.362 HYBRID OF DIGITALIS. 



ilutea parent were without a purple tinge; but careful examination 

 showed that at the base of the filaments, where attached to the 

 corolla, theie was a small, faint purple area. The yellow plant 

 obviously brought come influence or factor into the hybrid render- 

 ing a purple colouration possible, since if the white gloxiniceflora 

 had been crossed with another white plant of the same species all 

 the offspring would have been white. 



The general yellow colouration of In tea was transmitted con- 

 siderably to the hybrid, but the purple colour masked it greatly. 



In the intensity (34) of the purple colouration the hybrid was 

 nearly midway between the two parents with mean intensities of 

 77 and respectively. In spotting percentage the hybrid was only 

 •5, and consequently in this character it was nearer to lutea with 

 spotting than to gloxiniceflora with mean spotting of 20. In the 

 absolute length and breadth of the flower, and also in the ratio 

 of these two dimensions the hybrid was closer to gloxiniceflora than 

 to lutea (PI. IV, fig. 2). 



In lutea the lower lip exhibits no spotting, but it bears a 

 ■characteristic marbled pattern in gamboge or brown. In the 

 hybrid, faint traces of these brown marblings could generally be 

 found, especially on the lateral, internal surfaces. In addition to 

 this, the marbled pattern was frequently to some extent marked 

 out in purple. 



4. Hybrid (O) x glox. (cj 1 ) and Hybrid (Q) x lutea (cf) — 



Backcrosses. 



The hybrids were very sterile, and a microscopic examination 

 showed that the pollen was very defective. Numerous flowers of 

 different plants were pollinated with their own pollen, with pollen 

 from other hybrids, and with pollen from gloxiniceflora and lutea 

 plants. Several hundred pollinations were made. With self-fer- 

 tilisation, or with pollen from other hybrids, the flower did not 

 fall until faded, and the ovary grew but very little. Nevertheless 

 even with such pollination the ovaries became somewhat more 

 swollen than when no pollination had been attempted. With 

 gloxiniceflora pollen, in the case of some of the hybrids, the flowers 

 dropped quickly and the ovaries became greatly swollen. Lutea 

 pollen acted still more energetically, and exceptionally large 

 capsules were produced. In the case of all the hybrid plants, 

 except in one, these capsules on ripening contained nothing but a 

 few shrivelled seeds which seemed incapable of germination.! 

 In one hybrid plant the pollen was found to be better in quality, 

 judging from microscopical appearance, and in this plant only, out 

 of the 32 plants, the majority of the capsules resulting from pol- 

 lination with lutea and with gloxiniceflora pollen contained some 

 seeds of fairly normal appearance. The seeds showed some varia- 



f Note. — Subsequent investigation has shown that some of these seeds 

 possessed the power of germinating and producing healthy seedlings. 

 April, 1922. 



