420 BOTANICAL GAZETTE [December 



The extra soil necessary to fill these containers was in all cases 

 a mixture of two parts of garden loam and one part of fine gravel. 

 It had a wilting coefficient of 10.7 per cent, and was kept rather 

 uniformly at a water content of about 20 per cent. The methods 

 used throughout were the same as those already described. Original 

 weighings were not taken until after the trees had been in the 

 containers two weeks. At this time rabbits attacked the deciduous 

 trees and injured all but five of them. 



On July 20 the remaining trees in their containers, 19 in number, 

 were weighed to the nearest half gram. As during the preceding 

 interval, they were then placed in a row in the soil and in such a 

 position that all were shaded for a few hours in the afternoon. 



TABLE X 



Average transpiration losses per unit area from 

 several species of tree seedlings 



At the end of 26 days, on August 15, they were again weighed, the 

 leaf areas determined, and the losses calculated as in the preceding 

 experiments. These data are shown in table X. 



These data show that the losses from the various species of 

 conifers, growing in the open but under conditions where all were 

 shaded for a portion of the day, are very similar, and only about 

 half as great as those of the broad-leaved species.^ The midsummer 

 average daily loss per square decimeter from the yellow pines and 

 Engelmann spruce at the Alpine Laboratory, when compared with 

 that of the same species during the autumn at Lincoln, was found 

 to be only about half as great at the former station. Because of 

 the limited number of plants used, however, these data should be 

 considered indicative rather than conclusive. 



