Apr. a8. X923 Toxicity and A ntagonism of A Ikali Salts in Soil 321 
same as that of the carbonate, but there were 10 different quantities 
of each individual sulphate in the triangular grouping. In addition, 
there were 12 extra tumblers containing the four added substances 
alone, at their maximum concentrations. A '‘trial’' consisted of 10 
of these treatments in which the quantity of carbonate and sulphates 
added ranged from 1,000 to 10,000 parts per million. Three trials 
were first carried out with these materials in sand, Greenville loam, 
and West Logan clay loam. Then two more trials were made 
with Greenville loam and the same sulphates but with the substitution of 
sodium chlorid in one case and sodium nitrate in the other in place 
of sodium carbonate. In the last two trials the concentrations of the 
salts and also the sulphates ranged from 500 to 5,000 parts per million. 
Mechanical and partial chemical analyses of the soils used in these 
experiments have already been published by one of the writers {13). 
All the soils contained a large quantity of calcium carbonate. The 
sand was coarse and low in organic matter. The two heavier soils 
were nearly identical in texture, but the Greenville soil contained a 
little more organic matter. 
A critical study of the results of this preliminary test has shown that 
the experimental error in the individual tumblers may frequently be 
large enough to destroy the regularity of the variations due to the grad¬ 
ual interchanging of the three sulphates so that a presentation of the 
data in full detail would be confusing. The triangular formation has 
therefore been divided into seven regions, namely, 3 ie center, the three 
comers, and the middles of the three sides. The value for each region 
has been found by averaging three to six tumblers. It is fully realized 
that no two tumblers had exactly the same “alkali” treatment, and, 
accordingly, this mode of presentation may not seem strictly justifiable, 
but since it has been observed that the average for three glasses in a given 
area is usually rather close to the average of six or more tumblers at that 
place, it is felt that this objection is of minor importance. The results 
are reported on the basis of the average dry weight per plant. The 
use of this basis neglects the fact that the percentage of germination in 
the higher concentrations of alkali is very much reduced and therefore 
the curves do not fall as rapidly as they would on the basis of total 
dry-weight production. The results seem quite comparable, however, 
and, as the relations in the higher concentrations are thus more clearly 
brought out, it has seemed best to adopt this basis. The data are given 
graphically in figures 2 to 6, each of which consists of four charts. Chart 
A shows the results of the sulphuric acid treatment with and without 
sodium carbonate, while B and C give the corresponding data for calcium 
sulphate and sodium sulphate, respectively. Finally, the two and three 
component mixtures added to the carbonate soil are shown in Chart D. 
The data for the untreated carbonate soil is reproduced in all the charts 
as a heavy unbroken line. 
The results of the experiment with sand are given in figure 2. The 
beneficial action of calcium sulphate (fig. 2, B) and sulphuric acid (fig. 
2, A) on the carbonate soil is forcibly shown. The acid is somewhat 
more efficacious than the calcium sulphate, as would be expected from 
the fact that the gypsum is too insoluble to be added in solution and hence 
would react more slowly. Since the sand contained a large quantity 
of calcium carbonate, some calcium sulphate was doubtless formed on 
the addition of the acid, and it therefore seems likely that both correc- 
