the shortening of the antibrachium should be enabled to progress at an 
accelerated rate. But any opinion as to whether the brachium or anti- 
brachium should be shorter in some hypothetical aquatic mammal would 
be merely fanciful. 

Regarding the course of future development, we are justified in con- 
sidering that eventually a flipper of the character of that found in Plat- 
anista (fig. 36) will likely assume the essential characters of that of an 
ichthyosaur (fig. 34), in which the humerus, radius and ulna have been 
reduced to flattened ossicles of the same appearance as the carpal bones. 
This may or may not be the final goal of the pectoral limbs of mysticetes, 
sirenians and non-phocid pinnipeds. It is not improbable but is surely 
not a certainty. I will even go so far as to say that it seems to me un- 
likely that the true seals will ever assume this type of flipper char- 


acteristic of ichthyosaurs, chiefly for the reason that all but the manus 
is contained within the body contour, the function of the long bones 
appearing to be chiefly that of a scaffold upon which are hung some of 
the muscles of importance to swimming, rather than an integral part of 
a flipper. 

The part which the antibrachium plays in the economy of aquatic 
mammals diflPers considerably in various sorts. It is difficult to know 
in just what light it should be considered in the Phocidae. In the 
latter the position of the external axillary region is opposite the wrist, 
which is capable of extreme mobility, and hence the antibrachium cannot 
be considered as functioning with the manus. And yet the humerus 
acts as independently of the forearm as the exigencies of the position 
of both within the body contour will admit. Although in consideration 
of the functional differences the osteological similarity of the forearm, 
consisting of a broadening of the bones, in seals and sea-lions is really 
phenomenal, the two should be considered separately. In both of them 
the elbow joint is synovial but flexion of the forearm is apparently re- 
duced, especially in the sea-lion. 

In the seal (Phoca bispida) which I have dissected there is a remark- 
able broadening of the distal radius and proximal or olecranol part of 
the ulna, which does not affect the opposite ends of these bones. No 
useful reason for this specialization is positively known. As the anterior 
limb is used for neither propulsion nor as a specialized rudder, as shown 
by form of the manus, there could not be this stimulus for a broadened 
forearm, and besides, the recession of the segment within the body 
contour renders any such mechanical adjustment useless. Although 
there may be perfectly good obscure reasons for this condition there is 
only one apparent. The divisions of the phocid triceps muscle are truly 
enormous and in addition have phenomenal leverage by virtue of their 
attachments. Thus the longest division extends from the vertebral bor- 
der of the scapula to beyond the middle of the fore arm. In action it 
not only extends the antibrachium but sweeps the whole arm to the rear ; 
and the large, broad olecranon gives it just so much more leverage. The 
origin of the two extensores pollicis muscles from the lateral olecranon 
well away from the joint give to them a marked ability to supinate the 
manus. Upon the medial side of the bone an unusually large area is 
provided for the broad and powerful flexor digit, communis, with con- 
sequent power of digital flexion, and the broad ulna provides a better 
lever arm for the adductor action of the peculiarly developed abductor 
digiti quinti longus. I can see no definite muscular advantage in the 



broadened distal part of the radius save chiefly in providing better lever- 
age for the extensor metacarpi polHcis, and possibly to a slight extent of 
the pronator teres and supinator brevis. So far as we know there is no 
terrestrial nor natational use for such a specialized equipment for power- 
ful motions in several directions of the manus, and I am accordingly 
compelled to believe that it has other functions of exceeding speciali- 
zation, which the very shortness of the external arm facilitates. Perhaps 
the reported use of the manus of northern phocids for scratching holes 
through the ice is of more importance than might appear at first glance. 
But this function could not be of influence in the case of those tropical 
genera in which the nails are well developed, so perhaps the manus is 
particularly useful for some such purpose as scratching about on the bot- 
tom for echinoderms or similar fare. 

In the Phoca hisphia dissected there were very deep grooves (fig. 35) 
upon the distal radius for the passage of the tendons of the extensores 
metacarpi pollicis, digitorum communis and lateralis, and a shallower 
one for the extensor pollicis longus. These grooves are almost as well 
defined in P. fasciata but much less so in P. groenlandica, while that for 
the extensor metacarpi pollicis is the only one at all marked in the genus 
Monachus. Neither are they a character in Odohenus and they are 
entirely absent in the Otariidae. These grooves, when they occur, are 
for the purpose of fixing the position of the tendons concerned. They 
would indicate that contraction of the respective muscles, eff^ecting ex- 
tension of the whole manus as well as abduction of the pollex, is often 
repeated and rather strong; but this is perhaps incidental. Their real 
significance is to prove that the chief action of these muscles is effected 
at a time when the manus is already considerably extended, or, at the 
very least, on a line with the antibrachial axis; and further, I believe 
that the fact that these grooves are sunk into the bone not at a perfect 
right angle but facing somewhat in the ulnar direction is an indication 
either that the chief extensor action of the muscles is combined with 
some definite abductive movement of the manus, or else that the manus 
is usually definitely abducted during their action. By means of these 
grooves and ligaments above them the tendons are prevented from 
pulling away from the bone. This action of the wrist will be discussed 

In the sea-lion (Zalophus calif ornianus) which I dissected there is an 
even more marked broadening of the proximal ulna, while the breadth 
of the distal radius is no greater than in the seal. Unlike the latter 
animal, however, there is a purely mechanical need for a broadening of 



the sea-lion antibrachium which the former seems now to lack. In 
otariids the normal posture places the external axillary angle at about 
the middle of the fore arm, while in phocids it is situated opposite the 
wrist. Hence in the former this segment is influenced by the stimulus 
for becoming flatter and broader so as to oflFer less resistance, and has 
been more completely so in the past. The development of the seal manus 
seems to indicate that this same quality of stimulus has at no time been 
a very important factor in its evolution. Further, in the sea-lion the 
distal fore arm is essentially a mechanical part of the paddle (manus) , 
as it is not in the seal, and that the broadness of the flipper in the 
former pinniped has had some real influence in continuing such broad- 
ening onto the fore arm, thus aff^ecting the radius, while the correspond- 
ing broadening of the olecranol part of the ulna should rather be at- 
tributed to muscular stress. In view of this evidence I am of the opinion 
that the largely similar osteological specialization of the fore arm in 
these two pinnipeds is likely not ascribable strictly to the same stimuli, 
but rather to two (or more) rather diverse influences which have effected 
similar conformation or convergence. 

Thatj the sea-lion has need for an antibrachium broader than the 
bones alone have been able to supply may be shown by the fact that 
upon the radial edge there is a thickened, partly fibrous development of 
tissue which is continuous with the pectoralis profundus; or this may 
have been built up as a buffer by the action of water resistance. In 
otariids the purpose for which the flipper is used necessitates that it nor- 
mally be held more extended than the phocid finds advisable. Ob- 
servation of the segments during dissection, correlated with the position 
of the humeral condyles in relation with the axis of the bony shaft, 
indicates that the brachial-antibrachial angle in static posture is about 
125 degrees or less in Phoca, and 155 degrees m Zalophus. 

As in Phoca the lateral aspect of the olecranol surface of Zalophus 
is occupied by the very broad origins of the extensores pollicis longus 
and metacarpi pollicis, while the triceps longus and lateralis, for ex- 
tension of the fore arm and flexion of the whole limb, gain added lever- 
age by virtue of the broad olecanon. The two pollicis extensors are 
extremely broad in their tendinous parts and are well fitted for supinat- 
ing movements of the anterior flipper border. I believe, however, that 
the chief stimulus for the remarkable broadening of the proximal ulna 
of the Otariidae has been the origin of the palmaris longus (fused with 
the smaller head of the flexor carpi ulnaris) . This is a very specialized 
muscle, not only at origin but distad, where it extends in a tendon 25 



mm. broad at its narrowest point, thence widening to cover the whole 
wrist before sphtting in two sheets, the more robust going to the anterior 
and the lesser to the posterior border of the flipper. This effects a 
definite cupping action of the palmar surface as discussed later. 

In pinnipeds as well as Cetaceans there is no appreciable tendency 
for either the fusion of the radius and ulna, nor for the reduction of 
one of these bones at the expense of the other, which is as one would 
expect. The stimulus is for each bone to become independently broad, 
which is apparently facilitated by non-fusion, and ultimately after the 
entire abandonment of the land, for the formation of a roughly fibrous 
connection of all the bony elements, which gives a desirable amount of 
resilience. The antibrachial bones of the Sirenia, however, are anoma- 
lous among aquatic mammals of a high degree of specialization. In the 
dugong the radius and ulna are distinct and quite simple, without any 
real indication of broadening. In the manati there is some slight ten- 
dency toward flattening of the ulna and this bone is firmly fused at both 
extremities with the radius, although the shafts of the two bones are 
curved and quite wide apart. In Hydrodamalis there is little or no pro- 
pensity for flattening but the extremities not only are firmly fused but 
the shafts as well (fig. 40), save for one small interval that has doubt- 
less remained as a foramen for the passage of nerves and blood vessels. 
Hence one finds the unexpected condition that in this detail the Steller 
sea cow does not resemble its (presumably) nearest relative, Halicore, 
but the more distant Trichechus, and it must be inferred that in the first 
and last mentioned genera there is no need either for pronation or 
supination of the manus save what is possible through the carpus, while 
presumably in the dugong some little rotation of this segment is possible. 
Stronger extension of the antibrachium in the dugong is indicated by the 
better definition of the olecranon. A synovial elbow joint presumably 
of the normal sort is possessed by this order. 

In all known cetaceans, both living and fossil, the radius and ulna 
are distinct and neither is appreciably reduced in comparison with the 
other. Almost always the radius is slightly more robust than the ulna, 
which might be expected because of its exposed position upon the an- 
terior border of the arm, where it would encounter full water resistance ; 
but at times (as in Platanista), the ulna may be the larger. Apparently 
zeuglodonts retained the synovial character of the elbow joint, and this 
may also have been the case with some Miocene whales, for Etirhino- 
d el phis, and likely others, have articular surfaces upon the bones, al- 
though these are of lessened area and definition, so that it is more logical 


Figure 37. Areas of muscle attachment upon the pectoral limbs of a toothed 
whale {Monodon, to the left) and a whalebone whale {Balaenoptera bore- 
alis, to the right, redrawn from Schulte). Lateral view above, medial be- 



to presume that the joint had lost at least a part of its original mobility. 
As the antibrachium is essentially a mechanical part of the paddle the 
tendency is for an increasing flattening of both bones, and for their 
simplification, following the disappearance of functional antibrachial 

The only bony details that might be ascribable to present or past mus- 
cular stimulus is the olecranon. This was large and somewhat pinniped- 
like in zeuglodonts, and to a lesser degree in Eurh'modelphis. It is 
usually present in balaenopterids as a well defined process, which at 
times is of large extent. Schulte (1916) has shown that in the fetal 
state of at least one species this is truly of phenomenal size and as a 
distinct bony arm, stretching somewhat distad, may be half the length 
of the humerus (fig. 37). In Eubalaena it does not constitute a real 
process, although there is some projection of the ulnar head as a sub- 
stitute. Usually in odontocetes there is a slight process at this point, 
but occasionally (as Platanista) in the sorts in which the antibrachium 
is most reduced there is no indication whatever of an olecranon. 

The reason why the olecranon is of such enormous size in the fetal 
balaenopterid is unknown, but it is only natural that at this stage it 
should be larger than in the adult, in which the muscles of the lower 
arm are so reduced. The muscles most intimately concerned with the 
olecranon Schulte found to be a long and a short triceps head. He 
further found as indubitably present flexores ulnaris, radialis, and com- 
munis, and an extensor digitorum communis, as well as a ligamentous 
band which he considered as representing the biceps. This is essentially 
in accord with the findings of other investigators of balaenopterid mus- 
culature, except that he failed to encounter a plamaris longus, as re- 
ported by Carte and MacAlister, or a flexor sublimis as mentioned by 
Perrin. The details of these muscles do not concern us in the present 
connection. What is of interest is the fact that although they (except- 
ing possibly the long triceps) are either entirely or virtually nonfunc- 
tional, there are a number of them clearly present in the arm of mysti- 
cetes, and this constitutes one of the chief items of evidence that this 
group as a whole may be less highly specialized than odontocetes. 

Clearly recognizable fore arm musculature has often been reported in 
the Odontoceti. For instance Schulte (1918) stated that in Kogia 
there were triceps, extensor communis, flexores carpi ulnaris, digitorum 
radialis and digitorum ulnaris, as well as interossei. In my own dis- 
sections of Monodon and Neoweris, however, I failed to find any of 
these as clearly recognizable muscle remnants. In the former there 



was a tendinous band stretching from the humeral head to the olecranon 
to which I referred as a triceps ; but there was nothing else distad except 
fibers identical with those strengthening the tissue between the digits. 
In Monodon I distinguished and tentatively named by virtue of their 
position a triceps, biceps relic, flexor digitorum and extensor digitorum, 
but the two latter were rather loosely associated fascicular bundles and I 
cannot even be entirely sure without histological preparation that they 
really contained muscle fibers. It is thus indicated that at least in a great 
many odontocetes the intrinsic musculature of the flipper proper is 
considerably more atrophied than in mysticetes, and is on the rapid road 
to total disappearance. 

As mentioned there is a tendency in the Cetacea toward flattening of 
the antibrachial bones, and this is especially pronounced in those sorts 
in which this segment is the shortest. Like the case of the sea-lion this 
is entirely expected, for the antibrachium is a mechanical part of the 
flipper and is subjected to the same influences as those which have oper- 
ated to shape the manus. Also like the sea-lion there appears to be need 
for an antibrachium that is externally broader than the breadth of bone 
can furnish, and there is disposition of tough fibrous tissue not only on 
the radial border, but softer tissue in more generous amount upon the 
ulnar border as well. As previously mentioned the tendency in most 
sorts of odontocetes at least would seem to be toward the reduction of 
the radius and ulna to the size and conformation of the larger carpal 

In considering the manus of aquatic mammals we come to the critical 
part of the fore limb. All other parts of the arm are directly influenced 
by it or by certain body factors; hence the manus may be expected to 
reflect in most perfect degree the stimuli that have operated upon the 
pectoral appendage. In no two diverse sorts of aquatic mammals is this 
detail alike, for the reasons that in no two sorts has it been used in 
exactly the same manner, and that diverse kinds of aquatic mammals 
split off from the main stems and have constituted individually separate 
lines, each developing its own peculiarities, since before the manus be- 
came so highly specialized. In consequence it is unsafe to venture any 
broad generalities except that if the manus be used as a propulsive organ 
or an equilibrator it can be expected eventually to assume the character- 
istics of a paddle or flipper in which the bony element's become flattened 
and simple, without movable joints. Usually there will also be apparent 
a stimulus for a lengthening of the manus at the expense of the more 
proximal brachial elements. 






'^FLE)^.CAT\P\ RAD. 

Figure 38. Superficial musculature of the medial aspect of the left pectoral limb 
of a sea-lion {Zalophus) . 






ABB. D1&. S LONG,-- \— ' 


Figure 39. Superficial musculature of the medial aspect of the left pectoral 
limb of a seal {Phoca). 



There would also seem to be another character which the manus of the 
aquatic mammal may be expected ultimately to adopt. In most terrestrial 
mammals the bones of the metacarpus and digits consist of a shaft and but 
a single epiphysis. The latter is situated upon the the distal end of the 
shaft in the four lateral metacarpals, but at the proximal end of the first 
metacarpal and all the phalanges. Apparently there is a tendency in 
aquatic mamals for all of these bones to develop an epiphysis at each 
end. Kiikenthal (1891) appeared to think that this is attributable to 
the retarded rate of ossification in the elements of the manus. Double 
epiphyses are well differentiated in distal digital bones of at least 
most cetaceans, and Kiikenthal stated that this character is advanced in 
sirenians, and among pinnipeds in Mhounga, Hydrurga { = Stenorhyn- 
chus), Otaria, with indications in Odobenus, and in the pes only of 
Cystophora. But Flower (1876) said that they are double in the manus 
only of Mirounga, while Weber (1886) stated that in the pinniped 
manus the epiphyses are normal, being double only in the proximal 
elements of the pes. Kiikenthal also said that there are indications of 
double epiphyses in Ornithorhynchus, Hydromys, Hydrochoerus and 

All this is as may be. To establish the incontrovertible presence of 
double epiphyses is not always easy without extremely careful investi- 
gation, and without taking any especial pains in preparation I have been 
able to satisfy myself that this situation obtained only in the case of 
sundry Cetacea and a very few pinnipeds of the proper age to exhibit 
the character to good advantage. I might point out two likelihoods, 
however ; that it is extremely easy to interpret cartilaginous irregularities 
at the end of the phalanges as early stages of epiphyseal ossification, 
and, that if one searched for this condition among strictly terrestrial 
mammals with as much zeal as it has been sought in partially aquatic 
forms he might very well have equal success. 

It is almost an invariable rule that when a mammal first takes to the 
water the use of its anterior limbs as natatory organs will be merely 
incidental. The probabilities are that these members will be used with 
the hind limbs in swimming "dog-fashion" but the important function 
of all four legs is for terrestrial progression. As the hind feet are 
ordinarily larger than the anterior, these will have a tendency to increase 
in proportional importance for propulsion. In most cases I believe 
that the aquatic development of the manus will be extremely slow, for 
it has not sufficient area to be of decided import either in steering or 
swimming. But bodily conformation may introduce disturbing ele- 

[ 242 ] 


ments. Without a tail of fair proportions or with a body of unusual 
width (hippopotamus) the dog-fashion method of swimming will long 
be followed, the important function of the fore limb then being largely 
as an antagonist, to neutralize the oblique, sidewise impulse furnished by 
the separate kick of each hind foot. In such case the manus might 
be expected to acquire aquatic modification almost if not quite as rapidly 
as the pes. But still, each sort of mammal is a law unto itself, and 
accordingly it is necessary that each be considered separately, with some 
unavoidable repetition of facts. 

In the majority of insectivores and rodents the manus is totally un- 
modified. They appear to have no need for especial speed in the water, 
the food being mainly herbivorous or insectivorous, and the feet are 
presumably folded against the body so as to offer no resistance while 
swimming. They can have no use for propulsion and in the water 
can act only as grasping organs for securing prey. Manual activity will 
thus be in direct proportion to the amount of terrestrial locomotion 
indulged in, and as the latter decreases one would expect that a tendency 
for reduction of the limb would increase. 

Concerning various sorts of slightly aquatic mammals the statement 
is frequently made that the fore feet are partially webbed, when an 
examination of the specimens fails to indicate the fact. Apparently 
some authors have been prone to interpret the membranes stretching 
between the base of the toes as aquatically-adapted webbing when as a 
matter of fact this is no better developed than in a great number of 
comparable forms with habits that are exclusively terrestrial. Undoubt- 
edly in the desman, Australian water rats and some otters the webbing 
of the anterior appendage is slightly better defined than what may be 
considered as normal, but it is of very slight degree indeed and may 
be partly illusory because there is also a slight tendency for a lessened 
definition of the digits following decreased use of the limb. In Chimar- 
rogale, however, there is more of a hairy fringe to the manus than ter- 
restrial insectivores show, this matching a similar but far stronger de- 
velopment upon the hind feet. 

Here a question intrudes itself that should be accorded brief mention. 
Irrespective of the fact that hypertrophy of the posterior limbs is often 
accompanied by atrophy of the anterior ones (compensational develop- 
ment) , is there not a tendency, however slight, for the manus to adopt 
the same quality of equipment as that occurring upon the pes: the same 
quality of nails, webbing, or hairy fringes? Needless to say, // there 



be such a tendency it is often set at nought by diversity of function 
of the two members. 

At any rate the few manual changes in those mammals mentioned 
above are too sHght in degree either to be of much use to their possessors 
or to be of much significance. They may or may not be indicative of 
a hne of future progress. 

The development of the manus in the platypus is comparable in some 
respects to that of turtles, and in others to the sea-lion. Apparently 
there has been ample time since this queer beast took to the water for 
it to experiment fully and to adopt in marked degree those peculiarities 
best suited to its needs. Its body is definitely flattened and it is likely 
that this always has been a characteristic. Its tail probably assumed its 
present form not primarily as a rudder for swimming but to keep the 
animal on the bottom while feeding. Hence in order to overcome 
oblique movements which would be affected by the alternate kicking 
of one pair of limbs it probably, throughout a very long period, swam 
by diagonal kicks of all four feet. As a result it is likely that the 
webbing of these four members progressed for a while at a fairly uni- 
form rate. But it is the invariable rule among mammals that when 
aquatic specialization has attained some considerable degree of perfection 
the tendency is for a single member or pair of members to take over 
the primary means of propulsion. If for no other reason than that the 
fore limbs are better situated for acting in co-operation with the tail 
for keeping the animal near the bottom it is likely these would receive 
the greater stimulation for development. This might be either increased 
still further, or else hindered, according to the quality of the action 
involved, by the presumable function of removing the bottom debris 
and stones in search for food and of digging burrows. At any rate 
they have gained the evolutional ascendancy over the hind limbs. But 
the future of this creature is hardly predictable. The manus may or 
may not eventually assume the form of a flipper, or for that matter the 
tail might ultimately take over the function of a primary propulsive 

It is seen that the manus of the platypus (fig. 42) is very highly 
and peculiarly modified, the characters, in fact, being entirely unique. 
These, however, are not satisfactorily brought out in Burrell's photo- 
graphs. In a spirit specimen before me the membrane extends con- 
siderably beyond the nails, but projecting even farther are leathery 
thickenings of the membrane. There is but one of these to the first 
and fifth digits, extending in continuation of the digit. Each of the 


other three digits has a pair of such thickenings, however, which diverge 
very shghtly as they extend distad. Apparently they all have a sort of 
hinge at the base, for the predigital part of the membrane is folded 
back against the palm well out of the way when the animal is not only 
scratching and digging, which is accomplished chiefly by means of the 
well developed claws, but also during terrestrial progression. The 
claws have undoubtedly been retained for these purposes. Wood Jones 
has given the length of the digits as in the order of 4, 3, 2, 5 and 1, 
but it is seen that their development is to all intents bilaterally sym- 

It is not at all easy to determine from the literature the precise ac- 
tions of the platypus manus that are involved in swimming. From an 
examination of the skeleton it seems that the humerus must operate 

Figure 40. Bones of the right fore arm of the Steller sea cow {Hydrodamalis). 
from a specimen in the U. S. National Museum. 

chiefly in the transverse vertical plane, and the fore arm mostly parallel 
to the body axis and also in the vertical plane. Hence the manus prob- 
ably operates on the thrust-and-recovery method, rather than on some 
system which involves oblique action against the water. 

In some important respects the aquatic stimuli encountered by the 
hmbs of the capybara and the hippopotamus are similar. It is true that 
the former animal lacks the markedly broad beam of the latter, which 
makes it convenient for the hippopotamus to counteract the tendency 
for the alternate kicks of the hind legs to deflect its course first to one 
side and then to the other. But both lack a functional tail and in 
neither has the development of one pair of limbs gained decided ascend- 
ency over the other pair, which may be considered as somewhat unusual 
in mammals of their degree of aquatic predilection. Apparently, as 



stated, all four limbs are habitually used in diagonal alternation while 
swimming, but it is not meant to imply that an observer can never de- 
tect variations of action, just as a human swimmer may play about and 
keep afloat by several expedients. 

Neither fore nor hind feet of the hippopotamus are appreciably modi- 
fied in an aquatic direction, and it is readily seen that it would be ex- 
tremely difficult for a foot of its type, upon which such a ponderous 
body is imposed, to change into a flattened propulsive organ, at least 
as long as terrestrial locomotion is a necessity. The fore feet of the 
capybara are furnished with interdigital membranes which extend prac- 
tically to the tips of the toes, although these are indented, rather than 
straight, along their free margins. This is equally the case in the hind 

The fore feet of the otters are chiefly of interest in the present con- 
nection because I deem it likely that the ancestors of the Cetacea were 
beasts of somewhat similar conformation, swimming by the same meth- 
ods. It is usually stated that the manus of the common otter is partly 
webbed, but this character is not more pronounced than in many terres- 
trial mustelids. There is no webbing at all in the African clawless otter 
(Aonyx, fig. 42), while Allen (1924) shows that it is particularly ex- 
tensive and broad in the African Lutra maculicollis (fig. 42). It is 
to be expected that during swimming the fore feet of the latter animal 
are employed in a manner somewhat different from the case of the 
former. Other genera of otters are presumably variously intermediate 
between these two extremes. 

The common otter is very nimble and so given to sportiveness while in 
the water that it is very difficult to determine any really important uses 
to which the manus may be put while in this element. The anterior 
limbs are often used for grasping, but hardly as a definite aid to propul- 
sion, nor apparently for equilibration. Certainly they will eventually 
become more modified, but in the mean time they may be expected 
to exhibit some tendency toward reduction of size, as is indeed the case 
to a slight extent in the sea otter. 

Not a great deal of significance can be said about the manus of the 
Sirenia. They do not use the fore limbs as a direct aid to propulsion 
and probably never have since the days when they swam dog-fashion, 
if the present evidence is trustworthy: nor apparently are the limbs 
used as equilibrators for elevation and depression of the body after just 
the same fashion that the flippers of the whale are now employed. 
Rather should their action be compared to that of the seal, but to a more 


marked extent, being thrust to the side for making sharp turns or 
waved about as assistants to the accomphshment of a variety of rather 
languid evolutions. In addition they are said to be habitually employed 
for bringing herbage toward the mouth, and it is frequently stated that 
the females use the flippers for clasping the young. 

Externally the manus of the Sirenia is paddle-like though of a rather 
irregular and blunt form, rather than gracefully shaped like that of the 
sea-lion and most whales. I would not regard any of the known stimuli 
as particularly strong for the assumption of a fore limb of this char- 
acter and it seems likely that an immense space of time has been neces- 
sary for its attainment. Small blunt nails are present apparently in all 
forms of the manati except the African T rich ec bus i nun guts, and it 
is interesting to note that these are situated upon the flipper border where 
they would be most available for such acts as scratching, although they 
hardly project sufficiently to be very useful in this respect. Nails are 
absent in Halkore as was probably also the case in Hydwdamalis. 

The synovial character of the wrist joint seems to be completely re- 
tained in the Sirenia. As these mammals are entirely helpless out of 
water there is not the need for abduction to 90 degrees of the wrist 
that the sea-lion has, but yet this joint is said to be definitely and quite 
surprisingly mobile, probably because of feeding needs. The pisiform 
bone is absent in this order. In the dugong the carpal elements are 
astonishingly reduced to three, one subtending the radius, a second the 
ulna, and a third elongated bone distad with which the first four meta- 
carpals articulate. In the manati there are six carpal elements (Flower 
said seven) arranged in two transverse rows of three each. The sig- 
nificance of these carpal details is unknown. 

