Two thousand four hundred medical gloves were evaluated for leakage. Types of gloves examined (number of brands) included sterile latex (seven) sterile vinyl (four), nonsterile latex (six), and nonsterile vinyl (seven). Sampling was done from one box of each brand. Fifty gloves from each box were filled with 300 ml of water (the standard test used by the American Society for Testing and Materials). An additional 25 cm pressure was applied to water-filled gloves. Another 50 gloves of each box were donned and dipped into a basin that contained heparinized human blood. Only four brands of sterile latex surgeon's gloves proved nonpermeable to water and blood. Other brands showed leakage that ranged from 1% to 52%. Analysis of proportions of pairs of gloves permeable to water or blood indicated a strong statistical association of nonsterile packaging or packaging in suction kits with increased leakage rates. These findings affirm that gloves can be regarded only as a means of reducing the risk of gross soilage from blood or body fluids. Quality control standards to ensure more uniform glove quality are needed.