The Du Bois and Du Bois body surface area (BSA) equation is used widely to normalise physiological parameters. However, that only nine subjects were used in its derivation does not appear to be well known and does not justify its ubiquitous application. Furthermore, the derivation appears to be hampered by a lack of modern statistical methods and the omission of a large amount of available data. We have shown that the omitted data, obtained by measurement of the length of body parts, were identical to the data obtained by encasing subjects in moulds ¿BSA (moulds; cm2) = [1.00 (0.02)] x BSA (linear measurements) + [123 (347)]¿. Non-linear regression analysis of the BSA of all 42 subjects reported by Du Bois and Du Bois gave new values for the constants of the model ¿BSA (cm2) = 94.9 x [weight (kg)0.441] x [height (cm)0.655]¿. Although the original equation obtained by Du Bois and Du Bois was found to be adequate in adults, we recommend that it should not be used in daily practice, owing to the low number of subjects used in its derivation. The work presented here has placed the original results of Du Bois and Du Bois on a more robust statistical footing, yielding values for the model constants that would have been obtained if Du Bois and Du Bois had had access to modern statistical methods.