136 The Botanical Gazette. 
and fig. 9 is a good type of what I have found in 
stages of development in nearly all of them. This is aw 
commonly found in E. Americanum. The following not 
from BOTANICAL GAZETTE, II (June, 1877) may throw further 
light on this species. . 
‘*The normal E. albidum from Illinois to western Ic a 
noted for spotted broad leaves and very recurved petals, is (oF 
fined to groves; while a much smaller form from westt® 
Iowa to Kansas confined to prairies has narrow leaves, nev 
spotted, petals never recurved, not half the height of the 
former. The difference is doubtless due to soil and lo 
ity.— R. Burgess, Ames, Iowa.” 
The above seems to indicate that E. mesochoreum insteal 
of being a true species is but a variety marked by ceri’ 
variations constant in it, and found, but not constant, < 
albidum. Fi 
I took careful measurements of every part of the flows 
E. Americanum and E. albidum, and plotted the cue 
cording to the method used by Alfred Russell Wallace 8% 
mal variations. Comparing the results of thirty-four P” 
mens of each, we find the length and width of the ee 
length of stamens and ovary, and length and spread ots “* 
are closely correlated in E. Americanum. In E. albidum pS 
variations of the length of the petals and ovary af enti | 
oosely correlated. The other parts vary indepen 
sometimes seeming to follow the variations of thes © 
Oy 
for, varying not at all or even decidedly in the oppes in 
rection. The general variations of the two plants a 
opposite direction, toward each other. This is show? 7 , 
paring the mean measurements of the different parts 
variations from that, in the following manner: 
sama 
Number 
: bove 
Length of Spectes, Mean, Pom mean 
ED a oo ee 
Petal Americanum 32.65™™ 19 
Style Americanum 10.25™™ re 
Style albidam 11,.1=™ z 
Americanum § 
ary ge 
Ovary albidum 8.25™™ ” 
————————— 
