396 BOTANICAL GAZETTE [December 



former group were only about 25 per cent less than in the latter, 

 and concluded "that xerophytic leaf structure is not always incom- 

 patible with abundant transpiration, but sometimes exists only for 

 use in emergencies to protect the plant from injurious loss of 

 water." 



Hanson (9) has shown that the major water losses from the 

 crowns of isolated trees of Ulmus americana, Acer saccharinum, 

 and Fraxinus lanceolata occur from the peripheral branches, and 

 less than one-sixth from equal areas of leaf surface on shaded 

 branches. 



The present investigation was undertaken with a.double purpose 

 of obtaining data on the relative losses in summer and winter of 

 conifers and broad-leaves, and also to make a beginning on the 

 problem of winter killing of trees and shrubs. This latter project 

 is not included in this paper. 



Methods. 



In the spring of 19 18 seedling conifers of Pinus ponder osa 

 Dougl. and Pinus Banksiana Lamb, were obtained from the forest 

 nursery at Halsey, Nebraska; while those of Ahies grandis Lindl., 

 Pinus Murrayana Balf., Picea Engelmanni (Parry) Engelm., and 

 Pseudotsuga mucronata (Raf .) Sudw. were secured from the national 

 forests of northern Idaho. These seedlings, varying from two to 

 four years in age, were potted during May in 5-, 7-, and 8-inch pots 

 respectively, according to the size of the plants and the demands of 

 the root systems, and in soil consisting of two parts of rich garden 

 loam and one part of sand, thoroughly mixed and screened through 

 a one-fourth-inch-mesh sieve. The pots were placed on the lawn 

 near the greenhouse until needed in late summer, and were thor- 

 oughly watered every day and sometimes twice a day during the 

 driest periods. A few trees, mostly white fir and Engelmann 

 spruce, died. These, with numerous weaker individuals of other 

 species, were discarded, and only the very best plants, which showed 

 the most flourishing condition of growth, were used in the experi- 

 ments; in fact, only about half of the original stock was thus 

 selected and used. 



Galvanized iron containers with flat bottoms and straight walls 

 were used in the transpiration work, the size varying according to 



