The arthritogenic effect of dietary cow’s milk, egg albumin and soya milk were compared in Old English rabbits. The 12-week cow’s milk feeding regimen produced the highest incidence of significant joint lesions. Lesions were evident but mild at 5 weeks and the synovitis had resolved by 32 weeks. It is suggested that the transient nature of the synovitis may be due to the development of specific secretory IgA antibodies which were detectable in faecal pellet extracts. Sandy Lop rabbits were less susceptible to the arthritogenic effect than were Old English rabbits. Dietary ovalbumin was less arthritogenic than cow’s milk despite high titres of serum and synovial fluid antibodies and immune complexes. The rabbits were ‘tolerant’ to dietary soya due probably to pre-existing levels of soya protein in their diet. Lewis and Wistar strain rats, CBA, Balb/c and C57/BL6 mice fed on cow’s milk for 3 months did not develop serum antibodies or synovial lesions. It is suggested that this allergic synovitis is not a model for early rheumatoid joint disease because of the transience of the lesions and lack of stimulation of rheumatoid factor. It may well, however, be a model for the arthralgia seen in patients with certain food allergies.