24 



SCIENCE 



[Vol. LVI, No. 1436 



aggregates of soil material and then to go through 

 the operations by which the different sizes are 

 separated. The various methods employed have 

 used various means for deflooculation of the soil 

 material, but the method adopted as most satisfac- 

 tory by the Bureau of Soils has been that of sha- 

 king the sample of soil with water for a period of 

 about seven hours. Experiments recently carried on 

 have demonstrated that the colloid content of soils 

 is considerably greater than has been generally 

 believed. These experiments have shown that the 

 so-called clay separate obtained in the mechanical 

 analysis, consisting of material grains smaller 

 than .005 of a millimeter in diameter, is made up 

 partly of aggregates of colloidal material. These 

 aggregates have withstood the shaking operations 

 without breaking down, but it has been demon- 

 strated that they can be broken down and ob- 

 tained in a dispersed state hj rubbing lightly with 

 the finger or rubber pestle. These results indicate 

 the insufficiency of our present methods of me- 

 chanical analysis and show that in addition to the 

 separations as obtained at present this method 

 may have to be supplemented by a determination 

 of the amount of colloid contained in the soil and 

 the main separate groups, such as the clay and 

 silt. 



Studies on flavors, leverages and related prod- 

 ucts. 1. The determination of methyl anthran- 

 ilate:^ J. "W. Sale and John B. Wilson. A col- 

 orimetric method for the determination of methyl 

 anthranilate in genuine and imitation grape prod- 

 ucts which depends upon the formation of a red 

 azo dye, is described. The reagent used is 

 sodiuni-l-naphthol-2-sulphonate. Experimental data 

 are given which show that the method is quantita- 

 tive. Advantages of the method described over 

 others in current use are: first, the method is 

 quantitative; second, the test is applied directly 

 to distillate, thus avoiding possible loss of ester 

 through extraction and subsequent evaporation of 

 solvent; third, use of hydrazine sulphate in place 

 of urea for destroying excess of nitrous acid. 

 The method is recommended in the examination 

 of products which are believed to be sophisticated. 



Studies on flavors, h ever ages and related prod- 

 ucts. 2. Determination of methyl alcohol in ex- 

 tracts:- John B. Wilson and J. W. Sale. The 

 merits of a number of well-known tests for methyl 



1 Contribution from Water and Beverage Lab- 

 oratory, Bureau of Chemistry. 



- Contribution from Water and Beverage Lab- 

 oratory, Bureau of Chemistry. 



alcohol are discussed and data are given which 

 show their relative value and delicacy. The 

 Denigee and Lyons ' tests are found to be the most 

 satisfactory for examination of flavoring extracts 

 suspected of containing methyl alcohol. Flavor- 

 ing extracts contain a variety of esters and essen- 

 tial oils which interfere unless the analysis is con- 

 ducted properly. The sample to be analyzed 

 should be adjusted so that it will have a volume 

 of 100 cc and contain not more than ten per cent, 

 of ethyl alcohol. Interfering substances are elim- 

 inated by salting the sample, extracting it with 

 petroleum ether previous to distillation, and dis- 

 tilling, using a fractionating column. Thirty cc 

 of distillate are collected in which will be found 

 practically all the ethyl and methyl alcohols. The 

 colorimetric tests to be applied are described in 

 detail. 



JKeZatJons between the active acidity and the 

 lime-requirement of soils: Edgar T. Wherrt. 

 Lime requirement is stated in parts per thousand 

 of CaO, and, because of the ease with which rela- 

 tive values can be appreciated, active soil acidity 

 is stated in the form of specific acidity. The 

 ratio between these in a given soil may be ex- 

 pressed by a correlation coefficient C, obtained by 

 the equation: L.R. = C X (S-^- — !)• The value 

 of C is believed to be a measure of the adsorptive 

 power of the soil colloids for hydrogen-ion. The 

 coefficient C has been found to vary so widely 

 from one soil to another, from an untreated to a 

 limed soil, and even from one depth to another in 

 the same soil, that it is impracticable to calculate 

 lime-requirement from acidity determinations in 

 general, as has been proposed. Soils may be 

 roughly classified on the basis of the value of C, 

 a convenient ratio between classes being 3/10; 

 but only if some simple procedure is first devised 

 for classifying a given soil can there be obtained 

 from its specific acidity a value for its lime- 

 requirement. 



Characteristic proteins in high and low protein 

 corn: M. F. Showalter and E. H. Carr. The 

 protein content of corn is subject to wide varia- 

 tions by breeding and selection. The highest 

 protein ear the writers have been able to produce 

 contained 18.43 and the lowest 7.62 per cent, pro- 

 tein. A study has been made of the relative 

 abundance of the different proteins in corn, and 

 it was found that zein was the important protein 

 which varied most, averaging 50.28 per cent, in 

 high protein and only 31.85 per cent, in low pro- 

 tein corn. The averages for the other proteins 

 are: gluteUn 38.11, gloluUn 3.70, albumen 3.92, 



