FOOTNOTES:
[22] Hæmulon.
[23] Calamus.
[24] Clarion Island and Socorro Island.
[25] A few Mexican shore fishes, Chætodon humeralis, Galeichthys dasycephalus, Hypsoblennius parvipinnis, have been wrongly accredited to Hawaii by some misplacement of labels.
[26] The dragonets (Callionymus) are shore fishes of the shallowest waters in Europe and Asia, but inhabit considerable depths in tropical America. The sea-robins (Prionotus) are shore fishes in Massachusetts, semi-bathybial fishes at Panama. Often Arctic shore fishes become semi-bathybial in the Temperate Zone, living in water of a given temperature. A long period of cold weather will sometimes bring such to the surface.
[27] "Introduction to the Study of Fishes."
[28] Cyprinidæ.
[29] "A Comparison of Antipodal Faunæ," 1887.
[30] Galaxias, Neochanna, Prototroctes, and Retropinna.
[31] The shovel-nosed sturgeon (Scaphirynchus and Kessleria) and the paddle-fish (Polyodon and Psephurus).
[32] Esox lucius.
[33] Umrba, the mud-minnow.
[34] Leuciscus hakuensis.
[35] Eleotris fusca.
[36] Awaous genivittatus.
[CHAPTER XV.]
ISTHMUS BARRIERS SEPARATING FISH FAUNAS
The Isthmus of Suez.—In the study of the effect of the Isthmus of Suez on the distribution of fishes we may first consider the alleged resemblance between the fauna of the Mediterranean and that of Japan. Dr. Günther claims that the actual identity of genera and species in these two regions is such as to necessitate the hypothesis that they have been in recent times joined by a continuous shore-line. This shore-line, according to Prof. A. Ortmann and others, was not across the Isthmus of Suez, but farther to the northward, probably across Siberia.
The Fish Fauna of Japan.—For a better understanding of the problem we may give a brief analysis of the fish fauna of Japan.
The group of islands which constitute the empire of Japan is remarkable for the richness of its animal life. Its variety in climatic and other conditions, its nearness to the great continent of Asia and to the chief center of marine life, the East Indian Islands, its relation to the warm Black Current or Kuro Shiwo from the south and to the cold currents from the north, all tend to give variety and richness to the fauna of its seas. Especially is this true in the group of fishes. In spite of the political isolation of the Japanese Empire, this fact has been long recognized and the characteristic types of Japanese fishes have been well known to naturalists.
At present about 900 species of fishes are known from the four great islands which constitute Japan proper—Hondo, Hokkaido, Kiusiu, and Shikoku. About 200 others are known from the volcanic islands to the north and south. Of these 1100 species, about fifty belong to the fresh waters. These are all closely allied to forms found on the mainland of Asia, from which region all of them were probably derived. In general the same genera appear in China and with a larger range of species.
Fresh-water Faunas of Japan.—Two faunal areas of fresh waters may be fairly distinguished, although broadly overlapping. The northern region includes the island of Hokkaido and the middle and northern part of the great island of Hondo. In a rough way, its southern boundary may be defined by Fuji Yama, and the Bay of Matsushima. It is characterized by the presence of salmon, trout, and sculpins, and northward by sturgeon and brook lampreys. The southern area loses by degrees the trout and other northern fishes, while in its clear waters abound various minnows, gobies, and the famous ayu, or Japanese dwarf salmon, one of the most delicate of food fishes. Sculpins and lampreys give place to minnows, loaches, and chubs. Two genera, a sculpin[37] and a perch,[38] besides certain minnows and catfishes, are confined to this region and seem to have originated in it, but, like the other species, from Chinese stock.
Origin of Japanese Fresh-water Fishes.—The question of the origin of the Japanese river fauna seems very simple. All the types are Asiatic. While most of the Japanese species are distinct, their ancestors must have been estrays from the mainland. To what extent river fishes may be carried from place to place by currents of salt water has never been ascertained. One of the most widely distributed of Japanese river fishes is the large hakone dace or chub.[39] This has been repeatedly taken by us in the sea at a distance from any stream. It would evidently survive a long journey in salt water. An allied species[40] is found in the midway island of Tsushima, between Korea and Japan.
Faunal Areas of Marine Fishes in Japan.—The distribution of the marine fishes of Japan is mainly controlled by the temperature of the waters and the motion of the ocean currents. Five faunal areas may be more or less clearly recognized, and these may receive names indicating their scope—Kurile, Hokkaido, Nippon, Kiusiu, Kuro Shiwo, and Riu Kiu. The first or Kurile district is frankly subarctic, containing species characteristic of the Ochotsk Sea on the one hand, and of Alaska on the other. The second or Hokkaido[41] district includes this northern island and that part of the shore of the main island of Hondo[42] which lies to the north of Matsushima and Noto. Here the cold northern currents favor the development of a northern fauna. The herring and the salmon occupy here the same economic relation as in Norway, Scotland, Newfoundland, and British Columbia. Sculpins, blennies, rockfish, and flounders abound of the rocky shores and are seen in all the markets.
South of Matsushima Bay and through the Island Sea as far as Kobe, the Nippon fauna is distinctly one of the temperate zone. Most of the types characteristically Japanese belong here, abounding in the sandy bays and about the rocky islands.
