mations to the seeds of Gnetum gnemon and G. africanum, as proved by Emily M. Berridge and Mrs. Thoday.
The Williamsonian Tribe.—The boundaries of fossil groups are always obscure, or at least inferential. They change as discovery pro ceeds; but they are never settled. So, coexist ent with the Cycadeoidea, and without sharp As far as fossil preservation goes, Pier ophyllum and Zamites are the representative ancient lines, with Sphenozamites as a relative merely characterized by oddity of leaf form. And these old lines pass on to add their quota to the great Williamsonian group of the Rhwtic, the period of most rapid evolution. Culmina Lion occurs in the lower Jura; while extinction begins in the middle Jura and progresses stead ily to the upper Cretaceous, where the Cyca deoideas also find their final representation in both Europe and America.* Cycadeoid Relationships.— In comparing the stem and leaf structure of cycads and cyca deoids it is seen that the distinct differences generally have the nature of singularities ac centuated by separation in time. The existing cycads go to the Jurassic. Then, doubtless, they still retained their ramentum and various older features. The main gap is due to the or ganization of the sporophylls into cones in the cycads, into flowers in the cycadeoids; and that this severance did not occur later than the Per mian is an inference resting not merely on the general evidence, but on the cognate fact that the coniferous types also underwent much similar change in that age. The fossil trunks while these traces* may have a concen tric structure, the only cycadeoid cortical traces which are concentric pertain to peduncles. Thirdly, there are sometimes found in the cycad pith vascular cone domes due to the nearly or differ from the modern in four lesser, although distinct features, either of which affords a basis of separation. Firstly, in the cycadeoids the single heavy leaf trace of horse-shoe section arises from the lower angle of the leaf gap and passes directly out to the leaf base, while in the cycads the trace is double. In them, the two strands sometimes nearly girdle the trunk, although such girdle leaf traces are less in evi dence in Zamia and Stangeria. Secondly, incipiently terminal position of the enormous cones. As the vegetative axis reasserts itself the peduncle bundle strands and cylinder are grown round by the main woody cylinder. No parallel can arise in the cycadeoids, where the peduncles are all lateral, as indeed they are in most cycads. Fourthly, peduncular centrifugal wood, the remnant of the old cryptogamic wood, and still present in some modern cycads, is entirely lost in the cycadeoids, notwithstand ing their greater age. In fructification the cy cads and cycadeoids at first sight appear utterly dissimilar. But on closer consideration im portant reciprocal relationships appear. There is an Indian form suggesting a staminate whorl, rather than a disc; while there is no question of the strobilar nature of the ovulate fruit, whether to be regarded as a cone, or as an inflorescence made up of secondary axes of sterile scales clustered about a central seed stein, somewhat after the manner of the Pan demur cone. Furthermore, diffuse strobilar types, or those with more leafy sporophylls, are in evidence in Mesozoic times in both conifer ous and other gymnospermous types. Also, it
appears that in gymnosperms there are dis tinct gradations from carpellary leaves to cone scales, and that there is no impassable hiatus between cones, whorls of fertile leaves and discs. In cycads the microsporophylls are re duced to cone-scales, while the magasporo phylls are in the carpellary stage. In the cyca deoids exactly the opposite, or complementary course of change went on; the seed is simple stalked (as in Araucaria), and the microsporo phyll is frond-like. It has, too, that curious eared feature, which also appears primitively significant. Just as in various cycad cones both types of sporophylls have an eared or diceratoid character, so in the Cycadeoidea microsporophyll the same feature recurs; and the further fact is there learned for the first time that the position or insertion of the spore in the so-called dorsal or ventral position may depend on the mode of prefloration and reduc tion. Thus, owing solely to the manner in which the sporophylls of Cycadeoidea are or ganized into a disc, the synangia reach inser tion on the superior surface, although of cer tain derivation from pendent Maratfioid types. And this evidence bears directly on the origin of the ovuliferous scale of conifers. It can be readily seen that reduction of an eared mega sporophyll, preflorately deflexed, must lead to a complex ovuliferous scale simulating that of the modern conifers. Finally, in the article on Araucaria it was pointed out that a certain parallel with Cycadeoitiea exists. Aside from cycads, Araucaria must be considered the nearest relative, despite the floral parallel seer. in the Tumboa disc and the seeds of Gnetum. Fundamentally important is the fact that with the splitting up of the scalariform pits and loss of tracheidal end walls, vessels could arise, and that with medullar decrease and the develop ment of storage tissue, essentially dicotyled onous wood structures like those seen in Magnolia, Drimys and Trochodendron would be reached. Taken in conjunction with the floral plan, simulating that of Liriodendron and other Magnoliacem, there is indicated a possible, if not even a probable, mode of an giosperm derivation. Narrow indeed must have been the gap between upland and espe cially boreal cycadeoids and the •rue precursor or pro-angiosperms in early Mesozoic times. The only alternative hypothesis to angiosperm derivation from within or near to the cycadeoid stock, is a far-reaching homoplasy and paral lelism continuing throughout all post-Palzozoic time. See CYCADOPHYTA; PALEOBOTANY.
Bibliography.— The record of discovery and investigation of the Cycadales, existing and fossil, is to be found mainly in the scien tific writings of Buckland, von Mohl, Solms Miguel, Lignier, Zeiller, Fliche, Nathorst, Scott, Worsdell, Stopes, Seward, Ikeno, Web ber, Ward, Chamberlain and Wieland. The titles are given with essential completeness in the bibliography appended to Wieland's 'Amer ican Fossil Cycads' (Carnegie Institution of Washington Publication No. 34, Vol. I, 1906; Vol. II, 1916). For a still more recent account consult Seward's 'Fossil Plants' (Vol. III, Cambridge, England, 1917).