by Thomas Barbour (1916)
Matthew's thesis, in a few words, is that the permanence of the continents and ocean basins is a surely established fact, that cyclical climatic change has been the principal known cause of the present distribution of land vertebrates, and that this distribution has been effected by successive southward migrations from a holarctic center of dispersion, and that the impetus for these migrations is to be found in the theories of the "Alternations of moist and uniform with arid and zonal climates, as elaborated by Chamberlin." There is small occasion for me to review or criticise the great bulk of evidence which Matthew has presented, specially where he has drawn upon his profound knowledge of recent and fossil mammals. In the main his contentions are highly convincing, especially where he also draws conclusions from the mammals, a group for which geologic record is adequate in comparison with the fragmentary evidence regarding the history of recent birds, recent reptiles and amphibians. With some of these groups, as, for instance, Hylids and Cystignathids, it is hard to rid oneself of the belief that their origin was antarctic and not holarctic, for the northern outpost species seem to be so obviously the depauperate offshoots of the elaborate southern stock. Matthew, however, would argue by analogy with mammalian evidence that these species are [[p. 2]] reëntrants into the area of origin, and that the great result of speciation which we see now in the southern headquarters of Hylids and Cystignathids shows that this region was peripheral in relation to their area of origin. The steps of reasoning whereby Matthew arrives at these conclusions are carefully presented in his essay and there is no object in recounting them here.
It is of the general question of land bridges and of the relation which some islands bear to continents that I have been thinking for some time, and it is only because I have had some field experience and have given thought to these matters that I have the temerity to take issue with Dr. Matthew, knowing full well that many will maintain that his opinion outweighs mine--a possible assertion I am by no means ready to deny.
I take exception to statements such as this, where in speaking of land bridges (p. 179) Matthew says, "I can see no good reason why the only animals which availed themselves of such continental bridges should be the ones which might be accounted for in other ways, while those which would furnish conclusive proof are invariably absent." (Italics are mine.) I have maintained elsewhere that a waif fauna is easily recognized as such, and that the presence of burrowing amphibia, onychophores, cyprinodont fishes and many other groups of delicate organisms which are balanced to one particular environment cannot by any stretch of the imagination be distributed by "flotsam and jetsam" methods; and further that the element of the vast extent of geologic time does not in any way affect the probability of such dispersal, since it cannot be supposed ever to occur.
Again, on page 187, we read that Austromalaya is the debatable ground between the Oriental and the very distinct Australian region; but that the consensus of opinion classes it by preference with the Australian. It includes, we are told, Celebes, the Moluccas, Timor and the smaller islands and is separated from the Oriental region by "Wallace's Line." This is surely a step backward, for "Wallace's Line" marks the limit of but a small fraction of the whole species total of the Indonesian fauna, while the area from the Lesser Sunda Islands and Celebes on the one hand to Papuasia on the other represents a great transition zone, where a dominance of Malayan types may be found in the western part which merges into a predominance of Australian types in Papua. There is no real boundary line in the entire area and no reason to expect one.
Again Dr. Matthew in his "Summary of Evidence" (p. 308) states that "the continental and oceanic areas are now maintained at their different levels chiefly through isostatic balance and it is difficult to believe that they could formerly have been reversed in any extensive degree." [[p. 3]] Then, on page 309: "A rise of 100 fathoms would unite all the continents and continental islands, except perhaps Australia, into a single mass, but would leave Antarctica, New Zealand, Madagascar, Cuba and many smaller islands separate." These four areas Matthew believes to have been always isolated islands, and if we can show a probability that any one of them was continental, we can at least make more reasonable a proposition that they all were once united to some other continental land. This point will be returned to later on.