About the islands of Kiusiu and Shikoku, the semi-tropical elements increase in number and the Kiusiu fauna is less characteristically Japanese, having much in common with the neighboring shores of China, while some of the species range northward from India and Java. But these faunal districts have no sharp barriers. Northern fishes[43] unquestionably of Alaskan origin range as far south as Nagasaki, while certain semi-tropical[44] types extend their range northward to Hakodate and Volcano Bay. The Inland Sea, which in a sense bounds the southern fauna, serves at the same time as a means of its extension. While each species has a fairly definite northern or southern limit, the boundaries of a faunal district as a whole must be stated in the most general terms.
The well-known boundary called Blackiston's Line, which passes through the Straits of Tsugaru, between the two great islands of Hondo and Hokkaido, marks the northern boundary of monkeys, pheasants, and most tropical and semi-tropical birds and mammals of Japan. But as to the fishes, either marine or fresh water, this line has no significance. The northern fresh-water species probably readily cross it; the southern rarely reach it.
We may define as a fourth faunal area that of the Kuro
Shiwo district itself, which is distinctly tropical and contrasts strongly with that of the inshore bays behind it. This warm "Black Current," analogous to our Gulf Stream, has its origin in part from a return current from the east which passes westward through Hawaii, in part from a current which passes between Celebes and New Guinea. It moves northward by way of Luzon and Formosa, touching the east shores of the Japanese islands Kiusiu and Shikoku, to the main island of Hondo, flooding the bays of Kagoshima and Kochi, of Waka, Suruga, and Sagami. The projecting headlands reach out into it and the fauna of their rock-pools is distinctly tropical as far to the northward as Tokio.
Fig. 178.—Sacramento Perch, Archoplites interruptus Girard. Family Centrarchidæ. Sacramento River.
These promontories of Hondo, Waka, Ise, Izu, Misaki, and Awa have essentially the same types of fishes as are found on the reefs of tropical Polynesia. The warmth of the off-shore currents gives the fauna of Misaki its astonishing richness, and the wealth of life is by no means confined to the fishes. Corals, crustaceans, worms, and mollusks show the same generous profusion of species.
A fifth faunal area, closely related to that of the Black Current, is formed by the volcanic and coral reefs of the Riu Kiu Archipelago. This fauna, so far as known, is essentially East Indian, the genera and most of the species being entirely identical with those of the islands about Java and Celebes.
Resemblance of the Japanese and Mediterranean Fish Faunas.—It has been noted by Dr. Günther that the fish fauna of Japan bears a marked resemblance to that of the Mediterranean. This likeness is shown in the actual identity of genera and species, and in their relation to each other. This resemblance he proposes to explain by the hypothesis that at some recent period the two regions, Japan and the Mediterranean, have been united by a continuous shore-line. The far-reaching character of this hypothesis demands a careful examination of the data on which it rests.
The resemblance of the two faunal areas, so far as fishes are concerned, may be stated as follows: There are certain genera[45] of shore fishes, tropical or semi-tropical, common to the Mediterranean and Japan, and wanting to California, Panama, and the West Indies, and in most cases to Polynesia also. Besides these, certain others found in deeper water (100 to 200 fathoms) are common to the two areas,[46] and have been rarely taken elsewhere.
Significance of Resemblance.—The significance of these facts can be shown only by a fuller analysis of the fauna in question, and those of other tropical and semi-tropical waters. If the resemblances are merely casual, or if the resemblances are shown by other regions, the hypothesis of shore continuity would be unnecessary or untenable. It is tenable if the resemblances are so great as to be accounted for in no other way.
Of the genera regarded as common, only two[47] or three are represented in the two regions by identical species, and these have a very wide distribution in the warm seas. Of the others, nearly all range to India, to the Cape of Good Hope, to Australia, or to Brazil. They may have ranged farther in the past; they may even range farther at present. Not one is confined to the two districts in question. As equally great resemblances exist between Japan and Australia or Japan and the West Indies, the case is not self-evident without fuller comparison. I shall therefore undertake a somewhat fuller analysis of the evidence bearing on this and similar problems with a view to the conclusions which may be legitimately drawn from the facts of fish distribution.
Differences between Japanese and Mediterranean Fish Faunas.—We may first, after admitting the alleged resemblances and others, note that differences are equally marked. In each region are a certain number of genera which we may consider as autochthonous. These genera are represented by many species or by many individuals in the region of their supposed origin, but are more scantily developed elsewhere. Such genera in Mediterranean waters are Crenilabrus, Labrus, Spicara, Pagellus, Mullus, Boops, Spondyliosoma, Oblata. None of these occurs in Japan, nor have they any near relatives there. Japanese autochthonous types, as Pseudoblennius, Vellitor, Duymæria, Anoplus, Histiopterus, Monocentrus, Oplegnathus, Plecoglossus, range southward to the Indies or to Australia, but all of them are totally unknown to the Mediterranean. The multifarious genera of Gobies of Japan show very little resemblance to the Mediterranean fishes of this family, while blennies, labroids, scaroids, and scorpænoids are equally diverse in their forms and alliances. To the same extent that likeness in faunas is produced by continuity of means of dispersion is it true that unlikeness is due to breaks in continuity. Such a break in continuity of coast-line, in the present case, is the Isthmus of Suez, and the unlikeness in the faunas is about what we might conceive that such a barrier should produce.
