30 BULLETIN 1460, U. S. DEPARTMENT OF AGRICULTURE 
and digest for one hour. Cool, dilute with 200 cubic centimeters of 
distilled water, add a few pieces of granulated zinc or pumice stone. 
Next add sufficient sodium hydroxide and sodium thiosulphate or 
sodium sulphide solution to make the solution strongly alkaline, and 
to precipitate the mercury, pouring the alkali clown the side of the 
flask so that it does not immediately mix with the acid solution. 
Fifty cubic centimeters of solution is usually enough. 
Connect the flask with the condenser, mix the contents by shaking 
and distill until all the ammonia has passed over into a measured 
quantity of standard acid. The first 150 cubic centimeters of the dis- 
tillate will generally contain all the ammonia. Titrate with the stand- 
ard alkali. Calculate, first into terms of nitrogen, later into protein 
by multiplying the percentage of nitrogen found by 5.7. 
Reagents should be tested for absence of nitrogen by making a 
blank with sugar. The sugar partially reduces any nitrates present 
that otherwise might escape notice. Deduct any nitrogen found in 
the reagents from that found under the conditions of the test. 
SPECIAL POINTS FOR CONSIDERATION 
In making protein tests it appears that efficiency, speed, and inex- 
pensiveness are the most desired essentials. To attain speed an effi- 
cient source of heat is necessary. Lacking an efficient source of heat 
speed can not be obtained,, as catalytic agents offer only a limited 
degree of help in the digestion process. Heat controls the time a 
sample should be digested. 
The period for digestion must necessarily vary according to the 
heat intensity. With electric heaters, 600 watts capacity, a digestion 
period of from 40 to 45 minutes is sufficient. Heaters of 550 watts 
capacity develop sufficient heat to digest the sample completely in 50 
to 55 minutes. One hour digestion period is necessary when using 
heaters of 440 watts capacity. The best heat is developed from gas 
heaters when they are adjusted so as to evaporate 150 cubic centi- 
meters of water from 200 cubic centimeters of water in 20 minutes. 
The boiling should take place in a 550-cubic-centimeter Kjeldahl 
flask. 
The present practice of sampling as described should be carefully 
followed. Having obtained the sample, the matter of keeping 
its identity can not be overemphasized. If this point is not heeded, 
it is of little use to make further tests. 
As the protein test is usually given along with the numerical grade 
of sample of wheat, samples for analysis should be cleaned in the 
manner described. 
Practices which lead to the drying out of the grain should not be 
used, such as taking samples for test after they have been on the 
trading floor all clay subject to drying, handling, and contamination, 
and the placing of moist wheat in paper sacks preparatory to 
grinding. 
Great care should be taken to retain the moisture in the sample, 
as changes in moisture induce changes in the protein test results. 
Differences of 0.6 per cent in protein content have been noted in 
these investigations caused by changes in moisture content. 
Not less than 30 grams of wheat should be ground for protein 
tests to obtain results representative of the bulk sample. Errors 
