The Habitat of the Eurypterida/Chapter V

by Marjorie O'Connell

CHAPTER V

The Geological and Geographical Distribution of the Eurypterids and the Conditions of Migration

SUMMARY OF FACTS OBSERVED REGARDING THE DISTRIBUTION OF THE EURYPTERIDS

The anomalies in the geographic distribution of the eurypterids constitute one of the most difficult phases of the problem of the habitat. The facts which have been summarized in the tables on pages 37-49, and which have been discussed in various parts of the paper up to this point, clearly lead to the following generalizations: (1) There are many cases in which single individuals are found separated geologically and geographically from other known eurypterids or eurypterid-faunas. (2) The same or closely related species may occur in regions widely separated, although in the same horizon, in intermediate regions, either no eurypterids at all are found or else those which do occur are not related to those in the other localities. (3) Eurypterids are seldom found in the same chronofauna throughout the world, but appear suddenly, now in one place, now in another at different horizons, and continuous widespread faunas are entirely wanting.

(1) As illustrations of the scattered occurrence of single specimens of eurypterids may be mentioned: Strabops thacheri in the Upper Cambric, Echinognathus clevelandi from the Utica, Eurypterus prominens from the Clinton, Eurypterus boylei from the Guelph, Eurypterus microphthalmus from the Manlius, and Eurypterus douvillei from the Rothliegende.

(2) As an instance of the same species occurring in places many miles apart, Eurypterus remipes may be cited. This species has been found in Waterville, Oneida County, N. Y. in great numbers; at Jerusalem or Wheelock's Hill, Herkimer County; to the northeast (near Cedarville) and west (Paris Hill) of Jerusalem Hill, near Oriskany; at Cayuga Junction, Cayuga County; and possibly at Buffalo. In all of these localities it has been found in the uppermost part of the Bertie, but at Seneca Falls, Seneca County, specimens have been found in the Rondout Waterlime (which may be possibly of the same age as the Bertie). There are several cases of closely related species occurring in localities separated often by great distances. One example that may be cited is that of Eurypterus lacustris, E. remipes and E. fischeri. For a long time the Baltic form (E. fischeri) was identified with E. remipes and it was not until Eichwald pointed out the differences in surface sculpturing and certain other characteristics, that the species was made distinct. Clarke and Ruedemann conclude their discussion of the comparion of these two species by saying that, "Altogether, the differences are so small that Schmidt's suggestion that they are but geographical varieties is fully supported" (39, 172). They add, further, that E. remipes and E. lacustris "are more closely related to each other than either of them to E. fischeri, indicating that they had but lately separated. Their differences rest mainly in the shape of the carapace and they are duplicated by those between E. fischeri and E. laticeps, two forms associated in the same [Baltic] rocks" (39, 172). Eurypterus fischeri has been found in Oesel and in Podolia.

(3) The data on the distribution have brought out clearly the fact that at no geological horizon is there a widespread or continuous eurypterid fauna indicating passageways of migration. Even in the Upper Siluric, which marks the acme in all respects for the eurypterids, the fauna does not show that universality which would be expected of denizens of the sea or of organisms whose immediate ancestors were marine. The Bertie fauna of North America covers an area of not over 1000 square miles. The corresponding European chronofauna is found in the Baltic Isles and Russian Provinces in sediments similar in lithologic character to those in North America, but the areal extent is small and circumscribed. In the Upper Siluric of Bohemia and of Scotland the eurypterids occur within a very limited area. But in the adjoining undoubted marine formations which he in the path of migration by marine waters, the eurypterids are wanting. The graptolite fauna of the Ordovicic is known throughout the world, but the eurypterids are found only in the small area around Catskill, New York. Similarly, eurypterids are found in the Wenlock shales and limestones of Scotland, but not to the south in England, nor in other Niagaran formations at the same horizon throughout the world.

The tremendous importance of the geological and geographical distribution of the eurypterids has heretofore been overlooked except by Professor Grabau, who has dwelt upon it in the discussion of the most important occurrences, especially in North America. When the factors of distribution are considered throughout the Palaeozoic and on every continent, it will be seen that they constitute the gravest objection of all to any marine, lagoon, or estuarine theory of habitat that has been advanced. Again we must turn to a contemplation of the present, for we must believe that the laws which control the universe have always been undeviatingly constant and will always remain so. Our great difficulty in reading Earth history correctly lies in our failure to learn the laws; so much of the past appears, to our view not in the form of causes but of results. In the study of the phenomena of the present, we are usually privileged to see both the causes and the effects, and thus the opportunity is offered to ascertain the laws, although in many cases our lack of knowledge or our unreadiness, prevents us from taking advantage of this opportunity. Thus we fail to learn and to formulate the laws which are operative in every physical fact and phenomenon, visible or invisible. That man we call a master who has discerned the laws; he alone can interpret with truth the marvels of this world and of other worlds; he alone can prophecy, with a reasonable degree of certainty, the things which are to come; and he alone, if he be a geologist, can reconstruct along the lines of truth the former history of our earth. Therefore, it behooves us to become acquainted with the laws which may be studied today, before we attempt to formulate theories about the conditions which obtained in the past. If there were oceans during Palaeozoic time in which large accumulations of clastic material were forming, we are drawn to the reasonable conclusion that there were land masses from which this clastic material was derived. We must also conclude, if we view the matter rationally, that there must have been rivers on those ancient continents and that then, as now, they constituted the principal agents of transportation of material into the sea. And finally, we must believe that if there was any life in those rivers, it must have been subject to the same laws of dispersal as is the life in the rivers today. My statement does not say that because we have life in the rivers now there must have been life in the Palaeozoic rivers; that is obviously untrue. But if there was life in those rivers, then it was subject to the same laws which are operative now. It is advisable, therefore, to consider these laws and to formulate them that we may have certain definite principles for future reference.

MIGRATION AND DISPERSAL OF RECENT FLUVIATILE ORGANISMS

Of existing taxonomic groups, the fish have received more study than any other group of fluviatile organisms, and interesting as well as exceedingly pertinent data are at hand in regard to migration and dispersal of this group. Günther in his Study of Fishes, makes the general statement that: "The Freshwater fishes . . . . have been spread in circumpolar zones, and in but a limited degree from north to south. No family, much less a genus, ranges from the north to the south, whilst a number of families and genera make the entire circuit round the globe within the zone to which they belong. Not even the Cyprinoids and Siluroids, which are most characteristic of the freshwater fauna of our period, are an exception to this. Temperature and climate, indeed, are the principal factors by which the character of the freshwater fauna is determined; they form the barriers which interfere with the unlimited dispersal of the ichthyic type, much more than mountain ranges, deserts, or oceans" (97, 215).

A few illustrations of this widespread dispersal of fishes in circumpolar zones will show that the above statement is not merely theoretical. These illustrations are selected, but taken verbatim from Günther's work (97, 209–211).

A. Species Identical in Distant Continents. 1. A number of species inhabiting Europe and the temperate parts of eastern North America, as Perca fluviatilis, Gastrosteus pungitius, Lota vulgaris, Salmo salar, Esox lucius, Acipenser sturio, A. maculosus, and several Petromyzonts.

2. Lates calcifer is common in India as well as in Queensland.

3. Galaxias attenuatus inhabits Tasmania, New Zealand, the Falkland Islands, and the South American continent.

B. Genera Identical in Distant Continents. 1. The genus Umbra, so peculiar a form as to be the type of a distinct family, comprises two most closely allied species only, one of which is found in the Atlantic States of North America, the other in the river system of the Danube.

2. A very distinct genus of Sturgeons, Scaphirhynchus, consisting of two species only; one of these inhabits the fresh waters of Central Asia, the other the system of the Mississippi.

3. A second most peculiar genus of Sturgeons, Polyodon, consists likewise of two species only, one inhabiting the Mississippi, the other the Yang-tse-Kiang.

4. Amiurus, A siluroid, and Catastomus, a Cyprinoid genus, both well represented in North America, have a single species each, in temperate China.

5. Lepidosiren is represented by one species in tropical America, and by the second in tropical Africa (Protopterus).

6. Galaxias is equally represented in South Australia, New Zealand and the southern parts of South America.

C. Families Identical in Distant Continents. 1. The Labyrinthici, represented in Africa by 5, and in India by 25 species.

2. The Chromides, represented in Africa by 25, in South America by 80 species.

3. The Characinidae, represented in Africa by 35, and in South America by 226 species.

4. The Haplochitonidae, represented in southern Australia by 1, in New Zealand by 1, and in Patagonia by a third species.

The facts regarding the distribution of freshwater fish show that it is not uncommon for identical families, genera and even species to be found living in rivers on opposite sides of the world without any known relatives in the intervening rivers. There seems to be no limit to the distance which freshwater fish may migrate in the same circumpolar zone; while even mountains, deserts, or oceans, do not offer absolute barriers. It is thus easy to see that migrations which would be impossible for marine forms offer no difficulties to freshwater organisms, and localized occurrences which would be inexplicable for the former are easily understood for the latter. I do not mean to imply that migration of river forms all around the world can take place always in a single geological period. It is well known that certain related or identical species of fish which today are found in rivers thousands of miles apart have such a distribution because their ancestors in a former geological epoch when relations between land and sea were different had an opportunity to accomplish the migrations, all evidences of which have since been destroyed. In distributions observed today we see the result of migrations which may have taken place ten thousand or ten million years ago. Thus Günther observes that the present occurrence of the Dipnoi on the continents of Africa, South America and Australia is consequential upon their wide range in the Palæozoic and Mesozoic, while that of the Siluroids, which have an even greater range, is the result of their distribution during the Cenozoic. It may be well to refer here to the theory of the independent origin of specific characters, in widely dispersed organisms, which are, nevertheless, placed under similar or identical physical conditions. This theory has been especially applied to the fishes of South America by Hasemann (110), who has shown how further complications arise through the production of apparently identical though actually unrelated species in response to similar environmental complexes.

Summary. Observations upon freshwater fishes have brought out the following facts as to dispersal and migration:

1. Dispersal and migration take place in circumpolar zones the range of migration depending upon: (a) temperature, (b) climate, (c) euryhalinity or stenohalinity of species, genera, etc., (d) vitality of given individuals to withstand sudden changes in temperature, in salinity, or in the amount of available water and food supply.

2. The interlacing of the headwaters of mighty river systems oftentimes accounts for the occurrence in the lower reaches of rivers hundreds of miles apart of identical or closely similar genera and species. The case of the trout on the North American continent is a familiar illustration. In the interlacing headwaters of both the Columbia and Missouri rivers occurs the cut-throat trout, Salmo clarki. Various species are gradually differentiated away from the headwater region. Thus the nearest relatives of S. clarki are S. virginalis in the basin of Utah, and S. sternias of the Platte River. "Next to the latter is Salmo spilurus of the Rio Grande and then Salmo pleuriticus of the Colorado. The latter in turn may be the parent of the Twin Lakes trout, Salmo macdonaldi. Always the form next away from the parent stock is onward in space across the barrier" (Jordan, 134, 547). Migration from the headwaters of one system to those of another only a few miles distant is accomplished: (a) as a result of river capture, (b) by the accidental transportation of the eggs of fishes, by birds, from one stream to another, (c) by the temporary formation of connecting streams or lakes between two river systems in a period of torrential rains, (d) by the temporary or permanent shifting of the watershed between two systems by a slight geological change, (e) by actual migration of fishes over areas where there are not continuous waterways. "Some fishes, provided with gill-openings so narrow that the water moistening the gills cannot readily evaporate; and endowed, besides, with an extraordinary degree of vitality, like many Siluroids (Clarias, Callichthys), eels, etc., are enabled to wander for some distance over land, and may thus reach a watercourse leading them thousands of miles from their original home" (Günther, 97, 212).

3. A shallow body of salt water between two continents may, by a very slight negative eustatic movement, be drawn off and a dry land connection will be afforded which will enable easy migration for freshwater fishes from one continent to the other. A subsequent positive eustatic movement would conceal the route of migration and one would have to deal with some apparently inexplicable occurrences of identical species.

4. "From the great number of freshwater forms which we see at this present day acclimatised in, gradually acclimatising themselves in, or periodically or sporadically migrating into, the sea, we must conclude that, under certain circumstances, salt water may cease to be an impassable barrier at some period of the existence of freshwater species, and that many of them have passed from one river through salt water into another" (Günther, 97, 211).

These facts which have been found out in connection with the distribution of freshwater fish of the present are essentially true for those inhabiting the rivers of all earlier continents. They may, furthermore, be considered as equally true for the eurypterids who were highly organized gill-breathers and many of whom were powerful swimmers. While they lacked one of the modes of transportation from the headwaters of one river system to those of another in not having the possibility of accidental portage by birds they had, on the other hand, a far more important means, for they had walking legs, and it is possible that they might have been able to withstand exposure to the air for several hours. In passing from one stream to another their locomotion would be fairly rapid and their migration in this manner might not have been infrequent.

APPLICATION OF PRINCIPLES DEDUCED FROM MODERN FAUNAL DISTRIBUTION^[1]

By the discriminating use of the laws which have been observed to be potent in directing the migration of fishes and other organisms living in the rivers at present, and without making any unwarranted assumptions, it seems safe to postulate the following expectabilities in regard to the geological and geographical distribution which we should be able to find among the eurypterids, providing they lived in the rivers.

1. Unless, as some have supposed, but which is very improbable, there were no climatic zones in the Palaeozoic, and conditions of temperature were equable over the whole globe, related or identical species of eurypterids should be found in deposits geographically situated in a circumpolar zone, not necessarily the same as the climatic zones of the present.

