528 
Journal of Agricultural Research 
Vol. XXVIII, No. 6 
reduction with Devarda’s alloy (50 parts copper, 45 parts aluminum, and 5 parts 
zinc), conducted in an alkaline solution and now known as the Devarda alloy 
method (7, p. 456). The last procedure with some modifications has been found 
to be very satisfactory by many workers, and it is now becoming more generally 
used. Further work, together with minor improvements, will probably extend 
its application. 
EXPERIMENTAL WORK 
The writer, having occasion to make a large number of nitrate determinations 
in cured tobacco, desired to find a satisfactory method for this purpose. Previous 
experience had shown that if a distillation is made of tobacco with water and 
magnesium oxid according to the official procedure for ammoniacal nitrogen in 
fertilizers (1 , p. 10), ammonia is slowly evolved from the nitrogenous compounds 
present and comes over long after a sufficient volume of distillate has been col¬ 
lected to account for any ammonia compound that may have been present 
originally. 
The writer was most impressed with Strowd's method for the determination of 
nitrites and nitrates in plants (£) which is a modification of the Devarda alloy 
method. Strowd found that his procedure was more satisfactory for this purpose 
than any other method tried. One good point in this procedure is that it in¬ 
cludes a control determination with each sample, without the Devarda alloy. 
Strowd’s procedure was first tried on some aqueous extracts of cured tobacco 
leaf and found that with some samples it was not always possible to get satisfac¬ 
tory duplicates. It was thought possible that the prescribed conditions do not 
permit the hydrogen to completely reduce the nitrate during the preliminary 
heating and subsequent distillation. The modification described below was after¬ 
wards found to usually give higher and more concordant results than Strowd's 
method. The procedure used in this work and subsequently referred to as the 
modified method is as follows: 
THE MODIFIED STROWD METHOD 
Dilute two equal filtered portions of a cold-water extract of the sample to 
250-350 cc. in 1-liter Kjeldahl flasks. Add a small piece of paraffin, a few small 
glass beads or pieces of pumice stone and 2.5-3 gm. of sodium hydroxid in solu¬ 
tion (1:2) to each. To one solution add 2-3 gm. of Devarda alloy powder and use 
the other as a control. Place a rubber stopper with a Bunsen valve in the neck 
of each flask and let stand 24 hours, after which attach to a condenser carrying an 
improved form of distilling bulb (McHargue’s form (4) was used in this work). 
Heat with a low flame just under boiling for 2 hours and then distill over at least 
150 cc. The distillate is collected in a measured portion of 0.1 N sulphuric acid, the 
end of the outlet tube having been so placed as to dip beneath the acid from the 
beginning. Care should be taken that the control and determination be carried 
on in the same manner; that is, distilled at the same rate and same volume of 
distillate collected. The excess of acid in the receiving flask is then titrated with 
standard alkali, using an aqueous solution of sodium alizarin sulphonate as indi¬ 
cator. The difference between the two titrations corresponds to nitric plus nitrous 
nitrogen. 
In this work, any small amounts of nitrite which may have been present are 
disregarded. It is doubtful if any quantity was present that would have mate¬ 
rially influenced the nitrate results and besides this would not affect the compari¬ 
sons between the two procedures. If its separation is desired, this can be made 
by a preliminary treatment with aspartic acid according to Strowd’s directions 
and then following the above procedure. If relatively large amounts of nitrate 
