480 JOURNAL OF THE HOYAL HOETiCULTURAL SOCIETY:. 



constitution, having arisen from the same combination, viz., from an 

 ovule carrying singleness and a pollen-grain carrying doubleness. 



The practical outcome of this differential distribution among ovules 

 and pollen of the factors causing the production of normal flowers is 

 that the average output of doubles by an ever-sporting single remains 

 constant, and that we are unable by changes in the external conditions 

 to increase or diminish this output, just as we are unable by these 

 means to affect such characters as the surface of the leaves or the 

 colour of the flower. Singleness in the flower, hoariness in the leaves, 

 colour in the petals, are equally the result of the presence of certain 

 factors transmitted from parent to offspring, that is to say, they 

 severally belong to the category of inherited characteristics. 



In the case of the Ten-week races of Stocks the output of doubles 

 amounts on the average to from 53 to 57 per cent. In a sowing here 

 and there, especially if small, the number of doubles may fall con- 

 siderably below or distinctly exceed the average, but these deviations 

 from the mean are of the nature o'f accidental variations which with 

 larger counts tend to disappear. Once a pure strain is obtained, the 

 maintenance of a definite proportion of doubles in successive crops is, 

 in fact, automatically assured, provided that crossing with non-double- 

 throwing or cross-bred singles is prevented. 



Where the aim is to obtain as many doubles as possible interest 

 now centres round the question whether it is possible to distinguish, 

 before sowing, the seed which will give rise to double, and so, by 

 sowing only such seed, to eliminate the singles. For it is by seed 

 selection only that we can now hope to achieve progress in this 

 direction, since it has become clear that methods of cultivation have 

 no effect upon this character of the flower, the form of which, whether 

 single or double, is already predetermined in the seed. The statement 

 is often to be met with that a higher proportion of doubles can be 

 obtained from the lumpy irregular seeds than from those which are 

 flat and regular in shape. This belief has possibly arisen simply from 

 an association of ideas, connecting the regular symmetry of the single 

 flower with the more regular and symmetrical seed-form, and is con- 

 trary to such evidence as is at present available. The shape of the 

 seed is evidently mainly, if not wholly, determined by the dimensions 

 of the pod, and by the spacing of the ovules (fig. 177). In some strains, 

 as, e.g., the wallflower-leaved cream form 'Princess May,' the pods 

 may measure more than 6 inches in length and are wide in proportion. 

 Even when these pods contain from 70 to 80 ovules, and all are 

 fertilized, there is more than sufficient space for the seeds to lie flat 

 without being compressed through contact or actual overlapping. 

 Every seed in these pods is often quite regular and flat, yet there is 

 no deficiency of doubles from such strains. On the other hand, in 

 some of the short-podded strains the septum is so narrow, and the 

 seeds are so crowded together, that, of necessity, they are often 

 irregular and lumpy ; yet these types yield only the normal proportion 

 of doubles. Clearly selection of seed based upon shape cannot be 



