276 Evans : Hepaticae of Puerto Rico 



two modifications, both of which tend to hicrcase the size of the 

 water-sac. In the first the keel is so strongly arched that it is 

 approximately semicircular in outline (plate 19, figure 2). At 

 the same time the free portion of the lobule, which is involute 

 throughout its entire length and appressed by its surface to the 

 lobe, shows a distinct fold of involution almost as strongly arched 

 as the keel itself The result is that the water-sac becomes almost 

 spherical in shape and exhibits a very minute opening. The free 

 margin of such a lobule is rounded and the apical tooth shorter 

 than usual and nearly straight (figure 9) ; the hyaline papilla 

 maintains its usual position and appearance. With this first mod- 

 ification of the lobule, the lobe is sometimes as large as in ordinary 

 leaves, but is often considerably reduced in size. 



In the second modification of the lobule the pecuharities of the 

 first are exaggerated, and the water-sac is several times larger 

 than in normal leaves. Enlarged water-sacs of this type are de- 

 scribed as " utriculi '' and occur singly or in a pair at the base of 

 a branch (plate 19, figures i^ 4, 12, 14 ; plate 20, figure 4). 

 The keel is here so strongly arched that it forms rather more than a 

 semicircle and is quite a little longer than the free margin. The 

 free portion of the lobule is in the form of a small lamina, which 

 varies in shape from rectangular to trapezoidal and is usually cut 

 off square at the end. This lamina is revolute and concealed 

 within the water-sac. Sometimes it is slightly concave and helps 

 to form an oval opening into the sac ; sometimes it is strongly 

 concave and the opening becomes more nearly circular. The 

 apex and the hyaline papilla are essentially the same as in the sec- 

 ond modification. The free portion of the lobe is very greatly 

 reduced in size. 



Certain species constantly develop the normal type of lobule 

 or produce the spherical modification only in exceptional cases; 

 others show a preference for the spherical form but sometimes re- 

 vert to the normal ; still others exhibit both normal and spherical 

 lobules in almost equal number. In certain species the utriculi 

 are apparently never developed; in others they occur with 

 greater or less frequency. In all probability the development of 

 both spherical lobules and utriculi is dependent upon external 

 conditions, but at the present time no definite statements to this 



