508 
Journal of Agricultural Research 
Vol. XXXI, No. 6 
Soil-extract agar again shows its superiority, while the mannite- 
salts agar, although a fairly good substrate for some of the soil 
organisms, inhibits the development of a large group of bacteria ( 2 ) 
The counts on egg albumen were consistently lower than on soil- 
extract agar, as Waksman ( 18 ) also observed. His results show 
that soil-extract agar gave the highest count, casein agar came next, 
and this was followed by egg-albumen agar. If the counts obtained 
by him on soil-extract agar are recorded as 100, those on casein and 
egg-albumen agar equal 93 and 73 per cent, respectively. Despite 
this decided advantage in favor of soil extract, it was recommended 
that egg-albumen agar be adopted for the determination of the total 
counts of the microorganisms in soil, and that soil-extract agar should 
be discarded because of its being “not standard in composition.” 
Others have raised the objection that soil extract is of unknown 
composition, and that its variability would make it impossible to 
compare results obtained by workers experimenting with different 
soils. To test the validity of this objection, the following experiments 
were made. 
COMPARISON OF DIFFERENT SOIL-EXTRACT AGARS 
Ten soils, ranging from a very rich muck to a loamy sand, were 
selected for preparing the extracts. It was realized at the outset 
that these soils presented greater differences in character than would 
be found in field soils likely to be used for making extracts. But 
since so much weight has been attached to the assumed variability of 
soil extracts, these writers attempted to cover wide range in one 
experiment. The soils used were the following, with name and 
source of sample, and with the soil-survey party’s characterization 
when the sample was taken. 
Ontario, N. Y_...Very rich. 
Do. 
Clay loam, sand, and compost. 
Clay loam. 
Dark, heavy, poorly drained. 
Rich bottom land. 
Muck.... 
Peat loam....Walliston, Va 
Greenhouse soil.—...Washington, D. C 
Rich lawn.....do. 
Granby loam___Sodus, N. Y. 
Genesee silty clay loam...Lyons, N. Y- 
Silty clay loam...Arlington, Va... Field soil, low in nitrogen. 
Hollis loam...Ludlow, Vt—.. Light loam, fairly fertile. 
Dunkirk fine sandy loam__Ontario, N. Y_. Light, low in humus. 
Coloma loamy sand .—. Taunton, Mass___Very poor. 
The samples were all taken from cultivated fields, and air-dried. 
Extracts were made in a uniform manner, 500 gms. soil being boiled 
about 15 minutes over an open flame with 1,200 c. c. water, filtered 
through paper, and the volume made up to 1,000 c. c. Agar 2 per 
cent and K 2 HP0 4 , 0.05 per cent, were added to the extracts, and 
the reaction was adjusted to approximately Ph 7.0. 
Table V shows the results of plating 5 representative soil samples 
on soil-extract agars made from these 10 soil types as well as on 
Thornton’s mannite-salts agar and on egg-albumen agar. Samples 
1 and 2 were taken from a lawn (squares 25 and 26, Table IV), 
sample 3 came from a field at Arlington Experiment Farm, Rossyln, 
Va., samples 4 and 5 from two rows of pots (Nos. 34 and 35) in a 
greenhouse experiment, A dilution of 1:250,000 was used in plating 
samples 1, 2, and 3, but 1: 750,000 with 4 and 5. 