Now a word regarding isostasy. There is hardly a principle in geology concerning which there is greater uncertainty among geologists than the matter of isostatic balance. Only one thing is sure, isostasy must meet and conform to known or presumably known facts, and the fact that fundamental changes have taken place in the form of the earth's surface in recent geologic time is not to be denied. Such features as the Great Rift Valley of Africa and its continuation, the Red Sea and the Dead Sea, the Black Sea, the Basin of the Mediterranean, are held now by geologists to be the results of nothing but gigantic and not at all ancient down-thrown fault-blocks. For other examples of changes of land and sea level with relation to each other, the Valley of the Po and the Central Valley of California are good evidence. The argument of isostatic balance may probably be held to control the conditions in the Pacific Basin as a whole, but isostasy cannot be used effectively as an argument in a relatively small area anywhere. Professor R. A. Daly tells me that there is clear evidence of the fragmentation of a great land mass, including the Fiji Islands and New Caledonia, but that there is no evidence known at present of such a condition outside of a line joining Yap, in the Caroline Islands, the Fijis, Kermadecs and New Zealand. Besides this radiolarian ooze has long been known from Barbadoes, Trinidad, Aruba, Buen Ayre and Curaçao, supposedly only to be derived from the deep sea, but the origin of this series of deposits has been somewhat in dispute. Two recent papers by Dr. G. A. F. Molengraff, however, describe deposits of which there can hardly be any question whatever; one is "On Oceanic Deep Sea Deposits in Central Borneo,"3 while the other is entitled "Over mangaan Knollen in mesozoischen diepzeeafzettingen van Borneo, Timor en Rotti, hun beteekenis en hun wijzer van Opstaan."4 These papers show that on the islands of Borneo, Timor and Rotti, at an elevation of about 4000 feet, very extensive deposits occur which a microscopical examination shows to be composed of radiolaria, together with the manganese nodules so characteristic of the deep sea. In other words, Molengraff [[p. 4]] has found an extensive area of deep sea floor raised to 4000 feet above the present sea level. On the southeast coast of Africa, W. M. Davis noticed the truncation by the present shore line of extensive concentric terraces, traceable far inland, which could only mean the down-faulting of a gigantic block of material to bring the shore line into its present state. It will be said at once that some of these changes of level have taken place in zones known to be in incomplete isostatic adjustment, but this is a matter of no moment whatsoever in comparison with the fact that change of level may be found to have occurred in the very areas where the islands under discussion are to be found. Celebes does not lie upon the continental shelf and yet the island has an obviously continental fauna, and Dr. Matthew has told me himself that Celebes has been a source of no small worry to him. Cuba has similarly a large fauna, derived from the American continent, although it does not lie upon the continental shelf. Vaughan, a thoroughly conservative observer, believes (in litt.) that Cuba was quite possibly separated, by the down-faulting of blocks of material, from both Haiti and the mainland. Dr. Matthew (in litt.) says: "The fault block theory is of course a very familiar one; its application to continental movements is undoubtedly extensive, although it is just now somewhat of a fetish among stratigraphers, as folds were fifty years ago. But on land the great fault blocks are largely compensated by erosion, so that they do not involve so extensive a displacement of adjoining surfaces as one might at first suppose. Their application to explain submarine conditions where such compensation does not occur brings them into an apparent conflict with isostatic adjustments. Considering that we cannot possibly prove their responsibility for the sudden changes from shallow sea to abyssal depths in any case, I am inclined to avoid hasty ascription to such features as block-faulting. I have passed beyond the stage of immaturity when one is unreasonably certain about things." I can only add that I am as far from being unreasonably certain regarding isostatic adjustments in general as Dr. Matthew is regarding marine down faults. To the zoölogist these geologic problems seem so differently interpreted by different and equally gifted and trustworthy students that one is inclined to relegate them all to the limbo of where "you pay your money and take your choice."
Vastly different, however, is the matter of the zoölogic evidence presented by the faunas of some islands as indicative of the island's geologic or geographic history. Dr. Matthew lays great stress upon the importance of the mammalian element in the fauna. Here a word of caution is not amiss, for mammals act queerly upon islands and often have a way of being most strangely absent, as this is the group which has greatest [[p. 5]] difficulty in surviving in a limited area. Trinidad, a large island separated by a very narrow strait from Venezuela, has a reasonably full quota of mammalian inhabitants, while the large and heavily wooded island of San Miguel, just off the coast of Panama, has but a few small mammals, quite a contrast to Coiba Island, farther north off Honduras, where even a peculiar deer is known still to occur. Gorgona Island, off Colombia, also with luxuriant vegetation, has a peculiar Cebus (Cebus curtus Bangs), a peculiar Proechimys, and so far as known no other mammals. Yet these differences are all among islands on the shelf and near or fairly near the shore; and I could multiply the examples!
