118 BOTANICAL GAZETTE (FEBRUARY 
a characteristic which was not evident in the field, where owing 
to individual variation the stature of a plant is not a pronounced 
characteristic. Yet when the descendants of individual parent 
plants were brought together in groups under uniform conditions, 
the contrast in stature between the different groups was constant 
and very marked. 
The more important morphological characteristics in which the 
various types differed from each other were the shape of the leaves 
and flowers, and the form of the inflorescence, but the descendants 
of a single parent were entirely uniform with respect to these char- 
acteristics. Even such minor characteristics as the tint of the 
leaves and color or shade of the flower were uniform throughout 
the plants of each group. In some cases the same type had been 
selected more than once, so that some of the groups were identical. 
The plants in the field were studied during their entire growing 
season, and each individual was often examined, but among the 
several thousand plants no aberrant form occurred. 
The uniformity of the descendants of each of the parent plants 
indicates that the plants originally selected represented elementary 
species, for if the parent plants had been hybrids, splitting accord- 
ing to Mendel’s law, the splitting should have occurred in the 
generation of 1909. The absence of hybrids among the plants 
selected can be explained by the ease with which self-pollination 
takes place, and by the scarcity of pollinating insects. The 
flowers are slightly proterogynous, but even before the flowers are 
fully open the anthers begin to shed their pollen. As both pistils 
and stamens are about equal in length, self-pollination is easily 
accomplished. Pollinating insects seem to be rare. During the 
two seasons in which I spent much time in the tobacco fields, I 
observed only in a few instances honey bees and hawkmoths 
pollinating flowers. It seems very likely, therefore, that in the 
majority of cases the tobacco flowers in Cuba are self-pollinated. 
In order to continue the pure line cultures, a large number 0! 
plants of each type were bagged for seed in the usual way, but on 
account of my removal from Cuba in the spring of 1909, seeds from 
all the types were not obtained. Some clusters with mature seeds 
were found on 11 of the types. The seeds from 1o—15 plants of 
