528 



SCIENCE 



[N. S. Vol. XXXII. No. 824 



able to ascertain whether one or both of the 

 characters are of some selective value. 



Tables III. and IV. show the results for the 

 1908 material. The differences are not large, 



TABLE rv 



1908 Series. Selective Elimiiiation for Radial 



Asymmetry xoithin tlie four Loeular 



Composition Types 



Difference in Radial Asymmetry in 

 Character of Ovary. Elimination and Matured Ovaries. 



3 " even " — .017 



2 "even" + l "odd" — .019 



1 " even " + 2 " odd " — .034 



3 "odd" —.077 



but they are consistent throughout. This in- 

 dicates, I think, that both characters are of 

 some independent selective value. 



Change in Number of Locules per Ovary 

 due to Selective Elimination. — Normally in 

 Staphylea the fruit is three-celled, but not 

 infrequently (in some individuals especially) 

 those with two and those with four cells occur. 



To determine finally whether either of these 

 types, has better chances of surviving than the 

 others would require very large series of ob- 

 servations. 



So far as the results from 2,000 eliminated 

 and 3,000 matured ovaries are trustworthy 



TABLE v 



1908 Series. Showing Elimination of Dimerous 



Ovaries 



Table V. shows that dimerous ovaries are 

 more liable to elimination than tetramerous 

 ones. 



Summary. — From the constants in the fore- 

 going sections there can be little doubt con- 

 cerning the fact of a selective elimination of 

 the ovaries of Staphylea during their develop- 

 ment from flowering time to the maturing of 

 the fruit. By this selective elimination the 

 mean number of ovules is increased, the mean 

 radial asymmetry is lowered, the proportion of 



ovaries with odd numbers of ovules in one or 

 more locules is very stringently cut down, and 

 perhaps the mean number of locules per fruit 

 slightly raised. 



These results are, I am inclined to think, 

 of considerable importance from the stand- 

 point of morphology and physiology. They 

 show that a physiological unfitness — an in- 

 capacity for developing to maturity — ^is 

 coupled with certain definite morphological 

 characters. Personally I take it that we are 

 not to assume that low numbers of ovules, 

 high radial asymmetry and the presence of 

 " odd " locules are fundamental causes of the 

 incapacity for development, but rather that 

 both morphological and physiological pecu- 

 liarities are dependent upon some inherent 

 abnormality of the growing point which 

 morphologically finds its expression in the 

 structural features of the fruit and physiolog- 

 ically in its relative capacity for development. 

 These interrelationships between slight aber- 

 rations of structure and the capacity of organs 

 for performing their functions offer a most 

 attractive field for research. 



In their bearing on the problem of organic 

 evolution the results outlined in this paper 

 are of interest in showing that natural selec- 

 tion may act upon the organs of an individual 

 as well as upon the individual organisms of a 

 population. Without knowing whether the 

 characters we have investigated are inherited 

 it is impossible to say that this elimination 

 is a factor in maintaining the present type 

 of the species." And to argue that this kind 

 of natural selection has been of significance 

 in evolving the considerable degree of radial 

 symmetry found in the fruits of many species 

 of plants with compound ovaries would be 

 stepping too far from a secure pier of facts 

 into the uncertain bog of speculation. 



J. Arthur Harris 

 Cold Speing Haebob, N. Y., 

 September 8, 1910 



' It seems to me unlikely that we shall ever be 

 so fortunate as to find many eases of Darwinian 

 evolution going on in nature. That a constant 

 selection may maintain a type already secured, 

 and that one may be able to observe and measure 

 the intensity of this factor, seems much more 

 probable. 



