310 EVANS. 



the hairs were fairly numerous. When the wings are distinctly invo- 

 lute and the hairs abundant, a delicate weft is sometimes to be 

 observed between the contiguous margins, much as in M. dccrescens, 

 but the hairs are just as likely to extend irregularly in all directions. 

 Even when marginal hairs are present the thallus is often naked else- 

 where; in other cases ventral hairs can be demonstrated on the costa 

 and, still more rarely, on the wings. The hairs vary greatly in length, 

 the majority measuring perhaps 0.08-0.12 n in length; the diameter is 

 mostly 10-12 /j.. The costa, in ordinary well-developed thalli, is 

 bounded both dorsally and ventrally by two rows of cortical cells. 



In two specimens a few male branches were observed. They were 

 smooth and almost spherical, the costa being so strongly incurved that 

 the apex almost touched the base. Some of the branches were about 

 0.25 mm. in diameter but a few were somewhat larger, the largest one 

 seen measuring 0.4-0.35 mm. The single female branch demonstrated 

 was so disintegrated that its true features could not be determined. 

 It bore a young calyptra with crowded hairs in the upper part and 

 scattered hairs below the middle. 



In most of the material gemmae are present in large numbers, and 

 the gemmiparous branches (Fig. 9, A, B) show interesting modifica- 

 tions, comparable with those described under M. epiphylla but reach- 

 ing a more advanced type of specialization and approaching in this 

 respect the highly specialized gemmiparous branches of M.fruticulosa. 

 Even when the vegetative branches are strongly convex and prostrate 

 the gemmiparous branches are plane or nearly so and curve away from 

 the substratum. At the same time the wings become narrower and 

 narrower until, in extreme cases, they become reduced to a width of 

 only two or three cells. No cases have been noted, however, in which 

 the wings had entirely disappeared, and the growth of the gemmip- 

 arous branch often comes to an end while the wings are still four 

 cells broad or more. With the reduction in the width of the wings, 

 the cortical cells sometimes continue to show the usual arrangement 

 in four rows, but the rows sometimes become increased to as many as 

 six, both dorsally and ventrally, the rows under these circumstances 

 being irregular and the cells themselves considerably reduced in size. 

 Sometimes, especially when the formation of gemmae begins in a 

 juvenile thallus, the gemmiparous branches may lose their costae and 

 become reduced to narrow unistratose thalli only five or six cells 

 broad; or, if the vegetative thallus itself lacks a costa, the gem- 

 miparous branches may retain the same simple structure throughout 

 their entire length. In other cases the gemmiparous branches may 



