308 
have originated by a division of a single one. The sepa- 
ration between the endotesta en sclerotesta, the first indi- 
cations of which are observed amongst the Taxads, is 
accomplished in Ephedra and Gnetum, but has not taken 
place in Welwitschia. Though at first sight this seems a 
very simple explanation, it is still rather risky, as a regular 
splitting of one homogenous organ into two is a very 
uncommon feature. 
À tangential dédoublement, which happens with the 
stamina of some of the Rosaceae, also is observed some- 
times in the petals of Crucifers (Velenovsk\, 165). Though 
it is possible that here we have the same phenomenon, 
the question about the origin of the second integument is 
not to be solved entirely, and for the present we have to 
wait till perhaps more data are obtained from the investi- 
gations in palaeobotany. 
When the sporophylls enclose the ovule entirely and 
take over the function which has been performed by the 
sarcotesta and sclerotesta, the differentiation in the inte- 
gument may be left out. This case we find in the An- 
giosperms, where both integuments are thin and indifferen- 
ntiated, functioning as a endotesta. Owing to this an 
investigation upon a multiple origin in the ovules of the 
Angiosperms is encluded, because we have no starting- 
point here through the absence of layers as a sclerotesta 
on which ribs and sutures are most distincly to be seen. 
Also slips at the micropyle are wanting. It seems to me, 
however, that the integument of the Angiosperms is 
homologous yet with that of the Gymnosperms, though 
we have no true demonstration. The probably connec- 
ting links as the inner integument of ÆEphedra and 
Gnetum and that of Welwitschia which have exhibited 
an angiospermous type give this opinion a base of surety. 
On the other hand the ovule of Myrica Gale shows a 
feature resembling that of the pteridospermous seeds, the 
