The results of the experimental work are summarized as follows:

"The richness and acidity of the cream, size of the butter granules, temperature of and churning in wash water, method of salting, and amount of salt used do not materially influence the moisture content of the finished butter.

"Large churnings yield butter with a higher percentage of moisture than small churnings. Butter from raw cream contains more moisture than butter from pasteurized cream. High churning temperatures make butter retain more moisture than low churning temperatures. Working the butter in water regardless of temperature increases the moisture content of butter. The secret of moisture control lies in regulating the churning temperature and in adjusting the amount of water present during the working process according to the firmness of the butter, as determined by the chemical, physical, and mechanical properties of the butter fat, and in the systematic use of a reliable moisture test.

"Conditions that cause the formation of round, smooth butter granules, such as very thin cream held at a low temperature for a long time and which requires excessive churning and tends toward salviness of the butter, make moisture control more difficult and the results more uncertain than when the butter granules are irregular, flaky, and not too firm.

"The moisture is not evenly distributed throughout the churn. For this reason it is not safe to run too close to the 16 per cent limit and it is advisable to establish 15 per cent as the danger line. In order to secure a representative sample of the butter in churn, it is necessary to take small portions of butter from all parts of the churn. When sampling, care should be taken to avoid water pockets.

"Some moisture is lost during the transfer of the butter from the churn to the tub or box and when printing the butter. This loss tends to be greater during the winter months when the butter is firm than during the summer months when the butter is soft. A conservative estimate puts the average loss of moisture in packing at about 0.5 per cent.

"Considerable moisture is lost during the storage of butter. This loss is controlled by the salt content of the butter and by the thoroughness of moisture incorporation. Unsalted butter loses very little, if any, moisture in storage. The more salt the butter contains the greater is the loss of moisture in storage. Butter in which the moisture is properly incorporated loses less moisture than butter with a loose and leaky body.

"The accuracy of the results of moisture determinations by the butter maker depends on the preparation of the sample, the sensitiveness, condition, and manipulation of the balance, and the carefulness and judgment of the operator in making the test. Most of the moisture tests now available for the use of the butter maker are satisfactory and yield reasonably accurate results if manipulated according to directions."

The dairy school, L. A. BROwn et al. (N. Y. Produce Rev, and Amer. Cream., 35 (1912), No. 6, pp. 302-304).—A symposium by practical butter makers on the features of dairy school instruction which have been of most help to them. Introduction to the manufacture and handling of Emmental cheese, A. PETER and J. HELD (Praktische Anleitung zur Fabrikation und Behandlung des Emmentalerkases. Bern, 1910, 2. ed., pp. VIII+108, pls. 9, fig. 1).—A guide to cheese making, based on practical experience and scientific investigation.

Dairy statistics of Sweden for the fiscal year 1910, G. LILJHAGEN (Meddel. K. Landtbr. Styr. [Sweden], 1912, No. 176, pp. 68).—These figures, which contain reports of 133 dairies located in 13 counties, show a general increase in the

industry. The data include value of buildings and equipment, number of patrons and cows, methods of payment, net profits, and the production of milk, butter, and cheese.

VETERINARY MEDICINE.

Report of the bacteriologist, C. E. MARSHALL (Michigan Sta. Rpt. 1912, pp. 151-181).-This includes a report by Miss Rademacher on the commercial phases of the laboratory, viz, hog cholera serum, tuberculin, and legume cultures distribution, and 2 papers, as follows:

[The use of salt solution in the production of hyperimmune serum against hog cholera], by W. S. Robbins (pp. 151-178).-In manufacturing antihog cholera serum by the Dorset-Niles subcutaneous method the virus pigs are sacrificed for their blood, and this is only sufficient to hyperimmunize one pig of the same size as the virus pig. An attempt was made to reduce the cost of producing the serum by employing the principle set down by Craig and Madaus (E. S. R., 23, p. 185). Experiments were run parallel with the Dorset-Niles method. In the work 56 virus pigs were injected with salt solution (0.75 to 0.85 per cent at 37.5° C.) in amounts varying from 20 to 45 cc. per pound of body weight. The time in which the solution was allowed to remain in the abdomen varied and ranged between 3 and 14 hours. With the virulent salt solution taken from the abdomen 43 pigs were hyperimmunized, giving subcutaneous injections in amounts varying from 10 to 20 cc. per pound of body weight.