Now I do not believe, with Matthew, that the Antilles are oceanic islands--islands which have received their fauna by fortuitous transportal. My reasons for thinking as I do are these: First, I believe that the islands of the Antillean chain have too evenly distributed and homogeneous a fauna for it all to have been fortuitously derived; secondly, I consider the fauna to be composed of too many different animal phyla; and thirdly I believe that many of these elements are not of a nature to have withstood "flotsam or jetsam" dispersal. We must now consider Matthew's exposition of the natural raft hypothesis (p. 206 et seq.). He states: "1) Natural rafts have been several times reported as seen over a hundred miles off the mouths of the great tropical rivers such as the Ganges, Amazon, Congo and Orinoco. For one such raft observed, a hundred have probably drifted out that far unseen or unrecorded before breaking up." This is obvious and undoubted. But, and this is most important, these rafts, even the very large ones, float low in the water; they soon become soaked with salt water in a calm sea, rippled over or broken over if the sea be choppy or rough as it is in the trade wind or monsoon belts. Only organisms or their eggs which are encapsulated or otherwise naturally resistant can withstand these conditions. Molluscs are stimulated to activity by dampening, but most are killed by salt water--although some such as Cerion are resistant. Scincs and Gekkos show by their distribution that they may be carried about in this way. Amphibians, amphisbænians, naked gastropods, earthworms, fresh-water fishes or crayfishes, Peripatus and a host of such delicate creatures simply cannot withstand salt water. No such creatures have ever been observed upon any raft, of the very few recorded, and to transport cyprinodonts, ampullarias and the host of other fresh-water types one meets with in Cuban ponds, for instance, the raft would have to include a puddle, at least, of fresh water. Supposing that an amphisbænian, to take a good example, withstood an ocean voyage upon a raft, how would the landing take place? The raft would have to make a haven and then ground in [[p. 6]] such a way that a very delicate, blind and legless lizard would be enabled to reach a suitable environment on shore. I only ask the reader to tramp West Indian shores with this in mind. If a pair or a gravid female did not make the voyage the process would have to be repeated promptly. Now consider the number of rafts each of which would have to carry an amphibian or a pair of them and which would have to start on their journeys before one would reach shore so as to permit a landing such as I have indicated. Think of the number broken up at sea, and the still greater number broken up on a tropic beach--where the sun would instantly kill crawling amphisbænians--and we see at once how excessively improbable is a single occurrence such as this. But the important point is that five West Indian Islands support peculiar amphisbænians; two species occur upon Cuba, two others related to these two on Haiti and two others similarly related to the Haitian types on Porto Rico, while but one type is as yet known upon St. Thomas and Sta. Cruz. To account for the presence of these creatures, then, eight practically inconceivable voyages must be postulated, and I have only cited one improbability out of many hundred necessary to derive all the organisms practically or wholly incapable of such sea travel and which are found in the Greater Antilles. A few such cases as Amphisbæna settle the status of the greater Antillean fauna to my mind absolutely, paucity of mammals and possibly disputable geologic evidence to the contrary notwithstanding.
Matthew's second premise is: "2) The time of such observation of rafts covers about three centuries (I set aside the period of rare and occasional exploring voyages). The duration of Cenozoic time may be assumed at three million years (Walcott's estimate)." But is it not true that this multiplication of time or any other, of course, affects only the number of rafts and does not in any way alter the resistance to raft conditions of the creatures which I have already chosen above as examples or the possibility of their being able to swim. It really carries no weight in this connection.