The sirenian metacarpal and digital bones are somewhat flattened, 
especially in the manati, as is usual in aquatic mammals, the ungual 
phalanges are of irregular shape and particularly flat as in the sea-lion, 
and the pollex is definitely reduced. The fourth digit is the longest, 
this representing the tip of the flipper. Beddard (1900) has mentioned 
that in this order "hyperphalangy is also met with but to a very small 
extent." In most skeletons any predigital cartilaginous elements which 
might be present would be lost during cleaning, but in the mounted 
dugong in the National Museum the left manus (fig. 35) shows a small 
cartilaginous nodule, now shrunken, upon the tip of the third digit, this 
being the fourth phalangeal element. Furthermore this is of the exact 
character as the single premetatarsal element of the first, and the third 
of both the second and fifth digits, which is of significance in show- 



ing how much similarity may exist between actual phalanges and a pre- 
digital item. This is of much interest and will be discussed later. 

The manus of the seal and sea-lion could not be more utterly unlike 
were they totally unrelated, and they must be discussed from a different 

In the normal posture assumed by the seal the manus is the only part 
of the arm that projects beyond the body contour, the external axilla 
of the specimen dissected falling opposite the pisiform bone. The 
manus is not appreciably altered in relation to body length from the 
condition obtaining in such a terrestrial carnivore as the cat, but as so 
much of the arm is out of sight the hand appears unduly short. It 
is very broad, however, this being largely due to fatty tissue, skin, and 
hair. The digits are all completely connected by membranes, but these 
are narrow and do not permit spreading of the digits so as to present 
a broad surface to the water. In Phoca the first digit is the longest 
and heaviest, the others being successively slightly shorter, and all are 
furnished with heavy nails. And this is the case in most seals, although 
perhaps in none are the nails relatively heavier than in P. hispida, 
while a few have them slightly lighter, and Beddard (1900) has 
stated that in Ommatophoca they are quite rudimentary. In mounted 
Mirounga examined they are short and blunt as though worn down 
by much use. The conformation of the arm would permit of the nails 
being used for scratching over but a negligible area of hide, so they 
must have some other function, else they would almost certainly have 
become atrophied. This may consist either of scratching holes in the 
ice or prying about on the ocean floor. 

As there has been no rotation of the long bones of the pinniped arm 
the radial border of the manus is presented directly forward, and it can 
accordingly be pressed flat against the body without either pronation or 
supination from the normal. While dissecting Phoca it was found that 
the static posture of this segment is markedly abducted with respect to 
the antibrachial axis, this being through an arc of more than 45 de- 
grees. From this position adduction cannot be great, the chief in- 
hibitor being the peculiar abductor digiti quinti longus, but the latter 
helps in further abduction, which is possible to an angle with the anti- 
brachial axis of at least 90 degrees. Provision is apparently not made 
upon the articular surfaces of the phocid scapholunar and ulnare for 
extension of the manus to 90 degrees, but nevertheless such extension 
seems possible because of the looseness of the ligamentous connections. 
Of course these may not have been so loose in life, but it is likely that 



they were almost so. This same character allows of flexion of the manus 
after the integument and tissue has been removed to the excessive point 
where the metacarpals are actually parallel to the fore arm ; so the flexion- 
extension arc is through 270 degrees. This would hardly be possible 
in life because of the confining influence of the surface tissue. 

With all of the arm except the hand within the body contour and with 
the segments somewhat flexed at that, a relaxed position of the manus 
of the normal sort (i. e. on a line with the fore arm) would bring the 
palm almost against the body. In this position the body would accord- 
ingly prevent much flexion of this member (although if the arm be 
extended or rotated this would not be so and extreme flexion would 
then be possible), and in order that the manus might be thrust hori- 
zontally it would have to be extended to almost 90 degrees. Further- 
more, such extension would seem usually to be initiated from a normally 
abducted position of the manus (in respect to the antibrachium) . That 
this is not only so, but that such motions are strong and oft repeated, 
is indicated by the depth and somewhat lateral inclination of the extensor 
grooves upon the radial surface in this mammal, as already discussed. 

The change in posture of the phocid manus has brought about some 
alteration in the carpus. Its ulnar side is weak and has a mitred ap- 
pearance, because of the normal abduction of the rest of the manus. 
Carpale 2 is partly interposed between the scapholunar and carpale 1. 
Crowding has resulted in, or at least has been accompanied by, the as- 
sumption of a rather conical shape by the carpales, this being especially 
pronounced in the third, in which the apex is presented dorsad so that 
from this aspect it appears as a bony point. The ulnare (cuneiform) 
is reduced in size and the fifth metacarpal subtends that, rather than 
the lateral side of the unciform, which has the function purely of a 
fourth carpale. In consequence the fifth digit really appears to be more 
opposable than the thumb. The abducted posture of the manus places 
the pisiform at the side of the fifth metacarpal. 

It is seen (fig. 35) that there is a gradual reduction in size of the 
metacarpals from the first to the third, while the fourth and fifth are 
abruptly smaller, and this, in connection with carpal details, places the 
metacarpal-phalangeal articulations on a line. The first phalanx of the 
pollex is more than twice as long as the first two phalanges of any of 
the other digits, and this causes the pollex to be the longest. It appears 
that the assumption of an abducted posture by the hand is partly efl^ected 
not by pure adduction at the bases of all the digits, but only part of them. 
Thus there may be said to be pure abduction only of the fifth and 



fourth digits, while in the others this is compHcated by a flexional ten- 
dency progressively in a radial direction, so that the pollex is almost as 
much flexed as abducted. It is difficult to express this in words but 
the illustration (fig. 35) shows it quite clearly. In reality there has 
been partial rotation of the first digits and an increasingly lesser amount 
in the others, so that extension of the thumb really helps to adduct it 
(in relation to the antibrachial axis) . 

With the exception of the radial grooves or pulleys, as already noted, 
the extensor muscles of the phocid manus show nothing that may be 
deemed of much significance. The flexor muscles, however, exhibit 
several points of interest. The tendon of the palmaris longus broadens 
so as to cover the whole radial half, including the border, of the palm. 
The flexor carpi ulnaris tendon does the same to cover the entire palm. 
The peculiar abductor digiti quinti longus has already been mentioned. 
The origin of these three muscles from the broadened olecranol part of 
the ulna, in connection with the normally abducted posture of the 
manus, makes it likely that all three of them have more the function of 
effecting still more marked abduction of this member, than of flexors. 
Apparently the chief muscle for the accomplishment of flexion is the 
flexor communis, as indicated by the excessive breadth of its origin and 
of its common tendon, and it acts chiefly as a flexor of the manus as a 
whole, rather than of individual digits. 

In recapitulation it may be said that the anatomical characters of the 
phocid manus, and what they indicate, is precisely in conformity with 
what one would expect did one bind down his own arm in phocid pos- 
ture and attempt to move the manus similarly. Almost all mobility 
would be through the carpus, which would have to be adapted for an 
extreme amount of abduction and extension. Such mobility is useful 
in fiddling movements of the manus in making slight adjustments of 
posture while floating, just as in the case of a human swimmer, but in 
order to explain the degree of alteration which we find in this member 
of Phoca it is necessary to assume that it has some other important ac- 
tion, involving a considerable degree of strength, than any which is 
now surely known. 

As in the seal the radial border of the manus of the sea-lion is pre- 
sented directly forward so that it offers the least possible resistance to 
the water. The external part of the limb has the shape of a long, rather 
falcate paddle, thicker upon the anterior border and thinner upon the 
hinder edge. The external axillary angle is about at the middle of the 
fore arm, so that the distal half of this segment in reality constitutes 



a part of the flipper proper, for which it is fitted partly by the broaden- 
ing of the radius and partly by subcutaneous tissue deposited upon the 
radial and ulnar borders. 

The manus of the sea-lion differs in many respects from that of the 
seal. It is not carried in an abducted posture and there has been no 
crowding of the carpal elements. The scapholunar is large to corre- 
spond with the great width of the distal radius, and its proximal articu- 
lar surface, as well as that of the ulnare and the metacarpals, extends 
farther onto the dorsal surface than in the seal. This is so as freely 
to allow extension of the manus to an angle of 90 degrees with the anti- 
brachium, which is essential for terrestrial progression. The ulnare is 
not on a transverse line with the scapholunar but above it, to correspond 
with the more proximal position of the distal extremity of the ulna. The 
reason for this is obscure. The first carpale is very large to match the 
size of the first metacarpal, and the unciform upon which the fifth meta- 
carpal abuts, is on a line with the three carpales. The latter have no 
tendency to adopt a conical form as in the seal and are not crowded. 
Also there is perhaps less tendency than usual for the crowding of the 
proximal ends of the metacarpals. These latter are on a line in the 
first four, and somewhat more proximad in the fifth. The first digit 
is much more robust than the others and is the longest. This is not 
accomplished almost entirely by the elongation of the first phalanx, as 
in the seal, but the metacarpal is also markedly lengthened. Ulnad of 
the pollex the other digits are progressively shorter in gradual sequence. 
The increase in the robustness of the pollex is to be expected, in order 
that the anterior border of the flipper should be strengthened, for it is 
this part of the manus that furnishes the chief impulse in swimming, 
while the posterior border plays a more passive role. All the digital ele- 
ments are much more definitely flattened than in the seal, and this is most 
pronounced in the ungual phalanges, which are also of very irregular 
shape, this likely having followed the atrophy of the nails, for the same 
situation obtains in the Sirenia. 

There is a considerable deposit of fibrous tissue upon the entire 
radial border of the flipper and even more at the ulnar border proximad 
of the tip of the fifth digit, this helping to give a broader surface. This, 
and the leathery covering of the manus greatly hinders adductive and 
abductive action. After this tissue has been removed it is found that in 
static posture the axis of the manus is abducted from that of the anti- 
brachium by not more than 15 degrees. Thence practically no further 
adduction is possible, but abduction is permitted to about 90 degrees. 



I would deem that in life the latter figure might be cut to one third by 
the confining effect of the integument. Supination and pronation, even 
in the partially dissected specimen, was through only some 45 degrees 
with respect to the humerus (this therefore including rotation of the anti- 
brachium) . Flexion of the manus is through an arc of 90 degrees, which 
is more than one would expect because one cannot see that such an action 
would be useful. 

It has been seen that the effectiveness of the otariid paddle is in- 
creased by the extension of its border beyond the digital tips. Unlike 
the case of less modified aquatic mammals in which the dorsal and pal- 
mar integument of the digital webbing is in close contact with practi- 
cally no subcutaneous tissue intervening, there is a liberal amount of 



FiGURE 41. Outlines of cetacean and pinniped flippers: (a) Globiocephah 
(after Murie) \ {b) Orcinus (from a photograph by E. P. Walker) ; {c) se3 
hon {Zalophiis) ; and {d) seal {Phoca) . 

such tissue in the manus of the sea-lion and whale, as well as the 
sirenians. This has the effect of padding out the interdigital depres- 
sions so that the flipper presents a plane surface both above and below. 

In the Otariidae the effective length of each digit is increased by 
a cartilaginous rod, the extent of which can be determined in the entire 
animal by the distance from the respective nails to the flipper border. 
Presumably there has been a strong stimulus either for a flipper longer 
than the bony elements of the manus have been capable of supplying, 
or else for an elastic border. 

It has been mentioned that the predigital part of the membrane of 
platypus is strengthened by thickenings of the integument, and it is 
not unlikely that these may eventually be replaced by cartilage. But 


these are double in the three middle digits of the platypus and single 
in the sea-lion, so although the two are analogous in some respects they 
need not be homologous. In other words the same need has been ful- 
filled by somewhat different structures. But the chances are that in 
both sorts the membranous border of the manus extended well beyond 
the digits before there was much stiffening needed, although in the 
platypus the stiffening elements have now outstripped the intervening 

Presumably there must be a contrivance in the sea-lion for over- 
coming involuntarily the tendency for the water resistance while swim- 
ming to force the manus to extension with the fore arm through 90 de- 
grees ; in other words to prevent the manus from assuming its usual ter- 
restrial posture. It is likely that this need is provided for by special 
tonus of the palmaris longus. The latter is exceedingly broad and 
powerful, and partly tendinous throughout a considerable portion of its 
length, so that it could well perform such a static function. Further- 
more its palmar tendon splits into two parts, the more robust of which 
extends along the radial border of the hand and the other along the ul- 
nar border. Flexion while forcing the palm against the water would 
accordingly have a cupping action upon the palm, the force being more 
pronounced along the anterior border. The fourth head of the flexor 
digitorum communis and the flexor carpi radialis, in addition to the ex- 
tension of a part of the pectoral, are also well disposed to act upon the 
radial border of the flipper, so that the latter has an unusually powerful 
equipment for sweeping the flipper obliquely through the water after the 
manner that will prove most effective in propulsion. 

The nails of the manus in the sea-lion should receive further brief 
mention. As the part of the flipper distad of the bony termination of 
the digits cannot be folded back against the palm the nails can no 
longer be used in scratching the body or in any other effective way. 
Accordingly there is no stimulus for their retention and they are now 
reduced to insignificant nodules which do not project beyond their in- 
tegumentary pits. 

It is frequently stated that the manus of the walrus is of a character 
intermediate between that of the seal and the sea-lion, but this is not 
quite correct. In size this is so, but in characteristics it is merely a 
less developed otariid flipper that is shorter and broader, but with digi- 
tal cartilages and rudimentary nails also. 

Just as there is considerable variation in the details of the remainder 
of the cetacean arm, so is there corresponding differences in the manus. 



To attempt to interpret the reason for this lack of uniformity would be 
idle because we can have no conception of the precise manual needs of 
a particular sort of whale, nor what form would constitute an ideal 
equilibrator for its precise body form. I think one should work on the 
theory not that the cetacean pectoral limb has evolved from a single flip- 
per form, but that diverse kinds of whales split off from one or more 
stocks long before the manus had assumed a paddle shape, so that sub- 
sequently that of each has developed independently, with many some- 
what different lines of specialization according to individual needs. 

The width of the cetacean manus depends upon the disposal of the 
phalanges. In a broad flipper the digits are spread somewhat fanwise 
and are relatively short. The breadth of flipper is marked, according 
to Flower (1876) in Physeter, Hyperoodon, Monodon, Delphinapterus, 
Inia, Platanista, Orca, and Orcella. In the latter genus (fig. 41) this 
broadness in relation to moderate length is especially marked. A very 
narrow flipper is usually long, just as in the case of the wing of a bird. 
The limb of Globiocephala is especially noteworthy for this feature, 
and accompanying, and contributing to, the condition is the fact that 
the border digits are much reduced in length, while the second and 
third are remarkably elongated. The second may have about eleven 
elements in the adult and as many as nineteen in the late fetal stages, 
and a comparable situation obtains in at least some of the other por- 
poises (as Phocaena) . Leboucq, I believe, was the first to make this 
claim, later (1888) denied by Weber, but reaffirmed by Kiikenthal. 
This is, of course, evidence that the flipper of such porpoises is under- 
going reduction in length, but it must not be inferred that this is neces- 
sarily the case in all cetaceans. That great length of flipper is not in- 
variably associated with extreme hyperphalangy is shown by the case 
of the humpback {Megaptera) . This genus has the longest flipper of 
any living whale, the exposed part of the arm reaching a length ex- 
ceeding twelve feet, although in the mounted specimen in the National 
Museum there are in the second digit but 6 phalangeal elements in ad- 
dition to the metacarpal, while there are 5 such plalanges in the 
mounted Sibbaldus. These facts have aroused much speculation. It 
is commonly reported that the flippers of Megaptera are used for giving 
gargantuan and resounding love pats to the opposite sex, and it has 
been claimed that they are also employed for clasping during mating, 
as well as to help "herd" schools of fish into the cavity of the capacious 
mouth. This is all that we have to go on, unless we fall back on the 
theory that the great arm length has been purely in the nature of an 



It should here be noted that whereas the flippers of all other whales 
are flat, of a shape determined by the disposition of the tips of the 
digits, and without marked serrations of the borders or uneven pro- 
tuberances, that of the humpback has numerous sharply defined promi- 
nences of a warty appearance, as also occur upon other parts of its body, 
and one of these is always situated dorsad of the tip of each digit. 

But few remarks can be offered regarding the great variation in the 
shape of the cetacean flipper. Just as might be expected the extremes 
in shape are not encountered in the most speedy sorts. In the latter 
this member is always moderately falcate and moderately pointed, with 
graceful proportions. We can but presume that the excessively broad 
flipper of Orcella and the excessively narrow one of Globiocephala have 
developed in response to some particular need of the animal in question. 

Following the complete abandonment of the land the cetecean carpus 
lost the synovial character of the joints. At present the rate of manual 
ossification is much retarded, probably following this loss of synovial 
articulation, and the cartilages representing the various carpal elements 
are closely packed together in a flattened mosaic, without movement save 
that afforded by the elastic character of the tissues. In some sorts of 
odontocetes complete ossification of the carpal bones appears to be at- 
tained with age. 

In most sorts, however, it is even slower, and this is especially marked 
in mysticetes. In adults of the latter the carpal elements are mostly 
cartilage, while the bony centers are relatively insignificant and sunk 
within the hyaline substance, reaching the surface of the cartilage only 
with age.3 This is in contrast to the situation in the majority of highly 
modified aquatic reptiles, for in these the carpal bones appear to be as 
closely packed as possible. 

The homology of the cetacean carpal bones has often been discussed 
but without any very convincing success, and it is likely that the question 
can never be settled. Malm (1871) gave it much attention. Briefly, in 
the Odontoceti the pisiform, or what appears to be this bone, is occa- 
sionally represented. Besides this there are either five or six elements, 
or even additional ossicles which are believed to be supernumerary in 
character, for the number may not be uniform upon the two sides of 
the same animal. Usually because of their position a scaphoid, lunar 

"It should be noted that the drawings of mysticete limbs (fig. 36) have been 
made from mounted specimens in which the cartilage has been replaced by 
some modelling substance. Hence the cartilaginous elements of the carpus are 
not individually defined. 



and ulnare may be identified, but at times (see Platanista, fig. 36) the 
conditions are more uncertain. Thie second row of bones cannot be 
named with any feeling of confidence. Some of them are surely com- 
pound but their constituent elements can not be determined. In Glohio- 
cephala (fig. 36) the fifth metacarpal may be fused with the ulnare, 
thus in effect connecting this digit with the ulna, and in Platanista (fig. 
36) the same end is attained by the direct conjunction of the metacarpal 
with the ulna. Variations of this character might be enumerated at 
great length but without much profit. 

In the Mysticeti the carpus may have as many as nine elements (in 
Eubalaena, according to Holder, 1883) including a presumable pisiform 
extension, five in Sibbaldus, or even as few as three in Balaena, accord- 
ing to Flower (1873), but the latter is not clear as to whether this 
includes the pisiform, which is present. 

I do not see that the undoubted trend of the cetacean carpus can be 
determined. Is it in the direction of a bony mosaic, like it has been 
in marine reptiles, or toward a completely cartilaginous carpus, as the 
condition in Mysticeti might indicate .»* The decreased rate of ossifica- 
tion even in odontocetes might bear but the latter possibility, but for 
all that we know the two groups may ultimately attain opposite goals 
in this respect. 

In the Cetacea the metacarpals of those digits having considerable 
length are absolutely indistinguishable from the proximal phalanges. 
They all have precisely the same function, or rather practical lack of 
function, and hence the two elements may be considered to comprise a 
homogeneous complex. 

There are invariably five digits in the Cetecea except in the case of 
the rorquals, which have four. As might be expected the development 
of the border digits is largely dependent upon the conformation of the 
flipper as a whole. In those sorts having a pollex this is always short, 
but the fifth digit may be either practically as long as any of the others 
{Platanista) or extremely rudimentary in a flipper of such extreme 
falcate shape as that of Globiocephala. In those sorts with narrow 
flippers the digits are situated close together, while if these mem- 
bers are broad the digits are disposed somewhat fanwise. Invariably, 
I believe, the second one of the digits present is either the longest, or 
practically as long as any other. {Platanista) . Unlike the case of the 
Otariidae the digit upon the anterior border of the flipper is (as a rule) 
not markedly more robust than any other, although Flower and Lydekker 
(1891) have stated that in Sotalia the two outer digits are heavier than 




the other three. It can accordingly be deduced that if the flipper of an 
aquatic mammal be employed as an oar the digit upon its anterior border 
will likely be more robust than the others, while if it be used purely in 
equilibration there will probably be no digit markedly heavier than the 

As in the sea-lion and sirenians the subcutaneous tissue between the 
cetacean digits has been built up so that the flipper surfaces are plane. 
This interdigital tissue is very tough, and at least in adults of some 
sorts (Neomeris) is strengthened by a network of fibers which have a 
tendency to converge distad from the bases of the digits, suggesting the 
manner in which the heads of the femur and humerus are ordinarily 
strengthened osteologically. 

As there is such variation in the width of the cetacean manus it is 
only to be expected that there would occur a corresponding stimulus for 
digital differentiation, and there might conceivably be a tendency for 
digital reduction in one group of whales and the opposite tendency in 
another. Consequently, in spite of the fact that there are but four 
digits in the rorquals, it was not inconsistent with the possibilities for 
Kiikenthal to claim that at least the broad-handed beluga {Delphiuap- 
terus) exhibits a trend toward the polydactylous condition of the ichthyo- 
saurs (in which there occur as many as nine digits) and fish, because 
he discovered in the white whale that the two phalangeal elements (but 
not the metacarpal) of the pollex were double, one pair being situated 
beside the other. This is a possibility that can not be lightly discarded, 
but as the only beluga in the National Museum with articulated manus 
has a pollex entirely normal, and no one else besides Kiikenthal has 
ever encountered such a doubling in this or any other cetacean, 1 must 
presume that his material was pathologic. 

The same investigator (1893) argued with his customary vigor that 
it is not the pollex which is absent in the rorquals but the third digit. 
Attempting to carry this argument still further he made the remarkable 
statement that in Balaena mysticetus the metarcarpus of digit 1 is not 
really a part of the pollex but represents a prepollex, in spite of the fact 
that in Etibalaena the same detail is provided with two phalanges. His 
reason for this course of action was that he had "occasionally" discovered 
in Sibbaldus musculus a structure which he interpreted as constituting 
atrophied phalangeal segments lying between digits three and four, and 
entirely disassociated from the carpus. Such a situation would be quite 
astonishing. In the first place it would be extremely unlikely indeed that 
the more distal part of a digit could persist as a remnant in the connective 



tissue of the flipper while its metarcarpus and all carpal sign of its 
original situation had disappeared without leaving the faintest trace. In 
the second place supernumerary digits, of greater or lesser completeness, 
not infrequently occur upon the manus of divers sorts of mammals, 
even to the point where these are perfectly functional, and this does not 
necessitate the implication that derivation was from an ancestor which 
normally had six or seven digits. 

In the Cetacea the rate of ossification of the digital centers is very 
slow and one seldom encounters a specimen of sufficient age to show 
the bony parts closely approximated each to its neighbor. More fre- 
quently in the articulated but otherwise cleaned manus there is a con- 
siderable hiatus between each element, this space being occupied by 
cartilage that has shrunk as it dried. Each phalanx is flattened, truncated 
at the ends, and somewhat constricted in the middle. The proximal bone 
(metacarpus) of each digit is the longest and most robust, and there 
is a progressive decrease in size distad. Almost invariably these ossicles 
are longer than broad, but Flower and Lydekker (1891) stated that ex- 
cept for the proximal phalanges of digits two and three, the reverse is 
the case in Orcella. The shorter digits may have only two or three 
phalanges, but the longer ones are furnished with a greater number, 
there occasionally being as many as eleven (at least) elements in the 
adult, and a considerably greater number may occur in the young (Glo- 
biocephala, Phocaena) . This is the condition known as hyperphalangy, 
which occurs in no mammal other than the Cetacea, although it is charac- 
teristic of certain of the large, extinct, aquatic reptiles. Before con- 
sidering the possible explanations for the manner in which this has been 
brought about it will be well to examine the conditions now obtaining 
in cetacean digits. 

Flower (1876) has stated that in the Odontoceti the phalanges are 
often connected together by imperfect synovial joints. So far as I know 
no other investigator has independently made this claim and although 
it cannot be denied unqualifiedly, I would consider it as extremely un- 
likely that this is ever the case. There is not the faintest sign of a 
synovial cavity in the young odontocetes which I have had the oppor- 
tunity of examining and it is not likely that such would develop only 
in adult life. In a cetacean fetus of say one third term size, it is seen 
that each digit is composed of normal hyaline cartilage with a small 
center of ossification at its middle. Investing each cartilage is what 
appears to be perichondrium, and this covers the ends of the elements 
as well. There is not the least sign of an irregularity, cavity, or articular 



cartilage, and the latter is undoubtedly never differentiated at all in 
these mammals, so that there is no process of changing over of a synovial 
joint into a syndesmotic or fibrous one, but rather a later, direct altera- 
tion of the interphalangeal perichondrium or periosteum into fibrous 
tissue as there develops need for interphalangeal strengthening. In 
effect, then, the phalangeal articulations of the Cetacea are in the na- 
ture of synarthroidal (rather than synchondrosial) joints or sutures, 
differing from most sutures as we know them only in that they occur 
between cartilages rather than bones, at least until old age. 

The subject of hyperphalangy in the cetacea has been productive of 
a perfect furore of argument and the most violent partisanship for one 
theory or another. The theses that have been advanced in the endeavor 
to explain this condition may be said to number five, and to these may 
be added a sixth, which hardly seems promising, and a seventh consisting 
of the theory to which I incline. 

(i) Leboucq (1889) considered that the Cetacea are descended 
directly from swamp-inhabiting creatures on the order of amphibians, 
rather than from a terrestrial mammalian stock ; hence that hyperphalangy 
has been directly inherited from this ancestor. Presumably, then, he 
would consider that this condition can be explained in the same way 
that Howes has argued (in 3) ■ 

(2) Steinmann (1912) advanced the belief that the Cetacea are 
derived directly from the ichthyosaurs rather than from land mammals. 
It was therefore necessary that he consider that hyperphalangy has been 
inherited from these reptiles, which does not explain its origin. 

(3) Howes and Davies (1888) considered that hyperphalangy has 
been attained by means of the independent ossification of intercalary 

(4) Kiikenthal (1889) was persuaded that this condition was 
brought about through the initial separation and subsequent independent 
development of the phalangeal epiphyses, and a number of others have 
favored this view. 

(5) Weber, Bauer and others have held that hyperphalangy was 
initiated through the secondary division of a predigital strand of carti- 
lage, presumably on the order of those present in the Otariidae. 

(6) There may be added to this group the rather unlikely theory 
that supernumerary digital elements have occurred through the funda- 
mental division of the phalangeal anlage. 

(7) A further premise, and one which to me seems to hold the 
vr>n<it promise, is that hyperphalangy was initiated by the addition of 



one simple cartilaginous element after another, according to need, upon 
the tip of the normal digit. I claim no great originality for this. It has 
been implied, at least partially, in the past, but I know of no one who 
has really championed it as the method by which hyperphalangy has 
been developed. 

