From time to time, when a land-connection between North America and Eurasia was established, some of the successive equine genera migrated to the Old World, but they do not seem to have gained a permanent footing there until the end of the Miocene or beginning of the Pliocene, eventually diversifying into the horses, asses, and zebras of Africa, Asia and Europe. At about the same period, the family extended its range to South America and there gave rise to a number of species and genera, some of them extremely peculiar. For some unknown reason, all the horse tribe had become extinct in the western hemisphere before the European discovery, but not until after the native race of man had peopled the continents.

In addition to the main stem of equine descent, briefly considered in the foregoing paragraphs, several side-branches were given off at successive levels of the stem. Most of these branches were short-lived, but some of them flourished for a considerable period and ramified into many species.

Apparently related to the horses and derived from the same root-stock is the family of the Palaeotheres, confined to the Eocene and Oligocene of Europe, dying out without descendants. In the earlier attempts to work out the history of the horses, as in the famous essay of Kowalevsky ("Sur l'Anchitherium aurelianense Cuv. et sur l'histoire paleontologique des Chevaux", "Mem. de l'Acad. Imp. des Sc. de St Petersbourg", XX. no. 5, 1873.), the Palaeotheres were placed in the direct line, because the number of adequately known Eocene mammals was then so small, that Cuvier's types were forced into various incongruous positions, to serve as ancestors for unrelated series.

The American family of the Titanotheres may also be distantly related to the horses, but passed through an entirely different course of development. From the lower Eocene to the lower sub-stage of the middle Oligocene the series is complete, beginning with small and rather lightly built animals. Gradually the stature and massiveness increase, a transverse pair of nasal horns make their appearance and, as these increase in size, the canine tusks and incisors diminish correspondingly. Already in the oldest known genus the number of digits had been reduced to four in the fore-foot and three in the hind, but there the reduction stops, for the increasing body-weight made necessary the development of broad and heavy feet. The final members of the series comprise only large, almost elephantine animals, with immensely developed and very various nasal horns, huge and massive heads, and altogether a grotesque appearance. The growth of the brain did not at all keep pace with the increase of the head and body, and the ludicrously small brain may will have been one of the factors which determined the startlingly sudden disappearance and extinction of the group.

Less completely known, but of unusual interest, is the genealogy of the rhinoceros family, which probably, though not certainly, was likewise of American origin. The group in North America at least, comprised three divisions, or sub-families, of very different proportions, appearance and habits, representing three divergent lines from the same stem. Though the relationship between the three lines seems hardly open to question, yet the form ancestral to all of them has not yet been identified. This is because of our still very incomplete knowledge of several perissodactyl genera of the Eocene, any one of which may eventually prove to be the ancestor sought for.

The first sub-family is the entirely extinct group of Hyracodonts, which may be traced in successive modifications through the upper Eocene, lower and middle Oligocene, then disappearing altogether. As yet, the hyracodonts have been found only in North America, and the last genus of the series, Hyracodon, was a cursorial animal. Very briefly stated, the modifications consist in a gradual increase in size, with greater slenderness of proportions, accompanied by elongation of the neck, limbs, and feet, which become tridactyl and very narrow. The grinding teeth have assumed the rhinoceros-like pattern and the premolars resemble the molars in form; on the other hand, the front teeth, incisors and canines, have become very small and are useless as weapons. As the animal had no horns, it was quite defenceless and must have found its safety in its swift running, for Hyracodon displays many superficial resemblances to the contemporary Oligocene horses, and was evidently adapted for speed. It may well have been the competition of the horses which led to the extinction of these cursorial rhinoceroses.

The second sub-family, that of the Amynodonts, followed a totally different course of development, becoming short-legged and short-footed, massive animals, the proportions of which suggest aquatic habits; they retained four digits in the front foot. The animal was well provided with weapons in the large canine tusks, but was without horns. Some members of this group extended their range to the Old World, but they all died out in the middle Oligocene, leaving no successors.

The sub-family of the true rhinoceroses cannot yet be certainly traced farther back than to the base of the middle Oligocene, though some fragmentary remains found in the lower Oligocene are probably also referable to it. The most ancient and most primitive member of this series yet discovered, the genus Trigonias, is unmistakably a rhinoceros, yet much less massive, having more the proportions of a tapir; it had four toes in the front foot, three in the hind, and had a full complement of teeth, except for the lower canines, though the upper canines are about to disappear, and the peculiar modification of the incisors, characteristic of the true rhinoceroses, is already apparent; the skull is hornless. Representatives of this sub-family continue through the Oligocene and Miocene of North America, becoming rare and localised in the Pliocene and then disappearing altogether. In the Old World, on the other hand, where the line appeared almost as early as it did in America, this group underwent a great expansion and ramification, giving rise not only to the Asiatic and African forms, but also to several extinct series.

Turning now to the Artiodactyla, we find still another group of mammals, that of the camels and llamas, which has long vanished from North America, yet took its rise and ran the greater part of its course in that continent. From the lower Eocene onward the history of this series is substantially complete, though much remains to be learned concerning the earlier members of the family. The story is very like that of the horses, to which in many respects it runs curiously parallel. Beginning with very small, five-toed animals, we observe in the successive genera a gradual transformation in all parts of the skeleton, an elongation of the neck, limbs and feet, a reduction of the digits from five to two, and eventually the coalescence of the remaining two digits into a "cannon-bone." The grinding teeth, by equally gradual steps, take on the ruminant pattern. In the upper Miocene the line divides into the two branches of the camels and llamas, the former migrating to Eurasia and the latter to South America, though representatives of both lines persisted in North America until a very late period. Interesting side-branches of this line have also been found, one of which ended in the upper Miocene in animals which had almost the proportions of the giraffes and must have resembled them in appearance.

The American Tertiary has yielded several other groups of ruminant-like animals, some of which form beautifully complete evolutionary series, but space forbids more than this passing mention of them.

It was in Europe that the Artiodactyla had their principal development, and the upper Eocene, Oligocene and Miocene are crowded with such an overwhelming number and variety of forms that it is hardly possible to marshal them in orderly array and determine their mutual relationships. Yet in this chaotic exuberance of life, certain important facts stand out clearly, among these none is of greater interest and importance than the genealogy of the true Ruminants, or Pecora, which may be traced from the upper Eocene onward. The steps of modification and change are very similar to those through which the camel phylum passed in North America, but it is instructive to note that, despite their many resemblances, the two series can be connected only in their far distant beginnings. The pecoran stock became vastly more expanded and diversified than did the camel line and was evidently more plastic and adaptable, spreading eventually over all the continents except Australia, and forming to-day one of the dominant types of mammals, while the camels are on the decline and not far from extinction. The Pecora successively ramified into the deer, antelopes, sheep, goats and oxen, and did not reach North America till the Miocene, when they were already far advanced in specialisation. To this invasion of the Pecora, or true ruminants, it seems probable that the decline and eventual disappearance of the camels is to be ascribed.

Recent discoveries in Egypt have thrown much light upon a problem which long baffled the palaeontologist, namely, the origin of the elephants. (C.W. Andrews, "On the Evolution of the Proboscidea", "Phil. Trans. Roy. Soc." London, Vol. 196, 1904, page 99.) Early representatives of this order, Mastodons, had appeared almost simultaneously (in the geological sense of that word) in the upper Miocene of Europe and North America, but in neither continent was any more ancient type known which could plausibly be regarded as ancestral to them. Evidently, these problematical animals had reached the northern continents by migrating from some other region, but no one could say where that region lay. The Eocene and Oligocene beds of the Fayoum show us that the region sought for is Africa, and that the elephants form just such a series of gradual modifications as we have found among other hoofed animals. The later steps of the transformation, by which the mastodons lost their lower tusks, and their relatively small and simple grinding teeth acquired the great size and highly complex structure of the true elephants, may be followed in the uppermost Miocene and Pliocene fossils of India and southern Europe.

Egypt has also of late furnished some very welcome material which contributes to the solution of another unsolved problem which had quite eluded research, the origin of the whales. The toothed-whales may be traced back in several more or less parallel lines as far as the lower Miocene, but their predecessors in the Oligocene are still so incompletely known that safe conclusions can hardly be drawn from them. In the middle Eocene of Egypt, however, has been found a small, whale-like animal (Protocetus), which shows what the ancestral toothed-whale was like, and at the same time seems to connect these thoroughly marine mammals with land-animals. Though already entirely adapted to an aquatic mode of life, the teeth, skull and backbone of Protocetus display so many differences from those of the later whales and so many approximations to those of primitive, carnivorous land-mammals, as, in a large degree, to bridge over the gap between the two groups. Thus one of the most puzzling of palaeontological questions is in a fair way to receive a satisfactory answer. The origin of the whalebone-whales and their relations to the toothed-whales cannot yet be determined, since the necessary fossils have not been discovered.

