Skip to main content

Full text of "Distribution and origin of life in America"

See other formats











Ph.D., B.Sc. 

Author of "European Animals, their Geological History and Geographical Distribution. 

Corresponding Member of the Academy of Natural Sciences, Philadelphia; 

of the Senckeriberg Natural History Society, Frankfurt a] Main ; 

of the Linnean Society of Bordeaux ; and of the 

Anthropological Society of Paris 












THIS volume is the outcome of the second series of " Swiney 
Lectures " on Geology which I delivered at the Victoria and 
Albert Museum in London in 1908. The subject chosen for 
these lectures was the " Geological History of the American 
Fauna." Instead of publishing the lectures, like the previous 
ones on the " European Fauna," with little alteration, I decided 
to amplify and rewrite them completely. In doing so I was 
enabled to bring out more clearly certain points in the 
geological history of the American animals which seemed 
to me of general interest. 

It is always difficult to choose a suitable title for a work of 
this kind. The one I have selected is brief yet comprehen- 
sive. At the same time, it may be urged that the origin and 
distribution of Man might reasonably be expected from the 
title, whereas it has been omitted in the text. This omission 
is mainly due to the fact that the literature dealing with the 
subject, exclusive of Man, is very extensive, while the origin 
and distribution of human races in America is treated in quite 
a distinct set of periodicals and books. 

An important feature of my researches is the elucidation of 
some of the problems connected with the cause and nature of 
the " Ice Age " or " Glacial Epoch." There is a considerable 
amount of faunistic evidence that both the Atlantic and the 
Pacific oceans were closed simultaneously in the north, result- 
ing in a higher temperature of these oceans and a greater 
snowfall in the countries surrounding them in the north. 
Similarly, if the Glacial Epoch had been due to a closing of the 
Arctic Ocean, the higher temperature prevailing in the Arctic 



area during early Tertiary times must have been caused by a 
more abundant supply of warm currents in the Arctic Ocean 
at that time. Consequently North America would probably 
have been more isolated in the north than it is now, for the 
currents must have had ample facilities for passing freely into 
the Arctic Ocean. Palaeontological evidence, however, points 
to an intimate union of North America and the Old World 
during the greater part of the Tertiary Era. There are thus a 
number of interesting problems, all of which are fully dis- 
cussed in this volume, quite apart from many other important 
theories arising out of the subject of the origin and distribution 
of life in North America. In this connection the testimony 
of fossil animals, and that of the marine fauna, has been largely 
availed of. To a certain degree the distribution of plants has 
also been included. 

I have endeavoured, as far as possible, to quote all the more 
noteworthy works and papers dealing with the subject under 
consideration. Nevertheless I am conscious of having possibly 
overlooked a few important contributions ; and I only hope my 
colleagues will be good enough to draw my attention to any omis- 
sions that they may notice, so as to enable me to add them in 
case a second edition of this book should be called for. I can 
lay no claim to having made careful studies of the literature 
dealing with the Geology or Botany of the American continents. 
The works cited are those that I happened to be acquainted 
with, and no doubt many gaps might be filled by those more 
conversant with these sciences than I am. 

I have hesitated for some time before actually illustrating 
my views as to the former changes of land and water. Some 
geologists consider it inadvisable to publish palaeogeographical 
maps founded on zoogeographical data, since such maps are 
very liable to revision according to the constantly advancing 
state of our knowledge. The latter are also apt to lead to a 
good deal of adverse criticism, for one map can only represent 
a single short phase of a long series of geological ages during 
which repeated changes of land and water probably took 
place. In spite of these objections I have ventured to con- 
struct a few of these ideal maps of ancient geography as a 


general illustration of my views. I have felt that although I 
am open to criticism, my views will be more easily understood, 
and we shall eventually make more rapid progress in arriving 
at the correct conclusions as to the alterations in the configura- 
tion of the oceans and continents that actually occurred in 
the past than if such maps were altogether omitted. 

The immense advance that has been made in the United 
States in the study of zoology has greatly facilitated my work. 
And everywhere in America I found willing helpers ready to 
impart information. My thanks are particularly due in this 
respect to Prof. H. F. Osborn, Dr. Matthew and Dr. Hornaday 
of New York, to Prof. Morse and Mr. John Kitchie of Boston, 
to Dr. Henry B. Bigelow and Mr. T. Barbour of Harvard 
University, to the late Dr. A. E. Brown and Mr. Witmer Stone 
of Philadelphia, to Dr. Dall, Dr. T. W. Vaughan, Dr. Stejneger 
and Miss Kichardson of Washington, to the late Prof. Packard 
of Providence, to Dr. Ortmann of Pittsburg, to Dr. Van Den- 
burgh of San Francisco and to Dr. Kuthven of Ann Arbor. 
Several of my European friends also aided me in many ways. 
Among them particularly the following members of the staff 
of the British Museum : Dr. Smith Woodward, who was good 
enough to look through the proofs and to draw my attention 
to several errors, Dr. Andrews, Dr. Boulenger, Mr. Eegan, 
Mr. Pycraft and Mr. B. B. Woodward. Prof. Hull, Dr. Bruce 
and Dr. Stapf kindly pointed out to me various valuable 
sources of information. My most grateful acknowledgment, 
moreover, is due to the guardians of the Great Libraries, with- 
out whose ready assistance I should never have succeeded in 
accomplishing this work. The most generous facilities for 
study were given me by my friends Mr. Lyster of the 
National Library of Ireland, Mr. Praeger of the Koyal Irish 
Academy Library and Dr. Foord of the Royal Dublin Society's 
Library. I am indebted, too, for many services to Mr. Kappel, 
the Librarian of the Linnean Society, to Mr. Jones, Librarian 
of the Geological Society, Mr. Waterhouse of the Zoological 
Society of London, and Mr. Hinch of the National Library of 

I am under a great obligation to Messrs. Meiklejohn & Son 


for allowing me to reproduce two of their beautiful oroscopic 
maps, while the council of the Eoyal Irish Academy, Prof. 
Osborn, Dr. von Ihering, Dr. Ortmann, Mr. Thompson Seton, 
Dr. True and Dr. Hornaday were good enough to give me 
permission to copy maps illustrating their works. 

I can scarcely hope that all the conclusions I have arrived 
at in this book will be readily accepted, but I trust to have 
succeeded in impressing those who have not made a serious 
study of the science of geographical distribution of animals 
with the profound importance of the subject as an aid to 
palaeogeographical researches. 

B. F. S. 



November 11th 1911. 





Description of Greenland The reindeer and its origin The 
Arctic musk ox The lemming and Arctic hare On the 
ancient land connections between Greenland and the adjoining 
continents On the cause of the Ice Age The marine fauna of 
the North Atlantic cited as evidence in favour of a northern 
land connection Did the fauna of Greenland survive the Ice 
Age in the country ? The snails of Greenland quoted in support 
of this theory On the supposed polar origin of life General 
faunistic affinities of Greenland with other northern countries . 1-25 



Description of the general features of Labrador and its fauna The 
voles and jumping mice On some northern frogs The origin 
and distribution of the moose deer The fauna of the White 
Mountains and its origin Pre- Glacial relict colonies in the 
North-East Supposed effect of the Ice Age on the fauna 
Were the animals driven south by the cold ? On mild inter- 
glacial phases of the Glacial Epoch Foraminifera in the Glacial 
drift deposits The relict fauna of the Great Lakes The 
fishes of the Great Lakes and their origin Distribution and 
geological history of the fresh-water mussel Margaritana The 
past history of the Great Lakes The fauna of Newfoundland 26-57 



The Mackenzie basin and its fauna The lemmings and gophers 
On American ptarmigans The past history of the bison and 
its gradual destruction On the distribution and origin of 
the wapiti deer The Canadian tree - porcupine On the 
molluscan fauna of the north-west and its relationship . 58-73 





On the general features of Alaska, its glaciers, volcanoes and flora- 
Conditions of Alaska during the Glacial Epoch On the evidences 
derived from the fossil mammals as to the former climate of 
the country Evidences of an Asiatic invasion of animals into 
Alaska The faunistic affinities of Alaska On the former 
presence of the mammoth in Alaska The fresh-water fishes of 
Alaska The distribution and origin of the butterflies belonging 
to the genus Parnassius On the mammals peculiar to Alaska 
On the relationship of the marine mollusks of the northern 
Pacific in recent and past times Evidences of a former land 
bridge near Bering Strait 74-100 



The region of the Great Salt Lake Distribution and origin of the 
jnountain sheep and Rocky Mountain goat On the geological 
history of the American deer tribe the prong-horn antelope 
and its origin On the Eocky Mountain flora Accidental 
distribution as applied to butterflies Description of the 
American game preserves The great Tertiary deposits and 
their contents . . 101-122 



On the poverty of lizards in this region The snakes of the eastern 
states and their affinities The origin_and distribution of the 
garter snakes Rattlesnakes and their range On American 
tortoises and turtles The giant salamanders of America and 
Asia On newts and their allies The relationship of the 
Eastern plants The pine-barren flora The range of the 
star-nosed mole 123-143 



Description of the prairie region and its inhabitants On the extinc- 
tion of the horse in America and the geological history of the 
horse tribe Prairie dogs and their distribution On the origin 
and past history of the raccoon On caves and their contents as 
illustrating the geological history of the fauna The former 


presence of the reindeer south of its existing range Evidences 
of a southern fauna in the north and of its gradual southward 
retreat Helicina Turkeys and the Florida parrot On the 
dispersal of fresh- water mussels Eelicts of the past Bony-pike 
and bow-fin The blind fauna of America . 144-166 



On the vegetation and general features of Florida The geological 
history of Florida On the reptiles and amphibians of the south- 
east The south-east as a centre of dispersal The terrestrial 
mollusks of Florida On scorpions and dragon-flies Alligators 
and crocodiles, their origin and past history The fauna of 
Bermuda island The geology and flora of Bermuda The birds 
and their origin Terrestrial isopods, spiders and the land 
nemertean worm Geonemertes On the snails of Bermuda and 
their ancestors The origin of the Bermudan fauna . 167-195 



On the so-called " Sonoran Region " The palaeogeography of 
California and the adjoining States On the western insecti- 
vores On horned toads and burrowing lizards Discontinuous 
distribution of some western toads The snails of California 
The fauna of lower California and its Cape region On the 
European element of the Calif ornian fauna as illustrated by the 
mollusks, butterflies and ants The crayfishes of the south- 
west and their origin Various faunistic problems On mam- 
malian affinities The American rabbits and their relations 
The former Atlantic land connections between the New World 
and the Old, supported by the evidence derived from Tertiary 
mammals Conclusions 196-232 



The fishes of Central America and their origin On the geology and 
palaeogeography of Central America Importance of deter- 
mining the age of Central America Evidence derived from a 
study of the marine faunas On the faunistic affinities between 
North and South America The flora of Central America 
Monkeys and raccoons On birds, tortoises and snakes The 



fresh- water crabs and their dispersal The European element in 
the fauna of Central America -Oil discontinuous distribution 
of animals as a proof of their antiquity .... 233-260 



The geology of the Antilles Palaeography as illustrated by the dis- 
tribution of snails Antiquity and dispersal of snails, their 
relationship and European affinities On the theory of the 
ancient Atlantis The problem of the supposed permanence of 
ocean basins The marine mammals of the Antilles and their 
origin Evidences in favour of a mid- Atlantic land bridge 
The terrestrial mammals of the West Indian islands Evidences 
of former changes of level in the Antillean area The blind 
animals of Cuba On the ancient Antillean continent 261-294 



Darwin's visit to the islands and his views on the origin of their 
animals and plants On accidental dispersal The birds, 
tortoises and snails of the Galapagos islands Origin of the 
flora Baur's criticisms of Darwin's views Euphorbia and 
its past history Conclusions as to the origin of the Galapagos 
fauna The geological history of the Pacific Ocean On the 
supposed former Pacific continent Mr. Guppy's researches- 
Coral reefs and their history A circum-Pacific land belt 
Evidences of ancient lands westward of Central America On 
faunistic affinities between the Atlantic and Pacific oceans 
The Humboldt current . 295-335 



The geological history of South America The faunal elements of 
the continent Palaeontology Course of development of the 
continent during the Tertiary Era Peripatus and its distribu- 
tion in South America On the origin of Clausilia and its 
relations The South American bear On the tapirs and their 
origin The dwarf deer of South America On the supposed 
existence of former land connections Derivation and history 
of the fossil elephants The river Amazon . . . 336-362 





South American monkeys, their distribution and geological history 
Sloths The birds of South America On fossil mammals and 
Old World affinities The theory of an ancient land connection 
between South America and Africa Evidence derived from 
the distribution of reptiles, amphibians, fishes, mollusks and 
fresh- water crabs Dr. von Ihering's researches Fernando de 
Noronha and St. Helena The fauna of St. Helena and its 
antiquity Tristan da Cunha 363-394 



The fauna of Argentina Prof. Ameghino's discoveries of fossils 
and their importance Fossil edentates and rodents Faunistic 
affinities The llama and its origin Fossil elephants Ancient 
land connections between South and North America Fossil 
plants Floral affinities The oircum-Pacific land belt 
Antarctic problems On the Antarctic continent Affinities 
of the faunas of Patagonia and New Zealand On a Pacific 
continent The marine fauna of Patagonia The Falkland 
islands Antarctic expeditions The pendulation theory 395-435 


1. Map of North America (reproduced from Meiklejohn's oro- 

scopic maps by permission of Messrs. Meiklejohn & Son) . 1 

2. Map of the North Atlantic, showing the supposed conditions 

of land and water during later Pliocene times (reproduced by 
permission of the Council of the Eoyal Irish Academy) . 16 

3. Map of North America, showing the distribution of the Moose 

Deer (reproduced by permission of Mr. Thompson Seton) . 32 

4. Map showing the outlines of the White Mountain Plateau and 

Mount Washington, with an inset Map of North America 
indicating the geographical position of the White Mountains 36 

5. Map of North America showing the original distribution of the 

Bison, the range of the same species after the completion of 
the Union Pacific Railroad and its present range (reproduced 
by permission of Dr. Hornaday) 65 

6. Map of North America, showing the original distribution of the 

Wapiti deer, and its range in 1900 (reproduced by permission 

of Mr. Thompson Seton) ....... 67 

7. Map of the northern parts of the northern hemisphere, 

showing the supposed land connections and the extent of the 

sea in the early part of the Glacial Epoch .... 98 

8. Map of the World, indicating the distribution of the newt 

Spelerpes 138 

9. Map of North America, showing the distribution of the Star- 

nosed Mole (reproduced by permission of Dr. True) . .142 

10. Map of North America, showing the distribution of the Bein- 

deer (reproduced with some modifications by permission of 
Mr. Thompson Seton) 155 

11. Map of North and South America, showing the distribution of 

the snail Helicina 158 

12. Map of the World, showing the distribution of the family of 

fishes Cichlidae (reproduced by permission of Mr. Tate Eegan) 234 



13. Map of the West Indian Area, showing the position of the 

principal islands ......... 262 

14. Map of North and South America, indicating roughly the 

supposed conditions of land and water about the commence- 
ment of the Tertiary Era . . ... . .280 

15. Map of the World, showing the conditions of land and water 

in upper Cretaceous times according to Dr. Ortmann (repro- 
duced with the author's permission) . ... . . 292 

16. Map of North and South America, indicating roughly the 

supposed conditions of land and water about the middle of 

the Tertiary Era . . . . ... .,, .294 

17. Map of the World, showing the conditions of land and water 

during the Eocene Period, according to Dr. von Ihering 
(reproduced with the author's permission) . ,. . .314 

18. Map of South America (reproduced fromMeiklejohn'soroscopic 

maps, by permission of Messrs. Meiklejohn & Son) . . 336 

19. Map indicating the distribution of the snail Clausilia in the 

Mediterranean region, in Africa and America, with an inset 
figure of Clausilia (Nenia) ....... 349 

20. Map of the World, showing the conditions of land and water 

in lower Tertiary times, according to Dr. Ortmann (repro- 
duced with the author's permission) . . .... 382 

21. Map of the World, showing the conditions of land and water 

in late Cretaceous and basal Eocene times, according to 
Professor Osborn (reproduced with the author's permission) 419 

Highland over 1,000 Feet. 
Upland under 1.000 Feet. 
Lowland under 600 Feet. 

. M a p of North America. (Reproduced from Meiklejohn's Oroscopic 
Map, by permission of Messrs. Meiklejohn & Son.) 

[To face p. 1, 

Distribution and Origin 
of Life in America 



GREENLAND geographically belongs to arctic America 
rather than to Europe, and, as it no doubt formed part of the 
land bridge that once connected America and Europe, its past 
history contains chapters of the greatest interest. Scanty 
as the fauna and flora of Greenland are, they afford us many 
a clue as to former changes of land and water which th&t 
country has undergone. Their study enables us also to trace 
the origin of the animals and plants of the neighbouring por- 
tion of continental America, which is one of the objects of 
the present work. 

Greenland is now too well known to need a long description. 
Yet few readers realise the vast size of this stern and 
uninviting country, which covers an area considerably larger 
than the whole of France and Germany together. Three- 
quarters, at least, of this area being completely buried 
under an enormous glacier ice-sheet, or inland ice, only a 
comparatively narrow belt of partly barren rocky ground is 
left along the shore on which animal and plant life is possible. 
The broadest exposed strip of land on the west coast of Green- 
land is about a hundred miles .wide. Here and ther e two 
kinds of willows and the dwarf birch together form scrubby 
low- growing woods, the stems rarely rising more than a few 
feet from the ground. Thickets of alder, white birch and 
dwarf juniper likewise occur, while in sheltered nooks the 

L.A. B 


Lapland rhododendron, many kinds of saxifrages, cinquefoil, 
crowfoot, a willow-herb, a lousewort, the narrow-leaved arnica 
and other flowering plants for a time infuse variety into the 
brownish-grey landscape, giving it quite a gay appearance 
with the many brightly-coloured blossoms. During a couple 
of weeks in the short summer the natives may even have the 
advantage of gathering crow-berries (Empetrum nigrum), 
which grow in great abundance almost everywhere, just as 
they do in boreal America and Europe. "The bilberry and 
cowberry are also noticeable, but they are less common and 
do not seem to be used by the natives, nor is the cloudberry, 
whose fruit rarely ripens. All these are well-known American 
and European plants. The most valuable vegetable, which is 
largely used for food, is a tall angelica. It sometimes attains 
a height of six feet in favourable localities. 

During early Tertiary times many species of conifers and 
several of our common deciduous oaks, poplars, walnuts, 
maples, lindens, magnolias and beeches flourished in Green- 
land. Many of the same trees also lived in Iceland, Spits- 
bergen, Grinnell Land, boreal North America and the New 
Siberian islands. It is evident that the climate all round 
the Arctic Ocean must have been mild. Many causes have 
been assigned for the change from the former temperate 
climatic conditions to the arctic ones now prevailing in the 
eame region. A less restricted circulation of warm ocean 
water into the circumpolar area might, as has been suggested, 
help to bring about such a state. Yet it would not alone be 
sufficient. Sun and light for a longer period than now obtains 
would seem to have been necessary to ripen the wood of the 
trees and bring the fruit to maturity. At any rate, all this 
luxuriant flora must have disappeared from these northern 
regions long before the Ice Age. It is quite possible, and I 
even think it probable, that many of our common arctic and 
alpine plants originated in these remote Tertiary times. I 
cannot adduce any direct geological evidence for such a belief, 
since few of these plants possess properties which would 
enable the remains of their leaves to be preserved in a fossil 
condition. The assumption is founded chiefly on their wide 
geographical distribution. 

As I am dealing mainly with faunistic problems, the 


animals of Greenland have to be considered more in detail. It 
is especially the beasts or mammals that are of importance 
because their distribution and geological history are better 
known than those of the smaller animals. They also enable 
us to readily appreciate the more recent changes in the fauna, 
and the causes which have led to them. 

The most noteworthy and best known of the Greenland 
mammals is the reindeer. The term reindeer (Eangifer 
tarandus) is employed for the sake of convenience in its wide 
sense. In America it has been the custom for many years 
past to distinguish under the names of " barren-ground 
caribou," and " woodland caribou " two strikingly different 
forms which occur on the continent, the term " caribou " 
being probably derived from the Indian word " maccaribo." 
Not only have these two kinds of 'reindeer dissimilar habits, 
they also differ from one another in size, and in the structure 
of their skulls and antlers. More recently, smaller groups of 
reindeer have been discovered, showing distinctive characters 
in their antlers alone. Several of the nine forms thus 
recognised as American are founded upon comparatively 
slight differences. An illustrated account of these various 
American reindeer has been given by Mr. Madison Grant.* 

Mr. Lydekker f holds that they and the Old World forms 
are all referable to one single species, of which he distin- 
guishes six varieties. Whether these differences are to be 
regarded as specific or varietal is a matter of opinion, but, 
as th,e antlers of reindeer are so very variable, it has been 
suggested by Professor Camerano that the structure of the 
skull might be utilised to greater advantage in discriminating 
between the various varieties or species. In his researches 
on the Spitsbergen form, he argues that there are three 
species, viz., the reindeer of Spitsbergen (R. spitsbergensis). 
the one inhabiting Greenland, Norway, Lapland and part of 
Siberia (R. tarandus), and lastly the American reindeer (R. 
caribou) . Among the last two species he distinguishes again 
between the reindeer possessing cylindrical (cylindricornis) 
and those having compressed (compressicornis) antlers. 

It is worthy of note that the reindeer of Scandinavia and 

* Grant, Madison, " The Caribou." 

f Lydekker, E., " The Deer of all Lands," p. 33. 

B 2 


Greenland constitute a single variety, according to Professor 
Camerano,* while Mr. Lydekker separates the Scandinavian 
from the Greenland race. Following Professor Camerano's 
example, Dr. Lonnberg f prefers the nasal bones of the skull, 
to the antlers for specific diagnosis. He unites the Greenland, 
Spitsbergen and Scandinavian forms with the American 
barren-ground reindeer in one group, characterised by their 
flat nasals. ^The Finland and Siberian reindeer he holds to 
be quite distinct races, particularly the former, which, with 
its elevated nasal bones and somewhat flattened antlers, ap- 
proaches the American woodland caribou. Even Robert 
Brown J long ago commented on the likeness of the Green- 
land and European reindeer, remarking that some reindeer 
horns from Greenland could not be distinguished from 
European ones. Since Professors Camerano and Lonnberg, 
moreover, take both skull characters and antlers into con- 
sideration, we may assume that this view of the affinity of the 
Greenland to the north-west European race is likely to be the 
correct one. 

An anonymous writer in " Natural Science," (p. 358, 
1899) gave it as his opinion that there are no grounds for 
deciding whether the Old or the New World was the original 
home of the reindeer. No doubt he meant to convey that we 
know nothing of the ancestry of the reindeer. I ventured, 
on the contrary, to put forward the theory that the reindeer 
had originated in the Polar area, beyond the confines of the 
European Continent, while Professor Schlosser || argued that 
it was evidently of American origin. Neither of these views 
is supported by positive geological evidence ; nevertheless 
I cannot agree with the above critic that we possess no grounds 
for deciding the place of origin of the reindeer, as the results 
of Professor Camerano's and Professor Lonnberg's cranial 
researches seem to indicate the existence of a closer rela- 
tionship between the west European and the Greenland and 

* Camerano, L., " Renne delle Isole Spitzberge," p. 167. 
t Lonnberg, Einar, " Taxonomic Notes about Reindeer," p. 10. 
| Brown, Robert, "Mammals of Greenland," p. 352. 
Scharff, R. F., " History of European Fauna," p. 155. 
|| Schlosser, M., M Beziehungen d. ausgestorben. Siiugethier-faunen," 
p. 617. 


arctic American varieties than between the west European 
and the Siberian reindeer. The assumption that the species 
originated in Greenland or arctic America and thence spread 
along old land connections east and west to Europe arid Asia, 
is, therefore, not quite so unjustified ias Dr. Stejneger * seems 
to think. 

But we must not forget the possibility of the reindeer 
having originated in the Old World. Dr. Brown f argues 
that the barren-ground variety spread to America from 
western Europe by way of a Spitsbergen-Greenland land 
connection, while the other traversed Siberia, and crossed to 
the New World by Bering Strait. In the more southern 
parts of its range in America, new varieties may have 
arisen, perhaps owing to changes in the natural sur- 
roundings. Long cylindrical antlers were no doubt a dis- 
advantage to reindeer in wooded districts, hence the beam 
became reduced in length and increased in width in what is 
known in America as the " woodland caribou." Some of these 
varieties, or species as many American naturalists choose to 
call them, intergrade to such an extent as to-be difficult to dis- 
criminate from one another. Thus Rangifer stonei is a form 
of barren-ground reindeer living in Alaska, while Rangifer 
osborni is a woodland form. Dr. Allen, who first described 
them, considers them as well-marked species, while Mr. 
Osgood,J during his travels in Alaska, finds that the differ- 
ences between the two are all relative, that they are excessively 
variable and rather intangible. The mountain caribou (Ran- 
gifer montanus), originally described by Mr. Thompson 
Seton, is considered by Dr. Allen to be allied to the wood- 
land form, but distinguished from it in colour, size and shape 
of antlers. Mr. Seton, || who has published a most instructive 
map of the geographical distribution of the caribou in North 
America, which has furnished me with the materials for the 
map given in this work (Fig. 10), thinks that in future the 

* Stejneger, L., " ScharfE's History of the European Fauna," 
p. 112. 

f Brown, A. E., " North American Big Game," p. 87. 
J Osgood, W. H., " North American Fauna," 33, p. 17. 
Allen, J. A., "Mountain Caribou," pp. 8 & 9. 
|| Seton, E, Thompson, "Northern Animals, I.," p. 192, 


ten species now recognised in America will be reduced to 
four well-marked species. 

How long the caribou or reindeer has been in existence 
we do not know. Its remains have been discovered both in 
Europe and America in deposits believed to have been laid 
down during the Pleistocene Period, and as these occur far 
to the south of its present range, it has invariably "been 
assumed that the species was compelled, owing to unfavour- 
able climatic changes to abandon its more northerly habitat. 
When the climate became once more suitable to its require- 
ments, the reindeer is supposed to have returned to its, original 
home. This idea suggests that the reindeer originated in 
pre-Glacial times, and this view is, in my opinion, supported 
by the evidence of its occurrence, conjointly with the hyaena, 
in Irish cave deposits.* 

The whole subject of the influence of the Glacial Epoch 
on animals and plants will be dealt with later on, and need 
not be considered here. The fact of the occurrence of un- 
doubted reindeer remains far to the south of its present 
range certainly requires an explanation, and this is more 
easily given in conjunction with other facts to be stated 
in this chapter. While .the reindeer still lives in Europe no 
/further south than the fifty-second degree of latitude, in 
America it is found no less than seven degrees further south. 
In former times it inhabited Oregon and Kentucky. In the 
east it came down to the neighbourhood of the present site 
of New York City, whereas in Europe it advanced as far as 
Mentone on the shores of the Mediterranean, and penetrated 
to the north of Spain, i.e., to the latitudes of the thirty-eighth 
degree in America and of the forty-third degree in Europe.f 

The musk ox (Ovibos moschatus) is another even more 
arctic mammal than the reindeer. So called from the musky 
odour of its flesh, this species was believed to be more closely 
allied to the sheep than to the ox, which it resembles more in 
gize. In its geographical distribution it differs strikingly 
from that of the reindeer in so far as it is now quite confined 
to Greenland and arctic North America. It no longer inhabits 

* Scharff, R F., " European Animals," p. 112. 
t Brauer, A., " Die arktische Subregion." 


the Old World. Yet in Europe it penetrated as far as England 
and France in Pleistocene times, having apparently reached 
these countries from an eastward direction. While the rein- 
deer occurs even in southern Greenland, the musk ox 
frequents only the north and east of that country. 

According to Dr. Conwentz, it has only wandered east- 
ward from the north coast of Greenland in recent times. 
At any rate, Professor Nathorst alleges that Scoresby, in the 
year 1822, did not observe it in the neighbourhood of the 
Sound which bears his name, though it is now found there. 
Nevertheless, the discovery of some very ancient skulls 
between the seventieth and seventy-fifth degrees of latitude 
on the same coast make it probable that the musk ox lived 
in that region in more remo'te times, afterwards becoming 
extinct only to advance once more to its former habitat in: 
recent times.* 

The musk ox to most observers has appeared to be subject to 
little variation. Mr. Lydekker, however, described a form 
from eastern Greenland (Ovibos moschatus wardi) with a 
whitish face-patch and lighter general colouration, which 
Dr. Allen f has since raised to specific rank, showing that it 
also differed from the type in the structure of the horns and 
hoofs. The latter likewise pointed out that it was not con- 
fined to Greenland, but inhabited also Ellesmere and Grinnell 

Another form of musk ox, an almost black one, from the 
far north of Canada, was first noticed and described by Mr. 
Elliot. A renewed examination of musk ox skins and skulls 
was recently undertaken by Dr. KowarzikJ with rather re- 
markable results. He recognises two strikingly distinct 
groups of "musk oxen, a western and an eastern group, 
characterised by very important osteological and external 
features, which had apparently escaped the attention of pre- 
vious observers. In a private letter he admitted to me 
that the musk ox must have originated in North America, 
Bootherium being probably its direct ancestor. Yet he is 
evidently inclined to believe that, after spreading by way of 

* Conwentz, " Yerbreitung d. Moschusochsen," pp. 429 431. 
t Allen, J. A., " Musk-Oxen of Arctic America," p. 76. 
I Kowarzik, E., " Der Moschusochs," p. 120. 


Asia to Europe, it became extinct in America as well as in 

Part of one section of the old musk ox stock returned to 
North America by way of Asia, while another travelled by 
means of a north Atlantic land connection to southern Green- 
land, gradually finding its way along the eastern and western 
coasts to northern Greenland. Meanwhile the musk ox had 
entered Alaska again from Asia and crossed the Mackenzie 
Eiver into arctic Canada, where some of the peculiar forms 

Dr. Kowarzik promises us further contributions to this 
interesting problem from a palaeontological point of view. 
These will be of particular value in elucidating the question 
of the relationship between the European and the Greenland 
forms of the musk ox. I agree in so far with Dr. Kowarzik, 
as I feel convinced that the musk ox originated in North 
America in pre-Glacial times, but I differ in most of his other 
views. If land connections existed between North America 
and Asia, and between North America and Europe, they must 
have persisted through Pliocene to early Pleistocene times. 
I am of opinion, therefore, that the musk ox spread east- 
ward to Europe and westward to Asia as long as land 
bridges enabled it to do so. It is quite possible that the 
Asiatic section subsequently travelled westward to Europe, 
but neither of them succeeded in returning to their native 
lajid. The musk oxen we find in Greenland and arctic 
America probably survived the Ice Age. where they now 
live. This opinion has the support of Dr. F. Wright * and 
Professor Upham as the result of their special study of the 
geological history of Greenland. 

Not long ago the musk ox was still found in Alaska. Now 
it has been quite exterminated west of the Mackenzie Kiver. 
As Dr. Allen remarks, the genus Ovibos is a declining type, 
which has attained its last stronghold in the arctic barren 
grounds. Wherever it is within the grasp of man it will 
soon be a creature of the past. 

We possess no undoubted evidence of the existing species 
of musk ox having ever penetrated as far south as the United 
Sta;tes, nevertheless it is quite possible that the fragmentary 

* Wright, F. F., and Warren Upham, " Greenland Icefields," p. 332. 


remains which have been discovered in caves and other super- 
ficial (Pleistocene) deposits in Pennsylvania, Kentucky, Mis- 
souri, Iowa and Ohio may have belonged to varieties of Ovibos 
moschatus rather than to extinct species of musk oxen. 

There are two other arctic mammals which are of special 
interest to us, since both of them inhabit Greenland. They 
are the arctic hare (Lepus variabilis = L. timidus) and the arctic 
lemming (Dicrostonyx torqnatus = Cuniculus torquatus). 

Like the reindeer, the arctic hare is subject to great varia- 
tion over the vast area which it inhabits, and some of these 
varieties or races are now recognised by many zoologists as 
species. Mr. R. Brown * long ago felt the difficulty, nay, 
almost impossibility, in discriminating between the Green- 
land hare and the arctic or varying hare of Europe. The same 
embarrassment has confronted many subsequent authors. In 
American-he Greenland hare is now generally considered as 
specifically distinct from the arctic American hare, Mr. 
Rhoads f having pointed out some important features in its 
incisor teeth, while Mr. Nelson J finds that its excessively 
heavy wool -like coat of fur, the long claws and tapering upper 
mandible sufficiently distinguish it from its relatives on the 
American Continent to give it the rank of separate species. 
However, as Mr. Rhoads aptly remarks (p. 356) the peculiar 
dentition, long fur, and long claws are due partly to the 
scantiness and character of plant life, and partly to the depth 
and long continuance of the snow in Greenland. These fea- 
tures seem to have gradually impressed themselves on the 
Greenland hare in modifying it in the direction indicated. 
In Major Barrett-Hamilton's opinion the chief differences 
between the European arctic hares and the Greenland form, 
from an external point of view, are only of a su|b- specific 
value, while Dr. Winge || does not specifically distinguish the 
American or European arctic hares from the Greenland one. 
The same problem of affinity has recently been investigated 
from a new standpoint by Mr. Hinton. His researches have 

* Brown, R., " Mammals of Greenland," p. 351. 

t Rhoads, S. N., u Polar Hares of North America," p. 368. 

| Nelson, E. W., " Rabbits of North America," p. 68. 

Barrett-Hamilton, G. E. H., " The Variable Hare," p. 92. 

|| Winge, H., " Gronlands Pattedyr," p. 376. 


led him to a comparison of the fossil hare remains from the 
ossiferous fissures of Ightham in England with the recent 
polar or arctic hares as a whole. He has been exceptionally 
fortunate in disposing of large osteological collections, and 
his conclusions are of particular interest and of much value 
in elucidating some important zoogeographioal problems. 
Dwelling upon the close affinity existing between the English 
fossil and the Irish living arctic hare, he regards the latter 
as the direct descendant of the former. I should not have 
mentioned these particulars except that they give rise to far- 
reaching deductions. From the fact of the fossil English and 
recent Irish forms being the least specialized members of the 
whole group of arctic hares, Mr. Hinton concludes that it is 
highly improbable that the group can have had a boreal 

That the Irish arctic hare has always lived in Ireland under 
temperate conditions is an opinion which I expressed long 
ago in my works on the European fauna, and in this view Mr. 
Hinton concurs. But I also believe, as I shall endeavour 
to demonstrate later on, that even southern Greenland and 
all the lands surrounding the north Atlantic had a milder 
climate during the Pleistocene Period. It is principally the 
climate of arctic America and Greenland, I think, that be- 
came much more unfavourable within recent geological times, 
while that of the British Isles has undergone comparatively 
little alteration. Meanwhile specialization among the animals 
constituting the Greenland fauna probably proceeded at a 
more rapid rate than in Ireland, where the hare had no need 
to become adapted to different 'conditions of food 'and tempera- 
ture. In spite of Mr. Hinton 's -argument, I still believe in the 
arctic origin of the group in Pliocene times, mainly because I 
do not admit that we have any evidence for the assumption 
that Ireland was connected with Great Britain during or 
after the Pleistocene Period. Mr. Hinton thus differs from 
me in regarding Central Asia as the centre of origin of the 
arctic hares in Pleistocene times. 

I mentioned that the banded lemming (Dicrostonyx tor- 
quatus) was found in Greenland. Of late years it has been 

* Hinton, M. A. C., " The fossil hare of Ightham," pp. 263264. 


customary to distinguish the New World form (D. hud- 
sonius) from the Old World banded lemming (D. torquatus), 
but they are very closely allied. Several varieties of the 
former inhabit the mainland and islands of arctic America, 
including the north and east coasts of Greenland. The 
whole genus Dicrostonyx (Myodes), is confined to the Arctic 
Eegions. As in the case of the reindeer and other arctic 
species, we possess fossil testimony of a former southern 
extension of the range of the banded lemming in Europe.* 
It occurred in Central Europe, and also in England and 
Ireland, yet, as far as we know, it never penetrated into the 
United States in Pleistocene times. That the banded lemming 
is not a recent immigrant to Greenland, but has persisted 
there from pre-Glacial times seems to be indicated by the 
fact that Colonel Feilden f discovered its remains, with those 
of the reindeer and musk-ox, in post-Tertiary (Pleistocene) 
deposits from sea-level to an altitude .of 1,000 feet in northern 

There are a couple of other mammals in Greenland, viz : 
the arctic wolf and the arctic fox, which need not be specially 
considered here. Nevertheless, a significant factor in connec- 
tion with one of these carnivores has been pointed out by 
Major Barrett-Hamilton and Mr. Bonhote.J It is that the 
arctic fox (Vulpes lagopus) of Spitsbergen, Novaya Zemlya 
and Iceland agrees with that from eastern Greenland, form- 
ing a variety distinct from those of the European and 
American mainlands. 

All the mammals alluded to as inhabiting Greenland, live 
also in Europe in identical or closely allied forms, or did so in 
former times. Hence it is permissible to argue that a land- 
connection once bridged over the intervening ocean. The 
affinity between Greenland and arctic America in some 
respects is even closer than that between Greenland and 
Europe. Only the narrow Davis Strait and the still narrower 
Kennedy Channel separate the two countries. Another 

* Scharff, E. F., G. Coffey, and others, " Caves of Kesh," p. 196. 

t Feilden, H. W., and C. E. de Eance, "Geology of Arctic Coasts," 
p. 566. 

t Barrett-Hamilton, G. E. H., and I. L. Bonhote, " Sub-species of 
Arctic Fox," p. 288. 


sound of inconsiderable width, Bering Strait, divides arctic 
America from the Asiatic Continent. We can easily imagine 
what a comparatively small change in the configuration of 
land and water would thus bring Greenland into direct touch 
with Asia, by way of arctic America (compare Fig. 1). 

It has actually been suggested by several writers that the 
faunal relationship which exists between Greenland and 
Europe is due to former land bridges across Bering Strait 
and Davis Strait, facilitating the passages of animals from 
Europe by way of Asia. That such land connections really 
existed in recent geological times, I do not doubt, and they 
will be further described in this and one of my next chapters. 
Still, their existence does not preclude the possibility of 
Greenland having likewise been connected by land directly 
with Europe. A close relationship exists between some of the 
western European and Greenland mammals. In some in- 
stances that affinity is no longer apparent in the recent 
mammalian fauna, but can be demonstrated to have once been 
a prominent feature. The extinct Irish reindeer for example, 
clearly indicates, in the character of its skull and antlers, 
that it was closely related to the present Greenland and arctic 
American races of reindeer, thus suggesting the former 
existence of some more direct means of transit from Green- 
land to Ireland, or vice versa, than by the circuitous route 
across Europe, Asia and arctic America. Many authorities, 
indeed, have on other grounds insisted upon the former 
presence of a land bridge uniting Scotland, the Faroes, 
Iceland and Greenland, with America. 

The story of the extermination of the indigenous reindeer 
in Iceland in the twelfth century is too vague to be of much 
use in connection with these researches. It appears certain, 
at any rate, that no reindeer had inhabited Iceland for some 
hundreds of years before the end of the eighteenth century, 
when the ancestors of the present stock were brought over 
from Finmark. The assumption of the existence of such 
a land connection as that referred to, in Pliocene, and perhaps 
early Pleistocene times, or, in other words, just before and 
during the beginning of the Glacial Epoch, is generally based 
upon other ground than the distribution of mammals. I have 
recently collected the various sources of evidence in favour 


of the theory from the oceanographical, geological and 
biological standpoints.* 

A second land bridge apparently united northern Greenland 
and Lapland at about the same time. This I described on a 
previous occasion. It will not be necessary for me to repeat 
all the arguments I advanced for and against the hypothesis, 
and I must content myself with a short statement of the main 
facts. The theory of the Greenland-Iceland- Scotland land 
bridge (Fig. 2) had been put forward by quite a number of 
authorities on independent grounds. Yet while some maintain 
that it merely existed in early Tertiary times, others contend 
that it could .only have been made use of by the .members of our 
present fauna and flora after the Glacial Epoch, because the 
latter had not yet come into existence before that time. 

One of the most important facts in favour of the existence of 
this land bridge is presented, in my opinion, by the geographi- 
cal distribution of the land snail known as Helix hortensis. 
I have quoted many other instances in my paper on this 
subject, but this no doubt affords the most striking biological 
support for the belief in a land connection between Scotland, 
Iceland, Greenland and Labrador. 

Helix hortensis is a typically west European species, being 
quite unknown in the eastern parts of Europe or in Asia. 
Beyond the mainland of Europe we find it in Great Britain, 
in Ireland, the Shetland Islands, the Faroes, Iceland, in 
southern Greenland, Labrador, the islands off the north-east 
coast of North America, and part of the opposite mainland. 

Biologists are often too ready in invoking human agency 
when endeavouring to explain the occurrence of certain 
common European species in unexpected localities abroad. 
When this snail was first discovered in North America, its 
presence there was universally attributed to the action of man, 
and was often cited as a good instance of the facility with 
which terrestrial mollusks are introduced into foreign coun- 
tries and become established there. Until the year 1864 no 
other theory was even thought of. During that year, however, 

* Scharff, E. F., " On a former Land Bridge between Europe and 
North America." 


Professor E. Morse first discovered the shell of this snail 
among the contents of ancient ' k kitchen-middens," those 
peculiar refuse heaps of primitive man, on some of the islands 
off the east coast of Maine. He pondered over this singular 
mystery and finally came to the conclusion that the snail must 
have slowly wandered, during a long series of centuries, from 
the Old World to the New, by means of an ancient North 
Atlantic land bridge. Mr. Johnson,* to whose instructive 
paper I am indebted for this information, states that Dr. 
Binney and Professor Cockerell concurred in Professor 
Morse's opinion. He also informs us that the Rev. Winkley 
and he were of opinion that the arrival of the snail Helix 
hortensis in America must have taken place before the advent 
of the Glacial Epoch. 

This theory, as can be imagined, was by no means gene- 
rally accepted in America. All doubts, however, as to 
the claim of Helix hortensis being an indigenous American 
species are now set at rest, through the discovery by 
Dr. Dall f of the shell of this snail in undoubtedly Pleisto- 
cene deposits in the State of Maine. Some naturalists might 
still be inclined to urge that greater facilities for occasional 
transport across the Atlantic may have existed in those remote 
times than at present, and that the argument in support of 
a land bridge is not convincing. No evidence, however, in 
favour of an ocean current from Europe to North America in 
Pleistocene times has as yet been brought forward, while the 
conviction in the former presence of a land connection between 
north-eastern America and north-western Europe is based 
upon other biological observations. From a geological point 
of view we can scarcely hope to be able to determine the period 
or periods during which this land bridge existed. The bathy- 
metrical features of the north Atlantic, according to Pro- 
fessor Hull,J Dr. Spencer and Dr. Nansen,|| point to a pre- 
Glacial elevation of the land in northern latitudes. Dr. 

* Johnson, C. W., " Distribution of Helix hortensis," p. 73. 

t Dall, W. H., " Land and Freshwater Mollusks of Alaska," p. 20. 

J Hull, E., " Submerged Terraces and .River Valleys." 

Spencer, J. W., " Submarine Valleys," p. 224. 

|| Nansen, F., "North Polar Expedition," p. 192. 


Wright * and Mr. Upham, two well-known authorities on 
glacial phenomena expressed the view that the northern lands 
must have been gradually elevated in Pliocene times, be- 
coming continuous before the Ice Age. Further particulars 
on this subject are contained in Dr. Spencer's articlef on 
high continental elevation. 

Sir Henry Howorth opposes this view, urging that Green- 
land, Scandinavia, and North America were all at a much 
lower level in so-called Glacial times than they are now. If 
this be so, then the epeirogenic theory has no base to stand 
upon. I do not think that Sir Henry Howorth's J statement 
is applicable to southern Greenland or eastern North America. 
At any rate, it is not contended by anyone that the high level 
condition of these countries persisted during the whole of 
the Ice Age. 

The theory that the Ice Age or Glacial Epoch was brought 
about mainly by the closing of the Arctic Ocean to the influence 
of the Gulf Stream is a very tempting one. The temperature 
of the Atlantic Ocean would under such conditions have been 
higher than it is now, because its heat would not have been 
modified by cold arctic currents, as it is at present. Southern 
Greenland, Iceland, and the lands on both sides of the 
Atlantic should also have had a more favourable climate than 
obtains under existing circumstances, since the warm ocean 
would have had considerable influence upon their climate for 
a certain distance inland. There is evidence, on the European 
side of the Atlantic, that, at a not very distant geological 
period, and presumably at a time when the coasts of France, 
south-west of England and south of Ireland were still united, 
the southern fauna and flora crept steadily northward along 
the ancient shore-line. That this did not take place in very 
recent times is evidenced by the fact that the southern marine 
shore forms of mollusks, crustaceans and other invertebrates 
have a somewhat discontinuous range on the west coast pf 
the British Islands, and do not occur in the English Channel 
or in the southern part of the North Sea. 

* Wright, G. F., and Warren Upham, " Greenland Icefields " p. 331, 
f Spencer, J. W., " High Continental Elevation." 
t Howorth, H. H., "Ice or Water," Vol. I., p. 136. 


Hence they evidently wandered northward under different 
geographical conditions than exist at present, and only main- 
tain a slender hold in isolated localities, where they form 
relict colonies. I assume that this northward advance of the 
marine fauna took place in late Pliocene times when the 
northern Atlantic was closed, and the temperature of the 
ocean raised. As Dr. Dall wrote to me " the Pliocene all over 
the northern hemisphere was a period of warmer sea water 
than the Miocene or Pleistocene." My attention was also 
recently drawn by Professor Morse* to the fact that even 
littoral European species have extended their range across the 
ocean to North America. As long ago as 1855 he received 
specimens of the common European " Periwinkle " (Littorina 
litorea) from Chaleur Bay in the Gulf of St. Lawrence. Later 
on this 'mollusk was reported also to occur on the coast of 
Nova Scotia and Labrador. It is steadily advancing now in a 
southward direction, while it is becoming scarce in the north. 

Dr. Ash worth f tells me that the lugworm (Arenicola 
marina) which is common on the west coast of Europe is also 
met with on the coasts of the Shetland islands, the Faroes, 
Iceland, Greenland, Labrador and Nova Scotia. It is absent 
from the coasts of the Pacific Ocean. Hence he concludes 
that a continuous shore-line formerly existed between the two 
areas on each side of the Atlantic Ocean in which this worm 
is now found. Dr. Ashworth believes that the extensive 
stretch; of ocean at present existing between Europe and 
America forms just as much a barrier to the dispersal of 
the littoral lugworm as it does to that of the terrestrial Helix 
hortensis just alluded to. 

Commenting on this occurrence of European littoral marine 
mollusks on the north-east coast of North America, Mr. B. B. 
Woodward writes to me that no other theory than that of the 
previous existence and subsequent rupture of a land bridge I 
in the direction indicated can satisfactorily account for the/ 
present disjointed distribution of the two divisions of the/ 
boreal fauna. 

It was assumed by Professor NathorstJ that a portion of 
* Morse, E. S., "Dispersion of certain Mollusks," p. 8. 
Ashworth, J. H., Catalogue of Chaetopoda part I. 
| Nathorst, A. G., " Pflanzengeographie der Vorzeit," p. 267. 



o> * 



.-H O 



O jj 

"3 '3 



the arctic flora had originated in Greenland in pre-Glacial 
times and had been scattered east and west across the exist- 
ing land bridges on the advent of the Glacial Epoch, during 
which the maintenance of life was no longer possible in that 

Sir Joseph Hooker * long ago expressed the opinion that, 
although many Greenland plants were possibly destroyed 
during the Ice Age, the existing remnant of a much richer 
flora had survived in the southern parts of the country, whence 
it subsequently spread northward again. 

One of the biological arguments I adduced in favour of a 
former north Atlantic land bridge was derived from the well- 
known fact that deposits of dead marine shallow-water species 
had been dredged in deep water in various localities such as 
Eockall Bank and off the coast of Iceland. This peculiar 
circumstance has been applied by several authorities in sup- 
port of the theory of a gradual sinking of the land, the 
shallow- water species having thus been moved to a position 
in which they are no longer able to live. It is this part of 
the biological argument on the land bridge theory which has 
received most of the adverse criticism. Dr. Johansen,f for 
instance, pointed out that the evidence derived from the 
marine shallow-water shells is untrustworthy, because their 
presence in great depths in the northern Atlantic is not due 
to a sinking of the land, but mainly to various casual or 
accidental activities of transport. The theory of the north 
Atlantic land bridge, in so far as it is founded upon the occur- 
rence of shallow-water marine shells at great depths, does 
not, therefore, meet with his approval. 

Similarly, Dr. Appell^f J insists that the most recent dis- 
coveries on the " Faroe Bank " are of considerable zoogeo- 
Igraphical significance, since they are opposed in some 
measure to the theory of the sinking of the land. He informs 
us that among the thick layer of dead shallow shells found 
on the Faroe Bank living specimens of several of the species 
were met with. 

* Hooker, J. D., " Distribution of Arctic Plants," pp. 252255. 
t Johansen, A. 0., "On the Sinking of Sea-beds," p. 403. 
J Appeltyf. A., " Norwegian Fisheries," Vol. II., pp. 8389. 

L.A. C 


No doubt this discovery provides an argument for the op- 
ponents of the land bridge theory, yet we know how adaptive 
certain species are to a change of conditions, and how long 
they can maintain themselves under adverse circumstances. 
I am not, therefore, disposed to attach too much importance 
to Dr. Appell^f s discovery. In any case, the land bridge theory 
is not dependent on the evidence alluded to. 

Dr. Spethmann,* on the other hand, reiterates what we 
already know, that from a purely geological standpoint there 
are no positive proofs in favour of a former land bridge 
between Europe and Greenland. 

These seem to be the principal arguments that have been 
advanced in opposition to the land bridge theory, and they 
are, in my opinion, not very formidable ones. 

The question of the supposed survival of plants through 
the Ice Age in Greenland is closely connected with that of 
the land bridge alluded to. Whether any plants survived, and 
what proportion of those previously existing, largely depends 
on the nature of the Ice Age or Glacial Epoch and on the 
former extension of the glaciers in Greenland. Professor 
James Geikie f maintains that it is a fair assumption that 
the ice of Greenland in Glacial times completely buried the 
land and, perhaps, protruded beyond it. It has recently been 
very clearly demonstrated, however, by the leader of the 
German Greenland Expedition, Dr. E. von Drygalski,J that 
the strip of land now free from ice on the west coast of Green- 
land has never been entirely invaded by glaciers. No doubt 
it can be proved, he remarks, that the ice in past times had 
a greater extension. All the same, glaciers never reached the 
cliffs and rock pinnacles which abound on all parts of the 
coast land of Greenland. 

No special reason can be adduced, therefore, why the pre- 
sent flora of Greenland should not have survived the Ice Age 
in that country, particularly as we have some grounds for the 
belief that the land in parts of the Arctic Regions then stood 
higher than it does now, and that consequently more land was 

* Spethmann, H., " Aufbau d. Insel Island," p. 8 

t Geikie, J., " The Great Ice Age," p. 736. 

| Drygalski, E. von, " Grdnland Expedition," Vol. I., p. 385. 


available for plant life. Indeed, Professor Vanhoffen * who 
describes the plants and animals observed during the expedi- 
tion, adopts this attitude. He not only believes in the survival 
of the flora of Greenland through the Ice Age, but he also 
argues that the great mass of the fauna is indigenous to the 

Though he does not deny the possibility of organisms 
being accidentally carried by birds, and other occasional 
means of transport, he protests against the assumption that 
the fauna of Greenland, as a whole, owes its origin to such 
a mode of conveyance. Dr. Brehmf quite concurs in this 
view. He is convinced that the fauna of Greenland is 

It is very generally believed that the climate of Greenland 
was much colder during the Glacial Epoch than it is now, 
although we possess no direct evidence that it was so. A 
mild climate in South Greenland during the existence of the 
land bridge is implied by the fact that a number of Euro- 
pean species found in north-western North America, and 
which no doubt travelled by way of Greenland, have since 
become extinct in the latter country. 

Greenland must, therefore, have passed through a phase 
during which existence became impossible for these species. 
Yet, for all we know, it may be only quite recently that the 
climate of Greenland has grown so inhospitable. It has been 
suggested by Professor Whitney, and more recently by Sir 
Henry Howorth, that the Arctic Kegions enjoyed temperate 
conditions during the Glacial Epoch and are only now passing 
through the more severe post-Glacial stage. In view of the 
accumulation of evidence pointing to a southward advance 
of the Arctic fauna and flora in Pleistocene times, such a sug- 
gestion may seem contrary to biological evidence. I shall 
endeavour to show, however, in the next few chapters that 
we possess a good deal of valuable testimony, principally of 
a biological nature, in support of Professor Whitney's and Sir 
Henry Howorth's contention. 

With regard to the species which I believe to have made use 

* Vanhoffen, E., " Gronland Expedition," Vol. II., p. 174. 

t Brehm, V., " Entomostraken d. Danmark Expedition," p. 316. 

c 2 


of the ancient north Atlantic land connection, and are sup- 
posed to have subsequently become extinct on the intermediate 
stations between the two continents, particular attention may 
be drawn to the range of the " running beetles " of the genus 
Carabus. They are of great value in aiding us to solve pro- 
blems of this nature, because, being usually found under 
stones and clods of earth, they are not liable to oc<?asional 
transport by floods. Being wingless they cannot be carried 
to distant lands by winds ; and lacking any kind of means by 
which they might become attached to a mammal or bird they 
would not be conveyed in such an accidental manner from one 
locality to another. The great importance of the species of 
Carabus has been recognised, and their distribution brought 
to bear upon zoogeographical problems by Mr. Born.* He cites 
two of the species, viz., Carabus catenulatus and Carabus 
nemoralis, as evidences of a former land bridge between 
northern Europe and North America, although they no longer 
occur in Iceland or in Greenland. Both these running beetles 
are typically European species, being quite absent from Asia. 
The conspicuously ornamental Carabus memoralis is confined 
in North America, to Nova Scotia and New Brunswick. Hence 
it somewhat agrees in its American range with that of Helix 
hortensis. The other species of Carabus has a wider dis- 
tribution in boreal North America. 

Such instances lead us to believe, therefore, that the faunas 
of Greenland and Iceland were richer in pre-Glacial times 
than at present. They are certainly suggestive also of a sur- 
vival of species having taken place through the Ice Age within 
the glaciated area of North America. We possess no evidence 
that these beetles and the snail Helix hortensis, and many 
other animals belonging to the same group of European in- 
vaders, were pushed south during Pleistocene times into the 
United States, and that they then regained their former 
northern habitat, after having become extinct again in their 
more southern stations. 

The extinction of a large part of the former beetle fauna 
of Greenland may be inferred from the fact that Greenland 
only possesses forty-one species of beetles, while there are 

* Born, P., " Zoogeographisch-carabologische Studien," p. 8. 


ninety-one in Iceland.* Of the more hardy spiders the per- 
centage of survival is very different, for there are fifty-three 
species in Greenland and only twenty-four in Iceland. "j* 

The theory of the survival of species in Greenland may 
be tested by some other examples. Besides Helix hor- 
rfcensis about a dozen other kinds of land and fresh- 
water mollusks inhabit the country. J Eight of these either 
have their centre of distribution in Greenland or are 
quite peculiar to the country. These are Planorbis nathorsti, 
P. arcticus, Limnaea vahli, L. holbolli, Succinea groen- 
landica, Vitrina angelicae, Conulus fabricii and Pupa hoppii. 
The two species of Planorbis are also known from Labrador ; 
Succinea groenlandica occurs in Iceland ; P. arcticus has 
been met with in Scandinavia, Finland and Siberia. The 
first of the Limnaeas ranges from Greenland to Alaska, 
the other is peculiar to Greenland. The latter, however, 
is replaced in boreal North America by the closely-allied 
Limnaea retusa. Vitrina angelicae is not found in America, 
but occurs in Iceland and Norway. Conulus fabricii is pro- 
bably only a variety of the common Conulus f ulvus of northern 
Europe, Asia and America, while Pupa hoppii is confined to 
Greenland. With the single exception of Planorbis arcticus 
all these species live at present well within the glaciated area, 
that is to say, within that portion of the northern regions 
supposed to have been either wholly or partially buried by 
ice during the Glacial Epoch. As none but Planorbis arcticus 
have ever been found fossil outside that area, we may assume 
with some justification that most of them originated in Green- 
land, and that all, at any rate, survived the Ice Age in that 
country. Planorbis arcticus, as Mr. Kennard informs me, 
has been taken in Pleistocene deposits in Denmark and in 
the south of England. (Compare also Kennard and Wood- 
ward's paper.) 

It is more difficult to demonstrate that butterflies and moths 

* Poppius, B., " Coleopteren des Arktischen Gebietes/' p. 428. 
t Strand, E., " Arktische Araneae," p. 436. 

t Morch, 0. A. L., " Land and Freshwater Mollusca of Greenland." 
Kennard, A. S., and B. B. Woodward, " Extinct post-pliocene 
Mollusca of Southern England," p. 5. 


have survived the Glacial Epoch in Greenland, yet several 
authors who have written on the lepidopterous fauna of that 
'country maintain that the theory of a former land bridge 
with Europe and North America is quite consistent with the 
facts of distribution. Mr. Petersen,* for example, concludes 
from his studies that land connections in high northern lati- 
tudes enabled certain arctic butterflies and moths to spread 
from a polar centre to Europe and North America. After 
alluding to a large number of moths common to Canada and 
Europe, Mr. Girard f emphasises the significance of this fact 
as indicating the former existence of a land bridge between 
North America, Greenland and, Europe. 

Although no butterflies occur in Iceland, Coleas nastes, 
Argynnis chariclea and A. polaris are known even from the 
barest and most exposed districts of Greenland, where 
nothing but Dryas octopetala and some slender grasses grow. 
These species have a wide range in boreal America, Europe 
and Asia, but of the nine butterflies known from the arctic 
American archipelago five are European, while only two are 
met with in Asia, according to Dr. Pagenstecher. Many 
striking examples, showing the relationship between Green- 
land and the neighbouring continents occur among the moths, 
especially among the Noctuidae. Anarta melanopa inhabits 
only Colorado, the White Mountains, Labrador, Scandinavia, 
Scotland and the Alps. Anarta leucocycla and A. lapponica 
are found in Labrador, Greenland and Scandinavia. Mr. 
Grote J alludes to no less than twenty-eight other Noctuids 
that are common to Europe and North America, even ex- 
cluding those met with in Labrador or circumpolar species. 
All these facts tell in favour of the view I have endeavoured 
to elucidate. 

No matter what group of terrestrial invertebrates we 
choose, similar close relationships may be discovered between 
American and European species, which cannot be explained 
by the assumption of a former land connection across Bering 

* Petersen, W., " Lepidopteren-fauna d. arkt. Gebietes, p. 44. 
t Girard, Maurice, " L'Entomologie de PAmerique du Nord," 
p. 287. 

| Grote, A., "Noctuidae of North America," p. 313. 


Strait alone. Mr. Emerton * makes mention of several 
instances among spiders. 

Such cases can be traced among earth-worms, beetles, wood- 
lice, ants and other groups. In some cases the identity or. 
similarity of species occurring on the two continents may be 
due to the fact that the species originated in Greenland or 
some polar centre, no longer in existence, and subsequently 
travelled in different directions towards their present habitat. 
This conception, however, has very little in common with that 
of a polar origin of life which was first mooted by Dr. Allen. f 
He argued that the northern circumpolar lands may be looked 
upon as the base or centre from which have spread all the 
more recently developed forms of mammalian life. 

A few years later Dr. Haacke J directed attention to the 
peculiar circumstance that the most primitive orders of 
mammals and birds all have their living representatives 
in outlying areas of the southern hemisphere, such as 
Australia, New Zealand, Madagascar, South Africa and 
South America, notwithstanding that these animals are 
known to have formerly inhabited the northern hemis- 
phere. This, he thinks, implies that a southward re- 
treat has taken place of the more ancient forms before 
the advancing host of higher orders of life. It would 
seem, therefore, as if streams of more and more highly 
specialized orders of mammals and birds had been slowly 
evolved in the north and had constantly pressed the older ones 
southward. -This suggested to Dr. Haacke the idea of a polar 
continent from which the various orders had gradually been 
distributed across the continents. Dr. Wilser even assumes 
a polar origin for man. 

A similar theory was pronounced by Canon Tristram || in 
explanation of the present distribution of the higher groups of 
birds and their migrations. The migratory instinct, he con- 
tended, was due to their having originated in a polar centre 

* Emerton, J. H., "Spiders common to New England and Europe," 
p. 129. 

t Allen, J. A., "Geographical Distribution of Mammals," p. 37o. 

I Haacke, W., " Nordpol als Schopfungszentrum." 

Wilser, L., " Der Nordische Schopfungsherd," p. 134. 

|| Tristram, H. B., "Polar Origin of Life." 


and having subsequently been dispersed southward by different 
routes. He believed that the tendency of birds to migrate 
northwards was due to a natural instinct to return to the 
home of their ancestors. 

What I chiefly endeavoured to prove in this chapter was the 
existence in pre-Glacial and early Glacial times of a land 
bridge joining Scotland, Iceland, Greenland and Labrador. 
The evidence in favour of such a land connection must be 
largely, if not entirely, biological ; but the testimony, as far 
as it goes, leads me to believe that the theory is well founded. 
I shall allude to a similar land connection in another chapter 
which probably joined North America and Asia. If the 
climatic changes ushered in by the Glacial Epoch were pro- 
duced by the closing of these two highways to the Arctic' 
Ocean, it is evident that the preceding warm period must 
have been due to a greater flow of warm currents to the Arctic 

A few years ago I drew attention to the fact that the animals 
and plants found on the Faroes and Iceland in particular 
imply the existence of a former land connection between Scot- 
land and the latter country. The occurrence in Iceland of the 
European field-mouse (Mus sylvaticus), of the snail Arianta 
arbustorum, which also inhabits the Faroes, of the beetle 
Nebria gyllenhali, which likewise inhabits Greenland, and 
many other forms not likely to have been conveyed by acci- 
dental means, all favour the view that the fauna of Iceland 
owes its existence mainly to the land "bridge referred to. But, 
as I pointed out, Iceland also possesses a distinctly American 
element in the snail Succinea groenlandica and others, while 
some of the American plants have even invaded the continent 
of Europe by the Greenland -Iceland land bridge. One of the 
objections raised against this view is that the low tempera- 
ture in the north would have prevented any faunistic inter- 
change across the land bridge. Tihe temperature, on the con- 
trary, in Iceland, southern Greenland, Labrador and Scot- 
land would have been considerably higher under such geo- 
graphical conditions than it is now. If go, why should not the 
whole fauna of northern North America have streamed across 
this bridge to Europe and that of northern Europe to North 
America ? If we examine the fauna of Canada we find that it 


is largely composed of Siberian immigrants. It is possible that 
these only began to enter North America from the extreme 
west at the time when Labrador was already joined to Green- 
land. Moreover, we may assume that as soon as the North 
Atlantic land connection excluded the Gulf Stream from the 
Arctic Ocean, glaciers began to gather on all the mountain 
ranges surrounding the northern parts of the ocean, thus 
preventing many forms of animal life from taking advantage 
of the facilities afforded by the land bridge of passing from the 
one continent to the other. The greater warmth on the 
northern coasts led to greater evaporation of sea water and 
more ample condensation of moisture on the elevated lands 
surrounding the ocean. All this will be more fully discussed 
in the succeeding chapters. 

Apart from the authorities already cited who expressed 
themselves in favour of a north Atlantic land connection, 
between northern Europe and North America, Professor 
Lobley,* Dr. Brown,f Mr. Grant J and Mr. Krishtafovitch 
have all urged the same view on different grounds. 

* Lobley, J. L., " American Fauna and its Origin," p. 26. 

t Brown, A. E., " American Big Game," p. 87. 

| Grant, M., "Mammals of North America," p. 12. 

Krishtafovitch, N. J., "La derniere periode glaciaire," p. 296. 



CROSSING Davis Strait from southern Greenland to 
Labrador, we land in a country whose barren headlands look 
just as stern and uninviting as those on the opposite side of 
the Strait. But whereas the west coast of Greenland pre- 
sents a shore-line varying between rugged precipices and 
undulating plains, the whole sea-front of Labrador, for over 
a thousand miles, rises abruptly from sea-level to a height 
of about 1,000 feet or more. Labrador presents an irregular 
plateau with a general elevation of from 1,500 to 2,000 feet 
above ,sea-level (see Fig. 1). It forms part of the oldest 
known land on the continent of North America, and, so far 
as we know, it has never been entirely covered by the sea 
since very remote geological times. The rocks are largely 
metamorphic with ancient igneous intrusions greatly resem- 
bling those of Greenland, which, like Labrador, is a fragment 
of that ancient continent to which Professor Suess applied 
the name " Laurentia." * 

A cold current of water loaded with icebergs from the 
Arctic Ocean sweeps down the east coast of Labrador. Off 
the coast of Newfoundland it meets a branch of the Gulf 
Stream, thus producing the dense fogs so characteristic of 
the Grand Banks. This arctic current is mainly responsible 
for the inhospitable nature of the Labrador coast. 

The vegetation is mostly stunted in character. The dwarf 
birch (Betula nana), the mountain-ash, alder and some 
willows here and there form small woods, while many of the 
familiar Greenland flowers reappear. Yet the flora is dis- 
tinctly richer than that of Greenland. Even close to the coast, 
in sheltered ravines, occasional specimens of the white spruce, 
which is more hardy than the black spruce, are noticeable. 

* Suess, E., " Antlitz der Erde," Vol. III.2, p. 284. 


To some of those who landed on the coast and explored the 
nooks and valleys, the country seemed full of beauty, of 
attractiveness, and even of a rich and appealing fertility. At 
certain times it presents an amazing wealth of strikingly 
coloured flowers. So thickly sown are they that at certain 
seasons they remind one of a cultivated garden.* 

Once we leave the coast region and enter the interior of 
Labrador, the climate becomes less arctic in character and 
timber increases in quantity. In fact there are two distinct 
climates in Labrador, the arctic on the coast, the north tem- 
perate in the interior. 

According to Dr. Packard,f the Greenland and arctic forms 
of animal and plant life occurring on the coast are 'the 
remnants of the glacial or arctic fauna and flora which, being 
formerly spread over the entire territory of British America 
and the north-eastern United States, still retain their hold 
on the treeless and exposed islands and headlands of 
Labrador. In many respects the Labrador fauna and flora 
resemble those of trie far distant White Mountains in New 
Hampshire, as we shall learn later on (p. 35). 

When we survey the fauna of the coast of Labrador more 
closely we find that, besides the Greenland or arctic element, 
another much richer one has apparently invaded the territory 
previously occupied by the former. This new fauna becomes 
more and more abundant as we proceed westward and south- 
ward. Thus the existence of the barren-ground caribou of 
Labrador which resembles the reindeer of Greenland, is 
threatened by enemies such as the glutton or wolverine (Gulo 
luscus), and these do not penetrate farther north. Occa- 
sionally the Canadian porcupine (Erethizon dorsatus) has 
been noticed in the coastal territory. Besides the lemming 
(Dicrostonyx hudsonius), which is probably identical with 
the Greenland form, quite an assembly of distinct ground 
rodents make their appearance, among them Synaptomys 
innuitus, Microtus enixus, Microtus pennsylvanicus, Evo- 
tomys ungava, Evotomys proteus, Zapus hudsonius, and 
Peromyscus maniculatus, also the arctic fox, red fox, several 

* Grenfell, W. T., " Labrador," pp. 393395. 
t Packard, A. S., " The Labrador Coast," p. 194. 


small carnivores and a variety of the black bear (Ursus 

The small mouse-like -creatures belonging to the genus 
Synaptomys are aptly called " lemming-voles " because, ex- 
ternally like lemmings, their teeth approach those of voles. 
When Dr. Merriam undertook his revision of the species of 
Synaptomys, f seven kinds were known to him. A few more 
have since been discovered. The genus is entirely confined to 
the North American continent, and all the species except two 
have a limited range in the boreal region. One of the latter 
(Synaptomys cooperi) is found from Massachusetts westward 
to Minnesota and southward to North Carolina, the other is 
peculiar to Mount Washington. Almost all the other species 
are confined to Canada and Alaska. It does not seem, there- 
fore, as if Synaptomys innuitus, which inhabits the coast of 
Labrador, were a recent immigrant from the south. On the 
contrary, the genus Synaptomys, being almost confined to the 
north, is probably of boreal American origin, one adaptable 
species having advanced far southward along the east coast 
of America. 

The next two species alluded to, Microtus enixus and M. 
pennsylvanicus, are true voles, though the last-named animal 
is generally known in the States by the name of " meadow- 
mouse." Both of them belong to a sub-genus which is 
very widely spread in Europe, Asia and boreal North 
America. J Some migrations across ancient land connections 
must have taken place, no matter whether we assume that the 
sub-genus is of Old World or New World origin. It only 
remains for us to determine whether Microtus crossed the 
North Atlantic land bridge or the one supposed to have been 
situated at Bering Strait, assuming that there was such a one. 
Professor Tullberg thought the members of the genus Microtus 
had travelled to America in Pliocene times by means of the 
first land bridge. This view does not appear to me pro- 
bable, because the family is entirely absent from Ireland, 

* Bangs, O., " Mammals of Labrador." 

t Merriam, C. H., " Revision of Synaptomys." 

\ Bailey, V., " Revision of Microtus." 

Tullberg, Tycho, " System der Nagetiere," p. 499. 


Iceland and Greenland. There is no fossil evidence that voles 
ever lived in any of these countries, where we might expect 
them to have survived had any extensive migration taken place 
from Great Britain to North America. Neither in Europe nor 
in North America are there any fossil remains of the sub- 
genus Microtus older than Pleistocene, if we accept Mr. 
Barnum Brown's estimate of the age of the Potter Creek 
deposits.* Nevertheless, it is possible that these voles origi- 
nated in North America long prior to the Pleistocene Period. 
If so I believe they made use of the Bering Strait land con- 
nection rather than the North Atlantic one, in passing from 
the New World to the Old. 

Evotomys, another genus found in the coastal district of 
Labrador, has a range somewhat similar to that of Microtus. 
It is closely allied to it and scarcely deserves the name of 
"red-backed mouse " as the Americans call it, because it is 
distinctly a vole, without any mouse -like characters about it. 
The relationship between the American and Old World forms, 
as in the case of Microtus, must be du to the existence of a 
former land bridge across Bering Strait. 

The jumping-mice (Zapus), to which Zapus hudsonius 
belongs, are, in many respects, an interesting group of 
rodents. Somewhat kangaroo -like in their movements, they 
are almost entirely confined to boreal North America. The 
Labrador jumping-mouse is a variety of Zapus hudsonius, 
which ranges from Alaska to Labrador and New York. A 
single species of jumping-mouse (Zapus setchuanus) occurs in 
China. f The theory that a former land bridge across Bering 
Strait enabled its ancestors to traverse the northern Pacific 
seems quite evident in this case. Yet we must not forget that 
the American jumping-mice also have somewhat more dis- 
tant relations in the Old World, the jerboas, from which the 
remote ancestors of Zapus may possibly have descended. 

We need not at present deal with the other rodents or the 
carnivores found in the coast district of Labrador, as most of 
these will be referred to again in subsequent chapters. The 
animals that have just been alluded to show us that the 

* Brown, Barnum, " Conard fissure," p. 208. 

f Preble, E. A., " Revision of the Jumping Mice%" 


Labrador fauna is largely composed of indigenous American 
species, which have not penetrated to Greenland. This seems 
to suggest that the points of resemblance in the fauna of 
Labrador and Greenland may have been more marked in 
remoter times, before the influence of the continental fauna 
had impressed itself upon the outlying peninsula of Labrador. 
Considering the extreme probability of Labrador having been 
connected by land with Europe by way of Greenland and 
Iceland in Pliocene times, we might expect some Euro- 
pean mammalian types to have occurred in north-eastern 
North America. The little evidence we possess tends to 
show that the mamtaialian life of the extreme north of 
western Europe was always poor. We cannot assume 
that insurmountable barriers prevented European mammals 
from invading America, because certain species such as the 
reindeer and lemming seem to have passed from America to 
Europe across a North Atlantic land bridge. It might be 
argued that European animals did cross over, but were unable 
to maintain themselves in America, a fate which has largely 
befallen the American immigrants in Europe. For such a 
supposition, however, we still lack evidence. If no examples 
of European animals or plants were known from the American 
side of the water, we might assume the land bridge to have 
been a discontinuous one, as Mr. Hedley suggested to me, 
connecting Greenland alternately with Labrador and Scotland. 
But one of our strongest supports for the North Atlantic- 
land bridge is, as I said, the presence of the European 
Helix hortensis in North America. It still occurs in 
Labrador. From Labrador it travelled southward along the 
coast. Is it possible tfoat the whole strip of coast was at 
'that time cut off by some barrier from the interior of 
North America ? 

Supposing the western parts of the Labrador plateau had 
become covered by glaciers as soon as the North Atlantic land 
bridge was formed, it might have produced an effectual barrier 
against western invaders and yet have allowed eastern forms 
to reach Labrador. 

Later on I intend to return to this problem again. Mean- 
while, let us journey westward across the high plateau of 
Labrador towards Hudson Bay and Central Canada. As we 


enter the vast forest region, plant life as well as animal life 
become more and more abundant and differentiated. In- 
stead of the barren-ground caribou, we now meet with the 
woodland form accompanied by another large ungulate, the 
moose, while the flying squirrel, chipmunk, the ground 
squirrel, woodchuck, white-footed mouse, musk-rat, beaver, 
skunk, weasel, shrews, moles and many other beasts tenant 
the forests, meadows, and banks of rivers.* 

A few amphibians, even, have succeeded in surviving the 
rigours of the arctic winter of those regions, and have success- 
fully established the most northern outposts in eastern North 
America. The leopard frog (Eana pi pi ens), one of the 
commonest as well as one of the most brilliantly coloured of 
American frogs, is one of these. The pickerel (Kana palus- 
tris), also the northern wood frog (Eana cantabrigiensis) 
and the northern frog (Eana septentrionalis) have all been 
observed in the neighbourhood of Hudson Bay. The most 
interesting species is the swamp-tree frog (Chorophilus 
nigritus), whose northern variety has advanced into this in- 
hospitable region, though almost all of the other members 
of the tree frog family (Hylidae) are typically southern 
forms. f Whether newts occur in the Hudson Bay region is 
not definitely known, but the salamander (Plethodon 
cinereus), at any rate seems to have been met with. All these 
species are peculiar to America. 

No reptiles have been noticed. The distribution of the 
terrestrial mollusks, the snails and slugs, implies that an ad- 
vance in a northward direction, similar to that recorded in 
the case of mammals and amphibians, has taken place among 
some groups of invertebrates. The typically American snails, 
Polygyra monodon and Strobilops labyrinthica, have been 
collected near Hudson Bay. 

When we analyse the constitution of all these western and 
southern groups, and trace the relationship of the members 
more carefully, we notice that many of them are not of 
American ancestry. They all have lived, no doubt, long 
enough in America to have become thoroughly established, 

* Preble, E. A., "Hudson Bay Eegion." 

t Dickerson, Mary C., " The Frog Book," p. 158. 


being part of the indigenous fauna, yet we recognise that their 
ancestors must have entered the continent from Asia in com- 
paratively recent geological times. 

Let us take for example the moose deer (Alces americanus). 
Its range extends from, Bering Strait, in a broad tract of forest 
land eastward, along the northern shores of the Great Lakes 
as far as Nova Scotia on the Gulf of St. Lawrence. Only 
along the Rocky Mountains, as Mr. Thompson Seton has so 
clearly indicated in his map of the range (Fig. 3), does the 
moose occur further south.* 

There are a few historical records, and also some fossil 
ones, which indicate that the moose once penetrated further 
into the United States in various directions, but it evidently 
never diverged very much from its present range.f The bones 
of a couple of closely allied animals have been met with in the 
Pleistocene deposits of Washington territory, and the skeleton 
of a peculiar moose, somewhat resembling the Alaskan variety, 
has been discovered in the Pleistocene of New Jersey, and 
placed by Professor Scott into a distinct genus (Cervalces). 
Whether this animal was ancestral to the living moose, as 
has been suggested, or whether it represents an aberrant 
type which has come in from Siberia with the moose, as Mr. 
Grant seems to think likely, are problems which may be 
left to future research es.J Certain it is that when we cross 
Bering Strait into Northern Asia, we meet with a moose 
(Alces bedfordiae) which in its simple antlers somewhat 
resembles the young American moose. Further west as far as 
Scandinavia, we find another species (A. machlis) differing 
but slightly from Alces americanus. It seems almost as if 
the moose had originated in eastern Asia from some more 
generalised type like Alces bedfordiae, and had gradually 
produced the forms with more palmated antlers in America 
and Europe by a process of convergent evolution. In any 
case, we are led to assume that Bering Strait was dry land 
when the ancestors of the existing moose entered the New 
World. Even if we suppose the moose to have originated in 
America, a land bridge connecting the latter with Asia was a 

* Seton, Thompson, " Life Histories of Northern Animals, L," p. 151. 

t Grant, Madison, "Moose." 

% Grant, Madison, " Origin and Relation of Mammals," p. 23. 



necessity. This problem of the Bering Strait land bridge, 
however, will be more fully discussed later on (pp. 83 86). 

The moose, or elk as it is called in Europe, has been stated 
to occur in the Caucasus, but its range only extends to the 
forests north of this mountain range. Like the reindeer, it 
occurred much further south formerly, even as far as 
northern Italy, and yet no theories as to a former arctic 
climate are founded on this fact. In Caesar's time the elk was 
still abundant in the Black Forest in southern Germany, 
while it is now confined to certain parts of Russia and 

Many other American species resemble the moose in their 
range, except that they manifest more clearly their Asiatic 
origin. The further west we travel the more often do we 
meet with such types of animals. 

We have now become acquainted with species that seem 
to be of arctic American or Greenland origin. A few 
apparently travelled across from Europe. Others were partly 
of southern extraction having advanced northward from the 
United States, and partly derived from a western invasion of 
the continent. What we want to find out ne-xt is the geological 
age of some of these invasions into boreal America. The 
writings on this subject of the prominent biological authors 
are largely biassed by the prevalent geological opinions with 
regard to the nature of the Ice Age. 

It is currently believed that a climatic change towards the 
end of the Pliocene Period caused ice-masses to form in 
various Canadian centres, so as to produce continental 
glaciers of the type of the ice-sheet now covering Greenland. 
These glaciers are supposed to have spread from at least three 
centres in Canada, termed the Labradorean, Keewatin, and 
Cordilleran. During this time, viz., in our most recent 
geological age, the " Pleistocene Period," fully one half of 
North America is thus assumed to have been gradually buried 
beneath these vast sheets of ice. Ice is believed to have ex- 
tended practically all over the continent from Newfoundland 
in the east to Vancouver in the west, and from the Arctic 
Ocean to the present site of the City of St. Louis.* Belying 

* Russell, I. 0., " North America," p. 315. 
L.A. D 


on the accuracy of these data, on which the majority of 
geologists are agreed, biologists have endeavoured to work 
out the past history of the American fauna in strict accord- 
ance with the facts these phenomena are thought to reveal. 
" Throughout the growth of the great ice -mass, and its ex- 
tension from the north southward," says Dr. Merriam, " it 
is clear that the animals and plants that could not keep pace 
with its advance must have perished, while the steady push- 
ing towards the tropics of those that were able to escape to 
the rapidly narrowing land in that direction must have 
resulted in an overcrowding of the space available for their 
needs and a corresponding increase in the severity of the 
struggle for existence."* Immediately upon th,e close of the 
Glacial Epoch life began to reclaim the regions from which 
he thinks it had so long been shut out. 

Dr. Allen's views are very similar. " There is evidence," 
he remarks, " that towards the close of the Tertiary, a marked 
change in the earth's climate took place, culminating in the 
Glacial Period, during which the whole northern half of the 
northern hemisphere became covered with a heavy ice-cap, 
lasting for possibly thousands of centuries, and extending 
its chilling influence nearly to the northern tropic. The rise 
of the Glacial Period was of course gradual, and the south- 
ward progress of the great ice- cap drove before it all forms 
of life capable of any considerable powers of locomotion, while 
those unable thus to escape must have perished from cold. 
Finally the ice receded to its present limits and the whole 
north, under radically altered climatic conditions, became 
again available for occupation by the more or less modified 
descendants of th,e pre-Glacial exiles. "f 

The bog plant societies so graphically described by Mr. 
Transeau probably existed, he thinks, along the whole ice 
front. The bog and tundra types were eventually the first 
to push into the barren ground left by the retreating ice. 

Professor Adams takes a more independent attitude. He 
assumes that repeated glaciation had almost sterilised the 

* Merriam, C. H., " Life in North America," p. 45. 

t Allen, J. A., " Distribution of North American Birds," p. 100. 

| Transeau, E. N., " Distribution of Bog Plant Societies," p. 414. 


northern part of the continent, still he concedes that even 
during the Glacial Epoch, life of the tundra type may have 
flourished in Alaska.* 

Hence it is not unreasonable to argue from his point of 
view that life could also have existed in Greenland at that 
time, and this opinion I endeavoured to vindicate in the 
last chapter. Even Professor Adams does not venture to 
cast a doubt upon the correctness of the current geological 
theories, and speaks of three distinct belts of life in the 
vicinity of the ice margin (p. 56). The latter being fringed 
to the south by tundral biota (fauna and flora), next to which 
came the northern trans -continental coniferous forest belt 
and its associated fauna, and finally the deciduous forests. 
All these are assumed to have moved forward to the north on 
the disappearance of the ice. 

That the so-called " tundral " or what we might call arctic 
fauna and flora actually did advance far south of their present 
habitats can be demonstrated much more clearly than by the 
occurrence of a few stray fossil reindeer's antlers south of 
the area covered by glacial drift. 

Whether the past southward migrations of the reindeer 
were influenced by climatic changes or by other considera- 
tions, we cannot definitely assert. Since we are told that 
there was a refrigeration of the climate during the Glacial 
Epoch, we are apt to assume that this lowering of the tem- 
perature drove the reindeer and other arctic species to more 
southern localities. Tlie former occurrence of an animal of 
such a roving disposition as the reindeer in more southern 
districts may have been due to a natural expansion of its 
range, and this need not imply a change of temperature. 

The fauna and flora of the White Mountains has been cited 
as a living testimony of a former arctic climate in latitudes 
where temperate conditions now prevail. 

Surrounded by an entirely alien assemblage of animals and 
plants we find in the White Mountains of New Hampshire, 
not far from the city of Boston, an extraordinary gathering 
of species, many of which are only known elsewhere in 
Labrador and Greenland. A thousand miles away from their 

* Adams, Chas. 0., " Dispersal of North American Biota," pp. 55 58. 

D 2 


natural home, they form in the White Mountains a peculiar 
island colony on a bare plateau which lies about 5,000 feet 
above sea-level, and out of which project a series of conical 
peaks. These constitute what is known as the Presidential 
Range, the highest being Mount Washington (6,293 feet) 
(Fig. 4). The plateau on which this remarkable relict fauna 
and flora lives may be reached by the railroad that now takes 
the traveller to the top of Mount Washington or by following a 
very rough trail leading to it from the Pass called " Crawford's 
Notch " through the dense forest. It was the latter route I 
chose to gain the wind-swept ridge. As we approach the 
ridge, the fine spruce, balsam and paper birches are gradually 
replaced by diversified conifers, which become increasingly 
stunted by exposure in the more elevated parts of the range. 
At last, on emerging from the remnants of the forest, we 
have gained not only a wide expanse of open country, but we 
can imagine ourselves transplanted all of a sudden to the 
Arctic Regions. Here and there may be gathered specimens 
of Rhododendron lapponicum and Salix phylicifolia, growing 
among Arenaria groenla-ndica, Phleum alpinum, Diapensia 
lapponica, Campanula rotundifolia, Gentiana nivalis and 
hosts of others, few of which can be studied elsewhere nearer 
than Labrador. 

Moreover, as Dr. Scudder remarked, no State in the Union 
presents so striking an assemblage of animal life as New 
Hampshire, where the White Mountains form so conspicuous 
a feature. Swiftly running over the bare rocks of the high 
plateau we notice the black spider, Pardosa groenlandica, 
which, though occurring also in the Rocky Mountains, ,is 
otherwise confined to the White Mountains, Labrador and 
Greenland.* The grasshopper, Pezzotettix glacialis, is, I 
believe, peculiar to Mount Washington, while another species, 
Pezzotettix borealis, is a near relative of the North European 
Pezzotettix frigida. Fluttering among the arctic vegetation, 
we notice the butterfly Oeneis semidea, which has never been 
taken nearer than Labrador, while the moths Dasychira rossii, 
Arctia quenselii and Anarta melanopa, are all well-known 

* Chamberlin, E. V., " Eevision of North American Lycosidae," 
p. 200. 

FIG. 4. Map showing the outlines of the White Mountain Plateau (shaded) and 
Mount Washington Range (black), with an inset Map of North America, 
indicating the position of the White Mountains (in the centre of small circle). 

[To face p. 36. 


arctic species occurring on th<e White Mountain plateau.* 
Among the beetles, too, and other groups of invertebrates, 
there are many arctic forms, showing clearly the intimate 
relationship that exists between the faunas of the White 
Mountains and Labrador. f 

Even among mammals we have some most interesting 
boreal representatives, the White Mountain lemming- vole 
(Synaptomys sphagnicola) being peculiar to this region. J In 
alluding to the mammalian fauna of Labrador I specially 
dealt with this genus, and expressed the belief in its arctic 
origin and subsequent southward dispersal (p. 28). 

If merely a few arctic plants and insects inhabited this re- 
markable plateau, the argument might be permissible that 
they had been carried southward by wind currents from their 
northern home at great intervals of time and had successfully 
established themselves in this manner on the White Moun- 
tains, because the latter proved to be uninhabitable by the 
fauna and flora of the surrounding country. No one, however, 
who has seriously studied this congregation of animals and 
plants as a whole, can for a moment entertain such an idea. 
We must therefore take for granted that a fauna and flora 
similar to that now existing in Greenland, Labrador and on 
the White Mountains once extended over a large portion of 
Canada and at any rate the north-eastern United States. The 
problem to be solved is, what were the circumstances which 
led those animals and plants to extend their range so much 
southwards ? The prevalent theories regarding these move- 
ments have already been referred to. They are simple enough. 
Similar ones have been current in Europe for many years 
past. I shall quote Professor Adams again, as his views 
seem to me to express those generally entertained on 
this subject. After an allusion to the final northward retreat 
of the ice which he supposed to have crept down from the 
north " grinding to pieces everything beneath its awful 

* Scudder, S. H., " Distribution of Insects in Now Hampshire," 

pp. 331341. 

f Gardiner, F., " Coleoptera of the White Mountains." 

| Miller, GK, " Mammals of New Hampshire Mountains." 

Adams, Charles 0., " Post-glacial Origin of the life of North - 

Eastern United States," p. 309. 


weight," it once more permitted life from the south to move 
into the newly opened territory. The first wave of life, he 
thinks, which became dispersed over the glaciated region, as I 
have already mentioned above, was the barren-ground fauna 
and flora, the types of the far north. " Hardly a trace," 
he continues, " now remains in the eastern United States of 
this pioneer class, and this class is now restricted to very 
limited areas or mountain top ' islands,' as in the White 
Mountains. The present distribution of these arctic-alpine 
' islands ' brings up a number of very interesting points. How 
is it that these alpine plants and animals are now found so far 
south and only upon high mountain peaks ? This question 
could not be rationally explained until the influence of the Ice 
Age upon life became recognised, and would, indeed, be a 
difficult one to answer if we did not take into consideration 
past conditions of climate and topography. We must recall 
that the arctic conditions which now occur outside of the arctic 
regions only on high mountain tops at that time extended to 
the very base of the mountains, and as the ice retreated to the 
north this cold zone gradually moved up the sides of the 
mountains carrying with it a characteristic flora and fauna. 
Thus with the retreat of the ice the first wave of life had 
two evident possibilities before it: first, to follow the ice 
north, or to follow the cold zone up the mountains. By this 
means small colonies of arctic plants and animals became 
separated from the main body of forms, and thus became 
alpine. To be sure, this segregation could only occur where 
the mountain peaks are isolated. If the mountains had been 
of sufficient height and had extended far to the north, a high- 
way would have been retained to the original stock, and thus 
the distribution would not have become discontinuous. From 
the above interpretation it seems fair to conclude that the 
arctic forms which now occur at alpine heights are relicts of 
the former widespread glacial fauna and flora, whose dis- 
tribution has become discontinuous by a change of climate." 
Very similar ideas were held by Mr. Grote.* He thought 
the White Mountain butterfly (Oeneis semidea) was pushed 
southwards by the advance of the great northern ice-sheet. 

* Grote, A. E., " Effect of Glacial Epoch," p. 441. 


On the decline of the ice-sheet the butterflies turned north- 
ward, again returning to their native home. Some of the 
specimens strayed by the way and were destined to planlt 
colonies apart from their companions as, for example, on the 
White Mountains. 

Botanists entertain analogous views. Dr. Harshberger * 
argues that the tundra vegetation and other arctic species 
of plants occupied during the Glacial Epoch the southern 
margin of the great ice-sheet, and that when most of them 
migrated north, on the disappearance of the ice, some re- 
mained behind to form the vegetation of sphagnum bogs and 
alpine summits of the higher mountains. 

If these theories are correct, the Asiatic invasion and the 
much more insignificant one from Europe, of which Helix 
hortensis is one of the most striking representatives (p. 13), 
should both be more recent than the flora and fauna of the 
White Mountains, for the former have n6t penetrated beyond 
the lower slopes of these mountains. 

Helix hortensis does not occcur nearer the White Mountains 
than Portland in Maine, which is fully seventy miles to the 
east. I have traced Oniscus asellus, a wood-louse probably 
belonging to the same group, as far as the base of the White 
Mountains. The earth-worm, Lumbricus castaneus, which 
seems to have spread from wntinental Europe to the Faroes 
and Iceland, reappearing on the other side of the Atlantic 
in Canada and New England, may be a member of the same 
dispersal. At any rate, I feel sure there are a great many 
more of such species that have not spread to the higher parts 
of the White Mountains, and therefore proclaim themselves as 
more recent immigrants than those which are now in posses- 
sion of the high plateau referred to. The latter are likewise 
clearly older than the Asiatic immigrants, which will l)e more 
fully described later on. 

But since Helix hortensis occurs in the lower Pleistocene 
clays of Maine, it, as well as the whole group of European 
immigrants, are pre-Glacial in age, and in this opinion I 
concur with several of the authorities who have discussed, 
this problem (p. 14). The members of this group arrived 

* Harshberger, John W., " North American Plant Dispersal," p. 2. 


in America before the time when the glacial drift was de- 
posited, or, in other words, when the northern ice -sheets were 
supposed to have advanced to their southern limits. The 
glacial drift, I may mention again, is the mantle of clay, sand 
and boulders believed to have been left by the ice as it 
retreated northward. Whether this drift or boulder clay really 
is the product of immense glaciers, or whether it was deposited 
in the sea by floating icebergs, it is evident that wherever the 
country is covered by it the pre-existing fauna and flora must 
have been destroyed. The evidence seems to me all in 
favour of destruction rather than emigration. 

The idea of a gradual southward withdrawal of the fauna 
and flora, that they fled like a conquered army before an ad- 
vancing foe, sounds very plausible, but is there any foundation 
for such a belief ? 

No evidence can be adduced from fossil specimens that 
any members of what I have called the European invasion 
ever penetrated southward of the limits of the drift in North 
America. They do not seem to have been pushed south in front 
of the advancing masses of northern ice. Theoretically, they 
ought to have survived the Ice Age somewhere in south-eastern 
North America. If they did, they must subsequently have 
reoccupied the very parts, viz., Labrador, Newfoundland and 
the coast of New England, where they originally set .foot 
on American soil after completing their travels across the 
North Atlantic land bridge. But is it possible that they 
quitted the south-eastern States without leaving a trace of 
their former presence there ? As Dr. Harshberger * points 
out, not a single species of thirty-four plants characteristic 
of the area just south of the glacial drift deposits, is a native 
of Europe. On the other hand, of the plants growing on 
th,e drift itself, about one-third are common to northern 
Europe and America. All the available evidence, therefore, 
points to a survival of the European element within the 
glaciated area. 

Could any 'islands have existed in the midst of this glaciated 
area, where this assemblage of European plants and animals 

* Harshberger, J. W., " Comparative Age of Floristic Elements," 
p. 606. 


might have survived the Ice Age, without being affected by 
the Ice-sheets ? 

Labrador was one of the areas which was believed to have 
had an independent centre of glaciation, yet Professor Daly,* 
after inaking a special study of the geology of Labrador, 
remarked " nothing is more striking in the glacial geology of 
the southern part of the coastal belt than the almost com- 
plete absence of drift deposits." In the Torngat Mountains of 
Labrador no signs of glaciation were noticeable above 2,000 
feet. That there were considerable tracts of Labrador which 
were free from ice must be evident, and I presume the Euro- 
pean plant and animal migrants survived the Glacial Epoch 
there and also further south. The island of Newfoundland 
seems to have had quite a separate area of glaciation, and the 
same was probably true of Nova Scotia, according to Pro- 
fessors Chamberlin and Salisbury. f 

The two countries of Labrador and Newfoundland have 
many species of animals and plants in common, and in both 
no doubt a large part of the pre-existing fauna and flora sur- 
vived the Glacial Epoch. I have urged in the last chapter 
(p. 14) that the land probably stood at a much higher level 
towards the beginning of the Pleistocene Period than at pre- 
sent, the whole of the Bank of Newfoundland, and southward 
as far as Cape Cod, being raised high above sea-level. While 
I claim that the remainder of boreal North America has sub- 
sequently become largely submerged, these eastern tracts are 
likely to have remained above water, thus forming an asylum 
for the survival of the arctic and Old World fauna and flora. 
This opinion is confirmed by Professor Upham's J remark 
that the elevation of the fossilif erous marine beds lying on the 
glacial drift increases as we proceed north,-westward from 
Boston, that is to say, inland, while along the lower St. 
Lawrence it decreases again, so that in Nova Scotia actually 
on the sea-coast marine deposits are wanting. 

The current geological theories of the Ice Age or Glacial 

* Daly, E. A., " Geology of Labrador," pp. 245251. 
t Chamberlin, T. 0., and E. D. Salisbury, " Geology," Vol. III., 
p. 336. 

{ Upham, Warren, " Marine Shells near Boston," p. 140. 


Epoch do not seem to me to harmonise at all with the geo- 
graphical distribution of animals and plants. If we assume 
that an arctic climate prevailed at that time all over Canada 
and the northern United States, we are faced by numerous 
difficulties. The biological evidence favours the view that the 
climate in boreal North America, though much more humid 
than at present, so that it led to extensive glaciation on all 
higher mountain ranges, was not arctic but temperate, and 
that in many parts within the so-called glaciated area there 
existed islands where life was abundant and survived to the 
present day. 

Let us return to the animals and plants inhabiting the 
White Mountains. Their relationship is almost altogether 
with Lapland and Greenland, and yet that affinity has clearly 
been brought about at a much earlier date than that of the 
arrival of the European element in North America. 

During the Pliocene Period movements seem to have taken 
place resulting in an increased height of land. This need not 
necessarily have affected the whole of North America. It was 
probably more or less confined to the north-eastern and north- 
western parts. While the closing of the North Atlantic left 
the coastal districts open to the beneficial influence of the 
Gulf Stream, the temperate fauna and flora must have gradu- 
ally disappeared from the more inland boreal parts of the con- 
tinent, thus leaving room for the expansion of the arctic 
animals and plants in various directions. It was during the 
Pliocene Period, I think, or earlier, and, at any rate, long 
before the commencement of the Glacial Epoch, that the 
animals and plants from Labrador thus found their way south- 
ward to the White Mountains. However, 1 shall bring forward 
further evidence later on which will throw additional light 
on the problems I have discussed. 

The theory that the animals and plants were driven south 
of the ice foot or southern margin of the supposed great 
ice -sheet ought to be supported by biological evidence. 

Theoretically it is assumed that the barren-ground or arctic 
fauna and flora lived close to this margin, as already stated, 
and the temperate forms further south. The only fossil evi- 
dences we possess of arctic animals having actually lived south 
of the ice-sheet, or, as we might say, south of the limits of 


the drift, are rather problematical. Most, if not all, the occur- 
rences of reindeer and musk ox bones lie within the drift area. 
As already mentioned, animal remains have been met with in 
caves and other deposits, close to the limits of the drift, and 
supposed to belong to the Pleistocene Period, which seem to 
indicate a climate somewhat milder than the present one. 

The most noted Pleistocene fossiliferous strata within the 
drift area lie in Canada, and these contain largely the remains 
of plants. Since the first place as tests of climate has gene- 
rally been assigned to plants,* their testimony will be of 
particular value in our present enquiries. These deposits have 
been principally studied by Professors Coleman'f and Pen- 
hallow. J Some are in the- neighbourhood of Toronto near the 
shores of Lake Ontario, others further west near the Moose 
and Albany Kivers, both of which empty their waters into 
Hudson Bay. 

In dealing with the beds in the neighbourhood of Toronto, 
Professor Coleman reports that those of Scarboro' Heights 
contain mosses, diatoms, a few fresh-water shells and a con- 
siderable number of elytra of beetles. According to Dr. 
Scudder the latter, numbering twenty-nine species, are all 
extinct, and related to species occurring in Lake Superior and 
Hudson Bay regions, the fauna having a boreal aspect. The 
fossils from the Don Eiver deposits were found to be sur- 
prisingly different. They seemed to point to a climate as 
warm as that of Toronto, if not much warmer, while the forest 
trees suggest a temperature far from glacial. Not a trace of 
an arctic fauna or flora could be discovered. It was con- 
cluded, therefore, by Professor Coleman that both these series 
of beds were inter-glacial, that is to say, laid down during the 
mild phases which are supposed to have separated the in- 
tensely arctic ones from one another. 

The plant remains from the other deposits were like those 
of Scarboro' Heights and Montreal. They were essentially of 
the same character representing a vegetation similar to that 
of our own time, or perhaps even a little more severe. 

* Seward, A. C., "Fossil Plants as Tests of Climate," p. 10. 

t Coleman, A. P., " Glacial and Inter-glacial Deposits," pp. 625 640. 

I Penhallow, I). P., " Pleistocene Flora of Canada," p. 77 


Four years later a valuable report on the Pleistocene fauna 
and flora of Canada was read at the British Association Meet- 
ing at Bradford by a committee which had been appointed to 
investigate the subject. Of this committee, Professor Cole- 
man and Professor Penhallow were members, as well as 
Sir William Dawson.* The number of beetles brought to light 
from the Scarboro' Heights had now increased to seventy-two 
species, of which seventy were pronounced by Dr. Scudder to 
be extinct. The new species confirmed Dr. Scudder in the 
opinion, previously expressed, that on the whole the fauna 
has a boreal aspect, though by no means so decidedly boreal 
as one would anticipate. 

No less than eighty-three species of plants were studied 
from eighteen different localities, one of the plants, viz., Acer 
pleistocenicum, being extinct. The abundant occurrence of 
some species, such as the Osage orange (Maclura aurantiaca), 
the paw-paw (Asimina triloba) and others, point to the pre- 
valence of a much warmer climate than now prevails. On 
the other hand, the equally abundant occurrence of boreal 
types at Scarboro' Heights suggests the existence of a cooler 
climate at the time these deposits were laid down. 

Once more, in 1907, Professor Penhlallow f dwelt upon the 
results of his researches on the plant remains of the Don 
River beds, urging that the same flora must have characterised 
the entire region between Virginia and Ontario in Pleistocene 
times, whilst a much warmer climate than at present pre- 

If similar evidence were brought to light from any other 
deposit than the Pleistocene, there can be no doubt as to the 
conclusions that would be drawn from it. The climate in 
boreal North America during the Pleistocene Period, as re- 
vealed by the plant and animal remains, must have been on the 
whole a temperate one. Yet geologists maintain, in the face 
of this testimony, that all these plant and animal remains only 
represent the so-called interglacial phase of the Glacial 
Epoch, during which the climate was supposed to have been 
temperate or mild. The other phase of the Ice Age, they 

* Dawson, J. W., D. P. Penhallow and others, "Canadian Pleistocene 
Fauna and Flora," p. 334338. 

t Penhallow, D. P., "Pleistocene Flora of Canada," pp. 443-450. 



argue, has left no traces of animal or plant life in all the 
numerous deposits which have been examined. 

The whole conception of these interglacial phases of the 
Glacial Epoch has given rise to a good deal of animated 
discussion. Professors Chamberlin and Salisbury have 
adopted Professor James Geikie's view that there were 
six great advances and retreats of the ice -sheets, sepa- 
rated by five interglacial intervals, during which a mild 
climate prevailed. But the evidences for these alternate 
advances and /retreats of the glaciers are "by no means ad- 
mitted as valid by all geologists. Some maintain that there 
were only three such great advances and retreats. Others 
admit only two of them. Some authorities disbelieve alto- 
gether in mild interglacial phases, and admit only one ad- 
vance followed by a gradual retreat of the ice. Even after 
studying the Toronto clays, which Mr. Lamplugh * acknow- 
ledges impressed him strongly as affording the kind of evi- 
dence which he has sought in vain in Britain, he is still of 
opinion, as expressed in his address to "the Geological Section 
of the British Association, that there is no proof of mild inter- 
glacial epochs, nor even of one such epoch. 

My own conclusions as to the nature of the Glacial Epoch, 
and the causes that produced the glacial clays, being almost 
entirely based on the evidence derived from the past and pre- 
sent fauna and flora, I have no hesitation in agreeing with 
Mr. Larnplugh's views. There is no biological evidence in 
North America in favour of one or more interglacial phases. 
Everything moreover points to the fact that during the so- 
called Glacial Epoch there was no diminution of temperature, 
or if so only a very partial one, although the higher mountain 
ranges were covered by glaciers. In many parts of North 
America there was probably a higher temperature during the 
Ice Age than obtains at present. The first to advocate the 
idea of a higher mean temperature being compatible with a 
greater extension of glaciers was, I think, Professor Lecoq.*j~ 

Much more recently a similar theory was very ably main- 

* Lamplugh, G. W. , " On British Drifts and the Inter-Glacial Problem," 
p. 26. 

t Lecoq, H., " Des Glaciers et des Climats." 


tained by Professor Whjtney.* and was supported by weighty 
arguments, while Sir Henry Howorth f has contributed addi- 
tional testimony in favour of a mild climate having prevailed 
during the so-called Glacial Epoch or Ice Age. 

That extensive glaciers existed during that Epoch in 
western Europe and north-eastern North America is pri- 
marily due to the closing of the North Atlantic. This pro- 
duced a rise of temperature in the Atlantic Ocean and intense 
precipitation over western Europe, particularly Scandinavia, 
and also .over north-eastern America, especially Labrador. 
As Professor Whitney observes, the regions mentioned are the 
only two in the world in which the topographical and climatic 
conditions seem to have been considerably different during the 
Glacial Epoch from what they are at present. The prevail- 
ing opinion, however, among geologists is, that not only was 
there -a, centre of glaciation in Labrador, but also in the 
Keewatin district to the west of Hudson Bay. The latter being 
a perfectly flat region, without even a suggestion of a moun- 
tainous nucleus, constitutes, indeed, as Professors Chamberlin 
and Salisbury J acknowledge, one of the most marvellous fea- 
tures in ice dispersion. All our notions as to the behaviour of 
glaciers are derived from careful .observations on existing ones. 
But nowhere on earth can a glacder be seen which, having 
originated on level land and developed quite independently of 
any adjacent higher region, proceeds to invade the neigh- 
bouring areas. Hence a biologist may be excused for express- 
ing some hesitation in accepting so extraordinary an hypo- 
thesis in spite of the fa-ct that we are assured that the days 
of reasonable doubt as to the former existence of these almost 
inconceivably large ice-fields are past. 

To attempt even to discuss all the various lines of evi- 
dence which have led to the almost general acceptance of the 
land-ice theory, as understood at the present day, would be 
impossible in a work of this nature . I only wish to bring 
forward some of the chief reasons which have prevailed upon 
me to reject this theory. I c?an scarcely venture to hope that 

* Whitney, J. D., " Climatic Changes," p. 321. 

t Howorth, H. H., "Ice or Water," Vol. II., p. 492. 

J Chamberlin, T, C,, and B, D, Salisbury, " Geology," III., p. 332. 


my criticisms will be readily accepted, yet they may possibly 
help in clearing up some points which hitherto could not be 
satisfactorily elucidated by other methods. The vast drift 
deposits which shroud the country like a great mantle of clay, 
sand and stones, frequently contain the remains of the ex- 
clusively marine group of foraminifera. Mr. Joseph Wright, 
a well-known European authority, has shown that the species 
of foraminifera have a very wide distribution in European 
boulder-clays. And it appears, on the authority of Sir Henry 
Howorth,* that Mr. Wright has likewise identified foramini- 
fera from American glacial clays. Samples were .submitted 
to him by the late Dr. G. Dawson from Saskatchewan 
River, 1,850 feet above sea-level, from Selkirk in Manitoba, 
and from Ottawa. The sample from Saskatchewan contained 
specimens of foraminifera referable to recent species, one 
of which (Nonionina depressula) is also common in European 

It is quite possible that foraminifera may be found in many 
other localities in the same clays ; indeed, Sir Henry Howorth 
mentions that Sir William Dawson had found them generally 
diffused in the Pleistocene clays of Canada. This fact, there- 
fore, tends to support Colonel Feilden's f contention that all 
the glacial deposits which he had examined in Arctic and Polar 
lands, with the exception of terminal moraines now forming 
above sea-level, are glacio-marine beds. 