Sources of Faunal Resemblances.—There are two main sources of faunal resemblances: first, the absence of any barriers permitting the actual mingling of the species; second, the likeness of temperature and shore configuration on either side of an imperfect barrier. Absolute barriers do not exist and apparently never have existed in the sea. If the fish faunas of different regions have mingled in recent times, the fact would be shown by the presence of the same species in each region. If the union were of a remote date, the species would be changed, but the genera might remain identical.
In case of close physical resemblances in different regions, as in the East Indies and West Indies, like conditions would favor the final lodgement of like types, but the resemblance would be general, the genera and species being unlike. Without doubt part of the resemblance between Japan and the Mediterranean is due to similarity of temperature and shores. Is that which remains sufficient to demand the hypothesis of a former shore-line connection?
Effects of Direction of Shore-line.—We may first note that a continuous shore-line produces a mingling of fish faunas only when not interrupted by barriers due to climate. A north and south coast-line, like that of the East Pacific, however unbroken, permits great faunal differences. It is crossed by the different zones of temperature. An east and west shore-line lies in the same temperature. In all cases of the kind which now exist on the earth (the Mediterranean, the Gulf of Mexico, the Caribbean Sea, the shores of India), even species will extend their range as far as the shore-line goes. The obvious reason is because such a shore-line rarely offers any important barrier to distribution, checking dispersion of species. We may, therefore, consider the age and nature of the Isthmus of Suez and the character of the faunas it separates.
Numbers of Genera in Different Faunas.—For our purposes the genera must be rigidly defined, a separate name being used in case of each definable difference in structure. The wide-ranging genera of the earlier systematists were practically cosmopolitan, and their geographical distribution teaches us little. On the other hand, when we come to the study of geological distribution, the broad definition of the genus is the only one usually available. The fossil specimens are always defective. Minor characters may be lost past even the possibility of a guess, and only along broad lines can we achieve the classification of the individual fossil.
Using the modern definition of genus, we find in Japan 483 genera of marine fishes; in the Red Sea, 225; in the Mediterranean, 231. In New Zealand 150 are recorded; in Hawaii, 171; 357 from the West Indies, 187 from the Pacific coast of tropical America, 300 from India, 450 from the East-Indian islands, and 227 from Australia.
Of the 483 genera ascribed to Japan, 156 are common to the Mediterranean also, 188 to the West Indies and Japan, 169 to the Pacific coast of the United States and Mexico. With Hawaii Japan shares 90 genera, with New Zealand 62; 204 are common to Japan and India, 148 to Japan and the Red Sea, most of these being found in India also. Two hundred genera are common to Japan and Australia.
From this it is evident that Japan and the Mediterranean have much in common, but apparently not more than Japan shares with other tropical regions. Japan naturally shows most likeness to India, and next to this to the Red Sea. Proportionately less is the resemblance to Australia, and the likeness to the Mediterranean seems much the same as that to the West Indies or to the Pacific coast of America.
But, to make these comparisons just and effective, we should consider not the fish fauna as a whole; we should limit our discussion solely to the forms of equatorial origin. From the fauna of Japan we may eliminate all the genera of Alaskan-Aleutian origin, as these could not be found in the other regions under comparison. We should eliminate all pelagic and all deep-sea forms, for the laws which govern the distribution of these are very different from those controlling the shore fishes, and most of the genera have reached a kind of equilibrium over the world.
Significance of Rare Forms.—We may note also, as a source of confusion in our investigation, that numerous forms found in Japan and elsewhere are very rarely taken, and their real distribution is unknown. Some of these will be found to have, in some unexpected quarter, their real center of dispersion. In fact, since these pages were written, I have taken in Hawaii representatives of three[48] genera which I had enumerated as belonging chiefly to Japan and the West Indies. Numerous other genera common to the two regions have since been obtained by Dr. Gilbert. Such species may inhabit oceanic plateaus, and find many halting places in their circuit of the tropical oceans. We have already discovered that Madeira, St. Helena, Ascension, and other volcanic islands constitute such halting places. We shall find many more such, when the deeper shore regions are explored, the region between market-fishing and the deep-sea dredgings of the Challenger and the Albatross. In some cases, no doubt, these forms are verging on extinction and a former wide distribution has given place to isolated colonies.
The following table shows the contents, so far as genera are concerned, of those equatorial areas in which trustworthy catalogues of species are accessible. It includes only those fishes of stationary habit living in less than 200 fathoms. It goes without saying that considerable latitude must be given to these figures, to allow for errors, omissions, uncertainties, and differences of opinion.
