1. Parainfluenza infections of the airways cause dysfunction of inhibitory M2 muscarinic receptors on the pulmonary parasympathetic nerves. To distinguish the direct effects of virus from the effects of virus-induced airway inflammation on M2 muscarinic receptor function, guinea-pigs were depleted of leukocytes by pretreating with cyclophosphamide (30 mg kg-1, i.p. daily for 7 days) after which they were infected with parainfluenza virus type 1 (Sendai virus). 2. Guinea-pigs were anaesthetized, tracheotomized, and ventilated. The vagus nerves were isolated and cut, and the distal ends were electrically stimulated causing bronchoconstriction. In control animals, pilocarpine (1-100 micrograms kg-1, i.v.) inhibited and gallamine (0.1-10 mg kg-1, i.v.) potentiated vagally-induced bronchoconstriction by stimulating or blocking M2 muscarinic receptors on the vagus. These effects of pilocarpine and gallamine were almost completely lost in virus-infected animals, demonstrating loss of M2 receptor function. 3. Cyclophosphamide depleted peripheral blood leukocytes and inhibited the virus-induced influx of inflammatory cells into the lung. Depletion of leukocytes protected M2 receptor function from viral infection in some, but not all, guinea-pigs tested. 4. Among infected animals that had been depleted of leukocytes, the viral content (expressed as the log of the number of tissue culture infectious doses per g lung tissue) of those that retained normal M2 receptor function was 4.29 +/- 0.05 (mean +/- s.e. mean), while the viral content of those that lost M2 receptor function despite leukocyte depletion was 5.45 +/- 0.20 (P = 0.011). Thus the viral content of the lungs in which M2 receptor function was lost was 16 times greater than that of the lungs in which M2 receptor function was preserved. Viral content correlated with the inhibition of vagally-mediated bronchoconstriction after the maximum dose of pilocarpine (100 Microg kg-1; r2 = 0.81, P =0.0004).5. In antigen-challenged animals, inhibitory M2 muscarinic receptor function is restored when positively charged inflammatory cell proteins are bound and neutralized by heparin. However, heparin(2000 micro kg-1, i.v.) did not reverse virus-induced loss of M2 muscarinic receptor function, even in those guinea-pigs with a lower viral titer.6. Because leukocyte depletion protected M2 muscarinic receptor function only in animals with mild viral infections, it appears that viruses have both an indirect, leukocyte-dependent effect on M2 receptors and, in animals with more severe infections, a leukocyte-independent effect on M2 receptors. The failure of heparin to restore M2 receptor function demonstrates that the leukocyte-dependent loss of M2 receptor function is not mediated by positively charged inflammatory cell proteins.