Supposing the waters of the Arctic Ocean had risen, per- 
haps in consequence of the closing of the Atlantic Ocean, 
and had poured into Hudson Bay, overflowing its banks, 
and had then crossed the low-lying watershed separating 
this northern region from the depressions of the Great Lakes, 
the latter would soon have been filled with brackish water, 
killing or driving away many of those forms of life that were 
unable to adapt themselves to this change of conditions. I 
presume, of course, that troughs, not necessarily like the 
lakes now existing, already occupied the same region in pre- 
Glacial times. Such an hypothesis of this area having been 
invaded by the sea in Pleistocene times is supported by some 

* Howorth, H. H., "Ice or Water," Vol. II., p. 216. 

t Feilden, H. W., " Glacial Geology of Arctic Europe," p. 57. 


biological evidence, though it Is usually argued that the ocean 
crept inland through the St. Lawrence or Hudson Eiver 

Fifty years have elapsed since Professor Loven * first drew 
attention to the presence of several crustaceans, allied to 
marine forms, in Swedish lakes, and endeavoured to prove that 
the latter -must have been covered by the sea in recent geo- 
logical times. One of these crustaceans is the fresh-water 
shrimp Mysis relicta, closely related to the common Arctic 
marine form Mysis oculata. Since the latter does not occur in 
the Baltic, the theory that Mysis relicta is a recent immigrant 
from that sea is untenable. It seems much more probable 
that it gradually developed from its marine relative Mysis 
oculata, when the Arctic Ocean covered the lowlands of 
northern Eussia, Sweden and northern Germany. This view 
is confirmed by the fact that Mysis relicta has now been dis- 
covered, along with other marine organisms, in several lakes, 
all of which are situated within the area covered by the north 
European drift. I have also alluded to its occurrence in 
Lough Neagh in Ireland.f 

Now it is of great interest that this small freso.-wa.ter 
shrimp Mysis relicta, which is so closely related to a marine 
species, should also occur abundantly in Lakes Superior and 
Michigan. First discovered in the stomach of a Coregonus 
taken in Lake Michigan by Dr. Hoy, it was subsequently met 
with by Dr. Stimpson,J living in about fifty fathoms of water 
in the same lake. It has since been noticed in Lake Superior 
by Messrs. Smith and Verrill, but in no other of the Canadian 
lakes. This is in so far significant, as neither of these two 
lakes are supposed to have been invaded by the sea in Glacial 
and post-Glacial times. All that geologists acknowledge is 
that the sea crept up the Hudson Eiver and the St. Lawrence, 
as far as the western end of Lake Ontario. The latter and 
Lake Champlain had then a marine fauna, and it is there 
that we should expect Mysis relicta to occnr. 

But besides this fresh-water shrimp, another small 

* Loven, S., " Tiber einige Crustaceen." 

t Scharff, E. F., "European Animals," p. 155. 

J Stimpson, A., " Fauna of Lake Michigan," p. 403. 


crustacean, called Pontoporeia hoyi, allied to a marine form, 
inhabits the Lakes Michigan and Superior, while in the 
former still another marine form (Pontoporeia filicornis) and a 
remarkable fish called Triglopsis thompsoni have been met with. 
Professor Peschel first applied the term " relict lakes " * 
to lakes like those referred to, because they contain the 
relicts of a former sea, and the word has since been largely 
adopted in the sense in which it was first proposed. Never- 
theless, some authorities, notably Professor Credner,f who 
treated the subject most exhaustively, maintain that these 
so-called relict animals have either immigrated to the fresh- 
water lakes directly from the sea, or have been transportedi 
from their marine habitat by accidental means. He notes 
with surprise the absence from all the so-called relict lakes 
of marine mollusks or species allied to such, contending that 
some of them ought to be found in these lakes, if the latter 
had ever been covered by the sea. Professor Credner's argu- 
ments against the relict nature of such lakes as Superior and 
Michigan, seem at first sight very convincing. But zoo- 
logists are now well acquainted with 1 the fact that all marine 
mollusks have free-swimming larvae. These being altogether 
a prey to currents would infallibly be swept into the sea if 
placed into running fresh water. Thus the parent marine 
mollusks would be unable to propagate their kind in fresh 
water even if they could adapt themselves to it. This circum- 
stance seems to account in a satisfactory manner for the 
absence of marine mollusks from relict lakes. Quite apart 
from this objection to the relict theory, Professor Credner's 
contention of an. active migration of marine organisms to 
fresh water is particularly inapplicable in the case we have 
been considering, because none of the " relict species " 
alluded to occur in the lower lake of Ontario, which, moreover, 
is separated from the upper lakes by the impassable Niagara 
waterfall. It is of interest also to note that the only other 
locality in which the fish Triglopsis thompsoni has been 
taken, besides that of the Great Lakes, is in tidal pools on the 
west coast of Hudson Bay. 

* Peschel, O., " Vergleichende Erdkunde," p. 167. 
t Oredner, E., " Eeliktenseen," I., p. 105. 

L.A. B 


Other species of fishes related to marine forms live in the 
Great Lakes. But we need not suppose that typical fresh- 
water fishes are absent. The family of Cyprinids, those 
essentially fresh- water fishes to which the carp belongs, is 
well represented there. Hence if we suppose that they were 
destroyed during the invasion of that region by the sea, the 
lakes must, at a subsequent period, have been freely connected 
with rivers which were stocked with fresh -water fish. And 
it appears that this actually has been the case. 

Whether the St. Lawrence was choked by obstructing ice 
masses, or whether it was filled with detrital matter carried 
down from the neighbouring eastern heights, does not affect 
our present inquiries. It seems that after the ice, or, if we 
prefer it, after the retreating sea, had left the district, there 
were only three lakes instead of the five as at present. These 
three lakes, which have been called " Duluth," " Chicago " 
and "Maumee," were independent of one another. Lake 
Duluth emptied its waters into the Mississippi, Lake Chicago 
excavated an outlet which has since become the Chicago 
drainage canal, while Lake Maumee communicated with the 
Wabash River. All, of course, eventually helped to swell the 
waters of the Mississippi. Lake Maumee gradually developed 
into the huge Lake Warren. Later on the other lakes joined, 
and poured their waters into Ontario, which then communi- 
cated with the Hudson by means of the Mohawk River. It 
was only then that the sea finally broke down the barrier 
which had prevented an overflow into the St. Lawrence, and 
advanced as far as the western end of Lake Ontario, one 
marine channel communicating with the Atlantic by way 
of the Hudson Valley. 

We can easily imagine how the fresh-water fishes of the 
north, which were either killed or driven south, such as the 
redfin (Notropis cornutus) and Notropis atherinoides (it does 
not seem to possess a common name) managed to return to 
their northern stations after the Glacial Epoch. Yet I feel 
sure that a survival even of some fresh-water fishes took place 
within the drift area in some of the eastern mountain ranges. 
Near Freeport in Maine, quite an isolated colony of the 
beautiful red-bellied dace (Chrosomus erythrogaster) occurs. 
This is probably a pre-Glacial relict. Nor is it likely that 


the silver chub (Semotilus corporalis), which only frequents 
clear streams, could have spread in post-Glacial times from 
the south-eastern counties to the St. Lawrence basin. Two 
more examples might be mentioned that are practically, 
confined to the drift area and they are among the most 
interesting geographical puzzles of North America. 

The first of these is the mud minnow (Umbra limi), a 
small minnow-like fish frequenting muddy and reedy ponds 
near the Great Lakes as far east as Quebec. A closely allied 
species lives in streams and ponds in the eastern States from 
Connecticut to South Carolina. The only other relative of 
these two little fish, in fact the only other member of the, 
family of mud-minnows (Umbridae), is Umbra krameri found 
in Hungary. Dr. Gill* gives a most interesting account: 
of their habits. 

Blanding's pond tortoise (Emys blandingi) is the second 
example I have in my mind. It is a small species, the black 
carapace or shell being dotted all over with numerous yellow 
spots. It is entirely confined to the drift region from 
Wisconsin in the west to New Hampshire in the east. It 
agrees with Umbra in so far as its only relation inhabits 
Europe. The European species (Emys orbicularis) is 
common in central and southern Europe. A map of its range 
is given in my work on the European animals. f 

A more striking instance of a case of survival within the 
glaciated, or drift area, is afforded by the fresh-water pearl- 
mussel (Margaritana margaritifera). Being an immensely 
ancient species which no doubt originated in North America, 
it has taken advantage of old land connections to invada 
Europe and east,ern Asia. In North America it occurs within 
the ancient eastern land-mass in several localities, notably 
in Labrador, on Newfoundland and Anticosti Islands, in 
the Quebec province and in New England. J From its western 
station in the lower Saskatchewan it is separated by a 
tract of over one thousand two hundred miles in which 
it is unknown. This discontinuous distribution implies 

* Gill, Theodore, " The Umbras." 

t Scharff, E. P., " European Animals," p. 179. 

| Walker, Bryant, " Distribution of Margaritana margaritifera," p. 127. 



the destruction of the species in the drainage area of the 
Upper St. Lawrence. What other cause but the Glacial 
Epoch can be assigned for this partial destruction ? And 
in explanation of this and other similar cases it seems 
reasonable to conclude that certain areas within the drift 
region were entirely free from the devastating agency, be it 
land -ice or the sea, which destroyed so much of the pre- 
existing fauna and flora. We might argue, of course, that the 
pearl-mussel had advanced northward in two independent 
streams in post-Glacial times, one along the Mississippi valley 
and the other along the Atlantic coast, and that they had thus 
reached the northern habitats alluded to. But quite apart 
from the difficulties connected with pushing its way along 
the coast from river to river, there seem to me to be no 
grounds for such a supposition. The fact that Margariiana 
margaritifera is the only fresh-water mussel (Naiad) belong- 
ing to the Atlantic river system which is now found living 
within the Mississippi drainage area, shows how inconceiv- 
ably slowly these mollusks spread. The fresh-water mussel 
fauna, so characteristic of the Mississippi drainage area, is 
believed to have descended in a scarcely broken line from that 
of the Cretaceous Period, and many recent forms are only 
distinguishable with difficulty from those that lived in those 
remote times. 

However, the question of survival or non-survival in the 
drift area is not confined to the instance alluded to. Lampsilis 
(Unio) borealis, Lampsilis superiorensis, Anodonta mar- 
ginata, Anodonta kennicotti, and Anodonta pepiniana are all 
restricted in their range to the drift area, and are not known 
to have ever lived outside it.* 

There is no doubt ample zoological evidence for the assump- 
tion that many species of vertebrates and invertebrates 
survived the Glacial Epoch within the area commonly sup- 
posed to have been covered by vast ice-masses. But there 
is another important point which the geographical distribu- 
tion of the fresh- water pearly mussels may elucidate. Several 
times I have referred to the theory that in pre- Glacial times 
the land between northern Labrador and eastern Massa- 

* Simpson, C. T., " Synopsis of pearly freshwater mussels." 


chusetts was raised to a much higher level than ifc is now, 
and that it has gradually sunk to its present position. If this 
had been the case, the rivers on the western side of this, 
highland would have drained westward and would have 
formed part of the Mississippi drainage, while the eastern 
streams would have emptied their waters straight into the 
Atlantic. The latter ought therefore to show distinct signs 
in the composition of their fauna of having been isolated 
from the Mississippi drainage system. Dr. Simpson de- 
scribes the fresh-water pearl-mussels of the Mississippi 
drainage system as large, ponderous and ornamented with 
beautiful and odd patterns of colour and sculpture. The 
waters that drain into the Atlantic on the other hand are 
inhabited by a totally different set of mussels. They are 
moderate in size, frail and not remarkble for either colour 
or sculpture. Only very few of the eastern species extend 
westward to the head waters of the St. Lawrence.* 

In the interesting report on the Pleistocene fauna and flora 
already cited, there is a list of twelve species of fresh -water 
mussels that have been observed in the Don valley deposits. 
None of them belong to extinct species, though it is a signifi- 
cant fact that nearly one half of them no longer inhabit the 
St. Lawrence drainage, being now confined to the Mississippi 
and its tributaries.f 

This certainly implies that the Mississippi was more inti- 
mately connected with the upper St. Lawrence system in 
early Glacial and probably in pre-Glacial times than, it is now, 
for it is not unjustifiable to assume that these southern forms 
found their way northward before the commencement of the 
Glacial Epoch. It is possible that those southern species, 
found in the Don deposits, which still have a wide range in 
Canada, may have survived the Glacial Epoch in that country. 
Dr. Simpson J pointed out that almost all the Mississippi 
valley species of mussels now living in St. Lawrence drain- 
age area Jiave become more stunted and more simple in 
outline, while their characteristic sculpture and colours are 

* Simpson, 0. T., " On some Fossil TJnios." pp. 591592. 
t Dawson, J. W., D. P. Penhallow, &c., " Canadian Pleistocene Flora 
and Fauna," p. 331. 

J Simpson, C. T., "On Mississippi Valley Unionidae," p. 384. 


nearly obliterated. This change in form is so striking, 
says Dr. Simpson, that many of these northern races of 
Mississippi mussels have been described as distinct species. 
Dr. Simpson is of opinion that this remarkable change in 
form among the Mississippi species inhabiting the St. Law- 
rence drainage system has been produced since the Ice Age 
began to draw to a close, because, he argues, it is almost 
certain that all fluviatile and lacustrine life under the ice 
sheet was destroyed. 

The prevalent geological opinions as to the nature of the 
Ice Age thus dominate all biological thought in reference to 
problems of distribution. If we emancipate ourselves from 
these pre-conceived notions in our speculations on the origin 
of the existing fresh-water mussel fauna, we must arrive at 
different conclusions. 

Two almost diametrically opposed theories have been pro- 
pounded to explain the origin of the geographical features 
of the river system of north-eastern North America in pre- 
Glacial times. According to Professor Grabau,* Mr. Warren 
Upham and Professor I. C. Russell are of opinion that the 
greater part of the Laurentian basin formerly discharged its 
waters southward to the Mississippi. With this view Professor 
Grabau agrees. Much of the Archaean region in north-eastern 
North America must have stood greatly higher than at present, 
the slope of the land being toward the south-west. Two great 
rivers then existed, according to Professor Grabau, in place 
of the present Great Lakes, the Saginaw and the Dundas 
Rivers, both of them draining into the ancient Mississippi. 
Dr. Spencer is in agreement with Dr. Grabau in so far as he 
contends that the land stood at least two thousand feet higher 
than at present. He differs from the other writers in his 
belief that the lake region and the more north-easterly high- 
lands were equally affected by the uplift, so that the rivers 
which replaced the lakes in pre-Glacial times drained into 
the St. Lawrence just as the waters from the lakes do now.f 

The first of these two theories seems to me to explain the 
origin of the eastern fauna in a satisfactory manner. The 

* Grabau, A. W., " Geology of Niagara Falls," pp. 4254. 
T Spencer, J. W., "Falls of Niagara," pp. 400405. 


other does not. We may suppose that the species inhabit- 
ing the Mississippi during the period preceding the Ice Age 
were not precluded from advancing up stream towards the 
highlands of Labrador. But only the hardiest forms, those 
that could adapt themselves to waters probably brackish and 
laden with mud, might have succeeded in surviving, though 
not without becoming stunted in form and undergoing 
various other changes. Such forms as the pearl-mussel 
(Margaritana margaritifera), accustomed to pure mountain 
streams, only survived in isolated localities in the eastern and 
western parts of its range, becoming extinct in the central 
parts, where the conditions must have been less favourable for 
its survival. 

Let us now examine the land fauna of another part of 
this north-eastern Archaean land surface which is supposed 
by Professors Upham and Grabau to have stood at a rela- 
tively much higher level to the lake region than it does at 
present. Labrador has been amply dealt with, but New- 
foundland, which must have been completely isolated from 
the mainland for some time past, being a large island situated 
at the mouth of the mighty St. Lawrence River, ought 
to contain some interesting pre-Glacial relicts. We have no 
reason to assume that Newfoundland has been connected 
with the mainland since the passing away of the Ice Age, 
nor is there any evidence to show that mammals or other 
terrestrial vertebrates have reached the island by swimming 
across the Strait of Belleisle or down the St. Lawrence.. 
Geologists tell us that Newfoundland was not overridden 
by the huge Labrador ean glacier, but that it had a system 
of local glaciers quite independent of those of the mainland. 
Even if all the higher parts of the island had been buried 
in snow and ice, tracts of land near the coast must have 
remained free from ice, as in Greenland, and so have given 
shelter to the survivors from pre-Glacial times. This view 
is certainly strengthened by the fact that all the mammals 
hitherto observed on the island belong to well-marked varieties 
or species peculiar to it. The Newfoundland caribou, the 
only deer inhabiting the island, has antlers differing con- 
spicuously from those of other races of reindeer, and many 
authorities now recognise it as a distinct species under the 


name of Eangifer terraenovae. All reindeer are expert 
swimmers, yet the Newfoundland reindeer or caribou has 
never been known to cross over to the mainland. The follow- 
ing other gpecies of mammals are all peculiar to Newfound- 
land. The Newfoundland musk rat (Fiber obscurus), the 
Newfoundland vole (Microtus terraenovae), the Newfound- 
land arctic hare (Lepus bangsi), the Newfoundland lynx 
(Lynx subsolanus), the Newfoundland fox (Vulpes deletrix), 
the Newfoundland otter (Lutra degener), the Newfoundland 
marten (Mustela atrata). A bear and a wolf are also said to 

The fauna of Newfoundland, moreover, is characterised by 
the remarkable absence of all such mammals which we might 
reasonably expect to have come from Asia in comparatively 
recent geological times, such as the moose, wapiti deer, brown 
bear and many smaller species. The island does not seem 
to have been affected by the great stream of new-comers which 
poured into the country and reached other parts of eastern 
America, such as Nova Scotia. 

In speaking of the present land bird fauna of Nova Scotia, 
Dr. Trotter argues that it was derived from two faunal 
stocks, a more primitive boreal one which occupied the region 
from remote times, and a later transition fauna which in- 
vaded the peninsula since the re- elevation of the previously 
sunken isthmus. He speaks of this as a " venturesome state- 
ment," yet he suggests that many boreal types belonging 
to the genera Pinicola, Carpodacus, Loxia, Spinus, Sitta, 
Eegulus, Certhia and Parus may have occupied Nova Scotia 
even during the Glacial Epoch, since glaciers do not preclude 
forest growth, while food must have been abundant during 
the short breeding season.* 

It has been my intention in this chapter to briefly direct 
attention to the extreme north-east as one of the salient 
bio -geographical features of the American continent. More 
than fifty years ago Professor L. Agassiz spoke of " the 
zoological island of New England," which he described as 
encircled by a uniform combination of fresh -water animals, 

* Trotter, Spencer, " Land Bird Fauna of North -Eastern America," 
p. 226. 


being itself isolated in character and approaching more to 
that of the Old World.* He was then unaware that this zoo- 
logical island, which his keen insight and judgment had 
recognised, extended really far to the north of New England, 
that it formed part, in fact, of the north-western prolongation 
of the European continent in late Pliocene times, preserving 
until the present day many characters in common with the 
Old World. 

* Agassiz, L., " Fishes of the Tennessee," pp. 363364. 



LEAVING Newfoundland, the St. Lawrence and the region 
of the Great Lakes on our north-westward travels, we now 
enter the drainage area of the mighty Mackenzie River which 
conducts its muddy waters to the Arctic Ocean. We 
know from Mr. A. H. Harrison's * recent exploit that it 
is possible to journey down the one thousand eight hun- 
dred miles from the Athabaska landing and reach the Arctic 
Ocean entirely by boat. But we obtain a better idea of the 
fauna and flora of this vast Athabaska-Mackenzie region by 
crossing the country on foot, or by a perusal of the excellent 
report lately published by the biological survey of the United 
States Department of Agriculture. The task of furnishing 
this report was entrusted to Mr. Edward A. Preble,f a natura- 
list who already had the advantage of accomplishing the diffi- 
cult survey of the Hudson Bay region, and who had shown 
himself possessed of the necessary qualifications for such an 

The Mackenzie basin comprises a vast region of nearly 
700,000 square miles, or about six times the size of the 
British Islands. As is usually the case in regions covered 
by glacial drift, the country is studded with innumerable 
fresh-water lakes extending in a more or less connected system 
from Lake Superior to the Arctic Ocean. With the exception 
of a large area in the north, which is mainly outside the actual 
drainage basin of the Mackenzie, much of this region is 
entirely covered with forests. The principal trees are white 
and black spruce, the canoe birch, tamarack, aspen and balsam 

* Harrison, A. H., "In Search of a Polar Continent." 
t Preble, E. A., " Athabaska-Mackenzie Kegion." Compare also 
Macfarlane, K., " Mammals of North- West Territory." 


poplars, Banksian pine and balsam fir. With these are asso- 
ciated, generally in the form of undergrowth, a variety of 
shrubs, some of which have a continuous distribution through 
the forest zone, whereas others are more or less restricted 
in range. But the whole region between Hudson Bay and 
the mouth of the Mackenzie River is by no means all forested. 
North of a line drawn from the mouth of the Churchill River 
to the mouth of the Mackenzie, the country belongs to what 
are known as the " barren-grounds," which have already been 
alluded to in previous chapters. They are largely covered 
with short grass, moss and small flowering plants interspersed 
with patches of sedgy or peaty soil on which grow Labrador 
tea, crowberries, dwarf birches, and willows. Both flora 
and fauna, in fact, remind us vividly of certain parts of Green- 
land and Labrador. 

The only large mammals, as in Greenland, are the barren- 
ground caribou and the musk ox. The former is simply 
spoken of as the " deer " by the northern hunters. Zoologi- 
cally it belongs to the barren-ground form (Rangifer arcticus) 
which, as I remarked, seems to be closely allied to the Green- 
land reindeer. It is not so with the musk ox, as Dr. Kowarzik 
has recently shown. After a very careful and extended ex- 
amination of a number of skins and skulls of musk oxen 
from this region, he was able to show that they differed from 
those living elsewhere by the possession of a deep lachrymal 
pit and two mammary glands. The Greenland musk o<x, which 
belongs to the eastern group, has no lachrymal pit and four 
mammary glands. There are other minor differences clearly 
proving that Dr. Kowarzik's * Ovibos moschatus macken- 
zianus is much more than a mere race. It is, in fact, a species 
perfectly distinct from the on inhabiting Greenland. In 
spite of these differences, the casual observer is easily misled 
by the apparent external resemblances among all the musk 

The general similarity in the fauna of the Mackenzie 
region and that of Greenland is by no means super- 
ficial, and holds good to some extent even among the smaller 
kinds of beasts. The arctic hare seems much like the Green- 

* Kowarzik, E., " Der Moschusochs und seine Kassen," p. 120< 


land hare, and the lemmings resemble the creatures we have 
already met in the same country. Yet there are two perfectly 
distinct kinds of lemming in these barren-grounds. In winter 
they are easily discriminated since one of them does not 
change its rusty brown fur, while the other (Dicrosfconyx 
hudsonius) becomes white all over. The former we have not- 
met with before. It neither occurs in Greenland nor in 
Labrador, and appears to be a more recent immigrant 
to North America. This supposition is strengthened by 
the fact of the occurrence of this lemming (Lemmus 
trimucronatus) from Alaska as far east as Melville Penin- 
sula. In Alaska even another species of the same genus 
(Lemmus minusculus) occurs, and on the Pribilof Islands 
still another, the black-footed lemming (Lemmus nigripes) 
Beyond Bering Strait in eastern Asia other kinds make their 
appearance, whilst in Europe the common lemming (Lemmus 
lemmus) was well known long before the days of Linnaeus. 
Its remains having been discovered in many European caves 
along with those of other arctic animals, it is not unreason- 
able to assume that the genus Lemmus is of Old World origin 
and one of the many Asiatic immigrants that have invaded the 
North American continent in the past, though only a few of 
the latter have chosen the inhospitable barren grounds for 
their home. 

Another new form which we have not met with before 
is the ground squirrel known as the Hudson Bay spermophile 
(Citellus partryi). It is abundant all over the barren-ground 
region alluded to, the earth being occasionally riddled with 
their burrows. These " spermophiles " or " gophers " as they 
are sometimes called, occupy a somewhat intermediate posi 
tion between squirrels and marmots. Numerous species are 
known from North America, altogether about forty. In Asia 
and Europe there are fewer kinds of spermophiles, yet they 
do occur, and this circumstance, while tending to prove once 
more that Bering Strait was dry land, renders the search 
for the original home of the genus more difficult. The fact 
that in Europe Citellus (Spermophilus) only inhabits the east, 
having invaded the continent in Pleistocene times, and then 
retreated again towards the country whence it came, points 
to Asia as the home of the spermophiles. On the other hand, 


the remains of Citellus douglasi have been recorded from 
P,otter Creek cave in California, and those of Citellus tride- 
cemlineatus from the Conard fissure in Arkansas. Both of 
these deposits are regarded as Pleistocene in age.* If these 
caves were in Europe they certainly would be held 'to be pre- 
Glacial, for about fifty per cent, of the species contained in 
them are extinct, while Mastodon, Megalonyx and other 
ancient forms accompany the spermophiles in the Cali- 
fornian cave. The two spermophiles alluded to are still 
living in North America and have a wide range. If the genus 
had invaded Alaska from Asia in Pleistocene times, it is 
gcarcely credible that there would have been time for the 
development of the two species and their enclosure in these 
distant caves within the Pleistocene Period, especially when 
we consider that the whole of Canada is believed to have been 
buried under a thick mantle of ice. It is possible, of course, 
that there were two distinct invasions into North America of 
the genus Citellus, one having entered from the west, as I 
shall explain later on, in Miocene times, the later one from 
the north-west towards the end of the Pliocene Period. 

Among the carnivores of the barren grounds, the barren- 
ground bear (Ursus richardsoni), the wolverine (Gulo luscus), 
and the barren-ground wolf (Canis occidentalis albus) are 
the most noticeable. 

I have scarcely as yet attempted to bring the study of 
the geographical distribution of birds within the scheme of 
this present work, partly because our knowledge of their past 
history is so defective in comparison with that of many other 
groups, and partly because mountain ranges or narrow 
channels of the sea do not as a rule afford effective barriers 
to their dispersal. Occasionally, however, birds are extremely 
circumscribed in their range, and there can be no doubt that a 
study of their geographical distribution is well worthy of care- 
ful attention. If I had not a superabundance of better material 
for my purpose, I should gladly dwell a little longer on the 
problems suggested by their range. Many of them, however, 
are quite useless for our purpose, loons and auks, for instance, 
two groups of arctic birds attached to the sea coast. Many 

* Brown, Barnum, " Conard Fissure," pp. 166 167. 


others are circumpolar but too migratory. Only the genus 
Lagopus is of special interest, because it occurs in all the 
regions and countries so far discussed. It includes the 
American and Old World ptarmigans and the European 
grouse, all of which are more or less permanently resident in 
the countries they inhabit. 

The willow ptarmigan (Lagopus albus) breeds in the 
barren-grounds and further south. It has a very wide range, 
inhabiting northern Asia, northern Eussia, and northern 
and central Scandinavia. It is not found in Greenland or 
Newfoundland, being replaced there by the rock ptarmigan 
(Lagopus rupestris). The latter also lives in Iceland, 
western Asia, arctic America and Japan. The third American 
species is the white-tailed ptarmigan (Lagopus leucurus) 
which frequents the Eocky Mountain summits from Alaska 
to New Mexico. It is sometimes called the " white or snow- 
quail." Besides the willow ptarmigan, we have in Europe the 
common ptarmigan (Lagopus mutus), the red grouse (Lagopus 
scoticus), peculiar to the British Islands, and Lagopus hyper- 
boreus of Spitsbergen arid Franz Josef Land. Long ago I 
ventured to express the opinion that the genus Lagopus was 
of North American origin having thence spread to Europe 
and Asia.* 

Dr. Stejnegert has since pointed out that, by an unfortunate 
oversight, I had included Greenland in the range of the willow 
ptarmigan instead of the rock ptarmigan. He also argued 
that the willow ptarmigan of Scandinavia is more likely to 
have originated from the British red grouse than, as I thought, 
vice versa. Though I still believe that the genus had a 
North American ancestry, I quite concur with Dr. Stejneger 
in the view that the points he raised cannot be reconciled 
with the conclusions I formerly arrived at. The subject is 
certainly worthy of further careful study. At any rate, this 
instance shows clearly the much greater difficulties we have 
to contend with in tracing the geological history of birds 
than that of mammals. 

As we proceed southward from the barren-grounds we first 

* Scharff, E. F., " History of European Fauna," p. 336. 
t Stejneger, L., " Scharff's European Fauna," p. 105. 


enter a belt of more or less stunted timber before reaching the 
real forest. Yet even here quite a number of new forms of 
animal life are met with that are quite absent in the more 
northerly districts. They are mostly species occurring also 
in the forest zone of western Labrador. In referring to them 
in the last chapter, I alluded more fully to the geological 
history of the moose (pp. 32 33), but I dwelt also on the fact 
that the flying squirrel, musk rat, .chipmunk, woodchuck and 
others made their appearance as soon as we entered the 
forest from the east. And it is precisely these forms of animal 
life that make such a change in the Appearance of the fauna 
as we leave the Mackenzie region barren-grounds on our 
way to the great Canadian forest region. 

The flying squirrel (Sciuropterus sabrinus) of northern 
Canada and arctic America is also found further south. Other 
species of the same genus inhabit the United States. The 
skin at the sides of its body extends between the limbs in 
such] a manner as to act like a kind of parachute, butf nothing 
in the nature of actual flight takes place. The flying 
squirrel is rarely seen, being an entirely nocturnal crea- 
ture. The distribution of these flying squirrels is decidedly 
interesting. They are found from Florida to Alaska, and 
are only known as fossils from the Potter Creek Cave in 
California, They have almost certainly been in the country 
since Pliocene times. Hence Sciuropterus yukonensis, which 
is peculiar to Alaska, has, in all probability, survived the 
Glacial Epoch in that country. On the western side of Bering 
Strait, in Kamchatka, another species is found whose range 
extends westward as far as Scandinavia. Southward, as 
we cross China into Burma and India, we still encounter 
species of flying squirrels, and even on the islands of 
Java, Sumatra and Borneo. And all these belong to the same 
genus Sciuropterus. Such an extensive distribution implies 
that the genus must be an ancient one, and, indeed, we are 
acquainted with several species of Sciuropterus from the, 
Miocene and Pliocene deposits of France. This in itself is 
no proof that the flying squirrels originated in Europe. They 
may have spread there from Asia in later Tertiary times and 
have subsequently died out in Europe, only to be re-estab- 
lished more recently from a later Asiatic invasion. In any 


case, the presence of flying squirrels in America seems to be 
due to a migration from Asia, and that this took place at a 
time when a land bridge existed between the two continents 
is evident. 

Th,e musk rat (Fiber zibethicus) * derives its popular fame 
chiefly Irom the fact that thousands of its skins are annually 
exported from North America by the fur traders. Having 
adopted the more attractive name of "musquash" for these 
skins, the dealers have no difficulty in disposing of them. Like 
beavers, these creatures inhabit subterranean burrows on the 
banks of lakes and streams, and are expert swimmers, yet they 
are really more allied to the voles. Varieties of the common 
musk rat are met with from the Mackenzie River to Labrador 
and southward as far as New Mexico. In the Yukon district 
and westward Fiber spathulatus replaces the common musk 
rat ; and far to the east in Newfoundland the district dusky 
musk rat (Fiber obscurus) occurs. Only two other living 
kinds of musk rat are known to science. One of them 
inhabits the State of Oregon, the other the dismal swamp 
in Virginia. This genus, therefore, is one of the most 
typically North American we have noticed so far, and no 
doubt it has originated in North America. This view is 
entirely confirmed by fossil evidence, for no musk rat 
remains are known outside North America. The common 
musk rat has been observed in the Pleistocene deposits 
of South Carolina, New Jersey, and Pennsylvania, while 
the jaw of an extinct species (Fiber annectens) has been 
found by Mr. Barnum Brown f in the Conard fissure of 
Arkansas. The remarkable circumstance about this jaw is 
that the teeth it contains strongly resemble those of Neofiber, 
a sub-genus of Microtus. Hence this cave species forms a 
connecting link between the genera Fiber and Microtus, and 
this fact supports the opinion I ventured to put forward 
(p. 29) as to the American origin of the latter. 

Before describing some of the other small beasts, I must 
refer now to an animal which is not only the largest living 
terrestrial American mammal, but likewise one that we, from 

* Hollister, N., " Synopsis of Muskrats." 

f Brown, Barnum, " Conard Fissure," p. 197. 

FIG. 5. Map of North America, showing the original distribution of the Bison (lightly 
shaded), the range of the same animal after the completion of the Union Pacific 
Railroad (darkly shaded), and the present range in Yellowstone Park (black). 
(Reproduced with Dr. Hornaday's permission.) 

[To face p. 65. 


a European point of view, are accustomed to associate above 
all with North America, namely, the bison, or so-called buffalo. 

The bison (Bison bison) is now almost extinct in its wild 
state, yet here in the Mackenzie region, ,a little to the south of! 
the Great Slave Lake, are still found some wild herds of this 
magnificent creature, the last remnants of the millions that 
once roamed over the continent. 

Dr. Howard* contributed many years ago an interesting 
article to " Science " on the manner in which insects and 
other creatures are disseminated over the States through the 
agency of man. An equally instructive paper might be written 
on the manner in which man has been the means of destroy- 
ing a portion of our fauna. For no one can doubt that human 
agency alone is responsible for the rapid destruction of the 
bison and other animals. 

When the Spaniards landed in America in the year 1521, 
the bison was still plentiful in Northern Mexico. In the com- 
mencement of the following century the English found it in 
abundance in the neighbourhood of the present site of the city 
of Washington. No doubt the range of this huge ungulate 
extended over about one-third of the entire continent of North 
America. The extreme south-eastern limit was on "the coast of 
Georgia. The western boundary was in New Mexico. From 
these two southern localities to the shores of the Great Slave 
Lake in Canada, vast herds of bison were known to exist even 
in. the early parts of the last century. According to Dr. 
Hornaday's f graphic description, they lived and moved, as no 
other quadrupeds ever had, in great multitudes, like grand 
armies in review, covering scores of square miles at once. 
They were so numerous that boats were sometimes stopped 
by them in the rivers, and they threatened to overwhelm 
travellers on the plains. In later years they occasionally 
derailed locomotives and cars. One herd, seen by Colonel 
Dodge in 1871, only forty years ago, and described by him, 
extended for a distance of twenty-five miles and must have; 
included a million individuals. The Indians believed that 
these buffaloes issued from the earth continuously, the 

* Howard, L. O., " The Spread of Species by the Agency of Man." 
t Hornaday, W. T., " Extermination of American Bison." 

L.A- F 


supply being consequently inexhaustible. As long as these 
wild tribes only possessed the bow and arrow, bisons were 
comparatively safe from their depredations, in spite oJ the 
great drives that they were accustomed to organise. On the 
introduction of firearms, the Indians assisted civilised man 
to the best of their ability in his utterly wanton destruction 
of these helpless creatures. The building of railroads across 
the continent naturally hastened the process of extermination, 
the completion of the Union Pacific line dividing for ever 
the bisons of the United States into two great herds. These 
subsequently became known as the northern and southern 
herds. The great slaughter of the bison really only began 
in 1871. Four years later the southern herd had ceased to 
exist. The year 1881 witnessed a similar destruction of the 
northern herd, and at present this most picturesque and im- 
pressive member of the American fauna is practically extinct 
in its wild state, but for the small herds alluded to in northerii 
Canada (see Fig. 5). There are other small herds preserved 
in the Yellowstone Park and in some reservations in western 
Canada. In them the bison can no longer be said to live 
altogether in the wild state. 

I mentioned that the bison was looked upon as one of the 
most typically American species. Nevertheless, we have in 
eastern Europe a bison which is closely allied to the American 
species, and from the circumstance that it formerly roamed 
over a large part of that continent, it might be argued that its 
cousin from the New World is but a new-comer and in no way 
typical of America. A certain amount of support for that 
argument might be derived from the well-known fact of a 
Pliocene bison (Bison sivalensis) being known from India 
and Java and another from China. But in America there are 
likewise bison remains (Bison alleni) which were considered 
by Professor Marsh to belong to the Pliocene series, while 
Professor Cope described one even from Nicaragua and 
Southern Mexico (Bison scaphoceras) . More recently, how- 
ever, Dr. Lucas * has clearly demonstrated that Cope's bison 
is a sheep, and that Marsh's specimens are probably referable 
to the lower Pleistocene, so that it does seem likely after 

* Lucas, F. A., "Fossil Bisons of North America," pp. 756766. 


all that the ancestors of the American bison came originally 
from Asia. The exact geological period of this supposed 
Asiatic invasion will be discussed in the next chapter, when' we 
come to deal with Alaska. Some other extinct bisons are 
known from America besides those alluded to. Bison crassi- 
cornis, according to Dr. Lucas, is confined to Alaska. Mr. 
Lydekker, * on the other hand, considers it identical with 
Bison priscus, an extinct bison which ranged throughout 
arctic Siberia and a large part of Europe. 

South of the belt of stunted timber lie the vast forests 
of the Athabaska and Peace Eiver valleys, and a great zone 
of warmer country abounding in animal life, as we proceed 
in the same direction. A noble representative of the deer 
tribe which we meet here reminds us of the European red 
deer. In the States it is commonly known as the elk (Cervus 
canadensis), though it has no connection with the moose, 
which in Europe is known by that name. In Canada it is 
more generally called " wapiti deer." While the range of 
the moose seems to be on the increase in the Mackenzie 
Region, it is curious that the wapiti has become almost extinct 
in the northern parts of its former geographical range, and 
now only occurs there in small numbers. 

Mr. Thompson Seton f tells us that originally, that is to 
say about the beginning of the sixteenth century, the wapiti 
was found from the Mackenzie Region as far east as Boston, 
and as far south as Arizona and Alabama (see Fig. 6). Its 
destruction proceeded unchecked until the year 1895, when a 
change in public opinion took place. Henceforth the wapiti 
was protected ; and it is now actually on the increas ; e in 
Manitoba and along the Rocky Mountains from Alberta to 
the borders of New Mexico and also along the Pacific States. 

The wapiti has only been observed in a fossil state in 
Pleistocene deposits. We know from these records that its 
range extended in Pleistocene times southward as far as 
Florida and from North Carolina to New Jersey and Kentucky. 

The resemblance between the American wapiti and, at any 

* Lydekker, E., " Catalogue of Fossil Mammalia," II. , p. 24. 
t Seton, Ernest Thompson, "Life Histories of Northern Animals," I, 
p. 43. 



rate, the eastern European races of the red deer is so close 
that Dr. Caton * was sometimes at a loss to discriminate 
between them. The late Professor Nehringf also drew atten- 
tion to the resemblance of certain fossil deer antlers from 
central and eastern Europe to those of Cervus canadensis. 
As we proceed westward in Europe the distinctive cha- 
racters of the red deer become more prominent, and less so 
as we travel towards America through Asia. Dr. Tcherski,J 
I think, was the first to direct attention to the fact that the 
wapiti was a variety of the Siberian maral. And a similar 
opinion has more recently been expressed by Mr. Lydekker. 

It is many years since I argued that the Elaphus group of 
the genus Cervus, to which the Canadian wapiti and thie 
European red deer belong, originated in Central Asia, and 
I have not had any cause to modify my views in this respect. || 

Cervus canadensis differs from all other American deer and 
agrees with the great majority of the Old World forms in so 
far as the proximal portions of the lateral metacarpal bones 
persist. This is, as Sir Victor Brooke first pointed out, an 
important structural character. In the possession of these 
small leg-bones the Canadian wapiti stands entirely apart 
from its more distant relations of the American Continent, 
while clearly betraying its close relationship to the Siberian 
maral and the Old World deer generally. Its presence in 
America, therefore, forms one of the most powerful argu- 
ments in favour of the view that North America and Asia 
have been united by land within fairly recent geological 
times. We shall meet with so many similar examples during 
the course of this work, that one might almost be inclined to 
say that the matter requires no further proof, that it is an 
established fact. Notwithstanding, one of the most distin- 
guished authorities in America is not disposed to accept the 
view that there was such a recent land connection. Nor is 
it expedient for us to adopt the attitude of the modern 
glacialist who, in referring to the theory of the presence of 

* Caton, J. D., "Antelope and Deer of America," p. 214. 

t Nehring, A., " Tundren und Steppen," p. 203. 

\ Tcherski, J. D., " Das Janaland, &c.," p. 496. 

Lydekker, E., " Geographical History of Mammals," p. 315. 

'I Scharft', R. F., " Mammiferes de la region holarctique," p. 448. 


gigantic glaciers during the Ice Age, exclaims " this is no 
vain hypothesis or speculation founded on uncertain data, 
no mere conjecture which the light of future discoveries may 
explode. The evidence is so clear and so overwhelmingly 
convincing that we cannot resist the inevitable conclusion."* 

The Canadian wapiti deer is a typical example of a recent 
Asiatic immigrant. In some previous cases it was doubtful 
whether the species referred to was of truly American or 
Asiatic origin. Again in others we were able to trace a 
distinctly American ancestry, but not beyond the more recent 

I want now to mention an interesting creature whose 
lineage we can follow through several geological ages, and 
whose most remote ancestors had, as we might say, not 
a drop of foreign blood in their veins. The species I allude 
to is the Canadian tree porcupine (Erethizon dorsatus). It 
is in many respects a most striking and remarkable animal. 

In Europe, in fact everywhere in the Old World, porcu- 
pines are ground animals, and they usually dig their 
burrows in the earth. The American porcupines differ from 
them in many important anatomical characters. They, more- 
over, all live on trees, and generally have prehensile tails. 
We have here two entirely distinct groups, as among the deer, 
though externally all porcupines have a strong family like- 
ness. As far as trees grow right up to the barren-grounds the 
Canadian porcupine was once abundant, but yielding whole- 
some food and being easily killed it has been exterminated in 
many districts. Whenever it is undisturbed it is plentiful, feed- 
ing on leaves and bark, and it seems entirely unaffected by 
climatic extremes. The Canadian tree porcupine lives in 
Canada from the Rocky Mountains in the west to southern 
Labrador in the east, and enters the United States in Maine 
and New Hampshire. Formerly it occurred all along the 
Alleghenies, through Pennsylvania into Virginia and Ken- 
tucky to Arkansas. In these southern parts of its range it 
no doubt fell an easy prey to the Indian hunters who were 
always fond of its flesh. The destruction of forests on the 
arrival of the white man hastened its northward retreat. 

* Geikie, James, "The Great Ice Age," pp. 6768. 


There is no sign here of a post-Glacial centre of dispersal 
south of the drift area. The centre of dispersal, on the con- 
trary, lies in Canada and it is from there that it has spread 
southward. Yet the species had already come into existence 
when the sabre-tooth tiger and peculiar kinds of peccaries 
haunted the forests of Arkansas, for its remains have been 
found together with these extinct creatures in the Conard 
fissure.* It likewise lived in Pennsylvania at a time when 
the great Mylodon, Megalonyx and Mastodon still flourished 

Beyond the Mackenzie Eegion, in the far distant Alaska, 
there lives another porcupine very closely related to the 
Canadian species. This yellow-haired porcupine (Erethizon 
epixanthum), as it is called, ranges from Alaska through 
the Eocky Mountains and westward to the Pacific as far south 
as northern Mexico, thus exhibiting the same indifference 
to climatic conditions as its near relative. Both of these 
North American porcupines have short tails. In Mexico, 
Central and South America we meet with numerous species, 
all allied to Erethizon, but with prehensile tails, which 
considerably assist them in climbing trees. 

Now if the genus to which these tree porcupines belong had 
originated in Alaska or Canada, we should certainly expect 
it to have traversed Bering Strait into Asia while the wapiti 
deer and many other Old World forms poured into America. 
That it has not done so does not tend to disprove the assump- 
tion of the former existence of a Bering Strait land bridge. 
It only implies that the genus Erethizon is of southern 
origin, and has merely spread northward within recent geo- 
logical times. The south-western region, that vast country 
of mountains and plains which contains the most important 
centre of dispersal in North America, has no doubt given 
rise to the genus Erethizon. Of its past history we know 
nothing as far as North America is concerned. South of 
Mexico, as already stated, all tree porcupines possess pre- 
hensile tails, and are distinguished by other minor differ- 
ences from Erethizon. For these reasons they have been 
placed into the distinct genus Coendu. These southern forms 

* Brown, Barnum, "Conard Fissure," p. 166, 


of tree porcupines have apparently advanced northward along 
the isthmus of Central America and have reached Mexico 
within quite recent times. There are no grounds, however, 
for the supposition that Erethizon is a modified Coendu. 
When North and South America became joined by the exist- 
ing isthmus of Central America in Pliocene times, Erethizon 
no doubt was already an inhabitant of the northern continent. 
Since the two genera belong to the same family Coendidae, 
which differs fundamentally from the Old World family 
Hystricidae, we must assume that long anterior to the Plio- 
cene Period North and South America, or such portions of 
those continents which then existed, had already been united 
and then become disconnected again. These theories ,are 
not founded on zoogeographical data alone. Professor 
Osborn * urges on palaeontological grounds that North and 
South America were joined in Cretaceous and perhaps in 
early Tertiary times, and then separated again until the 

All this will be discussed in detail when we come to deal 
with Central America. I only mention the matter now 
because in the Santa Cruz beds of Patagonia the skeleton of 
a peculiar tree porcupine has been discovered and placed 
by Dr. Ameghino into the new genus " Steiromys." More 
recently, Professor W. B. Scottf re-examined these Steiromys 
remains, and noted the remarkable fact that they are more 
like the recent Erethizon of North America than any of the 
modern South American tree porcupines. He even argues 
that Steiromys is the direct ancestor of Erethizon. 

As this implies a former land connection between Pata- 
gonia and North America, independently of the remainder 
of South America, it largely aided me in the theoretical 
construction of an ancient land bridge between south.- western 
North America and Chile. J It is on this land bridge, I think, 
that the ancestor of the North American Erethizo-n wandered 
northward from Patagonia in early Tertiary times. All this 
will be more fully explained later on. 

* Osborn, H. F., "Progress in Mammalian Palaeontology," p. 99. 
t Scott, W. B., " Eeport of Princeton Expedition," V., pp. 413417. 
{ Scharff, E. F., " Early Tertiary Land-Connection," p. 521, 


It may be of interest now to inquire as to how far a group 
of loss mobile creatures than the mammals, the snails, for 
instance, can be classified into species whose ancestors were 
Asiatic, and such as were originally native to the soil. Of 
course we need only consider snails resident in the Atha- 
baska Mackenzie Region. A direct European influence 
among the fresh -water molluscan fauna is apparent, since 
certain species such as Limnaea stagnalis, L. truncatula, L. 
palustris, and Aplexa hypnorum inhabit the Mackenzie 
Region as well as Europe. How they have spread to America 
is not readily ascertainable. The points of resemblance 
between the two regions are clearly of long standing, but none 
of these species need necessarily have come from Europe 
direct, as they all inhabit Siberia as well as Canada. 

Among the land snails we also meet with forms familiar 
to the European conchologist, such as Pupa muscorum, 
Cochlicopa lubrica, Hyalinia radiatula, H. nitidula, Euconu- 
lus fulvus, Zonitoides nitidus, and others. All these are no 
doubt very ancient species, too ancient in fact to help us 
materially in our present inquiries* It is possible, moreover, 
that they have special facilities for accidental dispersal, that 
is to say for dispersal other than the ordinary mode of pro- 
gression on land, although my studies have not led me 'to 
believe in the efficacy of such a mode of conveyance in 
permanently stocking a country. 

The strictly American Pupa armifera and P. holzingeri, 
Vertigo ventricosa, V. ovata, Vitrina limpida, Patula solitaria, 
and P. striatella are all easily transported by flooded streams, 
and thus scattered far and wide. In spite of the fact that the 
land and fresh-water mollusks of the Mackenzie Region 
indicate that a distinct Old World influence is recognisable, 
nothing points definitely to a recent land connection with 
either Europe or, Asia. They do not confirm the view that any 
survival of the molluscan fauna through the Glacial Epoch 
has taken place in that region. On the contrary, the absence 
of species peculiar to the region implies, as among the 
mammals, that that part of America has only recently been 
stocked with animals from another portion of the continent. 
Whether we assume that gigantic glaciers covered the coun- 
try, or whether we argue that the ocean invaded it, the, 


ultimate effect would be somewhat similar. In reality it 
would not be quite the same, because in the latter case it is 
conceivable that the whole Rocky Mountain chain, which lay 
to the west of the flooded territory, could have supported a 
fairly rich fauna and flora. If a huge glacier really existed 
in the Keewatin district, it is inconceivable to suppose that 
any but the most scanty fauna and flora could have found a 
stray asylum on the few isolated " Nunatakkr " projecting 
beyond the Rocky Mountain ice-fields. 

The mollusks do not favour either theory. The mammals, 
on the other hand, show clearly a geologically recent invasion 
of Old World forms from the west, which is more compatible 
with what I may call the marine theory. 

Mr. Walker * contends that the Glacial Epoch was a great 
factor influencing the distribution of the North American 
molluscan fauna. I can only perceive signs of local destruc- 
tion. The distribution of the land and fresh -water mollusks 
in North America took place evidently long before this Glacial 
Epoch, and was not seriously influenced by the latter. 

Both theories will be more fully dealt with in the next 
chapters in which the animals, principally of Alaska, will be 
described. These animals afford us more material for our 
inquiries, being of great importance to the just appreciation 
and understanding of the varied origin and development of 
the American fauna. 

* Walker, Bryant, " Origin and: Distribution of Mollusca," p. 53. 



SINCE Alaska came into the possession of the United States 
its progress has been a rapid one. The sudden development 
of the resources of this vast country attracted public atten- 
tion and led to urgent demands for further exploration of its 
natural wealth. This work has been actively prosecuted 
within the last few years, so that we now possess some know- 
ledge of the geography, geology and mineral resources of the 
country. Situated in the north-westerly extremity of North' 
America, its area equals in size about one fifth of the United 
States (see Fig. 1). It is nearly three times as large as the 
German Empire. 

A broad mountain belt spreads along the Pacific coast of 
Alaska, its western prolongation being still recognisable in the 
chain of the Aleutian Islands. The north-eastern extension of 
the Eocky Mountains curves rapidly westward on reaching the 
Arctic Ocean and there forms a range almost parallel with the 
north coast of Alaska. Between these two mountain systems 
lies partly a plateau and partly a great plain drained by the 
Yukon Kiver, the latter sending its waters into the Bering 
Sea. About one quarter of Alaska lies within the Arctic 
Circle. Hence, from the standpoint of geographic position, 
it must be regarded as an Arctic province. The southern 
sea-board, on the other hand, which is exposed to the warm 
winds and waters of the Pacific Ocean, gives a comparatively 
mild climate to the entire southern portion of the territory. 
Thus Dr. Brooks * draws attention to the fact, which few of 
us probably realize, that the range of climate in Alaska is 
greater than that between Maine and Florida. 

The capital of the country, Sitka, has a slightly lower mean 

* Brooks, A. H., " Geography and Geology of Alaska," p, 11. 


temperature than Scotland, but it has a much greater rain- 
fall. This enormous precipitation on the Pacific coast has 
a striking effect on the vegetation. The whole of south-eastern 
Alaska is densely timbered with forests of spruce and hem- 
lock, also red and yellow cedar and other trees, their luxurious 
growth contrasting strongly with the open forests of the 
interior and the more northern treeless barren-grounds. Yet 
the lowlands of the interior are usually well wooded with 
spruce, cottonwood, birch, willow and alder. At a height of 
from about 2,000 to 3,000 feet above sea-level, forest growth 
ceases altogether. Here the climate becomes continental in 
character, with great extremes of heat and cold. In the 
northern parts of the country the surface everywhere is 
covered with dense growth of moss and grass, with an 
abundance of wild flowers, whereas the ground beneath is 
perpetually frozen. 

Alaska possesses many other features of unrivalled interest. 
The magnificent peaks clustering about Mount Fairweather 
as a centre, and containing Mount St. Ellas, Mount Logan 
and Mount McKinley, the latter with a height of 20,400 
feet being the loftiest mountain in North America, have 
excited the just admiration of those who have been fortunate 
enough to visit this region. Splendid glaciers here and there 
descend from the great mountain valleys down to sea-level, 
while further west we notice in Mount Wrangell a fine example 
of an active volcano. The vast glaciers have given rise to 
the popular idea that Alaska is a country of unlimited snow 
and ice. As a matter of fact they are almost entirely confined 
to the Pacific mountain system. 

Near the sea border the climate is, as I mentioned, much 
milder than in the interior of the country, yet an abundant 
precipitation from the warm moisture -laden winds here pro- 
duces the extensive snow-fields from which the glaciers are 
fed. On the land side of the Pacific Mountain range glaciers 
are very poorly developed. Further inland they are entirely 

Since a vast area of land in Alaska remains uncovered by 
ice, in spite of a low mean temperature, we must recognise the 
justice of Professor Whitney's assertion, that a still greater 
increase of cold would not cover those regions with ice, nor 


bring about the conditions of a Glacial Epoch. On the con- 
trary, it is evident that the opposite result would take place. 
The conditions of a Glacial Epoch, says Professor Whitney, 
are not to be sought for in a general diminution of tem- 
perature. " It is true," he argues, " that the regions in which 
the glaciers take their rise cannot be so warm that precipita - 
tion takes the form of rain ; but the required low temperature 
must be confined to a limited area, and be brought about by 
local causes, and not be something in which all the other 
portions of the earth are sharers. There must be copious 
precipitation, which, although locally in the form of snow, 
can in reality only be the result of a high mean temperature 
in other regions. In short, warmth, as a phenomenon of 
general occurrence, is more indispensably necessary than cold, 
because the latter may always be produced, where locally 
wanted, by increase of elevation or by the shifting of the 
ocean currents."* 

We now know that the glaciation of Alaska during the 
Ice Age must be regarded as an extension of the present 
system of glaciers. They were no doubt considerably larger, 
while glaciers also existed in the Endicott range in north- 
western Alaska. Nevertheless, no signs of former glaciation 
have as yet been discovered in the interior of the country. 
The latter is now generally believed to have been entirely 
unglaciated. If we adopted Professor Whitney's views, and 
endeavoured to picture to ourselves the conditions of Alaska in 
the Glacial Epoch, we must imagine the Pacific to have been 
warmer than it is now and the climate of Alaska much milder 
and more humid. A more luxuriant flora would, under such 
circumstances, have been able to flourish in the interior of the 
country, and consequently a richer and more varied fauna. 

And how much is this conception of the Glacial Epoch at 
variance with the theories generally propounded ! I may be 
excused, for example, for again citing Professors Chamberlin 
and Salisbury's f words, that one of the most marvellous 
features of the ice-dispersion during the Glacial Epoch was 

* Whitney, J. D., "Climatic Changes," p. 388. 

t Chamberlin, T. C., and E. D. Salisbury, "Geology," Vol. III. 
pp. 332333. 


the pushing out of the Keewatin ice-sheet from a low flat 
centre, without even a suggestion of a mountainous nucleus, 
to one thousand miles westward, while the Rocky Mountain 
glaciers were thrust eastward, but little beyond the foot- 

The ice from this great centre of dispersion is supposed 
to have reached the mouth of the Mackenzie River close to 
the borders of Alaska. Nevertheless, the same authors ac- 
knowledge that the plains of Alaska were apparently free from 
glaciation, even during the time when, two thousand miles 
further south, the waters of the Ohio and the Missouri were 
actually believed to have been turned from their courses by the 
encroaching ice-sheets. How can we reconcile the co-existenoe 
of these two extraordinary and altogether anomalous climatic 
conditions in adjoining parts of the same continent ? Surely 
there must be some mistake. At present there is far less pre- 
cipitation of moisture in the Keewatin region of Canada than 
in any of the western mountain ranges. The existing con- 
ditions of land and water must consequently have been entirely 
altered during the Glacial Epoch. Indeed, even our concep- 
tions of the nature of climates would have to undergo some 
change [before we can realize how this stupendous ice -sheet 
in the Keewatin region came to be built up, while Alaska was 
only able to form a few local glaciers. In a previous chapter 
(p. 46) I ventured to make some critical remarks on the 
supposed gigantic ice-sheets of the Glacial Epoch, and I hope 
to show now that the biological evidences are altogether op- 
posed to the views that the Ice Age was an epoch of excep- 
tional cold. My criticisms on the current beliefs in the land- 
ice theory are by no means new. We need only peruse the 
fascinating volumes entitled " Ice or Water " published by 
Sir. Henry Howorth, in which the claims of water as a power- 
ful agent in the formation of so-called glacial deposits are 
ably discussed. My object, however, in writing this work was 
not to investigate the origin of the Ice Age. This short 
digression into the domain of glacial geology merely serves 
to acquaint the general reader with s,ome of the special 
difficulties we have to contend with in explaining certain 
phenomena connected with the Alaskan fauna. 

Since the beginning of the last century it was known that 


fossil ivory might be gathered on the shores of Alaska. 
Several Kussian authorities reported on its occurrence there, 
animadverting to the fact that the remains of elephants had 
also been discovered on some of the Pribilof Islands. Within 
recent years several expeditions have been sent to Alaska 
from the United States with a view to discovering more about 
these and other remains of extinct animals. The first of these 
was dispatched by the Smithsonian Institution of Washington 
in 1904. Mr. Maddren, who had already visited the country 
several times and had travelled extensively in the interior, 
was charged with the expedition, and he issued an interesting 
report on his return. 

He contends that the lowest beds in Alaska in which- 
mammoth remains occur are the " lacustrine silts," which' 
form an extensively developed feature in the country. 
Scattered through these Pleistocene deposits we find remains 
of skeletons, isolated cheek-teeth, tusks and bones, the animals 
to which they belonged having probably died near the shores 
of the lakes in the bottom of which they became embedded. 
These fluvial and lacustrine beds of Alaska, with their occa- 
sional gravels, rest unconformably on the eroded surfaces 
of the .older formations. At the time the silts and clays were 
forming Alaska was for the most part, according to Mr. 
Maddren,* a low -lying country, characterised by enlarged 
rivers with slow drainage, and many lakes. 

The general conclusions arrived at by Mr. Maddren are, 
" that while remnants of the large Pleistocene mammal herds 
may have survived down to the recent period, and in some 
cases their direct descendants, such as the musk ox, have 
done so, most of them became extinct in Alaska with the close 
of the Pleistocene. The most rational way of explaining this 
extinction of animal life, says Mr. Maddren, is by a gradual 
alteration of the climate from more temperate conditions, per- 
mitting of a forest vegetation much further north than now, 
to the more severe climate of to-day. Recent changes, while 
checking the vegetation and thus reducing the food supply, 
have acted injuriously on animal life, only leaving those 
forms that were capable of adapting themselves to the new 

* Maddren, A. G., " Smithsonian Exploration in Alaska," pp. 25-^28. 


conditions. Mr. Maddren urges that there are no facts to sup- 
port the contention that the climates of the Arctic and 
sub-Arctic regions ever had been colder than they are at 
present. He thinks that there are no phenomena presented 
by the past history of these regions that require a more 
severe climate than that now existing to account for them. 
There are no ice deposits in Alaska, he says, except those of 
large glaciers, that may be considered of Pleistocene age. 
There are no ice-beds, he continues, interstratified with the 
Pleistocene deposits of Alaska. Finally he argues that the 
various forms of land ice, together with the deposits of peat, 
now existing throughout the Arctic and sub -Arctic regions 
of Alaska, belong to the Eecent Period, and that these deposits 
may be most conveniently and logically classified by their 
position with reference to the Pleistocene and Keeent forma- 
tions, and that the ice deposits cannot be differentiated satis- 
factorily into deposits of snow or water origin by their 
physical structure and character alone.* 

A second Smithsonian expedition to Alaska was dispatched 
a few years later under Mr. Gilmore. He noticed that the 
scattered remains of Pleistocene animals occurred throughout 
the unglaciated region of Alaska and the adjacent Canadian 
territory in three distinct deposits, viz., in what is locally 
known as " black muck," which is accumulated in the valleys 
of the smaller streams, then in the silt and clays already 
referred to, and finally in the more recent fluvial and alluvial 
beds. The best-preserved remains were obtained in the muck 
deposits, and in these occurred certainly several examples of 
primary entombment. A magnificent skull with tusks of the 
mammoth was discovered in the muck forty-two feet below 
the surface, near Dawson, in the Yukon territory in north- 
western Canada. Mr. Gilmore f is -convinced that the 
determinable species of mammals found are sufficient to 
show at once that they represent a typical Pleistocene 
fauna, some of them, such as the moose, caribou, musk ox, 
sheep, bear, and beaver, having persisted down to the present 

* Maddren, A. G., " Smithsonian Exploration in Alaska," pp. 65 66. 
t Gilmore, C. W., "Smithsonian Exploration in Alaska," pp. 2638. 


The complete list of mammals is as follows : 

Mammoth (Elephas primi- Extinct musk ox (Symbos- 

genius). tyrelli). 

Mastodon (Mastodon ameri- Bison (Bison crassicornis). 

canus). (Bison occidentalis). 

Horse (Equus sp.). ,, (Bison alleni). 

Moose (Alces sp.). Sheep (Ovis sp.). 

Reindeer (Rangifer sp.). Bear (Ursus sp.). 

Musk ox (Ovibos sp.). Beaver (Castor sp.) 

The results of these two expeditions excited so much interest 
in the United States that another was sent out to Alaska by 
the American Museum of Natural History under Mr. 
Quackenbush.* He was fortunate enough in discovering part 
of a mammoth skeleton which still retained the flesh, skin 
and hair similar to those found in northern Siberia. This 
discovery is of particular interest, as it elucidates an extremely, 
important feature connected with the preservation of the soft 
parts of mammoths in Siberia. It has been stated that the 
preservation of the bodies of mammoths is due to their being 
embedded in underground ice. Mr. Gilmore informs us that 
fossils in Alaska have never been found in pure ice masses. 
On the contrary, he makes it quite clear that the flesh, skin 
and hair of the mammoth were preserved to the present time 
because they had been pressed down into the soft mud by the 
weight of the rest of the^body. All that remained exposed 
of the hody rotted away ./The author agrees with the previous 
writers in considering uie mammoth and the mammalian re- 
mains as Pleistocene. He argues that the climate of this 
period must have been somewhat milder than that of the 
present time, because large trees had been met with associated 
with these mairdnalian remains in regions which are now tree- 
less. He also adds Bootherium bombifrons, the wapiti deer 
and the wolf to the Pleistocene species already known from 
the Yukon territory of Canada. 

From these three most valuable reports we can gather cer- 
tain remarkable facts which are apparently antagonistic to 

* Quackenbush, L. S., "Alaskan Mammoth Expeditions," pp. 125 


the theory of the existence of enormous ice -sheets in dis- 
tricts adjoining Alaska during Pleistocene times. All the 
writers agree that the numerous mammals .whose remains were 
discovered in Alaska lived there in the Ice Age during climatic 
conditions w'hich were vastly more favourable than those now 
prevailing. Evidences of a milder climate are furnished by 
the abundance of mammoth remains in Alaska. Horses, rein- 
deer and herds of bisons likewise roamed all over the country. 
Hence rich meadows and forests mus^b have been plentiful 
to support them all with food. Nothing whatsoever has been 
brought to light which would sustain the theory expressed 
by some writers that this fauna inhabited Alaska during a 
mild post-Glacial or inter-Glacial stage. No trace of the exist- 
ence in Alaska of a cold period preceding the supposed inter- 
Glacial or post-Glacial stage has been found. We must con- 
clude, therefore, that Alaska and the neighbouring countries 
during the whole of the Pleistocene Period had a temperate 

These heterodox views on the climate during the so- 
called Ice Age are not new. They were enunciated by 
Sir Henry Howorth and supported by able arguments 
more than twenty years ago. Very little was then known 
of Alaska, but the remains of innumerable large mammals 
had been unearthed in northern Siberia and even on the New 
Siberian Islands situated in the Arctic Ocean. These dis- 
closures naturally caused a good deal of speculation. They 
led to the conclusion that the cold of the Ice Age or Glacial 
Epoch was succeeded by a warm or temperate climate during 
which these animals were tempted to wander into the Arctic 
Regions. After carefully examining the problem, however, 
Sir Henry Howorth * expressed the conviction that this mild 
or temperate epoch did not follow the Ice Age, but pre- 
vailed during the whole of it. The creatures alluded to must, 
therefore, have lived in those northern regions during the 
Pleistocene Period, which was temperate in character. Since 
the close of that period Siberia has gradually become more 
and more inhospitable. 

A few years later the extremely valuable scientific results of 

* Howorth, H. H., " Mammoth and the Flood," p. 265. 
L.A. G 


Dr. Tcherski's * expedition to the New Siberian Islands were 
published by the Academy of Sciences in St. Petersburg ; and 
it is of interest to note that he arrived at conclusions similar 
to those of Sir Henry Howorth. He contended that the causes 
which produced glaciation in Europe acted in quite a different 
manner on the conditions of life in northern Siberia, where 
the climate remained as a direct continuation of that pre- 
vailing in Tertiary times, so that this country had a higher 
tejnperature during the Pleistocene Period than obtains at 

He looks upon Siberia as a country in which the processes 
of general refrigeration of the northern hemisphere and the 
deterioration of the conditions of existence for plant and 
animal life during the post-Tertiary Period took a regular 
and steady course. 

Baron von Toll concurs with Tcherski and other writers in 
their opinion that the large mammals whose carcases and 
skeletons are now unearthed in the extreme north of Siberia 
really lived there, but he urges that they did not arrive there 
until after the Glacial Epoch. Underground ice, as in some 
parts of Alaska, is extensively met with in northern Siberia 
and in the New Siberian Islands. Its occurrence, according 
to Baron von Toll, is due partly to a recent origin in crevices 
and lateral infiltration, partly to a former glaciation of Siberia, 
the ice having been preserved as " relict ice " from the Glacial 
Epoch. Being a follower of those who connect that period 
with a universal lowering of temperature, his contention is 
that northern Siberia enjoyed a milder climate in post-Glacial 
times, and that the mammoth, horse, rhinoceros and other 
mammals were thereby induced to visit the shores of the Arctic 
Ocean and the New Siberian Islands, which, according to 
Baron von Toll,f were then connected with Siberia and arctic 

The mammoth and many other animals, as I have indi- 
cated, invaded America from Asia. Consequently, if we can 
determine whether a land connection existed between the 
two continents, how long it lasted, and what species made use 
of it in their travels, some valuable data may be obtained as a 

* Tcherski, J. D., "Das Janaland," p. 469. 

t Toll, Eduard von, " Die fossilen Eislager," pp. 8085. 


basis for discussion of some of the problems referred to. 
Other attractive inquiries, such as the influence of the Glacial 
Epoch on animal migrations, might be elucidated. With the 
climatic problem alone are connected so many important geo- 
logical speculations that the subject is well worthy of a most 
careful scrutiny. 

At a meeting of the Biological Society of Washington about 
ten years ago the theme chosen for discussion was the sup- 
posed former land bridge between Asia and North America. 
Dr. F. A. Lucas spoke in favour of a geologically recent brief 
land connection, and supported his theory by the geographical 
distribution of the wild sheep and the brown bears of North 
America, both of which have their near relations in Asia. 
He also pointed to the remains of mammoth and bison in 
Alaska as evidence of the same view. An older land connec- 
tion, he thought, was indicated by the occurrence of a fossil 
Nemorrhaedus in Colorado and of Elephas columbi. Dr. 
Theodore Gill concurred with the last speaker, and suggested 
that the presence in America of such species of fishes as the 
pike and Scaphirhynchus tended in favour of the older land 
connection. Mr. F. V. Coville dwelt upon the importance of 
plants in sustaining the view of a transient and geologically 
recent land bridge. Additional testimony in favour of an 
old land connection of long duration was brought forward by 
Dr. Stejneger. He particularly emphasised the occurrence 
in Asia and America of the alligator and the salamander 
Cryptobranchus. Dr. W. H. Dall alone dissented from the 
opinions expressed, urging that such mammals as the 
mammoth might have crossed over Bering Strait by an ice- 
bridge which he thought might have joined North America 
and Asia during the Glacial Epoch.* 

Dr. Dall f gave no reasons for his aversion to the belief 
in a geologically recent land connection between North 
America and Asia ; although, in acknowledging that north- 
eastern Asia and Alaska have certain species of land and 
fresh -water mollusks in common, he recognised the intimate 
relationship existing between the two areas. 

* Lucas, F. A., Th. Gill, and others, "Former Land Connections." 
t Dall, W. H., "Mollusk Fauna of Alaska and Eastern Siberia," 
pp. 365366, 



The problem at issue is obviously not to be solved by the 
application of purely geological reasoning. Dr. Brooks, in 
his splendid contribution to the geology of Alaska, carefully 
avoids any discussion of a former land connection with Asia. 
In his brief statement that a land bridge between Asia and 
North America could only have been utilised for Glacial or 
post-Glacial migrations of Asiatic types, Professor Suess * 
does not make us acquainted with the reasons for his supposi- 
tion that the land connection did not also exist in Pliocene 
times. He alludes to the striking circumstance that the 
hypothetical land bridge was used apparently by Asiatic types 
only, and leaves us wondering why American types should 
not have taken a similar advantage of pouring into Asia. 
That, however, is part of the problem which we must reserve 
for a later stage of the discussion. 

Dr. Matthew f does not share Professor Suess' opinion, 
for he indicates a wide land connection, and Professor Osborn 
follows him in that respect, between North America .and! 
north-eastern Asia in the Bering Sea region in Oligocene, 
Miocene, Pliocene and Pleistocene times. 

As a rule the opinions of geologists on this problem are 
based on biological evidence alone. It is the latter, there- 
fore, that we have to rely upon. Curiously enough, Mr. 
Knopf maintains that all the stratigraphical testimony from 
which conclusions of some positiveness can be drawn, record 
only epochs of more widely-spread submergence and in- 
creased separation of the continents, although he recognises 
the strength of the palaeontological demands in favour of 
more or less continuous inter -continental communications. 
I shall endeavour, now, to critically examine the problem 
from that point of view. I have already had several oppor- 
tunities of expressing my own views on this subject (pp. 32 
and 68), and these are entirely in favour of a geologically 
recent Bering Strait land bridge (see Fig. 7). But we require 
to know more than this. We want evidence which will lead to 
the determination of the geological age of the bridge and the 
approximate date of its beginning and end. 

* Suess, E., " Antlitz der Erde," Vol. TIL, p. 764. 

t Matthew, W. D., " Hypothetical Outlines of Continents." 

| Knopf, A., " Probable Tertiary Land Connection," p. 419. 


Dr. Allen * considers the inter-relationship of the boreal 
mammals inhabiting North America and northern Asia so 
intimate that it could only have been brought about by the 
existence of a geologically recent land bridge connecting the 
two areas. He thinks that most of the more northern forms 
of mammal life on the two continents are probably the slightly 
modified descendants of types which formerly had a con- 
tinuous circumarctic distribution, and which have become 
slowly differentiated, mainly, no doubt, since the disruption 
of the former land connection at Bering Strait. 

Every naturalist who casually surveys the boreal fauna 
of North America and northern Asia must be struck by the 
apparently large predominance of Asiatic invaders in North 
America, such as the moose, bison, wapiti deer, American 
wild sheep, glutton, brown bears, the now extinct mammoth 
and many others, over the American element in Asia. In 
fact, we scarcely recognise any signs of an exodus having, 
taken place in the opposite direction. Typically American 
mammals, such as the common American musk-rat, the por- 
cupine, the black bear, and others, for example, have repre- 
sentatives in Alaska, nevertheless they are quite unknown 
on the western shores of the Bering Sea. At the same time 
Dr. Allen draws our attention to the occurrence along the 
Siberian and Kamchatkan coasts of certain mammalian types 
that are of distinctly American origin. These are a species 
of weasel (Putorius pygmaeus) closely related to the arctic 
American weasel (Putorius rixosus), and only remotely akin 
to any Eurasiatic species; a spermophile (Citellus buxtoni), 
closely allied to the boreal American Citellus paryi ; the Kam- 
chatkan bighorn (Ovis nivicola), which is more nearly related 
to American forms of sheep than to Asiatic ones, and several 

Whether the presence of these closely connected forms on 
both sides of Bering Strait proves the case of an American 
invasion of Asia appears to me still somewhat doubtful, 
especially as the sheep (Ovis) is, as far as we know, of Old 
World origin. A sheep is known from the Pliocene Forest 
Bed of England, while it only appeared in America in 

* Allen, J. A., " Mammals from North-Eastern Siberia," p. 183. 


Pleistocene times. Moreover, the Kamchatkan sheep (Ovis 
Iiivicola) is generally looked upon as a very near relation 
to another Siberian wild sheep also inhabiting part of Kam- 
chatka, and which has been found fossil by Tcherski in the 
New Siberian Islands, viz., Ovis borealis. To judge by the 
recent as well as by the fossil sheep in America, the genus 
Ovis, to which all sheep belong, must, I think, have pene- 
trated to North America, together with the mammoth and 
other mammals, in comparatively recent geological times. 
I shall return to the distribution of the American sheep 
later on. 

I should have thought the genus to which the musk ox 
(Ovibos) belongs was a better example of an American 
intruder into Asia. Although no longer inhabiting the Old 
World, its incursions into Asia and Europe must have taken 
place about the same time as the mammoth's advent in 
America. A still more striking instance of an American in- 
vader into Asia is the camel, although Professor Osborn's * 
statement, " in the Pleistocene the camels wandered into Asia 
from America, while the bears passed them en route .toj 
America," can scarcely be considered as strictly correct, 
since two kinds of camels are known from the Pliocene 
Siwalik deposits of India. The brilliant researches of 
American palaeontologists have long ago acquainted us with 
the fact that the camel family (Camelidae) inhabited America 
since the dawn of the Tertiary Era, while the Indian occur- 
rences alluded to are the earliest indications of camels having 
reached the Old World. It is in Pliocene times, therefore, 
or earlier even, that a land connection between America and 
Asia must have existed, for no one would venture to propound 
the theory that camels could have crossed from one continent 
to another on an ice bridge. 

Sir Henry Howorthf collected in 1892 some valuable testi- 
mony showing that the mammoth had lived in western Europe 
in pre-Glacial times. In the following year Dr. Tcherski J 
reminded us that a complete skeleton of the mammoth was 

* Osborn, H. F., " Faunal Eelations of Europe and America," p. 58. 
t Howorth, H. H., "The Mammoth and the Drift." 
I Tcherski, J. D., " Das Janaland," p. 474. 


discovered in an undoubtedly pre-Glacial deposit at Troizkoje, 
near Moscow. 

In Alaska the mammoth is the only elephant that has been 
found fossil. In the United States its remains are associated 
with those of two other elephants, viz., Elephas columbi and 
Elephas imperator, which, according to Professor Osborn,* 
are analogous to the European Elephas antiquus, and E. 
meridionalis. The ranges of these species overlap, yet their 
distribution seems to imply that the mammoth was the last- 
comer, or that it became modified in structure further south, 
giving rise to the two other species of elephants. The de- 
posits in which the mammoth occurs in the United States 
are generally looked upon as Pleistocene. But in the Potter 
Creek Cave in California it is associated with Mastodon, 
Megalonyx and other types belonging to extinct genera, f 
This cave contains nearly fifty per cent, of extinct species, 
and if situated in Europe would probably be classed among 
Pliocene deposits. 

Similar cases illustrating the invasion of Asiatic types by 
way of Alaska, and their apparent modification in character 
as they travelled southward, occur in almost all groups of 
animals. Sometimes, however, we meet with instances that are 
very difficult to explain by the assumption of their ancestors 
having utilised the land bridge alluded to. The magpie (Pica 
rustica), for example, is a handsome and strikingly coloured 
bird inhabiting the whole of Europe, North Africa, central 
and eastern Asia and western North America. Various races 
and varieties are recognisable among the Asiatic magpies, 
while the north. African form is likewise distinct. Curiously 
enough, the American variety resembles the European form 
more closely than it does any of the Asiatic ones, and it occurs 
from Alaska to Mexico, and eastward as far as the Missouri 
Eiver, whereas no magpies at all have been noticed in the 
extreme north-east corner of Asia. The distribution is there- 
fore somewhat discontinuous, the most easterly district in 
Asia where it is known being southern Kamchatka. In the 
extreme south-west of its American range in southern 

* Osborn, H. F., "Mammalian Palaeontology," p. 111. 
t Brown, Barnuni, " Conard Fissure," pp. 167 168. 


California quite a peculiar yellow-billed magpie has arisen, 
which is sometimes distinguished by the name of Pica nutalli. 

It is not at all evident that the ancestors of these American 
magpies entered America by way of Alaska, and yet if mag- 
pies had originated in America there would probably be many 
species there ; at any rate, they would inhabit the whole of the 
continent, and perhaps central America as well. Being con- 
fined to the west, and differing but slightly from Old-WorJd 
farms, their origin points to Asia. At the same time the fact 
that the American magpie is more like the European form 
than the Asiatic one is a feature which cannot be lost sight 
of. Dr. Diederich discusses the problem whether the 
American magpies arrived in America before or after the 
Glacial Epoch. In spite of the fact that magpies in America 
are entirely migratory and have lived on the continent suffi- 
ciently long to develop the peculiar Calif ornian race just 
referred to, Dr. Diederich * inclines to the belief that they 
entered the New World in post-Glacial times. Without any 
fossil or other evidence to guide us, it is difficult to express 
an opinion on this subject. But it seems to me that the 
magpie should be placed into one group with a good many 
other forms of animal life which reached North America long 
before that event, as I shall more fully describe in one of 
the subsequent chapters. 

Dr. Gill, as I mentioned above (p. 83), expressed the 
opinion that the occurrence of Scaphirhynchus in Asia pointed 
in favour of an older land connection between that continent 
and North America. The fish known in America as ;the 
shovel-nosed sturgeon (Scaphirhynchus platyrhynchus) is 
abundant in the river system of the Mississippi to which it 
is confined. It is now known that its nearest relatives are 
two species inhabiting central Asia which arena longer classed 
with Scaphirhynchus, but are placed into the distinct genus 
Kessleria. All these ganoid fishes belong to a very ancient 
group, most of which are extinct. The skeleton of the 
ganoids is generally cartilaginous, and they retain many 
other archaic characters in their anatomical structure. 
Another example is the paddle -fish (Polyodon spathula). It 

* Diederich, F., " Verbreitung der Elstern," p. 51. 


likewise lives in the Mississippi and its tributaries, while its 
next of kin 3 Psephurus gladius, occurs in the Yangtse Kiang 
and Hoang Ho rivers in China. These also are ganoid fishes 
and, like the true sturgeon and bony pike, are the only few 
remaining members of a very ancient and formerly abundant 
group of fresh- water fishes. The fact of their survival in 
these two continental centres might tempt us to adduce these 
instances in support of the theory stated above, that Asia 
and North America were recently connected with one another 
by land, whereas the migration of these fishes is an old story 
dating back to some remote geological period when, their ances- 
tors no doubt spread from one continent to the other by means 
of a land connection which probably lay further south. Their 
geographical distribution has obviously nothing to do with 
the subject under discussion, which concerns a geologically 
recent event. The cat-fishes or horned-pouts (Ameiurus) 
are possibly of a more recent origin. They swarm in every 
quiet stream and pond, especially in the eastern States pf 
America. A single species (Ameiurus cantonensis) is found 
outside the American continent, viz., in China. Finally 
I may mention that the " suckers " (Catostomus), which are 
very numerous in the American streams, are confined to 
North America, with the exception of one species (Catos,- 
tomus rostratus), which inhabits eastern Siberia. The last 
case certainly indicates a former land connection between 
North America and Asia, and consequently a dispersal from 
the former to the latter, at a geologically recent date. Still 
in the present state of our knowledge we are unable to deter- 
mine whether the passing across to Asia of these fishes coin 
cided with that of the great mammals whose dispersals I 
have endeavoured to elucidate. 

It is impossible to discuss the relationship of all the Asiatic 
to the Alaskan groups of animals, but no matter what class or 
order we examine with a view to testing this kinship, we 
find among almost all of them some indications of it. Dr. 
Horvath,* for instance, recently dwelt on the large number of 
species of Hemiptera (a group including bugs and allied 
insects) common to Europe and North America. He quoted no 

* Horvath, G., " Faunes hernipterologiques," p. 7. 


less than a hundred and twenty-two species found in temperate 
and northern Europe, and also in North America, reasoning 
that since the greater number of them also live in northern 
Asia they must have spread from one continent to another 
across Bering Strait. He does not definitely state that he 
considers the Strait to have been a land surface at the time 
these species are supposed to have passed across it. Still he 
emphasises the intimate relationship of the Alaskan and the 
opposite Siberian shore fauna, by citing the names of four 
species of Hemiptera which are confined to these two 

A rather striking example clearly illustrating the gradual 
advance of a genus across Asia, and its final immigration 
into North America, is shown by Parnassius, a genus 'of 
butterflies to which the European " Apollo " belongs. The 
Apollo group of butterflies originated, according to Mr. 
Austaut, towards the latter part of the Tertiary Era in Asia, 
at the time when some of the great mountain chains of that 
continent were being formed. I alluded to the European 
species of Apollo in my work on European Animals, gave a 
map of their distribution and enlarged upon their origin. I am 
in full agreement with Mr. Austaut's views as to the Asiatic 
centre of dispersal. The only point in which I differ is with 
regard to the period of its origin, which I. believe to have 
been considerably earlier than Mr. Austaut* thinks likely. 

In Europe we possess only three species of Parnassius, 
whereas there are thirty-five in Asia, Turkestan being the 
richest country in Apollos. Some species are peculiar to the 
Altai Mountains in Siberia, and as we go eastward new forms 
replace those with which we had become familiar. In Amur- 
land there are others, and yet when we cross Bering Strait to 
Alaska we still meet with an Apollo (P. nomion) which also 
occurs in the Altai Mountains, while the Calif ornian Apollo 
(P. clodius) is very closely related to Parnassius clarius of 
these same mountains. The Asiatic character has thus been 
retained by two of the American species. Two others, Par- 
nassius thor of Alaska and Parnassius sminthus, are quite 
distinct from any others, indicating that the passage from 

* Austaut, J. L., " Les Parnassiens," pp. 62 63. 


Asia to America of the genus is not a recent event. It 
must have taken place at a time sufficiently remote from 
the present to permit the gradual evolution of the two new 
species from the ancestral forms. This may possibly have 
coincided with the eastward advance of the mammals referred 
to. A land connection in the neighbourhood of Bering Strait 
would certainly have facilitated the dispersal of these butter- 
flies and other insects just as much as that of mammals. 

In a discussion on the relationship of the Asiatic and North 
American forms of another genus of butterfly (Vanessa), Dr. 
Standfuss * also supported the theory of the former land con- 
nection between Asia and North America in pre-Glacial times. 
The Glacial Epoch, he contends, subsequently segregated 
the butterfly fauna into insular districts in which many 
species survived, and whence they afterwards spread to other 

Before we consider the land bridge problem from the point 
of view of the marine fauna, some remarks on the general 
character of the present mammalian fauna of Alaska will 
be of interest. Both Labrador and Newfoundland, the two 
districts in the east which seem to have been little affected by 
the Glacial drift, and on which the still existing mammalian 
fauna probably survived from pre-Glacial times, were found 
to contain a certain number of peculiar species. We should, 
therefore, expect such a vast region as Alaska, which was 
also scarcely affected by Glacial drift deposits, to contain 
even a larger number of indigenous species of mammals that 
survived the Ice Age in the country. We do not positively 
know that any mammals survived the Ice Age in Alaska, but 
since we are unacquainted with any reasons why they should 
not have done so, that assumption is warrantable. The num- 
ber of mammals peculiar to the country is surprisingly large, 
and this alone implies that these animals inhabited the 
country for a sufficiently long time to develop characters 
distinguishing them specifically from those of the neigh- 
bouring parts in North America. The objection has been 
raised that American naturalists hold somewhat narrower 
views as to specific distinctions than are " current among 

* Standfuss, M., " Palaearktische Gross-Schmetterlinge," pp. 296298. 



European zoologists, and that many of their so-called species 
are nothing more than varieties or races. In many cases this 
criticism is not justified, but even if we regarded the forty- 
five species of mammals peculiar to Alaska merely as varieties, 
the fauna still retains a remarkably distinctive character which 
can only have arisen through long-continued isolation. I 
herewith give a list of the species of mammals confined to 
Alaska. In doing so it must be understood that I am not com- 
mitting myself to any opinion as to their specific distinct- 

Aloes gigas. 
Rangifer montanus. 
Oreamnos kennedyi. 
Sciuropterus yukonensis. 
Citellus osgoodi. 




Microtus operarius. 

,, unalascensis. 

,, kadiacensis. 


,, sitkensis. 


,, abbreviatus. 
Ochotona collaris. 
Tamias caniceps. 
Peromyscus sitkensis. 
Teonoma saxamans. 
Evotomys dawsoni. 

Evotomys wrangeli. 
Lemrnus minusculus. 


Fiber spatulatus. 
Lepus othus. 

Synaptomys wrangeli. 

,, dalli. 

Vulpes kenaiensis. 
,, beringensis. 
Ursus middendorffi. 
,, eulophus. 
,, kenaiensis. 
Mustela kenaiensis. 

,, arcticus. 
Sorex tundrensis. 
,, glacialis. 
,, pribilofensis. 

The late Professor Asa Gray was the first to draw attention 
to the relationship of the flora of America and that of Asia. 
He was struck particularly with the remarkable resemblance 
of some of the plants found in the eastern United States to 
those inhabiting Japan. This particular feature, however, 
will be discussed in one of the subsequent chapters. It was 
not till much more recently that botanists became acquainted 
with the intimate floristic relationship between Alaska and 


north-eastern Asia. Professor Engler * cites a large number 
of arctic-alpine plants, most of which are met with in these 
two regions, and extend southward in America along the 
Eocky Mountain chain. He argues that some of the species 
were probably capable of spreading from the one continent 
to the other under the existing conditions of land and water. 
Of others he does not consider that they could have drifted 
across the ocean by occasional means of transport, and he 
urges that only a former land connection in the north could 
satisfactorily explain their presence in Asia and North 

Finally a few observations on the land bridge problem from 
the marine faunal aspect may elucidate some doubtful points. 
Dr. Dallf divides the existing marine fauna of the north- 
eastern Pacific into three zones, viz., the Oregonian, Aleutian 
and Arctic. The Oregonian extends from Monterey to the 
Shumagin Islands. The Aleutian passes westward from the 
Shumagin Islands, includes the Aleutian chain, and reaches 
northward to the floating ice in Bering Sea, bounded by a 
depth of five hundred fathoms. The arctic zone stretches 
indefinitely southward beyond the five-hundred-fathom line, 
and is bounded in the north by floating ice, or water of thirity- 
two degrees. Some arctic species have crept southward to the 
northern islands of Japan and Cape Newenham, while a series 
of isolated arctic colonies occur from Cook's Inlet southward 
and eastward. 

Dr. Dall's views are mostly derived from a study of the 
marine mollusca, and it is interesting to note that in the 
main points they are confirmed by Miss Eathbun J after an 
examination of the northern Pacific Crustacea. She states 
that arctic species often continue southward through Bering 
Strait, along the west coast of Bering Sea, to Okhotsk Sea and 
the Kurile Islands. Some of them extend eastward to Puget 
Sound and even further south. Miss Eathbun noticed that in 
exceptional cases, for instance Philyra pisum and Cancer 

* Engler, A., " Entwicklungsgeschichte der nordl. Florengebiete," 
pp. 22 43. 

t Ball, W. H., " Marine Faunal Eegion of North Pacific," p. 206. 

J Eathbun, Mary J., " Decapod Crustaceans of North-West Coast," 
p. 6. 


amphiaetus, Japanese species are found to occur in approxi- 
mately the same latitude on the American coast, without 
obvious connection by way of Alaska. 

The Black Stream of Japan, the " Kuroshiwo," comparable 
to the Gulf Stream of the Atlantic Ocean, keeps outside the 
island chain of the east coast of Asia, sldrts the Aleutian 
Islands, and then makes itself felt on the south coast of 
Alaska. The fish fauna of the northern Sea of Japan has 
nineteen per cent, of species in common with the south coast 
of Alaska. With the Bering Sea the latter has about twice 
as many fishes in common. All of these are forms frequent- 
ing cold seas.* 

Dr. Dallf recently expressed the opinion, based on a 
study of Tertiary marine deposits, that the conditions indi- 
cated by the faunas of the post -Eocene Tertiary on the Pacific 
coast from Oregon northward are a cool temperate climate 
in the early and middle Miocene, a warming up towards the 
end of the Miocene, culminating in a decidedly more warm- 
water fauna in the Pliocene, and a return to cold, if not prac- 
tically arctic, temperature in the Pleistocene. Further south, 
on the Californian feoast, the Tertiary marine faunas, espe- 
cially those of San Pedro, have been very carefully studied 
by Professor R. Arnold. The Pliocene fauna, he remarks, 
though not quite similar to the fauna at present living off 
San Pedro, still contains many species which now only occur 
north of that locality. Many of these northern species are 
limited in range to the boreal waters north of Puget Sound. 
Hence he concludes that these Pliocene deposits were laid 
down in water much colder than that now found off San Pedro. 
In the lower beds of the Pleistocene, he continues, the cold 
climatic conditions prevalent during the later Pliocene were 
giving place to a warmer climate, which had its effect on 
the boreal species of San Pedro. Southern species gradually 
increase in number -while northern ones become scarcer. 
Finally the upper Pleistocene beds contain 14'2 per cent, 
of species that are only now found living south of San Pedro, 

* Schmidt, P., " Verbreitung der Eische im Stillen Ocean," p. 564. 
t Dall, W. H., "Climatic Conditions at Nome," p. 457. 


while 6*1 per cent, at present occur exclusively north of 
this locality. Professor Arnold* is thus led to the con- 
clusion that semi-tropical conditions prevailed during the 
deposition of the Pleistocene formation. He also emphasises 
the fact that the later Tertiary and Pleistocene faunas of 
Japan and the west coast of the United States resembled one 
another much more than the faunas of the two sides of the 
Pacific do at the present time. 

Nothing could be more contradictory than the two state- 
ments of Dr. Dall and Professor Arnold as to the climatic 
conditions prevailing in two portions of the Pacific coast of 
North America in Pliocene and Pleistocene times. It seems 
almost as if the deposits from which Dr. Dall derived his con- 
clusions were not contemporaneous with those that led Pro- 
fessor Arnold to pronounce the views just stated. It is 
scarcely possible to conceive that, while a warm -water fauna 
existed on the Oregon and Alaskan coasts in Pliocene times, 
California should have had a cold climate. Arctic conditions 
are then supposed to have supervened on the north Pacific 
coast. On the Californian coast, on the other hand, the cold 
Pliocene climate is stated to have been succeeded by a semi- 
tropical one during the Pleistocene Period. 

It is now generally recognised, I think, that Central 
America, in its present configuration, originated by a final 
union of pre-existing independent land-masses in Pliocene 
times. That an inter-oceanic current, now no longer exist- 
ing, might have produced altogether peculiar climatic con- 
ditions on the Californian coast in Miocene times but 
not later seems admissible. If we suppose that the Japanese 
"Kuroshiwo" current formerly sent part of its warm 
waters through a wider opening at Bering Strait into the 
Arctic Ocean, would it have had the effect of inducing 
the Mexican fauna to advance northward and the arctic 
fauna to pour southward towards the coast of Oregon ? 
I doubt, even under such geographical conditions, whether 
the Pleistocene faunas of California and Oregon could 
have differed to such an extent as described by Dr. Dall and 

* Arnold, Balph, " Marine Pliocene and Pleistocene of San Pedro," 
pp. 6567. 


Professor Arnold. At any rate, it is reasonable to infer that 
a rise to a higher temperature of the northern Pacific Ocean, 
coupled with an increased conveyance of Asiatic species to 
the American coasts and a northward advance of southern 
forms must have coincided with the closing of Bering Strait. 
And it was not until Pliocene times, according to Professor 
jJ.;P. Smith,* that the marine faunas of Japan and the western 
coast of America began to be remarkably similar, many species 
being identical. From this fact we must conclude that inter- 
migration between'the two continents had set along a northern 
shore line. During the preceding Miocene Period the .marine 
fauna of California consisted of endemic species mixed with' 
southern and circumboreal ones, but without any Asiatic 
admixture. Consequently there was probably a wide com- 
munication between the Pacific and the Arctic Ocean, favour- 
ing the entrance into the latter of a warm current which 
profoundly affected the Arctic Regions. The curious relation- 
ship to Pacific mollusks which is noticeable among some 
forms of the English Crag deposits may possibly date from 
this theoretical Miocene current, which may have carried 
marine species right across the Polar Seas to Europe. 

At the beginning of the Pleistocene Period, the same con- 
ditions existed, according to Professor Smith, as in the Upper 
Pliocene. As the waters of the Calif ornian coast gradually 
became warmer, he remarks, Mexican species began to .creep 
northward. But this, he says, does not mean that connection 
with Japan was cut off. The continuation of the conditions 
that permitted Japanese species to migrate to California, 
merely allowed marine animals to make their way up the 
American coast also. Here I must beg to differ from Pro- 
fessor Smith. If a change in the fauna of the upper Pleis- 
tocene of California took place as asserted, that change was 
in all probability due to a gradual sinking of the land in 
the north, for a moderate subsidence in northern Alaska at 
any rate has been recorded by Dr. Dall f during later Pleisto- 
cene times. A gradual modification was thus brought about 
in the disposition of land and water, the continents of Asia 
and North America slowly assuming their present shapes. 

* Smith, J. P., " Periodic Migrations," pp. 225226. 
t Ball, W. H., " Neocene of North America," p. 278. 


I have endeavoured to show in this chapter how the more 
striking instances of relationship among the animals of Asia 
and North America can be explained, how they arose and 
under what climatic conditions. That Alaska was once joined 
by land to the opposite shore of Asia has been almost uni- 
versally acknowledged by biologists and geologists. But for 
our purpose we required something more definite than a state- 
ment such as that of Professor Heilprin,* " it appears likely 
that the bears, swine, oxen, sheep, antelopes and elephants 
originated in the Old World, whence they were transplanted, 
by way of some land connection existing in the north^ into 
the New World." Even Dr. Arldt's f remark, "the bridge 
between Asia and North America remained until the Glacial 
Period/' is too vague. I think I have brought together suffi- 
cient data to show that a land connection existed in the 
region of Bering Strait during Pliocene times until the com- 
mencement, or perhaps until the end, of the early part of the 
Pleistocene Period. As far as I know, the only fact that can be 
brought forward against such an assumption is the discovery 
by Dr. Dall J of marine gravels at Nome, in Alaska, contain- 
ing shells which he believed to be of Pliocene Age. Consider- 
ing, on the other hand, that the gravels in question only con- 
tained one distinct species, it is quite possible that they may 
not be so old as Dr. Dall thinks. And even if their Pliocene 
age should be established by further researches, it only proves 
that the land connection was not so wide as .we are apt ,tq 
believe. These gravels do not disprove the existence of the 
land bridge, for Nome is situated to the south of Bering Strait 
on the shores of the Bering Sea. 

I have adduced testimony showing that Alaska, not only in 
Pliocene times, but also during part of the Pleistocene Period,, 
had a comparatively mild climate. Many of the larger 
mammals seem to have become exterminated within more 
recent times owing to a change of climate. There is no indi- 
cation of any wholesale destruction of the fauna during th-> 
Glacial Epoch followed by a recent introduction from the 

* Heilprin, A., " Geographical Distribution of Animals," p. 179. 
t Arldt, Th., " Entwicklung der Kontinente," p. 293. 
t Dall, W. H., " Climatic Conditions at Nome," p 457. 
L.A. H 


south. If such an event had happened we should not have had 
such a large percentage of peculiar forms of animal life in 
Alaska, and more southern forms ought to have found their 
way there, such as the American deer and many Cithers./* 

It would seem, therefore, as if both the Atlantic and the 
Pacific Ocean became closed in the north simultaneously and 
remained so for a considerable time (Fig. 7). The southern 
shores of both the (great land bridges were then under the direct 
influence of warm ocean currents resulting in favourable con- 
ditions for the growth of vegetation and the food supply 
for large mammals. The northern shores of the land 
bridges, on the other hand, were in immediate contact 
with a closed Arctic Ocean, whose waters would naturally 
have remained frozen for the greater part of the year. 
During winter the snowfall all round the northern Atlantic 
and northern Pacific Oceans was probably considerable. 
The land being, moreover, at a higher level, this would 
have resulted in the production of local glaciers. Marine 
transgressions from the Arctic Ocean then seem to have 
taken place across northern Russia, as I described in my 
work on the History of the European Fauna,* and across 
the lowlands of arctic Canada as indicated on pp. 46 49. 
My views on the Glacial Epoch and its nature are thus at 
variance with those held by most geologists of the present 
time. They agree with those put forward by Sir William 
Dawson,f and are more in accordance with the current 
opinions at the time when the Glacial Epoch was spoken of 
as the " Diluvial Age." 

It is very generally believed, as I mentioned before, that the 
climate in northern Europe and northern North America was 
very cold, and that all that vast region which is covered by 
the deposit known as " Glacial drift " had been invaded during 
the Glacial Epoch by thick masses of land ice, so as to destroy 
practically all life or drive it far southward of the southern 
limits of the drift. I stated in another chapter (p. 77) that I 
did not intend to make a special point in discussing the origin 
of the Glacial Epoch. I only incidentally bring forward a 

* Scharff, E. F., ''History of European Fauna," p. 172184. 
t Dawson, W., " Ice Age in Canada." 

V \ 

FlG. 7.- Map of Northern portion of Northern Hemisphere, with supposed land 
connections in early part of Glacial Epoch (shaded) and extent of sea (white). 

[To face p. 98. 


theory of its origin and nature which appear to me to agree 
better with the geological history of the boreal fauna and flora, 
as far as we are able to ascertain them, than the hypotheses 
which have been so widely accepted by scientific men. 

Professor Shaler * reminds us that the extension of rela- 
tively warm climates which has occurred at certain stages 
of the Earth's history is perhaps explicable in an equally 
simple manner as that of the Glacial Epoch. He expresses 
the belief that if Bering Strait were as rea'dily open to the 
warm stream of the Pacific or " Kuroshiwo " as the Atlantic 
is open to the Gulf Stream, the temperature of the region a*bout 
the North Pole would be lifted by at least thirty degrees above 
its present mean annual. Indeed, if the Glacial Epoch had 
been due to a simultaneous closing of the Arctic Ocean to the 
genial influences of both Atlantic and Pacific, a mild climate 
all over the Arctic Regions must have coincided with a more 
copious flow of the " Kuroshiwo " into the Arctic Ocean. And 
here apparently lies the great stumbling block to my theory 
on the origin of the Glacial Epoch. If the warmer tempera- 
ture in the Arctic Regions in Miocene and Oligocene times 
had been due to a wider Bering Strait, and, consequently, to 
the fact that a greater volume of the " Kuroshiwo " then 
poured into the Arctic Ocean, how are we to account for the 
faunistic affinities existing between Asia and North America 
during this part of the Tertiary Era ? There are numbers of 
animals in North America which have an Asiatic ancestry, hut 
could not have entered the Continent with the great invasion 
that I described as crossing the Bering Strait land bridge. 
(These and many other facts point to the existence of a land 
bridge between Asia and North America in early Tertiary 
times. All the same, several important features imply that 
before the Pliocene Period the " Kuroshiwo " really sent its 
warm waters altogether to the Arctic Ocean. 

As I mentioned above, no Japanese affinities are recognis- 
able in the American Miocene marine fauna. On the other 
hand, there is some evidence that the Miocene floras of 
Sakhalin and Japan were intimately related to the Miocene 
flora of North America. These apparently contradictory 

* Shaler, N. S., " Nature and Man in America," p. 143. 



evidences can only be due to the fact that a portion of the 
Pacific coast of North America was connected by land in Mio- 
cene times with Sakhalin and Japan, whereas the " Kuro- 
shiwo " flowed northward between these islands and the main- 
land of Asia. In the subsequent chapters the evidence which 
has led me to these conclusions will be more fully dealt with. 



FROM Alaska we retrace our steps to Canada, 'by that magni- 
ficient mountain range popularly known as the " Rockies." 
I have already mentioned that the Rocky Mountains prac- 
tically end near the shores of the Arctic Ocean. They begin 
in northern Mexico. Although we can approximately fix 
the beginning and end of this vast range of mountains and 
even its eastern border, the western boundaries are more 
vague and indefinite. In British Columbia, which is so famed 
for its grand and impressive scenery, its rugged mountains 
and great forests, the northern spurs of the Cascade Moun- 
tains appear to merge into the Rockies, so that it becomes a 
matter of some difficulty to discriminate clearly between the 
two. Further south the Rocky Mountains cross the high 
plateau of Wyoming, sometimes spoken of as the " Laramie 
region." We also meet here the complex mountain groups to 
which the name of " Stony Mountains " has been applied. 
South of the plateau the mountains again grow more irregular 
and lofty than to the north of it. Another great plateau covers 
part of southern Utah, western Colorado', New Mexico land 
northern Arizona. With a height of over 6,000 feet above 
sea-level, this region has suffered great erosion, and is deeply, 
trenched by fantastic gorges which intersect it in every direc- 
tion. The most famous of them, the Colorado Canon, is a 
clean-cut chasm, which, in the course of ages, has been slowly 
carved by the river to the stupendous depth of 6,000 feet 
in the horizontal strata. 

It is not only the lover of scenery, but particularly the 
naturalist and palaeontologist who appreciate the unrivalled 
attractions of the Rocky Mountains. These mountains, more 
over, have been the direct means of exposing what are probably 


the most valuable and extensive deposits of fossils in 

As we enter the United States, proceeding along the Rocky 
Mountains, we soon find ourselves in the midst of the Miocene 
and the lower Oligocene beds (White River) of Montana. 
Further south, in Wyoming, we come to the Wind River beds, 
while on our left to the feast lie the Wasatch deposits, both 
of which belong to the lower Eocene. Westward we cross 
into the middle Eocene Bridger and Washakie beds of 
Wyoming, and also the upper Eocene Uinta of Utah. Much 
further south we finally meet with the famous Puerco, Tor- 
rejon and Wasatch formations of the San Juan basin in New 
Mexico, which are held to be of basal and lower Eocene age. 
Owing to the labours chiefly of Leidy, Marsh, Cope, Scott, 
Osborn, Wortman, Matthew, Hatcher, and others, a most re- 
markable assemblage of fossils has been obtained among these 
immensely rich deposits. Our knowledge of the former 
inhabitants of North America has thus greatly increased 
within recent years, and has aided us in tracing the gradual 
changes of land and water that the continent has undergone 
in past times. Great efforts are now being made to work out 
the correlation of the North American mammal -bear ing 
horizons. I propose to return to this subject later on, and 
need not dwell on it any longer at present. 

Although glaciers have now almost entirely disappeared 
from the Rocky Mountains, abundant proofs have been left 
of their past presence in the shape of moraines, and polished 
as well as striated surfaces. These signs of former glaciation 
are very different from the thick mantle of drift that we 
noticed in Canada, and which is likewise attributed to the 
action of glaciers. Only the highest summits and the most 
elevated valleys of the Rocky Mountains were ever occupied 
by ice, and there does not appear to be any sign of a 'deposit 
in the whole range resembling the northern drift.f 

These glaciers, no doubt, owed their existence to a greatly 
increased precipitation of moisture in the Rocky Mountains 
during the Ice Age, for we possess quite an unmistakable 

* Russell, I. C., " North America," pp. 122136. 
t Whitney, J. D., " Climatic Changes," pp. 6472. 


record of a past humid period in the basin of the Great Salt 
Lake in Utah. All round the present lake we find a most 
interesting series of terraces at varying heights, which clearly 
represent ancient shore -lines. It has thus been established 
that in Pleistocene times, during its greatest development, 
the lake had the enormous area of more than 19,000 square 
miles, that is, nearly the size of Lake Michigan, with a depth 
of about 1,000 feet. This ancient " Lake Bonneville," as it 
has been called, has since been greatly reduced in size by the 
slowly increasing aridity of the country. Its diminutive 
descendant, the Great Salt Lake, is rapidly drying up, ita 
average depth being only twenty feet. Other large fresh-water 
lakes existed in the Great Basin during the Glacial Epoch. 

