147 
As the sporocarp ripens the outer cells become very hard, es- 
pecially the first layer of hypodermal cells (Fig. 12, b.), whose 
walls become finally so much thickened that the cell cavities are 
almost completely obliterated. The second hypodermal layer 
also has its walls more or less thickened, but not nearly to the 
Same extent. 
The ripe sporocarp is about 3 mm. in diameter, and the pe- 
duncle about as long and bent downward, so that the sporocarp is 
partially or completely buried in the earth. When perfectly ripe 
it splits into four valves corresponding to the lobes or ‘leaflets of 
which it is made up. This splitting follows the median line of 
the partitions in the sporocarp. 
A comparison of the foregoing statements with the corres- 
ponding points in the development of Jarsilia, so far as the latter 
is known, show, as might be expected, close resemblances. There 
seems no doubt that the sporocarp is simply a portion of a leaf, 
bearing much the same relation to the sterile part that the fertile 
pinne of Ophioglossum or Osmunda for example, do the sterile 
part of the leaf. We may perhaps more aptly compare it to such 
a fern as Onoclea, which is really more nearly related. The struc- 
ture of one of the spore-bearing leaf segments of O. senstbilts for 
instance, is very similar, indeed, to the sporocarp of Pilularia, ex- 
cept that the sporangia are borne upon the lower and not upon 
the upper side of the leaf. As the Marsiliacez are in all probabil- 
ity descendants of forms related to the Polypodiacez, the origin 
of these peculiar points is probably to be looked for in forms hav- 
ing fertile leaves of a type similar to Onoclca. 
On comparing Pilularia Americana with P. globulifera, of Eu- 
rope, less difference was found than is generally supposed.* Ex- 
cept the longer peduncle of the fruit of the former and a slightly 
diminished development of the wall of the macrospore, I could see 
no difference. In size my specimens were little, if any, inferior to 
specimens of P. globulifera studied by me in Europe, either as re- 
gards the leaves or the sporocarp; and, almost without exception, 
the sporocarp was four-chambered as in that species, instead of 
three-chambered as described in the text-books. The absence of | 
* Goebel, L. et 240. ndeewend, «Qur Native Ferns and Their Allies,” 3d. a a 
ed., pp. 126-127, Watson, “ Botany of California,” Vol. i ii, Pp» 352: 