Among the carnivorous mammals, phylogenetic series are not so clear and distinct as among the hoofed animals, chiefly because the carnivores are individually much less abundant, and well-preserved skeletons are among the prizes of the collector. Nevertheless, much has already been learned concerning the mutual relations of the carnivorous families, and several phylogenetic series, notably that of the dogs, are quite complete. It has been made extremely probable that the primitive dogs of the Eocene represent the central stock, from which nearly or quite all the other families branched off, though the origin and descent of the cats have not yet been determined.

It should be clearly understood that the foregoing account of mammalian descent is merely a selection of a few representative cases and might be almost indefinitely extended. Nothing has been said, for example, of the wonderful museum of ancient mammalian life which is entombed in the rocks of South America, especially of Patagonia, and which opens a world so entirely different from that of the northern continents, yet exemplifying the same laws of "descent with modification." Very beautiful phylogenetic series have already been established among these most interesting and marvellously preserved fossils, but lack of space forbids a consideration of them.

The origin of the mammalia, as a class, offers a problem of which palaeontology can as yet present no definitive solution. Many morphologists regard the early amphibia as the ancestral group from which the mammals were derived, while most palaeontologists believe that the mammals are descended from the reptiles. The most ancient known mammals, those from the upper Triassic of Europe and North America, are so extremely rare and so very imperfectly known, that they give little help in determining the descent of the class, but, on the other hand, certain reptilian orders of the Permian period, especially well represented in South Africa, display so many and such close approximations to mammalian structure, as strongly to suggest a genetic relationship. It is difficult to believe that all those likenesses should have been independently acquired and are without phylogenetic significance.

Birds are comparatively rare as fossils and we should therefore look in vain among them for any such long and closely knit series as the mammals display in abundance. Nevertheless, a few extremely fortunate discoveries have made it practically certain that birds are descended from reptiles, of which they represent a highly specialised branch. The most ancient representative of this class is the extraordinary genus Archaeopteryx from the upper Jurassic of Bavaria, which, though an unmistakable bird, retains so many reptilian structures and characteristics as to make its derivation plain. Not to linger over anatomical minutiae, it may suffice to mention the absence of a horny beak, which is replaced by numerous true teeth, and the long lizard-like tail, which is made up of numerous distinct vertebrae, each with a pair of quill-like feathers attached to it. Birds with teeth are also found in the Cretaceous, though in most other respects the birds of that period had attained a substantially modern structure. Concerning the interrelations of the various orders and families of birds, palaeontology has as yet little to tell us.

The life of the Mesozoic era was characterised by an astonishing number and variety of reptiles, which were adapted to every mode of life, and dominated the air, the sea and the land, and many of which were of colossal proportions. Owing to the conditions of preservation which obtained during the Mesozoic period, the history of the reptiles is a broken and interrupted one, so that we can make out many short series, rather than any one of considerable length. While the relations of several reptilian orders can be satisfactorily determined, others still baffle us entirely, making their first known appearance in a fully differentiated state. We can trace the descent of the sea-dragons, the Ichthyosaurs and Plesiosaurs, from terrestrial ancestors, but the most ancient turtles yet discovered show us no closer approximation to any other order than do the recent turtles; and the oldest known Pterosaurs, the flying dragons of the Jurassic, are already fully differentiated. There is, however, no ground for discouragement in this, for the progress of discovery has been so rapid of late years, and our knowledge of Mesozoic life has increased with such leaps and bounds, that there is every reason to expect a solution of many of the outstanding problems in the near future.

Passing over the lower vertebrates, for lack of space to give them any adequate consideration, we may briefly take up the record of invertebrate life. From the overwhelming mass of material it is difficult to make a representative selection and even more difficult to state the facts intelligibly without the use of unduly technical language and without the aid of illustrations.

Several groups of the Mollusca, or shell-fish, yield very full and convincing evidence of their descent from earlier and simpler forms, and of these none is of greater interest than the Ammonites, an extinct order of the cephalopoda. The nearest living ally of the ammonites is the pearly nautilus, the other existing cephalopods, such as the squids, cuttle-fish, octopus, etc., are much more distantly related. Like the nautilus, the ammonites all possess a coiled and chambered shell, but their especial characteristic is the complexity of the "sutures." By sutures is meant the edges of the transverse partitions, or septa, where these join the shell-wall, and their complexity in the fully developed genera is extraordinary, forming patterns like the most elaborate oak-leaf embroidery, while in the nautiloids the sutures form simple curves. In the rocks of the Mesozoic era, wherever conditions of preservation are favourable, these beautiful shells are stored in countless multitudes, of an incredible variety of form, size and ornamentation, as is shown by the fact that nearly 5000 species have already been described. The ammonites are particularly well adapted for phylogenetic studies, because, by removing the successive whorls of the coiled shell, the individual development may be followed back in inverse order, to the microscopic "protoconch," or embryonic shell, which lies concealed in the middle of the coil. Thus the valuable aid of embryology is obtained in determining relationships.

The descent of the ammonites, taken as a group, is simple and clear; they arose as a branch of the nautiloids in the lower Devonian, the shells known as goniatites having zigzag, angulated sutures. Late in the succeeding Carboniferous period appear shells with a truly ammonoid complexity of sutures, and in the Permian their number and variety cause them to form a striking element of the marine faunas. It is in the Mesozoic era, however, that these shells attain their full development; increasing enormously in the Triassic, they culminate in the Jurassic in the number of families, genera and species, in the complexity of the sutures, and in the variety of shell-ornamentation. A slow decline begins in the Cretaceous, ending in the complete extinction of the whole group at the end of that period. As a final phase in the history of the ammonites, there appear many so-called "abnormal" genera, in which the shell is irregularly coiled, or more or less uncoiled, in some forms becoming actually straight. It is interesting to observe that some of these genera are not natural groups, but are "polyphyletic," i.e. are each derived from several distinct ancestral genera, which have undergone a similar kind of degeneration.

In the huge assembly of ammonites it is not yet possible to arrange all the forms in a truly natural classification, which shall express the various interrelations of the genera, yet several beautiful series have already been determined. In these series the individual development of the later general shows transitory stages which are permanent in antecedent genera. To give a mere catalogue of names without figures would not make these series more intelligible.

The Brachiopoda, or "lamp-shells," are a phylum of which comparatively few survive to the present day; their shells have a superficial likeness to those of the bivalved Mollusca, but are not homologous with the latter, and the phylum is really very distinct from the molluscs. While greatly reduced now, these animals were incredibly abundant throughout the Palaeozoic era, great masses of limestone being often composed almost exclusively of their shells, and their variety is in keeping with their individual abundance. As in the case of the ammonites, the problem is to arrange this great multitude of forms in an orderly array that shall express the ramifications of the group according to a genetic system. For many brachiopods, both recent and fossil, the individual development, or ontogeny, has been worked out and has proved to be of great assistance in the problems of classification and phylogeny. Already very encouraging progress has been made in the solution of these problems. All brachiopods form first a tiny, embryonic shell, called the protegulum, which is believed to represent the ancestral form of the whole group, and in the more advanced genera the developmental stages clearly indicate the ancestral genera of the series, the succession of adult forms in time corresponding to the order of the ontogenetic stages. The transformation of the delicate calcareous supports of the arms, often exquisitely preserved, are extremely interesting. Many of the Palaeozoic genera had these supports coiled like a pair of spiral springs, and it has been shown that these genera were derived from types in which the supports were simply shelly loops.

The long extinct class of crustacea known as the Trilobites are likewise very favourable subjects for phylogenetic studies. So far as the known record can inform us, the trilobites are exclusively Palaeozoic in distribution, but their course must have begun long before that era, as is shown by the number of distinct types among the genera of the lower Cambrian. The group reached the acme of abundance and relative importance in the Cambrian and Ordovician; then followed a long, slow decline, ending in complete and final disappearance before the end of the Permian. The newly-hatched and tiny trilobite larva, known as the protaspis, is very near to the primitive larval form of all the crustacea. By the aid of the correlated ontogenetic stages and the succession of the adult forms in the rocks, many phylogenetic series have been established and a basis for the natural arrangement of the whole class has been laid.

