MASS. EXPERIMENT STATION BULLETIN 433 

 Figure 1 



Bureau of Agricultural Economics, to wh'ich are added rough estimates of the 

 more important losses by frost, winterkill, hail, fire, and storms, these being 

 little better than mere guesses in some cases.* Moreover, lack of data made it 

 impossible to allow for variations in losses due to insect injury from 3'ear to 

 year. It can hardly be doubted that these variations have been very considerable, 

 especially in the case of fruit worm losses, and may not have been closely related 

 to the weather conditions studied. It was necessary also to neglect the adequacy 

 of flower pollination except as this may have been governed by weather. Re- 

 gardless of these lacks, however, it is believed that the long and reliable records 

 of actual cranberry production in three substantial areas rather widely separated 

 geographically and climatologically have presented a unique chance to learn 

 something about the relations of weather to the production of fruit by a perennial. 

 For convenience the different weather elements were handled by months in 

 these studies. It is, of course, not to be supposed that they are naturally con- 

 fined in this way and the limits of their influence are in many cases hard to de- 

 termine precisely. It is believed that the method followed is good enough for 

 practical purposes. 



*The sources of information about these losses are the same as those given on page 25 of Bui. 

 402 mentioned above. 