Distribution of Shore Fishes.—
| A. Japan and the Mediterranean. | |
|---|---|
| Genera[49] chiefly confined to these regions | 2 |
| Genera of wide distribution | 77 |
| Total of common genera | 79 |
| Total in both regions | 399 |
| Genera above included, found in all equatorial regions | 55 |
| Genera[50] found in most equatorial regions | 11 |
| Genera more or less restricted | 13 |
| 79 | |
| B. Japan and the Red Sea. | |
| Genera[51] chiefly confined to these two regions | 2 |
| Genera of wide distribution | 109 |
| Total genera common | 111 |
| Total in both regions | 424 |
| C. Japan and Hawaii. | |
| Genera chiefly confined to these regions | 3 |
| Genera of wide distribution | 79 |
| Total genera common | 82 |
| Total in both regions | 396 |
| D. Japan and Australia. | |
| Genera chiefly confined to these regions | 13 |
| Genera of wide distribution (chiefly East Indian) | 122 |
| Total genera common | 135 |
| Total in both regions | 533 |
| E. Japan and Panama. | |
| Genera chiefly confined to these regions | 2 |
| Genera of wide distribution | 89 |
| Total genera common | 91 |
| Total in both regions | 499 |
| F. Japan and the West Indies. | |
| Genera chiefly confined to these regions | 5 |
| Genera of wide distribution | 108 |
| Total genera common | 113 |
| Total in both regions | 520 |
| G. The Mediterranean and the Red Sea. | |
| Genera confined to the Suez region | 0 |
| Genera of wide distribution (chiefly Indian) | 40 |
| Total genera common | 40 |
| Total in both regions | 295 |
| H. West Indies and the Mediterranean. | |
| Genera chiefly confined to the equatorial Atlantic | 11 |
| Genera of wide distribution | 59 |
| Total | 70 |
| Total in both regions | 373 |
| I. West Indies and Panama. | |
| Genera chiefly confined to equatorial America | 68 |
| Genera of wide distribution | 101 |
| Total genera common | 169 |
| Total in equatorial America | 376 |
| J. Hawaii and Panama. | |
| Genera chiefly confined to the regions in question | 3 |
| Genera of wide distribution | 74 |
| Total genera common | 77 |
| Total in both regions | 323 |
| K. Hawaii and the East Indies. | |
| Genera chiefly confined to Hawaii | 4 |
| Genera of wide distribution in the equatorial Pacific | 123 |
| Genera confined to Hawaii and the West Indies | 1 |
| Summary. | |
| Genera (shore fishes only) in the Mediterranean Sea. | 144 |
| Genera in the Red Sea | 191 |
| Genera in India | 280 |
| Genera in Japan (exclusive of northern forms) | 334 |
| Genera in Australia | 344 |
| Genera in New Zealand | 108 |
| Genera in Hawaii | 144 |
| Genera about Panama | 256 |
| Genera in West Indies | 299 |
Extension of Indian Fauna.—From the above tables it is evident that the warm-water fauna of Japan, as well as that of Hawaii, is derived from the great body of the fauna of the East Indies and Hindostan; that the fauna of the Red Sea is derived in the same way; that the fauna of the Mediterranean bears no especial resemblance to that of Japan, rather than to other elements of the East Asiatic fauna in similar conditions of temperature, and no greater than is borne by either to the West Indies; that the faunas of the sides of the Isthmus of Suez have relatively little in common, while those of the two sides of the Isthmus of Panama show large identity of genera, although few species are common to the two sides. Of the 255 genera recorded from the Panama region, 179, or over 70 per cent., are also in the West Indies, while 68, or more than 30 per cent. of the number, are limited to the two regions in question.
The Isthmus of Suez as a Barrier to Distribution.—With the aid of the above table we may examine further the relation of the fauna of Japan to that of the Mediterranean. If a continuity of shore-line once existed, it would involve the obliteration of the Isthmus. With free connection across this isthmus the fauna of the Red Sea must have been once practically the same as that of the Mediterranean. The present differences must be due to later immigrations to one or the other region, or to the extinction of species in one locality or the other, through some kind of unfitness. In neither region is there evidence of extensive immigration from the outside. The present conditions of water and temperature differ a little, but not enough to explain the difference in faunæ. The Red Sea is frankly tropical and its fauna is essentially Indian, much the same, so far as genera are concerned, as that of southern Japan. The Mediterranean is at most not more than semi-tropical and its fishes are characteristically European. Its tropical forms belong rather to Guinea than to the East Indies. With the Red Sea the Mediterranean has very little in common, not so much, for example, as has Hawaii. Forty genera of shore fishes (and only fifty of all fishes) are identical in the two regions, the Mediterranean and the Red Sea. Of those, every one is a genus of wide distribution, found in nearly all warm seas. Of shore fishes, only one genus in seven is common to the two regions. Apparently, therefore, we cannot assume a passage across the Isthmus of Suez within the lifetime of the present genera. Not one of the types alleged to be peculiar to Japan and the Mediterranean is thus far known in the Red Sea. Not one of the characteristically abundant Mediterranean types[52] crosses the Isthmus of Suez, and the distinctive Red Sea and Indian types[53] are equally wanting in the Mediterranean. The only genera which could have crossed the Isthmus are certain shallow-water or brackish-water forms, sting-rays, torpedoes, sardines, eels, and mullets, widely diffused through the East Indies and found also in the Mediterranean. The former channel, if one ever existed, had, therefore, much the same value in distribution of species as the present Suez Canal.