Very instructive series may also be observed among the Echinoderms and, what is very rare, we are able in this sub-kingdom to demonstrate the derivation of one class from another. Indeed, there is much reason to believe that the extinct class Cystidea of the Cambrian is the ancestral group, from which all the other Echinoderms, star-fishes, brittle-stars, sea-urchins, feather-stars, etc., are descended.

The foregoing sketch of the palaeontological record is, of necessity, extremely meagre, and does not represent even an outline of the evidence, but merely a few illustrative examples, selected almost at random from an immense body of material. However, it will perhaps suffice to show that the geological record is not so hopelessly incomplete as Darwin believed it to be. Since "The Origin of Species" was written, our knowledge of that record has been enormously extended and we now possess, no complete volumes, it is true, but some remarkably full and illuminating chapters. The main significance of the whole lies in the fact, that JUST IN PROPORTION TO THE COMPLETENESS OF THE RECORD IS THE UNEQUIVOCAL CHARACTER OF ITS TESTIMONY TO THE TRUTH OF THE EVOLUTIONARY THEORY.

The test of a true, as distinguished from a false, theory is the manner in which newly discovered and unanticipated facts arrange themselves under it. No more striking illustration of this can be found than in the contrasted fates of Cuvier's theory and of that of Darwin. Even before Cuvier's death his views had been undermined and the progress of discovery soon laid them in irreparable ruin, while the activity of half-a-century in many different lines of inquiry has established the theory of evolution upon a foundation of ever growing solidity. It is Darwin's imperishable glory that he prescribed the lines along which all the biological sciences were to advance to conquests not dreamed of when he wrote.



XII. THE PALAEONTOLOGICAL RECORD. By D.H. Scott, F.R.S.

President of the Linnean Society.

II. PLANTS.

There are several points of view from which the subject of the present essay may be regarded. We may consider the fossil record of plants in its bearing: I. on the truth of the doctrine of Evolution; II. on Phylogeny, or the course of Evolution; III. on the theory of Natural Selection. The remarks which follow, illustrating certain aspects only of an extensive subject, may conveniently be grouped under these three headings.

I. THE TRUTH OF EVOLUTION.

When "The Origin of Species" was written, it was necessary to show that the Geological Record was favourable to, or at least consistent with, the Theory of Descent. The point is argued, closely and fully, in Chapter X. "On the Imperfection of the Geological Record," and Chapter XI. "On the Geological Succession of Organic Beings"; there is, however, little about plants in these chapters. At the present time the truth of Evolution is no longer seriously disputed, though there are writers, like Reinke, who insist, and rightly so, that the doctrine is still only a belief, rather than an established fact of science. (J. Reinke, "Kritische Abstammungslehre", "Wiesner-Festschrift", page 11, Vienna, 1908.) Evidently, then, however little the Theory of Descent may be questioned in our own day, it is desirable to assure ourselves how the case stands, and in particular how far the evidence from fossil plants has grown stronger with time.

As regards direct evidence for the derivation of one species from another, there has probably been little advance since Darwin wrote, at least so we must infer from the emphasis laid on the discontinuity of successive fossil species by great systematic authorities like Grand'Eury and Zeiller in their most recent writings. We must either adopt the mutationist views of those authors (referred to in the last section of this essay) or must still rely on Darwin's explanation of the absence of numerous intermediate varieties. The attempts which have been made to trace, in the Tertiary rocks, the evolution of recent species, cannot, owing to the imperfect character of the evidence, be regarded as wholly satisfactory.

When we come to groups of a somewhat higher order we have an interesting history of the evolution of a recent family in the work, not yet completed, of Kidston and Gwynne-Vaughan on the fossil Osmundaceae. ("Trans. Royal Soc. Edinburgh", Vol. 45, Part III. 1907, Vol. 46, Part II. 1908, Vol. 46, Part III. 1909.) The authors are able, mainly on anatomical evidence, to trace back this now limited group of Ferns, through the Tertiary and Mesozoic to the Permian, and to show, with great probability, how their structure has been derived from that of early Palaeozoic types.

The history of the Ginkgoaceae, now represented only by the isolated maidenhair tree, scarcely known in a wild state, offers another striking example of a family which can be traced with certainty to the older Mesozoic and perhaps further back still. (See Seward and Gowan, "The Maidenhair Tree (Gingko biloba)", "Annals of Botany", Vol. XIV. 1900, page 109; also A. Sprecher "Le Ginkgo biloba", L., Geneva, 1907.)

On the wider question of the derivation of the great groups of plants, a very considerable advance has been made, and, so far as the higher plants are concerned, we are now able to form a far better conception than before of the probable course of evolution. This is a matter of phylogeny, and the facts will be considered under that head; our immediate point is that the new knowledge of the relations between the classes of plants in question materially strengthens the case for the theory of descent. The discoveries of the last few years throw light especially on the relation of the Angiosperms to the Gymnosperms, on that of the Seed-plants generally to the Ferns, and on the interrelations between the various classes of the higher Cryptogams.

That the fossil record has not done still more for Evolution is due to the fact that it begins too late—a point on which Darwin laid stress ("Origin of Species" (6th edition), page 286.) and which has more recently been elaborated by Poulton. ("Essays on Evolution", pages 46 et seq., Oxford, 1908.) An immense proportion of the whole evolutionary history lies behind the lowest fossiliferous rocks, and the case is worse for plants than for animals, as the record for the former begins, for all practical purposes, much higher up in the rocks.

It may be well here to call attention to a question, often overlooked, which has lately been revived by Reinke. (Reinke, loc. cit. page 13.) As all admit, we know nothing of the origin of life; consequently, for all we can tell, it is as probable that life began, on this planet, with many living things, as with one. If the first organic beings were many, they may have been heterogeneous, or at least exposed to different conditions, from their origin; in either case there would have been a number of distinct series from the beginning, and if so we should not be justified in assuming that all organisms are related to one another. There may conceivably be several of the original lines of descent still surviving, or represented among extinct forms—to reverse the remark of a distinguished botanist, there may be several Vegetable Kingdoms! However improbable this may sound, the possibility is one to be borne in mind.

That all VASCULAR plants really belong to one stock seems certain, and here the palaeontological record has materially strengthened the case for a monophyletic history. The Bryophyta are not likely to be absolutely distinct, for their sexual organs, and the stomata of the Mosses strongly suggest community of descent with the higher plants; if this be so it no doubt establishes a certain presumption in favour of a common origin for plants generally, for the gap between "Mosses and Ferns" has been regarded as the widest in the Vegetable Kingdom. The direct evidence of consanguinity is however much weaker when we come to the Algae, and it is conceivable (even if improbable) that the higher plants may have had a distinct ancestry (now wholly lost) from the beginning. The question had been raised in Darwin's time, and he referred to it in these words: "No doubt it is possible, as Mr G.H. Lewes has urged, that at the first commencement of life many different forms were evolved; but if so, we may conclude that only a very few have left modified descendants." ("Origin of Species", page 425.) This question, though it deserves attention, does not immediately affect the subject of the palaeontological record of plants, for there can be no reasonable doubt as to the interrelationship of those groups on which the record at present throws light.

The past history of plants by no means shows a regular progression from the simple to the complex, but often the contrary. This apparent anomaly is due to two causes.

1. The palaeobotanical record is essentially the story of the successive ascendancy of a series of dominant families, each of which attained its maximum, in organisation as well as in extent, and then sank into comparative obscurity, giving place to other families, which under new conditions were better able to take a leading place. As each family ran its downward course, either its members underwent an actual reduction in structure as they became relegated to herbaceous or perhaps aquatic life (this may have happened with the Horsetails and with Isoetes if derived from Lepidodendreae), or the higher branches of the family were crowded out altogether and only the "poor relations" were able to maintain their position by evading the competition of the ascendant races; this is also illustrated by the history of the Lycopod phylum. In either case there would result a lowering of the type of organisation within the group.

2. The course of real progress is often from the complex to the simple. If, as we shall find some grounds for believing, the Angiosperms came from a type with a flower resembling in its complexity that of Mesozoic "Cycads," almost the whole evolution of the flower in the highest plants has been a process of reduction. The stamen, in particular, has undoubtedly become extremely simplified during evolution; in the most primitive known seed-plants it was a highly compound leaf or pinna; its reduction has gone on in the Conifers and modern Cycads, as well as in the Angiosperms, though in different ways and to a varying extent.