Sometimes it is customary to include the western Sierra 
Nevada and Cascade Mountains under the term " Rocky 
Mountains." Asa Gray and Sir Joseph Hooker have even 
added to this area that of the Great Basin in their account 
of the Kocky Mountain flora. More recently the distinctive- 
ness of these two mountain systems is being more generally 
recognised. The forests of the Kocky Mountains are dis- 
tinguished from those of the eastern States by the prevalence 
of the pyramidal evergreen conifers, whereas in the east the 
trees are round-headed and mostly deciduous. The endemic 
flora, consisting of thirty-three per cent, of all the plants 
found in the Rocky Mountains, forms a prominent element.* 

We have noticed that some of the alpine plants inhabiting 
the White Mountains are unknown in western America, for 
instance, Diapensia lapponica and Loiseleuria procumbens. 
On the other hand, a large number of the alpine Kocky Moun- 
tain species do not occur in the east, whereas a few, such as 
Rhododendron lapponicum, Arctostaphylos alpina, Rubus 
chamaemorus and Veronica alpina, are common to both, point- 
ing to the great antiquity of this element in the flora.f 
Many of the plants occur also in Asia ; others are closely 
related to arctic-alpine species. Most of the endemic element 
of the Rocky Mountain flora has clearly been derived from that 

* Gray, A., and J. D. Hooker, "Vegetation des Eocky Mountain 
Gebietes," p. 267. 

t Eydberg, P. A., "Composition of Rocky Mountain Flora," p. 870. 


of the lower slopes of the mountains, and to it belong the 
typically American alpine species. They reach their greatest 
development in the southern portion of the Eockies. 

One of the most remarkable features, which I shall have 
occasion to draw attention to again in another chapter, is the 
occurrence in the Eocky Mountains of some species of plants, 
which reappear in the southern Andes, without being repre- 
sented either in the Mexican highlands or in the tropical 
Andes. Among these, Professor Engler * mentions the 
dwarf gentian (Gentiana prostrata), the mealy primrose 
(Primula farinosa), the hoary draba (Draba incana) and 
the alpine foxtail (Alopecurus alpinus). All these are 
common well-known European species. A few more have 
since been added by Professor Bray f This small group 
of species is of peculiar interest, not only because some- 
what analogous instances of distribution occur among insects, 
but also because this feature is by no means limited to 
mountain forms. In another chapter I shall allude to : a 
number of species inhabiting the arid regions of south- 
western North America that likewise are quite unknown in the 
moist tropical and sub -tropical belts of country to the south, 
yet reappear in the extreme south of South America. Whether 
all these species possess particular facilities for dispersal 
over wide areas or whether former conditions of land and 
water were more suitable than they are at present for dis- 
persal will be discussed later on (p. 414). 

The fauna of the Eocky Mountains resembles the flora in 
its composition, except that among the higher animals, at any 
rate, the species are almost all endemic. Two kinds of animals 
in particular are associated with the Rocky Mountains in 
sportsmen's minds, namely, the mountain sheep, or big-horn, 
and the Eocky Mountain goat. The former used to. be known 
as Ovis canadensis. But from the systematic mill of the 
zoologist, as Dr. A. E. Brown J puts it, have been produced 1 
seven new species and sub-species, all very closely re- 
lated to one another. As already stated, the Kamchatkan 

* Engler, A., "Entwicklungsgeschichte," II., p. 256. 

t Bray, W. L., " Eolations of North American Flora," p. 713. 

J Brown, A. E., " Zoology of North American Big Game," p. 69. 


sheep (Ovis nivicola) is more nearly akin to the American 
sheep than it is to other Asiatic ones, and the occurrence 
on both sides of Bering Strait of such near relations 
forms one of the strongest buttresses for the belief in 
a geologically recent land connection between Asia and 
North America in the neighbourhood of Bering Sea. The 
various forms of American sheep are entirely confined to the 
western mountain region, where they are found from the 
Alaska mountains to the mountains on the long peninsula 
of Lower California, and eastward as far as Yellowstone Park. 
The home of the big-horn is the loftiest rim-rock of the high 
mountain plateaux, or the most rugged and forbidding bad- 
lands of the middle altitudes. In summer, says Dr. Horna- 
day,* its favourite pastures are the treeless slopes above the 
timber-line, and in winter it paws through the snows of the 
mountain meadows to reach the tallest spears of grass. When 
the raging storms and deep snows of winter drive the elk and 
deer down into the villages for food and shelter, the mountain 
sheep makes no perceptible change in its habitat. Its agility 
is nothing short of marvellous, and, from its wariness and diffi- 
culty of approach, it is a favourite object of pursuit of the 
experienced hunter. 

If, as it seems likely, the American mountain sheep has 
entered North America from north-eastern Asia within recent 
geological times, the fact of its having spread to Lower Cali- 
fornia and developed several distinct forms is an argument 
in favour of a pre- Glacial immigration. That sheep had 
already penetrated to North' America in Pliocene times is 
also proved by the discovery of the horn cores of a sheep 
(Ovis scaphoceras) in northern Nicaragua. f 

The comparatively dull-witted Kocky Mountain goat (Ore- 
amnos montanus) shares with the big-horn the almost inac- 
cessible peaks and ridges of the Rocky Mountains, but, being 
clumsy and slow, it rarely ventures far from its usual haunts. 
Unlike the sheep, the Rocky Mountain goat has a very local 
and discontinuous range. It seems almost as if its original 
home had been in the coast ranges of Oregon and Washington, 

* Hornaday, W. T., " Notes on the Mountain Sheep," p. 77. 
t Lucas, F. A., " Fossil Bison of North America," p. 756. 
t Grant, Madison, "The Eocky Mountain Goat," p. 9. 


and that it had then spread northward to British Columbia 
and southern Alaska, and eastward to the Rocky Mountains. 
Its whole distribution is thus confined to the north-western 
United States and western Canada. <The term " goat " is really 
a misnomer. It is not a true goat, nor yet a true antelope. 
We might with more justification call it a " goat-antelope." 

In some respects it resembles the European chamois 
(Rupicapra), in others the serow and goral (Nemorrhaedus) 
and the takin (Budorcas) . The last three are confined to Asia, 
and all of them, though generically distinct from the Rocky 
Mountain goat, belong to the same group of goat-antelopes. 

A species of Nemorrhaedus occurs in the mountains of 
Japan and northern China, and we might, with Dr. Brown, 
be tempted to assume that Oreamnos has arisen in America 
from some Pleistocene immigrant of the genus Nemor- 
rhaedus.* But surely the genus Oreamnos, to which the 
Rocky Mountain goat belongs, must be a very ancient one. 
The peculiar discontinuous range of the whole group implies 
antiquity. The remains of the mountain goat have been dis- 
covered in Potter Creek Cave of California, while Mr. Cragin 
is said to have described a fossil species of Nemorrhaedus 
from the Pleistocene of Colorado. I have not been able to 
trace the description or further particulars. At any rate, 
the more remote origin of Oreamnos is shrouded in obscurity. 
We certainly have no definite evidence that its ancestors came 
from Asia in Pleistocene times. They probably reached the 
continent much earlier. 

The Rocky Mountain goat occasionally falls a victim to the 
cougar or puma, better known in the west as the " mountain 
lion," which, like its prey, is a typically American animal. 
From a distributional point of view it is one of the most 
remarkable mammals, as it exceeds all others in the great 
extent of its range. Its utter indifference to climatic condi- 
tions is shown by its occurrence from southern Patagonia 
right through the tropics to western Canada. It flourishes 
from the plains of Florida to the regions of the permanent 
snows in the Rockies and Andes. It is true that those who 
take note of small differences no longer believe in one species 

* Brown, A. E., " North American Big Game," p. 75. 


possessing this vast range. Like almost all other mammals, 
the cougar (Felis concolor), which in all its essential habits 
and traits remains the same, whether living in mountain, 
open plain or forest, under arctic cold or tropical heat, has 
yet been split up into several distinct species. 

The nearest Old World relations of the cougar are the lion 
and tiger, both of which differ from it very strikingly in size, 
habit and colour. It is perfectly obvious, therefore, that it 
is not a geologically recent immigrant from Asia. Two very 
closely allied species of large cats, moreover, have been dis- 
covered in Pleistocene deposits in Argentina ; while the 
cougar itself has left its remains, along with those of extinct 
members of the cat tribe, in the Conard Fissure. Another 
large cat (Felis hillianus) has been found fossil by Professor 
Cope in the Blanco formation of Texas, this being now looked 
upon as middle Pliocene. Hence it is probable that the 
ancestors of the cougar already flourished in North America 
as well as in the southern continent in Pliocene times. The 
facts of its recent distribution seem to point to its having 
entered North America from the south, and it may pos- 
sibly have done so in Pliocene times when the northern 
continent became definitely connected with South America. 
Further details as to its early history are still lacking. 

As we descend the mountains through the forest belt and 
finally reach the foot-hills, we meet with two large ungulates 
whose acquaintance we have not hitherto had an opportunity 
of making. Both of these are confined to the western States, 
and are well known to the hunters of the Eocky Mountain 
region. The black-tail, or mule deer (Odocoileus hemionus), 
as it is often called on account of its big prominent ears, : 
seems at first sight not to be very different from the American 
elk or wapiti, except in size. But the latter belongs to quite 
a different genus. If we examine the antlers of the two more 
carefully, we notice that the brow tines are lacking in the 
mule deer. There are also distinctions in the skull, while 
the lower parts of the meta carpal bones of the front limbs 
are retained in the mule deer. The wapiti deer is descended 
from an Old World stock which, as I explained (p. 68), 
crossed, over from Asia by a land bridge in Pliocene or early 
Pleistocene times. In it only the upper metacarpals remain, 


and it differs in other fundamental characters from the mule 
deer, which has apparently no near relations in the Old World, 
and which we may justly call the true American deer. 

Fossil remains of the mule-deer have been found in the 
Conard Fissure.* The fossil bones of other deer from the 
Pleistocene of Illinois, Iowa, Wisconsin and Indiana, have 
been referred to extinct species of the genus Odocoileus to 
which the mule-deer belongs. No Pliocene or older traces 
of this genus have as yet been discovered in North America, 
if we adopt the generally accepted view of the Conard Fissure 
being of Pleistocene age. Altogether there are three fairly 
distinct species of the type of the mule-deer in North America, 
namely the one I have just described, the white-tailed deer 
(Odocoileus virginianus) and the black-tailed Columbian deer 
(0. columbianus) . 

It is of considerable interest to know that this genus lives 
not only in Central America, but right to the southern ex<- 
tremity of South America in Chile. This fact alone is remark- 
able, for nowhere else in the world are deer found south of the 
Equator. They are entirely absent from Africa and Australia. 
However, it is by no means the only noteworthy circumstance 
about this American group of deer. Those who are ac- 
quainted with the habits and life history of the deer tribe 
know that the young of deer with large branching antlers< 
at first possess no antlers. Afterwards small, simple and 
unbranched processes appear on their heads. From year 
to year they are shed and new ones take their place, and these 
are always a little more complex than the previous ones. The 
gradual development of the race seems to follow that of the 
individual. It is only in the more recent geological periods 
that deer with branching antlers make their appearance. As 
we go back to earlier deposits the deer skulls only bear simple 
antlers with one or two branches. In still older strata we 
meet with deer that were devoid of anflers, while they gene- 
rally possessed long canine teeth which no doubt were useful 
as organs of defence. It has been rightly argued that the 
complex antlers have only been developed in comparatively 
recent geological times, and that deer with simple antlers 

* Brown, Barnum, "Conard Fissure," p. 205. 


which we find living here and there in isolated districts, are 
more ancient in point of origin. 

In South America we have still living at the present 
moment deer of the type of the mule-deer, only smaller, 
with simple forked antlers. Other still smaller deer possess 
merely minute spike-antlers. Extinct deer, moreover, with all 
the different kinds of antlers, have been observed in South 
American Pleistocene and Pliocene deposits. One species 
(Odocoileus avius), which, according to Professor Ameghino, 
belongs to the group with complex antlers, has even been 
noticed in the upper Miocene of Argentina. 

Both fossil and recent evidence thus clearly points to South 
America as the source of these true American deer. If we 
supposed that the ancestors of the North American species 
of Odocoileus had penetrated northward in Pliocene times, 
when Central America assumed its present shape, we should 
have a reasonable explanation for the fact that the genus 
has never spread to the peninsula of Alaska, nor into north- 
eastern Canada and Newfoundland. 

What prevents the general adoption of the theory of the 
South American origin of this group of deer ? Clearly the 
fact that while the deer family (Cervidae) is represented" 
from the Oligocene to the most recent deposits in Europe, 
it only makes its appearance in South America in the upper 
Miocene. The original home of the family is therefore 
believed to be in the northern hemisphere, and this assump- 
tion is strengthened by the circumstance that nowhere except 
in South America have deer penetrated to the southern hemis- 
phere. Since it is inadmissible to argue that mammals so 
near akin as the Old World and New World deer should have 
appeared quite independently of one another in two distinct 
centres, these affinities can only be explained by migration 
from the one centre to the other. According to most palaeonto- 
logists who expressed an opinion on this problem, the South 
American ideer could only, for the reasons stated, have entered 
South America from North America. Whether they were 
developed in the Old World or the New, it is evident, remarks 
Mr. Lydekker,* that the American deer originated in the 

* Lydekker, B., " Deer of all Lands," p. 245. 


northern hemisphere, and that they are comparatively modern 
immigrants into South America, where they now attain their 
maximum development. 

Mr. Lydekker gets over the difficulty of the simple-antlered 
South American deer, or brockets as they are called, by the 
admission that because they are unknown in North America, 
they are not ancestral forms. He thinks they should be 
regarded as degraded or arrested types of the group (p. 296). 

Professor Marsh* was inclined to look upon the North 
American Leptomeryx as the probable progenitor of the 
Cervidae. His suggestion led to further researches on the 
part of Dr. Matthewf who supplied a connecting link in the 
chain of ancestry of Odo,coileus in the Miocene Blastomeryx. 
That Dr. Matthew's view, however, is not generally accepted 
may be gathered from Professor Osborn's f recent remark in 
reference to the Pleistocene Period in North America, that 
among the newly entering northern forms are Odocoileus, 
Ursus and Erethizon. Professor Osborn's opinion is that 
the origin of the Cervidae will probably prove to be Asiatic. 
I quite concur in the view that they are of Old World origin, 
and yet I hold that the ancestors of the North American 
Odocoileus have invaded the northern continent from South 
America. The remote ancestors of Odocoileus must, there- 
fore, have penetrated from the Old World to South America 
without attaining North America. How they have done so is 
the problem I shall endeavour to solve. 

Later on, when we come to deal with the zoogeographical 
relationship of South America and Africa, I shall show that 
we possess valuable evidence for the belief in a former land 
connection across the southern Atlantic between these conti- 
nents. This, however, will not help us in explaining the deer 
problem, because no deer have ever been found fossil in Africa 
south of the Sahara, and those species which have succeeded in 
establishing themselves in northern Africa have clearly done 
so in recent geological times. Deer are absent from all the 

* Marsh, 0. 0., " Introduction of Vertebrate Life in America," p. 36. 

t Matthew, W. D., "Osteology of Blastomeryx," p. 535. 

J Osborn, H. F., " Cenozoic Mammal Horizons," p. 88. 

Osborn, H. F., "Ten Years Progress in Palaeontology," p. 107. 


remainder of the continent. On the other hand, there is living 
in Europe and Asia at the present time a genus of deer 
(Capreolus) which has several important characters in 
common with Odocoileus and its more primitive South 
American relations. The genus Capreolus, which includes 
the roedeer, is distinguished from all the other Old World 
deer in its tele-metacarpal front limbs, that is to say in the 
possession of the lower remnants of the lateral metacarpal 
bones. It resembles, as we already know, in this character 
the true American deer. Moreover, as Mr. Eorig * has 
pointed out, this is not the only feature of resemblance be- 
tween Capreolus and Odocoileus. The antlers of the former 
likewise agree with those of the New World deer, rather than 
with those of the Old World. Capreolus has the backwardly 
directed tine of Odocoileus, and lacks the brow tine of Cervus. 
Even in the period of renewal of the antlers, the roedeter 
agrees with the American deer, this change taking place in 
the winter months, while it occurs in all the other Old World 
deer in the spring and summer. Only three kinds of roedeer 
exist at present. In Miocene and Pliocene times, however, 
France and Germany were tenanted by quite a large 
assemblage of tele-metacarpal deer, all of which lacked the 
brow tine like Capreolus. The earlier history of these deer 
is largely obscured by the circumstance that only fragmen- 
tary parts of the skeleton are known. Thus the Miocene 
species of Dicrocerus, which is supposed to be related to the 
living Cervulus, possessed antlers that can be almost matched 
by some of the recent South American mountain forms of 
Odocoileus, whereas other South American forms (Blasto- 
cerus) remind one of the modern roebuck. f 

I venture to think that all the deer of South America have 
originated from one or more ancestors which invaded that 
continent direct from western Europe in early Tertiary times. 
Although it is true that we possess little palaeontological evi- 
dence in support of such a theory, a land connection must 
then have joined Europe with South America. The prob- 
able period of this migration from Europe to South America, 

* Eorig, Ad., " Wachstum des Qeweihes," p. 424. 

f Eorig, Ad., " Phylogenie des Cervidengeweihes," p. 542. 


as well as the nature of the land bridge which enabled the 
deer to reach western America rather than the east, will be 
discussed in one of the subsequent chapters. Some of the 
more primitive forms still survive in South America, where 
they have now been pressed into the mountain regions. The 
newer and more vigorous types must have passed into North 
America as soon as that continent became definitely connected 
with South America in later Tertiary times. 

In speaking of the western North American fauna, Pro- 
fessor Osborn * tells us that in middle Miocene the peculiarly 
American Hyper tragulidae disappeared, while the European 
Cervidae and the distinctly American Merycodontinae took 
their place. Professor Osborn alludes no doubt to Palaeo- 
meryx and Blastomeryx which seem to have originated in 
southern Europe, and spread subsequently eastward to India 
and onward to America. Both apparently became extinct in 
North America before the advent of Odocoileus from the 

The other large hoofed animal I alluded to as frequenting 
the lower slopes of the Rocky Mountains is one of the most 
peculiar creatures in existence. It is so different from other 
animals that it occupies the exclusive position of being the 
solitary member of a distinct family. Known among zoo- 
logists as the "prong -horn" (Antilocapra americana), and 
among hunters as the " antelope," this splendid animal 
possesses the graceful movements of the latter, while its horns 
have a superficial resemblance to the antlers of a deer. There 
is not the least real likeness, however, between the antlers 
of the prong-horn and those of a deer, for they are not solid, 
but hollow like the horns of a goat. The horn -sheaths,! 
like the antlers of a deer, are shed and reproduced at regular 
intervals. The prong -horn ranges from eastern Mexico to 
Saskatchewan in Canada and from the Missouri River in the 
east to the Cascade Mountains of Oregon in the west. Hence 
it is now a peculiarly western species, while it had crossed 
the Mississippi during the Pleistocene Period, and roamed 
about Illinois and Wisconsin. The closely-allied extinct 
genus Merycodus (Cosoryx) made its first appearance in the 

* Osborn, H. F., " Cenozoic Mammal Horizons," p. 77. 


" Deep Eiver " deposits of Montana in middle or lower 
Miocene, and persisted until the Pliocene. During the latter 
period several new genera, viz. : Capromeryx, Platatherium 
and Leptotherium, branched off from the ancestral stock, and 
made their way into newly-opened areas. The last two have 
occurred in recent beds in Brazil, the other in Nebraska. 
Thus the family Antilocapridae to which the recent prong- 
horn and all these fossils belong, originated in America and 
never left it. 

Among the birds of the Bocky Mountains the dippers are 
very characteristic. They are quite unrepresented on any 
of the eastern mountain systems, being also absent from 
Greenland and Labrador. In the Old World they are more 
or less confined to the mountains, but not nearly to the same 
extent as in America. In my " European Animals "* I have 
already cited Dr. Stejneger's interesting article on the geo- 
graphical distribution of the dippers (Cinclus), and I may 
again briefly summarise the results of his studies. He places 
the origin of the genus on the great plateau adjoining 
northern India. At the dawn of the Tertiary Era the species 
radiated from this centre east and west. America b'eing then 
connected by land with northern Asia, the ancestors of the 
present dippers had special facilities for crossing to the New 
World from Asia. They are supposed to have spread since 
along the Bocky Mountains and Andes to the very furthest 
end of South America. Dr. Stejneger's theoryf certainly ex- 
plains the existing range of Cinclus in a satisfactory manner, 
but there seems to me still another view of looking at the 
problem. If, as I believe, the Atlantic Ocean was bridged 
over by land in the direction of southern Europe in early 
Tertiary times, it is possible that the dippers may have been 
introduced into the New World in that manner. At any rate, 
the problem is worth considering from that point of view, 
when the dippers come to be worked ou,t in -a thoroughly 
systematic way. Among the older forms of both animals and 
plants there are in western America a surprisingly large 
number of groups which are closely related to European ones. 

* Scharff, E. F., "European Animals," p. 200. 

t Stejneger, L., " Geographical Distribution of Cinclus," p. 425. 

L.A. I 


In a later chapter I shall mention many of these. I would 
only here draw attention to a striking botanical example 
which seems to me due to direct migration from Europe to 
western North America and not by way of Bering Strait. 

Anemones are so much cultivated now in gardens that 
almost everyone is familiar with them. Their geological 
history no doubt has been a very remarkable one. Not a 
single fossil anemone is known to science, because the plant 
contains no part that might readily be preserved, and yet it 
can be asserted that the genus must have originated in very 
remote times. The occurrence of many species in isolated 
mountain regions, the extremely discontinuous and wide 
range of others, and especially the high percentage pf 
endemism, clearly imply that we have to deal with an ancient 
genus. Dr. Ulbrich,* to whom we are indebted for a splendid 
monograph of the genus Anemone, is of opinion that some 
of its sections were already developed in early Tertiary times. 
His view is, of course, entirely derived from what we might 
call circumstantial evidence, just as Dr. Stejneger's was in 
regard to the age of the genus Cinclus. One of these species 
of anemone (A. baldensis), a well-known alpine plant, is, 
according to Dr. Ulbrich, probably of Miocene age. It grows 
also in the Carpathians, the Apennines, the Pyrenees and 
northern Spanish Mountains, but nowhere in Asia. Never- 
theless, the same species occurs in the highest elevations of 
the Sierra Nevada in California, in the Cascade and Bocky 
Mountains. A very closely-allied form of anemone (A. teto- 
nensis) lives at a height of over 10,000 feet in Idaho, and 
another (A. jamesoni) at about the same elevation in the 
Andes of Ecuador. All these nearly related forms, there- 
fore, are confined to Europe and western America. 

Among all the older American forms both of animals 
and plants, other possible routes of migration besides the 
Bering Strait one have to be taken into consideration. Be- 
turning to the birds again, it seems to me that the genus 
Begulus to which the European golden-crested wren belongs, 
must have entered North America in Pliocene times along 
with the great mammals alluded to in the last chapter. In 

* Ulbrich, E., " Qeograph. Verbreitung d. Gattung Anemone," p. 325. 


North. America we, have Regulus cuvieri, only one specimen 
of which is known, Regulus satrapa and Regulus calendula. 
None of these have succeeded in penetrating southward 
beyond Guatemala in Central America, the genus having a 
wide range on the continent. Hence it is reasonable to 
suppose that the ancestors of Begulus invaded North America 
from Asia by way of the Bering Strait land connection. 

Among the invertebrates of the Rocky Mountains the 
beetles and butterflies are probably the best-known groups. 
They may be considered by some as of little importance in the 
solution of such problems as we have been dealing 'with,, 
because these insects are generally believed to be very liable 
to accidental dispersal. One of the most powerful distri- 
buting agents of insects subject to accidental dispersal is no 
doubt the wind. Nevertheless many naturalists, who have 
made a serious study of the geographical distribution of 
animals and plants, have come to the conclusion that neither 
wind nor other agencies of accidental dispersal are of such 
paramount importance as we are often led to believe. The 
species of a genus, even of butterflies or beetles, as a rule', 
are clustered round a centre from which we can easily 
imagine them to have been slowly dispersed in the course 
of time. Usually we can trace an intimate relationship 
between the species whose areas of distribution adjoin 
one another. The conditions of dispersal, in fact, even 
among winged insects, must be quite similar to those 
with which we are acquainted among the higher mam- 
mals. The latter spread gradually on land from their 
centre of origin. Sometimes we meet with allied groups 
of species among beetles and butterflies whose ranges are 
separated by extremely wide areas in which no near relations 
occur. We might be tempted to attribute such instances to 
accidental dispersal by wind. We might suppose that an 
exceptionally powerful storm had carried these frail insects 
a few thousand miles away to a spot, where on alighting they 
found the conditions for their future development favourable. 
When similar cases of distribution occur among mammals 
they are explained in a different manner. We then argue 
that the related, but now widely separated or " discontinuous," 
groups must long ago have had a perfectly continuous range 



and that for some reason or other they have subsequently 
been destroyed over wide areas, leaving at present only iso- 
lated colonies. In some instances this theory of the origin 
of discontinuous colonies of animals has been amply con- 
firmed by fossil evidence. Camels, for example, or we might 
say the family Camelidae to which they belong, are only 
represented in South America and Asia ; but since numerous 
fossil members of the family occur in North America, 
we possess decisive evidence that long ago the range of 
the camel family was continuous. Discontinuity of range 
among mammals is always looked upon by zoologists as 
an unmistakable sign of antiquity. It is only when a 
similar range occurs among the more easily dispersed in- 
vertebrates and plants that naturalists are in the habit 
of calling to aid exceptional forces of nature in explaining 
their origin. The undoubted facility with which human 
importations are scattered far and wide and become success- 
fully .established in districts remote from their original 
home seems to encourage and invite speculations as to 
the origin of dis continuously distributed invertebrates of 
all kinds. Actual records of seeds, insects, snails, etc., sowed 
broadcast by accidental agencies far from their native 
land, appear to confirm the theories derived from successful 
human transplantations. So much are these in vogue that 
the ordinary and normal mode of dispersal is almost for- 
gotten. Winds no doubt exert an influence in driving species 
in the direction in which they blow. But many animals, 
beasts as well as beetles or butterflies, possess the faculty 
in a high degree of detecting the presence of their own kind 
by the sense of smell. Winds would, and do, as we know, 
convey scents from one animal to another, thus tending to 
bring the sexes together. This has been very clearly estab- 
lished by Mr. Webster.* The insect, being apprised by scent 
of the presence of its own kind, would endeavour to travel 
in the direction contrary to that of .the prevailing wind in 
order to reach it. It may possibly be true that some insects 
are apt to spread in the direction of the prevailing wind, as 
Mr. Webster asserts. But, judging from the few examples 

* Webster, F. M., "Diffusion of Insects," p. 797. 


he quotes, I do not think there is sufficient evidence for the 
belief that insects as a whole are influenced in their dispersal 
by that agency. His statement that the influence of high 
winds on insects is illustrated by the great number of butter- 
flies that are sometimes encountered by ships at sea at long 
distances from land, is somewhat misleading. As a matter 
of fact, remarkably few species of butterflies have been 
observed far out at sea, and these always belong to species 
that are in the habit of migrating. It is a well-known 
phenomenon that certain species of butterflies and moths, 
such as the painted lady (Vanessa cardui), the milk- weed 
butterfly (Anosia archippus) and the moth Urania leilus, 
congregate into flocks or swarms and migrate in a body at 
certain times of the year. And it is such swarms that are 
occasionally scattered by storms and carried out to sea. 
These are, however, altogether exceptional instances, and we 
are not justified in drawing conclusions from them and apply- 
ing them to insects as a whole, very few of which possess any 
migrating instincts. , On the contrary, the facts of the 
geographical distribution of insects are, as a rule, quite 
in conformity of thpse of mammals. Even the distribu- 
tion of the strongly- winged Sphingidae in North America 
shows a distinct division into an Atlantic and Pacific sub- 

I have already alluded, on p. 90, to the range of the genus 
of butterflies Parnassius in North America, pointing out that 
it had apparently entered the continent in Alaska and had 
then spread along the Eocky Mountain chain. Altogether, the 
butterflies and moths of the Rocky Mountains show a close, 
resemblance to those of the Old World ; among them we meet 
with the familiar genera Colias, Argynnis, Erebia and 
Oeneis.* The main advance has apparently taken place in a 
southward direction from the north along the crest of the 

Much remains to be done before we can obtain even a 
general idea of the beetle fauna of the Rocky Mountains, but 
it is certain that many Old World genera and even species, 
have travelled southward along this chain for a considerable 

* Pagenstecher, A., " Lepidopteren d. Hochgebirges," p. 145. 


distance. On the other hand, typically American genera, like 
the oil beetles (Eleodes), have gained access to heights up to 
8,000 feet in the southern spurs of the Rocky Mountains. 

Of fossil insects, in spite of Mr. S. H. Scudder's * classic 
researches and Dr. Handlirsch's splendid treatise, we know 
comparatively little. But our knowledge has been particularly 
enriched by the discovery, in the midst of the Rocky Mountains 
of a deposit of shales containing a wealth of the most beauti- 
fully preserved specimens of insects and plants. During one 
of the volcanic eruptions, which were so frequent in Tertiary 
times throughout the greater part of the Rocky Mountains 
region, great masses of leaves and innumerable insects were 
entombed among the fine volcanic ash, and were thus readily 
preserved. Over six hundred species of insects are now known 
from these Florissant shales of Colorado, which, according to 
Professor Cockerell f are not of Oligocene age, as Mr. Scudder 
thought, but of Miocene age. The absence of mammalian 
remains, however, increases the difficulty of estimating the 
exact age of these deposits. It is possible, moreover, that some 
of the shales may be much older than others. 

I cannot leave the Rocky Mountains without expressing a 
few words of appreciation as to the wisdom and forethought 
of the Americans in preserving large tracts of country in the 
wild state. These large land reserves, as Mr. Roosevelt $ so 
forcibly reminds us, are mainly to keep the forests from 
destruction, but likewise to preserve, for future generations, 
the wild animals that live in them. 

The first and most fa/mous game preserve in the world was 
established in 1872 and set apart as a public park or pleasure 
ground for the benefit and enjoyment of the people. Congress 
provided against the wanton destruction of fish and game, or 
their capture or destruction for merchandise or profit. As a 
result of this wise enactment we find to-day thousands of 
deer of various kinds in this magnificent world-famed en- 
closure known as the " Yellowstone Park," the name being 
derived from Yellowstone, the largest tributary of the Missouri 

* Scudder, S. H., " Tertiary Insects of North America." 

t Cockerell, T. D. A., " Fauna and Flora of Florissant," p. 160. 

| Eoosevelt, Th., " Wilderness Eeserves," pp. 2324. 


River. I need not enlarge upon the fascinating spectacles of 
the hot springs or the geysers, waterfalls and other natural 
wonders which attract sightseers to this district. 

Since the opening of the Yellowstone Park, largely owing 
to the efforts of the Boone and Crockett Club and its founder 
Mr. Theodore Koosevelt, many other game and forest pre- 
serves have been established in the United States. The 
American Bison Society and the New York Zoological Society 
also worked incessantly towards the same end, so that at pre- 
sent ove'r seven million acres in the United Staites are devoted 
to the preservation of the native fauna and flora. The two 
largest enclosures are Ihe Yellowstone Park in Wyoming, and 
the Grand Canon Game Preserve in Arizona. This growth of 
sentiment in favour of protecting animals and plants from 
destruction has also spread beyond the borders of the States 
into Canada, and induced the authorities there to imitate these 
beneficent measures. In their enthusiasm to vie with their 
neighbours, Canadians have even provided game preserves 
which exceed in size the largest of those referred to, for the 
new Jasper Park in Alberta has an area of nearly three and a 
half million acres, while Kocky Mountain Park in Alberta 
has two million seven hundred thousand acres. Two others 
have over a million acres. 

Hidden game preserves of the past life of North America, 
as I mentioned before, lie among the vast accumulations of 
Tertiary rocks in the same mountains that shelter the modern 
representatives of the American fauna. In the beginning of 
the chapter I just alluded to the names of some- of the more 
important deposits and the geological formations they belong 
to. Enough is now known of the remains of the animals con- 
tained in these deposits to enable palaeontologists to compare 
their relationship with that of fossil assemblies of animals in 
other continents. Professor H. F. Osborn has recently pub- 
lished an excellent summary of our knowledge of these western 
beds and their mammalian fauna, and I cannot do better than 
quote some of his conclusions. 

The Eocene Tertiaries of the mountain region, lying in the 
Rockies and west of them, were partly formed by the post- 
Cretaceous or post-Laramie uplift, accompanied by great 
volcanic activity, lava flows and eruptions of volcanic dust, and 


by the formation of a series of lake, river and flood -plain 
basins, filled with volcanic and erosion sediments. During 
the first faunal phase of the Eocene Period a land connec- 
tion with South America seems to be indicated by the occur- 
rence of similar mammals in the upper Cretaceous or basal 
Eocene of Patagonia. Additional evidence of South American 
connection is afforded by the subsequent occurrence of animals 
related to the Edentata-Dasypoda in the American middle 
Eocene. A momentous change occurs, according to Professor 
Osborn, during the second faunal phase of the Eocene. 
Similar faunas appear almost simultaneously in south -western 
North America and in western Europe. In Professor 
Osborn's * judgment this remarkable circumstance is due to 
the gradual southward extension of the fauna from a hypo- 
thetical northerly American- Asiatic land mass. 

Not a single specimen of an Eocene mammal has been dis- 
covered in northern Asia or the northern parts of North 
America. Professor Osborn and many other authorities 
assume the Eocene existence of a great American- Asiatic 
land mass, because large tracts of land in the north certainly 
are very ancient, and must have been raised above the sea in 
Eocene times. That is about all the evidence we possess 
for the belief that the great similarity of the western European 
and western North American fauna during the Eocene Period 
was due to some land connection via northern Asia and 
northern North America. 

I should prefer to throw my hypothetical land bridge 
straight across the Atlantic from western Europe 'to North 
America. In another chapter I shall endeavour to show that 
we possess important zoogeographical evidence for the belief 
in such a transatlantic bridge in Eocene times. Professor 
Schlosserf contends that geological researches are alone 
capable of yielding information about former land connec- 
tions. He does not believe that much weight can be attached 
to ancient reconstructions of continents based on zoogeo- 
graphical or distributional data. I hold, on the contrary, that 
since certain old groups of animals, even genera and species, 

* Osborn, H. F., " Cenozoic Mammal Horizons," pp. 19 35. 

t Schlosser, M., "Uber Tullberg's System der Nagetiere," p. 748. 


have* originated in the far distant past and are now scattered 
here and there in isolated colonies, their present range indi- 
cates former conditions of land and water. By comparing 
the discontinuous distribution of such old groups in the dif- 
ferent continents, we come to certain conclusions as to the 
conditions of the continents during the time when their range 
was continuous. We have already become acquainted with 
several genera and species showing discontinuous distribu- 
tion. I argued, in the first and fourth chapters, mainly from 
distributional evidence, that North America was connected 
with north-western Europe and with north-eastern Asia in 
Pliocene times. We know of relict genera and species which 
represent the life and the geographical conditions of still 
more remote stages of the Tertiary Era, namely, the Miocene, 
Oligocene and Eocene, and even periods of the secondary or 
Mesozoic ^Era. We thus possess in our recent fauna an invalu- 
able adjunct to palaeontological research. That this state- 
ment is not a mere assumption will be amply demonstrated in 
the next few chapters. 

Returning to Professor Osborn's researches, he directs at- 
tention to the striking diminution of European types in the 
last stages of the Eocene deposits. This, he argues, might 
have been due to the existence of prolonged geographic or 
climatic barriers between the two continents. In Oligocene 
times the faunal community with western Europe once more 
becomes closer. It is important to note, Professor Osborn re- 
marks, that many American lower Oligocene types are repre- 
sented by more primitive forms of European upper Eocene and 
partly of north African types. This stage is followed by a long 
period of independent evolution and partial extinction of the 
same fauna to the close of the lower Miocene age. About the 
middle of the Miocene Period another profound change in 
the mammals of North America occurs. This is mainly due to 
the .sudden appearance of a large number of new forms of 
African and Eurasiatic origin, such as the elephants, which 
are believed to have come from Africa, and the rhinoceroses 
and the true ruminant animals, which are supposed to have 
had their home in Asia or Europe-. These North American 
middle Miocene deposits contain animals which first appear 
in the lower Miocene of Europe. Hence there is distinct evi- 


dence for the assumption that the general movement of the 
fauna has been from the Old World to the New. Finally, 
in the Pliocene beds of North America, we perceive clearly 
that an invasion of South American animals has taken 

* Osborn, H. F., " Cenozoic Mammal Horizons," pp. 42 82. 



WHEN a naturalist from western Europe crosses the Atlantic, 
and after landing at one of the great ports on the east coast 
of the United States, takes a stroll in the country on a fine 
summer's day, a great many novel features strike his eye. 
Let us suppose, for instance, that he is specially interested 
in reptiles and amphibians. Although the neighbourhood of 
New York and Boston abounds with ideal dry sandy banks 
which would be alive with lizards if situated in France or 
Germany, these reptiles seem to be completely absent. This 
is actually the case. No lizards have been observed in the 
neighbourhood of these cities. If he went further inland 
to ascertain whether America is really devoid of lizards, our 
visitor might succeed, after crossing the Hudson Eiver, in 
capturing some. Two kinds of lizard have been recorded 
from the western parts of the State of New York, viz., the 
blue-tailed lizard (Eumeces quinquelineatus) and the common 
swift (Sceloporus undulatus).* The first of these is a strik- 
ingly handsome species with five vividly yellow lines along 
the back and a tail of brilliant blue. It is altogether different 
from the ordinary European lizards, for the scales are shiny 
like those oif the slow-worm. "The blue-tailed lizard forms 
part, in if act, of the large family of skinks (Scincidae) most 
o'f which inhabit the tropical portions of the Old World. The 
genus Eumeces (Plestiodon) to which the American blue- 
tailed lizard belongs, is largely confined 1 to the south-eastern 
and southern States of North America. A few species are 
found in the south-west and in Mexico, while the genus is 
entirely absent from South America and Europe. Eight 
species are known from Asia. Now the most remarkable point 

* Eckel, E. C., and F. 0. Paulmier, " New York Keptiles," p. 390. 


about this blue-tailed skink is that a species, until recently 
considered absolutely identical with it, is found in Japan, 
being unknown on the mainland of Asia. Some differences 
have now been detected between these two skinks, but they 
are no doubt very closely related to one another. We are con- 
ironted, therefore, with the extraordinary problem how to 
account for the occurrence of two species, so nearly akin, 
in localities so distant from one another. It must be clear to 
anyone who is familiar with distributional problems that acci- 
dental dispersal within recent times either by man or by 
any other agency is out of the question. This is a case of 
geographical distribution which must be explained by the 
ordinary modes of migration. If it was quite a unique 
instance of such a remarkably discontinuous range, it might 
be a matter of some difficulty to discover a plausible explana- 
tion to account for it. But it is by no means the only example 
of such a range. Quite a number of instances are known. A 
still more striking one is that of the so-called ground lizard 
(Lygosoma laterale).* 

The ground lizard, with its minute limbs, thick tail arid 
sluggish movements, reminds one more of a salamander than 
a lizard. It lives, moreover, under the bark of trees or among 
rotten wood, and is thus altogether different in habits from 
the ordinary lizard. Now this peculiar ground lizard occurs 
in identically the same form in North America, in China and 
Japan. f The most searching comparison by the best experts 
has hitherto failed to elicit the slightest difference between 
the Asiatic and this North American ground lizard. 

It is interesting to note that the ground lizard and 
the blue-tailed lizard, both of which exhibit such a remark- 
ably East Asiatic relationship, are members of the family 
Scincidae. But, whereas we possess in America over thirty 
species of the genus Eumeces, to which the blue-tailed 
lizard belongs, there are only two American species of 
Lygosoma. We now have to ascertain whether these two 
genera Eumeces and Lygosoma, are of American origin, or 

* Cope, E. D., " Crocodilian, lizards and snakes of North America," 
p. 622. 

f Stejneger, L., " Herpetology of Japan," p. 219. 


whether they have come to America from some other part 
of the world. Lygosoma laterale is known from eastern 
Mexico in the south-west, and from all the southern and 
eastern States as far north as New Jersey. Altogether fifty 
species of the division of Lygosoma, to which the American 
species belong, have been described.* Five of them live in 
New Zealand, twenty in Australia and the adjacent islands, 
seven in the Pacific Islands, four in the Philippines and 
Borneo, seven in India, two in the Nicobar Islands, one in 
Mauritius, two in West Africa, one in Central America, while 
a single species, as far as we have learnt, is found in China, 
Japan and North America. The wide range of the species in 
North America shows that it has not been introduced. It is 
no doubt indigenous. Yet, to judge from the range of the 
genus Lygosoma, America is certainly not its home. We may 
also safely conclude, from its most discontinuous range, that 
it is of very great antiquity, although quite unknown as a 
fossil. New Zealand, according to Dr. Wallace,f received its 
flora and fauna during the latter part of the Secondary Era, 
and has not since been connected with any mainland. Since 
this view has been widely accepted, it would tend to show 
that the genus Lygosoma w,as already in existence in 
Mesozoic times, and that it possibly gained its present 
distribution towards the end of the Secondary or early in 
the Tertiary Era. 

The second genus Eumeces may help us in our enquiry 
as to the mode of entry into North America. There are about 
twenty species of Eumeces in North America, ranging from 
Mexico in the south to Minnesota in the north, and New 
Jersey in the east. Considerably over one-half of these inhabit 
the south-western States and Mexico. Certainly the centre 
of distribution in America lies in the south-west. The various 
species seem to have radiated from this centre in all directions 
except the south. The genus must have existed in this south- 
western centre for a very long time past, because one species 
peculiar to an island in the Bay of Campeche, another to the 
Island of Bermuda, and still another to southern Florida, are 

* Boulenger, G. A., " Catalogue of Lizards," Vol. III., pp. 253289. 
t Wallace, A. K, " Island Life," p. 506. 


known to science. Eumeces must have lived in North 
America possibly since early Tertiary times. The genus has 
also been observed in India. One of the species ranges from 
Baluchistan, right through Persia, Syria and Egypt to Tunis, 
having there probably given rise to the allied form confined 
to Algeria and Morocco. 

If Eumeces had already existed in south-western North 
America in early Eocene times, we should expect it to have 
travelled to South America during the supposed Eocene land 
connection with that continent. That it has not done so may 
be due to the fact that its original home is in south-eastern 
Asia. This assumption is strengthened by the circumstances 
that its nearest relations are the genera Tribolonotus of New 
Guinea, and Brachymeles of the Philippine Islands. 

Since Lygosoma, like Eumeces, also occurs in the south- 
western States, it may likewise have made its entry into North 
America in that region, though its original home seems to have 
been somewhere in the western Pacific. Being probably an 
older genus than Eumeces, both may, nevertheless, have taken 
the same route in reaching North America. Where that route 
lay and what were the geographical features of North America 
at the time will be discussed later on when more material has 
been gathered from other sources. 

I alluded above to another lizard which has penetrated as 
far north as New Jersey, namely, the swift (Sceloporus undu- 
latus) , so called from the great activity and speed of its move- 
ments.* Its scales are strikingly different from those of the 
skinks. They are large and coarsely keeled, terminating in 
sharp, bristly points. The swifts belong to the typically 
American family Iguanidae. All the Iguanidae, and there are 
many of them, are confined to North and South America, with 
the exception of one genus which inhabits the Fiji Islands 
and two others living in Madagascar. It is quite possible, 
however, as Dr. Gadowf suggests, that the Madagascar genera 
represent cases of convergent evolution from some common 

Although a species of Iguana has been described from the 

* Ditmars, E. L., " The Eeptile Book," p. 123. 
t Gadow, H., " Amphibia and Keptiles," p. 501. 


Eocene deposits of France and England, we have no reason 
to doubt that the family has originated in America. The 
Cretaceous genera Iguanavus and Chamops from Wyoming, 
have always been considered as belonging to the Iguanidae. 
Since the family scarcely enters eastern America even at the 
present day, the western States must always have been its 
headquarters. How the dispersal from the American con- 
tinent to the Fiji Islands was effected is another problem 
which I defer to a later discussion. 

As for the genus Sceloporus, to which the swift (S;. undu- 
latus) belongs, almost all the species inhabit Mexico, that 
country forming the centre of distribution. From there the 
genus ranges as far south as Nicaragua,. One species occurs 
in the western States, while the single northern species, 
Sceloporus undulatus, ranges from Guatemala to the eastern 
States. Both the nearest relations of this genus, Phrynosoma 
and Uta, have their headquarters in the south-western States 
and Mexico. We possess, consequently, satisfactory evidence, 
both from its recent as well as its fossil distribution, that 
the family Iguanidae is of North American origin, and that 
the latter can be traced to the end of the Mesozoic Era. 

In spite of the occurrence of these three species of lizards 
in some of the north-eastern States, these reptiles certainly 
are scarce in the east. It is not so with the snakes. No less 
than twenty-three different kinds have been observed in the 
vicinity of New York,* whereas in the whole of France there 
are only about half that number. The eastern States of 
America are, in fact, remarkably rich in snakes. Snakes have 
even penetrated to the eastern islands, for several species, in- 
cluding the dreaded rattlesnake, occur in Long Island. 

The commonest of these eastern and probably of all the 
American snakes, is the garter-snake (Thamnophis or Eutenia 
sirtalis). It is abundant from southern Canada to Florida, 
and I have met with it even on the edge of the White Moun- 
tains plateau at a height of 5,000 feet. Very closely allied 
to, arid scarcely distinct from the genus Tropidonotus, the 
garter- snakes (Thamnophis) form a most troublesome group 

* Eckel, E. C., and F. C. Paulmier, "New York Eeptiles," p. 356 


from a systematic point of view. Several of the species are 
exceedingly variable, and all are difficult to discriminate from 
one another. 

The genetic relationships of the various species and the 
causes which have given rise to the differentiation of the 
garter-snakes are most attractive subjects for study. Dr. 
A. E. Brown* has discussed the connection between moisture 
and variability, especially in the direction of colour intensity 
in this group of snakes. More recently an ingenious and 
novel method of carefully estimating the value of the cha- 
racters commonly held to be specific in snakes has been 
adopted by Dr. Ruthven. He shows that the reductions in the 
number of rows of dorsal scales as the girth ,of the body 
decreases in the individual snake, are brought about by the 
dropping of certain definite rows. This leads him to the con- 
clusion that specific variation in the scale rows follows the 
same sequence and is also correlated with the circumference 
of the body. Similarly, presence, absence or fusion of the 
labial scuta are dependent on the length of the head. Dr. 
Ruthven's f assumption is that the garter-snakes started in 
America with the maximum number of dorsal rows of scales 
known in the genus, and that the forms resulting from geo- 
graphical extension are mostly due to dwarfing in consequence 
of unfavourable environment. He then traces four lines of 
descent, which all emanate from northern Mexico as the centre 
of origin of the genus. 'The area inhabited by the nineteen 
species of garter-snakes includes all North America and south- 
ward as far as the southern boundary of Guatemala. The 
genus is evidently a geologically recent immigrant to Central 

It is of the greatest interest to the student of zoogeography 
that Thamnophis differs from its nearest American relative, 
Tropidonotus, by the absence of scale pits, and by the pre- 
sence of an undivided anal plate, for it seems almost certain 
that Thamnophis has originated in North America from some 
ancestral form of Tropidonotus (Natrix), the latter being 
clearly a much older genus. Tropidonotus has a vast range 

* Brown, A. E. Variations of Eutaenia." 

t Buthven A. G. "Variations of the Garter Snakes." 


from north Australia through the Malay Archipelago and 
northward to Japan. Westward it extends to India, Africa 
and Europe. In America the genus occurs principally 
in the eastern States, but there is also a species in Cuba 
and several in Central America. None have penetrated 
to South America. From Dr. Boulenger's * catalogue and 
Dr. Stejneger's description of Japanese forms,f it seems 
evident that the east Asiatic forms are not very closely 
related to the American ones. On the other hand, Tro- 
pidonotus validus, from Lower California, Arizona and 
Mexico, is nearly akin to the west European Tropidonotus 
viperinus, whose high antiquity is indicated by its occur- 
rence in Sardinia and Algeria. I may have another oppor- 
tunity later on of dealing with this interesting genus, but 
it must be conceded that although we possess no fossil 
evidence, the origin of Tropidonotus must date back at least 
to the commencement of the Tertiary Era. In another chapter 
I shall show how extraordinarily intimate is the relationship 
of the south-west American and the West European faunas. 
Evidence will then be given in support of the theory that 
southern Europe and the south-western parts of North 
America have been connected with one another by land, and 
that we probably owe the resemblance in the animals and 
plants of these two regions to that fact. The species of 
Tropidonotus, in America at any rate, are semi)- aquatic. They 
are so much attached to water that they are popularly known 
as " water-snakes," whereas the garter-snakes are much less 
bound to the neighbourhood of water. The latter are not 
uncommonly found on higher ground and in drier situations, 
a change in the climatic conditions of the south-west may thus 
originally have given rise to the Thamnophis branch from the 
original Tropidonotus stock. The headquarters of Tropi- 
donotus now lie in the eastern States, where the conditions 
for its existence are more favourable than in the south-west. 
All the same, it seems probable that the original centre of dis- 
persal was in the south-west, for when Tropidonotus arrived 

* Boulenger, GK A., " Catalogue of the Snakes in the British Museum,', 
Vol 1. 

t Stejneger, L., " Herpetology of Japan," pp. 264 294. 

L.A. K 


in North America the climate of that region must have been 
very different from what it is at present. 

I may just mention two other examples of innocuous snakes 
which frequent the north-eastern States, viz., the smooth 
green snake (Liopeltis vernalis), and the rough green snake 
(Cyclophis aestivus). The former is abundant in New York 
State and northward as far as south-eastern Canada. South- 
ward it ranges to the Gulf of Mexico and westward to New 
Mexico, becoming rarer as we approach the drier and warmer 
districts. The other does not extend nearly so far north. On 
the other hand, the rough green snake is found westward as 
far as northern Mexico and California. Both of them share 
the peculiarity of being the only members known in America 
of the genera to which they belong. That is not the only fea- 
ture of interest about their distribution. I have just urged 
that the ancestors of the American species of Tropidonotus 
must have come from Europe. We cannot claim the same 
origin for the American species of Liopeltis and Cyclophis, 
for neither of these genera inhabits Europe. Both of them are 
absent also from Africa. Their headquarters are in southern 
and eastern Asia, but they do not extend as far north as Japan. 
Formerly these snakes were classed among that insoluble 
zoogeographical enigma, namely, the group of animals and 
plants peculiar to eastern Asia and eastern America. Now 
we have advanced in so far as we have been able to trace 
some of the eastern forms to an originally western American 
range. It has been made easier, therefore, for those natura- 
lists who are in the habit of explaining anomalies of dis- 
tribution by the convenient flotsam -jetsam theory, to bring 
their views to bear upon problems such as those suggested 
by the two green snakes. That these snakes could have been 
floated across the Pacific Ocean on a raft by any possible 
chance, is to me inconceivable. That they should have utilised 
the Bering Strait land connection, and subsequently have 
become extinct all along north-eastern Asia and north-western 
North America does not appeal to me either as likely. We 
must only leave the consideration of the problem for the pre- 
sent, as was done in the case of the lizard genera Eumeces 
and Lygosoma, which also apparently had an east Asiatic 


From the popular point of view, as Dr. Ditmars * remarks, 
the venomous rattlesnakes are the most interesting of the 
American serpents. The Old World naturalist involuntarily 
associates America with rattlesnakes ; and the ominous warn- 
ing sound produced by the unique appendage at their tail 
alone exercises a strange fascination on everyone who has 
become acquainted with them. Not only is the common rattle- 
snake (Crotalus horridus) abundant in some of the hilly por- 
tions of New York and Massachusetts, it actually appears to 
be increasing in numbers in these populous States, owing to 
the decrease of its natural enemies. From a zoogeographical 
point of view rattlesnakes are of considerable importance, 
because, being strictly limited to the ground, they are less 
liable to accidental dispersal than the members of many 
other genera which are expert swimmers or fond of climbing 
trees. The rattlesnakes are generally divided into two genera, 
viz., the pigmy rattlesnakes (Sistrurus) and the rattlesnakes 
proper (Crotalus). Both genera range almost all over the 
United States, a couple of species even cross the borders 
of Canada. Southward, Sistrurus is also met with in Mexico. 
The true rattlesnakes (Crotalus), on the other hand, have a 
much wider distribution in America, one species (C. terrificus) 
having been found in Mexico as well as 'in Yucatan, Bolivia, 
Venezuela, northern Argentina, and southern Brazil. 

Remains of rattlesnakes have been noticed in a couple of the 
North American caves, otherwise they are unknown in earlier 
deposits. Their range is confined to America, and we have 
no reason to suppose that they have originated in any other 
continent. Their home is, no doubt, as Dr. Brown f suggests, 
in the south-western States in what he calls the Chihuahuan 
district. We have no means of estimating the age of the two 
genera of rattlesnakes. Yet Crotalus seems to be the more 
ancient, and, like its near relation Lachesis of southern Asia, 
Central and South America, has probably a remote ancestry. 
It may possibly have inhabited North America since early 
Tertiary times. 

I have drawn particular attention to the fact that the 

* Ditmars, E. L., " The Eeptile Book," p. 426. 

t Brown, A. E., " Texas Eeptiles and their Faunal Eolations," p. 558. 



great scarcity of lizards and the abundance of snakes are 
characteristic features of the north-eastern States of North 
America. But I have not yet alluded to the most remarkable 
feature of the reptilian fauna of that region, as it is not a 
character readily noticeable to the untrained naturalist. It 
is the great wealth of land and fresh-water tortoises. In 
England it is customary to distinguish the marine, paddle- 
limbed kinds of Chelonians as " turtles," all others as " land 
and water tortoises." In America the term " turtle " is 
usually employed quite indiscriminately to true tortoises and 
turtles. In adopting the common American names applied 
to these tortoises, I am only endeavouring to make my 
remarks quite clear to American readers. 

I have had occasion already to refer to one species of 
fresh-water tortoise (p. 51), sometimes called " semi -box " 
turtle (Emys blandingi), and to discuss the peculiar range 
of the genus, which is confined to parts of Canada and the 
northern States in America, and to central and southern 
Europe in the Old World. We possess no fossil remains pf 
any members of the genus except from recent geological 
deposits. Although we must assume that a migration has 
taken place either from North America to Europe, or vice 
versa, in more remote times, palaeontology gives us no clue 
as to the origin of Emys. We can only surmise that the 
genus and species of Emys are ancient from the fact that 
the genus Clemmys occurs in the Eocene of North America, 
while Chrysemys has been discovered in the Eocene of 
Europe. Both of these are still living genera of turtles and 
closely allied to Emys. They are classed among what are 
called " terrapins " in North America. 

The terrapins are characterised by their broad, flattened 
shell. The hind feet are extensively webbed, and they are 
good swimmers. Perhaps the best known of the terrapins 
is the "diamond-back" (Malacoclemmys* centrata)* which 
frequents the salt marshes of the Atlantic coast. It is in 
great favour in the States as an article of diet. The less, 
familiar spotted turtle (Clemmys guttata) is of greater zoo- 

* Ditmars, E. L., "The Eeptile Book." Siebenrock, F., "Synopsis 
der Schildkroten." 


geographical interest. Four species Oiccur in North America, 
three of them being limited in distribution to the eastern 
States. Only one, viz. : Clemmys marmorata, is quite con- 
fined to the rivers and ponds west of the Cascade and Sierra 
Nevada Mountains in Oregon and California. In Central 
and South America the genus is unknown, but in eastern 
Asia we find four species. One of them inhabits Japan, 
another the island of Hainan, and a couple of them China. 
Far away in western Asia we meet with another species in 
Persia and Mesopotamia, spreading across Asia Minor to 
Greece and Turkey, while finally Clemmys leprosa lives in 
Spain, Portugal and north and western Africa as far as Sene- 
gambia. The genus is thus quite confined to the northern 
hemisphere. We might suppose that the ancestor of the 
American Clemmys insculpta had spread northward from 
eastern Asia in Pliocene times and, after crossing the Bering 
Strait land bridge, had invaded Alaska and Canada, ulti- 
mately reaching the eastern States of America in that 
manner, and subsequently becoming extinct in the vast tract 
of country which now separates the Chinese from the eastern 
American species. The south-western Clemmys marmorata 
would thus have been the latest development of the genus 
in America. Such a supposition seems unlikely, owing 
to the unsuitability of the western States for such, 
turtles. The few that now inhabit the west are rather to be 
regarded as ancient survivals of a long-distant age when 
the climatic conditions were much more favourable for them 
than at present. The earliest member of the genus indeed 
(C. morrisiae), is found in a western Eocene deposit (Bridger 
beds), and Dr. Hay* expresses the opinion that North 
America is probably the original home of Clemmys. If so, 
the genus could not possibly have spread to Asia, and thence 
to western Europe in Pliocene times. The existing centres 
of distribution must have become established during much 
more remote geological periods. Anyhow, although the 
majority of the American species of Clemmys are now con - 
fined to the east, we have ample palaeontological evidence 
of its having formerly lived in the west. 

* Hay, O. P., " Fossil Turtles of North America," p. 290, 


Besides these so-called fresh-water turtles, we have in 
America even terrestrial ones. The box-turtles (Terrapene 
= Cistudo), for instance, are strictly confined to the land, where 
they live largely on vegetable matter and berries. The six 
species are distributed from Mexico in the south-west to the 
New England States in the north-east. In the more northern 
habitats they burrow to some depth in soft ground and pass 
the winter there. All the box turtles inhabit North America, 
the genus Terrapene to which they belong being, in Dr. Hay's 
opinion, of North American origin. The oldest fossil species 
(T. putnami) comes from a deposit in Florida stated to be 
of Pliocene age. Dr. Hay * is inclined to the belief that the 
genus arose in North America, having probably been derived 
from Emys. If this should be the case, the origin of Emys 
itself must have taken place at a much earlier date than the 
Pliocene. Indeed it is possible that Emys is one of the most 
ancient of Chelonian genera, for Dr. Boulenger f argued that 
it is in many respects the least specialized of the Emydidae 
and that it should be placed at the base of the family. It is 
of interest to note that although box turtles lived in North 
America at least since Pliocene times, they have not been 
able to spread beyond the confines of the continent. This 
gives us some idea of the slow rate of dispersal of these 

Of the family of snapping turtles (Chelydridae) a few 
fossil remains are known from the Jurassic and Cretaceous, 
though not of existing genera. To judge from their distri- 
bution the latter must nevertheless be very ancient. These 
large fresh-water Chelonians with their proportionately huge 
and sinisters heads, are, according to Dr. Ditmars, bold and 
aggressive fighters, their massive, keen-edged jaws causing 
them to be the terror of most of the aquatic and semi -aquatic 
creatures. There are three species in North America. One 
of these ranges from Mexico, east of the Bocky Mountains, 
to Canada, but far to the south of Mexico the same species 
(Chelydra serpentina) reappears in Ecuador. Such a very 
peculiar discontinuous range is not unknown among fresh- 

* Hay, 0. P., "Fossil Turtles of North America," p. 360. 
f Boulenger G. A., " Catalogue of Chelonia," p. 49, 


water animals, as we shall learn later on. A second species 
(Chelydra rossignonii) is confined to Mexico and Guatemala. 
The so-called alligator snapping-turtle (Macros! emmys tem- 
minckii), which belongs to another genus, frequents the 
rivers emptying into the Gulf of Mexico, being common in the 
Mississippi as far north as Missouri. Now the most striking 
zoogeographical feature of the family Chelydridae is, that 
right across the Pacific Ocean we meet with another snapping 
turtle (Devisia mythodes) in the Fly River of New Guinea, 
the same river that contains the unique members of the family 
Carettochelyidae. In many other respects New Guinea has 
acquired the reputation of being the home of remarkable and 
peculiar types of animals, and no doubt the island .once 
formed part of an ancient land, most of which has long since 
been submerged. 

Quite as instructive and important from a distributional 
point of view as the reptiles are the amphibians. We can 
roughly divide the latter into tailed amphibians, such as the 
newt and salamander, and tailless ones, of which the frog 
and toad are examples. All these are amply represented in 
the north-eastern States of North America, and we notice 
among them that same curious relationship between eastern 
America and eastern Asia which has given rise to so much 
comment among naturalists. 

Two kinds of giant salamanders are known to exist in the 
world. One of these, the so-called "hellbender" (Crypto - 
branchus allegheniensis), lives in the eastern States of North 
America, the other, (Cryptobranchus or Megalobatrachus 
japonicus), in the mountain streams of Japan. The hell- 
bender is a voracious lead-coloured slimy creature, living 
in the Allegheny and other eastern rivers, where it feeds on 
worms and fish. It grows to a length of two feet, while 
its Asiatic relative attains to double that size. Of the geo- 
logical history of the giant salamanders we know nothing 
as far as America and Asia are concerned, but in the Miocene 
of Switzerland a large amphibian was discovered, which 
appears to be closely allied, though now referred to the 
distinct genus Proteocordylus. To suppose that these relicts 
of bygone ages are no older than Miocene would scarcely be 
justified. In the absence of palaeontological evidence, we must 


depend upon the testimony based on the remarkably discon- 
tinuous ranges of the two recent species. Any theory as to 
the home of the giant salamanders should be founded on that 
of other animals possessing a similar range. 

The hellbender is not the only large amphibian frequent- 
ing tihe rivers of the eastern States. A somewhat eel -like 
creature, with feeble diminutive limbs and three pairs of 
bushy external gills, inhabits many of the larger streams 
and lakes in the north-eastern States.* It grows to about the 
same length as the giant salamander, but belongs to quite a 
different family. It is commonly known as the " Mud puppy " 
(Necturus maculatus), and does not occur west of the Rocky 
Mountains. Now in this case the nearest relation of the east 
American form does not live in Eastern Asia, but, like that 
of the mud minnow (see p. 51), in Eastern Europe. This 
European member of the family Proteidae has manifestly 
undergone a certain amount of degeneration. It possesses 
only three fingers and two toes, is completely blind, and is 
restricted to the subterranean waters of Dalmatia and the 
neighbouring provinces of southern Austria. The ''Olm" 
(Proteus anguineus), as it is called, is scarcely a foot long, 
and quite white except for the gill bunches which are 
brilliantly red in colour. It was believed that the Texas 
subterranean newt (Typhlomolge rathbuni) was related to 
the mud puppy. Miss Emerson,f however, has shown that 
it is a salamandrid. No fossil remains of any of these amphi- 
bians are known, and any theories as to the origin of the 
discontinuous distribution of the members of this ancient 
family, must be based on zoogeographical data. 

The family of true salamanders and newts (Salamandridae) 
likewise comprises certain members which seem to be of 
very ancient origin, such as the blunt-nosed salamanders 
(Amblystoma). Being only semi-aquatic animals, they 
possess greater facilities for dispersal than the purely fresh- 
water forms. Most of them are large species, the tiger sala- 
mander (Amblystoma tigrinum) growing to nearly a foot in 
length, and they only repair to the water in the spring to 

* Cope, E. D., " Batrachia of North America," p. 26. 
+ Emerson, E. T., "Anatomy of Typhlomolge," p. 72. 


deposit their eggs. The tiger salamander is variable in 
colour. Generally of a dark brown, it is marked with 
irregular yellow blotches, and is commonly met with under 
stones and in decayed hollow trees. At the time when the 
Spanish conquerors landed in Mexico, the natives were in 
the habit of eating roasted or boiled fresh-water creatures, 
which they called " axolotls." It is now known that these 
axolotls are nothing but the larval forms of the tiger sala- 
mander, which is found from Mexico to the city of New York. 
Under certain conditions, the larva, instead of losing its gills 
and turning into the terrestrial salamander, retains them, 
and continues its existence in the medium in which it was 
born, growing into a creature somewhat resembling the hell- 
bender, and breeding without leaving the water. 

The genus Amblystoma is almost confined to Mexico, the 
United States and Canada. Only one species (A. persimile), 
which most resembles Amblystoma jeffersonianum of the eastern 
States and Canada, inhabits the far distant mountains of Siam 
and Upper Burmah. This represents, therefore, another 
example of that curious relationship between eastern North 
America and Asia. In this case, however, it is with southern 
instead of eastern Asia, while the genus in North America occurs 
in the south-west as well as in the eastern States. 

Several smaller kinds of newts have been observed in 
the eastern States. Among them there are several belong- 
ing to the genus Spelerpes, which are worthy of special com- 
ment. They all display remarkably brilliant colours, and 
these make them more attractive than salamanders usually 
are. Quite apart from this feature, their method of feeding 
is interesting to watch. Like chameleons, they possess an 
enormously long tongue, ending in a soft sticky knob, which 
is shot out of the mouth with extreme rapidity at any insect 
coming within range and likely to he a dainty morsel. About 
twenty species of Spelerpes are known to science, the head- 
quarters of the genus being in Mexico. Dr. Gadow* dis- 
covered several of them at considerable heights. Spelerpes 
orizahensis and S. leprosus ascend to 12,000 feet, and S. 
chiropterus to 10,000 feet, above sea-level. 

* Gadow, H., " Mexican Amphibians and Eeptiles," p. 203. 


Their earliest centre of distribution lay in what has been 
called Sonoraland, or south-western North America, and 
from here the species have spread in Miocene times, according 
to Dr. Gadow, to the eastern States, to the island of Haiti 
and even to Peru. The latter occurrence is of particular 
importance, as we shall see later on, when we come to the 
consideration of the points of resemblance between the 
Mexican and the South American faunas. Still more remark- 
able is the fact that a single species of Spelerpes (S. fuscus) is 
known from some of the remaining fragments of the ancient 
Tyrrhenian continent in southern Europe (see Fig. 8). Even 
Dr. Gadow,* who shows little inclination for reconstructing 
ancient land bridges, does not suggest that this salamander 
could have crossed the Atlantic Ocean without their 
assistance. He thinks a land connection joining north- 
eastern North America with north-western Europe, by way 
of Greenland, might have enabled the south European 
Spelerpes to cross from the New World to the Old. 
I concur with Dr. Gadow in the belief of the former 
existence of a land bridge in the extreme North Atlantic, 
but I am of the opinion that it had not yet made its appear- 
ance at the time when Spelerpes undertook its journey to 

This short review of some of the characteristic north- 
eastern reptiles and amphibians has clearly revealed a 
relationship of some of the older forms with those of southern 
and eastern Asia and also of southern Europe. In several 
instances it was demonstrated that the eastern States were 
not the original home of the genera, but that the North 
American centre of distribution lay in the south-west. Hence 
it seems possible that the south-west was in remote times, say 
about the commencement of the Tertiary Era, the great centre 
from which reptiles and amphibians wandered eastward. 
Owing to subsequent changes in the climatic conditions of 
the south-west, some genera, and even families, probably 
became extinct there, thus obscuring the original relation- 
ship of that part of North America with Asia and Europe. 
Without fossil evidence to guide us, it would seem as if these 

* Gadow, H., " Mexican Amphibians and Keptiies," p. 244, 


speculations were based on somewhat unreliable foundation. 
We may therefore call to our aid another branch of the geo- 
graphical distribution of living organisms, namely that of 
plants, so as to test the validity of these theories. 

Professor Asa Gray * was the first to direct attention, in 
1859, to the striking similarity of the flora of eastern Asia 
to that of the eastern States of North America. In a popular 
raccount of the distribution of the North American flora, Sir 
Joseph Hooker again alluded to this feature more recently, 
stating that there is actually specific identity in about two 
hundred and thirty cases, and very close representation in 
upward of three hundred and fifty. What is most curious, he 
says, is that there are not a few very singular genera of which 
only two species are known, one* in east Asia, the other in 
east America. In some of these instances the Asiatic species 
is a widespread plant in east Asia, whilst the American is an 
extremely scarce and local plant. This and other conditions 
render it conceivable, according to Sir Joseph Hooker,f that 
the Asiatic element in east America is dying out. 

Still more recently Professor Engler discussed the same 
subject very fully. He believes that the number of species 
common to the eastern States and eastern Asia is far less 
than Sir Joseph Hooker thought. Some of these occur also 
in the north, others in western North America. Yet there 
are certain plants which exhibit extraordinarily discontinu- 
ous distribution, quite comparable to what we have noticed 
among reptiles. Monotropa uniflora and Phryma lepto- 
stachya, for instance, occur only in the eastern States, in 
Japan and the Himalayan Mountains. Professor Engler 
looks upon these as relicts of a flora which was uniformly dis- 
tributed in Tertiary times between the Himalayan Mountains 
and North America. Of the genera Liquidambar, Ostrya, 
Platanus, and Castanea, we know that they lived further north 
in Tertiary times than they do now. We have also learned 
from the Pliocene and Miocene beds of the Eocky Mountains, 
as Professor Engler points out, that the flora west of these 
mountains was formerly not so distinct from that of the 

* Gray, A., "Eelations of Japanese Flora." 

t Hooker, J, D., "North American Flora," p. 573, 


eastern States as it is at present. A more pronounced 
climatic differentiation between the two parts of North 
America supervened, he thinks, in later Tertiary times, and 
while many of the species became extinct in the western 
States owing to the gradually increasing dryness of the 
climate, new forms better adapted to the altered conditions 
arose. Hence certain types of plants are represented in the 
western States by many, and in the eastern by few species. 

It is evident that Professor Engler* favours the view that 
the western States were once much more nearly related in 
their flora to eastern and southern Asia than the eastern 
States, and that the present similarity between the latter 
regions has arisen as a secondary character. Professor 
Engler, moreover, believes and in this respect my views 
differ from his that Japan, western and eastern America 
were connected with one another in the north during the 
Tertiary Era and probably even in Cretaceous times, forming 
three great peninsulas of land joined at their northern bases. 
At any rate, the distribution of plants in America seems to 
offer a certain amount of support to the view suggested, that 
the relationship of the east American and east Asiatic faunas 
is due to the recent geological changes in south-western North 
America having obliterated the more striking features of 
resemblance between the latter and eastern Asia. 

In connection with the character of the north eastern flora 
just referred to, I might offer a few remarks on the subject 
of the supposed former eastward extension of the land. This 
subject was discussed towards the end of the second chapter. 
I then maintained that, although the north-eastern States 
had been under water in Pleistocene times, unsubmerged 
land existed to the eastward quite close to, and including 
portions of the present shore-line. 

Probably one of the best recognised and most characteristic 
elements of the eastern North American floras, as Mr. Hollick 
remarks, is the one generally known as the " Pine-barren 
flora," which is such a prominent feature throughout the 
eastern and southern parts of New Jersey and southward. 

* Engler, A., " Entwicklungsgeschichte d. Florengebiete," I., pp. 22 


Now this flora has a curiously discontinuous range further 
north-eastward. It disappears from the mainland almost 
entirely, but reappears on Staten Island and Long Island. 
Still further east comes a stretch, of eighty miles of sea, 
beyond which the pine -barren flora once more is in evidence 
on Martha's Vineyard and Nantucket Islands. On a limited 
stretch of the opposite mainland, near New Bedford, the same 
flora again makes its appearance. Further north, isolated 
members of the flora such as Magnolia glauca, from near 
Cape Ann, are known from certain coastal tracts. 

It might be argued that this discontinuous distribution is 
due to marine currents or winds, but both the prevailing 
winds and the currents set in from the opposite direction. 
It would not explain the fact, moreover, that the pine-barren 
flora is almost limited to the islands. Hence it seems more 
likely that Long Island was connected by land with Cape 
Cod, forming a continuous strip of land, which was separated 
from the mainland by a broad river or a lake, as Mr. Hollick* 
suggests. Mr. Rollick's theory not only explains the method 
of dispersal of the southern pine-barren flora, it gives us a 
clue to the problem why the northern Helix hortensis, which 
has evidently survived as a relict form, should be almost 
confined to the islands off the coasts of Maine and Massa- 
chusetts. However, while I believe that much of that land 
which lay off the Atlantic coast remained unaffected by the 
Glacial deposits, and that the southern flora survived the 
Glacial Epoch on these islands, Mr. Hollick considers the 
eastward extension of Long Island, and with it the pine- 
barren flora, of post-Glacial age. 

I have mainly dealt with reptiles and amphibians in this 
chapter, because they form a very characteristic feature of 
the north-eastern States. Besides no opportunity occurred 
of mentioning them in the earlier part of this volume. 
The mammals, on the other hand, scarcely need further 
comment here. Only comparatively few species are peculiar 
to this province. Nevertheless, there is an order which 
has not hitherto been alluded to, and which contains 
some remarkable form's confined to the eastern States. The 

* Hollick, A., " Plant Distribution," pp. 191201. 


order is that of the Insect! vora or insect-eaters. They are 
distinguished externally by their small size and soft dense 
fur, while many of them are adapted for an underground life 
and possess specially modified front limbs for the purpose 
of digging. 

Five different genera of moles (Talpidae) are known from 
North America, and three of these are restricted in distribu- 
tion to the eastern States. One of them, containing but a 
single species, the star-nosed mole (Condylura cristata), has 
its headquarters in the north-eastern States, extending north- 
ward as far as Hudson Bay, and southward to North Carolina 
(Fig. 9). The name was given to it from the fact that a 
ring of riband-like appendages surrounds the end of the 
muzzle, in the middle of which are situated the nostrils. Like 
its European relative (Talpa europaea), it constructs extensive 
galleries underground, throwing up a ridge of loose earth 
along the line of the tunnels. No fossil remains of the star- 
nosed mole having ever been discovered, palaeontology 
furnishes no evidence as to its past history, and we must 
assume that it has originated in north-eastern North America. 

What is often known as the common mole (Scalops 
aquaticus) in the States, is no near relation of the European 
species of that description. It is more appropriately called 
naked -tailed mole. The Latin name aquaticus was given to 
it because its webbed hind-feet led to the inference that it 
must be a water animal, whereas it actually lives underground 
in dry sandy soil. Two species of Scalops are known, one of 
them occurs from Massachusetts to Florida, the other further 
west, in the Indian Territory. 

Still another eastern species is the so-called hairy-tailed 
mole (Parascalops breweri), whose habits are very similar 
to those of the other moles, though it is readily distinguished 
from them by its thickly-haired black tail. Its range extends 
along the Atlantic coast, from New Brunswick in the north 
to North Carolina in the south, and inland to the shores pf 
the Great Lakes. It is even said to occur on Martha's Vine- 
yard Island off the coast of Massachusetts, which locality is 
of interest in connection with the view expressed above as 
to this island having formed part of an ancient land surface 
now partially submerged. 


As for the geological history of these forms, a number of 
insectivores with talpoid dentition have been met with in the 
middle Eocene (Bridger) of North America, and referred to 
the mole family (Talpidae), but, according to Dr. Matthew, 
most of them are incompletely known and of doubtful 
affinities.* The first undoubted member of the family (Pro- 
scalops) makes its appearance in the Oligocene beds of 
Colorado, being apparently related to the modern Scalops. 
Owing to their subterranean habits, the moles are rarely 
found fossil, except in fissures such as those of La Grive 
St. Alban in southern France. Palaeontology does not aid 
us materially in solving the problem as to the place of origin 
of the Talpidae, or the geological period during which the 
moles have wandered from one continent to the other. We 
are also entirely in the dark as to the route they have taken 
on entering the New World from the Old or vice versa. They 
may possibly have spread eastward from the western States, 
but a careful comparative study of the living American moles < 
seems more likely to elucidate this problem than palaeonto- 
logical research. 

* Matthew, W. D., " Carnivora and Insectivora of the Bridger," p. 536. 



BOUNDED in the east by what might collectively be termed 
the Atlantic Mountains and in the west by the Pacific Moun- 
tain .system lies the immense continental basin. It is open 
to the sea both in the north and south, extending in one 
continuous series of plains and plateaux from the Gulf of 
Mexico to the Arctic Ocean. The northern portion of this 
great interior basin has already been briefly described in the 
second and third chapters. The rivers of this part of the 
continent drain eastward to the Atlantic Ocean and northward 
to Hudson Bay and the Arctic Ocean. The drainage of the 
southern portion is supplied almost entirely by the Missis- 
sippi, and is thus discharged into the Gulf of Mexico. It is 
this southern section of North America and its fauna with 
which I propose to deal very briefly in this chapter. 

The low-lying and gently seaward-sloping belt of land 
bordering the Gulf of Mexico is known as the " gulf plains." 
It is here in this rich soil that sugar-cane, cotton and rice are 
cultivated. The west- central part of the continental basin 
is occupied by the " prairie plains." By the term " prairie " 
we recognise a level region, either a plain or a plateau, with- 
out forests, but clothed in a carpet of luxuriant grasses and 
flowering annuals. On their eastern and northern border 
these prairie plains merge into the adjacent forested plains, 
while in the west they gradually pass into the more elevated 
and drier high plains, where bunch grass, with bare intervals 
between the scattered tufts, takes the place of the continuous 
sod of the true prairies. 

There is a widespread popular belief in Europe that the 
whole of the vast American continental basin is one extensive 
prairie or pasture land. This is quite a mistake. As we 
approach the Mississippi River from the west we gradually 


pass from the treeless prairie to the forest region, which is 
continued eastward as far as the Atlantic Ocean. In the 
prairie region a struggle has been in progress for thousands 
of years between the conditions favouring tree growth and 
those adverse to them. The increase in the mean annual pre- 
cipitation from west to east is the determinant factor in forest 
production. The main cause, therefore, of the absence of trees 
in the prairies lies, according to Professor Kussell,* in the 
climatic conditions, and principally in the lack of sufficient 
rain during the long, hot summers. 

A thorough survey of the fauna of the prairie region has 
still to be made. Dr. Merriam f devotes only a short para- 
graph to it. Most other writers have confined themselves to, 
a description of one or two typical prairie forms. The sole 
attempt to give us a more lucid impression of the general 
'features of the vertebrate life of the region was made by 
Dr. Kuthven.J He noticed that the peculiar conditions of the 
prairie region had an effect on the fauna in modifying the 
species as they entered this region from the adjoining ones. 
Yet he thinks that there is a great difference in the extent 
to. which the species of eastern North America push westward,' 
or the plains-forms eastward, into the prairie region, before 
(becoming modified or checked. Dr. Euthven's studies lead 
him to the conclusion that the prairie region is an extensive 
area of transition between the plains and eastern forest 
regions, but he expresses the opinion that the conditions of 
environment are either not intensive or not extensive enough 
to mould the animals into a peculiar fauna. 

What was once the most characteristic animal of the prairie 
region is now practically extinct in the United States in its 
feral condition. I need no longer dwell on the history of the 
extinction of the bison, the animal I am alluding to, for it 
has been sufficiently described in the third chapter (pp. 
65 67). When discussing the question of the bison's origin, 
I suggested that its ancestors might have invaded North 

* Kussell, I. 0., "North America," pp. 8996. 
t Merriam, 0. H., " Life in North America," p. 20. 
I Euthven, A. G., "Faunal Affinities of Prairie Eegion," pp. 390 

L.A. I* 


America from Asia in pre-Glacial times. Long before the 
advent of the European conquerors in the New World, herds 
of another large ungulate, the horse, roamed about these same 
prairies and no doubt shared the abundant fodder with the 
bison. When the Spaniards landed in America in 1521 it was 
already extinct, and the natives had not any knowledge even of 
the former existence Of the horse in their continent. Yet even 
in Pleistocene times several different kinds of wild horses 
still lived in North America and were probably contem- 
poraneous with early man. One of these (Equus giganteus) 
seems to have exceeded in size any known race of horse either 
living or extinct.* What caused the sudden extinction of the 
wild horse all over America we do not know. Professor 
Osbornf suggests that a disease of the nature of the African 
" rinderpest " might have done it. The " tse-tse fly " renders 
thousands of square miles of Africa uninhabitable for horses, 
and the invasion of a similar pest into America might pos- 
sibly have swept away the whole of the equine stock in a short 
time. But the interest aroused among zoologists by the dis- 
covery of fossil horses in America was not only connected with 
their unexplained disappearance in modern times, it yielded 
what was thought to be absolutely demonstrative evidence of 
the theory of evolution. Fossil forms no doubt had already 
been discovered in Europe which seemed to indicate that 
the remote ancestors of the existing horse had five digits on 
every foot while intermediate stages with three fully deve- 
loped toes were known. In America, horses, or at any rate 
animals possessing all the essential characters of a horse, 
have been brought to light from very early Tertiary deposits, 
possessing four toes and a rudimentary fifth on the hind foot 
and short-crowned teeth. These are succeeded in Oligocene 
and Miocene strata by others with three toes and short- 
crowned teeth. In still more recent deposits, horses occur 
with three toes and long-crowned teeth which are finally 
followed by horses of the modern type with one toe and long- 
crowned teeth. 

* Gidley, J. W., " Ee vision of North American Species of Equus," 
p. 137. 

t Osborn, H. F., " Causes of Extinction of Mammalia," p. 835. 


In a revision of the American Eocene horses, Mr. Granger* 
distinguishes twenty-six species, all the three genera to which 
they belong being distinct from the early horses found in 
Europe. The American Eohippus appears to be closely re- 
lated to the Old World Hyracotherium, while Epihippus ap- 
proaches Lophiotherium. We thus have a somewhat parallel 
series in the two continents. 

In the Oligocene deposits the horses are still small, some 
of them less than eighteen inches high at the withers. 
Twenty-eight species, belonging to the two genera Mesohippus 
and Miohippus, have been described by Professor Osborn.f 
Sixty more species are mentioned by Mr. GidleyJ as having 
been procured in the Miocene and Pliocene beds, and over half 
a dozen more from Pleistocene strata. Thus we know from 
America already about one hundred and twenty different kinds 
of fossil horses. They gradually increase in size as we pro- 
ceed from the older to the newer deposits. The species with 
many toes are replaced by others with fewer toes, until we 
come to the highest form of specialization in the modern 
horse. All that remains of the outer toes is a splint-bone left 
on each: side of the single toe, while the teeth which originally 
possessed short crowns have now long ones. There is ap- 
parently a gradual evolution from smaller and simpler forms 
to larger and more complex ones, as we glance from the older 
horse remains to the recent ones. And yet not a single gradual \ 
transition from one genus to the other seems to be known. ; 
No wonder that one of our foremost palaeontologists exclaims : 
" The supposed pedigree of the horse is a deceitful delusion, 
which simply gives us the general process by which the tri- 
dactyle foot of an ungulate can be transformed in various 
groups into a monodactyl foot in view of an adaptation for 
speed, but this in no way enlightens us on the palaeontological 
origin of the horse. " 

Considering the extraordinary abundance of horse remains 
in North America, and even in the south of South America, 

* Granger, W., " American Eocene Horses," p. 233. 
t Osborn, H. F., "New Oligocene Horses." 

t Gidley, J. W., "Miocene and Pliocene Horses of North America." 
Deperet, C. H., " L' evolution des Mammiferes Tertiaires," OXL., 
p. 1517. 



it would seem as if the family Equidae had originated in 
America and had sent certain offshoots to the Old World 
during such times when America was connected by land 
with either Asia or Europe. Professor Deperet * certainly 
takes the view that both Anchitherium and Hipparion 
reached Europe by means of a land connection with America, 
and that the two continents were several times joined to one 
another by land during the Tertiary Era. This opinion is 
amply supported by the most weighty zoogeographical evi- 
dence, as will be shown in one of the succeeding chapters 
(pp. 226 231). There is, indeed, a very general agreement 
among palaeontologists on this point. The only difference of 
opinion concerns the exact location of the site of these ancient 
land bridges. 

One of the most characteristic animals of the great plains 
is the so-called " prairie dog " (Cynomys ludovicianus) . The 
name has been applied to it on account of the peculiar barking 
sound it emits when alarmed. Otherwise it has nothing to do 
with the dog family, being more nearly related to the ground 
squirrels and marmots. It loves the sunshine and a dry 
atmosphere, and becomes less and less numerous as we ap- 
proach the humid prairies from the west. The prairie dog is 
a social creature living in colonies, and these, according to 
Dr. Merriam,f are sometimes from twenty to thirty miles in 
length. The damage done to crops by these animals is enor- 
mous, while their increase is greatly favoured by the spread 
of agriculture. The cultivation of the soil enables them to 
support larger families, whereas the cultivator further pro- 
tects them by destroying their natural enemies. 

The prairie dog inhabits a vast area between Montana in 
the north and southern Texas in the south. Altogether seven 
species of Cynomys are known, some of which range into 
Arizona and Mexico, whereas none occur in the eastern or 
extreme western States of America. No fossil remains of 
Cynomys from Tertiary deposits have been identified, except 
from the Miocene Republican River deposits of Kansas and 
Nebraska, and even they only doubtfully belong to the genus. 

* Deperet, Ch., " Transformations of the Animal World," p. 313. 
t Merriam, C. H., "Prairie Dog," pp. 258263. 


Cynomys, however, is certainly of American origin, though 
some of its near relations, as I have shown, have probably an 
Asiatic ancestry. 

The coyotes or prairie wolves have been described as the 
most inveterate enemies of the prairie dog. They are small, 
graceful creatures hunting in packs like other wolves, but 
living in burrows on the plains. Not long ago only a single 
kind of coyote (Canis latrans) was recognised. More than a 
dozen species are distinguished now ; some of them on rather 
slender grounds.* All these occur west of the Mississippi. 
The presence o'f wolves in the Arctic regions of America has 
been alluded to (p. 11 and p. 61), but I have not hitherto 
made any remarks on their past history and origin. 

The dog tribe (Canidae), to which all wolves belong, is 
more widely spread in the world, that is to say, it has a larger 
geographical distribution than any, other family of carnivores, 
one species being even found wild in Australia. Judging 
merely from the extent of its range, the family Canidae should 
be a very ancient one, and this assumption is fully borne out 
by the knowledge we have obtained from fossil remains of the 
dog tribe. 

According to one of the most recent views, it would seem 
as if the Cretaceous ancestors of the Carnivora, the greaft 
order to which the dojg tribe belongs, were a group of small 
arboreal mammals resembling the opossum in size and 
habits, while more nearly allied to the primitive Insectivora. 
The most strictly terrestrial types, such as the Canidae and 
Hyaenidae, have departed widely from the primitive skeletal 
structure. In the Eocene we already find several families 
of the Carnivora fully developed, one of which, the Miacidae, 
is regarded by Dr. Matthew f as the precursor of the dog 

The latter originated in Oligocene times, but it was not until 
the Miocene , Period that, the genus Canis marked its first 
appearance in America and Europe simultaneously. Since 
it is highly improbable that the same genus should have arisen 

* Merriam, 0. EL, "Eevision of the Coyotes." 

t Matthew, W. D., " Carnivora and Insectivora of the Bridger," pp. 328 


independently in two different continents, the genus Canis 
must have originated either in America or Europe. Dr. 
Matthew* directs attention to the fact that the modern repre- 
sentatives of the Canidae living in the Oriental region and in 
South America are more akin to the Oligocene and lower 
Miocene species than are the true wolves, jackals and foxes. 
Assuming the original centre of evolution to have been some- 
where in Europe or North America, we might argue that the 
older types of dog -like creatures spread into distant parts and 
were preserved there, while they were superseded in their 
ancestral home by more modern types. 

But we are apt to forget that the two species of Canidae 
which live furthest from our hypothetical centre of origin 
are most nearly related to what we generally look upon as the 
most modern of the dog tribe. I am alluding to the wild dog 
of Australia '(Canis dingo) and to the Falkland island wolf 
(Canis antarcticus) . It is a most remarkable fact that the 
latter is not nearly related to a single South American species 
of the dog tribe, whereas it really belongs to the coyotes which, 
as we have noticed above, are confined to western North 
America. Similarly the Australian dog, which is now gene- 
rally considered a truly wild species and not a recent human 
introduction, is akin to one of the European Pleistocene dogs. 
There is a wild dog in the mountains of Java (Canis teng- 
gerana) which also appears to be nearly related to the dingo 
of Australia. These two anomalous cases do not seem to fit 
in with any of the existing theories. Dr. Wallace f main- 
tained that it must have been as far back as the Secondary 
Era of geological history that Australia was in actual con- 
nection with the northern continents, and received from the 
latter the ancestors of the present fauna. There was no 
subsequent land connection, according to Dr. Wallace, so that 
from that remote time until now the Australian lands have 
thenceforth evolved the various Marsupial and Monotreme 
types which we now find there. 

It is evident that Dr. Wallace did not believe in the indi- 
genousness of the Australian dog when he made these remarks. 

* Matthew, W. D., " Lower Miocene Fauna from Dakota," p. 180. 
t Wallace, A. K., " Geographical Distribution," I., p. 465. 


Yet even those who do are at a loss to account for its presence 
in Australia. Professor Weber * favours a very early human 
introduction, even in Pliocene times. I myself have been 
unable to form a definite judgment on this subject. 

The origin of the Falkland island wolf is in so far a very 
much more difficult problem to solve, as none of the species 
of the dog tribe living on the opposite mainland of South 
America are at all nearly related to it. Dr. Wallace'f and also 
Mr. Lydekker J express the opinion that the Falkland islands 
were evidently connected with the mainland at no distant date. 
Dr. Wallace believed that this wolf was closely allied to a 
Patagonian species. 

Later on (p. 430) I shall have some further remarks to 
make on this subject. I only mentioned these two instances 
of distribution to show the difficulties which we frequently 
have to contend with. 

The ra-ccoon (Procyon lotor) is by no means a typical in- 
habitant of the prairie, still as it occurs here and there and 
is very characteristic of North America it may as well be 
mentioned here. By nature a forest animal, the raccoon, with 
its omnivorous propensities and great adaptability, easily ac- 
customs itself to the most diverse surroundings, and as a 
rule thrives and breeds well in confinement. It inhabits the 
whole of the United States and southern Canada and belongs 
to a family (Procyonidae) which is quite confined to North 
and South America, and always has been. We need not hesi- 
tate in this case in attributing its origin to America, The 
raccoon family has the same ancestors (the Early Tertiary 
Miacidae) as the dog family, according to Dr. Matthew, one 
of the members of the former, Cercoleptes (Potos), being 
actually the nearest in its skeletal construction to the Eocene 
Miacidae. It has been stated by Dr. Matthew that the lower 
Miocene Phlaocyon from Colorado is approximately, though 
not exactly, ancestral to the raccoon ; but in view of the fact 
that both Dr. Ameghino and Dr. von Ihering || emphatically 

* Weber, M., " Der Indo-australische Archipel," p. 40. 

t Wallace, A. E., "Geographical Distribution," II., p. 49. 

t Lydekker, E., " History of Mammals," p. 140. 

Matthew, W. D., " Carnivora and Insectivora of the Bridger," p. 331. 

|| Ihering, H. von, "Siidamerik. Eaubtiere," pp. 159160. 


support its being one of the dog tribe (Canidae), we should 
hesitate before accepting the earlier opinion. 

The genus Procyon, to which the North American raccoon 
belongs, only makes its appearance in the Pleistocene Period. 
Nevertheless, it is quite possible that cave deposits, such as 
that described from McCloud River in California containing 
the new species Procyon sinus, may eventually be placed in 
the Pliocene series.* At any rate, the genus Procyon must 
have existed before Pliocene times, and it seems to me prob- 
able that it originated either in South America or in some 
western lands which have long since subsided. Dr. von 
Ihering f believes that the Procyonidae have undoubtedly 
come from eastern Asia. Why he should think so I cannot 
imagine, for neither recent nor fossil species are known from 
that continent. 

It is now generally admitted, as I mentioned already 
(p. 95), that Central America assumed its present shape and 
contours at some time during the Pliocene Period (compare 
p. 243). As soon as this land bridge became habitable for 
terrestrial animals, northern species are supposed to have 
poured across it into South America. We possess strong 
evidence certainly that a steady stream of southern animals 
invaded the northern continent in Pliocene and even in 
Pleistocene times and that northern ones succeeded in reach- 
ing the south. 

The group of the so-called toothless mammals (Edentata), 
comprising the ant-eaters, sloths and armadillos, are almost 
entirely confined to South America; and that continent no 
doubt is their original home. A few penetrated in some 
mysterious manner to North America in Eocene times, as I 
shall explain more fully in another chapter. Shortly after- 
wards they seem to have become extinct again in North 
America, for no traces of edentates have yet been discovered 
in the succeeding Oligocene deposits. It is only in the 
Miocene beds of North America that we again meet with 
examples of this curious group. They were representatives 
of the huge Megalonyx which is closely allied to a southern 

* Gidley, J. W., " Fossil Eaccoon from Calif ornian Cave." 
t Ihering, H. von, " Siidamerik. Raubtiere," p. 160. 


genus. Megalonyx and Mylodon, which follows in Pliocene 
times, were giant ground sloths almost the size of elephants, 
while Glyptotherium, another Pliocene species, had a great 
shield -like bony armour covering the whole body. The plio 
cene species were accompanied by a great peccary (Platy- 
gonus), a llama of a very large size (Pliauchenia) and a 
number of other interesting creatures, all of which have now 
completely vanished from the northern continent. 

In the succeeding deposits from the great plains and moun- 
tain regions, which have been classified by Professor 
Osborn * as belonging to the lower Pleistocene series, we 
notice the remains of two large elephants (Elephas columbi 
and E. imperator), a true camel (Camelus) and two other mem- 
bers of the same family, the great peccary Platygonus and the 
two large ground sloths, Mylodon and Paramylodon. The 
rivers were tenanted by beavers, otters, musk rats and a 
curious semi -aquatic creature about the size of a bear, called 
Castoroides. Belated to some of the groups of South American 
rodents, the latter suddenly makes its appearance in the 
Pleistocene beds of eastern North America. 

Now we come to the cave deposits, which I have alluded to 
already on several occasions, and which Professor Osborn 
includes in the middle Pleistocene or Glacial series. The 
Port Kennedy cave in Pennsylvania contains no less than 
four species of the great edentate Megalonyx, also a Mylodon, 
two kinds of sabre -tooth tigers, a Mastodon, four species of 
peccaries and a tapir. In the Potter Creek cave of California 
were discovered, among others, four species of Megalonyx, a 
camel and a Mastodon. Finally, the Conard fissure of 
northern Arkansas revealed two species of sabre-tooth tigers, 
three kinds of peccaries and numerous small animals. Apart 
from a few deer bones and the remains of the curious Symbos^ 
an animal allied to the musk ox, it contained no traces of large 
ungulates. Their presence in the district adjoining the fissure 
is, nevertheless, indicated by the sabre -tooth tigers. 

What I wish to make clear is that huge creatures requiring 
an abundance of vegetable food poured into North America, 
not only in Pliocene but also in Pleistocene times. Many 

* Osborn, H. F., " Cenozoic Mammal Horizons," p. 85. 


other mammals apparently had their original home in this 
continent. Peccaries and tapirs, which, as we know, require 
a hot and moist climate, lived as far north as Pennsylvania 
even during the time when vast glaciers were supposed to have 
covered the whole of Canada and a substantial slice of the 
United States. We are told that the fauna of .this period 
clearly reveals the state of the climate. If the remains of the 
animals above referred to indicate anything, they show us un- 
doubtedly that the climate was mild, with an abundance of 
vegetation and animal life. In common with most other geo- 
logists, Dr. Hay believes that the climate of the Glacial 
Epoch must have been cold in North America, because he 
assumes the certain existence of vast ice-masses at that time 
even in New York, in Indiana and in Missouri. If we deal 
with this climatic problem from an independent standpoint 
and endeavour to reconstruct the conditions prevailing during 
the Glacial Epoch from purely faunistic evidence, our con- 
clusions cannot point to the prevalence of an exceptionally 
cold climate. Proof of the existence of a cold climate in the 
United States during the Pleistocene Period seems to be fur- 
nished, says Dr. Hay,* by the occurrence of the three genera 
of mammals, Kangifer, Bootherium and Symbos. 

The name Bootherium is now applied to an extinct large 
sheep-like creature, viz., B. bombifrons, whose remains have 
been discovered in Pleistocene deposits of Kentucky. Accord- 
ing to Dr. Kowarzik (see p. 7), Bootherium was probably 
the direct ancestor of the northern genus Ovibos, which has 
never been found in any Pleistocene beds in the United States. 
Bootherium can scarcely be claimed as an exponent of a 
cold climate,. because it has never lived north of the United 
States. The latest discoveries seem to indicate that a number 
of sheep-like animals originated in the United States towards 
the latter part of the Pliocene Period, and left their remains 
in various parts of the country. Thus the extinct Eucera- 
therium, first identified by Dr. Sinclair and Mr. Furlong from 
a cave in California, and Preptoceras from another Californian 
cave, are both allied to Bootherium and Ovibos. Hence Ovibos 
is the sole member of this group which has survived, having 

* Hay, 0. P., " On the Changes of Climate," p. 372. 


succeeded in adapting itself to an arctic habitat. The avail- 
able evidence is all in favour of a gradual advance having 
taken place of those large sheep-like forms from a more 
southern to a northern habitat during late Tertiary times. 
Mr. Osgood * has now discovered another extinct relation of 
the musk ox in the Yukon Territory of north-western Canada. 
He first described it as Scaphoceros tyrelli (including Ovibos 
cavifrons of Leidy in the same new genus), and suggested 
that Scaphoceros may be ancestral to Ovibos. According to 
Mr. Barnum Brown, the name Symbos has now been substi- 
tuted for Scaphoceros. Hence Symbos is known from Indian 
Territory, Iowa, Missouri, Ohio, Pennsylvania, Kansas, 
Arkansas, from Yukon Territory and from Alaska. Yet even 
its former presence in Alaska cannot stamp Symbos as a cold- 
loving animal, for close to its remains were dug up those of 
a Mastodon, and who would be prepared to argue that the 
Mastodon is an indicator of a cold climate ? 

Lastly, Dr. Hay claims that the reindeer (Rangif er) having 
occurred so far south of its present habitat (Fig. 10) in 
Pleistocene times is a proof of the existence of a cold climate 
at that time in the United States. I have discussed this pro- 
blem once before (pp. 3 6), but I may add a few remarks. 
If the drift area of North America had been covered largely by 
the sea, as I believe it was, during part of the Glacial Epoch, 
the country which was still habitable for the reindeer must 
have been greatly reduced. Hence a southward emigration 
was the only possible chance of survival for some herds of 
reindeer. Driven out of their home by the stress of circum- 
stances, they would have passed into a district, even if the 
latter had been unsuitable to their requirements. At any rate, 
we know that reindeer can live perfectly well in a temperate 
climate and that they still inhabited Scotland in the twelfth 
century long after the Glacial Epoch had passed away. I 
cannot therefore consider its former presence in the United 
States a proof of a cold climate. That it could only have 
penetrated south in small numbers is indicated by its 
total absence .from all the North American caves hitherto 
examined except one, and from almost all the other Pleisto- 

* Osgood, W. H., " Scaphoceros tyrelli," p. 178. 


cene deposits. On the other hand, we have noted that pec- 
caries lived in the United States during the Pleistocene and 
the preceding .geological periods. They were not exterminated 
by the severity of the climate. Representatives of the peccary 
family not only survived the Glacial Epoch, they even showed 
their indifference to it by invading the area which had only 
just been forsaken by the supposed Wisconsin glacier, for 
their remains, as Dr. Hay tells us, were found in deposits 
overlying the Wisconsin drift at three different localities. 
In the single cave in which the reindeer occurred its re- 
mains were mingled with those of a species of peccary 
(Tayassus tetragonus) very closely allied to that still living 
in the Southern States and in South America. Nor was the 
Glacial Epoch any more trying to the great ground sloth,, 
Megalonyx, for it also survived it and invaded the area covered 
by the drift. The remains of a species of that giant edentate 
were found some years ago, according to Dr. Hay, in an old 
filled -up pond, just within the alleged outermost moraine of 
the Wisconsin glacier near Millersburg in Ohio. My Own 
views as to the nature of the climate prevailing during the 
Pleistocene Period, and particularly during that phase of it 
known as the " Glacial Epoch " or Ice Age, are derived from 
a careful scrutiny of the living and extinct fauna and flora. 
This study of the animals and plants does not reveal to me- 
that the Pleistocene Period was a period of extreme cold. 
On the contrary, as I remarked before, the climate seems to 
have been milder in a large portion of the northern 
hemisphere than it is at present. An apparent increase of 
temperature after the passing away of the " Ice Age " is 
supposed to he indicated by the appearance of forms of animal 
and plant life requiring a higher temperature than is com- 
patible with the arctic condition believed to have prevailed 
during the height of the Glacial Epoch. It is really due, I 
think, to that perfectly natural re-occupation of tracts of 
country on which both plants and animals had been destroyed. 
The destructive agent, in my opinion, was not ice but water. 
Glaciers no doubt existed on all the higher mountains near 
the Atlantic and Pacific coasts. They owed their presence, 
however, not to cold, but principally, as I mentioned before, 
to the higher temperature of the eastern and western oceans. 


Towards the latter part of the Glacial Epoch, when the 
existing geographical conditions of the northern lands were 
gradually brought about, the temperature of the Atlantic and 
Pacific Oceans decreased, causing a diminution of precipita- 
tion on the continents. With slight climatic oscillations the 
conditions almost all over the northern hemisphere gradu- 
ally seem to have grown less favourable for the survival of 
Tertiary animals and plants than they were during the Ice 
Age. Warmth and moisture-loving species are almost every- 
where being replaced by others that can support greater 
extremes of temperature, and the former only exist here and 
there in diminishing colonies as relicts of the past. 

In the United States we have evidence of such a course of 
events, not only among the higher groups such as the 
mammals; some of the more slowly-moving invertebrates 
are even more trusty indicators of the past geological history 
of the country. 

Three species of an operculate snail belonging to .the family 
Helicinidae inhabit the United States. One of them(Helicina 
chrysocheila) occurs in Texas near the mouth of the Rio 
Grande. Another (H. orbiculata) has a wide range from' 
Florida and Texas as far north as Arkansas and Tennessee. 
A third (H. occulta) lives in isolated colonies among loose 
leaf -mould in well-wooded districts from Carolina to Wis- 
consin and Minnesota. Though inhabiting States where 
severe winter frosts are common, it is amply protected against 
them by its mode of life. Mr. Cooke* maintains that all 
operculate land mollusks are exceedingly sensitive to cold, 
and that the whole group is undoubtedly a product of tropical 
or semi-tropical regions. This view is borne out by the range 
of Helicina. Far to the west of North America, beyond the 
Pacific Ocean, a, few stragglers occur in Burma and on the 
Nicobar islands. As we advance eastward they increase in 
number in certain parts of southern Asia. Almost throughout 
Polynesia we meet with some species, and also on the West 
Indian islands (Fig. 11). That the genus is a very ancient 
one must be evident from its geographical distribution. It had 
already reached America in early Tertiary times, for Dr. Dall 

* Cooke, A. H., "Molluscs," p. 24. 


describes a species from the Oligocene Silex beds of Tampa 
in Florida apparently related to a Helicina still inhabiting 
the Bahama islands. Considering that the genus Helicina is 
almost confined to tropical and semi-tropical countries, we 
may assume that it spread northward at a time when very 
mild climatic conditions prevailed in the northern United 
States, and that a few more hardy species have survived in 
isolated colonies wherever they could obtain sufficient pro- 
tection against frost. This view is confirmed by the fact 
that Helicina occulta, now an extremely rare shell, is abun- 
dant in the Pleistocene loess beds of Indiana, Iowa and 
Nebraska. Mr. Shimek * likewise expressed the opinion that 
the still existing northern colonies of Helicina occulta ap- 
pear to be the remnants of a once common race which is 
evidently dying out. 

A family which resembles the Helicinidae, in so far as it is 
largely confined to tropical and sub-tropical regions, is that 
of the Phasmidae.f They comprise orthopterous insects of 
the shape of a small twig, and hence are known as " walking 
sticks," also " prairie alligators " or " stick-bugs " in America. 
All the species found in the United States are wingless. They 
are thus not liable to accidental transport except perhaps by 
water. All are vegetable feeders, and over a dozen kinds 
inhabit the southern States. Among these walking-stick 
insects there is one which has a remarkably northern range, 
viz., Diapheromera femorata. I met with it on Goat Island, 
above the Niagara Falls. It has also been observed near 
Toronto, and several other places in southern Canada. We 
possess no fossil evidence of the geological history of the 
genus Diapheromera; nevertheless, the fact that the family 
Phasmidae was already represented in Jurassic times, accord- 
ing to Dr. Handlirsch,J justifies the assumption that Dia- 
pheromera originated and began to spread northward in pre- 
Glacial times, and that it may be regarded as a southern relict 
in its present northern habitat. 

Two well-known instances of survivals of southern species 

* Shimek, B., " Helicina occulta." 

t Caudell, A. N., " The Phasmidae of the United States," p. 874. 

| Handlirsch, A., " Die Fossilen Insekten," p. 1191. 

FIG. 11. Map of North and South America, showing the distribution (in black and 
within the area surrounded by small circles) of the snail Helicina. 

\Jfoface p. 158. 


of birds in northern habitats occur to me. Strictly speaking, 
they should not be quoted, because they are now extinct in 
their northern habitats. But as tlheir extermination happened 
within historic times, and has been caused by human interfer- 
ence, I may venture to include them in this group of southern 
invaders. They are the turkey and the Florida parrot. 

The wild turkey (Meleagris gallopavo) belongs to a dis- 
tinctly southern group of birds. In the time of the early 
settlers it was common as far north as Massachusetts, and 
extended westward to Colorado and southward to Mexico and 
Florida. Being a much-prized luxury of the pioneer hunter, 
it was soon exterminated in the more populous districts. It 
still occurs in some of the southern States, while an allied 
species is known from Central America. The genus Meleagris 
was already an inhabitant of North America in Oligocene 
times, for Prof essor Marsh described a species from the White 
River deposits in Colorado. No doubt it has lived in North 
America ever since those early Tertiary times. All we know 
from fossil evidence is that the remains of two other species 
were identified from Pleistocene deposits in New Jersey, while 
the bones of the wild turkey itself have been noticed in a cave 
in Pennsylvania. Like the edentates and peccaries this 
southern genus of birds flourished in the northern States 
throughout the Glacial Epoch and survived there until his- 
toric times. 