{1 and 2).I believe that the countless anatomical details of the Cetacea 
that are either fundamentally or precisely similar to those of terrestrial 
mammals constitute overwhelming evidence that the order is directly 
derived either from a strictly mammalian stock, or at least from the 
same group of mammal-like reptiles from which the land mammals took 
origin. I hasten to affirm that the latter statement is made purely in 
the line of argument, and that there is not as yet the slightest reason 
for believing it to be so. Furthermore, if one believed that the Cetacea 
were derived directly from amphibia, or "swamp-inhabiting creatures", 
then anatomical details render it obligatory also to embrace the hypothesis 
that all mammals had a similar direct ancestry, rather than from ter- 
restrial, mammal-like reptiles. The possibility that the Cetacea sprang 
from the ichthyosaurs is utterly untenable, in which opinion every anato- 
mist will agree. Before progressing to the next question, however, it 
is in order to examine more thoroughly the quality of the hyperphalangy 
that occurs in the manus of aquatic reptiles. This, of course, varied in 
direct degree to the completeness of aquatic adaptation, and relative to 
other considerations. Thus the flipper of the highly aquatic Geosanrus 
(fig. 34) had but one more phalanx than the normal mammal, but the 
single carpal element upon its anterior border, the first metacarpus and 
first phalanx of the pollex, are enormously more robust than the other 
comparable ossicles, which probably indicates that the flipper was long 
used as an efficient oar before it became so reduced in size. In Elasmo- 
saurus (a plesiosaur) hyperphalangy was about as far advanced as in 
those whales which show this to best advantage (although other elements 
of the arm differed considerably) . Some ichthyosaurs had as many 
as twenty-six phalanges in the third digit, and in these reptiles the bony 
elements of the entire arm were essentially similar, save for the slightly 
larger size of the humerus, radius and ulna, and were all closely packed 
in a bony mosaic. Williston (1925) mentioned that this condition could 
not be due to ossification and separation of epiphyses, for these reptiles 
had none of the latter. Whatever was the mode of their development, 
it is not necessary to infer that this was the same as in the Cetacea. Both 
groups encountered stimuli for an increase of phalangeal elements and 
both have exhibited this convergence, but it is by no means certain that 


the structures are entirely homologous. It seems odd that the turtles 
have apparently been unable to attain the condition of hyperphalangy, 
although of course it is entirely possible that the stimuli which they 
have encountered were of a different sort. Be that as it may, in spite 
of the fact that the manus of the most specialized turtles have the pha- 
langes very much elongated, no additional elements have been acquired. 
Presumably their flippers are entirely satisfactory as swimming organs 
except for the fact that the length of the phalanges renders the bones 
more liable to breakage, and signs of the previous fracture of the manus 
is frequently to be seen in skeletons. In such turtles as the leatherback 
the fact that the first digit is not much the heaviest, in spite of the flipper 
being used as an oar, is without great significance, for it is not elongated 
like the middle three. 

(3) Perhaps the majority of present-day cetologists incline to the 
theory of hyperphalangy by intercalary syndesmoses. Howes and Davies 
advanced this thesis as applying to the Cetacea after having studied the 
small accessory phalangeal element in several amphibians. Their con- 
tention was that the interosseous spaces are filled with fibrous tissue which 
takes the place of synovial joints. This structure might appear to be 
a derivative of the phalangeal mvesting tunic, but rather is it a differen- 
tiation of the mass from which the phalanges themselves are derived, 
although intimately related to the sheath; so that the phalanges and 
syndesmoses are, together with the investing sheath, differentiations of 
a continuous common blastema. Thickenings of the fibrous interosseous 
tissue might then become phalangeal elements. Thus is indicated a pos- 
sible intercalary origin, from articular syndesmoses, of the supernumerary 
phalanges of the Cetacea. This may have been associated with loss of 
the ungues in a manner similar to that in which elongation, by regular 
segmentation, of the cartilaginous rays in the paired fins of the Batoidei 
(skates and rays) would appear to have been connected with the dis- 
appearance of the horny fin rays. 

I have no criticism to make of the above facts and theories in so far 
as they concern amphibia, but they should not have been applied to 
the Cetacea. In the first place, as these intercalary bones are derived 
from fibro-cartilage they are in effect nothing but sesamoids and al- 
though occasionally these may reach a considerable size, there is no 
reason for believing that in mammals they can ever develop to the exact 
size and proportion of a phalanx derived from hyaline cartilage, nor 
that such could ever truly become interpolated as a part of the digital 
series. Even if they could do so in amphibians it is no proof that this 



would be the case in mammals. Furthermore, in spite of the fact that 
hyperyhalangy is much more marked in some whales than others, and 
even in some digits of a particular individual than in other digits, all 
phalanges are perfectly and uniformly graduated in decreasing size distad 
from the carpus. If the supernumerary phalanges had intercalary origin 
of this sort it is, to my mind, entirely inconceivable that all of these 
added elements could now invariably occur in perfect and complete de- 
velopment, being indistinguishable from the true phalanges. Rather 
would there frequently be encountered a number of elements throughout 
one particular phalangeal series, at least in some whales, which were 
markedly smaller and different in shape from the true phalanges. Fi- 
nally, what appears to render the theory incontrovertibly untenable 
insofar as concerns Cetacea is the virtual certainty that were it a fact 
there would be embryological evidence to support it. In the first place 
there would, during embryonic development of the arm, likely be a 
stage by no means short during which the intercalary elements would be 
relatively much smaller than the phalangeal ones. In the second — and 
this is the most important point of all — had any of the phalanges ever 
had origin by intercalary syndesmosis, the cartilage in which they are pre- 
formed ivould have to he fibro-cartilage, while that of the true phalanges 
would be hyaline cartilage: but all of them are uniformly hyaline. 

(4) Kukenthal argued with his customary ingenuity that hyper- 
phalangy in the Cetacea is attained by means of the separation and sub- 
sequent development of the phalangeal epiphyses. According to this 
theory, then, the retardation of the ossification of the phalangeal shafts 
allowed greater individuality of the epiphyses, and these, lacking the 
stimulus for conservativeness usually imposed by synovial joints, were 
permitted to increase in size, following a trend toward longer digits, until 
they had attained the exact conformation of the phalangeal shafts. The 
facts that this process could account for but eleven or twelve phalanges, 
and that the proximal ones still show double epiphyses did not deter the 
sponsor of this theory in the slightest; nay, it but stimulated him to 
claim that after this had once taken place, it occurred all over again a 
second time, or a third if necessary, giving birth to as many elements as 
needed. Unlike Zeus, however, these epiphyses could not spring forth 
full panoplied, as perfect phalanges, and if they had such origin there 
would surely be encountered various stages in their development, so 
that in a single digit of at least some whales there would not be a uni- 
form diminution in all the phalanges, but some would be abruptly smaller 
than their neighbors, while some phalangeal shafts would have epiphyses 


and some would not. This should invariably be more marked in the 
embryo, but no indications of it have ever been discovered. 

(3) Weber, Bauer and others have subscribed to the behef that 
hyperphalangy has been accomphshed by the segmentation of a predigi- 
tal strand of cartilage. It has been mentioned that in Ornithorhynchus 
the part of the webbing of the manus that projects beyond the nails is 
slightly stiffened by linear thickenings of the skin, running in continua- 
tion of the digits. These may be partly or wholly subcutaneous in 
situation, however. Of course skin and cartilage is very different in 
structure, but it is well known that where there is real need for the 
stiffening of a membraneous extension of the skin in mammals this will 
be accomplished, where practicable, by rods of bone or cartilages (as 
in flying squirrels and bats) . Hence it is reasonable to suppose that the 
predigital cartilages of the sea-lions had their inception in just this 
manner. And where there is well formed cartilage, ossification may, 
without difficulty, eventually take place. The developmental process 
involved in the initiation of predigital cartilages is, of course unknown, 
but it cannot reasonably be doubted that this was intimately correlated 
with the fact that there was no longer the stimulus furnished by repeated 
contact of the tips of the true digits with a hard surface. There were 
other factors concerned, however, as indicated by the accomplishment 
of digital extension in bats and marine turtles by means of elongation 
of the bones. There seems to be nothing particularly remarkable in this 
process. Rather would the incomprehensibility lie in the question of 
how the cartilaginous strands could become segmented so as to develop 
separate centers of ossification for the formation of phalangeal elements 
precisely like those of the true digits. One would reasonably expect 
that the ossicles situated distad of the latter would exhibit some abrupt 
transition in character, and even if this modification originally took place 
so far in the past that this transition was not apparent in the adult, it 
would almost certainly show to a striking degree in some embryonic 

(6) Unlike the case of the axial skeleton the bones of the appendages 
develop from somatic, mesenchymal condensations. Occasionally, from 
unknown causes, some of these anlage may experience reduplication. 
This is known because sometimes .one hears of a man, cat or pig with 
six or more well-formed fingers, or one phalanx more or less than 
normal upon each digit. This, apparently, is different from the process 
known as twinning, in which the tip of the finger is split, and even the 
second phalanx may be double. A sixth finger should not be looked 
[ 264 } 


upon as a splitting into two parts of any particular digit, but as an initial 
laying-down of six instead of the usual five digital elements. 

Why could not the hyperphalangy of 'the Cetacea be attributable to 
a similar supernumerary duplication of the phalangeal anlage? This, 
according to my notion, constitutes a theory that is plausible in more 
respects than that of either intercalary syndesmosis or epiphyseal differen- 
tiation, but nevertheless there are particulars which render it unsound. 
Duplication of the individual phalangeal anlage would not be sufficient 
to account for it. There would have to be reduplication not once but 
several times, in the case of the digits of those Cetacea in which hyper- 
phalangy is most pronounced, of the entire digital complex. And this 
would have to occur not side by side, as occasionally encountered among 
Mammalia, but in tandem, so that an entire new digit with three 
phalanges would be superimposed upon the end of the original digit. 
But this has never been reported in mammals. If, however, we presume 
for the sake of argument that such an occurrence be possible, the hy- 
pothesis can be entertained only until the embryological evidence can 
be examined. If it were fact then the embryological development of 
each phalanx would be obliged to progress at approximately the same 
rate as of all other phalanges. The entire digital complex would have 
been derived from the three original phalanges, and the ultimate cetacean 
phalanx would develop at the same rate as the third. But this is not 
the case. The differentiation of each phalanx is a trifle slower than that 
of the one next proximad, which effectively disproves the possibility 
that the supernumerary elements could have been derived by this sort 
of reduplication. 

(7) Beddard (1910) indicated an opinion that conditions in the 
Sirenia are more pertinent to hyperphalangy than those in the Otariidae, 
but he therefrom progressed to a consideration of the Amphibia, which 
is an entirely different question. 

Both Flower and Lydekker (1876, 1891) have stated that there are 
never more than the usual number of phalanges in the Sirenia, but some- 
times in an articulated manus (at least of the dugong) that has been 
properly cleaned there is to be seen a small, cartilaginous button upon 
the tip of the ungual phalanx (fig. 35) , as mentioned by Beddard. This 
I consider to constitute the initiation of hyperphalangy. 

The fact that the digits of the terrestrial Mammalia exhibit a quite 
remarkable conservatism in that they never develop phalanges in ex- 
cess of the normal complement is probably partly attributable to the con- 
servative influence supplied by the almost invariable frequency with 


which the tips are applied to some hard surface — usually the ground. 
If such habits be entirely relinquished it is only natural that some de- 
velopment should follow which one would never expect to encounter 
in the normal mammal. This should be further stimulated, or com- 
plicated, by the tendency for the disappearance of the nails, a decreased 
rate of ossification which seems to be characteristic of the mammalian 
flipper, trend toward the alteration in the character of the manual joints, 
and an unknown number of other factors. Certainly the reduction of 
the nails has been followed by an irregularity in the form of the ungual 
phalanges not only in the Sirenia, but the Otariidae as well. It does not 
seem at all remarkable to me that this should be followed in the former 
group by the addition of a cartilaginous, predigital nodule. The con- 
dition that has brought it into existence will doubtless stimulate it to 
further differentiation, accompanying what may be termed the simplifica- 
tion of the true phalanges: so there is doubtless a tendency for a conver- 
gence in the characteristics of the ungual phalanx and the predigital 
element. What can happen once can happen again, and a second pre- 
digital nodule could be added in course of time, as the first increased in 
size, and this would continue to occur as long as there was an activating 
stimulus. As a matter of course there would eventually appear a center 
of ossification in each nodule as it attained sufficient size. Presumably 
the start of this process occurred at a sufficiently remote period so that 
the transition in form between the true ungual phalanx and the first 
accessory element is now almost insensibly gradual. The growth of the 
distal digital elements is therefore very similar to the successive distal 
caudal elements, as they become defined after birth in a rodent or other 
mammal with unusually long tail. 

But there can be nothing to this theory unless it be upheld by em- 
bryological evidence. This should consist of a condition in which the 
digital bones are not all laid down at once, but after the embryological 
differentiation of the first four phalanges (including metacarpus), there 
should be a gradual addition, one after the other, of digital cartilaginous 
segments, in each of which ossification is slightly less pronounced than 
in the segment next proximad. And as this process progresses the flip- 
per length should gradually grow distad from the tip of the original 
limb bud. This, apparently, is exactly what happens, and this theory 
to account for hyperphalangy is the only one yet advanced to which 
the known embryological evidence supplies any confirmation. 

Presumably the difference between this sort of addition of cartilaginous 
nodules, one by one, to the digits, and the predigital cartilage of otariids, 


is rather basic. In the former case the cartilages are added as fast as 
there is room for them between the original digit tip and the slowly 
extending flipper border. In the sea-lion it appears probable that the 
flipper border extended considerably beyond the phalanges before there 
was any indication of cartilaginous stiff^ening. This suggests that com- 
paratively recently, geologically speaking, the latter animals were in the 
habit of folding back the predigital part of the flipper in precisely the 
same manner as they now fold back the predigital part of the pes. 

Leboucq (1889) claimed to have found a thickening of the cetacean 
epidermis above the digits well back from the flipper border. This he 
interpreted as the remnant of a nail, and Kiikenthal held the same view. 
I believe this is no longer considered as probable, if for no other reason 
than that after the disappearance of a true nail any thickening of the 
epidermis in the same situation would speedily be obliterated. Assum- 
ing that the vestige of nails still persists, however, there has been con- 
siderable argument over the question of where such rudiments would 
be situated — whether upon the flipper border of farther back. The an- 
swer to this hypothetical question depends entirely upon what one be- 
lieves to have been the origin of hyperphalangy. If one subscribe to the 
belief that this had intercalary origin of some sort then he must also 
believe that the true ungual phalanx, and therefore any rudiment of 
nail, is situated upon the flipper border. If he entertain the conviction 
that accessory phalanges were added distad of the original digit, then 
must he argue that any possible nail rudiment must be situated above 
the fourth phalangeal element distad of the carpus. 

[ 267 .1 

Chapter Eleven 

The Telvic Limb 

In some respects the tendencies followed by the hind limbs of aquatic 
mammals are similar to those shown by the anterior appendages, and in 
others they are very different. In at least one detail the stimuli for the 
aquatic specialization of the posterior limbs are less complicated than 
of the anterior pair for the reason that the latter are often, or at least 
for a long time, of great service in the securing and consumption of food, 
while the hind limbs never have this function. Thus there are three 
primary stimuli acting upon these members of an aquatic mammal; the 
influence introduced by the mode of terrestrial locomotion, aquatic loco- 
motion, and equilibration, but unless specialization be very far advanced 
the two latter are often very intimately combined. Terrestrial locomo- 
tion introduces a variety of factors, for when an animal first takes to 
the water its hind limbs have been fashioned not only for a particular 
mode of walking, but as a support, and differ according as the weight is 
heavy or light, the foot plantigrade or digitigrade, et cetera. Hence 
when a mammal first swims it already has an equipment that may pre- 
dispose it to certain definite motions of the limbs to accomplish swim- 
ming, and prevent or handicap it in indulging in other kinds of ac- 
tions. Naturally the difficulties which a shrew must overcome in evolv- 
ing a paddle from its feet are very different indeed from those which an 
elephant would be obliged to surmount. For these reasons it is entirely 
impracticable to unravel the separate factors and be sure in all cases 
that we know just what we are talking about. But average trends may 
be evaluated and probabilities advanced. 

It seems certain that in practically all of the terrestrial ancestors of 
living aquatic mammals strictly quadrupedal locomotion was employed, 
and that none of them was a leaper, or a flyer, or any other sort of ani- 
mal in which one pair of limbs was used chiefly in progression. Hence, 
during the first inefficient attempts at swimming all four limbs were 
kicked with as nearly equal forcfe as the muscular equipment of the 
animal would allow. In most mammals the hind legs are not only 
appreciably more powerful than the fore limbs, but the hind feet have 
a larger area. Hence from the very start these members had inherent 



ascendancy for potential development, which, coupled with the fact that 
the most efficient method for propelling a body through the water is 
from the rear, introduced the extreme likelihood that they would at 
some stage of aquatic specialization be of more importance for swimming 
than the fore limbs, //nless the conditions were complicated by other 

In general the function of the hind limbs of a mammal that is ex- 
clusively aquatic may be of three sorts: that of propulsion, of steering, 
or of no use at all. Naturally, partial terrestrial dependence will intro- 
duce complex difficulties that may vary in their influence from 100 per 
cent to zero, according to the individual case. In a mammal of the shape 
of a hippopotamus and without a respectable tail, it would be expected 
that always the development of all four limbs will be at a fairly even 
rate, neither pair at the expense of the other, as in the case of some of 
the aquatic reptiles of somewhat similar body conformation. This is so 
for the reason that the mud-turtle type of swimming is employed, as 
already discussed. To some extent this is also so in the case of the 
capybara, and is not so essentially dependent upon the fact that in both 
animals there is great dependence upon the land. 

Lack of ascendency, or at least delayed ascendency, in the development 
of the hind limbs may also be attributable to the fact that the tail was 
already of such a form that it was enabled to take over the function of 
chief swimming apparatus before the hind feet could become highly 
modified for this purpose, as seems to have been the case in the river 
otters and the insectivore otter (Potomogale) . 

If the body be not unduly broad (hippopotamus) or if the tail be not 
already large and powerful (river otter), the hind feet may be expected 
usually to gain ascendency over the fore, and, for a time at least, to be 
the principal means of swimming. This statement needs qualification, 
however. The anomolous exception is the sea-lion, and we are utterly 
incapable of stating whether the development of this order is because 
of some individual idiosyncracy in its make-up, or whether such a tail- 
less mammal as the capybara might be expected to follow the same 
course, if it should ever become pelagic, rather than that of the seal. 
I judge, however, that the sea-lion is an exception, chiefly for the reason 
that the most efficient means of propulsion is from the rear. 

In most sorts of mammals that become aquatic the tail is of goodly 
length but of small diameter at base, and in such there will always be 
ascendental development of the hind feet. This takes the form of en- 
largement of these members, and either a growth of stiff, bristly hairs 


Figure 43. Left hind feet of {a) platypus {Ornithorhynchus) ; (b) the in- 
sectivore Galemys ; (c) the marsupial Chironectes ; the otters {d) Aonyx. 
{e) Microaonyx, and (/) Lutra maculicoUis; {g) sea otter (En hydra) \ 
(h) beaver (Castor); and (/) muskrat (Ondalra): (d, e, and / redrawn 
from Allen). 



along the borders or webbing between the toes, or a combination of the 
two, and this speciahzation increases, theoretically, until the time when 
the tail shall have become sufficiently modified so that it is capable of 
furnishing greater propulsive power than the feet are capable of pro- 
ducing, when the latter will rapidly fall into disuse and tend to disap- 

If a mammal use its hind feet for swimming by any such method 
as that employed by the seal, with lateral motions in a sole-to-sole posi- 
tion, the action will be entirely symmetrical and the tendency will be 
for the feet to become symmetrical, with fifth and first toes of equal 
length and longer than the third, giving the foot a shape comparable 
to the tail of a fish. Any oar-like use of the hind feet will tend to pro- 
duce, or rather to accentuate, an asymmetrical foot, with either the first 
or the fifth toe the longest, according to the precise method in which 
the foot is used. Usually this manner of swimming consists in alternate 
strokes of the feet, these reaching either down or out so as to act upon 
relatively undisturbed water. There is no mechanical reason why an 
aquatic mammal should not swim by placing the broad part of the feet 
horizontally upon either side of the tail and oscillating the entire hinder 
end in the vertical plane, but no mammal is surely known to do this, 
although it is not unlikely that the sea otter may employ this motion. 
It is also not improbable that asymmetrical development of the feet, com- 
parable to that in the sea otter, might follow the phocid method of 
swimming provided there were also a tail of moderate length ; but then 
the feet would not be the sole means of propulsion, for the tail would 

The potentialities of the mammalian foot for furnishing alternating 
propulsive strokes of the character that is employed by most water birds 
seems to be limited, and the two can hardly be compared. If a jerboa 
(Dzpus), with its elongated, fused metatarsals now took to the water 
it might eventually evolve a propelling foot very similar to that of a 
duck or grebe, but so far as the evidence goes no aquatic mammal ever 
had a terrestrial ancestor with pedal equipment of this type. Nor is it 
probable that any highly aquatic mammal has evolved from the cursorial 
ungulate type with fused metatarsals, and there is probably no stimulus 
connected with a life in the water that would produce such fusion. 

In an aquatic bird such as a duck the propelling mechanism gains 
proper leverage by being upon the end of a tarsus of considerable length. 
The latter is relatively narrow and ofl^ers but slight resistance during 
recovery, while at the same time the foot collapses almost like a closed 



umbrella. Undoubtedly the important features, except for the webbing, 
of this equipment were evolved before the avian ancestry became aquatic. 
But in mammals the only terrestrial types roughly approximating this 
condition are saltatorial sorts of a desert habitat which would never be 
expected to become aquatic. It seems apparent that every mammal which 
has ever become highly aquatic originally had a shank but slightly nar- 
rower than the foot, and there seems to be no stimulus furnished by an 
aquatic existence for narrowing the shank. If the foot and shank in- 
crease in breadth together, as is frequently the case, so that a greater 
surface may be presented to the water, the shank would then be pro- 
ductive of too much retardation through recovery during swimming and 
the result would be impossibly inefficient. If the shank should become 
shortened or withdrawn into the body contour, as is now the case in 
the Pinnipedia, not only would even a greater amount of resistance 
during recovery take place, but the propelling feet would be so close 
to the body that they would lack requisite leverage and the animal would 
fiddle along at a slow pace. Hence, if an aquatic mammal is going to 
swim at speed by means of the hind feet it will have to evolve some 
other method of using them than by the alternate thrust-and-recovery 
of the duck. And this has been accomplished by the seals in the way 
in which they place the feet sole to sole and oscillate them from side to 
side. It may also be so after a different fashion in the sea otter. This 
method of swimming is the most effective of any that uses the hind 
limbs as the sole method of propulsion, but it could not be highly de- 
veloped in a mammal with a long tail for the reason that the latter 
would quickly outstrip the feet as a propeller. Just how and why the 
feet of the seals adopted this style of function while those of the sea-lion 
did not is a question about which one can only speculate. 

If a mammal that is highly aquatic swim chiefly by means of the tail 
the hind feet will not play any important part as equilibrators for the 
reasons that they they are too near the propulsive organ to be efficient, 
and as but one pair of equilibrators in the same plane seems necessary 
in mammals, the fore feet, being better situated for the purpose, will 
assuredly take over this function. The only mammal swimming chiefly 
by the fore feet and with a tail suitable for equihbration is the platypus. 
It is probable that its hind feet were at one time a greater aid to active 
propulsion than they are now. Whether these members are now really 
necessary for assisting the tail to function in equilibration is unknown. 
But it is requisite that a flat-bodied animal such as the platypus, or marine 
turtle, swimming by partly or chiefly vertical movements of the anterior 


limbs, should have an equilibrator or rudder for vertical steering at the 
hinder end. In the mammal the tail is of such form that it should be able 
to accomplish this unaided, while in the turtle, without an effective tail, 
the hind feet are specialized for this purpose. 

Attention may here be called to conditions in some of the extinct 
aquatic reptiles. In no highly modified reptile of this sort do the 
hmd limbs seem to have disappeared completely, although in the most 
specialized sorts, as Ichthyosaurus, they are considerably smaller than 
the anterior ones. In those reptiles in which the transition of the tail 
into a swimming organ was only partial, however, the hind limbs were 
practically as large as the forward pair. In some of the more primitive 
ichthyosaurs, for instance, such as Cymbospondylus, the body was more 
anguilliform, there was certainly no broadly spreading caudal fins, and 
all four paddles were about equally well developed. These details are 
suggestive that not only were the inherent variational capabilities of the 
reptiles diflferent from mammals, as already claimed, but that most of 
these large aquatic reptiles were descended from broad-beamed ances- 
tors, which obliged swimming after the mud-turtle method, somewhat 
like the case of the hippopotamus, for long ages. But the ichthyosaur 
of the more familiar type hardly has need for two sets of equilibrators 
in the same plane and the posterior feet were evidently in course of 
elimination when these creatures became extinct. 

Perhaps there are some who will question my conviction that in air 
breathing mammals, and presumably in reptiles, but one pair of equilibra- 
tors is necessary in the same plane, for they may recall that in many 
fish there are all sorts of accessory fins in a variety of situations. But 
this seems to be an entirely difi^erent matter. The equilibration of small 
fish is a much more delicate matter, and fish have precision apparatus for 
remaining in one exact position in spite of currents and other disturbing 
influences. In larger forms of difi^erent structure fins must be much more 
substantial (i.e. thicker), and furthermore, an air breathing vertebrate 
cannot spend an indefinite amount of time drowsing in one spot or 
threading its way delicately in and out of submarine growth deep be- 
neath the surface. Even in the sharks, as in some other large fish, there 
is the same tendency as exhibited in cetaceans and ichthyosaurs for the 
elimination of accessory fins and the development of a single pair of 
equilibrating paddles, in addition to a' dorsal fin. The tendency in large 
aquatic forms thus appears to be for the acquisition of three equilibrating 
devices spaced at more or less regular intervals. 



If this be so, as seems probable, it is only to be expected that when 
the tails of more primitive cetecea and sirenians had become sufficiently 
specialized so that they were capable of sustaining higher speed than 
the hind limbs alone were capable of attaining, the latter would be 
folded back against the body and, playing a decreasingly important part 
in the economy of the completely marine mammal, would finally suffer 
atrophy and disappear. 

For the purpose of the present study the posterior limbs of aquatic 
mammals will be discussed under several headings as follows: 

1. Hind feet the chief or exclusive swimming organs 

2. Hind feet used chiefly as equilibrators 

3. Hind feet absent or definitely subordinate in swimming 

In this simple classification there are naturally several forms placed under 
one heading that might with almost equal propriety be put under an- 


'■'■\ yj 

Figure 44. Dorsal outlines of the feet of a sea-lion (Zalophus), seal {Phoca), 
and elephant seal {Mirounga). 

1. Under this first heading are included (d) those forms in which 
the fore feet are used but very little or not at all in swimming but which 
have a tail of sufficient length to be of marked potential value in nata- 
tion, although the latter is not as yet sufficiently specialized to be the 
exclusive agent for propulsion; {b) all aquatic mammals which em- 
ploy all four feet in swimming and whose tails are of insignificant size ; 
and {c) such forms as swim exclusively by means of the hind feet and 
in which the tail is too small to be of use in this function. 

