ON THE NATURAL SYSTEM. 



7 



centre of the ring, by a single fissure extending from the apex to the base, and they agree also in converging as near as possible at their basis so that in Marattia eight or ten spore-cases are 

 sometimes attached almost by a point. Dion edule has an abundance of paraphyses at the base of its anther cells which may represent the filamentous bodies occurring among spore-cases ; and 

 the interrupted character of the arrangement of the spore-cases on the fronds is fully explained by comparing it with the rings of anther cells in Cycadece. In the female Cycas fertile and sterile 

 fronds are produced in alternate succession, just as in some genera of Polypodiacece e. g. Blechnum; Stangeria paradoxa has their forked venation and their usual mode of branching by division of 

 the terminal bud instead of axillary budding, occurs in Coniferce, a whorl of branches being given off at once. The prothallium may be regarded as a primary embryo-sac bearing minute portions 

 of anther lobes (the antheridia ;)* — the male prothallium may be a barren embryo-sac as regards archegonia ? Mr. Salter's discovery of pollen in the ovules of Passiflora (Trans. Linn. Soc. Vol. XXIV, 

 p. 143,) leads to this conclusion. As thus understood the contents of the anthers (spore-cases) are metamorphosed into ovules (the prothallia) and the antheridia they bear are partial reversions.f 

 The male prothallia of Lycopodiacece and Marsileacece I regard as the same kind of development that takes place in the pollen of Coniferce and as not showing any analogy to those of the Ferns. 



I have placed Lycopodiacece in relation with Gnetacece because of the singular resemblance of the short truncate stem of Isoetes to that of Welwitschia. Hofmeister compares it with Beta 

 and Cyclamen both of which are included in Table III. J Lycopodiacece also approach closely in habit to Crassulacece and Plantaginece, e.g. Plantago pusilla. If Cycadece are derived from Marattiacece 

 becomes it most probable that Conifers are also derived from them through the former, rather than from Lycopodiacece, because it appears impossible that Coniferce can be derived from both 

 Marattiacece and Lycopodiacece. May not the 3-celled spore-case of Bernhardia be the prototype of the 3-celled anther of Welwitschia ? 



The affinities between the vascular Cryptogams and the hypogynous Exogens are sufficiently clear to be regarded as undoubted ; but the relations between Bryacece and their allies and 

 the epigynous Exogens are not so complete on account of the difference in their fructification ; the difficulty may however I believe be superseded, see p. 13. It has been repeatedly suggested that 

 there may be an affinity between Gnetacece and Loranthacece and if so the Epigynous Division, Table II, might be a retrograde branch of the Gnetacece, but this would require neither alteration 

 in the arrangement of the Tables nor changes in the position of any of the families. (V. Tables II. and III.) This however appears to me impracticable and I prefer adopting the Muscal 

 hypothesis supposing some connecting link may perhaps be wanting. As the theory of two generations or the alternation of generations does not offer any explanation of the nature of the 

 structures composing the very variable infloresence of the higher Cryptogams, the usual rules of analogy and homology are our only guides. 



If Lycopodiacece are derived from Bryacece of which they have so much the appearance, might not the near relation between them be explained on the supposition that the former are an 

 hermaphrodite form of the latter ? The capsules of the former would be the thecae of the latter containing both male and female spores (the calyptra being absent or represented by the stipule 

 of Lycopodiacece ?), the fertilization taking place after their dispersion as in the Algacece recently observed (Nat. Hist. Rev. — 1865, p. 74 et seq.) ; and the male spores in their development may 

 be compared to that of the pollen of Coniferce and probably also to the cell-division which takes place in the androspores of (Edogonium and Bulbochcete (1. c. p. 74). Assuming that Bryacece are, 

 through J linger manniacece, derived from Algacece, Lycopodiacece would be Bryacece returning to their original character as far as regards fertilization taking place after the dispersion of the spores. 



The term polyembrynous refers to the number of embryos or germinating bodies produced by a single archegonium ; in the vascular Cryptogams it produces but one, but in Bryacece and 

 their allies it produces many, and from the mode in which this is effected they may be regarded as one degree lower in their organization. 



Miscellanea. 



Some of the families now regarded by some botanists as distinct will doubtless be finally included in others, but I am obliged for the present to place them as distinct or as sections of 

 others for the purpose of showing the position which from my own analysis I infer they should occupy. 



* The long vitality and continued growth of the prothallium has nearly its parallel in the embryo-sac of Nuphar lutea, which I have seen thickened, completely enclosing the cotyledons, and still in vigourous growth, six weeks after the 

 germination of the embryo when it had produced two or three leaves. The albumen was dead, as it was quite softened like pulp. The cells of the cellular tissue of the embryo-sac were large, and had the glistening appearance of the surface 

 of a young leaf. How long these sacs would have continued living I do not know, as the plants were destroyed. 



f Consistently with this view it appears from very recent observations made by M. Duval- Jouve since the above was written (Nat. Hist. Rev. — 1864, p. 562,) that archegonia usually one or more occur on the male prothallia of Eqtiisetum, 

 but they are not fertile. 



J Currey's Translation of Hofmeister on the Higher Cryptogamia, p. 372. 



