502 ‘Evans: HEpATICAE OF PuERTO Rico 
and intermediate thickenings. These are a trifle larger in L. ellip- 
tica than in L. exocellata, but the difference is not at all striking. 
The species of Leptolejeunea when dried do not always assume 
their natural appearance upon the addition of water. Even boil- 
ing water is not always effective, and it is sometimes necessary to 
treat them with potash solution or some similar reagent before we 
can gain an accurate idea of their cells. This is probably one 
reason why the cell-structure is not always accurately described. 
The absence of ocelli in the leaves of L. exocellata is empha- 
sized by Spruce, and it is from this supposed character that the 
species derives its name. There is present,‘however, at the base 
of the lobe a large hyaline cell about three times as long as the 
neighboring cells. This cell, which would naturally be interpreted 
as an ocellus, occurs not only in Puerto Rico specimens, but also 
in South American specimens distributed by Spruce himself. The 
ocellus is more or less overlapped by the adjacent part of the lobe 
on the side toward the keel. This part is distinctly convex, and aes 
can see the ocellus clearly only by focussing through the cells which 
lie over it (p/., f. g). Sometimes the overlapping goes so far that 
the ocellus looks as if it were a foreign body inside the water-sa©s 
and possibly it has been so interpreted. In reality it takes direct 
part in the formation of the water-sac. It is not only longer than 
the neighboring cells, but greater in diameter and projects con- 
siderably beyond them toward the substratum. Here it comes 
into contact with the free edge of the lobule, which remains plane 
or nearly so (f. 5), and in this way a cylindrical sac is form 
open only at one end. The capacity of the sac is of ee 
increased by the convexity of the lobe between the ocellus - 
the keel, A similar ocellus occurs in Z. edliptica, but it is usually 
less concealed by overlapping cells. Except for this basal ocells 
the leaf of L. exocellata is truly “ exocellate,’’ and lacks the inter 
rupted row of ocelli, which is characteristic of L. e//iptica. ‘, 
The differences in the subfloral underleaves, as noted above, t 
not seem to be constant. It is sometimes difficult to make OF 
these structures at all on account of the numerous rhizoids ba ; 
they bear, but when well developed they are essentially alike 
the two species. . — ae 
‘ to 
The remaining differences mentioned by Spruce seem 