{a) Under this heading should be placed the single aquatic marsu- 
pial Chironectes, all Insectivora except the insectivore otters Lhnnogale 
and Potomogale, and all aquatic rodents except the capybara. 



Enlargement of the hind feet frequently results from a hfe in the water 
but this character may develop slowly and one must be careful always 
to compare forms in which approximately the same method of swimming 
is employed. It would be entirely without significance to compare in 
this respect the feet of hippopotamus, seal, otter and beaver, because 
the stimulus for foot development is different in each. In comparisons 
it must also be remembered that various terrestrial activities tend to in- 
crease the size of the hind feet. Thus the percentage of the hind foot 
length to head and body length may approach 30 per cent in arboreal 
squirrels. Probably for the reason that it is slightly given to arboreal 
activity this percentage in the wood rat (Neotoma) may be greater than 
20. But in several terrestrial shrews of different genera the above per- 
centage varies around 16.5, in various terrestrial rodents, such as the 
common brown rat {Rattus norvegicus) it is about 17.5, although in the 
same species it may vary from 16 to 19. In various other non-aquatic 
shrews, however, such as Sorex pacificus, the foot may be from 19 to 
23 per cent of the head and body, but this is probably because of some 
definite stimulus of which we do not know. On the whole it seems 
safe to say that this percentage in most terrestrial insectivores of the 
shrew type, and of rodents, will usually be found to be less than 20. 
Even in a rodent with definitely aquatic propensities it may fall below 
this figure (as little as 17.5 in Anicola) . Selecting a series of aquatic 
forms that chance to be available it is found that in the slightly aquatic 
genus Dasymys, without other modification of the feet, this percentage 
is 19 to 21, in Neomys foidens about 21.5, in Atophyrax (Sorex ben- 
dirii) 22.5 Ichthyomys (one specimen) 23, Hydromys 22.5 to 25, 
Galemys 26 to 30, Ondatra 38.5, and Castor 38 to 40. These figures are 
only approximately correct as they may be based on but a single specimen 
and the collector's measurements may not be accurate. They do show, 
however, that in such small mammals as swim chiefly by means of the 
hind feet there is a tendency for the elongation of these members, in 
degree according to aquatic modifications in other respects. Such elonga- 
tion may be barely appreciable until after the toes have become webbed 
or fringed, and may finally result in a foot that is twice as long as in 
a more generalized, terrestrial genus. There is frequently an appreciable 
lengthening of the toes in relation to the foot proper, and a more marked 
spreading of these members so as to present a greater surface (by means 
of webbing or fringing) to the water. 

Another modification which is often encountered in aquatic rodents 
and insectivores is, according to the spirit specimens that have chanced 


to be available for examination, the alteration of the foot posture in 
the direction of pronation, or tendency for the elevation of the outer 
border of the pes. The aquatic shrews Neosorex and Atophyrax, the 
genus Galemys, and the muskrat {Ondatra) all show this pronation to an 
extent of about 30 to 45 degrees. This is accomplished not through 
the ankle joint, or even, apparently, in the tarsus, but through the meta- 
tarsus and digits, thus affecting mainly the part of the foot which fur- 
nishes the chief propulsive power, and is very noticeable in the alcoholi- 
cally hardened specimens. Just what degree of further pronation, or 
of supination, from this posture is possible in life I do not know. Nor 
is it possible to determine the precise reason for this modification. All 
that can be said at present is that it is an adaptation by means of which 
these animals can employ the feet in a more effective manner for the 
purpose of swimming, for the action of the feet against the water de- 
pends entirely upon the position in which the legs are held. If the axis 
of the femur be maintained in the vertical plane then the feet would also 
be kicked in this plane, and the pronated posture of these members 
would tend to force the hinder part of the body mediad at each stroke, 
thus operating to neutralize the propensity which the anterior end 
would otherwise have of swinging toward the opposite side after each 
vigorous kick. Or if the femur and pes be held largely in the horizontal 
plane during swimming, then the pronated foot posture would tend 
to force the rear of the animal in a downward direction, which would 
have to be overcome by some antagonistic body posture or action. 

The pronated foot posture of these mammals, and of Ornithorhynchus 
as well, results in a slight rotation of some of the toes so that when 
the digits are slightly flexed the nails of the more lateral toes have a 
tendency to point in a lateral direction. 

In feet of small aquatic mammals a comparatively early modification 
is the acquisition of webbing between the toes or a fringe of stiff, bristly 
hairs upon the sides of the foot and toes. The latter is more often en- 
countered, but such fringes would be of little or no help in the case of 
a body of large size and in such instances they are never encountered. 
It seems that it is easier for a small foot to develop fringes than webbing 
but it is utterly impossible to state the reason for this difference. Ap- 
parently it is just a case of two mammals being able to respond differently 
to the same stimulus, as seems to be the case so frequently in other details. 

There is often a great difference in the precise character of the bristle 
fringes. They are present in varying definition and form in Desmana, 
Galemys, Neosorex, Atophyrax. Neomys, Chimarrogale. Crossogale, Nec- 


togale, Ondatra, Ichthyomys, Kheomp and Anotoniys. They may occur 
as a single fringe upon the outer border of the foot {Destnana, and 
Galemys fig. 43, b), upon both borders, upon the toes only, or the 
entire foot, including toes, may be heavily fringed {Ondatra, fig. 43, i) . 
Webbing may be either absent or present. Needless to say this fringing 
either supplements webbing or else is a make-shift to take its place. 
I say make-shift for I can hardly believe that it can be as efficient as 
webbing, although it undoubtedly helps in propelling a small body 
through the water. 

In many sorts of terrestrial mammals the bases of the toes are joined 
by rudimentary webs, and it is only natural that as these are of such 
importance in the economy of an aquatic mammal they should increase 
in area. Evidently they are easily developed by all classes of vertebrates, 
but why they should be lacking in such an essentially aquatic animal as 
the muskrat remains a puzzle. 

The form and relative proportions of the toes of the above sorts of 
aquatic mammals differ. In some the plan of these members is essen- 
tially normal. In some there is lengthening of the hallux so that the 
foot is practically symmetrical {Chhonectes and Castor'), and this is 
undoubtedly the most suitable shape for such mammals as swim by mo- 
tions that involve actions resulting in equal water force being applied 
to both borders of the feet. But while it is true that most rodents and 
insectivores are not as yet sufficiently aquatic for there to have been very 
profound changes in the pedal equipment, it must not be assumed 
that an asymmetrical foot falls short of the animal's requirements, for it 
may employ some method of swimming that involves asymmetrical foot 

Nails of the normal sort are present in the aquatic mammals men- 
tioned and where differences occur these are undoubtedly attributable to 
some habit that has nothing to do with the aquatic life. In Ornithor- 
hynchus the nail of the first digit is straighter and more spike-like than 
the others. In Chhonectes the hallux nail is not claw-like as are those 
of the other digits but is of the same general shape as the nails in man 
and is little more than a callosity (fig. 43, c) . 

The pes of the beaver (Castor) appears to be a more efficient paddle 
than that of any existing rodent. It is large, broad and fully webbed, 
and is operated by powerful muscles and heavy limb bones. Parsons 
(1894) found that its calcaneal tendon is very readily separated into its 
component parts. In this animal the nails of the first and second pedal 
digits are slightly more slender than those of the others (fig. 43, h) . 





TLE.X. D>&. LONG,. 




-FLEX.,Dl&. LONG, 



Figure 45. Posterior view of the leg musculature of the seal {Phoca, left), and 
sea-lion (Zalophus, right). 



That of the former is subtended by a soft pad, and of the latter by a 
similar pad in addition to a serrated, horny growth which Bailey (1923) 
affirmed functions like a fine-toothed comb in dressing the fur. 

An additional modification in these as well as all other aquatic mam- 
mals is the tendency for the elimination of the tubercles or prominences 
upon the soles of the feet. 

Such slightly aquatic mammals as the marsh and swamp rabbits and 
water rat {Arvicola), as well as the genus Neofiber, need no further 
mention in this chapter, for there has been no change in the feet. The 
feet of the coypu {Myocastor) are moderately webbed. Those of the 
Australian water rats, Crossojnys, Hydromys and Parahydromys vary 
from somewhat webbed in the first to unwebbed in the second. 

The aquatic stimuli brought to bear upon the hind feet of all the 
mammals discussed in the present connection (not Omithorhynchus) 
seem largely the same or essentially similar. All have tails that are 
at least moderately long and I believe that without exception all swim 
by alternate kicks of the hind feet while the fore feet are rarely utilized 
to any important extent during natation. At one time (Howell, 19'24) 
I reported catching a glimpse of a water shrew {Neosorex) swimming 
by kicking the hind feet in unison after the manner of a frog, but I 
am now of the opinion that I was mistaken, either in identifying the 
creature so hastily seen or else in my observation, for no one else has 
ever reported similar action in a small mammal. 

As already partially discussed in a previous chapter, small mammals 
of this sort quickly abandon any swimming motions of the fore limbs, 
chiefly, it seems, for the reason that the area of their hands is so much 
less than that of the feet that not only could they furnish very little 
propulsive power, but for the same reason they could not act efficiently 
in neutralizing the asymmetrical action imparted by the hind legs acting 
alone. For this latter purpose there is an efficient equilibrator already 
at hand in the tail, which furnishes as much stabilization in the water 
as does the tail of a kite in the air. Presumably the precise action upon 
the water of the hind feet kicking in alternation varies in different mam- 
mals, but it is a point extremely difficult of determination. At any rate 
there is one uniform result in that as the feet are kicked the effort swings 
the hinder end of the body from side to side, thus imparting a sinuous 
motion to the tail. This movement should be involuntary at first, as far 
as concerns the tail proper, but as the muscles concerned develop with 
use it is capable of being employed as a definite help to locomotion. 
Evidence shows that as a result the tail in some manner is modified so 



that its lateral aspects become broader than the vertical ones, either 
by the acquisition of a hairy fringe above or below or a flattening of 
the member in the transverse plane ; and of course with such a specializa- 
tion the efficiency of the tail as a swimming organ is thereby increased. 
This is the rule, but there are some exceptions. Presumably Chironectes, 
Aiyocastor and the Australian water rats are sufficiently aquatic to have 
developed a flattened tail, but the latter is perfectly terete. Whether they 
actually swim by some slight variation in the movements of the hind feet 
that does not introduce a stimulus for tail flattening, or whether the 
caudal appendage merely has unusually inherent resistance to such change 
but will later accomplish it, is unknown. 

The speed which the hind feet operating upon the kick-and-recovery 
principle are capable of imparting to the body is very definitely limited, 
as previously discussed. In mammals with a tail of respectable length 
there will almost certainly come a time (save possibly in the case of 
the beaver) when the tail and its musculature will have become suffi- 
ciently specialized so that it can propel the body at a faster pace than 
can the hind feet alone. The latter will then quickly fall into disuse 
unless they be employed for other important purposes. It is not un- 
likely that the tails of Desmana and Ondatra are now almost sufficiently 
specialized to accomplish more speedy swimming than the feet, but 
for a very long time at least the latter will continue a necessity to both 
animals for terrestrial activities. 

Most of these small mammals discussed are not sufficiently modified 
for there to have occurred marked change in any part of the leg above 
the ankle. Dobson (1882), however, has remarked that in Desmana 
the femur is but little more than half the length of the tibia, and the 
superficial biceps and semitendinosus have broadened considerably (fig. 
48), both of these alterations thus being in the direction followed by 
the seal. The gluteal muscles are also unusually robust, as in so many 
other aquatic mammals. In all aquatic mammals having an enlarged 
hind foot one will almost certainly find that the shank and thigh muscles 
are correspondingly strong. 

(b) Next for discussion is the case of such aquatic mammals as em- 
ploy all four feet in swimming but have a short tail, involving the polar 
bear, capybara, aquatic rabbits, hippopotamus and tapir. All have great 
dependence upon the land and in all the tail is insufficient in size to 
act as a rudder or stabilizer, and in all but the rabbit the hind feet are 
not much larger than the fore feet. The latter may thus function ef- 
fectively to overcome the asymmetrical, wabbling motion that the kick- 



ing of the hind feet alone would otherwise impart to the virtually tail- 
less body, and the animal accordingly trots through the water. The 
aquatic rabbits may possibly constitute an exception to this rule. If 
already the hind feet are not held sufficiently close together during pro- 
pulsion so that no stabilizing action of the fore feet is necessary to over- 
come any lateral deviation that would otherwise tend to occur, this 
might possibly be an eventual adaptation. 

The tendency in these mammals, except the rabbit with its negligible 
aquatic specialization, is for all four feet to develop equally, limited 
naturally by inherent functional difference, as, for instance, in the moder- 
ate webbing of the capybara. But the dependence upon the land is 
still too great for there to have been much change. 

This four-square method of propulsion is far from efficient and 
really high speed can never be attained without some adaptation pro- 
viding for an oblique action against the water by all four feet, such, 
conceivably, as that now employed by the walrus. For the same reasons 
as in the walrus it is now utterly impossible to predict whether the an- 
terior pair will gain the functional, and hence evolutional, velocity, as 
in the sea-lion, or the posterior pair, as in the seal, if indeed they do 
not have constitutional limitations that would prevent them from fol- 
lowing either course to a satisfactory conclusion. 

(c) Under this heading (as of page 274) there should be included 
Enhydra (which will be left for the last) , the Phocidae, and the Odo- 
benidae partially. The latter need be given scant attention for in both 
conformation and function the hind feet of the walrus is fairly inter- 
mediate between the seals and the sea-lions. In shape these members 
partake of the characters of both. They may be placed flat upon the. 
ground but the definition of the posterior part of the astragalus indicates 
the probability that they are developing the limitations in this respect 
of the Phocidae. The hind feet are said to assist swimming by lateral 
oscillations like the seals, or they may be used purely for equilibration 
like these members in the sea-lions. Which pair of limbs will ultimately 
gain the ascendency is a matter for pure conjecture. 

In the seals (Phocidae) the external features of the part of the hind 
limb above the foot are very similar to those of the sea-lion, and for 
the reason that the anatomical differences are more significant when 
treated comparatively, this part of the phocid limb will be discussed in 
connection with the otariid. Just why the seal adopted, or could adopt, 
its present method of swimming it is extremely difficult to say. Its foot 
is really much less remarkable than that of the sea-lion. Externally the 

Figure 46. Lateral view of the right hind limb musculature of the sea otter 
{Enhydra), from a specimen in the U. S. National Museum. 



noteworthy features consist of the hairy sole, or at least that it is as 
hairy as the dorsal part of the foot; the retention in all (apparently) 
genera but Mirouttga of sharp nails of fair size when there seems to be 
no use to which they could be put, the fact that the first and fifth digits 
are longer than the others, while the third is the shortest, giving to the 
pes when expanded a lunate shape comparable to the tail of a fish or 
whale; and the expected fact that the webbing between the toes is 

The foot of the seal has retained its proportional length in relation 
to body length, as judged by such a terrestrial carnivore as the cat, and 
therefore has not suflPered the marked shrinkage experienced by the 
shank and, especially, the thigh. In a partly dissected specimen of 
Phoca hispida the tendency of the feet was to adopt a position with 
soles not strictly parallel but with the upper (outer) borders slightly 
converging, but in the live animal this seems to be overcome by muscle 
or other tissue tension so that the soles are perfectly parallel. From a 
trailing position of the feet, movement through the ankle joint proper 
is apparently only about 30 degrees, but greater freedom is allowed by 
means of the articulations of the astragalus and calcaneum with the 
centrale and cuboid, this amounting to about 65 degrees. As a result 
the tarsus may be flexed approximately at a right angle and by this es- 
sentially hand-like action one foot is enabled to follow through and aid 
by its medial motion the lateral movement of the opposite foot while 
executing a swimming impulse. This tarsal joint, as it may be called, 
also enables the animal to progress very slowly by a manner first called 
to my attention by Breder (MS) and since observed personally. With 
feet palm to palm these members are flexed through the tarsal joints, the 
tips of the digits remaining together. The feet are then abruptly ex- 
tended, imparting a gentle though definite forward impulse to the en- 
tire body. This cannot be strong enough, however, to constitute a use- 
ful accomplishment. 

No great significance can be attached to details of the phocid tarsal 
bones, other than that already mentioned and the extension of the 
astragalus, discussed later. It is somewhat narrow, however, and thus 
sidewise (horizontal to the sole) motion is facilitated through looseness 
of the tarsal articulations. This tarsal narrowing causes a crowded, some- 
what overlapping condition of the proximal metatarsi. Spreading of 
the toes, in this animal at least, seems to be facilitated by an arrangement 
of the joints whereby this is accomplished by a partially oblique ex- 
tension of the first digit and a slight flexion of the fifth, this being com- 



parable to conditions in the muskrat, desman etc. The hallux is elon- 
gated and considerably more robust than the other digits, while the fifth, 
although equally elongated, chiefly by lengthening of the first phalanx, 
is only very slightly more robust than the three middle toes. 

No mammal swimming by means of the hind limbs alone could at- 
tain by any thrust-and-recovery method the speed of which the seal is 
capable, but would be obliged to develop some method whereby the 
pedes act obliquely upon the water. Mechanically there appear to be 
only three ways in which such a mammal as the seal could attain this 
end. One is by employing the hind limbs in the same fashion that the 
sea-lion uses its anterior flippers, adducting sharply in the transverse 
plane. This would probably be an impossible specialization. Not 
only does it seem utterly out of the question for the adductor muscles 
of the hind limb to evolve into anything comparable in efficiency with 
the enormous pectorals of the sea-lion, but any such development would 
almost certainly place the greatest body diameter in the pelvic region, 
and this would introduce very profound complications in the mechanics 
of its swimming. The second possibility might be for the feet to be held 
in the horizontal p'ane, soles up, with their inner borders touching, 
and oscillated in the vertical plane, as suggested for the sea otter. The 
third course left to it would be to take the direction which it has actually 
followed, and to swim by oscillations from side to side of not only the 
adpressed hind feet but the entire posterior end of the body. 

Just how the animal was enabled to do this is a somewhat puzzling 
question. It may, however, have depended on some determinant which 
in itself was very simple. Thus any habits which did not tend to de- 
velop the back muscles in swimming and did tend to throw the greater 
work upon the fore limbs would be apt to predispose the sea-lion ancestor 
toward its present course of evolution. On the other hand some mode 
of swimming when all four feet were being used for the purpose, tend- 
ing to develop the back muscles, as they evidently are now being de- 
veloped in the tailed muskrat, possibly coupled with some slight ana- 
tomical detail which one might judge of very minor importance, should 
have been sufficient to start the seal on the road to its present state. 

I have searched the literature and written to everyone likely of whom 
I could learn, including Alaskans' who have spent years about the Aleu- 
tian Islands, in an endeavor to ascertain precisely how the sea otter 
swims. No two accounts have been the same, for probably none of 
these informants has himself been sufficiently close to a sea otter moving 
through clear water at speed to be sure of the motions, or having ob- 



served, was satisfied with determining that there was movement of the 
feet. Consequently we must content ourselves for the time being with 
a judgment of the swimming method according to anatomical evidence. 

A brief examination of the external features of the sea otter will show 
that the fore feet are of small area but that the hind feet are much en- 
larged, with the fifth toe the longest and the others successively shorter, 
while the sole is densely haired. The remainder of the hind limb is 
unusually short. The tail is of moderate length, extending to the rear 
about the same distance as the tips of the toes when the limbs are trailed, 
and is said to be slightly expanded in a lateral direction. 

The anterior feet of Enhydra are too small to be of any practical 
use for swimming, and for the same reason the tail is not capable of act- 
ing alone. The hind feet are, however, sufficiently modified so that we 
can be sure that they furnish the chief propulsive force. Precisely how 
fast the animal can swim I have no idea, but according to accounts sev- 
eral men in a light skiff have all they can do on a long pull to catch 
up with one. It seems that they may be able to travel as fast as ten 
miles an hour, possibly with higher speed for a short distance. One 
can be sure that they are at least reasonably speedy swimmers, too, be- 
cause otherwise a pelagic mammal of this size would quickly be ex- 
terminated by sharks and killer whales, even though kelp beds are their 
favorite habitat. 

If this probability be conceded, then they cannot swim by direct 
thrust-and-recovery movements in alternation by the hind feet, for, as 
previously described, even moderately high speedj cannot be attained 
thereby. On the contrary some motion must be employed involving 
oblique action of the feet against the water, comparable to the swimming- 
actions of the seal. Can the sea otter swim in the same way as the seal, 
with lateral motions of the feet held palm to palm ? Indications are that 
it does not do so for the following reasons. In the seal the feet are 
shaped like the tail of a fish with lunate posterior border, and for the 
same reasons. Both first and fifth digits are elongated to accomplish this. 
If the sea otter swam in a similar fashion it is sufficiently specialized for 
it to show definite indications of the same development. This, however, 
might be complicated by the presence in this animal of a tail of moderate 
length. The tail of the seal is too short to be of any consequence, but 
it is possible that in the sea otter this member, of the same length as 
the feet, might in a mechanical sense take the place of the uppermost 
toe held against it in swimming posture. But the fifth toe is the longest 
and in order to act with the tail so that the posterior border of the swim- 



ming organ (feet and tail) would present the requisite lunate form 
the feet would have to be held back to back, instead of palm to palm, 
and this is either impossible or would at least require changes in posture 
which would be conspicuous externally. Of course there may be some 
idiosyncrasy about the swimming of this animal for which a non-lunate 
propulsive organ, and one with a longer upper than lower border would 
be advantageous; and this must be entertained as a possibility, but it 
is hardly likely. 

If the above arguments be correct, it seems that the only way left for 
the sea otter to swim efficiently with its existing equipment is to place 

Figure 47. Lateral view of the left hind limb musculature of the insectivore 
otter Potomogale, redrawn from Dobson. 

the feet horizontally to the rear, one upon either side of the tail, sole 
up, and oscillate them in the vertical plane on the same principle that 
the whale uses its flukes. It is true that no other mammal is known 
ever to have employed this method of swimming and accordingly there 
are doubtless those who will claim that it is illogical, but on the con- 
trary it appears to be a more logical method of swimming than that 
now employed by the seal. The external details of the sea otter support 
this presumption. The long fifth toes would form the outer borders 
of the swimming organ, as they should, and the tail, of almost if not 
quite precisely the length of the extended feet, would comprise a 
strengthening, central element. The whole would accordingly present 



in a satisfactory degree the lunate rear border theoretically desirable in 
such a propulsive apparatus. 

Fortunately it is possible to present some myological evidence in this 
discussion of the sea otter. There is available from the National Mu- 
seum the skinned carcass of one of these animals that was brought from 
Copper Island by Dr. Stejneger in 1897. This was evidently dried out 
at least once, for the muscles are leathery and the nerves brittle. Never- 
theless it is entirely suitable for investigating those muscles which 
might be expected to shed some light on the method of swimming em- 
ployed. Accordingly the more significant muscles of the hind limb are 
described and figured (fig. 46). Nothing further is attempted for I 
have no wish to encroach upon the chosen field of E. R. Hall in investi- 
gating the comparative anatomy of the Mustelidae. 

Reference to the illustration shows that the muscle usually termed 
caudofemoralis is large and powerful. It arises from the dorsal fascia, 
superficial except caudally where origin is deep to the semitendinosus. 
Insertion is upon the distal half of the femur and to the knee. At least 
in the majority of cases this muscle is merely the chief part, usually the 
whole, of the true gluteus maximus, for it is served, in the case of the 
sea otter as in others, by the inferior gluteal nerve. And what is so 
often called the gluteus maximus, where the term caudofemoralis is 
also employed, is proven by the innervation to be the gluteus medius. 

The tensor fasciae latae arises by tough aponeurosis from the ilium, 
and mediad from over the gluteus medius to an aponeurotic insertion 
upon the femur for a considerabe distance distad of the greater tro- 
chanter. It seemed to be innervated by a nerve thread coming from deep 
between the main gluteal mass and the quadriceps. This was undoubtedly 
a branch of the superior gluteal nerve, as usual, but was so brittle that 
it broke and I was unable later to pick it up again. 

As is so often the case the main gluteal mass was separable only with 
difficulty, the whole being fused toward insertion. It was so deep (23 
mm.) and hardened that the only way the innervation could be deter- 
mined was to pick the muscle to pieces, fiber by fiber. Three parts of 
origin could be detected. Gluteus A arose thinly from the tip of the 
ilium. The portion B arose from the dorsal fascia both craniad and 
mediad of the ilium. Both A and B increased rapidly in thickness and 
soon fused with the deeper and more robust part C, which arose from 
the entire lateral aspect of the ilium, the whole inserting very strongly 
upon the greater trochanter. The arrangement of the fibers was slightly 
multipenniform and the mass constituted a muscle of great strength. 



The deeper and most of the superficial parts were indubitably served by 
the superior gluteal nerve, but whether or not the caudal border of B 
may have received twigs from the inferior nerve cannot be stated with 
any certainty. If so then this part would have gluteus maximus affinity. 

The superficial biceps femoris was highly modified much in the man- 
ner that it is in the Phocidae (fig. 23). It arose from the tuber ischii 
and spread fanwise to a fascial insertion over the shank from the knee 
almost to the instep. The deep biceps was almost an exact counterpart 
of this muscle in both seal and sea-lion. It was very slender, arising 
from the deep dorsal fascia alone and perhaps a trifle anterior to the 
acetabulum, and passed to the vicinity of the outer malleolus of the 

It is frequently difficult to establish the homology of the semitendi- 
nosus and semimembranosus for the reason that they often shift about 
and the innervation is not particularly distinctive. Hence homology 
can seldom be indubitably established and one must name them as best 
he can, judging by generalities. In the sea otter the apparent semi- 
tendinosus is phenomenally developed, to a greater extent than in any 
mammal known to the writer. Origin is very broadly by fascia, the 
muscle increases in thickness and most of it passes to the medial side of 
the leg, where it inserts broadly along the shank, distad to its middle 
and partly under cover of the gracilis. Oddly enough, however, the 
fibers along the cranial border do not pass mediad of the shank with the 
remainder, but to the lateral aspect, inserting as a caudal extension of 
the superficial biceps. Accompanying the hypertrophy of the semitendi- 
nosus is atrophy of the semimembranosus, which is a weak slip of a 
muscle, some ten millimeters broad, beneath the other. 

There is a peculiar development of the sartorius. This arises from 
the tip of the ilium, with origin extending broadly onto the ventral 
border of the bone. As a broad, thin muscle it inserts for 40 mm. onto 
the knee and a bit distad. In continuation still farther distad over the 
proximal part of the medial shank is the insertion, for another 40 mm., 
of the gracilis. 

