452 STATE BOARD OF AGRICULTURE. 



procedure in soil bncteriologv, hocnuse it jrives the baciorinl dovelopment 

 at various time intervals, enipliasizing the earlv period where the most 

 rapid changes take phice. The writer has tried to point out the advisa- 

 bility of this in a paper concerning the usefulness of curves.* It may 

 be stated here that some of the conclusions in the present paper would 

 have been lost entirely if not all stages of bacterial development had 

 received equal attention. 



The pa])er is divided into four chapters : 



I. The extensive experiment covering tlie ammonia formation by 



B. mycoides, in five different soils, three different sands, in so- 

 lutions and soil extracts, and the interpretation of the data. 



II. A more detailed study of the mechanism of the ammonia pro- 

 duction. 



III. Experiments with other aerobic bacteria ; and 

 TV. The requirements of anaerobic organisms. 



I.— Bacillus Mycoides in Soils^ Sands and Solutions. 



Object. — The first and largest experiment shows the ammonia forma- 

 tion from peptone by B. mycoides in various media. Five different soils 

 were used, namely, surface soil and subsoil from field 11 of the Experi- 

 ment Station farm, two soils from adjacent fields, but of a widely dif- 

 ferent fertility, and fresh, brown peat from a swamp north of the college. 



To compare these soils at Ihe same moisture content would be unreason- 

 able. To compare them at % or % of their water capacity, as Stevens 

 did, is far better, yet arbitrary. The best results could be obtained only 

 by testing each soil at various moisture contents and comparing the 

 optimum conditions. This increases the extent of the experiment very 

 much, but it was accepted as the least objectionable policy. 



The amount that could be handled in one series was two soils at five 

 different moisture contents and a quartz sand check, also at five different 

 moisture contents, making a total of 30 distillations for each determina- 

 tion on account of the duplicates. All soils have also been tried with, 

 and without "fertilizer," i. e., potassium phosphate and calcium 

 chloride. 



Method. — The soils were air-dried, and the moisture content of the air- 

 dry soils was determined by drying at 105°. The sands were heated with 

 diluted hydrochloric acid, washed several times, first with tap water, 

 then with distilled water, and then dried at 1G0° C. 



The amount of soil corresponding to 50 g. of oven-dry soil was calcu 

 lated and the corresponding weights were filled into liter flasks, 12 flasks 

 for each moisture content of each soil to allow six duplicate determina 

 tions at different limes. Since the ex]>eriment was to be cai'ricd on with 

 strictly pure cijltures, it seemed not advisable to keep all the soil of one 

 kind in one container taking a saniple of it whenever a determina- 

 tion was made, as this is generally done with non-sterilized soils. It 

 would be almost impossible to avoid a conlamiuntion. It became, there- 

 fore, necessary to weigh out, sterilize and inoculate seiKiratcly each lot 

 used for a single dcloiiiiination. This comjilicatcd the experiment very 

 much, since it required for dui)licate determinations on six days for each 



♦Technical Bui. 6. (1910), Mich. Agr. College Exp. Station. 



