Age-specific mortality levels off at advanced ages in many species; one explanation for this phenomenon is provided by the population heterogeneity theory. Although mortality at advanced ages can be well fit by heterogeneity models, population heterogeneity remains theoretical, lacking much direct evidence to support the existence of unobserved heterogeneity. Here, we provide direct evidence to support the heterogeneity theory by using isogenic population of worms of Caenorhabditis elegans. We measure the ability of individual worms to respond to a heat stress using an HSP-16.2 promoter that has been attached to GFP, a fluorescent marker that can be assessed in living animals. Worms differ substantially in their response; worms with high response have a long lifespan, and worms with low response to stress have a short life. Each of these classes results from a mix of two distinct, heterogeneous classes of worms and the addition of more classes does not result in a better fit.