The lumbar part of the erector spinae was very robust, the powerful 
iliocostal part causing a flaring outward of the ilium, to which it is at- 
tached. The hypaxial lumbar muscles were, however, definitely weak 
and no larger than one would expect in any terrestrial mammal of equal 

In considering the above muscles it cannot be claimed that any particu- 
lar swimming method is clearly indicated. Because the interaction of the 


muscles in any strong motion is so intricate it is not safe to state that the 
muscular equipment is not fitted for the same swimming method as em- 
ployed by the seal. On the other hand, there are considerable differ- 
ences, which contribute evidence to the probability that the sea otter 
swims by vertical oscillations of the expanded feet. In such actions 
the muscles of the back are involved to elevate the pelvis. The erector 
spinae is certainly fitted for this. The hypaxial muscles are not fitted 
for depression, but the strong rectus abdominis can accomplish it, with 
a more favorable lever arm. The lesser gluteal complex is much 

Figure 48. Lateral view of the left hind limb musculature of the desman {Des 
fnana), redrawn from Dobson. 

Specialized for a purpose not particularly clear, for its action depends 
upon the habitual posture of the femur and the way in which other 
muscles co-operate. The gluteus maximus is powerful for further ex- 
tension of the femur after it is partly extended, and the superficial 
biceps, semitendinosus and gracilis form an extraordinarily efficient com- 
plex for elevation of the shank when the feet are extended to the rear, 
thus accomplishing upward thrusts oi these members. On the whole 
the muscular details do not detract anything from the above theory of 
the way in which this animal swims, but, on the contrary, contribute 
considerable coroborative evidence. 



The partial binding down of the hind limbs effects a static posture 
with these members more trailing than would otherwise be the case, 
and for the same reason the terrestrial posture must be with the back 
more arched, so that the sacrum is more elevated toward the perpendicu- 
lar in a manner suggestive of the sea-lion's posture. The evidence clearly 
shows that the feet are placed in plantigrate position with ease, although 
I have seen it stated in a letter that this is impossible. 

In examining a skeleton of Euhydra in comparison with one of Lutra 
it is seen that although the proportions of the ilium in regard to post- 
acetabular dimension is the same, the entire innominate is somewhat 
heavier, to accommodate stronger musculature, and the ilium flares con- 
siderably in a lateral direction, for a stronger iliocostal attachment. Tay- 
lor (1914) has stated that the pelvis is more nearly parallel to the ver- 
tebral column also and mentions that there is no evidence of a liga- 
mentum teres, as there is in Lutra. He also showed by drawings reduced 
to the same size that the thigh and shank bones of the sea otter are the 
heavier and that their relative proportions are not appreciably altered. 
The greater trochanter is slightly broader, because of the specialization 
of the lesser gluteal complex, and more seal-like. The metatarsals and 
phalanges are slightly flattened, and manipulation of the hardened spirit 
specimen seems to show that there is unusual freedom of movement in 
the articulation of the astragalus and calcaneum with the centrale and 
cuboid — a functional tarsal joint as in the seal, making the foot more 
handlike to facilitate oscillating movements when the pes is held on a 
line with the shank. 

Why the sea otter should have developed a swimming method essen- 
tially different from that seemingly in course of elaboration by the river 
otters is not entirely clear, but the fundamental reason which appears 
mose likely is that in the former animal the tail has always been too 
short and light to be used for propulsion alone. Presumably the hind 
feet will continue their specialization and become still more efficient, 
while the tail may either continue to function as an accessory, central 
stiffener for the swimming organ formed by the hind feet, as it now 
appears to do, or become shorter as the feet have less need for a support 
of this kind. Conceivably, however, the tail may ultimately increase in 
power and take over the function of primary propeller. 

2. Aquatic mammals which use the hind feet chiefly for equilibra- 
tion include Ornithorhynchus, the Otariidae or sea-lions and fur seals, 
and the Odobenidae or walruses partly. The latter need no further 
consideration here save to mention once more that their feet partake of 



the character of those of the sea-hon, with cartilaginous extensions of the 
digits, and in form somewhat resemble those of the seal. Apparently 
they are rather intermediate in their uses as well. 

The hind feet of the platypus are much less modified than the anterior 
ones, and in a somewhat different way. In all probability all four feet 
were formerly used for propulsion, on the mud-turtle principle, until 
the tail had become flattened and by more frequent use the fore feet 
were enabled to gain developmental ascendancy. Now the hind feet ap- 
parently are but rarely used to assist propulsion, but rather is their ac- 
tion comparable to that in the marine turtles — mainly for vertical and 
to a lesser extent for horizontal steering, in the former direction aiding 
the tail in depressing, or in elevating the body. They are fully webbed 
and articulation of the four lateral toes is such that these digits flex in a 
direction more toward the outer border of the foot and away from the 
hallux than in a strictly palmar direction. A variation in this flexional 
peculiarity has been noted in the case of Galemys, Sorex, Phoca and 
Ondatra, the advantage evidently being either for attaining a slightly 
different angle of the membranes with respect to the articulation of the 
ankle, or more probably so as more easily to secure the maximum ex- 
pansion of the membranes. 

For the purpose of more readily evaluating the differences between the 
pelvic limbs of seals and sea-lions these two groups of pinnipeds will be 
discussed together under the present heading. But first it should again 
be mentioned that seals swim by oscillations from side to side of the 
hind feet, placed palm to palm, and of the entire hinder end of the body, 
and that the hind feet are useless in terrestrial progression for the reason 
that the sole cannot be placed flat on the ground. The sea-lion, however,, 
can place the sole upon the ground, but only by curving the sacral ver- 
tebrae in a position perpendicular to the surface. In the water the feet 
are trailed, usually sole to sole much as in the seal, but they seem 
never to be used in the slightest degree as an aid to purposeful propul- 
sion. When playing or lolling about in the water the feet may be flexed 
in all manner of positions, assisting to alter the posture of the body, and 
I have watched a young fur seal progress very slowly by repeatedly roll- 
ing the entire body over and over, the hind feet thus acting to some ex- 
tent in a cork-screw manner as a propeller; but such inconsequential 
actions need not be considered. 

In scrutinizing the influences to which the pelvic limbs of the Pinni- 
pedia have been subject one must begin with the spinal musculature. 
In the Otariidae the latter is not appreciably modified and need not be 

C 291 1 


considered. In the Phocidae the spinal muscles, both apaxial and hypax- 
ial, are enormously developed and play a more important part in swing- 
ing the feet from side to side than do the intrinsic muscles of the limbs. 
The chief effect which this development has had upon the pelvis of this 
group is the fact that the ilium has been projected sharply to the side 
(figure 49) so that it may accommodate a powerful attachment of the 
iliocostalis lumborum. In other respects the innominates of these two 
groups exhibit many resemblances, chiefly brought about by the fact 
that in both the legs are always maintained in a traihng position. In 
both, also, a significant result of the latter posture has been a binding 
down of the shank so that the "crotch" consisting of the tissue joining 
the limb to the vertebral column, falls opposite the heel. As a con- 
sequence there is little independent movement of the limb bones, which 
means that flexion and extension of the thigh is very definitely circum- 
scribed, and that although rotation and abduction may be strong, they 
can be only through a short distance. This has resulted in reduction in 
the length of the femur, and of the ilium, or pre-acetabular part of the in- 
nominate, which accommodates the lesser gluteal muscles. The latter 
is but 16 per cent of the entire innominate length in a species of Phoca, 
and 32 per cent in Zalophus (an otariid), while in such a terrestrial 
carnivore as the cat this percentage is 59- In compensation the post- 
aretabular measurement, or pubo-ischial portion, is 74 per cent of the 
innominate length in Phoca, 55 in Zalophus, and 33 in a cat. Appar- 
ently the stimulus for the increase of this posterior part of the pelvis has 
been the increase in importance of the function of the muscles attached 
thereto as adductors of the shank. Especially is this true in the case of 
the seal, and parts of the ischium and of the pubis have been projected 
in a dorsal and a ventral direction respectively, which furnishes greater 
leverage chiefly to the superficial division of the biceps femoris and the 
gracilis (fig. 49). Ryder (1885) has claimed that the pinniped pel- 
vis is degenerate but this statement indicates an incorrect viewpoint. It 
is permissible to view the ilial part as having sufl^ered degeneracy in 
length, and the sea-lion innominate is definitely less robust in some par- 
ticulars, but the posterior part has experienced relative hypertrophy and 
the whole is little or no shorter in relation to body length than in the cat 
(21 per cent in Zalophus, 26 in. Phoca, and 25 in a cat). Neither is 
the sacropelvic connection weaker than in many terrestrial mammals. 

Accompanying and intimately involved with the shortening of the 
pinniped ilium is shortening of the femur. In Zalophus this segment is 
now 11 per cent of the body length, 12 in Phoca and 35 per cent in a 

en ^ £ < U=^^ cc 
S ^. ^ 1 I 



cat.* There is a complex of interacting conditions involved in this situa- 
tion, some constituting cause and some effect. As the heel is bound down 
to the body a long femur is not desirable unless really needed for other 
purposes, and this is not the case, for the binding down of the leg pre- 
vents great movement of the thigh from serving any useful purpose. 
Furthermore, a short femur can be held abducted without projecting 
from the side to an undesirable extent. Especially is this true of flexion, 
and atrophy of the flexor muscles accompanies the shortening of the 
ilium and the femur. With the flexors serving little useful purpose there 
can be but slight need for muscles accomplishing only extension. Ac- 
cordingly those extensors which usually insert near the knee have shifted 
their insertions to other situations and because of the increase in post- 
acetabular length of the innominate, the origins of the adductores, and 
of the obturator externus, have shifted so as more efl?ectively to accom- 
plish what extension of the femur is needful, in addition to the regular 
adduction and rotation. Because the actions involved in swimming by 
the seal comprise relatively simplified but constantly repeated motions 
there is apparently need only for simplified adductors, and these are re- 
duced to two, whose position in respect to the external obturator is so 
intimate that the three have previously been mistaken for one single 
obturator. Probably for the reason that the sea-lion has need for very 
intricate, though slight, movements of adjustment for steering, the ab- 
ductores now occur in six slips, and the sartorius is double. As a re- 
sult of the above conditions the arc of movement of the pinniped femur 
in those specimens dissected seemed to be only through about 25 or 30 

In this order the femur is relatively very broad transversely and nar- 
row in the sagittal plane. The breadth may be attributable to the fact 
that good leverage is needed through a greater trochanter which ex- 
tends well away from the head, and a broad distal end for articulation 
with the shank. Its thinness is permitted on the one hand by the fact 
that little mechanical strength in flexion and extension is needed, and 
promoted because a flat femur can accomplish a sharper angle of flexion 
with the shank. The latter point appears to be of lesser consideration 
in the seals. 

*I regret to say that in the last paragraph of page 124, Howell, 1929, an 
error was made in stating the percentage of femoral length in relation to body 
length, and the figures 22 and 29 as given are really those for the tibia. They 
should read 11 and 12 respectively, as was stated on page 36. 



It will be noted (fig. 52) that the sea-lion has a lesser trochanter de- 
fining insertion of the psoas magnus and iliacus (so-called), while in 
the seal these muscles have fascial insertion over the knee and there is 
no lesser trochanter. In this illustration it may also be noted that 
while in the sea-lion the inner condyle of the femur extends but a 
trifle more distad than the outer, in the seal this discrepancy is very 
marked indeed. The result of this is that when the femur is held at a 
right angle to the body axis the position of the condyles will cause 
greater supination of the pes, so that the soles may be maintained paral- 
lel with greater ease. If the femur be extended beyond a right angle, 
then the condylar conformation will cause the feet to be held more ele- 

A peculiarity which affects function in the pinniped knee is the length 
of the tibial collateral ligament. The distal attachment of this is not 
at approximately one-tenth the distance from the knee to the ankle, 
as in the cat, but at a point about one-quarter or two-fifths this distance. 
This allows for considerable rotation through the knee, with the fibular 
ligament as the pivot. In most other mammals such rotaton of the 
shank can be accomplished through the hip joint, but because of the po- 
sition of the short femur in the Pinnipedia such action is impossible in 
this order without adductive-abductive movements of the thigh. 

A further accompaniment of the above conditions of the femur and its 
more intimate musculature is that the function of those muscles which 
normally act as extensors of the thigh through an attachment in the vi- 
cinity of the knee joint are enabled to alter their function. Their inser- 
tions have either migrated wholly or in part distad, and accordingly 
they now operate to bind down the shank and accomplish its adduction 
or elevation, according to posture. There is considerable variation in the 
form of these muscles as between the seal and the sea-lion (figs. 22 
and 23), but in both animals the gracilis, semitendinosus and semimem- 
branosus act especially in this capacity and their insertions not only do 
not approach the knee but are extended onto the heel by fascia in the 
case of the sea-lion, and in the seal by a more tendinous structure pass- 
ing to the plantar fascia. In the former animal the biceps acts with par- 
ticular efficacy as a binder, but not to the same extent in the seal. The 
conformation of these muscles has naturally affected the position of the 
crotch, or posterior integumental border between the leg and body, or in 
this case the tail, so that in both animals this lies approximately at the 
heel and all of the hind limb except the pes is contained within the body 



The pinniped shank is considerably shortened (length of tibia being 
22 per cent of the body length in Zalophus, 29 in Phoca, and 36 per cent 
in a cat) , and because it is so bound down by muscles and other tissue 
its possible movement, judging by the specimens dissected, seems to be 
through an arc of only about 15 degrees. In this order the fibular head 
really is not concerned with the knee joint and its conformation has been 
affected by the posture habitually assumed. In the Phoca studied the 
angle which the femur formed with the shank was a trifle more than 90 
degrees, while in the Zalophus it was considerably less — apparently about 
45. Accordingly in the former animal the fibular head slopes moder- 
ately, and in the latter quite steeply, enabling consequent sharper bend- 
ing of the knee without mechanical interference. But there is some 
slight generic and even specific variation in this detail. 

Upon examining the distal ends of the shank bones (in fig. 13 of 
Howell, 1929) it will be found that there are no deep grooves in Zalo- 
phus, but several in Phoca. Upon the lateral aspect of the latter there 
is a deep one for the tibialis anticus and another for the peroneus 
longus ; upon the medial aspect there are three, for the posterior tibialis, 
flexor digitorum longus, and peroneus digiti quinti. As in the case of 
the radius these grooves have doubtless resulted from the constantly re- 
peated oscillations of the feet, made during swimming. 

The significance of the absence of the soleus in Phoca and of the lat- 
eral gastrocnemius in Zalophus is unknown. In both, the plantaris does 
not join the tendo calcaneus but passes mediad thereto. In the Phoca, 
but not the Zalophus, the tendon of the strong flexor hallucis longus 
passes over a posterior extension of the astragalus, as already described, 
and it is the tension of this tendon alone which prevents the foot from 
assuming an angle of 90 degrees with the shank. But this does not 
necessarily mean that if it were not for the tension of this muscle the seal 
could walk like a sea-lion. To do this it would have to bend the verte- 
bral column so that the sacrum is vertical to the ground, and it might 
find such a posture inconvenient, if not actually impossible of assump- 

In bony details the foot of the sea-lion is of especial interest in the 
present connection in having the first and fifth toes approximately equal 
in length and longer than the middle three, as in the seal, and the bony 
elements are definitely flattened, especially the hallux. The question of 
epiphyses has already been discussed in chapter ten, for the situation in 
this respect in the manus applies equally, with minor variations, to the 

[296] ■ 


In two specimens of sea-lion examined, one fresh and one preserved, 
the relaxed posture of the foot was latero-craniad at an angle of about 
45 degrees with the body axis. In life, however, an angle of approxi- 
mately 90 degrees may frequently be assumed, and in young fur seals 
observed the latter was the invariable posture. This facilitates the plac- 
ing of the feet sole to sole when in the water. In the Otariidae the 


Figure 50. Pelvic bones of sirenians. Above is illustrated progressive develop- 
ment according to the fossil evidence: (a) Eotherium (Middle Eocene) ; 
{b) Eosiren (Upper Eocene); (f) Halitheriutn (Middle Oligocene) ; (d) 
Metcixytherium (Middle Miocene); {e) Halicore dugong (recent); (/) 
Halicore tabernaculi (recent); and {g) Trichechus latirostris (the manati) 
(redrawn from Abel). 

interdigital membranes are capable of practically no expansion, and the 
foot is correspondingly narrow as compared with that of a seal. It is 
relatively very long, however, because of the predigital cartilages, as in 
the manus. But in the foot, these can be folded back against the palm 
so that the body may be scratched with the slender nails of the three 
middle digits. The nails of the first and fifth digits, are, however, ru- 
dimentary like those of the manus. 

[ 297 ] 


In the sea-lion manus the terminal border of the foot is practically 
straight, but in the pes it is deeply serrated, because the predigital car- 
tilages extend considerably beyond the interdigital membrane borders. 
The length of foot including cartilages, in the fur seal (Callorhinus) 
appears to be relatively greater than in the sea-lion (Zalophus). Also, 
judging by but two specimens measured, this length is proportionately 
greater in young of the latter genus than in an adult, which may indi- 
cate that a longer rudder was needed in the past, with somewhat less 
perfected swimming apparatus, than is now requisite. 

The development of the pes in the Otariidae is somewhat puzzling in 
certain respects, chiefly because of its length in respect to breadth, the 
former augmented by the peculiar predigital cartilages. It may really 
be said to be more highly specialized than either the manus or the pes 
of the Phocidae. At one time I was inclined to entertain the possibility 
that at a not very distant geologic date it might even have been of some 
definite, active aid in locomotion, but this I have abandoned. Evidently 
the stimulus for a long neck that is highly mobile is a strong one, and 
this mobility necessitates a highly specialized and highly efficient rudder 
fully capable of compensating for abrupt movements of the head, say 
in snatching at prey, that would otherwise tend to deflect the animal 
from its desired course. Such a rudder should project well to the rear, 
and does, but its "hinge" should be at the body junction, and is. Hence 
the leg has been shortened and is contained within the body covering 
while the foot has been lengthened by predigital cartilages, either be- 
cause these fulfill some need which longer digital bones could not ac- 
complish or because they were more easily developed than longer 

3. Under the third heading mentioned on page 000 will be dis- 
cussed those aquatic mammals in which the hind feet are absent or de- 
finitely subordinate in swimming or steering. Under this will be in- 
cluded the mink for the sole reason that there is nowhere else to place 
it without making a separate heading for its reception. It swims chiefly 
by all four feet, as far as I can learn, but it should not be placed with 
most other mammals employing this method of propulsion, for its ten- 
dencies are different. It is clearly what may be called an incipient otter 
which may be expected to develop its swimming abilities along pre- 
cisely the same lines as has the latter, and accordingly it will be given no 
further attention. 

For discussion under this heading there are the aquatic Tenrecidae, 
Llmnogale, and Potomogale, the Lutrinae or river otters, the three re- 
cent families of the Sirenia, and the Cetacea. 


I am unable to discuss Limnogale with any profit for I have never 
seen a specimen and have been able to find nothing of significance 
in the literature save that the toes are webbed and the distal part only 
of its powerful tail is laterally compressed. Potomogale, the insecti- 
vore otter (fig. 4) is peculiar in that there is no present evidence that 
the hind feet were ever used for efficient aquatic propulsion, for the legs 
are short and the toes unwebbed. In the case of other aquatic insecti- 
vores, as well as rodents, the feet are used for swimming pending the 
modification of an adequate tail, which is theoretically the more efficient 
propulsive organ, and during this time the former are confidently ex- 
pected to become quite notably altered. Hence it is requisite to assume 
that even from the first this insectivore had a phenomenally powerful 
tail, robust at base, and consequently somewhat different from this mem- 
ber in most representatives of- the order; and that these characters, 
coupled with an elongated body, enabled it to swim quite well by lateral 
oscillations with little or no help from the pedes. 

The only alteration in the pes of this animal is that there is a thin, 
cutaneous extension of the lateral border. This is believed to be for 
the purpose of enabling the feet to present fewer inequalities when they 
are folded back against the base of the tail during swimming. As a 
matter of fact it is doubtful if anyone knows just how the feet are then 
held and these lateral extensions of the foot may have some entirely dif- 
ferent function. Dobson (1882) has shown that in this genus there is 
a very remarkable specialization of the hip musculature. Evidently the 
femur is fixed by its flexor muscles, and with this bone thus acting as 
an origin the caudofemoralis or gluteus maximus extends far back- 
ward onto the tail (fig. 47), thus helping in the lateral oscillations of 
this member. There has been no definite reduction in the length of 
the femur as in the desman, and Dobson stated that although in gen- 
eral form the pelvis resembles that of Centetes it is narrower and with- 
out a true symphysis, interpubic connection being ligamentous. 

The river otters are of particular interest in the present connection 
because I regard it as most likely that the terrestrial ancestors of the 
Cetacea were beasts of very similar body form, namely, with long, 
sinuous trunk, short legs, and powerful tail, especially at base. The ot- 
ter is such a playful creature that it is often difficult during observa- 
tion to be sure which actions are concomitant to greatest efficiency in 
swimming, and which are attributable to exuberance of spirits. In the 
common genus Liitra the hind limbs are frequently kicked in a sort of 
galloping action, seldom entirely rhythmically but at odd intervals. I 



have gained the impression that this use is not invariably so much for 
the purpose of speeding progression as it is for maintaining proper 
posture in the water. Frequently when the greatest speed is attained the 
limbs are not used at all. It is highly probable that at one time in the 
past the tail of the otter was a less efficient swimming organ than is now 
the case, and undoubtedly the limbs then constituted the chief propul- 
sive apparatus, as now in the mink. But in all generalized mustelids 
the body is long and sinuous and the legs short. Hence, body motions, 
such as arching of the back, are important in terrestrial locomotion and 
it is natural that a similar use was made of the body and tail when the 
animal first took to the water. Because of this and the shortness of the 
limbs it is not improbable that the latter did not respond so readily to 
the stimulus for aquatic modification of the extremities, or else that 
this stimulus was not as strong as in the case of most other mammals, 
so that the tail more readily gained the ascendancy. 

The hind feet of otters (fig. 43) vary greatly from an unwebbed con- 
dition to one in which the interdigital membranes are broad and full. 
But as nothing is known about the finer points of the swimming meth- 
ods of these less familiar sorts the subject cannot be pursued further. 

One would hardly expect much change in the otter as long as it fol- 
lows its present mode of life. If it should increase in size and take to 
coastal waters then its tail should broaden into a more efficient propeller 
and its hind limbs tend to become reduced as it may become more inde- 
pendent of the land. Eventually its pectoral limbs should assume the 
form of equilibrating paddles and it is likely that superficially at least it 
should finally become very much the same sort of beast as is the ceta- 

The subject of the sirenian pelvic element is one which I approach 
with considerable hesitancy. Flower (1876) has said, the pelvis is ru- 
dimentary, composed in the dugong of two slender bones on each side, 
placed end to end, commonly ankylosed together. He considered that 
the upper, attached to the vertebral column by a ligament, represents 
the ilium and the lower the ischium, or ischium and pubis combined. 
He also stated that although there was a vestigeal femur in Halitheriuni 
this bone is not represented in any living member of the order. But 
other observers, possibly working with different species, have reported 
otherwise. The innominate of the dugong is usually figured as a single, 
slender bone, made up of two terminal elements, situated with the long 
axis vertical, connected by ligaments dorsad with the transverse process 
of one vertebra, and ventrad to its fellow of the opposite side. Abel 



(1906) has illustrated some of the steps in its evolution by the fossil 
evidence (fig. 50) whereby it is indicated that the ilium has become more 
slender but has not been shortened, the pubis has suffered suppression, 
and the ischium slight elongation and alteration in shape. 

Figure 51. Pelvic bones of Cetacea: {a) the zeuglodont Basilosaurus ; (b) Sib- 
baldus ; {c and d) Balaenoptera physalus; (e) Megaptera; (/) Physeter; 
(g) Mesoplodon; and {b) Lagenorhynchus: {b, d, and e are redrawn from 
Augustin, and c, g, f, and h from Abel). These are supposed to be in right 
latero-ventral aspect. 

Abel's figure of this region in a manati from Surinam (fig. 50) indi- 
cates that in this family there has been a fundamentally different proce- 
dure. He claimed that a vestigeal femur is present, attached near the 
anterior end of the pelvis ; hence that the ilium is greatly reduced. The 



pubis has probably been eliminated and the ischium shortened and con- 
torted. This situation is broadly similar to what is indicated in a zeu- 
glodont, as discussed shortly, but very different indeed from what has 
taken place in the dugong. Clearly the reduction in the hind limb of 
these two families of sirenians has been upon entirely different prin- 
ciples, likely involving fundamental difference in the uses to which the 
hind limbs were previously put, and it seems unsafe to hazard any opin- 
ion in regard to just what particulars of the disappearing hind limbs 
may have been involved. 

In other respects the modications of the Sirenia are much more 
puzzling than in the Cetacea and it is correspondingly difficult to recon- 
struct the probable steps that were followed. In the case of the whales 
the salient characteristics are such as to point clearly, to me at least, to 
numerous hypotheses that have a high degree of probability. This is 
not the case m the Sirenia. It is difficult for me to envision any sort of 
primitive mammal of proboscidean affinity taking to the water in the 
sluggish manner which one would naturally connect with the sirenian 
prototype and developing the tail which we now see in connection with 
the total disappearance of the hind limbs. Evidently the latter situa- 
tion followed abandonment of any use of these members, but more I do 
not feel like saying at the present time, for further statement would be 
pure, unsupported speculation. 

Indubitably many groups of the Cetacea have been distinct from one 
another for a great length of geologic time, presumably since the period 
when they had external hind limbs which were quite well formed. As 
the latter tended toward more and more complete disappearance there 
naturally occurred differences in precise details. Hence it is entirely ac- 
cording to expectation that in examining the anatomy of this order one 
finds hardly any two distinct sorts with the details of the pelvic muscu- 
lature entirely the same. In fact the differences are so great that it is 
often difficult to correlate the pelvic muscles of two sorts of whales, and 
utterly impossible yet to homologize them with muscles of terrestrial 
mammals for the reason that the nerves are so greatly altered, and in the 
present state of our knowledge it is unwise even to attempt to do so. 

