Sites 225, 277, and 228 in the central and southern-central Red Sea show classical diffusion gradients, approaching saturation in interstitial NaCl with depth as halite beds are approached. Site 228 did not actually encounter halite, but interstitial gradients suggest that it was only several tens of meters deeper than the bottom of the hole. K and Mg were characteristically depleted with depth until the evaporitic layers were approached. Then these constituents and also B and Li increased sharply. At Site 227, intra-halite pore fluids were strongly enriched in Mg (17 g/k Mg) and Ca (6.7 g/kg), indicating the presence of late stage evaporites, dominated by tachyhydrite, in the salt beds. The brines associated with evaporite beds resembled the brines in the hot brine metalliferous deeps only in total dissolved solids (about 250 g/kg); chemical composition was notably different. This confirms other evidence that the hot brine deep fluids do not have a local origin but probably rise from deeper layers through fracture channels. Site 229 showed a relatively small increase in chlorinity (21.5 g/kg at sea floor to 27 g/kg at 210 m). However, considering the extremely rapid rate of accumulation of Quaternary and Pleistocene sediments, this gradient is regarded as sufficient to suggest the presence of salt at depth here also. In contrast, an evaporitic episode in the Red Sea during the Wisconsin glacial maximum, as suggested by Friedman (1972), is not supported by the pore fluid data. Some evaporation and enhancement in salinity in the Red Sea may have occurred, but conditions sufficiently extreme to lay down evaporites in the sea as a whole probably did not occur. The hot brine deep sediments yielded interstitial fluids roughly comparable with previous analyses both of the hot brines themselves and of associated interstitial waters.