THE STRUCTURE OF CERTAIN NEW HYBRIDS. 159 
examples measure 2! to 2,*, in. long. They are light yellowish-green, and 
rather pleasantly scented. The outer wall, beneath the somewhat tough, 
green, easily separable skin, is } to } in. thick, and is white, soft, and 
spongy. From it pass numerous soft threads, once or twice branched, 
connecting the outer wall with the thin membranous sae which bears the 
placentas internally. The greatest numbers of seeds found were 258 and 
232, the former in a fruit resulting from fertilisation by pollen of another 
flower of the same plant, and the latter in a self-set fruit. The seeds are 
elliptical, ,3, in. long, black, and obscurely pitted. The viscid yellowish 
fluid which accompanies them is bitter and unpleasant to the taste, and 
gives a disagreeable odour to the fruit when it is opened. 
Comparatively few experiments in cross-fertilising this species have 
been carried out. Success followed pollination by P. Buonapartea and 
P. edulis respectively (infra, p. 165). The application of pollen of P. Con- 
stance Elliot has led to variable results. As an outcome of experiments 
in 1896, the fruit in one case developed fairly well, and bore numerous 
immature seeds ; in another, a somewhat nearer approach tu success was 
reached ; in a third, a fully developed fruit was obtained, containing 180 
seeds of all sizes, but only three or four appeared to be normal. A fourth 
swelled rapidly, and promised well, but was accidentally broken off when 
nine days old. 
During the present season (1899) a good fruit was obtained by the 
same cross. It contained 200 seeds, all more or less badly developed, 
except one which alone was embedded in yellow fluid, and possessed the 
very deep brown colour characteristic of a healthy condition. All the seeds 
were sown, but only one—the good one—has germinated. The seedling, 
still young, appears to have a vigorous constitution. It bears a close 
resemblance to the plants of the reciprocal cross, P. Constance Elliot x 
P. alba (p. 166) when these were of corresponding age. tis interesting 
to note that the pistil of the flower fertilised was four-rayed, and that the 
fruit secured possessed four, instead of three, placentas. 
The additional organs are doubtless due to the splitting of one of 
the normal three. The numbers of seeds borne by the four placentas 
of a self-set fruit of P. alba bear out this hypothesis, these being 65, 76, 
22,39. The total of the two latter numbers represents a good average 
number for one placenta. 
The application of the pollen of Tacsonia Van Volxemi unexpectedly 
caused the ovary to ewell to a length of 2 in., but no further develop- 
ment ensued. 
No swelling took place when pollen of the following Passion-flowers 
was used: P. Decaisneana (4 experiments), P. suberosa (2), and P. 
Watsoniana (2). 
PASSIFLORA ALBA x P. BUONAPARTEA. 
An experiment in crossing P. alba by P. Buonapartea, carried out in 
June 1896, proved successful. The fruit ten weeks after fertilisation was 
so ripe that a very light touch caused it to fall. It was of good size, 
and in general structure did not differ from the fruit of P. alba, as 
described above. The placentas bore respectively 24, 24, and 25 good 
