No. 563] STUDIES ON DIFFERENTIAL MORTALITY 685 



one of these physiological experiments, in as far as they 

 bear upon the questions of the existence of a differential 

 mortality and of its consequences in the population. The 

 evidence which they afford concerning the causes under- 

 lying the differential death rate is a question too com- 

 plicated both biologically and statistically to be discussed 

 in the limits of this paper. 



For a fair understanding of the portions of the data 

 which are placed before the reader, it will be necessary, 

 however, to state briefly the general purposes which led 

 to the adoption of the particular methods employed. 



On the assumptions that the vigor of the seeds in- 

 creases from the lower to the higher weight classes, 5 one 

 might expect a mortality of seeds in the lower portion of 

 the range of variation due to innate incapacity for devel- 

 opment. One must then seek some other factor to ac- 

 count for the mortality of the heavier seeds. 



One of the simplest a priori assumptions is that the 

 larger seeds require longer to germinate and that they 

 are in consequence longer exposed to the vicissitudes of 

 germination — to deatli by excessive moisture or by exces- 

 sive draught before or shortly after expanding their 

 leaves. 



Now nothing whatever is here stated or implied in 

 favor of any of these suggestions. For the present, they 

 stand purely and simply as the first of a series of hypoth- 

 eses to be tested in the quest of the true interpretation of 

 an observed phenomenon. They are mentioned here solely 

 to explain why a particular series of experiments was set 

 up in the way in which it was. 



II. Methods 



The first thing needful in testing these hypotheses is to 

 determine the relationship between the size of the seed 

 and the time required for its germination. To do this, 

 while at the same time securing data for a further test of 



5 The chief evidence in support of this view is that afforded by the results 

 planted and the characters of the plant produced. But of course this cor- 