The data, reported in tabular form, indicated that the virulence of the salt varied greatly with the amount of injection as well as the time retained in the abdomen. The percentage of injected solution recovered varied greatly with the size and age of the pig as well as with the time in the abdominal cavity. Salt solution injected in amounts not exceeding 30 cc. per pound of body weight and allowed to remain in the abdomen not less than 5 hours, or for a longer time, was found to be efficient in hyperimmunizing animals.

Subsequently, about 100 pigs were hyperimmunized, using 6 hour, 30 cc. virulent saline solution, together with virus blood, and all produced very potent sera. "Thus these results look suggestive of a considerable saving in the cost of manufacturing serum."

Avian tuberculosis, by W. Giltner (pp. 180, 181).-New outbreaks of tuberculosis in chickens were noted during the year. The tests reported did not indicate that cohabitation surely or quickly resulted in the transmission of the disease.

Seven White Leghorn hens kept in company with tuberculous hens, and later in uncleaned pens, showed no lesions of tuberculosis after 7 months. Some more recent tests are also reported on negatively. The disease could not be transmitted to white rats by subcutaneous injections of cultures or keeping the animals in cages occupied by tuberculous birds. No satisfactory results were obtained from making a microscopic examination of the swabbings from the cloaca or from droppings.

Some apparent relationships between the etiologic organism in Jöhne's disease of cattle and that of avian tuberculosis led to a test of the effects of the avian germ on calves. "On May 17, 1911, twin calves about 2 weeks old were fed in milk the finely chopped organs of a tuberculous hen. The organs were seriously affected. This material was entirely eaten only after 2 days. Slight diarrhea supervened and rapidly disappeared. On July 5, the male calf died suddenly from acute tympanites. No effects of the ingestion of tubercle germs could be seen at autopsy. On July 16 and 17 the female calf was tested with avian tuberculin. A normal calf was also tested at this time. The tuber

culin was made by cultivating the avian tubercle germ in potato broth for 8 weeks and evaporating to one-fifth its original volume. . . . The laboratory calf had a slight diarrhea during the reaction. At 2 p. m. on the next day, the temperature was 103.2°." The test has not yet been verified by autopsy, but this is expected to be done later.

Flax experiments, J. W. INCE (North Dakota Sta. Rpt. 1911, pt. 2, pp. 129-138).-At least 32 cattle died in North Dakota during 1911 as the result of eating flax screenings.

Continuing previous work (E. S. R., 26, p. 86), analyses were made of a plat of flax grown at the station, and also of other samples. The percentage of prussic acid in the bolls was found to be, on a moisture-free basis, 0.094 on July 15, 0.017 on August 7, and 0.035 on September 9, and in the leaves on the same dates 0.337, 0.04, and 0.029. The stems contained 0.287 per cent July 15, 0.039 August 7, 0.03 September 9, and 0.02 October 20. Leaves from another source, sun-cured, contained 0.01 per cent. The young flax plants, ground whole, contained September 21, when 14 in. high, 0.289 per cent; September 27, when 3 in. high, 0.277 per cent; October 4, when 4 in. high, 0.215 per cent; October 10, when 5 in. high, 0.211 per cent; October 16, when 5+ in. high, 0.2 per cent; October 20, when frosted slightly, 0.189 per cent; and October 26, when frosted severely, 0.161 per cent.

Other analyses showed a considerable amount of prussic acid present in flax, the percentage being the highest in young plants. Much apparently depended upon the time of the year that the sample is taken and the exposure of the flax to the influence of the weather. Outdoor curing, either the action of the rain or sun, lowered the amount of poison. Soaking in water green flax straw and straw which has been shocked in the field and exposed to the sun, rain, and snow for several months also showed a diminished amount of poison. "The green straw gave a decided test for prussic acid while the bleached straw gave none."

...

