The above simple tabulation will serve to present concisely the information available in our records on the history of dams and daughters. However, for the sake of completeness, the animals concerned are further classified. (Table 6).

Forty-five non-reacting dams produced 64 daughters that were under observation to the time of their calving. Thirtyfive, or 54.7 percent, of these were non-reactors, 28, or 43.8 percent, were reactors, and one was doubtful at the time of first calving. Ten of the gestations of 28 reacting daughters terminated in abortions.

Thirty-three reacting dams produced 39 calves, of which 28, or 71.8 percent, were non-reactors and 11, or 28.2 percent, were reactors at first calving. Three of the reacting cases terminated in abortions.

Five doubtful reacting dams produced five daughters of which three were non-reactors and two were reactors at the first calving. No abortions occurred.

Five dams whose reactions were unknown produced six daughters, four of which were non-reactors and two reactors at first calving. No abortions occurred.

No abortions were observed in any of the non-reacting first calf cows.

These data would seem to show conclusively that there is no indication whatsoever of persistant latent infection in daughters of infected dams. In fact, fewer of the daughters of infected dams became reactors immediately before or during the first gestation than those of the non-infected dams. This may suggest greater resistance to infection by daughters from infected dams than from uninfected, but such a conclusion should not be arrived at hastily. Herd B contributed a large majority of the negative daughters from positive dams. As this herd during the observation period showed evidence of stationary or suppressed infection, the results appear to be more favorable in the aggregate for the reacting than for the non-reacting dams. Since Herd B contributed eleven non-reacting and seven reacting daughters from negative dams, this interpretation is further supported. This subject is again referred to in the first paragraph on page 300.

The data in Tables 5 and 6 are provided by Herds A, B, D, H, and I. The following is a list of the animals studied, in which the first number is that of the dam and the numbers opposite are those of the daughters. Dam No. 17, for instance, has two daughters, 124 and 152. Reactions are given as 0 (non-reacting); (reacting); ? (doubtful); or U (unknown). The letter identifies the herd.

Note-See Key to Table in preceding paragraph

It will be noted from the record of the dams and daughters that twenty dams have produced two or more calving daughters, as follows: B 17, B 552, H 9, D 600, D 601, D 602, D 5, I 15, A 200, A 202, A 203, A 209, A 211, A 414, A 602, A 224 A 409, 404, A 407, and A 207. Of these twenty dams, five had reacting progeny, four had non-reacting progeny and eleven had both reacting and non-reacting progeny. This ratio of 1:1:2 further strongly points to the supposition that all daughters born into an infected herd are equally subject to the hazards of abortion, except as influenced by individualistic qualities.

V. HISTORY OF PURCHASED COWS.

In Bulletin 93 it was stated that 32.6 percent of the cows purchased were reactors when introduced into the herds.

To the end of 1923, the records of 179 purchased cows are available. Of these animals 118 were non-reactors, 9 were doubtful, and 52, or 29.1 percent, were clearly positive. This is significant in that it indicates a general widespread condition of infectious abortion.

Forty-nine of the 118 non-reacting cows became infected after being introduced into the several herds that are under observation, and developed a reaction to the test. This constitutes 41.5 percent of the original non-reactors.

Twenty-eight, or 57 per cent, of the 49 newly infected cows have aborted, most of the abortions occurring in the first gestation following infection. This indicates a susceptibility among newly exposed milking cows perhaps as great as among heifers.

CONCLUSIONS.

Including both the day of service and the day of calving in the summaries, the average length of a normal gestation period of 640 uninfected cows was 281 days. The 462 reacting cows which carried their calves more than 265 days had almost exactly the same length of gestation. By allowing only one full day for both the day of service and the day of calving, the normal gestation period may be taken as 280 days. The calvings are well distributed through the time range of 266 to 300 days, with a concentration of 37 per cent between the 281st and 285th days.

Only 2.87 percent of the gestations of non-reacting cows terminate before the 266th day. In frequency distribution a sharp division line occurs at this point, and gestations terminating before the 266th day may properly be looked upon as abortions, due, with relatively few exceptions, to Bact. abortus infection.

Many of the permanently reacting cows are able to carry calves to full time, but over a period of years 26.23 percent of the calvings terminate before the 266th day. A much higher percentage of reactors than this abort during the gestation in which they first become infected, and only a few reacting cows escape abortion entirely if their career extends over any considerable time.

The average length of gestation of the 629 reacting cows was 266.9 days, as compared with 279.2 days for the 671 nonreacting cows. This point is of much economic importance.

From these studies it would seem that only about one in ten of premature calvings may be attributed to causes other than Bact, abortus.

Abortions may occur at any time during the gestation period, but with greatest frequency between the 220th and 265th days. There seems to be no definite period of concentration except within these rather broad limits.

Twins seem to be carried in utero a shorter time than single calves.

There was one case of retained afterbirth out of 9.8 calvings in Herds A and B. These records are spread over eight years. Retention apparently occurs as frequently with nonreacting cows as with reactors, and this fact should eliminate Bact. abortus as the chief factor. It would seem that infection other than the Bang bacillus may operate as a causative agent. The abortion rate in this class is high among the non-reactors as well as among the reactors.

In retained afterbirth the majority of cows that suffer conceive and calve again, the average number of services to produce conception for those conceiving being slightly less than two. The processes of reproduction are undoubtedly impaired.

The strong tendency to repetition with many cows minimizes the importance of management and feeding in retained afterbirth and points to infection or physiological weakness as a causative factor. This suggests the advisability of selling ordinary cows which retain the placenta, since this condition so frequently suppresses milk production.

Calf scours, infectious mammitis and calf pneumonia have been so little encountered in these herds which have been badly infected with Bact. abortus that it would seem as if no additional evidence were necessary to mark these diseases as independent of the Bang abortion disease.

From a study of the calving and blood reaction records of the 64 daughters of non-reacting dams, 39 daughters of reacting dams, 5 daughters of doubtful dams and 6 daughters of dams of unknown reaction, the conclusion is reached that calves are not permanently infected in utero, nor during early calfhood, and, regardless of whether or not the dam is infected, all of the calves, aside from individual peculiarities in the matter of suceptibility, run the same hazards and share the same fortunes when they start their reproduction career. This clear