The seed offers another striking example; the Palaeozoic seeds (if we leave the seed-like organs of certain Lycopods out of consideration) were always, so far as we know, highly complex structures, with an elaborate vascular system, a pollen-chamber, and often a much-differentiated testa. In the present day such seeds exist only in a few Gymnosperms which retain their ancient characters—in all the higher Spermophytes the structure is very much simplified, and this holds good even in the Coniferae, where there is no countervailing complication of ovary and stigma.

Reduction, in fact, is not always, or even generally, the same thing as degeneration. Simplification of parts is one of the most usual means of advance for the organism as a whole. A large proportion of the higher plants are microphyllous in comparison with the highly megaphyllous fern-like forms from which they appear to have been derived.

Darwin treated the general question of advance in organisation with much caution, saying: "The geological record... does not extend far enough back, to show with unmistakeable clearness that within the known history of the world organisation has largely advanced." ("Origin of Species", page 308.) Further on (Ibid. page 309.) he gives two standards by which advance may be measured: "We ought not solely to compare the highest members of a class at any two periods... but we ought to compare all the members, high and low, at the two periods." Judged by either standard the Horsetails and Club Mosses of the Carboniferous were higher than those of our own day, and the same is true of the Mesozoic Cycads. There is a general advance in the succession of classes, but not within each class.

Darwin's argument that "the inhabitants of the world at each successive period in its history have beaten their predecessors in the race for life, and are, in so far, higher in the scale" ("Origin of Species", page 315.) is unanswerable, but we must remember that "higher in the scale" only means "better adapted to the existing conditions." Darwin points out (Ibid. page 279.) that species have remained unchanged for long periods, probably longer than the periods of modification, and only underwent change when the conditions of their life were altered. Higher organisation, judged by the test of success, is thus purely relative to the changing conditions, a fact of which we have a striking illustration in the sudden incoming of the Angiosperms with all their wonderful floral adaptations to fertilisation by the higher families of Insects.

II. PHYLOGENY.

The question of phylogeny is really inseparable from that of the truth of the doctrine of evolution, for we cannot have historical evidence that evolution has actually taken place without at the same time having evidence of the course it has followed.

As already pointed out, the progress hitherto made has been rather in the way of joining up the great classes of plants than in tracing the descent of particular species or genera of the recent flora. There appears to be a difference in this respect from the Animal record, which tells us so much about the descent of living species, such as the elephant or the horse. The reason for this difference is no doubt to be found in the fact that the later part of the palaeontological record is the most satisfactory in the case of animals and the least so in the case of plants. The Tertiary plant-remains, in the great majority of instances, are impressions of leaves, the conclusions to be drawn from which are highly precarious; until the whole subject of Angiospermous palaeobotany has been reinvestigated, it would be rash to venture on any statements as to the descent of the families of Dicotyledons or Monocotyledons.

Our attention will be concentrated on the following questions, all relating to the phylogeny of main groups of plants: i. The Origin of the Angiosperms. ii. The Origin of the Seed-plants. iii. The Origin of the different classes of the Higher Cryptogamia.

i. THE ORIGIN OF THE ANGIOSPERMS.

The first of these questions has long been the great crux of botanical phylogeny, and until quite recently no light had been thrown upon the difficulty. The Angiosperms are the Flowering Plants, par excellence, and form, beyond comparison, the dominant sub-kingdom in the flora of our own age, including, apart from a few Conifers and Ferns, all the most familiar plants of our fields and gardens, and practically all plants of service to man. All recent work has tended to separate the Angiosperms more widely from the other seed-plants now living, the Gymnosperms. Vast as is the range of organisation presented by the great modern sub-kingdom, embracing forms adapted to every environment, there is yet a marked uniformity in certain points of structure, as in the development of the embryo-sac and its contents, the pollination through the intervention of a stigma, the strange phenomenon of double fertilisation (One sperm fertilising the egg, while the other unites with the embryo-sac nucleus, itself the product of a nuclear fusion, to give rise to a nutritive tissue, the endosperm.), the structure of the stamens, and the arrangement of the parts of the flower. All these points are common to Monocotyledons and Dicotyledons, and separate the Angiosperms collectively from all other plants.

In geological history the Angiosperms first appear in the Lower Cretaceous, and by Upper Cretaceous times had already swamped all other vegetation and seized the dominant position which they still hold. Thus they are isolated structurally from the rest of the Vegetable Kingdom, while historically they suddenly appear, almost in full force, and apparently without intermediaries with other groups. To quote Darwin's vigorous words: "The rapid development, as far as we can judge, of all the higher plants within recent geological times is an abominable mystery." ("More Letters of Charles Darwin", Vol. II. page 20, letter to J.D. Hooker, 1879.) A couple of years later he made a bold suggestion (which he only called an "idle thought") to meet this difficulty. He says: "I have been so astonished at the apparently sudden coming in of the higher phanerogams, that I have sometimes fancied that development might have slowly gone on for an immense period in some isolated continent or large island, perhaps near the South Pole." (Ibid, page 26, letter to Hooker, 1881.) This idea of an Angiospermous invasion from some lost southern land has sometimes been revived since, but has not, so far as the writer is aware, been supported by evidence. Light on the problem has come from a different direction.

The immense development of plants with the habit of Cycads, during the Mesozoic Period up to the Lower Cretaceous, has long been known. The existing Order Cycadaceae is a small family, with 9 genera and perhaps 100 species, occurring in the tropical and sub-tropical zones of both the Old and New World, but nowhere forming a dominant feature in the vegetation. Some few attain the stature of small trees, while in the majority the stem is short, though often living to a great age. The large pinnate or rarely bipinnate leaves give the Cycads a superficial resemblance in habit to Palms. Recent Cycads are dioecious; throughout the family the male fructification is in the form of a cone, each scale of the cone representing a stamen, and bearing on its lower surface numerous pollen-sacs, grouped in sori like the sporangia of Ferns. In all the genera, except Cycas itself, the female fructifications are likewise cones, each carpel bearing two ovules on its margin. In Cycas, however, no female cone is produced, but the leaf-like carpels, bearing from two to six ovules each, are borne directly on the main stem of the plant in rosettes alternating with those of the ordinary leaves—the most primitive arrangement known in any living seed-plant. The whole Order is relatively primitive, as shown most strikingly in its cryptogamic mode of fertilisation, by means of spermatozoids, which it shares with the maidenhair tree alone, among recent seed-plants.

In all the older Mesozoic rocks, from the Trias to the Lower Cretaceous, plants of the Cycad class (Cycadophyta, to use Nathorst's comprehensive name) are extraordinarily abundant in all parts of the world; in fact they were almost as prominent in the flora of those ages as the Dicotyledons are in that of our own day. In habit and to a great extent in anatomy, the Mesozoic Cycadophyta for the most part much resemble the recent Cycadaceae. But, strange to say, it is only in the rarest cases that the fructification has proved to be of the simple type characteristic of the recent family; the vast majority of the abundant fertile specimens yielded by the Mesozoic rocks possess a type of reproductive apparatus far more elaborate than anything known in Cycadaceae or other Gymnosperms. The predominant Mesozoic family, characterised by this advanced reproductive organisation, is known as the Bennettiteae; in habit these plants resembled the more stunted Cycads of the recent flora, but differed from them in the presence of numerous lateral fructifications, like large buds, borne on the stem among the crowded bases of the leaves. The organisation of these fructifications was first worked out on European specimens by Carruthers, Solms-Laubach, Lignier and others, but these observers had only more or less ripe fruits to deal with; the complete structure of the flower has only been elucidated within the last few years by the researches of Wieland on the magnificent American material, derived from the Upper Jurassic and Lower Cretaceous beds of Maryland, Dakota and Wyoming. (G.R. Wieland, "American Fossil Cycads", Carnegie Institution, Washington, 1906.) The word "flower" is used deliberately, for reasons which will be apparent from the following brief description, based on Wieland's observations.

The fructification is attached to the stem by a thick stalk, which, in its upper part, bears a large number of spirally arranged bracts, forming collectively a kind of perianth and completely enclosing the essential organs of reproduction. The latter consist of a whorl of stamens, of extremely elaborate structure, surrounding a central cone or receptacle bearing numerous ovules. The stamens resemble the fertile fronds of a fern; they are of a compound, pinnate form, and bear very large numbers of pollen-sacs, each of which is itself a compound structure consisting of a number of compartments in which the pollen was formed. In their lower part the stamens are fused together by their stalks, like the "monadelphous" stamens of a mallow. The numerous ovules borne on the central receptacle are stalked, and are intermixed with sterile scales; the latter are expanded at their outer ends, which are united to form a kind of pericarp or ovary-wall, only interrupted by the protruding micropyles of the ovules. There is thus an approach to the closed pistil of an Angiosperm, but it is evident that the ovules received the pollen directly. The whole fructification is of large size; in the case of Cycadeoidea dacotensis, one of the species investigated by Wieland, the total length, in the bud condition, is about 12 cm., half of which belongs to the peduncle.

