September 19, 1913] 



SCIENCE 



403 



rate is selective. The difficulties, however, of 

 thfr problem of natural selection, the least in- 

 vestigated of all of the primary factors of 

 organic evolution, demand the collection of 

 evidences from every possible source. 



The purpose of this letter is to call attention 

 to inter-varietal and inter-specific competition 

 as a source of information on natural selection, 

 to illustrate the point by one or two recently 

 published observations, and to urge the accu- 

 mulation of more (and more precise) data of 

 this kind by those field naturalists and experi- 

 mentalists who have the opportunity for this 

 sort of work. 



The kind of studies to which I refer are 

 illustrated by Brimley's interesting account^ of 

 the capture of Ealeigh, 'N. C, by the wharf rat, 

 Mus norvegicus.. Up to 1909, the roof rat, M. 

 alexandrinus, was the only species seen during 

 a residence of twenty-five years. Since then it 

 has been nearly or entirely replaced by the 

 wharf rat." 



For a second illustration turn to botanical 

 material. 



Varieties of plants are generally believed to 

 differ in their susceptibility to disease. An in- 

 teresting demonstration of this is furnished by 

 researches on the potato fungus, Phytophthora 

 infestans. Jones, Giddings and Lutman have 

 shown* that there is a correlation between the 

 percentage growth of the fungus on various 

 varieties in the test-tube and the percentage of 

 rot observed in field trials on clay and sandy 

 soil by Stuart. They find for laboratory 

 growth and loss on clay a correlation of 

 r = .584 ± .059 and for laboratory and sandy 

 soil a correlation of r = .594 ± .055.' For con- 



= Brimley, C. S., ' ' Capture of Ealeigh by tlie 

 Wharf Rat," Journ. Elisha Mitchell Sci. Soc, 

 28: 91-94, 1912. 



'Dt. Hatai tells me that when he placed white 

 rats on an island inhabited by the brown rat, M. 

 norvegicus, the two forms at once established dif- 

 ferent centers and began fighting each other. See 

 also Year Booh Cam. Inst. Wa^h., 10: 83-84, 1912. 



* Jones, L. R., N. J. Giddings and B. P. Lutman, 

 Bull. Vt. Agr. Exp. Sta., 168: 74-81, 1912. 



° These are calculated on grouped data. I have 

 recalculated without classing and find results 

 agreeing within less than half the probable error. 



venient comparison I have also worked out the 

 correlation between the percentage rot of the 

 same varieties on clay and sandy soil, using 

 Stuart's data as quoted by Jones and his asso- 

 ciates. For the ungrouped material r = .707 rb 

 .045. 



Clearly enough there is a pronounced indi- 

 viduality in the varieties with respect to their 

 capacity for resisting disease. The interest of 

 such a result from the standpoint of natural 

 selection is clear, for in free competition more 

 susceptible strains would be rapidly weeded 

 out, and the morphological or physiological 

 characteristics to which their inferiority is 

 due would be lost. 



Now of course in neither of these cases do 

 we know why (i. e., because of what peculiar 

 characteristics) one variety or species was less 

 capable than another of survival. Nor can we 

 know until the questions are more intensively 

 studied. But one can not doubt that these 

 problems will yield to proper and persistent 

 observation. 



My point is merely that this sort of work 

 may (if carried out extensively and inten- 

 sively enough) have a most important bearing 

 upon the two fundamental questions of natu- 

 ral selection. First, is the death rate random 

 or selective ? Second, if selective, what weight 

 has each individual character in determining 

 the chances of survival of the individual? In 

 seeking the answer to the second question it 

 may be desirable to deal with characteristics as 

 strongly contrasted as possible — with varietal 

 or specific differences instead of with intra- 

 varietal variations — in order that the proxi- 

 mate causes of the differential mortality may 

 be more easily recognized. 



The value of such work for the problem of 

 natural selection will be quite supplementary 

 to that for which it was primarily carried out. 

 May we not, therefore, have more observa- 

 tions of this kind, carried out in such detail 

 that they may be of value to the evolutionist 

 seeking to ascertain the selective value of 

 individual characters? 



J. Arthur Harris 



Station foe Experimental Evolution, 

 August 20, 1913 



