130 CARNEGIE INSTITUTION OF WASHINGTON. 



SELECTION AND PURE-LINE WORK. 



This field of study deals with problems of the greatest present-day 

 interest. Doctors Harris, Banta, and MacDowell are carrying on this 

 work. 



Studies on Selective Mortality, J. A. Harris. 



For several years the problem has been investigated whether in a lot 

 of planted seeds of one species there is a selective elimination of any 

 type or whether each occurs strictly at haphazard. Recently Dr. Harris 

 has planted about 46,000 individually weighed seeds of the bean 

 {Phaseolus vulgaris), treated them alike, and classified their subsequent 

 history into three groups: (1) seeds that germinated normally; (2) seeds 

 that germinated by producing more or less abnormal seedlings ; (3) seeds 

 that failed to germinate. Comparing the weights of the planted beans 

 that had respectively these three fates, it is found that the average 

 weight of all the varieties of beans that germinated normally and of 

 those that failed to germinate was nearly the same, although in two 

 varieties of bean there is a distinct superiority in the mean weight of 

 the survivors, while in one other variety there is a pronounced inferi- 

 ority in the mean weight of the survivors as compared with those that 

 failed. But, on the other hand, when absolute variabilities of the three 

 sets of progeny are compared, it appears at once that seeds that germi- 

 nate vary less around the mean than those that fail to germinate. When 

 variability is expressed in units of the naean weight it appears still more 

 clearly that the seeds that germinate are a selected lot — selected for 

 their closeness to the mea n weight of all the bean seeds of their variety ; 

 the seeds of aberrant weight are less apt to germinate. In so far as the 

 weight-aberrancy is an inheritable trait, it is easy to see that the greater 

 mortality among the aberrant seeds will tend to hold the race to the 

 mean or '' typical" seed weight. In general, selective mortality operates 

 conservatively, tending to preserve pure the specific characters. These 

 studies have been extended to peas, and in these, also, the seeds 

 that germinate are on the whole less variable than those that fail to 

 germinate. In comparing the weight of the seeds that germinated 

 promptly with those that developed slowly, it became necessary to 

 get the exact time of germination of these 46,000 seeds; and that 

 involved observing at night as well as in the day. As a by-product, 

 the law was established that, in beans, the heavier the seed the more 

 slowly it germinates. 



Relationship between the Weight of the Seed Planted and the Characteristics 

 of the Plant Produced, J. A . Harris. 



The fact that a certain quality (such as aberrancy of seed weight) is 

 associated with exceptionally great mortality is no evidence that the 

 first quality is the cause of the second. It may be that the first cause 



