-160 STATE BOARD OF AGRICULTURE. 



The' increase from 0.22 to 0.31 and from 0.18 to 0.24 (0.34) is very 

 slight in comparison with the increase of ammonia, which is about four 

 to five times as large after 20 days as it is at two days. The 

 error due to a variation in culture is only (0.31 — 0.22) =0.09 cc. in soils 

 and sands, and 0.06 to 0.16 cc. in solutions, which compares very favor- 

 al>ly with the analytical error, 0.22 and 0.18 cc. respectively. The de- 

 velopment of duplicate pure cultures seems to be quite regular in soils as 

 well as in solutions. It must be remembered, however, that the culture 

 had been grown for many generations in the standard peptone solution, 

 and was more homogenous, perhaps, than the ones employed in most 

 laboratoiw exjx^riments. 



A comparison of the average error of the different series shows a con- 

 tinuous decrease. 



Series 

 Average Error 



Whether this is due to an improvement in the technic or to a more 

 complete acclimatization of the culture cannot be stated. There was a 

 difference of about three months between the inoculation of the first and 

 last series, and during all this time and about two weeks previous to the 

 first inoculation the culture had been grown in peptone solution. 



The average error is also a good test for the even distribution of bac- 

 teria in soils. The average error of the different soil types compares as 

 follows: 



AVERAGE ERROR OF DIFFERENT SOIL TYPES. 



All quartz sands ±0.24 cc. All soils A ±0.23 cc. 



All surface soils ±0.33 cc. All soils B ±0.38 cc. 



All subsoils ±0.20 cc. All peats ±0.25 cc. 



All solutions ±0.22 cc. 



Evidently Soil B gave the greatest difficulties in mixing and in dis- 

 tributing the bacteria, while Soil A and the sands allowed an almost 

 perfect mixture, being nearly equal to that in solutions which must be 

 considered perfect. 



The conclusions from the errors are made a little doubtful, however, 

 by the results of the last series, where the average error of the fine sand 

 cultures is smaller than that of the coarse and medium sand cultures. 

 This is greatly improbable since it took much shaking to distribute the 

 inoculum in the fine sand, while in the coarse and medium sand no diffi- 

 culties were found. 



Variation BeUceen the Different Series. — While the duplicate deter- 

 minations of the same series check quite closely, there is a considerable 

 deviation in the development of the same culture in the different series. 

 This comes partly from the variations of the ammonia in the blank de- 

 Iciiniiiations which is subtracted from the amiuoiiia found in tlie cultures. 

 Hut that alone does not account for the dilTcrence, and it must be as- 

 sumed that other factors, as yet unknown or at least beyond control, 

 leave their trace in the variation of the cultures. The cultures in pep- 

 tone solution and in medium sand carried on in each series give the data 

 for comparison. (Table II.) 



