August 20, 1903] 



NATURE 



3^1 



true " fern-fronds " in structure as well as in external 

 aspect. The vascular bundle traversing the petiole, for 

 example, is of the " concentric " type characteristic of 

 ferns, and any differences there may be are in details only. 



A section of the stem, however, bears at first sight no 

 resemblance to that of a fern ; outside the pith we find, in 

 all mature specimens, a broad zone of wood and bast with 

 its cells arranged regularly in radial series, like that of an 

 ordinary " exogenous " tree, and in detail approaching 

 especially the cycadean structure. .^t the border of the 

 pith there are distinct strands of wood, and this region, 

 which was laid down before the radially arranged zone, 

 recalls the structure of an Osmunda. The bundles in the 

 cortex of the stem, on their way out to the leaves, have, in 

 this part of their course, exactly the structure of the strands 

 in the leaf-stalk of a Cycad — a structure found, in this 

 form, in no other living plants. 



The roots, when young, resembled those of certain ferns 

 (Marattiaceae), but as they grew older they also formed 

 radially arranged wood and bast like the roots of Gymno- 

 sperms. 



On the ground of this remarkable combination of 

 structural characters, it was inferred that Lyginodendron 

 could not have been a true fern, but must have occupied a 

 position intermediate between the ferns and the cycadean 

 type of Gymnosperms. 



k similar association of diverse anatomical characters has 

 now been proved to e.xist in various other quasi-ferns of 

 PalcTBOzoic age. In Heterangium, for example, also investi- 

 gated by Williamson, leaves and roots resemble those of 

 the previous genus, but the stem is more obviously fern-like, 

 agreeing in its earlier stages with that of a Gleichenia, but 

 acquiring, with advancing age, a zone of secondary wood 

 and bast of the cycadean type. This plant likewise bore 

 foliage of the Sphencpteris form (S. elcgans). 



In Medullosa, on the other hand, to which the Aletho- 

 pteris and Neuropteris foliage belonged, the original ground- 

 plan of the tissues in the stem is like that of a complex 

 fern, but the structure of leaves and roots, and the secondary 

 structure of the stem itself, is almost purely cycadean. 

 We might continue the list much further. Wherever 

 one of these quasi-ferns has been examined anatomically, a 

 similar combination of characters has been found. It may 

 be pointed out in passing that, while many of these inter- 

 mediate forms lead on towards the Cycadophyta themselves, 

 others approach more nearly to the extinct family 

 Cordaite.-t, and indicate that they also, though so different 

 from ferns in habit, may yet have sprung from the same 

 stock. 



But so far the positive evidence has been wholly 

 anatomical, and botanists are not yet altogether in agree- 

 ment as to the value of anatomical characters. The 

 anatomist very naturally thinks that there is nothing like 

 anatomy, but the pure systematist will not be satisfied with- 

 out the characters en which he has been accustomed to 

 rely, and his faith in which has been so amply justified, 

 those, namely, draxyn from the reproductive organs. 

 Darwin, however, who neglected nothing, was fully alive 

 to the importance of anatomical evidence ; he expresses his 

 interest in an anatomical character in an amusing way in 

 one of his lately published letters (1861), saying, " The 

 destiny of the whole human race is as nothing compared to 

 the course of vessels in Orchids ! " 



Until the present year, we had no satisfactory knowledge 

 of the fructification in any one of the Cycadofilices, as we 

 now call them, of the Palaeozoic period. There is, it is 

 true, some reason to believe that a form of fructification 

 with long tufted spore-sacs belonged to Lyginodendron, 

 but we know nothing as yet as to the details — it may prove 

 to represent the male reproductive organs of the plant. 

 Among the unidentified seeds of the Coal-measures, there 

 are some — the great seeds known as Trigonocarpon — which 

 are not only associated with Medullosa, but which show 

 a certain structural resemblance to some of its tissues. 

 But still the indications were slight — so slight that Prof. 

 Zeiller, of Paris, than whom there is no higher authority, 

 has recently expressed a doubt whether these Cycadofilices 

 were, after all, anything more than a peculiar group of 

 ferns. 