The Florida parrot (Conuropsis carolinensis) is the only 
example of the large parrot tribe indigenous to the United 
States. It is now restricted to the comparatively small area 
of the Gulf States and the lower Mississippi Valley. Yet the 
early settlers noticed this bird even near the shores of the 
Great Lakes, and occasionally it was observed near the cities 
that were springing up in the eastern States. No doubt it 
survived in these northern districts from remote times, al- 
though we possess ,no fossil evidence of this fact. It cannot 
be contended that the parrot left its former habitat through 
persecution ; nevertheless, man in his agricultural pursuits, 
must have interfered with it, possibly by reducing the birds' 
food supply. 

I wish now to make a few remarks on the inhabitants of 
the mighty river and its tributaries flowing through the 


continental basin, as they are of such importance in tracing 
the geological history of the fauna. The Mississippi lies 
wholly within the boundaries of the United States, and drains 
more than two -fifths of their area. Originating in Lake Itasca 
in Minnesota, the Mississippi receives during its long course 
four great tributaries, the Missouri, Ohio, Arkansas and Red 
River, and a large number of smaller ones. The two prin- 
cipal groups of animals inhabiting this great river system are 
the fishes and fresh-water mussels. Some of the fishes are 
able to live in brackish water, others spend part of their 
lives in the sea, so that they are not of such extreme im- 
portance from a zoogeographical point of view as the fresh- 
water mussels. 

Fresh-water mussels, or Naiades as they have been called, 
all die quickly if immersed in salt water or if removed to the 
land. Their distribution being world-wide, they have been 
looked upon by some naturalists as among the best indicators 
of former changes of land and water over the globe. Others 
have urged that the wide range of these mussels may be due to 
accidental conveyance by birds or fishes. It was thought 
that the eggs or the newly-hatched fry of the mussels had 
been thus transported. Many species immediately after their 
fry has been hatched from the eggs, develop booklets on the 
temporary shell, by which the young mussels can attach 
themselves to foreign objects. It has been argued that such 
larval mollusks might become attached to the feet of aquatic 
birds and be carried by them in their flight from the fresh 
waters of one region to those of other regions and there be set 
free. Theoretically, such an accidental transport would seem 
quite a possible one from time to time, certainly much more 
likely than a similar conveyance of the fry by fishes from one 
river sys'tem to another. In a country like North America, 
where millions of migratory birds pass annually north and 
south, and to some extent east and west, the effects of a con- 
veyance such as suggested should be clearly discernible in the 
composition of the North American fresh -water mussel fauna. 
Yet although there are over four hundred different kinds of 
fresh-water mussels in the Mississippi drainage area, some 
of them having existed there almost unchanged since Cre- 
taceous times, the fauna to the east and west of that area is 


entirely different.* The geographical distribution of fresh- 
water mussels in North America thus constitutes a practical 
demonstration of the correctness of the view so ably main- 
tained by Dr. von Ihering,f and supported by Dr. White, 
that these mollusks are not appreciably affected by chance or 
accidental dispersal. 

The family Unionidae, to which all the North American 
fresh -water mussels belong, first appeared during the Triassic 
Age. Their principal diffusion over the globe may possibly 
have been effected in Secondary or Mesozoic times. During 
the closing period of the Mesozoic Era, the Cretaceous, the 
family attained an extraordinary development, particularly 
in the Laramie strata. The remarkable feature is that many 
of the species in these beds are so nearly like the living 
species that according to Professor Whitfield J they are to all 
intents a,nd purposes the same. There were at that time 
(the Cretaceous Period) two great land masses in place of the 
North American continent viz., one in the east, the other 
in the west. The fresh-water mussel fauna occupied then 
as far as we know, mainly the eastern flank of the western 
land-mass. The latter was probably connected, as I shall 
endeavour to show later on, with some old land-masses on 
the west coast of South America. South America may thus 
have acquired its Unionidae in Mesozoic times. 

The two great families of fresh-water mussels, Unionidae 
and Mutelidae, have been recognised for some time past, also 
the restriction in distribution of the latter family to Africa 
and South America. It was not until the year 1891 that 
Dr. von Ihering made the striking discovery that all the 
Unionidae begin their existence on hatching from the egg, as 
so-called " glochidium " larvae, while the Mutelidae have an 
entirely different "lasidium" larva. The Unionidae, of 
which about one thousand species are now known, have since 
been subjected to a thorough critical revision by Dr. Simp- 

* White, Charles A., " Ancestral Origin of North American Unionidae," 
pp. 7779. 

t Ihering, H. von, " Najaden von S. Paulo," pp. 133140. 

J Whitfield, E. P., "Fossil Unionidae from Laramie Clays," p. 624. 

Ihering, H. von, " Anodonta and Glabaris." 

L.A. M 


son,* as a result of which we now recognise sixty-one 

To return to the strictly North American fresh-water 
mussels, it has heen found that a common assemblage 
inhabits the entire Mississippi drainage basin, and that a 
considerable number of the species have a distribution cover- 
ing the greater part of this area, as well as <the whole of 
Texas and even parts of eastern Mexico. The streams which 
fall into the Atlantic are peopled by an entirely different set 
of forms, the Appalachian chain seeming, according to Dr. 
Simpson, to act as a sharp barrier between the two regions. 
In the greater part of Mexico and Central Am'erica a totally 
different fresh-water mussel fauna is found. Two Unios, 
one Margaritana and some half-a-dozen Anodons, are all that 
have hitherto been credited to the immense region on the 
Pacific slope of North America. One of the Unios, says 
Dr. Simpson,t has been recorded in error, the other is a 
form of the most abundant and most widely distributed Unio, 
viz., U. luteolus. The causes which led to this striking 
difference between the fresh -water mussel fauna of the 
Central basin and that of the Pacific slope will be discussed 
in another chapter. I may only mention that a somewhat 
similar disparity between the two faunas has been observed 
among the fresh-water fishes. Before dealing with these 
there is one other matter of importance that I should like 
to refer to in connection with the geographical distribution 
of fresh-water mussels. 

The far-reaching results of the study of the geographical 
distribution of such a group as the fresh-water mussels is 
exemplified in a striking manner by the following physio- 
graphic problem. In discussing the origin and recent history 
of the physical features of the southern Appalachians 
Messrs. Hayes and Campbell advocated the theory that the 
upper Tennessee river, now a tributary of the Ohio, formerly 
flowed into the Gulf of Mexico by way of the existing Coosa 
and Alabama Rivers. The conclusions were based entirely 

* Simpson, C. T., " Synopsis of the Naiades." 

t Simpson, 0. T., " Relationship and Distribution of Unionidae," 
pp. 354358. 


upon physiographic evidence such as the character of the 
Tennessee-Coosa divide, the nearness of the gorge below 
Chattanooga, and the general arrangement of the drainage 
lines. * Now it is a remarkable fact that Dr. Simpson has quite 
independently come to a similar conclusion from a study of 
the fresh -water mussels. The upper Tennessee and also the 
Alabama River abound in species of the genus Pleurobema, 
which is quite absent in the lower Mississippi. The species, 
moreover, found in these two rivers are very closely allied, so 
that this and other characters led Dr. Simpsonf to the con- 
viction that at some time in the middle or later Tertiary, the 
Tennessee River must have flowed southward into some of 
the streams of the Alabama drainage, discharging its waters 
in this manner direct into the Gulf 0:f Mexico. 

In the fourth chapter (p. 88) I cited some ganoid fishes 
of the Mississippi in illustration of the zoological relation- 
ship existing between eastern North America and eastern 
Asia. Two other well-known and very remarkable ganoid 
fishes live in the Mississippi basin, viz. the bow-fin (Amia 
calva) and the bony-pike (Lepidosteus osseus). The former 
is the sole surviving member of the family Amiidae. Long 
ago, in early Tertiary times, the genus Amia inhabited middle 
and western Europe, while it is amply represented in the 
Eocene (Bridger) deposits of Wyoming. Bony-pikes lived in 
Europe from Eocene to Miocene times. In America they 
likewise appeared in the Eocene period, and persisted until 
the present day. It is evident that both the bow-fin and bony- 
pike are extremely ancient types, which have managed to 
survive a great many geological changes of the American 
continent. Their ancestors must have travelled to Europe 
in the dawn of the Tertiary Era, assuming of course that 
North America was the birthplace of these genera. Did 
they travel from river to river and from lake to lake across 
North America and Asia to Europe, or did they utilise the 
fresh-water streams of a shorter direct land bridge to 
Europe ? These are problems to be solved. The zoological 

* Hayes, 0. W., and M. E. Campbell, " Eelation of Biology to Physio- 
graphy," p. 131. 

t Simpson, 0. T., " Evidence of Unionidae," pp. 134135. 

M 2 


affinity between Europe and North America is so strong, 
and already so many instances of this relationship have been 
referred to, that nothing short of a wide and convenient land 
bridge with lakes, rivers and mountains will suffice to ex- 
plain the meaning of the palaeontological facts. All we know 
is, that in early Tertiary times these fishes multiplied and 
migrated from their original centre of dispersal. The genus 
of the bow-fin is all but extinct. Only a single species 
remains. The bony-pikes exhibit a little more vitality, no 
less than three species being still living. The common form 
(Lepidosteus osseus) ranges from the Great Lakes to Vermont 
in the east, and from there south-westward as far as Mexico. 
A much larger bony-pike lives in the southern. States, in 
north-eastern Mexico and in the island of Cuba, while 'a 
smaller species (L. tropicus) has been observed in Tabasco, 
Guatemala, Nicaragua and Panama.* If Lepidosteus were a 
fish directly limited to fresh water, we might argue that at 
some remote time in the past, a land mass extended from 
Mexico to Cuba and southward to Panama, but being occa- 
sionally met with in brackish water it is possible that bony- 
pikes can traverse short distances by sea. We cannot for this 
reason base any conclusions regarding minor changes of land 
and water on the present distribution of these fishes. Never- 
theless it is a significant fact that the western States '.are 
devoid of bony-pikes and perches as they almost are of fresh- 
water Unios. 

Besides the Mississippi fauna, the curious dwellers of sub- 
terranean waters in the Mississippi drainage area throw a 
certain amount of light on the past conditions of the country. 
A brief account of them at any rate will be of interest before 
concluding this chapter. 

I believe Mr. Putnam f was the first to exhibit a collection 
of blind fishes and crustaceans from the Mammoth Cave to a 
scientific meeting. In doing so he expressed the opinion 
that most of the animals inhabiting the cave are of compara- 
tively late introduction, since they are closely allied to forms 

* Eegan, 0. Tate, "Biologia Centrali- Americana," p. 181. 
t Putnam, F. W., "Mammoth Cave and its Inhabitants," pp. 194 


living in the vicinity of the cave. Nevertheless he claimed 
for the blind fishes and some of the invertebrates a different 
origin, because the former had no immediate relations among 
fresh-water forms, while the lernean fish parasite was a 
more decidedly marine than fresh -water form. He took these 
facts to indicate that part of the great cave system was 
supplied by marine life. 

Professor Packard * makes no allusion to Mr. Putnam's 
view in his account of the origin of the subterranean, fauna 
of North America. Mr. Putnam's theory indeed appears to 
be scarcely tenable. His remark that the blind fishes of the 
Mammoth and other caves have no immediate relations among 
fresh-water forms has to be modified in accordance with our 
existing knowledge of fishes. The blind fishes, all of which 
belong to the family Amblyopsidae, are no doubt a very 
ancient group, and, as Drs. Jordan and Evermannf suggest 
they are probably descendants of the eyed genus Chologaster, 
or at least forms very closely allied to it. Now one species 
of Chologaster inhabits swampy marshes in the southern 
States, and two others live in the subterranean streams of 
Tennessee, Kentucky and Illinois. The most typical blind 
fishes, Typhlichthys subterraneus and Amblyopsis spelaeus, 
are met with in the underground streams of Indiana and 

Professor GarmanJ expressed the opinion that the blind 
species observed in the. caves were already blind prior to the 
formation of the caves, and that they only collected there 
from various directions owing to the favourable conditions 
for their requirements. Whether this theory is based on sound 
evidence need not be discussed, but his statement that 
Typhlichthys subterraneus has a very wide range has been 
questioned by Professor Eigenmann, who showed that the 
apparently identical species from Missouri is really quite 
distinct from that of the Mammoth Cave. Professor Eigen- 
mann points out that we have to deal with a remarkable and 

* Packard, A. S., " Origin of subterranean fauna." 
t Jordan, D. S., and B. W. Evermann, " Fishes of North America," 
Vol. L, p. 702. 

J Garman, H., " Origin of Cave Fauna," pp. 240241. 
Eigenmann, C. H., "A Case of Convergence," p. 281. 


instructive case of convergence. The two very distinct forms 
have converged because of the similarity of their environ- 
ment, and especially owing to the absence of those elements 
in their environment that lead to external protective 

The family Amblyopsidae is confined to North America, 
and its nearest relations are no doubt the Poeciliidae, a group 
of fresh-water fishes with a wide distribution in America, 
southern Europe, Asia and Africa. There are therefore no 
grounds for the supposition that the blind fishes of the sub- 
terranean waters of North America are descended from 
marine ancestors. 



THE inhabitants of the south-eastern States, which I shall 
endeavour to describe in this chapter, form, in many respects, 
a great contrast to those of the continental basin. The greater 
humidity of the Atlantic States produces that characteristic 
wealth and profusion in floral life which constitutes so at- 
tractive a feature as we approach the ocean from the west. 
And no one can fail being struck by the change in vegetation 
even in travelling southward through the Atlantic States from 
the north. The oaks, chestnuts and hickories become more 
varied, evergreens of all kinds increase in number, new and 
magnificent magnolias make their appearance, while vines 
and creepers mingle their foliage with that of shrubs and 
trees. The splendid white pine of the north is replaced in 
the south by the long-leaved yellow pine, whose hard, strong 
and durable wood serves such a wide range of uses. We also 
notice an entirely new conifer, the cypress, which, like the 
tamarack, sheds its leaves in the autumn. Further south in 
Florida, still greater changes await us, and in the extreme 
tip of that peninsula we are surrounded by tropical vegeta- 
tion. Low fan palms and the palmetto grow even further 
north, but here we meet for the first time with the royal palm, 
which for height and grace of shape is unequalled, and many 
other characteristic! denizens of the tropics. Indeed, as Mr. 
Brendel * points out, the flora of southern Florida should not 
be looked upon as part of the North American flora but as a 
link between it and that of the West Indies. Over two hundred 
and thirty species of plants do not extend north of Tampa, 
whereas southern Florida has one hundred and eighty seven 
in common with the West Indian Islands. 

This affinity between the floras of southern Florida and 

* Brendel, F., " Notes on the Flora of Florida," p. 449. 


the Antilles is generally believed to be due to accidental dis- 
persal. The seeds of these plants are supposed to have been 
conveyed to southern Florida from the West Indian Islands 
by winds, ocean currents or migratory birds. Professor 
Engler,* for instance, argues that a direct land connection 
between the West Indies and North America by way of Florida 
could never have eixsted, because the latter had been sub- 
merged beneath the sea until the end of Tertiary times, 
and that seeds are easily conveyed to Florida by the branch 
of the Gulf Stream sweeping along the northern shore of 
Cuba in a north-eastward direction, carrying quantities of 
vegetable matter and often even tree trunks. Winds could 
only transport such seeds that are specially adapted for long 
flights. As regards migratory birds, which are popularly 
believed to carry seeds to great distances in their crops 
and among their feathers, it may be pointed out that 
the main highway for birds travelling between the eastern 
States and South America is by way of north-western 
Florida and Cuba, and not by southern Florida. f If 
birds had any special influence in the transport of seeds, not 
the southern portion of Florida but the northern one should 
show affinities in the flora with the West Indies. If the 
resemblance in the vegetation of southern Florida and the 
Antilles were mainly due to the Gulf Stream, we should expect 
to find the most pronounced similarity between the two floras 
among the strand plants of Florida. This is not the case. 
The great majority of the flowering plants now known to be 
common to the West Indies and North America occur in what 
is called the " hummocks " of the southern extremity of 
Florida. These hummocks consist of isolated groups of hard- 
wood trees, shrubs and vines, and are scattered like islands 
in the everglades and pine forests, instead of being surrounded 
by the ocean. Moreover, the flora of southern Florida is by 
no means exclusively West Indian in character. About forty 
species of plants are peculiar to southern Florida, and over 
twenty are found elsewhere only in Mexico. J 

* Engler, A., " Entwicklungsgeschichte d. Florengebiete," II, p. 215. 
t Cooke, W. W., "New Facts about Migration of Birds," p. 376. 
| Harshberger, J. W., " Floristic Elements of Eastern North America," 
p. 612. 


A West Indian fauna, too, is associated with the flora in this 
portion of the United States, as will be shown later on. The 
problem presented by the origin of this tropical element in 
the North American flora at any rate is not quite so simple 
as It appears at first sight. Dr. Harshberger is of the opinion, 
that the hummock lands on which the tropical flora principally 
grows represent part of an ancient system of islands which 
existed at a time when the Gulf Stream passed fight across 
the then submerged portion of northern Florida. We might, 
therefore, make further enquiries as to whether this theory is 
supported by geological or other evidence, before taking for 
granted that the tropical element in the fauna and flora of 
Florida is of purely accidental origin. 

Professor Shaler * informs us that along the coast of 
Florida, both on the eastern or Atlantic and the western or 
Gulf side, there arise from beneath the sea a number of 
submarine springs. They thus discharge great tides of fresh 
water, originally gathered on the land, through openings on 
the floor of the ocean. He argues that these springs probably 
shed their waters along the margin of the sea above high water 
level, and remarks, " I cannot conceive any such under- 
ground waterways to have been produced under the existing 
conditions of land and water." He assumes consequently, that 
Florida, or a certain part of it at any rate, must have stood 
at a higher level in relation to the sea than it does now within 
comparatively recent geological times. 

Through the discovery of a submerged system of drainage- 
valleys off the coasts of Florida and the Antilles, Dr. 
Spencer is led to believe in an elevation of this area during 
the earlier part of the Pleistocene Period to the extent of 
8,000 to 12,000 feet or more. During a subsidence which 
followed, according to the same author, the. greater part of 
the existing peninsula of Florida was submerged. Dr. 
Spencer f does not specify what parts of it remained above 
water, but presumably the whole of the southern Florida 
which is low-lying was included in the submerged portion. If 

* Shaler, N. S., " Nature and Man in America," pp. 104106. 
t Spencer, J. W., " Becon struct ion of Antillean Continent," pp. 128 


the tropical flora had gained admission to southern Florida 
during its supposed land connection with the West Indies 
in lower Pleistocene times, it would have besn all destroyed 
again subsequently. If Dr. Spencer's theory were substan- 
tiated, the tropical flora of Florida should owe its origin to 
accidental transport. 

When Dr. E. A. Smith * visited Florida in 1880 he made 
some geological notes on the peninsula which do not bear out 
Dr. Spencer's views. He maintains that Florida was elevated 
nearly to its present height above sea-level after the deposi- 
tion of the Vicksburg limestone, that is to say, after the 
Eocene Period, and that this elevation persisted until the 
Pleistocene, when the country was partly submerged. 

No one, however, has done more practical geological work 
on the peninsula than Dr. Dall,f who, in his monumental 
treatises on the Tertiary fauna of Florida, presented us with 
a masterly survey of the past life of that portion of the United 
States. His opinion on the geological history of Florida 
deserves, therefore, most serious consideration, and it may be 
stated at onc'e that he is strongly opposed to Dr. Spencer's 
views, declaring them to be " incompatible with every geologic 
and palaeontologic fact of South Florida which has come to 
my knowledge." 

As the result of his researches Dr. Dall expresses the 
opinion that the peninsula of Florida, together with the larger 
Antillean Islands and the Middle American highlands, were 
uplifted, and the two Americas thus united in OligO'Cene 
times, that is to say, during the early part of the Tertiary 

Professor Gregory J had a similar idea, except that he did 
not specify any geological period. 

When Florida again became disconnected from this Antil- 
lean continent is not clearly stated, but Dr. Dall thought 
that it formed a peninsula of the southern continent as it 
does now of the northern. Florida, according to the same 
author, became definitely united to North America towards 

* Smith, E. A., "Geology of Florida," p. 306. 

t Dall, W. H., "Tertiary Fauna of Florida," IV., p. 1546. 

I Gregory, J. W., " Geology of West Indies," p. 305. 


the end of the Miocene Period, while a slight depression 
occurred in Pliocene times, and little change since. Whether 
the tropical flora of southern Florida, or part of it at any rate, 
is the relict of an Oligocene invasion from the south is left 
undetermined by Dr. Dall's researches. Yet, from the fact 
that a species of the terrestrial mollusk Glandina occurs in 
the Pliocene Caloosahatchie beds of south-western Florida, 
we might be led to infer that other southern forms might 
have survived on the peninsula till Pliocene and possibly 
recent times. 

Dr. Hill concurs in so far with Dr. Dall's conclusions 
as he establishes in Jamaica signs of a tremendous orogenic 
movement in late Oligocene or Miocene times, resulting in an 
uplift whereby many of the West Indian Islands and pos- 
sibly an insular southern portion of Florida became united 
with one another. In Miocene or early Pliocene the islands, 
according to Dr. Hill,* were severed from one another by 
submergence, assuming gradually their present outlines which 
they have since retained. 

That some kind -of union of the island of Florida with the 
West Indies took place in Miocene times is likewise indicated 
by Dr. Matthew f in his attempt to delineate the hypothetical 
outlines of the continents in Tertiary times, for he distinctly 
unites southern Florida and Cuba by land. 

According to Dr. Vaughan,J the Florida plateau already 
existed in pre -Oligocene times, but it was only towards the 
end of the Oligocene Period that a portion of the plateau rose 
above the sea and apparently remained so ever since. 
Although this constituted 'Only a small island (" Orange 
island " as he calls it), the deposits of the whole plateau are 
full of sand and arenaceous material implying proximity of 

From the opinions cited on the geological history of Florida 
it is evident that there is nothing distinctly antagonistic to 
the view that part of the tropical flora of southern Florida 
is a relict from Tertiary times, many of the pecies being 

* Hill, Eobert T., "Geology of Jamaica," p. 224. 

t Matthew, W. D., " Continents in Tertiary Times," p. 366. 

t Vaughan, T. W., " Geologic History of Floridian Plateau." 


probably the descendants of those which passed northward 
from the Antilles at a time when a land connection (possibly 
in Oligocene times) joined the latter with the old island of 
Florida. The fauna of Florida lends some support to this 
view. At any rate, it gives us more solid foundations for 
estimating the nature of the physical changes which the 
peninsula has undergone within more recent geological times. 
In the last chapter I alluded to a species of parrot 
(Conuropsis carolinensis) which in historic times still ex- 
tended its range as far north as the Great Lakes, and which 
seems at present to be retreating towards its original centre 
of dispersal in the south-east. Now this Carolina parrot, 
as it is generally called, has no near relations. Its 
closest ally, Conurus, is a genus of parrot ranging from 
Paraguay and eastern Bolivia in the south to Mexico 
and the West Indian Islands in the north. The fact 
of its occurrence in the West Indies alone implies that it is 
an ancient genus, for these islands have undoubtedly been 
separated from the continent for a long time. Moreover, 
Cuba, Haiti, Mona Island off Porto Rico, Jamaica and St. 
Thomas all possess distinct species of Conurus. It seems not 
unlikely that Conuropsis is an eastern offshoot of the older 
Conurus, just as Rhynchopsittacus has originated from it in 
the west. The occurrence of this parrot on the mainland does 
not necessitate the former existence of a land bridge to the 
West Indies, but, if other facts point to it, the presence of the 
Carolina parrot in the south-eastern States and its subsequent 
invasion of the northern States is more readily explained 
by it. 

I also made allusion before to the newt Spelerpes, a genus 
confined to America and southern Europe. One of its peculiar 
characters is that its tongue can be jerked out to a considerable 
distance. Only one other newt peculiar to Florida and Caro- 
lina, viz., Manculus, agrees with Spelerpes in the nature of 
its tongue, and, like it, must be an ancient genus. Another 
very peculiar amphibian is Amphiuma means, an eel -like 
creature with tiny limbs, inhabiting the ditches of rice-fields 
and swamps of the south-eastern States. No fossil remains 
of this interesting species are known, but there can be no 
doubt that it is a very old form. Pseudobranchus lateralis, 


another most primitive amphibian, is likewise confined to 
the south-eastern States, while Siren lacertina has its head- 
quarters in the same region. A very striking amphibian is 
the Florida tree frog (Hyla gratiosa), the largest tree frog of 
North America, which is common in Florida, extending from 
there across the borders of Georgia and Mississippi. 

The most remarkable member of the fauna of Florida, from 
a zoogeographical point of view, is one of the worm-lizards 
(Amphisbaenidae). The Florida worm-lizard (Rhinema 
floridana) is a limbless, blind, worm-like creature which 
spends its entire existence under ground. It is no doubt of 
immense antiquity, and the only member of the family known 
from North America. 

Whether the curious glass -snake (Ophisaurus ventralis) has 
originated in the south-west or south-east is a difficult pro- 
blem which future researches may help to solve. Professor 
Cope * speaks of an eastern and western type, and it may 
possibly have spread northward from two independent centres. 
I need scarcely mention that the glass -snake, like the Euro- 
pean slow- worm, is a limbless lizard. Anyone may convince 
himself of this fact by examining the eyes, which possess 
well-developed eyelids, while the presence of ear openings 
also distinguishes these creatures from true snakes. The 
particular point of interest in the presence of the glass-snake 
in America lies in the circumstance that a closely allied 
species (0. apus) inhabits Marocco, the Balkan Peninsula 
and Asia Minor, while a second glass -snake is known from 
the eastern Himalayas and Burma. This extremely discon- 
tinuous range denotes great antiquity. The glass -snakes, 
moreover, live principally underground, and, like the worm- 
lizard, are not liable to accidental conveyance by any of the 
occasional means of dispersal that we hear so much of. Their 
occurrence on both sides of the Atlantic gives great weight 
to the evidence, cited in previous chapters, of the former exist- 
ence of a land bridge right across the middle of the Atlantic. 
I need not discuss the subject any further now, because it will 
be amply dealt with in some of the succeeding chapters, when 
new arrays of facts in support of my contention will be sub- 

* Cope, E. D., " Crocodilian;?, Lizards and Snakes," p. 496. 


mitted. We may as well collect the facts tending to support 
this theory as we proceed, since a great accumulation of 
material in one place might appear wearisome. 

The distribution in America of one of the genera of earth- 
worms (Diplocardia), on which glass-snakes largely live, 
somewhat resembles that of the glass-snakes in America. 
There is a species in Florida and three in Carolina. Further 
north the genus appears in Illinois and Nebraska and we also 
have a couple of species in Mexico and Lower California. 

It was Professor Adams,* I think, who first directed atten- 
tion to the presence of two very distinct and powerful centres 
of dispersal, one in the south-east and one in the south-west. 
Although I am by no means so impressed, as Professor 
Adams and Dr. Brown are, by the significance of the south- 
eastern centre of dispersal as compared with the south- 
western one, which is incomparably more important, I quite 
concur in their opinion that the former faunistic centre is 
perfectly recognisable. These writers moreover discuss' the 
problem which I have dwelt upon so many times above, viz., 
that of the climate during the Ice Age. 

Having adopted the current views of the existence of giant 
glaciers in the northern United States accompanied by 
an arctic climate Professor Adams f and Dr. Brown J were 
obliged to search for suitable " biotic preserves^' where the 
pre-Glacial fauna could have safely weathered the Ice Age. 
These they discovered in the two centres of dispersal alluded 
to, in the south-east and south-west, and from them they sup- 
pose the waves of migrants to have streamed forth northward 
after the Glacial Epoch was over. The presence of a few 
stray remains of northern animals south of their present 
habitat lent a certain amount of credence to the theory in a 
southward extension of the arctic climate. But we must 
remember that these northern creatures, when actually 
pressed out of their boreal home by a restriction of 
their habitats, found themselves in the northern United 

* Adams, C. C., " South-Eastern States as a Centre of Distribution,' 
p. 121. 

+ Adams, 0. C., " Post-Glacial Dispersal of North American Biota." 
t Brown, A. E., " Post-Glacial Nearctic Centres of Dispersal." 


States among an almost semi-tropical fauna of colossal 
sloths, peccaries and other southern forms that have long 
since vanished, and which lived through it all in close proxi- 
mity to the supposed ice-sheets and arctic climates. The great 
majority of the Pleistocene deposits in the north indicate that 
the country had a milder climate during the Ice Age than at 
present, and this is particularly shown by those containing 
plant remains. As plants are supposed to be more trust- 
worthy guides than animals, as indicators of former climatic 
conditions, I will give one more example of a Pleistocene 
deposit from the southern States which has come to my know- 

A Pleistocene deposit in north Carolina examined by Pro- 
fessor Berry * yielded no boreal or even cool temperate 
plants. Hence it may safely be concluded, he thinks, that 
the temperature of the Pleistocene Period in the same latitude 
was not lower than it is now. If anything, he says, it was 
slightly higher. Additional facts pointing to the same general 
conclusion are the former more northward extension of the 
cypress (Taxodium distichum) and of Planera aquatica. That 
these plants did not flourish during. mild inter-Glacial phases 
of the Glacial Epoch, remarks Professor Berry, is indicated 
by their being associated in Maryland with ice-borne boulders 
of considerable size. 

To return once more to southern Florida, we find that what 
we observed among plants, namely, the tropical element, is 
likewise recognisable in the fauna. According to Dr. 
Merriam f the semi-tropical insect fauna of southern Florida 
comprises in all not less than a thousand species of Antillean 
insects, half of which are beetles. 

Among the mollusks there are a number of Antillean genera 
represented in southern Florida, such as Chondropoma, 
Liguus, Cepolis, Varicella, and others spoken of by Dr. 
Pilsbry as Mexican genera, such as Eglandina, Praticolella 
and Drymaeus (dormani type). Dr. Pilsbry regards only the 
last group as genuine natives of the soil. He believes that 
their ancestors entered Florida at the close of the Miocene 

* Berry, E. W., " Pleistocene Flora of Carolina," p. 347348. 

t Merriam, C. H., " Distribution of Life in North America," p. 53. 


Period by a land passage, but does not state clearly whether 
they came by a more direct route than exists at present. All 
the other snails are considered by Dr. Pilsbry * to be waifs 
and strays derived from Cuba and the Bahama islands, by the 
agency of hurricanes, drifting trees and the like. 

I do not know why Dr. Pilsbry should make this reserva- 
tion in favour of Drymaeus, as one of the species found in 
southern Florida (D. dominicus) is also known from Haiti 
and Cuba, besides the Mexican habitat. On the other hand, 
we must not forget that Dr. Dall f discovered a number of 
species of the land-snails Bulimulus and Cepolis in the 
Oligocene Silex beds of Tampa in Florida. Both of these have 
come from the south, for Bulimulus, though extinct in 
Florida, still lives on the island of Fernando de Noronha in a 
species almost indistinguishable from one of the Floridian 
ones. Cepolis still inhabits Florida, but is not found else- 
where in the United States. It has its headquarters in the 
West Indies, and was a European resident, according to 
Sandberger, in early Tertiary times. Of the land snail genus 
Oxystyla, allied to Drymaeus, the species 0. undata has a wide 
range in the West Indies, and is also known from southern 
Florida. Yet the Floridian specimens both belong to varieties 
peculiar to the peninsula. J Similarly, the Cuban varieties 
of Liguus fasciatus are not the same as occur in Florida. 

The tropical forms of mollusks alluded to by Dr. Pilsbry 
as inhabiting southern Florida are by no means the only ones 
that have been collected there. The southern genera 
Choanopoma, Truncatella, Microceramus, Cerion and Veroni- 
cella (Vaginulus) have also entered this region. One of the 
Urocoptidae lived in Florida already in Oligocene times, and 
it is quite possible that Microceramus pontificus and M. flori- 
danus, which are peculiar to southern Florida, have existed 
there ever since. Among many groups of invertebrata long 
specific persistence is much more common than is generally 
realised. The fact that some of the Floridian species are 

* Pilsbry, H. A., " Origin of Molluscs of South Florida," p. 193. 
t Dall, W. H., " Tertiary Fauna of Florida," Part IV., p. 1565. 
| Pilsbry, H. A., "Manual of Conchology (Pulmonata)," Vol. XII., 
pp. 109110. 


identical with West Indian ones does not necessarily imply a 
recent introduction. Some mollusks, at any rate, seem to have 
preserved their specific characters unchanged through several 
geological periods. On the other hand, although there cannot 
be the slightest doubt that a certain number of species intro- 
duced by human agency thrive in other localities besides their 
native homes, I am not convinced that mollusks spread across 
any wide expanse of sea by other accidental transport. With 
Mr. Bryant Walker * I prefer to attribute the tropical land 
mollusks of Florida largely to a former land connection 
between the then island of Florida and a larger southern land- 
mass. I cannot, however, agree with Mr. Walker's view that 
this event took place in comparatively recent times. Dr. 
Simpson urges that the Floridian area must have been joined 
to the greater Antilles by way of the Bahamas in Eocene 
times. Nevertheless, he does not derive the tropical species 
of Florida from the southern invasion which must have taken 
place at that time. He favours a recent colonisation by acci- 
dental transport. The rich fauna, of the Bahama islands seems 
to him entirely derived from the greater Antilles in that 
manner. f 

There are certain geological grounds for the supposition 
that an ancient Archaean land-mass trending north-eastward 
from the northern end of the Andes once existed, and that 
traces of it are still recognisable in Guatemala, Cuba and 
Haiti. J Much of this early land may still have stood above 
sea-level in early, and perhaps middle, Tertiary times, form- 
ing a centre from which the North American continent de- 
rived part of its present fauna. 

Dr. Ortmann demonstrated in a very convincing manner 
that the fresh-water crayfish belonging to the genus Cambarus 
originated in Mexico, spreading from this centre of dispersal 
into the United States at the beginning of the Tertiary Era. 
The centres for the more advanced forms of the sub-genus 
Cambarus, and for the sub-genera Faxonius and Bartonius, 

* Walker, Bryant, " Origin of American Mollusca," p. 56. 
t Simpson, C. J., " Land and Freshwater Mollusks of West Indian 
Eegion," pp. 447448. 

t Frazer, T., "History of Caribbean Islands," p. 398. 

Ortmann, A. E., " Affinities of Cambarus," pp. 124125." 

L.A. N 


are situated in the southern States of North America. It is 
suggested by Dr. Ortmann that the south-eastern centre of 
the early forms of Cambarus originated from the more ancient 
south-western one by; a process of migration across the present 
continent. May not these early forms have travelled eastward 
from Mexico towards Cuba and Florida when the latter were 
connected directly by land with Central America ? A species 
of Cambarus still lives in the rivers of Cuba. 

Instead of mollusks or crayfish we may take almost any 
group of North American invertebrates and readily discover 
among them certain ancient forms, which are either confined 
to small areas in the south-eastern States or have evidently 
spread northward from a south-eastern centre. 

Scorpions, for instance, are universally acknowledged to be 
a very ancient group. The genus Centrurus has its head- 
quarters in the West Indies and Central America. Now 
four species of Centrurus are known from Florida, viz., C. 
gracilis, C. carolinianus, C. margaritatus and C. hentzi, the 
latter being peculiar to Florida. The allied genus Tityus, 
which is rather more southern in distribution than Centrurus, 
has one endemic species in Florida, viz., T. floridanus.* 
Altogether the southern part of Florida shows marked affini- 
ties with the West Indies. There are also some species in 
Florida such as Cupiennius sallei, Keys, not yet recorded from 
the Antilles, which are known from Central America. Quite a 
similar southern relationship has been noticed among many of 
the Floridian Orthopteraf and the Coleoptera.j: One of 
the most noteworthy genera of beetles recorded from Florida 
is Khopalomesites. It has spread northward as far as Dela- 
ware, and reappears across the Atlantic in western and 
southern Europe. 

As regards the dragon-flies and their allies (Neuroptera), a 
few species are common to the West Indies and the northern 
continent. Some of these may have flown, or have been con- 
veyed by a storm, from one region to the other. Such a method 
of colonisation, however, cannot Jiave been usedj by Enallagma, 

* Banks, Nathan, " Arachnida of Florida," p. 142. 

t Kehn, J. A., and M. Hebard, " Orthoptera of Florida." 

J Schwarz, E, A., " Coleoptera of Florida." 


according to Prof. Kolbe,* on account of its feeble development 
and weak flight. Taking into consideration the general dis- 
tribution of ^this genus, he favours the view of a former 
ancient land connection between the Antilles and North 

In its butterflies and moths Florida is united to Central 
America and the Antilles, and it is from the latter that such 
genera as Heliconius, Dione and Agraulis gained a foothold in 
the northern continent.f 

An exceedingly ancient group are the woodlice (Isopoda) or 
" sowbugs " as they are sometimes called in America. In 
eastern Europe at Odessa, and again in north-western Africa, 
an interesting very peculiar species occurs which is known 
as Tylos latreilli. The same species has been, met with in 
southern Florida and in the Bermuda islands. A case of that 
kind is generally set down at once as an instance of human 
importation. The possibility of the survival of such a species 
through several geological ages is not even discussed. But 
in southern Florida a second species of Tylos has been dis- 
covered which occurs elsewhere only in Bermuda. This 
species, known as Tylos niveus, is white in colour and pos- 
sesses structural differences distinguishing it from the other. 
Of the genus Cubaris, which is widely distributed through the 
Antilles sand westward to Mexico and California, a single 
species (Cubaris pisum) occurs in Florida, and is peculiar 
to it.f 

All these, it may bp objected to, are inconspicuous creatures 
that might have been overlooked elsewhere. Too much im- 
portance, it might be urged, should not be placed on their 
occurrence in Florida. We may return, therefore, to more 
conspicuous objects. 

Everyone in America knows, or has heard of, the alli- 
gator (Alligator mississippiensis), though it is only found 
in the southern states of North America. Its distribu- 
tion in fact is rather limited. From Florida it extends 
northward to Carolina and westward as far as tlie Eio 

* Kolbe, H. J., " Neuroptera der Antillen," pp. 157158. 

t Pagenstecher, A., " Verbreitung der Schmetterlinge," p. 359. 

\ Eichardson, H., " Isopods of North America." 



Grande. Only one other species of alligator is known, viz., 
Alligator sinensis from the Yangtse River in China.* This 
enormously discontinuous range is significant, and implies 
great antiquity. Fortunately we possess most valuable 
palaeontologies! evidence as to the alligator's antecedents. 
Even the most pronounced advocate of accidental dispersal 
would not venture to apply the usual methods of wind, waves 
or hurricanes to explain the origin of this example of dis- 
tribution. The generally accepted theory, I believe, is that 
some ancestor of the American alligator has travelled north- 
ward, and succeeded in crossing the former land bridge across 
Bering Strait to north-eastern Asia, thence wandering south- 
ward to China. We possess no fossil evidence for such a 
belief. All we know is that the rather generalised alligator 
Diplocynodon lived already at the very commencement 
of the Tertiary Era both in North America and in Europe, 
and that it persisted in Europe until Miocene times. Henoe 
it seems likely that the modern genus Alligator originated in 
early Tertiary times either in Europe or North America, and 
spread thence to Asia. That America was probably the centre 
of dispersal is indicated by certain characters the Chinese 
alligator has in common with the South American caimans. 

An equally remarkable fact of distribution is that the true 
crocodile has succeeded in obtaining a footing on the North 
American continent in one single small area, namely, in that 
in which I have already signalled so many tropical speciee, 
in southern Florida. We are apt to associate crocodiles with 
Africa. Yet they have a much wider distribution. The genus 
Crocodilus occurs in Africa, Syria, India and eastward as far 
as northern Australia. Westward it reappears in South 
America, the West Indies and Central America. The species 
alluded to (Crocodilus americanus) is the only member of the 
family inhabiting the West Indian islands, and it also occurs 
in Central America, Columbia, Ecuador and Venezuela. 
Remains of crocodiles found in the Eocene of Wyoming 
and the eastern States have been referred by Leidy, Cope 
and others to the genus Crocodilus. The presence of croco- 
diles in America dates back, then, to the very beginning of 
the Tertiary Era, and it seems surprising that they have not 

* Barbour, Th., " Chinese Alligator." 


spread more widely in America. Although the American 
crocodile is fond of salt marshes, and some of the eastern 
crocodiles actually frequent the sea coast, I cannot for a 
moment believe in the possibility of a crocodile crossing an 
ocean such as the Atlantic. Only a land connection between 
America and the Old World in early Tertiary times can 
explain its present geographical distribution. 

Of West Indian mammals none have so far been observed in 
Florida, unless we include the raccoons among them. It is a 
most remarkable fact that a raccoon (Procyon maynardi) 
inhabits New Providence, one of the Bahama islands. At 
the first impulse we might think of a chance introduction by 
floating trees from the mainland of Florida. But Florida lies 
one hundred and seventy miles west of New Providence, and is 
separated from it by a swift current flowing northward. 
Raccoons occur nowhere else in the West Indian islands, 
and the New Providence species differs from that of the main- 
land. In the last chapter (p. 151) I contended that the 
whole family Procyonidae was certainly of American origin, 
and I thought the genus Procyon might have arisen in Plio- 
cene times. Possibly it is much older, though we possess no 
palaeontological evidence in support of such a supposition. If 
the Bahama islands had still been joined to Florida when the 
latter had already become a peninsula of North America, we 
should expect a good many of the smaller mammals to have 
crossed over to the Bahamas, which they have not done. So 
far, the occurrence of this species of raccoon in the Bahamas 
is a complete puzzle to me. 

Before concluding this brief survey of the south-eastern 
fauna I should like to dispel the impression I may have given 
that the tropical element forms any great share in the verte- 
brate fauna of Florida. On the contrary, the majority of the 
genera and species undoubtedly poured into the peninsula 
from the north and north-west, and they appear to be gradu- 
ally displacing the southern element. At any rate,. I look upon 
the latter as a relict of the Tertiary Era. 

There are many species of mammals peculiar to Florida, 
but not so many as we might expect from the favourable 
climatic conditions of the country. On the whole, the 
mammalian fauna of the peninsula bears the impress of a 


newly emerged land which has been populated from the north 
within recent geological times. The most noteworthy species 
that might be mentioned is the Florida water vole (Microtus 
alleni) . It is not alone the only North American water vole, 
but also the sole member of the sub-genus Neofiber, and thus 
forms a connecting link between the genera Microtus and 
Fiber, that is to say, between the meadow voles and the 
musk rats.* 

I have endeavoured to describe in almost every chapter the 
range of some typically 'North American mammal. Some have 
not yet been alluded to, among them the opossum. Two 
species are now often distinguished in North. America,f one of 
which is supposed to extend its range into South America. As 
a matter of fact the northern <and southern forms are 
extremely difficult to discriminate from one another, and 
many authorities are of opinion that they merely represent 
varieties of one species (Didelphys marsupialis) which thus 
has an enormously extensive distribution in North and South 

The existing marsupials, or pouched animals as we may call 
them, are generally regarded as modern survivors of one of the 
ancient groups of mammals, which apparently spread almost 
all over the globe before the superior beasts of more recent 
times had made their appearance. It is believed that at the 
time of their prime, Australia became separated from the 
mainland of Asia, so that this southern continent now forms 
the headquarters of the group, while in other parts of the 
world they succumbed in the struggle with superior and better 
fitted animals. Yet in America, where competition with the 
more highly developed and more aggressive beasts ought to 
be very keen, opossums, which belong to the marsupials, 
are by no means on the verge of extinction. On the 
contrary, they hold their own perfectly against the more 
modern competitors. The genus Didelphys is much more 
abundantly represented in South America than on the 
northern continent, and other genera of marsupials occur 
there besides opossums. To judge from these modern repre- 

* Bangs, Outram, " Mammals of Florida and Coast Eegion." 
+ Allen, J. A., " North American Opossums." 


sentatives of the marsupials, it would not be unreasonable to 
argue that the opossums had invaded North America from the 
southern continent. Mr. Lydekker's * conclusion is that 
opossums are only recent immigrants from the south, 
although he does not believe that South America was their 
original home. He (p. 55) selects south-eastern Asia as the 
birthplace of the opossum family (Didelphyidae), urging that 
the latter scattered 'from this .centre towards Europe and North 
America. He also contends that the allied family Dasyu- 
ridae (Originated in southern Asia, spreading thence to 
Australia, and by an antarctic land connection from there to 
South America. I shall return to this subject in somo of the 
subsequent chapters (p. 283 and p. 366). 

There is still another problem of exceptional interest which 
I wish to enlarge upon, namely, that of the origin of the 
Bermudan fauna. The island of Bermuda has certain 
faunistic affinities with Florida, and we may therefore con- 
sider the origin of its fauna as an appendix to this chapter. 
It consists in reality of a series of about one hundred 
islands and islets, their total area being less than twenty 
square miles. The island of Bermuda, as we may call it for 
the sake of brevity, lies approximately seven hundred miles 
eastward of North Carolina, being apparently surrounded on 
all sides by a depth of from 1,500 to 2,000 fathoms. 
Dr. Wallace,f who gives us a brief description of the 
fauna and flora of the island, concludes that Bermuda 
furnishes us with one of the most instructive facts as to the 
power of many groups of organisms to pass over seven 
hundred miles .of open sea. There is no doubt whatever, 
he remarks, that all the indigenous species have thus reached 
the island. 

I may as well say that my own views differ entirely from 
those of Dr. Wallace as regards the origin of Bermuda and of 
its indigenous fauna and flora. I believe the island to have 
formed part of a wide belt of land, which extended northward 
from the West Indies, joining the mainland of North America 
somewhere near Massachusetts, at a time when most of the 

* Lydekker, E., " Geographical History of Mammals," p. 108. 
t Wallace, A. E., "Island Life," p. 2713. 


existing coast line of the Atlantic States south of Massachu- 
setts was submerged (Fig. 14). I look upon the indigenous 
animals and plants of Bermuda as relicts of an ancient fauna 
and flora. I hold, moreover, that Bermuda furnishes us with 
one of the most instructive facts of the exceedingly slow 
change that many species of animals and plants undergo 
through successive geological ages, and that it does not sup- 
port the theory advocated by Dr. Wallace that many organisms' 
possess the power of crossing seven hundred miles of open 

The geology of Bermuda, so far as the visible structure is 
concerned, is identical with that of the Bahamas, except that 
the coral reefs are of greater importance in the latter. The 
rocks in both are limestone, and red clays resulting from its 
decomposition. Nearly all the rocks of Bermuda above 
sea-level, and to a considerable depth below it, are made up of 
wind-drifted shell sand with very little material derived from 
corals and other organisms. These materials, according to 
Professor Verrill, when consolidated, form a true aeolian lime- 
stone. The island is surrounded by coral reefs in such a 
manner as to give it the appearance of an atoll of the Pacific 
Ocean. It was actually regarded as such by Professor Rice.* 

The greater Bermuda or " Pliocene Bermuda " as it has 
been called, which was once dry land, had an area of about two 
hundred and thirty square miles. That this greater Bermuda 
represents an older land surface was revealed during the exca- 
vations made in 1870 for harbour worts. These extended to 
over fifty feet below sea-water level. At a depth of forty- 
six feet, as Mr. Jones f tells us, a stratum of peat and red 
earth two feet thick was found, containing the vertical stumps 
of cedar trees. This again rested on hard aeolian limestone, 
containing fossil land shells of the genus Poecilozonites. 
There is clear evidence, therefore, of a subsidence of the 
land to the extent of at least fifty feet. A re-elevation to 
that extent would nearly restore the island of greater 
Bermuda. Beyond this, in a south-westward direction, several 
shallows have been detected, all being surrounded by great 

* Eice, W. N., "Geology of Bermuda," p. 9. 

t Jones, J. M., "Recent Observations in Bermudas," p. 262. 


depths, yet indicating that they represent a range of drowned 

According to Professor Verrill it is now generally admitted 
that Bermuda is the flattened and greatly eroded summit of a 
vast submarine volcano, and he thinks it most reasonable to 
suppose that its last activity corresponded in time with the 
last great volcanic eruptions of the nearest American main- 
land. This, as he remarks, would imply that the Bermuda 
volcano was formed or completed during the Triassic Period 
or at its close. Immense outbursts of volcanic material took 
place all along the eastern coast of America at that time, 
giving rise to enormous trap-dykes. In Nova Scotia these 
dykes have a nearly north and south direction, and they may 
have had some direct relation with the volcano of Bermuda. 
It is estimated that the latter had a height of about 15,000 feet.* 

The whole surface structure of Bermuda reminds us vividly 
of the Bahamas. The latter owe their configuration to the 
same process of waste which has been going on during their 
subsidence. The coral reefs surrounding the Bahamas form 
but an insignificant part of the topography of the islands. 
The same aeolian rocks as in Bermuda cover all the visible 
parts of the Bahamas, and we find an intercalation of similar 
red earth. Altogether there is, as Professor Agassiz points 
out, clear evidence of the comparatively recent subsidence 
of at least three hundred feet of the Bahama Bank.j- 

A slightly greater elevation would have had the effect of 
shutting out the Gulf Stream from the northern Atlantic, 
for it now pursues its swift northern course through the 
shallow channel lying between the Bahama Bank and Florida. 
Now it is interesting to note that the ancestral Gulf Stream 
did not flow where it does now, but across northern Florida, 
thus separating the northern from the southern portion of the 
peninsula. Not only were northern Florida and Georgia sub- 
merged. Tertiary marine deposits are known even as far north 
as New Jersey. The sea covered a vast area of the present 
southern Atlantic States. That a strong current flowed 
through the channel of north Florida is evidenced by the fact 

* Verrill, A. E., "Bermuda Islands Geology," XII., pp. 4782. 
t Agassiz, A., " Reconnaissance of Bahamas," p. 7. 


that during early Tertiary times Antillean species were carried 
even as far as New Jersey. At no succeeding Epoch, says Dr. 
Dall, do we find such tropical and semi-tropical molluaks 
extending northward to such a distance from their present 
range. All these Tertiary deposits cease north of the Hudson 
estuary, and I have shown in ,a previous chapter (p. 41) 
that in later Tertiary times, at any rate, the coasts of New 
Yor"k, Massachusetts and Maine extended far out into the 
present Atlantic. The hypothesis of the latter land extension 
having once joined Bermuda and the Bahamas, etc., seems to 
me supported by a variety of circumstances which I shall 
allude to later on. This would have excluded the Atlantic 
Ocean either partially or wholly from the Gulf of Mexico and 
the southern Atlantic States. Some time during the Miocene 
Period, or earlier, a sudden influx of northern species into the 
area hitherto occupied by southern forms occurred. Dr. Dall 
and Mr. Harris * endeavoured to account for this phenomenon 
by the supposition that the course of the Gulf Stream was 
gradually turned more off shore than it was before or is at 

If we assume, however, that a belt of land such as above 
described had hitherto existed, the gradual breaking down of 
its northern portion might have admitted the Atlantic waters 
into the sea which covered the southern States and have 
brought with it the new fauna, which had meanwhile deve- 
loped in the northern Atlantic Ocean (see Figs. 14 and 16). 
For a time these northern, cooler inshore waters were even 
able to penetrate into the Gulf of Mexico. Even if we grant 
the correctness of Messrs. Dall and Harris's supposition of the 
altered course of the Gulf Stream, the cause of this deflection 
is more likely to have been produced by a change in the con- 
figuration of the northern land-masses than by that of Florida. 

Let us now study the flora and fauna of Bermuda, and 
endeavour to ascertain whether it supports in any way the 
theory I have advocated. f It is quite evident that the existing 
flora of Bermuda is only a remnant of the original one, before 
the early settlers, accompanied by hogs and rats, played havoc 

* Dall, W. H., and G. D. Harris, "Correlation Papers," pp. 185187. 
t Heilprin, A., " The Bermuda Islands." 


with it. The old records speak of thousands of palmetto trees 
that were cut down, and of cedars used for the construction 
of ships and buildings. The palmetto (Sabal blackburniana) 
is a species of palm much like that of southern Florida, but 
differing sufficiently to form a distinct species. All the islands 
were once thickly covered with the cedar (Juniperus bermu- 
diana), and it also occurs in a semi-fossil condition in the 
red earth. Among the one hundred and fifty six kinds of land 
plants now considered native to Bermuda, about fifty species 
are very restricted in their distribution, most of them being of 
West Indian origin. Altogether one hundred and eighteen 
species are native of the West Indies. Of these ninety are 
also found in Florida, the remainder being only met with in 
the West Indies. About the same number of plants as are 
common to Bermuda and the West Indies only, are also 
peculiar to Bermuda and continental North America. A few, 
such as the blue-eyed grass (Sisyrinchium bermudianum), 
have their nearest relatives in the north.- eastern States. Most 
of the botanists who have investigated the Bermuda flora, 
especially Mr. Hemsley, Mr. Moseley and more recently Pro- 
fessor S. Brown, Dr. Harshberger and Prof essor Yerrill * are 
agreed that all the native plants of the island have been intro- 
duced by natural agencies prior to the advent of man. Pro- 
fessor Verrill is of opinion, and I feel sure his view is 
almost generally accepted, that migratory birds have prob- 
ably always been the chief agencies for these introductions 
of plants, especially those from north-eastern North America. 
Currents, floating timber and hurricanes are also supposed 
to have had their share in transporting plants from various 
parts of America to Bermuda. I am among rthe few. who do not 
recognise the potency of these agencies of transport. That 
seeds are occasionally cast by currents upon the shores of 
lonely islands like Bermuda has been amply proved, but there 
is no evidence available to show that migratory birds distribute 
seeds on siuch islands, or that hurricanes carry seeds across 
seven hundred miles of sea and safely deposit them on an 
island. These theories are very widely accepted, but I think 
on insufficient grounds. 

* Verrill, A. E., "Bermuda Islands," XI., pp. 587588. 


Dr. Wallace * makes the statement that the few species of 
birds which are resident in Bermuda and breed on the island 
are so constantly crossed by individual migrants of the same 
species from the mainland that none of the former have 
acquired any .special peculiarity constituting even a distinct 
variety. It is perfectly true that the resident land birds 
are very few in number, but the opinions, of most recent autho- 
rities differ very considerably from those enunciated by Dr. 

Only ten species of land birds are resident on the island, 
and three of these have been introduced by man. The indi- 
genous fauna includes, therefore, seven "kinds of birds, 
namely, the ground dove, the Florida gallinule, the crow, the 
white-eyed Vireo, the blue bird, the cat bird and the cardinal. 

Of these the Bermuda ground dove (Columbigallina bermu- 
diana) is described as a species very distinct from the 
American C. passerina. The Florida gallinule (Gallinula 
galeata) does not seem to differ from the continental form. 
The crow of Bermuda is probably separable from the American 
crow, but has not been sufficiently studied. The white-eyed 
Vireo (Vireo bermudianus) is described by Messrs. Bangs 
and Bradlee as very different in all its ways from its shy, 
retiring, continental relative, V. noveboracensis.f The blue 
bird is not usually recognised as distinct from the continental 
form, but Dr. Sharpe of the British Museum held that its 
characters were certainly striking enough to deserve specific 
recognition, and he described it as Sialis bermudianus. Now I 
may mention that the blue bird belongs to one of those genera 
which seem to have retained their generic characters for very 
long ages past, probably throughout several geological periods. 
We know nothing of its past history from palaeontological evi- 
dence, but its present geographical distribution, and that of- 
its nearest relative, are so peculiar as to suggest their being of 
great antiquity. Sialis, with its three closely-allied American 
species, has its nearest relative (Grandala) in the Himalayan 
Mountains. I only mention this fact so as to show that the 
apparent specific identity of the Bermudan and the continental 

* Wallace, A. E., " Island Life," p. 269. 

t Bangs, 0., and T. S. Bradlee, " Birds of Bermuda," pp. 249257. 


blue bird need not necessarily be adduced as an argument 
for the recent geological origin of the former. "The Bermuda 
catbird (Galeoscoptes bermudianus), though closely resem- 
bling the continental species, is considered distinct by Messrs. 
Bangs and Bradlee and by Dr. Sharpe. Finally, the Bermuda 
cardinal (Cardinalis bermudianus) can at once be identified 
from its continental relative by its characteristic bill. 

Dr. Wallace alludes to the Bermudan lizard as being the 
only vertebrate animal which exhibits any peculiarity. But 
much more can be said about it. It is a member of the family 
of ,skinks (Scincidae), which is of cosmopolitan range and 
undoubtedly of great age. The Bermudan skink (Eumeces 
longirostris) is said to be nearly related to the American blue 
tailed lizard (Eumeces quinquelineatuts), still we must remem- 
ber that the latter has long been considered identical with the 
Japanese skink (Eumeces latiscutatus) from which it can 
be separated only by very careful study.* It may safely be 
argued, therefore, that the Bermudan skink has originated in 
early Tertiary times (compare pp. 123 126). 

Dr. Rehn f records twenty-eight species of Orthoptera, two 
of them, viz., Paroxya bermudensis and Gryllus bermudensis, 
being peculiar to the island. If we exclude the cosmopolitan 
and circumtropical forms which the author regards as prob- 
ably introduced by commerce, the remainder are more closely 
related to continental than to Antillean species. 

Only a single ant is peculiar to Bermuda, viz., Prenolepis 
kincaidi, because some of the more active recently introduced 
warlike species have no doubt exterminated the greater part 
of the older resident ant fauna. J 

Mr. Pocock recognised two species of myriopods from 
Bermuda as doubtfully distinct, though he described them as 
new species. These are Lithobius provocator and Lithobius 
bermudensis. Altogether he believes that only six species 
of centipedes ,and millipedes are native to Bermuda, and that 
three of them belong unquestionably to the Mediterranean 

* Garman, Sam, " Eeptiles of Bermuda," p. 287. 
t Behn, J. A. GK, " Orthoptera of Bermuda," p. 3. 
% Wheeler, W. M., "Ants of Bermuda," p. 347. 

Pocock, E. L, " Myriopoda of the Challenger Expedition," pp. 123 


fauna. This is a point of considerable interest and import- 
ance, and requires to be more closely studied in a recon.- 
sideration of the true relationship of the members of other 
groups of invertebrates. 

This relationship is clearly recognisable among the ter- 
restrial Isopods or wood-lice, which have been described 
by Miss Eichardson. Tylos latreilli, as already quoted, 
is a typical Mediterranean species, which has been dis- 
covered in southern Florida, where also another species, 
Tylos niveus, occurs. Both of these are now recorded from 
Bermuda. The European genus Porcellio is represented by 
two species, one of them (Porcellio parvicornis) new to 
science. Metoponorthus sexfasciatus, a typically Mediter- 
ranean species, also found in the Canaries and Azores, occurs 
in Bermuda, but nowhere else in America. Another species 
peculiar to Bermuda (Philoscia bermudensis) is closely 
related to the west European Philoscia couchi. The 
American affinities are likewise surprisingly interesting. 
Uropodias seems to be related to the West Indian genera 
Haplarmadillo and Sphaeroniscus, and this strikingly distinct 
genus is quite confined to Bermuda with the one species, Uro- 
podias bermudensis. The truly archaic Actoniscus ellipticus, 
which is only known from Bermuda and from the coast of the 
mainland near New Haven and Long Island Sound, is one 
of the only two members of the family Trichoniscidae. Its 
distribution is suggestive of a former land connection towards 
north-eastern North America. The other Actoniscus is con- 
fined to California. Finally, Leptotrichus granulatus, also 
peculiar to Bermuda, may be mentioned as the only occurrence 
of a very ancient Old World genus in America.* 

The only native spider which, according to Dr. Verrill,f was 
mentioned by 'the early writers, was the great silk spider 
(Nephila clavipes). The enormous webs which this spider 
constructs between trees at a distance of fifty feet from one 
to another excited their admiration, and suggested to them 
that the threads might be used in the manufacture of silk 
tissues. This has actually been done in Brazil, I believe, 

* Eichardson, Harriet, "Isopods of Bermuda." 
t Verrill, A. E., "Bermuda Islands,": XL, p. 829. 


where this spider occurs. The silk spider is known from 
Brazil, from Central America as far north as Texas, and from 
Southern Florida. Professor Dahl * argues that the group to 
which this species belongs already had a wide range in Cre- 
taceous times. Its nearest relation, Nephila clavata, is pecu- 
liar to China and Japan. It is quite possible, therefore, that 
the American form may have existed since early Tertiary 
times. But, as Dr. Marx f observes, the spiders introduced by 
human agency have not only acclimatised themselves ; they 
havo also, in a more or less marked degree, driven away and 
exterminated the indigenous spider fauna. Yet among the 
remnants of that ancient fauna we recognise four species that 
are peculiar to Bermuda. Most of the remainder are identical 
with American forms. 

One of the most noteworthy animals that might be adduced 
in favour of the theory that Bermuda represents the remnant 
of a large land surface once connected with the mainland is 
the land nemertean Geonemertes agricola. This very pecu- 
liar worm is found in abundance at several distinct localities 
in Bermuda. It might be said that this worm is not of much 
zoogeographical value, since it is able to resist immersion in 
sea -water for some time. On the other hand, it must be re- 
membered that land nemerteans have an extremely discon- 
tinuous range on ancient land surfaces. Hence their distribu- 
tion is suggestive of great antiquity. Nine species of the 
genus Geonemertes are now known from Australia, New Zea- 
land, New Guinea, Pelew Islands, Eodriguez Island, the 
Seychelles and Bermuda. To suggest that Geonemertes 
agricola owes its presence in Bermuda to an accidental intro- 
duction would be quite inadmissible, because it has neveu 
been found elsewhere. J 

Of greater importance than any of the foregoing inverte- 
brates are the land and fresh-water snails inhabiting 
Bermuda. Dr. Pilsbry distinguishes autochthonous species, 
drift waifs from the West Indies, and snails imported by the 
agency of man. The last group we need not consider here, 

* Dahl, F., " Verbreitung d. Spinnen," p. 278. 
t Marx, G., " Spiders of Bermuda," p. 100. 
| Coe, W. E., " Geonemertes agricola," p. 534. 


although some .members of it are probably indigenous. The 
remainder include several snails, only once recorded, and 
not again found on the island in recent times. iThe 
so-called waifs from the West Indies are of greater signific- 
ance. Dr. Pilsbry identifies Succinea bermudensis with 
S. barbadensis, yet acknowledges that the shells of this 
genus are peculiarly uncharacteristic, and that species of 
different regions frequently resemble each other. All oon- 
chologists, however, are agreed that the semi-amphibious 
amber-snail (Succinea), with its almost world -wide distribu- 
tion, must be a very an,cient one. The mere fact of several 
Bermudan species being identical with West Indian ones 
is no proof that they were conveyed to Bermuda by accidental 
means of transport. I have argued this point again and again, 
but it is a widespread assumption which can only be effectu- 
ally disproved by palaeontological evidence. No such evidence 
is available in the majority of cases. Yet of some of these sup- 
posed accidentally and recently introduced species of Bermuda 
I might mention Rumina decollata. It is certainly native in 
the Mediterranean region, where it exists, as a relict o the. past, 
and I have given a map of its range in my work on European 
animals.* The family Stenogyridae, to which it belongs, is 
an entirely tropical one. Rumina decollata has adapted itself 
to the European climate, though its shape has remained un- 
changed since Oligocene times. It has been known to exist 
outside Europe in Cuba, South Carolina and Bermuda. Are 
we justified in the assumption that this exceedingly old 
member of a tropical family of snails has been accidentally 
introduced into these localities ? I think not, and yet this 
surmise is received by almost everybody as an established 

The really interesting members of the Bermudan fauna of 
mollusks are those styled "autochthonous" by Dr. Pilsbry, 
viz., Helicina convexa, Thysanophora hypolepta, all the 
species of Poecilozonites, and the slug Veronicella schivelyae. 

I have already dwelt upon the distribution of the genus 
Helicina (p. 158), and on its occurrence in the Oligocene 
Silex beds of Tampa in Florida, and have indicated that it 

* Scharff, E. F., "European Animals," p. 222. 


represents a relict of early Tertiary times. The ancestors of 
Helicina convexa, which is peculiar to Bermuda, reached the 
island presumably in the Oligocene Period, when we know 
that the genus inhabited the island of Florida. 

Thysanophora is a member of the large family of Helicidae. 
It is most characteristic of the West Indian region and Central 
America, and Dr. Pilsbry * declares that it represents the 
oldest stock of the West Indian Helix fauna. The ancestors 
of Thysanophora hypolepta may well have wandered to the 
Bermudan area from the West Indian region in early Tertiary 
times. This suggestion is strengthened by the fact that along 
with Succinea and Poecilozonites it is found fossil in the 
" red earth." The genus Poecilozonites, with its four 
Bermudan species, is peculiar to Bermuda. Recently Pro- 
fessor Boettger f again drew attention to the intimate rela- 
tionship existing between this genus and the German lower 
Miocene Helix imbricata, in spite of Dr. Pilsbry's plea that 
the German savant was only chasing an " ignis fatuus." Dr. 
Pilsbry J prefers to derive Poecilozonites from an old 
American stock, arguing that it is related to Gastrodonta. To 
trace the exact affinity of a recent to a fossil form is always 
a matter of considerable difficulty, but since several other 
European palaeontologists have dwelt upon the remarkable 
and undoubted relationship existing between the present West 
Indian fauna and the fauna of the European early Tertiary 
deposits, and as I shall be able to bring forward additional 
evidence in favour of their views, I am certainly on Professor 
Boettger's side in this controversy. That Poecilozonites is 
not a recent arrival from elsewhere, quite apart from the fact 
that it is unknown outside the Bermudan area, is likewise 
indicated by its fossil occurrence in the red earth, which' 
is held to be of Miocene or Pliocene age. Dr. Gulick records 
eleven species and sub-species of Poecilozonites from the red 
earth, as well as Succinea bermudensis and a number of 
genera that have since become extinct. 

* Pilsbry, H. A., " Manual of Conchology," Vol. IX., p. 56. 
t Boettger, O., " Helix- Arten aus d. Tertiar Europas," p. 103. 
I Pilsbry, H. A., "Mollusks of Bermuda," pp. 491509. 
Gulick, A., " Fossil Land Shells of Bermuda." 
L.A. O 


Mr. Vanatta * recently made the interesting discovery that 
several typical fresh-water species inhabit Bermuda. Among 
them there is a Physa, two Planorbis, one Ancylus, a Palu- 
destrina and a Pisidium. All are new to science. 

The genus Veronicella (Vaginula), of which a species (Y. 
schivelyae) inhabits Bermuda, includes slug-like creatures 
which should certainly not be suitable for accidental dispersal 
by the usually quoted agencies. It is of interest, therefore, to 
note that Messrs. Robbing and Cockerell f record a variety of 
the Bermudan Veronicella from the Bahama islands, and 
direct attention to its close relationship with the Mexican 
Veronicella moreleti. 

From the foregoing brief survey of the Bermudan fauna it 
is manifest that although that most destructive of all creatures 
" man," has played havoc with the native animals and plants, 
largely exterminating them, a recognisable residue has sur- 
vived from remote times. As far as the ancestry of this relict 
fauna can be traced, it seems to be derived from the south- 
eastern and north-eastern States of America, from the West 
Indies and even from southern Europe. Many species, par- 
ticularly such as are unfit to take advantage of accidental 
modes of transport, namely, the Bermudan skink, the 
terrestrial nemertean worm, and various kinds of terrestrial 
isopods and snails, are peculiar to Bermuda. Is there any 
reason for the supposition that these animals, all of which 
bear the impress of vast antiquity on them, were conveyed 
to Bermuda in the past by accidental means ? We have no evi- 
dence whatever that terrestrial invertebrates are transported 
alive across seven hundred miles of sea and subsequently pro- 
pagate their kind except through the agency of man. Why 
should we assume, therefore, that Bermuda has received its 
entire fauna and flora by accidental means such as winds, 
hurricanes or ocean currents ? It is evident that the great 
depths of the ocean surrounding the little group of islands is 
the chief, if not the sole, stumbling block to the acceptance 
of the survival theory. Yet in view of the fact vouched for by 

* Vanatta, E. G., "Bermuda Shells," pp. 668672. 
t Bobbins, W. W., and T. D. A. Cockerell, "Veronicella," p. 383 


Professor Suess, that a subsidence of land amounting to 
10,000 feet has taken place between the island of Khodes and 
the mainland of Syria in post-Pliocene times, whereas marine 
deposits of late Pliocene age occur in the southern Andes at 
a height of 5,000 feet, is it really such an unlikely theory that 
the supposed ancient belt of land connecting Bermuda with 
a southern land-mass has subsided to a similar extent within 
comparatively recent geological times ? I need not dwell any 
longer on this problem. When I come to deal with the 
Antilles and the origin of their fauna further allusions to it 
will be made. 




THE region to be dealt with in this chapter is of enormous 
extent. To speak of it vaguely as south-western North 
America is apt to give the impression as if only Texas 
and the adjoining western States "were being considered. 
What is described here under that designation contains almost 
the whole of the western States with Mexico in addition. 
There is material enough for several chapters. Nevertheless, 
the profound zoogeographical importance of this portion of 
North America will be more readily realized, I think, if the 
leading facts are compressed into one brief essay. The Rocky 
Mountain fauna, moreover, which enters largely into the 
south-western region, has already formed the substance of a 
special chapter. 

The term " Sonoran Region " was first applied by Pro- 
fessor Cope to part of this important south-western life area, 
because he recognised the striking difference it presented as 
compared with other parts of the United States. In one of the 
previous chapters I alluded to the scarcity of lizards in the 
eastern States in comparison with western Europe. I might 
have said " western States of North America " instead of 
western Europe, for lizards especially, and reptile life in 
general, abound in the region we are now considering. 

The term " Sonoran,"* derived from Sonora, a province of 
northern Mexico, is now often applied to the fauna of the 
whole of North America excejpt the boreal portion ; and Pro- 
fessor Carpenter f supports the contention that the continent 
can be readily divided into two faunistic sections, one of which 
is distinctly autochthonous, the other derived from northern 
Europe and Asia. I have already put forward my reasons for 

* Merriam, 0. H., "Distribution of Life in North America," p. 26. 
t Carpenter, GL H., "Nearctic or Sonoran ?" pp. 5357. 


abandoning faunistic regions in a work of this kind, and I 
feel there is little advantage in discussing the merits of one 
regional system above another. I only wish to direct attention 
to the fact that the distinctness and importance of this south- 
western part of North America had long ago been recognised 
by zoologists. Professor Carpenter urges that two distinct 
faunas exist in America. I can trace even more than two, 
for a South American element is very prominently diffused 
throughout a large portion of the States. I cannot concur 
at all in Dr. Merriam's view* that "except for the. 
presence, chiefly in the southern United States, of a compara- 
tively few forms derived from the tropical region, the fauna 
and flora of North America are as distinctive and indepen- 
dent of the existence of this area as if separated from it by 
the broad ocean." To place ourselves within fixed and 
strictly limited boundaries at all seems to me a mistake. I 
do not limit myself in any way to political frontiers, and 
if I had set myself a boundary, I should: have been inclined to 
place it across the isthmus of Tehuantepec. At this point 
North America was evidently separated for some time from; 
Central America by a marine channel, though this division 
was not so effectual in keeping two great faunas distinct as 
one might expect. All this, however, will be discussed at 
greater length when we come to deal with the fauna of Central 

Although vast tracts of south-western North America are 
nothing more than deserts, there is, on the whole, an extra- 
ordinary abundance and variety of animal life. I have 
repeatedly pointed out in previous chapters that the roots 
of certain groups of eastern animals must be looked for in 
the south-west or west. In drawing attention to the strange 
affinities of some apparently very ancient east-American 
forms, such as the smooth and the rough green snakes, and 
east-Asiatic snakes, I urged that they must originally have 
spread eastward from south-western North America (p. 125). 
Among the tortoises, too, certain eastern groups can be 
traced to a remote western origin, although no longer, 
resident there. Thus there has arisen a comparatively modern 
repetition of that dissimilarity between the eastern a,nd 

* Merriam, 0. H., " Distribution of Life in North America,'' p. 37. 


western faunas which was the characteristic feature of the 
two ancient land-masses during Mesozoic times when the 
latter were separated from one another by a wide ocean. 
Dr. Pilsbry * once advocated the view, from a conchological 
standpoint, that the Sierra Nevada divided North America 
into two primary f aunal provinces, a division which he thought 
commenced in Jurassic times with the upheaval of this moun- 
tain range, the latter having proved insurmountable to most 
land snails. The true autochthonous American fauna east 
of the Sierra Nevada range seemed to him to have developed, 
with very little or no trace of Old World influence, from an 
early period, probably the Cretaceous. 

This -difference between the east and west is in many 
respects less pronounced when we include Mexico in our 
western district. Few species of the characteristically 
eastern snail Polygyra are found in California. Further 
south, however, in Mexico, their numbers again increase. The 
family of perches (Percidae) are quite absent from the western 
States. In Mexico they again appear. Among the garpikes 
(Lepidosteus) and the Unionidae we notice a similar distribu- 
tion, but it is not discontinuous. Their range passes gradu- 
ally from the Gulf States into Mexico. Among other groups 
or genera a greater discontinuity of distribution is noticeable 
between the Mexican and the east American centres of dis- 
persal. The tribes that are affected by desert conditions, 
and many of them are, would find the arid regions lying 
directly to the north of Mexico effectual barriers to dispersal 
in that direction. But we have good reasons for the belief that 
these desert conditions are comparatively modern develop- 
ments, and have little to do with the origin of the distribution 
referred to. The geological history of the south-western 
States may throw further light on this problem. 

From Cambrian to Jurassic times the greater part of Cali- 
fornia was raised above sea-level. Towards the end of the 
Mesozoic Era there existed in western North America a 
broad strip of land running north and south being bounded 
on each side by the sea.f Mexico must have extended further 

* Pilsbry, H. A., " Check-list of Land Shells," pp. 194195. 
t Schuchert, Charles, " Paleogeography of North America." 


west at that time, since no late Mesozoic. deposits are known 
from the western parts of that country. If we supposed that 
western Mexico had then been connected with some other 
land surface, a faunistic interchange could have taken place 
between the latter and western North America. 

In early Tertiary times the central sea, which formed the 
eastern boundary of the western belt of land referred to, had 
almost disappeared from the interior of America, but large 
tracts of western California were still under water (see 
Fig. 14). Professor Smith* argues that a temporary con- 
nection must have existed during the Eocene Period between 
the Atlantic and the Pacific Ocean, because the. cha- 
racteristic Atlantic shell Venericardia planicosta had been 
met with in the Eocene deposits of California and Oregon. 
In Oligocene and Miocene times the whole of the eastern 
borders of Mexico were submerged, while the sea was at 
first retreating from western California and then again 
invading it. During the latter part of the Miocene Period 
the sea even encroached on western Mexico, f All the 
same, certain parts of the coast ranges in western Cali- 
fornia never seem to. have been entirely submerged during 
Tertiary times and probably formed part of the Pacific land 
belt which has now almost entirely vanished. I think 
the alternative union and disruption of these western Cali- 
fornian land-masses with the mainland of North America 
must have played an important role in the origin and 
development of the American fauna. It seems as if Mexico 
had at first formed the stepping-stone to North America for 
new immigrants, and later on western California. I have 
endeavoured to represent this idea on two maps (Figs. 14 and 
16), but how the changes were actually brought about has not 
been made quite clear to us through geological research. 

It has been suggested by Messrs. Ordonez and Aquilera that 
the Cape portion of lower California really forms the western 
continuation of the Mexican Sierra del Sur.J But the very 
important question now arises from a zoogeographical point 

* Smith, Perrin, " Geological History of California," pp. 347348. 
t Arnold, Ealph, " Tertiary Pectens of California." 
I Suess, E., " Antlitz d. Erde," III 2 , p. 487. 


of view, how far westward and southward of the Gulf of Cali- 
fornia did this land extend ? No Cretaceous or Tertiary 
deposits have yet been discovered on the greater portion of 
the Pacific coast of Central America. Geological evidence is, 
therefore, rather in favour of the supposition that this western 
land formerly extended further south. 

As regards the present arid semi-desert conditions of 
Arizona and New Mexico, they supported in early Tertiary 
times, as I have already mentioned, a wealth of animal 
life. In the neighbouring state of Colorado the wonderfully 
preserved impressions of insects in the volcanic tuffs of 
Florissant, which have been described by Dr. Scudder and 
more recently by Professor Cockerell,* leave no doubt as to the 
former climatic conditions of that part of America and its 
suitability for plant and animal life. The vast outpouring 
of lava and general volcanic disturbance in the Rocky Moun- 
tain region continued through Miocene and partly through 
Pliocene times. All the same, the immigration of tropical 
types of mammals from South America into the Western 
States at that time would seem to imply the existence in the 
latter of a luxuriant flora. Even in Pleistocene times an 
abundance of large mammals, such as elephants and masto- 
dons, existed in southern California, and probably in the 
neighbouring States, to judge from the number of sabre- 
tooth tiger remains recently discovered in the asphaltum 
beds of Rancho la Brea near Los Angeles. The gradual 
desiccation noticeable in some of the south-western States 
is obviously a recent development, though the abundance 
and diversity of cactuses and of reptiles adapted to a desert 
life imply that local arid areas must have existed for long 
ages past. 

If the geological history of the extreme south-west of North 
America has been correctly interpretated in this very brief 
summary, we should certainly find relicts of ancient animal 
and vegetable types in some of the western areas that have 
remained unsubmerged during Tertiary times. For although 
most animals would tend to spread from these old centres as 
new land became available for their dispersal, some of the 

* Cockerell, T. D. A., " Fossil Fauna and Flora of Florissant." 


less active and less vigorous types must have remained in their 
ancient western habitats. 

Among the order Insectivora which is now on its way 
towards extinction, several peculiar mammals, such as 
Atophyrax, Notiosorex and Scapanus are entirely western 
in range. The most noteworthy is what has been called the 
mole-shrew (Neurotrichus), on account of its shrew-like look 
combined with its digging habits. The single species Neuro- 
trichus gibbsi is a western rather than a south-western 
animal, being confined to northern California and Washington 
State. It has no near relations in America. Its nearest akin 
inhabit eastern Asia and Europe, but it has not yet been 
(Satisfactorily established to which of its two Old World 
tranches it has most affinity. The whole group to which all 
these insectivores belong is known from the European 
Eocene onward. Nowhere else have fossils been found. 
Palaeontological evidence would, therefore, point to Europe 
as the centre of evolution. Possibly the ancestors of Neuro- 
trichus may have originated somewhere in the Mediter- 
ranean region, have crossed the Atlantic by a mid-Atlantic 
land bridge to California, as I shall explain later on, and have 
then passed to Japan, where a closely related genus occurs. 

The mole-mice (Onychomys), which are typical rodents, 
constitute a section which has evidently originated in the 
south-west, and has since spread northward and eastward so 
as almost to reach the Mississippi, though none of them have 
crossed this river. The wood-rats, belonging to the genera 
Nelsonia, Xenomys, Neotomodon, Teanopus and Hodomys 
are all confined to small areas in the south-west. The four- 
toed and the five-toed kangaroo-rats (Dipodomys and Pero- 
dipus) are almost all peculiar to the south-western States. 

Among the carnivores with a similar range may be men- 
tioned the raccoon foxes (Bassariscus) and the western skunks 

The most striking examples tenanting the ancient south- 
western land areas must be looked for, as I remarked, among 
the more sessile and slow-moving creatures. The reptiles and 
amphibians yield good instances. Among the former the most 
noteworthy is the so-called " Gila monster," a repulsive, stout, 
thick-tailed lizard. It has gained the unenviable notoriety 


of being the only poisonous lizard in existence. The genus 
Heloderma, to which it belongs, ranges from Mexico, through 
Arizona, to western Texas. No near relations of the Gila 
monster are known, but it possesses some distant affinity to 
a lizard living in Borneo. 

Better known probably are the horned toads (Phrynosoma). 
Owing to their greater activity they have been able to spread 
much further north and east of their original centre of dis- 
persal. The horned-toads are lizards belonging to the large 
and important family Iguanidae which has a peculiar dis- 
tribution suggestive of great antiquity, as I have already indi- 
cated (p. 126). I mentioned that the family lived in America 
as far back as Cretaceous times, and that beyond that con- 
tinent it was only known from the Fiji Islands and Mada- 
gascar. Leaving the latter out of consideration for reasons 
stated (p. 126), we have only to discuss the origin of the Fiji 
members of the family. The geological history of the Poly- 
nesian fauna will be fully dealt with in a subsequent chapter. 
I may mention, however, that I believe in the former exist- 
ence of an ancient circum -Pacific belt of land which was 
joined to south-western North America (Fig. 14), and that 
the Iguanidae passed across this land during their wanderings 
from America to Fiji or vice versa. 

Let us turn from the active Iguanidae, and take, as an 
example, a slowly-moving creature such as the Californian 
limbless lizard Anniella. It inhabits barren sand-dunes, lying 
buried in the sand and exposing only the anterior part of the 
head.* Not only is it a reptile which spreads very slowly, but 
it is eminently an animal requiring a continuous land surface 
for its dispersal. As might be expected, the genus Anniella 
is quite peculiar to the south-western States. No other 
member of the family Anniellidae is known, though it is 
closely related to the Anguidae, which are almost confined in 
their distribution to America and Europe. 

We also possess a single species of tha,t remarkable family 
of burrowing lizards, the Amphisbaenidae, in lower California 
and Mexico, viz., Euchirotes biporus. It is a significant fact 
that two peculiar genera of that ancient family occur in the 

* Coe, W. E., and B. W. Kunkel, " Californian Limbless Lizard," 
pp. 350351. 


two southern corners of North America, one (Rhineura) 
in Florida, the other (Euchirotes) in Lower California and 
Mexico. The only North American members of the blind 
snakes (Glauconiidae) are limited in their range to the south- 
western States. One of them (Glaucoma dulcis) lives in 
Mexico, New Mexico and Texas, the other (Glaucoma humilis) 
ranges from California as far as Arizona. These degenerate 
worm-like creatures are entirely subterranean, and feed on 
earthworms and larvae of insects. Hence their distribution 
is of great zoogeographical value. Besides the districts re- 
ferred to the family is found in the Lesser Antilles, Central 
and South America, south-western Asia and Africa. That 
these snakes should have passed all through the continent of 
North America and through northern Asia in spreading from 
South America to Africa or vice versa without leaving a trace 
of their former wanderings seems to me very unlikely. 
Yet geographical distribution of that kind is frequently ex- 
plained by the supposition of a former Bering Strait land 
bridge offering the only means of land communi cation between 
the Old World and the New. There being no fossil evidence 
to guide us, we must judge such cases altogether from the 
present distribution, and it appears to me that an ancient 
land bridge across the mid-Atlantic explains the latter more 
satisfactorily than the other hypothesis. 

The theory of the former existence of such a, land bridge is 
not built upon a single instance of distribution. I have 
mentioned many others in previous chapters, and I shall 
allude to several in subsequent ones. One other striking 
example may appropriately be mentioned here, viz., the dis- 
tribution of the boas (Boidae) . These are mostly large and 
active snakes. Nevertheless, they are related to the small 
and slowly-moving blind snakes, because, like them, they pos- 
sess rudiments of a hip-bone and hind limbs. Boas inhabit 
all tropical and sub-tropical countries. Only in two dis- 
tricts do they pass into temperate climates, viz., in the south- 
western States of North America and in south-eastern 
Europe. In these countries are found the two closely-allied 
genera. Lichanura and Eryx. One of the North American 
boas (Lichanura trivirgata) is confined to the extreme south 
of Lower California, another to southern California and 


Arizona, while a third, the rubber boa or silver snake actually 
passes into Washington State and central Nevada. 

The recent increase of dryness in the south-west has no 
doubt affected the original fauna considerably. Semi-aquatic 
forms and aquatic ones have either been destroyed or forced 
to take refuge in the neighbouring States where climatic con- 
ditions were more favourable. To this cause may be attri- 
buted the present scarcity of amphibians in most of the south- 
western districts, where many of them, we may presume, 
originally had their headquarters. The only American mem- 
ber of certain toads, which on account of the peculiar shape 
of their tongues have been called Discoglossidae, occurs in 
Washington State in western North America. This family 
has always been looked upon with particular interest, because 
to it belongs the solitary amphibian known from New Zealand. 
This toad (Liopelma) must have reached New Zealand, ac- 
cording to Dr. Stejneger,* before Cretaceous times by means 
of a very ancient land connection with the north. The same 
author expresses the opinion that the Himalayan Mountains, 
or rather the region to the south-west of them, was the ori- 
ginal home of these discoglossoid toads, and that they spread 
from there to New Zealand and North America. On the latter 
continent we still find, as I remarked, a single genus of the 
Discoglossidae (Ascaphus). The present centre of dispersal 
of these discoglossoid toads is southern Europe, since three 
genera are found there, viz., Discoglossus, Alytes and Bom- 
binator. The first two are strictly European, whereas a single 
species of Bombinator also inhabits northern China and Korea. 
Considering the fact that these toads do not occur in south- 
western Asia, and that both Discoglossus and Bombinator 
have been found in European Miocene beds, the Mediter- 
ranean Region seems more likely to have been the original 
centre of dispersal than south-western Asia. At any rate, 
that event leads us, no doubt, to the dim and distant past of 
the early part of the Mesozoic Era. 

The allied family Pelobatidae is likewise of great faunistic 
interest, as the two genera Scaphiopus of North America and 
Pelobates of Europe are only distinguished by slight differ - 

* Stejneger, L., "Distribution of Discoglossoid Toads," pp. 91 93. 


ences. The spade -foot toads (Scaphiopus), as they have been 
called, have their headquarters in Mexico and the south- 
western States, whence they have spread northward along the 
Pacific coast and eastward to the Atlantic States.* One 
species, the hermit spade-foot (S. holbrooki) is worthy of 
special note, owing to its occurrence on Martha's Vineyard 
Island, in view of the remark I previously made (p. 183) 
that this and other north-eastern islands are the remnants 
of an ancient land mass formerly connected by land with the 
mid-Atlantic land bridge. 

The faunistic relationship between Europe and south- 
western North America deserves an ample discussion, being 
one of the most important and noteworthy features of the 
American fauna. Besides the points already referred to 
there are a few others which throw light on the nature and 
origin of this relationship. 

I do not intend to dwell on the merits of the various classi- 
fications which have been adopted for the land mollusks. Even 
if we base our system' upon purely anatomical grounds, dif- 
ferent points of view may sometimes give rise to considerable 
differences in classification. Thus Dr. von Ihering f contends 
that the family Helicidae, as far as North America is con- 
cerned, is entirely confined to the western States. Dr. 
Pilsbry,J on th,e other hand, in including Polygyra among 
the Helicidae, naturally records for that family a very 
different range. 

Both authors agree in attaching great importance to the 
molluscan fauna of the western and south-western States, 
and both are inclined to attribute its origin to immigration 
from Asia. That Dr. Pilsbry supposes the mollusks to have 
wandered across an old Bering Strait land bridge, while Dr. 
von Ihering argues in favour of a much more southern 
Oligocene land connection between eastern Asia and Cali- 
fornia, is not perhaps of such fundamental importance. The 
general agreement is all the more remarkable, as both 
authors have formed radically different views on the interpre- 

* Gadow, H., " Mexican Amphibians and Eeptiles," p. 205. 
f Ihering, H. von, " Yerbreitung der Heliciden." 
t Pilsbry, H. A., " Manual of Conchology," Vol. IX., p. xxxii. 
Ihering, H. von, " Verbreitung der Heliciden," p. 442. 


tation of some anatomical facts. Dr. von Ihering supports 
the opinions, upheld by almost all the prominent European 
concholOgists, that the west American shells, which so greatly 
resemble the European Ariantas and Campylaeas, are really 
genetically connected with the latter. He even contends that 
no anatomical distinction worth mentioning exists between 
some of the American Epiphragmophoras and the European 
Arianta. Dr. Pilsbry, on the other hand, does not recognise 
any very intimate anatomical relationship between the Euro- 
pean and west American Helicidae, except in so far as he con- 
siders both to be branches of an ancient south Asiatic stock. 
As I shall explain later on, I concur with Dr. von Ihering, 
Dr. Sandberger, Professor Oppenheim, Professor Boettger and 
Dr. Kobelt in the opinion that many of the Antillean and 
west American forms are intimately related to extinct and 
recent European Helicidae. The great resemblance of the 
shells of Arianta and some of the Epiphragmophoras alone 
seem to imply a near relationship between them. 

The genera Ashmunella, Sonorella and Oreohelix are quite 
peculiar to south-western America, a large number of species 
having been described by Dr. Pilsbry and Mr. Ferris * in a 
series of articles contributed within the last few years to the 
Proceedings of the Academy of Natural Sciences of Phila- 
delphia. The same region contains many other distinct genera 
of land mollusks. The Urocoptidae, a family of snails some- 
what resembling the European Clausilia in shape, and having 
their headquarters in the Antilles, are represented by two 
genera, viz., Holospira and Berendtia, peculiar to south- 
western North America. 

Of some interest is the occurrence in southern Mexico, in 
lower California, and on the intervening islands,, of many 
species of the large and ponderous genus Bulimulus. The 
difficulty of satisfactory identification unfortunately is such 
that there is little agreement among the various authorities as 
to the limits or range of species, nor is the nomenclature of 
these shells in a more satisfactory position. All the same 
as I shall mention later on, we can gather some useful hints 
from their distribution. 

* Pilsbry, H. A., and J. H. Ferris, " Mollusca of the South- Western 


The point, however, I wish to direct special attention to 
is the fauna as a whole inhabiting the southern part of the 
long peninsula of Lower California. This peninsula, for which 
many people retain the Spanish name " Baja California," is 
a narrow strip of broken mountainous land, nearly eight hun- 
dred miles long, and averaging about fifty miles in width. On 
the western side of this region there are a series of isolated 
peaks or ridges rising to a height of about 2,000 feet. An 
older range of mountains, composed of granite and gneiss, lies 
along the eastern flank of the peninsula. Most of this vast 
country is bare and desert-like, with a scanty flora, water 
being difficult to obtain.* As soon as we enter what is called 
the " Cape Region " all this is changed. Running water can. 
be procured throughout the whole year, in the larger canons 
at any rate, and this gives rise to an extremely rich and 
exuberant vegetation. While almoist the whole of the fauna 
and flora of the peninsula are merely a continuation of those 
of Calif ornia and Arizona, having evidently passed into it from 
the north and east, the southernmost tip, or Cape Region, has 
an entirely different set of animals and ; plants. I have already 
alluded to the occurrence here of Euchirotes biporus, a very 
peculiar burrowing lizard belonging to the ancient family of 
Amphisbaenidae, and of the boa Lichanura trivirgata. Another 
reptile, which is quite peculiar to the same district, is Cteno- 
saura hemilopha. This great lizard, of nearly two feet in 
length, possessing a high crest along its back, a green head 
and pale yellow body spotted with olive, brown and black, is a 
most remarkable and striking object. The only near relations 
of the burrowing lizard, and of the great active Ctenosaura, 
live in south-western Mexico. But the Mexican forms do not 
belong to the same species. We cannot, therefore, assume that 
these lizards have been accidentally carried across the Gulf 
of California, which has a width of about one hundred miles 
near the tip of the peninsula. The genus Ctenosaura, it 
may be mentioned, is related to Cyclura of the West Indies, 
and Brachylophus of the Fiji and Friendly Islands, these 
genera all belonging to the great family Iguanidae, which 
has inhabited America since Cretaceous times (p. 202). f 

* Merrill, Q-. P., "Lower Calif ornia." 

t Denburgh, J. van, " Herpetology of Lower California," p. 78. 


The flora of the Cape Region is sub-tropical, a considerable 
proportion of the plants being West Indian. There are also a 
few genera which do not occur elsewhere on the American 
continent or the neighbouring islands. A small tree abun- 
dant in the Cape district belongs to the genus Albizzia, which 
is otherwise confined to Australia and the warmer parts of 
Asia and Africa.* Although the lower Californian species is 
quite distinct, its occurrence there might be attributed by 
some naturalists to accidental distribution by marine currents 
during some former period. No other occasional means 
of transport could be thought of. But Albizzia is by no 
means an isolated instance of floristic relationship between 
the countries bordering the west and east sides of the Pacific 
Ocean. Difficult problems of distribution of that nature are 
apt to be looked upon as instances of accidental dispersal. 
Yet these puzzling cases of distribution often supply us with 
valuable clues with reference to possible changes of land and 
water that may have taken place. That the Cape Region of 
Lower California is really a fragment of an ancient land -mass 
is suggested by the occurrence there of the burrowing lizard 
Euchirotes, of two species of the fresh- water oligochaet worm 
Kerria, and by a good many other faunistic features. Kerria 
is only met with in that region, in the West Indies and in 
southern South America. 

Among the most interesting members of the Cape fauna 
are the land shells of the genus Bulimulus above referred 
to. Dr. Cooper, f in his series of valuable papers, only 
mentions a few species found in that region, but he 
alludes to the noteworthy fact that two of the Bulimuli 
only live on the peninsula of Lower California, and in a similar 
situation on the coast of southern South America, though quite 
absent from the intervening moist tropical region. Dr. Dall,J 
and more recently Dr. Pilsbry, have shown, however, that the 
Cape species are not identical with the South American ones, 
though extremely like them in general appearance, and that 
they, together with those of southern Mexico and the interven- 
ing islands, form a group by themselves. About twenty species 

* Brandegee, T. S., "Flora of Baja California," p. 222. 
t Cooper, J. Gr., " Molluscs of Lower California," p. 99. 
t Ball, W. H., " Bulimulus in Lower California." 


of this genus of large snails are peculiar to the Cape Region, 
several others being identical with Mexican ones. Some of 
them have spread along the peninsula northward, yet it 
seems certain that the Lower Californian centre of dispersal 
lies in the Cape district, for no species occurs in California 
proper. If we examine the range of the genus Bulimulus as a 
whole, we find that its headquarters are in the West Indies. 
From there it has travelled to eastern South America possibly 
across a very ancient land surface. Another branch has gone 
westward and populated the distant Galapagos islands. A 
third stream apparently invaded the coast lands of South 
America from the west, for we find numerous groups of Buli- 
muli in certain western areas in Bolivia, Peru and Chile. 
Still another section has struck directly westward across 
southern Mexico to the southern parts of Lower California. 
A couple of species have passed into Texas and further 
east from this Mexican centre. 

Considering that the closely related genus Placostylus 
inhabits New Zealand, as well as a series of archipelagoes 
between it and the Solomon Islands, it must be admitted that 
in Bulimulus we have to deal with an almost archaic genus, 
and that the land-masses and islands on which it is now found 
are probably fragments of ancient continuous lands. Acci- 
dental dispersal of shells has formed the subject of Mr. 
Medley's * special study. Yet he shows that the species, par- 
ticularly of the southern Placostylus, being heavy massive 
shells, are singularly unfitted for crossing distant seas by 
occasional means of dispersal. Hence he arrives at the con- 
clusion that the Solomon and Fiji islands, the New Hebrides, 
Loyalty Islands, New Caledonia, Norfolk Island, Lord Howe 
Island and New Zealand form a zoological province, and are 
to be considered as the fragments of a shattered continent. 

A similar argument might be applied to Bulimulus, which is 
so closely related to Placostylus. It seems permissible to 
argue that the West Indies in early Mesozoic times were a 
large united land-mass, that the latter was continued south- 
eastward so as to join the archaic lands of eastern Brazil, that 
it swept westward across what is now Central America out 
* Hedley, 0., " Placostylus," A Study in Ancient Geography, pp. 337 

L.A. P 


into the Pacific Ocean, and that wherever in Bolivia, Peru and 
Chile we find groups of Bulimuli, the land on which they live 
represents some portion or fragment of that ancient land 
which once occupied part of the eastern Pacific Ocean. If we 
had no other data in support of such theories, they would be 
based on very slender evidence. But all this will be more 
amply discussed when we come to deal with the general fauna 
of the Galapagos Islands and western South America. 

The only other point bearing on the westward extension of 
the ancient "Antillea," as the supposed West Indian continent 
has been called, is the singular occurrence here and there 
of traces indicating land westward of California. Along the 
coast of southern California lies a little cluster of eight islands 
which have lately attracted the attention of naturalists. 
Now, as Mr. Greene informs us, on Santa Catalina Island 
may be found Crossosoma, a plant so peculiar as almost to 
represent a distinct natural order, and more closely related 
to the Asiatic and Australian Dilleniaceae than to any plant 
on the American Continent, except a single congener recently 
discovered in south-eastern California. Speaking of Santa 
Cruz, another of this group of small islands, he comments on 
its unique feature in possessing no less than forty-eight 
species of plants not found on the opposite mainland and 
expresses the opinion that the whole set of islands must have 
a very peculiar geological origin and history. Species belong- 
ing to distinctly Calif ornian genera, like Dendromecon, Esch- 
scholtzia, Thysanocarpus and Zauschneria, abound in such 
numbers as to suggest the question whether it was no\ from 
these islands that the mainland plants of the same genera 
wer'e derived. The remarkable fact that Lyonothamnus, 
Hazardia and the Lavateras, all of which have Asiatic or 
Australian affinities, occur on these islands in several distinct 
species, being quite unknown on the American continent, 
makes Mr. Greene* wonder whether this group of islands 
has not formerly been connected by land with some other 
continent than America. 

Surprising and apparently inexplicable as this feature may 
seem, California and south-western North America generally 

* Greene, E. L., '< Botany of Santa Cruz Island," pp. 377388. 


reveal a still more extraordinary phenomenon in distribution. 
We may imagine, and many people actually do, that the power 
of wind and wave's to carry objects from one part of the world 
to another is almost unlimited. One of the most striking 
and remarkable characters of California, however, is its very 
close faunistic affinity with western Europe. Even the most 
enthusiastic supporter of the flotsam- jetsam theory will feel 
that some other cause than this must have been responsible in 
producing this effect. A former land connection between the 
two regions, without others being affected, seems out of the 
question. Yet almost all- those who have endeavoured to 
explain the origin of this western fauna have preferred 
to choose the old Bering Strait land connection as offer- 
ing a safe passage to European animals, rather than 
disturbing the general arrangement of the existing oceans 
and continents. If the faunistic resemblance of Cali- 
fornia to western Europe had really been caused by a 
migration of animals from one area to the other across the 
whole Asiatic continent, eastern Asia ought surely to show 
affinity with California to a much more pronounced degree 
than western Europe does. As a matter of fact, certain 
groups in California are distinctly eastern Asiatic in affinity, 
a f s I have just mentioned, while others are just as clearly; 
isouth and west European in character. I have given a few 
instances already of these faunistic relationships, and further 
evidence will now be adduced in support of this statement. 

I think it was the snail Arianta arbustorum, so prominently 
alluded to in my work on European animals, which first 
drew the attention of American zoologists to this relation- 
ship, for, as already remarked, a snail extremely similar in 
appearance lives in California. Even Dr. Pilsbry * admits 
that the resemblance in shell characters of the Calif ornian and 
European species is astonishing, although he adds that it 
is due to a purely secondary modification that these shells have 
been moulded to a deceptive likeness, the genitalia having 
been left unchanged to tell more faithfully the story of their, 
lineage. Having made an anatomical study of this Californian 
Arianta -like group, Dr. Pilsbry arrives at the conclusion that 

* Pilsbry, H. A., " Manual of Conchology," IX., p. 196. + 

P2 ''' 


the latter belongs to the same great family Helicidae as the 
European Arianta, but that otherwise it is quite unrelated 
to it. He places the Californian forms, therefore, into the 
genus Epiphragmophora, contending that its nearest rela- 
tions are the Helices of Japan. In another place (p. 46) he 
adds the remark that it is unnecessary to throw land bridges 
across the depths of the Atlantic and Pacific Oceans to account 
for the distribution of Helices. Such hypotheses, he thinks, 
are contrary to many facts indicating that such groups of 
snails as are common to America and Europe have radiated 
from an Oriental centre westward to Europe and eastward, 
by way of a former Bering Strait land bridge, to America. 
Yet Dr. Pilsbry's conclusions, as I have already mentioned, 
are contested by Dr. von Ihering likewise on anatomical 
grounds. Although Dr. Pilsbry maintains that his opponent 
bases his deductions on "figures and not dissections (p. 195), 
Dr. von Ihering, in a recently published paper, again insists 
that, after having made a careful anatomical investigation of 
Arianta, (or Helicigona as he calls it,) and the American 
Epiphragmophora, he could perceive no difference worth men- 
tioning between the two. Hence Dr. von Ihering's* opinion 
is that the American Helices, which are now generally known 
under the name of Epiphragmophora, and which are entirely 
confined to the Pacific coast of America, possess their nearest 
relations not in Asia but in western Europe. 

Let us take another group, that of the well-known Euro- 
pean family of slugs, the Arionidae. In 1896 Messrs. Pilsbry 
and Vanatta f showed by anatomical investigations that the 
American slugs Ariolimax and Aphallarion belong to this 
family. Later on the same writers added the genera 
Anadenulus, Hemphillia, Hesperarion and Prophysaon to 
this list. The whole of this great assembly of Arionidae 
is quite confined to the Pacific region between British 
Columbia and southern California. No other slug of this 
family has as yet been discovered anywhere in the New 
World, except one or two European species in the north- 
eastern States, which may either have been introduced 

* Ihering, H. von, " System der Heliciden," p. 422. 
t Pilsbry, H. A., and E. G. Vanatta, "Revision of North American 


or have crossed over by a former North Atlantic land bridge. 
In western Europe and north Africa we have the four genera 
Arion, Geomalacus, Ariunculus and Letourneuxia. Only the 
first of these has a wide range, one species extending through 
Russia to northern Siberia. In Asia only a single genus 
(Anadenus) of the family is known, inhabiting the Hima- 
layan Mountains and China. The Arionidae thus have a most 
discontinuous range, and their origin and dispersal form an 
interesting problem. Since Geomalacus and Letourneuxia 
are confined to the western borders of Europe and North 
Africa, and Arion obviously has its headquarters in the same 
region, Professor Simroth * argued that the European Ario- 
nidae had either originated on a sunken land which lay out 
in the Atlantic, or wandered across an ancient Atlantis from 
western America. Dr. Pilsbry f also is in favour of an 
American genesis for the Arionidae, but he believes that 
the ancestors of the Old World genera just alluded to 
must have crossed over to Asia by means of a former 
Alaskan land bridge, and then have wandered along, dropping 
Anadenus on the way, until they finally reached western 
Europe. Of the two theories Professor Simroth's appears 
to me the more plausible one. And his hypothesis is 
strengthened by the occurrence in south-western North 
America of a species of slug either identical or very closely 
related to a European one, viz., Amalia hewstoni. The Euro- 
pean Amalia (Milax) gagates must be regarded as an exceed- 
ingly ancient species, certainly dating back to beyond Tertiary 
times. Geological evidence for such a belief there is none, 
nor could we expect to find the remains of slugs in ancient 
deposits. In a matter of that kind we have to rely purely on 
distributional evidence. That Amalia gagates could be trans- 
ported to any island from the mainland by the ordinary means 
of dispersal is out of the question. It can only have 
reached its present habitat on many very remote islands by 
human introduction or by former land connections with the 
mainland. It is by no means a slug that congregates near 
human habitations, nor is it commonly found among vegeta- 
bles like some of the Limaces and Arions. Yet it occurs in 

* Simroth, H., " Nacktschnecken Russlands," p. 60. 
t Pilsbry, H. A., " Phylogeny of Arionidae," p. 103. 


variously modified forms on the islands of Madeira, Sicily, 
Sardinia and Teneriffe, on the Azores, the Cape Verde islands, 
St. Helena, Tristan d'Acunha and Bermuda. It has never 
been reported from any of the West Indian islands, Central 
America or the southern States of North America. In western 
North America it has a wide range, being known as Amalia 
hewstoni, from British Columbia to Mexico. On the east 
coast it occurs only along the coast of Massachusetts and New 
York, exactly where so many other members of an ancient 
fauna reside. In Asia, except Asia Minor, no Amalia has 
hitherto been discovered. In Africa it is only met with along 
the north coast and in Cape Colony (A. ponsonbyi). From 
south Australia a similar form has been described as Amalia 
pectinata, while others occur in New Zealand and the Sand- 
wich islands. Most zoologists will insist that such an enor- 
mously wide and discontinuous range of a species, though 
exhibiting distinctive characters in its various habitats, can 
only be due to artificial introduction by man. My own opinion 
is that it has reached all the localities referred to by the 
natural means of progression on land during a long series of 
geological periods, and Professor Simroth,* the best living 
authority on slugs, shares the same view. As far as its occur- 
rence in western North America is concerned, Professor Sim- 
roth's Atlantis, a land bridge connecting Portugal with the 
West Indies and the lands beyond it across the Atlantic, 
would suit our purpose. The slug's absence from the Antilles 
might be due to the fact that it became extinct there during 
the extensive submergence which the West Indian area has 
undergone in Tertiary .times. 