The cetacean innominate is now so different from the generalized con- 
dition that very little can be proven about the homology of its parts, nor 
does the fossil material, except in the Archaeoceti, help us to any ex- 
tent. Usually one area of the bone is thicker and broader than the rest, 
and this, presumably, marks the previous situation of the hip joint, but 
there may be exceptions. In some cases existing projections from this 



area may be processes recently and secondarily developed for some 
specialized purpose, and in others an anterior and a posterior process may 
actually be remnants of both ilium and ischium, as suggested by the 
skeleton of a young sperm whale in the National Museum. In this there 
are two small pelvic bones on each side, one before the other, which 
one would expect to have fused at a later age. Again, in those whales 
in which the pelvis is broader at the anterior end there has probably 
been virtual elimination of the ilium as there seems to have been in 

In shape the cetacean pelvis varies greatly. In what seems to be its 
simplest condition it takes the form of a short, practically straight, 
slightly flattened bone (in many porpoises) . Or it may be bent some- 
what after the fashion of a boomerang, but with one end broader than 
the other (Megaptera), or quite broadly triangular (Sibbaldus), or nar- 
row at one end and abruptly expanded at the other (some individuals 
of Balaenoptera, Monodon, etc.) . In a large rorqual it will not exceed a 
length of 18 inches, and in a small porpoise may measure about one 
inch. At least those which are more elongate occur practically parallel 
with the body axis and almost directly above the genito-excretory ori- 
fice. It has been forced ventrad of its usual situation, evidently by 
the hypertrophy of the hypaxial musculature and has no direct (bony) 
connection with the vertebrae. In its simplest form it is doubt- 
ful whether the cetacean pelvis should really be looked upon as 
constituting a disappearing remnant of the hind limb, but rather that 
the hind limb has already disappeared and what remains is merely a 
necessary anchor for the perineal musculature, thus, possibly as greatly 
reduced as it will ever become. As these muscles are the more power- 
ful in the male the pelvic bones average larger in this sex than in the 

There has been considerable controversy over the question of whether 
the cetacean pelvis represents all, or but one, of the elements of the nor- 
mal innominate. Flower (1876) stated that this bone ossifies from one 
center and probably represents the ischium. Struthers was of substan- 
tially the same mind, while Delage (1885) argued that all three bones 
are represented. Whether the bone actually develops from three centers 
of ossification seems to me a purely academic question of not much im- 
port in the present connection. The chief thing is that there has been 
marked reduction, almost if not quite to the point of elimination, of 
certain elements of the normal pelvis, and relative increase in the im- 
portance of at least one other. The important point at issue is whether 
the present conditions in this regard can be interpreted. 
[ 303 ] 

Figure 52. Left femur of (Z) sea lion {Zalophus) and (P) seal (Phoca), show- 
ing areas of muscle attachment: anterior aspect above and posterior below. 


I think the evidence is clear that there has been reason for retention 
of a part of the ischium, for perineal attachments, and it may reasonably 
be conceded that a considerable part of the present pelvis — say at least 
half — is of such derivation. // another pelvic element has been retained 
in appreciable amount it probably consists of the ilium, for the muscles 
usually attached to the pubis have disappeared or relinquished pelvic 

This brings us back to conditions in the Sirenia. It is incontrovertible, 
I think, that the plan of pelvic reduction in the Trichechidae has been 
on a basically different principle from that in the Halicoridae. If this 
can be so in the Sirenia it can apply with equal logic to the Cetacea, and 
accordingly it must be postulated that in the whales the pelvis may have 
suffered reduction on at least two, and possibly more, basically different 
schemes, each perhaps varying in minor details according to the group 
concerned. Of those conditions involved in a possible retention in 
definite degree of the ilium nothing can be told, but considerable can be 
inferred about that plan by which the ilium may have been eliminated, 
just as in the manati, by analogy to the pinnipeds and zeuglodonts, as 

The pelvis of the zeuglodont Basilosaurus (fig. 51) was considerably 
less advanced in specialization toward simplification than in any living 
whale. This is a roughly quadrangular bone with an indubitable acetabu- 
lar depression upon its extreme anterior part, for attachment of the ru- 
dimentary femur, and posterior to this, a foramen which is obviously a 
relic of an obturator foramen. One can hardly escape the conclusion 
that this situation represents an extreme atrophy of the ilium and the 
entire disappearance of all but possibly that part forming the anterior 
margin of the acetabulum. Evidently the zeuglodont ancestor had be- 
gun to trail the hind limbs and there ensued a reduction in the length of 
the ilium, just as is now to be seen in the pinnipeds, and further atrophy 
of the hip muscles arising from this bone and relinquishment of its 
iliocostalis attachment would naturally be followed, in the course of 
time, by the entire suppression of the pre-acetabular part of the pelvis. 
Accompanying the reduction of the muscles attached to the ilium would 
be a similar atrophy of those of the hip and thigh that were attached to 
parts of the ischium and pubis. There would naturally follow atrophy 
of these parts of the innominate bone, but there would not be the same 
relative degree of atrophy of the ischial portion which afforded attach- 
ment to the perineal muscles, and these would increase in proportionate 
importance. As a matter of course the form of the complex would 



change in an unpredictable manner. And, this, apparently is just what 
has transpired in the case of this zeuglodont. 

A change which was fundamentally of this sort has doubtless oc- 
curred in some of the Cetacea, and in these there must be virtual if not 
complete elimination of the original ilium, as well as the pubis, but in 
other sorts there may be an entirely different situation, which the existing 
pelvis is too greatly altered to indicate with any clarity. 

According to Kellogg (1928) no zeuglodont is yet known having 
more of a hind limb than a rudimentary femur. In no member of the 
Odontoceti except the cachalot (Physeter) has any sort of a femoral ru- 
diment been found, but vestiges, in varying degree, of the pelvic limb 
have been reported from most if not all species of Mysticeti examined, 
although the frequency which which it has been stated as entirely lack- 
ing in certain specimens leads one to suspect that this may be an indi- 
vidually variable character in some sorts. At least in Balaenoptera phy- 
salus it apparently is never more than very rudimentary indeed, and Abel 
(1908) illustrated this detail in a number of individuals, mostly col- 
lected from the literature. Struthers (1887-88) reported that the hump- 
back (Megaptera) has a cartilaginous femur inclosed in fibrous tissue, 
and indicated that he considered that the pubis constituted the greater 
part of the pelvis. Eschricht also reported a femur in this genus. 
Flower found that in the blue whale (Sibbaldus) it occurred as a bony 
nodule, which Beddard stated was attached to the pelvis by one anterior 
and two posterior ligaments in which there were a few muscle fibers. It 
is in the balaenid whales, however, that the cetacean limb is least re- 
duced, and several authors have reported both a bony femur and a car- 
tilaginous tibial head present in the Greenland or bowhead whale 
(Balaena mystic etus) . The former was stated by Beddard to be from 4 to 
9 inches in length. According to Struthers (1880-81) the femur is 
flexed forward and the tibial rudiment is horizontal, while there is a 
synovial bursa between the femoral head and the pelvis, and another 
between the femur and tibia. He (1893) has also stated that there 
are three muscle slips from the femur to the pelvis. 

That the loss of the external hind limbs in the Cetacea occurred at a 
rather remote time is indicated by embryologic conditions. Guldberg 
and Nansen (1894) found that in a 17 mm. fetus of the porpoise 
Phocaena the pelvic buds were one-third mm. in length, while in one 
of 7 mm. these were three-quarters mm. long (the fore limbs being one 
and one-half mm.) and shaped like an oval leaf. Histologically they 
consisted of undifferentiated mesoderm without sharp separation from 



the epidermis, and were entirely comparable, except in size, with the 
hind hmb buds of other mammals. Apparently their lesser definition 
in the 17 mm. Phocaena, and still less in a 27 mm. Lagenorhynchus, in- 
dicates that a reduction in absolute size begins as the external cetacean 
characters (chiefly the relative increase in tail size) are initiated. In 
another Phocaena fetus, of 18 mm., the limb was represented merely by 
a papilla, which is the penultimate step in external disappearance. 

Before summarizing the discussion of the cetacean pelvic limb it may 
be well to consider briefly the case of some of the extinct aquatic rep- 
tiles. In almost all the larger sorts known the pes was practically as 
large as the manus, even though adaptation had progressed to the point 
where hyperphalangy was already far advanced. It has previously been 
argued herein that this was probably attributable largely to a difference in 
the methods of swimming employed by these reptiles compared with 
the case in mammals, partly because of difference in bodily conforma- 
tion as well as equipment and inherent tendencies. It was only when 
aquatic adaptation had become very far advanced indeed, to the point 
comparable with that now attained by the Cetacea, that the hind limbs 
had begun really to shrink in size, as illustrated by Ichthyosaurus. In 
bodily form these reptiles seem to have been as highly modified for an 
active aquatic life as cetaceans and it is not likely that the former 
had need for retaining the pedes as balancers. Hence it ap- 
pears necessary to believe either that the external hind limbs of aquatic 
reptiles were much slower to disappear, because employed for a much 
longer time for active swimming in a four-limbed manner, or else that 
the Cetacea were phenomenally amenable to influences which resulted 
in elimination of the pedes, for which there is not the slightest evi- 

To complete this chapter it is only necessary to offer a brief summary. 
If a mammal be well modified for an aquatic life but without present 
evidence that its hind limbs have ever been used to an important ex- 
tent for either swimming or steering, these may remain relatively un- 
altered in most respects, or become peculiarly specialized in an unpre- 
dictable manner (as in the gluteus maximus of Potomogale). If these 
members are used as the chief method of swimming or steering and 
are highly altered accordingly, one may expect with a considerable de- 
gree of confidence that there will be a shortening of the femur, fre- 
quently accompanied by shortening of the ilium. If the latter does not 
occur the lesser gluteal mass should be very robust, and if it does, the 
gluteus maximus may be very powerful. The crotch will tend to mi- 


grate toward the rear and there will very often be a broadening of the 
superficial division of the biceps femoris, usually to cover most of the 
shank. It may, or more often will not, be correspondingly broad at 
origin. There will frequently occur a great increase in the power of 
the semitendinosus, possibly accompanied by a similar development of 
the gracilis, and semimembranosus, the insertions of all three extend- 
ing farther distad toward the heel. 

If the animal be large and the feet need not be used for extensive ter- 
restrial progression the pes will usually be webbed, and symmetrical 
providing it be used in a symmetrical fashion, in which case it may be 




5E5An01D — 
TAR5 A\_E5— 

Figure 53. Dorsal view of left tarsus and metatarsus of (Z) sea-lion {Zalophus) 
and (P) seal {Phoca). 

expected that the first and fifth toes be of equal length and longer than 
the middle three. If the animal be of small size the pes may be either 
webbed or fringed, is less apt to be symmetrical, and the tendency will 
most frequently be for the gradual enlargement of the member. No 
matter in which active manner the Cetacea once employed the hind 
limbs for swimming the present result could have been attained pro- 
viding they were equipped with a relatively long and rather heavy tail 
and that swimming methods entailed curvature of the backbone (and 



tail) in the vertical plane. It is not probable that an actively predaceous 
mammal of the sort that it is likely the whale has always been would 
evolve a tail expanded in the horizontal plane for the purpose of keep- 
ing it near the bottom, as seems to be the case in the platypus, but 
rather that it developed along the same lines that the tail of the otter 
is now following. The speed of development of such a tail should be 
materially assisted by the presence of short, and relatively weak, rather 
than long and powerful, pedes, and such small feet would, presumably, 
disappear more rapidly than would those which had become large and 
highly specialized. Hence there seems to be considerable evidence that 
the Cetacea could most readily and quickly have evolved from some sort 
of terrestrial ancestor having many of the chief bodily characteristics now 
occurring in the common river otter. 


Chapter Twelve 

Other Soft Tarts and Physiological features 

1 HE WRITER has not concerned himself, save very incidentally, with 
any part of the physiology of aquatic mammals other than of the mus- 
cular system, nor with the internal organs. Nevertheless, the present 
contribution would be very incomplete without including some consid- 
eration of these features, because some of the most interesting problems 
concerning the Cetacea are involved with them. Accordingly it is aimed 
to offer brief mention of what seem to be the most important of these 
questions, and to point out some of the possibilities, without any at- 
tempt at an exhaustive consideration. These points have frequently been 
discussed with Doctors E. K. Marshall and G. B. Wislocki of the Johns 
Hopkins Medical School, who are working on some of them, and to 
whom I am accordingly obligated. 

There is a veritable host of interesting problems concerned with the 
physiology of the more highly modified aquatic mammals, especially 
the Cetacea, which are particularly difficult to solve largely for the rea- 
son that it is not easy to experiment on the live animal, and because 
they involve alterations in quality from those encountered in ourselves, 
so that a very real obstacle is our inability to compare conditions with 
what we know obtains in the case of terrestrial mammals. Perhaps the 
most important, or at least spectacular, of these questions is concerned 
with the ability of some whales to descend to depths in excess of a mile, 
and to remain thereabouts for more than an hour. How do they with- 
stand the pressure? How do they take down sufficient oxygen.^ More 
important still, how do they get rid of the carbon dioxide in the blood } 
If we hold our breath for a minute we are in distress, not because we 
need more oxygen at once, but because there is imperative need to dis- 
pose of the accumulation of carbon dioxide. 

Another point, of particular interest to students of the human brain, 
is that although the porpoise apparently has less need for intelligence 
than almost any other living mammal, and little if any more than had 
the ichthyosaurs, the convolutions of its brain are more marked than in 
man, thus indicating the probability that this character is not as signifi- 
cant of intelligence as many now believe. Other details of the cetacean 



brain are its large cerebellum, and the fact that it is larger in transverse 
than in sagittal measurement, the latter feature undoubtedly being a re- 
flection of the peculiar skull development. 


It has already been remarked that aquatic mammals are particularly 
prone to develop fatty tissue of one sort or another. An extensive blub- 
ber layer irrefutably fulfills a physiological need in aiding the retention 
of body heat by those sorts which inhabit cold waters. This does not 
apply with equal force to tropical and subtropical sirenians and cetaceans. 
Apparently they could make other provision for withstanding the very 
moderate temperatures to which they are subjected, and yet in spite of 
their abundant fat the former are particularly susceptible to any slightly 
unusual chilling of their aquatic habitat. It thus seems not unlikely that 
the blubber layer may have other uses than purely that of insulation. 

Seals, at least, often have a great deal of intestinal fat, and the Odonto- 
ceti usually have accessory fatty equipment in the form of a frontal 
adipose cushion of a specialized quality of fat, and other deposits about 
the angle of the jaw. Of greater significance still is the enormous and 
unwieldly spermaceti organ of the cachalot, and what I have previously 
referred to as a sort of circulatory system for oil. This, apparently, is 
present to a very slight degree in the porpoise Neomeris, but has not 
been reported in other cetaceans. In a fetal narwhal (Monodon), how- 
ever, which I dissected, the subcutaneous tissue, and much of the inter- 
muscular tissue, was given the appearance of a rubber sponge by an 
amazingly intricate and extensive system of oil ducts, intercommunicat- 
ing with each other and with larger subcutaneous oil reservoirs. Even 
some of the larger muscles had small oil sinuses throughout their sub- 
stance. About the neck at least these oil ducts were very numerous, of 
very small diameter, and apparently had intimate relation with the retia 
of the vascular system. How extensive the flow of oil through these 
ducts from one part of the body to another may be in the live animal 
is of course unknown, but while dissecting the back, if one pressed upon 
the dorsal region oil welled up and out in copious amount. In life mus- 
cular pressure could attain the same result, and there would accordingly 
be an irregular, and possibly somewhat limited, flow of oil from one 
area to an adjoining one. 

This specialization of the narwhal cannot be fortuitous any more than 
can the spermaceti organ of the cachalot. Any such unique and strongly 
marked character must have been developed to fulfill some definite phys- 



iological need. What this function can be is entirely unknown, but it 
it only logical to presume that as specialized fat organs such as are being 
discussed occur only in the Cetacea they have something to do with the 
ability of this order either to hold the breath for a phenomenal length of 
time, or to assist in withstanding great pressures — an explanation which 
seems unlikely. 

In what way could this oil system of the narwhal assist the animal to 
hold its breath.? The only possibility that occurs to me is that there 
might be sufficiently close connection between capillaries of the oil sys- 
tem and vascular capillaries for there to be possible interchange of car- 
bon dioxide by diffusion. The process then would be for the oil to take 
over a part of the carbon dioxide as this accumulated in the blood dur- 
mg prolonged submergence. When at the surface once more some eight 
or ten breaths might then enable the oil to give up its excess of carbon 
dioxide by the same process reversed. And it is well known that after a 
lengthy submergence a whale must spend an approximately equal period 
of time at the surface, frequently inflating the lungs, before it can again 
repeat its dive. The above is a possibility, I say, but it should be 
emphasized that it is not probable and it is offered merely as the only 
explanation that presents itself. 

If this detail of the narwhal has been developed to facilitate long sub- 
mergence one would naturally suppose that this animal can remain be- 
low the surface for a longer period than can the majority of porpoises, 
but it is not known whether this is actually the case. It appears only 
reasonable, however, that cetaceans of the order of the narwhal and be- 
luga which occur where they must often be hemmed in by extensive ice 
fields should have unusual ability for lengthy submergence, and it will 
be of interest to ascertain whether the latter has also developed this oil 

The sperm whale or cachalot is believed to have the ability to sub- 
merge for a longer period than any other cetacean, and it is the only one 
which has developed a spermaceti organ of phenomenal size. I am 
unable to believe that these two circumstances have no connection, but 
there is no clue to the manner in which they may be dependent. The 
fact that the fatty equipment of the narwhal and cachalot are so very 
diflferent may be without much significance, for they were undoubtedly in- 
dependently developed and could fulfill the same need. The difficulty 
lies in the statement that the spermaceti organ is a "closed system," and 
no mention is made of any marked vascular networks. 

Through the efforts of H. C. Raven and H. I. Wordell I secured two 



samples of sperm oil and E. K. Marshall kindly determined for me their 
power for absorption of carbon dioxide. One hundred cubic centime- 
ters of each dissolved respectively 92.8 and 94.6 cc. of carbon dioxide. 
This was left standing for 72 hours and the samples then refused to ab- 
sorb more carbon dioxide, indicating that none of the original gas taken 
up had suffered change. One hundred cc. of water, alcohol, and petro- 
leum will absorb respectively 54.6, 230, and 82 cc. of carbon dioxide; 
so it is seen that these experiments do not show that sperm oil is re- 
markably solvent of this gas. It should be pointed out, however, that 
the question is by no means settled. The samples used were of old 
sperm oil which had been altered by the action of the free fatty acids 
which it contains in considerable amount. The experiment should be 
repeated with fresh material, and under pressures increasing to one ton 
to the square inch. 

In abandoning this topic it should again be emphasized that we know 
nothing whatever about any physiological process, other than for the 
conservation of heat, which the fatty tissues of aquatic mammals may 
have. All that is now possible is to advance tentative and unsupported 
hypothesis as a basis from which to work. 


The digestive system of pinnipeds and cetaceans at least is called upon 
to furnish an unusual quantity of raw material for the production of 
both blood and fat; but it is difficult to know which features constitute 
phylogenetic inheritances from terrestrial ancestors and which may have 
been developed by an aquatic life. The stomach of the Cetacea, both 
Odontoceti and Mysticeti, is very complicated and with numerous di- 
visions. It is usually stated that most forms have four of these with a 
fifth dilation of the duodenum, but ziphioid whales may have "9, 10, 
even 13 or 14" divisions, according to Beddard, quoting other authors, 
and in the latter group the structure suggests that the usual first division 
may be missing. In this character of complexity the cetacean stomach is 
comparable to that of ruminants, but there are fundamental differences 
and their resemblances are probably attributable to convergence rather 
than to relationship. Apparently conditions may be interpreted as fol- 

The typical porpoise stomach consists of a relatively simple division 
which appears to be a dilation of the esophagus. It communicates by a 
passage in its upper wall with the second division, whose walls are 
heavily plicated and of great thickness. It is here that the gastric juices 



are secreted, and apparently these are regurgitated into the first division. 
Indigestible matter, like cuttlefish beaks, is probably vomited forth, from 
time to time, but the gastric juices are doubtless capable of dissolving 
fish bones. At any rate the stomach contents must be thoroughly liquefied 
before leaving the second division for its exit is minute and of little 
greater diameter than that of a stout probe. The third and subsequent 
division appears to be nothing but a specialization of the pylorus, con- 
torted and with several constrictions. 

The cetacean intestinal tract is not usually remarkable and is of mod- 
erate length. Owen (1868) quoted Hunter to the efi^ect that in a 
Balaenoptera acutirostrata 17 feet long the total intestinal length was 
93-3/4 feet. But on the other hand the same author stated that in 
Hyperoodon the whole intestinal canal was sacculated in a surprizing 

The stomach of the dugong is remarkable for being very thick and 
muscular, and for having two caecal appendages or diverticula. Owen 
gave the intestinal length of a half-grown individual as 27 feet. The 
caecum of this animal is single, but it is bifid in the manati. 

The stomach of the Pinnipedia is not particularly noteworthy but the 
intestinal tract is of phenomenal length. Engle (1926) has recorded a 
sea-lion (Eumetopias) with intestinal length of 264 feet, which is over 
twice as long as has been reported for a fur seal (107 feet). Stones 
of considerable size and of an aggregate weight of several pounds have 
been reported as found in the stomach of sea-lions. It has been sug- 
gested that this is for the purpose of weighting down the animal, but I 
would regard this added weight as insignificant to a body weighing sev- 
eral hundred pounds. More attention should be paid to the time of 
year at which stones are or are not present, and if both sexes are so 
equipped. If they should prove to occur only in males during the mat- 
ing season then it might be inferred that they are for the purpose of pre- 
venting undue atrophy of the stomach, through functioning as a sort of 
gastric "chewing gum," during the many weeks that this sex is without 
food while guarding the harem. 


Potentially there is much to be done on the urinary system of the ma- 
rine mammals. A fundamental question is how do these creatures get 
the water which they must drink. If they obtain it by drinking sea water, 
then how do they eliminate the excess of saline matter, which they would 
be obliged to accomplish by some specialized means? Actually, how- 



ever, there are practically no details of this system so far known that are 
surely attributable to an aquatic life. It is usually stated that the ceta- 
cean kidney is lobulated, but this is a misleading term. That of most if 
not all large terrestrial mammals is lobulated, but in the Cetacea it takes 
a different form, each kidney appearing to be an aggregation of a mul- 
titude of small kidneys closely packed and contained within a single 
envelope. The significance of this condition is unknown. 

Although of doubtful significance aquatically, there are many inter- 
esting features of the gonads of highly adapted aquatic mammals, some 
of which may here be accorded brief mention. Engle (1926) has re- 
ported on some of the equipment of the males as follows: 

Seminal vesicles Prostate gland 

Odobenus none present none 

Eumetopias none present none 

Phoca none present none 

Halicore 1 pair one none 

Phocaena none paired none 

Dephinapterus . . none present none 

Sibbaldus none present none 

As previously mentioned there is a tendency coupled with the aquatic 
life to eliminate the scrotum together with other bodily protuberances. 
Accordingly, among Pinnipedia, Sirenia, and Cetacea, the Otariidae and 
Odobenidae are the only ones which retain this feature. In the Cetacea 
the testes are not truly abdominal, but are situated in "a pouch near the 
inguinal ring." As Meek (19'18) has stated, in the porpoise the penis 
is greatly modified, the copulatory part being differentiated from the 
rest and separated by a joint which allows a wide range of movement. 
This jointed character is not present, or at least apparent, in Mysticeti. 
The cavernous body of the penis proper is single in the Cetacea and 
double in Sirenia. In both sexes of the Cetacea the uro-genital and anal 
orifices are situated within a common sulcus bordered by a pair of labia. 


The respiratory system and its appurtenances is perhaps the most criti- 
cal factor in the cetacean equipment, for every activity of an aquatic 
mammal is governed by the necessity of renewing as often as necessary 
the supply of air. This need is often antagonistic to many of its other 
activities, as it very seldom is in the case of a terrestrial mammal. 

During his investigations of the lungs of the porpoise Tursiops, G. 


B. Wislocki (1929) found that the cartilaginous armature is such as to 
give unusual strength and incompressibility. In the smaller bronchioles 
possessing a diameter of less than 0.5 mm. there are muscular sphincter 
valves. Moreover in the respiratory bronchioles there is a complete 
lining of flattened respiratory epithelium instead of a partial one as in 
other known mammals. There is a tremendous development of elastic 
tissue, and double capillary beds of interalveolar septa instead of single 
ones, as in terrestrial mammals. The respiratory bronchioles do not 
possess sacculations and alveolar ducts are lacking. 

The above sphincters probably close at the end of inspiration and 
ordinarily remain so until expiration begins, preventing the gradual 
collapse of the air spaces as outside pressure increases, thus acting in 
antagonism to the elastic tissue of the lungs. The individual sphincters 
are weak, but there are probably several million of them, and each 
imprisons but a minute amount of air, so that all of them working to- 
gether are doubtless capable of preventing the escape of air even into 
the trachea under any pressure to which the animal cares to subject it- 
self. Thus external and internal pressure can be equalized, as it must 
be in any animal which experiences an external pressure of a ton to the 
square inch. No thorax could resrst this stress without collapsing: and 
this is probably the chief physiological adjustment necessary for deep 
diving. Others are necessary, to prevent such things as bleeding at the 
eyes, and likely some alteration in the action of the heart, but it seems 
that these should be of a more minor character. The epitheleal equip- 
ment presents an increased surface area subserving the function of 
respiration, and the elastic tissue forms a powerful mechanism for empty- 
ing the lungs in a minimum of time. 

Pick (1907) has stated that the lungs of the dugong agree with those 
of the Cetacea in having a cartilaginous armature (which, however, is 
partially calcified) extending to the smallest bronchioles, a tremendous 
quantity of elastic tissue, and in the size of the air-sacs and thickness of 
the septa; but the distribution of muscular tissue differs. 

Schulze (1906) made an estimate of the number of air cells and their 
respiratory surface in the sloth, cat, man and porpoise (Phoceana com- 
munis, which reaches a length of about five and a half feet), as follows: 

Air cells Respiratory surface 

Sloth 6,250,000 5 square meters 

Cat 400,000,000 20 square meters 

Man 150,000,000 30 square meters 

Porpoise 437,000,000 43 square meters 



He considered that the lung capacity of the porpoise was about the 
same as in man, but it is Hkely that in a cetacean of strictly comparable 
size it is somewhat greater. At least all investigators are apparently in 
agreement that aquatic mammals have a greater lung capacity than ter- 
restrial ones of the same size. 

The above facts indicate that not only can the porpoise retain air in 
the lungs entirely independent of the muscles of the thorax or dia- 
phragm, but the oxygen can be more completely utilized by the lungs 
themselves. Not only this, but the blood of cetaceans and pinnipeds at 
least is richer in hemoglobin than usual, and hence can store more oxy- 
gen. In addition the intrinsic musculature of the lungs and the character 
of the diaphragm enable them to empty the lungs more completely, and 
the pressure at considerable depths also facilitates the absorption of all 

The physiology of normal respiration is too involved to present here 
and may be obtained from any good text book. Certain points must be 
considered briefly, however. The respiratory apparatus of man is not 
particularly efficient. Thus, about 20 per cent of the air which we 
breathe is composed of oxygen, but in normal, unlabored breathing we 
utilize only about a quarter of this amount. If nitrogen be added so that 
the oxygen be reduced to 10 per cent the respiratory rate will be slightly 
raised, and pulse accelerated, but there is no definite discomfort felt. 
When the oxygen content is gradually reduced to 6 or 7 per cent, how- 
ever, consciousness is lost, often without much of a premonitory symp- 
tom. Thus, in spite of rapid respiration, we are incapable of making 
much use of one-third of the oxygen that is furnished us. It is likely 
that whales can use almost all of this. 

The total lung capacity of the average man is supposed to be about 
4,700 cc. During quiet respiration only about 500 cc. is taken in, and 
in spite of our best efforts we are unable to empty the lungs of their 
final 1,000 cc. It is likely that the whale almost completely fills the 
lungs at each inspiration, and disposes of practically all the residual air 
at each expiration. 

