Microfilarial periodicity displayed by dogs infected with Dirofilaria immitis was analyzed using the methematical model proposed by Sasa and Tanaka (S.E. Asian J. Trop. Med. and Pub. Health 4, 1974, 518-536) except that Pearson's product-moment correlation test was substituted for their Chi-square analysis to estimate goodness of fit. Samples were drawn at 3 hour intervals for periods of 24 or 48 hours in a series of 10 studies conducted on 4 dogs. No significant differences were observed in microfilarial concentration between samples of venous and cutaneous capillary blood. Significant correlation of observed microfilarial densities with the values predicted by the model were observed in 7 of the 10 studies with r2 values ranging from 78.8 to 95.3%. The periodicity index ranged from 28.33 to 53.50 in these 7 studies and thus fell in the subperiodic category. The estimated hour of peak microfilarial density was not confined to the late afternoon as generally believed; for example, values of 04:00 and 22:00 hours were observed. Thus D. immitis microfilaremia cannot accurately be characterized either as diurnal or as nocturnal. Observation of the behavior of the experimental dogs led to an hypothesis that emotional and environmental stresses placed on the host may induce transients in the microfilarial density that may overwhelm or obscure the periodicity function.