TRANSPIRATION STUDIES. 9 



In general it may be said that the evaporating surfaces, both external and 

 internal, are largest in the plant which has carried on its entire development 

 in the green-house , next in the plant which has been in the green-house for 

 one year after germination, then in the open for six weeks, and least in the 

 adult tree. Of the different parts of the tree examined, the leaflets lose 

 water by evaporation most easily, the rachises next, and the branches least 

 easily. 



TRANSPIRATION STUDIES. 



METHODS. 



The best method yet devised for measuring transpiration is doubtless 

 the determination, by weighing, of losses sustained by sealed potted plants. 

 In the case of Parkinsonia this method was not practical, since the plants 

 raised in the green-house differ from those growing in the open, both in leaf- 

 structure and in the size of stems and leaves. The influence of the two 

 environments on the size of seedlings is shown in plate I B and on the size 

 and structure of the leaves in figure 1. It proved almost impossible to 

 transfer small seedlings from their natural position in the ground to pots, 

 there being only one survival out of some thirty which were transplanted . It 

 thus seemed necessary to use a method for the measurement of water-loss 

 in situ. Cannon's method* is adequate for a rough estimate of comparative 

 water loss from different plants under similar conditions, and seems to have 

 several advantages over Stahl'sf method, yet for the purposes of the present 

 study, which require accurate quantitative measurements of water loss 

 under natural conditions, a more accurate method had to be sought. The 

 older methods, based on the absorption of moisture by means of sulphuric 

 acid or calcium chloride, have many of the same objections as the polymeter 

 method, 1 e., the plant is not under normal conditions of humidity, wind, 

 and sunshine ; moreover, the gain in weight of the absorbing material may 

 not represent the actual amount of water given off by the plant, but is 

 equal to this amount plus or minus the loss or gain of moisture in the air of 

 the bell-jar during the experiment. If the beli-jar is a large one and the 

 part of the plant under the jar is losing a small amount of water, this error 

 may amount to as much as 100 per cent. 



A method was finally worked out which contains some of the features of 

 both Cannon's method and the absorption methods. By a reference to fig. 

 2 it will be seen that the plant or part of the plant to be measured passes 

 through a platform and is covered by a bell-jar, within which are an open 

 dish of calcium chloride, a maximum thermometer, and a dew-point appa- 

 ratus, the last being so arranged that the dew-point within the jar may be 

 taken without any disturbance of the conditions within the jar. 



*Cannon, W. A., A new method of studying the transpiration of plants in place, Bull. 

 Torr. Club, vol. xxxii, p. .51.5. 



fStahl, E., Einige Versuche iiber Transpiration und Assimilation, Bot. Zeit., in, p. 117, 

 1894. 



