667 



should be kept of the temperature and of the amouTit of moisture in the 

 soil of these different formations, and of tlic loss of heat from radiation 

 and evaporation. The observations should be taken in connection with 

 the ordinary meteorological observations of an observing station. The 

 work should be comparative so that deductions may be made from the 

 observations. It is not sufficient, for example, that observations be 

 taken in a soil well suited to tobacco, but they should also be taken in 

 a soil near by, where the conditions are'known to be unfavorable for 

 the growth or proper develo]>ment of the tobacco plant. 



A method is needed for the determination of the juoisture in the soil 

 without removing the sample from the field, as will be referred to later, 

 and the ordinary form of soil thermometers needs to be im]>roved n])(m. 

 These records should show the actual conditions of moisture and heat 

 within tlie soil, for, as we see in greenhouse culture, these conditions 

 largely determine the development of plants. Sut^h observations luive 

 shown plainly the cause of the local distribution of cotton and tobacco in 

 some of the Southern States. 



The relation of soils to heat is so dependent upon the moisture of the 

 soil that it is difficult to give a true interpretation of soil tein])eratures 

 or to show their relation to plant developnu^nt unless accompanied with 

 a record of the soil moisture. The relation of soils to water should be 

 carefully studied in the laboratory and in pot experiments. Samples 

 of the soils should be taken with care for laboratory work, and for pot 

 experiments to confirm the laboratory results. 



The laboratory work should be based upon the mechanical analysis 

 of the soil, as this shonld show the texture of the soil. The method of 

 mechanical analysis should be further perfected so that more separa- 

 tions could be conveniently made, especially of the very smallest- sized 

 grains of sand and clay. The clay grou]) has an important value, due to 

 the extremely small size of the grains. There should be a uniform scale 

 of separations for the comparison of soils from different localities. 



Tlie relation of soils to water resolves itself into two lines of investi- 

 gation, the forces wliich move the water and the conditions which 

 determine the relative rate of flow. 



Tlie forces which move the water within the soil are gravity and the 

 tension or contracting poAver of the exposed water surface. The 

 approximate extent of the water surface can be calculated from the 

 mechaiucal analysis of the soil. The surface tension and the eflect of 

 manures and fertilizers on the surface tension can be found by the ordi- 

 nary metluxl of tlie rise of li(]uids in capillary tubes, using as a solvent 

 pure water or extracts of the soil, rex>resenting as nearly as possible the 

 ordinary soil moisture. What little data we have show that the diflerent 

 fertilizing materials have a very marked eff'ect on this pulling ]>ower of 

 water. The same class of substances may differ widely in their eftect. 

 Kainit, for instance, increases the surface tension of pure water, but 



