Further Observations on the Selective Ehmination of Ovarieb in Staphylcn. ijs^ 



inflorescence, and the fertility characters of the fruit. But since we 

 have before us this problem of selective elimination we must do the 

 best we can with the characters as nature supplies them, depending 

 upon large series of obser\'ations to smooth our data. 



Recognizing the limitations of our material we must demand 

 moderately concordant results from different series of observations, 

 and not read our coefficients to too many decimal places. With these 

 precautions I think our constants are quite trustworthy enough for 

 present purposes. 



The material consists of three series of mature fruits: the first 

 of 2059 pods collected in 1906, the second of 4033 pods gathered in 

 1908 1), and the third of 2082 pods taken in 1909. All come from a 

 series of numbered shrubs in the North American Tract of the Missouri 

 Botanical Garden. The 1908 series includes the matured series of the 

 three upon which the problem of selective elimination was worked out. 



The characters considered are: 



Position of the fruit on the inflorescence axis (p). The position 

 of the fruit is recorded from the proximal to the distal end of the 

 inflorescence. The position of the node, not the actual serial order 

 of the fruit is the character taken into account. For example, if the 

 flowers of the first node failed to produce a fruit while these of the 

 second did, the position would be recorded as two, not as one. In 

 the correlation tables / is always to be thought of as a distance of 

 so many nodes from the beginning of the flower-bearing part of the 

 inflorescence. 



Number of fruits per inflorescence (11). This character can be 

 determined with perfect accuracy. Its interrelationship with the others 

 is primarily of physiological interest and has relatively little bearing 

 upon the problem of selective elimination. 



Number of ovules per locule (0). 



Number of seeds developing per locule (s). 



Radial Asymmetry of fruit (a). This is measured by the coefficient 

 of aftjTTimetry employed in the memoir on selective elimination. 



Locular composition (c). This is the composition of the fruit in 

 "odd" and "even" locales. It is measured here as in the former 

 paper by the number of locules with an odd number of ovules. 



The only method which I know to be applicable to our problems 

 is analysis by means of the biometrical formulae associated with the 



1) A severe frost in the spring of 1907 killed such a high proportion of the 

 ovaries that any investigation of number or position seemed idle. 