Geological Evidence of Submergence of the Isthmus of Suez.—Yet, from geological data, there is strong evidence that the Isthmus of Suez was submerged in relatively recent times. The recognized geological maps of the Isthmus show that a broad area of post-Pliocene or Pliocene deposits constitutes the Isthmus and separates the nummulitic hills of Suez from their fellows about thirty miles to the eastward. The northern part of the Isthmus is alluvium from the Nile, and its western part is covered with drifting sands. The Red Sea once extended farther north than now and the Mediterranean farther to the southeast. Assuming the maps to be correct, the Isthmus must have been open water in the late Pliocene or post-Pliocene times.
Admitting this as a fact, the difference in the fish fauna would seem to show that the waters over the submerged area were so shallow that the rock-loving forms did not and could not cross it. Moreover, the region was very likely overspread with silt-bearing fresh waters from the Nile. To such fishes as Chætodon, Holocentrus, Thalassoma of the Red Sea, or to Crenilabrus, Boops, and Zeus of the Mediterranean, such waters would form a barrier as effective as the sand-dunes of to-day.
Conclusions as to the Isthmus of Suez.—We are led, therefore, to these conclusions:
1. There is no evidence derivable from the fishes of the recent submergence of the Isthmus of Suez.
2. If the Isthmus was submerged in Pliocene or post-Pliocene times, the resultant channel was shallow and muddy, so that ordinary marine fishes or fishes of rock bottoms or of deep waters did not cross it.
3. It formed an open water to brackish-water fishes only.
4. The types common to Japan and the Mediterranean did not enter either region from the other by way of the Red Sea.
5. As most of these are found also in India or Australia or both, their dispersion was probably around the south coast of Africa or by the Cape of Good Hope.
6. In view of the fact that numerous East Indian genera, as Zanclus, Enoplosus, Toxotes, Ephippus, Platax, Teuthis, Acanthurus (Monoceros), Myripristis occur in the Eocene rocks of Tuscany, Syria, and Switzerland, we may well suppose that an open waterway across Africa then existed. Perhaps these forms were destroyed in European waters by a wave of glacial cold, perhaps after the Miocene. As our knowledge of the Miocene fish faunæ of Europe is still imperfect, we cannot locate accurately the period of their disappearance. About half the species found in the Eocene of Italy belong to existing genera, and these genera are almost all now represented in the Indian fauna, and those named above with others are confined to it.
The study of fishes alone furnishes no adequate basis for mapping the continental masses of Tertiary times. The known facts in regard to their distribution agree fairly with the provisional maps lately published by Dr. Ortmann (Bull. Philos. Soc., XLI). In the Eocene map (Fig. 179) the Mediterranean extends to the northward of Arabia, across to the mouth of the Ganges. This extension would account for the tropical, Eocene, and Miocene fish fauna of Southern Europe.
The Cape of Good Hope as a Barrier to Fishes.—The fishes of the Cape of Good Hope are not well enough known for close comparison with those of other regions. Enough is known of the Cape fauna to show its general relation to those of India and Australia. The Cape of Good Hope lies in the South Temperate Zone. It offers no absolutely impassable barrier to the tropical fishes from either side. It bears a closer relation to either the Red Sea or the Mediterranean than they bear to each other. It is, therefore, reasonable to conclude that the transfer of tropical shore fishes of the Old World between the Atlantic and Pacific, in recent times, has taken place mainly around the southern point of Africa. To pelagic and deep-sea fishes the Cape of Good Hope has offered no barrier whatever. To ordinary fishes it is an obstacle, but not an impassable one. This the fauna itself shows. It has, however, not been passed by many tropical species, and by these only as the result of thousands of years of struggle and point-to-point migration.
Relations of Japan to Mediterranean Explainable by Present Conditions.—We may conclude that the resemblance of the Mediterranean fish fauna to that of Japan or India is no more than might be expected, even had the present contour of the continents been permanent for the period of duration of the present genera and species. An open channel in recent times would have produced much greater resemblances than actually exist.
The Isthmus of Panama as a Barrier to Distribution.—Conditions in some regards parallel with those of the Isthmus of Suez exist in but one other region—the Isthmus of Panama. Here the first observers were very strongly impressed by the resemblance of forms. Nearly half the genera found on the two sides of this isthmus are common to both sides. Taking those of the Pacific shore for first consideration, we find that three-fourths of the genera of the Panama fauna occur in the West Indies as well.
This identity is many times greater than that existing at the Isthmus of Suez. Moreover, while the Cape of Good Hope offers no impassable barrier to distribution, the same is not true of the southern part of South America. The subarctic climate of Cape Horn has doubtless formed a complete check to the movements of tropical fishes for a vast period of geologic time.
Unlikeness of Species on the Shores of the Isthmus of Panama.—But, curiously enough, this marked resemblance is confined chiefly to the genera and does not extend to the species on the two shores.
Of 1400 species of fishes recorded from tropical America north of the Equator, only about 70 are common to the two coasts. The number of shore fishes common is still less. In this 70 are included a certain number of cosmopolitan types which might have reached either shore from the Old World.
Fig. 179.—Map of the Continents, Eocene time. (After Ortmann.)
