TESTING WHEAT FOK PEOTEIX 
25 
Table 17 — Normality figures given to a sample of sulphuric acid by 34 collab- 
orators 
Number 
of 
collaborator 
Normal- 
ity sub- 
mitted 
Number 
of 
collaborator 
Normal- 
ity sub- 
mitted 
Number 
of 
collaborator 
Normal- 
ity sub- 
mitted 
Number 
of 
collaborator 
Normal- 
ity sub- 
mitted 
1 
0. 1276 
.1286 
.1290 
.1268 
.1297 
.1276 
.1290 
.1285 
.1286 
10 
.1274 
.1276 
.1282 
.1288 
. 12775 
. 12852 
.1294 
.1296 
.1280 
20 
. 0. 1268 
34 
0. 12826 
2 
11 
: 21 
.1274 
. 12735 
.1277 
.1288 
.1282 
.1286 
. 13033 
. 12771 
36 .. 
.1288 
3 . 
12.... 
22 
37 
.1268 
4 
13 
14 
23 
39 
.1295 
5 
24 
27 
42 
. 1270 
6 
15 
16 . 
45 
Xi 
. 1250 
7 
29 
31 
. 12974 
8 . 
17 
9 
18 
33 
1 United States Bureau of Standards. 
Average normalit y found 0. 12823 
Maximum normality found . 13003 
Minimum normality found . 12560 
Range in normality found 0. 00443 
Protein equivalent , per cent .35 
The results obtained by 31 of the collaborators show that a de- 
cided effort should be made to standardize the manner of determin- 
ing the values of normal solutions. Eight distinct reagents were 
used to determine the normality of alkali solutions, namely, ben- 
zoic acid, constant boiling point hydrochloric acid, hyrochloric and 
sulphuric acid solutions of known normality, oxalic acid, potas- 
sium-acid-phthalate, potassium bitartrate, and succinic acid. In 
many cases the protein results are calculated direct from the values 
of the standard alkali. In other instances, however, a cross titra- 
tion is made and the results are calculated on the value of standard 
acid. Such a practice usually results in different normality values, 
especially when compared with results obtained by precipitation 
and weighing of either hydrochloric or sulphuric acid. 
The Bureau of Standards in submitting their results stated in 
connection with standardizing solutions volumetrically : 
The limit of accuracy of volumetric analysis is about 1 part in 500 unless 
special precautions are taken. All apparatus used at this bureau is cali- 
brated here and the tolerance permitted in the calibration will permit errors 
of this magnitude. Uncalibrated apparatus has been found to give errors 
very much in excess of this ratio. 
A solution standardized at 20° C. and used in correctly calibrated burettes 
will be in error 1 part in 1.000 if used at 25° C. and no account is taken of the 
expansion of the solution due to rise in temperature. 
The accuracy in reading the burette exerts another limiting effect upon the 
accuracy of the result as well as the ability of the analyst to bring the solu- 
tions titiated to ihe same degree of neutrality. When all of these condi- 
tions are considered, accuracies of 1 part in 500 are considered satisfactory and 
the expression of results to a greater degree of accuracy has no significance. 
It is suggested that careful consideration be given to standardiz- 
ing solutions with potassium acid phthalate. This salt can now be 
obtained in very pure form, is nonhygroscopic, and can be very ac- 
curately weighed on account of its high molecular weight. Fur- 
thermore, the end point in titration with sodium or potassium hy- 
droxide is remarkably clean cut.„ The use of this salt as a standard 
in volumetric analysis has been fully discussed by Hendrixon 1S and 
18 Hexdrixsox, W. S. acid potassium axd acid sodium phthalates as STAXDARDS IX 
acidimetry axd alkalimetry. Jour. Anier. Cbem 1 . Soc, vol. 37, p. 2352. 
