The main arguments used to support the concept that komatiites form by melting of hydrous mantle are as follows: (1) Water reduces liquidus temperatures from extreme values to lower, more “normal” temperatures. (2) Some komatiites are pyroclastic and some contain vesicles, features that have been attributed to magmatic volatiles. (3) It is claimed from experimental studies of peridotite melting that the chemical composition of komatiite requires the presence of water, as does their characteristic spinifex textures. Counterarguments are the following: (1) Loss of volatiles as hydrous komatiite approaches the surface should produce degassing textures and structures, which, though not unknown, are rare in komatiites. Degassing should produce a highly supercooled liquid that partially crystallizes to porphyritic magma; komatiites commonly erupt as phenocryst-poor, highly magnesian lavas. (2) Chemical and isotopic compositions of most komatiites indicate that their mantle source became depleted in incompatible elements soon before magma formation. Such depletion removes water, leaving a dry source. (3) The experimental data are at best ambiguous; neither the chemical composition of komatiites, nor the crystallization of spinifex, requires the presence of water. We conclude that although some rare komatiites may be hydrous, most are dry.