The currently known astronomical, chemical and magnetic data are not uniquely indicative of an extensively and globally molten Moon. We argue here for an accretional layering in the Moon, but at temperatures below solidus. The excess mass in the near side of the Moon compatible with a 2 km displacement in the center of mass relative to the centre of figure and the moment of inertia data is considered to be due to Fe-FeS liquid formation and inhomogeneous segregation. These Fe-FeS bodies, termed ‘fescons’, are shown to be capable of accounting for the presently available magnetization data, by acting as small regenerative dynamos with a time-stability less than that of the terrestrial equivalent. The chemical characteristics of the highly differentiated materials (KREEP, ‘granite’ etc.) are considered to be due to small scale localized melting caused by collisional events, from sources in which accessory phases play a significant role. Mare basalts are considered to be melts in the overlying material produced at a later time by40K radioactivity in the fescons. Some consequences of the present hypothesis are suggested. We conclude that these and other characteristics of the lunar materials are reconcilable with a ‘cold’ Moon such as discussed by Urey over the past two decades.