2. Eurypterid remains should be expected to occur in deposits of limited areal extent marking lake sediments, flood plain deposits, or littoral deposits in the sea at or near the mouths of rivers.

3. Eurypterids which inhabited the streams of one river system would be more closely related than those living in the tributaries of different and entirely distinct systems, and in general this would mean that forms which lived in the rivers of one continent in any period, would constitute a group of related genera and species, while those living in the rivers of another continent would constitute a distinct group, the individuals of which would be related; and if the different continents should remain unconnected for a long time, geologically, distribution and evolution would continue on each land mass, but we would not expect any of the individuals from one continent to migrate to another, so that succeeding faunas should not show intercontinental affinities, though phylogenetic relations should be discernable on each continent. It must be remembered, however, that remains of faunas from both continents might be carried into basins which received the simultaneous drainage of rivers from each.

4. In deposits which, from the study of their lithogenesis can be shown to have come from the same Palaeozoic continents, should be found remains of eurypterids in circumscribed areas as stated in "2" above, and the genera and species, while not necessarily having any near relatives in adjoining deposits, may be identical with forms whose remains are found in a formation perhaps two or three thousand miles distant, but on the same ancient continent. Such relationships are to be accounted for by migration from a common source where the headwaters of two or more river systems interlace (see p. 205 above).

5. The distribution of eurypterids would not have had any necessary connection with those organisms living in marine chronofaunas, and consequently, except when eurypterid-bearing deposits merge into thalassigenous ones, or when fragments or stray eurypterids have been washed out to sea, when intercalation between marine deposits would give the age, eurypterids would not serve as good index fossils.

6. Eurypterids would not suffer rapid changes in evolution, since it is a well known fact, that fluviatile types are often persistent for a long period of time. Thus the cray-fish Cambarus primaevus Packard of the Green River beds (Eocenic) of Wyoming, is a near relative of the modern C. affinis of the same region, a similarity due no doubt to the persistence of the type in essentially the same river basin during the interval.

Zoölogists and palaeontologists who have made detailed studies of the distribution of modern freshwater faunas are thoroughly agreed that accurate results are not to be obtained merely from observations on present distribution. It is an absolute necessity to study the fossil faunas and especially the palæogeography. The reason for this will be evident after a very little thought. If in the Lower Cretacic when there existed the Nearctic continent, comprising most of North America, and continuing across the North Atlantic through Greenland and western Europe, and including the Scandinavian mass, a family of some fluviatile organisms had arisen in the central Canadian area, quickly spreading from one river system to another and finally reaching Europe, we would find in the rocks of that period, that many of the genera on the two modern continents were the same, and that there would be quite a goodly number of identical species. The descendants of these Lower Cretacic organisms would develop on the two continents, (i.e., the two sides of this old nearctic land mass), and the species in the lower reaches of the rivers would diverge in their characters more and more from the parent stock. Those forms which came under the same environmental conditions might, and experience shows that they would develop along parallel lines, appearing in later geological times as similar or even what might be called identical species. In the course of centuries emigrants from an earlier home centre of distribution would pass from the headwaters of one stream to those of another, and soon these forms which had been passing through their individual modifications under one set of environmental factors would migrate down the rivers and mingle with those, forms which had in an earlier period sought the lower reaches of the rivers where a different complex of environmental factors obtained, and there the old immigrants and the new, would come to live in the same waters. A single family, in this way, would give rise to a certain number of primitive genera, some of which would migrate far from the original centre of distribution. The descendants of these early immigrants might, after a long time and after having suffered profound morphological changes, return to mingle with the descendants of the provincial forms which had never left the ancestral region. Now let us think of such inter-changes going on across the Nearctic continent all through the Tertiary until at the close Europe was separated from North America by an advance of the sea. At once we have two separate continents and two river faunas. Were one to try to account for the distribution of the fiuviatile forms now living in the rivers by a study of the present geography, one would be in despair to account for the similarity or seeming identity of many species on opposite sides of the dividing waters. Evidently the only mode of attack is by the study of successively earlier and earlier fossil faunas and by the slow reconstruction of the palaeogeography for each of those periods. One need not search far to find the application of these hypothetical statements to the eurypterids. If they were river-living organisms then it is clearly impossible to explain their distribution in any particular period without considering their distribution in each immediately preceding period. No one has ever done this because each writer tried to account for eurypterid occurrences on a hypothesis of marine distribution.

The results of migration are very different for marine organisms, because of the fundamental difference between the continuity of the seas and the discontinuity of the lands. Marine faunas, especially, the vagrant benthos of the littoral zone and the pelagic ones, tend to be widespread, for they have greater freedom in the size of life districts available, and in the lesser competition, as compared with the lineal extent of rivers and the great straggle for existence, particularly between crastaceous animals. For instance, Ortmann has pointed out that freshwater crayfishes existed in Southeastern Asia, the Malaysian Islands, India, and Madagascar in the Middle Cretacic. In the Upper Cretacic the freshwater crabs (which are geologically younger than the crayfishes) arrived (or originated) in Lemuria and "extended into Southern Asia and the Malaysian Archipelago, everywhere exterminating the crayfishes, namely, in India, South- eastern Asia (Farther India and China) and on the islands. They not only acted as a check to the distribution of the crayfishes, but directly annihilated them" (Ortmann 201, 391). As a result, no crayfishes are today found in the rivers of central and south Asia or on the Malaysian Islands.

We have previously seen that in river faunas the number of individuals is large but the number of genera and species is small, while in marine faunas genera, species, and individuals are abundant. The factor, then, of relative numbers of taxonomic groups would favor marine organisms in widespread migrations. Pelagic and vagrant benthonic organisms, living in the sea, have on the whole rather favorable conditions for migration. With river forms the factors of distribution are more accidental and much depends upon the individual. In the region of interlacing headwaters, streams of different systems are temporarily connected at times of flood and perhaps only two or three individuals of a certain species will change from one system to another, and then, when the connection is broken, the distribution of that species depends entirely upon the ability of the individual to contend with all of the new factors in the environment, and it is pure survival of the fittest which brings about the distribution of that species. In the sea, on the other hand, whole groups migrate or are carried by currents, and the chances are good that a large number or at least enough for populating a new region will survive, whatever vicissitudes befall. Thus, to sum up, distribution of river forms over broad areas is more precarious and fortuitious than is the case with marine organisms.

When we apply such considerations to fossil faunas, to a class of organisms wholly extinct, where we have no facts of modern distribution to help us, no facts of present habitat to point past modes of life, we can see that the criteria which we apply to such fossil faunas in the determination of relationships and migrations must be quite different from the ones applied to marine fossil faunas. We can now understand that a fauna may be made up of individuals which show a fairly close relationship with faunas in neighboring areas, but may contain one species which is identical or nearly so with a species in a fauna three thousand miles distant. If these were marine fossils we could not understand such a thing, because marine faunas show whole groups of species in one region related to groups in another, and contemporaneous marine deposits in the path of migration show similar related groups. But the routes of migration for river forms would almost never be shown to us in the rocks, because rivers in their upper and middle portions degrade and would continually be carrying away the traces of their history which would be recorded only in deltas or flood plains. Thus, contemporaneous and related fluviatile faunas would appear geographically at the outer ends of the spokes of a great wheel which has its hub at the centre of dispersal. The remains of synchronous faunas would of necessity appear scattered over the face of the earth, without any apparent connection; a fact which would be inexplicable if the faunas were interpreted as marine. The only way to solve the problem of the distribution of those forms would be through a study of the palæogeography of the period in which they occurred and of all preceding periods in so far as was possible.

When stratigraphers come fully to appreciate the value of continental deposits and faunas, they will have taken a big step toward the unravelling of the palæogeography of our earth. No one would attempt to restore the conditions of land and sea in the Tertiary without making use of the migrations of mammals and other terrestrial organisms, for it is evident that while a study of marine faunas will show the position of the oceans and epicontinental seas in any period, the exact configurations of the continents, the exact location of land barriers and connections can only be determined by the migrations of the animals and plants living on the land or in the rivers. This applies with as great truth to the Palaeozoic as to the Tertiary, and while the aid of plants cannot there be invoked until the end of the period, I hope to show before concluding this paper that the eurypterids will be of vast service in helping to locate Palaeozoic rivers and routes of migration from one continent to another.

MIGRATION AND DISTRIBUTION OF THE EURYPTERIDS

Theory of Early Marine Habitat and Routes of Migration. As I stated above, the anomalies in the distribution of the eurypterids have not usually been given much consideration, though they are of the utmost importance. There is a current opinion that has somehow been formed about the bionomy of the eurypterid faunas and no one thinks of challenging it. When a eurypterid fauna has been found in a place where a marine fauna was not expected, it has had to be made to fit in with the preconceived opinion about the bionomic facies in which eurypterids are supposed to occur. It has been spoken of as a "most unusual occurrence;" "one which is most interesting because found in beds formerly supposed to be devoid of marine fossils," and so on. Again we read of the clear evidence of a marine passage between the Buffalo region and the Baltic area, because two almost identical species of eurypterids are found in these localities. Formations are declared to be marine because they contain eurypterids, and eurypterids are held to be marine, because they occur in formations considered on a priori grounds to be marine. Every writer seems to feel it necessary to fit the eurypterids into a marine or estuarine habitat; where the facts refuse to fall into line, they are cited as ineresting because they fail to, or else they are consciously suppressed or carelessly overlooked. The prevailing opinion as to the bionomy of the successive eurypterid faunas is as follows: Until well on in the Siluric the eurypterids were purely marine forms living in the seas and, inferentially, associated with the marine organisms therein. Toward the middle of the Siluric, the eurypterids all over the world left the seas and migrated into the various brackish water bodies then existing, seeking the mouths of rivers, the bays, lagoons and interior cut-off arms of the sea. From that time until the end of the Palaeozoic, they are supposed to have sought water of ever-decreasing salinity until they became entirely freshwater denizens. Their geographical distribution is accounted for by an assumed migration from one estuary or lagoon to another along the shores of various Palaeozoic continents.

Objections to Marine Habitat Theory. If this succession of events is the correct one, then the following question arises in connection with the distribution: If the eurypterids lived in pools or in marginal lagoons on the seashore, in estuaries, bays or cut-offs how did they get there to begin with?

The question is generally answered by the statement that the eurypterids originally lived in the sea and then migrated to the various marginal water bodies and estuaries where they and a few peculiar crustaceans constituted a brackish water fauna. I have already shown (p. 70) that a "brackish water" fauna consists of modified marine and freshwater euryhaline organisms with a preponderance of marine types, and that the latter show particular characteristics such as dwarfing and thinning of the shell, but that such a fauna has representatives of nearly all invertebrate phyla and is not made up of a single class of organisms. But let us assume for the sake of argument that the eurypterids and a few other arthropods did form a brackish water fauna; then another assumption is necessary, for, if a class of organisms as a whole, such as the eurypterids, should in any given geological period migrate from the sea to estuaries or other brackish water bodies and at the same time should no longer be able to live in the sea, and should not, on the other hand, become adapted to river water, then the remains of such a class of organisms should be restricted to the geological period in which the migration took place, for the class could not persist unless the estuaries persisted from period to period in the same locality (see objection to this on p. 215 below).

But since the class is known to have persisted from period to period, as indicated by the occurrences of their remains in the rocks, we are forced to conclude, on the assumption that the organisms migrated from the sea to the estuaries, that there was a persistent marine stock to repeople each successive estuary. But, if that were true, then eurypterid remains of the same or allied species should be found entombed with the marine organisms of the period in the marine equivalents of the estuarine or other brackish water deposits, and the eurypterids should have constituted a part of the typical marine fauna. But it has been shown again and again that in the contemporaneous marine deposits with typical and undoubted marine faunas, no eurypterids are found, as, for instance, in the marine Wenlock of England, or the marine limestones of the Famennian of Germany. If there is no indication of such a persistent marine stock, then there must have been a persistent stock in the rivers to repeople the estuaries in the successive geologic periods. These arguments may be applied specifically to the Siluric and Devonic of North America. During the Lower Siluric (Niagaran), the eurypterids are supposed to have lived in the sea. During the remainder of the Siluric they are assumed to have lived in or along the shore of a shallow, epicontinental sea having a connection with the Atlantic or other waters to the east. In this restricted sea terrigenous deposits were formed, well represented by the Shawangunk delta. In the pools along shore, where, on account of the more sheltered conditions, only muds were accumulating, the young eurypterids lived. The larvæ were hatched in these pools and the early stages in the ontogeny were passed through, then the mature individuals sought the deeper littoral waters. Thus do Clarke and Ruedemann explain the presence of the abundant fauna composed almost entirely of young individuals in the Shawangunk shales at Otisville, New York, and, during the same period, the closely related but mature individuals in the Pittsford shales at Pittsford, New York.

A comparison, species by species of the forms from the Pittsford and Shawangunk will be given below (p. 225), and it will be seen to show that the two faunas are very closely related, indeed, almost identical except in the size of their individuals, and in the presence, in the Pittsford, of a species of Eurypterus related to a Bertie form to be considered presently. Such similarity might, if taken alone, seem to substantiate the "lagoon" theory. But it is usually impossible to draw very accurate or very far-reaching conclusions from the consideration of faunas or of deposits in a single circumscribed area or at a single horizon; one must take into account the palæogeographic conditions in neighboring regions and finally throughout the whole continent if not, indeed, the whole world, and one must consider the source of supply of sediments, the possibilities of migrations of faunas and the absolute necessity of a fauna to have a medium in which it can live from one period to another, unless we wish to revert to the belief in special creations. Thus, bearing these things in mind, we must account for the origin of the sediments of the Pittsford and Shawangunk and of the succeeding formations, the various waterlimes, which contain eurypterids. It has been demonstrated on pp. 100–6, that the conglomerates and shales of the Shawangunk and the shales of the Pittsford must have come from Appalachia, carried northwards by various rivers.

