31 



Wid© variations occur in th« amount of water required 

 to produce a given weight of «py matter in different plants. 

 In the two columns where corn oceurs, less water is required 

 than for any other plants given. Millet occurs only in 

 Wollny's results, and requires 74.25 % more water than corn. 

 Peas, on which data is given in the table, shows 235, 292, 

 477, and 477 parts, respectively for one part dry matfeer. 

 The figures given by Lawes and Gilbert show a greater 

 economy of water than the others. Is this eeonoay due to 

 the moist climate of England, to the manner ef conducting 

 the experiment, or to some other cause ? The results from 

 the German investigators dif'fer widely. 



King (2u. Soils, 26) conducted experiments on trans- 

 piration from the corn plant on four different soils. The 

 soil was placed in barrels and a constant supply of water 

 maintained one foot from the surface. Evaporation from 

 the soil surface was included in the transpiration results. 

 When the eorn made a very poor growth, the proportion of 

 water required was about three times as great as whea the 

 growth was good. Manifestly, where the evaporation is 

 included in the results, the data is untrustworthy as showing 

 the transpiration. Yet the data hag practical value and 

 supports King's general observation, that where the trans- 



