The figures presented underestimate the true prevalence of infection with M. paratuberculosis, especially since not all suspect cases were submitted for culture or PCR. However, they do show that M. paratuberculosis appears to be spreading in farmed deer in New Zealand and highlight the possibility that Johne's disease is emerging as a potential major problem affecting this species. Identification of the organism by bacterial culture or PCR is required in many cases to distinguish lesions in mesenteric lymph nodes and lymph nodes of the head caused by M. paratuberculosis from those caused by M. bovis and M. avium.

Mycobacterium bovis was cultured from nine of 25 (36%) tracheal washings but not from any of 38 urine and 38 faecal samples from tuberculous possums cross-sectionally sampled from the wild. One of three tracheal washings, one of three urine samples and one of three faecal samples from terminally ill possums were culture-positive. The respiratory route is implicated as the major route of excretion of Mycobacterium bovis from naturally infected possums in horizontal transmission. Tuberculosis was observed in two young possums and was evidence of probable pseudo-vertical transmission via the respiratory route or ingestion of milk. Discharging fistulae were present in 22 of 71 (31%) cross-sectionally sampled tuberculous possums and were associated with relatively advanced disease. Although the frequent involvement of superficial lymphocentres in early stage disease could not be explained satisfactorily, the respiratory route was implicated as the main route of infection from indirect evidence.

Culturing of pools of lymph node samples detected a significant proportion of M. bovis-infected ferrets that would otherwise have gone unnoticed based on samples that had only macroscopic lesions. Culturing of samples pooled from up to 30 different ferrets could provide significant cost savings in surveys of wildlife for the presence of M. bovis infection without any apparent loss of sensitivity.

Tuberculous lesions were identified over a 2-year period in 36 clinically normal red deer from a single herd. The lesions were only present in the retropharyngeal lymph nodes and lymph nodes draining the intestinal tract, indicating infection by the oral route. Mycobacterium avium was isolated from 27 of 29 lesions examined by bacterial culture. Grossly and histologically, the lesions were indistinguishable from those caused by Mycobacterium bovis. DNA restriction endonuclease analysis revealed that all the 26 M. avium isolates available for examination had identical cleavage patterns. These patterns were identical to a New Zealand M. avium serotype 2 isolate from a pig and were very similar to a reference strain of M. avium serotype 2. The DNA examinations indicated that the deer were infected from a common source that was not identified.

The finding of M. bovis infection in possums with non-visible lesions identified a potential deficiency of declaring possum populations free of M. bovis on the basis of absence of macroscopic lesions. The culturing of pools of selected tissues from multiple animals without visible lesions can be used to reduce laboratory costs of possum surveys without a major reduction in the ability to detect M. bovis infection.