Now, assuming for the sake of argument that the succession of events during Salina time was that outlined above (p. 212) then the following conditions are implied: (1) The Pittsford and Shawangunk faunas must have constituted the ancestral stock for the Bertie fauna of Erie and Herkimer counties. (2) Throughout the long period from Pittsford to Bertie time, one or several rivers must have occupied approximately the same position, so that the Pittsford and Shawangunk faunas could escape into the estuaries when the Salina sea became too salt, and could remain there in the brackish water part of the estuary until Bertie time, when they appeared in two localities, at Buffalo, 75 miles west of Pittsford, and around Herkimer, 130 miles east of Pittsford. Taking up the first condition, we are confronted with a grave difficulty if we try to think of the Pittsford and Shawangunk fauna remaining in the Salina "lagoon" or at the mouths of estuaries flowing into that inland body of water during Vernon, Syracuse, and Camillus time, for it is evident that we must consider the Pittsford-Shawangunk eurypterids as the ancestors of those found in the Bertie, if we believe in this estuarine theory. In the succeeding pages, where I shall consider every species of eurypterid as an entity and as a member of a faunule, unless it be an isolated form, and where I shall take up the possible modes and routes of migration of species and of faunas, I shall show that the Pittsford-Shawangunk eurypterids were not the ancestors of the Bertie forms, and therefore the first condition which I mentioned at the beginning of this paragraph as a logical deduction from the "lagoon-estuary" theory is impossible, in which case it would appear that the Bertie eurypterids had no ancestors. Let us suppose, however, for the sake of argument, that the Pittsford-Shawangunk fauna did constitute the ancestral stock for the Bertie fauna and that in the dry and at times uncomfortably saline conditions of Salina time the eurypterids left their lagoon and went into the estuaries and even part way up the rivers, seeking proper salinity of water; then we should look for estuarine deposits of mud or perhaps coarser clastics in the Salina of central and western New York, and for the remains of marine organisms which are characteristic of such deposits. (For criteria of estuarine deposits see p. 77 above.) But we search in vain for estuarine, or delta, or flood-plain deposits in that region. Following upon the Pittsford are the Vernon barren red shales with their evidences of subaërial deposition with thorough oxidation (Grabau, 84, 86a, 87), and then the Syracuse salt deposits. All of this has been discussed before, and the evidence is clear that there existed no estuaries in the area under question in which the early Siluric eurypterids might have sought refuge. Thus, descendants of early Salina "lagoon" species had no place of retreat during later Salina time, and must have perished of drought, and we see that the Bertie eurypterids were doubly deprived of ancestors if they had to depend upon the Pittsford-Shawangunk fauna.

Theory of River Habitat. To pursue this marine-lagoon theory to its logical conclusions in every case would use up many pages of print and would always lead to absurdities, impossibilities or contradictions. Therefore, without dwelling longer on the perplexities and inconsistencies attendant upon this theory, I shall pass at once to the development and exposition of the theory of river habitat. Throughout the Palæozoic there were in existence in the northern part of the western hemisphere three continents which, though varying much in size from period to period, often becoming confluent and at times even being largely covered by the epicontinental sea, nevertheless preserved a marked degree of integrity. These three continents were (1) Appalachia, which occupied what is now the eastern border of North America, and constituted the northward projection of the land mass now known as South America, and which supplied the greater part of the clastic materials deposited in eastern North America throughout the Palæozoic; (2) Rockymontana, which lay for the greater part of its length on the present continental mass extending from Mexico to Alaska, a palæocordilleran chain, from which clastic sediments were derived which were deposited on the western border of North America and along the continental shelf; (3) Atlantica, the great northern North American and northwestern European continent one portion of which, the Canadian shield, was formerly supposed to have been the source of nearly all of the Palæozoic clastic deposits over what is now the United States. Throughout the Palæozoic this Canadian area was usually connected with the Scottish and Scandinavian masses by a broad strip of land extending across the North Atlantic (see map, fig. 8). There was a fourth and smaller continent, Mississippia, occupying the area of the Mississippi valley, which at times was entirely covered by the sea, and again formed a part of Rockymontana. Each of these continents had its own river systems, the organisms living therein being subject to the laws of migration and dispersal which are seen to be operative now. Furthermore, the fiuviatile fauna of each continent would be distinct as a rule. If, however, migration in circumpolar belts occurred and fluviatile organisms from one continent passed to another, these migrant forms would yet show their closest affinity not to species living in the rivers of the continent to which they were immigrant, but to those in the rivers of the continent from which they emigrated. In any given period faunas which can be shown to have come from rivers on the same continent should be more closely related than faunas coming from rivers of different continents, but there may be single cases of a family, a genus, or even a species which occurs in sediments from one land mass, which is nearly related to or identical with one in sediments coming from another land mass. In such a case, to determine true relationships one must compare the whole of each fauna, species by species, and must in addition study the ancestors of each fauna and of each species in the preceding periods wherever possible.

THE EURYPTERID FAUNAS CONSIDERED BY CONTINENTS

The Eurypterid Faunas of Appalachia. Let us turn now to the placing of the various pre-Siluric eurypterids. Strabops thacheri from the Cambric is too primitive and morphologically undifferentiated to be looked upon as more than an ancestral form approaching the prototype and from which several branches of the eurypterid tree diverged. The first prolific eurypterid fauna in North America, the first to offer sufficient material and a large enough representation in genera and species to make it possible to state what are the general affinities of the fauna as a whole, is the newly discovered one in the Normanskill shales at Catskill, New York, which has so far been found to contain six species, included in five genera, but undoubtedly many more will be discovered as the material is worked over. On account of the fragmentary nature of the abdomina found, and because the carapaces are usually dissociated from the rest of the body, generic determinations have been provisional and comparisons with related species difficult. Yet the fauna shows a pronounced and altogether surprising similarity to that of the Schenectady beds (Trenton) despite the difference in age. In the case of Pterygotus? (Eusarcus) nasutus, Clarke and Ruedemann "have been unable to distinguish the Schenectady and Normanskill types," (39, 412); and have referred a number of carapaces from the Normanskill beds to P. nasutus, a species described originally from material from the Schenectady shales. Eusarcus linguatus from the Normanskill is very similar to Pterygotus? (Eusarcus) nasutus.^[2] Eurypterus chadwicki, Dolichopterus breviceps, and Stylonurus modestus are not well enough represented for relational comparisons to be made, so far as species are concerned. The finding of several Stylonurus carapaces, attached abdomina, and one with a portion of a leg, so early in the Ordovicic in muds derived from Appalachia is most suggestive. In the succeeding Schenectady beds in the same general region, in muds also washed down from Appalachia, occur a number of specimens which, in the shape of the carapace, position of the eyes, etc., suggest their generic reference to Stylonurus and have been described by Clarke and Ruedemann as S.? limbatus. They have furthermore found a number of body segments "which have the form and ornamentation of the Otisville species Stylonurus myops" (39, 296). Although it is a little out of chronological order to bring in the Utica species before taking up the Schenectady fauna, this, nevertheless, is the logical place for its discussion. Echinognathus clevelandi was described from a single endognathite which has shown two diagnostic characteristics, namely, an extreme spinosity, and a peculiar and distinctive type of surface sculpture. Clarke and Ruedemann state that this species "was either closely related to Stylonurus or had a convergent development to that genus as far as the two characters mentioned are concerned" (39, 322). It may quite properly be asked why it is that if the single endognathite known, shows only two diagnostic characteristics, and these two are recognized as definitive of Stylonurus, the species does not belong to that genus, or at least is it not more than likely, if more specimens are discovered, showing other parts of the body, they will be found to represent Stylonurus? It seems to the author that the geographical and geological position of E. clevelandi alone would suggest the greater possibility of the form being a Stylonurus. To be sure, this is somewhat speculative, but it is a suggestion for future work and consideration; it is sufficient that the Utica species is at least closely related to the genus Stylonurus which was found at earlier periods and also in the Siluric and Devonic, always in deposits derived from Appalachia. This statement includes the Utica beds just mentioned, for it is now recognized that, as Professor Grabau first pointed out, the muds were carried down from Appalachia and were merely the eastern near-shore facies which replaced that of the Trenton limestone facies (Grabau, 84, 231–232). Passing on to the next time in the history of North America when the genus Stylonurus is known to occur, we find S. (Ctenopterus) multispinosus in the Pittsford and two well defined species of this genus, as well as many fragments specifically indescribable though evidently distinct in the Shawangunk, both of which formations have been interpreted on stratigraphic grounds and on a comparison of the two faunas inter se, but not for any phyletic reasons, as derivatives from Appalachia, the Pittsford constituting the upper part of the Shawangunk (see p. 101 above). The problematic form from the Portage sandstone referred to S.? wrightianus is too incomplete to be of much value. It probably belongs to Stylonurus; it certainly occurs in an otherwise unfossiliferous deposit which has been interpreted by Grabau as partly of river floodplain and partly of wind-blown perhaps loess-like origin (87, 553, 569). Finally, the Upper Devonic yields two species of Stylonurus: one, S. beecheri described from a single individual, none too complete, from the Chemung sandstones of Warren, Pennsylvania; the other S. (Ctenopterus) excelsior from two specimens from the Catskill beds of New York and Pennsylvania. This latter species is related in many respects to S. (Ctenopterus) cestrotus from the Shawangunk, both belonging to the same sub-genus. The Catskill is a continental deposit whose material as shown first by Grabau (86) and later by Barrell was derived from Appalachia. The specimens of S. excelsior were beyond a doubt washed out into the Chemung sea, since all of the species of Stylonurus so far known from North America came from Appalachia, as has just been demonstrated.

Having now followed the history of this one genus from Ordovicic through Devonic time and found that it always lived in the rivers of Appalachia, let us return to the genus Dolichopterus in the Normanskill beds, and trace its subsequent occurrences. As in the case of Stylonurus, the specific relations of the Normanskill form cannot be determined, for only a single small carapace is known, but the point of especial interest is the occurrence thus early of a Dolichopterus. This genus is represented by two species certainly, and one doubtfully, in the succeeding Schenectady beds. Two specimens described under the new species of D. latifrons by Clarke and Ruedemann agree "closely with D. otisius" from the Shawangunk in the posterior contraction of the carapace (39, 270). The carapaces and metastomes of D. frankfortensis (Schenectady) do not seem to show close relationship to other species of the genus, though one metastoma "has been found which recalls that of D. macrochirus" from the Bertie (39, 269). A few fragments having certain Dolichopterus and certain Eurypterus characteristics have been referred to E.? (Dolichopterus?) stellatus, but they are of no value in the present discussion. Thus it is seen that in the meagre, unsatisfactory material from the Normanskill representative of Dolichopterus, one species shows affinities to a Shawangunk form, and one specimen of a second species recalls characteristics of a Bertie species. The evidence is frail, and yet it might seem a little disconcerting to have an individual which came from Appalachia, as we think, showing relationship to one coming from Atlantica, but I shall have a suggestion to make when I come to the Bertie that will do away with even this slight difficulty, which is, after all, entirely negligible, since the only specimen showing relation to a Bertie form is a single metastoma which bears only a suggestion of similarity. Continuing in chronological sequence, the next formation in which a Dolichopterus occurs is the Shawangunk grit where there are two species D. otisius with a large representation of carapaces none of which retain more than two body segments, and D. stylonuroides of which three carapaces and one more complete individual have been found. The former species has certain characters in common with D. macrochirus (Bertie), the young of both being even more alike than the adults. If the two species are phylogenetically related, then the adult D. macrochirus has kept the ancestral characteristic of a broad frontal lobe on the carapace, for this is found in the young of both species and retained throughout the ontogeny of the Bertie form, while D. otisius in the adult shows a development of this lobe into an angular extension. In this one characteristic, then, D. macrochirus would show retardation. The second species in the Shawangunk is rare and shows no close relationship to any known species.

