A helium leak test for canned foods was developed to provide a more sensitive method than existing ones and to reduce dependence on operator judgment for detecting leaking cans. The test forced helium through holes into cans filled with foods. Can headspace gas was sampled, and the helium content was measured by gas chromatography. An approximately linear (r = 0.81) relationship existed between the helium content of the headspace and the hole size in the can. The method detected holes as small as 1 μm in diameter (the smallest hole tested). Measurements of direct holes and dented seam holes in metal cans indicated that most holes tested smaller upon retesting. However, some seam openings did become larger. Some holes closed completely, probably as a result of clogging by particles. Cans with direct holes of known size were processed in a canning retort, and the cooling water was inoculated with ca. 106 bacteria/ml. Cans with holes greater than 5 μm became contaminated as exhibited by gas production during...