22 CALIFORNIA ACADEMY OF SCIENCES. [Proc. 3D Ser. 



this was the only instance of the kind found, it is not pos- 

 sible to say what the origin of this tissue was ; but it prob- 

 ably may be considered as mainly due to the development of 

 the endosperm. So far as I know, such a development of 

 the prothallial tissue in Angiosperms, except as a result of 

 fertilization, has not been recorded. 



The general development of the carpel has been treated 

 fully by Magnus, but he did not study some of the minute 

 details herewith given. In Naias jlexilis the carpel rapidly 

 grows up beyond the enclosed ovule, and just above this 

 point there is a marked thickening of the walls (fig. 37), 

 so that the pistil is closed above except for a narrow canal 

 in the middle, which communicates with the cavity of the 

 ovary. Very early, too, the margin of the carpel shows 

 unequal growth, resulting in the formation of four lobes 

 (fig. 40), of which two are decidedly longer than the others 

 (indeed, in the case figured, one of the smaller lobes is 

 scarcely visible). The two longer lobes finally have the 

 terminal cell transformed into a thick-walled brown spine 

 (fig. 46). The two shorter lobes between the spiny ones 

 are the stigmatic lobes; these have the terminal cells more 

 or less papillate. In N. major there are three papillate 

 lobes, and in N. minor and most other species, two; but 

 none of these species has the spiny lobes (Magnus, 1870, 

 p. 20, PL II). Two Brazilian species, according to Magnus 

 (1. c, p. 21) have the same structure as N. jlexilis. There 

 are also several Mexican species which sometimes show a 

 similar structure, but they are extremely variable, while in 

 N. jlexilis this structure is usually very constant. 



Finally, there must be noted the occurrence of a second 

 envelope in several species (Magnus, 1870, p. 22, PI. Ill, 

 figs. 1-3) , among which may be mentioned N. ancistrocarpa. 

 Here the outer envelope bears several spiny teeth, while the 

 inner one (carpel) bears two stigmatic lobes. This outer 

 envelope is perhaps to be considered as a floral envelope 

 and the possible homologue of that surrounding the stamen. 

 The case of N. jlexilis suggests that the two envelopes are 

 of the same nature and that this species is intermediate 