In considering both the Schenectady and Shawangunk faunas, we have seen that there was a species of Dolichopterus in the latter and a single specimen in the former which showed a more or less close relationship to certain species of the same genus in the Bertie. From purely stratigraphic reasoning it is known that specimens of the first two formations were derived from Appalachia, while those of the Bertie came from Atlantica. The question might be raised whether the stratigraphic facts do not conflict with my biological theories, for I have been trying to show that eurypterids found in sediments which were transported by rivers on the same continent should show genetic relationship and should for the most part be distinct from those which lived in rivers on different continents; but the species of Dolichopterus do not seem to conform to this law. In the case of this particular genus with its distribution in time, there would be no apparent physical objection to the accounting for its affinities on the very simple assumption that the Cambric or earlier generic ancestors lived in rivers either on Appalachia or on Atlantica and that during one of the peri ods when these continents were connected by a strip of land the eurypterids were widely dispersed on both continents; the later consanguinity is thus easily understood. But I think that such an assumption is unnecessary and my reason will be readily apparent when we consider the Bertie species of Dolichopterus, D. macrochirus, D. testudineus and D. siluriceps. The relationship of the first to a species in the Shawangunk has just been discussed. Concerning the second, Clarke and Ruedemann remark: "This species, as represented by the single carapace, is quite similar to D. otisius. It differs from the latter mainly by the greater extension of the frontal portion and by the more pronounced posterior contraction of the carapace. The frontal transverse ridge or fold observed in the species is also seen in D. otisius" (39, 275). If the two species were genetically related, this more pronounced extension of the frontal portion of the carapace would be predicable in the Bertie species, for according to the laws of recapitulation and tachygenesis a morphological character found in the adult of any species will appear at an earlier and earlier stage in the ontogenetic development of its descendants, and since the apparently orthogenetic tendency in the Shawangunk species D. otisius showed a progressive modification from rounded to angular and extended frontal margin, the late Bertie species D. testudineus should show a more protruding frontal rim than is found in the adult D. otisius. The third Bertie species is D. siluriceps of which a single poorly preserved carapace is known, and which cannot be compared to any other species save a small form from the Shawangunk. The genus Dolichopterus is not known from any other country, nor has it been found in beds of later age than the Bertie. Even the three species in the Bertie are so poorly represented that one wonders what happened to the fauna. Of the genotype, D. macrochirus, four incomplete though excellently preserved specimens are extant; of each of the other two species there is a single carapace. If these eurypterids lived in the Bertie "pools" of authors, it is inconceivable that not more individuals (or exoskeletons) should have been preserved; if they lived in the rivers coming from Atlantica, this scarcity is accounted for. But the study of the phylogeny of this genus leads me to think that Dolichopterus was confined to the rivers of Appalachia throughout its whole racial history. (Its occurrence in so fragmentary a condition in the Bertie suggests that the few remains were transported from the debouchure of some river of Appalachia and carried into the Bertie muds). There is as yet too little evidence, too many pages in the history are still unread, for a reasonably definite conclusion to be drawn, but I think that such a geographical development and confinement more satisfactorily accounts for the facts which are known than any others. We are not obliged to believe that Dolichopterus always lived in the rivers of Appalachia; the facts of distribution and relationship could be accounted for otherwise; but this belief requires fewer special conditions than the assumption of very early dispersal by rivers on the two continents, while a marine habitat is entirely out of the question. One of the strongest reasons for my conclusion that Dolichopterus was restricted to Appalachia lies in the evidence offered by the origin of the sediments. In the study of any problem if the lithogenesis of the formations concerned points overwhelmingly to one and only one history for those formations, then slight palaeontological incongruities should not be accepted as vitiating the history pointed by the facts of lithogenesis; the apparent incongruities can generally be turned into confirmatory bits of evidence if a broad enough knowledge and a scientifically guided imagination can be brought into play. Thus, when the nature of the outcrops, the lithological characteristics of the rocks, and, most important of all, the consideration of possible sources of supply for material, all point to the continent of Appalachia as the region whence the Normanskill, Schenectady, and Shawangunk deposits must have come, while these same considerations point just as conclusively to Atlantica for the Bertie deposits, then, if a fragment of a eurypterid in the Schenectady shales shows a faint similarity to a form in the Bertie, and if half a dozen specimens in the Bertie waterlime bear a slight or even pronounced resemblance to species in the Shawangunk, we must attempt to visualize the conditions obtaining on the North American continent during the early Palaeozoic and we must seek the most rational explanation, the one most in accord with our knowledge of the laws operating at present, to account for these seeming anomalies. And we should never forget that the geological record has revealed but a few specimens of most species of eurypterids, and that sometimes even a genus is described from a single individual, and that when a writer describes a new species he compares it with the ones already known, drawing analogies where he can; but species which may seem to be very much alike when one has, say, a single member, a carapace, or a claw, of each to compare, might, if a large quantity of perfect material were available, be discovered to be so different that kinship would be found to be entirely lacking where formerly it had been confidently pointed out.

From the great mass of detail which it has been necessary to give, we are at length able to reach two conclusions: (1) Stylonurus from its earliest appearance in the Normanskill beds (Black River or Basal Trenton) to its last appearance in the Chemung was an inhabitant of the rivers of Appalachia; (2) Dolichopterus also first known in the Normanskill, but so far not known from beds later than the Bertie, is most rationally to be considered as restricted in its habitat to the rivers of Appalachia, although the paucity and the condition of the specimens make this conclusion not absolutely certain.

This much being determined, we may consider the remaining faunas which have been found in sediments which from other lines of reasoning are recognized as coming from Appalachia. In the Schenectady shales eleven species are recorded, of which four have already been discussed. Of the remaining seven, Eurypterus ruedemanni and E. pristinus are represented each by a single carapace neither of which is of use in comparisons, and the same may be said of the doubtfully determined form Euscarcus (?) longiceps of which a few incomplete carapaces are known. Nine carapaces of Eusarcus triangulatus have been found, and these Clarke and Ruedemann state "have in common the broad, short, subtriangular form; and the forward position of the marginal lateral eyes bears a close resemblance to the carapace of E. scorpionis from the Bertie waterlime" (39, 258). Yet the figures, measurements and descriptions of these two species given by the above mentioned authors do not bear out this "close resemblance." In reference to E. triangulatus they say that the carapace is "twice as broad as long (length of type, 20 mm., width 43 mm.)" and of E. scorpionis they say that the carapace is about "as broad as long" (p. 234), while in the measurements which they give of this species the length is to the width (in millimeters) as 18:22, 60:66, and 56:59, respectively. For comparison I give outline drawings of the restoration of the carapace of E. scorpionis and of the actual carapace of the type of E. triangulatus (Figs. 19a and b). One of the commonest species in the Schenectady shales is Hughmilleria magna, known from a number of carapaces, some abdomina, and a half complete individual. "This exhibits a form of the preabdomen corresponding to H. socialis," but the swimming leg is "relatively longer than that of H. socialis" (Pittsford) (39, 342). Several detached body rings have been found regarding which Clarke and Ruedemann say: they "exhibit a type of ornamentation, consisting of transverse lines near the anterior margin, known to us only in H. shawangunk, the Otisville representative of the genus" (p. 342). Thus the nearest affinities of this Schenectady species are to forms from the Pittsford and Shawangunk, which it has been suggested might themselves be merely growth stages and not "species." Only comparatively young (for the most part nepionic or neanic) individuals are known from the Shawangunk, but it is significant that many of these are almost identical in

Fig. 19a. Eusarcus triangulatus. Clarke and Ruedemann. ✕ ${\tfrac {1}{2}}$.
(After Cl. & R. 1912, pl. LXXXIV, fig. 7)
Fig. 19b. Eusarcus scorpionis. Grote and Pitt. ✕ ${\tfrac {1}{2}}$.
(Outline after Cl. & R. 1912, pl. XXVII, fig. 1, restoration)

form of the carapace and the position of the eyes with larger, neanic or ephebic individuals from the Schenectady, indicating relationship by this recapitulation of characteristics in ontogeny (see fig. 20.) The two remaining species of the Schenectady fauna, Pterygotus (Eusarcus?) nasutus and P. prolificus are unlike any other species known from this country, so that their comparative value is small. Although

Fig. 20a. Hughmilleria shawangunk Clarke. Nepionic Individual. ✕ 8.
(After Cl. & R. 1912, pl. LXIV, fig. 2)
Fig. 20b. Hughmilleria magna. Clarke and Ruedemann. ✕ ${\tfrac {1}{2}}$.
(After Cl. & R. 1912, pl. LXXXV, fig. 11)

the last mentioned species is the most profuse in the Schenectady beds, yet the variability in the shape of the carapaces is so great that one might easily be led astray in drawing conclusions. On the whole, the study of this fauna reveals it to be rather unsatisfactory. For one thing, it is made up for the most part only of carapaces and these are often fragmentary; besides, the forms are so distinctive, possessing such unique and specialized characteristics, that with our present slight knowledge of the various faunas we are unable to perceive relationships which very possibly exist. Two species, however, do show kinship with known forms. Hughmilleria magna has characters in common with H. socialis and H. shawangunk from the Pittsford and Shawangunk, respectively, while Dolichopterus latifrons agrees "closely with D. otisius from the Shawangunk in the posterior contraction of the carapace." Thus, whatever relationship is indicated between the species of the Schenectady fauna and those of later faunas, is to species in the Pittsford and Shawangunk, all three of which formations have elsewhwere been shown to have had their origin in the sediments from Appalachia. Once again, it appears that rivers coming from the same continent have successive faunas more closely related than those from diverse continents.

Comparison of Pittsford and Shawangunk Faunas. The study of the lithogenesis of these two formations has shown that the Pittsford shale is of the same age as the shales in the upper part of the Shawangunk (p. 101 above), for which reason it is fitting to consider the faunas of the two formations at the same time, especially since the sediments are known to have come from Appalachia in both cases. A comparison of the Pittsford and Shawangunk faunas shows that the two most common species, Hughmilleria socialis from the former and H. shawangunk from the latter are almost identical. In the shape of the body and form of the head the two species closely resemble each other, while the telsons of the two are identical. As one reads through the description of the Shawangunk form he is struck with the constant similarities in the anatomy between this and the Pittsford species. For instance, Clarke and Ruedemann say in regard to H. shawangunk: "The metastoma has not been seen well preserved in position, but we refer several metastomas to this species because they possess on the one hand, the form of that in H. socialis, and on the other, exhibit a peculiar, striated ornamentation apparently characteristic of H. shawangunk" (39, 345). Again, "The crawling legs appear to have been both short and slender as in H. socialis" (39, 344). Because of these similarities, it seems not improbable that H. socialis might represent a mature H. shawangunk, especially since no specimen longer than 8 cm. is known from the Shawangunk, and no specimen so short as that from the Pittsford, where the individuals are up to 15 cm. in length.

Of the other five species in the Pittsford, Pterygotus monroensis is of small importance for it is represented by a single carapace and two fragments. The carapace looks as though it might well belong to a large Hughmilleria; at any rate it has no close affinities to any other species. Similarly, little of correlative value can be deduced from Stylonurus multispinosus which is known only from a group of endognathites. In their characteristics they are different from anything in the Bertie (39, 297), and are of little relational value.

It will be shown below (p. 232) that E. pittsfordensis is closely related to E. lacustris in the Bertie, but as will be seen, this is entirely expectable.

In the Shawangunk fauna the most abundant species is Hughmilleria shawangunk whose relationship has been discussed under the Pittsford fauna. The very rare forms, Eusarcus (?) cicerops, Dolichopterus stylonuroides, Stylonurus myops, and Pterygotus globiceps, represented by only a few fragments, show no particular relation to species in any other fauna. Indeed, a comparison of the young of E. scorpionis with the young of E. cicerops shows that the cephalon was very different in outline and the position of the eyes was not at all similar (Clarke and R., 39). Similarly, Stylonurus cestrotus, found only in a fragmentary condition, "stands apart from all its allies in a number of characters that show it to be an aberrant form" (39, 291) Eurypterus maria, of which many young and one or two mature individuals have been found, is "greatly different from all its American congeners," (39, 190). The relations of Dolichopterus otisius have already been pointed out, and it has been shown that while it agrees in certain characteristics with one species, in others it agrees with a different one, so that its affinities cannot be said to be with any particular fauna.

Summarizing the evidence offered in a comparison, species by species, it becomes clear that the dominant, most abundant species in the Pittsford and Shawangunk faunas are alike and that there is only one form in either of these which shows relationship to a Bertie species.

Summary of Facts of Distribution on Continent of Appalachia. The following points may be briefly recapitulated: 1. In the sediments which it has been demonstrated (by myself or others), were with more or less certainty derived from Appalachia, the eurypterids are either unique, showing no relation to known species in North America or other continents, or else they show phyletic relationship inter se, the species of later faunas having certain characteristics in common with those of (generally the mature forms of) earlier faunas.

2. Three genera, Stylonurus, Echinognathus, and Hughmilleria were restricted to the rivers of Appalachia.

A far greater interest must attach to the vast northeastern continent of Atlantica which stretched across the north Atlantic and formed a land bridge of vital importance in the migration of the eurypterids. The organisms living in the rivers of this continent were not geographically restricted like those in the rivers of Appalachia, whose remains were washed out occasionally into the surrounding ocean waters, but which were prevented from migration to European fresh waters by the broad expanse of the Palæozoic Atlantic; more fortunate by far were the fluviatile inhabitants of Atlantica, for this continent, we may feel sure, was fairly permanent throughout the Palæozoic, even though the ocean at times encroached over much of the southern part; it was the northern portion that would be vital for the interlocking headwaters of different river systems, and as we shall see there is overwhelmingly convincing evidence pointing to such an intimate relation between the river systems of the periods from the Upper Siluric through the Devonic. Not only were the geographical position and extent of Atlantica more favorable for the widespread dispersion of the eurypterids than were the same physical features of Appalachia, but the sediments derived from the former continent were for the most part of the particular lithological character most favorable to the preservation of organic remains, while those from Appalachia were quite often coarser, being prevailingly sandstones and conglomerates, with only thin beds of intercalated muds. The early differentiation in the character of the clastic deposits from these two continents reflects the still earlier difference which had existed between them in the matter of elevation, for, whereas during Ordovicic and Lower Siluric (Niagaran) time the Canadian area, already peneplaned, had been largely covered by the sea, as indicated by the remnants of Niagaran limestones, and whereas during the same period the Baltic region and that area now forming the southern shore of the Gulf of Finland had likewise been covered by a shallow sea in which coral reefs flourished, the continent of Appalachia on the contrary, had jutted up from the Atlantic with lofty mountain ranges of crystalline rocks. Thus it came about that the rivers in their slow but efficient work of denudation brought into the waters bordering the continent of Atlantica sediments that were calcareous and usually fine-grained (waterlimes) while the rivers of Appalachia carried highly siliceous materials of medium or coarse grain (sandstones and conglomerates) and the winds transported siliceous sands.

The Eurypterid Faunas of Atlantica. The eurypterid-bearing formations which, mainly on lithological grounds, are thought to have come from Atlantica are (1) Bertie, (2) Rondout, (3) Manlius (4) Siluric waterlimes of Oesel, (5) Waterlimes of Gotland, (6) Wenlock of Scotland, (7) Old Red sandstone. The faunas of these various formations will be taken up in detail with a view to determining the relations between individual species and between the faunas inter se.

