DOUBLE FLOWERS. 



481 



relied upon as a means of obtaining an increased proportion of doubles. 

 Only in the case of one strain, so 'far as I am aware, is it possible by 

 the method of seed selection to ensure an abnormally high yield of 

 doubles, selection in this case being based not on the shape but on the 

 colour of the seed. The strain in question is the curious form known 

 as sulphur- white, a particularly interesting form, ,for it has the 

 peculiarity of being ever-sporting in respect of a combination of 

 characters. The single individuals of this race are white, but though 

 se?/- fertilized they invariably yield creams as well as whites ; they also 

 yield a mixture of singles and doubles, the singles being white, the 

 doubles almost all cream. In these respects material obtained from 

 any source is found tO' behave ahke. In regard to the seed characters, 

 however, there is not always the same uniformity. Seed supplied by ' 

 some growers is uniform and has the small brown, crumpled appear- 

 ance characteristic of a pure-breeding white strain (fig. 176), while 

 seed obtained from other firms is larger, more symmetrical, fiat, 

 and mixed as to colour, about half being rather browner in tint, and 

 half yellowish like the seed of such a pure-breeding cream form as 

 * Princess May, ' in which the seeds are also large, flat, and regular 

 (fig. 176). This diversity of seed character points to a different origin 

 in the two cases. It is not unlikely that the sulphur-white form 

 may have arisen afresh from time to time by intercrossing, and 

 that the immediate progenitor was not derived in exactly the same 

 way in eveiy case. Be this as it may, the fact that certain sulphur- 

 whites yield a mixed seed which can be sorted according to colour 

 furnishes us with a means of distinguishing among such seed those 

 that will give rise to singles and doubles respectively. For the 

 yellowish seeds are those which give rise to the creams, and, as stated 

 above, the creams in this strain are all double. Hence, by selecting 

 only the yellowish seeds for sowing we can obtain a crop consisting 

 entirely, or almost entirely, of doubles, complete success depending 

 upon the seed having ripened well, so that the distinction of colour is 

 well-marked. In this way we can, if we wish, effect a fairly complete 

 separation of the doubles intended for planting out, and the singles to 

 be retained as seed-bearers, though a doubtful seed placed in the 

 wrong category may result in a rogue " here and there. 



In the early literature on this subject the statement is frequently 

 to be met with that seed which has been kept for a considerable period 

 before it is sown yields a higher proportion of doubles than seed more 

 recently harvested. My own experiments have confirmed this view 

 in so far that it was found that if seeds are kept until almost all have 

 lost the power of germinating the few survivors often give an excep- 

 tionally high proportion of doubles. This, however, must not be taken 

 to indicate that the lapse of time has caused doubling to result in an 

 mdividual which, if raised at once, would have been single, as was 

 probably the idea of the early writers; but rather that the seeds 

 destined to give rise to doubles possess, on the average, a sHghtly 

 greater vitality than those which give rise to singles, and hence 



VOL. XXXVIII. I I 