In man the expired air usually contains about 4 per cent of carbon 
dioxide. When the CO, content of the inspiration is raised to even 1 
or 2 per cent increased breathing results to a marked degree, and at a 
content of 10 per cent there is great distress experienced and the face 
becomes congested, while the respiration is multiplied quantitatively sev- 
eral fold. It has been said that the breath cannot be held voluntarily 
after the carbon dioxide content of the lungs has passed 7.5 per cent at 



which time the oxygen in the lungs has been reduced to a point of from 
9 to 1 1 per cent. Whether whales differ in their tolerance of free car- 
bon dioxide in the lungs is as yet unknown, but it is assumed that they 
must be much more tolerant, which would be accomplished at least par- 
tially by alteration in the sensitivity in this respect of the respiration cen- 
ter of the brain. 

But there is another aspect from which this gas must be considered. 
In dogs the venous blood has a CO, content of about 45 cc. per 100 cc. 
of blood. Not more than from 2 to 2.5 cc. of this can be held in 
physical solution, while the remainder must be in chemical combination. 
Carbon dioxide can enter the corpuscles and react with alkalis combined 
with the hemoglobin to form a bicarbonate which in the lungs breaks 
down again to liberate CO,. Not only should the efficiency of this 
process be quantitatively facilitated in marine mammals, with their re- 
ported increased hemoglobin content, but there may be qualitative aug- 
mentation as well. Pressure may be of critical import in this connec- 
tion, as it might well be if it should prove that the oil of whales can 
actually take up some of the carbon dioxide from the blood. I regard 
it as certain that any such physiological processes cannot operate with an 
equal degree of efficiency at surface-water pressure and at the pressure 
of a ton to the square inch. For this reason the possibility does not ap- 
pear fantastic that the cachalot may not be able to hold its breath for 
an inordinate length of time — say fifteen minutes — when at the surface, 
while the increased pressure at the depth of one mile might so alter 
its physiological processes that it would have no difficulty in remaining 
below for an hour. There is no evidence whatever in support of such 
a theory, but it must be taken into consideration in planning future 

It is not always easy to determine the length of time during which a 
mammal may hold its breath. Thus I have taken every opportunity 
to observe seals and sea-lions but have never yet seen one submerge 
for much more than a couple of minutes. And yet it seems certain that 
at least the boreal seals, which spend much of their time beneath the ice, 
must very greatly exceed this, say to the extent of at least ten and very 
possibly twenty minutes. 

A man may hold his breath without undue discomfort for one 
minute — I have just done so. The reader may exceed this, and if he 
first violently and completely empty his lungs several times he may be 
able to last as long as two and a half minutes. But this is a very mod- 
erate accomplishment. Beebe (1926) has stated that a sloth breathes 



about once per second, and yet he has seen one recover from 30 minutes 
of complete submergence, while he had heard of one doing so after 
45 minutes. Burrell (1927) found that the adult platypus can remain 
under water for the duration of 5 to 10 minutes, while the new born 
you'iig can survive a submergence of three and a half hours. Parker 
(1922) reported that a large manati at rest was in the habit of staying 
down for a period varying from 7 minutes to 16 minutes and 20 seconds, 
then arising to the surface for as much as three minutes and taking sev- 
eral breaths. A younger animal arose oftener, and fishermen told him 
that when these beasts are being hunted they may stay below for half 
an hour. 

An experienced whaler whom I consider trustworthy has assured me 
that when it "sounds" a humpback will ordinarily stay down from 
8 to 12 minutes, with observed maximum of 23 minutes; finback, 8 to 
12, with 28 as a maximum; bowhead, 12 to 15, with 30 minutes as a 
maximum ; and cachalot from 30 to 60 minutes with an observed maxi- 
mum of one hour and 45 minutes. Andrews (1916) has quoted an 
instance where a blue whale sounded for 50 minutes, reappeared to 
spout 20 times and then disappeared for another 40 minutes. Cer- 
tainly there is abundant evidence that the cachalot can easily submerge 
in excess of an hour, but records of over two hours should be viewed 
with suspicion. Certain other of the larger Odontoceti (as Hyperoodon) 
are notable for the speed with which they seek great depths when har- 
pooned, and for the length of time that they can stay there. The 
Mysticeti are evidently less gifted in this respect, and it is doubtful 
whether this ability is as well developed in porpoises with littoral pre- 

That lengthy submergence is not fatal to the sloth and young platy- 
pus is doubtless chiefly attributable to their low rate of metabolism, and 
armadillos and ant-eaters should be expected to exhibit the same accom- 
plishment. This has also been advanced not infrequently as the pos- 
sible explanation of why whales are also able to suspend breathing 
for so long. There is no evidence to show that in any marine mammal 
the rate of metabolism is low, and several particulars indicate that it 
is high, among which are details of the lungs, blood and the tempera- 
ture. In Turstops the latter is about 36 degrees Centigrade. So in the 
Cetacea I am inclined to believe for the present that what may be desig- 
nated as the normal rate of metabolism is not low. But some fish are 
known to be able to contrive, by some obscure means, to lower their 
rate of metabolism very quickly when it is advantageous for them to 



do SO. The same may almost be said of mammals which hibernate. 
This state, apparently, is almost entirely determined by temperature in 
the case of reptiles, but in mammals this is only partially so, and there 
seem to be other factors involved. With them metabolism almost ceases, 
and although the process is not well understood, it must be accomplished 
in a relatively simple manner, for some species habitually hibernate 
while other closely related species, in other climates, never do so. 
Furthermore, some rodents regularly aestivate during the hottest part 
of the year, when food is scarcest. 

In view of the facts so far available it seems that in any study of 
the whale's ability to withstand lengthy submergence, account will have 
to be taken of the possibility that this order, or at least many of its rep- 
resentatives, may have some apparatus whereby its rate of metabolism 
is lowered temporarily in a more or less voluntary fashion. So little 
is yet known of this general subject that there is no way of predicting 
the probability of this being the case. It may actually be either quite 
high or zero ; but it should be considered, nevertheless. 

No mention has yet been made of the fact that during deep sub- 
mergence by a human diver the pressure saturates the blood with nitro- 
gen, and unless decompression be very gradual, nitrogen bubbles will 
form in the vascular system, afflicting him with what is known as the 
"bends" and often causing death. Whales often ascend with rapidity 
from great depths and at first thought it might seem that they would 
be obliged to have some provision for overcoming such a disagreeable 
situation. It must be remembered, however, that a human diver gets 
a new supply of nitrogen at every breath, while a whale has only the 
initial supply of this gas which he has taken down with him. Very 
likely this is not sufficient in amount for saturation of the blood to the 
point where the latter would give off bubbles when the animal again 
reaches the surface. 

Very little is surely known about the depths to which marine mam- 
mals habitually descend. Almost nothing is known in this regard about 
pinnipeds. Reports are sufficiently frequent for us to believe that the 
cachalot often feeds near the bottom where the depth is in excess of 
a mile. How much deeper it can go is unknown. It is also believed 
that the other toothed whales of larger size have this ability to a pro- 
nounced degree. I have been told by a whaler of a finback which was 
harpooned and at once sounded vertically to the depth of 275 fathoms, 
where its neck was broken by impact with the bottom, and the carcass 
was then hauled straight up to the surface. Other reports exceed this 


to some degree. But it is probable that no mysticete can submerge as 
deeply as some favored odontocetes, and it does not seem probable that 
the majority of httoral porpoises are so phenomenally gifted in this 

Of particularly vital import in the function of respiration by marine 
mammals is the diaphragm, and it is only to be expected that there 
should occur some change in its details. In a pronograde mammal, as 
discussed by Jones (1913) , the fixity of the fore limbs allows the muscles 
passing therefrom to the thorax to assist in breathing by raising the ribs, 
while in orthogrades this function has become obsolete. It should be 
still further eliminated in those marine mammals which seldom or 
never use the fore limbs for pressing against a hard surface. An increase 
in the muscular character of the diaphragm of aquatic mammals indicates 
that this is probably so. 

Evidently in all aquatic mammals that are very highly modified for 
such a life the diaphragm is more sloping, or tending to assume a posi- 
tion more nearly parallel with the body axis, than in terrestrial mammals, 
and this is said to be so even in the otter. Certainly it is a character 
of all pinnipeds, cetaceans and sirenians, in the latter order reaching 
its greatest alteration. It is indicated that this character increases with 
ontogenetic development, for Beddard (1900) has stated that in an 
adult porpoise the ventral and dorsal extent of the thoracic cavity showed 
a proportion of 1 to 2.25, while in a young individual this was as 1 to 
1.75. I cannot state this proportion in the sirenians but Murie's de- 
scriptions and figures show that in the manati the diaphragm extends 
from the much reduced sternum below, quite to the last thoracic vertebra. 
As this mammal has but two lumbar vertebrae this means that the dia- 
phragm almost reaches the posterior end of the abdominal cavity and is 
as nearly parallel with the body axis as it could well be. 

Muller (1898) believed that the Mysticeti breathe more with the 
thorax and less with the diaphragm, for the latter is less muscular in 
this group than in the Odontoceti, but cetacean conditions are difficult 
to interpret. In the Sirenia it seems that the rigidity of the costal ar- 
ticulations would largely inhibit much mobility of the thorax while 
breathing. The reduction of the sternum and the costal cartilages, how- 
ever, does point to the probability that these features do facilitate mobil- 
ity of the diaphragm, and this must be of extraordinary efficiency. 

It seems that the pronounced slope of the diaphragm in marine mam- 
mals may have been assisted by an increase in the lung capacity, and an 
advantage thereby gained is that the levitation supplied by the inflated 



lungs is shifted by just so much toward the center of the body. In 
such a mammal as man the lungs tend to raise the anterior portion of 
the body, while there is a corresponding depression of the posterior end, 
in the water. This is not as yet entirely overcome in the pinnipeds, 
but in sirenians and cetaceans the natural position by flotation is hori- 
zontal, largely permitted by the alteration of the diaphragm and lungs. 


There are many points of interest connected with the vascular system 
of certain aquatic mammals, but it is as yet uncertain just what applica- 
tion these may have to life in the water. It is known that pinnipeds 
and cetaceans are abundantly supplied with blood. The accounts of 
sealers frequently make reference to this fact. Whether this character 
is more pronounced in the Pinnipedia is unknown, but it may well be 
so, for several authors, notably Murie, have shown that the seal, sea- 
lion and walrus are equipped with a capacious dilation of the vena cava 
in juxtaposition with the liver, absent in the Cetacea. I have verified 
this particular for the seal and sea-lion, and have found that in a young 
individual of the latter it was notably less marked, which seems to indi- 
cate that it is an ontological development. The capacity of this hepatic 
sinus is quite astounding and, as Murie has remarked, it would seem 
to occupy as much space when expanded as the liver itself. It cannot 
be doubted that it functions as a blood reservoir. 

The cetacean heart is noteworthy in many respects as enumerated in 
the literature. Apparently it is larger than in a terrestrial mammal of 
equal mass, for G. B. Wislocki tells me that the heart of a Tursiops, 
weighing perhaps 600 pounds, is larger than that of an ox. Owen noted 
that the heart of a large whale may be more than a yard broad and not 
much less in length. The axis of the heart has also shifted, with apex 
more dorsal than in other mammals. The papillary muscles are usually 
said to be enormous and the organ of such conformation as to indicate 
great potential power. 

Kiikenthal (1922), quoting various sources, has remarked the strong 
enlargement of the spinal meningeal arteries within the vertebral canals 
of the Cetacea. He considered that by this provision the blood supply 
of the brain escapes the effect of great pressure when the animal dives 
deeply. This attitude is scarcely justified, however, for it is hardly pos- 
sible that during long submergence the pressure of the blood could be 
much less in one part of the body than another. But something of im- 
portance seems indicated, for Kiikenthal stated that in baleen whales 



the transverse foramina of the cervical vertebrae become reduced, ac- 
companying ehmination of the internal carotid and vertebral artery, 
in most toothed whales these foramina are still more reduced, and in 
Physeter and Hyperoodon they are entirely lacking, this detail thus ap- 
parently being dependent upon the possible depths to which whales 
descend. In Sirenia the transverse foramina are also rudimentary. 

The most spectacular detail of the vascular system of aquatic mam- 
mals is the extent to which retia mirabilia, or networks of vascular 
anastomoses, occur. They are found extensively in the Cetacea, Sirenia 
(depicted with especial advantage in Murie's figures), in the Phocidae, 
and to a lesser degree possibly in the Otariidae and Odobenidae. But 
this is tiot in itself an aquatic adaptation, for extensive retia have also 
been reported in monotremes, some marsupials, some lemurs, ant-eaters, 
sloths, armadillos, some rodents, the Manidae, and others. Whatever 
the function may be we are justified in assuming that retia are not a 
secondary adaptation in aquatic mammals, but rather that such as ex- 
hibit this character have lacked the stimulus for fusion of the retia 
into larger blood vessels and that at least those of a diffuse pattern have 
been retained from the primitive ancestral condition, or rather that the 
embryonic condition is retained throughout life, as pointed out by von 
Baer (1835). 

It is not easy to determine the extent of retia without specially in- 
jected material. They were not striking in the sea-lion which I dis- 
sected, while they were, over certain areas, in the seal. They may occur 
in the arterial or venous system or both. Retia may be gathered in 
single, sheathed bundles, which also convey lymphatic trunks, as in 
sloths and armadillos, or a diffuse pattern as in monotremes and siren- 
ians. They may occur in different areas dependent upon the type of 
mammal considered. Many authors have noted the intracranial retia 
of the internal carotids at the base of the skull in ruminants, located at 
the sides and back of the sella turcica. These are said to be better de- 
veloped in grazers than browsers, and least so in the giraffe. Some 
vascular clusters seem clearly to function as reservoirs for blood, as in the 
case of the psoadic plexus of Cetacea. This takes the form of numerous, 
transverse, separated blood vessels posterior to the kidneys. 

Vrolik believed retia to be connected with aboreal habits, and Carlisle 
that they were correlated with slow movements, but the circumstance 
that they are present in some agile, nonarboreal mammals refutes these 
hypotheses. Hunter and Cuvier assigned to those of the arterial system 
the function of storing oxygenated blood, and Wilson, the storing of 



carbonized blood by the venous ones, Turner and Milne-Edwards of re- 
tarding the flow, and Murie was of the opinion that they facilitate inter- 
change of substances with the lymphatic system. Hyrtl (1854) ad- 
vanced the hypothesis that the diffuse pattern of retia is associated with 
animals doing heavy muscular work but of an agile character, as with 
the burrowing armadillo, while the cluster pattern is characteristic of 
mammals in which movement is slow and posture prolonged, such as 
the sloth. Wislocki (1928) has discussed this question at considerable 
length and the reader desirous of further information is referred to 
his paper. With the attention that the subject of retia is now attracting 
it is to be hoped that we may soon know more about their useful func- 
tions. At present, however, there is no strong evidence indicative of 
what these may be. There is no evidence whatever that a retial condi- 
tion of the blood vessels is useful for the storage of reserve blood. In 
itself it would not assist in retarding the flow, nor is it likely that there 
is interchange of substances with the lymphatic system within the retial 
bundles, where these occur. In fact it seems that very little can yet be 
said except that possibly a difl^use type of retia might largely overcome 
any interruption of the blood flow that muscular or other pressure is 
capable of producing. 


Before terminating the present contribution brief consideration should 
be accorded the subject of the mammary equipment of aquatic mammals. 
The way in which these are used is usually of slight consequence to a 
terrestrial mammal, for except in particular cases as in some ungulates, 
the young may nurse from any position. It is of critical import, however, 
that an aquatic mammal shall be able before it forsakes the land en- 
tirely to contrive a method whereby the young may suckle with reasonable 
comfort while in the water. If this cannot be accomplished the creature 
must either retain its connection with the land, with consequent impli- 
cation of reduced ability to develop the highest aquatic modifications, 
or else become extinct. 

Probably in no mammal has the aquatic life caused any definite altera- 
tion in the position of the nipples, a possible exception being in the 
Sirenia. In Cetacea and Pinnipedia there is one pair of mammae situ- 
ated inguinally. In Sirenia there is a pair of axillary ones, but they now 
occur practically upon the posterior border of the flippers. The coypu 
(Myocastor) has two pairs that are situated almost upon the back, one 
pair being just behind the shoulders and the other near the haunches. 
They are largely similar in the capybara but rather less elevated, as is 


also the case in some terrestrial species with octodont affinity. The coypu 
and capybara are said frequently to swim about with the young perched 
on the back, and apparently the latter are able to nurse from this position. 

Sirenians are reported to nurse while the parent maintains an upright 
posture with the young clasped by the flippers, although I do not see 
how the latter could reach the nipple from this position. 

There are numerous reports on the mammary equipment of the Ceta- 
cea. The gland itself is apparently of the usual histological character but 
there are numerous and rather large galactophorous sinuses which open 
into a lacteal duct or reservoir of generous proportions. Engle (19'27) 
found that in a female humpback 44 feet long the mammary glands, in 
full function, had a length of about 6 feet and the greatest diameter of 
18 inches. The lacteal capacity of a large whale must surely be aston- 
ishing and a great many gallons. The nipple is retracted within a slit- 
like orifice on either side of the genito-excretory labia. 

There is a great deal of uncertainty regarding the way in which the 
cetacean mammillae are employed. The great storage capacity should 
indicate that the milk is removed at a rapid rate. The presence super- 
ficially of a part (at times disconnected from the main sheet) of the 
panniculus carnosus muscle makes it not unlikely that this may assist 
in the voluntary and forcible ejection of the fluid. And the situation 
of the nipples introduces a further element of uncertainty. I have been 
told of a female finback in full milk whose nipples protruded to the 
extent of one foot when she was inflated with air, and there are a very 
few published accounts of protruding nipples. The condition is so sel- 
dom encountered, however, that I regard it as doubtful whether it ever 
occurs after death save in the event that an animal in full milk has been 
inflated to usual degree by introduced air or by gases of decomposition, 
and possibly the sudden death of the calf and. consequent engorgement 
of the mammary glands might have the same effect. On the other 
hand the securing under water of milk by the young would be so diffi- 
cult without temporary protrusion of the nipple that I regard it as 
highly probable that this takes place, and this could easily be accom- 
plished when necessary by contraction of the mammillary smooth muscle. 

There has been much speculation regarding the cetacean position for 
nursing. Obviously if the dam rolls over on her side so that the blow- 
hole of the nursing young is above the surface the female's blow-hole 
is submerged. If hers is elevated "then the young is completely below 
the surface; but Scammon (1874) illustrated the act in this position. 
Perhaps either posture is employed. 


Literature Qited 

Abel, O. 1906. Die Milchmolaren der Sirenen. Neues Jahrb. f. Miner., Geol. 

u. Palaeont., vol. 2. 
. 1908. Die Morphologic der Hiiftbeinrudimente der Cetaceen. 

Denkschr. k. Akad. Wiss. math.-nat. Kl., Wien, vol. 81 (1907), pp. 139-195. 
. 1924. Die Eroberungsziige der Wirbeltiere in die Meere der Vor- 

zeit. G. Fischer, Jena, pp. 1-121. 
Allen, H. 1888. Materials for a memoir on animal locomotion. Extracted 

from Report on Muybridge work at the Univ. of Pa., pp. 35-104. 
Allen, J. A. 1924. Carnivora collected by the American Museum Congo Ex- 
pedition. Bull. Amer. Mus. Nat. His., vol. 47, pp. 73-281. 
Andrews, R. C. 1909. Observations on the habits of the finback and hump- 
back whales of the eastern North Pacific. Bull. Amer. Mus. Nat. Hist., vol. 

26, pp. 213-226. 
■ . 1916. Whale hunting with gun and camera. D. Appleton & Co., 

N.Y., pp. 1-333. 
Anthony, R. 1903. Introduction a I'etude experimental de la morphogenie. 

Bull. Soc. Anthr. Paris, ser. 5, vol. 4, no. 2, pp. 119-145. 
. 1926. Les affinites des Cetaces. Ann. Inst. Oceanog., vol. 3, 

pp. 93-134. 
Baer, K. E. von. 1835. Ueber das Gefassystem des Braunfisches. Nova. Acta 

Acad. Caes. Leop. Carol, naturae curiosorum, vol. 17, p. 395. 
Bailey, Vernon. 1923. The combing claws of the beaver. Journ. Mamm., 

vol. 4, pp. 77-79. 
Beddard, F. E. 1900. A book of whales. John Murray, London, pp. 1-320. 
Beebe, W. 1926. The three-toed sloth. Zoologica, vol. 7, no. 1, pp. 1-67. 
Benham, W. B. 1901. On the larynx of certain whales. Proc. Zool Soc. 

London, pp. 278-300. 
Breder, C. M., Jr. 1926. The locomotion of fishes. Zoologica, vol. 4, no. 5, 

pp. 159-297. 
BuRRELL, Harry. 1927. The platypus. Australia, pp. 1-227. 
Cabrera, A. 1925. Genera mammalium. Insectivora., Mus. Nac. Cien. Nat., 

Madrid, pp. 1-232. 
Chapman, H. C. 1875. Observations on the structure of the manatee. Proc. 

Acad. Nat. Sci., Phila., pp. 452-462. 
Delage, Y. 1885. Histoire du Balaenoptera musculus echone sur la plage de 

Langrune. Arch. Zool. Exper. Mem., vol. 1, pp. 1-152. 
Denker, a. 1902. Zur Anatomie des Gehororgans der Cetacea. Anat. Hefte, 

vol. 19, pp. 423-445. 
DoBSON, G. E. 1882. A monograph of the Insectivora, systematic and ana- 
tomical. London, pp. 1-172. 



Dubois, E. 1924. On the brain quantity of specialized genera of mammals. 

Koninkl. Akad. v. Wetensch. Amsterdam, vol. 27, nos. 5 and 6, pp. 1-8. 
DuvERNOY, G. L. 1822. Recherches anatomique sur les organes du movement 

du phoque commun, Phoca vitulina, L.. Mem. Mus. Hist. Nat. Paris, vol. 9, 

pp. 49-70, 165-189. 
Engle, E. T. 1926. The intestinal length in Steller's sea-lion. Journ. Mamm., 

vol. 7, pp. 28-30. 
. 1926. The copulation plug and the accessory genital glands of 

mammals. Journ. Mamm., vol. 7, pp. 119-126. 

-. 1927. Notes on sexual cycle of the Pacific Cetacea of the genera 

Megaptera and Balaenoptera. Journ. Mamm., vol. 8, pp. 48-54. 
Flower, W. H. 1869. On the value of the characters of the base of the 

cranium in the classification of the order Carnivora, and on the systematic 

position of Bassaris and other disputed forms. Proc. Zool. Soc. London, 

pp. 4-37. 
. 1876. An introduction to the osteology of the Mammalia. Mac- 

millan & Co., London, pp. 1-344. 
and R. Lydekker. 1891. An introduction to the study of mam- 

mals, living and extinct., pp. 1-763. 
Garrod, a. H. 1877. Notes on the manatee (Manatus americanus) recently 

living in the Society's garden. Trans. Zool. Soc. London, vol. 10, pp. 

GiDLEY, J. W. 1913. A recently mounted zeuglodon skeleton in the United 

States National Museum. Proc. U. S. Nat. Mus., vol. 44, pp. 649-654. 
Gray, A. A. 1905. Anatomical notes upon the membranous labyrinth of man 

and of the seal. Journ. Anat. Physiol., vol. 39, pp. 349-361. 
GULDBERG, G. and F. Nansen. 1894. On the structure and development of the 

whale. Bergens Mus., vol. 5, pt. 1, pp. 1-70. 
Hay, O. p. 1928. Pleistocene man in Europe and in America. New York 

Herald-Tribune. July 1, pp. 1-8 (reprinted Wash., D.C., July 30, 1928, with 

slight changes). 
HiNTON, M. A. C. 1925. Report on papers left by the late Major G. E. H. 

Barrett-Hamilton relating to the whales of South Georgia. Crown Agents 

for the Colonies, London, pp. 1-209. 
Holder, J. B. 1883. The Atlantic Right Whales: a contribution, etc. Bull. 

Amer. Mus. Nat. Hist., vol. 1, no. 4, pp. 99-137. 
Home, E. 1796. A description of the anatomy of the sea-otter. Phil. Trans. 

Roy Soc. London, pp. 385-394. 
Howell, A. Brazier. 1924. The mammals of Mammoth, Mono County, Cali- 
fornia. Journ. Mamm., vol. 5, pp. 25-36. 
. . 1926. Anatomy of the wood rat. Monog. no. 1, Amer. Soc. 

Mamm., pp. 1-225. 

1927. Contribution to the anatomy of the Chinese finless por- 

poise Neomeris phocaenoides. Proc. U. S. Nat. Mus., vol. 70, art. 13. 
pp. 1-43. 



1929. Contribution to the comparative anatomy of the eared 

and earless seals (genera Zalophus and Phoca). Proc. U. S. Nat. Mus., 

vol. 73, art. 15, Jan. 26 (1928), pp. 1-142. 
Howell, W. H. 1921. A text-book of physiology. Saunders Co., ed. 8, pp. 

Howes, G. B. and A. M. Davies. 1888. Observations upon the morphology 

and genesis of supernumerary phalanges, with special reference to those 

of the Amphibia. Proc. Zool. Soc. London, pp. 495-511. 
Hyrtl, J. 1853. Das arterielle Gefassystem der Edentaten. Denkschr. d. k. 

Akad. d. Wissensch. Wien, vol. 6, p. 21. 
Jackson, H. H. T. 1928. A taxonomic review of the American longtailed 

shrews (genera Sorex and Microsorex). North Amer. Fauna, No. 51, pp. 

Jennings, H. S. 1924. Heredity and environment. Scientific Monthly, vol. 19, 

pp. 225-238. 
Jones, F. Wood. 1913. The functional history of the coelom and the dia- 
phragm. Journ. Anat. (London), vol. 47, pp. 283-318. 

— ■ . 1923. Mammals of South Australia. Adelaide, pp. 1-458. 

Kellogg, R. 1925. Structure of the flipper of a Pliocene pinniped from San 

Diego County, California. No. 5. Additions to the Tertiary history of 

the pelagic mammals of the Pacific Coast of North America. Contr. 

Palaeont., Carnegie Inst. Wash., Publ. 348, pp. 97-116. 
— — . 1928. The history of whales — their adaptation to life in the 

water. Quart. Rev. Biol., vol. 3, pp. 29-76, 174-208. 
Kernan, John D., Jr., and H. von W. Schulte. 1918. Memoranda upon 

the anatomy of the respiratory tract, foregut, and thoracic viscera of a 

foetal Kogia breviceps. Bull. Amer. Mus. Nat. Hist., vol. 38, pp. 231-267. 
Knauff, F. 1905. Ueber die Anatomie der Beckenregion beim Braunfisch (Pho- 

caena communis Less). Jenaischen Zeits. f. Naturwis., pp. 253-318. 
KiJKENTHAL, W. 1889. Die Hand der Cetaceen. Denkschr. d. med. naturwiss. 