Of the above mentioned formations and their contained faunas, the first three, which are North American, are quickly disposed of. The Rondout waterlime has thus far yielded but a single species, and this is the same as one from the Bertie, namely, Eurypterus remipes. Similarly, only one species is known from the Manlius, a number of specimens, for the most part poorly preserved, having been found in various localities ranging from Albany, Herkimer, Madison and Onondaga counties, New York, to Put-in-Bay, Lake Erie, where it occurs in the stratigraphically older early Monroe beds. Only one specimen has been found in which the abdomen is preserved, the remaining occurrences being only of carapaces, and even these are often poorly preserved. In outline of carapace and lack of ornamentation thereon, this species more closely resembles E. brewsteri from the Arbroath paving stones than any form known from North America, though the similarity to E. lacustris from the Bertie is not to be overlooked. Thus, the only known eurypterid from the Rondout is the same as a species from the Bertie, and the single species from the Manlius and the lower Monroe shows affinities to one from the Bertie and to one from the Old Red sandstone. With these two so easily dismissed, we may turn to a detailed discussion of the Bertie fauna in which connection it will be necessasy to establish the complete affinities of each species by a detailed morphological and phylogenetic comparison with species in preceding and contemporaneous faunules in America and Europe; the centres of dispersion and the routes of migration must be carefully studied, and the possibilities of fluviatile and marine distribution must be weighed. More deductions can be drawn from the study of the Upper Siluric faunas than from that of any other, because of the abundant data available, the appearance of chronofaunas in widely separated localities and the relative abundance of individuals and species in several of the faunules. Because it is impossible to draw correct deductions regarding the mode of distribution of organisms in any one period from the observation of the distribution visible at that time (see p. 208 above), and since the truth is to be arrived at only by the consideration of former land and sea connections or barriers, it is necessary in discussing the Bertie faunule to take into account the palæogeography of preceding periods and the distribution of earlier faunas.

It has seldom been our good fortune to find two succeeding eurypterid faunas in the same locality, so that not many opportunities have been available to trace direct descent; but New York State has been favored in this respect and too much importance can not be attached to the relational values of the Pittsford, Shawangunk, and Bertie faunas.

Comparison of the Pittsford, Shawangunk, and Bertie faunas. As a matter of fact, the Bertie fauna in neither "pool" shows any very marked affinities with the Shawangunk fauna or with the Pittsford, with one exception, already noted, and more fully discussed, below. Of the fourteen species known from the Bertie, there are only four in which even a slight resemblance can be seen to the upper Niagaran forms, and this resemblance in each case (with the one exception noted) is so very small that it cannot be said to constitute a proof of genetic relationship. For instance, Dolichopterus (?) testudineus from the Herkimer "pool" is represented by a single uncompressed carapace which in outline, general proportions, and the position of the eyes is quite similar to one of the specimens of D. otisius in the Shawangunk. But while this general resemblance to one carapace of the Shawangunk form has been pointed out by Clarke and Ruedemann, attention should also be called to the fact that it is very different from one of the best preserved, most typical Shawangunk carapaces of the same species. The sub-elliptical shape of D. testudineus is quite distinct from the sub-quadrate one of D. otisius, and it does not seem to the author that any genetic relation is indicated between these two forms. To overcome this difficulty of lack of relationship between the Shawangunk and Bertie faunas, it might be argued that the latter was derived from the Pittsford alone. But the only species in the latter which is supposed to have even a semblance of relationship to a Bertie species is Pterygotus monroensis which has been compared with P. cobbi. The former species was founded on a single carapace, and two other fragments are now known. One of these is the fragment of a free pincer of the chelicera which is thought to belong to P. monroensis. This shows one long, rounded tooth at the extremity, then a short tooth, another long tooth but not so long as the first, four short teeth alternating in size, followed by a long tooth, then three shorter ones. The chela of P. cobbi from the Bertie shows three long teeth at the end, two short ones, a long one nearly as long as the one at the extremity, then three fairly short ones followed by another long one. The teeth in the chelae in these specimens are similar neither in size nor arrangement, so that no particular relationship is set up between Pterygotus monroensis and P. cobbi (Fig. 21 a and b).

Not only, then, do the species themselves offer no indication that the Bertie fauna was derived from the Pittsford alone, but, furthermore, it seems impossible to believe that the five Pittsford species included in four genera should give rise to the profuse Bertie fauna of fourteen species included in four genera, two of which are different. The four Pittsford genera are: Eurypterus, Pterygotus, Eusarcus, and

Fig. 21a. Pterygotus monroensis Sarle. Fragment of Free Chela. ✕ ${\tfrac {1}{2}}$.
(After Cl. & R. 1912, pl. LXX, fig. 3)
Fig. 21b. Pterygotus cobbi Hall. Free Chela of Chelicera. ✕ ${\tfrac {1}{2}}$.
(After Grote & Pitt. 1878, fig. p. 301)

Dolichopterus. Clearly, with the exception of E. pittsfordensis which will be considered presently, the Pittsford-Shawangunk fauna does not supply the ancestors for the Bertie fauna which is thus left without progenitors on the basis of the "lagoon-estuarine" theory usually advanced. There is also another difficulty. Stylonurus has representatives in North America in the Pittsford and in the Devonic and Carbonic, but none existed in the Bertie waters which should, according to the generally accepted views, have been the one place for the perpetuation of the race of the eurypterids in the late Siluric.

There is yet one other difficulty arising if the Pittsford-Shawangunk fauna was ancestral to the Bertie. How can the many points of similarity between certain Bertie and European species be accounted for? It has already been pointed out that Eurypterus remipes from the Herkimer region is so closely related to E. fischeri of the Baltic area that the latter for a long time was identified with the former. Schmidt was the first to suggest that the differences between the two species were only geographical variations arising through migration. With this idea Clarke and Ruedemann have concurred. In fact, they point out in many places the close similarity between the Bertie and Oesel fauna, and especially between the two commonest species in both, E. remipes and E. fischeri. Now, if the Bertie fauna was an estuarine one, preserved from Pittsford time in various brackish water bodies, when and how did migrations take place to the Baltic sea of the Upper Siluric? The answer will undoubtedly be that the members of the marine stock in the Lower Siluric which were not caught or did not voluntarily seek refuge in the "lagoon" or remnant of Niagaran sea in New York State, migrated along the shore of Atlantica, passing from estuary to estuary until they reached, the island of Oesel. This might seem like a very happy solution, if the British Isles did not intervene between America and Oesel, and if they did not have a very clear record to show that no such migration took place. In the discussion of the faunas of the various Palæozoic continents, given below, it will be shown that the Wenlock, Ludlow, and Lanarkian faunas of Great Britain offer no indications of migrations along the neritic zone during those periods and that in many cases new genera as well as new species arose suddenly without, apparently, having a genetic relationship to corresponding taxonomic units in other countries.

Since the assumption that the early Salina eurypterids lived in a "lagoon more or less cut off from the sea," leads to such difficulties, we must seek another theory. Let us assume that they lived in the rivers, and draw the logical deductions. It has been shown from their lithogenesis that the Pittsford and Shawangunk deposits must have been derived from Appalachia, while the Bertie was derived from Atlantica. Rivers, whether existing at the same time or at different geological periods, would carry related forms if coming from the same continent, but unrelated or only distantly related forms if coming from two different continents. Thus the Pittsford and Shawangunk eurypterids would be near relatives to say the least; while the fact that the larval forms from the latter are merely the young of those from the former is all the more according to our expectations.^[3] Likewise the absence of close relationship between the Shawangunk and the Bertie, and with one exception between the latter and the Pittsford, is easily understood and entirely to be expected. The baffling break in the phylogenetic history of Stylonurus is also explained. First found in the Pittsford, it came from Appalachia; on that continent its evolution continued through the remainder of Siluric time, its remains not being found because the continental, chiefly river flood-plain, deposits from Appalachia during the Upper Siluric and the Lower Devonic are unknown on the North American continent. The perplexities which were so detrimental to the "lagoon" theory are completely removed by the river theory. But if the latter be accepted, a new objection arises—only one, to be sure, yet at first it seems to demolish the theory altogether. How does it come to pass that E. pittsfordensis so closely resembles E. lacustris as to seem almost certainly the direct ancestor? In specimens approximately the same size the two species are found to be almost identical in the proportions of the cephalon (i.e., length:width = 2:3), and in the position and shape of the eyes. On the other hand, the posterior portion of the cephalon flares out in E. pittsfordensis or at least broadens out in a hyperbolic curve, while E. lacustris is marked by the nearly parallel sides of the cephalon. E. lacustris is not so broad a species as E. pittsfordensis, but otherwise does not differ especially in form. The telson in the latter species is unusually long, being nearly equal in length to the rest of the postabdomen.

An immature, but complete individual in the Buffalo Society of Natural Sciences Museum measures as follows:^[4]

	mm
Length of head	21
Length of body	68
Length of telson	57
	——
Total length	146

In another specimen which is incomplete, the telson measures 11.5 cm., while in E. lacustris, in an individual of about the same size, it measures only 6.5 cm. In spite of these differences, however, the species are very much alike, though not so closely related as E. lacustris, E. remipes, and E. fischeri, which can be understood from the fact that the three latter belong to the same horizon, while the former precedes them by a long period. I am quite prepared to agree with Clarke and Ruedemann that E. pittsfordensis is the ancestor of the Bertie forms. Not only this; it actually came from the same region as the later types. For it must be apparent that the rivers of Atlantica, which furnished the deposits of the Bertie, were also in existence during Pittsford time and must have mouthed into whatever remnant there was of the Niagaran sea. It is not particularly likely that the ancestors, if so we may call them, of the Upper Siluric rivers occupied precisely the same location as the Bertie or Herkimer rivers, but undoubtedly they existed in somewhat the same general region. Therefore, what is more likely than that during Pittsford time these southward-flowing rivers from the continent of Atlantica should bring down the remains of organisms living in them? These rivers could not themselves have supplied the muds of the Pittsford shales, for they came from a limestone region, and whatever sediments they carried must have been of the nature of waterlimes. If such calcareous deposits were spread out on the flood plains of those rivers they are now no longer visible, for subsequent erosion has removed all traces of deposits of Pittsford age in Canada; but there is where a eurypterid fauna would be expected to occur, just as in Bertie time when waterlimes were deposited farther south the fine eurypterid fauna is found. This explanation makes it entirely clear why E. pittsfordensis is related to no form yet known from the Shawangunk, but has characteristics showing that it was ancestral to forms in the Bertie. New discoveries have corroborated this theory.

Professor C. J. Sarle has discovered the Pittsford fauna at a new locality in New York State. The details of this have not yet been published, but it is known that both Eurypterus pittsfordensis and Hughmilleria are common. The rock is a gray shale and the material was undoubtedly supplied by the rivers of Appalachia. Since Hughmilleria is otherwise known only from deposits derived from Appalachia it is reasonable to assume that the same rivers which carried in the muds also brought in the Hughmilleria. The abundance of E. pittsfordensis is not surprising, for if the rivers from Atlantica emptied into the Pittsford basin there is no reason why they should not bring as abundant a fauna as did those from Appalachia. If, as is to be expected, the basin in which these deposits were laid down was at times a fresh water lake, the eurypterid faunas of both river systems may have met and lived for a time in this water body. They were then killed by the sudden incursion of the Guelph sea which brought with it the remnant of the Guelph fauna found in the intercalated limestone.

Further corroboration is offered by Van Ingen's discovery in Oneida county already referred to. In the concretionary block obtained from that locality and determined from lingulas and orbiculoideas in it to come from dark gray shales with intercalated waterlimes and dolomite beds 21 feet below the base of the red Vernon shale,^[5] were found three carapaces and fragments of a eurypterid, which Clarke and Ruedemann have named Eusarcus vaningeni. They state that "the outline of the body . . . . the visual surface . . . . the appendages, so far as seen, are like those of E. scorpionis. The tergites and sternites have the form and relative dimensions of E. scorpionis. . . . . The ornamentation is that of E. scorpionis, but the scales are smaller and more clearly arranged" (39, 420, 421). It is also somewhat related to E. cicerops from the Shawangunk, but the relation is generic rather than specific. That this species of Eusarcus, more closely related to a species in the Bertie than to a contemporary species in the Pittsford, should be found in the waterlime facies of the Pittsford rocks in a region but a few miles distant from the mouth of the subsequent Herkimer river, is a most unusual corroboration of our theory. It is exactly what could have been prophesied. How such an occurrence is to be explained on the lagoon theory is puzzling.

If the river hypothesis is the correct one it must account for the migration of the eurypterids from the Buffalo region to the Baltic during the Salinan or early Monroan. If we assume the existence of two rivers flowing from the rather low and flat limestone-covered country to the north, into a sea which had its shore extending through New York, as indicated on the map (fig. 8), it would not be difficult to understand that the shed exoskeletons of arthropods inhabiting the waters of these rivers and occasionally dead or even living individuals would be carried down stream, and become embedded in the fine lime sediment of the two neighboring deltas or in the interstream areas. Probably the eurypterids themselves were seldom carried down to the debouchures, since it is their molted exoskeletons which are generally found. To account for the similarity of the Buffalo and Herkimer faunules, it is necessary to postulate the interfingering of the headwaters of the Bertie and Herkimer rivers. The physical and faunal conditions would then be analogous to those existing at the present time in the Columbia and Missouri rivers, as outlined on p. 205 above. If we assume such a mode of distribution by rivers for the eurypterids, it would explain the close relationship which exists between forms isolated, but in neighboring localities; that is, Eurypterus lacustris of the Buffalo area, and E. remipes of the Herkimer area, nearly related species, but occurring in two isolated localities. But besides, these two occurrences, the river hypothesis must account for the close relation of both of these species to the one in the Baltic region (see below p. 235). There is good stratigraphic reason for believing that in Siluric time there was a continental mass (the Atlantica of Grabau), which as already outlined occupied much of the present North Atlantic and extended from northern North America entirely across to eastern Europe. According to Walther, several high mountain chains extended across this land connection (294, 251), and undoubtedly large rivers came down from these. Their headwaters would very probably interlace, as do those of all large rivers on the various continents at present.