Matthew's third point: "3) Living mammals have been occasionally observed in such records of natural rafts. Assume the chance of their occurrence (much greater than of their presence being noted) at one in a hundred." We readily agree to the assumption and know that during the few years of human observation rafting mammals have been observed. It occurs to us, however, that multiplying the three hundred years' time of human voyages by the ten thousand necessary to occupy even the short Cenozoic period and then with this condition met, we find mammals infinitely rarer upon all islands than they should be if rafted according to Matthew's postulate. Similarly we find many of the reptiles most [[p. 7]] capable of withstanding transport by rafts conspicuously absent in the West Indies; for example, Basiliscus, almost semiaquatic, is absent in the Antilles, as Varanus is absent upon Madagascar: they were possibly derived in common with many other of the islands' original continental inhabitants and have failed to survive. We know that some edentates and a few rodents have become extinct in the West Indies, some within the last half century; why, in some cases, we cannot guess; probably more have gone the same way, of which we have unfortunately no fossil remains. So also the boa-constrictor, which has been observed to be carried once to St. Vincent from the Orinoco and which has probably come comparatively often, has never succeeded in establishing itself. Why some types fail to survive upon islands while others of apparently similar habit flourish is one of the enigmas for which no mite of answer has appeared. Matthew's arguments, which he numbers 4 and 5, may be considered together: "4) Three hundred miles drift would readily reach any of the larger oceanic islands except New Zealand. Assume as one in ten the probability that the raft drifted in such a direction as to reach dry land within three hundred miles. 5) In case such animals reached the island shores and the environment afforded them a favorable opening, the propagation of the race would require either two individuals of different sexes or a gravid female. Assume the probability of any of the passengers surviving the dangers of landing as one in three (by being drawn in at the mouth of some tidal river or protected inlet), of landing at a point where the environment was sufficiently favorable as one in ten, the chances of two individuals of different sexes being together might be assumed to be one in ten, the alternate of a gravid female as one in five. The chance of the two happening would be 1/10 + 1/5 = 3/10. The chance of the species obtaining a foothold would then be 3/10 + 1/3 + 1/10 = 1/100." He then continues, "If then we allow that ten such cases of natural rafts far out at sea have been reported, we may concede that 1000 have probably occurred in three centuries and 30,000,000 during the Cenozoic. Of these rafts, only 3,000,000 will have had living mammals upon them; of these only 30,000 will have reached land, and in only 300 of these cases will the species have established a foothold. This is quite sufficient to cover the dozen or two cases of Mammalia on the larger oceanic islands.
"I have considered the case only in relation to small mammals. With reptiles and invertebrates, the probabilities in the case vary widely in different groups, but in almost every instance they would be considerably greater than with mammals. The chance for transportation and survival would be larger and the geologic time limit in many instances much longer. Wind, birds, small floating drift and other methods of accidental [[p. 8]] transportation may have played a more important part with invertebrates, although they cannot be invoked to account for the distribution of vertebrates. The much larger variety and wider distribution on infra-mammalian life in oceanic islands is thus quite to be expected. And the extent and limits of such distribution are in obviously direct accord with the opportunities for over-sea transportation in different groups."
In the estimate which Matthew has made there seems to be an obvious error, for should we postulate 1000 rafts in 300 years we would have 10,000,000 rafts during the 3,000,000 years of Cenozoic time, not 30,000,000 as Matthew has it, and the chances of the whole concatenation of events are reduced by 66 per cent. But even this reduced estimate would if true bring more mammals to most islands than we find. Let us, however, for the sake of argument, admit that some mammals might be transported in this way, is the premise true that other creatures will be more easily carried? Some will, and these types by their haphazard occurrence can now be recognized easily; others most certainly will not. Matthew has not realized the enormous sum total of different species which go to make up the fauna of such islands as Cuba and Haiti. Such a vast number of species would require squadrons of rafts at frequent intervals, even if only ancestral stocks were transported from which many species arose after coming to the island by some sort of adaptive radiation. Another important point has also been missed. Almost all of these isolated groups of individuals have grown to be well differentiated island species. Distinct from the related forms of the mainland and neighboring islands, they represent types evolved in complete isolation; an occasional raft bearing individuals from the parent stock would by preventing breeding in, at least in some cases, prevent speciation by isolation taking place.
