240 S. W. Johnson on the Soil-analyses of the 
of the value of the soil upon analysis. _Some of the analyses . 
agree sufficiently to show that accordant results’ are possible i 
uniform material be taken; but the grand result of the investi- 
gation is that the difficulties of getting a uniform material are 
exceedingly great. Again, we must remember that in the case 
fore us, the three examinations of each soil were made upon 
ortions of one carefully mixed sample. What. would have 
fees the result had each chemist received a sample collected 
net ited from all the others, and from different parts of the 
el 
Dr. Peter mentions these analyses of the Landes Collaevum, 
and quotes a few of the results on page 187 of the 8d Kentucky 
He believes however that these discordant results do 
not invalidate soil analyses when made as they may be made 
with “means and appliances now at the service of the analytical 
chemist” and thinks “this statement however hazardous it may 
seem will be found to be sustained” in his Report. 
In the Report before us however we do not find anything to 
sustain Dr. Peter’s view. He gives, so far as we have discovered, 
no duplicate analyses, to show what accuracy his methods admit 
of on the same sample, much less does he prove by analyses of 
specimens separately gathered from the same field, that it 1s easy 
to procure an average material for analysis. Until this prool 18 
preceded him in the analysis of soils from Davy and Sprengé 
down, evidence that the best endeavors in this line of res¢ 
are entirely incommensurate with the desired results. 
t may Se objected to this criticism of the analyses tnat 
loss or gain must be distributed among the twelve ingredients 
determined. It is true that there is a probability that sucb A 
tribution would be just; but this is by no means certain, Le 
is equally true that this being done there is still force in ™° 
is equ E 
eriticism—for the four-tenths per cent of the soil which 9° 
ee 
Hers 