A few others invade brackish or fresh waters and may possibly have found their way, in one way or another, across the Isthmus of Nicaragua. Of fishes strictly marine, strictly littoral, and not known from Asia or Polynesia, scarcely any species are left as common to the two sides. This seems to show that no waterway has existed across the Isthmus within the lifetime, whatever that may be, of the existing species. The close resemblance of genera shows apparently with almost equal certainty that such a waterway has existed, and within the period of existence of the groups called genera. How long a species of fish may endure unchanged no one knows, but we know that in this regard great differences must exist in different groups. Assuming that different species crossed the Isthmus of Panama in Miocene times, we should not be surprised to find that a few remain to all appearances unchanged; that a much larger number have become "representative" species, closely related forms retaining relations to the environment to those of the parent form, and, finally, that a few species have been radically altered.
This is exactly what has taken place at the Isthmus of Panama with the marine shore fishes. Curiously enough, the movement of genera seems to have been chiefly from the Atlantic to the Pacific. Certain characteristic genera[54] of the Panama region have not passed over to the Pacific. On the other hand, most of the common genera[55] show a much larger number of species on the Atlantic side. This may be held to show their Atlantic origin.
Of the relatively small number of genera which Panama has received from Polynesia[56] few have crossed the Isthmus to appear in the West Indian fauna.
Views of Earlier Writers on the Fishes of the Isthmus of Panama.—The elements of the problem at Panama may be better understood by a glance at the results of previous investigations.
In 1869 Dr. Günther, after enumerating the species examined by him from Panama, reaches the conclusion that nearly one-third of the marine fishes on the two shores of tropical America will be found to be identical. He enumerates 193 such species as found on the two coasts; 59 of these, or 31 per cent. of the total, being actually identical. From this he infers that there must have been, at a comparatively recent date, a depression of the Isthmus and intermingling of the two faunas.[57]
Catalogue of Fishes of Panama.—In an enumeration of the fishes of the Pacific coast in 1885,[58] the present writer showed that Dr. Günther's conclusions were based on inadequate data.
In my list 407 species were recorded from the Pacific coast of tropical America—twice the number enumerated by Dr. Günther. Of these 71 species, or 17½ per cent., were found also in the Atlantic. About 800 species are known from the Caribbean and adjacent shores, so that out of the total number of 1,136 species but 71, or 6 per cent. of the whole, are common to the two coasts. This number does not greatly exceed that of the species common to the West Indies and the Mediterranean, or even the West Indies and Japan. It is to be noted also that the number 71 is not very definitely ascertained, as there must be considerable difference of opinion as to the boundaries of species, and the actual identity in several cases is open to doubt.
This discrepancy arises from the comparatively limited representation of the two faunas at the disposal of Dr. Günther. He enumerates 193 marine or brackish-water species as found on the two coasts, 59 of which are regarded by him as specifically identical, this being 31 per cent. of the whole. But in 30 of these 59 cases I regard the assumption of complete identity as erroneous, so that taking the number 193 as given I would reduce the percentage to 15. But these 193 species form but a fragment of the total fauna, and any conclusion based on such narrow data is certain to be misleading.
Of the 71 identical species admitted in our list, several (e.g., Mola, Thunnus) are pelagic fishes common to most warm seas.
Still others (e.g., Trachurus, Carangus, Diodon sp.) are cosmopolitan in the tropical waters. Most of the others (e.g., Gobius, Gerres, Centropomus, Galeichthys sp., etc.) often ascend the rivers of the tropics, and we may account for their diffusion, perhaps, as we account for the dispersion of fresh-water fishes on the Isthmus, on the supposition that they may have crossed from marsh to marsh at some time in the rainy season.
In very few cases are representatives of any species from opposite sides of the Isthmus exactly alike in all respects. These differences in some cases seem worthy of specific value, giving us "representative species" on the two sides. In other cases the distinctions are very trivial, but in most cases they are appreciable, especially in fresh specimens.
Further, I expressed the belief that "fuller investigation will not increase the proportion of common species. If it does not, the two faunas show no greater resemblance than the similarity of physical conditions on the two sides would lead us to expect." This similarity causes the same types of fishes to persist on either side of the Isthmus while through isolation or otherwise these have become different as species.
This conclusion must hold so far as species are concerned, but the resemblance of the genera on the sides has a significance of its own.
In 1880[59] Dr Günther expressed his views in still stronger language, claiming a still larger proportion of the fishes of tropical America to be identical on the two sides of the continent. He concluded that "with scarcely any exceptions the genera are identical, and of the species found on the Pacific side, nearly one-half have proved to be the same as those of the Atlantic. The explanation of this fact has been found in the existence of communications between the two oceans by channels and straits which must have been open till within a recent period. The isthmus of Central America was then partially submerged, and appeared as a chain of islands similar to that of the Antilles; but as the reef-building corals flourished chiefly north and east of these islands and were absent south and west of them, reef fishes were excluded from the Pacific shores when the communications were destroyed by the upheaval of land."
Conclusions of Evermann and Jenkins.—This remark led to a further discussion of the subject on the part of Dr. B. W. Evermann and Dr. O. P. Jenkins. From their paper on the fishes of Guaymas[60] I make the following quotations:
"The explorations since 1885 have resulted (1) in an addition of about 100 species to one or other of the two faunas; (2) in showing that at least two species that were regarded as identical on the two shores[61] are probably distinct; and (3) in the addition of but two species to those common to both coasts.[62]
"All this reduces still further the percentage of common species.
