1923] Variation of Protein Content of Corn 57 



minutes, ten grams of potassium sulphate, C. P. is added. Digestion 

 is continued five minutes after clearing. Time required for digestion, 

 twenty-six to forty minutes. 



In view of statements made in standard books,- that in cereal 

 analysis digestion for four hours is recommended, the longer digestion 

 was employed in a number of experiments. 



In analysis F', given below, the usual method, five minutes diges- 

 tion after clearing, was employed. The average per cent of nitrogen 

 is 1.532^. In analysis F", the digestion was continued three hours 

 after clearing. The result was almost identical with the mean of F'. 

 Again, in analysis G, the mean is 1.572^ nitrogen. C, digested three 

 hours after clearing, gave 1.587% of nitrogen, an increase of only 

 0.01% nitrogen. This confirms our former statement that 26 to 40 

 minutes will suffice to give complete results on nitrogen. 



A method involving so many operations as the Kjeldahl, must give 

 occasion for errors in analysis, yet the variation observed in some of 

 the analyses reported below, seemed too great. Analysis B shows a 

 range of 0.037% of nitrogen, equivalent to 0.23% protein. In E 

 there is a range of 0.032% of nitrogen, and in one set of analyses not 

 included in this paper, the variation amounted to 0.05% of nitrogen. 

 Our blanks also showed a similar puzzling variation. These results, 

 with the blank suggested as the cause of the irregularity, the one fac- 

 tor not constant, namely, the quantity of alkali employed to neutralize 

 the acid in the Kjeldahls. On investigation it was discovered that 

 even C. P. sodium hydroxide contained appreciable amounts of am- 

 monia. Since we used more alkali in neutralizing the acid in the 

 blanks than in neutralizing the acid in the corn samples, the correc- 

 tion factor would be too large. Furthermore, as no effort was made 

 to take the same amount of alkali in the blanks, the results could not 

 be expected to agree. In the later analyses, this was tested by using 

 a constant quantity of 50% alkali in all experiments. Our blanks 

 were much closer, and there was less variation in the per cent of 

 nitrogen in the corn, analyses. Note the close results in A'. For ac- 

 curate work, the analyst should employ equal quantities of alkali both 

 in the blank and in the samples to be determined. 



The ' ' White Plume ' ' analysis reported in paper I, was made from 

 a different sample from that actually planted. A sample of the corn 



■Technical Methods of Analysis: Griffin, page 65. 