The general arrangement of the organs is manifestly the same as in a typical Angiospermous flower, with a central pistil, a surrounding whorl of stamens and an enveloping perianth; there is, as we have seen, some approach to the closed ovary of an Angiosperm; another point, first discovered nearly 20 years ago by Solms-Laubach in his investigation of a British species, is that the seed was practically "exalbuminous," its cavity being filled by the large, dicotyledonous embryo, whereas in all known Gymnosperms a large part of the sac is occupied by a nutritive tissue, the prothallus or endosperm; here also we have a condition only met with elsewhere among the higher Flowering Plants.

Taking all the characters into account, the indications of affinity between the Mesozoic Cycadophyta and the Angiosperms appear extremely significant, as was recognised by Wieland when he first discovered the hermaphrodite nature of the Bennettitean flower. The Angiosperm with which he specially compared the fossil type was the Tulip tree (Liriodendron) and certainly there is a remarkable analogy with the Magnoliaceous flowers, and with those of related orders such as Ranunculaceae and the Water-lilies. It cannot, of course, be maintained that the Bennettiteae, or any other Mesozoic Cycadophyta at present known, were on the direct line of descent of the Angiosperms, for there are some important points of difference, as, for example, in the great complexity of the stamens, and in the fact that the ovary-wall or pericarp was not formed by the carpels themselves, but by the accompanying sterile scale-leaves. Botanists, since the discovery of the bisexual flowers of the Bennettiteae, have expressed different views as to the nearness of their relation to the higher Flowering Plants, but the points of agreement are so many that it is difficult to resist the conviction that a real relation exists, and that the ancestry of the Angiosperms, so long shrouded in complete obscurity, is to be sought among the great plexus of Cycad-like plants which dominated the flora of the world in Mesozoic times. (On this subject see, in addition to Wieland's great work above cited, F.W. Oliver, "Pteridosperms and Angiosperms", "New Phytologist", Vol. V. 1906; D.H. Scott, "The Flowering Plants of the Mesozoic Age in the Light of Recent Discoveries", "Journal R. Microscop. Soc." 1907, and especially E.A.N. Arber and J. Parkin, "On the Origin of Angiosperms", "Journal Linn. Soc." (Bot.) Vol. XXXVIII. page 29, 1907.)

The great complexity of the Bennettitean flower, the earliest known fructification to which the word "flower" can be applied without forcing the sense, renders it probable, as Wieland and others have pointed out, that the evolution of the flower in Angiosperms has consisted essentially in a process of reduction, and that the simplest forms of flower are not to be regarded as the most primitive. The older morphologists generally took the view that such simple flowers were to be explained as reductions from a more perfect type, and this opinion, though abandoned by many later writers, appears likely to be true when we consider the elaboration of floral structure attained among the Mesozoic Cycadophyta, which preceded the Angiosperms in evolution.

If, as now seems probable, the Angiosperms were derived from ancestors allied to the Cycads, it would naturally follow that the Dicotyledons were first evolved, for their structure has most in common with that of the Cycadophyta. We should then have to regard the Monocotyledons as a side-line, diverging probably at a very early stage from the main dicotyledonous stock, a view which many botanists have maintained, of late, on other grounds. (See especially Ethel Sargant, "The Reconstruction of a Race of Primitive Angiosperms", "Annals of Botany", Vol. XXII. page 121, 1908.) So far, however, as the palaeontological record shows, the Monocotyledons were little if at all later in their appearance than the Dicotyledons, though always subordinate in numbers. The typical and beautifully preserved Palm-wood from Cretaceous rocks is striking evidence of the early evolution of a characteristic monocotyledonous family. It must be admitted that the whole question of the evolution of Monocotyledons remains to be solved.

Accepting, provisionally, the theory of the cycadophytic origin of Angiosperms, it is interesting to see to what further conclusions we are led. The Bennettiteae, at any rate, were still at the gymnospermous level as regards their pollination, for the exposed micropyles of the ovules were in a position to receive the pollen directly, without the intervention of a stigma. It is thus indicated that the Angiosperms sprang from a gymnospermous source, and that the two great phyla of Seed-plants have not been distinct from the first, though no doubt the great majority of known Gymnosperms, especially the Coniferae, represent branch-lines of their own.

The stamens of the Bennettiteae are arranged precisely as in an angiospermous flower, but in form and structure they are like the fertile fronds of a Fern, in fact the compound pollen-sacs, or synangia as they are technically called, almost exactly agree with the spore-sacs of a particular family of Ferns—the Marattiaceae, a limited group, now mainly tropical, which was probably more prominent in the later Palaeozoic times than at present. The scaly hairs, or ramenta, which clothe every part of the plant, are also like those of Ferns.

It is not likely that the characters in which the Bennettiteae resemble the Ferns came to them directly from ancestors belonging to that class; an extensive group of Seed-plants, the Pteridospermeae, existed in Palaeozoic times and bear evident marks of affinity with the Fern phylum. The fern-like characters so remarkably persistent in the highly organised Cycadophyta of the Mesozoic were in all likelihood derived through the Pteridosperms, plants which show an unmistakable approach to the cycadophytic type.

The family Bennettiteae thus presents an extraordinary association of characters, exhibiting, side by side, features which belong to the Angiosperms, the Gymnosperms and the Ferns.

ii. ORIGIN OF SEED-PLANTS.

The general relation of the gymnospermous Seed-plants to the Higher Cryptogamia was cleared up, independently of fossil evidence, by the brilliant researches of Hofmeister, dating from the middle of the past century. (W. Hofmeister, "On the Germination, Development and Fructification of the Higher Cryptogamia", Ray Society, London, 1862. The original German treatise appeared in 1851.) He showed that "the embryo-sac of the Coniferae may be looked upon as a spore remaining enclosed in its sporangium; the prothallium which it forms does not come to the light." (Ibid. page 438.) He thus determined the homologies on the female side. Recognising, as some previous observers had already done, that the microspores of those Cryptogams in which two kinds of spore are developed, are equivalent to the pollen-grains of the higher plants, he further pointed out that fertilisation "in the Rhizocarpeae and Selaginellae takes place by free spermatozoa, and in the Coniferae by a pollen-tube, in the interior of which spermatozoa are probably formed"—a remarkable instance of prescience, for though spermatozoids have not been found in the Conifers proper, they were demonstrated in the allied groups Cycadaceae and Ginkgo, in 1896, by the Japanese botanists Ikeno and Hirase. A new link was thus established between the Gymnosperms and the Cryptogams.

It remained uncertain, however, from which line of Cryptogams the gymnospermous Seed-plants had sprung. The great point of morphological comparison was the presence of two kinds of spore, and this was known to occur in the recent Lycopods and Water-ferns (Rhizocarpeae) and was also found in fossil representatives of the third phylum, that of the Horsetails. As a matter of fact all the three great Cryptogamic classes have found champions to maintain their claim to the ancestry of the Seed-plants, and in every case fossil evidence was called in. For a long time the Lycopods were the favourites, while the Ferns found the least support. The writer remembers, however, in the year 1881, hearing the late Prof. Sachs maintain, in a lecture to his class, that the descent of the Cycads could be traced, not merely from Ferns, but from a definite family of Ferns, the Marattiaceae, a view which, though in a somewhat crude form, anticipated more modern ideas.

Williamson appears to have been the first to recognise the presence, in the Carboniferous flora, of plants combining the characters of Ferns and Cycads. (See especially his "Organisation of the Fossil Plants of the Coal-Measures", Part XIII. "Phil. Trans. Royal Soc." 1887 B. page 299.) This conclusion was first reached in the case of the genera Heterangium and Lyginodendron, plants, which with a wholly fern-like habit, were found to unite an anatomical structure holding the balance between that of Ferns and Cycads, Heterangium inclining more to the former and Lyginodendron to the latter. Later researches placed Williamson's original suggestion on a firmer basis, and clearly proved the intermediate nature of these genera, and of a number of others, so far as their vegetative organs were concerned. This stage in our knowledge was marked by the institution of the class Cycadofilices by Potonie in 1897.