"Of the 110 species obtained by us, 24, or less than 21 per cent., appear to be common to both coasts. Of these 24 species, at least 16, from their wide distribution, would need no hypothesis of a former waterway through the Isthmus to account for their presence on both sides. They are species fully able to arrive at the Pacific shores of the Americas from the warm seas west. It thus appears that not more than eight species, less than 8 per cent. of our collection, all of which are marine species, require any such hypothesis to account for their occurrence on both coasts of America. This gives us, then, 1,307 species that should properly be taken into account when considering this question, not more than 72 of which, or 5.5 per cent., seem to be identical on the two coasts. This is very different from the figures given by Dr. Günther in his 'Study of Fishes.'
"Now, if from these 72 species, admitted to be common to both coasts, we subtract the 16 species of wide distribution—so wide as to keep them from being a factor in this problem—we have left but 56 species common to the two coasts that bear very closely upon the waterway hypothesis. This is less than 4.3 per cent. of the whole number.
"But the evidence obtained from a study of other marine life of that region points to the same conclusion.
"In 1881, Dr. Paul Fischer discussed the same question in his 'Manual de Conchyliologie,' pp. 168, 169, in a section on the Molluscan Fauna of the Panamic Province, and reached the same general conclusions. He says: 'Les naturalistes Américians se sont beaucoup preéoccupés des espèces de Panama qui paraissent identiques avec celles des Antilles, ou qui sont représentatives. P. Carpenter estime qu'il en existe 35. Dans la plupart des cas, l'identite absolue n'a pu être constantée et on a trouvé quelques caractères distinctifs, ce qui n'a rien d'ètonnant, puisque dans l'hypothèse d'une origine commune, les deux races pacifique et atlantique sont séparée depuis la periode Miocène. Voici un liste de ces espèces représentatives ou identiques.' Here follows a list of 20 species. 'Mais ces formes semblables,' he says, 'constituent un infime minorité (3 per cent.).'
"These facts have a very important bearing upon certain geological questions, particularly upon the one concerning the cold of the Glacial period.
"In Dr. G. Frederick Wright's recent book, 'The Ice Age in North America,' eight different theories as to the cause of the cold are discussed. The particular theory which seems to him quite reasonable is that one which attributes the cold as due to a change of different parts of the country, and a depression of the Isthmus of Panama is one of the important changes he considers. He says: 'Should a portion of the Gulf Stream be driven through a depression across the Isthmus of Panama into the Pacific, and an equal portion be diverted from the Atlantic coast of the United States by an elevation of the sea-bottom between Florida and Cuba, the consequences would necessarily be incalculably great, so that the mere existence of such a possible cause for great changes in the distribution of moisture over the northern hemisphere is sufficient to make one hesitate before committing himself unreservedly to any other theory; at any rate, to one which has not for itself independent and adequate proof.'
"In the appendix to the same volume Mr. Warren Upham, in discussing the probable causes of glaciation, says: 'The quaternary uplifts of the Andes and Rocky Mountains and of the West Indies make it nearly certain that the Isthmus of Panama has been similarly elevated during the recent epoch.... It may be true, therefore, that the submergence of this isthmus was one of the causes of the Glacial period, the continuation of the equatorial oceanic currents westward into the Pacific having greatly diminished or wholly diverted the Gulf Stream, which carries warmth from the tropics to the northern Atlantic and northwestern Europe.'
Fig. 180.—Caulophryne jordani Goode and Bean, a deep-sea fish of the Gulf Stream. Family Ceratiidæ.
Fig. 181.—Exerpes asper Jenkins and Evermann, a fish of the rock-pools, Guaymas, Mexico. Family Blenniidæ.
"Any very recent means by which the fishes could have passed readily from one side to the other would have resulted in making the fish faunas of the two shores practically identical; but the time that has elapsed since such a waterway could have existed has been long enough to allow the fishes of the two sides to become practically distinct. That the mollusks of the two shores are almost wholly distinct, as shown by Dr. Fischer, is even stronger evidence of the remoteness of the time when the means of communication between the two oceans could have existed, for 'species' among the mollusks are probably more persistent than among fishes.
"Our present knowledge, therefore, of the fishes of tropical America justifies us in regarding the fish faunas of the two coasts as being essentially distinct, and believing that there has not been, at any comparatively recent time, any waterway through the Isthmus of Panama."
It is thus shown, I think, conclusively, that the Isthmus of Panama could not have been depressed for any great length of time in a recent geological period.
Conclusions of Dr. Hill.—These writers have not, however, considered the question of generic identity. To this we may find a clue in the geological investigations of Dr. Robert T. Hill.
In a study of "The Geological History of the Isthmus of Panama and Portions of Costa Rica," Dr. Hill uses the following language:
"By elimination we have concluded that the only period of time since the Mesozoic within which communication between the seas could have taken place is the Tertiary period, and this must be restricted to the Eocene and Oligocene epochs of that period. The paleontologic evidence upon which such an opening can be surmised at this period is the occurrence of a few California Eocene types in the Atlantic sides of the tropical American barrier, within the ranges of latitude between Galveston (Texas) and Colon, which are similar to others found in California. There are no known structural data upon which to locate the site of this passage, but we must bear in mind, however, that this structure has not been completely explored.
