14 SIPHONACE^. 



distichous, and on others tristichous or quadrifarious. A further step brings us to 

 C. Selago, C. Lycopodium, and their allies, in which the branches are thickly- 

 set with imbricating ramenta ; and the highest development of this type is reached in 

 C. obscvra, C. Mnelleri, and C. hi/pnoides, where pinnate and imbricated characters are 

 combined. Another group of species, like our C. paspaloides, is characterised by having 

 pinnate or multifid ramenta ; and in another, the ramenta are baglike, either round, 

 pyriform, or topshaped. Of this type we have an American example in C. clavigera, 

 one of the most widely dispersed and most variable of the species. By depressing the 

 apex of a baglike ramentum it becomes top-shaped, and by further depression peltate, 

 and this form distinguishes C. chemnitzia and C. peltata ; and again, peltate ramenta 

 becomes perfoliate in C. nummularia and C. stellata by the development of young 

 ramenta from the centre of the discs. In such species as C. ericifoUa and C. cupressoides 

 the gradual evolution of ramenta from mere prominent points of the frond is illustrated; 

 and such species lead us to C. Freycinetii where the ramenta remain in this rudimentary 

 condition. And thus we are conducted, by almost insensible gradations, through a 

 considerable number of forms, back to those from which we started, and which had 

 naked fronds destitute of ramenta or marginal incisions. And so, after a survey of all 

 the species, we become more reconciled to the generic group as limited by Lamouroux, 

 than if we had merely compared together such extreme forms as C piroUfera ai\(\. C. pas- 

 paloides. 



We have ali'eady said that the structure of all these plants is essentially the same. 

 It remains to describe more particularly what that structure is. I am not aware that 

 any observer has yet noted the early development of the frond, nor is the mode of re- 

 production as yet clearly made out. The spores are presumed to be similar to those of 

 Bryopsis, and to be formed in any portion of the grumous matter that fills the frond, 

 and most probably from that of the ramenta. When we take a fully formed frond 

 distinguishable into creeping stem, roots, upright branches and ramenta, we find that 

 it is every Avliere coated or encased in a homogeneous, hyaline, tough membrane des- 

 titute of further structure than this; that it may be seen in the thicker parts to be 

 composed of several layers of cellulose, equally dejiosited one within another, as in the 

 wood-cells of higher phmts. There is no septum throughout the plant, and no appear- 

 ance of cellular structure in the membrane of the walls. The frond, with all its 

 ramifications, is strictly ^^ continuous" forming a closed sac ; and so far as we know 

 it is formed by the evolution of a single cell, extending itself indefinitely without cell- 

 division, and showing in excess the same structure as we find in a minor degree in 

 such plants as Botrydium, Bryopsis and the like. This closed sac, frond or cell, in 

 Caiderpa, is filled as in Bryopsis, with a semi-fluid, semi-gelatinous, bright-green 

 endochrome containing starch-grains mixed with what seem to be oily particles, and 

 obviously highly organized, but its chemical comj)osition remains to be examined. 

 Most probably it is highly nitrogenous, for it bears considerable resemblance in sub- 

 stance to the glairy semi-fluid of many sponges ; and hence i)robably the reason of 

 Lamouroux's supposition that these plants were of a semi-animal nature. If the 

 structure of Caiderpa were mei-ely wliat we have described, a closed membrane filled with 

 grumous matter, it would not essentially differ from that of Codium and Bryopsis. 



