1905] LIVINGSTON—TRANSPIRATION AND GROWTH 195 
ological properties of the leaves which control the amount of 
water lost per unit of leaf surface. The water loss per unit area of 
leaves is practically uniform throughout the different treatments; 
therefore the variations in total transpiration exhibited are due not 
to any difference in structure or activity of the leaves, but simply to 
the differences in extent of leaf surface developed. 
In making use of this criterion of transpiration for the comparison 
of different nutrient media, it must be borne in mind that, as in all 
other biological experimentation, there will occur unexplained vari- 
ations, and the truth must be attained by the summation of the 
tesults of a number of experiments. Many of the unexplained dis- 
crepancies of the experiments just described might not have occurred 
had the number of plants used been larger. The individual varia- 
tion among wheat plants is found to be great, so that the larger the 
number of plants used the nearer would the result approach the true 
average. 
Also, it must be remembered that if total transpiration is 
decreased by temperature, atmospheric conditions, etc., the differ- 
ént members of a series will approach each other in the amount of 
water lost; were transpiration checked completely, all the members 
would agree. It is thus necessary to have good conditions for evap- 
oration from the leaves where such experiments are carried on, in 
: order to magnify the differences in transpiration and keep them well 
above the limits of experimental error. 
The method for comparison of plant growth here provisionally 
established for wheat is found also to hold for the grasses generally. 
This doubtless rests on the fact that the leaves of these plants are 
far and of continuous basal growth. Whether or not it can be 
adapted to other groups of plants is not yet determined. 
THE UNIVERSITY oF CHICAGO. 
