EFFKf'T OK CLIMATIC CONDITIONS. 



15 



collou'''' Itiiildiiij^s. 'V\n> door oi)ciis into :i hall which coiniuuiiiciites 

 with the ottices, choiuical I:iborat()ry, and the hasoniont. At Ann 

 Arhoi- the seeds were stored in the 1)otanieal hd)oratory, with sliohtlr 

 var3nng conditions, they being near a window which was frecjuently 

 open durinj^ the summer, and at irreoular intervals durino- the early 

 part of the suuuuer the i)acka«res were i)laced in the window so as to 

 receive the direct rays of the sun. The seeds stored at Ann Arbor 

 served partially as controls for those sent to the various other places, 

 and, in addition to the last-named series, seeds from the ori«;inal 

 packaj^es, as received from D. M. Ferry & Co., were kept in a dry 

 and comparatively cool closet on the fourth floor of the botanical lab- 

 oratory. These seeds served as checks for the complete set of exper- 

 iments, and are desionated throughout this paper as "Control." 



The samples were sent out to the al)Ove-named places in February, 

 1!H)0. The tirst comi)lete set was returned in June, or early July, of 

 that year. The second complete set was allowed to remain throughout 

 the entire summer, and was returned in Octol)er and early November 

 of the same year. The average time of treatment for the two series 

 of experiments was 128 and 251 days respectively. When the seeds 

 were returned, germination tests were made as soon as possible. The 

 length of time that the seeds were in the various places and the vitality 

 as shoAvn by the germination tests are given in Tables I and IT. In 

 both tables the columns from left to right, beginning with Mobile, 

 Ala., are in the order of the degree to which the seeds were injured. 



Table I.— Effect of climate on dtalitij, (ts shown by licrcentage of germination— first test. 



From Table I it will be seen that the loss of vitality in the case of 

 seeds stored at Mobile was much greater than in those stored at any 

 of the other places. The greatest loss in the samples tested was in the 



