Petrogenesis of basalts from the Mid-Atlantic Ridge, 26° N

Abstract

Basalt genesis at the Mid-Atlantic Ridge, 26(DEGREES) N, can be explained by variations in the degree of partial melting of a pyrolitic mantle followed by shallow crystal fractionation of olivine, plagioclase and minor Cr-spinel. Application of a ranked-euclidean method of cluster analysis can identify genetically related basalt data that more closely approach statistically valid estimates of "true" composition. Two supergroups are recognized on the basis of distinctly different trace-element and slightly different major-element abundances. There is no way to chemically relate Supergroup 1 to Supergroup 2 by simple crystal fractionation models. Major-element least-squares modelling shows that Supergroup 1 can be derived by 25% partial melt of a pyrolitic mantle leaving a residue of 77% olivine and 23% orthopyroxene. Ratioing of the trace-elements between Supergroup 1 and Supergroup 2 shows that Supergroup 2 can be derived by a 13% partial melt leaving a residue of 1-3% plagioclase in addition to approximately 76% olivine and 23% orthopyroxene. Within Supergroup 1, dredge 18, cluster 17A and cluster 17B, and within Supergroup 2, cluster A and cluster B can be genetically related by crystal fractionation involving olivine, plagioclase and minor Cr-spinel. More precise and accurate trace-element analyses of the gabbros in the same area are consistent with a single-injection magma chamber model. The REE data, within experimental error, do not display any crossing patterns. Reanalysis of TiO(,2) by XRF using matrix corrections shows a significant order of magnitude increase over previously reported values.