Winifred E. Brenchlf.y 



•235 



the relative rates of increase (efficiency indices) week by week are shown 

 in tlie following table, and are seen to fall into two distinct groups: 



Mean difference in efliciency indices of "changed" 



The insignificant differences obtained in the early weeks emphasize 

 the fact that during the first period of growth and for perhaps a short 

 time longer the plant gains no benefit from the constant renewal of the 

 nutrient solutions, probably because during this time the demands on 

 the external sources of nutrition are comparatively small, as some supply 

 is still available in the seed and the manufacture of fresh food by as- 

 similation is not yet in full swing. When once this period is over the 

 variation in food supply makes itself felt at once, and this continues 

 throughout the life-history. No constant change in the difference of the 

 rate of increase is observable, but the variations are irregular through- 

 out. This is somewhat contrary to expectation, as it was surmised that 

 the drop in the efficiency indices in the unchanged series in late life 

 would be far greater than that in the changed sets where no starvation 

 effects were manifest. Apparently the natural fall in the indices that 

 occurs late in life in all cases was more or less parallel to that induced in 

 the semi-starved plants. 



The correlations between relative rate of increase (efficiency index) 

 and the three environmental measurements for the "unchanged" series 

 are as follows: 