The eminent authority on butterflies and moths, Dr. Pagen- 
stecher,f tells us that California possesses an independent 
character among the North American fauna, in being more 
nearly related to Europe than to the eastern States. With 
the exception of Ctenucha and Agarista all the genera are 
European, while Pyrameis atalanta, P. cardui, Vanessa 
antiopa, Arctia ca'ja, Phragmatobia fuliginosa, Scoliopteryx 
libatrix, Brachionycha nubeculosa, Amphipyra pyramidea and 
Agrotis exclamationis are common to California and Europe. 

* Simroth, EL, " Pendulations Theorie," p. 92. 

t Pagenstecher, A., " Verbreitung d. Schmetterlinge," p. 359. 


Among the most abundant beetles in California there 
is a group of flightless forms which when touched will emit 
a pungent, oily secretion. Owing to their curious habit of 
elevating their bodies when alarmed they have received in the 
west the popular name of " circus bugs." Now this section 
Eleodinae of the large family Tenebrionidae are, in North 
America, quite peculiar to the south-west. A few species have 
spread northward into Washington State and eastward as far 
as the Mississippi. The great mass of these thick-bodied oily 
creatures lives in Mexico and California, where, according to 
Mr. Blaisdell,* they have probably originated. The only near 
relation of this south-western group in the New World is the 
genus Nycterinus, which is peculiar to Chile. But throughout 
the Mediterranean Region we meet with a very large number 
of similar beetles, commonly called " oil beetles," and 
although many of them are also found in Persia, Turkestan, 
Central Asia and even China, south-eastern Europe must be 
looked upon as the headquarters of the genus Blaps, to which 
they belong. 

The same family Tenebrionidae also offers instances of 
intimate relationship between the Antilles or Central America 
and southern Europe. I need only recall the American 
Gnathocerus maxillosus which likewise inhabits Madeira, the 
Canary and Mediterranean islands, as well as Sitophagus 
hololeptoides of Central and South America, the Antilles and 

Dr. Kolbe f mentions similar examples of distribution 
among the coprophagous lamellicorn beetles. Thus Oniti- 
cellus and Grlaresis are quite peculiar to the south-west in 
North America, whereas in Europe they are confined to the 
Mediterranean Region. Only the former has spread further 
southward into Africa and eastward into Asia. 

Of all the insects, faunistically the most interesting are 
the ants. In my work on European animals I alluded to 
Stenamma westwoodi as one of the members of the Lusitanian 
fauna, which penetrated as far north as south-western Ire- 
land. The same species is met with in America in a few 
varieties. One lives in California, the other in British 

* Blaisdell, F. E., "Eleodinae of the United States," pp. 2829. 

t Kolbe, H. J., " Verbreitung d. Coprophagen Lamellicornier," p. 499. 


Columbia, the third in the north-east.* Honey-ants, at least 
certain members of their colonies, possess the faculty of 
storing quantities of honey within their bodies, which swell 
up to a great size, and on that account are somewhat remark- 
able objects. In North America two species of honey-ants 
occur, namely, Myrmeocystus melliger and M. mexicanus, 
both being confined to Mexico, Arizona and the neighbouring 
regions. The other species of the genus Myrmeocystus in- 
habit the Mediterranean Region and further east as far as 
central Asia.f In southern Europe a soft, velvety ant is 
found belonging to the genus Liometopum. The same species 
was likewise supposed to inhabit California, but Professor 
Wheeler J found that although the two forms are very 
closely related, the American differs slightly in shape and 
even in its habits from the European species. Whereas the 
former constructs its nests underground, the latter utilises 
for that purpose the abandoned burrows of beetles under the 
bark of trees. Hence he calls the American ant, Liometopum 
apiculatum. It is limited in its range to California and 
Mexico. Only two other species of Liometopum are known. 
One, (L. microcephalum), as I mentioned, lives in southern 
Europe, the other (L. lindgreeni) in Assam. It is not often 
that we actually become acquainted with the extinct ancestors 
of such interesting invertebrates as these. But fossil forms 
of Liometopum have been discovered, according to Dr. Hand- 
lirsch in the Lower Miocene of Croatia and the Oligocene of 
Colorado. Even in Tertiary times the genus seems to have 
occupied much the same localities as it does to-day ; it only 
spread formerly further north no doubt owing to the more 
favourable climatic conditions then prevailing. 

Many other similar instances might be quoted. I will only 
allude to one more. The two well-known European ants 
Formica cinerea and Formica rufibarbis had been reported 
from the south-western States of North America, but Pro- 
fessor Emery doubted the correctness of the identification 
until Professor Wheeler sent him American specimens. Both 

* Emery, C., " Nordamerikanische Ameisenfauna," p. 299. 
t Wheeler, W. M., "Honey Ants," p. 347. 
J Wheeler, W. M., " North American Liometopum," p. 321. 
Handlirsch, A., " Die Fossilen Insekten," p. 870. 


species must have been established in America for long periods 
past, as their range extends inland in a north-eastward direc- 
tion beyond the Rocky Mountains. 

Instances of specific identity in the two widely separated 
regions of California and southern Europe are to be found 
in .several groups. Professor Kraepelin * reports, for example, 
that the scolopendrid Theatops erythrocephalus is such a 
case. From California it has spread as far as Oregon. 
In southern Europe it is known from Portugal, Italy, 
Dalmatia and Hungary. The genus Theak>ps is confined to 
North America, including the Sandwich islands, and southern 

A member of the order Palpigradi, a minute creature 
somewhat resembling a scorpion in shape, has been observed 
in Texas and named Koenenia wheeleri after Professor 
Wheeler. Two other members of the genus are known from 
South America, one from Chile, the other from Paraguay. 
But the nearest relations of the North American form ar,e evi- 
dently Koenenia mirabilis of Sicily and Tunis, and K. draco of 
the Balearic islands. The only two other species live in Siam.f 
We thus have in this ancient group again the same intimate 
affinity between southern Europe and south-western North 
America as in some ants, beetles and butterflies, showing 
clearly that the same potent cause, which is certainly not acci- 
dental distribution, has contributed to bring it about. In 
order, however, to make quite sure that these are genuine 
instances of migration on a land surface, and not due to 
occasional or accidental transport, let us now examine care- 
fully the range of a large fresh-water form and endeavour to 
trace its origin. 

Fresh-water crayfishes, as Dr. Ortmann remarks in his 

excellent account of them, do not possess any exceptional 

means of dispersal. They are restricted to fresh water and 

cannot exist out of it, neither in salt water nor on land. 

Moreover, they do not possess, during any stage of their life 

history, means or devices which might favour their passive 

transport from one fresh -water system to another. The whole 

character of their range is opposed to the assumption that 

* Kraepelin, K., " Revision der Scolopendriden," pp. 6466. 

t Hansen, H. J., "On Koenenia." 


their dispersal has been caused by any exceptional or abnormal 
means of transport. Few zoologists will consequently dis- 
agree with Dr. Ortmann's * assertion that fresh-water cray- 
fishes are among the most important animals in so far as the 
study of their distribution elucidates past changes of land 
and water over the globe. The crayfishes have been brought 
into great prominence by Professor Huxley's well-known 
treatise on the subject. More recently it is principally in 
America that their structure and distribution have been 
studied with great assiduity. Dr. Faxon f was the first to 
recognise that besides the American genera Potamobius 
(Astacus) and Cambarus, there is a third genus of fresh- 
water crayfishes which inhabits north -eastern Asia. The 
latter, it is true, is only 'considered a sub-genus of Pota- 
mobius by Dr. Faxon and also by Dr. Ortmann, but, as 
Mr. Stebbing J has pointed out, its intermediate position 
between Potamobius and Cambarus entitles it tp rank as the 
distinct genus Cambaroides. 

The geographical distribution of these crayfishes (Pota- 
mobiidae) is very suggestive and interesting. Europe is the 
headquarters of the old and well-known genus Astacus, which 
name, in the unfortunate search for priority, has had to 
give way to Potamobius. The genus ranges practically 
throughout Europe, from north to south and from east to 
west, and only very little beyond it. Beyond the Caucasus 
it crosses into Transcaucasia, Turkestan and western Siberia. 
It is quite unkno.wn in the remainder of Asia. In the rivers 
of eastern Asia, in Korea, Japan and eastern Siberia we meet 
only with members of the small group Cambaroides referred 
to. The somewhat close resemblance of this Asiatic genus 
to the American Cambarus does not point to blood relation- 
ship, according to Dr. Ortmann, merely to convergence. In 
America .we find not only Cambarus, but also Potamobius, the 
European crayfish, the latter genus being in America entirely 
confined to the western States. Professor Huxley and Dr. 
Faxon both urged that the American species of Potamobius re- 
sembled the European crayfishes much more than the Asiatic 

* Ortmann, A. E., " Distribution of Freshwater Decapods," pp. 315316. 
f Faxon, W., "Revision of Astacidae." 
J Stebbing, T. E. E., " Crustacea," p. 208. 


ones, yet, like Dr. Ortmann, they maintained that Europe was 
supplied with its crayfish fauna from the East. In his lucid 
essay on this subject, Dr. Ortmann argues that a primitive 
group of Potamobiidae, ancestral to all the living ones, must 
formerly have existed in eastern Asia, which region should 
be regarded as the centre of dispersal of the family. This 
ancient group, he thinks, sent one branch westward to Europe 
and another eastward across the old Bering Strait land bridge 
to western North America. Thus, three centres of dispersal 
gradually originated 1 in which the old stock developed on 
independent lines. The middle one changed to Cambaroides, 
jvhile the two branches retained the ancient characters. From 
the American branch eventually originated Cambarus, which 
spread eastward into the eastern States of North America 
(see pp. 28991). 

I accepted Dr. Ortmann 's explanation in my work on Euro- 
pean animals as an hypothesis, which satisfactorily accounted 
for the present distribution of the Potamobiidae. Doubts, 
however, have since arisen in my mind as to whether there 
is not a better theory. The more I studied the problem the 
less I felt disposed to agree with Dr. Ortmann's explanation. 
Why should the old stock, for instance, have become modified 
into Cambaroides in its original centre of dispersal, while still 
flourishing in two centres enormously distant from one 
another ? And these two new centres were reached after many 
struggles and vicissitudes, after long and weary travels, prob- 
ably through hundreds of miles of unsuitable ground. One 
would imagine the two distant branches to have become 
more and more unlike one another. Five species of the 
old .stock Potamobius still inhabit the streams of western 
America, from California in the south to Alaska in the north. 
If Dr. Ortmann's theory were the correct one, the centre of 
dispersal of the more modern genus Cambarus, which has 
developed from some member of the old stock, ought to be in 
north-western America. Everything, nevertheless, points to 
the conclusion that the new genus Cambarus originated in 
Mexico, and Dr. Ortmann (p. 291) is of that opinion, having 
recently supported it by means of many additional facts of 

* Ortmann, A. E., "Affinities of Cambarus." 


It is to south-western North America, therefore, that we 
must look for the original home of the ancestral group of 
Potamobius. They still inhabit that area, and may have 
ispread from there northward to Alaska, and even further 
to north-eastern Asia, eventually giving rise to Cambaroides. 
Taking these and many other remarkable facts of dis- 
tribution into consideration, it appears to me quite pos- 
sible that the presence of the crayfish Potamobius in 
Europe and North America, and its occurrence in the 
western parts only of the latter continent, may be due to 
an ancient land connection which, as already suggested, 
joined western Europe and Mexico by way of the West Indies. 
Whether the family originated in North America or in Europe 
will have to form the subject for future researches. That this 
migration took place in very remote times, is implied by the 
fact that Cambarus primaevus (which Dr. Faxon believes to 
be a Potamobius), occurs in the Eocene beds of western 
Wyoming. If such a land bridge as that alluded to actually 
existed in early Tertiary or late Mesozoic times, it may be 
asked why do we not meet with any members of the genus 
Potamobius in the streams of the West Indian islands ? To 
this we may answer that geologists are practically agreed 
that in post- Eocene, or even during Eocene times, the whole 
area of the Antilles was greatly submerged, so that we may 
suppose that the ancient fauna that wandered across that area 
from either Europe or North America was largely extermi- 
nated. That the islands were subsequently again connected 
with the mainland we may assume from the presence of 
Cambarus cubensis, a crayfish peculiar to the island of Cuba. 

My views as to the nature and extent of that mid-Atlantic 
land bridge will be more fully explained in the chapter dealing 
with the West Indies. The presence or absence of such a land 
connection, however, is of such vital importance to the eluci- 
dation of the phenomena of distribution, that I may he ex- 
cused for quoting still further examples of animals whose 
range throws light on the solution of this problem. 

In the beginning of this chapter (p. 205) I mentioned the 
fact that the spade-foot toads (Scaphiopus) have their head- 
quarters in Mexico and the south-western States of North 
America, and that their nearest relations are the members 


of the genus Pelobates which inhabit Europe. Only a 
single species (P. syriacus) has been discovered in Asia 
Minor. The others extend from western Europe to the 
Caspian. Another genus (Pelodytes) of the family Pelo- 
batidae has one species in Portugal and a second in the 
Caucasus. All the other genera of the family are found in 
the East, and are more remotely akin to Scaphiopus than 
Pelobates is. Hence we may conclude that the distribution 
of these two genera is distinctly favourable to the suggested 
trans -Atlantic land bridge. 

Another interesting amphibian that I have alluded' to 
(p. 137) is the newt Spelerpes, which, with the exception of 
a single European species, is peculiar to America (see Fig. 8). 
The centre of dispersal certainly lies in Mexico, from which 
country various sections have spread northward into the 
States, southward as far as Peru, and eastward to the island of 
Haiti. That this discontinuous distribution could not have 
been brought about under existing geographical conditions is 
evident, nevertheless, since no fossil Spelerpes are known, we 
can only judge of the age of the genus from its distribution. 
Dr. Gadow * suggests the Oligocene Period, and thinks that a 
north -Atlantic land connection, such as the one I have de- 
scribed in the first chapter of this work, from Labrador to 
Scotland via Greenland, might have brought about the exist- 
ing range of Spelerpes. Since the single European species 
inhabits only Sardinia and 1 the mountains bordering the Gulf 
of Genoa, while most suitable ground for its existence is found 
further north, I cannot admit that Spelerpes fuscus reached 
Europe in that manner. The land bridge by which it crossed 
the Atlantic must have lain much further south. 

I have already alluded to the fact (p. 173), that there 
is apparently a south-western and a south-eastern form 
of the American glass snake (Ophisaurus ventralis), and 
that both of them extend northward for considerable dis- 
tances. The only other members of the genus are the 
European glass snake (Ophisaurus apus), which inhabits the 
Mediterranean region and is very like the American, and the 
Asiatic species, which is found from the eastern Himalayas 
to Burma. 

* Gadow, H., " Mexican Amphibians and Eeptiles," p. 2-44. 


Dr. Buthven's valuable researches into the origin and dis- 
tribution of the garter-snakes (Thamnophis = Eutenia) have 
been discussed in an earlier chapter (p. 128). I need only 
restate that he traces the home of the genus to northern 
Mexico. And Thamnophis is clearly an of shoot from the 
older water-snakes (Tropidonotus), which have almost a 
world -wide range. As in the case of Potamobius and 
Cambarus, both genera seem to have spread northward 
from their south-western centre, the south-western Tropi- 
donotus validus having its nearest relation in the Sar- 
dinian Tropidonotus viperinus. Dr. Brown * recognised per- 
fectly that the affinity between such forms as the European 
and American species of Tropidonotus necessitated the exist- 
ence of a former land bridge between the two continents. 
He also urged that the existence of this bridge must have 
coincided with a warm climate in the north, for he naturally 
assumed that only in the extreme north could there have been 
such a land connection. Its geological age he fixes at about 
the early Miocene, though he believes many of the present 
genera to have been in existence even in Eocene times. 
My objection to Dr. Brown's theory is that we Tiave no evi- 
dence in Europe of a southward advance of Tropidonotus 
from a former northern centre of distribution, nor are the 
northern species in both continents more closely related to 
one another than the southern species. The former existence 
of a more southern trans -Atlantic land bridge in early 
Tertiary times, on the other hand, is supported, by such a 
number of palaeontological facts, as we shall learn later on, 
that the evidence is overwhelmingly in its favour. 

It is quite possible that the western tortoise Clemmys 
marmorata, or its ancestors, for the genus, as I mentioned 
(p. 133), has inhabited south-western North America since 
Eocene times, has spread across the same mid- Atlantic land 
bridge to western Europe, a near relation (Clemmys leprosa) 
being peculiar to Portugal, Spain and north-western Africa. 

That all these animals living in south-western North 
America and western Europe which show close relationship, 
are relicts of very remote geological times is rendered prob- 

* Brown, A. E., " Post-Glacial Nearctic Centres," p. 466. 


able, partly from palaeontological evidence, and partly from 
other methods of reasoning. That most of them trace their 
origin in America or Europe to the former existence of a 
direct land bridge across the mid-Atlantic seems also obvious. 
Yet many geologists are very strongly opposed to a theo- 
retical bridging of the Atlantic. Nothing short of a well- 
marked mammalian affinity between the two areas alluded to 
will satisfy them. We can produce little of such evidence, be- 
cause most groups of mammals have changed very rapidly 
during the course of the Tertiary Era. A few, howev;er, 
such as the rodents, appear to have the faculty of preserving 
their ancestral characters for longer periods, and some, 
apparently, have undergone little change since remote times. 

There are two instances, one among the voles, the other 
among the hare family, that seem to point to the existence of 
the land connection just discussed, and these cases may pos- 
sibly throw more light on the age of the land bridge than the 
invertebrates or reptiles can do. Professor Tullberg * con- 
tends that the meadow voles (Arvicola Microtus) only entered 
Europe in Pliocene times from Asia, where they had already 
existed for some time previously. He also expressed the 
opinion that these voles subsequently crossed over to America 
from Europe by a north- Atlantic land connection, which I 
presume must be the Scotland-Greenland-Labrador bridge. 
We know very little of fossil meadow voles. If Professor 
Sehlosser is correct in his assertion that only geological re- 
searches can give us any clues as to former changes of 
land and water, and that zoogeography cannot do so, the 
meadow voles can teach us very little. 

Yet if we examine the present range of one of the sub- 
genera of meadow voles, such as Pitymys, we find it very 
remarkable and instructive. One species lives on the Mont 
d'Or in France, at a height of over 4,000 feet, another on one 
of the southern Alpine spurs, at a height of 6,000 feet. De 
Selys long ago described a species (Pitymys incertus) from 
the St. Gothard mountain, a locality which is over 7,000 
feet high ; another related form inhabits the Pic du Midi in 
the Pyrenees, and still another a mountain in Sicily. Quite 

* Tullberg, T., " System der Nagetiere," p. 499. 


a number of species are now known from southern Spain and 
Portugal, owing to the researches of Mr. Thomas, Mr. Miller 
and Dr. Forsyth Major. To judge from its recent distribu- 
tion, these voles of the sub-genus or genus Pitymys ought 
to be of very great antiquity, yet not a single fossil specimen 
has ever been found. The most remarkable fact in the dis- 
tribution of Pitymys is that it is entirely confined to Europe 
in the Old World, whereas across the Atlantic, in Mexico, 
we again meet with a member of this group known as the 
Jalapa meadow vole (Pitymys quasitor). It lives there at an 
altitude of about 5,000 feet. Only two other species are 
known from North America, one (P. nemoralis) from the 
Boston Mountains in the Indian Territory, the other from 
certain areas in the eastern States. The latter (P. pine- 
torum) occurs from southern Florida to Carolina, a variety 
of it on the Allegheny Mountains, and another from Long 
Island to the borders of Illinois. The range of the three 
American species is disconnected, and confined to Mexico and 
the United States. What is the relationship of these species 
to one another, and which is the oldest, will have to be deter- 
mined by future researches, also whether the extinct species 
discovered in Pennsylvania by Professor Cope really belongs 
to Pitymys or Microtus proper. At any rate, there is nothing 
in the range of Pitymys that might lead us to suspect that it 
entered North America from the north-west, no member of 
the group having as yet been found in any part of Canada or 
Alaska. In Europe Pitymys is unknown in the north-west, 
whereas a number of species inhabit the south-west. Henoe 
the American group of Pitymys may possibly have been de- 
rived from one or more species which crossed the Atlantic 
on the land connection above referred to. 

The hare family (Leporidae), as a whole, has a very wide 
distribution in Europe, Asia, Africa and America, but some 
of the sub-genera, which are gradually being raised to the 
higher dignity of genera, are confined within certain circum- 
scribed limits. In his study on the recent and fossil Lago- 
morpha, Dr. Forsyth Major * comes to the conclusion that this 
family might conveniently be divided, according to the osteo- 

* Major, Forsyth, " Fossil and recent Lagomorpha," pp. 514 515. 


logical characters of its members, into two groups, viz., the 
Caprolagus and Lepus groups. The first of these, which is of 
particular interest, contains the four genera Caprolagus, Neso- 
lagus, Oryctolagus and Sylvilagus. Caprolagus lives in Asia, 
and is likewise represented in European Pliocene beds. The 
second is peculiar to the island of Sumatra. The third, which 
includes the European rabbit, has an extremely discontinuous 
range, being confined, according to Dr. Major, to the Mediter- 
ranean Region, Western Europe and South Africa. Sylvi- 
lagus is peculiar to America. From its very scattered range 
the Caprolagus group of hares and rabbits would seem to 
be of ancient lineage. In a more recent study of the hares and 
their allies, Dr. Lyon recognised nine genera of this group. 
He separates the South African form of rabbit from the 
European, raises several of Dr. Forsyth Major's sub-genera 
to 'the rank of genera, and adds the new genus Pentalagus 
from the Liu Kiu islands near Japan. Whereas Dr. Forsyth 
Major recognised four sub-genera of the American Sylvi- 
lagus, Dr. Lyon places only two sub-genera under this genus. 
Although Dr. Lyon gives a fuller and more complete 
account of all parts of the skeleton in the different species of 
Leporidae than Dr. Forsyth Major. He was moreover able to 
detect certain prominent structural characters, which appar- 
ently escaped Dr. Major's notice. While the latter believed 
in the close relationship of the European and the Cape rabbits, 
Dr. Lyon * demonstrates that the nearest relation of the 
former is the American Sylvilagus. The skulls of the two 
genera, taken as a whole, are not markedly different, and the 
teeth are essentially alike. 

The European rabbit also possesses marked affinities with 
Limnolagus, another American genus of rabbits, but not with 
any of the Old World genera of the family. The genus 
Oryctolagus, with its single species the European rabbit 
(0. cuniculus) is confined to the countries round the Medi- 
terranean. Only in France, that is to say, in western Europe, 
is there a tendency to a northward extension of this range, 
and in that it resembles a great many other typically Medi- 
terranean animals. Its relation, Sylvilagus, ranges from the 

* Lyon, M. W., " The Hares and their Allies," p. 406. 
L.A. Q 


northern border of the United States, all over the States 
and Central America. In South America its exact distribu- 
tion is still unknown, but it possibly reaches Patagonia. The 
cotton-tails and brush rabbits (Limnolagus) occur only in 
the southern States. We thus note the remarkable fact 
that rabbits of close relationship are separated in one 
direction by a comparatively short intervening space of sea 
water, in the other by a very much longer area of almost 
uninterrupted land, which is tenanted almost exclusively by 
the more distantly related hares. The other American 
rabbits, the pigmy rabbit (Brachylagus), and the Popocatepetl 
rabbit (Romerolagus), are confined to Mexico and western 
North America. The south-western region must, therefore, 
be looked upon as the centre of dispersal of the rabbits.* 

The fossil history of the American Leporidae is meagre in 
the .extreme. Only a few Oligocene species of the extinct 
Palaeolagus are known, and these, according to Dr. Major, 
seem to be ancestral to the modern genus Lepus. No fore- 
runner of the existing Sylvilagus and Oryctolagus has yet been 
found. It possibly lived in south-western North America 
in early Tertiary times. Palaeolagus already possesses in- 
cisors of the modern type, and Dr. Matthew thinks that we 
may look among Eocene rodents, or even in the fauna of 
Cretaceous deposits, for guidance as to the manner of evolu- 
tion of the teeth of the Lagomorpiha-f 

This leads us back once more to the general consideration 
of the American Tertiary deposits and the affinities of their 
fauna. These deposits, above all, ought to yield indications 
as to whether there was a direct land connection between 
south-western North America and western Europe across the 
mid-Atlantic, such as the one I advocated. 

We are confronted in America by two grand problems, 
says Professor Osborn,J one being the chronological correla- 
tion of the purely fresh -water horizons with one another, the 
other the chronological correlation of American horizons with 
Eurasiatic vertebrate horizons. When these are worked out, 
continues the same writer, we shall be able to establish a 

* Nelson, E. W., " Eabbits of North America." 

t Matthew, W. D., " A Horned Eodent from Colorado," p. 307. 

t Osborn, H. F., " Cenozoic Mammal Horizons," pp. 2930. 


complete and very accurate geological time scale for the entire 
Tertiary Era, and to speak with precision regarding the time 
of successive migrations. Much still remains to be done to 
establish the approximate ages of the various Tertiary de- 
posits in Europe and America before we can definitely say 
whether a certain group of mammals made its first appearance 
in Europe or in America. Possibly we may not reach such 
a state of knowledge for a long time to come. 

In New Mexico what are known as the " Puerco " and the 
" Torrejon " formations have yielded remains of archaic 
mammals, some of which are related to mammals occurring 
in late Mesozoic or early Tertiary deposits of Patagonia, others 
indicating affinity with those of a similar age in France. 
Now, as Professor Deperet tells us, it is inadmissible to argue 
that mammals so nearly akin to each other could have arisen 
independently in three distinct centres, in Patagonia, in 
south-western North America and in France. We can only 
explain these palaeontological affinities by migrations from 
one area to the other. Professor Deperet * believes that the 
cradle of these early placental mammals was in North 
America, and that they crossed over to Europe by utilising 
the " territories of the North Atlantic which had risen from 
the sea." This is a somewhat vague statement, yet it indi- 
cates clearly that the great French palaeontologist had in his 
mind the existence, in these remote geological times, of some 
kind of north Atlantic land bridge. He does not explain how 
the North American mammals reached Patagonia, but Pro- 
fessor Osborn f speaks distinctly of a, contemporary, that is 
to say, early Eocene or previous (Cretaceous) land connection 
between North America and South America, and he actually 
places it in the same position as it occupies at present (see 
Fig. 21). I propose to discuss it later on. At present we may 
confine ourselves strictly to the European affinities of North 

A momentous change occurs, according to Professor 
Osborn, in the succeeding Wasatch formation of New Mexico 
and Wyoming. The parallelism of similar stages in the 
archaic mammals of western Europe and south-western 

* Deperet, C., "Transformations of the Animal World," pp. 308309. 
f Osborn, H. F., " Cenozoin Mammal Horizons," p. 33. 



North America remains as before. The new feature is the 
sudden appearance of true carnivores, ungulates, primates 
and rodents. I have already referred to Eohippus 
(p. 147) as an early representative of the hoirke family. 
Another modem family which traces its origin back to 
these remote times is that of the tapirs, for the Eocene 
Systemodon has all the characters peculiar to the recent 
Tapiridae. Whether the early primates were lemurs or true 
monkeys is as yet undecided. The rodents all belong to the 
extinct Ischyromyidae, which Dr. Matthew * believes to 
have been arboreal creatures, somewhat resembling squirrels 
in shape, although more nearly related to the peculiar and 
typically west American sewellel (Aplodontia). 

This sudden and simultaneous appearance of modern 
families of mammals, along with several extinct ones, in 
western Europe and south-western North America is very 
striking, and has to be accounted for. To begin with, we 
have to determine the origin, or original centre of dispersal 
of this fauna. Professor Osborn feels certain that this fauna 
originated neither in Africa nor in South America. There 
remain, he thinks, four possible sources. They may have 
come from the Great Plains Region of North America, from 
the more northerly American Mountain Region, from the 
northerly Eurasiatic Region, or from the American-Asiatic 
land -mass. He is in favour of the last theory, namely, that 
of the intermediate or North American -Asiatic source of 
this fauna. Still he believes that the actual origin of this 
modernised fauna will not be determined until Eocene fossil 
mammal beds in the northern portions of America and Asia 
shall have been discovered. Such beds have not yet been met 
with, nor is there any reason to suppose that they will be. 
Have we any geological evidence for the supposition that 
there actually existed any such large and intimately connected 
northern land-mass at this stage of the geological history of 
the earth as Professor Osborn f implies ? No doubt it is 
generally assumed that Alaska and north-eastern Asia were 
joined by land in Eocene times, and Professor Schuchert,J 

* Matthew, W. D., " Osteology of Paramys," pp. 6469. 
t Osborn, H. F., " Cenozoic Mammal Horizons," pp. 35 36. 
J Schuchert, C., " Paleogeography of North America," Plan 96. 


in his palaeontological maps, unites America in this manner 
with Asia, and by way of Greenland with Europe. But Pro- 
fessor Suess contends that, although these affinities in the 
Eocene faunas of America and Europe imply united conti- 
nents in the north, this land connection was probably not in 
the vicinity of Bering Sea. On the contrary, he .rather favours 
a more direct land bridge between North America and 

During the latter stages of the Eocene Period, while the 
Wind River, Bridger and Uinta beds were being laid down, 
the descendants of the archaic and of the modernised 
mammals gradually evolved, and we may suppose that the 
archaic mammals finally succumbed in the struggle for exis- 
tence. At any rate, they slowly disappeared, and during 
the process of their elimination, the fauna of America assumed 
a more independent aspect, the affinities with Europe becom- 
ing less pronounced. This need not necessarily imply a cessa- 
tion of the intimate geographical relationship between the two 
continents. The growth of an impenetrable forest, like that 
in the interior of Brazil, the development of local desert 
conditions, or the existence of temporary volcanic distur- 
bances on the supposed trans-Atlantic land connection, would 
have been sufficient to faunistically isolate the two continents 
from one another. In the succeeding period, the Oligocene, 
the faunal resemblance of western North America and 
western Europe once more became conspicuous. The land 
area available for the development of mammalian life cer- 
tainly increased in America during early Tertiary times, while 
a corresponding decrease may have taken place on the trans - 
Atlantic land connection, thus bringing a renewed influx of 
strange forms to the New World. Professor Osborn f tells us 
that in the White River, John Day and other American Oligo- 
cene formations, sixteen new families of mammals made their 
appearance, most of them still existing, and that a very 
similar modernisation occurred in western Europe. Six new 
families appear, apparently simultaneously, in both areas. It 
is worthy of note that the opossu'm family (Didelphyidae) and 
the rhinoceroses (Rhinocerotidae) now make their first entry 

* Suess, E., " Antlitz der Erde," Vol. III. 2 , pp. 764765. 
f Osborn, H. F., " Cenozoic Mammal Horizons," pp. 58 59. 


into North America^ while the monkeys (Primates) disappear 
for ever from the continent. 

In his restoration of Oligocene conditions in North 
America, Professor Schuchert still depicts North America 
as being joined in the far north by wide land bridges with 
Asia and Europe, while practically submerging the whole of 
the West Indies. As we shall learn later on, an intimate 
relationship exists between the shallow water marine forms 
of early Tertiary European and Antillean deposits, and this 
has given rise to the suggestion that a land bridge must then 
have united Europe and the Antilles. If 'my view should be 
substantiated, that the resemblance in the Oligocene faunas of 
Europe and south-western North America is due to the exis- 
tence in Oligocene times of a mid-Atlantic land bridge, the 
West Indian area, of course, could not have been submerged 
at that time. 

After a short phase of independent evolution, during which 
the Oligocene deposits of western North America insensibly 
pass into Miocene ones, the succeeding Middle Miocene beds 
are characterised by the appearance of a large number of new 
forms, among which the elephants (Proboscidea) deserve 
special mention. Some of these new immigrants are, ap- 
parently, of African, others of Eurasiatic origin. The 
Miocene beds of Europe and of America are remarkable for 
the similarity of their fauna. The conclusion deduced from 
this fact by Professor Osborn * is that the North American 
middle Miocene formations contain animals which first appear 
in the lower Miocene of Europe, just as the American lower 
Miocene contains animals that first appear in the upper 
Oligocene of Europe. 

Now it is quite possible that while the faunistic inter- 
change between western Europe and western North America 
took place by means of one land connection during early 
Tertiary times, this land bridge was replaced later on by an 
entirely different one. Professor Deperet f had some such 
idea in his mind in expressing the view that the Miocene and 
Pliocene migrations from Europe to America probably 
arrived by way of Asia and the Bering Strait, while the earlier 

* Osborn, H. F., "Cenozoic Mammal Horizons," p. 76. 

t Dep6ret, C., "Transformations of the Animal World," p. 314. 


ones came direct from Europe. Professor Suess,* on the 
other hand, favours a north Atlantic land connection in high 
latitudes between Greenland, Iceland and Scotland, and he 
thinks its origin dates back to the beginning of the Miocene 
period. Considering that there is such unanimity in favour 
of some kind of land bridge between the Old World and the 
New at this time, it is strange that Professor Schuchert f 
should entirely isolate North America in his palaeogeographi- 
cal maps from the Old World, not only in Miocene, but even 
in Pliocene times. 

Although the Pliocene fauna, according to Professor 
Osborn (p. 81), is as yet only imperfectly and sparsely 
known, being characterised by hosts of southern invaders 
which now flood the continent, a direct land bridge between 
North America and the Eurasiatic continent must likewise 
have existed, unless we assume that the mighty elephant 
Tetrabelodon (Dibelodon) which is represented by several 
species in the New World, reached North America by way 
of South America. Whether the still existing more hardy 
stock of Asiatic immigrants came to North America towards 
the end of Pliocene, as I argued in a previous chapter 
(p. 97), or during Pleistocene times, is a question which had 
not hitherto, I think, been seriously debated. That their pre- 
sence in North America is due to a wide land bridge across 
Bering Strait (see Fig. 7) seems to me evident. 

In this very brief survey of the past faunas of NortH 
America I have endeavoured to show that an important centre 
of evolution and dispersal existed in western North America 
in the past, just as it exists to a lesser degree at the present 
day. The strong faunistic affinity between western North 
America and western Europe, which we can still trace at this 
moment among many of the invertebrates and lower verte- 
brates of the two areas, seems to be the outcome of an ancient 
direct geographical communion between these land-masses. 
Certain features in mammalian palaeontology appear to 
strengthen my views, which are primarily based on a study 
of the modern fauna. In the succeeding chapters I shall 
bring forward further evidence showing that North America 

* Suess, E., " Antlitz der Erde," Vol. III.2, p. 765. 
t Schuchert, C., " Paleogeography," PI. 98100. 


and Europe were joined in the south, while northern, Asia was 
in all likelihood disconnected from the former. The principal 
point which I think has been clearly demonstrated is that the 
south-western centre of dispersal has exerted a powerful in- 
fluence on the development of the living faun,a of North 
America. There still remains one other feature that I wish 
to allude to before concluding this chapter. 

I have mentioned above that, while eastern Mexico was 
either wholly or partially submerged by the sea during later 
Mesozoic and early Tertiary times, most of western Mexico 
and a portion of the Californian coast remained dry land, as 
far as we know, all through geological history. It ought on 
that account to be a preservation ground for all kinds of 
relicts of bygone ages. And so it is. I have cited a number 
of them already, although the country can scarcely be con- 
sidered as being exhaustively explored. The fresh-water 
fishes more than any other group show what zoological riches 
may still be discovered there. The Lerma river system, for 
example, in south-western Mexico has a fish fauna, accord- 
ing to Dr. Meek,* which is quite as distinct and character- 
istic as if it were on an island in the sea. Of the forty -nine 
species of fishes now known to occur in this area, not a single 
one is found elsewhere. These forty-nine species belong to 
seventeen genera, ten of them being quite peculiar to this 
region. This result is all the more surprising, as it has often 
been argued, by advocates of accidental introduction, that the 
eggs of fishes are apt to adhere to the legs or feathers of water 
birds, being thus easily transported to other river systems 
or isolated lakes. The extremely distinct and isolated 
character of the Lerma river area implies that dispersal of 
fishes is not affected by such agencies of accidental transport. 
Fishes only migrate from one river to another when a change 
of drainage occurs in the head waters, or when the stream 
itself shifts its course. 

* Meek, S. E., " Fishes of Mexico," pp. 775784." 



THE problems presented by a study of the fauna of Central 
America are of the utmost significance in our investigations 
into the origin of the American fauna as a whole. Even 
the casual observer cannot fail to perceive that certain animals 
from each of the two great continents lying to the north and 
south of Central America, tend to spread along the narrow 
isthmus, and thus intermingle with one another. When we 
look at Dr. Wallace's map of what he calls the " Neotropical 
Region," and notice how the South American fauna has 
apparently invaded the whole of Central America, and even 
crept northward along the lowlands of Mexico, so that the 
triangular table-land of that country, with its northern in- 
habitants, looks as if it had been forced like a wedge between 
the two wings of the southern army of invaders, the whole 
history of events seems to be clearly unfolded before our eyes. 
Apparently, quite a simple zoogeographical problem, and one 
that is easily soluble by a study of the distribution of existing 
animals. Thus it seemed to Dr. Wallace. Of the geology of 
Central America nothing was known when he wrote his 
famous work on the distribution of animals. Nevertheless, 
he argues (pp. 10 13), from the sudden appearance in post- 
Tertiary times of numerous South American forms of 
edentates in temperate North America, and from such facts 
as the occurrence of some identical species of sea fish on the 
two sides of the Central American isthmus, that the union 
of North and South America must be a comparatively 
recent event, and that these continents must have been sepa- 
rated during Miocene and Pliocene times by a wide arm of the 
sea. When the evidence of both land and sea animals support 
each other as they do here, adds Dr. Wallace,* the conclusions 

* Wallace, A. E., " Distribution of Animals," Vol. II., pp. 5759. 


arrived at are almost as sure as if we had geological proof 
of these successive subsidences. The author of the "Geo- 
graphical Distribution of Animals " thus not only claims that 
zoogeography is of great value as an interpreter of geological 
phenomena, but he feels confident of having indicated the 
correct solution to the problem of the faunistic relationship 
between the two continents. There are one or two points, 
however, in the fauna of South and Central America, which 
seem to throw doubt on Dr. Wallace's interpretation of the 
facts of distribution. He states (p. 58) that when the final 
union of the two continents took place, the tropical climate 
of the lower portion of Guatemala and Mexico would have 
invited rapid immigration from the south, while some 
northern forms would have extended their range into and 
beyond the newly elevated territory. That the fauna of South 
America has advanced across the isthmus, and has even in- 
vaded the State of Texas, appears probable from a study of 
the existing fauna. We need not even cite the presence of 
the South American fossils in North America in favour of 
this view. All sections, indeed, of the fauna seem to have 
taken part in this northward advance, even "the fishes. 

Take for example the members of the family Cichlidae. 
Over one hundred and fifty species of this strictly fresh- 
water group are known from America, the great majority 
being confined to South America. Mr. Regan * is of opinion 
that the Central American and Mexican species are more 
specialized than the South American ones, and that the 
former have certainly been derived from the latter. No 
member of this family of fishes is known from the Mexican 
plateau, and only three or four species extend northward 
on either side of it, North America being almost devoid 
of Cichlidae (see Fig. 12). All the available evidence 
thus points to South America as the place of origin of the 
family, as fax as the New World is concerned ; and thence it 
has spread northward, like hosts of other groups of animals. 
This advance must, of course, have taken place as soon as a 
connected land passage with suitable lakes and rivers enabled 
them to proceed northward. Dr. Wallace thinks that the 
sudden appearance of the large South American edentates 

* Regan, C. Tate, " Fishes of Central America," pp. xiii- xvi. 



in North America occurred in post- Tertiary times, for he 
believes Central America to have been still submerged during 
the Pliocene Period. The term " post-Tertiary," I may men- 
tion, is synonymous with Quaternary or post-Pliocene. Any 
geological deposits more recent than Pliocene come within 
the meaning of that term. But the skeletons of the great 
South American M^lodon_and Megalonyx certainly occur in 
Texas in true Pliocene beds. I do not think there is any geo- 
logist in America now who would uphold the Pleistocene, or 
even less the post-Glacial age of these deposits. My conten- 
tion, therefore, is that the northward advance from South 
America is a pre-Glacial or pre-Pleistocene event. 

Now one of the most remarkable and astonishing features 
of that faunistically so peculiar continent of South America 
is that, whereas its tropical fauna has very little affinity 
with the fauna of North America, the more remotely placed 
Chilean and Patagonian faunas present in some groups of 
animals a striking resemblance to it. This character will be 
more fully dealt with in another chapter (pp. 410 419). It 
may be mentioned, however, that numerous groups, and even 
species, of northern plants are met with in Chile, which are 
wholly, or almost entirely, absent in the intervening region, 
occupying an area of thousands of miles. Northern genera of 
butterflies and beetles, such as Argynnis, Colias and Carabus, 
all of which are almost unknown in the countries immediately 
south of Mexico, reappear in numbers in the extreme southern 
tip of South America. Dr. Wallace thought that this south- 
ward migration of northern forms of animal life must have 
been effected mainly during successive Glacial epochs, when 
the mountain-range of the isthmus of Panama, if moderately 
increased in height, might have become adapted for the 
passage of northern forms, while storms might often have 
carried insects from peak to peak, over intervening forest 
lowlands, or narrow straits of sea. Dr. Wallace's idea that the 
mountains all along Central America were formerly higher 
than they are now and sustained northern forms of animal 
life is not supported by any evidence. Considering that he 
imagined the long isthmus to have been slowly rising from the 
sea since pre-Glacial times, Dr. Wallace's suggestion that 
the mountains were so much higher during the Glacial Epoch 


than they are now, hardly seems well founded. Another still 
stronger objection to Dr. Wallace's * theory is that the 
northern forms alluded to as occurring in Chile and Patagonia 
belong almost all to different species, sometimes even to dif- 
ferent genera, from their northern relations. If storms had 
anything to d'o with this distribution they could only have 
acted during very long intervals of time so as to produce such 
specific and generic differences. Moreover, how could winds 
or storms affect the distribution of Carabus, which is a flight- 
less ground insect living under stones ? How could these 
agencies have transported fresh-water species across the im- 
mense tropical area, for several Chilean fresh -water forms 
exhibit a similar northern affinity ? These are some of the 
problems that present themselves to us. There are numbers 
of others. Why should the family of tortoises, Dermate- 
mydidae, which is known to have inhabited the North 
American continent since Cretaceous times, have become ex- 
tinct there and be now confined to Central America ? 

The scarcity of land and fresh-water fossils in Central 
America obliges us to resort to zoogeography and to the meagre 
geological information we possess in elucidating these and 
other problems. Before dealing with the general faunistic 
features of Central America, a few remarks on some of the 
more important geological characters will be of interest. 

The long neck of Central America from the isthmus of 
Tehuantepec to Panama, which joins North and South 
America to one another, has a length of about one thousand 
five hundred miles. We are sometimes apt to forget that it 
does not lie in a north and south direction, but almost east and 
west. Nearer South America the neck of land starts in a 
due westerly direction and only gradually turns somewhat 
towards the north and finally north-westward. Very little of 
this immense stretch of land has as yet been geologically sur- 
veyed. Nevertheless, some valuable hints as to its geological 
history have been gathered. In his essay on the geology of 
the isthmus of Panama, Professor Hill f tells us that, pos- 
sibly before the vast accumulations of more modern igneous 
and sedimentary rocks of Tertiary and post-Tertiary age were 

* Wallace, A. &., " Distribution of Animals," Vol. II., p. 45. 
t Hill, R. T., "Geological History of Panama," pp. 241257. 


laid down, a foundation of granitic rocks, occurring in an east 
and west arrangement, existed on the site of Central America. 
Apparently parallel granitoid ridges extended from the longi- 
tude of Trinidad directly across the path of the main con- 
tinental trends through forty degrees, as far north as Acapulco 
in Mexico. These fragmentary data,' he says, are sufficient to 
indicate that in pre-Tertiary times there may have been a 
basement barrier of granitic rocks forming an oast and west 
arrangement which outlined the Central American region, and 
constituted an ancient buttress against or upon which the 
higher mountain folding has originated. Professor Suess,* 
too, speaks of the mountain chains of Yucatan and Guate- 
mala as the western continuations of the Antillean system, 
contending that North and South America are to be looked 
upon as two radically distinct continents, separated from one 
another by a third element, that of Central America and the 
Antilles. All this agrees to some extent, as we shall see later 
on, with the conclusions we can draw from a study of zoogeo- 
graphy. Central America and the Antilles, which are collec- 
tively spoken of sometimes as the remnants of an ancient 
" Antillean Continent," possess a distinct and peculiar fauna 
quite apart from the South American one which has invaded 
this area. 

Towards the end of the Mesozoic Era parts of this Antillean 
continent must have begun to subside. About that time signs 
of the coming volcanic activity appeared all along the Central 
American region. During the successive igneous eruptions in 
early Tertiary times, which have been continued with varied 
intensity to the present day, the Atlantic Ocean seems to have 
invaded the existing area of Central America and submerged 
portions of it. At any rate, Professor Hill f states that 
biological and geological evidence led him to the con- 
clusion that a shallow marine transgression must have existed 
somewhere in Central America during Eocene times, although 
there is nothing to show with certainty that the isthmus 
of Panama was the exact site of this inter-oceanic con- 
nection. All the same, Professor Schuchert J seems to be 

* Suess, E., " Antlitz der Erde," Vol. I., p. 700. 

t Hill, E. T., " Geological History of Panama," p. 265. 

{ Schuchert, 0., " Paleogeography of North America," pp. 9697, 


satisfied that the isthmus of Panama was submerged daring 
the Eocene Period, and that the submerged area of Central 
America greatly increased in Oligocence times. Professor 
Hill, who has dealt with various lines of enquiry in regard to 
the geological history of Central America, contends that they 
all give evidence for the belief that no connection has existed 
between the Pacific and Atlantic Oceans since the close of the 
Oligocene (p. 270). A very careful re-examination of the 
fossils of Gatun, near the Panama Canal, by Professor Toula, 
however, throws considerable doubt on the supposed Eocene 
age of the deposits containing these fossils. He gives reason 
for his belief that they are of upper Miocene, possibly even of 
lower Pliocene age, and estimates the depth of water of the 
marine channel uniting the two oceans at about a hundred and 
fifty feet.* 

In conjunction with Dr. Bose, the same author also in- 
vestigated the Tertiary deposits of the isthmus of Tehuantepeo 
in southern Mexico with very noteworthy results. The fossil 
mollusks contained in them indicate that the sea covered the 
land to a depth of from fifty to two hundred fathoms, that is to 
say from three hundred to one thousand two hundred feet. 
As the isthmus does not rise much beyond eight hundred 
feet, the existing land, even as far north as this region, 
was submerged by a shallow sea. Dr. Bose is inclined to attri- 
bute to this fauna an early Miocene age, while Professor 
Toula believes it to be younger, in fact distinctly " jung- 

Now it has been argued, and the argument appears most 
reasonable, that we are able to check these results indicating 
a submergence of parts of Central America by means of two 
other tests, viz., the distribution of living animals and the 
palaeontology of North America. The first deals mainly with 
the amount of affinity existing between the marine animals 
of the Atlantic and Pacific sides of Central America. From 
the nearness or remoteness in relationship of the species on 
the two opposite coasts it was thought we might determine at 
what particular geological period, if at any, the Atlantic and 
the Pacific Oceans were joined to one another across Central 

* Toula, F., " Jungtertiare Fauna von Gatun," pp. 744745. 

t Bose, E., and F. Toula, "Fauna von Tehuantepec," pp. 221 and 273. 


America. The second test is that of the fossil mammals of 
North America. It was pointed out that, as no trace of typi- 
cally South American species occur in North American Oligo- 
cene or Miocene deposits, the two continents must have been 
separated by a sea during the period in which these beds were 
laid down. Let us examine these evidences more closely. 

A good many writers have discussed the problem of the 
former junction of the Pacific and Atlantic Oceans across 
Central America, from a purely biological point of view. Pro- 
fessor Gregory,* who last summarised the subject, came to the 
conclusion that the waterway across Central America was 
finally closed in the Lower Miocene, or possibly even in the 
Upper Oligocene. Among others he alluded to the researches 
of the two Agassiz, father and son, stating that, from a study 
of the sea-urchins, they proposed to date the junction of the 
two oceans much further back. These zoologists referred the 
separation of the two oceans and the formation of the Gulf 
Stream, to the period at the close of the Cretaceous, that is to 
say, to the end of the Secondary Era. Professor Verrill f 
finds that no species of corals are identical on the two sides 
of Central America, while even the genera and families show 
remarkable contrasts. The numerous genera and families of 
reef -building corals, so abundant on the Atlantic side, are 
wholly wanting on the Pacific, with the exception of Porites, 
which is represented by three or four small species. 

Mr. Belt J maintained that the marine mollusks on the two 
coasts separated by the narrow isthmus of Darien were almost 
entirely distinct. So remarkably distinct are the two faunas, 
he says, that most zoologists consider that there has been no 
communication in the tropics between the Pacific and Atlantic 
Oceans since the close of the Miocene Period. An apparently 
rather remote junction of the two oceans is suggested by 
all these authors, but none of them express any very strong 
convictions on the subject. Dr. Ortmann, on the other hand, 
states that the affinities of the Decapod fauna of the Atlantic 
and Pacific are unmistakable, and that we have ample and con- 
vincing evidence of a former connection between these oceans. 

* Gregory, J. W., " Palaeontology of the West Indies," pp. 304305. 
t Verrill, A. E., " Comparison of Coral Faunae," p. 500. 
| Belt, Th., " Naturalist in Nicaragua," p. 264. 


Seven species of crabs and their allies are identical to the two 
oceans, while many others are closely related. The fauna as a 
whole implies that the union of the seas cannot have been of 
a very recent date. Accordingly Dr. Ortmann * is of opinion 
that there was no communication since Miocene times. 

In 1880 Dr. Gunther f declared that the genera of fishes, 
with scarcely any exception, are identical on the two sides 
of Central America. Of the species found on the Pacific side, 
nearly one half, he states, have proved to be the same as those 
of the Atlantic, and he upholds the view of the existence o'f 
straits and channels between the two seas until a geologically 
recent period. Since Dr. GunJther expressed these views, the 
fish fauna of this region has received renewed attention and 
study, over a thousand species now being known from the 
coasts of the two seas. Of this large number only seventy - 
one species, or six per cent., are common to the two oceans, 
according to the latest researches of Professor Jordan. This 
great discrepancy between Dr. Giinther's and Professor 
Jordan's results arises, of course, to a large extent from the 
different views these observers hold as to specific limits. 
To account for the remarkable identity of genera and the 
divergence of species, Professor Jordan J suggests that the 
isthmus of Panama was depressed in or before Miocene times, 
that the channel was shallow, and that the currents set chiefly 
westward, thus favouring the transfer of Atlantic rather than 
Pacific types. In the case of the Medusae, the affinity between 
the littoral faunas of the two sides of Central America is 
much closer according toDr.Bigelow than in the higher groups. 

It has been suggested by Messrs. Gill and Bransford || that 
the occurrence of marine fishes in the great lake of Nicaragua 
is to be looked upon as an indication that the sea once flooded 
the area now occupied by these lakes. That these fishes are 
true " relicts " of the past, representing the survivors of a 
marine fauna, has recently been reaffirmed by Dr. M'eek.^f 

* Ortmann, A. E., "Distribution of Decapods," p. 398. 

t Gunther, A., " Study of Fishes," p. 280. 

J Jordan, D. S., "Study of Fishes," Vol. I., pp. 272280. 

Bigelow, H. B., " Medusae," p. 228. 

|| Gill, Th., and J. F. Bransford, " Fishes of Lake Nicaragua," p. 179. 

f Meek, S. E., " Fishes of Nicaragua," p. 99. 


The species in question are, in the first instance, two sharks, 
viz., Carcharhinus nicaraguensis and Pristis antiquorum, as 
well as Pomadasis grandis and others whose ancestors are 
marine forms. The antiquity of the region is indicated by 
the occurrence both in Lake Nicaragua and Lake Managua 
of the Central American gar-pike (Lepidosteus tropicus). 

The studies of geographical distribution, particularly those 
of marine species, have thus led to the conclusion that tha 
Central American land bridge has been in existence certainly 
since pre-Glacial times. No zoologist has suggested even the 
possibility of a submergence of Central America having taken 
place as late as the Pleistocene Period. No biological support 
can be given, therefore, to the theory which has beien advanced 
by some writers, that the Glacial Period was due to a diversion 
of the Gulf Stream across the isthmus of Panama. On the 
other hand, Central America was certainly submerged by a 
shallow sea in two or more places in early Pliocene or Miocene 
times. Another very important character, which is revealed 
both from a study of the recent and fossil marine fauna, is the 
Atlantic facies of the animals. The movement of the species 
seems to have taken place from the West Indian area towards 
the Pacific Ocean, thus implying the existence of a strong 
current in that direction. Speaking of the Tertiary deposits 
of Tehuantepec, Dr. Bose * remarks that the main mass of the 
species contained therein are related to Atlantic forms. Only 
very few show affinities with Pacific types. Similar views 
were expressed by Professor Jordan and Dr. Ortmann an 
regard to the recent marine fauna. 

A further complication, as Professor Gregory, f pointed 
out, oremains to be considered. It is not at all certain, he 
thinks, that when the Isthmus of Panama was submerged 
there was free communication between the Atlantic and the 
Pacific Oceans. The Caribbean Sea, he argues, may then have 
been a gulf of the Pacific, separated from the Atlantic by the 
land area of the hypothetical " Antillia." As I shall show in 
the next chapter, which deals with the origin of the West 
Indian fauna, the Caribbean Sea could not have been' .separated 
from the Atlantic Ocean at so late a geological period as that 

* Bose, E., and F. Toula, " Fauna von Tehuantepec," p. 220. 
t Gregory, J. W., Palaeontology of the West Indies," p. 305. 
L.A. B 


during which the S6a spread across Central America. The 
general opinion seems to be that the submergence of Central 
America was due to extensive subsidence in the Caribbean 
area and likewise in the Gulf of Mexico. The Mediterranean 
character of the West Indian marine fauna, moreover, implies 
the probability of a free migration from the one area to the 
other along some ancient shore-line. A land bridge joining 
North and South America along the chain of the Lesser 
Antilles, alluded to by Professor Gregory, may possibly have 
existed; but it must, I think, already have been destroyed 
at the time of the submergence of the Isthmus of Panama. 
And yet I concur with Professor Gr.egory in thinking that 
when the latter was submerged there need not necessarily 
have been free communication between the Atlantic and the 
Pacific Oceans. We must remember that all efforts have 
hitherto failed to discover any traces of Tertiary sediments on 
the sea-board betweeoi southern Mexico and Panama. This 
seems to imply that land lay to the west of Central America, 
and that the Pacific Ocean was formerly situated further 
westward than at present. What would appear as connections 
between the two oceans may have been merely shallow bays in 
the land referred to. Certain peninsulas would then have 
projected eastward from this old Pacific land towards those 
parts of Central America that were then in existence (see 
Fig. 16). Without giving further evidence, these theories 
may appear somewhat visionary, but as the subject will be 
more fully discussed later on (p. 408), I need not enlarge 
upon it at present. 

A comparison of the marine faunas of the two shores of 
Central America does not yield such satisfactory results in 
establishing the geological age of the submergence, because 
we have as yet little idea of the length of time during .which 1 
animals may retain their specific characters. The evidence 
derived from the first appearance in North America of dis- 
tinctly South American mammals would seem to give us a 
better clue as to the date of the formation of the present 
Central American land bridge. 

This appeal to the past dispersals of mammals in recon- 
structing former conditions of land and water has been utilised 
in several of the previous chapters, and in this case many bio- 


legists have attested their strong belief in the validity of the 
argument. A very early Tertiary or late Mesozoic influx of 
mammals from South America into North America has already 
been alluded to. After this event long periods of time elapsed, 
during which the two continents were seemingly separated 
from one another. Th,en southern mammals once more 
appeared in the north. This later invasion is proved from 
the contents of the deposits in Texas. Here we meet 
with gravigrade edentates, and these deposits have now 
been definitely placed by Professor Osborn * to the middle 
Pliocene. Hence Central America in its present form and 
shape would be of Pliocene origin. Although Mr. Lydekker f 
places this event at the end of the Miocene Period, Professor 
Deperet J and Dr. Smith Woodward concur in the opinion 
expressed by Professor Osborn which is in conformity with 
that elicited by Professor Toula. The latter bases his evidence 
on the fossils contained in th.e Panama and Tehuantepec 

Since the testimony derived from the recent marine fauna 
also agrees fairly well with the above conclusions, it seems 
reasonable to conclude that Central America in its present 
outlines, forming a highway for intercommunication between 
North and South America, came into existence about the 
beginning of the Pliocene Period. Thus one of the problems 
alluded to at the beginning of the chapter is apparently 

Yet still another difficulty has suddenly arisen owing to 
the recent most surprising discovery of true edentate re- 
mains of Megalonyx type in the Mascall beds of Oregon, 
which are of Middle or Lower Miocene age.|| If the 
Gatun deposits near Panama are really, as Professor Toula 
affirms, of Upper Miocene age, how can we reconcile the 
submergence of Panama, and probably also of the isthmus of 
Tehuantepec, with this latest discovery in Oregon ? The whole 
problem is evidently much more complex than it at first ap- 

* Osborn, H. F., " Cenozoic Mammal Horizons," p. 82. 

t Lydekker, E., " History of Mammals," p. 119. 

| Deperet, 0., " Transformations of Animal World," p. 282. 

Woodward, A. Smith, "Palaeontology," p. 429. 

|| Osborn, H. F., " Age of Mammals," p. 289, 


peared. In the Oligocene deposits of the northern continent, 
as above mentioned, there are no distinctly South American 
species. Yet, curiously enough, when we come to still earlier 
strata, we again meet with remains of animals that exhibit 
characteristically South American features. In the Puerco 
formation, in north-western New Mexico, a number of 
peculiar mammalian bones have been discovered, which were 
referred by Professor Cope to the extinct order Tillodontia, 
whereas Dr. Wortman endeavoured to show that Cope's 
genera Psittacotherium, Onychodectes and Conoryctes were 
ancestral to the Gravigrada or ground sloths of South America. 
Dr. Wortman * proposed that they be placed in a new sub- 
order of the edentates, which he named Ganodonta. But he 
did not look upon these animals as immigrants from South 
America. He thought this order of primitive mammals must 
have actually arisen in North America, and have thence emi- 
grated to South America before the close of the Eocene 
Period. Although these Ganodonta are no longer considered 
as ancestral to the ground sloths, the same Puerco formation 
has yielded other mammals which show distinctly South 
American or rather Patagonian affinities. Dr. Wortman's 
theory as to the North American origin of the Ganodonta 
has not found favour. Dr. Osborn, in fact, urges that a 
direct land connection with South America is indicated at 
this stage of geological history in order to account for the 
South American features in the North American fauna. This 
view has been amply confirmed by the remarkable discovery in 
Wyoming, in a deposit of Middle Eocene age (Bridger), of 
the remains of a true armadillo closely related to the modern 
armadillos, but exhibiting a few more primitive characters. f 

Since Dr. Ameghino's researches in Patagonia have brought 
to light such a wealth of edentates from the earliest 
Tertiary, and probably even from Mesozoic deposits, scarcely 
anyone can doubt that South America is the original home 
of that group of mammals, and that they have passed from 
there during the Eocene Period and earlier into North 
America, and not vice versa, as Dr. Wortman suggested. 
But very few would assert that the physical geography of the 

* Wortman, J. L., " Psittacotherium," pp. 259262. 

i Osborn, H. F., " An Armadillo from the Eocene," p. 163. 


New World was the same then as it is now, and that Central 
and South America had already been evolved in anything like 
the present outlines. As I shall endeavour to demonstrate 
later on, South America did not then exist as a distinct great 
continent. A large land-mass evidently lay in the neigh- 
bourhood of the existing State of Brazil and another further 
south. And as far as we Imow, the southern land-mass was 
the original home of the edentates. Between it and Central 
America on the site of the present South American continent 
there were one or more broad marine channels, or oceans, as 
we might call them. And yet the edentates succeed cd in 
attaining North America. I do not now wish to discuss my 
reasons for the supposition that western Mexico was then 
united by a direct land bridge with Chile. I only mention it 
in order to indicate that the appearance of edentates in the 
Eocene of North America does not afford a proof of the ex- 
istence of Central America at that time, nor during the 
Miocene Period. 

Professor Osborn, as I mentioned before, indicates the 
nature of the land connection between North and South 
America in the dawn of the Tertiary Era on a map which he 
kindly allowed me to copy (Fig. 21). However, he expresses 
the belief that already in early Eocene times, that is to say 
almost before the appearance of the above-mentioned arma- 
dillo in North America, the land bridge had ceased to exist. 
We are too apt, I think, to look upon South America as exclu- 
sively the home of edentates, forgetting that many other mam- 
mals may have originated there too. We may not all agree 
with Dr. Ameghino * in attaching the importance he does 
to that continent as a source of the Tertiary mammalia, 
but I believe we possess other evidences of a faunistic inter- 
change between Chile and Patagonia on the one hand, and 
western North America on the other, during the ages that 
passed between the Lower Eocene and the Miooene. 

In southern Africa we meet with a group of small blind 
subterranean creatures, the golden moles (Chrysochloridae) 
which are among the most primitive mammals in existence. 
They are quite confined at present to South Africa. But 

* Ameghino, Fl. " South America, the Source of Mammalia." 


within recent years the remains of a small insectivore have 
been met with in the early Tertiary Santa Cruz deposits of 
Patagonia. This mammal, according to Professor Scott,* is 
genetically related to the African golden moles. The South 
American Necrolestes, as it has been called, is certainly more 
primitive in structure than its African relations. Moreover, 
in western North America fossil mammals have been dis- 
covered both in Oligocene and Miocene strata which have been 
referred to the same family Chrysochloridae. It is true that 
Dr. Matthew f is now inclined to place the Oligocene Apter- 
nodus among the nearly related family Centetidae, but the two 
other genera Xenotherium and Arctoryctes are still looked 
upon as true chrysochlorids. A possible land connection 
between Africa and South America will be discussed later on. 
Whether Patagonia was the original home of the chryso- 
chlorids or South Africa we cannot tell, but the family may 
possibly have spread from South America to North America 
by utilising the hypothetical land bridge that I think lay to 
the west of the present continent. 

A few words of appreciation of the splendid work that has 
been done in making us acquainted with the rich fauna and 
flora of Central America are due to Mr. Godman. For years 
he and Mr. Salvin laboured with great industry and at con- 
siderable expense in bringing together an immense collection 
of vertebrates and invertebrates, subsequently publishing th& 
series of beautifully illustrated volumes of the " Biologia 
Centrali- Americana " in which the results of their studies were 
made known to the scientific world. In the volume describing 
the botany of Mexico and Central America there is an excel- 
lent summary dealing with the constituents of the flora and 
their relationships. No such summary has been attempted in 
the other volumes, so that Mr. Hemsley's account of the plants 
of Mexico and Central America is of particular value to those 
who are engaged in a study of the zoogeography of that region. 

The first item of interest is one which we have noticed 
occasionally among apparently very ancient groups of North 
American animals. Genera like the amphibians Spelerpes 
and Amblystoma, which have their headquarters in Mexico, 

* Scott, W. B., "Eeport of Princeton Expedition," Vol. V. 
t Matthew, W. D., "The Skull of Apternodus," p. 35. 


are either absent or very rare in the western United States of 
North America, while abundant in the eastern States. Some 
of these bear an impress of their antiquity in their wide and 
discontinuous range. Now Mr. Hemsley enumerates the fol- 
lowing genera of trees as occurring in southern Mexico and in 
the Atlantic States of North America, though they are absent 
from the Pacific forests of the western United States: 
Magnolia, Asimina, Tilia, Bobinia, Liquidambar, Ilex, Dio- 
spyros, Bumelia, Ulmus, Celtis, Morus, Ostrya, Carpinus and 
Carya. Even species of plants from southern Mexico and the 
Atlantic States of North America are sometimes identical, 
such as Liquidambar styraciflua, Ostrya virginica and Car- 
pinus americana. And yet only four out of the fourteen 
genera referred to extend even to northern Mexico. Of some 
of these we possess fossil evidence that they lived in Europe 
already in early Tertiary times, and we may safely assume that 
the whole group is of great antiquity. The flora of Guatemala 
is essentially of the same composition, according to Mr. 
Hemsley, as that of southern Mexico, though apparently less 
rich in specific diversity. Some of the trees just alluded 
to, such as limes (Tilia) and elms (Ulmus), are unknown 
in Guatemala; others, for instance sweet gums (Liqui- 
dambar), mulberries (Morus), lever-wood (Ostrya) and horn- 
beams (Carpinus), occur in that country. The southern floral 
province of Mr. Hemsley comprises Nicaragua, Costa Bica 
and Panama ; and, as might be expected, these countries ex- 
hibit a much closer relationship with the South American 
tropical flora than Guatemala or Mexico do. The endemic 
generic element of the whole of Mexico and Central America 
is rather inconspicuous, but the southern floral province is by 
far the poorest of the three into which the region has been 
divided. One of the most curious features in the constitution 
of the flora of Mexico is one which I have already briefly, 
referred to, namely, the presence there and in the extreme 
south of South America of certain northern genera of plants 
which are absent or only represented in a few scattered dis* 
tricts in the intermediate region. Mr. Hemsley assumes that 
such plants have spread southward in remote times. There 
are likewise genera of distinctly southern origin with a simi- 
larly discontinuous range in a northward direction. I need 


only mention Fuchsia, which is abundant in Chile and Pata- 
gonia, reappears more sparingly in Colombia, and is almost 
unknown further north until we come to Guatemala and 
Mexico. The extreme antiquity of this genus is indicated 
by its occurrence on the island of Haiti (F. triphylla), while 
it is altogether absent from the rest of the West Indian 

In his analysis of the flora of the whole region, Mr. 
Hemsley * informs us that the northern province, that is 
to say the plateau of Mexico, is the focus of a xerophilous 
flora extending into the dry regions of south Mexico and into 
the territories north of Mexico. The central province, com- 
prising southern Mexico and Guatemala, possesses a mingling 
of northern and southern types exhibiting an extraordinarily 
rich production of local species. The southern province is 
merely an outlier of the American tropical flora. The tropical 
element of the whole flora is more closely allied to that of 
eastern South America than to that of the West Indies. 

The relationship of the Mexican and Central American 
floras to those of other parts of the world is also of import- 
ance. Thus Mexico and Central America have a large number 
of genera in common with eastern Asia, with Africa, Mada- 
gascar and Europe. They likewise exhibit affinities with the 
Galapagos islands, with New Zealand, Australia and Polynesia. 

Now, if we compare the results of the study of botany with 
those derived from an examination of the mammalian fauna, 
we are struck at once by the fact that the plants among which 
the mammals live must be of much greater antiquity than the 
latter. As Mr. Alston f tells us, the mammals of Mexico 
and Central America are composed, partly of southern species 
not found southward of Mexico, of a few that extend as far 
south as Panama, and chiefly of South American ones which 
have spread across Central America. There are also a number 
of autochthonous species in this region. Yet very few of the 
genera are found in other parts of the world than America. 

Among our new acquaintances we meet for the first time 
with the kinkajous, one species of which (Potos flavus) has 
succeeded in reaching the State of Vera Cruz in Mexico. 

* Hemsley, W. B., " Botany of Central America," pp. 306315. 
t Alston, E. R., "Mammals of Central America." 


From my previous remarks (p. 152) it would appear as if the 
raccoon family (Procyonidae) had originated in some western 
land in America, and yet the genus Cercoleptes (Potos), which 
belongs to this family, is certainly an invader from the south. 
Indeed, when we examine the range of the members of this 
typically American family of Procyonidae, we notice the 
peculiar feature that almost all the species are confined to the 
Pacific coast. The raccoon (Procyon lotor) no doubt has ex- 
tended its range to the eastern States, while the allied species 
Procyon maynardi, as we have learnt, is even confined to the 
Bahama islands, and one, the coati (Nasua rufa), has a wide 
distribution in South America from Bolivia eastward. Almost 
all the other members of the family, however, inhabit 
curiously disconnected areas in the vicinity of the Pacific 
Ocean. Bassaricyon lives in Ecuador, Panama and Costa 
Rica. One species of Bassariscus is peculiar to the island 
of Espiritu- Santo near lower California, another ranges from 
Mexico to the western States, a third occurs in western 
Mexico, Guatemala, Costa Rica and on Mount Chiriqui, in 
western Panama, at a height of 6,000 feet. Of the coatis 
(Nasua), one species passes from Mexico northward to Cali- 
fornia and southward to Costa Rica, another is confined to 
the island of Cozumel, a third lives in the Ecuador mountains 
at a height of 7,000 feet, whereas Nasua olivacea is met 
with in Santa Fe de Bogota and in the Merida of Venezuela 
at heights up to 12,000 feet. Altogether it looks as if the 
members of the family Procyonidae had spread from various 
western foci. Some of them may have retained their original 
distribution, while the more adaptable genera sent outposts 
eastward into the great continents. The early stages of this 
evolution must have taken place before either Central America 
or South America had become consolidated into anything like 
their present shapes. Later on I shall have occasion to dis- 
cuss other similar cases of discontinuous distribution occur- 
ring among the lower vertebrates. All of these appear to be 
due to the same peculiar features in the physical geography 
of Tertiary America. 

In eastern Mexico we make our first acquaintance with 
monkeys. In early Eocene times, as already mentioned, 
monkeys, belonging to extinct groups, probably entered the 


United States from the south. They seem to have had a short 
existence in North America, for no trace of them has been 
noticed in later deposits. Ever since renewed facilities for a 
northward advance occurred an entirely new invasion has 
taken place, and one of the " howlers " can now be heard at 
night even in the forests of Vera Cruz in eastern Mexico. A 
spider monkey ( Ateles vellerosus) may be seen in the, same dis- 
trict gracefully swinging from branch to branch. These 
monkeys differ in distribution from the raccoon family in so 
far as they have their centre of distribution in Brazil and 
northern South America, from which they no doubt invaded 
Central America in more recent geological times. There are 
other families of mammals which we meet here for the first 
time. We need not dwell on them any longer, as we shall 
have occasion to become acquainted with them in subsequent 

The wealth of new bird life in Central America is very 
striking, and is vividly described 'in Mr. Belt's * delightful 
book of travels in Nicaragua. Among the more noteworthy 
families are the toucans (Rhamphastidae), with their enormous 
gaily-coloured bills, the humming-birds (Trochilidae), a great 
variety of parrots, the peculiar curassows and many others. 
Some of these of course, like the humming-birds, enteir the 
United States largely during their northward migrations, and 
to some extent are resident in the southern States. I should 
also mention the most beautiful of all birds, the quesal 
(Trogon resplendens), or royal bird of the Aztecs as it has 
been called, with its delicately tinted plumage of metallic 
green and blue, and its long waving plumes. The Trogonidae 
now have their headquarters in Central and South America, 
but Trogon gallicus occurs in France in Miocene deposits ; 
and this seems to suggest in what manner the early members 
of the family crossed over to Africa and the Oriental Region, 
where some genera are still found living ; the discontinuous 
range corroborating the palaeontological evidence of the great 
antiquity of the group. As Messrs. Salvin and Godman f 
remind us, the avifauna of Central America is essentially 

* Belt, J., "Naturalist in Nicaragua." 

t Salvin, 0., and F. D. Godman, " Birds of Mexico and Central 


neotropical, with certain peculiar endemic forms, the greater 
number among 'the latter being confined to Mexico arid 

I alluded above to the fact that the family of tortoises, 
Dermatemydidae, is entirely confined to Central America. 
The only species of the family (Dermatemys mawi) inhabits 
precisely that part of Central America which we have reason 
to believe to be one of its oldest parts, viz., Guatemala and the 
neighbouring Honduras and Yucatan. Several members of 
the family formerly lived in North America. They first ap- 
peared in Cretaceous times, and still inhabited the south- 
western States during the Eocene Period. It is possible they 
may then have spread to Guatemala, becoming subsequently 
extinct in their centre of dispersal.* Another family which 
I mentioned before (p. 134), the " snapping turtles " (Cheli- 
dridae), occur in eastern North America from Mexico to 
Canada. Southward of their range they are found only in 
Guatemala and Ecuador. We have noted examples of a 
similar discontinuous range before. Others will be cited in 
the next chapter. 

More important from a faunistic point of view are such 
creatures, as for instance the snake-like limbless amphibians, 
belonging to the family Coeciliidae. They live in moist ground, 
and lead altogether a burrowing life. Their distribution ought 
to give us, therefore, some valuable hints as to former changes 
of land and water. Dr. Sarasin f looks upon this family as a 
pre-Cretaceous relict, on account of its peculiar discontinuous 
range, namely India, the Seychelles, east Africa, west Africa, 
South and Central America. Whether we agree With him or 
with Dr. Alcock,J who believes that the family wandered 
along a continuous land surface from India across Africa to 
northern South America in early Tertiary times, there can be 
no doubt at all as to these subterranean amphibians being 
exceedingly ancient. The American home of this circum- 
tropical family, as Dr. Gadow points out, is South America. 
No members are known from the West Indies or the Galapagos 

* Hay, O. P., " On Fossil Turtles," p. 32. 

t Sarasin, F., " Geschichte d. Tierwelt von Ceylon," pp. 7476. 

t Alcock, A., " Description of Apodous Amphibian," p. 270. 

Gadow, H., " Distribution of Mexican Amphibians," pp. 199200. 


islands. Dr. Gadow next proceeds to argue that because one 
of these coecilians inhabits eastern Mexico, its ancestors must 
have travelled slowly across the whole neck of Central America 
since the close of the Miocene Epoch, when he assumes the 
isthmus to have been first opened up for southern immigrants. 
This argument is in so far faulty, as the coecilians need not 
necessarily have passed through Central America. The inti- 
mate relationship that exists among many ancient species of 
Central America to those of northern South America suggests 
the existence of some far older link between these countries. 
In very remote times species were, I believe, able to reach 
certain areas such as Guatemala and western Mexico long 
before the present Central America had com into existence, 
that is to say long before Pliocene times. Dr. Gadow himself 
urged that the Isthmus of Panama is but the last vestige of a 
former much broader land connection between North and 
South America (p. 243). In my opinion this should read 
" the Isthmus of Panama contains some vestiges of a former 
much broader land connection." 

To the uninitiated the Typhlopidae would seem nearly 
related to the coecilians. Both are snake-like burrowing crea- 
tures, and yet the former are true snakes and, therefore, 
reptiles, while the others are merely limbless amphibians. On 
close examination the true burrowing-snakes (Typhlopidae) 
are found to be covered with minute cycloid scales, and to 
exhibit other reptilian characters. Their distribution is ex- 
tremely discontinuous and extensive, and they are largely 
confined to solitary islands. That they possess no special 
facilities for accidental dispersal across the ocean is evident, 
and yet it is held by some zoologists that their presence on 
islands, such as Christmas island for instance, can only be 
due to such a cause. A't any rate, the family exhibits all the 
signs of antiquity, and, in the absence of any positive evidence 
of accidentally distributed species, I am firmly convinced that 
they spread by the usual method of slow migration on land. 
Dr. Sarasin * places the dispersal of the family into pre- 
Cretaceous times, in spite of the fact that we possess no 
palaeontological evidence of their antiquity. All the same he 

* Sarasin, F., "Tierwelt von Ceylon," p. 75. 


is perfectly justified in doing so, because it can be shown from 
various other tests that the great continents must have been 
united by certain land connections, and that the present dis- 
tribution of the Typhlopidae can be readily explained if we 
assume that their dispersal took place at the same time. The 
presence of Typhlops tenuis in Guatemala and Mexico, on that 
account, points to an ancient union between these countries 
arid to a survival in them of a relict fauna during a long course 
of geological ages, in which the two countries may have been 

I suggested above (p. 131) that the rattlesnakes, belong- 
ing to the genus Crotalus, were of early Tertiary origin. 
The centre of dispersal lies in the south-western States of 
North America, all the species except one being confined 
to that continent. The black-tailed rattlesnake (Crotalus 
terrificus) is the only one which, according to Mr. Boulenger,* 
ranges from south-western North America to South America. 
Whether the South American species is identical with the 
North American one is a matter in which authorities disagree. 
All concur, however, in the view that the genus Crotalus 
reappears in South America in one or more forms, differing 
but slightly from those inhabiting Mexico. And these South 
American forms are not, as we might expect, confined to the 
west coast, but have even penetrated to eastern Brazil, if 
Mr. Boulenger's records are reliable. At the same time it 
seems as if their range in South America was very discon- 

The fresh-water fish fauna of Central America is described 
by Professor Eigenmann f as poor, the genera south of the 
Isthmus of Tehuantepec being practically all South American. 
The North American fauna is entirely distinct from that of 
tropical America, the former not having contributed a single 
element to the fresh-water fish fauna of South America. It 
is interesting to compare this view with those founded on 
other fresh-water groups such as the mollusks and crus- 

* Boulenger, G. A., "Catalogue of Snakes," Vol. III., p. 573. 
t Eigenmann, C. H., " Freshwater Fishes of South America," pp. 521 


I suggested above (p. 161) that the fresh- water mussels 
(Unionidae) probably effected their principal dispersal during 
the Mesozoic Era, and that this circumstance might account 
for the fact that we possess distinct proofs of a migration of 
species from North to South America. The great genus Unio 
has recently been subdivided by Dr. Simpson into numerous 
genera. One group of Unio (Plagiola), ranging from Mexico 
to the Mississippi drainage "basin, reappears southward in 
Nicaragua, another (Lampsilis) is known from Guatemala 
to Yucatan. Other groups of Unio are confined to South 
America. Unio-Tetraplodon occurs in Ecuador, having spread 
from there into the Amazon valley. Unio-Castalina lives in 
southern Brazil, Unio-Castaliella in Surinam and so forth. 
Finally Unio-Diplodon principally inhabits Chile, Argentina 
and Patagonia, while it reappears right across the Pacific 
in New Zealand and Australia. 

The range of these groups of Unio is apparently very com- 
plex in South America. 'Nevertheless, I quite concur with Dr. 
Simpson * in the belief that they all are the descendants of 
certain members of the family Unionidae, which wandered 
slowly from one river system into another, during the Triassic 
or some later Mesozoic Period, from North America to South 
America. To judge from the general distribution of the Unio- 
nidae in South America, they entered that continent from the 
west and only reached the eastern States subsequently. The 
group Unio-Hyria, as Dr. von Ihering f tells us, is nothing 
but a modified Unio, which has comparatively recently pene- 
trated from Guiana into Brazil. The most surprising fact 
which is so strongly brought out in that author's remarkable 
researches is, that, while these Unionidae live in company with 
other families of fresh-water mussels in eastern South 
America, in Central America, Ecuador, Peru and Chile, that 
is to say westward of the Andes, Unios alone occur. This con- 
firms the opinion I expressed several times in previous chap- 
ters, that the faunistic interchange between North and South 
America took place between the western portions of the two 

* Simpson, C. J., "Synopsis of the Najades," p. 507. 

t Ihering, H. von, " Archhelenis und Archinotis," p. 122. 


The fresh-water crabs (Potamonidae) must have invaded 
Central America from the south. The family is confined to 
southern Asia, southern Europe, Africa, South and Central 
America. Except for a few species in Mexico, fresh-water 
crabs are entirely absent from North America, nor do we 
possess any evidence of their ever having lived there.* A 
comparison with the range of Unio is, therefore, of particular 
interest. The two South American groups of fresh-water 1 
crabs apparently spread westward from eastern South 
America, that is to say in a direction opposed to that taken 
by the Unios. We need only consider the northern group 
which, to judge from its range, is much the oldest. .'Dr. 
Ortmann f distinguishes the three genera Kingsleya, Epilo- 
bocera 'and Pseudothelphusa. Kingsleya only occurs in 
Guiana, while Epilobocera is peculiar to the Greater 
Antilles. The third genus, Pseudothelphusa, ranges from 
the Amazon through Guiana, Venezuela and Colombia north- 
ward as far as Mexico, and south-westward to Ecuador, Peru 
and Bolivia. There are quite a number of endemic species of 
fresh-water crabs in Central America. Yet are we to conclude 
from this fact that a slow migration took place across the 
long isthmus since Pliocene times ? On the contrary, if, as 
Dr. Ortmann suggests, th genus Epilobocera arose in the 
West Indies from some ancestral Central American Pseudo- 
thelphusa, that event must have happened in much more 
remote times. It is customary to assume that the great mass 
of the South American fauna, including mammals, birds, 
reptiles, fishes and invertebrates all surged across th newly 
opened highway towards Mexico in the Pliocene Period. If 
Epilobocera succeeded subsequently in crossing from Central 
America on a land bridge to Cuba, Haiti and Portorico, how 
can we account for the fact that the existing faunas of Central 
America and the Greater Antilles do not show more affinity 
to one another than they actually do ? As compared with 
Central America the mammalian fauna of the West Indies 
is .strikingly distinct and poor in species. We have also to 
take into consideration that certain species of Pseudothelphusa 

* Eathbun, Mary J., " Freshwater Crabs of America." 
t Ortmann, A. E., "Distribution of Freshwater Decapods," pp. 306 


possess a remarkably discontinuous range. Pseudothelphusa 
colombiana inhabits the United States of Colombia. In the 
extreme western end of Panama, in the Chiriqui region, it 
occurs at a height of 4,000 feet, and a thousand miles north 
of this locality it is met with in Mexico. All along the great 
isthmus, on the -other hand, there are isolated localities in 
which distinct species are found. Both the ranges of Unio 
and Pseudothelphusa in Central America seem to me more 
in agreement with the theory of the existence to the west of 
Central America of an ancient continuous land surface uniting 
Colombia and certain parts only of Central America, while the 
rest of the latter was still largely submerged. 

One of the chief zoogeographical features in the range of the 
land and fresh-water mollusks of Central America consists in 
the intermingling of South and North American forms within 
the limits of this area, rather than in the endemic species, and 
in the relationship of the molluscan fauna to that of the West 
Indies.* The affinity of the Antilles with South America, 
however, is much more pronounced than with North America, 
if we exclude Mexico. The distinctly endemic elements of 
Central America and the West Indian islands appear to have 
invaded both North and South America to some extent. 
The family Helicinidae for instance, as I mentioned (p. 157), 
is probably of semi-tropical origin, its range indicating that it 
has reached America from the west, establishing itself 
primarily in Central America and the West Indies at a very 
remote time of geological history. The operculate land mol- 
lusks, to which the Helicinidae belong, are of great faunistic 
interest. Let us take for example the large family Cyclo- 
phoridae, which has its headquarters in southern Asia and 
southern Europe. In America the family is almost limited to 
Central and South America and the West Indian islands. 
Only in Mexico does it touch North American territory. As 
the geographical distribution of the family is now fairly well 
known, a great deal of interesting information can be derived 
from its study. The genus Cyrtotoma is peculiar to southern 
Mexico. Its nearest relations are Buckleyia of Colombia and 
Ecuador, and Crocidopoma of Haiti, Cuba and Jamaica. The 

* Martens, L. von, " Mollusks of Central America," p. xiv. 


larger genus Amphicyclotus has its centre of dispersal in 
Colombia and Ecuador. From there it pressed eastward 
through Venezuela into Guiana and across Trinidad to the 
island of Martinique, which must have been connected for 
some time with the southern mainland. In Central America 
the genus has a discontinuous range. A few species occur 
in Costa Rica, Guatemala and southern Mexico. Not a single 
species is known from the Greater Antilles. 

There is one very important feature in the fauna of Central 
America which I have scarcely dwelt on as yet, and that is its 
affinity with Europe. It is not at all striking. Yet it does 
exist. The large group Diplommatininae belonging to the 
family Cyclophoridae is almost entirely confined to southern 
Asia, the Pacific islands and Australia.* Only the single 
genus Adelopoma occurs in America. Its wide and extremely 
discontinuous range in the New World marks it as a very 
ancient immigrant, for it is likewise known from Argentina, 
Peru, Guatemala and Trinidad. Now the Guatemalan species 
(Adelopoma stolli) has its nearest relation in the Miocene 
beds ,of Oppeln in Silesia, for Professor 'Andreae f informs 
us that the fossil Adelopoma martensi occurring in these 
deposits is scarcely distinguishable from a species inhabit- 
ing Central America. Our first impulse on hearing of 
this extraordinary discovery is to attribute it to con- 
vergence or even misidentification. But the identifica- 
tion has been confirmed by Professor Boettger, one of the 
most eminent of European specialists, while Adelopoma 
stolli is not by any means the only Central American 
land mollusk that possesses European affinities. Many other 
instances will be alluded to in the next chapter (p. 265). 
Those who wish to interpret all cases of intimate relationship 
between American and European forms, as arising from a 
remote migration across a hypothetical Bering Strait land 
bridge, will find it difficult to reconcile this particular occur- 
rence with the fact that no Adelopomae inhabit any part of 
Asia or North America. 

Professor Stoll,J who made a special study of the mites of 

* Kobelt, W., "Cyclophoridae." 

t Andreae, A., " Binnenconchylienfauna von Oppeln," II., p. 23. 

t Stoll, 0., " Zoogeographie d. Wirbellosen," pp. 1920. 

L.A. S 


Guatemala during a residence there of several years, states 
that the mite-fauna of Central America is composed of the 
same genera as that of middle Europe. In some cases even 
the species are identical. It is true that most of these genera 
also inhabit Asia, though Linopodes, Haplophora, Nicoletiella 
and Uropoda have not yet been recorded from that continent. 
As some of these are known from the Oligocene Baltic amber 
deposits, it is probable that the genera date back to at least 
early Tertiary times. 

Of the nearly related spiders and harvestmen, Mr. Pickard- 
Cambridge * remarks, that while the majority of the Central 
American species are peculiar to the Nearctic and Neo- 
tropical regions, the genera are in many cases identical with 
those of the Mediterranean region, India, Australia or Africa. 
There are not now in existence any land connections, he con- 
cludes, which an arachnidal fauna could take advantage of 
in order to pass from Africa, Australia or Europe to North 
or South America, and we can only suggest that at some 
period or other these now widely separated regions must 
have been linked together by land which has long since dis- 

The various groups of Central American insects are as yet 
far from well known, and their relationships with the insects 
of other regions have been but imperfectly determined, not- 
withstanding the fact that Messrs. Salvin and Godman have 
published a series of fine volumes about them. The consensus 
of opinion of the writers, who have contributed to the 
" Biologia Centrali-Americana," is that Central America 
is .essentially a part of South America. The purely North 
American forms are few in number, their southern limit being 
as a rule in Mexico. This verdict agrees also with that of 
Dr. Calvert,f who in his recent researches on the dragon - 
flies (Odonata) of Mexico and Central America, comes to the 
conclusion that in the endemic as well as in the non-endemic 
species, the South American element is much the strongest. 

Before concluding this chapter, I might again emphasize 
the fact that the occurrences of northern genera and species 
in Central America are almost all discontinuous. I need only 

* Pickard- Cambridge, O., "Araneidaof Central America." 

t Calvert, P. P., " Odonate Fauna of Central America," p. 467. 


mention among northern plants, the oak (Quercus) and the 
plane (Platanus), among animals, the newt Spelerpes, the 
snapping turtles (Chelidridae) and the fresh- water mussels 
(Unionidae). Many other northern animals and plants have 
a similar disjointed range in the midst of an entirely alien 
southern fauna. These, and the northern groups that are 
altogether absent from Central America, but reappear much 
further south, are, I think, part of a very ancient invasion of 
South America. If it is true that the South American animals 
and plants which have taken almost complete possession of 
Central America, poured into the latter from the south in 
early Pliocene times, surely the northern groups must be 
immeasurably older! In their discontinuous range and fre- 
quent isolation on mountain tops they show every sign of 
antiquity, and it seems likely that they are being crowded 
out rapidly by their stronger southern rivals, which are better 
fitted to support the present climatic conditions of this region. 
Dr. Wallace * suggested, as I mentioned above, that the 
northern faunistic affinity of South America which is so cha- 
racteristic, especially of Chile, but which we find to some 
extent all along the Pacific coast, is mainly due to the Glacial 
Epoch. He fancied that a migration, aided by gales and hurri- 
canes during successive Glacial Epochs, when the mountain 
range of the Isthmus of Panama, if moderately increased in 
height, might have become adapted for the passage of northern 
forms, would explain all these extraordinary features of distri- 
bution. Dr. Wallace thought the resemblance only consisted 
in a few plants and insects. He had very little idea of the real 
extent and character of the northern element that has actually 
penetrated into Central and South America. It is unlikely 
that he would have supported these views had he known 
of the range of the newts, turtles, freshwater mussels and 
other groups. Moreover, we also have evidence of ancient 
migrations of mammals and other animals, as well as plants, 
in an inverse direction from the south to the north. It is 
perfectly certain, therefore, that long anterior to the southern 
invasion into the existing area of Central America, two 
streams of animals and plants passed between the great 

* Wallace, A. E., " Distribution of Animals," II., p. 45. 



continents, leaving certain traces of their transit in the 
more ancient portions of the country. And yet I believe, and 
shall produce ample evidence in support of my contention, 
that only certain fragments of Central America formed part 
of that land which long ago served as the highway between 
North and South America. This fact is not so readily re- 
vealed from a study of the Central American animals and 
plants. All we can gather from our present researches is 
that there are certain ancient elements in the fauna and flora 
of Central America exhibiting affinities with North America, 
Asia, Europe, Africa, the West Indies and South America, and 
that these older elements are being dispossessed or driven into 
the more inaccessible parts by the members of the new and 
most recent invasion from the south which traversed the newly 
formed Central American isthmus. It is believed that this 
must have commenced in Pliocene times. Of the two marine 
barriers which previously prevented this southern advance, 
one was in the neighbourhood of the Panama Canal, the other 
at the Isthmus of Tehuantepec, but to judge from the animals 
and plants of Central America, the former had already dis- 
appeared when the more northerly one was still in existence. 



NORTH and South America are to be regarded, according 
to Professor Suess,* as two essentially distinct land-masses, 
between which is interposed, as a third element, the area of 
Central America and the Antilles. This geological distinct- 
ness of Central America and the Antilles from the two neigh- 
bouring continents is scarcely recognisable in the fauna of the 
great isthmus. But the West Indies are comparable to a 
wedge driven in between two faunistically, more or less, in- 
dependent and distinct land masses. Almost everyone who 
has dealt with the fauna or flora of the West Indian islands 
has expressed his surprise at this fact. In position, says Dr. 
Walla ce,f the Antilles form an unbroken chain uniting North 
and South America, in a line parallel to the great Central 
American isthmus. Yet instead of exhibiting an intermixture! 
of the productions of Florida and Venezuela, they differ 
widely from both these countries, possessing in some groups 
a degree of speciality only to be found elsewhere in islands far 
removed from any continent. 

One other important feature which strikes the visitor to 
the islands is their extreme poverty in the higher groups of 
animal life. It is not that the Antilles are climatically un- 
favourable to animal life. On the contrary, they are excep- 
tionally favoured by nature to support a luxuriant and varied 
fauna and flora. Their temperature is high and uniform, 
there is an abundance of moisture, the soils are very fertile, 
while high mountains as well as gentle plains abound, at least 
in the larger islands. Cuba, the largest of them, exceeding 
Ireland in size, and being far more favourably situated, has 

* Suess, E., " Antlitz der Erde," I., p. 700. 

f Wallace, A. E., "Distribution of Animals," II., p. 61. 


not half-a-dozen species of land mammals, while Ireland pos- 
sesses nearly three times that number. There are only two 
ways in which we can account for this great deficiency in the 
higher animal life on the Antilles. We may suppose that 
either the islands have not been connected with the mainland 
since early Tertiary times, or if they have, that their fauna 
was largely destroyed since their isolation. In the latter 
case the apparent poverty of the fauna might be due to great 
destruction of animal life during a submergence of the land, 
and the consequent reduction of the habitable area. On the 
other hand, we should expect the relict fauna of the islands to 
exhibit marked affinities with that of the two great continents 
lying to the north and south respectively. However, as I 
have mentioned, the fauna on the whole is essentially dis- 
similar from that of North and South America. The problem 
of the origin of the West Indian fauna, therefore, like that of 
Centra] America, is much more complex than it would at first 
sight appear. I alluded to the apparent poverty of the fauna 
because, although it does appear very poor in the higher groups 
some of the lower forms of animals are represented by a large 
number of species. The land-snails, in fact, are extremely 
varied in character, and the abundance of species is one of the 
most remarkable features of the West Indian fauna. A study 
of their distribution and their relationship will probably 
give us a better insight into the origin of the fauna as a whole 
than the higher vertebrates, which are so poorly represented 
on the islands. 

Before describing the molluscan fauna of the Antilles, a 
few preliminary remarks on the islands may not be out of 
place. The islands as a whole form a natural breakwater or 
barrier between the Atlantic Ocean on the one hand, and the 
Gulf of Mexico and the Caribbean Sea on the other, which 
lie on the opposite side of it (Fig. 13). A mere glance at the 
map is sufficient to show that we have to deal with two distinct 
sets of islands. A series of large ones belong together, viz., 
Cuba, Haiti or San Domingo, Jamaica and Portorico, with 
mountain crests running in an east-westward direction, while 
a chain o,f the much smaller, Anguilla, Guadeloupe, Dominica, 
Martinique, St. Vincent, Barbados and others, placed in north 
and southward position, constitute quite an independent 


group. The former are always spoken of as the ''Greater 
Antilles," and the latter as the "Lesser Antilles." With 
Professor Suess * we may look upon the mountain ranges of 
Yucatan and Guatemala, which trend in a west-easterly direc- 
tion, as the western continuations of the mountain system of 
the Greater Antilles. The latter, as well as a few of the 
northern Lesser Antilles, are composed of sedimentary rooks 
of Mesozoic and Cainozoic, possibly even of Palaeozoic age, 
while many of the remaining smaller islands, which cluster 
together in a concave arc, seem to be of comparatively 
recent volcanic origin. The Bahamas, and some of the 
more southerly flat islands, including part of Barbados, are 
apparently of young Tertiary age. The whole of the main 
series of the Antilles, from Cuba through Jamaica, Haiti and 
Portorico to Barbados, is composed of similar rocks. Granite, 
older eruptive rocks, serpentine, glauconitic sandstone and 
cretaceous limestone, form the visible remnants of a once 
connected mountain range. Westward the latter divides into 
several branches. One of them passes from southern Haiti 
through Jamaica to Honduras, another by way of Cuba to 

There is some evidence, according to Professor Hill, that 
the east coast of North America lay far eastward of its pre- 
sent site in pre- Cretaceous times, whereas some faunistic 
facts point to a continuation of this condition until the 
Tertiary Era. The Pacific marine fauna transgressed east- 
ward during the Jurassic Period, probably across the Mexican 
plateau, fossils of Pacific type having been found in western 
Cuba. This implies that the barrier separating the Atlantic 
from the Pacific in, those remote times must have been situated 
to the east of Cuba. Professor Hill f argues that the chain 
of low-lying islands between Florida and north-eastern South 
America represent the remnants of this ancient Jura-Cre- 
taceous isthmus between the two great continents. Whether 
such a land bridge existed is difficult to determine from 
faunistic evidence, but the Atlantic waters seem to have 
entered the Caribbean Sea in Lower Cretaceous times. 
During part of the Eocene and Oligocene Periods, extensive 

* Suess, E., " Antlitz der Erde," I., pp. 700707. 
t Hill, E. T., " Geology of Jamaica," pp. 200-216. 


subsidences drowned the Antilles to such an extent, accord- 
ing to Professor Hill, that only the higher summits of Cuba, 
Haiti and Jamaica remained above sea-level as small islands. 
The West Indian islands were subsequently raised into a large 
continuous and connected land. In late Miocene and Pliocene 
times the gradual and final dismemberment of the Antillean 
lands took place. Still more recently a further elevation 
occurred, not sufficient, however, to establish a united 
Antillean continent. Whether Professor Schuchert* supports 
Professor Hill's hypothesis of a wide land connection be- 
tween Florida and Venezuela in late Jurassic times is not 
clearly indicated in his maps. But during the Cretaceous 
Period all the West Indian islands except the Bahamas are 
represented as being entirely submerged. In Eocene times 
the greater part of Cuba was above sea-level. In the succeed- 
ing Oligocene Period all the islands, except the Bahamas, 
once more disappeared. Thenceforth all the Greater Antilles 
retained their present outlines. Only during the Plio- 
cene Period was there a land connection between Cuba and 
Yucatan. All these writers thus concur in the view that some 
time during the earlier part of the Tertiary Era there was a 
very profound and widespread subsidence of almost the whole 
of the Antillean area. Yet, according to Professor Schuchert, 
the Bahamas, or some land area in the position of the 
Bahamas, if I correctly interpret his maps, remained above 
sea-level practically from the earliest Palaeozoic ages to the 
present day. The idea that there was once a land connection 
between North and South America along the chain of the 
Lesser Antilles, Cuba, the Bahamas and Florida is also 
advocated by Professor Gregory, f though he admits that 
the area of the Windward islands was submerged at the period 
when the oceanic deposits of Barbados were laid down. There 
is no adequate evidence, he thinks, to show that there was 
more land at any subsequent time in this region than there 
is at present. 

Now as for the light thrown on these various problems 
by a study of the geographical distribution of the West Indian 

* Schuchert, Charles, " Paleogeography of North America," Maps 89 
t Gregory, J. W., " Geology of the West Indies," p. 305. 


land mollusks, it may b urged, in the first instance, that the 
value of the evidence is greatly impaired by the possibility 
of occasional or accidental dispersal. That a certain per- 
centage of the fauna of an island is due to importation by 
human agency is, I think, well established. That a small 
proportion of the West Indian fauna may possibly have been 
carried from island to island and from the mainland by other 
accidental means is likewise possible. But I quite concur 
with Dr. Simpson in his belief that we are not justified in 
explaining the whole distribution of the terrestrial mollusks in 
the West Indian islands by such an hypothesis. The very 
nature of the West Indian fauna, and its distinctness from 
that of the neighbouring continents, precludes the assumption 
of any extensive accidental dispersal. Hence we may take for 
granted the correctness of the theory that the main mass of 
these mollusks have migrated from island to island when the 
latter formed a united land surface. This belief is vouched 
for, moreover, by the fact that the study of practically all 
other groups of West Indian animals brings us to precisely 
the same conclusion. In their general agreement with the 
results arrived at from a geological study, all these groups 
tend to show that the study of geographical distribution is a 
science of profound importance as an indicator of former 
changes of land and water. 

The facts adduced by Dr. Simpson, that the operculate 
species form so large a proportion of the Antillean land- 
snail fauna, that a majority of the genera is found on two 
or more of the islands and the mainland, while nearly every 
species is absolutely restricted to a single island, is a strong 
testimony in favour of a former general land connection in 
this area. Dr. Simpson * very carefully compared the mol- 
luscan fauna of the various islands with one another and with 
that of the mainland, and bases his conception of the geolo- 
gical history of the Antilles mainly on the results so derived. 
He believes a considerable portion of the species inhabiting 
the Greater Antilles to be ancient, and to have developed on 
the islands where they are now found. Probably some time 

* Simpson, C. T., " Distribution of Mollusks in West Indies," pp. 447 



during the Eocene Period the Greater Antilles were at a 
higher level, so that the islands were united with one another 
and with Central America. This resulted in an exchange of 
species between the two regions. A land connection is also 
indicated between Cuba, the Bahamas and Florida. At this 
time, he thinks, the more northern isles of the Lesser Antilles 
were not yet elevated above the sea, or, if so, they have since 
probably been submerged. After this period of elevation 
there followed one of general subsidence. During it Jamaica 
was the first island to be separated, then followed Cuba, and 
afterwards Haiti and Portorico. The connection between 
the Antilles and the mainland was broken, while the subsi- 
dence continued until only the summits of the mountains of 
the four Greater Antillean Islands remained above water. 
Eventually there was another period of elevation which lasted, 
no doubt, until the present time. The Bahamas gradually 
emerged and were populated by forms drifted from Cuba and 
Haiti. In this last conclusion Dr. Simpson unfortunately 
fails to appreciate the full value of his own studies. If all 
the snails, amphibians and reptiles known to inhabit the 
Bahamas and many of them are peculiar to these islands 
had reached them by accidental dispersal, this mode of con- 
veyance must be of enormous importance. In such a case 
I think we should scarcely be justified in basing our theories 
of the geological history of the Antilles on the geographical 
distribution of animals. 

Now among the families of snails alluded to in Dr. Simp- 
son's paper as being abundant in the West Indies, there are 
some that have received special attention by American 
malacologists. One of these, the Urocoptidae, or Cylindrel- 
lidae as they were formerly called, comprises small snails 
with an elegant fusiform shell composed of many narrow 
whorls. They are found in the Antilles, in southern Florida, 
along the northern coasts of South America, in Central 
America and in Mexico, as well as the adjacent parts of the 
United States. The centre of distribution thus lies in the 
West Indies. 

Since it has been proved by fossil evidence in Jamaica that 
numerous sub-generic groups of land snails, in essentially 
their modern forms, were established before the close of the 


Oligocene Period, generic differentiation probably dates from 
a much earlier time. Indeed, Dr. Pilsbry * is of opinion that 
the first radiation of the Antillean group of the Urocoptidae 
may have occurred on a Mesozoic Antillean land area. The 
succeeding Eocene depression, he thinks, isolated various 
branches of the existing stocks, western Cuba being pro- 
bably the first fragment to be dismembered. It was probably 
not until near the close of the Tertiary that continuity of 
land was restored with east Cuba. Haiti and Jamaica would 
seem to have remained united after both western and eastern 
Cuba had seceded. Finally, these islands were widely 
separated by the subsidence culminating at the end of the 
Eocene, or in the beginning of the Oligocene Period. This 
depression was again followed by an elevation in later Oligo- 
cene times, and it is likely that there was a transitory connec- 
tion between Jamaica and Haiti. Between the latter and 
Cuba the land connection probably lasted longer, thus pro- 
ducing the homogeneous distribution of several groups. It 
is likely, says Dr. Pilsbry, that during this mid-Oligocene 
elevation, the Haitian land included Portorico, the Virgin 
islands and the islands of the Anguilla bank, that is to say, 
some of the northern group of the Lesser Antilles. Dr. 
Pilsbry argues that the presence of large fossil mammals of 
South American type (Amblyrhiza and Loxomylus) in Plio- 
cene deposits of Anguilla demonstrates that the whole Carib- 
bean chain of islands was elevated into a ridge connected 
with South America during the Pliocene Period. He likewise 
expresses the opinion that the genus Brachypodella, one of 
the Urocoptidae, extended its range westward to Yucatan. 
Nevertheless, he contends that there is but scanty evidence of 
any direct land connection between the Greater Antilles and 
the mainland of Central America during the whole of Tertiary 

Thus, while differing from Dr. Simpson on several minor 
points, Dr. Pilsbry's careful researches confirm his view, 
and that of many geologists, that originally there was a large 
area of land of which the Antilles are the last remnants, and 
that some time during the Tertiary Era almost the whole of 

* Pilsbry, H., "Manual of Conchology," XVI., pp. xx. xxiv. 


this old land was submerged having since gradually regained 
its present position. 

The great age of the West Indian fauna and the inter -re- 
lationship between the islands and the mainland is well exem- 
plified by the ancient family of operculate snails the 
Cyclophoridae.* The genus Neocyclotus inhabits principally 
northern South America and the Antilles. From this 
apparently very old centre of dispersal some members of 
the genus have pushed southward as far as Peru in the west 
and Rio de Janeiro in the east. A few have entered Central 
America. One distinct group (Plectocyclotus) has no less 
than thirty-two species in Jamaica and only one in Portorico. 
Another genus (Crocidopoma) is entirely confined to Jamaica, 
Haiti and eastern Cuba. This indicates strikingly the re- 
lationship of the three Great Antilles and their distinctness 
'from western Cuba, which was already pointed out by Dr. 
Pilsbry, while geologists maintain that western Cuba was 
submerged quite independently from the remainder of the 
islands. It also illustrates the extreme slowness with which 
the dispersal of these mollusks takes place. 

Still more instructive is the whole group to which 
Crocidopoma belongs. With Cyrtotoma, Amphicyclotus and 
Buckleya, it forms, as already mentioned (p. 256), a group 
of closely related genera of operculate snails. I alluded 
also to the fact that three of them had a discontinuous range 
in Central America, and that Amphicyclotus had apparently 
travelled eastward from Ecuador, invading Venezuela and 
Guiana, and had thence passed into the islands of Martinique, 
Guadeloupe and Dominica, when the latter were connected 
with one another, and with the mainland. It might be 
urged that accidental dispersal is responsible for their 
presence on these islands. 'But we have no reason for such a 
supposition, because the species occurring on the islands are 
quite distinct from one another and from those of Venezuela. 
Some evidence is afforded by these snails for the belief that 
the Lesser Antilles are remnants of older land which ex- 
tended northward from Venezuela, although all the visible 
parts of the islands are covered by modern volcanic deposits. 

* Kobelt, W., " Cyclophoridae." 


We are top apt to argue that the fauna of an island covered 
by recent volcanic deposits must necessarily have been derived 
by occasional means of dispersal. I need only cite the case of 
the Galapagos islands, which are entirely volcanic, and have 
seemingly risen from the floor of the ocean. Nevertheless, 
it can be demonstrated from a faunistic point of view, as I 
shall show later on, that they have once formed part of an 
ancient continuous land surface. 