"Even though it was granted that the coincidence of the occurrence of a few identical forms on both sides of the tropical American region, out of the thousands which are not common, indicates a connection between the two seas, there is still an absence of any reason for placing this connection at the Isthmus of Panama, and we could just as well maintain that the locus thereof might have been at some other point in the Central American region.
"The reported fossil and living species common to both oceans are littoral forms, which indicate that if a passage existed it must have been of a shallow and ephemeral character.
"There is no evidence from either a geologic or a biologic standpoint for believing that the oceans have ever communicated across the Isthmian regions since Tertiary time. In other words, there is no evidence for these later passages which have been established upon hypothetical data, especially those of Pleistocene time.
"The numerous assertions, so frequently found in literature, that the two oceans have been frequently and recently connected across the Isthmus, and that the low passes indicative of this connection still exist, may be dismissed at once and forever and relegated to the domain of the apocryphal. A few species common to the waters of both oceans in a predominantly Caribbean fauna of the age of the Claiborne epoch of the Eocene Tertiary is the only paleontologic evidence in any time upon which such a connection may be hypothesized.
"There has been a tendency in literature to underestimate the true altitude of the isthmian passes, which, while probably not intentional, has given encouragement to those who think that this Pleistocene passage may have existed. Maack has erroneously given the pass at 186 feet. Dr. J. W. Gregory states 'that the summit of the Isthmus at one locality is 154 feet and in another 287 feet in height.' The lowest isthmian pass, which is not a summit, but a drainage col, is 287-295 feet above the ocean.
"If we could lower the isthmian region 300 feet at present, the waters of the two oceans would certainly commingle through the narrow Culebra Pass. But the Culebra Pass is clearly the headwater col of two streams, the Obispo flowing into the Chagres, and the Rio Grande flowing into the Pacific, and has been cut by fluviatile action, and not by marine erosion, out of a land mass which has existed since Miocene time. Those who attempt to establish Pleistocene interoceanic channels through this pass on account of its present low altitude must not omit from their calculations the restoration of former rock masses which have been removed by the general levelling of the surface by erosion."
Fig. 182.—Xenocys jessiæ Jordan and Bollman. Galapagos Islands. Family Lutianidæ.
In conclusion, Dr. Hill asserts that "there is considerable evidence that a land barrier in the tropical region separated the two oceans as far back in geologic history as Jurassic time, and that that barrier continued throughout the Cretaceous period. The geological structure of the Isthmus and Central American regions, so far as investigated, when considered aside from the paleontology, presents no evidence by which the former existence of a free communication of oceanic waters across the present tropical land barriers can be established. The paleontologic evidence indicates the ephemeral existence of a passage at the close of the Eocene period. All lines of inquiry—geologic, paleontologic, and biologic—give evidence that no connection has existed between the two oceans since the close of the Oligocene. This structural geology is decidedly opposed to any hypothesis by which the waters of the two oceans could have been connected across the regions in Miocene, Pliocene, Pleistocene, or recent times."
Final Hypothesis as to Panama.—If we assume the correctness of Dr. Hill's conclusions, they may accord in a remarkable degree with the actual facts of the distribution of the fishes about the Isthmus. To account for the remarkable identity of genera and divergence of species I may suggest the following hypothesis:
During the lifetime of most of the present species, the Isthmus has not been depressed. It was depressed in or before Miocene time, during the lifetime of most of the present genera. We learn from other sources that few of the extant species of fishes are older than the Pliocene. Relatively few genera go back to the Eocene, and most of the modern families appear to begin in the Eocene or later Cretaceous. In general the Miocene may be taken as the date of the origin of modern genera. The channel formed across the Isthmus was relatively shallow, excluding forms inhabiting rocky bottoms at considerable depths. It was wide enough to permit the infiltration from the Caribbean Sea of numerous species, especially of shore fishes of sandy bays, tide pools, and brackish estuaries. The currents set chiefly to the westward, favoring the transfer of Atlantic rather than Pacific types.
Fig. 183.—Channel Catfish, Ictalurus punctatus (Rafinesque). Illinois River. Family Siluridæ.
Since the date of the closing of this channel the species left on the two sides have been altered in varying degrees by the processes of natural selection and isolation. The cases of actual specific identity are few, and the date of the establishment as species, of the existing forms, is subsequent to the date of the last depression of the Isthmus.
We may be certain that none of the common genera ever found their way around Cape Horn. Most of them disappear to the southward, along the coasts of Brazil and Peru.
While local oscillations, involving changes in coast-lines, have doubtless frequently taken place and are still going on, the past and present distribution of fishes does not alone give adequate data for their investigation.
Further, it goes without saying that we have no knowledge of the period of time necessary to work specific changes in a body of species isolated in an alien sea. Nor have we any data as to the effect on a given fish fauna of the infiltration of many species and genera belonging to another. All such forces and results must be matters of inference.
The present writer does not wish to deny that great changes have taken place in the outlines of continents in relatively recent times. He would, however, insist that the theory of such changes must be confirmed by geological evidence, and evidence from groups other than fishes, and that likeness in separated fish faunas may not be conclusive.
Fig. 184.—Drawing the net on the beach of Hilo, Hawaii. Photograph by Henry W. Henshaw.