188 NATUBAL SCIENCE [March 



1.3.1.3. 



1.3.2.1. 



1.3.2.2. 



1.3.2.3. 



1.3.3.1. 



1.3.3.2. 



1.3.3.3. 

 It will be noted that the greater number of " the observed possi- 

 bilities, 20 out of 35 (57 per cent.), occur when the calyx whorl is 

 at the mean. Of course, recollecting that in (practically) 96 per 

 cent, of the specimens there was no variation from the mean in the 

 calyx, such would naturally follow. Of the remaining 4 per cent. 

 — 41 specimens — 15 have purely chance variations, a rather less 

 proportion than that of the whole series — 36 per cent, of 41 speci- 

 mens against 43 per cent, of 1000. 



I have now shortly discussed correlation of variation in calyx, — 

 calyx and corolla, — stamens and carpels, — and lastly, — correlation of 

 variation in all four whorls collectively. To summarise the results, 

 we found in the analysis of the calyx that increase in corolla was in 

 every case accompanied by increase in the stamens and carpels in- 

 dependently of the calyx. Therefore, as regards the calyx on the 

 one hand, and corolla, stamens, and carpels collectively on the other, 

 the data obtained seem to show that the number of members of the 

 calyx has no effect whatever on the other three whorls. In this 

 instance at least, any variation is pure chance. It may be, how- 

 ever, that this conclusion should be modified on account of the small 

 number of specimens affecting it. 



Calyx and Corolkc 

 The results obtained here are rather perplexing, but in view of 

 the number of specimens discu.ssed (202), I presume, must be 

 accepted. 