Still another group of operculate land snails contains the 
two genera Megalomastoma and Tomocyclus. The centre 
of dispersal is Cuba, chiefly the western part of the island. 
From there Megalomastoma reached Haiti, Portorico and the 
Virgin islands, which lie close to the latter. Tomocyclus 
inhabits only southern Mexico and Guatemala. Thus it 
seems probable that Cuba, Guatemala, and southern Mexico 
were connected with one another by land in very remote 

Let us take as another example, that of the ancient and 
large family Bulimulidae. In another chapter I have dwelt 
on the great age of this family of snails _(p. 209) and its 
general range. It contains mostly large, ponderous snails 
with somewhat conical shells. Anatomically they are re- 
lated to the Helicidae. Although represented by a great 
many species, only a few genera enter the West Indies. One 
of these (Plekocheilus) inhabits almost exclusively Guiana, 
Venezuela, Colombia, Ecuador, Peru and Bolivia, that is to 
say the northern and western States of South America, where 
the different species are often found at great heights in the 
mountains. Only two species enter the West Indies, viz., 
P. aurissileni, which is peculiar to St. Vincent, and P. aula- 
costylus, which is only met with in the island of St. Lucia. 
These islands are two of the most southern group of the 
Lesser Antilles. The ancestors of the two species of Pleko- 
cheilus have probably entered these islands when the latter 
were connected with one another and with the mainland of 
Venezuela, and, as the species are very distinct from one 
another, this could not have happened within very recent 
geological times. 

The species of the genus Bulimulus, as I have already had 
occasion to state, are exceedingly difficult to discriminate 


from one another. Dr. Pilsbry divided them roughly into 
three groups, only one of which inhabits the Antilles. To 
trace the relationship of the various species to one another 
seems to he a task even beyond the powers of this distin- 
guished American conchologist. The minor sections being 
arranged geographically in Dr. Pilsbry's work, it does not 
enable us to draw any conclusions as to their former dis- 
persals. One interesting fact, however, has been brought to 
light, which proves not only the antiquity of this family but 
the .extraordinary persistency of specific characters among 
some of its members. The Oligocene Bulimulus americanus 
of Florida is practically identical with B. ridleyi, still living 
on the island of Fernando de Noronha. 

The genus Drymaeus (OtostomusJ, which is as difficult to 
classify as Bulimulus, has mainly a continental range. 
Only a few species live in the West Indies. Among these 
we likewise notice a remarkable persistency of specific 
characters. Drymaeus dormani, for instance, occurs in 
southern Florida, while the closely allied D. dominicus is 
resident in Haiti, Cuba, Florida, Yucatan, Nicaragua and 
Mexico. At the first thought we might feel inclined to attri- 
bute such a strikingly discontinuous range to accidental dis- 
tribution ; but a careful study of the whole family has im- 
pressed me with the conviction that we have to deal with a 
set of very ancient and very persistent types of mollusks. 

Even less satisfactory to identify than Bulimulus and 
Drymaeus are the Orthalicinae, another large group of Buli- 
mulidae. Dr. Strebel * has recently undertaken their revision, 
utilising several characters which had hitherto not been em- 
ployed in the discrimination of the species. It is an attempt, 
at least, to trace the complex relationship of the innumerable 
closely allied forms, although the author does not seem to 
realize the great antiquity of the group. His arguments in 
favour of wholesale accidental dispersal, even from western 
Mexico to the Antilles, are quite at variance with the lesson 
the study of the West Indian fauna has taught us. The species 
of Simpulopsis are mostly South American, but a small aber- 
rant group with smooth apical whorls is confined to Portorico, 

* Strebel, H., " Eevision der Orthalicinen." 


Haiti and Mexico. The genus Gaeotis is quite confined to 
Portorico, while Amphibulima is limited to the Lesser 

All these studies yield the same fundamental results, 
namely, the presence within the Antillean area of an ex- 
tremely ancient stock distantly related to that of Central 
America and northern South America. After this fauna 
had developed to some extent, a widespread destruction 
apparently took place, due probably to submergence, fol- 
lowed by a reimmigration from the south and west. A 
more precise knowledge, however, of the minor physical 
changes which the West Indian islands have undergone 
can be acquired when we compare the results derived 
from the sedentary or slowly moving mollusks with those 
drawn from the more active members of the fauna. But 
there is one more element of importance in the Antillean 
molluscan fauna which has not yet been considered. It 
is especially in view of the remarks I made in the ninth 
chapter, on the striking relationship of the south-western 
American fauna to that of Europe, that I wish to direct par- 
ticular attention to the European affinities of the Antillean 

I think it was Dr. Kobelt f who first animadverted on the 
resemblance between the European molluscan fauna and that 
of ,the Antilles and Central America. At first sight, as he 
remarks, the two faunas seem to be as distinct as any disciple 
of the principle of multiple centres of creation could wish. 
After a careful study, however, we certainly perceive distinct 
traces of relationship. The carnivorous snail Glandina, which 
is represented in the Mediterranean region of Europe by a 
single species, has its headquarters in the Antilles and in 
the surrounding States. It cannot be looked upon as a recent 
introduction to Europe, because its discontinuous range be- 
between the Caucasus and Algeria is altogether opposed to any 
such supposition. Several species of Glandina, moreover, 
occur in French and English Eocene, Oligocene and Miocene 
deposits, some of them being almost identical with still 

* Pilsbry, H. A., "Manual of Conchology," Yols. X. XII. 
f Kobelt, W., " Verhaltniss d. Europaischen Landmolluskenfauna, &c.," 
p. 145148. 


existing American forms. The operculate genera Tudorella 
and Leonia of the Mediterranean region are certainly related 
to West Indian forms, though not so closely as was formerly, 
believed. These and other considerations induced Dr. Kobelt 
to postulate a land connection between Europe and North 
America, which was only severed in Miocene times. The 
German Miocene genus Subulina, the large Glandinae and the 
early Tertiary European Oleacinae, are considered by Pro- 
fessor Boettger * as the nearest relations or direct ancestors 
of Central America or West Indian genera and species. The 
same authority also alludes to the American facies of the 
Tertiary flora of Europe, and concludes that a land bridge 
right across the Atlantic Ocean existed up to early Miocene 
times. Finally Professor Andreaef supports the same theory, 
on account of the occurrence of the West Indian Boltenia, 
Strobilus and Pleurodonte in the Miocene deposits of Silesia 
in Germany. And he was the first to definitely fix the posi- 
tion of the land connection as one uniting western Europe 
with the Antillean area. 

Still other features of relationship between these regions 
remain to be considered. Long ago Mr. Woodward J pointed 
out that the presence of the European genus Clausilia in the 
West Indies and in northern South America (see Fig. 19) 
implied the existence of a former more direct land way across 
the Atlantic than would be afforded by the land connection 
which was supposed to have once united the boreal regions 
of Europe and North America. Since that time others have 
repeated his assertion. Clausilia is now known to have lived 
in Europe since Cretaceous times, and has probably originated 
there. It is partial to high altitudes, large numbers of species 
being found in the Alps, the Dalmatian mountains and the 
Caucasus. Only a single European species resembles the 
American group (Nenia), namely Clausilia pauli of the 
western Pyrenees, and it is so closely related to the members 
of that group that both Mr. Bourguignat and Mr. Locard, 
two French conchologists of the " nouvelle ecole," failed to 
find any satisfactory difference between them. It is quite 

* Boettger, 0., " Verwandschaftsbeziehungen d. Helix Arten," p. 116. 
t Andreae, A., " Binnenconchylienfauna d. Miocans," II., p. 31. 
\ Woodward, S. P., " Manual of the Mollusca," p. 112. 


true that another group of Clausilia, known as Garnieria, from 
tropical and eastern Asia also approaches the American group 
closely. Yet the fact that no Clausilia, either fossil or recent, 
has ever been discovered in northern Asia or North America, 
although many species, as I remarked, thrive in high altitudes 
and cold climates, speaks strongly against the supposition of 
the ancestors of the West Indian and South American stock 
having wandered through Asia and North America to these 
regions. The geographical distribution of Clausilia thus 
offers one of the most cogent arguments in favour of a direct 
land bridge between the Mediterranean and the Antillean 
regions. Opponents of this view may urge that Clausilia is 
merely represented by a single species in the West Indies, 
while none are known from Central America. We. may explain 
this curious circumstance by the fact that the West Indies 
were submerged probably before the newly immigrated Clau- 
silias had time to gain possession of the higher eminences, 
so that most of them would have been destroyed. In Central 
America fewer traces of European affinity have been detected 
than in western North America or western South America, 
because in the latter 1 regions the faunas remained more or less 
isolated for long periods, while the great rush of South 
American invaders, combined with climatic changes, swept 
all before them in Central America. Why Clausilia has suc- 
ceeded in reaching western South America without attaining 
western North America will be discussed in the neixt chapter 
but one. 

Dr. Simroth suggests that the ancestors of the American 
Bulimulidae, at least the group of Orthalicinae, may be of 
European origin, while I venture to think that the curious 
Bumina decollata, which is supposed to be a human importa- 
tion in Cuba, may be indigenous there, since it is now known 
to have lived unchanged in the Mediterranean region since 
Oligocene times. In view of my remarks in previous chapters 
on the relationship of the North American snake Tropido- 
notus, of the crayfish Potamobius, of the slugs belonging 
to the family Arionidae, of the glass snake, of all that remark- 
ably European assemblage of animals in the south western 
States, of the snail Adelopoma in Guatemala and numerous 
other instances, such; as the range of the flamingoes, it need 

L.A. T 


not 'be surprising that I am a strong adherent of the theory 
just stated, that a land bridge existed right across the Atlantic 
between the Mediterranean and the Antillean regions, and 
that the European element of the fauna made use of it in 
passing to America (Fig. 14). This land connection, of 
course, was quite independent of the one I described (p. 13) 
as having once joined Labrador and Scotland by way of 
Greenland. The latter may possibly have come into existence 
when the other had already crumbled away. At any rate, the 
two are quite distinct as to age and position. 

Before I had an opportunity of making this more thorough 
study of the North American fauna, I was under the impres- 
sion that the " Southern Atlantis," as we may call this land 
connection, joined Africa with South America, and that there 
was no other land bridge across the mid- Atlantic.* I am still 
an advocate, as I shall explain more fully later on, of what 
Dr. von Ihering calls " Archhelenis," the hypothetical con- 
tinent of the southern part of the Atlantic Ocean. But I 
maintain that a more northerly land bridge likewise existed, 
and that the two were completely separated by a wide ocean. 

The disciples of Dr. Wallace will exclaim, " What about the 
permanence of ocean basins, a theory which receives such 
weighty support from some of the most eminent geologists 
of the day ? " This question of the permanence of ocean 
basins, and we may say of continental areas too, really lies at 
the root of most of our inquiries into the past changes of the 
earth and its animal inhabitants. The internal characters 
of the rocks we see around us, remarks Sir Archibald Geikie,f 
point unmistakably to deposition in comparatively shallow 
water. "Their abundant . intercalations of fine and coarse 
material, their constant variety of mineral composition, their 
sun-cracks, ripple-marks, rain-pittings and worm-tracks, 
their numerous unconformabilities and traces of terrestrial 
surfaces, together with the prevalent facies of their organic 
contents, combine to demonstrate that the main mass of the 
sedimentary rocks of the earth's crust was accumulated close 
to land, and that no trace of really abysmal deposits is to be 
found among them." From these considerations, says Sir 

* Scharff, E. F., " Atlantis Problem," p. 279. 
t Geikie, A., "Text Book of Geology," p. 911. 


Archibald Geikie, we are led to the conclusion that the pre- 
sent continental areas must have been terrestrial regions 
of the earth's surface from a remote geological period. 
Subject to repeated oscillations, continues Sir Archibald 
Geikie, so that one tract after another has disappeared 
and reappeared from beneath the sea, the continents s 
though constantly varying in shape and size, have yeit 
maintained their individuality. So far, I think, most 
geologists will agree with Sir Archibald Geikie. It is his 
inference, that the existing ocean basins have probably always 
been the great depressions of the earth's surface, which has 
not met with such general approval. Dr. Wallace supports 
Sir Archibald Geikie's view on the grounds, not only of the 
enormous depths and great extent of the oceans, and of the 
circumstance that the deposits now forming in them are 
distinct from anything found upon the land surface, but also 
owing to the supposed extraordinary fact that the countless 
islands scattered over their whole area (with one or two ex- 
ceptions) never contain any Palaeozoic or Secondary rocks, 
that is, have not preserved any fragments of ancient con- 
tinents, nor of the deposits which must have resulted from 
their denudation during the whole period of their existence ! 
The exceptions alluded to by Dr. Wallace* are New Zealand 
and the Seychelles islands, both situated near to continents 
and, according to the same writer, not really oceanic. "The 
vast areas of the Atlantic, Pacific, Indian and Southern 
Oceans are thus left almost without a solitary relict of the 
great islands or continents which some naturalists believe to 
have sunk beneath the waves of these oceans." Thus writes 
Dr. Wallace. Another argument in favour of the permanence 
of ocean basins has recently been brought forward by Pro- 
fessor Joly.f It is based on the facts of solvent denudation. 
He regards the sodium in the O'Cean as the key to the history 
of solvent denudation, arguing that it was derived from the 
igneous rocks of the earth by the processes of weathering and 
solution progressing throughout geological time. He shows 
that the quantity of oceanic sodium agrees with the sediments 

* Wallace, A. E., " Island Life," p. 105. 
t Joly, J., "Radioactivity and Geology," pp. 127131. 



as we find them upon the existing* continents, and concludes 
that there cannot be yet other continents with their own 
burdens of sediment hidden beneath the ocean. That former 
continents of any antiquity or magnitude are not hidden 
beneath the waves, says Professor Joly, seems certain, 
unless the estimates of sediments are quite erroneous. 
Whether the estimates are correct is a question which must 
be left to authorities in physical chemistry ; but it has been 
argued by Professor Carthaus * that the waters of the 
ocean, as well as those of continents, were originally rich in 
sodium chloride, and that fresh-water organisms only came 
into existence comparatively recently, that is to say in late 
Mesozoic times. When we consider the enormous area of 
North America that was under water in Cretaceous times for 
instance, less than an equivalent strip of land in the shape of a 
trans -Atlantic land bridge would be all that is required for our 
purpose. We need not call it a continent. 

As for the arguments in favour of the permanence of con- 
tinents and ocean basins raised by Sir Archibald Geikie and 
Dr. Wallace, they are based on the following facts and 
assumptions, viz., great ocean depths, absence of abysmal 
rocks on present land surfaces, and absence of older Mesozoic 
or Palaeozoic rocks (with one or two exceptions) on oceanic 
islands. It has been shown, however, by Professor Suess that 
great depressions on the surface of the earth's crust are not 
necessarily old or permanent. Quite near the south-west 
coast of Asia Minor, and close to the mighty Ak Dagh (10,000 
feet high), a depth of over 10,000 feet has been recorded. This 
depth is all the more remarkable when we consider that fresh- 
water Pliocene beds of the mainland are continued across to 
the neighbouring island of Ehodes, thus showing that the 
latter was, until such a recent geological period as the Plio- 
cene, still joined to the continent. Altogether Professor 
Suessf inclines to the view that geological evidence does not 
prove, nor even point to a permanence of the great depths, 
at least in the oceans of the Atlantic type. The next point 
which has been raised in favour of the view of the permanence 

* Carthaus, E., " Klimatische Verhaltnisse der Geologischen Vorzeit." 
t Suess, E., " Are Great Ocean Depths Permanent ? " pp. 182186. 


of oceans and continents is the supposed absence of abysmal 
deposits on any existing land surface. The discovery, how- 
ever, of true deep-sea ooze in Barbados and Cuba shows that 
this argument is no longer valid.* There only remains one 
other argument in favour of this theory, and that is the 
structure of the oceanic islands. With two exceptions, re- 
marks Pr. Wallace, they do not contain any Mesozoic or 
Palaeozoic rocks, being mostly volcanic. Hence he concludes 
that these islands must be of , modern origin. How illusive 
this conception is may be gathered from Dr. Blanford'sf re- 
marks on this subject: " If Africa, south of the Atlas, sub- 
sided 2,000 fathoms, what would remain above water? So 
far as ,our present knowledge goes, the remaining islands 
would consist of four volcanic peaks, th Camaroons, Mount 
Kenia, Kilimanjaro, and Stanley's last discovery, Kuwenzori, 
together with an island, or more than one, containing part of 
the Abyssinian tableland, which, like the others, would be 
composed of volcanic rocks, but, unlike them, would consist of 
horizontal or nearly horizontal lava flows, probably of 
of Mesozoic age. In southern Africa, too, the peaks of the 
Stormberg and Drakensberg, though not rising, or scarcely 
rising above 10,000 feet, are the highest in the country and 
consist of volcanic rocks. The same is the case with the 
highest peaks in Madagascar, in Mexico, in the Caucasus, in 
the Elbruz chain south of the Caspian, and in many other 
parts of the world ; though the case of Africa is perhaps 
the most remarkable." 

The question whether all the numerous oceanic islands that 
are scattered about the Pacific and Atlantic Oceans are really 
modern volcanic products or of recent organic origin, has 
also received some attention recently. According to Dr. 
Arldt,J Archaean rocks have been noticed on the Solomon 
islands, gneiss and allied rocks occur on the New Hebrides 
and New Caledonia, while even from the Marquesas have 
been recorded ancient rocks. 

So far I have discussed only a very small portion of the 
purely zoological aspect of the trans -Atlantic land connection 

* Gregory, J. W., " Geology of the West Indies," p. 307. 
t Blanford, W. T., " Anniversary Address," pp. 3435. 
J Arldt, Th., " Entwicklung der Kontinente," pp. 457458. 


problem. This problem has been approached from many 
other points of view. On the evidence of the fossil marine 
mollusks of the West Indian and the Mediterranean regions, 
Mr. Guppy* concluded that a migration must have taken place 
right across the Atlantic along an ancient shore-line. More 
recently, Professor Gregoryf dwelt upon the intimate affini- 
ties that exist between the fossil sea-urchins of the West 
Indian and Mediterranean areas, and urged that it could only 
be explained by the assumption of a belt of shallow water 
across the Central Atlantic in, at latest, Miocene times. A 
few years later he adduced evidence from the fossil corals of 
Barbados, that the West Indian fauna is only a fragment 
of that of the Mediterranean Miocene, having received 
nothing from the Pacific. That this fauna did not follow 
along the shores of the North Atlantic basin is shown 
by its absence from the northern Miocene of Europe and 
America.} Mr. Guppy has lately renewed the discussion of 
this subject and once more affirmed his adherence to the 
theory he expressed long ago, which has meanwhile received 
so much additional support. Even the recent marine fauna of 
the Antilles is intimately related to that of the Mediterranean. 
Some of the faunistic marine affinities between the two re- 
gions referred to might just as well have been produced by a 
dispersal along a land bridge between Africa and South 
America. Such, for instance, is the occurrence in early Ter- 
tiary deposits of the aquatic snake Pterosphenus in Egypt 
and Alabama. Dr. Andrews, || indeed, thought it yielded an 
argument in favour of the more southern land connection. 

Let us now examine what light the recent marine mammals 
inhabiting the Caribbean Sea and the Gulf of Mexico throw on 
the problem. On the south-east coast of Florida we meet with 
one of the most curious of American mammals. With 
its seal-like head and flattened tail it is at once recognised as 
something quite distinct from other marine creatures. The 

* Guppy, E. J. L., " West Indian Geology," p. 501. 
t Gregory, J. W., " American and European Echinoid faunas," 
pp. 101108. 

t Gregory, J. W., " Geology of the West Indies," p. 307. 

Guppy, E. J. L., " Geological Connexions of the Caribbean Eegion." 

II Andrews, C. W., " Tertiary Vertebrates of the Fayum," p. xxv. 


manatee (Trichechus manatus), as it is called, does not in- 
habit the open ocean. It frequents shallow bays and lagoons, 
where it browses peaceably on seaweeds, just as cattle graze 
on land. This northern manatee lives also near the coasts 
of Cuba, Haiti, Jamaica and other islands, as well an along the 
shores of Central America and northern and eastern South 
America. Curiously enough, a second species (Trichechus 
inunguis) seems to be confined to the upper reaches of the 
Orinoco and the river Amazon. Still more remarkable is 
the fact that a third species (Trichechus senegalensis) is 
confined to the coasts and rivers of West Africa, for 
since the open ocean is to the manatees just as much 1 
a barrier to migration as it is to terrestrial mammals, the 
distribution of these manatees implies the existence of a 
former shore-line across the Atlantic. It is quite true thait 
in early Tertiary times manatees have lived much further 
north than they do now, but the European ones, at any rate, 
belonged to different genera from those now living. We 
possess no evidence, therefore, for the supposition that the 
ancestors of the American species passed along the eastern 
shores of North America and crossed to northern Europe 
along the ancient Greenland-Iceland land bridge, thus even- 
tually reaching Africa. Another theory, even less probable 
I think, is that suggested by Professor Osborn.* He thought 
a migration might have taken place from Africa by ; way of 
the Pacific coasts of Asia and North America, the ancestors 
of the West Indian manatees entering the Atlantic through a 
strait, which is supposed to have connected that ocean with 
the Pacific, in mid-Tertiary times. He considers this cir- 
cuitous route a more probable one than the trans -Atlantic one. 
Yet he does not clearly explain how the close relationship 
between the West African and eastern South American forms 
was brought about. 

This, however, by no mean completes all the evidence de- 
rived from the manatees in favour of the theory of a trans- 
Atlantic land -bridge. Dr. Dilg f has pointed out that the 
molar teeth of the adult living manatees resemble those of 

* Osborn, H. F., " Age of Mammals," pp. 493494. 

t Dilg, Carl, " Morphologic des Schadels bei Manatus," p. 139. 


the Eocene Prorastomus, which must be looked upon as the 
ancestor of Trichechus. I mentioned before that Proras- 
tomus has been discovered in the Eocene of Jamaica. The 
teeth of another form (Prorastomus veronensis) are 1m own 
from Italy. Thus the affinity still existing between the South 
American and West African forms was apparently fore- 
shadowed already during the Eocene Period by the relation- 
ship of the two species of Prorastomus, the ancestors of the v 
modern manantees. Dr. Smith Woodward, however, informs 
me that the relationship of these species is too uncertain to be 
used as evidence in favoutfof an Eocene land bridge. 

All the seals inhabiting the North Atlantic, both on the 
European and North American side, belong to the genera 
Halichoerus or Phoca. As soon as we enter the Antillean 
region these genera disappear, their place being taken by the 
genus Monachus. On the opposite shores of Europe it is just 
the same. As far south as Portugal we still find the common 
seals, but as we enter the Mediterranean we again meet with 
the same genus Monachus. The Antillean form (Monachus 
tropicalis), like the Mediterranean Monachus albiventer, 
seems to be on the verge of extinction. The former was once 
common off Florida and near most of the islands. Now it is 
only noticeable in the neighbourhood of Cuba and some of the 
islands near Yucatan. We can hardly believe, remarked 
Messrs. Sclater,* that these creatures could easily traverse 
the whole Atlantic. The hypothesis of a former barrier of 
land between Africa and America, which we know to be sup- 
ported by other facts of distribution, would alone explain the 
difficulty, according to these authors. They only had the sup- 
posed land connection between Africa and South America in 
mind, but what strengthens the evidence in support of another 
more northerly mid-Atlantic land bridge between the Antilles 
and the Mediterranean region is the fact that the only locali- 
ties outside the Mediterranean where Monachus albiventer 
occurs are on the coasts of Madeira and the Canary islands. 

Among the terrestrial species of vertebrates and inverte- 
brates of the Antilles, as I observed, the affinity with Europe 
is less marked perhaps than it is in the south-western States 

* Sclater, W. L., and P. L. Sclater, " Geography! of Mammals," p. 217. 

FIG. 14. Map of North and South America, indicating roughly the supposed 
conditions of land and water about the commencement of the Tertiary Era. 
The ancient land is slightly tinted. 

[To face p. 280. 


of North America. I explained this by the supposition that 
the widespread submergence of the Antilles destroyed most of 
the immigrants from Europe. It ought not to have affected 
the emigrants to Europe from the Antilles to the same extent, 
I mean those forms which spread from the Antillean centre, 
because they would have had more time to adapt themselves 
to the more elevated regions in the West Indies, and would 
thus have had more chance of .surviving the submergence 
which did not entirely cover the islands. 

One of the most noteworthy examples of that kind, though 
not a very conspicuous onq, is the newt Spelerpes. [ alluded 1 
to its range in North America on several occasions (pp.137 
138 and p. 221), pointing out that its headquarters were in 
Mexico, while a single species had succeeded in reaching the 
Mediterranean region. We may assume, therefore, that 
certain members of the old American Spelerpes stock 
emigrated, in early Tertiary or even in Mesozoic times, by 
means of the trans -Atlantic land bridge, that extended from 
the Antilles to a land area which covered part of the western 
Mediterranean (Fig. 14). That Spelerpes long ago existed all 
over the Antilles is indicated by the fact that a single species 
(Spelerpes infuscatus) still inhabits the island of Haiti. 

Another instance I alluded to (p. 173) in support of 
the mid- Atlantic land bridge theory, is the glass -snake 
family (Anguidae). The genus' Ophisaurus, w'hich is found 
in the Mediterranean region, does not occur on the Antilles, 
but several species are known from the mainland of North 
America. The genus Anguis is quite confined to Europe, 
western Asia and north Africa. On the other hand, numerous 
relations live in the West Indies. Sauresia and Panolopus are 
confined to Haiti, whereas Celestus (Diploglossus) inhabits 
Portorico, Haiti, Jamaica, Cuba, Central America, Mexico 
and northern South America. The headquarters of the 
Anguidae certainly are in the West Indies and Central 
America, and it is from there that they must have spread to 
Europe and beyond, as far as the Himalayas, when a land 
bridge across the Atlantic permitted them to do so. 

Let us now return to the investigation of the geological 
history of the Antillean area. Among the vertebrates, the 
mammals perhaps are of the greatest importance, in so far 


as they are not supposed to be subject to accidental dispersal. 
The West Indian mammals * consist of a mixture of exceed- 
ingly ancient and of apparently much more modern types, and 
yet all are distinct enough from mainland forms to exclude the 
idea of recent land connections of the Greater Antilles either 
with Central America or the two neighbouring continents. 

The most ancient mammal found in the West Indian 
islands is the curious insectivore Solenodon. It is the sole 
genus of the family Solenodontidae, whose nearest living 
relations are the Centetidae of Madagascar and West Africa. 
The two Antillean forms (S. paradoxus and S. cubanus) f are 
in general quite similar. Yet they differ somewhat in size, 
colour and dentition, as well as in the shape of the skull, and 
for that reason are perfectly distinct species. The first is con- 
fined to Haiti,: the other to Cuba. Professor Leche J expresses 
the view that Madagascar lost its continental land connection 
already during the Eocene Period. Hence the Centetidae may 
be of early Tertiary or even Mesozoic age. Professor Leche 
believes in a former land connection between Madagascar and 
Africa, and in another between Africa and Brazil. Both of 
these must have existed about the same time, and they were 
used presumably by the ancestors of the Centetidae and 
Solenodontidae in passing from Madagascar to South 
America, and thence to the West Indies, or vice versa. 

The only large West Indian mammals, Capromys and 
Plagiodontia, belong to the rodents. The hutias, as they are 
called, remind us somewhat of the great rat-like South 
American coypu, but the tail is longer and they possess 
arboreal habits and certain structural characters differing 
from the latter. The two genera of hutia are quite confined 
to the West Indies. Three species of Capromys are known 
from Cuba, one from the Bahamas and one from Jamaica. 
Still another Capromys inhabits the small Swan island, in the 
Gulf of Honduras, mid-way between Jamaica and Central 
America. Nevertheless, the genus is quite unknown from the 
mainland. The other genus (Plagiodontia) only occurs on 

* Allen, Glover M., "Mammals of West Indies." (This work was 
received too late for discussion.) 

t Allen, J. A., "Notes on Solenodon paradoxus," pp. 507 515. 
t Leche, W., " Centetidae, Solenodontidae, &c.," pp. 132139. 


Haiti. Of the past history of these hutias we only know 
that one extinct species (Capromys columbianus) has been 
discovered in the Pleistocene deposits of Cuba. The only near 
relation living is Procapromys geayi, from the neighbourhood 
of -Caracas in the mountains of Venezuela,. This implies 
southern affinities of the genera Capromys and Plagiodontia. 
Proceeding further south in search of their possible ances- 
tors we meet with another nearly related genus (Matyoscor) 
in the Pleistocene of Bolivia. Still further south we find 
on some of the islands of the Chilean coa,st, as well ais in Chile, 
Peru and Argentina, the coypu (Myocastor coypus), already 
alluded, to as resembling the hutias. Finally, in the Pliocene, 
Miocene and Eocene beds of Patagonia, various ancestral 
types of these modern forms have been discovered by Pro- 
fessor Ameghino. Thus the available evidence points to a 
remote Patagonian origin of the hutias. The question then 
arises, have the ancestors of these West Indian mammals 
proceeded northward through eastern or western South 
America ? The testimony we possess is distinctly in favour 
of the latter theory, though it is mainly of a negative cha- 
racter. In the west we have the living coypu and the extinct 
Matyoscor, while the former has only invaded Brazil in recent 
geological times. The Venezuelan Procapromys might lead 
us to 'believe that the ancestors of the hutias had gained 
admittance to the Antillean region by an old land connection 
across the lesser Antilles. But since no trace of the former 
presence there of any coypu -like mammal has been, dis- 
covered, and as a species of hutia exists on Swan island, the 
hypothesis that the ancestors of these mammals reached the 
West Indies directly from some western lands seems to me the 
most probable. 

Except the Bahama raccoon (p. 181) and a species of 
opossum (Didelphys marsupialis), which inhabits the islands 
of Trinidad, Dominica, Grenada and St. Vincent, all other 
Antillean mammals are small and inconspicuous. The 
raccoon may possibly be very ancient, but we know too little 
about its geological history to enable us to speculate on 
its origin in the Bahama islands. The presence of the 
opossum on some of the Lesser Antilles seems to indicate 
that they had been connected with one another and with Vene- 


zuela within comparatively recent geological times, a con- 
clusion which perfectly agrees with the results obtained from 
other faunistic evidence. 

The groups of smaller mammals which have representa- 
tives in the West Indies are the rice-rats (Oryzomys) and 
the musk-bearing rats (Moschophoromys=Megalornys). Like 
the hutias and solenodons, the latter are confined to this 
region, but they are only known from the Lesser Antilles, 
Moschophoromys desmaresti being peculiar to Martinique, 
whereas M. luciae is only found on St. Lucia. The occurrence 
of this peculiar genus of small mammals on the Lesser 
Antilles implies that their connection with the mainland must 
be of longer standing than is generally supposed. The 
opossum (Didelphys marsupialis), is one of the few species 
of terrestrial mammals having a wide range in both North 
and South America, and must, therefore, be considered as 
a persistent specific type of great antiquity. It may possibly 
have retained its specific characters, while the ancestors of 
the musk-bearing rats rapidly became modified on the islands. 
That the latter are not recent introductions is likewise in- 
dicated by the fact that an extinct species of Moschophoromys 
is known from Barbados. 

The rice-rats (Oryzomys) inhabit both the islands and the 
mainland. Their very wide and discontinuous range in North 
and South America at once suggests an ancient origin. One 
species (0. antillarum) was so abundant in Jamaica, and did 
such damage to the crops, that the mongoose, a small carni- 
vore, was imported from India for its destruction. For a 
time the experiment appeared to be, quite successful. After 
some ten years, however, when the rice-rats became scarce, 
the mongoose began to pay attention to snakes, lizards, 
tortoises and birds. As it developed a taste for them they 
in their turn became scarce, until it was found that 
the mongoose was worse than any other animal pest, for x 
owing to the great destruction of the useful members of 
the fauna, insects and ticks had increased to an alarming 
extent. The disturbance of the due balance of nature by 
man is always risky, and thus sometimes leads to incalculable 

A second species of rice-rat (0. victus) is peculiar to the 


Lesser Antilles. A few others live in Trinidad. It is im- 
probable that Jamaica was at any time connected by land 
with the Lesser Antilles, independently of the larger islands. 
It is possible, therefore, that the genus, being mainly South 
and Central American, entered the Antilles in two independent 
streams, one from the south, the other from the west. This 
may have taken place when the other large islands had already 
become isolated from Jamaica. 

We have still to consider the remains of some extinct 
mammals which have been traced in the Antilles. To judge 
from recent cave researches in Cuba, we are only now begin- 
ning to acquire a knowledge of the past history of this fauna, 
though certain indications enable us to speculate as to its 
general nature and origin. Dr. Spencer * alludes to the dis- 
covery in a Cuban cave of large edentate remains belong- 
ing to Myomorphus cubensis. This creature, which Dr. 
Matthew informs me is now known as Megalocnus, was about 
the size of a small bear, being a peculiar aberrant specializa- 
tion apparently derived from some Miocene (Santa Cruz) 
ground sloth. Professor Ameghino f describes a new species 
of man (Homo cubensis) from a cave in Cuba; and from a 
{preliminary report read at the International Congress of 
Geology at Stockholm by Professor de la Torre, we are led to 
infer that other edentates, and also several kinds of rodents, 
lived on this island within comparatively recent geological 
times. Dr. Matthew writes to me that the mammalian re- 
mains from Cuba sent to the American Museum of New York 
are awaiting determination. Dr. Spencer J also makes refer- 
ence to several large species of rodents of the genus Amblyr- 
hiza which were found in the phosphate beds of the small 
island of Anguilla. The Amblyrhiza remains are now thought 
to belong to a single species (A. inundata), closely related to 
Castoroides, a giant beaver, which made its first appearance 
in North America during the Pleistocene Period. The former 
presence in Anguilla and Cuba of large mammals naturally 
suggests that these islands had some continental connection. 
Dr. Spencer, indeed, quotes these fossils in support of the 

* Spencer, J. W., " Antillean Continent," pp. 128138. 
t Ameghino, Fl., " Nueva especie del g6nero Homo," p. 5. 
t Spencer, J. W., "Fossil Mammals in Cuba," pp. 512 513. 


theory, largely based on the supposed drowned valleys of the 
Antillean area, that the two American continents were united 
with one another by means of a West Indian land bridge 
during the Pleistocene Period. He thought this land must 
have had the enormous altitude of from 8,000 to 12,000 feet. 
Against this view it may be urged from a general faunistic 
study that such a land bridge is out of the question, 
because the Antillean fauna would be quite different from 
what it actually is. A very small part of southern Florida 
may have been still joined, by way of the Bahamas, to Cuba 
and Anguilla in Pliocene times. But southern Florida must 
then have been separated from North America by a marine 
channel. Nor could there have been any junction of the 
greater Antillean land with South America, by means of the 
Lesser Antilles, at any time during the Tertiary Era. In all 
probability the West Indies, while still partly united with one 
another, had some kind of land connection with Central 
America, while the latter was cut off by the sea from both 
North and South America. Hence the mammalian supply came 
mostly from the west. Still, it may be asked, how did this 
fauna reach Central America ? Almost all the Antillean mam- 
mals, both recent and extinct, can be traced to a southern 
ancestry, except Castoroides, of which we only know that it 
appeared in the north along with a host of southern invaders. 
I believe the cause of the mystery connected with the occa- 
sional entrance during various periods in Tertiary times of 
South American forms into North America is now buried 
beneath the waves far out in the Pacific Ocean (see Fig. 14). 
I have frequently suggested the theory that a large land 
eurface, mainly tenanted by a South American stock, lay 
westward of Central America. A good deal of the next 
chapter will be devoted to its consideration. Meanwhile, 
I can only state my conviction that the animals alluded "to, 
and many others, had their origin on that western Pacific 
land whence they streamed into the neighbouring continents 
whenever physical conditions offered an opportunity for 
doing so. 

The main conclusion which this brief study of the mam- 
malian fauna of the Antilles has revealed is that the larger 
islands were formerly united with one another, and with the 


Bahamas, by land. How far eastward this land extended 
cannot be determined from the mammals, but it certainly 
must have reached beyond the Virgin islands to Anguilla, one 
of the most northerly islands of the Lesser Antilles. Many 
islands of the Lesser Antilles may have been completely 
submerged at that time. Later on Cuba, the Bahamas and 
Jamaica were connected with the mainland, after the other 
islands had already been separated from the Antillean land- 
mass, and lastly, Cuba and Jamaica were independently joined 
to Central America before the existing physical features were 
brought about. 

Many more details are required before we can construct 
an approximate history of the geological events which led to 
the present conditions of the fauna and flora of the Antilles. 
In his chapter on the evolution of middle America Dr. 
Gadow * deals with the problem from the aspect of the 
reptilian and amphibian fauna. He does not attempt to recon- 
struct the geological history from a study of this fauna alone. 
His maps are derived and abstracted from the opinions ex- 
pressed by various geological authorities. They are in so far 
of importance, as he thinks, that the general conclusions 
based on a geological study appear to fit those founded on 
faunistic considerations. Several striking points may be 
gathered at a glance from his maps. At no time since the 
Cretaceous Period were the Antilles directly connected with 
the main continent of North America. Only in Miocene times 
was " Antillea" (the term used for the united Antilles) joined 
to southern Florida, the latter being then disconnected from 
North America. Antillea had only one continental connec- 
tion, namely, during the Miocene Period by means of Central 
America. It follows, therefore, that Antillea never was directly 
joined either to North or South America, at least not since 
Cretaceous times. If Dr. Gadow's maps are correct, the whole 
of the Antillean fauna and flora, in so far as they are derived 
from migration on land, should be of Miocene age or older. 
Dr. Gadow's conclusions would have been of greater value had 
he trusted more to the evidence based on geographical dis- 
tribution and less on that of geology. As pointed out by 

* Gadow, H., " Mexican Amphibians and Eeptiles," pp. 234237, 


Dr. Stejneger,* a few Antillean genera, such as Ameiva, 
Amphisbaena, Typhlops and Alsophis, possess a (Southern rela- 
tionship. The species of burrowing snake, Typhlops lumbri- 
calis, is even confined to the West Indies and British Guiana 
without entering Central America. Dr. Stejneger does not 
actually speak of a former land connection of Antillea with 
South America, but I explained above that some geologists 
advocate an ancient union of the two continents by way of 
the Lesser Antilles. If such a connection existed, I think 
it must have been in the pre-Cretaceous times. Most zoo- 
logists would deride the idea of any terrestrial species 
having been transmitted to us unchanged from pre-Cretaceous 
times to the present day. We have no geological evidence of 
such a fact, but Dr. Sarasin expresses the opinion that the 
genus Typhlops, at any rate, is older than the Cretaceous 
Period. We gather from Dr. Stejneger's researches that the 
great majority of the Portorican species are western in their 
affinities, and that Portorico and the more easterly Virgin 
islands certainly must have been connected with one another 
in comparatively recent geological times. 

As regards Jamaica, Mr. Barbour gives us some interesting 
details as to its reptiles and amphibians. He emphasises the 
fact, which I have already alluded to in speaking of the 
mammals (p. 285), that the distinctive characters of this 
island are not shared by the rest of the Antilles^Jamaica is 
related faunistically to Central America and Haiti, while the 
relationship with Cuba is much more remote/ Mr. Barbour 
thinks that the early separation of Jamaica from the main- 
land and from Haiti would account for the absence of types 
having such a distribution as Bufo and Amphisbaena, which 
may have reached Haiti from Central America by way of 
Cuba. If fortuitous dispersal has played practically no part in 
providing the Antilles with a fauna, as Mr. Barbourf is in- 
clined to think, the Bahamas certainly must have belonged to 
Antillea and cannot have been submerged since. Sixteen 
species of reptiles and amphibians are described by Mr. 
BarbourJ as occurring in the Bahaman archipelago. Several 

* Stejneger, L., " Herpetology of Porto Eico," pp. 561563. 
t Barbour, Thomas, " Herpetology of Jamaica." pp. 277 285. 
t Harbour, Thomas, "Batrachia and ReptiJia of Bahamas." 


of these are peculiar to the group. And yet in spite of this 
feature, in spite of the fact that two kinds of mammals live on 
the islands, and that numerous mollusks and other inverte- 
brates are quite confined to the Bahamas, we often read the 
statement that the fauna of these islands is entirely a product 
of accidental transport by winds, waves or human agency. 

How little influenced the Antilles are by accidental intro- 
ductions is well shown by a study of such groups as the 
dragon-flies and their allies (Neuroptera and Pseudoneu- 
roptera), and the butterflies and moths (Lepidoptera). If 
winds played any active part in their dispersal, we should 
find many species from North America in the Greater Antilles, 
whereas the affinities agree almost precisely with those of 
the other animals already considered. Several genera of 
dragon-flies, known both -from North and South America, such 
as Chauliodes, Corydalis, Mantispa and others, are wanting 
in the West Indian islands, thus showing that the migration 
between the two continents did not take place across the 
Antillean area. The great majority of the genera are South 
or Central American. Only two genera of dragon-flies ,are con- 
fined to the Antilles. On the other hand, many continental 
ones are absent. The Lesser Antilles contain some South 
American species which are wanting in the Greater Antilles.* 
Similarly the Lepidoptera exhibit affinities with northern 
South America, Central America and Florida. Several cha- 
racteristically southern families, nevertheless, are wanting in 
the Antilles. The four large islands are characterised by the 
absence of many neotropical genera, and the presence of a 
genus and many species that are peculiar to them. Jamaica 
shows some special features of interest. f 

The birds exhibit similar peculiarities in their geographical 
distribution. One whole family, the todies (Toditae), are con- 
fined to the larger islands of Cuba, Jamaica, Haiti and Por- 
torico. They bear a certain resemblance to the kingfishers, 
but differ in colour, whereas in habits and shape of bill they 
resemble the flycatchers. In reality they are related to the 
Central American motmots. We also have the same striking 

* Kolbe, H. J., " Neuroptera und Pseudoneuroptera d. Antillen," 
pp. 154161. 

t Pagenstecher, A., " Verbreitung der Schmetterlinge," pp. 391 394. 
L.A. U 


number of absentees and peculiar species among the birds, 
as in other groups. Dr. Chapman* is evidently puzzled how 
to account for the avifauna. Nevertheless, he makes the 
remarkably suggestive statement that if it could be shown 
that Central America was cut off from both continents at the 
time when it was joined to the West Indies, the origin of the 
island fauna could be explained in a satisfactory manner. 
That is precisely, I think, what did happen, as I have indicated 

Cuba presents some special aspects of interest. We have in 
the first instance quite a remarkable assemblage of fossil 
mammals. Western Cuba, with its extensive limestone dis- 
tricts, abounds in caves, many of which contain stagnant 
water or running streams. One of the most remarkable fea- 
tures of these subterranean waters is that they are tenanted 
by two kinds of fishes belonging curiously enough to a family 
of deep-sea forms (Brotulidae). These species, Stygicola 
dentatus and Lucifuga subterraneus, moreover, are blind, 
as might be expected from their habitat. Another typically 
marine genus, Atherina, possesses a single fresh -water species 
(A. evermanni) which is only known from western Cuba. 
Whether the presence of these marine species in fresh water 
implies that western Cuba had been submerged below sea- 
level is a debatable question that need not be further pur- 
sued. Other Cuban fresh-water fishes are forms often found 
in brackish water, or marine forms migrating into fresh water. 
Besides these there are two species of Heros, a Symbranchus, 
a Lepidosteus and one Agonostomus, all of which belong to 
purely fresh -water groups that could only have reached Cuba 
by means of a continuous system of lakes and rivers between 
the mainland and the island. Heros tetracanthus and Heros 
nigricans are members of the family Cichlidae, which has a 
wide range in South and Central America, only a single 
species entering the United States. Agonostomus monticola 
has a wider distribution in the West Indies, and is also met 
with in Mexico. Lepidosteus tristoechus lives in Mexico and 
the southern States, while Symbranchus marmoratus is widely 
distributed in the fresh waters of tropical America. If, as 

* Chapman, F. M., " West Indian Bird Life," p. 330. 


geologists claim, western Cuba has been deeply submerged, it 
seems equally certain that it must have been subsequently 
joined to Central America. As Professor Eigenmann* re- 
marks, this Cuban fauna has greater affinity with that of 
Mexico than with that of Florida, showing that the fishes 
probably reached Cuba by way of Yucatan. 

Western Cuba possesses, among other curiosities, blind 
crustaceans (Palaemonetes). They are what we might call 
monster prawns, and have probably had marine ancestors. 
An allied species is known from the subterranean waters of 
Texas. Faunistically they are not of great importance. Not 
so, however, the fresh -water crayfishes, which are justly 
regarded by Dr. Ortmann f as among the most valuable means 
for demonstrating former changes of land and water. All the 
species of Cambarus but one inhabit either Mexico or the 
United States. Only a single crayfish occurs in the isolated 
position of Cuba. This species (C. cubensis) is closely related 
to Cambarus mexicanus, whereas the United States species are 
more distant relatives. Dr. Ortmann assumes on that 
account a former land connection between Mexico and Cuba. 

The fresh-water crabs, as I already indicated, advanced 
northward from the south, eventually invading the Central 
American territory and the West Indies. Thus three species 
of Pseudothelphusa have been discovered in Cuba. One of 
these (P. americana) ranges from Mexico through Cuba to 
Haiti. Another (P. terrestris) is confined to Mexico and 
Cuba, while a third (P. affinis) is restricted to Cuba. More- 
over the genus Epilobocera is peculiar to the Greater 
Antilles, viz., Cuba, Haiti and Portorico, and some smaller 
neighbouring islands. A species of Pseudothelphusa closely 
allied to a Venezuelan form is likewise known from some of 
the Lesser Antilles. 

From these facts Dr. Ortmann (pp. 344 347) concludes 
that the first immigration of fresh-water decapods represented 
by Epilobocera, into the Greater Antilles, belongs to the end 
of the Cretaceous or the beginning of the Tertiary era. 
Cambarus cubensis, he suggests, possibly belongs to it also. 

* Eigenmann, C. H., "Freshwater Fishes of Cuba," pp. 220221. 
t Ortmann, A. E., "Distribution of Freshwater Decapods," p. 315. 



Although it is a primitive form, he prefers to put its immigra- 
tion in the Tertiary rather than in the Cretaceous. The history 
and development of the Central American and West Indian 
region he imagines to have been as follows : " Central America, 
the West Indies and the northern margin of South America, 
formed in the Mesozoic Period (certainly during the Jurassic 
and Cretaceous) a continental mass (Antillean Continent), 
which was bounded by sea to the north and south (Fig. 15). 
This continent broke up at the end of the Cretaceous, 
the chief factor in its destruction being the formation 
of the Caribbean Sea. The northern remnant of this 
continent, consisting of the Greater Antilles and parts 
of the present Central America, probably remained a unit 
up to the Eocene." At the end of the Eocene, he 
continues, and during the Oligocene and Miocene Periods, 
the connection between the Greater Antilles and the main- 
land was severed, being subsequently re-established toward 
the end of the Tertiary Era (Fig. 20) and again destroyed in 
recent times. 

My own views are in some respects similar to those of Dr. 
Ortmann, yet they differ in a few important points. If we 
begin, say with later Cretaceous times, when a wide marine 
channel still separated eastern from 'western North America, 
northern South America was submerged, and could not have 
formed part of the supposed Antillean Continent. Towards 
the end of the Cretaceous Period, all the Antilles, except the 
Bahamas, were entirely covered by the sea, according to Pro- 
fessor Schuichert's paleogeographic maps (Map 95). Yet 
although the peninsula of Yucatan was then submerged, no 
deposits of this age are known from Guatemala or Honduras, 
nor have any Mesozoic or Tertiary beds been discovered in 
these countries. The North American relationship of the 
relict fauna inhabiting these countries indicates that they 
were connected by land with western Mexico. I shall also 
show later on that this region was united by land with western 
South America before the Isthmus of Panama came into exis- 
tence. Possibly the Greater Antilles were not so completely 
covered by the sea as is assumed. They may have been repre- 
sented by small islands, and these would have possessed frag- 
ments of an ancient fauna and flora. Cuba, at any rate, 


emerged at the beginning of the Tertiary Era, and must then 
have been joined by land with Mexico. Some time during 
the Oligocene Period Professor Schuchert again records a 
complete submergence of all the West Indian islands except 
the Bahamas. But such ancient types as Typhlops, Soleno- 
don and many others, could not have been destroyed. We 
might suppose that they took refuge on the Bahamas, and 
thus repopulated the other islands subsequently. Such a 
theory, however, is exceedingly unlikely. A much more pro- 
bable explanation is that the Antilles were reduced to small 
islands, and retained their old animals and plants. In early 
Miocene times all the Greater Antilles were certainly raised 
above the sea, and must have been then connected with one 
another. Jamaica was joined to Guatemala, and Cuba to 
Mexico, but Jamaica must have separated early from Haiti. 
While the islands were joined to one another, an interming- 
ling of the more active ancient types occurred, the less pro- 
gressive ones being forced to the higher altitudes by the new 
arrivals from Mexico and Central America. During the whole 
of Miocene times Yucatan was apparently below sea-level. 
When it rose in the Pliocene Period, it may have had a short 
land connection with the Antilles by way of western Cuba. 
An opportunity was then afforded the mammals of South 
American type like Amblyrhiza, Megalocnus, and others to 
spread to the islands. The smaller West Indian mammals 
came earlier. Central America, during the existence of this 
Yucatan land bridge, may still have been separated from 
North and South America. At this time the Lesser Antilles 
probably had an independent land connection with Venezuela ; 
but that there was an Antillean Continent connected with the 
mainland in Pleistocene times, as suggested by Dr. Spencer, 
when Central America had already been invaded by the North 
and South American immigrants, is entirely opposed to the 
results derived from a study of the fauna and flora. 

As for the trans -Atlantic land bridge, we cannot assume 
that it oscillated up and down like the Antillean area, where 
peculiar local conditions produced exceptional changes of land 
and water. Some time during the Eocene Period it must 
have enabled European types to travel right across the 
Antilles and whatever portions of Central America then 


existed, to western North America. I have endeavoured to 
roughly represent on a map my conception of the geographical 
conditions then prevailing in North and South America 
(Fig. 14). Then followed a complete cessation of the migra- 
tory stream to North America. This must have corresponded 
with the depression of the Antilles. During the Oligocene 
Period Professor Osborn recorded a re-establishment of the 
faunal resemblance of western North America with western 
Europe. Here our correlation of Antillean geology with that 
of the continental mammalian deposits fails us completely. 
The supposed Oligocene beds of the Antilles may really belong 
to the Eocene Period. Another phase of independent evolu- 
tion then succeeded this faunal interchange with western 
Europe. This has been represented on another map (Fig. 16), 
although I am fully aware of the extreme difficulties and 
pitfalls connected with such an attempt. Later on, in 
Miocene times, a fresh invasion occurs from the Old World. 
This, I think, certainly did not come from the same quarter 
as the others. We may conclude, therefore, that the land 
bridge between the Antilles and Europe ceased to exist some 
time during the Oligocene Period. It is quite possible that 
long after that time large islands may have marked the former 
site of the land bridge, while temporary land connections be- 
tween the Azores and Europe, and between the Canary islands 
and Africa, may have involved certain fragments of the older 
trans-Atlantic land connection. 

FIG. 16. Map of North and youth America, indicating roughly the supposed 
conditions of land and water about the middle of the Tertiary Era. The 
ancient land is slightly tinted. 

[To face p. 294. 



RIGHT on the Equator, off the west coast of South America, 
in the Pacific Ocean, lies a small group or archipelago of 
islands known as the "Galapagos islands." There are five 
large islands and twelve small ones, the most extensive 
(Albemarle) being about eighty miles long and of very irre- 
gular shape, while the four next in importance (Chatham, 
Indefatigable, James and Narborough) are each about twenty 
or thirty miles in length, and of a rounded or elongate form. 
They are best seen on the map representing the range of 
Helicina (Fig. 11) being situated within the circle nearest 
South America. The climate of the islands is cooler than 
that of any other equatorial land at sea-level, because the 
temperature of the surrounding water is singularly low. This 
is caused by the Humboldt current, which strikes the islands 
during its north-westerly course on its way from the cold 
southern regions. The lower slopes of the islands suffer from 
want of rain. At a height of a thousand feet and upwards 
sufficient moisture is condensed to provide for a fairly 
luxuriant vegetation. 

The Galapagos islands differ in two important respects from 
many of the West Indian islands. They are entirely formed 
of volcanic rocks, according to Mr. Darwin, who gave us the 
first reliable description of them, and they are situated at the 
great distance of seven hundred and fifty miles from the main- 
land of South America. The nearest land, apart from the 
small Cocos islands, is Ecuador, to which state the Galapagos 
archipelago now belongs. In the case of the Antilles almost 
all authorities are agreed that the nucleus of the fauna re- 
presents survivals from the time when the islands were joined 
to the mainland, arid that accidental distribution has not 
greatly modified the original fauna. In the Antilles we meet 


with a number of ancient types of animals and plants. In the 
Galapagos islands similar relicts of bygone ages are found, yet 
their presence on the islands is not generally considered to 
be due to a former land connection with the mainland. But 
the question whether the animals and plants inhabiting this 
archipelago are the descendants of purely accidental visitors 
or not, is of more than usual importance in this case. Both 
Mr. Darwin and Dr. Wallace expressed the opinion that the 
Galapagos islands are of the oceanic type, and that they 
originated from sub-oceanic volcanoes, having never formed 
part of a continent or any large land-mass. Many of the more 
recent observers have followed the lead of our two great 
naturalists. I need only mention Professor Agassiz, Dr. 
Wolf, Mr. Salvin, Dr. Dall and Dr. Stearns. On the other 
hand, Professor Baur was the first to seriously contest the 
view of these authorities. His researches led to a reconsidera- 
tion of the problem presented by the fauna and flora of the 
Galapagos islands and a rejection by many naturalists of the 
theory so ably maintained by Darwin and his followers. 

Islands, as I have already indicated, possess many advan- 
tages over continental areas for the study of the laws and 
phenomena of distribution. The islands we are now dealing 
with, moreover, are of the most profound importance in con- 
nection with the geological history of the two great American 
continents. We might almost say that the problem of 'the 
Galapagos fauna presents the key to the solution of the most 
complex and intricate questions concerning the American 
fauna with which we are confronted. It is for this reason 
that the whole of this chapter will be devoted to an exhaustive 
survey of the Galapagos islands and their inhabitants. 

When Darwin landed on the archipelago in 1835, he noted 
in the first instance the fact that all the islands were formed 
of volcanic rock and that they contained numerous craters. 
The craters he found to consist either of lava and scoriae, or 
of finely-striated sandstone-like tuff. All the latter, which he 
examined, had the southern sides either much lower than the 
other sides, or quite broken down and removed. He concluded 
from this fact that all these craters were formed beneath the 
sea, and that their broken state was due to the action of the 
waves produced by the southern trade winds. On the other 


hand, Darwin's* remark that he noticed a few fragments of 
granite curiously glazed and altered by heat among the ejecta- 
menta, would seem to imply that the bases of the craters are 
composed of older forms of eruptive rocks. This supposition 
is strengthened by an observation made by Professor Suess,f 
that on the whole of the Pacific coast of America only a single 
mountain range comes to an abrupt termination on the Pacific 
coast, namely, the Central American continuation of the An- 
tillean Cordillera, and that precisely at the point where we 
might imagine the arcuate prolongation of this chain to meet 
the principal Sooith American mountains lie the volcanic, Gala- 
pagos islands. At any rate, as Dr. Blanford J has pointed 
out, the rocks of an island may be entirely volcanic, although 
the island may nevertheless be a remnant of a continental 
mass. Except that some of the craters have their southern 
faces broken down, which may be due to some other cause 
than that suggested by Darwin, the Galapagos islands could 
just as well represent the mountain tops of sunken land as 
the summits of originally submarine volcanoes. Neither of 
these two theories is supported by strong geological evidence. 
No one was more impressed by this fact than Darwin himself, 
and he bases his theory of the origin of the Galapagos fauna 
and flora almost entirely on the nature, composition and dis- 
tribution of the animals and plants he found on the islands. 
His conclusions were that all the animals and plants must be 
derived from accidental transport by sea-currents or by birds, 
except for a few recent immigrants which were introduced by 

The natural history of the islands, as Darwin truly remarks, 
is eminently curious and well deserves attention. Of terres- 
trial mammals, he says, there is only one which must be con- 
sidered as indigenous, namely a mouse (Mus galapagoensis). \ 
A. rat also is sufficiently distinct from the common kind to have 
been named and described, " but," continues Mr. Darwin, 
" as it belongs to the Old World division of the family, and as 
this island had been frequented by ships for the last hundred 
and fifty years, I can hardly doubt that this rat is merely a 

* Darwin, C., " Journal of Eesearches," pp. 270271. 
t Suess, E., " Das Antlitz der Erde," II., p. 263. 
| Blanford, W. T., " Anniversary Address," p. 34. 


variety, produced by the new and peculiar climate, food and 
soil to which it has been subjected." Hence Darwin * con- 
sidered the rat as imported by man, and he adds that the 
transportation of the mouse in a vessel from the mainland 
is not improbable. 

Dr. Wallace concurs with Mr. Darwin in the belief that the 
Galapagos islands agree with all other oceanic islands in so 
far as they are devoid of indigenous mammals. He acknow- 
ledges that the mouse, which had since been found to belong 
' to the American genus Hesperomys, differs somewhat from 
other known species. Yet he remarks that we can scarcely 
consider the species to be indigenous, firstly because these 
creatures have been little studied in South America, and that 
there may still be many undescribed mice there, and in the 
second place because even had it been introduced by some 
European or native vessel, there would have been ample time 
in two or three hundred years for the very different conditions 
to have established a marked diversity in the characters of the 
species. Dr. Wallace f reiterates in conclusion Darwin's 
statement that the islands are really destitute of truly in- 
digenous mammals. 

Now since these great naturalists wrote about the Gala- 
pagos archipelago, and stated their opinions on the mam- 
malian fauna of the islands, the latter have been visited by 
many competent observers, who brought their collections 
home for careful study. Thus the mammals collected by 
Dr. Baur on the islands were described by Dr. Allen, who 
recognised among them a typical brown rat (Mus decu- 
manus). He also thought that a skeleton which was picked 
up on South Albemarle belonged to the black rat (Mus rattus). 
There was likewise a bat which he identified as closely allied 
to a Chilean form, but sufficiently distinct to deserve the 
separate name of Atalapha brachyotis. Dr. Allen, J moreover, 
points out that the Galapagos mouse does not belong to the 
Old World genus " Mus " nor to " Hesperomys," but agrees 
with Oryzomys, the well-known American rice-rats. Besides 

* Darwin, C., " Journal of Researches," p. 275. 

t Wallace, A. E., "Island Life," p. 278. 

t Allen, J. A., " Mammals from the Galapagos Islands." 


Oryzomys galapagoensis he found a second distinct form 
which he named Oryzomys bauri. 

A more important natural history survey of the islands 
was made by the Hopkins Stanford Expedition in 1898. In 
addition to the mammals already known from the archipelago, 
it was discovered that Narborough, Indefatigable and Sey- 
mour islands are inhabited by rice-rats, which show such 
structural distinctness in the interorbital portion of the skull 
that Mr. Heller * proposed their being placed into the 
new genus " Nesoryzomys." Four mouse-like mammals are 
thus known to be peculiar to the Galapagos islands, viz., 
Oryzomys galapagoensis; 0. bauri, Nesoryzomys indefessus 
and N. narboroughi. The only other terrestrial mammal not 
found elsewhere is Atalapha or, as it is now called, 
Lasiurus brachyotis. 

It is interesting to note that the genera to which these 
species belong are characteristically American. Both Ory- 
zomys and Lasiurus have a wide range in North and South 
! America as well as in the Antilles. We have learned that 
Oryzomys antillarum is confined to Jamaica. Another species 
lives on St. Vincent, and several more on Trinidad, while a 
species (0. nelsoni) is peculiar to the Tres Marias islands off 
the west coast of Mexico. A genus ranging from New Jersey 
in the north to Tierra del Fuego in South America, with a 
large number of species, must be a very ancient one; and 
yet not a single species occurs outside the New World. 

I have never hitherto alluded to bats because they are so 
generally regarded as of little value in faunistic problems. 
Specialists of this group, such as Dr. Knud Andersen,f in- 
form us, however, that this assumption is by no means correct, 
and that in most cases bats are as good and reliable zoogeo- 
graphical guides as other small but non-flying mammals. 
The fact that although Lasiurus is found from Canada to 
Patagonia it has nevertheless preserved its original habitat 
in the New World, and has not even crossed Bering Strait into 
Asia, would seem to support Dr. Andersen's contention, while 
its antiquity is vouched for by the occurrence of distinct 
species of Lasiurus on several ^ remote American islands. 

* Heller, Edmund, " Mammals of Galapagos Archipelago." 
t Andersen, K., " Geogr. Distribution of Khinolophus," p. 652. 


Lasiurus has likewise succeeded in reaching one other group 
of islands besides the Galapagos archipelago, namely the 
Sandwich islands. The latter are inhabited by Lasiurus 
semotus, which is peculiar to them, and cannot therefore be 
considered as a recent human importation. 

Our present knowledge of the mammalian fauna of the 
Galapagos islands has added one factor of extreme importance 
to those hitherto known, namely the certainty that there are 
indigenous mammals on the archipelago. For this reason the 
Galapagos islands should be excluded from the oceanic 
islands, and be placed among the islands which once formed 
part of a larger land-mass or continent. It is not certain, 
however, that Darwin and Wallace would have taken that view, 
even had they been convinced that the mammals alluded to 
were truly indigenous in the islands. Some naturalists hold 
that even mammals can be successfully transported across 
the ocean on tree trunks and floating islands. Dr. Stearns,* 
in alluding particularly to the Galapagos fauna, remarks 
that a single tree of large size might carry with it not only 
molluscan and insect life, but also living individuals of many 
vertebrate forms that found refuge or safety upon it. Thus, 
he continues, if the environmental conditions were at all 
favourable, colonies of many animal forms could be trans- 
planted to distant regions. The possibility of such an acci- 
dental transportal must have been carefully considered and 
rejected "by Darwin and Wallace. The Humboldt current I 
alluded to as striking the Galapagos islands does not come 
from the coast of Ecuador nor from Central America. It 
originates in the far south, and, passing northward, skirts the 
coast of Chile and southern Peru, and then leaves the land in a 
north-westward direction. The tree trunks spoken of by Dr. 
Stearns as carriers of all kinds of animal life would have had 
to travel several thousand miles, no doubt experiencing stormy 
weather on the way, before they could have safely deposited 
their loads of vertebrates on the shores of the Galapagos 
islands. If these floating trees are responsible for the present 
mammalian fauna of these islands, how is it that they have 
brought nothing new to them since their occupation by man ? 

* Stearns, E. E. 0., " Mollusk-fauna of the Galapagos Islands," p. 366. 


Man has brought the brown and the black rats apparently. 
Why should the agencies of accidental transport have been 
so much more potent in the distant past than they are 
now ? We might add also, is it possible that these same 
agencies should be able to select the most ancient forms of 
life as more suitable for transport than more modern pro- 
ductions ? Of course, these ideas of mammals being carried 
across a vast expanse of ocean and safely landed on a distant 
shore are mere conjectures unsupported by any evidence. 

" Of land-birds," says Mr. Darwin, " I obtained twenty-six 
kinds, all peculiar to the archipelago and found nowhere else, 
with the exception of one lark -like finch (Dolichonyx oryzi- 
vorus) from North America." Of waders and water-birds he 
succeeded in capturing eleven kinds, only three of them being 
new species. In 1875 the number of land-birds known to Mr. 
Salvin had increased from twenty-five to thirty-six species. 
Most of the genera to which they belong are of very wide 
distribution ; seven, however, are confined to continental 
America, leaving five peculiar to the islands. Mr. Salvin * 
expresses no doubts as to the correctness of Darwin's inter- 
pretation of the origin of the Galapagos islands. Hence he 
concludes that the birds now found on the islands, being 
related to American birds, must have emigrated from' America 
and have become modified by the different climatic condi- 
tions with which they were surrounded. A later review of the 
Galapagos birds was undertaken by Mr. Bidgway f in 1897, 
when the number of land-birds recorded from the islands was 
nearly doubled. He notes the exact distribution of the various 
species and varieties in detail, and shows how circumscribed 
their range is. Of the five genera peculiar to the islands 
only two, viz., Nesomimus and Nesopelia are of evidently 
American relationship. The remaining three, he thinks, have 
so obvious a leaning towards certain Hawaiian dicaeidine 
forms, that the possibility of a former land connection with 
the Sandwich islands, either continuous, or by means of inter- 
mediate islands as " stepping stones," becomes a factor in 
the problem of their origin. "It may be," he adds, "that 
the resemblance of Cocornis, Cactornis and Camarhynchus 

* Salvin, 0., " Avifauna of Galapagos Archipelago." 

t Eidgway, E., " Birds of Galapagos Archipelago," p. 467. 


to the above-mentioned Hawaiian forms, is merely a super- 
ficial one, and not indicative of real relationship. Drs. Roth- 
schild and Hartert, indeed, believe that this is the case, but 
as a certain affinity between these widely separated islands 
undoubtedly exists in some other groups, Mr. Ridgway's view 
may possibly turn out to be the correct one. Moreover, as 
Mr. Ridgway points out, two of the Hawaiian Procellariidae, 
viz., Aestrelata phaeopygia and Oceanodroma cryptoleucura 
occur in the Galapagos islands. 

Drs. Rothschild and Hartert * derive the whole avifauna 
from America. All the same, their studies do not lead them to 
adopt Darwin's theory. They remain uncertain whether the 
fauna points to a former land connection with the mainland 
or whether it owes its existence to accidental dispersal. That 
there should be this doubt in their minds is rather surprising, 
considering their own statement that the geological evidence 
is entirely opposed to the theory of a former land bridge be- 
tween the islands and America. 

The most recent study on the birds of the Galapagos archi- 
pelago is that of Messrs. Snodgrass and Heller, f Taking 
the general results of all these papers, there can be no 
doubt that the avifauna is mostly related to South and Central 
America and the West Indies, thus agreeing in that respect 
with the mammalian fauna. 

The reptiles of the Galapagos islands certainly offer the 
greatest difficulty to the acceptance of Darwin's theory that 
the islands have risen from the floor of the ocean and have 
never been connected with one another or with the continent. 
Darwin observed only one small lizard belonging to a South 
American genus, two different kinds of the great lizard 
Amblyrhynchus a genus confined to the islands one kind 
of snake and a land tortoise. He noticed that the one kind of 
Amblyrhynchus lived altogether on land, the other being 
semi-aquatic. The latter he described as a hideous looking 
black creature up to four feet in length, with a laterally 
flattened tail and webbed feet. Sluggish in their movements 
on shore these huge lizards are expert swimmers, darting 

* Rothschild, W., and E. Hartert, " Ornithology of the Galapagos 
Islands," pp. 137139. 

t Snodgrass, E. E., and E. Heller, "Birds of the Galapagos Islands." 


among the rock -pools and feeding there on seaweed. The 
terrestrial species has no webs on the toes, and is of about the 
same size as its aquatic relative, reddish brown above and 
yellow underneath. It inhabits burrows, which occur in such 
numbers that the soil is constantly giving way as one walks 
over them, much to the annoyance of the tired traveller, as 
Mr. Darwin * remarks. Some of the tortoises, according to 
Darwin, grow to an immense size, yielding as much as two 
hundred pounds of meat. Like the two lizards they are 
entirely vegetable feeders. 

When Dr. Gunther f described the reptilian fauna of the 
islands he distinguished five species of lizards instead of 
the three mentioned by Darwin. The gigantic tortoises, 
moreover, he thought, were referable to five distinct forms, 
every one of them inhabiting a different island. He did not 
speculate on their origin or past history, but mentioned that, 
although large land-tortoises are now restricted to the two 
widely separated regions of the Galapagos islands on the 
one hand, and the Mascarenes and Seychelles on the other, 
they formerly had an extensive range. In the Miocene Period 
they lived in India, southern Europe and in North and South 
America. We may consequently look upon the few survivors 
as relicts of Tertiary times. J 

More recent researches on the reptiles show that there are 
far more distinct forms than was anticipated even by Dr. 
Gunther. Mr. Garman alludes to twenty-two species, among 
them several geckos. He does not enter into a discussion as 
to their origin ; he merely states that two theories have been 
advanced to explain the affinities of the fauna. 

The Hopkins- Stanford Expedition of 1898 made such ex- 
tensive collections on the islands that many additional forms 
were discovered among the material sent to the Museum of the 
Stanford University. Mr. Heller, who describes the reptiles, 
mentions twenty-five species and several varieties. Nine of 
these are giant tortoises, fifteen are lizards, and there is one 
snake. Among the lizards there are six geckos (Geckonidae) 

* Darwin, C., " Journal of Eesearches," pp. 278284. 

t Gunther, A., " Zoological Collections from Galapagos," p. 66. 

t Gunther, A., " Gigantic Land-tortoises," p. 253. 

Garman, S., " Eeptiles of Galapagos Islands." 


belonging to the two genera Gonatodes and Phyllodactylus. 
The single species of Gonatodes is most closely allied to a 
species found in the West Indies, while the genus also occurs 
along the coast of Ecuador and Peru. Of the five species 
of Phyllodactylus, four are peculiar to the Galapagos archi- 
pelago and more or less nearly related to P. tuberculosus, the 
non-peculiar species being distributed along the west coast 
of Mexico, Central and South America as far as Ecuador. 
All the species of Tropidurus, a genus belonging to the family 
Iguanidae, are peculiar and closely related forms, most of 
them being confined to different islands. The genus is prin- 
cipally found in Peru and Ecuador. The two very large lizards 
which Darwin observed really belong to distinct genera. Two 
species of Conolophus are now recognised, while the aquatic 
form still retains the original name of Amblyrhynchus cris- 
tatus. , Both genera of Iguanidae are quite peculiar to the 
archipelago and distantly related to a South American genus. 
Finally, the snake Dromicus biserialis, which is met with on 
seven of the islands, is nearly akin to D. chamissonis of Chile 
and Peru. 

The reptiles thus exhibit affinity with those inhabiting 
western South America, Central America and the West Indies. 
If they had reached the Galapagos islands by means of marine 
currents, as has been alleged, one may be tempted to ask the 
same question as I put before Why should practically all the 
species belong to ancient types ? Giant tortoises seem to have 
become extinct on the continents in Miocene times, while the 
lizards of the archipelago belong to the two families Gecko - 
nidae and Iguanidae. Although no fossil representatives of the 
Geckonidae are known, the resemblance of their vertebrae to 
those of the Palaeozoic Microsauri has been commented upon 
as indicating their ancient lineage. The genus Phyllodactylus 
occurs in tropical America, in Africa, Madagascar, Australia 
and on Norfolk and Lord Howe's islands in the Pacific. A 
single species inhabits the western Mediterranean region. The 
distribution of the genus, therefore, is suggestive of a very 
remote origin. Of the Iguanidae, as I have already pointed 
out, (p. 127), we know that they lived in America in Creta- 
ceous, and in Europe in Eocene times, while at present they 
are confined to North and South America with the outlying 


Tres Marias, Galapagos and Chiloe islands. Beyond this 
centre of distribution there are only two remote stations, one 
in Madagascar, the other on the Fiji islands. We could not 
have a better example of discontinuous distribution, which, 
as Dr. Wallace has always urged, is a proof of antiquity. 
Lastly, the Galapagos snake belongs to the ancient family 
Colubridae, which also includes Tropidonotus, a genus 
specially alluded to (pp. 128 and 222). 

In spite of these facts, Mr. Heller * expresses the opinion 
that the archipelago is mostly of Tertiary age, and that the 
fauna has been chiefly derived from material carried by ocean 

Still another expedition has lately been sent to the archi- 
pelago by the California Academy of Sciences. Dr. Van 
Denburghf who was asked to report on the reptiles feels quite 
convinced that the islands must all at some former period 
have formed parts of a single land-mass. In a letter which he 
addressed to me, he implies that several species of snakes 
inhabit the islands. He states that "the closest relatives 
of the serpents of the Galapagos archipelago are native to the 
Bahamas, Greater and Lesser Antilles, Costa Eica and the 
whole of South America. This being true, the snakes of these 
localities must have had a common origin. Either the West 
Indian and Galapagos snakes have been derived from South 
America, or else all must be descendants of species occupy- 
ing a great central land-mass which has sunk below the level 
of the sea, leaving mere remnants in Central America, 
northern South America, the Antilles and the Galapagos. 
Either view implies a former land connection and a con- 
tinental origin of the Galapagos ophidian fauna. 1 cannot 
bring myself to share the opinion of those who believe that 
the fauna of the Galapagos has reached these islands by the 
more or less accidental agency of the winds and ocean 

When Dr. Stearns J reported upon the land and fresh-water 
snails inhabiting the Galapagos archipelago, he alluded to one 

* Heller, E., " Hopkins-Stanford Expedition : Keptiles," pp. 46 48. 
t Denburgh, John Van, " Preliminary Descriptions of Land-tortoises," 
p. 1. 

I Stearns, E. E. C., " Mollusk-fauna of the Galapagos," pp. 359370. 
L.A. X 


species of Hyalinia, a large number of Bulimuli, two Pupae, 
one Leptinaria, a Succinea and a Helicina as having been 
collected on the islands by previous visitors. He then dilates 
upon the various accidental means of transport, and concludes 
that the agency of winds and ocean currents is amply sufficient 
to explain the origin of the molluscan fauna. He acknow- 
ledges that the archipelago is situated within an almost wind- 
less area, where storms are of rare occurrence, yet maintains 
that during the course of ages the climatic features may have 
been different from those obtaining at present. He also 
emphasises the fact of the tenacity of life of mollusks and 
the length of time they are able to subsist without food, claim- 
ing that a theory which will fairly explain a good portion of 
the phenomena of distribution by such agencies and operating 
directly under our eyes, is preferable to those, however plau- 
sible and attractive, involving conjectural and remoter con- 

In his endeavour to support the theory of accidental dis- 
tribution as applied to the origin of the Galapagos fauna, Dr. 
Stearns seems to forget that no observation has ever been 
made clearly indicating the conveyance of mollusks or their 
eggs in drift-timber. Mollusks have never been found 
in the crevices or under the bark either of trees encountered 
on the sea or of those stranded on foreign coasts. Very 
shortly after the trees are stranded, they are no doubt invaded 
by all sorts of creatures seeking moisture and shelter ; but the 
theory of accidental distribution, so ably advocated by many 
eminent naturalists still involves a good deal of the conjecture 
which Dr. Stearns is so anxious to avoid. Although Dr. 
Ball * fully acknowledges that we possess no actual proof 
in support of the drift-theory, he also concurs with Dr. 
Stearns in the opinion that there can be little doubt of the 
land snails having been introduced to the Galapagos islands 
in that manner. In his review of the molluscan fauna of the 
islands, he mentions thirty-three species of Bulimuli, two 
species of Vitrea and one of Conulus, two Pupae, four 
Succineae, one Tornatellina (Leptinaria), one Helicina and 
finally one Endodonta. Of these forty-five species, only a 

* Dall, W. H., "Insular Land-shell Faunas." 


couple are probably common to the islands and the main- 
land. All the others are restricted and quite peculiar to the 
Galapagos archipelago.* 

The genus Bulimulus, to which most of the Galapagos snails 
belong, comprises mostly species with ovate large heavy shells, 
which would seem to be peculiarly unfit for accidental trans- 
port by marine currents, though many live almost habitually 
on shrubs. As defined by Dr. Pilsbry, the genus is restricted, 
as I have already indicated, to the western borders of South 
and Central America and beyond the latter as far north as 
Mexico and Lower California. Eastward it passes along the 
northern States of South America to Brazil, it extends to the 
Antilles, and from Mexico to the southern United States. 
Thus it is limited to North, South and Central America, the 
West Indies and Galapagos islands. The Melanesian genera 
Liparus and Placostylus are closely related. The range of the 
large family Bulimulidae is very similar, one group being con- 
fined to America, the other to the opposite side of the Pacific, 
from Tasmania in the south through Australia to China north- 
ward, and eastward to the Society islands. One is almost 
tempted to invoke a Pacific continent in explanation of this 
curious discontinuous distribution. It seems as if the family 
had originated in some central Pacific area and had taken 
refuge on the nearest lands east and west on the disappear- 
ance of its ancient habitat. But we need not consider this 
problem at present. 

There are three groups of apparently very closely related 
Bulimuli in Chile and Peru, in the Galapagos islands and in 
Lower California. The shelLin these groups certainly is very 
similar, being narrow and elongated, and possessing the same 
texture. All the same, Dr. Pilsbry f is of opinion that this 
similarity in appearance is not to be looked upon as indicating 
close relationship, but as a case of convergence produced by 
similar environments. He recognises three sections of Buli- 
mulus according to the sculpture of the apical whorls of the 
shell. The first of these lives in Argentina, Chile, Peru and 
Bolivia, the second in tropical America, the Galapagos islands 

* Ball, W. H., " Insular Land-shell Faunas : Additions." 
t Pilsbry, H. A., "Manual of Conchology," (Pulmonata), X., p. 127. 



and West Indies, the third in Mexico and the United 

The two species of Vitrea (Hyalinia) seem to b related to 
the Vitrea radiatula group, which, like Conulus, has a very 
wide range in the Old World and- the New. One of the Gala- 
pagos species of Pupa is apparently found in Ecuador also, 
the other is peculiar to the islands. The genus has a world- 
wide range. All the species of Succinea, a genus which has 
an enormous range, are confined to the islands. Tornatellina 
chathamensis is closely related to the Hawaiian Achatellina 
and belongs to a typically Polynesian genus. Helicina, as 
already pointed out (pp. 157 158), has evidently invaded 
America from the Pacific. Considering that it is an operculate 
shell, and therefore supposed to be specially fitted for trans- 
port across the sea by marine currents, it is surprising that 
only a single species of Helicina should have taken advantage 
of these facilities (^compare Fig. 11). Lastly, one species of 
Endodonta (E. helleri) has been observed on the archipelago. 
Of the Galapagos mollusks this is one of the most in- 
teresting, because the extremely primitive genus Endodonta 
is almost confined to the Pacific region, including New Zea- 
land, Tasmania, Australia, New Guinea and the Philippine 
islands. Only St. Helena, where it is also found, lies entirely 
outside the area. It is significant that not a single species is 
known from the American or Asiatic: mainlands or the West 

I cannot help thinking that the attempt to derive the Gala- 
pagos molluscan fauna from America is a mistake. The 
family Bulimulidae is Pacific in origin, rather than American. 
Succinea and Pupa have a wide range in the Pacific islands, 
Tornatellina and Endodonta are altogether Pacific, being un- 
known in America, Helicina is mainly Pacific, and it has 
entered America from the west. Only Vitrea and Conulus may 
be looked upon as typically American and Old World genera of 
great antiquity. With their exception all the other genera 
except two are more likely to have passed into America from 
the direction of the Galapagos islands than vice versa. The 
two others stopped short at the Galapagos archipelago and 
went no further east. 

Too little is as yet known of the insects of the islands and 


their distribution to be of much service in this study, but 
Dr. Giinther records a woodlouse (Cubaris galapagoensis), 
which is apparently related to a West Indian form. The genus 
Cubaris, which is otherwise confined to the West Indies, 
Florida, Central America, California and South America, is 
the American representative of Armadillo. The latter has its 
headquarters in southern Europe. Cubaris, no doubt, owes'its 
presence in America to the fact that some ancestor common 
to it and the European ^Armadillo crossed over by the trans- 
Atlantic land bridge described in the last chapter. 

On the whole, remarks Dr. Wallace,* the flora agrees with 
the fauna as indicating a moderately remote origin, great 
isolation, and changes of conditions affording facilities for 
the introduction of organisms from various parts of the 
American coast, and even from the West Indian islands and 
the Gulf of Mexico. As in the case of the birds, he continues, 
the several islands differ considerably in their native plants, 
many species being limited to one or two islands only, while 
others extend to several. 

A point of great importance in connection with the theory 
of dispersal of seeds by marine currents is the condition of 
the coasts of the Galapagos islands. The dried up desert 
aspect of the islands on landing has been commented upon by 
Darwin. Dr. Wolff gives us his impression of the same 
islands during the rainy season. The lower zone up to six 
hundred feet is only very sparsely covered by vegetation. The 
whole landscape has a greyish-brown appearance. A .Lantana, 
a couple of Crotons and a few Euphorbiaceae and Compositae 
are the principal members of this wretched flora. The only 
plants that thrive in this district are two giant Cactuses, 
reaching to a height of about twenty feet. They seem to 
select the driest and roughest spots, where nothing else will 
grow. Herbaceous plants are restricted in this lower zone 
to a few parched grasses and a few shrivelled diminutive 
plants. There are very large districts, moreover, where 
nothing whatsoever will grow, the ground being paved, as it 
were, with enormous blocks of lava. At a height of about 

* Wallace, A. E., " Island Life," p. 289. 

f Wolf, Th., " Die Galapagos Inseln," pp. 257260. 


seven hundred feet the scenery changes completely. A fresh 
and moisture-laden wind greets us from the south-east, the 
ground is covered with short dense grasses, evergreen, shady 
little woods are dotted about here and there, and the flora 
as a whole is quite different from that below. Those ac- 
quainted with the flora of Ecuador will feel as if transplanted 
to the woods of the Paramo region, which flourish at a height 
of about 9,000 feet. The most striking character of the 
flora is its strong endemism, every one of the larger islands 
having its peculiar species, which do not range to the 
others even when the islands are in close proximity to one 

From Dr. Wolf's graphic description we can gather valu- 
able information. In its strongly endemic character the flora 
of the archipelago agrees perfectly with the fauna. As among 
the animals so it has been noticed among the plants that the 
various islands all possess their own species, though the latter 
are related to one another. But, in answer to our inquiries 
into the causes of this singular distribution, we are told by 
Dr. Wolf that he concurs with Darwin in the belief that the 
islands received their plants, as well as their animals, by acci- 
dental means of transport. He does not specify these means 
of transport. Darwin* makes it perfectly clear to us that 
winds could not have played any part in it. " As the archi- 
pelago is free to a remarkable degree," he says, "from gales 
of wind, neither the birds, insects, nor lighter seeds, would 
be blown from island to island." If the seeds were carried 
by marine currents from the mainland to the archipelago, 
how were they afterwards conveyed from the inhospitable 
shores of the islands across the almost absolute desert of 
the lowlands to the higher level ? We are unacquainted with 
any forces except wind, which would carry the seeds to a 
height of seven hundred feet, but Darwin expressly tells us 
the islands are remarkably free from wind. Moreover, we 
should expect the plants found at that height in the Galapagos 
islands to agree to some extent with the flora of the lowlands 
of the continent. Few, if any, seeds carried down by rivers 
would come from the highlands of the Andes. Yet the Gala- 

* Darwin, C., "Journal of Researches," p. 290. 


pagos flora possesses affinities with the highlands rather than 
the lowlands. 

In this brief account of the main features of the fauna and 
flora of the Galapagos archipelago I have endeavoured to state 
mainly the opinions of those who agree with the theory of 
elevation a,s propounded by Darwin. Some authorities formu- 
lated no decisive views on the subject. It is currently believed 
that the first criticism of this theory was brought forward by 
Professor Baur * in 1890, in a paper on the variation of the 
genus Tropidurus in the Galapagos islands. Although it was 
the first serious criticism, Mr. Andrew Murray f remarked 
much earlier that Darwin's observations on the Galapagos 
fauna had led him to an entirely different conclusion. " The 
American type of the whole group," he says, " speaks pri- 
marily of connection with the continent. The family facies 
of the group inter ,se, speaks of a period when the whole 
islands were separated from America, but united to each other. 
The endemic peculiarity of the species of each individual 
island tells of subsequent separation and change wrought 
in each, probably at the same time, by the alteration of climate 
from continental or terrestrial, to isolated and oceanic." 

It was not until 1891, when a more popular account of 
Professor Baur's views appeared in the " American Natu- 
ralist," that his opinions led to considerable discussion both 
in America and in Europe. It had seemed as if Darwin's 
theory, supported as it was by Wallace, Hooker, Agassiz and 
many other naturalists of less note, was unassailable. Never- 
theless, Professor Baur's careful reasoning induced many sub- 
sequent writers to ad/>pt his views in preference to older 
ones. He urged with Dr. Wallace that all islands may be 
divided into Continental and Oceanic ones. The first have 
developed from continents or larger bodies of land through 
isolation or subsidence. The second have not been so formed, 
but have arisen from submarine portions of the earth by eleva- 
tion. He thought that the fauna and flora of the first group of 
islands would be more or less harmonic, that is to say, the 
islands would be like satellites of the continent from which 
they took their origin. The fauna and flora of the second 

* Baur, G., " Variieren der Eidechsen-Ofattung Tropidurus." 
t Murray, A., " Distribution of Mammals," p. 17. 


group, however, ought to be disharmonic, for it will be com- 
posed of quite a different mixture of animals and plants, since 
they have been introduced accidentally from other land areas. 
Accidental immigrants will also reach the continental islands 
eventually in a similar manner. Thus continental islands 
may be composed of two f aunal and floral elements, an original 
and a secondary one. Oceanic islands, on the other hand, can 
only contain a secondary faunal and floral element, and will, 
therefore, not exhibit harmonic distribution. Since the dis- 
tribution of the animals and plants on the Galapagos islands 
is harmonic, just as it is on the West Indian islands, both of 
them, according to Professor Baur, represent the remains 
of land-masses that have been split off from the adjoining 
continent, and they have not risen from the floor of the 
ocean. Though agreeing in his final conclusions with Pro- 
fessor Baur, I am unable to follow him in regard to his defini- 
tions of " harmonic " and " disharmonic." Some of the 
Pacific islands are looked upon by Professor Baur himself as 
remnants of a Pacific continent, others, I believe, are not. 
Where is the exact distinction between them in the nature of 
their fauna ? 

A few years later Professor Baur * once more returned to 
the subject in order to answer the objections brought against 
his theory by numerous critics. He also intended to make some 
remarks on the geological history of the Pacific Ocean, and 
from his preliminary observations it is evident he believed 
in its comparatively recent age ; but before he could complete 
his interesting series of articles his sudden death unfortu- 
nately deprived us of further communications. From his 
writings, however, we know that Professor Baur believed that 
the Galapagos archipelago represented the remnants of a large 
peninsula which joined the West Indies by way of Cocos 
island and Central America. As some of his critics maintain 
that in his views Professor Baur stands almost alone, and is 
opposed by most of those who have discussed the fascinating 
problem of the origin of the Galapagos archipelago, a few 
remarks may be permissible in support of this theory. 

Alluding to the gigantic land-tortoises which have survived 

* Baur, G., "New Observations on the Galapagos Islands." 


on the Galapagos and various other islands, Dr. Gadow * asks 
the pertinent question Where did all these creatures come 
from, and how did they get to these oceanic islands ? Acci- 
dental transport is out of the question, as land-tortoises are 
drowned within a few hours after immersion in water. Since 
none of their kind inhabit the great continents now, although 
they lived there formerly, we are constrained to assume that 
the existing species are the survivors of tortoises which pro- 
ceeded to the islands when they were connected by land with 
the adjoining continents. In another place Dr. Gadow quotes 
the Galapagos tortoises as supporting the hypothesis of an 
Oligocene extension of land, considerably to the west and 
south of the present Central America. In his maps illus- 
trating the palaeogeography of Middle America, he actually 
joins the Galapagos islands with the Antilles, as Professor 
Baur had done, by way of Cocos island. f 

Professor Boettger J approves of the results obtained by 
Professor Baur's methods of investigation, arguing that the 
facts of distribution are much more easily explained by the 
assumption of a former land connection between the Gala- 
pagos archipelago and Central America than by the theories 
hitherto accepted. 

While in general agreement with the same views, Dr. 
Sarasin suggests that the ancient land bridge lay in a north- 
eastward direction, that is to say, from the Galapagos islands 
towards Mexico rather than to Ecuador. 

The fossil marine mollusks and their distribution are em- 
ployed by Dr. von Ihering || in support of the same theory. 
He draws attention to the remarkable fact that the species 
of the Tertiary Chilean deposits only appear on the Cali- 
fornian coast in Pleistocene times, and that similarly the 
Californian Tertiary forms only reached the coasts of Chile 
about the same time. On these grounds alone Dr. von Ihering 
bases his theory of a Tertiary peninsula extending westward 
into the Pacific Ocean and embracing the Galapagos islands, 

* Gadow, H., " Amphibia and Keptiles," p. 373. 

t Gadow, H., "Mexican Amphibians and Keptiles," p. 211. 

J Boettger, O., " Baur's Differentiation of Species," p. 462. 

Sarasin, F., " Die Fauna der Galapagos Inseln," p. 293. 

|| Ihering, H. von, "Fauna der Neotropischen Eegion," p. 296. 


which, thus acted as a barrier in separating the North 
American from the South American fauna (see Fig. 17). 

That the distribution of reptiles and mollusks on the Gala- 
pagos islands points to a former land connection with the 
American mainland is especially urged by Dr. Arldt,* but, 
like Dr. Ortmann, he thinks the islands must have been iso- 
lated since the end of the Mesozoic era. The land connection 
referred to is shown in Dr. Ortmann's f map illustrating the 
distribution of land and water during the Upper Cretaceous 
Period (Fig. 15). 

As regards the flora, Messrs. Kobinson and Greenman con- 
tend that no species to be found on the different islands illus- 
trates the noteworthy racial divergence in related forms better 
than Euphorbia viminea. This species differs markedly in 
foliage from any other known member of the large genus 
Euphorbia and is characteristic of the Galapagos archipelago. 
Being essentially a desert plant, it can subsist even upon those 
islands of the archipelago which are of low altitude and do 
not attain the upper regions of moist atmosphere. The most 
cursory inspection of the forms from the different islands 
discloses marked variation in the contour, size, thickness, 
rigidity and colour of the leaves, as well as in the length of 
the internodes, colour of the stems and so forth, whereas more 
careful examination shows that these are not mere individual 
differences, due to chance, state of development, or individual 
environment, but that each form appears in general to be 
restricted to a single island. 

The question consequently presents itself, according to 
Messrs. Kobinson and Greenman, J " If this archipelago is 
composed of islands of elevation, built up from the sea-floor 
independently by volcanic action, how has such a distribution 
been effected? " If the vegetation has been derived from the 
mainland by chance transportation of seeds, it is quite impos- 
sible to believe that each island has received a slightly different 
form of the same species, and we are forced to the much more 
natural assumption that racial and varietal divergence has 

* Arldt, T., " Entwicklung der Kontinente," p. 116. 
t Ortmann, A. E., " Distribution of Decapods," p. 381. 
t Robinson, B. L., and J. M. Greenman, " Galapagos Flora," pp. 135 


come about after the introduction of the species on the islands. 
Continuing the supposition that these are islands of elevation, 
the seeds of Euphorbia viminea must have reached them in 
one or two ways. Either every one of the nine islands where 
we know the species now to occur must have received its seed 
directly from the mainland, or, what is much more natural, 
seed must have reached one or more of the islands and from 
these have spread to the rest. That the same species should 
have reached all these islands presupposes a considerable 
facility of transportation. But as soon as this is granted, 
it is impossible to understand the highly individual develop- 
ment of the forms on the different islands. For relative or 
complete isolation seems necessary to account for the racially 
divergent floras of the islands, and especially for the occur- 
rence of only one form on each island. On Dr. Baur's 
assumption of a former union between the islands, and sub- 
sequent separation by subsidence, the authors maintain that 
not only is an explanation of the facts possible, but the exist- 
ing flora of the archipelago is just that which would most 
naturally result from such an origin. A former union of the 
islands would account at once for the occurrence of identical 
ancestral species upon the different members of the group. 
The subsequent separation would give the needed isolation for 
varietal and racial divergence, while the latter could not have 
come about if a continual interchange of seed were taking 
place from island to island. 

Messrs. Robinson and Greenman's careful reasoning is 
just as well applicable to the birds of the genera Geospiza, 
Certhidea and Nesomimus, to the reptilian genera Tropidurus 
and Testudo, and to the snail Bulimulus as it is to Euphorbia 
viminea, and from a study of any of them we should come to 
precisely the same conclusion as these authors. Mr. 
Hemsley * thinks the biological data which we possess from 
the Galapagos islands are strongly in favour of Professor 
Baur's views, and he supposes the area on which the islands 
stand to have been continued eastward to the mainland of 

Professor Stewart does not produce any new data for or 

* Hemsley, W. Botting, " Insular Floras," VI. (A), p. 299. 


against the theory of a former land connection between the 
islands and the mainland. He scarcely discusses the problem 
from a botanical point of view, but assumes that the Gala- 
pagos islands are of oceanic origin.* 

Personally, I am in agreement with Mr. Andrew Murray's 
and Professor Baur's theory that the archipelago originated 
through subsidence of a land-mass which was connected with 
the mainland. From the latter the islands received the prin- 
cipal members of the fauna and flora, and these progressed 
from the one to the other on a solid land surface. Subse- 
quently, possibly owing to volcanic disturbance, the land sub- 
sided, leaving only the mountain tops with their animals and 
plants as fragmentary relicts of a rich and varied fauna and 
flora. So far I am in agreement with Professor Baur's theory, 
but I believe that the geological history of the Galapagos 
archipelago forms only part of a much larger and far more 
complex problem. In the first place, it is evident that the 
whole fauna and flora of the Galapagos islands cannot have 
originated in Central America or the West Indies. A large 
proportion of the animals and plants seem to have their 
nearest relations in Ecuador and Colombia. Others point 
to a still more southern or northern origin. If we derive the 
fauna and flora principally from a former land connection 
with the mainland of America, that connection must have had 
several branches. The affinities of the archipelago appear 
to be chiefly with Central America and Ecuador and much 
less with Chile. On the other hand, a certain relationship, for 
instance among the Bulimuli, exists with the southern ex- 
tremity of Lower California. All these different areas of 
North, Central and South America may have been joined 
with the Galapagos land-mass by peninsulae, though one or 
more of them may have become disconnected long before the 
others. Dr. Arldt f indicates something of that kind in his 
map illustrating the conditions of land and water in Creta- 
ceous times. At least, he connects western Mexico and 
Ecuador by a broad land bridge westward of Central America 
and across the area of the Galapagos archipelago. I ventured 

* Stewart, A., " Botanical Survey of Galapagos Islands," p. 239. 
t Arldt, J., "Entwicklung der Kontinente," Karte 19. 


to propound the theory that North and South America might 
have been joined by means of a land bridge between Mexico 
and Chile across the Galapagos islands at a time when the 
greater part of South America was still submerged.* But 
that land connection is evidently one which must have ceased 
to exist in its entirety, at least in early Tertiary times (see 
Pig. 16). 

I have likewise alluded above to a supposed affinity of some 
of the Galapagos birds to species found in the Sandwich 
islands, pointing out that in other groups this relationship 
undoubtedly exists, although only to a slight extent. I have 
mentioned examples among the mollusks. Among plants the 
composite Lipochaeta is a good example. The genus is con- 
fined to the Sandwich islands, except one species which in- 
habits the Galapagos archipelago. Instances of affinity of 
the Sandwich islands with Mexico and Central America also 
occur, though as a rule the relationship is a remote one. 
Still, it has been tentatively suggested by several writers that- 
some time or other in the past Mexico or California were 
joined by a 'land bridge to these islands. It might be con- 
sidered somewhat beyond the scope of this work to discuss a 
faunistic relationship of this nature, but it really forms part 
of the great problem as to the origin of the older Asiatic stock 
on the American continent. 

What I described above (p. 97) as the latest or most recent 
immigration of Asiatic types into North America took place, I 
think, in Pliocene and perhaps early Pleistocene times. 
A large proportion of the more northerly animals and plants 
now living in North America readily reveal their Asiatic origin 
as slightly modified descendants of a stock once common to 
both .Asia and North America. These we must regard as 
the offspring of the latest Asiatic invasion. The American 
species which thus indicate their Asiatic lineage were des- 
cribed as Asiatic invaders principally in the early chapters. 
In the fourth chapter the time and the place of their invasion 
across Bering Strait were more fully discussed. Throughout 
the subsequent chapters I drew attention to the existence 
of severally geologically more ancient invasions from Asia, ex- 

* Scharff, R. F., "Early Tertiary Land-connection," p. 525. 


pressing my belief, contrary to the opinion of almost all those 
who had occupied themselves with this problem, that these 
earlier Asiatic immigrants took an entirely different route 
from the later ones. The opinion I had formed was founded 
on a variety of circumstances. 

If the theory I had formulated is correct, namely that the 
effects presented by the Glacial Epoch were due to the 
simultaneous closing of the Arctic Ocean from the Atlantic 
and Pacific by the formation of two land bridges, one 
of which connected North America with Europe in the 
east, while the other joined North America and Asia in 
the west (see Fig. 7), the warmer climate in pre-Glacial 
times must have been due to the fact that the Arctic 
Ocean then received a greater amount of heated water 
than it does now. That the two land bridges must have existed 
at nearly the same time seems to be amply demonstrated from 
biological evidence. During the Miocene Period the climate 
in the extreme north of Europe and North America must have 
been much milder than in Pliocene times. The evidence from 
northern Europe is against the supposition that the Gulf 
Stream entered the Arctic Ocean more freely during the 
Miocene Period than it does at present. The Japanese 
"Kuroshiwo," or possibly some even more powerful marine 
current, must have passed, therefore, entirely into the basin 
of the Arctic Ocean across Bering Strait. I remarked before 
(p. 96) that the appearance of certain Pacific species of 
mollusks in the English Crag deposits may perhaps be due 
to this cause. It was not until Pliocene times, according to 
Professor J. P. Smith,* that the marine faunas of Japan 
and the western coast of North America began to be re- 
markably similar, thus implying that a migration at that time 
took place along a continuous shore-line. The Californian 
Miocene marine deposits seem mostly to contain endemic 
species with a slight admixture of southern and circumboreal 
ones. All this evidence favours the view of an open Bering 
Strait in Miocene times, and a closed one during the Pliocene 
Period. Dr. Dall f states that the Kenai leaf beds in Alaska 
(now generally considered of Eocene or Oligocene age) are 

* Smith, J. P., " Periodic Migrations," pp. 225226. 
t Dall, W. H., "Correlation Papers Neocene," p. 251. 


succeeded by Miocene beds with a marine fauna. Similarly 
he describes a series of Miocene marine mollusks from sand- 
stones obtained at the head of the Gulf of Penjinsk on the 
opposite shore of Asia. As in Alaska, these Miocene sand- 
stones are apparently resting on leaf-bearing lignites, thus 
strengthening the assumption of a wide and freely open 
passage in the north, between the Pacific and Arctic Oceans.* 
Professor Schuchertf separates North America from Asia by 
a marine channel throughout Miocene and Pliocene times in 
his maps illustrating North American paleogeography. All 
the evidence, says Mr. Knopf,J from which conclusions of 
some positiveness can be drawn, record only epochs during 
the Tertiary Era of more widely spread submergence and 
increased separation of the continents. And yet his exten- 
sive studies of the mammalian fauna led Professor Osborn 
to the conclusion that the emergence of continents progressed 
during the Miocene Period and that North America was 
broadly united with eastern Asia. How can wo reconcile these 
diametrically opposite views ? I have shown that in Oligo- 
cene times, or during part of that Period,; a trans- Atlantic land 
connection probably enabled the Old World types to travel to 
North America. A Bering Strait land bridge is not essential, 
therefore, in explaining existing Oligocene or Eocene affini- 
ties between the Old World and the New. But I have given a 
large number of instances among North American plants as 
well as animals, indicating a direct migration either from Asia 
to North America or vice versa, in early and late Tertiary, at 
any rate in pre-Pliocene times. I need only allude again to 
the close relationship of the hellbender of the eastern States 
to the Japanese giant salamander, of the blue-tailed skink 
of the eastern States and Japan, and of the absolute identity 
of the American and Japanese ground lizards (Lygosoma 
laterale) . The only living relation of the American alligator 
inhabits the Yangtse river in China ; the nearest akin to the 
American green snakes (Liopeltis and Cyclophis) reside in 
south-eastern Asia. The family of snapping turtles (Chely- 

* Ball, W. H., "Miocene Fauna in Arctic Siberia," p. 473. 

t Schuchert, C., " Paleogeography of North America," Maps 98100. 

J Knopf, A., " Tertiary Land-connection," p. 419. 

Osborn, H. F., "The Age of Mammals," p. 244245. 


dridae) are confined to America and New Guinea, the Aglypha, 
a division of the Colubrine snakes, inhabit south-eastern Asia 
and Papuasia, except one genus which is restricted to Central 
America. The pythons are mainly Australian and southern 
Asiatic, but a single species (Loxocemus bicolor) lives in 
southern Mexico. The American land-snail Bulimulus, as 
we have seen, has its nearest relation (Placostylus) in 
Australia and the islands of the western Pacific. The geogra- 
phical distribution of the land isopods belonging to the group 
of Spherilloninae, from New Zealand, Australia, Polynesia, 
south-eastern Asia and Japan to south-western North 
America, seemed to Mr. Budde-Lund * so very peculiar that 
he invoked a Pacific continent to explain it. As a matter of; fact 
we have very numerous similar instances indicating a striking 
faunistic affinity between North America and eastern Asia. 
But these examples belong to an entirely different category 
from the animals and plants referred to as forming the later 
Asiatic immigration. Certain closely allied or even identical 
forms such as the alligator, the blue-tailed skink and the 
ground-lizard in America and eastern Asia apparently point 
to a geologically recent faunistic interchange between these 
countries. But alligators are known even from the earliest 
Tertiary deposits, while the range of the ground-lizard 
(Lygosoma) from New Zealand, Australia and southern Asia 
as far north as Japan, indicates that the genus is probably of 
Mesozoic origin, and that we have to deal with a remarkably 
persistent ancient type. 

When we take a general survey of the range in North 
America of these older Asiatic animals (and the plants per- 
fectly agree with them), we find that their number decreases 
as we go north -westward, but increases towards the south- 
west and Central America. The whole stream of these ancient 
Asiatic immigrants seems to have issued forth from the south- 
west, precisely from the same part of North America that 
also yielded the European colonists. It is in south-western 
North America, I think, that we have to search for indications 
of the older land connection with Asia, and not in the Bering 
Strait. Another point that strikes us during our general 

* Budde-Lund, G., "Kevision of Crustacea Isopoda Terr.," II., p. 40 


survey is that the American and Asiatic forms related to one 
another are in Asia mainly confined to the south-eastern 
border generally, or to Japan. Australia, New Zealand and 
New Guinea also possess a large number of species and genera 
very closely allied to west American ones. These seem rarely 
to extend further north in Asia than Japan. Are we to throw 
a bridge across the Pacific ocean from Japan in order to find a 
possible explanation of this former invasion of south Asiatic 
and Australian types into western America, or does any other 
theory meet all the facts of the case ? 

Since Dr. Augustus Gould * first mooted the hypothesis of 
a former Pacific continent about sixty years ago, the idea has 
been widely discussed by biologists and geologists. Mr. 
Murray's f attention seems to have been first drawn to the 
subject by the occurrence of the beetle Meristhus scrobinula 
in Mexico and China, and by the presence of the mole Urotri- 
chus in California and Japan. The Japanese and Californian 
moles are no longer placed into the same genus, but no one 
doubts that the American Neiirotrichus and Japanese Urotri- 
chus are very nearly related to one another and that they 
must have had a common ancestor. 

The supporters of the theory of the permanence of our great 
ocean basins explain such cases by means of a former Bering 
Strait land bridge, but, as already stated, the solution of this 
problem must be sought elsewhere. That several of the pre- 
mises on which the theory of the permanence of ocean basins 
is founded are incorrect has been demonstrated (pp. 274 
277). Darwin's theoretical considerations on the formation 
of coral reefs and atolls, which demanded a long continued 
subsidence of the mid-Pacific region, have been amply veri- 
fied. To put his subsidence theory to a practical test Darwin 
suggested that a boring should be made into one of the cores 
of an atoll. Through the perseverance aoid energy of Pro- 
fessor Sollas J and Professor Edgworth David a boring on 
Funafuti atoll was carried to a depth of one thousand one 
hundred and fourteen feet, where cores were obtained showing 

that the whole mass of rock was composed of pure coral. Since 

* G-ould, A., " Kemarks on Mollusks," p. 78. 

t Murray, A., " Geographical Eelations of Coleopterous Faunas," p. 37. 

t Sollas, W. J., "The Atoll of Funafuti." 
L.A. Y 


the organisms that form coral reefs are unable to live at a 
greater depth than one hundred and fifty feet, it is manifest 
that the floor of the ocean must have subsided very slowly 
and continuously, thus enabling the reef-building corals to 
raise their structures to the requisite depth of water. The 
verdict of Funafuti is thus clearly and unmistakably in favour 
of Darwin's theory of subsidence. The fact that a slight local 
elevation seems to be taking place in some parts in no way 
detracts from the truth which has been so firmly established. 

