that the agglutination and complement fixation tests for infectious abortion constitute an indispensable part of an intelligent study of the abortion problem and of any effective system of eradication. Evidence of the value of such tests may be sought in the following ways:

(1) By establishing a definite relation between the serological reactions and the calving incidence.

(2) By observing close agreements between the results of the

two tests.

(3) By testing the same animals again and again and noting agreements between the data obtained at different times.

I. RELATION BETWEEN THE SERUM REACTIONS AND THE LENGTH OF THE GESTATION PERIODS

It is a well-known fact that many cows which are infected with Bact. abortus carry their calf to full term. Therefore, positive reactions are not necessarily indications of premature calving. This is not due to any defects in the tests, but to the ability of some cows to retain the fetus to full time, in spite of Bang bacillus infection.

On the other hand, if the tests are of real merit very few nonreacting heifers and cows should abort, providing of course that no other bacteria or agencies are operating to cause premature calving. In the work of the Storrs Station the results have been such throughout the extended investigation as to preclude other organisms, except perhaps in a few isolated instances.

The numbers of reacting and non-reacting cows which calved normally and those which aborted are summarized in the following tables:

It will be seen at once that the figures in Table I are of no significance except that they show that normal calving animals may and often are reactors to both the agglutination and fixation tests. While in a few instances the positive animals constitute at least 45 per cent. of the given groups, and in two cases 66.6%, in most of the groups the negative heifers and cows greatly preponderate. It should be stated that in each of the herds recorded here infectious abortion had been thoroughly established, and that a large proportion of the animals were reactors to the serological tests.

Table II brings out a strong correlation between abortions and positive serological reactions, and demonstrates the specificity and diagnostic value of the blood tests.

In herd A in the years 1916 and 1917, the reactors constituted only 25% and 33 %% respectively of the aborting animals. No explanation can be offered for this except that other organisms may have played some important part in the bringing about of the abortions. In these two years the numbers of animals tested were so small as not to affect the total calving and reacting records materially. In the 1917 figures are included a premature fetus which was decidedly deformed.

TABLE I. Giving Blood Reactions of Normal Calving Animals.

Year

Herd A.

Herd B.

Herd C.

Herd D.

Pos. Neg. % Pos. Pos. Neg.% Pos. f'os. Neg. % Pos. Pos. Neg. % Pos.

NOTE: The differences in the proportions of reacting and non-reacting animals noted from year to year are not due to changes in individual animals, but to removal of old and addition of new animals.

TABLE II. Blood Reactions of Aborting Animals (1914-1923).

TABLE I. Giving Blood Reactions of Normal Calving Animals.

Pos. Neg. % Pos. Pos. Neg.% Pos. Pos. Neg. % Pos. Pos. Neg. % Pos. Pos. Neg. % Pos.

TABLE II. Blood Reactions of Aborting Animals (1914-1923).

In two groups of aborting animals 80% were positive; in one 85.7%; in three the reactors constituted 87.5%; in one 88.8%; in another 90%; in two 91.7%, and in 26 of the groups, 100%. In only four groups did the per cent of reactors fall below 80. These results in themselves furnish ample evidence of the diagnostic value of the agglutination and fixation tests. Further data on this point will be found in Storrs Bulletins 112 (1923) and 123 (1924) by the present authors.

II. CORRELATION BETWEEN THE RESULTS OF THE AGGLUTINATION AND FIXATION TESTS.

The following figures are compiled from the blood reaction records covering the ten year period, 1914 to 1923. All possible relationships between the two serological tests are shown here. The reactions are concentrated in the double plus, double negative, and double doubtful columns; outside of these they are thinly scattered through the other columns, under headings indicating disagreement.

Doubtful reactions to one test alone or the other, or even an occasional positive-negative or negative-positive combination, should not be surprising. Some animals may continue to react in an indecisive manner over considerable periods of time, and may never become confirmed reactors, but of course such instances are relatively few.

That heifers and cows which have recently been infected should react sooner by one test than by the other must be expected also; and it should not be surprising that young calves when they become non-reactors should not become negative by both methods at the same time. The same will apply to older or mature animals which happen (although infrequently) to revert to negative.

III. CONSTANCY OF REACTIONS OVER PERIODS of time.

An important check on the efficiency of any method is regularity in results when the tests are applied to the same test object in duplicate, or better still at widely different times. The following table (IV) is but one of many similar charts that may be constructed from the records of tests conducted on the various herds.

These records have been chosen for illustration purposes because they include three phases of reaction which are met with in any extensive serological work on infectious abortion, namely (1) the continued negative or positive type of reactions, (2) the transition from negative to positive, and (3) the change from positive to negative.

Cows No. 237, 238, 240, 434, 436, and 137 continued to give negative reactions to both tests as long as they were tested. Nos. 240 and 137 did so in spite of an attempt to infect them artificially by application of B. abortus emulsion to the scarified vulva.