THE NATURE OF TRANSPIRATION 



13 



in errors, which render the method unsuited to exact 

 observation. 



In order to eliminate these sources of inexactness, 

 recourse was made to the arrangements shown in the 

 figure. The results obtained by this method are displayed 

 in the following table. The difficulties of keeping the flow l 

 of gas exactly constant through the apparatus, and other 

 experimental errors, lead to variations between the indi- 

 vidual observations often amounting to 10 per cent. ; but 

 by multiplying these observations, an approximation to 

 the actual alteration in the rate of transpiration has been 

 obtained. The numbers here given are the means of a 

 large number of observations. In each case, the branches 

 experimented on were from a bush of Syringa vulgaris, 

 except in the experiments where effect of ether vapour 

 was observed. In these, branches of Cytisus laburnum 

 were used. 



If the amount transpired in air be taken as 100, the 

 amounts transpired in the other gases are as follows. 

 These figures may be said to denote the specific transpiration 

 for the gases : 



Table 1. 



The first source of error affecting these experiments, 

 and one which it seems hard to eliminate, arises from the 

 fact that the effects of the different gases may be more or 

 less rapid. Thus it is very certain that the light gases 



1 It is to be observed the rate of flow of the different gases will be 

 different, even if the pressure be the same. 



