I 



WrNFFRKD K. UlfKNOHI.KV 221 



(a) Relation of lohil groirlJi lo sinisliiitc cnid fempernture. 



Under ordinary field conditions of cultivation the amount of water 

 supplied as rain and the action of wind in drying out the soil and causing 

 rapid alterations of temperature would influence the direct effect of 

 temperature and sunlight. Under greenhouse conditions the first two 

 factors are eliminated and the enquiry can therefore be narrowed down. 

 In order to work out the relationship of total growth to sunshine and 

 temperature it is essential that the growth should have occurred over 

 equal periods, and for this reason only the first eleven weeks of each 

 series are considered. One series, 0, died after seven weeks, and this 

 will be specially considered later on. The results are most clearly demon- 

 strated by Fig. 3, in which are graphed for each series the total growth 

 of the plant, the total number of hours of sunshine over the 11 week 

 period, and the ranges of the mean weekly maximum and minimum 

 temperatures over the same period. As would naturally be expected 

 the total grow^th in summer was much in advance of that in winter. 

 The three series C, E, G, started in October, December and January, 

 show a steady increase in the total growth, G being more than twice as 

 heavy as C. The available sunshine in all three cases was the same, but 

 the range of maximum temperature was successively rather higher, 

 though the greatest difference was only six degrees (60°-66° F.). In the 

 dead season of the year, therefore, slight differences in maximum tem- 

 perature are very important in determining the amount of growth and 

 they act independently of the available sunshine. The ranges of minimum 

 temperatures were similar in all three cases and therefore did not in- 

 fluence the result. 



The optimum time for growth under the experimental conditions 

 was spring and early summer. In the presence of an abundant and un- 

 stinted food supply series K, started in March, made five times as much 

 growth as G, started two months earlier. Double the amovmt of sunshine 

 was available and the temperature ran up steadily from a rather low 

 level until it almost reached the maximum weekly mean attained during 

 the season. The series M, started a month later, did not do quite as w^ell, 

 but this may have been due to the fact that though the highest mean 

 temperature was the same as K the means for the last three weeks were 

 comparatively low^ 



A striking point emerged w^hen an attempt was made to grow a 

 series at the beginning of July. The weather w^as bright and sunny 

 throughout, the seedlings were started at 80° F., for only one week did 



