Seventy male mice (Kunming strain) were randomly divided into 7 groups (10 mice per group), each mouse was orally inoculated with 30, 25, 20, 15, 10, 5 or 3 muscle stage larvae of Trichinella spiralis, respectively. All infected mice were sacrificed 6 weeks post-inoculation, number of larvae per gram (LPG) of diaphragm were counted by compression method (trichinelloscopy), the carcass was digested by artificial digestion method and LPG was counted. The larval detection rate by trichinelloscopy and digestion method was 100% (10/10) in all mice infected with 30, 25, 20, 15 or 10 larvae, but 70% (7/10) and 100% in mice infected with 5 larvae, respectively. No larva was found by either method in mice infected with 3 larvae. There is a positive correlation between the larval burden (of diaphragm and muscle) and the infecting dose (r = 0.759, P < 0.05; r = 0.638, P < 0.05), and the minimum infection dose to a mouse is 5 larvae.