472 JOURNAL OF THE ROYAL HORTICULTURAL SOCIETY. 



sown in each pot, as far as possible at equal distances apart, and covered 

 to an equal depth with fine sifted soil. The members of one pot- 

 family are thus under conditions as nearly similar as possible, and can 

 be allowed to grow undisturbed for a considerable time. In order 

 to render the comparison as stringent as possible, only seedlings from 

 the same pot were compared, these all being derived from the seed 

 of one pod. When it becomes necessary to turn out the pots, the 

 seedlings forming each pot-family can be sorted and labelled in order of 

 their size. When the plants come into flower the following season the 

 total number of doubles from the whole culture may approximate 

 very closely to the expected 53 or 57 per cent., but if the plants 

 belonging to each grade are tabulated separately the distribution 

 of the doubles among the different grades shows a marked relation 

 to degree of vigour. Among the most vigorous plants doubles were 

 found to be greatly in excess of expectation ; in each successive grade, 

 as we descend the scale, the proportion of doubles steadily diminished, 

 until among the least vigorous plants the singles were often found to 

 be in the majority, the deficiency of doubles in these lower grades being 

 approximately counterbalanced by the great preponderance occurring 

 in the higher grades. If then, as in practice was found to be most 

 convenient, the group containing the most vigorous individual from 

 each pot-family is numbered i, that containing the second largest 

 individual from each family 2, and so on in order, we get a series in 

 which, as a rule, the higher the number of the group, the smaller the 

 proportion of doubles, and vice versa.* 



Hence it appears that, when the vegetative period is sufficiently 

 prolonged to render the method of selection by vigour practicable, 

 this method may be successfully employed by the gardener, and will 

 enable him to show in his beds a proportion of doubles far in excess 

 of the natural output of the plant. Furthermore, I now feel httle 

 doubt that in this greater vigour of development of the doubles at 

 an early stage we have the real explanation of the high percentages 

 of doubles quoted by the French horticulturist Chate,! and that the 

 70, 80 or 85 per cent, which he claims to result from reduction in 

 the number of pods allowed to mature, and rejection of seed from the 

 upper region of the pods, is, in fact, not due to these methods of 

 conscious selection, but to an unconscious selection in favour of the 

 doubles at the time of pricking or planting out. 



* I hope to publish full statistical results on this point shortly, 

 t loc. cit. 



