Morphological Observations. 301 



heing formed by the endocarp. It is this structure which probably 

 has as a rule l:)een looked upon as the whole capsule, but. as we 

 have seen, this is erroneous, and as at this stage the placenta is free 

 above, we may also look to this stage as the source of the incorrect 

 descri{jtion of the placentation. 



The seeds again require our attention, as they have also been 

 hitherto described incorrectly. Sender (I.e. p. 382) describes them 

 as winged. Bentham and Hooker (I.e. I., p. 157) as "angular or 

 laterally compressed, 3-winged or nude." Pax (I.e. p. 57) says: — 

 " Seeds angular or compressed. 3-winged or without wings." but 

 though the seeds differ in slight details in the different species, they 

 are again surprisingly uniform generally speaking. They all arise 

 from cami)ylotr()t)OUs ovules, in which the chalazal portion is some- 

 what longer than the micro})ylar region. The result is that both 

 •ovules and the seeds are obliquely club-shaped. They are covered 

 with very minute papillae, which in some species are a trifle largei 

 than in others, and it is just possible that these may have given rise 

 to the view that the seeds are winged. With reference to the 

 ■endosperm in Portulacaceae, Pax mentions that it is scanty (I.e. p. 

 51). I have sectioned only the seeds of one species, A. filameniosa. 

 and find that it is devoid of any endosperm. If this is the case in 

 other species also, the genus Anacampscros should be placed in 

 a rather isolated position among.st Portulacaceae, to which some of 

 its other characters mav also entitle it. 



