MINUTE STRUCTURE OF PLANT HYBRIDS. 



231 



Petals. — The external surface of the petals in 1 shows relatively large, straight - 

 walled, or slightly sinuous cells; the surface of 3 shows small cells with zigzag walls, though 

 towards the base they become nearly straight- walled ; in 2 the cells are nearer the latter 

 in shape, but nearer the former in size, though it is extremely hard to determine with 

 accuracy how far these may not be intermediate. Long delicate hairs grow out from the 

 epidermis in 1, none occur in 8, while a few are encountered in the hybrid. The inner 

 surfaces resemble the outer except that all the cells are prolonged into tubercular 

 outgrowths. 



Stamens. — The pollen cells (Plate V. figs. 10, a, b, c) in both parents are, as a rule, 

 extremely good, but those of 3 exhibit considerable diversity in size. This I regard as a 

 matter of extreme importance in the discussion of sexual potency, and some attention will 

 be given to it later on. The pollen cells of 1 are 27 fi across, and those of 3, 32 /*, while 

 those of the hybrid are very bad, consisting largely of small, irregular, shrivelled cells con- 

 taining little protoplasm. Here and there, however, one encounters a normal-looking cell, 

 filled with finely granular protoplasm as in the parents, and one of these is illustrated in 

 fig. 10 b. Another figured beside it would probably prove inferior in pollinating action. 



Pistil. — Muller has pointed out that R. nigrum is almost habitually self-pollinated, 

 that R. Grossularia is not only greatly frequented by appropriate insects, but that 

 arrangements in the flower itself favour cross pollination. There are one or two pretty 

 contrivances which favour the latter view. 



In 1 the style is deeply split to a point, from which downwards a dense coating of long 

 hairs grow out so arranged that they form an insect guard to prevent ingress of small insect 

 thieves. In 3 the style is simple, and ends in a bifid stigmatic knob ; its surface also is 

 glabrous. The style of the hybrid is split to half the extent, and the style hairs are, as 

 nearly as one can estimate, half in size and number those of the first parent. 



The receptacular surface above the ovary in 1 has each cell elevated into a fine papil- 

 lary hair, and these aid the style hairs in excluding small insects. The corresponding pro- 

 cesses in 3 are very minute papillae, which one inclines to regard as functionless, though 

 caution is needed in assuming even this. The hybrid has papillae of half the 

 length. 



The ovarian surface of 1 is rather sparsely covered with long narrow hairs mixed with 

 club-shaped hairs ; that of 3 is densely covered by fine, short, curved hairs, with here and 

 there a sessile gland hair ; that of 2 exhibits all the four types just noticed. 



Large loose cells filled with brown contents are found in the mesophyll of the ovarian 

 wall in 1 ; the corresponding cells of 3 are all small, nearly uniform in size, and quite so 

 in colour ; in the hybrid numbers of the brown cells derived from the former parent are 

 readily traced. 



As with other hybrids, I have not yet attempted to examine the ovules minutely, but 

 while working at other parts opportunities for observing these occurred, and they seemed 

 to be well grown and to contain an egg-cell. Against perpetuation of the plant, however, 

 as a pure hybrid is to be reckoned the bad quality of the pollen, while Mr Culverwell, 



1. Ribes Grossu- 



laria. 



2. Ribes Culver- 



■wellii. 



3. Ribes nigrum. 



I 



