1019] SOILS — FERTILIZERS. 215 



Tests of commercial cultures for legume inoculation, H. A. Noyes and 0. O. 

 Obomkb (Roil Sol., i; (1918), No. /, pp. 89-79, figs. t). Investigations of legume 

 inoculation made at the Indiana Experiment Station are described. Pot ex- 

 periments Included a comparison of soil obtained from fields growing the re- 

 spective legumes and four commercial cultures applied to each seed, planted In 

 either its exad proportion of the commercial culture, or its proportion of 

 bacteria in the weight of soil used per acre, together with a study of the 

 inoculating ability of 1 lb. and of 0.5 lb. quantities of soil per acre. 



In addition greenhouse plat tests were undertaken to study the effect of 

 fertilization upon the percentage of inoculation obtained with a specific culture. 

 Soy beans, sweet clover, cowpeas, and hairy vetch were employed in the pot 

 experiments, and 3 per cent hydrogen peroxid was used as a sterilizing agent 

 for the seed. Air-dry Wabash sandy loam soil was placed in greenhouse pots 

 8 in. deep and S in. in diameter, and both the soil and pots were sterilized by 

 dry heat. Nine seeds were planted in each pot and the plants thinned to three 

 per pot. Sterile, distilled water was used in watering the pots until the plants 

 were harvested seven weeks later. All treatments were made in triplicate. 

 Inoculated and uninoculated soy bean seeds were seeded in a bank sand and in 

 a brown silty loam soil in greenhouse plats without previous sterilization. The 

 plats had been twice cropped to lettuce and fertilized seven months and again 

 four months previously with various combinations of acid phosphate, sodium 

 nitrate, potassium chlorld, and manure. 



Both soil and commercial cultures gave 100 per cent inoculation with sweet 

 clover in the pot exp< riments, while in the case of the other legumes all treat- 

 ments failed to give satisfactory inoculation. Since both quantities of soil 

 failed to produce inoculation with three of the legumes the double quantity 

 could not be regarded as superior to the single quantity. A commercial culture 

 applied to soy bean seeds at a double rate produced an average inoculation of 

 75 per cent, as compared with 20 per cent for plants sown to uninoculated seed. 

 Sodium nitrate tended to reduce the percentage of inoculation secured. 



It is concluded that "larger quantities of commercial cultures and soil than 

 those used in these tests would be necessary to furnish satisfactory inocula- 

 tion over the entire area for which the culture was put up." 



Report on the examination of commercial cultures of legume-infecting bac- 

 teria, C. R. Feij.ers {Soil Sci., 6 (1918), No. 1, pp. 53-67).— The results of an 

 examination of about 30 official samples and of some 20 unofficial samples of 

 commercial legume cultures made at the New Jersey Experiment Stations are 

 noted. The number of organisms contained in the culture was determined, 

 as well as the purity of the cultures and their efficiency in nodule production. 



Only two cultures were classed as "poor," indicating that less than one 

 nodule per plant was produced, and four as "partly poor." Two of the latter 

 were classed as "good" in all tests except those with the pea bean. In most 

 cases the purity and general condition of the culture was found to be very 

 good. Soy beans appeared to be harder to inoculate than most of the common 

 legumes, many of the cultures failing to give satisfactory results with this 

 plant. The soil-transfer method of inoculation is recommended for soy beans 

 unless the commercial cultures are known to be of good quality. 



Soil or muck cultures are said to be excellent carriers of legume bacteria. 

 The plate method of testing pure cultures gave a good indication of the infect- 

 ing ability of the organisms, although it is thought best to verify this test by 

 growing the plants and examining the roots for nodules. The establishment of 

 a standard for a lower limit of " bacteria per acre-size sample " is not deemed 

 justified at the present time, and possibly not at all because of the variability in 

 the physiological efficiency of the organisms themselves. 



