Bowen's Unsolved Problem: The Mechanics of Differentiation

The compositional diversity of the igneous rocks is well explained by Bowen's Reaction Series. Residual melts generally become increasingly silica-rich with increasing crystal content and diminishing temperature. F. A. Fouque came upon this effect in the 1870's when he separated and analyzed the groundmass of an intermediate lava on Santorini. In lieu of any knowledge of silicate phase equilibria, the bimodality reflected by apparent preponderance of basalts and granites was, until Bowen, inferred to reflect a system of phase equilibria dominated by two eutectic-like `pole' compositions in a binary system. Building from Day's plagioclase work, Bowen soon found that `when the ternary [with solid solutions] comes in at the door, the binary goes out at the window.' His geological common sense, especially his understanding of compelling field relations, and his analytical use of physical chemistry allowed him to see every igneous suite as a simple result of fractional crystallization. He became a master at constructing a path through a ternary to yield the liquid line of descent of any cooling magma chamber. At a purely chemical level, with the spatial context of each analyzed rock not a part of the problem, this style of interpretation became a fundamental truth. This is still true to this day. But because the senior petrologists in Bowen's early years were field based, he still had to prove two things: First, that crystals could really settle in magma, and second, that the latest, most siliceous melt could be reasonably extracted. He solved the first problem with his famous picrite experiment. Doubt continued, even to the extent that Wager felt the need to prefaced his work on layered intrusions with Bowen's picture of his thin sections through the picrite charge. That crystals settle in magma is, indeed, a fundamental truth. But the second problem, of collecting siliceous melt by fractional crystallization, haunted Bowen throughout his career. For example (1915): "We now return to the commonly observed association of diabase and its salic differentiate. During the period when the magma has largely crystallized, `the straining-off or squeezing out of the residual fluid magma" [Harker, 1909] is the most important process.' Later he specifically described the squeezing as possibly happening as a result of tectonic deformation, but then later still withdrew this idea when he realized the difference in time scales of deformation and crystallization. As voluminous volcanic (Kilauea) and magmatic systems (Skaergaard) were found to exhibit little silicic differentiates, he suggested (1947) that perhaps "cooling has been at such a rate that fractionation is ... unable to effect the separation of liquid at intermediate stages" or that the magma did not retain its volatiles and was too dry to stabilize phases like biotite and hornblende which are strong silica fractionators. The fact is the problem of how crystal fractionation takes place is still with us. Magmatic bodies initially free of crystals solidify to uniformity inward from the margins via ever-thickening crusts or solidification fronts (SFs). There are no or very tiny crystals in the interior, precluding conventional fractionation. And the residual melt only starts getting siliceous beyond a crystallinity depth of about 50percent where the matrix has significant strength. The processes to study are spatial ones that operate within SFs and vertically extensive and integrated mush column magmatic systems. It is time to move on from repetitive chemical studies of orchestral crystal fractionation.