Popular Science Monthly/Volume 80/February 1912/The History of Gymnosperms
| THE HISTORY OF GYMNOSPERMS |
By Professor JOHN M. COULTER
UNIVERSITY OF CHICAGO
NO great group of plants seems to have left so continuous and full a record of itself, through so long a stretch of time, as have the Gymnosperms. Further work may uncover equally extensive records of certain other vascular groups, but our knowledge of the history of Gymnosperms is at present more complete than that of any other group of plants.
Several things have contributed to the completeness of this knowledge of Gymnosperms. They have always been abundant in the flora of every period that has left a plant record, and they are still abundant. This has given continuity and wealth of material throughout the whole history of vegetation, so far as that history is known. Those who picture an historical succession of the great plant groups are not representing the records in sight, for our earliest records show a vegetation as varied and complex as that of to-day, so far as the great groups are concerned. Although Angiosperms are probably relatively modern, seed-plants were represented early in the Paleozoic by Gymnosperms. This means that the evolution of the plant kingdom, in all its essential outlines, had been attained at least as early as any known records of vegetation.
Another fact which has contributed to the completeness of our knowledge of the group is what may be called the renascence of paleobotany as a morphological subject. Not only did this involve the comparative study from sections, of the essential structures, but also it enormously extended the range of structures used in indicating relationships, by including the vascular system in the evolutionary scheme. The incorporation of vascular anatomy into modern morphology was significant not merely because it supplied another line of attack, but also because it deals with the most completely recorded structure of vascular plants and really gives continuity to the paleobotanical record.
The Gymnosperms were represented during the Paleozoic by two great groups, Cycadofilicales and Cordaitales. They were not merely members of the Paleozoic flora, but they were conspicuous members, together constituting the seed-plant vegetation. Neither of these groups has been traced with certainty into the Mesozoic, so that even though a few lingering forms may be found at a later period, they are essentially restricted to the Paleozoic, and our knowledge of them has been derived 
chiefly from material obtained from the Coal-measures. Although the two groups are equally ancient, so far as our records go, having been recognized well into the Middle Devonian, it is quite evident from comparative structures that the Cycadofilicales are the more primitive, and presumably the more ancient. Whether they are actually the most ancient seed plants or not, they are at least the most primitive seed plants of which we have any knowledge.
The presence of this great group of primitive seed plants (Cycadofilicales) in the Paleozoic was obscured for many years by the impression that they were ferns. A very large percentage of the Coal-measure vegetation consisted of these fern-like plants, and so the coal period was pictured as a time of luxuriant fern vegetation, rivalling our present tropics in that feature. Approximately ten years ago these fern-like plants were observed to bear seeds, and the Cycadofilicales became established as the most fern-like group of Gymnosperms. All of the great Paleozoic "fern" groups were found involved in the seed-bearing habit, until now the residuum of real ferns in the Paleozoic seems to be quite small. In any event, it has been made clear that the Cycadofilicales were derived from ferns; and if so, probably all the other Gymnosperms. It should be understood that the ordinary ferns of to-day are relatively modern, and are quite unlike those very ancient ferns which gave rise to the Cycadofilicales, and which have received the general name Primofilices.
This ancient group of Gymnosperms resembled ferns in every important particular except in the seed-bearing habit. Whereas in ordinary ferns the sporangia are borne in groups or so-called "fruit dots"
(sori) on the fronds, in Cycadofilicales some of the sori were replaced by seeds, which makes a seed the morphological equivalent of a sporangium or a group of sporangia (a sorus). The bearing of seeds necessitated also the presence of structures corresponding to stamens, and producing pollen. These pollen-bearing structures remained like the fern sporangia (in sori), and for a long time confirmed the notion that these fern-like plants were really ferns. To say that a fern-like leaf must belong to a fern might be unsafe, but to say that such a leaf bearing sporangia in sori must belong to a fern seemed absolutely safe. And still many of these "fern sporangia" have turned out to be the pollen sacs of seed plants.
If the bearing of seeds distinguished Cycadofilicales from ferns, the absence of cones distinguished them from other Gymnosperms. The seeds and pollen sacs were borne as freely on the fronds as are the sporangia on the fronds of ferns. In the later groups of Gymnosperms, the seed-bearing leaves and pollen sac-bearing leaves (both kinds called "sporophylls") became distinct from the ordinary foliage leaves, and were finally compactly organized into the cone-like structure (strobilus) characteristic of most Gymnosperms. But among the Cycadofilicales the strobilus stage was not reached.