Let us for a moment consider the Antillean chain as a whole; it is utterly impossible that ocean currents could now or in the past have brought rafts with equal frequency to all parts of this island arc, and yet the same types reappear upon island after island all the way from Cuba to Grenada. Rafting from island to island could certainly not have occurred, since there could never have been large rivers on them had they always retained their present size. The fauna is far larger in number of species upon the Greater Antilles than upon the Lesser, as the conditions favorable for the survival of species are obviously better upon the large islands with their luxuriant vegetation than upon such barren islets as, for example, Sombrero or Redonda. The types, however, which have been able to survive upon Sombrero or Saba are just those which are found, along with many others, upon Cuba or Haiti. In my "Herpetology of [[p. 9]] Jamaica" and "Zoögeography of West Indian Reptiles," I have gone into this matter in detail and there is no need of repeating what has been said there. This homogeneity of the fauna is the best possible proof that winds (tornados, hurricanes, etc.), birds, small floating drift, etc., have played no considerable part in populating the island by carrying eggs or adults, since it is inconceivable that by these means the same improbable choice of passengers would be carried to so many islands.
Matthew, upon the basis of the mammalian fauna of Madagascar entirely, he believes, derivable from a few waifs, and from the fact that the island is not upon the present continental shelf, concludes that it is an oceanic island. We may grant that all the lemurs have radiated from a single type, and this may have been a waif type--all this for the sake of argument--but what does the rest of the fauna show? We find abundant amphibians of many different families, as well as a great host of other land and fresh water organisms which cannot by any stretch of the imagination be considered as more probably capable of surviving raft transport than mammals, nor in very many cases of possibly surviving such transportation at all. Yet such types as these are most abundant upon Madagascar, in individuals and in species--species representing wholly unrelated mainland stocks and not those which might possibly have arisen after coming to the island.
My friend Dr. G. M. Allen has contributed the following note regarding Madagascar which is interesting in this connection. He writes me: "The total absence from Madagascar of any native species of the typical Murinæ seems to be a striking bit of negative evidence against a chance population of the island. All the nine genera of indigenous rodents are Cricetine in their affinity, though now considered to represent a special subfamily by themselves--Nesomyinæ. The Cricetine-like rodents are abundant still in the Americas, less so in number of species in northern Eurasia. The African Lophiomys is nearly related. If we consider the more specialized typical Murinæ as representing a later development of the Muridæ, it is easy to account for their absence from America, if formerly, as now, their northward range did not extend to the East Siberian region, whence they could have crossed by land bridges if such existed. That no member of so widespread and successful a type in the Old World as Mus (in the broad sense) has reached Madagascar, it seems evident that it is because none have been able to cross the intervening water. If nine distinct genera of Cricetine-like rats or their ancestor or ancestors could have reached Madagascar by chance methods, it seems inconceivable that no single Murine could have done so, despite the great adaptability and abundance of the representatives of this group. The most attractive [[p. 10]] explanation of this fact is that the more primitive Cricetine-like rodents reached Madagascar by land connection from Africa and that they were subsequently isolated there before the advent of the more specialized and successful Murinæ, which have now totally replaced them on the mainland."
So much for the question of rafting--some creatures can be carried and some--many more--cannot survive such conditions. We may recognize by their haphazard distribution and by their habits in the field those waifs which can withstand raft transport. Yet even these resistant types are very often strangely absent; there are no Varanids in Madagascar, and yet we should naturally suppose that they would be among the very first immigrants by raft carriage.
This whole question is really but a side issue with Matthew. He is far more absorbed in other problems; hence it is only fair to say that this island question is of secondary interest to him. Dr. Matthew's masterly handling of his chapters dealing with mammals is beyond praise. He has surely shown that the present distribution of most if not all of the recent mammal groups may be plausibly explained without having recourse to postulating extensive changes of land forms. But Matthew deals with some other matters as one without authority, and one feels that his opinions would be different had he seen and not merely read about the rafts, and the landings of the rafts, of which he must perforce write to explain his ideas.
Many will notice trifling inaccuracies in the text, such as the statement that the large ground-birds of modern times are "to-day peculiarly inhabitants of arid regions." There is the New Zealand Kiwi in the rank damp fern forest and the host of different cassowaries in Ceram, the Papuan Islands, New Britain, Queensland, and anyone who has ever tried to hunt cassowary knows how well they are adapted to getting about in the densest jungle in the world. However, such points are of so small import that it is hardly worth while mentioning them. My final word is not to advise but to adjure everyone who aims at a wider knowledge of natural history to read Dr. Matthew's paper.
1. Manuscript received by the Editor
22 October, 1915. [[on p. 1]]