22 



IQ2I — The correlation of temperature aiid flowering is obvious 

 for this season. Sometimes the gap between tlie 1921 date and 

 the nearest one to it is as great as three weeks as in the case of 

 No. 21. This flower is not uncommon here and the discrepancy 

 can not be attributed to chance. 



Even in the case of a few^ species that flowered earUer in an- 

 other year, in some cases at least that fact is correlated witli 

 a lowering of 192 1 temperature. Nos. 55, 58, 59, 60, 67-72 can 

 all be explained either by the absence of a 192 1 record or by 

 the fall of 1 92 1 temperature to normal or below for the period 

 Alay 6-19. 



