I'iO Ph]ii<iolofj\c(il Pr< -ihtiinnnnfion 



The following table gives both the air-diy weight and the absolute 

 dry-weight of 1000 seeds ripened in each case. 



* Calculated from three achenes, all that wero produced from 1000 plants. 

 t Calculated from 600 achenes. 



It is seen that the largest average seed-w^eight is never obtained in 

 full light, but with Cannabis saliva in \ light, with Saponarut officinalis 

 in I light, wdth Sinapis arvensis in f light, wdth Amaranthus retrojlexxs 

 in f light, and with Chenopodium albion in -1 light. 



This somew^hat striking result shows that the largest seeds are 

 produced under conditions of light-intensity which are only just sufficient 

 to allow the production of fruit at all, i.e. with the minimum light inten- 

 sity. The case of Cannabis saliva illustrates this point very clearly; 

 the maximum average weight of single achenes was obtained in \ light, 

 as stated above; in i light only two out of 1000 plants produced achenes, 

 to the number of three in all. The average w^eight of these, however, 

 falls only very little short of the average weight of those obtained in 

 large numbers with | light^. The seeds from Combes' harvest were tested 

 for germination; in result it was found that the highest percentages of 

 germination were obtained with seeds developed under light-intensities 

 w^hich were most favourable for the production of large seeds, in other 

 words, the heaviest seeds gave the highest percentages of germination'-. 



It is to be expected that many other environmental conditions 

 acting through the parent plant would greatly affect the size of the seed 



^ The number of fruits per individual plant and tlH> number of seeds per fruit diminisli 

 from the full light to 1/9 light. 



- Similar results have been obtained for twelve different .species of economic and 

 garden plants by Lubimenko (50), who, however, only submitted fruits during their 

 development to different light -intensities. 



Under this heading attention must be drawn to tlie oI)servations of Boorkor ('2) men- 

 tioned above (p 117, footnote). 



