364 HYBRID OF DIGITALIS. 



The question arises as to what interpretation is to be placed 

 on the differences in the seed produced by the same plant under 

 the action of the two kinds of pollen. 



The mere act of pollination and the growth of the pollen 

 tubes in the style obviously exeit a stimulus on the mother-plant, 

 and induce the fall of the flower and growth of the ovary. 



After fertilisation, and when the development of the embryo 

 and endosperm commences, a continued influence is exerted on the 

 mother-tissue (cf. plant and animal galls), and it is of interest to 

 enquire as to how far such an influence is specific and depends on 

 the nature of the fertilisation. With lutea pollen the resulting 

 mature seed c tended to have smaller integument cells, persistent 

 outer cell-walls and lighter coloured cell-walls than when 

 gloxiniceflora pollen was used. In these characters the backcross 

 seeds tended to resemble ordinary hi tea seeds or gloxiniceflora 

 seeds according as to whether lutea pollen or gloxiniceflora pollen 

 was used for fertilisation. The results obtained indicate a more 

 specific influence being exerted by the growing embryo and 

 endosperm on the mother-tissue than might have been expected. 

 Such an influence is, of course, not heredity in any ordinary sense. 



It might have been anticipated that the general size and shape 

 of the backcross seeds would depend on the size and shape of the 

 embryo and endosperm mass ; but the shape of the seed produced 

 by the cross, gloxiniceflora (Q) x lutea (cf) was not intermediate 

 between the shapes of the seeds of the two species, and in fact it 

 resembled neither. This was also the case in the seeds obtained by 

 the backcross fertilisations. The differences in size and shape were 

 perhaps mainly due to the fact that there was an arrest in the 

 general growth of the seed owing to a cessation in the growth of 

 the embryo and endosperm, resulting through uncongenial fertili- 

 sation. When such an arrest took place the growth of the integu- 

 ment tended always to exceed the requirements, and the formation 

 of an internal space was prevented by the hypertrophy of the 

 tapetal layer producing the peculiar compact tissue which occurred 

 in all of the undersized seeds. 



6. Gloxiniceflora (Q) x Hybrid (cf) and lutea (Q) x 

 Hybrid (cf)- 



The pollen of the hybrid was tried on a considerable number 

 of gloxiniceflora plants, but there was little response. In one case 

 some 6 or 7 seeds were obtained in a capsule, and from these only 

 one plant was raised. The cross between the fertile hybrid and 

 lutea was tried, but no seeds were obtained. 



It is clear that the cross between the hybrid as male and either 

 original species as female was less easy to obtain than the cross 

 with the hybrid as female and either species as male. 



7. The Gametic Constitution of Hybrid and Backcrosses. 



The hybrids were sterile with each other, and with their own 

 pollen,* but with reference to the factorial hypothesis and the 



— In the subsequent season more success has been attained in self in g 

 hybrids. April, 1922. 



