208 



BOTANICAL GAZETTE 



[SEPTEMBER 



not as vigorous as could have been desired. The spHtting is prompt 

 and the divisions follow so quickly that the roof and cells of the 



pore margin are soon completed (fig. ii). 

 In Conocephalus also the formation 

 of the air chamber is rapid (figs. 12, ij) 

 and the details are not as regular and 

 therefore not as easy to follow as the 

 Fig. 10.— Mar chant ia poly- Marchantia. After the initial cleft has 



!!'{!?' .Ir!!l'^ extended to the surface {fig. 12, ^) the 



expansion of the tissues at first outstrips 

 the divisions which are to produce the roof, so that the chamber is 

 open (fig, 12, 3^^, fig, ij^ 3,4)^ Later the roof closes the chamber, 



perhaps intermittently (fig. 12, ^, fig. ij, *' ^), which only becomes 



margin, with complete closure. 



fn- 



\ 



i 



Fig. II. — Lunularia vulgaris: a, 



Fig. 12. — Conocephalus conicus: a, apical 



apical cell; b, c, successive segments; cell; i-6, air chambers. 

 /, 2, J, air chambers. 



JJ 



6,7 



). 



ih 



(figs 



more or less completely, leaving only traces of their rhomboidal out- 



lines on the surface.^ 



very 



I 



" Campbell found no e\^dence of air chambers in D. irichocephala, which we have 

 not seen. It would be very remarkable if the early stages were wholly wanting. 



