Further Notes on Seed Structures. 149 



sonianus, and is difficult of study in the silicified American 

 specimens. Again, the French Paleozoic seeds are scarcely 

 known to us in that closer detail which present methods of 

 fossil seed study are rapidly bringing to light. 



But meanwhile further interesting comparisons with ancient 

 seeds have been made possible by the instructive interpreta- 

 tions of Paleozoic seed structures and noteworthy study by 

 Oliver and Salisbury (11) of the two species (Jonostoma 

 oblong um and C. anglogermanioum. 



In the species of Conostoma no cupule obscures the exosar- 

 cal features, but the seeds have the small size of those of 

 Cycadeoidea, in consequence of which they present much 

 difficulty of study due to dependence on chance sectional 

 planes. Nevertheless, by plotting about a score of sections of 

 Conostoma anglo-germanioum and nine sections of Conostoma 

 oblongum, and then constructing models from the estimated 

 planes of the sections, nearly concise restorations of both 



B. — More nearly transverse and basal section than the preceding in which 

 the dark staining mass clinging to the lower left cotyledon must be an endo- 

 sperm remnant rather than a tissue like the convolute extension just noted 

 in A. 



O. — Sublongitudinal section. Passes through lateral hilum and emerges 

 at the seed " shoulder," traversing nearly the entire length of the embryo 

 with its two cotyledons (c) and their lamina bundles (&). 



D. — More nearly longitudinal section than the preceding, cutting chalazal 

 traeheid stem or goblet, and passing beyond the " shoulder" near to base of 

 micropylar tube. The embryo is here especially instructive as showing the 

 forking (6) of the cotyledonary bundles in both lamina as well as the apical 

 proembryo folds and endosperm or albuminous remnants (a). Taking this 

 section at W the characteristic proportion and appearance of the seed wall 

 layers is indicated distinctly. On the inside next the embryo to which the 

 nucellar sack clings is a considerable empty space left in part by shrinkage 

 of the embryo, in part by collapse or failure of preservation of the spongy 

 enclosing cells of the nucellus, and in part finally by the invariable flatten- 

 ing out of the cells of the inner flesh, indicated by the inside solid black 

 line. 



The inner flesh clings closely to the middle sclerenchyma, also shown in 

 solid black but notched on the outer border, this being the artist's method of 

 indicating the cup-shaped appearance of the component cells due to preser- 

 vation of the thick inner or basal wall and thinning out of the lateral walls 

 without easily detected conservation of the very thin tangential wall. Exte- 

 riorly comes the outer flesh, or more exactly the "blow off layer" strictly 

 homologous to that of Carboniferous Cycadofilicinean seeds, Physostoma, 

 etc. But in strobili of trunk 131 and in most American Cycadeoidean seeds 

 the very thin cell walls of this layer are indistinct, though their characteris- 

 tic palisade-droop, so pronounced in Physostoma, cf . figure 4 D, is plainly in 

 evidence. Finally, the strongly lined outer basal tissue is that of the corti- 

 cal husk. 



[These sketches of seed sections are diagrammatic in the best sense, since 

 they are in reality the artist's excellent drawings just as he saw them in 

 these particular seeds, but to be seen over and over in seeds of other strobili 

 cut from trunk 131. This trunk was referred by Professor Ward to the 

 species C. Wielandi as best represented by Yale Cycads Nos. 77 and 393, 

 but cannot be of that species. Not only are there leaf base differences but 

 in the seeds of 77 and 393 the "blow off layer " is virtually eliminated.] 



Am. Jour. Sci.— Fourth Series, Vol. XXXII, No. 188.— August, 1911. 