Under such conditions we can see that the common ancestor of Eurypterus lacustris, E. remipes and E. fischeri could have lived in the headwaters of one of those rivers, and that getting farther away from the point of origin, the various species derived from it would be differentiated. E. lacustris and E. remipes were developed in two neighboring streams, but the forms connecting E. remipes and E. fischeri which must have lived in the rivers of Atlantica, are now buried under the waters of the North Atlantic Ocean. The more distant relationship of these species suggests that there were intermediate forms, though these have not yet been found, and are probably nowhere preserved, though it is not impossible that Siluric strata with such intermediate species may exist beneath the ice cap of Southern Greenland. In this great system of rivers, which to all appearances characterized the continent of Atlantica, the Bertie and Herkimer Rivers were not very far apart, so that the faunules of each were very similar. In fact, the deltas spread out at the mouths of the two rivers may have become confluent in their outermost or seaward portions, though the waterlime now known would, as above explained, represent only the inshore facies. It may have happened that in times of flood the river waters flowed out over a broader area near the debouchures until some of the distributaries became for a time confluent, thus allowing some of the species from one river to be carried over into the area of deposition of the other. Thus might the presence of Pterygotus cobbi in both regions be accounted for.

The Upper Siluric Faunas of the Baltic Region. Let us next consider the fauna from Oesel, Gotland, and the Baltic provinces of Russia. On Oesel three species and two varieties of eurypterids are known: Eurypterus fischeri Eichwald, E. fischeri var. rectangularis Schmidt, E. laticeps Schmidt, Pterygotus osiliensis,^[6] Schmidt, and P. osiliensis var. laticauda Schmidt. From Gotland the same Pterygotus species is reported, but no Eurypterus has yet been found. In Podolia a few specimens of Eurypterus fischeri, fragments of Pterygotus osiliensis occur, and Schmidt reports a few broken pieces of shell referable to the latter species in Galicia. From Livland, Pt. osiliensis has been reported by Eichwald. It is thus seen that in the Baltic Isles and West Russian provinces three species and two varieties of eurypterids occur. The close similarity, approaching identity, of Eurypterus fischeri to E. remipes and E. lacustris from the Bertie has been dwelt on at length (p. 230 above); the variety E. fischeri rectangularis naturally has its closest affinities with the Bertie forms. Schmidt described E. laticeps from two carapaces and did not compare it with any other form. There is no species in the Siluric fauna of Great Britain to which it shows any relationship, and so far as I am aware it cannot be compared with any other European form; but it shows considerable resemblance to E. microphthalmus from the Manlius waterlime. The largest specimen of the latter species measures 30 mm. long by 45 mm. wide, while one of the two known carapaces of E. laticeps shows corresponding measurements of 40 mm. and 60 mm., the ratio in both cases being as 2 to 3. The form of the eyes corresponds quite closely in the two species, but whereas in E. microphthalmus the distance between the eyes is almost equal to that between the eye and the lateral margin, in E. laticeps, on the other hand, the eyes are more widely spaced so that the distance between the eyes is one and a half times as great as between each eye and the margin (Schmidt, 248, 63). No ornamentation has been observed on the carapace of E. microphthalmus, but on E. laticeps a series of black dots occur in rather regular arrangement between the eyes, extending forward toward the frontal margin and posteriorly a shorter distance. Since both of these species are as yet so little known, it is not safe to draw conclusions as to their relations. The fact of chief interest is that the Baltic form is more closely related to the Manlius species than to any other. There remain the specimens of Pterygotus osiliensis and its variety, laticauda.

In the lower beds of the Old Red sandstone are two species of Pterygotus, bilobus and anglicus to both of which P. osiliensis shows some similarity, though the stronger affinity is to the latter of the two from Great Britain. A comparison of Schmidt's restoration of P. osiliensis and of an actual specimen of P. bilobus var. inornatus

Fig. 22a. Pterygotus osiliensis Schmidt. Restoration
(After Schmidt. 1883, p. 72, fig. 1 A)
Fig. 22b. Pterygotus bilobus var. inornatus (Salter). ✕ ${\tfrac {1}{2}}$.
(After H. Woodward. 1878, pl. X, fig. 1)
Fig. 22c. Chela of Same Species. ✕ ${\tfrac {1}{2}}$.
(Ibid. fig. 2)

(fig. 22), brings out the similarity in general form, the correspondence of the telsons especially in their bilobate character, the agreement between the pincers and the arrangement of teeth in the chelae, the similarity in the shape of the carapace and in the size and form of the swimming paddles. The abdomen of P. bilobus is not so narrow nor so gracefully tapering as is that of P. osiliensis; the proportions of the carapace likewise differ, that of the former species being longer than that of the latter. The similarities, however, are pronounced, and it is not to be denied that P. osiliensis finds its nearest relative in P. bilobus. Aside from this species, there are several others belonging to the subgenus Erettopterus, and we cannot dismiss our comparative study without calling attention to them. They are: Pterygotus (Erettopterus) grandis, and globiceps from North America; but they are so very distinct from the Baltic form that genetic relationship is in no way indicated.

The variety laticauda of P. osiliensis is founded not without certain misgivings on the part of Schmidt for an exceptionally large metastoma and a similarly large telson found associated with the P. osiliensis specimens. So far as the present problem of the determination of the relations between faunas is concerned, this variety would be classed along with P. osiliensis, and needs no separate discussion.^[7]

The species of faunas of the Upper Siluric waterlimes of Oesel, Gotland, Livland, Podolia, and Galicia are thus seen to show very close relationship either to species in the Bertie waterlime, as in the case of Eurypterus fischeri, and var. rectangularis, or to species found in Great Britain at the end of the Siluric or the beginning of the Devonic. That is, they show affinities to the faunas occurring in deposits which for reasons other than faunal ones were judged to have been derived from the continent of Atlantica.

The Fauna of the Wenlock. There now remains only the discussion of the eurypterid-bearing deposits of Great Britain, (6) the Wenlock of Scotland, and (7) the Old Red sandstone. In the Wenlock beds there is a large fauna represented by at least twelve determinable species of eurypterids, and one would expect to be able to attain to some critical knowledge of the relationship of the forms there occurring to those in North America and Europe; but while the fauna lacks not in the number of species and of individual remains, complete or even nearly complete specimens are not to be found, and one is forced to attempt to draw conclusions concerning relationships from fragments of legs, carapaces, or body segments, an attempt which is not only difficult but altogether unsatisfactory because of the probable errors attending it. Let us, however, consider the species seriatim, bearing in mind that details in structure are in most cases unavailable and that consequently genetic relationships are obscured. Of the genus Stylonurus, three species have been described by Laurie: S. elegans, S. macrophthalmus, and S. ornatus. The first species has been placed by Clarke and Ruedemann into the subgenus Ctenopterus, together with S. cestrotus Clarke, and S. multispinosus Clarke and Ruedemann; the former from the Shawangunk, the latter from the Pittsford, the subgeneric characters being the relatively greater length of the second and third pairs of legs when compared to the first, and the presence on the former of more than two pairs of long, slightly curved spines, which are vertical on the lower side of the segments (Clarke and Ruedemann, 39, 286–287). The Scottish species is so different from the two American forms grouped with it that the author is tempted to take exception to their being placed in the same subgenus, particularly because the very characteristics which are mentioned as diagnostic are not always observable. My reasons for objecting to the subgeneric grouping of this form under Ctenopterus are as follows: (1) It is unsafe to base a taxonomic group of such great value as a subgenus upon the characteristics of one set of organs alone, as for instance, the legs. Nothing at all is known of the body of S. multispinosus and very little about that of S. elegans; only that of S. cestrotus having been found in good enough preservation to allow of restoration. (2) Single identical morphological characters do not of themselves establish specific relationship and, therefore a fortiori they cannot be used to unite their possessors into groups of higher taxonomic value for it is a law of palaeontology which is coming more and more to be recognized, that the same morphological characters crop out in many diverse phyletic groups and their presence in no wise indicates genetic relation. Thus, a modification in the proportions of the legs or in the number of spines cannot be considered characters of subgeneric rank. (3) The length, breadth, general form and grouping of the spines on the second aud third pairs of legs are not at all similar in S. elegans and S. cestrotus (fig. 23). The comparatively short spines of about equal length, regularly spaced, and projecting at almost right angles from the walking legs, in S. excelsior (provided the restoration of this species is correct) and in S. cestrotus, together with the greater length in the second and third pairs of legs as compared with the first pair, might allow of these two species being placed in the same subgenus, and with them quite probably S. multispinosus. S. elegans, however, is too distinct, it seems to me, to be considered even subgenerically related, while specifically this species must certainly stand alone. This is especially evident when we consider (4) That one of the two diagnostic characters of the subgenus Ctenopterus depends upon the comparison of the lengths of the first three pairs of legs, the particular comparison

Fig. 23

a. Stylonurus elegans Laurie. Secon and third legs on right side. (After Laurie, 1900, pl. II, fig. 12.)
b. Stylonurus cestrotus Clarke. Second and third legs on left side. (After Clarke and Ruedemann, 1912, pl. XLIX, fig. 4.)

being made between the first, and the second and third pairs, but in S. elegans the first pair is unknown.

Stylonurus ornatus and S. macrophthalmus are in some respects quite closely related to species occurring in the later Scottish horizons in connection with which they will be spoken of again. Here it is sufficient to note that there are no North American species which have the characters of the genus Stylonurus (sens. str.) namely, the first three pairs of legs relatively short and stout, with only two short, curved spines on each segment, as in Drepanopterus and Eurypterus.

From the Wenlock, Laurie has also described three species of Eurypterus: E. conicus, E. minor and E. cyclophthalmus. These are three small species which are not very well represented and which are primitive or retarded in development. They are not related to any American forms nor do they appear to fill ancestral positions, for the British species of the Upper Siluric and Devonic. In some one characteristic a Wenlock species seems to foreshadow a later one, but phyletic lines are difficult, if not impossible, to trace. The exceedingly large eyes in the single known specimen of E. cyclophthalmus, and in E. conicus, and the small size as well as the general form suggest that these two species are larval forms. Clarke and Ruedemann consider that E. minor also is either immature or has had its development arrested. They think that such is especially the case in Bembicosoma pomphicus Laurie, a small, stunted form with large head, rapidly tapering body, and "warty texture of skin."

The genus Drepanopterus Laurie is now placed by Clarke and Ruedemann as a subgenus of Stylonurus. To this group belong Laurie's three species: D. bembicoides, D. lobatus, and D. pentlandicus, which need not be discussed in detail since they show no affinities either to American or to continental European forms. The species described by Laurie as Eurypterus scoticus has since been revised by Clarke and Ruedemann who recognized its affinities to Eusarcus. In the American faunas it finds its nearest representative in E. scorpionis from the Bertie. Because of the impossibility of making accurate measurements of the proportions of different parts of the bodies and of obtaining exact outlines to show the form, one is unable to make careful comparisons.

The only remaining species in the Wenlock eurypterid fauna is Slimonia dubia Laurie, a small individual, much broken and without appendages. Laurie has included in this species a second individual which shows a portion of the telson. Since the genus at present comprises only two species, the one just mentioned, and S. acuminata from the Ludlow, there is no opportunity to trace relationships over broad areas. The main reasons for making a new species of the Wenlock Slimonia, were, the difference in geologic age between the two forms, and the fact that the Pentland Hills individual was gradually tapering instead of abruptly contracted in the seventh segment into a telson.

Summary of the Wenlock Fauna. A survey of the entire eurypterid fauna of the Wenlock of Scotland must be made with the realization beforehand that all of the material is so fragmental, dismembered, macerated, and poorly preserved that detailed descriptions, accurate measurements, and unimpeachable determinations are things beyond the power of anyone to obtain, and that, furthermore, until discoveries of new faunas at earlier horizons in Great Britain shall be made, the ancestry of the Wenlock species must remain obscure. Many new genera appear suddenly in this Lower Siluric horizon, and we are unable to do more than say that such and such genera came from a common ancestor. It is unfortunate, indeed, that the Ordovicic of Great Britain has not yielded such faunas as it has in America. Yet, keeping these points in mind, we are still struck by the provincial character of the Wenlock fauna. There is not a species in it which is closely related to any of the North American species except Eusarcus scoticus which foreshadows in certain respects E. scorpionis from the Bertie.

The Fauna of the Ludlow. The Ludlow of Lanarkshire has yielded nine species of eurypterids. Slimonia acuminata Salter has just been mentioned in connection with S. dubia, the two being very similar. S. acuminata, Clarke and Ruedemann state, "has all the features of a local and aberrant type," (39, 130). Pterygotus (Erettopterus) bilobus with the four varieties: acidens, crassus, inornatus, and perornatus is found abundantly at Lesmahagow, the last variety, however, being very rare. As was pointed out in the discussion of the Baltic provinces faunas, there is closer relationship between P. bilobus inornatus and P. osiliensis than there is between either of these forms and a species in any other fauna (p. 238 above). Stylonurus logani belongs to the revised Stylonurus sens. strict., having the second and third pairs of legs short, thick, and with two pairs of spines in each segment (see Woodward, 312, 131). There are no known species on any other continent to be compared to this form which is not even very much like any of the Wenlock species, with two of which it agrees in its subgeneric characters, but with neither of which it has specific similarities. Indeed, it is quite unlike S. macrophthalmus which is characterized by the peculiar ear-shaped epimeral expansions, the long parallel-sided metastoma, the rounded cephalon, and very short second pair of legs. It is a little more like S. ornatus which has a slightly more squarish cephalon than S. macrophthalmus, and which has not such pronounced ears on the epimera, although these are extended posteriorly to a much more pronounced degree, than in S. logani.

