In December 1987, we investigated an increased number of cases of herpetic whitlow in medical intensive care unit nurses who routinely gloved for secretion contact. One particular brand of vinyl examination glove had been used in the medical intensive care unit. Restriction endonuclease mapping established the similarity of employee isolates with one patient isolate of herpes simplex virus type I. When initial viral assay demonstrated 2.5% to 10% penetration of herpes simplex virus type I across unused gloves, an evaluation of glove quality was undertaken. In a 300-mL watertightness test, seven brands of vinyl gloves failed 4% to 28% (average, 11.1%; 132/1200), while seven brands of latex gloves failed 0% to 2.6% (average, 1.4%; 24/1750). The brand of vinyl glove that had been in use in the medical intensive care unit failed 28% of the time. Watertight gloves were then tested for permeability to herpes simplex virus type I. None of the latex gloves failed (n = 1726), while only 10 of the vinyl gloves failed (n = 1068, 0.95%). Extreme variability in glove quality was observed. However, gloves made from intact vinyl may provide similar protectiveness as those made from intact latex. As the demand for gloves increases, emphasis should be placed on the production of plentiful, better quality latex and vinyl gloves.