No. 563] STUDIES ON DIFFERENTIAL MORTALITY 687 



pletely sunk in the sand so that the series number and 

 the weight of the seed were quite unknown at the time of 

 recording the results. Thus personal equation as far as 

 it implies any bias with regard to the material was 

 absolutely excluded. 



At the outset, I must emphasize the fact that this tech- 

 nique (which I still believe is the best possible under 

 all the requirements) falls far short of what one would 

 desire. The germination of bean seeds under glass on a 

 large scale is a rather difficult process. If a sufficient 

 supply of moisture can be held in the soil from the begin- 

 ning to the end of the experiment and the temperature 

 be kept fairly high, the problem of good germinations in 

 the greenhouse is solved. But when one is doing the work 

 during the period of hot days and cool nights coming 

 in the early fall, the question of maintaining proper soil 

 moisture and temperature is a very serious one. It is 

 remarkable how heterogeneous the environment of a 

 single section of a greenhouse system is ! This is espe- 

 cially noticeable in the drying out of the pots in sand cul- 

 tures. Just here lies one of the greatest difficulties. The 

 germinating bean seedling is very sensitive to watering, 

 especially in connection with low temperature. I imagine 

 this is particularly true of old seeds which have nearly 

 lost their viability. Probably the considerable irregu- 

 larity in our percentages of germination is very largely 

 due to the impossibility of controlling closely enough the 

 soil moisture. 7 



In classifying, three groups were recognized: (A) 

 seeds germinating normally, (B) seeds germinating but 

 producing seedlings more or less abnormal, (C) seeds 

 failing to germinate. 



On general grounds, the recognition of the three classes 

 seemed desirable ; for purposes other than those of this 