The only species which has epimera approaching in size and form those of S. macrophthalmus is S. scoticus from the Old Red sandstone (see p. 251 below), from which, however, it differs in certain important features. It is closest to a second species found in the Old Red sandstone, S. powriei, which it resembles in the tapering form of the body the long, narrow telson, the subquadrate outline of the head (this is decidedly square in logani) and in the great length of the fourth appendage. In details, on the other hand, these two species differ considerably, so that S. logani must remain a rather separated species until new discoveries reveal its relatives. It is of great interest to have reported from the Ludlow fish bed in one of the tributaries of Greenock Water (see p. 164 above), Stylonurus ornatus associated with the typical Lanarkian (Downtonian) fishes and with Eurypterus dolichoschelus, a Ludlow and Lanarkian species, together with Ceratiocaris, Dictyocaris, plants, etc. (p. 164 above). S. ornatus, then, evidently persisted from Wenlock into and through Ludlow time. In this case one is again confronted with an anomalous geographic and geologic distribution. The Pentland Hills are less than thirty miles distant from the Lanarkian inliers, and the two areas are approximately on the same line of strike. In two thin beds but a few inches in thickness and extending only a few yards laterally S. ornatus occurs in the Wenlock in Lanarkshire; but in the Pentland Hills this species occurs in none of the many Ludlow eurypterid horizons until the fish bed is reached and there a few specimens are found. If the eurypterids lived in the Wenlock sea as they are commonly supposed to have done, then the supporters of this view must account for the limited vertical and horizontal distribution of the merostome remains, since it is absolutely inconceivable that members of a marine neritic fauna should be confined to an area a few square yards in extent. It is equally inconceivable that a marine fauna should be perpetuated for so great a period of time as from the Wenlock through the Ludlow, the members of the later fauna in some cases showing resemblance to members in the earlier, while in others they are entirely distinct and apparently arise suddenly, there being, besides, no indications of a persistent marine stock to furnish decendants from the Wenlock fauna, nor yet any trace in the maine Ludlow of the incursion from other regions of new genera and species of eurypterids. Moreover, we can not understand why one species of the Wenlock recurs in the Upper Ludlow, but does not occur in the beds of intermediate age, although there are many such with good marine molluscan faunas and even with fragments of other eurypterid species. Such a perpetuation, to repeat, would be impossible if the eurypterids were not having a continuous existence in the sea. But their remains are at all times spasmodic in appearance, being altogether wanting in certain horizons, especially where the typical marine fauna is abundant. The fact that they occur in a given band which, when traced even a short distance laterally, shows no lithological change, but only an absence of eurypterids, indicates that migrations along shore were non-existent; while the fact that new species and even new genera appear at horizons far separated from underlying and overlying eurypterid horizons seems to deprive "marine" eurypterids of ancestors or descendants, while to account for a marine Stylonurus ornatus in the Wenlock of Lanarkshire and in the uppermost Ludlow of the Pentland Hills, is not within the inventive powers of the author. But, on the other hand, the conditions of bionomy in rivers are eminently satisfactory to account not only for the persistence of a species for a long period of time without morphological modifications of specific rank, but also for the development of new species and genera, and for their sudden appearance. This takes place, because they have been developing either in other river systems, whence they have migrated to the headwaters of the river at the mouth of which their remains are found, or because they have been traversing a great distance in longitude, automatically suffering specific variation in their progress. In this way, would I account for the anomalies in distribution just dwelt upon (see also p. 203 et seq.).

Of this Ludlow fauna there still remain four species to be considered. There are three species of Eusarcus which may be taken up at the same time: E. scorpioides, E. obesus, and E. raniceps. The last species may be quickly dismissed, since it is represented by a single specimen showing only the carapace and a part of the abdomen, enough, indeed, to place the individual generically; but specific comparisons are impossible. E. scorpioides is represented by one almost entire individual, a large, robust form in many respects similar to E. scorpionis from the Bertie waterlime. The length and width of the appendages, the number and disposition of spines thereon, the ratio of length of carapace to the remainder of the body, and the proportions generally agree in the two species. Even more like E. scorpionis is the single known specimen of E. obesus from the same Lesmahagow horizon, and it has been suggested by Woodward, who described the species, that E. obesus may possibly represent the young of E. scorpioides; certainly E. obesus looks very much like a young individual of E. scorpionis figured by Clarke and Ruedemann from the Bertie (Figs. 24 and 25). Thus there is close relationship between the two species from Lanarkshire and the one from the Bertie.

The last species from the Ludlow fauna, and the only Eurypterus yet found therein is E. lanceolatus Salter. As Sarle, Clarke, and

Fig. 24. Eurypterus obesus H. Woodward. ✕ ${\tfrac {9}{10}}$. (After Woodw. 1878, pl. XXX, fig. 8)		Fig. 25. Young of Eusarcus scorpionis Grote and Pitt. ✕ ${\tfrac {1}{2}}$. (After C. & R. 1912, pl. XXXVI, fig. 1)

Ruedemann have pointed out, this species has many points in common with Hughmilleria and either belongs to that genus or is transitional to it. The form of the body, shape of the carapace and of the telson, marginal position of the eyes, the relative proportions of the somites, and details in the appendages, all point to affinities with Hughmilleria socialis Sarle, from the Pittsford (figs. 26, 27). Such a relationship seems a little disconcerting at first, in view of the fact that the Pittsford sediments and fauna came from Appalachia, while the Ludlow was a derivative from Atlantica and should have a fauna essentially distinct from the former. Indeed, with the exception of this one species, the members of the Ludlow fauna show no relationship to any species from the faunas of Appalachia. We have here, as a matter of fact, one of the "anomalies" of distribution which may occur among fluviatile organisms, but are inexplicable for marine forms. In the upper Niagaran in North America H. socialis occurs by the hundred in the Pittsford shale and the closely related H. shawangunk, which may be only the young of the former species, occurs in the synchronous shales of the Shawangunk, but in no other part of the world at that time, so far as we know, were there any representatives of Hughmilleria. It appears that the genus originated

Fig. 26. Eurypterus lanceolatus Salter. ✕ ${\tfrac {1}{2}}$. (After Woodw. 1878, p. 142, fig. 44)		Fig. 27. Hughmilleria socialis Sarle. ✕ ${\tfrac {1}{2}}$. (After Cl. & R. 1912, pl. LIX, fig. 1)

in the rivers of Appalachia. Curiously enough, in the Upper Ludlow, that is, lower Upper Siluric, of Scotland, Eurypterus lanceolatus appears, showing a striking resemblance to Hughmilleria socialis. The prolific Scottish fauna of the Wenlock has revealed no possible ancestors for this distinctive Eurypterus (or Hughmilleria?) and one naturally wonders how it arose. Since Hughmilleria was restricted in occurrence in the Niagaran, any migrations which took place must have been effected during the Salina period. The important Salina break or period of emergence has recently been recognized by Grabau as affecting all countries bordering on the North Atlantic and has been recorded for North America, Scotland, Oesel (by the author), and even in what was formerly supposed to be the continuous section in England (89b). There was a widespread diastrophic movement at the end of the Niagaran marking a broad expansion of continental areas during Salina time so that perhaps nowhere are there preserved to us the marine sediments of that period. Certainly the North American eurypterids were cut off from marine routes of migration with which most authors like to provide them, and yet migration seems to have gone on. The Salina in North America was a period of aridity west of the mountain mass of Appalachia, but that chain died out northward and probably merged into the continent of Atlantica, there being no northeast Atlantic sea-lobe at that time. Several possible lines of fluviatile migration were open and nothing is more probable than that emigrants from Appalachian north and northeast flowing rivers should have entered some one of the tributaries of the systems on Atlantica. The exact mode of transit can not be determined, but many routes were open. Indeed, it is possible that migration occurred even in Pittsford time from the rivers of Appalachia into one of the Pre-Bertie rivers which we have seen probably existed in the western New York region even during the Niagaran (p. 113 above). This much we may conclude: There were many routes and possibilities of migration open to eurypterids living in the rivers of Appalachia during the Lower and Middle Siluric, but continuous marine paths to Europe were non-existent. Furthermore, the distinctness of the Ludlow fauna as a whole from any of the faunas of Appalachia, but the close relationship of one species from the former to two from the latter is inexplicable for members of a marine fauna, but normal and expectable for members of fluviatile faunas.

The Old Red Sandstone Fauna. The last of the European formations which is believed to have been derived from the continent of Atlantica is the Old Red sandstone. Most the eurypterids occur in the beds in various localities in Forfarshire. By far the most abundant species is Pterygotus anglicus which finds it nearest relatives in Pterygotus bufaloensis and P. macrophthalmus from the Bertie, and P. osiliensis from the Baltic region. The various points of similarity are so well known that it is not necessary to take them up. Pterygotus minor is a small form found associated with P. anglicus, but it is a unique form, and shows a marked divergence from congeneric forms throughout the world. The telson is elongate, spatulate, with a pronounced median keel, which is represented on the last three segments of the postabdomen as a long spine, rather than a ridge (Woodward, 312, 199, Pl. X, fig. 2; 195, p. 35, Pl. I, fig. 4). There can be little doubt that this species, which is represented by a single, nearly entire individual, represents a neanic stage of some form, the adult of which probably is not known. Only two and a half inches long, it has the large eyes slightly removed from the border, a feature which is so characteristic of neanic Pterygoti; but it is difficult to account for the pronounced spines on the body segments, and for the high keel, features, which in associated species are less developed at so early a stage. The shape of the carapace and the position of the eyes suggest P. macrophthalmus from the Bertie, but the spines on the epimera of the last five segments of the postabdomen, the median spine on the last three, the very marked median keel on the telson as well as the proportions of the telson indicate a specialization far beyond that observable in the species just mentioned, particularly when it is borne in mind that all of these features are observed in an undoubtedly young individual, which means that they would be much more marked in maturity. This species has all of the appearances of an aberrant form, the relations of which it is impossible to determine from the one known specimen, but it certainly has characters which unite it with Bertie species and with formr which occur in the Baltic region.

The Stylonuridæ of the Old Red sandstone are represented by four species: S. scoticus, S. powriei, S. ensiformis, and S. symondsii. The first, represented only by a head and by one nearly entire individual, is yet so remarkable, so entirely distinct from the typical Stylonurus that it has been set apart by Clarke and Ruedemann as the representative of a new subgenus, Tarsopterus. These two authors have dwelt upon what they consider the close similarity between S. scoticus and S. myops from the Shawangunk, stating that "it seems probable, therefore, that S. myops, when fully known, will prove a representative of the subgenus Tarsopterus of which S. scoticus is the type" (39, 303). The reasons which they cite are: occurrence of "spurlike epimera of equal relative size," the "outline of carapace," and "the approximate position of the eyes and the sculpture of the tergites." Since it is my purpose in the present section of this paper to marshal all of the evidence provided by the relationship existing between the genera and species of different faunas in order to determine from which continents these were derived, it is evident that a claim of close similarity between a species in the Shawangunk fauna, derived as I believe from Appalachia, and a species in the Old Red sandstone derived from the continent of Atlantica, as I hope to prove, must be carefully investigated. Therefore, I proceed to the points enumerated, always bearing in mind that certain types of similarity are of more value than others. In the beginning I may state that S. myops is known only from immature specimens, most of which are carapaces alone, and that only one entire specimen has been found and this is but 55 mm. in length (see pl. 52, fig. 6, Clarke and Ruedemann). The largest carapace of S. myops observed measured 19 mm. in length by 27 mm. in width; the only carapace of S. scoticus known measured 16 cm. in length and 19 cm. in width; the single, entire individual known measured 3 feet, 4 inches in length. A most profound difficulty arises at once, namely, that of comparing neanic and nepionic specimens of a mid-Siluric species with a gerontic, or perhaps a late ephebic individual of the Lower Devonic. But granting that such comparisons are possible or even allowable, let us turn to the characters which would justify placing these two species in the same subgenus. First, there is the outline of the carapace. It must be admitted even by Clarke and Ruedemann that, with all due allowance for compression, the carapaces of S. myops display a most unusual amount of variation in outline, some, were it not for the position,of the eyes, being easily referable to Eurypterus. The carapaces show a strong tendency to grow narrower posteriorly, showing the greatest width in the anterior third, whence the lateral margins slope gently backwards; the nearly parallel sides shown in the carapace of S. scoticus are usually not present in S. myops, while the frontal

species	length of carapace in mm.	breadth of carapace	ratio: length: breadth
S. myops, smallest carapace observed	003.5	005.5	0.63
S. myops, type	012.3	016.5	0.74
S. myops, largest carapace observed	019.0	027.0	0.70
S. macrophthalmus (Ludlow)	051.0	0061.0	0.83
S. powriei	050.0	062.0	0.80
S. scoticus (separate carapace)	160.0	190.0	0.84
S. scoticus (carapace attached to body)	204.0	242.2	0.83

margin is quite as likely to be curved as to be nearly flat (as in S. scoticus). A comparison of the proportions of length to breadth of carapace in these two species and in two others with which relationship might more readily be established will, when taken in connection with the illustrations, show that S. scoticus in so far as its carapace is concerned, is far more nearly related to associated forms in the Old Red and to others in the Ludlow, than to the Shawangunk forms.

