Evans : Hepaticae of Puerto Rico 277 



effect can be made. Of course, under some circumstances the 

 lobules fail to develop or present a very rudimentary appearance. 



The underleaves are attached by an almost straight line of 

 insertion and often show a rudimentary disc In the basal region,^ 

 from which the rhizoids develop. They vary In shape from orbic- 

 ular to broadly reniform, and the base varies from cuneate to 

 deeply cordate. In the majority of the species the underlcaves 

 are bifid ; in a very few species they are undivided. The margin 

 is commonly entire. 



The inflorescence in the genus seems to be almost equally di- 

 vided between autoicous and dioicous. The archegonial branch 

 varies greatly in length in most of the species, but is invariably 

 subtended by one or two innovations, which are themselves often 

 floriferous. The bracts are usually connate with the bracteole 

 and sometimes with each other (plate 20, figure 17). They are 

 rarely much larger than the leaves and are often perceptibly 

 smaller. 



The perianth is perhaps the most striking feature of the genus. 

 It is more or less obovoid in shape and abruptly narrowed into a 

 short but distinct beak. In the lower part It is nearly terete ; in 

 the upper part the antical surface is plane or nearly so, but there 

 are two distinct lateral keels and two postical keels. Each of 

 these four keels grows out beyond the beak in the form of a hol- 

 low horn, which sometimes remains short and sometimes becomes 

 nearly as long as the remainder of the perianth (plate 19, figures 

 I, 12 ; PLATE 20, FIGURES I, 2, 2i). Such a hoHi Is formed by the 

 rapid growth and division of the cells in the upper part of the keel 

 and shows at maturity a group of small and thin-walled cells at its 

 apex, very different from the thick-walled cells found elsewhere In 

 the perianth. On account of the delicacy of these cells, the horns 

 often become irregularly torn at the apex but it is doubtful if they 

 are ever normally toothed or spinose. In rare instances a fifth 

 horn grows out from the base of one of the other horns and is sup- 

 plementary to it. Of course, a five-horned perianth of this type is 

 very different from that found in such a genus as Leptolejeunca^ 

 where the fifth horn is always antical in position. 



Specialized organs of vegetative reproduction have not yet been 

 observed in Ceratolejeunca, It is not unusual, however, for a 



