From measurements of subcutaneous fat temperature (Tsf) at known depths below the surface, skin surface temperature (Tsk), and direct skin heat flux (H), the superficial shell isulation (Iss) of the thigh (fat + skin) was calculated as Iss (degrees C.m2.w-1) = (Tsf - Tsk)/H in nine male subjects immersed head out in a well-stirred water bath. Also, at critical water temperature (CWT = 28-33 degrees C), eight of the subjects rested for 3 h, enabling overall maximal tissue insulation (It,max) to be calculated as It,max (degrees C.m2.W-1) = (Tre - Tw)/(0.92 M +/- delta S), where Tre is rectal temperature, Tw is water temperature, M is metabolic rate, and s is loss or gain of body heat. Five subjects performed up to 2 h of mild leg cycling, preceded and followed by 60 min of rest, and both thigh Iss and overall It were measured during exercise. Iss increased from minimal values in Tw greater than 33 degrees C to maximal values (Iss,max) at CWT or below. Iss,max was linearly related to tissue thickness (d) in millimeters of fat plus skin, Iss,max (degrees C.m2.W-1) = 0.0048d-0.0052; r = 0.95, n = 37, and was not influenced by leg exercise up to a metabolic rate of 150 W.m-2 in CWT despite large increases in Tsf and H and large decreases in overall It. The slope of Iss,max vs. depth, 0.0048 degrees C.m2.W-1.mm-1, is almost identical to thermal resistivity of fat in vitro, suggesting that the superficial shell is unperfused in CWT at rest or during mild exercise. When maximal superficial shell insulation (It,ss,max) for the whole body was calculated with allowance for differing fat thicknesses and surface areas of body regions, it could account for only 10-15% of overall It,max at rest and 35-40% of overall It in mild exercise. We suggest that the poorly perfused muscle shell plays a more important role as a defense against cooling at CWT than does the superficial shell (fat + skin), particularly at rest.