Nothing, however, was known of the organs of reproduction of the Cycadofilices, until F.W. Oliver, in 1903, identified a fossil seed, Lagenostoma Lomaxi, as belonging to Lyginodendron, the identification depending, in the first instance, on the recognition of an identical form of gland, of very characteristic structure, on the vegetative organs of Lyginodendron and on the cupule enveloping the seed. This evidence was supported by the discovery of a close anatomical agreement in other respects, as well as by constant association between the seed and the plant. (F.W. Oliver and D.H. Scott, "On the Structure of the Palaeozoic Seed, Lagenostoma Lomaxi, etc." "Phil. Trans. Royal Soc." Vol. 197 B. 1904.) The structure of the seed of Lyginodendron, proved to be of the same general type as that of the Cycads, as shown especially by the presence of a pollen-chamber or special cavity for the reception of the pollen-grains, an organ only known in the Cycads and Ginkgo among recent plants.

Within a few months after the discovery of the seed of Lyginodendron, Kidston found the large, nut-like seed of a Neuropteris, another fern-like Carboniferous plant, in actual connection with the pinnules of the frond, and since then seeds have been observed on the frond in species of Aneimites and Pecopteris, and a vast body of evidence, direct or indirect, has accumulated, showing that a large proportion of the Palaeozoic plants formerly classed as Ferns were in reality reproduced by seeds of the same type as those of recent Cycadaceae. (A summary of the evidence will be found in the writer's article "On the present position of Palaeozoic Botany", "Progressus Rei Botanicae", 1907, page 139, and "Studies in Fossil Botany", Vol. II. (2nd edition) London, 1909.) At the same time, the anatomical structure, where it is open to investigation, confirms the suggestion given by the habit, and shows that these early seed-bearing plants had a real affinity with Ferns. This conclusion received strong corroboration when Kidston, in 1905, discovered the male organs of Lyginodendron, and showed that they were identical with a fructification of the genus Crossotheca, hitherto regarded as belonging to Marattiaceous Ferns. (Kidston, "On the Microsporangia of the Pteridospermeae, etc." "Phil. Trans. Royal Soc." Vol. 198, B. 1906.)

The general conclusion which follows from the various observations alluded to, is that in Palaeozoic times there was a great body of plants (including, as it appears, a large majority of the fossils previously regarded as Ferns) which had attained the rank of Spermophyta, bearing seeds of a Cycadean type on fronds scarcely differing from the vegetative foliage, and in other respects, namely anatomy, habit and the structure of the pollen-bearing organs, retaining many of the characters of Ferns. From this extensive class of plants, to which the name Pteridospermeae has been given, it can scarcely be doubted that the abundant Cycadophyta, of the succeeding Mesozoic period, were derived. This conclusion is of far-reaching significance, for we have already found reason to think that the Angiosperms themselves sprang, in later times, from the Cycadophytic stock; it thus appears that the Fern-phylum, taken in a broad sense, ultimately represents the source from which the main line of descent of the Phanerogams took its rise.

It must further be borne in mind that in the Palaeozoic period there existed another group of seed-bearing plants, the Cordaiteae, far more advanced than the Pteridospermeae, and in many respects approaching the Coniferae, which themselves begin to appear in the latest Palaeozoic rocks. The Cordaiteae, while wholly different in habit from the contemporary fern-like Seed-plants, show unmistakable signs of a common origin with them. Not only is there a whole series of forms connecting the anatomical structure of the Cordaiteae with that of the Lyginodendreae among Pteridosperms, but a still more important point is that the seeds of the Cordaiteae, which have long been known, are of the same Cycadean type as those of the Pteridosperms, so that it is not always possible, as yet, to discriminate between the seeds of the two groups. These facts indicate that the same fern-like stock which gave rise to the Cycadophyta and through them, as appears probable, to the Angiosperms, was also the source of the Cordaiteae, which in their turn show manifest affinity with some at least of the Coniferae. Unless the latter are an artificial group, a view which does not commend itself to the writer, it would appear probable that the Gymnosperms generally, as well as the Angiosperms, were derived from an ancient race of Cryptogams, most nearly related to the Ferns. (Some botanists, however, believe that the Coniferae, or some of them, are probably more nearly related to the Lycopods. See Seward and Ford, "The Araucarieae, Recent and Extinct", "Phil. Trans. Royal Soc." Vol. 198 B. 1906.)

It may be mentioned here that the small gymnospermous group Gnetales (including the extraordinary West African plant Welwitschia) which were formerly regarded by some authorities as akin to the Equisetales, have recently been referred, on better grounds, to a common origin with the Angiosperms, from the Mesozoic Cycadophyta.

The tendency, therefore, of modern work on the palaeontological record of the Seed-plants has been to exalt the importance of the Fern-phylum, which, on present evidence, appears to be that from which the great majority, possibly the whole, of the Spermophyta have been derived.

One word of caution, however, is necessary. The Seed-plants are of enormous antiquity; both the Pteridosperms and the more highly organised family Cordaiteae, go back as far in geological history (namely to the Devonian) as the Ferns themselves or any other Vascular Cryptogams. It must therefore be understood that in speaking of the derivation of the Spermophyta from the Fern-phylum, we refer to that phylum at a very early stage, probably earlier than the most ancient period to which our record of land-plants extends. The affinity between the oldest Seed-plants and the Ferns, in the widest sense, seems established, but the common stock from which they actually arose is still unknown; though no doubt nearer to the Ferns than to any other group, it must have differed widely from the Ferns as we now know them, or perhaps even from any which the fossil record has yet revealed to us.

iii. THE ORIGIN OF THE HIGHER CRYPTOGAMIA.

The Sub-kingdom of the higher Spore-plants, the Cryptogamia possessing a vascular system, was more prominent in early geological periods than at present. It is true that the dominance of the Pteridophyta in Palaeozoic times has been much exaggerated owing to the assumption that everything which looked like a Fern really was a Fern. But, allowing for the fact, now established, that most of the Palaeozoic fern-like plants were already Spermophyta, there remains a vast mass of Cryptogamic forms of that period, and the familiar statement that they formed the main constituent of the Coal-forests still holds good. The three classes, Ferns (Filicales), Horsetails (Equisetales) and Club-mosses (Lycopodiales), under which we now group the Vascular Cryptogams, all extend back in geological history as far as we have any record of the flora of the land; in the Palaeozoic, however, a fourth class, the Sphenophyllales, was present.

As regards the early history of the Ferns, which are of special interest from their relation to the Seed-plants, it is impossible to speak quite positively, owing to the difficulty of discriminating between true fossil Ferns and the Pteridosperms which so closely simulated them. The difficulty especially affects the question of the position of Marattiaceous Ferns in the Palaeozoic Floras. This family, now so restricted, was until recently believed to have been one of the most important groups of Palaeozoic plants, especially during later Carboniferous and Permian times. Evidence both from anatomy and from sporangial characters appeared to establish this conclusion. Of late, however, doubts have arisen, owing to the discovery that some supposed members of the Marattiaceae bore seeds, and that a form of fructification previously referred to that family (Crossotheca) was really the pollen-bearing apparatus of a Pteridosperm (Lyginodendron). The question presents much difficulty; though it seems certain that our ideas of the extent of the family in Palaeozoic times will have to be restricted, there is still a decided balance of evidence in favour of the view that a considerable body of Marattiaceous Ferns actually existed. The plants in question were of large size (often arborescent) and highly organised—they represent, in fact, one of the highest developments of the Fern-stock, rather than a primitive type of the class.

There was, however, in the Palaeozoic period, a considerable group of comparatively simple Ferns (for which Arber has proposed the collective name Primofilices); the best known of these are referred to the family Botryopterideae, consisting of plants of small or moderate dimensions, with, on the whole, a simple anatomical structure, in certain cases actually simpler than that of any recent Ferns. On the other hand the sporangia of these plants were usually borne on special fertile fronds, a mark of rather high differentiation. This group goes back to the Devonian and includes some of the earliest types of Fern with which we are acquainted. It is probable that the Primofilices (though not the particular family Botryopterideae) represent the stock from which the various families of modern Ferns, already developed in the Mesozoic period, may have sprung.

None of the early Ferns show any clear approach to other classes of Vascular Cryptogams; so far as the fossil record affords any evidence, Ferns have always been plants with relatively large and usually compound leaves. There is no indication of their derivation from a microphyllous ancestry, though, as we shall see, there is some slight evidence for the converse hypothesis. Whatever the origin of the Ferns may have been it is hidden in the older rocks.

