292 
tot een kegelvormig mutsje verdikt is. Het integument 
wordt bij het rijpen opgeheven en omgeeft alleen maar 
het bovenste derde deel van het zaad” (p. 14) !). 
This is not correct, the membrane surrounding the 
endosperm, splits as to say in two, at one third from the 
micropyle, it being out of question that one of the two 
membranes should be freely attached to the other, like for 
example the calyptra on a sporogone of the mosses. Here 
also the nucellus is concrescent with the integument to 
»shoulderheight” as is the case in nearly all Gymnos- 
permous seeds. The membrane surrounding the endosperm 
is formed by a narrow fusion of the nucellar wall and 
the integument, wheras both remain free from one another 
at their apex, even in the ripe condition of the seed. 
Lignier et Tison (89) mention the occasional occur- 
rence of bracts at the base of the outer envelopment. 
,»Exceptionnellement, dans cinq fleurs qui toutes se trou- 
vaient près de la base des cônes, nous avous observé, 
au-dessous de l'enveloppe ailée, deux petites bractées 
situées dans le plan tangentiel” (p. 129). 
The occurrence of these bracts in the lowermost part 
of the strobilus, where generally, as Ï have pointed out 
in Ginkgo, the possibility is greatest to observe a more 
primitive stage, may perhaps indicate that the absence of 
these bracts in other cases is a secondary feature. In this 
way they strengthen the opinion, to regard the strobilus 
as a compound one, for these bracts must necessarily be 
borne on an axis of the second order. 
To form a clear idea about the structure of the female 
strobilus, and the value of the different segments, we now 
have to pay our attention to the male strobilus. 
This consists of an axis with four rows of decussate 
bracts like the ovulate one does, but it is smaller, the 
l) J cursivied. 