It would lead me too far from my main object to allude to 
the numerous papers that have been written on the Pacific 
Continent controversy. After Gould and Murray it was 
Captain Hutton,* I think, who again revived the theory, which 
he later on discussed in his presidential address to the Philo- 
sophical Institute of Canterbury in New Zealand. His idea 
was that New Zealand, eastern Australia and India formed one 
biological region in early Mesozoic times. In Lower Cre- 
taceous times a large Pacific Continent extended from New 
Guinea to Chile, and from the latter a long lobe of land 
stretched southward to New Zealand. This Pacific Continent, 
in his opinion, supported plants, insects, snails, frogs, some 
lizards, perhaps snakes and a few birds, but no mammals. 
Later on, during the Cretaceous Period, New Zealand became 
separated, while the Pacific Continent broke up. 

More recently Dr. von Iheringf alluded to a Pacific Con- 
tinent which he believes to have gradually subsided during 
the Mesozoic Era, but without going into further details as 
to its nature and size. I may mention that the supposed 
antarctic land connection between Patagonia and New Zea- 
land is a subject which I am not dealing with at present. 
Dr. Pilsbry assailed the problem entirely from the point of 
view of the molluscan distribution. He points out that many 
genera of land-snails reach back to the Oligocene Period 
unchanged save in specific characters, and that the modern 
family groups of these snails undoubtedly diverge far back 
in Mesozoic time. Now it is a most significant fact that the 
Pacific islands are almost entirely tenanted by the most primi- 

* Hutton, F. W., " Origin of Fauna and Flora of New Zealand." 
t Ihering, H. von, "Eelations between New Zealand and South 
America," p. 444. 


tive and oldest groups of land-snails. And, as Dr. Pilsbry 
facetiously puts it, "it is very easy to show that snails may 
have been carried from place to place by a hurricane, a float- 
ing tree or ' floating island/ or that their eggs may find room 
in the pellet of earth clinging to a bird's feather, but it is in- 
cumbent upon the theorist who peoples the mid-Pacific islands 
by such means to show why such dominant groups as the Heli- 
cidae, Bulimulidae, Bhytididae, Streptaxidae in fact the 
whole Holopoda and Agnathomorpha, with the higher mem- 
bers of the aulacopod families, as well as the higher opercu- 
lates should have utterly failed to take advantage of these 
means of transport." Instead of being a faunal dependency of 
the Australian or Oriental regions, Polynesia has every ap- 
pearance, says Dr. Pilsbry, of being a region which started 
with a fauna long antedating the present Australian and 
Oriental faunas, developing along its own lines, retaining old 
types because they did not come into competition with the 
higher groups of animal life. Dr. Pilsbry's conclusion is that 
a Pacific Continent existed, which was finally separated from 
other lands as early as the middle of the Mesozoic Era, and 
that the northern portion became disconnected when the 
remainder was still joined to the mainland.* 

A careful review of the distribution of the ants and lizards 
in particular led Professor Baur f to formulate the theory 
of a former Indo-Pacific Continent extending from Malaysia 
to the west coast of America. He looked upon the Pacific 
islands as the last remnants of this continent, which still 
existed, he thinks, until the commencement of the Miocene 

Mr. Hedley,$ who took part in the famous Funafuti Ex- 
pedition, and has had the advantage of studying the problems 
of dispersal on the spot, altogether disbelieves in a Pacific 
Continent in the sense of Baur, Pilsbry and Button, but he 
suggests that New Zealand was formerly connected with Aus- 
tralia by way of New Caledonia, the New Hebrides, Solomon 
islands and New Guinea. Even the Fiji islands come within 
this scheme. Some years earlier he had already demonstrated 

* Pilsbry, H. A., " Genesis of Mid-Pacific Faunas," pp. 569578. 
| Baur, G., "New Observations on the Galapagos Islands," p. 869. 
| Hedley, C., " Zoogeographic Scheme," pp.400 405. 



the essential unity of the area inhabited by the snail Placo- 
stylus,* thus regarding New Zealand and the archipelagoes of 
Solomon, Fiji, New Hebrides, Loyalty, New Caledonia and 
Lord Howe as portions of a shattered continent which he 
called the " Melanesian plateau." The short review on the 
subject by Dr. Holdhausf in favour of a Pacific Continent in 
Tertiary times adds little to the arguments already advanced 
by Professor Baur. 

A most determined and thoroughgoing onslaught on the 
theory of a supposed former Pacific Continent was recently 
made by Mr. Guppy. It seems scarcely fair to compare the 
results of his studies with those just alluded to, because 
he derives his arguments almost altogether from the flora 
of the Pacific islands, while the others were founded on 
faunistic data. Still, Mr. Guppy' s works J contain a great 
deal of personal observation, and his careful labours in 
this particular field of enquiry will be sure to attract 
the serious attention of the student of geographical dis- 
tribution. His discussion of the subject is disappointing! 
in some respects. " If the distribution of a particular 
group of plants or animals does not accord with the pre- 
sent arrangement of the land," he remarks, " it is by far 
the safest plan, even after exhausting all likely modes of 
explanation, not to invoke the intervention of geographical 
.changes." A little further he explains " I scarcely think 
that our knowledge of any one group of organisms is ever 
sufficiently precise to justify a recourse to hypothetical altera- 
tions in the present relations of land and sea." In reading 
such passages one wonders whether Mr. Guppy can have 
become acquainted with the science of geology, or with the 
principles that underlie the geographical distribution, for 
example, of mammals. His opinions differ radically even from 
those of Dr. Wallace, who cannot be said to have been unduly 
rash in any of his conclusions as to former changes of land 
and water. One would also expect from Mr. Guppy an abun- 
dance of important facts concerning the dispersal of seeds 
by the various means of accidental transport. But he tells us 

* Hedley, C., " Eange of Placostylus," p. 339. 

t Holdhaus, K., " Argumente f. d. Existens eines pazif. Kontinents." 

t Guppy, H. B., " A Naturalist in the Pacific," I., p. 380. 


that the story of plant distribution in the Pacific is bound 
up with the successive stages of decreasing activity in the 
dispersing agencies. He thinks that the area of active dis- 
persion, as illustrated by the non-endemic genera of plants, 
at first comprised the whole of the tropical Pacific.* After- 
wards it was limited to the southern Pacific, and finally to the 
western Pacific only. The birds that carried seeds all over this 
ocean became more and more restricted in their range, pro- 
bably, as Mr. Guppy suggests, on account of increasing 
diversity of climatic conditions. The plants of necessity re- 
sponded to the ever narrowing conditions of bird-life in this 
ocean, the differentiation of the plant and bird taking place 

Mr. Guppy dislikes the idea of hypothetical alterations in 
the present relations of land and water, and yet what an 
amount of hypothesis he has to resort to in his endeavour to 
explain the theory which he is so anxious to support ! There 
is not a scrap of evidence for the belief that dispersing 
agencies have practically ceased at the present time, nor that 
the migrations of birds have diminished. On the contrary, 
ornithologists have done their utmost to trace the origin of 
all bird migrations to the gradually increasing diversity of 
climatic conditions during the Tertiary Era, which they sup- 
pose to have culminated in the Glacial Epoch. Although we 
have little evidence for such a belief, it seems a reasonable 
supposition. But what can be said in favour of Mr. Guppy's 
theory, that, owing to the increasing diversity of climatic con- 
ditions, the seed-carrying birds have become more restricted, 
that is to say, less migratory ? 

A careful perusal of Mr. Guppy 's work must make it evident 
to anyone, that, although plants have far greater facilities for 
accidental transport, and have in many cases actually been 
thus conveyed from one land surface to another, they agree 
on the whole perfectly with animals, in so far as the Pacific 
islands are largely tenanted by very ancient types. If we adopt 
the theory of accidental dispersal for the origin of the Pacific 
island flora, we must apparently assume that the means of 
occasional conveyance were far more efficient in former times 

* Guppy, H. B., "A Naturalist in the Pacific," IL, pp. 519520. 


than they are at present. Even Mr. Guppy, however, some- 
times relents and partly withdraws from his precepts when 
confronted by really difficult cases of distribution. Thus he 
acknowledges that the conifer Dammara vitiensis, which grows 
on the Fiji islands, is unfitted for accidental dispersal by any 
of the known modes of conveyance. The genus Dammara is 
confined to New Zealand, eastern Australia, New Caledonia, 
the New Hebrides and Fiji islands. From this region it ex- 
tends westward to Java and Borneo, the centre of distribution 
being in the western Pacific. The absence of the genus from the 
neighbouring Samoan and Tongan groups is, as Mr. Guppy * 
remarks, very significant, and it is evident that the ordinary 
agencies of dispersal by birds, winds or currents have here 
failed to extend Dammara over a few hundred miles of sea. 
For once Mr. Guppy concedes, therefore, that the present 
relations of land and sea do change sometimes, and that, 
" nolens volens," we must admit that Dammara may well be 
cited in support of any continental hypothesis affecting the 
western Pacific. Later on, in fact, he expresses the opinion 
that the Fiji islands mark the site of a Mesozoic continental 
area in this region. 

There is thus a certain amount of distributional evidence 
in favour of the theory of the existence of a large land sur- 
face in the western Pacific. Whether the remainder of that 
ocean was ever completely occupied by land is a more difficult 
question to answer. But even on the distant Marquesas 
islands granites and gneisses occur, as I mentioned before. 
In the tuffs of the Kermadec islands numerous boulders of 
hornblende granite have been found. New Caledonia consists 
of an ancient series of mica schists and slates with a general 
north-easterly strike. There are also shales containing fossils 
identical with those of the New Zealand Trias, followed by beds 
of coal of Jurassic age. Gneisses, crystalline limestones and 
serpentines, like those of New Caledonia, are reported from the 
New Hebrides. Crystalline schists, granular limestone, 
granite, diorite and gabbro have been discovered on the Fiji 
islands. The occurrence on the Tonga group of fragments 
of garnet, tourmaline and uralitic gabbro suggests the close 

* Guppy, H. B., " A Naturalist in the Pacific," II., pp. 297306, 


proximity of an area of metamorphic rocks, while a series of 
plutonic rocks have recently been described from Tahiti. 

Thus, according to Mr. Speight,* there appears to be geo- 
logical evidence of the former extension of continental con- 
ditions over a large area of the mid-Pacific region. As he 
remarks, it is highly probable that many volcanic islands 
classified as oceanic will ultimately have to be looked upon 
as built up on the remnants of a continental area. We may 
imagine that a large land area or continent covered the greater, 
part of the present Pacific Ocean in Palaeozoic and early 
Mesozoic times, and that there was a subsidence during later 
Mesozoic and Tertiary times with more recent local elevations. 

Professor Haug,f discusses the Pacific problem from 
another point of view. His studies of the geosynclinals, which 
he calls the essentially mobile regions of the earth's crust, 
led him to the conclusion that the circumpacific geosynclinal 
implied the former existence of a continent in place of the 
present Pacific Ocean. 

The well-known parallelism of the different groups of Pacing 
islands has likewise been utilised in support of the same 
theory. It may be explained by the supposition that these 
islands are either the remnants or the initial stages of a 
series of mountain chains. J The Funafuti boring results seem 
to point to the first of these as the more likely assumption. 

That Darwin's theory of subsidence still meets with a good 
deal of determined opposition by the believers of the per- 
manence of ocean basins may be gauged from Sir John 
Murray's writings on the structure and origin of coral reefs. 
I think it is unnecessary for me to discuss the bearings of his 
arguments on the American problems raised in this chapter, 
because, in the first place, it seems probable that both Mur- 
ray's hypothesis of elevation and Darwin's of subsidence may 
be applicable to certain cases, and, secondly, because a Pacific 
continent in the sense of Hutton, Pilsbry and Baur cannot 
evidently be cited in support of most of the older Tertiary 
affinities between Asia and North America that I have alluded 

* Speight, E., " Petrological Notes on Eocks from Kermadec Islands," 
pp. 244250. 

f Haug, E., " Geosynclinaux et aires continentales," p. 646. 
J Arldt, T., " Parallelismus d. Kiisten v. Sudamerika." 


to. If a Pacific continent existed, and I quite concur with 
those who are of that opinion, it must have largely subsided 
before the Tertiary Era. It seems to me as if the central 
part of it had broken down gradually, the margins slowly 
following suit, both on the eastern and western Pacific, only 
leaving here and there a few remnants which either remain 
as isolated pillars far out in the ocean or have become joined 
to more recent land-masses. I imagine that the latest pre- 
Pliocene land connection between North America and Asia was 
not the Pacific Continent, but merely its margin, which per- 
sisted probably until Oligocene or Miocene times. In a geo- 
logical sense, remarks Dr. von Drasche,* it is more correct 
to draw the western boundary of the Pacific Ocean through 
Kamchatka, Japan, the Philippines, New Guinea and New 
Caledonia, because they all possess old crystalline or ancient 
sedimentary rocks. But the oceans, as Professor Walther f 
has pointed out, are areas of depression surrounded by folds or 
flexures which give rise to extravasation of eruptive material. 
The chain of the volcanic Aleutian islands lie in such a fold. 
Near the east coast of Japan the depth greatly increases. On 
the eastern side of the Pacific, in western North America, 
the igneous rocks skirt the coast for some distance, whereas 
in the south-west the volcanic centres lay far inland, justi- 
fying the assumption that the Tertiary coast-line extended 
some distance inland, which is fully established by geological 
observation. Although the Pacific is known to have invaded 
Californian territory, there is no evidence that the coast hills 
and outlying islands were covered by the sea ; and these pro- 
bably remained as part of the marginal land which skirted the 
west coast of North America. It is from this old land, I think, 
which contained Asiatic immigrants, that North America 
received its ancient Tertiary fauna from Asia. I suggest, 
therefore, that in early Tertiary times a belt of land, possibly 
representing the margin of the more ancient Pacific Continent, 
extended from the south-west coast of North America in a 
great curve to Japan and further south (see Fig. 14). The 
extraordinary similarity of the east Asiatic, Mesozoic and 

* Drasche, E. von, " Palaeozoische Schichten auf Kamtschatka," p. 268. 
t Walther, J., " Uber den Bau der Flexuren, &c." 


early Tertiary marine faunas to those of California would 
thus receive a satisfactory explanation without invoking a 
land connection across Bering Strait. As soon as the marine 
channel which separated the coast hills in California from the 
rest of the country disappeared, a number of Asiatic immi- 
grants entered North America. But the flora, especially of the 
small islands lying off the coast of California, still bears the 
impress, as Mr. Greene* has pointed out, of belonging phyto- 
geographically to another continent than America. 

I also mentioned 'that the European invasion of North 
America, which travelled by the trans -Atlantic land bridge, 
had ultimately entered the Continent from the south-west. 
The two elements, the Asiatic and the European, must have 
joined there eventually. To judge from purely faunistic testi- 
mony, that was evidently the course of events (compare 
p. 211). Somewhere about the Miocene Period extensive sub- 
sidence of the land west of California must have compelled 
the fauna and flora to seek refuge on the continent with which 
the Pacific belt of land seems to have become united. Palaeon- 
tological evidence gives us reason for such a supposition. Take 
for example the great land-tortoises. Their sudden appearance 
in south-western Miocene deposits suggests that they came 
from the west with other new-comers. This hypothesis like- 
wise throws light on their survival near at hand in the Gala- 
pagos islands, which no doubt once formed part of the Pacific 
belt of land alluded to. There are such a variety of problems 
connected with this theory that I shall defer the further dis- 
cussion of it till the next chapter. In conclusion, a few addi- 
tional remarks on the nature of the supposed extension of land 
west of Central America will facilitate the comprehension of 
the scheme of land connections that have only been roughly 
outlined so far. 

When I described the remarkable fauna and flora of the 
Cape region of Lower California (p. 207) and their marked 
affinities to those of the opposite coast of Mexico, I made no 
reference to the fact that this interesting assemblage of 
animals and plants is living in a hilly district being separated 
from the nearest mountains to the north of it by a wide extent 

* Greene, E. L., " Botany of Santa Cruz Island," pp. 377388. 


of a low -lying plain. This district, moreover, is entirely 
granitic and composed of a number of high ridges running 
parallel in an east- westerly direction, the remainder of the 
great peninsula being largely formed of calcareous rocks with 
mountain ranges running in a north and south direction. 
Between the two lies a great plain several hundred miles long 
with a height of scarcely one hundred and fifty feet above sea- 
level. The two mountain ranges manifestly belong to entirely 
different systems, and the junction between the two must have 
been a comparatively recent geological event. Mr. Eisen* 
was so much impressed by the supposed severity of the climate 
during the Glacial Epoch that he believed the whole Cape 
region was at that time wrapped in snow and ice and devoid of 
animal life. But he also contends that it must have been an 
island and that during its rise animals and plants gradually 
reached it from the mainland by accidental transport. That 
the Cape Region has only recently become part of Lower Cali- 
fornia is highly probable. To judge from the fauna and flora, 
it must have been connected by land with some part of Central 
America or southern Mexico, though it possesses affinities, too, 
with Asia and the Pacific islands (compare, p. 208). Rather 
more than half-way across the sea between the Cape Region 
and the south coast of Mexico lies the small group of tihe 
Tres Marias islands, and it might be argued that they 
had once formed the connecting link between the mainland 
and that faunistically so remarkable Cape Region of Lower 
California. The animals and plants of these islands, how- 
ever, although clearly showing that the islands have been 
joined to one another and to southern Mexico, exhibit no 
near relationship to those of the Cape Region.} Hence it 
is probable that the faunistic and floristic affinity between 
the Cape Region and southern Mexico is due to the fact 
that both regions have acquired their animals and plants, in 
more remote times, from the same source in Central America. 
I suggested in a former chapter (p. 287) that the moun- 
tains of Guatemala had once extended further westward. 
Guatemala certainly seems to have been a land surface 

* Eisen, G., " Explorations in the Cape Region," p. 735. 
t Nelson, E. W., L. Stejneger, and others, " Natural History of the 
Tres Marias Islands." 


since very remote geological times, while its ancient moun- 
tain system, trending in a west-eastward direction, abruptly 
terminates on the Pacific coast. I ventured to explain 
the many instances of curiously discontinuous distribution 
in Central America by the supposition that an ancient land 
occupied the adjoining portion of the Pacific, and that the 
present Central America is partly formed of the remnants of 
that land having eventually become moulded together by geo- 
logically recent volcanic deposits. If we assume that the Cape 
Kegion of Lower California belonged to that Pacific land 
which really formed the southern continuation of the north 
Pacific belt of land alluded to, a reasonable explanation is 
advanced for the faunistic affinity of the Cape Kegion to 
Central America, the West Indies and Asia, and its dissimi- 
larity with the rest of Lower California. The hypothesis of a 
former westward continuation of portions of Central America 
is not founded purely on zoogeographical considerations. 
In his reference to the older Tertiary sediments of Central 
America Professor Hill* states that it is impossible to avoid 
the conclusion that they were derived from a near-by 
land which existed at the time of their deposition. The only 
hypothesis, he says, that can fit the condition of their 
present lay and arrangement is that this land existed 
towards the Pacific coast or in the area now covered by the 
Pacific waters of the Isthmian Eegion. The " old land " 
or early representative of the isthmus of Panama, remarks 
Mr. Hershey,f lay mainly south of the present isthmus. 
That it was a land-mass of considerable extent, he con- 
tinues, is indicated by the heavy beds of conglomerate 
formed from it, and he argues that the peninsula of Azuero 
which projects out into the Pacific is a remnant of this land, 
while on its northern border were laid down the more recent 
formations which make up the main body of the isthmus. 
Geologically there is, therefore, some evidence, too, for the 
assumption of a former extensive land surface on the Pacific 
side of Central America. It must be remembered that while 
Central America as a whole has a south-east and north-west 

* Hill, E. T., " Geological History of Panama," p. 263. 

t Hershey, 0. H., " Geology of Isthmus of Panama," p. 249, 


trend the isthmus of Panama runs in an east-westward direc- 

That this Pacific land persisted to some extent until com- 
paratively recent geological times, seems to be indicated by 
several notable features. When Mr. Agassiz surveyed the 
ocean floor from the steamer " Albatross," he found the Pacific 
side of Panama faunistically poor compared with the Atlantic 
side, but he attributed that condition to the absence of a great 
oceanic current on the south side of the isthmus. The absence 
of deep-sea corals on the Pacific side was most striking. Yet, 
he remarks that there is on the west coast of Central America, 
oven in deep water, a considerable fauna, which finds its 
parallel in the West Indies and recalls later Cretaceous times 
when the Caribbean Sea was practically a bay of the Pacific. 
This assumption that the Caribbean Sea was once a bay of 
the Pacific has been put forward by other writers. The 
faunistic resemblance, or parallelism, as we might call it, 
between the seas on each side of Central America may be due 
to a former westward extension of the Atlantic Ocean just 
as much as to an eastward extension of the Pacific. However, 
one of the most remarkable features brought to light by the 
" Albatross " expedition was the condition of the ocean 
floor. There was not a station between Acapulco, on the 
coast of Mexico, and the Galapagos islands, according to 
Mr. Agassiz, of which the bottom could be characterised as 
(strictly oceanic. At the most distant points from the 
shore the bottom specimens invariably showed some trace of 
admixture of terrigenous material. All the way, even to a 
depth of 2,000 fathoms, the trawl became filled with a sticky 
mud containing logs of wood, branches, twigs and decayed 
vegetable matter. Being a firm believer in the permanence 
of ocean basins, Mr. Agassiz * naturally attributed this extra- 
ordinary condition of the sea floor to the existence of currents, 
which, striking Central America from north and south, are 
reflected in a westward direction. He likewise argues from 
this discovery that it offers a very practical object lesson re- 
garding the manner in which the Galapagos islands received 
their fauna and flora. The peculiar condition of the ocean 

* Agassiz, A., "Keportson 'Albatross' Expedition," pp. 1177. 


floor may, I think, be explained in an entirely different way. 
We might possibly expect a condition such as Mr. Agassiz 
describes off the mouths of the Amazon or Orinoco rivers, but 
there are no vast rivers anywhere on the Pacific coast to fur- 
nish all this vegetable detritus he speaks of. A similar condi- 
tion, moreover, occurs only to a very limited extent on the 
Atlantic side of the isthmus. That the ocean floor is covered 
with tree trunks, twigs and other vegetable detritus may be 
due to the existence within recent geological times of a well- 
timbered land between Central America and the Galapagos 
islands which has since vanished beneath the ocean. 

And yet that all was not land on the Pacific side of Central 
America is proved by many distinct lines of evidence. 

Mr. Guppy * tells us that Laguncularia racemosa, Ehizo- 
phora mangle, Anona paludosa and Conocarpus erectus, all 
of which are plants of the mangrove formation, occur not 
only on the Pacific and Atlantic coasts of America, but also 
on the west coast of Africa. We cannot suppose that such 
characteristic sea-shore species can be conveyed across a land 
area by any known means of accidental dispersal. Moreover, 
since that vegetation does not occur south of the Bay of Guaya- 
quil, it could not have reached the coast of Ecuador from 
eastern South America by travelling round Cape Horn. Thus 
the most likely explanation of the occurrence of the flora on 
both sides of Central America is to assume the existence of 
a former continuous shore line between Ecuador and Vene- 

In the list of stalk-eyed Crustacea found on the coast of 
Peru, Miss Eathbun f reports that a few species of crabs, like 
Acanthonyx petiverii, Micrqphrys platysoma, Panopeus ber- 
mudensis and Geograpsus lividus, are common to both sides 
of Central America, while a most striking feature is the 
faunistic uniformity of the marine area between the Cape 
Kegion of Lower California and the coast of Chile. Dr. 
DallJ divided this whole fauna into two provinces, viz., the 
Panamic, extending from Lower California to Guayaquil, and 
the Peruvian, from the latter to the island of Chiloe in Chile, 

* Guppy, H. B., " A Naturalist in the Pacific," p. 498. 
t Eathbun, M. J., " Stalk-eyed Crustacea of Peru." 
t Ball, W. H., "Report on the Shells of Peru," p. 185. 


but the two have many forms of mollusks in common, and 
should be regarded perhaps as sub-divisions of one faunistic 
marine area. According to Dr. von Ihering,* Chile received, 
in early Tertiary times, certain tropical genera of mollusks 
which never succeeded in attaining the North American 
coasts, yet are represented also in Patagonia, while others, 
such as Conus, Purpura, Oliva, Concholepas, Cassis, Cypraea 
and Bissoa are absent from the latter country. They are 
supposed to have travelled along the north coast of South 
America to Ecuador, Peru and Chile by means of a Central 
American marine channel. Certain species even of that an- 
cient marine migration have persisted to the present day, 
not only on the coast of Chile, but on the west coast of Africa 
and in the Mediterranean. Even in Miocene times the in- 
fluence of the Caribbean and European marine faunas was 
felt on the coa-st of Peru, according to Dr. Ortmaam.f 
Certain northern species of the genera Saxidomus and 
Chlorostoma, says Dr. von Ihering (p. 524) did not reach the 
coast of Chile until the Pleistocene Period. Thus it seems 
manifest that during practically the whole of the Tertiary 
Era there was no Humboldt current sweeping northward 
along the west coast of South America, as it does at present. 
On the contrary, there is sufficient evidence to show that 
whatever current there existed flowed in the opposite 

This investigation has resulted in two very important 
results, viz., firstly, the demonstration that the Humboldt 
current formerly did not exist, and secondly, that its absence 
must have been caused by profound differences in the condi- 
tions of land and water from those now prevailing. Of the 
nature of these changes I have foreshadowed already enough 
to enable anyone to reconstruct them. When the currents 
issued from the Caribbean Sea into what is now the Pacific, 
they must have been faced by land westward and northward. 
They could only have flowed southward. But the land which 
lay south-westward between Central America and the Gala- 
pagos islands extended probably far southward, parallel to 

* Ihering, H. von, " Mollusques fossiles de 1' Argentine," pp. 514516. 
t Ortmann, A. E., " Tertiary Invertebrates of Sta. Cruz," p. 320. 


the present coast. How far it reached will be more fully dis- 
cussed in the subsequent chapters. It may only compara- 
tively recently have subsided. Ma.y not the remarkable cir- 
cumstance, alluded to by Dr. Bigelow,* that to the south-west 
and west of the Humboldt current the sea is almost entirely 
devoid of surface as well as bottom life, be due to this 
cause ? 

* Bigelow, H. B., " Albatross Expedition Medusae," p. 222. 



THE great continent of South America (Fig. 18), with its 
roughly triangular shape, is not much smaller than its 
northern relative. The basis of the triangle facing northward 
is bounded by the Caribbean Sea, and the two other sides 
by the Atlantic and Pacific Oceans respectively. There is 
a very general impression that the west coast of South 
America is mountainous, the remainder being a vast plain. 
This is not quite correct. A northern continuation of 
the long chain of the Andes skirts also the southern shores 
of the Caribbean Sea, while to the south-east of it lies the 
highland of Guiana and further south the Brazilian highland. 
The two eastern highlands that of Guiana and Brazil are 
broken into two parts by the Amazon valley. The space 
between these eastern highlands and the Andes is occupied 
by the central lowlands. We might say, therefore, that the 
three great natural regions of South America are the eastern 
highlands, the central lowlands and the western Cordillera. 
The Amazon, Orinoco and La Plata the three great streams 
all drain eastward. 

The main part of the continent has a tropical climate. 
Sub-tropical and temperate conditions prevail in the south. 
The climate of the Brazilian region is exceptionally favour- 
able for a luxuriant development of the fauna and flora, 
whereas certain parts of what Dr. Wallace called " the Chilean 
sub-region " are arid, and inimical to expansion and specific 
differentiation of animals and plants. 

Although our acquaintance with the fauna, flora and 
geology of this vast continent is naturally much less complete 
than that of North America or Europe, we possess unmis- 
takable evidences of the fact that South America has under- 
gone considerable alterations in the distribution of land and 

Highland over 1500 R 
Upland over 600 F* 

FIG. 18. Map of South America. (Reproduced from Meiklejohifs Oroscopic 
Map, by permission of Messrs. Meiklejohn <k Son.) 

[To face p. 336. 


water sine the close of the Mesozoic Era. The testimony is 
of a threefold character. It is founded on our knowledge of 
geology, of palaeontology, and of the geographical distribu- 
tion of living animals and plants. Eichness, as Dr. Wallace 
observed, combined with isolation, is the predominant feature 
of neotropical zoology. Nevertheless, he thinks that early 
during the Tertiary Era, the zoological differences between 
the Nearctic and Neotropical, that is to say, between the North 
and South American regions, were probably even more radical 
than they are now. South America, he argues, was then a 
huge island or group of islands a kind of Australia of the 
New World chiefly inhabited by the imperfectly organised 
group of edentate mammals. Dr. Wallace * believes, more- 
over, that there must have been one or more ancient Ian4 
connections between the two continents, perhaps in Eocene 
or Miocene times, admitting ancestral types of monkeys and 
the members of the camel-tribe (Llamas) from the north to 
South America. 

Dr. Wallace's opinions, expressed thirty-five years ago, 
were founded entirely on the distribution of living animals. 
Eapid strides have been made since that time in our know- 
ledge of the fauna of South America. The geology of certain 
districts is being worked out. Botanists have made great 
progress in mapping out the distribution of plants, while the 
most astonishing discoveries have been disclosed principally 
among the past inhabitants of the continent. Thus we are 
now in a very different position from that of Wallace, when 
he pronounced upon the physical changes of South America 
during the past, on the strength of his zoological know- 

We are particularly indebted to T)r. von Ihering's re- 
searches on the fauna and flora of South America, which he 
conducted during many years of devoted labour, that our in- 
formation on the main features of distribution has advanced 
so rapidly. During his long residence in southern Brazil he 
collected, and is still collecting, data bearing principally orx 
the question of the geological history of the continent; and 
since his scattered papers have recently been reprinted in' 

* Wallace, A. E., " Distribution of Animals," II., p. 58. 

L.A. Z 


book form we are better able to appreciate the advances in our 
knowledge that are due to his investigations. His researches 
were not confined to recent zoology. He has also dealt with 
the subject from a palaeontological point of view, and has 
even subjected the main features of the flora to a detailed 
revision. All the same, he acknowledges that there are several 
important problems connected with the geological history of 
South America for which no satisfactory solution has as yet 
been suggested. More than twenty years ago, Dr. von Ihering 
announced that the close relationship of the Brazilian with 
the African fresh -water mussels implied the former existence 
of a land connection between South America and Africa. 
But, whereas America as a whole is one of the richest regions 
for fresh-water mussels in the world, Chile an'd Peru are 
among the poorest. While east of the Andes there is a 
luxuriant fauna of fresh-water mussels largely related to that 
of Africa, we find only the genus Unio (in its wide sense) re- 
presented on the western slope of the mountains. Dr. von 
Ihering explains this and other similar facts by the supposi- 
tion that in the east and the west there were originally similar 
faunas, but while eastern South America obtained a rich stock 
of immigrants across the land bridge from Africa, the great 
mountain chain of the Andes, which was then commencing 
to rise, prevented any further influx westward. Considering 
that the fresh- water mussels (Unionidae) are well represented 
even in Jurassic deposits, Dr. von Ihering* recognised that the 
geographical distribution of these mollusks may be utilised 
to advantage as indicators of very ancient, especially 
Mesozoic, conditions of land and water. 

In the course of his studies the same author came to the 
conclusion that from the Cretaceous to the end of the Pliocene 
Periods, South America must have been completely separated 
from North America. A real South American continent 
existed only since Oligocene times. It then consisted of two 
parts united by the narrow isthmus of the newly formed 
Andes. He named these two old original constituents of 
South America " Archiguiana " and " Archiplata." The 
former comprised the highlands of Guiana and Venezuela, 

* Ihering, H. von, " Archhelenis and Archinotis," p. 79. 


the latter the remainder of South America, the two parts 
being separated by a broad ocean except for a narrow western 
land bridge. Each of these great islands had its own peculiar 
fauna and flora, but Archiguiana must have been connected 
by land with Africa until Oligocene times, and Archiplata 
with New Zealand and Australia during the Mesozoic 
Era, ! ; I I . 

With the gradually increasing knowledge of palaeontology 
Dr. von Ihering's original ideas naturally became subject to 
various modifications. Thus in a map representing the con- 
ditions of land and water during the Eocene Period, and pub- 
lished in 1907 (Fig. 17), not Archiguiana but Archiplata is 
connected by land with Africa and also India, the whole of 
this ancient continent being called " Archhelenis." Archi- 
plata is still joined at this time to Australia by means of the 
antarctic continent " Archinotis," while Archiguiana is 
united with the West Indies and parts of Central America 
into a large land-mass which stretched forth westward to the 
Sandwich islands, and was called " Pacila." Quite recently 
the same author brought forward testimony in favour of a 
Miocene land bridge between Central America and eastern 
Asia. I have already alluded to it in the previous chapter. 
Dr. von Ihering now tells me that he will shortly publish 
a revised palaeogeographical map in the " Neues Jahrbuch 
fur Mineralogie und Geologie," in which these features are 
indicated. i 

The same problem, studied from the point of view of the 
distribution of fresh-water crabs and crayfishes, led Dr. 
Ortmann* to somewhat different conclusions. At the end 
of the Mesozoic Era he recognises the existence of the island 
of Brazil, which had previously been connected with Africa, 
while Guiana was still joined to western North America on the 
one hand and Africa on the other. The independent Chilean 
tract of land was connected with Australia by means of the 
supposed antarctic continent (Fig. 15). At the commence- 
ment of the Tertiary Era South America had assumed its 
present shape, except for an elongated bay extending inland 
from the Atlantic Ocean into the valley of the Amazon. In 

* Ortmann, A. E., "Distribution of Decapods," pp. 379381. 



the south it was still joined to the Antarctic Continent, which 
had then become separated from Australia. In the north it 
had lost its land connection with North America, which it did 
not regain until later Tertiary times. 

To a certain extent the views of the two authors agree, at 
any rate, in the assumption that the continent of South 
America is composed of several originally independent land- 
masses, one of which was joined to Africa. The most striking 
difference in their opinions, apart from the geological period 
during which the various elements are supposed to have be- 
come fused together, lies in Dr. Ortmann's conception of 
three totally distinct land-masses, while Dr. von Ihering only 
recognises two. Nevertheless, even the latter acknowledges 
the faunistic division of his " Archiplata " into a northern 
and southern portion, although his nomenclature is apt to be 
somewhat confusing. Dr. von Ihering informs us (p. 177) 
that the old Archiplata fauna has no close relationship to 
that of the rest of South America. 

A third contribution to the geological history of South 
America is furnished by another group of fresh-water animals, 
namely, the fishes. The tropical American fresh-water fauna, 
having its centre of greatest diversity in the middle Amazon 
basin, says Professor Eigenmann,* is attenuated northward 
till it reaches the vanishing point just on the borders of 
the United States. Southward it extends to somewhere south 
of Buenos Aires. The Patagonian and North American 
faunas are entirely different from the tropical American fauna 
and from each other. The results of his studies are that the 
existing distribution of the fresh-water fishes can only have 
been brought about by the supposition that tropical America 
in early Tertiary times consisted of two land areas (" Archi- 
guiana " and " Archamazona "), separated by the lower valley 
of the Amazon, which was submerged by the sea. There was 
a land-mass between Africa and South America, possibly 
joining Guiana and tropical Africa. But this connection, he 
urges, must have ceased to exist before the origin of the 
present genera, and even before that of some of the families. 

* Eigenmann, 0. H., "Freshwater Fishes of South America," pp.|517 


Several other writers have discussed the theory of former 
land bridges between South America and Africa and between 
South America and other continents without dealing with the 
geological history of South America itself. Their views will 
be presented later on. 

The only author who, to my knowledge, has treated the 
subject under review from a purely palaeontological stand- 
point is Dr. Ameghino. His views were brought for- 
ward in several important technical papers. As they may 
have undergone slight modification in the course of time, I 
prefer to quote his opinions from a recent semi-popular 
account. It may be mentioned that Dr. Ameghino's * 
theories are founded on the known distribution of fossil 
mammals. North of the Equator, he says, there was, at the 
end of the Mesozoic Era, a vast ocean containing many 
islands. In the south lay a great continent, united in the 
east with Africa and in the south with Australia and New 
Zealand by way of an antarctic land bridge. Gradually, in 
early Tertiary times, the northern islands became fused and 
joined to the southern land -mass, while Australia was dis- 
connected. During all this time North and South America 
remained entirely separated. Mammals, however, were able 
to pass from the latter continent through Africa into Europe, 
and across a land bridge in northern latitudes to North 
America. The sea, which had covered the Andean region in 
Mesozoic times, disappeared early in the Tertiary Era, but 
certain bays of the Pacific Ocean still remained and ex- 
tended eastward to the far side of the Andes. During the 
Eocene Period the Atlantic Ocean advanced from the south, 
covering part of Argentina and practically separating 
Brazil from the rest of the continent, while the former land 
connection with Africa ceased to exist. Towards the end of 
the Oligocene Period the " Archhelenis " land bridge once 
more rose to the surface for a brief period and then finally 
subsided entirely, save for a few scattered islands. 

It would have been particularly valuable to obtain Pro- 
fessor Osborn's opinions on the geological evolution of South 
America. We possess only his palaeogeographical maps 

* Ameghino, Fl., " Geologia de la Republica Argentina," pp. 916. 


founded upon the distribution of fossil mammals. It is sur- 
prising that in these maps South America in late Cretaceous 
and basal Eocene times, is represented as almost precisely 
what it is to-day, except that it is continued southward across 
an antarctic continent to Australia. In the middle Eocene, 
South America differs only in so far as a long bay of the 
Atlantic has entered the Amazon valley. There are no indica- 
tions of any land bridges at that time, South America being 
completely isolated from all other continents. During the 
Oligocene Period it still remained so, but the sea made 
further inroads on the Amazon valley, it encroached on the 
valley of the Parana river and flooded a large part of Argen- 
tina, reducing southern Chile to a few islands. It is only in 
Miocene times, according to Professor Osborn,* that South 
America became divided into two parts by a broad gulf ex- 
tending from the Atlantic to the Pacific across the Amazon 

Geologists, except Dr. Katzer and Professor de Lapparont,f 
have as a rule dealt with the problem in a less comprehensive 
manner. The ideas of the latter differ from the authors cited 
in so far as the main permanent land -mass at the end of the 
Mesozoic Era in South America was confined, in their opinion, 
to the east. They suppose the highlands of Guiana, eastern 
and southern Brazil to have been united. All the rest of the 
continent was then under water. At the commencement of 
the Eocene Period, according to Professor de Lapparent, 
Central America had come into existence, but disappeared 
again shortly after, while a broad marine channel stretched 
from the Pacific to the Atlantic between northern Chile and 
Argentina. Dr. Katzer's views are somewhat similar. He 
dees not believe in the Atlantic Ocean having invaded South 
America from the east. In the beginning of the Mesozoic Era 
the area of archaean rocks and later palaeozoic deposits of 
Guiana and Brazil formed a large connected land-mass. In 
Upper Jurassic times, he says, the old land connection be- 
tween South America and South Africa on the one hand, and 
between South America and Australia, still existed. An old 

* Osborn, H. F., "Age of Mammals" Maps, pp. 64, 137, 183 and 245. 
t Lapparent, A. de, "Traite de Geologic," 4th ed., pp. 1376 and 1455. 


ocean strait extended from the north side of Guiana across 
Venezuela and Colombia to Peru during the Cretaceous 
Period. During all this time, and even in early Tertiary 
times, the waters from the old eastern land continued to 
drain westward towards the Pacific. The persistent rise of 
the newly formed Andean mountain chain resulted at first 
in the formation of a vast lake covering the entire lowlands 
of the Amazon valley area. Eventually, in Miocene times, 
according to Dr. Katzer,* the drainage was reversed, with the 
result that the Amazon river flowed for the first time into the 
Atlantic Ocean. Concurrently North and South America be- 
came united through the Central American land bridge. 

Professor Koken's f palaeogeographical maps were con- 
structed as the outcome of a combination of geological and 
palaeontological studies. South America, he remarks, had 
already assumed its present shape and form in Cretaceous 
times, though it did not extend so far west as at present except 
in Ecuador and Colombia. It was separated from all other 
continents but Africa. In early Tertiary times South America 
became entirely isolated. Argentina and southern Chile were 
largely flooded by the sea, while a long gulf filled the whole 
valley of the Amazon as far east as the Andes. 

Dr. Arldt, J who included the distribution of living animals 
and plants as well as palaeontology within the sphere of his 
studies, gives a series of highly complex maps which cannot 
readily be described. His conception is that South America 
in Lower Cretaceous times was somewhat like that described 
by Professor Koken, viz., an extension of land eastward as far 
as Africa and a simultaneous submergence of the west coast. 
Towards the end of the Mesozoic Era, that is to say at the end 
of the Cretaceous Period, a complete change in the conditions 
of land >and water supervened. South America was then 
divided into two parts by an interoceanic connection across 
the Amazon valley. The northern portion, consisting of 
Colombia, Ecuador and Guiana, is supposed to have extended 
westward across the Galapagos islands as far as the Sandwich 

* Katzer, F., " Geologie des Amazonengebietes," pp. 239262. 

t Koken, E., " Die Vorwelt," Maps 1 and 2. 

I Arldt, J., " Entwicklung der Kontinente," Maps 19 and 20. 


islands, thus forming a great peninsula of land which like- 
wise was joined to lower California and western Mexico. The 
southern part of South America was then still united by a 
narrow land bridge with Africa, while the western side of it 
now became joined to a great belt of land extending right 
across the Pacific Ocean to New Zealand and Australia. In 
early Tertiary times South America became separated from 
North America and the Sandwich islands, while the two 
sections of the continent fused in the west. The African 
and Australian land connections still persisted in a modified 

We can gather from all these expressions of opinion as to 
the past geological history of South America that there is 
comparatively little general agreement on the subject. Some 
points, however, seem to be fairly well established. All 
authorities concur in the belief that the Eastern highlands of 
Guiana and Brazil have been land surfaces since the begin- 
ning of the Secondary Era, at any rate, and on these, there- 
fore, we ought to find relicts of a Mesozoic fauna. All the 
writers quoted also agree that at some time or other during 
the Tertiary Era there was either a complete interoceanic 
connection along the Amazon valley or a long gulf of the ocean 
extending for some distance inland. Yet there is an im- 
portant difference of opinion as to whether this gulf belonged 
to the Atlantic or the Pacific Ocean. But since most of the 
writers contend that the central portion, at any rate, of the 
Andes is made up largely of Jurassic and Cretaceous marine 
deposits, while the eastern parts of South America were land 
in Mesozoic times, it seems more reasonable to assume, with 
Dr. Katzer, that the Pacific Ocean extended eastward as far 
as the archaean highlands of Brazil and eventually retreated 
so as to leave only a Pacific gulf on the site of the existing 
upper Amazon valley. In view, however, of the fact that the 
Pacific Ocean must have been completely shut out from South 
America by the western belt of land above alluded to, .it was 
really the waters of the Atlantic Ocean which flooded western 
South America as far east as the highlands of Brazil (com- 
pare Fig. 14). 

Several of the authors cited recognise a faunistic relation- 
ship between Australia as well as between Africa and South 


America, and some contend that Chile has been wholly or par- 
tially above the surface of the ocean since very early geological 
times. As Professor Suess * has pointed out, the extreme 
south of the continent is composed of a peculiar and probably 
very ancient system of rocks. This Patagonian mountain 
chain is quite independent of the Andes, of which it does not 
form a continuation. Further north in Chile it lies outside or 
westward of the Andes, constituting the coast Cordillera. 

Still further north, in Peru, only fragments of this Cor- 
dillera are recognisable here and there close to the coast, 
while in Ecuador the ancient coast Cordillera again appears 
to be more clearly represented. There is some evidence, 
therefore, of the existence of a very old land -mass not only on 
the eastern but also on the western side of the continent. And 
it seems as if the Mesozoic rocks forming a large part of the 
Andes had been deposited in a great trough along the eastern 
shores of this ancient land. When the Andes were subse- 
quently raised, the older western land presumably subsided, 
leaving only here and there along the coast some traces of its 
former existence. 

Let us now pursue this subject from a purely zoogeo- 
graphical point of view, and endeavour to utilise groups with 
whose distribution we are fairly well acquainted, in elucida- 
ting the more obscure points in the geological history of 
northern South America. Groups of animals of which we 
have reason to believe that they are very ancient, possibly 
of Mesozoic age, will be best for this purpose. Such a one for 
example is that of the Onychophora (Protracheata). It 
contains creatures caterpillar-like in appearance, with a beau- 
tifully soft skin and unjointed limbs. Internally they contain 
air-tubes so characteristic of insects ; at the same time, their 
excretary organs resemble those of worms. These very pri- 
mitive features, taken together with a wide and most discon- 
tinuous distribution, have always caused these creatures to 
be looked upon as survivors or relicts of very remote geological 
ages. Formerly they were all classified under the genus 
Peripatus. In recent years many new forms have been dis- 
covered and subjected to careful study, with the result that 

* Suess, E., " Antlitz der Erde," I., pp. 666690. 


quite a number of genera are now recognised. The most 
exhaustive account of these remarkable creatures is, no doubt, 
the monograph recently published by Professor Bouvier. He 
divides the group into seven genera, of which Peripatus is 
confined to tropical and sub-tropical America and tropical 
Africa, and Opisthopatus to Chile and South Africa. This is 
a most astounding discovery. That a group of these creatures 
found in tropical South America should be more closely related 
to another occurring in tropical Africa than to that of Chile, 
and that the latter should exhibit a more intimate affinity with 
South African forms than with tropical American ones, is of 
great zoogeographical interest. Professor Sedgwick,* however, 
does not share Professor Bouvier's opinion with regard to 
the intimate relationship supposed to exist between the 
tropical South American and tropical African, and between 
the Chilean and South African groups. He thinks they are 
perfectly distinct from one another. On the other hand, he 
agrees with Professor Bouvier in the recognition of a group of 
Onychophora, limited to tropical and sub -tropical America, as 
distinct from the Chilean group. And this is really the prin- 
cipal point I wish to draw attention to. 

Professor Bouvierf claims that the Andean species of Peri- 
patus are the most primitive members of the whole family. 
He believes that the ancestral stock inhabited a former Pacific 
continent, and that their immediate descendants took refuge 
on the eastern and western land areas when their original 
habitat vanished. The whole genus Peripatus, as defined by 
Professor Bouvier, I may mention again, is found from 
Mexico in the north, throughout Central America, the West 
Indies ,and South America as far south as Bolivia. The 
Chilean species belongs, according to the same authority, to 
the distinct genus Opisthopatus. 

The genus Peripatus is readily divisible into two sections, 
the Andean and the Caribbean one. The twelve species be- 
longing to the former all inhabit the Pacific side of the Andes, 
except Peripatus eiseni and Peripatus goudoti, which live in 
Mexico nearly two thousand miles north-westward of the other 

* Sedgwick, A., "Distribution of Onychophora," pp. 383 406. 
t Bouvier, E. L., " Onychophores," I., pp. 64 79. 


members of the Andean group. This discontinuous distribu- 
tion is a very noteworthy fact, for it cannot be explained by 
the supposition that some member of this Andean group may 
still exist and have been overlooked in the intermediate vast 
tract of country, because many specimens of Peripatus have 
been discovered in Costa Bica, Nicaragua and Panama, all 
belonging to the Caribbean group. Accidental dispersal, such 
as marine currents, cannot be invoked as being responsible for 
this distribution. It is due, in my opinion, to a former direct 
land connection between western Mexico or Guatemala and 
some part of the west coast of Soutli America. That the 
mountain system of Guatemala suddenly terminates at the 
edge of the Pacific, and that it formerly had a westward 
continuation, has been alluded to, and I have mentioned also 
several cases of discontinuous distribution that I thought 
were due to the existence of an ancient land, more or less in- 
dependent of Central America. The newt Spelerpes is one 
of these. Its headquarters seem to be in Mexico. A few occur 
in Guatemala, Costa Eica and Chiriqui. Further south we 
meet with the genus again in Colombia, Ecuador and northern 
Peru, but nowhere else in South America. The tortoise 
Chelydra rosignoni occurs in ^Guatemala. It is absent from 
the rest of Central America, yet in Ecuador we find an isolated 
colony. Another tortoise, Geoemyda punctularia, inhabits 
Guatemala and Mexico. Southward it is only known from 
Ecuador, Colombia, Venezuela and further east. The gecko- 
like genus Eublepharus is probably an exceedingly ancient 
one, its range being most peculiar and strikingly western. 
One species occurs in California, another in Mexico, still 
another in Panama, and lastly one in Ecuador. All the re- 
maining species, which show great resemblance to the 
American ones, are confined to southern Asia. We probably 
have to deal in this case with a persistent type which through- 
out many geological periods has retained the same characters 
and has died out in the still existing land fragments of the 
ancient Pacific continent, whence it originally spread east 
and west after its subsidence. There are numerous other 
examples, particularly among plants, implying that the land 
which I described as lying westward of Central America 
once touched the South American continent, probably near 


Ecuador (compare Fig. 14). As this western belt of land 
was in direct communication with the West Indies by way 
of Central America, it follows that the West Indies and 
Ecuador were able to enter into a faunistic exchange. Many 
instances might be quoted showing this relationship between 
the West Indies, Central America and western South 
America, or between parts of these areas. The head- 
quarters of the snake-like limbless amphibians, known as 
Coecilia, are in Ecuador. From there they have spread 
eastward through Guiana to Brazil, and northward through 
Colombia as far as Panama. No species is actually known 
to occur in Central America beyond Panama. I alluded 
to the family Anguidae on several occasions, those generally 
limbless lizards to which the so-called glass -snake belongs. 
One genus with well -developed limbs (Diploglossus), inhabits 
chiefly the principal islands of the West Indies, viz., Cuba, 
Haiti and Jamaica. In Central America it occurs in Guate- 
mala and Costa Kica, while in South America it inhabits only 
Ecuador and Brazil, having apparently spread into the latter 
state from the west. 

Attention has been drawn to the fact that while the fauna 
and flora of the Galapagos islands are principally Central 
American and West Indian in character, they also are related 
to those of western South America. A mere fragment only 
of the animals and plants that passed across the lands of 
which these islands formed part could have been preserved 
there. Thus the Streptaxidae, a family of carnivorous snails 
almost restricted in America to the southern continent, do 
not occur in the Galapagos islands, although a few species 
have penetrated to Guatemala, and one even to Haiti. The 
genus Martinella is peculiar to Ecuador, whereas two other 
genera, viz., Guestieria and Systrophia appear to have spread 
from an Ecuadorian centre of dispersal to Peru, Colombia 
and Bolivia. 

A most interesting and important case of discontinuous dis- 
tribution is that of Clausilia, a genus of snails which I men- 
tioned when dealing with the origin of the West Indian fauna 
(p. 272). I then stated the reasons for my belief that 
Clausilia had travelled across the mid-Atlantic land bridge 
from southern Europe to the West Indies rather than by a 

.5 O 

03 CU 

~ a 

It s 

o ^ ^ 

-I | 

S -S ^ 

I - 

es 2 42 

^ -5 Q 


land connection between eastern Asia and Central America 
as suggested by Dr. von Ihering.* The American species 
all belong to the sub -genus Nenia, being extremely like the 
Pyrenean Clausilia pauli. So far only a single species is 
known from the West Indies, while the genus has not 
yet been discovered in Central America; (Fig. 19). No 
trace of Clausilia, either recent or fossil, has been met with 
in North America or in northern Asia. If we assume that 
Clausilia passed across the Atlantic Ocean to the West Indies 
and thence to South America, it should have traversed Central 
America. It may still be found in Guatemala, or it may 
recently have become extinct. At any rate, I think it passed 
from. Central America to the western or Pacific land bridge con- 
necting Guatemala with Ecuador. The latter state, together 
with northern Peru, must be looked upon as the centre of 
dispersal of the South American species of Clausilia. Over 
fifty Clausilias are now known from South America.f A few 
are found northward of Ecuador as far as the province of 
Bogota in Colombia. The great mass, however, is confined to 
the highlands of Ecuador and Peru. South of Bolivia the 
genus is unknown. Is there any possible explanation for the 
presence of this isolated colony of mountain snails in western 
South America except by migration on a former land connec- 
tion from Europe or eastern Asia ? A dispersal of the ances- 
tral Clausilia in early Tertiary times to South America by 
means of either land bridge is possible, but I greatly favour 
the Atlantic one (compare Fig. 14). It also explains the 
presence of a species of Clausilia in Portorico better than the 
other theory. 

Besides this instance of discontinuous distribution, we more 
over have a large number of others exhibiting ancient rela- 
tionship between the faunas of the west coasts of North and 
South America. These will be more fully dealt with when we 
come to the consideration of the Chilean fauna. I may only 
mention one example, that of the three very closely related 
genera of scorpions, Hadrurus, Hadruroides and Caraboc- 
tonus. The first inhabits California, having spread from there 

* Ihering, H. von, " Verbreitung der Heliciden," p. 450. 
t Boettger, O., "Die Nenia Arten." 


into the adjoining state of Arizona. The second lives in 
Ecuador and Peru. The third is only met with in Chile and 
southern Peru. They all are entirely confined to the west 
coasts. A somewhat more distant relationship exists between 
the centipede Newportia of the West Indies and Central and 
South America, and Plutonium, which is confined to Sardinia 
and Corsica. This affinity implies the presence of a former 
mid-Atlantic land bridge between the Mediterranean and 
Antillean Eegions and between the latter and northern South 

My principal aim, in alluding to the fauna of Ecuador 
and the adjoining areas, was to point out the more im- 
portant features of the former as an ancient land-mass 
somewhat independent from the rest of South America, 
and the affinity of its animal inhabitants with those 
of Central America, the Antilles and even southern 
Europe. The antiquity of the region can be demonstrated by 
many examples. Even birds show it clearly. Among one 
hundred and eighteen genera of humming-birds (Troohilidae) 
known to science, Androdon, Eutoxeres, Damophila, Uro- 
chroa, Phaeolaema, Agapeta, lonolaema, Eugenia and others 
are quite confined to this region.* Among mammals I need 
only mention the remarkable discovery by Mr. Oldfield 
Thomas f of Caenolestes. This small rat-like mammal in- 
habits Ecuador and the province of Bogota in the adjoining 
State of Colombia. It is, therefore, quite confined to north- 
western South America. The only other known genera of the 
family Epanorthidae were found in the Santa Cruz deposits 
of Argentina. The presence of a still living member of this 
otherwise extinct family is of considerable interest. Even 
more important is the fact that Caenolestes is the only living 
American herbivorous marsupial mammal more nearly related 
to the kangaroo of Australia than to the American representa- 
tives of the order the opossums. 

The only South American representative of the family of 
bears (Ursidae) known as the spectacled bear (Tremarctos 
ornatus), on account of the yellow rings surrounding its eyes, 

* Hartert, E., " Trochilidae." 
t Thomas, O., " On Caenolestes,' 1 


is confined to the Andes between Colombia and Chile, that is 
to say, altogether to western South America. This range sug- 
gests that the bears are not originally a South American family, 
otherwise we might expect their having a wider distribution. 
The only alternative is that they migrated from some other 
part of the world to the part of South America where they 
still maintain themselves. Dr. von Ihering * argued that the 
ancestors of the South American bear originally came from 
Asia, and that they wandered across on a Miocene land con- 
nection which united eastern Asia with Central America with- 
out touching North America. In that case they would be older 
than the North American bears which, according to Professor 
Osborn,! belong to the much later Pleistocene Eurasiatic inva- 
sion. As a matter of fact, the geological history of the bears in 
America has never received the careful attention it deserves. 
Bears, it is well known, are entirely absent from Africa south 
of the Sahara, that is to say, from what is known as the 
Ethiopian Region, and no fossil remains of any members of 
the family have ever been discovered there. Hence it is un- 
likely that Africa was the source of the Ursidae or that they 
invaded South America by means of a direct land bridge from 
that continent as suggested by Dr. Ameghino. J The absence 
of bears from the whole of eastern South America indicates, 
moreover,, that bears have made their; entry from the west. Thte 
bears of South and North America are not closely related. 
They appertain to different genera. A fossil bear (Arcto- 
therium vetustus), belonging to a genus closely related to or 
identical with Tremarctos, has been recorded by Dr. Ameghino 
from the Entrerios deposits of Argentina which are of Mio- 
cene age. It would appear on that account as if Dr. von 
Ihering's suggestion that the South American bears were older 
than the North American ones, and had quite an independent 
Asiatic origin, was borne out by palaeontological evidence. 
Since Arctotherium also occurs fossil in some Pleistocene 
beds of North America it must have spread northward in 
recent times and subsequently have become extinct there. 

* Ihering, H. von, " Greschichte der Siidamer. Raubtiere," p. 1*79. 

t Osborn, H. F., "Age of Mammals," p. 438. 

J Ameghino, Fl., " Tetraprothomo argentinus," p. 230. 


Among living bears the nearest relation of the Andean species 
seems to be the Malayan bear (Ursus malayanus) inhabiting 
the Malay peninsula and neighbouring countries. But the 
European Miocene Ursus boeckhi and the Pliocene Ursus 
etruscus are members of the same group, and it appears to me 
possible that the South American Tremarctos and the Ursus 
malayanus groups may have had a common ancestor which 
passed from southern Europe to South America by way of the 
mid-Atlantic land bridge and the Antilles in Oligocene times. 
Although tapirs have a much wider range in South America 
than bears, their American distribution also suggests that, 
like the bears, they are immigrants either from Europe or 
Asia. Their general range is in so far comparable to the 
distribution just cited as the only living tapirs are confined 
to South America and southern Asia. Tapirs are often alluded 
to as among the most striking and familar instances of what 
is called " discontinuous distribution." But we know a good 
deal more of the geological history of tapirs than of bears. Of 
the two South American tapirs the smaller one is confined to 
the Andes between Colombia and Peru, while the other 
(Tapirus americanus) has probably spread eastward from a 
western centre of dispersal, for it occurs from eastern Peru to 
Brazil, Venezuela, Guiana and to northern Argentina. Two 
other tapirs live in Central America. The genus is only known 
fossil from Pleistocene South American deposits, and it might 
appear as if it were a recent immigrant from North America. 
But in the latter continent only a single fossil species 
(Tapirus haysi) has been discovered, and that likewise in 
Pleistocene beds. Professor Osborn asserts that a tapir, un- 
distinguishable from the living South American species 
(Tapirus americanus), invaded North America together with 
Mylodon and Megatherium in Pleistocene times. I am not 
aware of any reliable osteological characters distinguishing the 
living South and Central American species. If there are such, 
Professor Osborn's * statement may possibly refer to one of 
the latter. Still, it is quite evident that the genus Tapirus 
could not have come across any Bering Strait land connection 
in Pleistocene times and have travelled to Argentina before 

* Osborn, H. F., "Age of Mammals," p. 472. 


the end of the Pleistocene Period, nor does Professor Osborn 
suggest such a mode of origin. The problem, therefore, still 
remains unsolved. Mr. Earle * pointed out that tapir-like 
creatures or tapiroids arose about the same time in Europe 
and North America. In the light of more recent researches 
it would appear that the Eocene Systemodon and Isectolophus 
are confined to North America, while the European tapiroid 
remains belong to the related family Lophiodontidae. The 
true tapirs, to which the American genera belong, do not 
make their appearance in Europe until the Oligocene Period.- 
According to Professor Osborn,f the existing Malayan 
tapir is almost identical with the Pliocene tapir of southern 
Europe (T. arvernensis), and I cannot help thinking that the 
genus Tapirus has evolved in the Mediterranean region from 
American ancestors much earlier than is generally supposed, 
the modern tapirs having spread west and east from this centre 
of dispersal at a time when the mid-Atlantic land bridge was 
still in existence. 

If we pass from Ecuador southward along the chain of the 
Andes, we meet with a number of new forms of animal life, 
all of which are more or less confined to this great mountain 
range. In certain districts in Peru at high altitudes there 
are immense colonies of curious little squirrel-like rodents 
with very large ears and grey fur of extreme softness. Like 
the prairie-dogs and other North American rodents, these 
chinchillas, as they are called, live in burrows. There is a 
larger kind, too, which has still longer ears and great black 
whiskers, differing sufficiently from Chinchilla to deserve 
recognition as the distinct genus Lagidium. Both genera 
inhabit exclusively the high mountains between Peru and 
Chile. A third member of the same tribe, the viscacha, lives on 
the plains of Argentina, and will be more fully described later 
on. These three genera included in the family Viscaciidae 
(Lagostomidae) have, to judge from their distribution, pro- 
bably originated from one or more western ancestors. But 
Dr. Ameghino J has described quite a number of genera 

* Earle, C., "Fossil Mammalia of Europe," p. 115. 

t Osborn, H. F., "Age of Mammals," p. 315. 

J Ameghino, Fl., " Formations sedimentaires," p. 428. 

L.A. A A 


obviously of the same family from the Patagonian and Argen- 
tine Eocene, Oligocene, Miocene and Pliocene beds, so that 
it would seem, as he indeed suggests, as if South America 
had been the original home of the family. Professor Schlosser 
quite agrees that these are the undoubted ancestors of the still 
existing South American genera, but he thinks that Dr. 
Ameghino is mistaken about the age of the South American 
deposits in which these rodents occur. He believes them to 
be not earlier than Miocene, while more primitive ancestors 
of these rodents are to be found in European Oligocene beds. 
The Theridomyidae of Europe which, according to Professor 
Schlosser, have given rise 'to Chinchilla and its relatives, are 
more primitive in structure than the latter and nearly allied 
to them. The same writer acknowledges that other groups of 
mammals also entered South America from Europe, although 
he does not give us the least clue as to the means they em- 
ployed in doing so. Whether Chinchilla and its relations are 
descended from European ancestors, or whether the European 
Theridomyidae have been derived from South America, is 
really comparatively immaterial to our present enquiries. The 
important point at issue is the recognition, by such an autho- 
rity as Professor Schlosser,* that the two continents have had 
a faunistic interchange about Oligocene times in which North 
America took no part. 

Ecuador possesses such a number of quaint archaic forms 
of animal life that I have scarcely been able to give a general 
sketch even of the main features of the fauna, but before 
dealing with the causes that produced them I must mention 
one more instance. It is the most noteworthy perhaps of the 
whole deer-tribe, namely, the little pudu (Pudua mephisto- 
pheles). No larger than a hare, this tiny creature, with its 
simple unbranched spike-antlers, reminds us of some of the 
early progenitors of the deer-tribe. The genus Pudua has two 
species, both entirely confined to western South America, like 
the bear and Chinchilla. One of these lives in the mountains 
of Ecuador, the other in Chile and on the island of Chiloe. 
I need not repeat the remarks made in an earlier chapter 

* Schlosser, M., "Tullberg's System der Nagethiere," pp. 741 


(pp. 107112), when I argued that the South American 
deer had not originated in North America, as is generally 
assumed, but in South America from European ancestors. 

I venture to think that most palaeontologists will agree with 
my contention, which is by no means a new one, that there is 
quite a remarkable affinity between the living western South 
American groups of mammals and those of the early European 
Tertiaries. It is my interpretation of the causes which pro- 
duced this striking feature that will not so readily commend 
itself. The faunas of the West Indies and Central America 
form the chief difficulty to the acceptance of my theory. I 
acknowledge that comparatively few traces remain in these 
countries of the vast migration that swept across them. In 
the West Indies, I presume, the subsequent submergence must 
have destroyed the principal part of the original fauna, while 
Central America in its present form did not exist at the time 
when the mid-Atlantic land bridge spanned the ocean. Com- 
petition with newer arrivals, moreover, must have been very 
keen, so that Central America became unfitted for the survival 
of European relict forms. This explanation does not appear 
altogether satisfactory. But the crux of the problem is North 
America. By what possible system of land bridges can 
western South America have received part of its fauna from 
Europe and have exchanged certain groups in return without 
North America having become affected ? This seems all the 
more puzzling considering that I drew special attention in 
Chapter IX. to the conspicuous faunistic relationship between 
southern Europe and California. The faunas of western 
North America and western South America as a whole are 
strikingly different, and yet I have indicated certain points 
of resemblance, especially between some of the more ancient 
members of the two faunas. If we supposed the mid- 
Atlantic land bridge of early Tertiary times to have been 
connected at first with both western North America and 
western South America, while disconnected at all other points 
with these continents, certain very ancient points of resem- 
blance between the two continents and with Europe might 
thus receive a satisfactory explanation. If the same land 
bridge had then become entirely separated from North 
America, remaining united with South America, the faunistic 

A A 2 


interchange would have continued only between Europe and 
South America. If the land bridge had then become discon- 
nected with South America, and joined to western North 
America, while the Antilles were submerged, the latest Euro- 
pean emigrants would have taken refuge in California as the 
last remnants of the old land sank into the Pacific. That 
something of this kind actually took place I feel convinced, 
although the details of these events must be founded largely 
on geological studies which unfortunately are as yet insuffi- 
ciently known. My own knowledge, moreover, of the geo- 
logical features of the regions alluded to is only fragmentary. 
Nevertheless, the little I have been able to gather does not 
tend to contradict the general scheme of the theories I now 

In later Cretaceous times a broad sea, as I have pointed 
out on several occasions, separated western from eastern North 
America, while the Pacific Ocean flooded a large portion 
of the western States, so as to leave only a comparatively 
narrow strip of land between the two seas. The Cretaceous 
deposits can be traced all along the Pacific coast as far almost 
as the extreme tip of Lower California. Here they suddenly 
stop. The Cretaceous sea evidently did not cover the in- 
teresting Cape Region of Lower California nor any part of 
western Mexico. There are reasons for the belief that even 
at this time the western part of the coast ranges of California 
were not submerged, thus suggesting the existence of another 
land-mass to the west of the Californian Sea. This land may 
have been connected with the Cape Region of Lower Cali- 
fornia, and thus with Mexico. By the end of Cretaceous time, 
says Professor Smith,* the subsidence and erosion of the 
western part of the continent had almost established a con- 
nection between the Pacific Gulfs in California and Oregon 
and the old Mediterranean Sea in the Mississippi valley. The 
intervening isthmus was covered by extensive marshes. Pro- 
fessor Smith tells us that the geographical conditions re- 
mained the same in Eocene times as in the Upper Cretaceous, 
except that the sea encroached still further on the land. Now 
it is precisely at this time that we notice a striking affinity 

* Smith, J. Perrin, " Geological History of California," pp. 347348. 


between the western mam'malian fauna of North America and 
those of Europe on the one hand and South America on the 
other. Since south-western North America was then prac- 
tically isolated and separated from the remainder of North 
America by great ocean belts, how can we imagine these Euro- 
pean and South American affinities to have been brought 
about ? Surely only by some land connection that lay to the 
south. I suggest that it was from western Mexico that these 
earliest mammals invaded south-western North America. 
Then followed a time when the Gulf of Mexico and the Pacific 
Ocean probably communicated with one another, thus separat- 
ing the supposed mid-Atlantic land bridge from North 
America. Professor Smith speaks only of a temporary con- 
nection between the oceans, accounted for by the occurrence 
in Oregon, as well as in California., of the Atlantic marine 
mollusk Venericardia planicosta. Before the Miocene Period 
this Atlantic connection had ceased, and the faunas of the 
later Tertiary were wholly of the Pacific type, continues Pro- 
fessor Smith.* He does not allude to Oligocene deposits, but 
it is not long since that these were recognised at all outside 
Europe. At any rate, after the Eocene follows a time during 
which the Pacific recedes from the wes.t coast, thus giving 
full opportunities for an invasion of animals from the theo- 
retical western land. We may suppose that this corresponded 
with the Oligocene Period and with the time when, as Pro- 
fessor Osborn tells us, there was a re-establishment of the 
faunal resemblance of south-western North America with 
Europe. Possibly Chile, which was connected at an earlier 
period with this same western belt of land, became separated 
from it. This again was succeeded by a period of marine 
transgression in the west. Even northern Mexico was largely 
covered by the sea, as well as both sides of Lower California 
and a large portion of western California. All this time 
western South America must have risen gradually above the 
sea, and I presume that certain fragments of land, like 
Peru, became joined to the long peninsula which stretched 
far southward running parallel with the newly formed west 
coast of South America. Thus while North America no longer 

* Smith, J. Perrin, " Geological History of California," p. 348. 


received its quota of European emigrants, they found suitable 
accommodation on the newly-formed land of north-western 
South America. The mid- At] an tic land bridge now ceased 
to exist, and far-reaching geographical changes super- 
vejied almost everywhere. Owing possibly to the extensive 
subsidences in the Pacific Ocean, the land that I assume to 
have occupied part of the eastern Pacific, was gradually re- 
duced in size. But, being no longer tenanted by the more 
vigorous and more advanced European types, South American 
forms now commenced to occupy this land, thus first reaching 
the Antilles and then North America as it again became fully 
connected with the western land. Central America in its pre- 
sent shape had not yet come into existence. Not long ago it 
was thought that the Central American isthmus must have 
been first utilised as a highway at the beginning of the Plio- 
cene Period, and that then the great edentates began to pour 
into North America. Now the surprising evidence has come 
to hand that true edentates of the Megalonyx type occur in the 
Middle Miocene Mascall beds of Oregon. Dr. Matthew writes 
to inform me that the Megalonyx remains only doubtfully 
belong to this horizon. Professor Osborn,* on the other hand, 
believes in the discovery, and expresses the opinion that it 
tells in favour of my theory of the former existence of a 
western land connecting North and South America inde- 
pendently of Central America. Still, he points out that it is 
inconsistent with the fact that other animals did not pass 
south or north. 

Let us examine some of the other new arrivals in North 
America in Miocene times, and endeavour to trace their geo- 
logical history. One of the most noteworthy of these is the 
first appearance in North America of elephants (Mastodon). 
Since Dr. Andrews' f surprising discoveries in northern 
Africa were published, Africa is generally looked upon as the 
original home of elephants, the earliest form being Moeri- 
therium, from which later on Palaeomastodon and its more 
modern relations arose. The first of the latter was the Mas- 

* Osborn, H. P., "Age of Mammals," pp. 289292. 
t Andrews, C. W., "Tertiary Vertebrates of the Fayum," pp. xvi 


todon called Trilophodon on account of the three transverse 
rows of cusps on its intermediate grinding teeth. Now Trilo- 
phodon arrived in Europe and in North America at about the 
same time during the Miocene Period. Professor Osborn 
assumes that these mammals came from Asia, although we 
possess no evidence of their having reached the northern or 
eastern parts of that continent. We might be tempted to 
invoke a direct land connection between Africa and South 
America in Oligocene times, but, as we shall see later on, that 
connection must have disappeared at a still earlier period. 
However, these and other problems will be considered in the 
next chapter. 

Several important zoogeographical features of western and 
northern South America still remain to be considered. Special 
researches among the Cretaceous rocks and their fossils in 
Peru have shown that during Lower Cretaceous time, that is to 
say, towards the latter part of the Mesozoic Era, the greater 
part of the country was buried deeply beneath the ocean. 
From Bolivia and Chile, even as far south as the Strait of 
Magellan, Lower Cretaceous deposits have been discovered. 
North of Peru they occur in Colombia and Venezuela. The 
most surprising circumstance connected with these South 
American beds, however, is the great number of species that 
are either identical with or closely allied to, such as occur iq 
the Cretaceous deposits of north Africa, the south of France, 
Switzerland and the neighbouring countries.* More than 
sixty years ago D'Orbigny already drew attention to this fact, 
and argued from it that a land connection across the mid- 
Atlantic must have enabled species to cross the ocean by 
travelling along a continuous shore-line. On the other hand, 
scarcely any affinity exists between the Cretaceous of Vene- 
zuela and that of Mexico or Texas, thus clearly implying the 
presence of a land barrier between these two areas. The old 
highland of Guiana east of Venezuela was long ago a penin- 
sula of the archaean highlands of Brazil in the south. There 
is reason to believe that the great mountain chain of the 
Andes gradually emerged out of this sea. During this process 
some of the newly-formed islands probably became attached 

* Paulcke, W., " Kreideformation in Sudamerika," pp. 305 308. 


to parts of pre-existing western lands. All through Tertiary 
time the mountains must have continued to rise, though our 
knowledge of later geological history is still meagre. We know 
that Tertiary marine deposits occur in the Orinoco valley, and 
it is likely that a narrow marine channel still separated north- 
western South America from the rest of the continent during 
the earlier part of the Tertiary Era. The Amazon valley no 
doubt was at that time a hay of the Pacific ; still, I am unaware 
of any geological or zoogeographical evidence for Professor 
Osborn's supposition that north and middle South America 
were completely divided in Miocene times by a wide sea. That 
the Orinoco and Amazon valleys were in communication with 
one another for a long time is shown by the fact that one of 
the species of manatees (Trichechus inunguis) and the fresh- 
water turtle Podocnemis expansa are confined to the upper 
portions of these two great rivers. A most surprising 
confirmation of the theory that an ocean bay extended to the 
neighbourhood of the Andes has been discovered near the small 
town of Pebas, on the upper Maranon, more than twenty 
degrees of longitude west of the mouth of the Amazon. Pro- 
fessor Boettger described deposits from this locality contain- 
ing typically brackish water mollusks which could only have 
lived in the neighbourhood of the sea. He naturally came to 
the conclusion that the Atlantic then had invaded the Amazon 
valley so as to extend near to the foot of the Andes. But 
Dr. Katzer's view, already alluded to, according to which the 
Amazon drainage only changed eastward in later Tertiary 
times, appears to me to agree better with the zoogeographical 
features of eastern South America. Professor Boettger * 
looked upon the Pebas beds as being of Oligocene, possibly 
Eocene age. The fresh-water fish fauna of the Pacific slopes 
of southern Ecuador still exhibits such affinity to that of the 
Amazon that the Ecuador mountains could only have had a 
slight elevation until comparatively recent geological times. 
Hence we may assume that the Pacific extended to the neigh- 
bourhood of Pebas when thes;e brackish water beds were laid 

An interesting zoogeographical demonstration of the 

* Boettger, 0., " Die Tertiarfauna von Pebas," p. 503. 


gradual elevation and consequent slow change of the marine 
character of an ancient lagoon is afforded by Lake Titicaca 
on the borders of Peru. This lake, with a length of eighty 
miles, lies in a mountain valley over 12,000 feet above sea- 
level, and occupied not long ago a much larger area. To sup- 
pose that this region should have risen from sea-level to such a 
height, and still preserve the remnants of an ancient marine 
fauna dating back to the period when it was a gulf of the 
Pacific, would seem a very bold theory. Professor Suess,* 
indeed, expresses the opinion that the presence of a marine 
fauna in Lake Titicaca cannot be regarded as a sufficient proof 
of the theory that the lake was at sea -level within recent 
geological times. I quite concur with Professor Suess in so 
far as the assumption of a recent elevation is concerned, but 
we have reason to believe that certain ancient forms of animal 
life, particularly among aquatic groups, have transmitted 
their specific characters unchanged to their modern descen- 
dants. It is conceivable, therefore, and even possible, that the 
striking affinity of the fauna of Lake Titicaca to that of the 
Pacific coast may have been preserved, although the actual 
junction of the lake with the sea took place perhaps as far back 
as early Tertiary times. The theory of the recent elevation was 
first mooted by Mr. A. Agassiz,f owing to the discovery in the 
lake of eight species of the marine amphipod Allorchestes, one 
of which (A. dentatus) differs but slightly from a form still 
inhabiting the Strait of Magellan. That fact alone might 
be attributable to accidental dispersal, although the enormous 
difference in height between the sea and the lake, and the 
circumstance of there being eight different species of Allor- 
chestes, would be difficult to explain on that theory. But 
besides this marine crustacean other members of a marine 
fauna have been shown to exist in Lake Titicaca, and thus 
the case against accidental dispersal has assumed a stronger 
position. The fish fauna consists of a catfish (Pigidium 
rivulaturri), belonging to a genus which is very widely distri- 
buted all over South America, and several species of Orestias. 
The latter genus is quite confined to Lake Titicaca, and since 

* Suess, E., " Antlitz der Erde," I., p. 693. 
t Agassiz, A., " Lake Titicaca," p. 287. 


its nearest relations are all marine forms, its ancestors, as 
Professor Eigenmann * remarks, could only have entered the 
area when it was still a gulf of the sea. 

In the Eocene Period " Archiguiana," as Dr. von Ihering 
named the ancient highland of Guiana and eastern Venezuela, 
was supposed by this writer to have been isolated from the 
highland of Brazil. And, indeed, the mountain plateau of 
Guiana contains a very large number of archaic and most 
peculiar types, some of which seem to spread westward into 
Venezuela and Colombia rather than into Brazil. Yet the 
great majority of these ancient forms of Guiana also occur 
southward in eastern Brazil. One of the most noteworthy 
birds, the hoatzin (Opisthocomus hoatzin), whose young 
climb about among the branches of the trees by means of 
well -developed claws on their wings, and which have been 
placed into a distinct order by themselves, range southward 
as far as Bolivia. The chatterers (Cotingidae) comprising 
some of the most ornate and peculiar birds of South America, 
are almost equally divided between Brazil and Guiana. The 
familiar umbrella bird (Cephalopterus ornatus), the bell bird 
(Casmorhynchus niveus), the bald-headed crow (Gymnoce- 
phalus calvus), and the cock-of-the-rock (Rupicola crocea), all 
belong to genera which have spread westward from Guiana 
rather than into Brazil. In all these cases we have to deter- 
mine what was the original centre of dispersal. The singular 
genus of snails Ampullaria probably spread across South 
America from a Brazilian centre, and so did the fresh -water 
crab (Pseudothelphusa) and the whole family of fishes called 
Cichlidae. The snail Strophocheilus, the fresh-water mussel 
Unio, the archaic arthropod Peripatus, the family of tortoises 
Cinosternidae and others, have apparently entered Brazil from 
the north and west. As I shall endeavour to show in the next 
chapter, many of the forms that have spread from the Brazi- 
lian highlands have near relations in Africa, while among the 
northern and western immigrants into Brazil scarcely any 
have succeeded in crossing the Atlantic area to Africa. 

* Eigenmann, C. H., " Freshwater Fishes of South America," p. 521. 



WHEN we consider that Brazil covers an area larger than 
that of the whole of Europe, merely the roughest outlines of 
the general zoological features of that vast country can be 
attempted in a work of this nature. The highlands of Brazil, 
as I have mentioned already, are confined to the east, being 
surrounded by the ocean on one side and by a continuous 
tract of lowlands on the others. It is believed that the 
mountains were once continuous with those of Guiana 
right across the present mouth of the river Amazon. 
The whole of the area between the two great highlands 
of Brazil and Guiana is drained by the mighty Amazon, 
whose waters exceed in quantity that of any river in the world. 
The Paraguay and its tributary the Parana drain the south- 
western parts of Brazil. Practically the whole of the country 
lies in the tropics. The combined influence of high tempera- 
ture and abundance of moisture thus produces that exuber- 
ance of animal and plant life which is so characteristic of 
Brazil. The great tropical forests are truly bewildering in the 
wealth and variety of their vegetation, and it is here that we 
realize more than anywhere else the keen struggle for exist- 
ence as applied to plants and animals. In their constant efforts 
to reach the light, and in their endeavour to obtain a full share 
of it, the trees crowd and press upon one another until the 
weaker succumb. In their turn they are invaded again by 
others, while hosts of enemies, parasitic plants and climbers, 
twine round and strangle the less sturdy long before they 
succeed in attaining their object. We are thus presented with 
a scene of savage warfare among plants. It is less evident 
among animals. But the ringing sound of the multitude of 
voices issuing out of the forest reminds the traveller that in 
this teeming profusion of nature there must be an acute 


struggle for life among animals as well as among plants. 
And yet, except birds, insects and lizards, animals do 
not appear to be present in unusually large numbers. The 
large class of mammals, which elsewhere form such a con- 
spicuous feature of a fauna, seem almost to be absent. This 
is largely due to the fact that Brazil is really poor in terrestrial 
mammals. Those that do inhabit the country are chiefly of 
arboreal habits, and thus escape attention. Mr. Bates * 
believes that the South American fauna has been slowly 
adapted to an arboreal life, and that extensive forests 
must always have existed since the region was first peopled 
by mammalia. 

Among these arboreal mammals the family of the capuchin 
monkeys (Cebidae), whose unusually prehensile tail gives 
them peculiar facilities for climbing, are the most noteworthy. 
They range all over tropical America, being most abundant 
in the dense forest regions of Brazil. In some monkeys, like 
the howlers (Mycetes), the end of the tail underneath is devoid 
of hair, and thus acts with even greater efficiency than in 
capuchins (Cebus). The sakis (Pithecia) and squirrel-mon- 
keys .(Chrysothrix) have non-prehensile tails. The spider 
monkeys, with their long limbs and long prehensile tail, are 
the most admirably adapted creatures for a purely arboreal 
life. A second family (Hapalidae) includes the smallest of all 
monkeys, the marmosets. Mostly very active little squirrel- 
like creatures, with arboreal habits, they are almost confined 
to Brazil and north-western South America. In many respects 
these two families of South American monkeys are closely 
related to one another, whereas they differ from the Old World 
species, especially in their dentition. Since no members of 
the Cebidae and Hapalidae have ever been found fossil outside 
South or Central America, it is believed that they form a 
branch distinct from the Old World monkeys, having had a 
separate origin from lemur-like creatures. Dr. Ameghino has 
described quite a number of lemuroid remains from the Cre- 
taceous of Patagonia, and he maintains that these early fore- 
runners of monkeys and man originated in the ancient vast 
territories of southern South America. I am fully aware that 

* Bates, H. W., "Naturalist on the Amazons," p. 32. 


the views of Dr. Ameghino * as to the age of these beds have 
received a considerable amount of adverse criticism. Still, 
if we assume the correctness of his arguments and the former 
existence of a land bridge between South America and Africa, 
these lemuroid mammals might have passed from Patagonia, 
as Dr. Ameghino supposes, to Africa and thence to Europe, 
and lastly, from there to North America. I do not think that 
this was the history of events. One distinct branch may 
have travelled from Patagonia to Chile, and thence direct to 
North America by a western land connection (compare 
Fig. 14), which I have already mentioned and which will be 
further discussed in the next chapter. From North America it 
may have passed into Europe by the mid- Atlantic land bridge. 
I doubt whether a separate branch reached Africa from South 
America by a land bridge, which Dr. Ameghino contends 
joined these two continents. However, it is this very problem 
of the zoological affinity between South America and Africa 
and its origin which will be dealt with in this chapter. 

What we have to consider principally, therefore, is 
whether there are really such affinities between the living 
faunas of the two continents as to make it probable that 
the latter were once connected with one another by land. 
The capuchin and marmoset families, which are quite con- 
fined to South and Central America, have probably originated 
there in the remote past and have not been able to pass into 
any other continent. All we know of their immediate ancestry 
is that in the Eocene deposits of Patagonia a new family 
of monkey-like creatures arose, possessing certain marks of 
resemblance to the two recent South American families. They 
were named " Homunculidae " by Dr. Ameghino. As these 
also are quite unknown outside South America, it would 
appear as if Patagonia had become isolated during the course 
of the Eocene Period from the rest of the world. We cer- 
tainly have no evidence of any Tertiary land connection 
between the southern portions of South America and Africa 
from the distribution of monkeys. 

Another typically Brazilian arboreal mammal is the sloth, 
which lazily and cautiously moves from branch to branch 

* Ameghino, Fl., " Formations sedimentaires de Patagonie," p. 289. 


supported by the long claws of its toes. Two genera of this 
edentate mammal are known, viz., Bradypus and Choloepus, 
both of them confined to South and Central America, and, like 
the monkeys, absent from Argentina, Chile and Patagonia. 
In spite of their absence from the latter country, it is there, 
according to Dr. Ameghino, that we find the earliest traces 
of the sloth-tribe in the Eocene beds. The Eocene Entelops 
and Trematherium have been placed into the same family 
with the modern sloths. Certain sloth-like remains have even 
been traced back to the Upper Cretaceous of Patagonia. 
Sloths of the arboreal type have not been found fossil out- 
side South America. The remains of a supposed sloth (Brady - 
therium) were discovered by Mr. Grandidier ten years ago 
on the island of Madagascar. According to Dr. Smith Wood- 
ward they belong to a lemur, yet there are so many other 
points of affinity between South America and Madagascar 
that a former direct land connection between the two regions 
has been suggested. I return to this subject more fully in 
the next chapter. 

In a previous chapter (p. 70) I explained that, although 
all American porcupines are arboreal in distinction to the 
Old World species which live on the ground, only the 
South and Central American forms have prehensile tails. 
This gives them greater facilities for dispersal in Brazil, and 
greater protection from their enemies. These South 
American porcupines agree in their general range with the 
monkeys and the sloths, and like them are absent from the 
southern states of South America, although their ancestral 
home was seemingly in Patagonia. Some species of Coendu 
are known from the Brazilian caves, still, the centre of origin 
lay manifestly further south. 

Of the pouched or marsupial mammals we have noticed that 
the opossum has a wide range in North America. In Central 
and South America the same North American species 
(Didelphys marsupialis) occurs widely distributed. Such an 
enormously extensive range must be due to the fact of its 
being a persistent mammalian type. The genus, or one 
closely related to it, certainly was already represented in the 
Lower Eocene of North America and the Upper Eocene of 
France. Only one genus of these marsupial mammals occurs 


in North America. From South America three others are 
known. The general range of the family Didelphyidae points 
to South America as the centre of dispersal. Although the 
genus Didelphys or Peratherium has been met with in the 
Eocene of North America and France, while it first appears 
in South America in the Miocene Period, if Dr. Ameghino is 
correct, the earliest member of the family (Proteodidelphys) 
occurs in the Lower Cretaceous beds of Patagonia. Even if 
we look upon these beds, with Professor Osborn, as really of 
Eocene age, the more primitive characters of Proteodidelphys 
point to South America as the ancestral home of the family, 
and on this continent no doubt the genus Didelphys has 
originated a,nd not in south-eastern Asia, as suggested 
by Mr. Lydekker.* I think the geological history of the 
opossums, though dating further back than that of the South 
American monkeys, followed much upon the lines of the 
groups just considered, at any rate, they seem to have entered 
Brazil about the same time. 

Among the birds of South America we have precisely 
similar examples, except that in their case we know unfor- 
tunately very little of their past history from palaeontological 
evidence. The wonderful family of humming birds (Trochi- 
lidae) is comparable in distribution with the opossums, in so 
far as it ranges all over South and Central America. It 
has in all likelihood entered North America in later geological 
times. Not a single species of humming bird is known be- 
yond the confines of America. It is of importance to note that 
of the one hundred and eighteen genera admitted by Dr. Har- 
tert,f the great majority are confined to the west coast. Some 
of them inhabit Chile, others Peru, Bolivia, Ecuador, 
Colombia, Central America and Mexico. A few (Oreotro- 
chilus) live at enormous heights, up to 20,000 feet. Others 
are limited in their range to the Antilles and Brazil. Only 
the single genus Avocettula, with one species, is peculiar to 
Guiana. This seems to suggest that the family originated in 
western South America, and has only gradually spread east- 
ward on the mainland. The West Indian area no doubt was 

* Lydekker, E., " Geographical History of Mammals," p. 112. 
t Hartert, E,, " Trochilidae," 


invaded by the humming birds in earlier times from the west, 
since several very distinct genera with well marked characters 
have originated there. The humming birds constitute a 
sharply denned family, whose nearest relatives are the tree- 
swifts (Macropteryx) of southern Asia and western Polynesia. 
The common ancestors of these two groups may possibly 
have inhabited part of the supposed ancient Pacific continent, 
and may have diverged from it east and westward when it sub- 
sided. As Dr. Wallace* aptly remarks, no naturalist can 
study in detail this single family of birds without being pro- 
foundly impressed by the vast antiquity of the South 
American continent, its long isolation from the rest of the 
land surface of the globe, and the persistence through count- 
less ages of all the conditions requisite for the development 
and increase of varied forms of animal life. 

The chatterers (Cotingidae) which have already been alluded 
to as comprising some of the most beautiful and noteworthy 
of American birds, are almost confined to South and Central 
America and the West Indies. Only a few species enter North 
America in Mexico. The curious toucans (Rhamphastidae) 
with their huge bills and strange texture and coloration of 
their plumage, as well as several other families, are restricted 
to South and Central America and have no near relations in 
the Old World. 

If South America had been connected by land with Africa 
in later Tertiary times, we should expect, at any rate, such 
mobile creatures as birds and also butterflies to indicate by 
their distribution the former existence of such a land bridge. 
Among the most conspicuous butterflies in the Brazilian 
forests, Mr. Bates mentions the Heliconiidae, which are 
readily recognisable by their long narrow black wings, 
variously ornamented with white, crimson and yellow dots 
or stripes. Yet the family is almost purely South American. 
A single species only (Heliconius charitonia) passes into 

The Morphidae likewise range from Argentina to Mexico, 
some of the larger species, with their dazzling metallic 

* Wallace, A. E., " Distribution of Animals," II., p. 9, 
t Stichel, H., and H. Biffarth, "Heliconiidae," 


colours, being characteristic of the Brazilian forest fauna. 
The family is unknown in the Old World, but, as in the case 
of the humming birds, we have closely allied groups (Ama- 
thusidae, etc.) in western Polynesia and southern Asia, so 
that the Morphidae or their ancestors may possibly have 
entered South America, as so many other groups have done, 
from the west. 

The large and mostly active lizards belonging to the family 
Teiidae are spread all over South America, ranging northward 
as far as California and Texas and through the West Indies. 
They appear to have only extended their range into North 
America in comparatively late geological times, although they 
are known from early Tertiary deposits in South America. 
Nevertheless the Teiidae are quite unknown in the Old World. 
Instances of that kind might easily be cited from every group 
of vertebrates and invertebrates. 

Yet although these examples seem to show that the faunistic 
community between South America and Africa is so slight 
as not to necessitate the hypothesis of a former land connec- 
tion between these continents, we possess other very striking 
features of distribution which are strongly opposed to such 
a conclusion. I may draw attention again to the most re- 
markable fact, pointed out by Professor Bouvier,* that the 
genus Peripatus is peculiar to America, with the single ex- 
ception of Peripatus tholloni of West Africa. All the other 
Peripatus -like creatures have been proved by Professor 
Bouvier to belong to different genera or groups. 

During his travels in the Amazon region, Mr. Bates dis- 
covered a curious archaic spider-like creature, which was 
named Cryptostemma westermanni. More recently the same 
species has been recorded from Sierra Leone and from the 
Cameroons in West Africa. Those who allege that accidental 
dispersal is responsible for such cases of distribution, pro- 
bably accept the discovery of Cryptostemma westermanni 
on both sides of the Atlantic as a notable instance of the 
facility with which species are wafted across the ocean. But 
we must remember that there are thousands of species of 
animals and plants inhabiting the coasts of West Africa and 

* Bouvier, E. L., " Monographie des Onychophores," I., p. 91. 
L.A. B B 


Brazil that are specifically and generically distinct from one 
another, although more fitted for accidental dispersal across 
the ocean than Cryptostemma is. The instances of such 
specific or generic identity of animals and plants, in fact, are 
scarce. Moreover, Cryptostemma is the single survivor that 
has yet been brought to light of the otherwise extinct order 
Meridogastra, which occupies a somewhat intermediate posi- 
tion between the true spiders and what are called "harvest- 
men."* For this reason it must, like Peripatus, be looked 
upon as an exceedingly ancient relict member of our fauna. 

Let us take another group of apparently very ancient 
animals, the worm-like and limbless coecilians, which, live 
underground. Dr. Boulengerf tells us that of the genus 
Dermophis three species are found in Central and South 
America, one in West Africa and another in East Africa, and 
that the genus Herpele is confined to Gaboon in West Africa 
and Panama. Dr. Sarasin argues that the dispersal of 
Herpele at any rate dates from pre-Cretaceous times. It 
seems possible, therefore, that these few archaic creatures 
indicating faunistic relationship between South America and 
Africa, have obtained their present range during some very 
remote geological period, when the conditions of land and 
water were entirely different from what they are at present, 
and that they are not to be regarded as instances of accidental 
dispersal across the Atlantic. The suggestion that South 
America and Africa were once united by land is not a new one. 
It has been made, as we have learnt, by various authorities on 
entirely different grounds. Considering the contradictory 
nature of the evidence, however, the problem requires close 

I have already stated that Dr. Ameghino had expressed the 
opinion, based on the evidence of the fossil mammals, that 
South America and Africa were joined by a land bridge during 
the whole of Upper Cretaceous times. During the Eocene 
Period this land connection, he thinks, became more restricted 
or narrowed down, while it still persisted incompletely as a 
chain of islands, until middle Miocene times. Dr. Ame- 

* Karsoh, F., " tJber Cryptostemma," pp. 2529. 

t Boulenger, G. A., "Synopsis of apodal Batrachians," pp. 404409. 


ghino's * arguments in favour of this union of South America 
and Africa are not founded on any palaeontological resem- 
blance between these two continents, but rather on the affinity 
of the fossil mammals of South America to those of Europe, 
Asia and North America. Since I have shown that until 
about Oligocene times southern Europe was connected by 
land with western South America, by way of the West Indies 
and part of Central America, there does not seem to be any 
necessity for a second land bridge further south in order to 
account for the mammalian affinities existing between South 
America and Europe as well as Asia and North America. 
Professor Osborn does not recognise the existence of any 
former land bridge during the age of mammals between South 
America and the Old World except by way of North America. 
Mr. Lydekker f argues that the only marked community 
between the Ethiopian and Neogaeic( South American) faunas 
as regards mammals, relates to the hystricomorphous rodents, 
but he thinks this community is a very marked one and diffi- 
cult to explain on any other hypothesis than that of a land 
connection between the two areas. The Hystricomorpha are 
a section of the rodent mammals, well distinguished by mor- 
phological characters from the other sections of that order. 
It is of the greatest importance to note that this hystrico- 
morphous section is now confined to Africa and America, 
with the exception of a couple of genera which range into 
southern ^Europe and Asia. The centre of distribution is 
no doubt South America. If they had passed from there to 
Africa, we .should expect them to be found in Brazil and 
western Africa. Let us examine the two closely related 
families of Octodontidae and Ctenodactylidae, which are in- 
cluded in the section Hystricomorpha. The first is con- 
fined to South America, the other to Africa. Of the former 
it is the sub-family Octodontinae which is nearest related 
to the Ctenodactylidae, and almost entirely confined to 
Argentina and the west coast of South America. Only a few 
species like Ctenomys brasiliensis and Ct. minutus really 
enter Brazil. The extinct species are all but one confined to 

* Ameghino, F., " Formations s6dimentaires," pp. 281 287. 
t Lydekker, B., " History of Mammals," p. 127. 

BB 2 


Argentina. It seems manifest, therefore, that the ancestral 
members of this sub -family have travelled along the west 
coast of South America probably from a southern centre of 
dispersal. We find no trace of them in Central America or the 
Antilles, but it is in North Africa where we meet with 
Ctenodactylus with its peculiarly modified inner toes. Here 
in the Mediterranean region, and not in Africa proper, must 
have been the Old World centre of dispersal, for we find the 
allied extinct genus Pellegrinia in the Pleistocene of Sicily, 
and Ruscinomys in the Pliocene of southern France. A 
recent relation of Ctenodactylus (Massoutiera) has passed 
southward towards the Senegal. The main branch, however, 
has apparently invaded eastern Africa from the Mediterranean 
region, giving rise to the genera Pectinator, Thryonomys and 
Petromys. Only a single species (Thryonomys swinderianus) 
has gained the west coast of Africa. Of the sub-family 
Echimyinae, which largely inhabits Brazil, Africa possesses 
no near relations. The only African family of the hystrico- 
morphous rodents, that of the Cape jumping hares (Pede- 
tidae), occupies a more isolated position, its exact relationship 
being still somewhat obscure. But in any case, I fail to 
deduce sufficient evidence from the distribution of these 
hystricomorphous rodents, in favour of a direct land connec- 
tion between South America and Africa, although there must 
have been one between the Mediterranean region and western 
South America by way of the West Indies and Central 
America (see p. 280 and Fig. 14) ., 

Apart from the cape jumping hares (Pededitae), there are 
in South Africa certain mammals which indicate a distant 
relationship with South American ones. The peculiar pig- 
like African edentate Orycteropus occurs in Africa, while 
another edentate, the pangolin (Manis), inhabits Africa and 
the Indian region. Dr. Tullberg thought that these and other 
features implied that south-western Africa must have been 
joined by land to South America during a time when the 
former was completely severed from the rest of Africa. But 
even this land bridge ceased to exist, according to Professor 
Tullberg,* at the beginning of the Tertiary Era, at latest in 

* Tullberg, T., " System der Nagetiere," p. 498. 


Eocene times. He thought that south-western Africa later on 
became united with eastern Africa, thus enabling some of 
these ancient mammals of South American origin to spread 
northward towards Europe and eastward into Asia. But both 
the Manidae and the Orycteropidae are represented in the 
Eocene deposits of southern Europe, while the genus Orycte- 
ropus lived in Samos and on the mainland of Greece in Miocene 
times. Moreover, neither Manis nor Orycteropus are confined 
to southern Africa. Both are distributed north-eastward as 
far as Kordofan and Senaar. It seems more likely, therefore, 
that these edentates, like the hystricomorphous rodents just 
alluded to, have originated in the Mediterranean region from 
South American ancestors and have spread southward subse- 

The same problem has also been discussed by Dr. Andrews* 
as the result of his remarkable discoveries of fossil mammals 
in the Fayum of Egypt. All the Garni vora he found there 
belonged to that archaic group known as the " Creodonta." 
He argues that the presence of these creodonts in Africa would 
account for the existence of the " Sparassodonta " in Pata- 
gonia, if we assumed that during the remote period when these 
ancient groups originated, Africa and South America had 
been joined to one another by land. He also brings forward 
another testimony in support of his suggestion, which I shall 
deal with later on. As for the Creodonta, they are so amply 
represented in the early Tertiaries of both North America and 
Europe, that the assumption of a mid- Atlantic land bridge is 
sufficient to explain their presence in these continents as well 
as in Egypt, while Patagonia must have had some land con- 
nection with North America in late Cretaceous or early 
Eocene times. 

The distribution of birds does not give us many definite 
suggestions as to former land connections between South 
America and Africa. The range of the parrots (Psittacidae), 
however, points to an affinity between these continents. A 
more striking example seems to be that of the ostrich and 
rhea. The former inhabits exclusively Africa and Arabia, the 
other Argentina and Brazil. Hence we might be tempted to 

* Andrews, C. W., " Tertiary Vertebrates of the Fayum," p. xxii, 


explain their present range by a direct land bridge between 
the two continents. The American ostrich (Rhea) is only 
known fossil from superficial deposits, but the ostrich (Stru- 
thio) occurs in the Miocene or Lower Pliocene of the Siwaliks 
of India and of the island of Samos. Since the ostrich once 
lived in the Mediterranean region, it is possible that the 
common ancestor of the two families may have utilised the 
'mid-Atlantic land bridge to travel from the Old World to 
the New or vice versa. 

The reptiles and amphibians, as Dr. Blanford * has pointed 
out, indicate a much more pronounced faunal relationship 
between South America and Africa than the birds or mammals 
do. Professor Pfeffer f endeavoured to explain this very inge- 
niously by the assumption of a former sub -universal, or 
almost universal, distribution and a subsequent extinction on 
the northern continents. He admits that some form of 
land bridge was necessary, of course, yet almost all inter- 
continental communication must have passed, according 
to his views, across a Bering Strait land bridge. He 
quotes a number of instances of groups which are now 
confined to the southern hemisphere, but have once also 
extended to the northern continents, and because they 
have done so he contends that they must have had a 
sub-universal distribution. All those examples which are 
not found fossil in the northern hemisphere are nevertheless 
supposed to have had a similar range and to have gained their 
present southern distribution in different countries by wander- 
ing from one to the other almost by way of the North Pole and 
then south again. Nothing but a careful general study of 
existing distribution can convince us of the fallacy of such 
an assumption. 

Let us take, for instance, the family of fresh- water tortoises, 
the Pelomedusidae. It is confined to Africa, including Mada- 
gascar, and South America. Curiously enough, one of the 
genera of this family, viz., Podocnemis at present inhabits 
only northern South America and Madagascar. But, as Pro- 
fessor Pfeffer tells us, the genus is known as far back as 

* Blanford, W. T., " Anniversary Address," pp. 70 71. 

t Pfeffer, G., " Zoogeographische Beziehungen," pp.417 418, 


the Cretaceous of Patagonia and New Zealand, the Eocene 
of Egypt, England and India, and the Miocene of Egypt and 
Malta. The faunistic relationship between Patagonia and New 
Zealand will be explained later on, yet it has by no means been 
produced by a sub-universal distribution. The only fossil oc- 
currence of Podocnemis we know of from Africa is that from 
Egypt, and even that, with Malta and England, is rather sug- 
gestive of a radiation from a Mediterranean centre. The 
dispersal from an originally Patagonian birthplace may have 
taken place westward to New Zealand arid northward along 
the ancient land connection, which I described as extending 
by way of Central America to southern Europe. It is impor- 
tant to bear* in mind this very peculiar case of faunistic rela- 
tionship between South America and Madagascar, exclusive 
of the African continent, because similar instances are met 
with among many groups of animals and plants. I have 
already alluded to one, and it was first suggested by Dr. 
H. 0. Forbes, I think, that this range was due to a former 
direct land connection between Patagonia and Madagascar. 
I drew attention in a former chapter (p. 173) to the 
members of the family Amphisbaenidae as furnishing, on 
account of their subterranean habits, important evidences 
of former changes of land and water. I may men- 
tion again that they are limbless, wormlike liazrds, and 
that many of them live underground in ants' nests. They 
are not liable, therefore, to accidental dispersal. Their 
distribution is most interesting. They inhabit mostly Africa 
and South America. A few occur in the West Indies and 
the Mediterranean region. Some have even spread into the 
southern States of North America, and we possess a few 
remains from Oligocene deposits. Otherwise we know nothing 
of their past history. We might suppose that some early mem- 
ber of this family had spread across the mid-Atlantic land 
bridge to the Mediterranean region and thence colonised 
Africa with Amphisbaenidae. But in this instance such a 
land bridge cannot help us, because the genus Amphisbaena 
occurs in Africa and South America, while the Mediterranean 
region is inhabited by the genus Blanus. The latter, no doubt, 
may owe its origin to a migration across this mid- Atlantic 
land bridge, still we can scarcely imagine that, coming from 


the north, and after having given rise to Blanus in southern 
Europe, the American ancestor reverted again to Amphisbaena 
when it reached Africa. Amphisbaena, moreover, is represented 
in Brazil by about a dozen species, some of which are very 
closely allied to those of Sierra Leone and Liberia in West 
Africa. If we attribute such a range to a former sub-universal 
distribution, as Professor Pfeffer suggests, then practically 
all animals with a wide southern range must at some time or 
other have passed through 'North America and Asia or Europe 
to roach their present habitats, and for such an assumption 
we have no palaeontologies! evidence. I think a case of this 
kind, as Dr. Blanford remarks, implies a former land con- 
nection between South America and Africa. We possess 
among the Amphisbaenidae an even more striking example of 
faunal relationship between these two continents. There are 
two species of Anops, one of which (Anops kingi) ranges 
from Brazil to Argentina, while the only other species (A. 
africanus) is confined to West Africa. 

Of particular interest are the skinks (Scincidae), as I 
explained (p. 124) that this family must be of very great 
antiquity. Some of the genera certainly date back to Meso- 
zoic times. Now, while skinks are abundant in North 
America, the southern continent is almost entirely free 
from these lizards. Those that live there, at any rate, 
belong to a genus distinct from the northern one. Southern 
Asia and Africa are the headquarters of this large family. 
Only the single genus Mabuia has spread into South 
America. Mabuia occurs principally in southern Asia, 
Madagascar and Africa. A few species are known from 
the Cape Verde islands on the west coast of Africa, 
while one (Mabuia punctata), very closely allied to them, 
has been met with on the island of Fernando de Noronha 
off the east coast of South America and also in Guiana. 
Three species are found in Brazil. A couple have passed 
northward penetrating into Central America and one even 
to the Greater Antilles. We have clear evidence here of 
a former land connection between West Africa and South 
America and of its having been utilised by the skink, Mabuia, 
in traversing the Atlantic in a westward direction. 

The worm-like amphibian Dermophis has not yet been 


noticed in eastern South America, but to judge from the cir- 
cumstance that one species lives in east Africa, another in west 
Africa, a third in Ecuador, and a fourth, in Central America, 
according to Dr. Boulenger,* we may conclude that the genus 
will probably turn up in Guiana. 

The frogs and toads are divided into two groups, one of which 
(a very small one) contains those that are devoid of a tongue, 
while all others possess this useful organ. The tongueless 
toads are generally looked upon as the most primitive of the 
tailless amphibians. Three genera of these creatures are 
known to science, two (Xenopus and Hymenochirus) inhabit- 
ing Africa south of the Sahara, the other (Pipa) eastern 
South America. The only member of the last genus is the 
remarkable Surinam toad (Pipa americana) whose skin on 
the back of the female is provided with pits in which the 
eggs are placed and the young undergo their whole metamor- 
phosis. These tongueless toads (Aglossa) are not known 
as fossils. We have thus to conjecture the origin of their dis- 
persal from their present distribution which is highly sug- 
gestive of a former direct land bridge between Africa and 
South America. This view is considerably strengthened by 
the fact that the West African Hymenochirus is in several 
respects more closely related to Pipa than to Xenopus. 

The toothless frogs (Dendrobatidae) appear at first sight 
to be typical examples of a group which has obtained its pre- 
sent range by means of a former land bridge between South 
America and Africa. They occur only in northern South 
America, in west Africa and Madagascar, a discontinuous dis- 
tribution which betokens antiquity, and which is not unknown 
among other animals and plants. Yet Dr. Gadowf argues 
that the Old World and New World genera were evolved inde- 
pendently from toothed frogs, that this is, in fact, a case of 
convergence. It may be so, but his arguments are by no 
means convincing. 