The author "failed to find free prussic acid liberated in germination of flax. Chloroform and ether both acted upon young green plants and a good deal of hydrocyanic acid was liberated. . . As before stated, the amount of the poison in the flax falls off gradually, finally becoming almost negligible. Why then, should the samples of flax screenings have caused the death of so many cattle the past year? Our theory is this: The season was such as to prolong the growing period or to postpone the maturity of the flax. The plant remained green longer and perhaps the glucosid was arrested by frost, so that considerable of the prussic acid was available for liberation under proper conditions. These green parts of the plant, the leaves and stems, found their way into the screenings, and therefore caused the poisoning. Examination of the samples of screenings submitted showed that there was considerable green material present."

The hydrocyanic acid content of other materials, such as bitter almonds, peach, plum, and cherry kernels, cherry laurel leaves, vetch, etc., is also determined and reported.

The diagnosis of anthrax by Ascoli's method, F. FISCHOEDER (Ztschr. Infektionskrank. u. Hyg. Haustiere, 12 (1912), Nos. 1, pp. 84-97; 2, pp. 169–182).— Ascoli's method will give positive results where anthrax is present, but, on the other hand, it may show positive results in other cases as well. It should not, therefore, be used alone but in conjunction with some other method.

Precipitation in anthrax and hog erysipelas, M. DECLICH (Ztschr. Infek tionskrank. u. Hyg. Haustiere, 12 (1912), No. 5, pp. 434–454).—The thermoprecipitin reaction in general is considered an excellent aid for diagnosing anthrax and erysipelas in hogs. Although it is one which is easily conducted,

it should not be employed by the ordinary practitioner without an adequate control. Other methods of diagnosis should be used.

Comparative investigations of the trypanosomes of East Prussian Beschälseuche and Algerian dourine, H. MIESSNER and WEBER (Mitt. Kaiser Wilhelms Inst. Landw. Bromberg, 4 (1912), No. 3, pp. 188–224; abs. in Berlin. Tierärztl. Wchnschr., 28 (1912), No. 49, p. 915).—The investigations here reported have led the authors to conclude that Beschälseuche and dourine are identical. A bibliography of 35 titles is appended.

The action of salvarsan on the causative organism of dourine (Beschälseuche), F. FAVERO (Clin. Vet. [Milan], Rass. Pol. Sanit. e Ig., 35 (1912), No. 4, pp. 150–155; abs. in Berlin. Tierärztl. Wchnschr., 28 (1912), No. 39, p. 721).— This paper relates to the treatment of dogs, infected with Trypanosoma equiperdum, with salvarsan.

It was found that salvarsan prevented the development of the trypanosome when injected simultaneously with it or during the period of incubation. When injected after the appearance of the trypanosomes in the blood, they disappeared the following day.

East Coast fever in the Colony of Eritrea, M. CARPANO (Clin. Vet. [Milan], Rass. Pol. Sanit. e Ig., 85 (1912), No. 19–22, pp. 821-862, pls. 3, fig. 1).—This paper deals with the morphology and biology of Theileria parva. Rhipicephalus appendiculatus is said to be the tick which transmits the disease in Eritrea. A bibliography of 47 titles is appended.

Tests with the new serodiagnostic method for glanders, W. PFEILER and G. WEBER (Berlin. Tierärztl. Wchnschr., 28 (1912), No. 43, pp. 785–788).—This is a preliminary report on a study of the conglutination reaction (E. S. R., 23, p. 785), with 8 horses, 6 of which were glanderous and 2 healthy.

For determining the activity of the conglutinating system decreasing amounts of bovine serum, previously heated to 54° C. are mixed with 0.1 cc. of fresh horse serum and 3 drops of a 0.5 per cent sheep blood corpuscle suspension. The mixture is then filled up to the 1 cc. mark with physiological salt solution and placed in the incubator for about 1 hour. After this period conglutination should be absent in other tubes which serve as controls, i. e., those containing either fresh horse serum, bovine serum, or sodium chlorid solution acting upon sheep's blood corpuscles. The valuation of glanders bacillus extract, which is used in the test, is ascertained in the same manner.

For the actual test decreasing amounts of the serum from suspected animals are taken, and to these are added a titrated bacilli extract, fresh healthy horse serum, physiological salt solution, and finally bovine serum and sheep corpuscles. If no conglutination takes place, the reaction is positive, or in other words, anticonglutinins are present.

