1884. ] The Theory of Sex and Sexual Genesis. 785 
curred to me that since in this plant the receptacle of the umbel- 
let is somewhat projecting in the center, and since the apex of 
the axis in growing plants is generally the point of most rapid 
growth, it might be the case that this central point of the recep- 
tacle would in some instances bear fertile flowers when parts fur- 
ther out and intermediate in position would not. And on more 
careful examination I found in many instances a single pistillate 
flower occupying this position, but showing this remarkable pecu- 
liarity, that it was sessile upon the apex of the receptacle; and 
from the considerable number of flowers in the umbellet, twenty 
to thirty, and the fact that cll the others were carried out beyond 
this one, by their extended pedicels, so as to partially conceal it, 
Thad previously overlooked it in making out the distribution of 
male, perfect and female flowers on the umbellet. 
As has just been shown, there is a strict correlation between 
the size of the umbellets of a compound umbel and the sex of the 
flowers they bear, the smallest umbellets bearing male and the 
largest predominantly female flowers. In Indian turnip (Arisema 
triphyllum) there is a similar correlation between size and sex for 
the whole plant, the largest plants being females, the smallest 
(that produce flowers) males, and many of the intermediate ones 
monæcious. So striking is this the case that I found that when 
€ plants are in bloom the sex can be inferred, with a great 
degree of certainty, from the comparative size of the plants alone. 
At this time the spathe so incloses the spadix on which the flowers 
are borne, that it cannot be determined by direct observation 
whether a plant bears male or female flowers, or both, except by 
opening the spathe. The plants reproduce by buds from corms, 
and by bulbs dropped from the receptacle, and are thus found 
neat frequently in groups, the plants in the same group varying 
steatly in size. The large buds from the older corms produce 
est plants, and the bulbs of the previous year the small- 
oo latter, however, do not produce flower-stalks the first 
be Prove the correlation between sex and size, I made exami- 
“og at one time of 213 plants, and at another of 137 p 
‘tested were in bloom. An inference was first made as tot ~ 
with le sex of each plant, judging from its size as compar 
A the others of the same group, or as compared with the 
© TAE size for the locality, each plant being pronounced either 
