The 1.032 is the specific gravity of average milk; 18 the number of grams of milk delivered by the 17.6 c. c. pipette; .999 the specific gravity of 30 per cent cream. The result, 17.4 grams, shows the weight of 30 per cent cream contained by a 17.6 c. c. pipette.

By making a second proportion:

17.4 : 30 : : 18 : X

We find that if 17.4 grams gives a reading of 30 per cent, that 18 grams should give a reading of 31 per cent. The error

in testing 20 per cent cream by measure, therefore, is 1 per cent. The same method was used to find the error for the different per cents of fat for which the specific gravity had been found.

Afterwards a large number of trials were made by testing the same cream by weight and by measure, and the results compared with the error as calculated by proportion. It was found that by careful work it was possible to test cream by measure, adding the calculated correction to the reading and to quite closely approximate the test of the same cream by weight. The error in the reading after correction should be, and in most cases was, as small or smaller in relation to the whole than an error of .1 per cent on average milk.

The following table shows the specific gravity, the theoretical difference between tests made by weight and by measure as found by proportion, and the average difference as found by testing the same cream by weight and by measure.

The following table designed for practical use, is made up from the above. The error for each 5 per cent is divided equally and changed to the nearest tenth of a per cent.

It is to be noted that the corrections are only applicable to cream that is practically free from air bubbles, and therefore cream fresh from the separator, containing more or less air, will show more difference than the table represents. The liability to errors increases with the thickness of the cream. So the corrections are more certain to apply on cream of average thickness or less, than on that containing a larger per cent of fat. If it is desired to test considerable quantities of rather thick cream, as is often the case in factories which are the center of skimming station systems, it is undoubtedly best to depend upon weighing the samples for testing.

TESTING SOUR CREAM.

It is often desirable to test cream after it has reached a sour condition, and the question arises regarding the accuracy of

such tests.

It is well known that the production of acid has no effect on the fat content, so if sour cream is weighed for testing there will be no difference on account of the fermented condition. If there be any effect on the reading of the test, it must be when the amount is measured out with a pipette.

Sour cream is more viscid than sweet and consequently holds bubbles of air longer than sweet. If samples of sour cream are not mixed with care, enough air may become incorporated to seriously affect the result. Although of minor importance, it is probable that occasionally enough gas may be developed during the souring to lower the reading of the test slightly. However, as a rule, sour cream can be tested nearly as accu

rately as sweet. The following figures show the result of testing cream by measure after standing five hours to allow the air to escape, and again after being allowed to sour for twenty-four hours.

TESTING CREAM SOON AFTER SEPARATION.

It is sometimes important to know if cream will test the same after standing some hours as it does fresh from the separator. A number of tests were made on this point and the conclusion is that as a rule cream will test slightly higher after standing several hours. A certain amount of air is mixed into it during separation and this gradually escapes. The difference is at no time very much, usually less than 1 per cent. It varies with the separator used and the thickness of the cream. Cream that shows a fine foam on the surface after standing, contains air and will give too low a test fresh from the separator.

TESTING PASTEURIZED CREAM.

There has also been some inquiry regarding the effect of pasteurization on the accuracy of cream testing.

If a sample of cream be tested by measure, then pasteurized under the usual conditions (158° F. for 15 or 20 minutes) and another test made, the later will as a rule show a higher reading, usually from .5 to 1.5 per cent. A test by weight made before and after pasteurization shows practically the same increase in reading. This indicates that the increase under the first conditions is due to the same cause as in the latter, which seems to be the evaporation of a certain amount of water by the heating. The error of testing by measure, due to the low specific gravity, remains practically the same as when testing raw cream.

*

THE SPECIFIC GRAVITY OF CREAM.* The cream for the determination of specific gravity was secured under conditions as uniform as possible. About a quart of cream was taken, always from the same separator, and with a per cent of fat somewhat higher than desired for the particular test. Skim milk was taken at the same time and a small amount of formaline added to each to prevent fermentation. After standing five hours after separation a test for fat was made from the cream and the necessary amount of skim milk added to bring the per cent of fat to the desired point. On account of the possibility of variation in condition of samples, it was considered best to use a rather large quantity in finding the specific gravity. A measuring flask holding something over 500 c. c. was weighed empty and again when filled with distilled water a 17.5° C. The flask was then filled with the cream to be tested at the same temperature with as little agitation as possible and weighed. It was almost impossible to fill the flask exactly to a mark with cream, on account of the bubbles of air making the surface line indistinct. In order to overcome this, the flask was fitted with a rubber stopper containing a small glass tube. As the stopper was forced into the neck to a certain mark, the cream raised in the tube and overflowed, allowing a close measurement. Three trials were made for each 5 per cent from 10 to 50 per cent as shown in the table following. No two samples of the same per cent of fat were from the same day's supply of cream.

*The writer wishes to acknowledge his obligations to Dr. J. B. Weems, of the Depart ment of Agricultural Chemistry, for suggestions in regard to methods of carrying out this work.

THE TEST BOTTLE LAW AND ITS INEFFICIENCY.

The following section of the code has been in force since 1897 and the several dairy commissioners holding office in the time since the law went into effect, have made special efforts to induce creameries to comply with it. So far about half the creameries of the state have complied with the law, so far as to get from the dairy commissioner the certified bottle and pipette. But it is very much doubted whether any large number of creameries of the state have really complied with the law in its spirit:

SEC. 2523. Milk test. Any person or corporation, or the employe of such person or corporation, who operates a creamery or cheese or condensed milk factory, and uses a chemical milk test to determine the quantity of butter fat in milk purchased, used or received, shall so use only such tests as shall be clear oil, free from any foreign substance, and produce correct measurements of butter fat, and every such person or corporation using a milk test shall procure from the dairy commissioner for each factory so operated one standard tube or bottle, and one standard measure or pipette, for testing milk, certified and marked by him as in this chapter provided, which shall be kept for inspection by the patrons and used by such person or corporation in testing or verifying test tubes or bottles and milk measures or pipettes used. In any action arising between any such operator and patron, the burden of establishing the use of reliable tests and the results therefrom, equivalent to the standard herein provided, shall be upon the operator.

Another section authorizes the dairy commissioner to sell at cost, one certified bottle and one certified pipette to each creamery having in use a milk tester to which the director of the Iowa experiment station at Ames had given his approval. All but two of the machines in use in the state have received such approval.

It was expected that the creamery men would hasten to avail themselves of the provisions of this law, as a means of convincing skeptical patrons that the creamery was actually using correct apparatus. It was supposed that every creamery operator would do as the law requires, at least so far as to use the certified glassware for the purpose of calibrating the bottles and pipettes in use, and would discard the bottles that were incorrect, and that he would be shrewd enough to advertise the fact among his patrons.

With this thought in mind, a card of instructions was issued with every certified bottle and pipette so that the receiver of it would know exactly how to proceed. As a matter of fact