304 
been possible that hardly anywhere the great number of 
triangular seeds has been mentioned. 
Sometimes as is the case in Callitris the layers in the 
integument have not been differentiated, probably in con- 
nection with adaptations to the dispersion of the seed. Here 
the testa is dry and membranous, whereas the nucellar 
tissue is soft and well developed, in Thuya, however, 
having a differentiated testa, the nucellus on the other 
hand becomes dry and papery. Thus it is obvious that the 
differentiation of the integument has a clearly physiological 
meaning, it is the leading tendency amonst all seeds. The 
dense haircovering in Physostoma forms the first indication 
of a better enclosure and protection of the germinating 
macrospore. The cupules of Lagenostoma and Sphaero- 
stoma accomplish the function elsewhere performed by 
the differentiation of a sarcotesta, which is also the case with 
the cupules of Torreya and Gnetum. 
The arillus of T'axus is homologous to all appearances 
with the cupule of T'orreya. The first is less developed 
in comparison to the latter and in connection herewith 
the testa in Taxus is still differentiated in the three 
layers, wbhilst in Torreya the cupule has performed the 
function of the sarcotesta, an outer fleshy layer not being 
formed by the integument. In both cases the megasporophyil 
is not to be observed, whereas in Cephalotaxus, having 
no cupule, it is very distinct. Therefore the suggestion 
may be ventured that both enveloping organs have been 
formed by two decussate sporophylls, each bearing origi- 
nally one ovule. One of the ovules has been reduced and 
suppressed, so that the other has become terminal, invested 
by the two fusing sporophylls. 
The two opposite vascular bundles in the cupule of 
Torreya (having a normal orientation in contrast with 
those in the sarcotesta of Cephalotaxus) and the case I 
found in T'axus, in which the aril was divided into two 