Within the last few months, however, an altogether new 

 light has fallen on our subject. Among the seeds discovered 

 by Williamspn in the English Coal-measures were three 

 NO. 1764, VOL. 68] 



species which he placed in his genus Lagenostoma. These, 

 as we shall see, are characteristic seeds of complex struc- 

 ture. One of them, named L. Lomaxi by Williamson, 

 though not described by him, has lately been reinvesti- 

 gated, in the first instance by my friend Prof. F. W. 

 Oliver (see Nature, June 4). The great peculiarity about 

 it is that the seed itself was borne in a little calyx-like cup, 

 fitting loosely round it, just as a hazel nut is borne in its 

 husk. The cup, or cupule, which is deeply lobed, bears very 

 peculiar glandular bodies, usually with a short thick stalk 

 and a round head which is empty, as if the secretory tissue 

 had broken down. These glands, on the cupule of the seed, 

 have been found to agree exactly in dimensions, form, and 

 structure with the glands borne on the leaves and stems 

 of the particular form of Lyginodendron Oldhamium with 

 which the seeds are associated. 



Suppose that in some tropical forest where the trees were 

 too lofty for their leaves and fruits to be reached, seeds and 

 leaves and twigs were found scattered together on the 

 ground, and that they all proved to bear exactly similar 

 glandular outgrowths of a kind unknown elsewhere. 

 Suppose, further, that the structure of the envelope of the 

 seed turned out to agree in other respects with that of 

 the vegetative fragments, should we hesitate to conclude 

 that the seeds belonged to the same plants as the leaves 

 and twigs, though we had never seen them actually in 

 connection? Such is the argument with regard to the 

 relation of the seed Lagenostoma - to the plant Lygino- 

 dendron. Short of finding the vegetative and reproductive 

 organs in continuity, the proof is as strong as it can be, 

 and I think we need not hesitate to conclude that the one 

 belonged to the other. 



But, if this be so, the question as to the nature of the 

 Palaeozoic Cycadofilices is settled, at least as regards one 

 member of the group. Lyginodendron was already a seed- 

 bearing plant. The seeds are highly organised, and, 

 broadly speaking, of the cycadean type. The integument 

 and central body of the seed are closely joined to near the 

 tip and along the line of junction run the strands which 

 conveyed the water-supply. The upper part of the integu- 

 ment has a curious chambered structure — the central body 

 terminates in a large pollen-chamber of peculiar bell-shaped 

 form, in which the pollen-grains are sometimes found. The 

 neck of the pollen-chamber fits into the opening of the 

 integument and reaches the surface. The centre of the seed 

 is occupied by the large megaspore or embryo-sac, in which 

 remains of prothallial tissue can sometimes be detected. 

 The seed, in fact, is as highly differentiated as any seed of 

 its period, lacking only an embryo, as do all its con- 

 temporaries. 



But if Lyginodendron, with all its fern-like characters, 

 was thus a true seed-plant, we cannot doubt that other 

 quasi-ferns of that period, exhibiting a similar combination 

 of characters, had also entered the ranks of the Spermo- 

 phyta, and we may confidently expect that, one by one, 

 many of the as yet unowned Palaeozoic seeds will be traced 

 to their fern-like possessors. 



Further positive indications of this are already presenting 

 themselves. For example, there is a specimen in the 

 British Museum collection showing a cast of a branched 

 rachis accompanied by a multitude of ribbed seeds, many 

 of which are in clear connection with the rachis itself. At 

 one place we see a leaflet of Sphenopteris obtusiloba, a well- 

 known Coal-measure " fern," and everything indicates that 

 we have here the fertile, seed-bearing rachis of that species. 

 There are other specimens which point in the same direc- 

 tion, and now that the eyes of collectors are opened to 

 the possibility of their so-called " fern-fronds " bearing 

 seeds — an idea which before seemed too improbable to be 

 entertained — more of such specimens will doubtless find their 

 way into our museums. 



The present position, then, of our question is this. Some, 

 probably many, of the fern-like plants of Palaeozoic age 

 bore seeds of the same general structure as those of the 

 Cycads among living Gymnosperms. The plants in ques- 

 tion were not merely fern-like ; their anatomical structure 

 proves them to have had so much in common with true 

 ferns that there can be no doubt of their affinity with them. 

 In fact, apart from the newly discovered seeds, these plants, 

 for the most part, show a balance of characters on the fern 

 side. 



The evidence thus points unmistakably to the conclusion 



