276 
level of separation of endotesta and sclerotesta no trace 
of it could be seen and in the upperpart the cells com- 
posing the epidermis were coloured very faint in their 
outer walls. Therefore it is not evident that both layers 
are fused separate ones, but it is more reasonable that the 
innermost layer of one integument has become free inits 
upperpart to cover the nucellus. 
This feature is also established as we will see later on 
by the other Taxads viz. Torreya and Taxus, and there- 
fore it is an other caracteristic of this group contrary to 
the other Gymnosperms. 
In the ripe seeds two ribs are found on the sclerotesta 
and accordingly the seed splits open when pressed. Within 
the stony coat the nucellus is lying entirely free, enveloped 
by a papery membrane, representing the inner fleshy layer, 
which has become entirely separated from the testa. 
Some authors amongst whom Celakovsk\ and 
Worsdell supposed the outer fleshy sarcotesta to be 
homologuous with the arillus of Taxus. Whether it is 
allowed to make this comparison, Î will discuss in the 
following chapter. 
On both sides of the seed, just above the ribs on the 
stony coat two vascular bundles are running in the 
sarcotesta, their structure being investigated by Stras- 
burger (147) and Worsdell (176) The former has 
found that the orientation of the vascular bundles is 
inverted, the xylem lying outside, whereas the phloem is 
situated in close proximity to the sclerotesta. The same 
inverted orientation is present in the ligule of the 
Selaginellaceae and the Jsoetaceae, and in comparison 
with this, he supposes the outer fleshy layer to be a 
ligular structure. 
Worsdell has amplified this investigation, and has 
indicated, that there is a well-developed centripetal xyvlem 
extending along the whole tangential face of the centri- 