From these figures we may conclude that Clarke and Ruedemann find the approximate ratio of length to breadth of carapace in S. myops to be as 2 : 3, but it is evident that in S. scoticus it is 4 : 5. It is not to be denied that the ratio changes from that in the young of S. myops where it is 2 : 3 to that in the type where it is nearly 3:4, and perhaps it might be conceded that in larger forms the ratio might approach 4 : 5; but we cannot be sure. There is in the Ludlow, however, a species which has a carapace proportioned exactly as in S. scoticus and even in the Old Red is a species, S. powriei, with proportions almost the same. Thus there is no need to form conjectures about what might be possible relations to a Middle Siluric species from Appalachia when there are forms which actually show the similarity in formations derived from the same land-mass.

A second point of supposed similarity between S. scoticus and S. myops was the occurrence of long and pronounced epimera in both species. I have in another part of this paper discussed the significance of spinous prolongations on the epimera, but I shall call attention to the arguments again, since they are not universally recognized. Beecher has assembled a wealth of illustration from all branches of the animal kingdom to show that the appearance of spines as a modification of any morphological character marks degeneration in respect to that character, and, when extreme spinosity is accompanied by certain other easily recognizable and similarly degenerate characters the species, genus or family, all members of which show like degeneration, is doomed to decline and extinction. But not only that; as Beecher, Hyatt, and a few present-day palaeontologists, notably Grabau, have shown and have demonstrated by countless illustrations which have led to the most certain deductions, the formation of spines is a homeomorphic character, not in the least indicative of genetic relationship in forms which develop such spines, but marking only an onto- or phylogenetic stage. Spines may and do appear in end-members of totally distinct phyletic groups and are of absolutely no diagnostic value in determining true relations. The Eurypteridae offer many new illustrations of this law which is so simple, which so strongly makes its appeal to the reason, and which yet is so constantly ignored. The Carbonic species of Eurypterus develop spines wherever possible; the surface scales are produced into pointed wedges or spines; the ends of the epimera grow out to a great length; spines develop on the appendages not only in rows along the various segments but also on the lines of junction between segments: showing that the final expression of morphological characters in the eurypterids was the development of spines which was followed by extinction. Such a development has seemed expectable to many authors for the species living in the late Palæozoic, in the Mississippic, and Carbonic; but there is really nothing to prevent these phylogerontic characters from appearing much earlier. And so, to apply all of these general statements to the case in question, I would say that the epimeral spines observable on S. myops indicate that the line which that species represented was on the decline even in the Siluric, at a time when the majority of eurypterids were at their acme. A glance at the illustration of S. scoticus (Woodward 312, Pl. XXII) will show to the reader that this species has a typically gerontic appearance. Its epimeral prolongations do not in the least resemble those in S. myops, but are most like those of S. macrophthalmus from the Ludlow.

Two points remain as supposedly indicative of relation between these two species. The position of the eyes is, it seems to the author, the only feature of marked similarity, but certain of the British forms also show such a position, so that it is not of striking importance. As for the ornamentation of the tergites, I can see little to warrant the statement that the sculpture is similar in the two species.

The species Stylonurus (Tarsopterus) scoticus has now been compared in detail with S. myops and it has been shown that they are not closely related and consequently the presence of the first genus in the Old Red sandstone not only does not militate against my thesis that the faunas living in rivers coming from the same continent and in the same latitude should be most alike, but it is actually an additional proof, for S. scoticus is most nearly related to Ludlow and Old Red species, though it shows phylogerontic characteristics which somewhat obscure its relations.

The three remaining species of Stylonurus from the Old Red may be quickly dismissed. S. symondsii, from England, is represented by a single apparently complete carapace which is almost as long as wide, but is distinctly narrower posteriorly than anteriorly. There is a possibility that the marginal fold has been destroyed in the posterior portions, but Woodward thinks that the specimen is entire, and that the fold did not pass all the way around the carapace. S. ensiformis is described from a single broken tail spine which, it seems to the author, is hardly sufficient for the founding of a new species, and certainly is of no use in determining the affinities of the fauna. S. powriei, represented by a single individual, has a carapace very similar in form and identical in proportions to S. scoticus, from which species it differs most noticeably in having the last pair of appendages long and tapering, not short and broad. Woodward has suggested that it probably had epimeral prolongations which have not been preserved, because only the internal mold in sandstone has been found, and the epimera would be likely to remain with the actual integuments; for the same reason none of the surface markings are visible. The tail is extremely long and narrow, quite similar to the telson of S. logani from the Ludlow, which form it also resembles in the character of the last pair of appendages. Both species belong to the provisional group of Stylonurus s. st. recognized by Clarke and Ruedemann.

Completing the Old Red sandstone fauna are two species of Eurypterus: E. brewsteri and E. pygmaeus. The first consists of a carapace, a portion of a thoracic segment slightly displaced, and an ovisac containing more than twenty ova (Woodward, 312, 151). Woodward says that "this species agrees most nearly in general form with E. lacustris" from the Bertie, while Clarke and Ruedemann have pointed out a close similarity to E. microphthalmus from the same horizon (39, 195). But since both authors make their comparison on the form, proportions of length or width, and position of eyes, and since the actual figures do not support either statement, I find it impossible to agree with them.

	length of carapace	width	ratio l:w
	mm
Eurypterus brewsteri	01.48	05.52	0.27
E. lacustris	44.00	63.00	0.70
E. microphthalmus, type	15.50	22.00	0.70
E. microphthalmus, best preserved specimen	17.50	27.40	0.64

E. pygmaeus is a small form found near Kington, England, and though represented by very young individuals, yet has characters which point to its affinities with E. remipes (Fig. 28).

Summary of Facts of Distribution on Continent of Atlantica. We are now enabled to bring together all of the many lines which we have been following in tracing the affinities of the faunas which for other reasons were supposed to have come from the continent of Atlantica, and here, as in the case of the faunas of Appalachia, the great weight of evidence shows that the Bertie, Rondout, Manlius, Ludlow, Lanarkian, Baltic, and Old Red faunas are more closely related inter se than they are to the faunas which from the study of the petrogenesis of the formations in which they occur, were believed to have come from other continents.

The Eurypterid Faunas of Mississippia. So far only a single fauna is known from the continent of Mississippia, and therefore it is not possible to institute any comparisons between the species found in that fauna and those from other faunas on the same continent, as was possible in the case of Atlantica and Appalachia; the most

Fig. 28. Eurypterus pygmaeus Salter. ✕ 1
(After Woodw. 1878, pl. XXVIII, fig. 5)

that can be expected is that we shall find the Kokomo eurypterids distinct from all those which lived in rivers on other continents. As we shall see, the theoretical expectations are fully borne out by the facts.

The Eurypterid fauna of the Kokomo waterlime is distinct from any of the known North American eurypterid faunas. The material is never well preserved and the number both of species and of individuals is small. "Stylonurus (Drepanopterus) longicaudus," says Clarke and Ruedemann, "is a unique form among the American eurypterids being the sole representative thus far found on this continent of this rare and phylo-genetically interesting genus. From its Scottish allies, it is readily distinguished by its slender and elongated postabdomen and the long, clavate telson." (39, 320) Four specimens are known, two young and two mature individuals, and though they are in sufficiently good condition to enable Clarke and Ruedemann to make a restoration of the species, they do not approach the perfection of preservation found in the Bertie material. The characters are clearly enough shown to make it a certainty that this form has no relatives in the American faunas, so far known. Five specimens of Eusarcus newlini are known. This species, though attaining the gigantic size of E. scorpionis of the Bertie, shows marked differences in the proportions of the body. There is a general shortening up and broadening throughout. A set of figures taken from Clarke and Ruedemann's discussion will bring out this fact; some of the figures are only approximate.

Lengths in millimeters
	cara- pace	preab- domen	postab- domen	last post abdominal segment	telson	ratio of carapace to rest of body
E. scorpionis	53	67	146	40	62	0.17 : 1
E. newlini	58	57	112	34	43	0.23 : 1

It will be noted from these figures that although E. newlini, in the specimen measured, had a carapace 5 mm. longer than that of E. scorpionis the remainder of the figures for the other portions of the body are considerably less, showing that the proportions throughout are different. The ratio of the length of the carapace to the length of the rest of the body in the two species shows that in E. scorpionis it is as 0.17 : 1, while in E. newlini it is as 0.23 : 1. The cephalothoracic appendages are much stouter in E. newlini, with longer and stouter spines. Since the Bertie and Kokomo species of Eusarcus are the only ones in this country which are well enough preserved to allow of careful description, they are the only ones which can be compared and it has been shown that they do not show close relationship. The Kokomo fauna has yielded further two species of Eurypterus which are very similar, namely, E. (Onychopterus) kokomoënsis, and E. ranilarva. Of the difference between these two species Clarke and Ruedemann say: "It is possible that these differences are only those of sex, a point that at present cannot be determined since the opercular appendages of E. ranilarva are not distinctly shown" (39, 211). The proportions between the length and width of the cephalon in the Kokomo and Bertie forms are quite different. In E. ranilarva the ratio is as 7.1 : 10; in one specimen of E. kokomoënsis it is as 8 : 10, in another as 8.4 : 10, but in E. dekayi the ratio is only as 6 : 10 in one specimen and is even as low as 5.3 : 10 in another. From these figures it appears that the Kokomo forms had cephala which were much more nearly square than rectangular. A set of comparative figures for the proportions in the different parts of the three species brings out the differences clearly.

Lengths in millimeters
species		carapace	preabdo- men	postabdo- men	telson	ratio of cara- pace to rest of body
E. scorpionis		31	40.4	56.0	53.0	20.7:100
E. ranilarva	$\left\{{\begin{aligned}\\\end{aligned}}\right.$	35	40.0	50.0	35.0	27.8:100
		33	43.5	41.5	36.0	27.0:100
E. kokomoensis		28	32.3	43.4	30.4	25.9:100

The same relations hold here between the body proportions of the Kokomo and Bertie species as held in the case of Eusarcus. A comparison of the figures for E. dekayi and the first specimen of E. ranilarva shows that though the carapace of the latter is longer, all of the other parts of the body are shorter. Thus, the Eurypterus species as well as the one of Eusarcus are relatively shorter and broader forms than the ones found in the Bertie.

The Kokomo eurypterid fauna as a whole is quite distinct from any other American fauna, a fact which is difficult to explain on the theory of marine habitat for these organisms. If, as Clarke and Ruedemann have stated, the Kokomo is of Lockport age, and belongs to the marine fauna of that time, it is greatly to be wondered at that there should be no eurypterid fauna in the succeeding Guelph beds in the same locality or in adjoining regions. Yet the only Guelph form that has ever been found is the single specimen of Eurypterus (Tylopterus) boylei from Ontario, a form which shows not the slightest resemblance to any of the Kokomo eurypterids. If the Kokomo is to be considered of Monroan age, for reasons which have been given in full on p. 118 then, on the marine theory, the Kokomo forms should show relationship to the Bertie, and their area of deposition should constitute merely another "pool" cut off from the Monroan sea. But it has just been shown that the Kokomo fauna is quite distinct from the Bertie and that the two faunas have no species in common, a fact difficult to explain on the ground that the forms lived in neighboring "pools" where faunas were segregated from a once widespread marine fauna.

On the other hand, these peculiarities are easily understood, if we consider the eurypterids as fluviatile organisms. It is quite evident that the Kokomo deposits have a distinctly different source from those of the Bertie. If then, these eurypterids belong to a distinct river system, developed upon a separate land mass, it would indeed be surprising if they were not wholly distinct specifically from those of the rivers of Atlantica which were responsible for the Bertie waterlime deposits. The alternation of beds with marine fossils with beds carrying only eurypterids and ceratiocarids, suggests that the Kokomo deposits may have approached those of some modern estuaries in which we have an alternation of marine and fresh-water deposits. The map, Fig. 8, shows the position and general extent of these late Siluric river systems.

---

1. The importance of distinguishing between dispersal, the passive and migration, the active distribution of organisms has been insisted upon by Grabau (Principles, p. 1041).
2. Clarke and Ruedemann point out this similarity, but claim also that E. linguatus "strongly suggests the Eusarcus vaningeni" from the Salina, in position of eyes and shape of carapace (39, 414). A close examination of their descriptions and of all the figures they give does not reveal any marked similarity.
3. It should be noted here that the adult individuals of the Shawangunk were preserved only as unrecognized fragments, the young forms alone, by virtue of their small size, escaping the destruction which was meted out to their progenitors, as was discussed on p. 101.
4. These measurements were kindly furnished to me by Mr. Henry R. Howland.
5. I shall refer to these shales hereafter as the Farmer's Mills shales, from the locality near which they were found.
6. While it is not the intention of the author of this paper to revise or emend any generic or specific appellations of other authors except in so far as is necessary in the discussion of the problem at hand, it is advisable to call attention to the fact that Pterygotus osiliensis belongs to the subgenus Erettopterus established by Huxley and Salter for the Pterygoti which have bilobed telsons.
7. The author, however, questions the propriety of the creation of a new variety for the two specimens found. Undoubtedly the metastoma which Schmidt cites is larger than the two which he considers belong to the typical P. osiliensis; on the other hand, it is only slightly larger than would be required to fit with the operculum or with the thoracic segment which he figures on Plate V, figs. 1 and 3. The three metastoma are so similar in form and ornamentation that it seems rather dangerous to use mere variation in size, particularly when that is so expectable, and when various parts of the body indicate a species of no mean dimensions.