Notes: [Auszug] Ocean-floor metamorphism, which overprints the pseudo-stratigraphy of ophiolite complexes in southern Chile, is best explained by a combination of hydrothermal and contact metamorphism associated with igneous activity at a spreading centre. Such metamorphism may be more important than igneous ...

Notes: Abstract The Sarmiento and Tortuga complexes are two mafic complexes located in southern Chile that represent the remnants of the mafic portion of the floor of an Early Cretaceous extensional back-arc basin. Basaltic dikes and lavas within each complex exhibit tholeiite differentiation trends whereby FeO*, FeO*/MgO, TiO2, P2O5, Zr, and Y increase together without significant increases in SiO2. In both complexes, as FeO*/MgO increases, REE abundance increases without significant change in Ce/Yb ratio, but with an increasing negative Eu anomaly. The Sarmiento complex contains intermediate icelandites and silicic dikes and lavas which are conspicuously absent in the Tortuga complex. These non-basaltic compositions have higher Zr, Y and REE contents than the associated basalts, but similar Ce/Yb ratios, suggesting co-genetic origin. Thick cumulate gabbro sequences in both complexes suggest shallow level crystal-liquid fractionation as a major cause of the observed wide range of chemical variations. Significantly, in basalts from the Tortuga complex, incompatible elements (Zr, Y, REE) increase in abundance more rapidly with increasing FeO*/ MgO than in the Sarmiento complex. The rapid increase of incompatible elements relative to FeO*/ MgO in the Tortuga complex is best modeled by fractionation within an open magma chamber steadily replenished with new batches of undifferentiated magma. The observed chemical variations within the Sarmiento complex are best modeled by a magma chamber replenished only a limited number of times by a continuously decreasing volume of undifferentiated magma, followed, subsequent to the last input of new parental magma, by closed system fractionation which results in the formation of ferro-basalts, icelandites and silicic differentiates. Ferro-gabbros (FeO* 〉20 wt °/00) found within the gabbro unit of the Sarmiento complex closely approximate in composition the calculated crystal extracts required to evolve ferro-basalts into icelandites and the more silicic differentiates. The difference between the nature of the postulated magma chambers within the spreading centers at which the Sarmiento and Tortuga complexes originated suggests that the zone of magma intrusion from the mantle may have been diffuse in the region where the Sarmiento complex formed and more localized in the region where the Tortuga complex formed. This is consistent with other geochemical and field evidence suggesting that the Sarmiento complex represents a less developed stage of evolution than the Tortuga complex of the mafic floor of the Mesozoic back-arc basin in southern Chile. The apparent decoupling of major and trace element variations in ocean floor basalts may be explained by shallow level igneous fractionation without involving large proportions of clinopyroxene if the magma chambers within spreading centers at midocean ridges behave as open systems periodically replenished with batches of undifferentiated parental magma as is inferred for the Tortuga complex in southern Chile.