iSS Harris. Further Observations on the Selective Elimination of Ovaries etc. 



be placed upon the calculated probable errors. I therefore lay no 

 stress whatever on the exact numerical results. Bearmg in mind 

 probable errors, they are very consistent throughout. 



3. The correlations between the position of the fruit on the 

 inflorescence and the characters of the fruit are low. Sometirnes it 

 is even impossible to be sure of the sign of relationship. 



4. The data and analysis described above throw no doubt upon 

 the conclusions drawn in an earlier memoir on the selective elimination 

 occuring during the development of the ovary, but tend to make 

 them more significant by showing that apparently they are not to be 

 explained by a combination of such simple factors as a differentiation 

 of the ovaries associated with their position on the inflorescence and 

 a proportionately higher but random elimination in the more distal 

 regions of the inflorescence. 



Probably, however, the slight differentiation of ovaries with 

 respect to position on the inflorescence does account for some of the 

 difference between eliminated and matured ovaries. The difference 

 in mean number of ovules may be in part due to this cause. There 

 is no evidence that the results announced for radial asymmetry or 

 locular composition can be explained in any other way than that 

 ovaries with the higher grades of asymmetry and with an excess of 

 "odd" locules have a lower capacity for development. 



Cold Spring Harbor, L. I., U. S. A., Januarj- 15, 1911. 



