~ 153 - 



tliG pore space here tabulated a few suggostions are rrade . 

 It apoears that getting total weight and deterir.ining mois- 

 ture and specific gravity, and rrom this data calculating 

 the pore space, is more dependable than the direct method 

 by adding w.n.ter. Only ore cylinder is needed thus ob- 

 viating errors due to lack of exact BeaaureiEents . .--ince 

 the variability is sniall it is essential that errors be 

 elininated as far as possible. CompoGlte cylinders of 

 soil should be v/eighed in the field and th© sample for 

 moisture and specific gravity determination taken froro 

 the whole of the samples after thoroughly mixing. Tf striking 

 variation occurs probably this should be eliminated, as 

 any coir.pacting factor, such as the step df a horse, for 

 instance, would cause a sample not to be comparable 5 If 

 this elimination Is not resorted to a larger number of sam- 

 ples should be taken from the plats when much variation 

 occurs . 



— RATE OF EVA.J^PORATION MD LOSS FHO]^ DRAINAGE — 



Soil from samples taken for moisture determi- 

 nation Nov. 20th. (Table "B" ) was used. Platinum dishes 

 were filled, packed by jarring uniformly, and struck. 

 The samples were placed in a roon fe-jr feet from the 

 radiator. The results follow in table "E": 



