The question, whether obesity is associated with an increased incidence of positive axillary nodes at mastectomy for breast cancer, was studied in two quite different hospital populations; one from a large urban teaching hospital (Montefiore) and one from a medium-sized Air Force medical center (Malcolm Grow). In the Montefiore population, the answer was “yes”; 67% of the node-positive patients, but only 31% of the nodenegative patients were obese (20% or more above ideal weight) (p < 0.05). In the Malcolm Grow population, the answer was “no”; 20% of the node-positive and 20% of the node-negative patients were obese. The different answers, we believe, are due to the biological differences between the populations; the Montefiore population was shorter (by an average of 1.7 inches), heavier (by an average of 20 lbs), and more obese. The incidence of obesity was about three times as high in the Montefiore population (52% versus 20%; p < 0.02) and it contained a statistically distinct subpopulation of obese patients, while the few obese patients in the Malcolm Grow population constituted merely the upper tail of a unimodal log-normal distribution of weight in that population. We propose that it is possible to demonstrate a relationship of obesity to node-positivity in the Montefiore population but not in the Malcolm Grow population because obesity was highly prevalent in the former and almost nonexistent in the latter. It seems self-evident that it is not possible to demonstrate an effect of obesity in a population if that population manifests no significant obesity, statistically speaking, but disregarding this principle, we believe, may account for the controversy in the literature about whether obesity is a risk factor in breast cancer.