The value of the bacilli-conglutination method for diagnosing glanders, W. PFEILER and G. WEBER (Berlin. Tierärztl. Wchnschr., 28 (1912), No. 47, pp. 873-875). In the above work it was shown that the sera from glandered horses could be differentiated from healthy sera with the conglutination reaction. The authors now report on some tests made with the bacterial-conglutination reaction.

It is believed that the phenomena of bacterial conglutination are highly interesting from a scientific point of view, but it is doubtful whether the method in its present form will be of value in actual practice for diagnosing glanders. Accordingly the blood corpuscle conglutination test is considered the better one. Comparative investigations of the sera from 100 horses with the agglutination, complement fixation, and conglutination tests for diagnosing glanders, W. PFEILER and G. WEBER (Ztschr. Infektionskrank. u. Hyg. Haustiere, 12 (1912), No. 5, pp. 397-415).-The first group of animals examined consisted of

45 sound horses selected on the basis of agglutination and complement fixation tests. All of these animals showed no inhibition of the conglutination reaction. Fifty-four other animals which were found to be glandered with either the agglutination or complement fixation test were examined with the conglutination test. It was found that all which reacted with the complement fixation test also gave an anticonglutination test. On the other hand, 23 of the animals which gave the agglutination reaction could not be detected with the conglutination test.

In a third group, 5 animals were examined in which no reaction was obtained with either the agglutination or the complement fixation test. These animals were, however, condemned on the basis of the clinical findings, etc. Three of the animals gave a positive conglutination test and 2 were negative.

Another set of tests was made with horses not affected with glanders, which received subcutaneous injections of mallein for the purpose of determining whether or not the anticonglutination reaction was specific. The sera of these animals were tested by means of the agglutination, complement fixation, and conglutination reactions before and after malleinization. All the animals with one exception gave positive agglutination, complement fixation, and conglutination reactions 14 days after instillation of the mallein.

Evidently anticonglutinins are produced by injecting portions of the glanders bacillus into the organism.

Some tests in regard to the use of the proposed methods for diagnosing glanders in sound horses, A. DEDJULIN (Ztschr. Infektionskrank. u. Hyg. Haustiere, 11 (1912), No. 5, pp. 365–377).—In order to determine how often the biological methods show glanders in sound horses, a study of the complement fixation, agglutination, malleinization, and precipitation tests was made. In the case of the complement fixation test the author distinguishes between the SchützSchubert and the Bordet-Gengou method.

In all 245 sound horses which showed no clinical manifestation of the disease, and 6 horses affected with glanders were used for the tests. The animals were located in 4 different districts but in one of the groups were horses which had passed through a course of pleuropneumonia. Both sexes were represented. Of the 245 sound horses 143 gave negative results and the remaining 102 showed positive or other doubtful results. Of these 102 animals 74 gave only one kind of reaction; 23, 2 reactions; and 5, 3 reactions. In none of the horses was 4 or 5 reactions positive or doubtful. Of the 74 cases where only 1 of the 5 reactions was obtained (positive or doubtful), 5 reacted to subcutaneous malleinization, 50 gave precipitations, 15 agglutinations, and 4 complement fixations. Coincidental reactions were noted in only 25 horses, in which 2 reactions agreed 19 times and 3 reactions 6 times. The agglutination reaction agreed with the precipitation reaction 12 times. In 2 cases the reaction after subcutaneous malleinization agreed with the agglutination test, and in another 2 cases the subcutaneous reaction gave the same results as the complement fixation test. In 1 case the ophthalmo reaction and agglutination reaction agreed, and in another case the complement fixation reaction agreed with the precipitation reaction.

If the doubtful reactions are left out of consideration, the results show that the ophthalmo reaction and complement fixation test did not show positive in a single instance and the subcutaneous malleinization was positive in only 4 cases; a positive precipitin test was obtained with 11 horses. A positive agglutination reaction was noted in 4 cases.

With the 10 horses which went through a cycle of pleuropneumonia 4 gave a doubtful precipitin reaction and 1 a doubtful agglutination test. The reac