Considerable insight into the evolution of back-arc spreading centers has been gained as the result of Leg 135 of the Ocean Drilling Program in the Lau Basin in the western Pacific, as well as from recent study of Mariana trough magmas. In an overview discussion of the geochemical and petrologic results of the drilling, Hawkins and Allan [1994] found that significant crustal heterogeneity existed within the Lau Basin, and that MORB-type lavas, arc-type lavas, and lavas transitional between the two exist in close spatial and temporal proximity. These associations are also found at other backarc basins, and lead those authors to suggest that the lavas found in back-arc basins formed from supra-subduction zone magmas that are neither true arc nor true MORB, but, depending upon the geologic history of their source, show varied chemical characteristics of each type. This interpretation is in contrast to earlier views of back-arc basin magmatism which required that a spatial relationship exist between MORB-type and arc-type magmas at back-arc spreading centers, with arc-type magmas occurring towards the subduction side of the back-arc basin. Mixing of melts from a multiply-depleted mantle source, which subsequently received variable re-enrichment with a subduction component, is the mechanism presented as the primary control on the generation of the spectrum of compositions erupted at back-arc spreading centers.
Study of spinel compositions from lavas from three different sites in the
Lau Basin confirm the close spatial association of parental magmas of both
normal-MORB affinity and arc affinity [ Allan, 1994]. Using observed
mineral assemblages and compositions, Bryan et al. [1994] inferred
that many of the Lau Basin lavas crystallized from magmas with elevated
H
O-abundances. Gaetani et al. [1994] used experimental
constraints obtained from H
O-saturated phase equilibrium experiments
on a natural basaltic andesite composition from Hole 839B to infer
that pre-eruptive H
O contents for lavas from Hole 839B were
4 wt.%. Farley [1994] demonstrated that the backarc
glasses recovered from the Lau Basin are unique among submarine
tholeiitic glasses with respect to their oxygen fugacity
(fO
) (higher and more variable) and their sulfur abundance
variations (coupled with fO
but not with FeO), and suggested
that the source of the backarc magmas was affected by a hydrous fluid
with high fO
and high alkali and alkaline earth element compositions.
Working with submarine basaltic glasses from the Mariana trough,
Stolper and Newman [1994] determined that H
O abundances ranged
from 
0.5 wt.% to 2.0 wt.% and found
that they positively correlated with U, K, Ba, Cl, Sr and 
Sr, and
negatively
correlated with Na
O, TiO
, V, Y, Zr, and 
Nd. They
developed a
quantitative model that explains most of the variations in composition among
primitive
basalts from the Mariana back-arc trough by melting mixtures of a
normal-MORB mantle
source and an H
O-rich component, provided that the degree of melting
is positively
correlated with the proportion of the H
O-rich component in the
mixture. A strong
feature of this model is its ability to quantify the relationship between
the amount of
the H
O-rich component added to the source and the resultant increase
in extent of
melting.
Acknowledgments. The author thanks J. Reynolds, Dr. D. Christie and two anonymous reviewers for useful input, as well as Ken Verosub and the editorial staff for their assistance.