Gesell., Jena, vol. 3, pt. 2, pp. 23-69. 
. 1891. Ueber die Anpassung von Saugethieren an das Leben im 

Wasser. Zool. Jahrb., Abtheilung System. Georg. Biol. Thiere, pp. 373-399. 
1891. On the adaptation of mammals to aquatic life. Ann. Mag. 

Nat. Hist., ser. 6, vol. 7, pp. 153-179. 

1893. Vergleichend-anatomische und entwicklungsgeschichtliche 

Untersuchungen an Walthieren. Denkschr. med. -naturwiss. Gesell., Jena, 

vol. 3, pt. 2, pp. 224-338. 
Leboucq, H. 1887. La nageoire pectorale des cetaces au point de vue phylo- 

genique. Anat. Anz., vol. 2, pp. 202-208. 
. 1889. Recherches sur la morphologie de la main chez les mammi- 

feres marins. Archiv. Biol., vol. 9, pp. 571-648. 
Le Danois, E. 1910. Recherches sur I'anatomie de la tete de Kogia breviceps 

(Blainv.). Arch. Zool. Exp. Gen., vol. 6, p. 149. 



LiLLiE, D. G. 1910. Observations on the anatomy and general biology of some 
members of the larger Cetacea. Proc. Zool. Soc, London, pp. 769-792. 

Malm, A. W. 1871. Hvaldjur i sveriges museer. Kongl. Svenska Vetenskaps- 
Akad. Hand!., vol. 9, pp. 1-104. 

Meek, A. 1918. The reproductive organs of the Cetacea. Journ. Anat., vol. 52, 
pp. 186-210. 

Miller, G. S., Jr. 1923. The telescoping of the cetacean skull. Smiths. Misc. 
Colls., voL 76, no. 5, pp. 1-62. 

. 1929. The gums of the porpoise Phocoenoides dalli (True). 

Proc. U. S. Nat. Mus., vol. 74, Art. 26, pp. 1-4. 

MivART, St. G. 1885. Notes on the Pinnipedia. Proc. Zool. Soc. London, 
pp. 484-501. 

MiJLLER, O. 1898. Untersuchungen uber die Veranderungen welche die Res- 
pirationsorgane der Saugethiere durch die Anpassung an das Leben im 
Wasser erlitten haben. Jena. Zeitschr. f. Naturwis., vol. 25, pp. 95-230. 

MURIE, J. 1870. Researches upon the anatomy of the Pinnipedia. Part 1. On 
the walrus (Trichechus rosmarus Linn.). Trans. Zool. Soc. London, voL 7, 
pp. 411-464. 

. 1870. Note on the anatomy of the walrus. Proc. Zool. Soc. Lon- 
don, pp. 544-545. 

1872. Researches upon the anatomy of the Pinnipedia. Part 

Descriptive anatomy of the sea-lion (Otaria jubata). Trans. Zool. Soc. Lon- 
don, vol. 7, pp. 527-596. 
. 1872. On the form and structure of the manatee (Manatus ameri- 

canus). Trans. ZooL Soc. London, vol. 8, pp. 127-202. 

1873. On the organization of the caaing whale, Globiocephalus 

melas. Trans. Zool. Soc. London, vol. 8, pp. 235-301. 

1874. Researches upon the anatomy of the Pinnipedia. Part 3. 

Descriptive anatomy of the sea-lion (Otaria jubata). Trans. Zool. Soc. 
London, vol. 8, pp. 501-582. 
. 1885. Further observations on the manatee. Trans. Zool. Soc. 

London, vol. 11, pp. 19-48. 
Nichols, J. T. 1915. On one or two common structural adaptations in fishes. 

Copeia, voL 20, pp. 19-21. 
OSBURN, R. C. 1906. Adaptive modifications of the limb skeleton in aquatic 

reptiles and mammals. Ann. Acad. Sci. N.Y., vol. 16, pp. 447-479. 
. 1903- Aquatic adaptations. No. 1. Adaptations to aquatic, ar- 
boreal, fossorial and cursorial habits in mammals. Amer. Nat., vol. 37, 

pp. 651-665. 
Owen, R. 1866, 1868. On the anatomy of the vertebrates, vols. 2 and 3. 

Longmans Green & Co., London, pp. 1-592, 1-915. 
Parker, G. H. 1922. The breathing of the Florida manatee (Trichechus latiros- 

tris). Journ. Mamm., vol. 3, pp. 127-135. 
Parsons, F. G. 1894. On the morphology of the tendo-achilis. Journ. Anat. 

Physiol., vol. 28, pp. 414-418. 



Pick, F. K. 1907. Zur feineren Anatomic der Lunge von Halicore dugong. 
Arch. f. Naturgesch. 73 Jahrb., voL 1, p. 245. 

PococK, R. L 1921. On the external characters and classification of the Lutrinae 

(Otters). Proc. ZooL Soc. London, pp. 535-546. 
. 1921. On the external characters and classifications of the Mus- 

telidae. Proc. Zool. Soc. London, pp. 803-837. 

PoucHET, G. and H. Beauregard. 1885. Note sur I'organe des spermaceti. 
Compt. Rendu Soc. Biol. Paris, vol. 37. 

Putter, O. 1902. Die Augen der Wassersiiugetiere. Zool. Jahrb., Abth., f. 
Anat. Jena, vol. 17, pp. 99-402. 

Roux, W. 1883. Beitriige zur Morphologic der functionellen Anpassung. L 
Structur eines hochdiflferenzirten bindegewebigen Organes. Arch. Anat. 
Physiol., pp. 76-162. 

Rowley, J. 1929. Life history of the sea-lions on the California coast. Journ. 
Mamm., vol. 10, pp. 1-36. 

Ryder, J. A. 1877. On the laws of digital reduction. Amcr. Nat., vol. 11, 
pp. 603-607. 

. 1885. On the development of the Cetacca, together with a con- 
sideration of the probable homologies of the flukes of cetaceans and sirenians. 
Rept. U. S. Comm. of Fisheries, pp. 427-485. 

SCAMMON, C. M. 1874. The marine mammals of the north-western coast of 
North America. San Francisco, pp. 1-319. 

Schulte, H. von W. 1916. Anatomy of a foetus Balaenoptera borealis. 
In Monogs. of Pacific Cetacea. 2. The Sei whale (Balaenoptera borealis 
Lesson). Mem. Amer. Mus. Nat. Hist., new ser., vol. 1, pt. 6, pp. 389-502. 

and M. deF. Smith. 1918. The external characters, skeletal 

muscles, and peripheral nerves of Kogia breviceps (Blainville). Bull. 
Amer. Mus. Nat. Hist., vol. 38, pp. 7-72. 

SCHULZE, F. E. 1906. Beitrage zur Anatomic der Saugctier lungcn. Sitzber. 
Kgl. preuss. Akad. d. Wissenschaft. Berlin, pp. 225-243. 

SoNNTAG, C. F. 1922, 1923. The comparative anatomy of the tongues of the 
Mammalia. Proc. Zool. Soc. London, pp. 639-647 (Cetacea, Sirenia), pp. 
145-151 (Pinnipedia). 

Steinmann. 1912. Ueber die Ursache der Asymmetric der Wale. Anat. Anz., 
vol. 41, pp. 45-54. 

Struthers, J. 1880-81. On the bones, articulations, and muscles of the rudi- 
mentary hind-limb of the Greenland Right-Whale (Balaena mysticetus). 
Journ. Anat. Physiol., vol. 15, pp. 141-176, 301-321. 

- — . 1887-88. On some points in the anatomy of a megaptera longi- 

manna. Journ. Anat. Physiol., vols. 22, 23. 

. 1893. On the rudimentary hind-limb of a great fin-whale (Balaen- 
optera musculus) in comparison with those of the humpback-whale and 

the Greenland Right-Whale. Journ. Anat. Physiol., vol. 27, pp. 291-335. 


Taverne, L. 1926. A propos de I'orientation difference de la nageoire caudale 
chez les Cetaces et chez les poissons. Bull. Mus. Nation. Hist. Nat. (Paris), 
p. 260. 

Taylor, W. P. 19l4. The problem of aquatic adaptation in the Carnivora, 
as illustrated in the osteology and evolution of the sea-otter. Univ. Calif. 
Publ. Geol., vol. 7, no. 25, pp. 265-295. 

Thomas, A. and R. Lydekker. 1897. On the number of grinding teeth pos- 
sessed by the manati. Proc. Zool. Soc. London, pp. 595-600. 

Thompson, A. 1869. On the difference in the mode of ossification of the 
first and other metacarpal and metatarsal bones. Journ. Anat. Physiol., vol. 

3, p. 131. 

Thompson, D'Arcy W. 1917. On growth and form. Cambridge Univ. Press, 

pp. 1-793. 
Todd, T. W. 1922. Numerical significance in the thoracicolumbar vertebrae 

of the Mammalia. Anat. Rec, vol. 24, pp. 261-286. 
Townsend, C. H., 1914. The porpoise in captivity. Zoologica, vol. 1, no. 16, 

pp. 289-299. 
Trutat, E. 1891. Essai sur I'histoire naturelle du desman des Pyrenees. 

Toulouse, pp. 1-107. 
Turner, Wm. 1883. Cervical ribs, and the so-called bicipital ribs in man, 

in relation to corresponding structures in the Cetacea. Journ. Anat. Physiol., 

vol. 17, pp. 384-400. 
• . 1888. Report on the seals collected during the voyage of the 

H.M.S. Challenger in the years 1873-1876. Reports on the Sci. Results 

of the Voyage of the H.M.S. Challenger. Zoology, vol. 26, pp. 1-240. 
VON Huene, F. 1922. Die Ichthyosaurier des Lias and ihre Zusammenhange, 

4. Jahresversammlung der Palaeontologischen Gesell. im Aug. zu Tubingen, 
Verlag von Gebriider Borntraeger, BerHn, pp. 1-114. 

Warren, E. R. 1927. The beaver, its work and its ways. Amer. Soc. Mamm. 

Monog. no. 2, pp. 1-177. 
Weber, M. 1886. Studien iiber Saiigethiere. Jena, pp. 1-252. 

. 1927-28. Die Saugetiere. Jena, pp. 1-444, 1-898. 

Wilder, B. G. 1875. On a foetal manatee and cetacean, with remarks upon 

the affinities and ancestry of the Sirenia. Journ. Sci. & Arts, vol. 10, 10 pp. 
Wilder, H. H. 1923. History of the human body. Henry Holt, pp. 1-623. 
WiLLiSTON, S. W. 1902. On certain homoplastic characters in aquatic, air- 
breathing vertebrates. Kans. Univ. Sci. Bull., vol. 1, pp. 259-266. 
. 1914. Water reptiles of the past and present. Chicago Univ. 

Press, 61 pp. 
Wilson, H. S. 1880. The rete mirabile of the narwhal. Journ. Anat. Physiol., 

vol. 14, pp. 377-398. 
WiNGE, Herluf. 1921. A review of the interrelationships of the Cetacea. 

(Translation by G. S. Miller, Jr.) Smiths, Misc. Colls., vol. 72, no. 8, 

pp. 1-97. 



WiSLOCKi, G. B., 1928. Observations on the gross and microscopic anatomy 
of the sloths (Bradypus griseus griseus Gray and Choloepus hoffmanni 
Peters). Journ. Morph. Physiol., vol. 46, pp. 317-397. 

. 1929. On the structure of the lungs of ■ the porpoise (Tursiops 

truncatus). Amer. Journ. Anat., vol. 44, no. 1, pp. 47-77. 

WORTMAN, J. L. 1921. On some hitherto unrecognized reptilian characters 
in the skull of the Insectivora and other mammals. Proc. U. S. Nat. Mus., 
vol. 57, pp. 1-52. 



acoustic sense, 68 

adipose cushion, 103, 311 

Alces, HI 

Amblonyx, 30 

Amblyrhynchus, 12 

Anotomys, 24, 37, 185, 277 

antibrachium, 233 

Aonyx, 30 

aquatic reptiles, 20, 138, 191, 273, 307 

Archaeoceti, 25, 41 

Arctonyx, 127 

Argyrocetus, 117 

Artiodactyla, 38 

Arvicola, 24, 36, 185 

Atophyrax, 24, 28, 185, 276 

Aves, 13 

Balaena, 87, 126, 256, 306 

balaenid whale, 134, 153 

Balaenoptera, 57, 79, 96, 134, 172, 303 

baleen, 86 

baleen whale, see Mysticeti 

Basilosaurus, 12, 41, 147, 198, 216, 305 

Bathyergidae, 127 

Batoidea, 13 

bear, sea, see Callorhinus 

beaver, see Castor 

beluga, see Delphinapterus 

Berardius, 172 

Bison, 178, 206 

blackfish, see Globiocephala 

blow hole, 95 

blue whale, see Sibbaldus 

blubber, 57, 311 

bowhead, see Balaena 

Bradypus, 4 

brain, 125, 310 

cachalot, see Physeter 
carpus, 247 
Callorhinus, 32, 127 

Carnivora, 12, 30 

capybara, see Hydrochoerus 

Castor, 24, 37, 192, 277 

caudal vertebrae, 196 

Caudata, 12 

Centetes, 299 

cervical vertebrae, 138 

Cetacea, 12, 41, 112, 145, 170, 192, 

216, 255, 302 
Cetotherium, 120 
clavicle, 213 
coypu, see Myocastor 
Chelonia, 12 

Chimarrogale, 24, 28, 185, 276 
Chironectes, 24, 26, 185, 277 
chevron bones, 179, 204 
Choeropsis, 24, 38 
Choloepus, 4 
Cricetidae, 37 
Crocodilia, 12, 21 
Crossogale, 24, 28, 185, 276 
Crossomys, 24, 36, 185, 279 
Cymbospondylus, 273 
Cystophora, 242 

Dasymys, 24, 36, 185 
Delphinapterus, 57, 148, 258 
Delphinidae, 148 
Delphinus, 103, 178 
Dermochelys, 230 
Desmana, 24, 27, 185, 286 
Diaphorocetus, 124, 136 
diaphragm, 321 
Didelphis, 26 
digestive system, 313 
digits, 248, 296 
diverticula, nasal, 97 
dorsal fin, 58 
dugong, see Halicore 

ear, 68 

eared seal, see Otariidae 



earless seal, see Phocidae 

Elasmosaurus, 261 

elephant seal, see Mirounga 

Enhydrinae, 13, 31, 192, 285 

Eosiren, 111 

Equidae, 90 

Erinaceus, 126 

esophagus, 82 

Eubalaena, 126, 134, 256 

Eumetopias, 314 

Eurhinodelphis, 85, 113, 148, 226 

external features, 48 

eye, 61 

Ichthyomys, 24, 37, 185, 277 
Ichthyosauria, 12, 21, 192, 307 
ilium, 290, 300 
Inia, 178, 254 
innominate, 290, 300 
Insectivora, 12, 27 
insectivore otter, see Potomogale 
intestinal tract, 314 
ischium, 292, 300 

kidney, 315 

killer whale, 68, 155 

Kogia, 100, 151 

femur, 292, 301 

fibula, 296 

finback whale, see Balaenoptera 

flipper, 207 

flukes, 198 

foot, 296 

fur seal, see Callorhinus 

Galemys, 24, 28, 185, 276 

Geosaurus, 261 

Globiocephala, 56, 113, 178, 254 

gonads, 315 

Grampus, 148, 177 

gray whale, see Rhachianectes 

gular folds, 81 

hair, 52 

hair seal, see Otariidae 

Halicore, 25, 40, 173, 196 

Halitherium, 300 

head, 55, 149 

heart, 322 

Hippopotamus, 12, 24, 38, 53, 153, 185, 

humerus, 218 

humpback whale, see Megaptera 
Hydrictis, 30 

Hydrochoerus, 24, 35, 325 
Hydrodamalis, 25, 40, 196 
Hydromys, 24, 36, 185, 279 
Hydrurga, 242 
hyena, 206 

Hyperoodon, 105, 172, 319 
hyperphalangy, 259 

lachrymal duct, 64 
Lagenorhynchus, 178, 204, 307 
Lagomorpha, 38 
larynx, 91 
leg, 268 
Leporidae, 38 
Leptonychotes, 18 
limb, pectoral, 206 

pelvic, 268 
Limnogale, 13, 24, 28, 185 
lips, 78 

lumbar vertebrae, 161, 168, 180 
lungs, 91, 316 
Lutra, 30 

Lutrinae, 30, 192, 246, 299 
Lutrogale, 30 

Macropus, 129 

mammae, 324 

manati, see Trichechus 

manubrium, 161, 168, 173 

manus, 239 

marine turtle, 212, 262 

Marsupialia, 13, 24, 26 

mechanics of swimming, 9 

Megaptera, 51, 79, 134, 149, 172, 215, 

254, 303 
Mesoplodon, 148, 178 
mesosaur, 21 
metabolism, 320 
metacarpus, 247 
metatarsus, 298 
Microaonyx, 30 
Microtus, 24, 36 



Mictomys, 83 

mink, see Mustela 

Mirounga, 33, 47, 242, 283 

Mixosaurus, 191 

Monachus, 234 

Monodon, 57, 64, 84, 92, 148 

Monotremata, 12, 25 

mosasaur, 21 

mouth, 76 

mud turtle, 19 

Muridae, 36 


abductor digiti quinti longus, 233 
adductores (of leg), 294 
apaxialis, 180 

atlantoscapularis, 144, 165, 214 
biceps brachii, 238 

femoris, 280, 288, 295 
caudofemoralis, 287 
cephalohumeralis, 108, 142, 164, 214 
clavoacromiodeltoideus, 221 
cleidomastoideus, 156 
coracobrachialis, 220 
deltoideus, 215, 229 
depressor anguli scapulae, 216 
digastricus, 81 
episubscapularis, 226 
erector spinae, 170, 178, 288 
extensor digitorum communis, 238 
digitorum lateralis, 234 
poUicis, 233 
flexor digitorum communis, 233, 238 
digitorum longus, 296 
digitorum sublimis, 238 
hallucis longus, 296 
radialis, 238 
ulnaris, 235, 238 
gastrocnemius, 296 
geniohyoideus, 81 
gluteus, 189, 280, 287, 299 
gracilis, 288, 295 
gular, 81 

humerotrapezius, 110, 142, 165, 214 
hyoglossus, 81 
hypaxialis, 170, 180, 288 
iliacus, 295 
iliocaudalis, 189 
iliocostalis, 165, 180, 288 

infraspinatus, 215, 228 
ischiocaudalis, 189 
interossei, 238 
intertransversarii, 182 
latissimus dorsi, 156, 165, 214, 227 
levator caudae internus, 170 
longissimus dorsi, 165, 180 
longus colli, 142 
masseter, 108, 119 
mastohumeralis, 214 
mastoscapularis, 214, 227 
maxillonasolabialis, 88 
mylohyoideus, 81 
nasolabialis, 88 
obturator externus, 294 
palmaris longus, 235, 238 
panniculus carnosus, 164, 168, 227, 

pectoralis, 144, 155, 164, 173, 176, 

214, 253 
peroneus digiti quinti, 296 
longus, 296 

plantaris, 296 

platysma, 82 

pronator teres, 234 

psoas, 180, 295 

quadratus lumborum, 170, 180 

rectus abdominis, 170, 176, 182, 289 
capitis, 125, 132, 136 

rhomboideus, 108, 214 

sartorius, 288, 294 

scalenus, 125, l40 

semimembranosus, 288, 295 

semispinalis, 108, 131, 136, 145, 165 

semitendinosus, 280, 287, 295 

serratus magnus, 214 

soleus, 296 

sphincter colli, 81, 168 

spinotrapezius, 165 

splenius, 108, 165 

sternohyoideus, 145, 173 

sternomandibularis, 81 

sternomastoideus, 108, 156, 174 

subscapularis, 215, 228 
subscapulocapsularis, 226 
supraspinatus, 215, 228 
supinator brevis, 234 
temporalis, 108, 119 



tensor fasciae latae, 287 

teres major, 215 

tibialis anticus, 296 

trachelomastoideus, 108 

transversalis, 176 

trapezius, 156, 213 

triceps, 218, 238 
muskrat, see Ondatra 
Mustek, 30, 192 
Mustelidae, 30 

Myocastor, 24, 35, 185, 279, 324 
Mysticeti, 43, 79, 118, 216 

nails, 247, 253, 267, 277, 297 
narwhal, see Monodon 
neck, 138 

Nectogale, 24, 28, 185, 276 
Neobalaena, 151, 172, 178, 204 
Neofiber, 24, 37, 185 
Neomeris, 54, 119, 146 
Neomys, 24, 28, 185, 276 
Neosorex, 24, 28, 185, 276 
Neotoma, 275 

facial, 123 

glossopharyngeal, 125 

hypoglossal, 125 

inferior gluteal, 287 

infraorbital, 88, 123 

olfactory, 75 

spinal accessory, 12 5 

superior gluteal, 287 

trigeminal, 123 

vagus, 125 
nictitating membrane, 64 
Nilopegamys, 24, 36, 185 
nose, 87 
nostrils, 101 

Octodontidae, 35 

Odobenidae, 12, 32, 49, 77, 141, 281 
Odontoceti, 25, 43, 112 
olfactory sense, 74 
Ommatophoca, 248 
Ondatra, 24, 36, 185, 276 
Orca, 254 
Orcella, 254 

Ornithorhynchus, 25, 68, 76, 192, 291, 
319, 344 

Otaria, 242 

Otariidae, 13, 18, 32, 53, 102, 156 

otter, insectivore, see Potomogale 

river, see Lutrinae 

sea, see Enhydrinae 

Parahydromys, 24, 36, 185, 279 

Paraonyx, 30 

Patriocetus, 120 

pectoral limb, 206 

peduncle, 196 

pelage, 52 

pelvic limb, 268 

pelvis, 290, 300 

penguin, 13, 212 

Perissodactyla, 24 

pes, 296 

Phoca, 107, 194, 291 

Phocaena, 85, 148, 173, 254, 306 

Phocidae, 13, 33, 53, 90, 102, 141, 156, 

Phocoenoides, 86 

Physeter, 43, 99, 117, 126, 150, 306 
physiology, 310 
Pinnipedia, 12, 31. 52, l4l, 156, 192, 

210, 291 
Pisces, 12 

Platanista, 43, 113, 146, 215 
platypus, see Ornithorhynchus 
Plesiosaurus, 12, 139 
polar bear, see Thalarctos 
porpoise, see Delphinidae 
Potomogale, 12, 24, 28, 89, 155, 185, 

Procyon, 6 
Pseudorca, 148 
Pteronura, 30, 195 
pubis, 292, 301 

racoon, see Procyon 
radius, 233 
Rangifer, 142 
rat, water, 36 
■ Rattus, 275 
reptiles, aquatic, 20, 138, 191, 273, 307 
Reptilia, 12 
respiratory system, 315 
retia mirabilia, 323 



Rhachianectes, 51, 80, 134, 146, 172 
Rheomys, 24, 37, 185, 277 
Rhizomyidae, 127 
right whale, see Eubalaena 
ribs, 161, 166, 177 
river otter, see Lutrinae 
rorqual, 51, 202, 256 
Rodentia, 12, 35 
rostrum, 56 
Salientia, 13 
scapula, 215 
sea cow, see Sirenia 
seal, see Phocidae 
earless, see, Phocidae 
elephant, see Mirounga 
fur, see Callorhinus 
hair, see Otariidae 
sea-lion, see Otariidae 
sea otter, see Enhydrinae 
sense, acoustic, 68 
olfactory, 74 
visual, 61 

Serpentes, 12 

shank, 296 

Sibbaldus, 118, 134, 149, 172, 303 

Sirenia, 12, 40, 53, 76, 110, 140, 165, 
192, 216, 246, 300 

skin, 52 

skull, 107 

sloth, 4, 154, 318 

Sorex, 275 

Soricidae, 28 

Sotalia, 256 

Spalax, 127 

spermaceti organ, 99, 104, 311 

sperm whale, see Physeter 

Steller sea cow, see Hydrodamalis 
sea-lion, see Eumetopias 

Steno, 204 

Stenodelphis, 57, 85, 172 

Stenorhynchus, 242 

sternum, l6l, 166, 173 

stomach, 313 

sumpfotter, see Mustela 

swimming methods, 9 

Sylvilagus, 24, 38 

tail, 183 

Talpidae, 27 
Tapiridae, 24, 90 
tarsus, 283 
Tatusia, 126 
teeth, 82 
Tenrecidae, 28 
Teonoma, 107 
Thalarctos, 12, 24, 30 
thorax, 153 
tibia, 296 
tongue, 77 

toothed whales, see Odontoceti 
Trichechus, 25, 40, 83, 196 
trunk, 153 
Tursiops, 57, 90, 99 
turtle, marine, 212, 262 
mud, 19 

ulna, 233 

urogenital system, 314 
Ungulata, 12, 38 
Ursidae, 30 

vascular system, 322 
vertebrae, caudal, 196 

cervical, 138 

lumbar, 161, 168, 180 

sacral, 179 

thoracic, 157 
visual sense, 61 

walrus, see Odobenidae 

water opossum, see Chironectes 
rat, 36 
shrew, 28 

Weddell seal, 18 

whalebone, 86 

whale, balaenid, 134, 153 
baleen, see Mysticeti 
bottlenose, see Hyperoodon 
blue, see Sibbaldus 
bowhead, see Balaena 
finback, see Balaenoptera 
gray, see Rhachianectes 
humpback, see Megaptera 
killer, 68, 155 
right, see Eubalaena 



whale, xiphoid, 173 

sei, see Balaenoptera , _, . 

r.1 , yapok, see Chironectes 

sperm, see Physeter ' ^ ' 

sulfurbottom, see Sibbaldus Zalophus, 107, 142, 194, 234, 292 

toothed, j^e Odontoceti Zarhachis, 56, 113 

whalebone, see Mysticeti zeuglodont, 41, 114, 139, 198 

white, see Delphinapterus Ziphiidae, 80, 117, 148, 177 

ziphioid, 85, 117, 177 



oAquatic Mammals 

was set, printed and bound by The Collegiate Press 
of Menasha, Wisconsin. The Cover Design is by 
The Decorative Designers of Chatham, Neiv jer- 
sey. The type face is Garamont. The type is set 
10 on 11 point (Linotype). The type page is 
24 X 40 picas. The end sheets are 20 x 26 — 63 
White Antique Buckeye. The jacket is 24 x 36 — 
80 Gray Antique construction. The text paper is 
]Winnebago English Finish, 38 x 30 — 120 — 60 
Pound. The inserts are printed on 70 Pound Poly- 
chrome Enamel. The binding is Holliston 
Extra Colors 213. 

With THOMAS BOOKS careful attention is given 
to all details oj manufacturing and design. It is 
the Publisher's desire to present books that are 
satisfactory as to their physical qualities and artis- 
tic possibilities and appropriate for their particular 
use. THOMAS BOOKS will be true to those laws 
of quality that assure a good name and good will.