The moon is covered by a blanket of porous impact rubble (megaregolith) that probably averages about 2--3 km in thickness, and has a thermal conductivity of roughly 0.20W m-1 K-1, i.e., about 13 ¿ lower than the mean conductivity of the lunar lithoaphere, and 5--10 ¿ lower than the mean conductivity of the subjacent anorthositic crust. We employ finite difference models to study the effects of insulation by the megaregolith on lunar thermal evolution. Results indicate that the megaregolith has two important influences on heat flow: (1) Because the megaregolith is exceptionally thin in mare regions, heat passes more readily through them than through highland regions, and even flows laterally from the highland toward the mare. As a result, heat flow is exceptionally high along a boundary between highland and mare regions. The two lunar heat flow determinations were made along highlands/mare boundaries. Our calculationsindicate that heat flow at the Apollo 17 site is enhanced over the global mean by a factor of roughly 1.15--1.20. The effect is similar at the Apollo 15 site, which is probably further unrepresentative due to a local enrichment of (regolith) U and Th contents to 3.3¿ the global mean. Adjusting the Apollo 17 heat flow for the boundary effect, our best estimate for global mean heat flow is 12 mW m-2. (2) On a global scale, megaregolith insulation combined with lithosphere insulation causes slow cooling, which as a cumulative effect results in high present-day mantle temperatures and heat flow. Assuming that the global mean megaregolith thickness is 2 km, a heat flow of 12 mW m-2 is best matched by models with bulk moon U contents of 20--21 ng/g. Independent constraints on lunar internal temperatures derived from magnetic and tectonic data are best matched by models with ≈14 ng/g U. Thus our best estimate for the bulk moon U content is roughly 17 ng/g. These results imply that the bulk moon contents of U, and related refractory lithophile elements such as Th, A, Ca, etc., must be considerably lower than commonly assumed. ¿American Geophysical Union 1987