22 



THE PHYSIOLOGY OF STOMATA. 



was only 3.6 per cent in the most extreme case. The data in table 14 are 

 by way of comparison, not without interest. 



TABLE 14. Experiment 19. Covillea mexicana, June 19. 



In the last instance the positive error was 3.7 per cent. 



The figures displayed in tables 13 and 14 indicate that the error in oco- 

 tillo is higher in general than in the other plants tried, a difference which is 

 referable, it is not unlikely, to the absorbing power of the water-storage 

 tissue, which designation may be applied to indicate the function, in part, 

 of the cortex, or wholly, perhaps, the function of certain clear islands of 

 tissue within the cortex. 



Evidence of this kind, it must be said, does not prove what another plant 

 or piece will do, so that it would remain an untrustworthy method for certain 

 purposes, including my own (and must remain so in the case of quantitative 

 problems involving comparisons of root activity, absolute transpiration of 

 plants, etc.), were it not for the system of controls afforded by a good number 

 of the same kind of experiments carried on synchronously. This condition, 

 in potometer work, must not be overlooked. If from 6 to 10 experiments, 

 running side by side, give identical results, the degree of probability that their 

 behavior is normal is very high. Especially when sudden changes in the 

 environment are followed by sudden changes in the rate of transpiration may 

 importance be attached to the data. Thus, if the rate falls immediately on 

 the plant being placed in darkness, this fall is due to the change, and not to 

 irregularities of flow in the stem. The truth of this will appear clearer upon 

 the inspection of the curves presented in Part I. 



STOMATAL ACTION. 



For the study of stomatal action none of the means heretofore used appeared 

 adequate. To assume the condition of the stomatal openings from the 

 behavior of either the physical or chemical indicators of water-vapor is to 

 assume what is to be proved, if it be the case, namely, that there is a con- 

 stant relation between the size and number of the openings and the rate of 

 transpiration (Burgerstein, 1904, p. 32), and upon this point I take issue with 

 Burgerstein. Merget's statement, made as a result of the palladium chlorid 



