The infectivity of Plasmodium-infected humans in western Thailand was estimated by feeding laboratory-reared Anopheles dirus Peyton and Harrison mosquitoes on venous blood placed in a membrane-feeding apparatus. Between May 2000 and November 2001, a total of 6,494 blood films collected during an active malaria surveillance program were checked by microscopy for the presence of Plasmodium parasites: 3.3, 4.5, and 0.1% of slides were P. falciparum- (Pf), P. vivax- (Pv), and P. malariae (Pm)-positive. Venous blood was collected from 70, 52, 6, and 4 individuals infected with Pf, Pv, Pm, and mixed Pf/Pv, respectively, with 167 uninfected individuals serving as negative controls. Only 10% (7/70), 13% (7/52), and 0% (0/6) of membrane feeds conducted on Pf-, Pv-, and Pm-infected blood yielded infected mosquitoes. One percent (2/167) of microscope-negative samples infected mosquitoes; however, both samples were subsequently determined to be Pf-positive by polymerase chain reaction. Gametocytes were observed in only 29% (4/14) of the infectious samples. All infections resulted in low oocyst loads (average of 1.2 oocysts per positive mosquito). Only 4.5% (10/222) of mosquitoes fed on the seven infectious Pf samples developed oocysts, whereas 2.9% (9/311) of mosquitoes fed on the seven infectious Pv samples developed oocysts. The probability of a mosquito becoming infected with Pf or Pv after a blood meal on a member of the human population in Kong Mong Tha was estimated to be 1 in 6,700 and 1 in 5,700, respectively. The implications toward malaria transmission in western Thailand are discussed.