L rCOPODIA CE.E. ^I 3 



from the ventral side of the creeping stem without any fixed rule ; they fork 

 when they have attained a length of 3 or 4 centimetres, but probably not until 

 they come in contact with the ground. Their plane of dichotomy stands (as in Se/a~ 

 ginella la'igata and cuspidata) at right angles to the longitudinal axis of the stem 

 (in Isoetes it is, on the contrary, parallel to it) ; the succeeding branches are either 

 also dichotomous, or sympodial ; in the latter case the real or apparent lateral 

 branches appear distributed either in decussating pairs or singly with a divergence of 

 \ or \. The position of these lateral roots has not been correlated with the arrange- 

 ment of the fibro-vascular bundles of the mother-root. This, together with the 

 circumstance that the young ramifications are densely crow^ded at the end of the 

 mother-root, appears to exclude the supposition that the branching is monopodial ; 

 and in any case the process is more like that which occurs in Selaginella and 

 Isoetes. Above ground these roots are of a bright green colour. It is very difficult 

 to prove the existence in them of an apical cell ; yet Nageli and Leitgeb conclude 

 that one is present, having the form of a four-sided pyramid, the segmentation of 

 which would influence the peculiar position of the root-branches. 



The Sporangia exhibit considerable diff'erences in the difi"erent genera of the 

 class, both in their position on the fertile branch, and in their development and 

 mature form. But they agree in a single sporangium being always formed in the axil 

 of a leaf; and they are distinguished by their size from those of all other Cryptogams. 



The sporangia of Isoetes arc sessile in the fovea of the leaf-sheath, to which 

 they are attached by their dorsal line (Fig. 306 A). They are unquestionably 

 products of the leaves ; the outer leaves of the fertile rosettes produce only macro- 

 sporangia, the inner ones only microsporangia, the former containing a large number 

 of macrospores. Both kinds of sporangia are imperfectly segmented by threads of 

 tissue {Trabeailcc) which cross from the ventral to the dorsal side. The sporangia 

 do not dehisce, but the spores escape by the decay of the wall. 



In Selaginella the sporangia are shortly stalked roundish capsules, their origin 

 being still doubtful, whether from the base of the leaf or from the stem itself, per- 

 haps variable in the different species. The macrosporangia contain usually four, 

 less often two or eight macrospores. In the division of Articulatse the lowermost 

 sporangium only of a spike produces macrospores ; in the other divisions there 

 are several macrosporangia. 



The remaining genera have, as has been mentioned, only one kind of sporan- 

 gium, the contents of which bear a greater resemblance, in Lycopodium to the 

 microspores of Selaginella, in Psilotum to those of Isoetes. The sporangia originate, 

 in Lycopodium, from the leaf itself ; they consist, as in Selaginella, of only one 

 compartment, and split into two valves at the apex or on the anterior surface. In 

 Psilotum. the sporangia are described as trilocular, and as placed in the axil of a 

 bipartite leaf; Juranyi's recent researches seem, however, to show that three spo- 

 rangia are here seated round the end of a short branch, which at the same time 

 forms below them two leaflets on the outer side. In Tmesipteris the elongated 

 sporangium is seated on a stalk (shoot .?) bearing two leaves right and left of it. 



The history of development of the sporangia is still incomplete in many points. 

 It is important to observe at the outset that Hofmeister considers that the sporan- 

 gium of Lycopodiacece arises from a single cell of the leaf or stem, and that he 



