The objectives of this experimental study were: (1) to develop a reproducible skin xenograft mouse model of canine demodicosis, and (2) to test the hypothesis that lymphocytes affect 'Demodex canis ' populations 'in vivo'. This study compared the healing of full- and split-thickness canine skin xenografts, developed canine peripheral blood lymphocyte mouse chimeras and recreated canine allogeneic skin graft rejection in the murine model. Canine blood lymphocytes survived transfer to immunodeficient mice and produced variable amounts of canine IgG, up to 6.0 mg/mL. Transferred lymphocytes mediated allogeneic skin graft rejection. The full-thickness skin xenografting techniques developed led to well-haired, relatively large, canine skin xenografts. Four types of genetically immunodeficient mice (scid/bg, ICR scid, tg[epsilon]26 and 'Rag2') were found to support canine skin xenografts and 'D. canis' graft infections; however, development of the "leaky" phenotype and/or low survivability limited the use of scid/bg, ICR scid, or tg[epsilon]26 mice for modeling demodicosis. To directly test the lymphocyte hypothesis, 'D. canis' infected skin grafts on 'Rag2' null mice were treated with syngeneic canine lymphocytes and then graft mite numbers were compared. Grafts received either 25 * 106 unstimulated lymphocytes or 15 * 106 lymphocytes that were stimulated 'in vitro' with phytohemagglutinin and human recombinant interleukin-2. Skin xenografts grew abundant hair and did not develop gross lesions after ' D. canis' infection or lymphocyte transfer. Inflammation was not associated with 'D. canis' infected follicles. Ninety days post infection, the mean (±SEM) calculated number of mites per xenograft sample was significantly higher after treatment with stimulated lymphocytes, 15,330 (±3,583), than with unstimulated lymphocytes, 6,582 (±1,118) (P = 0.016), or with saline 8,931(±1,716) (P = 0.049). Canine IgG, measured by ELISA, was significantly higher in mouse sera after treatment of mite infected grafts with stimulated lymphocytes (mean ±SEM, 34.01 ±4.19 [mu]g/mL), than with unstimulated lymphocytes (P < 0.001). In conclusion, ' D. canis' mites proliferated to high numbers on xenografts, confirming the importance of systemic dog factors in controlling mite populations. ' D. canis' did not induce lesions of demodicosis in the absence of inflammation. Treatment with 'in vitro' stimulated lymphocytes was associated with increased numbers of mites; this was an unexpected finding. Furthermore, the methodology applied herein demonstrates the applicability of the skin xenograft mouse model in veterinary dermatology research.