102 MR. R, P. GREGORY ON THE 
Taking into consideration only the sex of the offspring, the above crosses 
give, in the cases where both sexes occur among the offspring, 67 $, 27 d, 
or 2:48 : 1. 
Combining the results of all the above crosses in which both sexes occur 
among the offspring, the numbers obtained were 151 9, 59 g (including 
both long- and short-styled), or 2:56 : 1. 
(3) Hermaphrodite plants were only obtained in those crosses in which one 
offthe parents was of that character. In one case the hermaphrodite plant 
was used as the female parent; one seedling only, an hermaphrodite, was 
raised. In the other case the hermaphrodite plant was the male parent ; 
the offspring were 15 9, 2 9 approaching the hermaphrodite form (see 
p. 93, and figs. 10, 11), 2 8,5 & L.,1 d doubtful (died without producing 
a sufficient number of flowers ; those examined were short). 
The most interesting fact shown by the above results is the large pre- 
ponderance of females which occurs in the offspring. In each type of cross 
where the two sexes appear among the offspring, this preponderance is in 
the proportion of, roughly, 2:5 : 1. This ratio may, however, be affected by 
the power which this species appears to possess of producing seed in the 
absence of pollination. In the course of these experiments 30 inflorescences, 
borne upon 20 plants, were covered to prevent accidental pollination, and 
were left without artificial pollination. The majority of these spikes 
produeed nothing, but from 5 of them 62 fruits were obtained, many of 
which indeed appeared empty, but others appeared good, and 11 seedlings 
have been raised from this harvest. 
EXPLANATION OF PLATE 8. 
The figures are from free-hand sketches drawn to scale from measurements of the flowers 
represented. Тһе scale is the same for all the figures except Nos. 5, 6, 19, < 20. In 
figs. 5 and 6 the scale is twice, and in figs. 19 and 20 it is one half, that employed for the 
other figures. For the sake of clearness no attempt is made, in the figures showing flowers 
which have been dissected, to represent the hairs which occur in the corolla-tube. 
Figs. 1 & 2. Female flowers in side view. 
Fig. 3. Young female flower in side view; the style is not yet elongated. Fig. За. The 
same flower scen after the corolla has been split up from the side. 
Fig. 4. Diagram of a typical female flower as seen directly from above. 
Figs.5 & 6, Female flowers displayed by splitting up the corolla from the side, to show 
the variation in the form of the anther rudiments. (The magnification is twice 
that used for the other figures of female flowers.) 
