The Honolulu Volcanics comprises small volume, late-stage (post-erosional) vents along rifts cutting the older massive Koolau tholeütic shield on Oahu, Hawaii. Most of these lavas and tuff of the Honolulu Volcanics have geochemical features expected of near-primary magmas derived from a peridotite source containing Fo_87–89 olivine; e. g. 100 Mg/(Mg + Fe²⁺) >65, >250 p. p. m. Ni, and presence of ultramafic mantle xenoliths at 18 of the 37 vents. Consequently, the geochemistry of the alkali olivine basalt, basanite, nephelinite and nepheline melilitite lavas and tuff of the Honolulu Volcanics have been used to deduce the composition of their mantle source and the conditions under which they were generated by partial melting in the mantle.

Compositional trends in 30 samples establish that the magmas were derived by partial melting of a garnet (<10 per cent) Iherzolite source, which we infer to have been carbon-bearing, from analogy with experimental results. This source was isotopically homogeneous (Sr, Lanphere & Dalrymple, 1980; Pb, Sun, 1980; Nd, Roden et al., 1981), and we infer that the source was compositionally uniform in all major-element oxides except TiO₂, in compatible trace elements (Sc, V, Cr, Mn, Co and Ni), and in highly incompatible trace elements (P, Th, La, Ce). However, the source appears to have been heterogeneous in TiO₂, Zr, Hf, Nb, and Ta, elements that were not strongly incompatible during partial melting. Some nepheline melilitite samples may be derived from a source with distinct Sc and heavy-rare-earth-elements (REE) abundances, or which had a phase or phases controlling the distribution of these elements.

The relatively limited abundance range for several elements, such as Ti, Zr, Nb, is partly a consequence of the low degrees of melting inferred for the series (2 per cent for nepheline melilitite, 11 per cent for alkali olivine basalt), which failed to exhaust the source in minor residual phases. We infer that these residual phases probably included phlogopite, amphibole, and another Ti-rich phase (an oxide?), but not apatite.