The Cycadofilicales seem to have given rise to two great branches of Gymnosperms, both of which are represented in the present flora. One of them includes the Cycads, and therefore have been called the Cycadophytes; the other includes the Conifers, and therefore may be called the Coniferophytes. The Coniferophytes differentiated from the Cycadofilicales earlier than our records of vascular vegetation, for the Paleozoic representative (Cordaitales) of Coniferophytes is distinct from the Cycadofilicales as far back as records go. On the other hand. the Cycadophyte branch is not distinguishable until the Mesozoic.
Historically, therefore, the Cordaitales must be considered as the second group of Gymnosperms. Their connection with an ancient fern stock is evident in their structure, but they have lost many of the fern characters that were retained by Cycadofilicales. The fact that Cordaitales are much further from ferns than are the Cycadofilicales is perhaps the best proof that they have come from the ferns by way of the Cycadofilicales. The combination of changes involved in their structure is all in the direction of the later Conifers, as, for example, the branching stem (constituting what is called "the habit") with its thick cylinder of secondary wood, the narrow and entire leaves, and the cones (strobili). It would not fit the purpose of this presentation to include the changes in the more intimate structures, since their nature and significance can be appreciated only by the special students of the group, but they are just as striking as the more obvious changes mentioned.
The Gymnosperm vegetation of the Paleozoic, therefore, comprised two great genetic groups: the Cycadofilicales, representing the primitive Gymnosperm stock that differentiated from the ferns; and the Cordaitales, representing the primitive Coniferophyte stock that differentiated from Cycadofilicales more ancient than those we know.
In the Mesozoic flora the Gymnosperms were represented by four great groups, evidently derived from the two Paleozoic groups. As stated above, the Cycadophyte branch became distinct, and for a long time all of its representatives were thought to be Cycads. For this reason, the Mesozoic has been called the "age of Cycads," so far as the vegetation is concerned. It is one of the triumphs of American paleobotany that it has put on a firm basis our knowledge of the great Cycadophyte group of the Mesozoic, and has shown that it is quite different from the modern Cycads. The group is called Bennettitales, and although a few forms from foreign localities have been known for a long time, it remained for Mesozoic deposits of the United States and Mexico to reveal a remarkably rich display of forms in admirable preservation. The investigation of this material has been carried on chiefly by Dr. G. E. Wieland of the Yale Museum.
The Bennettitales, therefore, are the so-called "fossil Cycads" of the Mesozoic. So far as the records show, they are restricted to the Mesozoic, so that they represent an extinct Mesozoic group, just as there are two extinct Paleozoic groups. Of course it is not only conceivable, but also probable that the Paleozoic Cycadofilicales, from which Bennettitales were derived, continued into early Mesozoic; and that the Mesozoic Bennettitales began to differentiate in late Paleozoic. The external appearance of Bennettitales justifies their early assignment to the Cycads, for the whole habit is Cycadean. The stems are either tuberous or cylindrical, and crowned by a rosette of large, fern-like leaves, giving to the cylindrical forms the appearance of tree ferns. The remarkable feature of Bennettitales, however, is the cone (strobilus), whose structure is unique among Gymnosperms. These cones, instead of being solitary and terminal, in the midst of the rosette of leaves, as in most Cycads, are lateral on dwarf branches which arise in profusion from the stem. But this is a small feature as compared with the fact that the cone is "bisporangiate." In other Gymnosperms the ovules (and of course seeds) and stamens are in different cones, and often these cones are on different plants. In Bennettitales, the rosette of stamens, which resemble small fern fronds bearing sporangia, subtends the more or less extended axis-bearing ovules, and both stamens and ovules are encased by enveloping bracts. This bisporangiate character, and the relation of stamens to ovules in the cone, are so suggestive of such an Angiosperm flower as that of magnolia that some botanists would see in Bennettitales the ancestral forms of Angiosperms. It is certainly true that the Bennettitales were abundant and wide-spread during the Mesozoic, and it seems to be true that the Angiosperms originated during the Mesozoic.
