A sero-epidemiological study on canine leptospirosis was conducted in house, stray, farm and hunting dogs, as well as in suspect cases of clinical canine leptospirosis. Serum samples were collected from apparently healthy (vaccinated and nonvaccinated), house dogs. A questionnaire was administered to the owners to elicit information on risk factors for leptospirosis. The microscopic agglutination test was used to screen for leptospirosis using 17 international serovars. Reciprocal titres of between 100 and <800 were considered as evidence of past exposure while reciprocal titres of 800 or greater were classified as suggestive of acute/current infection. Of a total of 419 serum samples tested, 61 (14.6%) were seropositive for Leptospira agglutinins, 23 (5.5%) had mixed infections and 16 (3.8%) had current infection. Amongst 50 suspected cases of clinical leptospirosis, 24 (48.0%) were seropositive and only 13 (26.0%) had current infection compared with 10 (6.3%) and three (1.9%) of 160 apparently healthy house dogs respectively. The difference was statistically significant (P < 0.05; v 2 ). Twelve (25.5%) of 47 hunting dogs, 10 (20.4%) of 49 farm dogs and five (4.4%) of 113 stray dogs were seropositive (P < 0.05; v 2 ). Overall, a total of nine serovars were detected with serovars mankarso, icterohaemorrhagiae RGA, autumnalis and copenhageni being involved in 29 (47.5%), 20 (32.8%), 25 (41.0%) and 10 (16.4%) respectively in 61 seropositive dogs (P < 0.05; v 2 ). Serovar mankarso was most predominant in seropositive apparently healthy dogs, 37.8% (14/37), suspected clinical cases of leptospirosis, 62.5% (15/24) compared with serovar icterohaemorrhagiae with a frequency of 21.6% (8/37) and 50.0% (12/24), the difference being statistically significant (P < 0.05; v 2 ). Although all vaccines used for prevention of canine leptospirosis in the country contain serovars canicola and icterohaemorrhagiae, serovar mankarso was mostly associated with infection and disease and may be a good candidate for inclusion in the vaccine used locally. The public health risk posed to owners of dogs infected with Leptospira cannot be over-emphasized considering the zoonotic nature of the disease.