SELECTION WITHIN PURE LINES OF PESTALOZZIA 177 



means consistent. The first set of selections produced a higher number of 

 spines in the group selected for low spine number than in the group selected 

 for high spine number. It is unfortunate that at this point a change was 

 made in the method of selection. The author attributes the effect secured 

 in further selections to the use of a better criterion for the selection. The 

 effect may have been due to independent chance fluctuations within the 

 two groups and not to selection at all but the change in procedure prevents 

 this being detected. It would have been very interesting to know whether 

 or not such divergences would have occurred had the original method been 

 continued. Five further selection periods showed an increase in spine 

 number in the group selected for high spine number and a decrease in the 

 group selected for low number. The difference between the high and the 

 low series, however, does not show a steady increase, but fluctuates from one 

 selection period to the next. When the divergence between the two groups 

 was at a maximum, selection was stopped, so that we do not know whether 

 the divergence would have decreased or not had selection been continued. 



On the cessation of selection the divergence between the groups de- 

 creases very little for one period. In the next period the divergence falls 

 to almost nothing while in the next following period it again rises to nearly 

 half that secured at the end of the six selection periods. It is approximate- 

 ly equal to that found in the second selection period. In the second selec- 

 tion period the increase in divergence is attributed to selection. To what 

 is it due in this non-selection period? It seems logical to assume that it 

 is due to the same cause in both cases, probably to the independent 

 chance fluctuations of the two groups. 



Further selection for one period in the high line resulted in a slightly 

 larger number of spines in the group selected for low spine number than 

 in that selected for high spine number. That this is due to selection is, 

 of course, impossible. 



Three selection periods were effective in that they developed two diver- 

 gent lines within the low line established by the early selection. Here 

 again the divergence fluctuates from period to period. After selection 

 ceases the divergence becomes greater than it was during the selections. 

 This again indicates that the divergence may be totally unrelated to the 

 selection. In experiment 1 of this paper may be found similar divergences, 

 which were only temporary. 



In another experiment HEGNER practiced selection for diameter and 

 spine number for a period of nine days. During this time distinct high 

 and low lines were produced. In a non-selection period of thirteen days 

 following the selections the divergence between the high and low series 



GENETICS?: Mr 1922 