It has, however, been held that certain other Cryptogamic phyla had a common origin with the Ferns. The Equisetales are at present a well-defined group; even in the rich Palaeozoic floras the habit, anatomy and reproductive characters usually render the members of this class unmistakable, in spite of the great development and stature which they then attained. It is interesting, however, to find that in the oldest known representatives of the Equisetales the leaves were highly developed and dichotomously divided, thus differing greatly from the mere scale-leaves of the recent Horsetails, or even from the simple linear leaves of the later Calamites. The early members of the class, in their forked leaves, and in anatomical characters, show an approximation to the Sphenophyllales, which are chiefly represented by the large genus Sphenophyllum, ranging through the Palaeozoic from the Middle Devonian onwards. These were plants with rather slender, ribbed stems, bearing whorls of wedge-shaped or deeply forked leaves, six being the typical number in each whorl. From their weak habit it has been conjectured, with much probability, that they may have been climbing plants, like the scrambling Bedstraws of our hedgerows. The anatomy of the stem is simple and root-like; the cones are remarkable for the fact that each scale or sporophyll is a double structure, consisting of a lower, usually sterile lobe and one or more upper lobes bearing the sporangia; in one species both parts of the sporophyll were fertile. Sphenophyllum was evidently much specialised; the only other known genus is based on an isolated cone, Cheirostrobus, of Lower Carboniferous age, with an extraordinarily complex structure. In this genus especially, but also in the entire group, there is an evident relation to the Equisetales; hence it is of great interest that Nathorst has described, from the Devonian of Bear Island in the Arctic regions, a new genus Pseudobornia, consisting of large plants, remarkable for their highly compound leaves which, when found detached, were taken for the fronds of a Fern. The whorled arrangement of the leaves, and the habit of the plant, suggest affinities either with the Equisetales or the Sphenophyllales; Nathorst makes the genus the type of a new class, the Pseudoborniales. (A.G. Nathorst, "Zur Oberdevonischen Flora der Baren-Insel", "Kongl. Svenska Vetenskaps-Akademiens Handlingar" Bd. 36, No. 3, Stockholm, 1902.)

The available data, though still very fragmentary, certainly suggest that both Equisetales and Sphenophyllales may have sprung from a common stock having certain fern-like characters. On the other hand the Sphenophylls, and especially the peculiar genus Cheirostrobus, have in their anatomy a good deal in common with the Lycopods, and of late years they have been regarded as the derivatives of a stock common to that class and the Equisetales. At any rate the characters of the Sphenophyllales and of the new group Pseudoborniales suggest the existence, at a very early period, of a synthetic race of plants, combining the characters of various phyla of the Vascular Cryptogams. It may further be mentioned that the Psilotaceae, an isolated epiphytic family hitherto referred to the Lycopods, have been regarded by several recent authors as the last survivors of the Sphenophyllales, which they resemble both in their anatomy and in the position of their sporangia.

The Lycopods, so far as their early history is known, are remarkable rather for their high development in Palaeozoic times than for any indications of a more primitive ancestry. In the recent Flora, two of the four living genera (Excluding Psilotaceae.) (Selaginella and Isoetes) have spores of two kinds, while the other two (Lycopodium and Phylloglossum) are homosporous. Curiously enough, no certain instance of a homosporous Palaeozoic Lycopod has yet been discovered, though well-preserved fructifications are numerous. Wherever the facts have been definitely ascertained, we find two kinds of spore, differentiated quite as sharply as in any living members of the group. Some of the Palaeozoic Lycopods, in fact, went further, and produced bodies of the nature of seeds, some of which were actually regarded, for many years, as the seeds of Gymnosperms. This specially advanced form of fructification goes back at least as far as the Lower Carboniferous, while the oldest known genus of Lycopods, Bothrodendron, which is found in the Devonian, though not seed-bearing, was typically heterosporous, if we may judge from the Coal-measure species. No doubt homosporous Lycopods existed, but the great prevalence of the higher mode of reproduction in days which to us appear ancient, shows how long a course of evolution must have already been passed through before the oldest known members of the group came into being. The other characters of the Palaeozoic Lycopods tell the same tale; most of them attained the stature of trees, with a corresponding elaboration of anatomical structure, and even the herbaceous forms show no special simplicity. It appears from recent work that herbaceous Lycopods, indistinguishable from our recent Selaginellas, already existed in the time of the Coal-measures, while one herbaceous form (Miadesmia) is known to have borne seeds.

The utmost that can be said for primitiveness of character in Palaeozoic Lycopods is that the anatomy of the stem, in its primary ground-plan, as distinguished from its secondary growth, was simpler than that of most Lycopodiums and Selaginellas at the present day. There are also some peculiarities in the underground organs (Stigmaria) which suggest the possibility of a somewhat imperfect differentiation between root and stem, but precisely parallel difficulties are met with in the case of the living Selaginellas, and in some degree in species of Lycopodium.

In spite of their high development in past ages the Lycopods, recent and fossil, constitute, on the whole, a homogeneous group, and there is little at present to connect them with other phyla. Anatomically some relation to the Sphenophylls is indicated, and perhaps the recent Psilotaceae give some support to this connection, for while their nearest alliance appears to be with the Sphenophylls, they approach the Lycopods in anatomy, habit, and mode of branching.

The typically microphyllous character of the Lycopods, and the simple relation between sporangium and sporophyll which obtains throughout the class, have led various botanists to regard them as the most primitive phylum of the Vascular Cryptogams. There is nothing in the fossil record to disprove this view, but neither is there anything to support it, for this class so far as we know is no more ancient than the megaphyllous Cryptogams, and its earliest representatives show no special simplicity. If the indications of affinity with Sphenophylls are of any value the Lycopods are open to suspicion of reduction from a megaphyllous ancestry, but there is no direct palaeontological evidence for such a history.

The general conclusions to which we are led by a consideration of the fossil record of the Vascular Cryptogams are still very hypothetical, but may be provisionally stated as follows:

The Ferns go back to the earliest known period. In Mesozoic times practically all the existing families had appeared; in the Palaeozoic the class was less extensive than formerly believed, a majority of the supposed Ferns of that age having proved to be seed-bearing plants. The oldest authentic representatives of the Ferns were megaphyllous plants, broadly speaking, of the same type as those of later epochs, though differing much in detail. As far back as the record extends they show no sign of becoming merged with other phyla in any synthetic group.

The Equisetales likewise have a long history, and manifestly attained their greatest development in Palaeozoic times. Their oldest forms show an approach to the extinct class Sphenophyllales, which connects them to some extent, by anatomical characters, with the Lycopods. At the same time the oldest Equisetales show a somewhat megaphyllous character, which was more marked in the Devonian Pseudoborniales. Some remote affinity with the Ferns (which has also been upheld on other grounds) may thus be indicated. It is possible that in the Sphenophyllales we may have the much-modified representatives of a very ancient synthetic group.

The Lycopods likewise attained their maximum in the Palaeozoic, and show, on the whole, a greater elaboration of structure in their early forms than at any later period, while at the same time maintaining a considerable degree of uniformity in morphological characters throughout their history. The Sphenophyllales are the only other class with which they show any relation; if such a connection existed, the common point of origin must lie exceedingly far back.

The fossil record, as at present known, cannot, in the nature of things, throw any direct light on what is perhaps the most disputed question in the morphology of plants—the origin of the alternating generations of the higher Cryptogams and the Spermophyta. At the earliest period to which terrestrial plants have been traced back all the groups of Vascular Cryptogams were in a highly advanced stage of evolution, while innumerable Seed-plants—presumably the descendants of Cryptogamic ancestors—were already flourishing. On the other hand we know practically nothing of Palaeozoic Bryophyta, and the evidence even for their existence at that period cannot be termed conclusive. While there are thus no palaeontological grounds for the hypothesis that the Vascular plants came of a Bryophytic stock, the question of their actual origin remains unsolved.

III. NATURAL SELECTION.

Hitherto we have considered the palaeontological record of plants in relation to Evolution. The question remains, whether the record throws any light on the theory of which Darwin and Wallace were the authors—that of Natural Selection. The subject is clearly one which must be investigated by other methods than those of the palaeontologist; still there are certain important points involved, on which the palaeontological record appears to bear.

One of these points is the supposed distinction between morphological and adaptive characters, on which Nageli, in particular, laid so much stress. The question is a difficult one; it was discussed by Darwin ("Origin of Species" (6th edition), pages 170-176.), who, while showing that the apparent distinction is in part to be explained by our imperfect knowledge of function, recognised the existence of important morphological characters which are not adaptations. The following passage expresses his conclusion. "Thus, as I am inclined to believe, morphological differences, which we consider as important—such as the arrangement of the leaves, the divisions of the flower or of the ovarium, the position of the ovules, etc.—first appeared in many cases as fluctuating variations, which sooner or later became constant through the nature of the organism and of the surrounding conditions, as well as through the inter-crossing of distinct individuals, but not through natural selection; for as these morphological characters do not affect the welfare of the species, any slight deviations in them could not have been governed or accumulated through this latter agency." (Ibid. page 176.)