I may once more be permitted to draw attention here to 
the opinion long ago expressed by Dr. Gill J that fishes are 

* Boulenger, G. A., " Synopsis of apodal Batrachians," p. 404. 

t Gadow, H., "Amphibia and Eeptiles," p. 272. 

I Gill, T., " Principles of Zoogeography," pp. 2930, 


among the best indicators of past continental changes, 
and that the relations of the several southern continents as 
to their ichthyio faunas can be best understood by the assump- 
tion that, at some remote epoch or epochs, there was distribu- 
tion of land and water which eventually permitted an emigra- 
tion and immigration of types from one into another. Dr. 
Giinther * enunciated similar views. He thought that the 
existence of so many similar forms of fishes on both sides of 
the southern Atlantic supported the supposition that they are 
the descendants of a common stock which had its home in a 
region now submerged under some intervening part of the 
ocean. Since the days when these two great authorities pro- 
nounced upon this subject, the ichthyology of both Africa and 
South America has become better known, and new light has 
been thrown on the relationship between the various groups 
of fishes. We can deal with the problem now in the light of 
the latest researches. Dr. Boulenger has made a special study 
of the African fish fauna, while Mr. Regan and Professor 
Eigenmann have paid particular attention to the ichthyology 
of South America. In his interesting address to the British 
Association on the distribution of African fishes Dr. Bou- 
lenger mentions Protopterus as the most noteworthy fish of 
Africa. In some respects it approaches the amphibia. It 
possesses a double lung as well as gills, and is able to live for 
a long period in dried mud. Anatomically there are many 
characters in which it differs from almost all other fishes. 
Its only living relation, Lepidosiren of Brazil, is very similar 
in form and so closely allied that the two are placed into 
the same family Lepidosirenidae. Professor Pf effer f con- 
tends that since this family is represented in the Permian 
and Trias of western North America, in the Trias of India 
and South Africa, in the Cretaceous of Patagonia and the 
Trias and Jurassic of Europe it must have had a universal dis- 
tribution. Hence he argues that the occurrence of Pro- 
topterus in Africa and of Lepidosiren in South America can 
have no significance in supporting the theory of a former 
land connection between these two continents. But recent 

* Gunther, Albert, " Study of Fishes," p. 233. 

t Pfeffer, G., " Zoogeographische Beziehungen," p. 433. 


researches have shown that these extinct forms all belong 
to the genus Ceratodus, that all are members of the family 
Ceratodidae.* The only fossil member of the Lepidosirenidae 
is Protopterus libycus of Egypt. There is really no reason, 
therefore, as Dr. Boulenger f puts it, why the latter family 
should not have passed from one of the two continents into the 
other when they were connected by land. 

Another striking feature of the South American fresh-water 
fish fauna is the extraordinary number and variety of forms 
of the Characinidae, unquestionably, according to Dr. Bou- 
lenger, one of the most lowly and generalised groups of ex- 
clusively fresh-water teleosts. They are likewise abundantly 
represented in western and central Africa, yet not a single 
genus is common to both continents. Hence Dr. Boulenger J 
concludes that it is quite legitimate to explain the distribu- 
tion of this family by the assumption of a land bridge con- 
necting Africa and South America about the end of the 
Cretaceous Period. 

A similar argument is used by Mr. Began to account for 
the large fresh-water family Cichlidae (compare Fig, 12) in 
these two continents, except that he is of opinion that the 
land bridge was still in existence in early Eocene times. 

Professor Eigenmann || states that there is no known means 
by which these two families could have crossed the existing 
gap between Africa and South America. There has been no 
exchange of species in recent times, for there is no species 
or genus common to the two continents. The South American 
and African elements of these two families must have been 
derived from some intermediate land-mass, or must have 
travelled from one continent to the other over a land bridge. 
That this connection must have been obliterated before the 
Tertiary Era is evidenced, he remarks, by the fact that the 
Tertiary deposits of Taubate and Parana include existing 
genera, and that many South American families of fishes 

* Goodrich, E. S., " Cyclostomes and Fishes," p. 258. 
t Boulenger, G. A., " Distribution of African Fishes," p. 5. 
J Boulenger, G. A., " Distribution of African Fishes," p. 6. 
Eegan, T., " Fishes of Central America," p. xiv. 
|| Eigenmann, 0., " Fishes of South and Middle America," pp. 525 


which are absent from Africa have arisen in South America 
from the Characinidae and Siluridae since the separation of 
the two continents. 

Dr. von Ihering's early recognition of the importance of 
fresh-water faunas as an aid to the palaeogeographical studies 
of the Palaeozoic and Mesozoic Eras, has led to very remark- 
able results. Some of these I have indicated already. His 
noteworthy discovery that some of the fresh-water bivalves 
of South America have a " lasidium " larva, while all the 
Unionidae possess a " glochidium," placed the family Mute- 
lidae into quite a different position. The Mutelidae are a 
family essentially South American and African, demonstrat- 
ing clearly, as Dr. von Ihering * admits, the existence of a 
land connection between South America and Africa during the 
Mesozoic Era. 

The family Achatinidae includes some of the largest and 
most conspicuous land-snails, Achatina achatina growing to a 
length of seven inches. Their distribution, which is well 
known, indicates an African centre of evolution. The early 
members of the family are unknown and should be looked for 
in mid-Mesozoic deposits, according to Dr. Pilsbry.f The 
same authority believes that some of the sections o'f the Acha- 
tinidae migrated to South America before the interruption of 
the land connection across the tropical Atlantic. Dr. Pilsbry, 
who in his earlier volumes condemns the practice of throwing 
hypothetical bridges across the oceans, now almost takes it as 
a matter of course that Africa and South America were once 
united by land. The mollusks, as most other groups of 
animals, yield facts of distribution that are quite unexplain- 
able by the theory of accidental dispersal, and many of those 
who at first were firm believers in the immutability of our 
ocean basins, have entirely changed their opinions after a 
careful study of zoogeography. It is the smaller kinds of the 
Achatinidae which show the affinity between the two con- 
tinents most clearly. Thus the genus Subulina is confined to 
Africa and tropical America while Opeas and Pseudopus have 
spread to other parts as well. 

* IherinT, H. von, " Archhelenis and Archinotis," pp. 125 145. 
t Pilsbry, H. A., " Manual of Conchology," Vol. XVIIL, p. vi, 


It is now fifty years since Mr. Andrew Murray * first 
directed attention to the fact that the beetle fauna of Old 
Calabar in West Africa presented certain affinities with that 
of South America. He even then suggested that some sort of 
communication must once have existed between these two 
regions of the earth. A few years later he returned to the 
same problem, expressing the opinion that this communica- 
tion consisted of an actual land bridge of which the only re- 
maining vestiges are the islands of Ascension, St. Paul's, St. 
Helena and Tristan da Cunha.f 

Great stress is also laid on this remarkable relationship 
between the southern continents by Professor Kolbe,J but he 
explains it by the assumption of a land bridge far to the 
south of the Equator. 

Dr. Packard was good enough to inform me some years 
ago that the distribution of the Lepidoptera was distinctly 
in favour of the theory of a former union between South 
America and Africa. He alluded in particular to a family of 
moths known as the Saturnidae, stating that their general 
range confirmed the view arrived at from other sources, that 
perhaps at the close of the Cretaceous Period and through 
the early part of the Tertiary Era the two continents were 
connected with one another by land. 

The importance of the fresh-water crabs in the solution of 
problems of this nature has been emphasised, as I mentioned 
before, by Dr. Ortmann,|| who showed that the west African 
Potamoninae are geographically most closely approached by 
the South American Potamocarcininae, and that this suggested 
a former union of these regions. This land bridge in its full 
extent, he thinks, existed during the Jurassic and in early 
Cretaceous time. In the middle of the Cretaceous Period the 
southern Atlantic advanced northward and gradually invaded 
the east coast of South America extending as far as the 
Amazon valley. Guiana still remained joined to West Africa 

during the remainder of the Cretaceous Period, and was not 

* Murray, A., " Coleoptera of Old Calabar," pp. 453454. 
t Murray, A., " Coleopterous Faunae," p. 15. 
| Kolbe, H. J., " Die Coprophagen Lamellicornier," p. 503. 
Packard, A. S., " Larval Forms of Moths," p. 280. 
|| Ortmann, A. E., "Distribution of Decapods," pp. 350351. 


finally destroyed until early Tertiary times (compare Figs. 15 
and 20). Quite recently, however, Dr. Ortmann * argued that 
this " Archhelenis " of Dr. von Ihering must have sunk 
beneath the waves of the ocean by the end of iJie Mesozoic Era. 

Some time ago Mr. Pocock f pointed out that the scorpions 
of the genus Opisthacanthus were confined to tropical Africa, 
Madagascar and South America. Our knowledge of the dis- 
tribution of scorpions has increased considerably since he 
wrote his essay, but the range of Opisthacanthus has not been 
greatly extended. It occurs in west, east and south Africa 
and Madagascar, while it reappears on the opposite side of the 
Atlantic in Colombia and the island of Haiti. It is not known 
from either Brazil or Guiana and may possibly have become 
extinct there. On, the other hand, Damon, one of the Tarantu- 
lidae, occurs in west and east Africa and from Brazil to Pata- 
gonia. Professor Kraepelin | assumes that because the African 
Damon variegatus in a specifically identical form actually 
inhabits South America, it must have been introduced into the 
latter continent, yet in another place he urges that several 
of the centipedes (Scolopendridae) probably maintained their 
Bpecific characters from archaic times to the present day. Like 
some of the species of Otooryptops, Rhysida and Scolopendra 
and the curious arachnid Cryptostemma, Damon variegatus 
may possibly represent one of these ancient species which 
has retained its morphological characters unchanged since 
Mesozoic times. 

The numerous examples, cited above, indicate that there 
exists a perfectly recognisable faunistic relationship between 
tropical Africa and tropical South America. Nevertheless this 
relationship rarely extends to genera and species. If so it is 
confined to forms like Anops, 'Mabuia, Subulina, Opistha- 
canthus, Cryptostemma and others, of which there is reason 
to believe that they are extremely ancient types. From some 
cause or other they may have retained their specific or generic 
characters throughout a series of geological ages. The faunas 
of Africa and South America as a whole are thoroughly dis- 

* Ortmann, A. E., " Tertiary Archhelenis," p. 242. 
t Pocock, E. I., " Distribution of Arachnida," p. 229. 
| Kraepelin, K, " Scorpiones and Pedipalpi," p. 240. 


tinct. All the more modern families of these continents show 
scarcely any traces of relationship to one another. Conse- 
quently we must conclude with Dr. Ortmann and several other 
authors who have definitely expressed themselves on the geo- 
logical age of the former land bridge between South America 
and Africa, that the latter ceased to exist before Tertiary 

Dr. von Ihering* did not limit his studies in South America 
to the fauna, he likewise urged, contrary to the opinion 
held by most botanists, that the floras of South America and 
Africa pointed unmistakably to the existence of his " Arch- 
helenis." His arguments were so convincing that Professor 
Englerf adopted his views after a very careful and critical 
examination of the problem from a botanical aspect. He ac- 
knowledged that for a long time he had looked upon the theory 
of a former land connection between South America and Africa 
with scepticism, chiefly on account of the great intervening 
ocean depths and the marked endemism in the flora of the two 
continents. After discussing the various means of accidental 
dispersal, and eliminating those plants which might possibly 
owe their presence in both continents to some of the known 
modes of occasional transport, he was impressed by the cir- 
cumstance that there was still a residuum of species, genera 
and families which must have had another origin. He urged 
that the distribution of the Strelitzioidae, whose fruits cannot 
be dispersed by wind, also that of some of the water plants of 
Africa, which have very near relations in South America, and 
others might be accounted for by the supposition of the former 
existence of a series of large islands in the Atlantic Ocean 
separated by narrow channels. All the same he prefers a 
complete land bridge between the two continents. 

Dr. Aj-ldt J discusses the problem in a very effectual manner 
both from a biological and a geological point of view. Like 
Dr. Ameghino, however, he is led to the conclusion that even 
in early Tertiary times South America was still directly joined 

* Ihering, H. von, " Das Neotropische Florengebiet." 
f Engler, A., " Floristische Verwandtschaft zwischen Afrika and 
Amerika," pp.50 51. 

| Arldt, Th., " Entwicklung der Kontinente," p. 451, 


to Africa, whereas I favour the theory that such a connec- 
tion could only have persisted indirectly by way of the Antilles 
and southern Europe. 

It has been urged that if a land bridge had once existed 
between Brazil and west Africa we ought to find some traces 
of sedimentary rocks on the few islands that are scattered 
about in the intermediate area, like Fernando de Noronha, St. 
Paul's, St. Helena, Ascension and Tristan da Cunha, whereas 
all these are believed to be composed of volcanic rocks. Pro- 
fessor Schwarz,* nevertheless, affirms that although Ascen- 
sion consists of volcanic rock the substratum is granitic. St. 
Paul's island, he thinks, is mainly metamorphic in structure, 
and there are signs of its having been subjected to considerable 
earth movements. Continental types of rocks occur on Tristan 
da Cunha. Consequently Professor Schwarz concludes that 
these Atlantic islands furnish some geological testimony for 
the belief in a Mesozoic Continent in the position of the 
southern Atlantic lasting until the beginning of the Tertiary 

The island of Fernando de Noronha lies about two hundred 
miles east of Cape San Koque on the coast of Brazil. Since 
its discovery in 1503 both the original fauna and flora of the 
island have been laid waste by early settlers, so that only 
vestiges of these are left. The splendid trees that once clothed 
the island are gone and with them probably the greater part 
of the fauna. A mammal, at any rate, existed on Fernando de 
Noronha when it was discovered, but it has quite vanished. 
Mr. Eidley, who visited Fernando de Noronha in 1887, noticed 
that quite a number of species of animals and plants had evi- 
dently been introduced by human agency. Many others, he 
thinks, have been transported to the island from the main- 
land by winds, although he noted that the smaller birds and 
a large proportion of the smaller insects are endemic. A few, 
he remarks, may possibly have been brought over on the feet 
of wading birds. There still remains a small group of indi- 
genous species which are difficult to account for, on the hypo- 
thesis of accidental distribution. The subterranean lizard 

* Schwarz, E. H. L., "Land-connection between Africa and ^uth 
America," pp. 8190. 


Amphisbaena ridleyi, for example, is peculiar to the island, 
its nearest relation living in the West Indies. The skink 
(Mabuia punctata) occurs on the island as well as in British 
Guiana. The snail Bulimulus ridleyi, as I have already 
pointed out (p. 176), is the only living representative of the 
Oligocene group of species which once inhabited Florida. 
There are, moreover, two species of Opeas identical with West 
Indian forms, while a Pupa seems closely allied to a Cuban 
species. The single fresh-water species (Planorbis noron- 
hensis) is endemic. The noteworthy relationship with the 
Antilles is also noticeable among the plants as well as among 
the marine forms living on the snores of the island. As might 
be expected, this is attributed by Mr. Ridley * to the action 
of sea-currents. No other theory of the origin of the fauna 
and flora of Fernando de Noronha than that of accidental 
dispersal by wind or waves ever occurred to him. Yet both 
Mabuia and Amphisbaena are genera which, as I have shown, 
can be cited with some justification as evidences of the exist- 
ence of a former land connection between South America and 
West Africa. And these, with a gecko of enormously wide 
range, constitute the only members of the reptilian fauna. 
Why should these ancient reptiles inhabit Fernando de 
Noronha and not any of the more modern groups ? This 
and other pertinent criticisms of Mr. Eidley's theory have 
also been urged by Dr. von Ihering,f who shows, especially 
from a botanical point of view, that accidental dispersal has 
not played a very important part in the origin of the island 
flora. On the contrary, he maintains that it is part of the 
ancient Archhelenis which once united South America and 

The small island of Trinidad, not to be confounded with the 
island of the same name in the West Indies, lies much further 
south, and about seven hundred miles east of the coast of 
Brazil. It is now uninhabited, though for a time it contained 
a small colony of settlers who left a few goats and pigs on the 
island, which continued the destruction of the native fauna 
and flora commenced by their late masters. The ground 

* Eidley, H. N., " Zoology of Fernando Noronha," pp. 473 502. 
t Ihering, H. von, "Fernando de Noronha," p. 6. 

L.A. C C 


moreover, is riddled with the burrows of myriads of land 
crabs, which have aided the domestic animals, to the best of 
their ability, in this work. Trinidad seems to. have been 
densely wooded formerly with tree ferns, palms, etc., whereas 
living trees are now confined to the mountain tops. Among 
the ferns collected on the island by Dr. Copeland * there was 
one (Asplenium compressum) hitherto only observed in St. 
Helena. Of the fauna nothing appears to be known. 

Ascension island lies in about the middle of the southern 
Atlantic. When Darwin visited the island during his famous 
voyage round the world, he found little there of interest. A 
principal mound in the centre looks as if i't were the jyarent of 
the lesser volcanic cones studded round the island. The sur- 
face is parched and barren, and everything presents a scene 
of utter desolation. Yet not only is Ascension island in- 
habited, the scanty pasture has even been made serviceable 
for a number of sheep, goats and cows. Ascension stands on a 
submarine bank about two hundred miles in length. It is 
probable on that account that the existing island only repre- 
sents the last remnant of a cluster of volcanic cones which 
rose high above the surrounding land. The surviving fauna 
and flora are extremely poor. Many of the species, like the 
centipede Scolopendra morsitans, are almost cosmopolitan in 
their range, a few also occur in Africa or South America. 
But no effort has as yet been made to conduct a thorough 
zoological and botanical survey of the island, and it is to be 
hoped that this will be accomplished before all traces of the 
ancient relicts that might still occur have been superseded 
by recent importations. At any rate, nothing of any impor- 
tance can be gathered from the little information we possess. f 

St. Helena, like Ascension island, is wholly volcanic in 
structure, and being, moreover, surrounded by depths up to 
17,000 feet, Dr. Wallace claims that we ought to be satisfied 
as to its being a true oceanic island and as to its owing none of 
its peculiarities to a former union with any continent or other 
distant land. But, as I have pointed out once before, the exis- 
tence of such a great depth of the ocean does not prove that it 

* Copeland, R., " Insel Trinidad," pp. 274277. 

t Gimther, A., and others, " Collection made in Ascension Island." 


always was so. Geological evidence, says Professor Suess,* 
does not prove, nor even point to, a permanence of the great 
depths, at least in the oceans of the Atlantic type. No longer 
can we, therefore, subscribe to Dr. Wallace's statement that 
St. Helena is necessarily what he calls a "true oceanic 
island." The island was densely covered with a luxuriant 
forest vegetation when it was discovered in the year 1501. 
Human occupation has almost wholly destroyed this in- 
digenous vegetation, and with it no doubt the greater part of 
the fauna. The rich soil, no longer protected by the covering 
vegetation, has been swept away by tropical rains, leaving a 
vast expanse of bare rock or sterile clay. Fortunately the 
fauna and flora of St. Helena have received greater attention, 
than those of Ascension, and although a mere fragment only of 
what originally inhabited this area, it allows us to draw some 
conclusions as to their origin. There are neither mammals, 
amphibians nor reptiles on the island. The only indigenous 
bird is a small plover (Aegialitis sanctae-helenae), closely allied 
to a species found in South Africa. Among the invertebrates 
the beetles in particular have been carefully studied by Mr. 
T. V. Wollaston.f Out of two hundred and three species 
collected on the island, he considers one hundred and twenty- 
nine to he indigenous. Only one of these is found else- 
where. These species belong to thirty-riine genera, of 
which no less than twenty-five are peculiar to St. Helena, 
most of them being weevils (Ehynchophora). Since the 
greater number of weevils are woodborers, we conclude 
from this fact alone that the island once possessed a luxuriant 
forest vegetation. Many of the beetles show no close affinity 
with any existing insects. A small number only are more 
or less remotely related to European and South African 
species. These features imply, as Dr. Wallace truly remarks, 
that the beetle fauna of the island is extremely ancient, dating 
back to at least the Miocene age. Dr. Wallace acknowledges 
that at present the marine currents flow towards St. Helena 
from the region of the Cape of Good Hope ; nevertheless he 
contends that in former geological periods the currents may 

* Suess, E., " Are Ocean Depths Permanent ? " p. 186. 
t Wollaston, T. V., ' Coleoptera Sanctae Helenae." 

c c 2 


have varied, and that drift-wood might have been carried to 
the island from different directions. Great alterations of 
wind, he thinks, were probable during the Glacial Epoch, so 
that these two most potent agencies in the accidental trans- 
port of species may have varied very much during the periods 
in which the island received its colonists from other countries. 
Of course, Dr. Wallace* firmly adheres to the belief that all 
animals and plants found on the island owe their existence 
there to some means of occasional transport. 

Thirty species of bugs (Hemiptera) have been recorded by 
Dr. White from St. Helena, of which five have certainly, and 
one probably, 'been introduced. This leaves twenty-six species 
indigenous to the island. Of the twenty-one genera of Hemip- 
tera eight are peculiar to St. Helena, but the general 
distribution of these insects was so little known at the time 
that Dr. White was unable to trace their affinities, except that 
most of them have a wide range, and several are known from 
Miocene deposits. Only one genus (Megarhaphis) has dis- 
tinctly African affinities. Dr. White f argues from the 
general aspect of the fauna and flora, from the non-existence 
of mammals and reptiles, from the large number of endemic 
species, and from the great depth of the surrounding sea, that 
St. Helena at no time could have been joined by land with 
Africa or South America. A careful consideration of all the 
known facts led him to believe that the colonists did not arrive 
all in a body, but that colonisation was spread over a con- 
siderable period. He rejects the theory of a continuous land 
surface, contending that the fauna and flora arrived from the 
north in the direction of the Cape Verd islands. Stepping- 
stones in the shape of islands, now disappeared, may have 
existed formerly, thus facilitating dispersal, whilst the 
marine currents were probably reversed. 

Mr. Pickard-CambridgeJ informs us that forty-four species 
of spiders are known from the island, some of them being also 
found in Europe, and two in. Egypt. The rest are endemic, 
but most of them show European relationship. In a former 

* Wallace, A. E., " Island Life," pp. 294303. 

t White, F. B., "Hemiptera of St. Helena," pp. 446460. 

t Pickard-Cambridge, O., " Spiders of St. Helena," p. 210. 


ncte he recorded two scorpions (Lychas maculatus and L. 
americanus). These have since been shown to be both forms 
of Isometrus maculatus, a cosmopolitan species of the tropical 
and sub-tropical zone. In Europe it occurs only in the south 
of Spain.* 

The total number of land-shells indigenous to St. Helena 
is estimated by Mr. Smith as amounting to twenty-seven 
species. With regard to their origin, he likewise rejects the 
theory of a possible land connection of the island with either 
Africa or South America, or, indeed, any other land surface. 
All the same, he points out that there is a greater resemblance 
between the molluscan fauna of St. Helena and South 
America than was recognised by previous observers. It is 
especially a species resembling the exclusively Brazilian 
genus Tomigerus that seemed to Mr. Smith to point to South 
America as the likely source of some of the indigenous species. 
He does not feel tempted to probe further into this mystery, 
for he exclaims, " How they were transmitted is a hopeless 
problem to solve, and although drift-wood, carried by oceanic 
currents, is doubtless answerable for a good deal in the way 
of distribution the subject must apparently ever remain one 
of mere speculation." Nevertheless Dr. Kobelt f regards it as 
highly probable that this molluscan fauna is the last remnant 
of that of an ancient Mesozoic continent. 

About half of the species of the snails of St. Helena are 
now extinct and occur only in a sub-fossil condition. Hyalinia 
cellaria, H, alliaria and a few others, along with Amalia 
(Milax) gagates are supposed to have been introduced with 
plants. It is quite possible that some of them were brought 
to St. Helena in that manner, still it is very remarkable that 
the only slug thus conveyed should be one of our raxest 
British species, and the one which has the widest range of all 
slugs. It occurs in California, New Zealand, and other remote 
localities, while none of our common Arions, Agriolmiaces 
or Limaces have found their way to the island. I am fully 
convinced that Amalia gagates is one of the most ancient of 
our European terrestrial species, as I explained in a previous 

* Pickard-Cambridge, O., " Spiders and Scorpions of St. Helena," p. 544, 
f Kobelt, W., " Zoogeographische Stellung von St. Helena," p. 201. 


chapter (p. 213), and that it might well form part of the 
indigenous fauna of the island. There are twelve species, 
some of which greatly resemble Polynesian forms, as Mr. 
Smith pointed out, and, indeed, Dr. Pilsbry has placed them 
provisionally with Endodonta, a genus otherwise confined to 
Australia and Polynesia. There are three species of Succinea, 
a genus which seems to be particularly unfit for transport 
on drift-wood, while the most striking members of the snail- 
fauna are a species of Bulimus, seven species of Bulimulus, 
and one provisionally placed by Mr. Smith* in the South 
American genus Tomigerus. 

More recently, Dr. Pilsbry demonstrated that these 
apparently bulimoid snails are really members of the family 
Achatinidae, and that they all belong to the genus Chilonopsis. 
Since all the nearest relations of Chilonopsis are African 
genera, the affinity of this group is with Africa rather than 
with South America. But the great antiquity of Chilonopsis 
is indicated by the fact that Trichodina, to which it is closely 
allied, has a most remarkably discontinuous range, occurring 
only on some islands off the coast of West Africa and on the 
Comoro islands in East Africa. Chilonopsis, moreover, is 
more primitive than the genera of the large African Achati- 
nidae, and Dr. Pilsbryf expresses the opinion that it, as well 
as the related genera, originated from a common ancestor 
which lived in Mesoaoic times, when Africa and South 
America were united by land. None of the species have any 
affinity with South African ones, as was supposed to be the 
case among the Hemiptera. 

Although Dr. White's Megarhaphis has now been shown 
to belong to Macrorhaphis, a genus confined to Africa, 
two species have since been taken in west Africa. 
The supposed slight affinity of the St. Helena bugs to 
those of the Cape has therefore entirely disappeared. As 
regards the pronounced European element in the St. Helena 
insect fauna, it is probably to a large extent of American 
origin. The beetle Zophobas morio lives, outside St. Helena, 
only in west Africa, Central and South America and 

* Smith, E. A., " Land-shells of St. Helena," pp. 259269. 
t Pilsbry, H., " Manual of Conchology (Pulmonata) " (2), Vol. XVII., 
pp. 171173. 


the Antilles, while all other members of the genus are 
American. The only relation of the bug Metacanthus con- 
color of St. Helena occurs in Europe, whereas the three most 
closely tallied genera inhabit America. The genus Nysius 
has a world-wide range, and is known from American Miocene 
deposits. Cardiastethus occurs in Europe as well as in St. 
Helena, but the genus is mostly American, and from there 
ranges across the Pacific to New Zealand. Nabis capsiformis 
occurs in southern Europe, and also all over Africa and 
America, while the genus has a world-wide range. The allied 
Vernonia of St. Helena has near relatives in the West Indies 
and the western Pacific region. Salda is a distinctly northern 
genus, although a few species reappear far southward of 
the others in Chile, New Zealand and St. Helena. That such 
a range implies great generic antiquity is evident, and, 
indeed, Salda is known in several species from the lower 

Of the Curculionidae which are so largely represented in the 
fauna of St. Helena, and which Dr. Wallace thought might be 
of Miocene age, many recent genera are now known from the 
Lower Oligocene and they are probably much older even than: 
that, for, according to Dr. Handlirsch, over two hundred 
species are now known from Oligocene deposits. The genus 
Homalota which Dr. Wallace fancied was exclusively Euro- 
pean, is represented in the Oligocene of North Am'erica. 
Philonthus, Xantholinus and Oxytelus occur in the American 
Miocene and European Lower Oligocene. The supposed 
European genera of insects inhabiting St. Helena are thus 
mostly groups of very wide range, or such of which 
we actually possess palaeontological evidence of their having 
existed since early Tertiary times. 

However ancient the insect fauna may be, remarks Dr. 
Wallace, the flora must be more ancient still. Of the fifty 
truly indigenous flowering plants, about forty are peculiar 
to St. Helena, and of twenty-six ferns, about ten. The re- 
lationship of this flora is mainly African, according to Sir 
Joseph Hooker, whereas Mr. Bentham maintains that the 
Compositae have their affinities for the most part with South 
America. Sixteen species of ferns are common to St. Helena 


and other countries. Dr. Wallace,* however, acknowledges 
that specific identity among ferns does not necessarily imply 
a recent origin. According to Professor Engler f the five 
endemic genera are somewhat related to American ones, 
whereas some species of Pelargonium, Mesembryanthemum, 
Wahlenbergia and Oteospermum must have originated in 

Dr. von Ihering J discusses all the various modes of dis- 
persal, by birds, winds and ocean currents, in connection 
with the origin of the flora of St. Helena, and rejects them 
as quite inefficient. In his opinion, only a land connection 
with Africa and South America could have produced the flora. 

Although St. Helena is surrounded by great ocean depths, 
and possesses neither mammals, reptiles nor amphibians, we 
cannot assume, as Dr. Wallace has done, that the whole fauna 
and flora owe their origin to accidental dispersal by birds, 
winds and ocean currents. If such were the case, there is no 
reason why these modes of transport should not have been 
equally effective in Pliocene and Pleistocene times than in the 
older geological periods, and why there should not be many 
species differing but slightly from American or African ones. 
Everything, on the other hand, points to the animals and 
plants being almost all of extreme antiquity. We have either to 
suppose, therefore, that during some very remote period the 
modes of transport alluded to were far more efficient than they 
are now, or that St. Helena was connected by land with South 
America and Africa. For many reasons, already fully dis- 
cussed in this chapter, the latter hypothesis appears to me 
the only one that is at all probable. If we assume that there 
was such a land bridge, it is quite possible that the actual 
volcanic area now forming St. Helena was not in existence, 
but originated after portions of the land connection had 
already subsided. A mere remnant of the fauna and flora 
would eventually have taken refuge on the rock of St. Helena, 
whilst the surrounding country disappeared beneath the 
floods. Whether in early Tertiary, or even in later Tertiary 
times, there was a short and independent land bridge between 

* Wallace, A. K., " Island Life," pp. 305307. 

t Engler, A., " Entwicklungsgeschichte der Florengebiete," II., p. 179. 

I Ihering, H. von, " Das neotropische Florengebiet," p. 50. 


St. Helena, Ascension island, the Cape Verd islands, the 
Canaries and Madeira, is a matter which must be left to future 
observations. Certain faunistic features appear to be in 
favour of such a theory. 

Tristan da Cunha is almost as large as St. Helena. It 
lies considerably further south, being bleak and inhospitable 
in the extreme, owing to its proximity to the northern boun- 
dary of the antarctic ice-drift. The flora is quite different 
from that of St. Helena, as it possesses antarctic affinities. 
The genera Nertera, Uncinia and Acaena inhabiting antarctic 
America, New Zealand or Australia occur here also. It has 
Lomaria alpina, in common with South America, Kerguelen 
island, New Zealand and Australia, while Spartina arundi- 
nacia occurs in Tristan da Cunha and St. Paul, and is related 
to a South American species.* The most remarkable member 
of the fauna is the genus Tristania which, with its two species, 
is confined to the island. It is related to th snail Opeas which 
inhabits both Africa and South America. 

Almost three hundred miles further south there is still an- 
other volcanic island, of which scarcely anything was known 
until it was visited by Dr. Bruce, the leader of the Scottish 
National Antarctic Expedition. First called " Diego 
Alvarez " according to Dr. Brown,f its name was later on 
changed to " Gough island." The island differs from some 
of the other islands referred to, in being densely covered with 
vegetation, which is similar in character to that of Tristan 
da Cunha. While showing the same general affinities, Diego 
Alvarez seems to have a stronger American element than the 
islands forming the Tristan da Cunha group. The land birds 
comprising two species of bunting (Nesospiza) and a galli- 
nule (Porphyriornis) are peculiar to the island, but related 
to species living on Tristan da Cunha. 

To briefly review the results of our studies so far, it appears 
highly probable that Brazil was entirely separated from 
western South America in Mesozoic and part of Tertiary 
times, but was connected with the highland of Guiana across 
the mouth of what is now the Amazon river, which flowed 

* Engler, A., " Entwicklungsgeschichte der Florengebiete," II., p. 159. 
t Brown, E. N. E., " Diego Alvarez," pp. 910. 


westward. During part of Secondary times eastern Brazil 
was most likely united by land with West Africa. During the 
Cretaceous age already, the southern Atlantic rapidly advanced 
northward, invading eastern Brazil, so that the land bridge 
then joined only northern South America and a small tract 
of West Africa. In Eocene times the southern trans -Atlantic 
land connection had ceased to exist, yet the South American 
continent must have extended still far eastward of its present 
coast-line. Eastern Brazil continued to receive new additions 
to its fauna from 1 the northern States of South America while 
its area extended gradually westward. When Brazil finally 
became joined to Argentina a faunistic interchange took 
place, although Brazil and Chile remain to the present day 
two fundamentally distinct countries. 



ALTHOUGH the natural southern boundaries of Brazil appear 
to be the great river Plate and the Paraguay river (Fig. 18), 
the two smaller republics of Uruguay and Paraguay are inter- 
posed between Brazil and this river system, while Argentina 
has also claimed a large slice of territory on the northern side 
of it. As we proceed southward from eastern Brazil the moun- 
tains gradually pass into hills which again are replaced by 
gently undulating plains. Magnificent pasture lands abound 
in Uruguay, accompanied by a mild and healthy climate, so 
that the country is well adapted for the staple industry, 
which consists in raising horses and cattle. The spurs of 
the eastern highlands of Brazil are still recognisable in the 
northern provinces of Argentina, where they are covered by 
extensive forests of algaroba, cedar and other trees, thus 
greatly adding to the natural resources of that wealthy 
country. The surface of the remainder of the great Argen- 
tine Republic is largely occupied by somewhat monotonous 
tracts of low and nearly level land, here and there relieved by 
small groups of mountain ranges which seem mostly to have 
some connection with the foothills of the Andean mountain 
chain. The whole of western Argentina again is very moun- 
tainous and well wooded. The highest summit of the Andes, 
Aconcagua (23,080 feet), actually lies in Argentine territory. 
Beyond the crest of this immense mountain chain, occupying 
an elongated narrow strip of country along the shores of the 
Pacific Ocean, lies the Republic of Chile. The upheaval of 
the Cordillera of the Andes, separating the two states, must 
have been the result of movements of the earth's crust which 
occurred long after the formation of the rocks composing the 

With these alterations in the topography and climate as 


we proceed southward, the fauna and flora gradually undergo 
a complete change. On leaving the forest, monkeys, sloths 
and other animals, bound to an arboreal life, are replaced by 
new types, unknown in Brazil, such as the Patagonian hare 
and vizcacha. A very characteristic faunistic feature of the 
Argentine " pampas," as the humid grassy country is called, 
is the presence there of the American representative of the 
ostrich. This rhea is quite absent from the vast forests of 
Brazil. In the northern parts of that country, however, where 
we again meet with grass lands, this flightless bird reappears 
as a memento of the times when north and south were joined 
by continuous prairies. We thus have some noteworthy 
instances of discontinuous distribution suggesting a former 
period when the drying up of vast lakes, perhaps, may have 
given rise to grass lands, which enabled certain southern types 
to push northward. Other Argentine animals, such as the 
coypu (Myocastor coypus), one of the largest of the rodent 
tribe, have apparently invaded Brazil in more recent times, 
for they have only entered the southern and western pro- 
vinces. To those unacquainted with the past history of South 
America it must come as a matter of surprise that the only 
ungulate which is noticed in these vast pampas of Argen- 
tina is a deer, closely related to the true North American deer 
(Odocoileus virginianus) and its allies. I have expressed my 
belief above (p. Ill) that the true American deer owe their 
origin to one or more ancestors which passed into South 
America directly from Europe, and it is important to note 
that the centre of dispersal of the various groups in South 
America lies in the western states. Only a couple of species 
are peculiar to Brazil. All the others are more or less 
confined to the western states. Even the distribution of the 
South American wolves led Dr. von Ihering * to the similar 
conclusion that these animals entered the continent from the 

Mr. W. H. Hudson f gives us a vivid idea of the life in the 
pampas of Argentina, the striking poverty of its fauna as 
compared with Brazil, and all the more characteristic features 

* Ihering, H. von, " Siidamerikamsche Raubtiere," p. 162. 
t-Hudson, W. H., " Naturalist in La Plata." 


of that remarkably fertile country, which has lately become 
a centre of attraction, not only from a point of view of 
agriculture, but as a treasure-house of palaeontological 

Argentina must have undergone very notable physio- 
graphical and climatic changes within the Tertiary Era. As 
Dr. White * has pointed out, the whole of the coast-line from 
Eio de Janeiro southward appears to present evidences of 
submergence. Eivers, bays and islands exhibit an aspect 
of drowning similar to that shown by the rivers and bays of 
Nova Scotia, Cape Breton and Newfoundland; and the amount 
of coastal depression seems to increase southward, judging 
by borings for harbour works at Bio de Janeiro and at 

The presence of terrestrial and fresh -water deposits to a 
depth of nearly a thousand feet below the city of Bu,enos 
Aires implies, as Dr. Ameghino f remarks, that here also the 
continent extended formerly much further eastward. The 
same author contends, in fact, that the whole of the southern 
Atlantic is of Tertiary age. I am not prepared to concur in 
this view, but it can scarcely be doubted that Argentina had 
in later Tertiary times at any rate a much greater area than 
at present. 

It is now more than twenty years since Dr. Ameghino first 
made known to the world that an extraordinary wealth of 
animal life once tenanted the vast plains of Argentina. Not 
only in early Tertiary deposits; even in late Mesozoic beds 
were found the bones of mammals belonging to many different 
groups. The origin of life, at least of the higher animals, was 
always looked for in the north. Dr. Ameghino's discoveries, 
which certainly rank among the most noteworthy that have 
ever been made in palaeontology, riveted attention for the 
first time to the southern hemisphere. No wonder that Dr. 
Ameghino J in his enthusiasm pronounced Argentina to be 
the original home of all the mammals of the world. 

* White, I. C., "Relatorio final de estudos das minas," p. 3. 
f Ameghino, PL, " Formations sedimentaires, p. 29. 
| Ameghino, PL, "South America as the Source of Mammalia," 
p. 260. 


Naturally his views gave rise to a considerable amount of 
discussion jand criticism among geologists and palaeontolo- 
gists ; and many visited the scenes of these startling dis- 
coveries. Almost all of them felt convinced that Dr. 
Ameghino had wrongly interpreted the age of the deposits 
alluded to. They thought that the fossiliferous beds described 
must be much more recent than Dr. Ameghino supposed, and 
that at any rate South America was certainly not the source of 
all the mammalia. One of the richest of the deposits exposed 
in different parts of Patagonia, especially in the region of the 
Santa Cruz river, and hence spoken of as the " Santa Cruz 
beds," is considered by Dr. Ameghino as of Upper Eocene 
and Lower Oligocene age. Almost all other authorities believe 
them to belong to the Miocene. The latter view seems to be 
strengthened by the results of the study of the marine shells 
contained in beds of similar age which were pronounced by 
Dr. Ortmann * to be of Miocene Age. As Professor Scott f 
expresses the opinion that Dr. Ortmann has clearly demon- 
strated the Miocene age of the Santa Cruz beds, I should like 
to direct attention to a more recent work dealing with this 
problem. Its author, Dr. von Ihering,J has made a special 
study for many years past of the marine mollusks of South 
America. Hence his opinion is of particular value. After 
a long and careful study of a larger series of fossil mollusks 
than was available before, he came to the conclusion that 
the marine Cretaceous deposits of Patagonia pass very gradu- 
ally into those of the next formation, which is the one 
described by Dr. Ortmann, the latter scarcely containing 
five per cent, of living species. He is inclined on that 
account to agree with Professor Ameghino as to the Eocene 
and Oligocene age of these marine beds, and thus indirectly 
with his general scheme of correlation. Dr. von Ihering's 
arguments appear to me more convincing than those of Dr. 
Ameghino's opponents, and I shall therefore adhere to the 
latter's nomenclature of the Argentina deposits. Not only 
has Dr. von Ihering adopted Dr. Ameghino's views, they 

* Ortmann, A. E., "Princeton University Expedition," Vol. IV., 
p. 317. 

t Scott, W. B., "Mammalian fauna of Santa Cruz," p. 241. 
| Ihering, H. von, "Mollusques fossiles de PArgentine," p. 95. 


have partially been accepted also by Dr. Koth,* who has 
personally studied the problems on the spot. 

As regards the mammalian remains contained in the Santa 
Cruz terrestrial beds, Professor Scott f was greatly struck 
by the strangeness of the assemblage. Not a single genus 
occurs in any part of the northern hemisphere. Some of the 
orders even of mammals are distinct from those of the 
northern faunas. Thus the beds have yielded no carnivores, 
no modern groups of ungulates nor elephants, while the 
rodents all belong to the section Hystricomorpha. The place 
of the carnivores was taken by carnivorous marsupials, some- 
what resembling the Tasmanian wolf (Thylacinus). Numbers 
of small plant-eating marsupials, of which Caenolestes 
(see p. 350) is an interesting survival, likewise occur. One of 
the largest, most varied and most characteristic elements of 
the Santa Cruz fauna are the edentates. They are repre- 
sented by the Dasypoda or armadillos, the greater part of 
whose skin is strongly ossified, the scutes forming a great 
shield over the body, and by the Glyptodontia and the Gravi- 
grada. The last two groups are now extinct. The glyptodonts 
resembled armadillos, except in so far as the bony scutes were 
joined into a solid mass like the shield of tortoises, while the 
Gravigrada or ground sloths were extraordinarily varied and 
numerous. Only a single genus (Necrolestes) of Insectivora 
has been obtained in the Santa Cruz beds, as already alluded 
to (p. 246). At present this order is quite unknown in 
South America. The ungulates belong to the extinct groups 
Toxodontia, Astrapotheria and Litopterna. The toxodonts 
were represented by the genus Nesodon which somewhat re- 
sembled a rhinoceros in shape and had similar teeth. Of the 
second group very little is as yet known, while the Litopterna 
are the most remarkable of all the hoofed animals. Without 
being in any way related to the horse-tribe, certain genera 
have paralleled the structure of the horse-foot in a most 
wonderful way, giving the latter a striking and deceptive re- 
semblance to that of the ancient Hipparion. The animals 
afford, indeed, as Professor Scott remarks, one of the most 

* Roth, S., " Sedimentablagerungen in Patagonien." 

t Scott, W. B., " Mammalian fauna of Santa Cruz," pp. 242247. 


remarkable and instructive examples of convergent develop- 
ment among mammals. Finally the Santa Cruz beds contain 
the remains of monkeys of South American type. The earliest 
traces of mammalian remains in Patagonia were supposed to 
have been found together with the bones of dinosaurs. Since 
these reptiles are characteristic of the Mesozoic Era, the state- 
ment that they were contemporaneous with rather advanced 
types was at first treated with little credence, especially as 
the actual deposits were not examined by Dr. Ameghino * 
but by his brother. Dr. Both, however, has since re-examined 
the localities in question and has met with mammalian re- 
mains partly mixed with those of dinosaurs, and partly resting 
actually below the latter, so that there is scarcely any doubt 
as to the correctness of the original observation. The 
mammals belong to peculiar mastodon-like ungulates, having 
been placed in the order Pyrotheria which is now extinct and 
quite confined to Patagonia. Others, such as Notostylops, be- 
long to the Toxodontia above alluded to. 

I think there is a general agreement now that, at any rate 
at the dawn of the Tertiary Era, a number of rather 
specialized groups of mammals lived in Patagonia, and, 
although some of them became extinct, others continued 
to inhabit the country until recent geological times. A few 
of the largest edentates, like the giant ground sloth Mega- 
therium, which was about the size of an elephant, and is 
supposed to have dragged down trees in order to feed on the 
leaves, rather than climb up like its modern diminutive rela- 
tion the tree-sloth, still roamed about the country in 
Pleistocene times. f A few years ago the dried skin, with hair 
still attached to it, of a huge creature was discovered in a cave 
in southern Patagonia, near the boundary between Argentina 
and Chile. It proved to belong to the ground sloth Neomy- 
lodon, now known as Grypotherium listai.J Later on, the 
bones of the animal were disinterred, along with those of an 
extinct horse and a large carnivore. Traces of a fire were 
also noticed and an enclosure with cut hay. From these evi- 

* Ameghino, Fl., "Formations sedimentaires," p. 80. 

t Lankester, Bay, " Extinct Animals," p. 172. 

t Moreno, F. P., and A. Smith Woodward, "Neomylodon listai." 


dences of man's presence, along with the remains of the 
ground sloth, it was concluded that the cavern was an old 
corral in which the ground sloths had been kept and tended 
by some primitive human race.* 

The survival of many of such large creatures until recent 
geological times implies that, as in the northern hemisphere, 
there must have been ample food available in Argentina 
during the Pliocene and Pleistocene Periods to nourish these 
mammals, the climatic conditions having since become more 
varied and unfavourable. Compared with the Santa Cruz 
edentates, the less ancient groups were mostly larger. 

In spite of the fact that the fossil mammalian fauna of 
Argentina presents so many features pointing to long isola- 
tion, the relationship of certain forms to those found in far 
distant regions is of extreme interest and importance from a 
zoogeographical point of view, as elucidating the geological 
history of the South American continent. The diversity in 
shape and character between the Santa Cruz armadillos 
(Dasypoda), for example, is very notable, according to Pro- 
fessor Scott,f no less than three families and seven genera 
having been described so far. And yet a genus of armadillo 
(Metacheiromys) makes its appearance in the Middle Eocene 
beds of western North America. Dr. Wortman's theory that 
the edentates were of North American origin, having sub- 
sequently spread to South America, has not been adopted by 
any later authors. Dr. Wortman assumed that their earliest 
appearance in South America did not antedate the Santa Cruz 
epoch, whereas Professor Scott points out that they also 
occur in the oldest known Tertiary and possibly even pre- 
Tertiary deposits of Patagonia, and that there is every 
appearance of their having been indigenous in that region. 
It is, in fact, generally assumed now that South America was 
the original home of the edentates. If it is correct, as I have 
endeavoured to show, that Central America has only come into 
existence in comparatively recent geological times, and that 
the whole continent of South America in the dawn of the 
Tertiary Era consisted of several distinct masses, Patagonia 

* Woodward, A. Smith, " Grypotherium listai," p. 64. 
t Scott, W. B., " Princeton Expedition to Patagonia," Vol. V., p 7. 
L.A. D D 


at that time must have had a direct land connection with 
western North America almost independently of the rest of 
South America (Fig. 14). At a still earlier stage there was 
even a more marked affinity between Patagonia and south- 
western North America ; and from this Professor Osborn con- 
cluded that the northern and southern continents were con- 
nected by land. But the points of resemblance are not alone 
with Patagonia and south-western North America. Professor 
Gaudry * expressed his astonishment at the striking faunistic 
relationship between the Patagonian Notostylops fauna on the 
one hand, and the faunas of the Torrejon in New Mexico and 
Cerney in France on the other. That a land bridge, discon- 
nected at certain intervals, extended between western North 
America and southern Europe I have urged again and again 
in the preceding chapters ; and it should be borne in mind 
how, even in these remote times, special facilities existed for 
the passage of species from Europe to the extreme south of 
South America, which no doubt were taken advantage of by 
several groups then inhabiting the Old World. 

Until recently it was thought that North and South America 
could have had no land connection subsequently to these early 
events until the end of the Miocene or the beginning of the 
Pliocene Periods. Professor Osborn, f however, has shown 
that there is now evidence for the existence of true edentates 
of the Megalonyx type in the Mascall beds of Oregon, which 
are of Middle Miocene age. During the Miocene Period 
Central America in its present shape had not yet come into 
existence. Hence we may assume that even in Miocene times 
there was a land connection between western North America 
and some portion of South America by means of a route 
which, as I argued, lay to the west of that continent. 

The rodents of the Santa Cruz fauna, as previously men- 
tioned, all belong to the section Hystricomorpha. They are 
very closely allied, according to Professor Scott, to recent 
South American genera. Yet all are extinct and many of 
them have left no successors. Nevertheless, though the Santa 
Cruz rodents are more primitive, the skull structure is nearly 

* Gaudry, A., " Fossiles de Patagonie," p. 105. 
t Osborn, H. F., " The Age of Mammals," p. 289, 


the same as in the recent ones. There is reason to believe, 
says Professor Scott,* that several of the genera represent 
the direct ancestors of existing forms. Viscacha was probably 
derived from Prolagostomus, Dolichotis from Schistomys, 
Erethizon and Coendu from Steiromys. As regards the latter, 
it is interesting to note that it resembles the recent Erethizon 
more than it does the modern South American genera 
of the same family. It seems to me probable, therefore, 
as I have already suggested (pp. 69 71), that the South 
American and North American tree porcupines originated 
from two distinct branches of ancestral Steiromys-like rodents, 
thus favouring the view I advocated of a direct land connec- 
tion between Patagonia and south-western North America. 
That the genus Erethizon, to which all the North American 
porcupines belong, should not be known from pre-Pleistocene 
deposits presents, no doubt, a difficulty to the acceptance of 
this theory, but its ancestors may have remained on the last 
remnants of the land which once existed westward of North 
and South America until compelled to leave that land in Plio- 
cene times, when it finally subsided. 

The Santa Cruz fauna likewise reveals an affinity with the 
fauna of Australia and Tasmania. The Patagonian marsupials 
are referable to three families, remnants of which survive in 
widely separated parts of the world. The Thylacinidae, now 
confined to Tasmania, where the Tasmanian wolf represents 
the family, formerly inhabited both Patagonia and Australia. 
As we might expect, the Santa Cruz thylacines are of a more 
primitive type than the Tasmanian wolf, but Professor Sinclair 
expects that the common ancestor of these two will probably 
be found among the marsupials occurring in still older Pata- 
gonian deposits. The opossums (Didelphyidae), among which 
Microbiotherium is the best known, are met with in several 
genera in the Santa Cruz beds. The Cretaceous Proteo- 
didelphys suggests, as I have already mentioned, that South 
America must be looked upon as the original home of the 
family whence some members passed into North America and 
Europe. I have briefly alluded to the occurrence in Ecuador 


* Scott, W. B., "Princeton Expedition," Vol. V., pp. 384386 and 
p. 413. 

D D 2 


of the marsupial Caenolestes and to its being the only living 
member of the extinct family Epanorthidae. In the Santa 
Cruz beds this and another family of herbivorous marsupials 
are represented by several genera. Professor Sinclair * in- 
clines to the belief that the South American Epanorthidae and 
the Australian phalangers have had common ancestors. From 
all these facts he concludes that a land connection between 
Patagonia and the Australian region must have existed not 
later than the close of the Cretaceous Period or the beginning 
of the Tertiary Era, and that the interchange of marsupials 
may have been effected between the two continents at this 

Of all the similarities between the South American and 
Australian extinct faunas none perhaps is more striking, re- 
marks Dr. Smith Woodward,f than the essential identity of 
the extinct Miolania in the two regions. He thinks there can 
be no doubt that this creature was a truly terrestrial or marsh 
chelonian. The theory of a former land connection between 
South America and Australia see mis, therefore, to receive 
weighty support from the new discovery of this curious 

It has also been urged that the Patagonian fossil fauna 
indicates traces of an affinity with South Africa. Professor 
Scott, J with some hesitation, expresses the opinion that the 
resemblance of the solitary Patagonian insectivore Necrolestes 
to the golden mole Chrysochloris of South Africa may be due 
to genetic relationship. It is significant, in view of the early 
Tertiary faunistic relationship of Patagonia and western 
North America, to note that certain creatures resembling the 
golden moles made their appearance in North America in 
Oligocene and Lower Miocene times. I have had occasion to 
allude to Xenotherium and Arctoryctes already (p. 246). The 
affinities between these various forms are not very marked. At 
any rate, it would be unsafe to draw any conclusions as to the 
existence of a former direct land bridge between Patagonia 
and South Africa from the presence of Necrolestes in the 
Santa Cruz deposits. 

* Sinclair, W. J., " Marsupial Fauna of Santa Cruz," pp. 7681. 
t Woodward, A. Smith,, "Extinct Eeptiles from Patagonia," p. 182. 
J Scott, W. B., " Princeton Expedition," Vol. V., p. 379. 


Professor Tullberg's* reference to a pre-Tertiary land 
bridge between South America and south-west Africa is not 
very definite. From his remarks it is not clear that he is in 
favour of the existence of an independent connection far to 
the south of the one I discussed in the last chapter. 

These are among the more important results derived from 
the study of the fossil fauna of Argentina as to the affinities 
presented by this part of South America to the more distant 
parts of the world. We have still to consider the faunistic 
kinship between Argentina and the neighbouring states of 
South America. Dr. von Ihering f pointed out long ago that, 
whereas America as a whole is the richest part of the world 
in the variety of genera and species of fresh-water mussels, 
Chile and Peru belong to the poorest districts, since, at any 
rate west of the Andes, only the genus Unio occurs. Similarly 
Ampullaria and many other typically American fresh-water 
genera are absent from Chile. On the other hand, the Unios of 
Chile are most of them nearly related to those of the La Plata 
region. Dr. von Ihering J concludes from these very peculiar 
zoographical features that, while the whole of southern South 
America ( Archiplata) formed a united land-mass in Secondary 
times, the elevation of the Andes afterwards prevented a 
faunistic interchange between the two districts. The fresh- 
water Crustacea tell us a very similar story. The fresh-water 
crayfish Parastacus is met with in eight species in South 
America. None of them occur north of southern Brazil, 
although several ,of the Chilean species are closely related to 
Brazilian ones. The fresh-water crab Aeglea laevis, no doubt 
an exceedingly ancient form and the only representative of the 
family Aegleidae, lives in identically the same species on both 
sides of the Andes. The absence of almost all the leading 
genera of Brazilian fishes from Chile and Patagonia, || empha- 
sises the noteworthy distinctness in the fresh-water fauna of 
the two regions. On the other hand, Patagonia and Chile 
present traces of a relationship, as I intend to show later on, 

* Tullberg, Tycho, " System der Nagetiere," p. 495. 

t Ihering, H. von, *' Verbreitung der Ampullarien," p. 106. 

t Ihering, H. von, " Archhelenis and Archinotis," p. 57. 

Ortmann, A. E., "Distribution of Decapods," pp. 292296. 

|| Eigenmann, C., "Freshwater Fishes of Patagonia," pp. 227229. 


with western North America, with southern Africa and with 
New Zealand and Australia. 

As we ascend the Andes from northern Argentina, we meet 
with a familiar animal which we generally associate with Peru 
and its inhabitants, namely, the llama. The llama is only a 
domestic animal. But its wild ancestor the huanaco (Lama 
huanachus) still exists plentifully in the Andes of Peru and 
Chile, as well as in the plains of Patagonia. Its western dis- 
tribution, like that of the bear, some of the deer and many 
vertebrates and invertebrates, is thus very marked and de- 
mands an explanation. For the benefit of those who may 
be unacquainted with the pedigree of this creature I may men- 
tion that the huanaco and its domesticated relations the llama 
and alpaca, are essentially camels in structure. There are 
certain differences in the teeth, the huanaco being with- 
out a hump and smaller than the camel, while the head is 
more like that of a sheep. But the shape and structure of the 
skull and bones and the general anatomy are very similar 
in the huanaco and camel, implying that they have descended 
from the same ancestor. The occurrence of the camels in Asia 
and of the llamas in South America is a case of widely dis- 
continuous distribution of two related genera and thus points 
to an ancient origin. The geological history of the family is 
of the greatest interest. The first member of the camel tribe 
makes its appearance in Argentina in a deposit known as the 
Lower Pampean, which Dr. Ameghino considers of Pliocene 
age, though some authorities think it is more recent. Here 
occurs Palaeolama, an extinct ancestor of the huanaco. In 
the upper portions of the same deposit the remains of quite 
a number of llama-like creatures have been discovered such 
as Protauchenia, Hemiauchenia, Stilauchenia and Mesolama, 
as well as the huanaco itself. That is about all we know of the 
past history of these camel -like animals of South America. 
The fact that no members of the family are known from earlier 
deposits than Pliocene, and that the extinct genera are nearly 
related to the living one, suggests that the ancestors of the 
huanaco were not originally native to South America but in- 
vaded the continent from elsewhere. Africa and Europe no 
longer were connected with South America. North America 
is thus the only possible source whence the ancestral huanacos 


or llamas could have originated. And that is precisely what 
happened, as already mentioned (p. 86). Towards the 
latter end of the Eocene Period there appeared four genera 
in western North America, all of which exhibit decidedly 
camel -like characters, and Professor Osborn * believes that 
one of these, the diminutive Protylopus, may possibly repre- 
sent the most remote ancestor of the grand American phylum 
of camels. Other genera occur in Oligocene and Miocene beds 
of North America. During the latter period camels had ap- 
parently spread in great herds over the continent. It is 
thus probable that they then extended their range to other 
parts of the world. Some of them, like Pliauchenia, had 
assumed llama-like characters, and as the .western Pacific land- 
belt was then in communication with California the ances- 
tors of the South American llamas were able to pass south- 
ward. According to Professor Osborn llamas survived in 
North America until Pleistocene times. They were then 
becoming extinct in the eastern States, lingering on in Cali- 
fornia where the great sabre -tooth tiger no doubt stalked tihem. 
In the Siwalik beds of northern India camels first appear in 
the Pliocene, as in South America, and it is generally assumed 
that the ancestors of the Old World camels crossed over to 
Asia by the Bering Strait land bridge. But as it was probably 
in Miocene times that these early camels wandered westward 
from North America, the Bering Strait land bridge had not 
yet come into existence. They must have utilised the more 
southern bridge, which I think replaced it in earlier Tertiary 
times (Fig. 16). 

Kecently we have received clear proof of a migration of 
mammals from Asia to North America, which I think must 
have taken place across the same Pacific land bridge. One 
of the most remarkable discoveries among the many note- 
worthy ones in North American palaeontology is that by Dr. 
Matthew and Mr. Cookf of Asiatic antelope remains in 
western Nebraska. The American invasion by true Asiatic 
antelopes was brilliantly and amply confirmed, according to 
Professor Osborn, by Mr. Merriam's discovery in Nevada, 

* Osborn, H.' F., " Age of Mammals," p. 170. 

t Matthew, W. D., and H. J. Cook, " Pliocene Fauna from Nebraska." 


Rather earlier arrived in North America the mastodons, 
and no doubt by the same Pacific land connection. They in- 
vaded the continent from Asia and remained until Pleistocene 
times, being thus co-existent with the early races of man in 
America. That they travelled beyond North America, pene- 
trating far into the southern continent, has been clearly de- 
monstrated. A large number of apparently distinct forms have 
been described from South America, most of them, as we 
should expect, from the west coast, though it is very doubtful 
if more than a few species lived there. Dr. Nordenskiold 
thinks that Mastodon chilensis, M. bolivianus and M. andium 
all belong to one species. In North America, as in Europe, 
both the trilophodont and tetralophodont types of mastodon 
have been discovered, that is to say, animals which possessed 
intermediate molar teeth with either three or four ridges. 
In the South American Mastodon andium, at any rate, the 
molars are in the transition stage between the trilophodont 
and tetralophodont types. These extinct elephants made their 
first appearance in Argentina in the Lower Pampean deposits. 
Since they had thus penetrated so far south in early Pliocene 
times they must have left North America before Central 
America had come into existence. They could only have 
wandered southward along the southern continuation of the 
Pacific land bridge and have entered the South American 
continent from the west during the time the bridge was joined 
to the latter. Dr. Nordenskiold * argues that, having no more 
efficient competitors in South America, the mastodon probably 
lived longer there than in North America. 

Hitherto the theory has been quite generally accepted that 
the invasions of animals from North America to the southern 
continent, during later Tertiary and Pleistocene times, took 
place across the narrow isthmus of Central America as soon 
as its formation was completed. Some doubts have latterly 
been raised as to the nature of the barrier which prevented 
the interchange of the two faunas in the earlier portion of 
the Tertiary Era. Yet until the discovery of gravigrade sloth 
remains in the Miocene Mascall beds of Oregon was announced 
by Professor Sinclair, the assumption seemed perfectly justi- 

* Nordenskiold E., " Saugetier-fossilien des Tarija Tals," pp. 1425. 


fied, from all available data, that North and South America 
became joined at the commencement of the Pliocene Period. 
It was supposed that the advance of northern types, such as 
the mastodon and hosts of others, towards the southern con- 
tinent must have coincided with the opening up of the new 
land of Central America. The occurrence of edentate remains 
in North American Miocene deposits upsets this theory, 
because, if mammals were able to reach North America from 
the south during the Miocene Period, northern species must 
have had equal facilities for invading South America at this 
time. If there is geological evidence that Central America was 
not available as a safe land bridge between North and South 
America in Miocene times, some other land connection 
must have united the two continents. When I advanced the 
theory of the former existence of a Pacific land bridge between 
North and South America, westward of Central America,* I 
was unaware of Professor Sinclair's interesting discovery 
among the Mas call beds of Oregon. My theory was largely 
founded on zoogeographical data on curious instances of 
discontinuous distribution of ancient groups in North and 
South America. Professor Osborn f regards my theory as 
inconsistent with the fact that the Pacific land bridge should 
only have been used by these gravigrade sloths. If such a 
land connection really existed why was it not more exten- 
sively used ? I think it was used by other animals, such as the 
mastodons and the ancestors of the llamas to pass southward, 
and by the ancestors of the North American tree porcupines 
in entering North America. 

Since I wrote my essay on the problem of a former land 
connection, other than the Central American one, between 
North and South America, I have had opportunities of study- 
ing the subject more at my leisure. I find that the affinity 
existing between south-western North America and the 
extreme south of South America among some of the more 
ancient groups of animals is greater than I thought. Let 
us examine this curious relationship between the two 
widely separated faunas a little more closely. I explained 

* Scharff, E. F., " Early Tertiary Land-connection." 
t Osborn, H. F., "Age of Mammals," p. 292. 


(p. 120) that the mammals inhabiting Patagonia and south- 
western North America in late Mesozoic and early Tertiary 
times indicated the existence of a direct land bridge between 
these two areas. But the Patagonian land-mass must have 
included Chile or such parts of it as were then above sea- 
level. We know very little of the past fauna of Chile. It', 
as I endeavour to prove, a portion of our living fauna is of 
great antiquity, some animals should reveal unmistakable 
signs of this early affinity of Chile with that of south-western 
North America. I imagine, as previously stated, that the early 
Tertiary land bridge connected Chile directly with western 
Mexico and California. Hence we might expect that not only 
Chile and Patagonia, but the outlying islands of California, 
as well as the Galapagos islands, ought to show a faunistic 
relationship with one another. 

I have already alluded to the fact that three groups of the 
snail Bulimulus inhabiting Chile and Peru, the Galapagos 
islands and Lower California are very similar in their narrow 
elongated shape, so that they were formerly considered to be 
very closely related. Dr. Pilsbry* is now of opinion that this 
similarity in appearance is not any evidence of near relation- 
ship but a special parallel modification of different Bulimulus 
stocks, or, as we might say, a case of convergence. He attri- 
butes these cases to be products of similar environments. Yet 
is the environment of Chile really so similar to that of the 
Galapagos islands or Lower California ? Climatically and geo- 
logically I should think there must be a good deal of difference 
between these districts. 

The relationship between the western North American snail 
fauna and that of western and south-western South America 
is well illustrated by the distribution of the genus Epi- 
phragmophora. This is a group of snails entirely confined to 
Central America and the Pacific borders of North and South 
America. It is important to note also that the majority of 
the snails inhabiting the islands off the coast of California 
belong to this genus. Cerros, Guadelupe, Santa Barbara, 
San Clemente and other islands all have their peculiar species 
of Epiphragmophora. Now the most interesting point about 

* Pilsbry, H. A., "Manual of Conchology (Pulmonata)," X., p. 126. 


the members of this genus is their astonishing resemblance, 
as mentioned before, to certain species of the European group 
Helicigona, including the well-known Arianta arbustorum. 
Some of the American shells are externally so much like the 
latter that they were placed into the same genus until recently. 
Dr. Pilsbry,* however, maintains that the European Heli- 
cigona differ anatomically from Epiphragmophora, and that 
the two should be placed in quite distinct groups. Even on 
anatomical grounds opinions may differ very radically, for 
Dr. von Ihering claims that the American Epiphragmophora 
and the members of the group Helicigona are genetically con- 
nected. After careful examinations of the European Arianta 
arbustorum and the Argentine Epiphragmophora tucuma- 
nensis he could perceive no noteworthy anatomical difference 
between the two, and placed them both into the same genus. 
If Dr. von Ihering's f observations are correct, this is another 
example of the same extraordinary relationship between Euro- 
pean and west American forms that I have had occasion to 
point out in previous chapters. Occasionally this relation- 
ship manifests itself only in Europe and south-western North 
America. In other cases it exists simply between European 
and western South American forms. In this case both the 
American western centres are related to the European centre 
of dispersal. It clearly illustrates, as stated before, that 
Europe was once joined to a mid-Atlantic land bridge which 
communicated directly with a belt of land lying to the west of 
America. On the disappearance of the latter many of the 
animals inhabiting that belt of land took refuge in the western 
parts of America and still persist there. 

Epiphragmophora, all the same, does not clearly reveal the 
special affinity that exists among the older forms of animal 
life of south-western North America and southern South 
America. The wingless insects known as " bristle tails "(Thy- 
sanura) are generally looked upon as a very ancient group. 
Among these the genus Japyx, which is easily recognisable 
by its forceps -like tail -appendage, has a noteworthy distribu- 
tion. Japyx solifugus occurs in the Mediterranean region, 

* Pilsbry, H. A., "Manual of Conchology (Pulmonata)," IX., p. 195. 
t Ihering, H. von, " System der Heliciden," p. 422. 


whence it has spread as far north as southern Germany. 
Several other species are peculiar to southern Europe and 
Madeira. One Japyx is found in India. Across the Atlantic 
Japyx subterraneus has been observed in the Mammoth Cave 
of Kentucky, while a second species (Japyx saussurei) inhabits 
only Mexico and Chile.* 

The curious scorpion-like Koenenia, the only genus of the 
order Palpigradi, is confined, as already mentioned, to 
southern Europe, Texas and Chile. As far as America is 
concerned, it is limited in its range to the two regions alluded 
to. The family of spiders, Mecicobothriidae, according to 
Mr. Pocock,f occurs only in the western States of North 
America and in Argentina, the genus Hexura being found in 
the former and Mecicobothrium in the latter part of America. 

The ancestral form of the evidently very ancient family of 
earthworms, Megascolecidae, seems to be Notiodrilus. This 
genus is met with in America only in Mexico, Guatemala, 
Chile, Argentina, Patagonia, Tierra del Fuego and the Falk- 
land islands, while another genus, namely, Kerria, inhabits 
Lower California and southern South America, being absent 
from the intermediate area. The slug Philomycus has a wide 
range in North America, occurring in Mexico, Guatemala and 
Costa Rica, and then turns up again far away to the south 
in Chile. Similarly the salamander Plethodon, which in North 
America inhabits principally the western States, is found in 
a single district in South America, namely, in Argentina. 
There are a large number of similar examples among such 
forms of insects that we have reason to believe date back 
to at least early Tertiary times. The best known example, per- 
haps, is that of the common northern genus Carabus. In 
Mexico there are still two species. Neither in Central 
America, nor in northern nor .middle South America has it 
been discovered, while in Chile and Tierra del Fuego there 
are nine species. That the latter are now considered suffi- 
ciently distinct to be placed into the separate genus Cero- 
glossus does not alter the fact that they are closely related 
to northern Carabi, whereas they have no affinity at all with 

* Karsch. F., "Neue Fundorte von Japyx," p. 154. 

t Pocock, R. I., " Geographical Distribution of Spiders," p. 346. 


any South American forms. The genus of Tenebrionidae 
Apocrypha is quite confined to California and Chile. That 
all these instances indicate the existence of a former direct 
land connection between Chile and California, independent 
of the rest of South America, as I have explained before, is 
indicated by another example derived from the same family 
of beetles. The three closely allied genera Arthrocomus, 
Stomion and Eurymetopon occur respectively in Chile, the 
Galapagos islands and California. 

Under this heading also comes an instance of distribution 
which had not hitherto been noted and which was pointed out 
by Mr. McLachlan.* He remarked that the family Limno- 
philidae, a family of insects the larvae of which manufacture 
those cases of twigs and straws, so abundant in northern 
ponds and ditches, is not known to occur south of Mexico, 
except in Chile and the Falkland islands. 

It was Dr. Wallace,f I believe, who first drew attention, 
as already noted in a previous chapter (p. 235), to the 
remarkable fact that a large number of European and North 
American genera, such as the butterfly Argynnis and the 
running beetle Carabus, reappeared far south of the tropics 
in Chile, Argentina and Tierra del Fuego. I also alluded 
to his explanation of the manner in which this surprising 
phenomenon had been brought about. He was under the im- 
pression that this migration across the tropics had been 
effected mainly during successive Glacial Epochs, when the 
mountain range of the isthmus of Panama, if moderately in- 
creased in height, might have become adapted for the passage 
of northern forms, while storms would often carry insects 
from peak to peak over intervening forest lowlands or narrow 
straits of sea. Improbable as this theory may appear, it might 
still be defended as long as we had to deal merely with the 
occurrence in southern South America of a few northern 
insects. But the phenomenon is, as we have seen, a much 
more widespread one. It applies to earthworms, slugs, sala- 
manders and even mammals, and it is evidently the result 
of a dispersal which occurred long before the Glacial Epoch. 

* McLachlan, E., "Insect Fauna of Chile," p. 162. 

f Wallace, A. E., " Geographical Distribution," Vol. II., p. 45. 


It took place, in fact, in very remote times, probably during 
the dawn of the Tertiary Era; and only such forms as were 
capable of preserving their specific and generic characters 
till the present day clearly reveal their northern origin. 

The flora of the New World, as I remarked in the paper 
just referred to, retains even more pronounced traces of that 
curious relationship between the south-western areas of its 
two continents.* As among the fossil mammals so do we 
find also among the fossil plants, a remarkable affinity in late 
Mesozoic deposits between species from Argentina and from 
western North America. Professor Berry tells us that in 
mid-Cretaceous times seventy-five per cent, of the known 
plants of Argentina were characteristic types of the Dakota- 
group flora of North America. During a period of geological 
history when a large section of the existing western part of 
South America was under water, there was this extraordinary 
similarity between two regions lying at such a great distance 
from one another. Professor Berry justly argues that the 
surprising affinity of these floras to one another points to a 
community of origin. In these ancient plant deposits of 
Argentina ,all the familiar northern genera such as Lirio 
dendron, Liquidambar, Cinnamomum and Sassafras are met 
with. Even Platanus, Populus, Quercus and other modern 
genera are represented. No wonder that Professor Berry f 
came to the conclusion that a geographical connection must 
have existed between North and South America during mid- 
Cretaceous times. During Cretaceous and early Tertiary 
times the genus Sequoia, to which the Californian red-wood 
and big-trees belong, likewise ranged from North America to 
Chile. And it is now held by many botanists that the fossil 
Sequoia langsdorfi is identical with the still living big-tree 
(Sequoia gigantea) of California. We possess no fossil testi- 
mony of the occurrence of the smaller deciduous plants in 
those remote times, but to judge from the fact that many of 
the Mesozoic genera of trees still survive to the present day, 
certain persistent deciduous species presumably did so too. 
Mr. Engelhardt J records a number of plant remains from the 

* Scharff, E. F., " Early Tertiary Land-connection," pp. 523526. 
f Berry, E. W., " Mid- Cretaceous Geography," p. 510. 
I Engelhardt, H., " Tertiarpflanzen von Chile," p. 635. 


west coast of Chile which he considers to be of early Tertiary 
age. Of these the greater portion, he thinks, belongs to still 
existing species, though he hesitates, in the absence of further 
evidence, to adopt the recent specific names. At any rate, 
almost all are very closely related to species now living in 
the West Indies, Central America and Brazil. 

There exists a very widespread assumption that no species 
of animals or plants survive to this day from the Mesozoic 
Era. As our knowledge of fossil animals increases, instances 
accumulate of very close resemblance of Mesozoic species of 
mollusks to those now living, while some genera certainly 
date back to Palaeozoic times, so that we can no longer place 
implicit faith in the old traditional belief. My own convic- 
tion is that many species of the less highly organised groups 
of animals have survived unchanged even from those remote 
times to the present day. We know that certain species of 
plants such as Sequoia langsdorfi, and some of the Unios 
among animals, have remained practically unaltered through- 
out a series of geological ages. It seems equally possible 
that others of whose geological history we know nothing have 
likewise done so. 

Professor Asa Gray and Sir Joseph Hooker * long ago 
directed attention to the unexpected feature that many genera, 
and even species of North American plants, recur in the dis- 
tant regions of southern South America. More recently 
Professors Engler,f Bray and Hackel have dwelt on this re- 
markable phenomenon, and have speculated on the problems 
connected with it. The flora of the Rocky Mountains, 
including the Sierra Nevada Mountains above the transition 
zone, and the mountains of Chile and Argentina, though 
separated from one another by a stretch of some ten degrees of 
latitude of moist tropical country, abound in northern genera 
of plants, such as Ranunculus, Anemone, Geranium, Spiraea, 
Geum, Rubus, Saxifraga, Vaccinium, Gentiana, Hieracium 
and others. The greater number of such plants occurring in 
the southern continent are endemic, pointing to long -con- 
tinued isolation. Yet certain species even of the Rocky 

* Gray, Asa, and J. Hooker, "Vegetation des Eocky Mountain Ge- 
bietes," p. 292. 

t Engler, A., " Entwicklungsgeschichte der Florengebiete," .II., p. 2*>6. 


Mountain arctic alpine region reappear in the extra-tropical 
Andes, being, so far as is known, wholly absent from the 
Mexican Mountains as well as from the tropical Andes. 
Among these Professor Bray mentions particularly Gentiana 
prostrata, Trisetum subspicatum, Primula farinosa and its 
variety magellanica, Draba incana, Alopecurus alpinus, Saxi- 
fraga caespitosa, Polemonium microcanthum and Collomia 
gracilis. Dr. Stapf kindly drew my attention to another 
instance, namely, to the occurrence on the mountains of 
Argentina, at a height of 10,000 feet, of the typically arctic 
genus of grasses Phippsia. The similarity in the flora of the 
two regions is by no means confined to mountain forms. 
Some species of the western plains of North America also 
reappear in the extreme south of South America. Thus the 
monotypical saxifrage Lepuropetalon, a peculiar and some- 
what abnormal genus, grows in damp low-lying meadows in 
Texas. Yet far south on the coast of Chile the same species, 
Lepuropetalon spathulatum is to be met with. 

A few of these plants may possibly have been casually in- 
troduced from the northern locality to the southern. But in 
most cases, Professor Bray * thinks, we have to deal with 
forms which were connected by a remote ancestry, and which 
flourished at a time and under conditions permitting a more 
general distribution. 

What these conditions were like he does not venture to sug- 
gest, but a direct land bridge between western North America 
and Chile probably existed, as I suggested, in late Cretaceous 
and early Tertiary times. To it, I think, the relationship 
of the floras of these two widely separated areas to one 
another, is mostly due. Whether many species of plants 
have persisted to the present day from such remote times 
we do not know. Some no doubt have, and, as already 
stated, others, among them those alluded to, may have 
done so. A large portion of this old western land, with' 
its mixture of a southern and northern fauna and flora, 
evidently remained above sea-level until a much later geo- 
logical period. The evidence derived from certain relict 
land -areas of this Pacific land belt clearly shows that a 

* Bray W/L., " Belations of North American Flora," pp. 709716. 


southern flora passed over them, for of course southern 
plants penetrated northward just as animals did. Mr. Sereno 
Watson,* for example, in referring to the flora of the small 
island of Guadalupe off the west coast of California, remarks 
that the presence of so many South American types suggests 
that this, and the similar element which characterises the 
flora of California, may be due to some other connection be- 
tween these distant regions than the one now existing. He 
expresses the opinion, too, that all the peculiarities in the 
western floras of both continents had a common origin in an 
ancient flora which, prevailed over a wide, now submerged 
area, and of whose character the former are the partial 

As regards 'the question whether the northern plants 
growing in Chile and Argentina can really be regarded as 
indigenous or introduced, Professor Hackel has recently dealt 
with this problem from a new point of view. He shows that 
the species which are either identical with European or North 
American plants, or such as may be considered as varieties or 
sub-species of them, increase rather than diminish as we go 
further south. Thus he records fifty-one plants from 
southern Patagonia and Tierra del Fuego belonging to this 
group, which he regards as undoubtedly indigenous to these 
countries. Thirty others may possibly have been introduced 
by man. Professor Hackel thinks that their southward 
advance could only have taken place step by step on a land 
surface. If these plants had passed across Central America 
and along the Andes, we should find relict colonies of the 
species, or at least their modified descendants, scattered on 
their route of migration. Only very few of such occur in 
Bolivia, Peru and Ecuador, and these have the appearance 
of having reached these localities from the south rather than 
from the north. When we examine the composition of this 
remarkable flora in the Magellan district, we notice, according 
to Professor Hackel,f that the majority of the species belong 
to the grasses (Gramineae) and the sedges (Cyperaceae). The 
dicotyledons only comprise fifteen species. This he believes 

* Watson, Sereno, " Flora of Guadalupe Island," p. 112. 
t Hackel, E., " Flora der Magellanslander," pp. cxi cxv. 
L.A. E B 


to be due to the greater antiquity of the grasses and sedges, 
and he argues in favour of an ancient direct land connection, 
long since disappeared, on which the slow southward advance 
of these forms took place, the identical species having pre- 
served their specific characters throughout a long series of 
ages. This is precisely the conclusion I have come to in 
regard to many species of animals of very wide range, con- 
trary to the view generally held that most of such forms are 
to be regarded as human importations scattered throughout 
the world by commerce. 

I have mentioned on several occasions that California ex- 
tended considerably further westward in former times, and 
that the numerous little islands such as Guadalupe, Cerros, 
Santa Catalina, Santa Eosa and others are the visible remains 
of that ancient Pacific land belt. The fact that they are in- 
habited by fourteen species and varieties of reptiles and 
amphibians* not known from the mainland, suggests that the 
islands have been separated from the latter since at any rate 
pre-Glacial times. 

Now on the coast of Chile we have similar evidence o'f a 
westward extension of land in former times. The tiny deer 
known as the "pudu" occurs in Chile and Chiloe island. 
Its only near relative lives in Ecuador. The small mouse-like 
mammal Acodon brachyotis of Chiloe island and the Chonos 
archipelago is peculiar to these islands, and so are many other 
species of the lower groups of animals. There is, in fact, 
quite a considerable assemblage of animals and plants on 
these islands, indicating a former westward extension of the 
mainland. Even the far distant Juan Fernandez island, 
which lies five hundred miles from the mainland, possesses a 
peculiar species of humming-bird (Eustephanus fernan- 
densis) and no doubt other indigenous forms of animal life. 
It is situated on the tract of the old land belt which I believe 
to have once extended from the west coast of southern Chile 
to south-western North America, and from there eastward 
to Europe (Fig. 14). In early Tertiary times already parts of 
this old land bridge had disappeared, so that Chile and south- 

* Denburgh, J. van, " Keptiles and Amphibians of Pacific Coast 
Islands," p. 4. 


western North America became disconnected. Yet the eastern 
shores of the great peninsula, which long afterwards stretched 
southward from the west coast of California, joined the con- 
tinent of South America further north, probably in Ecuador 
and Peru (Fig. 16). Animals and plants arriving from the 
north, like the mastodon above alluded to, no longer reached 
Chile directly, though still able to enter South America by 
this second route, which added many new forms to the fauna 
of the continent. 

This is all I have to say on the origin of the f'aunistic and 
floristic relationships between southern South America and 
North America as well as Europe. I have also referred 
above to a kinship existing between the Santa Cruz fauna of 
Argentina and the living fauna of Australia. This is a pro- 
blem of even greater interest than that just discussed. At 
any rate, the testimony in its favour has converted many of 
those who had hitherto looked upon great changes in the 
disposition of continents and ocean basins during the Tertiary 
Era as altogether visionary speculations. Some, like Professor 
Osborn,* who contends that the Atlantis hypothesis is highly 
improbable, regard the hypothetical reconstruction of a great 
southern continent (Fig. 21) uniting South America with 
Australia as one of the greatest triumphs of recent biological 
investigation. The theory of the former land connection be- 
tween southern South America and New Zealand or Australia 
has been discussed from almost every point of view, so that we 
possess at present quite an extensive literature on the subject. 
A few years ago Dr. Ortmann summarised the more important 
papers dealing with this subject, but many other essays have 
since been published throwing further light on this fascina- 
ting problem. The great majority of those who have endea- 
voured to account for the undoubted faunistic or floristic affi- 
nities of southern South America and the Australian region 
have come to the conclusion that it is due to the existence of a 
former antarctic continent. A few, however, argue that there 
was a more direct land connection between the two areas 
across the southern Pacific, while some believe that the resem- 
blance in the fauna and flora of the countries referred to has 

Osborn, H. F., " Age of Mammals," p. 75. 

E E 2 


arisen partly by accidental dispersal and partly by migration 
through the existing continents, when they were united by 
short temporary land bridges in the far north. 

Sir Joseph Hooker seems to have been the first to formulate 
the theory of a former land connection of southern lands from 
the occurrence of identical species of plants in widely 
separated areas. The first to mention an antarctic continent 
in this connection was Professor Ruetimeyer. This was 
twenty years later, in 1867. Six years elapsed after that until 
Professor F. W. Hutton again brought forward the theory, 
expressing his belief in the former existence of an antarctic 
land-mass and its union with the southern extremities of the 
present continents. A few years later Dr. Gill declared him- 
self in favour of a large continental mass embracing Africa, 
South America and Australia, his views being based on the 
geographical distribution of fishes. The fact that Dr. Wallace 
repudiated the opinions expressed by his predecessors need 
not be taken too seriously. Since his great work on the geo- 
graphical distribution of animals was issued, facts have been 
brought to light which are so overwhelmingly in favour of 
some kind of continuous land connection between South 
America and the Australian region that he could not at the 
present day assert, as he did then, that the affinities alluded to 
are mostly due to accidental dispersal across the ocean, aided 
by the presence of an antarctic land-mass unattached to other 
continents. Professor Hutton subsequently abandoned his 
theory of a land-bridge between the two regions by way of an 
antarctic continent in favour of a land connection right across 
the South Pacific towards the end of the Mesozoic Era.* 

In the year 1891, an important contribution to the subject 
was published by Dr. von Ihering.f He points out that the 
species of fresh-water Unio from New Zealand and adjoining 
countries have their nearest allies exclusively in Chile and 
southern Brazil, and that the fresh-water crayfishes (Para- 
stacidae) show the same character. The conclusions he de- 
rives from these and other facts are that the southern portions 

* Ortmann, A. E., " Origin of Antarctic Faunas and Floras," pp. 139 

t Ihering, H. von, " Relations between New Zealand and South 
America," p. 442. 


of South America (Archiplata) extended southward in form 
of a South Pacific antarctic continent, which kept this area 
in communication with the Pacific continent during the whole 
of Mesozoic times (Fig. 17). From the latter became separated 
first a number of Polynesian islands, then New Zealand and 
finally Australia and New Guinea. Dr. von Ihering deals with 
the same problem in several of his papers, some of which have 
now been reprinted in his recent work entitled " Archhelanis 
and Archinotis." * After giving a brief review of the main 
features pointing in favour of a former antarctic continent 
connecting the Australian region with South America, Pro- 
fessor Jacobi f expresses himself in favour of the theory. A 
great deal of additional evidence in support of the same view 
was collected by Dr. H. O. Forbes. J He cites many examples 
among birds, the southern pikes (Galaxiidae), which occur in 
Chile, Patagonia, the Falkland islands and New Zealand, the 
family Cystognathidae among amphibians, and numerous 
plants f as illustrating the intimate affinity existing between 
the faunas of the Australian and South American regions. 
He also traces a faunistic relationship between these regions 
and the Madagascar area, and concludes that all these 
southern land-masses were once joined to an antarctic con- 
tinent. It is easy enough to vaguely describe such a land 
connection. Dr. Forbes was not satisfied to do this. He 
was the first to give a map clearly indicating how he 
conceived that all the areas referred to were united with one 

Mr. Beddard briefly discussed the problem in 1895. 
While criticising Dr. Forbes' evidence in so far as it relates 
to the struthious birds and penguins, he admits that there 
are very strong arguments in favour of a former land connec- 
tion between Patagonia and the Australian region, particu- 
larly when we take into consideration the distribution of 

As Mr. Hedley aptly remarks, "we may compare the 
shattered biological monuments of Tasmania and South 

* Ihering, H. von, " Archhelenis and Archinotis." 

t Jacobi, A., " Biogeographische Gebiete," pp. 204 207. 

t Forbes, H. 0., "The Chatham Islands." 

Beddard, F. E., "Zoogeography," pp. 161172. 


America to the broken columns found by Oriental travellers 
in the ruined and deserted cities of a vanished civilisation. 
And as an archaeologist may restore from such fragments the 
fallen temples or disused aqueducts, so may a naturalist trace 
the missing arches of life that once spanned the gap." Mr. 
Hedley * favours the theory of a direct land connection in 
Mesozoic or early Tertiary times between Tasmania and 
Tierra del Fuego across the South Pole, while New Zealand 
then reached sufficiently near this antarctic land without join- 
ing it, to receive by flight or drift many animals and plants. 
He thinks the faint affinity of Antarctica to Africa would be 
explicable on the supposition that before either America or 
Australia had united with the former, Africa had already 
broken away from it. 

A very memorable discussion on this question took place 
during the fourteenth annual meeting of the American 
Society of Naturalists in Philadelphia. Reviewing the geo- 
logy of the antarctic regions, Professor Heilprin f stated that 
in its relation to the other continents there was reason to 
believe that Antarctica, whether as a continent or in frag- 
mental parts, had a definite connection with one or more of the 
land -masses lying to the north, and that the suspicion could 
hardly be avoided that such connection was, if with nothing 
else, with New Zealand (and through it with Australia) and 
Patagonia. The facts of palaeontology are best explained, 
according to Professor Scott J on the assumption that the 
antarctic land-mass has at one time or another heen connected 
with Africa, Australia and South America, all of which once 
radiated from the South Pole, just as North America and 
Eurasia now do from the North Polar area. 

Although Professor Britton cited many examples of 
astonishingly close relationship between plants of Australia, 
southern South America and South Africa, it is unnecessary 
in his opinion to invoke as an explanation a former land con- 
nection across the antarctic region. 

Arguing from the geographical distribution of the fishes, 

* Hedley, C., " Surviving Eefugees in Austral Lands," pp. 36. 
f Heilprin, A., " Geology of Antarctic Eegions," pp. 306307. 
t Scott, W. B., " Antarctica Palaeontology," p. 310. 
Britton, N. L., "Origin of Antarctic Flora," p. 311. 


reptiles and amphibians, Dr. Gill* maintained that the evi- 
dence in favour of a former antarctic continental area is 
strong, and, in view of the affinities of the species of the now 
distant regions, the conclusion is logical that the time of 
disruption was not remote in a geological sense. 

So far as existing mammals and birds are concerned, said 
Dr. Allen,f there seemed to be very slight need for calling in 
the aid of a former antarctic continent to explain their pre- 
sent distribution. The distribution of marsupials alone 
among mammals gave a hint of a possible former land connec- 
tion between South America and Australia. 

Mr. Ancey J gives a valuable contribution to the subject 
at issue, founded on the distribution of the terrestrial 
mollusks. The results of his enquiries are that the evident 
relationship of the faunas of South America and the Austra- 
lian region can be satisfactorily explained by the assump- 
tion of a former antarctic continent uniting these land areas. 

As the antarctic problem has been widely discussed from 
an ornithological standpoint, and since Dr. Forbes and Pro- 
fessor Milne-Edwards have based their theories of former 
land connections largely on the past and present range of the 
flightless ratite birds, it is right to mention that Professor 
Burckhardt disagrees entirely with their conclusions. He 
does not dispute the possibility of former antarctic land con- 
nections, especially between New Zealand and South America. 
But he does not consider that the flightless giant birds of the 
southern hemisphere can in any way be regarded as furnish- 
ing evidence in favour of the theory of a former antarctic 

From a zoogeographical point of view, inconspicuous 
groups of invertebrates, such as the earthworms, are of far 
greater value than birds are. When the older writers dis- 
cussed the antarctic problems we knew as yet practically 
nothing of their distribution. Within recent years several 
zoologists have made a special study of this group, and have 
found that earthworms are of the greatest value in the study 

* Gill, Th., " Fishes, Eeptiles and Amphibians of the Antarctic," p. 315. 

t Allen, J. A., "Birds and Mammals of the Antarctic," p. 317. 

J Ancey, C. F., "Faunes malacologiques australes," p. 27. 

Burckhardt, E., " Problem des antarktischen Schopfungscentrum." 


of palaeogeography. Professor Benham, moreover, has the 
advantage of residing in New Zealand, so that his views on 
these problems are of peculiar interest. To the south of New 
Zealand, between the latter and the antarctic regions, there 
are two groups of islands, viz., the Auckland and Macquarie 
islands. Now it is a specially noteworthy fact, remarks Pro- 
fessor Benham, that the species of earthworms from Mac- 
quarie island, and those occurring in the Auckland islands, are 
more nearly allied to South American and South African 
species than to those of the mainland of New Zealand. The 
evidence derived from the distribution of earthworms is, ac- 
cording to Professor Benham,* strongly in favour of the 
theory that New Zealand, South America and South Africa 
were once connected with one another by land. 

A study of the fresh-water crayfishes yields an even more 
remarkable testimony in support of the same view, at least 
of the theory that New Zealand and southern South America 
were connected by land. Dr. Ortmann points out that the 
family of crayfishes known as Parastacidae is confined to 
Australia, including Tasmania, New Zealand, southern South 
America and Madagascar. This family has thus an extremely 
discontinuous range and must be of great antiquity. Dr. 
Ortmann f believes that the Parastacidae existed in Australia 
in Upper Cretaceous times and thence spread by a land con- 
nection into Antarctica, from which Parastacus reached Chile. 
Subsequently Parastacus extended its range to Argentina and 
southern Brazil. Madagascar, he thinks, was not connected 
with the Antarctic Continent, and received its fresh -water 
crayfishes by another route. Besides this affinity in the fresh- 
water crayfishes of the Australian region and temperate South 
America, there are other crustaceans showing similar features. 
Mr. Geoffrey Smith J tells us that one of the commonest 
plankton organisms in the mountain lakes and tarns of 
Tasmania is the copepod crustacean Boeckella. Now this 
genus nowhere spreads into the tropics, yet reappears in 
New Zealand and southern South America. 

* Benham, W. B., " Geographical Distribution of Earthworms," 
pp. 329335. 

t Ortmann, A. E., " Distribution of Decapods," p. 340. 
{ Smith, Geoffrey, " Naturalist in Tasmania," p. 137. 


Professor Kolbe contributes a careful review of the subject 
from the point of view of the beetles (Coleoptera). After 
illustrating the intimate relationship that exists between 
southern South America and the Australian region, by means 
of the distribution of some of the Carabidae, Lucanidae, Scara- 
baeiidae, Buprestidae, Elateridae and other families of beetles, 
he comes to the conclusion that an Antarctic Continent 
have existed in early Tertiary times. Professor Kolbe* be- 
lieves that this continent was joined to the southern parts of 
South America, New Zealand and Australia, and that some 
genera passed from the latter to America, whilst others 
travelled in the opposite direction. Although the affinities 
of South Africa and Madagascar with South America and 
Australia are much less pronounced, Professor Kolbe never- 
theless urges that these countries, too, were to some extent 
connected by land with the Antarctic Continent. 

Dr. Arldtf recently gave us an excellent summary of the 
most striking points in favour of the theory that southern 
South America and the Australian region were once con- 
nected by means of an Antarctic Continent. The theory meets 
with his entire approval, but he concurs with Professor 
Hutton and others in the belief that a second means of com- 
munication between Australia and South America was fur- 
nished by a mid-Pacific land bridge. 

Finally, Professor ChiltonJ reviewed the problem from the 
point of view of New 'Zealand and the neighbouring su.b,- 
antarctic islands and their fauna and flora. His conclusions 
are strongly in favour of a former land connection by way of 
the Antarctic Continent between South America and New 

1 have still to state my own opinion on this subject. That 
there was some kind of a direct land connection between Chile 
and New Zealand and Australia appears to me obvious. The 
strongest arguments in its favour are supplied by the distri- 
bution of the fresh-water mussels and the fresh-water cray- 
fishes. The genus Diplodon, one of the Unionidae, inhabits 

* Kolbe, H., " Coleopteren der Magalhaensischen Sammelreise," 
pp. 19 30 (compare also "Die Siidpolarkontinenttheorie "). 
f Arldt, Th., " Bedeutung der Antarktis," p. 370. 
t Chilton, C., " Biological Eelations of Sub-antarctic Islands," p. 806. 


the rivers of western and southern South America. No living 
Diplodon has ever been discovered in Central or North 
America, where other Unionidae are found in the greatest pro- 
fusion. No Diplodon has ever been discovered fossil either in 
Central or North America, although a great many fossil 
Unionidae are known to science. Hence it seems certain that 
Diplodon has never lived north of its present habitat. Yet 
in New Zealand, in Tasmania and Australia this genus re- 
appears. Further north in Asia it has never been met with. 
Some zoologists are of opinion that affinities such as the one 
alluded to can be interpreted by the supposition of a former 
sub-universal distribution and a subsequent extinction in all 
but the present habitats. We do not possess a shadow of any 
evidence for such a belief, in so far as the range of the fresh- 
water mussel Diplodon is concerned. 

Let us take the second case, that of the fresh-water cray- 
fishes. Crayfishes are abundant in the streams of North 
America and Asia, but they all belong to the family Potamo- 
biidae, while southern South America is inhabited by quite a 
different family, the Parastacidae. Crayfishes of the latter 
family are again met with on the other side of the Pacific, but 
only in the extreme south, in New Zealand, Tasmania and Aus- 
tralia. Neither of these instances can be due to convergence, 
nor to a passage from one continent to the other by way of 
the northern continents. A direct land bridge becomes abso- 
lutely essential, yet this need not necessarily have lain in the 
direction of the antarctic regions. Its position might have been 
further north, as suggested by Professor Hutton,* although 
the latter has more recently revised his theories, in so far 
as he now advocates two south Pacific land bridges, instead 
of a single one as previously maintained. In 1905 he 
announced that, having reconsidered his former conclusions, 
he believed that an Antarctic Continent existed in Jurassic 
times which connected South America with New Zealand and 
South Africa. He thinks that this continent sank in the Cre- 
taceous Period, Antarctica never having since been connected 
with northern lands. Subsequently, either during Cretaceous 
or early Tertiary times, a Pacific Continent must have united 

* Hutton, F. W., " Ancient Antarctica," p. 245. 


New Zealand with Chile, which, as he maintains, ceased to 
exist at the close of the Eocene Period. 

Certain geological features no doubt point to a former west- 
ward extension of Chile. Dr. Burckhardt * showed that in 
western Chile there are enormously thick deposits of porphy- 
ritic conglomerates which become more attenuated and com- 
posed of lighter sandy material as we proceed eastward. 
Hence he argues that these deposits were laid down on an 
ancient shore-line of a vast western land-mass of which the 
existing coast cordillera of Chile is the last remnant. He 
advocates, in fact, nothing short of what we might call a 
Pacific Continent which lay mainly to the westward of Chile. 
That land formerly extended in that direction I have en- 
deavoured to demonstrate from purely faunistic evidence, but 
I believe that it was part of a great circum-Pacific belt of 
land which stretched mainly northward, communicating from 
time to time with Central America and the Antilles, and also 
with Mexico and western California, and then eventually bend- 
ing across to eastern Asia in a great loop and thus joining 
New Guinea, Australia and New Zealand. That Central 
America and western North America must have been thus 
connected by land with the Australian region has been urged 
above on several occasions when dealing with the North 
American fauna. It might be argued that if such a northern 
land connection once existed, uniting Chile with the Austra- 
lian region, there would no longer be any necessity for postu- 
lating an antarctic land bridge. Professor Dahl has recently 
taken up this attitude,* illustrating his adverse criticism against 
the antarctic theory by means of the distribution of spiders. 
He does not adopt, of course, my view of a northern semi- 
circular land belt because this theory has never before been 
published and is entirely my own. His contention is that the 
continents and oceans have remained within the lifetime of 
the present fauna what they are now, that is to say, within 
the more recent geological periods, except that the great land- 
masses were joined in, the north. A powerful centre of dispersal 
existed in the arctic regions, according to Professor Dahl.f 

* Burckhardt, 0., " Traces d'un ancien continent," pp. 186 190. 
t Dahl, F., " Die Verbreitung der Spinnen." 


and from it streamed forth practically all life, which eventually 
reached the most southern tips of the continents. We thus not 
only find there the oldest forms congregated together, accord- 
ing to Professor Dahl, but these would naturally resemble one 
another in all the southern parts of the continents. It is 
a simple theory, and at first sight, as I mentioned before, 
when discussing the similar views of Dr. Haacke, it seems 
to put before us the solution of all the problems of dispersal 
in a satisfactory manner. No one, however, who has taken 
the trouble to read this book, can for a moment, I venture to 
think, perceive any justification for holding such a belief as 
that of Professor Dahl's. There may possibly have been such a 
centre of dispersal as that suggested, but it was long before 
any geological records existed. Any forms of animal life that 
we are acquainted with show clearly and unmistakably that 
there have been scores of great centres of dispersal in the 
world, and that from them streamed forth new forms in every 
available direction. Northern animals advanced southward 
and southern forms northward, aided, no doubt, by the ever- 
changing conditions of climate and the gradual evolution of 
oceans and continents. 

To return to the question at issue, let us examine what light 
the marine fauna throws on the problem. The geographical 
distribution of the elephant seal (Macrorhinus leoninus), 
which occurs on the shores of southern South America, of 
the antarctic lands, of New Zealand and Tasmania, will not 
contribute much to its solution, but some of the inverte- 
brates are of more importance. The Eocene beds of Pata- 
gonia, as well as those of New Zealand, contain several species 
of marine mollusks common to both, such as Cucullaea alta, 
Scalaria rugulosa and Turritella ambulacrum. Other Pata- 
gonian species, for example, Brachydontes magellanica, Tur- 
ritella patagonica and Crepidula gregaria only appear in New 
Zealand in later deposits. Besides these species the older 
Tertiaries of the two countries have a large number of genera 
common to both. Dr. von Ihering * thus concludes with con- 
siderable justification that the geographical conditions during 
early Tertiary times must have been favourable for a dis- 

* Ihering, H. von, "Mollusques Fossiles de L' Argentine," p. 499. 


persal of marine mollusks between Patagonia and New Zea- 
land. It is interesting to note that, whereas this faunistic 
interchange took place between Patagonia and New Zealand, 
the early Tertiary Chilean fauna exhibits scarcely any traces 
of relationship with that of New Zealand. 

If the ancient land connection between southern South 
America and New Zealand had been a westward extension of 
the coast of Chile, we should expect a greater resemblance 
between the old Tertiary fauna of the latter country and New 
Zealand than between Patagonia and New Zealand. The evi- 
dence derived from the Eocene mollusks of Patagonia and 
New Zealand is, therefore, in favour of a southern land con- 
nection by way of the antarctic regions. I cannot believe that 
an Antarctic Continent, in the sense of Forbes or shorn, 
could have existed at that time. If the affinity of the marine 
faunas of Patagonia and New Zealand was due to a con- 
tinuous shore -line between the two countries, Chile must 
have extended southward beyond Tierra del Fuego and then 
have curved across to New Zealand in a great loop. The Pata- 
gonian fauna spread to New Zealand on the south side of the 
latter. Hence the South Polar region must have been largely 
covered by sea at that time. At any rate, I believe in the 
former existence of two land connections between Chile and 1 
New Zealand, one by way of North America and eastern Asia, 
the other directly across the South Polar area. 

This leads us 'to the third problem connected with that 
extraordinarily interesting region of southern South America, 
namely, its direct affinities with the anntarctic regions, and in- 
cidentally with South Africa and Madagascar. The southern 
tip of Tierra del Fuego bends eastward. At its eastern ex- 
tremity lies Staten island, and beyond it the great Burdwood 
Bank. North of the latter, at no great distance, are situated 
the Falkland islands. Nowhere is there a greater depth of 
water than a hundred fathoms between the latter arid the 
mainland. Consequently all this presumably was once part of 
South America. 

The Falkland islands cover an area of about half the size 
of Ireland, lying three hundred and fifty miles east of Pata- 
gonia. The country consists of undulating moorland, with an 
abundance of peaty soil yielding an ample supply of wiry 


grass with a growth of bushes here and there. As Darwin 
pointed out long ago, almost the whole of the islands are com- 
posed of sedimentary rocks, such as clay-slate and sandstone, 
while the hills are formed of white granular quartz. Darwin* 
considered the sandstone of Silurian age, whereas Mr. 
Newton,f who has recently examined the fossils brought back 
by Dr. Bruce, shows that they belong to the Devonian Period, 
many of them being identical with South African species of 
similar age. 

No clue, however, can be obtained from the geological fea- 
tures as to the age of the islands, or, as we might say, the 
time when the latter became separated from the mainland. 
We are thus obliged to seek for information from the exist- 
ing animals and plants. Mr. Vallentin's J account of the 
fauna and flora only deals with the general features, such as 
a casual visitor might notice. He informs us that no indi- 
genous mammals occur on the Falkland islands, as the peculiar 
wolf (Canis antarcticus) which once lived there is now ex- 
tinct. It is still customary to connect this Falkland island wolf 
with some of the mainland species, although Mr. Lydekker 
pointed out long ago that it differs markedly from all the dog- 
and wolf-like creatures (Canidae) of South America, being 
closely allied to the North American coyote (Canis latrans). 
Dr. von Ihering || places it, along with the coyotes, into the 
sub-genus Lyciscus, whereas the remaining South American 
species of the family belong to Chrysocyon, Carcinocyon and 
other genera. Neither of these writers dwell upon the causes 
of this altogether extraordinary range. Are we to regard this 
as an instance of convergence ? I scarcely think that any 
mammalogist will maintain such a view. And yet how are we 
to account for the fact of the existence in these remote islands 
of a mammal whose nearest relatives are all confined to North 
America ? It is an extremely interesting problem. I have 
alluded to the noteworthy fact more than once that Chile, and 
even Tierra del Fuego and Argentina, exhibit a very surprising 

* Darwin, 0., "Journal of Kesearches," p. 142. 

t Newton, E. T., " Fossils from Falkland Islands," p. 251. 

1 Vallentin, E., " Voyage to the Falklands," p. 352. 

Lydekker. E., " Geographical History of Mammals," p. 140. 

|| Ihering, H. von, " Verbreitung d. Siidamerik. Eaubtiere," p. 153. 


faunistic relationship with western North America, but this 
affinity seems to be confined to very ancient groups of verte- 
brates and invertebrates. I ventured to allude to the North 
American porcupines as being possibly the descendants of 
Steiromys, which I suggested had utilised the early Tertiary 
land connection from Chile to Mexico or California. Stei- 
romys, however, is long since extinct. If we suppose the 
coyotes to have taken advantage of that land connection in 
early Tertiary times, we might assume a member of that early 
fauna to have succeeded in surviving in these lonely islands 
until recent times. But the difficulty in accepting such an 
hypothesis is that it would give the coyotes a much longer 
ancestry than any palaeontologist would care to admit. Never- 
theless it is the only theory I can suggest. Another question 
that occurs to me is If the wolf was the only mammal in- 
habiting the Falkland islands, what did it live on ? Wolves are 
carnivorous, as everyone knows, usually feeding on smaller 
mammals. Yet Darwin states that it is the only quadruped 
native to the islands ; he adds, however, in a footnote, " I have 
reason to suspect there is also a field mouse." This is probably 
a species of Keithrodon or some allied genus. Still this 
mouse -like creature could scarcely have constituted the sole 
food of such a large animal as the wolf. In all likelihood, 
some larger rodent existed on the islands and may have 
become extinct before the advent of man. The wolf soon after 
this event shared the same fate. 

An examination of the earthworm fauna, which is probably 
of very great antiquity, reveals the fact that there are living 
on the Falkland islands two species of the genus Noteo- 
drilus (N. bovei and N. falclandicus) which otherwise is con- 
fined to South and Central America, New Zealand, Australia, 
Cape of Good Hope, Madagascar and several of the antarctic 
islands. A second genus, viz., Chilota, which is represented on 
the Falkland islands by Chilota daleti, inhabits only southern 
South America and the Cape of Good Hope, with the exception 
of a single species living on the Cape Verd islands. The third 
genus, Yagansia, has not yet been found on the Falkland 
islands. It is almost peculiar to Chile and Argentina. Only 
one species of Yagansia has been met with in South Africa.* 
* Michaelsen, W., " Verbreitung d. Oligochaeten," p. 74. 


Thus there appears to he a distinct faunistic affinity between 
southern South America as a whole, including the Falkland 
islands, and New Zealand, as well as South Africa and Mada-