When Brucella melitensis strain Rev 1 vaccine (Rev 1) is administered by the standard method (1-2 x 10(9) viable bacteria injected subcutaneously), it may induce long-lasting serological responses and/or cause abortion in pregnant animals. The conjunctival route considerably reduces these drawbacks. In the present experiment a 1 x 10(8) CFU dose for both ewes and goats conjunctivally vaccinated at mid-pregnancy was tested for innocuousness (outcome of pregnancy, contamination of unvaccinated contact animals, duration of serological responses) in comparison with 3 x 10(8) CFU (ewes and goats), 1 x 10(9) and 3 x 10(9) CFU (ewes) doses. No reaction was observed at the time of vaccination, and the risk of environmental contamination with Rev 1, due to the conjunctival administration of the vaccine, is negligible. Abortions occurred later at surprisingly severe rates (over 60% of pregnant vaccinated animals), except in the 1 x 10(8) CFU ewes group (20%). Moreover, the serological reactions of the 1 x 10(8) CFU ewes which normally lambed were negative again as early as 12 weeks after vaccination. Although the dose of 1 x 10(8) CFU Rev 1 was safer for pregnancy than the standard dose mainly in ewes as compared to goats, the innocuousness was not yet sufficient to propose the former dose to indiscriminately vaccinate sheep and goats by the conjunctival route, whatever the age or physiological status.