This is a sufficiently liberal concession; Nageli, however, went much further when he said: "I do not know among plants a morphological modification which can be explained on utilitarian principles." (See "More Letters", Vol. II. page 375 (footnote).) If this were true the field of Natural Selection would be so seriously restricted, as to leave the theory only a very limited importance.

It can be shown, as the writer believes, that many typical "morphological characters," on which the distinction between great classes of plants is based, were adaptive in origin, and even that their constancy is due to their functional importance. Only one or two cases will be mentioned, where the fossil evidence affects the question.

The pollen-tube is one of the most important morphological characters of the Spermophyta as now existing—in fact the name Siphonogama is used by Engler in his classification, as expressing a peculiarly constant character of the Seed-plants. Yet the pollen-tube is a manifest adaptation, following on the adoption of the seed-habit, and serving first to bring the spermatozoids with greater precision to their goal, and ultimately to relieve them of the necessity for independent movement. The pollen-tube is constant because it has proved to be indispensable.

In the Palaeozoic Seed-plants there are a number of instances in which the pollen-grains, contained in the pollen-chamber of a seed, are so beautifully preserved that the presence of a group of cells within the grain can be demonstrated; sometimes we can even see how the cell-walls broke down to emit the sperms, and quite lately it is said that the sperms themselves have been recognised. (F.W. Oliver, "On Physostoma elegans, an archaic type of seed from the Palaeozoic Rocks", "Annals of Botany", January, 1909. See also the earlier papers there cited.) In no case, however, is there as yet any satisfactory evidence for the formation of a pollen-tube; it is probable that in these early Seed-plants the pollen-grains remained at about the evolutionary level of the microspores in Pilularia or Selaginella, and discharged their spermatozoids directly, leaving them to find their own way to the female cells. It thus appears that there were once Spermophyta without pollen-tubes. The pollen-tube method ultimately prevailed, becoming a constant "morphological character," for no other reason than because, under the new conditions, it provided a more perfect mechanism for the accomplishment of the act of fertilisation. We have still, in the Cycads and Ginkgo, the transitional case, where the tube remains short, serves mainly as an anchor and water-reservoir, but yet is able, by its slight growth, to give the spermatozoids a "lift" in the right direction. In other Seed-plants the sperms are mere passengers, carried all the way by the pollen-tube; this fact has alone rendered the Angiospermous method of fertilisation through a stigma possible.

We may next take the seed itself—the very type of a morphological character. Our fossil record does not go far enough back to tell us the origin of the seed in the Cycadophyta and Pteridosperms (the main line of its development) but some interesting sidelights may be obtained from the Lycopod phylum. In two Palaeozoic genera, as we have seen, seed-like organs are known to have been developed, resembling true seeds in the presence of an integument and of a single functional embryo-sac, as well as in some other points. We will call these organs "seeds" for the sake of shortness. In one genus (Lepidocarpon) the seeds were borne on a cone indistinguishable from that of the ordinary cryptogamic Lepidodendreae, the typical Lycopods of the period, while the seed itself retained much of the detailed structure of the sporangium of that family. In the second genus, Miadesmia, the seed-bearing plant was herbaceous, and much like a recent Selaginella. (See Margaret Benson, "Miadesmia membranacea, a new Palaeozoic Lycopod with a seed-like structure", "Phil. Trans. Royal Soc. Vol." 199, B. 1908.) The seeds of the two genera are differently constructed, and evidently had an independent origin. Here, then, we have seeds arising casually, as it were, at different points among plants which otherwise retain all the characters of their cryptogamic fellows; the seed is not yet a morphological character of importance. To suppose that in these isolated cases the seed sprang into being in obedience to a Law of Advance ("Vervollkommungsprincip"), from which other contemporary Lycopods were exempt, involves us in unnecessary mysticism. On the other hand it is not difficult to see how these seeds may have arisen, as adaptive structures, under the influence of Natural Selection. The seed-like structure afforded protection to the prothallus, and may have enabled the embryo to be launched on the world in greater security. There was further, as we may suppose, a gain in certainty of fertilisation. As the writer has pointed out elsewhere, the chances against the necessary association of the small male with the large female spores must have been enormously great when the cones were borne high up on tall trees. The same difficulty may have existed in the case of the herbaceous Miadesmia, if, as Miss Benson conjectures, it was an epiphyte. One way of solving the problem was for pollination to take place while the megaspore was still on the parent plant, and this is just what the formation of an ovule or seed was likely to secure.

The seeds of the Pteridosperms, unlike those of the Lycopod stock, have not yet been found in statu nascendi—in all known cases they were already highly developed organs and far removed from the cryptogamic sporangium. But in two respects we find that these seeds, or some of them, had not yet realised their possibilities. In the seed of Lyginodendron and other cases the micropyle, or orifice of the integument, was not the passage through which the pollen entered; the open neck of the pollen-chamber protruded through the micropyle and itself received the pollen. We have met with an analogous case, at a more advanced stage of evolution, in the Bennettiteae, where the wall of the gynaecium, though otherwise closed, did not provide a stigma to catch the pollen, but allowed the micropyles of the ovules to protrude and receive the pollen in the old gymnospermous fashion. The integument in the one case and the pistil in the other had not yet assumed all the functions to which the organ ultimately became adapted. Again, no Palaeozoic seed has yet been found to contain an embryo, though the preservation is often good enough for it to have been recognised if present. It is probable that the nursing of the embryo had not yet come to be one of the functions of the seed, and that the whole embryonic development was relegated to the germination stage.

In these two points, the reception of the pollen by the micropyle and the nursing of the embryo, it appears that many Palaeozoic seeds were imperfect, as compared with the typical seeds of later times. As evolution went on, one function was superadded on another, and it appears impossible to resist the conclusion that the whole differentiation of the seed was a process of adaptation, and consequently governed by Natural Selection, just as much as the specialisation of the rostellum in an Orchid, or of the pappus in a Composite.

Did space allow, other examples might be added. We may venture to maintain that the glimpses which the fossil record allows us into early stages in the evolution of organs now of high systematic importance, by no means justify the belief in any essential distinction between morphological and adaptive characters.

Another point, closely connected with Darwin's theory, on which the fossil history of plants has been supposed to have some bearing, is the question of Mutation, as opposed to indefinite variation. Arber and Parkin, in their interesting memoir on the Origin of Angiosperms, have suggested calling in Mutation to explain the apparently sudden transition from the cycadean to the angiospermous type of foliage, in late Mesozoic times, though they express themselves with much caution, and point out "a distinct danger that Mutation may become the last resort of the phylogenetically destitute"!

The distinguished French palaeobotanists, Grand'Eury (C. Grand'Eury, "Sur les mutations de quelques Plantes fossiles du Terrain houiller". "Comptes Rendus", CXLII. page 25, 1906.) and Zeiller (R. Zeiller "Les Vegetaux fossiles et leurs Enchainements", "Revue du Mois", III. February, 1907.), are of opinion, to quote the words of the latter writer, that the facts of fossil Botany are in agreement with the sudden appearance of new forms, differing by marked characters from those that have given them birth; he adds that these results give more amplitude to this idea of Mutation, extending it to groups of a higher order, and even revealing the existence of discontinuous series between the successive terms of which we yet recognise bonds of filiation. (Loc. cit. page 23.)

If Zeiller's opinion should be confirmed, it would no doubt be a serious blow to the Darwinian theory. As Darwin said: "Under a scientific point of view, and as leading to further investigation, but little advantage is gained by believing that new forms are suddenly developed in an inexplicable manner from old and widely different forms, over the old belief in the creation of species from the dust of the earth." ("Origin of Species", page 424.)

It most however be pointed out, that such mutations as Zeiller, and to some extent Arber and Parkin, appear to have in view, bridging the gulf between different Orders and Classes, bear no relation to any mutations which have been actually observed, such as the comparatively small changes, of sub-specific value, described by De Vries in the type-case of Oenothera Lamarckiana. The results of palaeobotanical research have undoubtedly tended to fill up gaps in the Natural System of plants—that many such gaps still persist is not surprising; their presence may well serve as an incentive to further research but does not, as it seems to the writer, justify the assumption of changes in the past, wholly without analogy among living organisms.