The Cycads (Cyadales) constitute a second group of Mesozoic Gymnosperms, associated with the Bennettitales of common origin, but apparently not a conspicuous part of the vegetation. It is very likely true that Bennettitales and Cycadales represent two independent Mesozoic derivatives from the Paleozoic Cycadofilicales: that the Bennettitales attained a dominant place in the Mesozoic flora; and that the Cycadales, much less conspicuous during the Mesozoic, persisted until the present day as the only living representatives of the Cycadophytes.
The Cycadophyte line is characterized by the retention of many of the fern-like features of the Cycadofilicales. In habit, in foliage, in stem structure, in sporangia, in reproductive habits, the features of Cycadofilicales were continued; so that the living Cycads, although relatively modern from the standpoint of history, are structurally the most primitive of living Gymnosperms, because they most resemble the historically ancient Cycadofilicales.
The two other groups of Mesozoic Gymnosperms were derived from the Paleozoic Cordaitales, the Paleozoic member of the Coniferophyte branch. They are known as Ginkgoales and Coniferales, the former being nearly or quite extinct to-day, and the latter comprising the present conspicuous Gymnosperm vegetation of the temperate regions.
The Ginkgoales were abundant during the Mesozoic, but apparently remained quite constant in characters, so that they can be represented structurally by a single line extending from the late Paleozoic to the present time. They are really a Mesozoic type, and their single representative in the present flora has probably continued to exist simply because it is a tree kept in cultivation in the temple grounds of China and Japan. Of course Ginkgoales continued all the features of Cordaitales that looked towards Conifers, such as the branching habit, relatively simple leaves, and thick vascular cylinder. The monosporangiate cones were also continued, but the most notable feature is the continuation of the swimming sperms ©f Cordaitales. Cordaitales had continued the swimming sperms of Cycadofilicales and ferns; in fact all the vascular plants of the Paleozoic had swimming sperms. One of the most primitive features of the Cycadophyte branch is that it retained throughout this primitive, fern type of male cell. In the present Gymnosperm flora, therefore, Cycads and Ginkgos are distinguished by having swimming sperms, the former having continued them directly from the Cycadofilicales, the latter obtaining them indirectly from the same source, and directly through the Cordaitales. To state the situation in other terms, it may be said that the Cycads have continued primitive vegetative and reproductive structures, while the Ginkgos have retained primitive reproductive structures and have changed the vegetative structures, a change initiated by the Cordaitales.
All of this serves to emphasize the position of the Coniferales, the fourth Mesozoic group, and the dominant Gymnosperm group to-day. It not only continued to change the vegetative structures, but it also abandoned the primitive features of reproduction in abandoning the swimming sperm, which became a relatively passive cell conducted to the egg by a pollen tube. Few persons realize that Gymnosperms in general, in terms of great groups, have swimming sperms, which the pollen tubes do not conduct; for this situation is overshadowed by the fact that the single overwhelming group of Gymnosperms to-day has passive sperms conducted by pollen tubes.
In connection with this change of reproductive habit, the Coniferales during the Mesozoic differentiated into the six great families or tribes recognized to-day, so that it is the one great group of Gymnosperms that developed an extensive range of forms. The historical interest connected with Conifers, therefore, is not so much the origin of the group as a whole, for that seems to be traced clearly to the Paleozoic Cordaitales, as the origin and relative antiquity of its tribes. This question has been answered by the study of vascular anatomy, chiefly by Jeffrey of Harvard, supported by the morphology of the reproductive structures, and by history. The conclusion is that the tribe containing the pines is to be regarded as including the modern representatives of the most ancient Conifers. The only possible contestant for this honor is the tribe comprising the araucarians of the southern hemisphere. The four other tribes (podocarps, taxads, taxodiums and cypresses) are clearly relatively modern. The general conclusion, therefore, is that the Conifer stock became differentiated from the Cordaitales with features characteristic of the pine tribe, and that from primitive stock the other tribes separated later, and the pine tribe itself became more modern.
The history of Gymnosperms is not complete without mention of its seventh great group, the Gnetales. They comprise three genera, extremely unlike in habit, habitat and geographical distribution, but held together by certain important characters. They are so different from other Gymnosperms that no phylogenetic connection is clear, and there is no sure record of them as fossils. This seems to indicate that they are very modern, but their dissimilarity and their wide separation from one another geographically make it certain that if they are genetically related they must have a history that remains to be discovered. Such suggestions of connection as can be obtained indicate as a possibility that Gnetales are derivatives from the prolific Coniferales stock.