1917] 
GATES—THE GENUS TRILLIUM 83 
of some of the germinal materials has taken place, but the 
nature of that redistribution is at present unknown. 
Such wide variations in Trillium as the formation of long 
petioles from sessile-leaved species, and the multiplication 
of the number of the leaf whorls, with internodes between 
them, are, however, not obviously teratological; and in the 
former case they are similar to ordinary specific differences 
in the genus, while in the latter they, if constant, might well 
serve as the basis of a distinct genus. Thus the long petioles 
and short stems of T. petiolatum furnish its most striking 
distinction from such species as Т. sessile, and it is very 
tempting to assume that T. petiolatum was derived from a 
sessile-leaved, long-stemmed species in the same way that the 
typical condition of the variety variegatum now apparently 
arises from Т, grandiflorum. These suggestions may seem 
to systematists bold, but we have reached a point where our 
experimental knowledge of variation must be applied directly 
in any discussion of the phylogeny and relationships of par- 
ticular species. The known variations of species of Trillium 
furnish a more reasonable basis for an evolutionary recon- 
struction than hypothetical continuous variations which ex- 
periment seems to show are not usually inherited. 
PARIS L. 
The Eurasian genus Paris is mentioned here on account 
of its close affinities to Trillium. Just as Trillium is chiefly 
North American, with a few species closely related to 7. 
erectum in northeastern Asia and one species (T. Govaniana 
Wall.) in the Himalayas; so Paris is almost entirely Asiatic, 
with one species (P. quadrifolia L.) extending into Europe. 
The genus Paris was probably derived from the ancestors of 
the group of four species, relatives of Trillium erectum, occur- 
ring in Japan, Manchuria, and eastern Siberia. Some 30 
species have been described, mostly from China, but includ- 
ing 3 from Siberia, 3 from Japan, and 2 from Thibet. 
Bearing in mind the probable origin of the genus, its dif- 
ferences from T'rillium are of much interest. The two may 
be compared as follows: 
