Chapter 4

How the mantle melts


McKenzie’s 1984 paper presents simple models of the process of partial melting, in which a melt and its residue form. By considering how a mixture of solid and liquid would deform – two phase flow – using fundamental physical laws, McKenzie finds general principles relevant to real geological situations. He establishes that a reasonable average temperature for the mantle could produce enough partial melting at ridges to build the ocean crust, and that a hot spot within the mantle under an ocean plate could produce the Hawaiian Islands, as Tuzo WilsonTuzo WilsonJohn Tuzo Wilson (1908-1993) worked at Princeton and was the first to suggest that island chains such as Hawaii arose as plates moved across a hot spot in the mantle below. In the early 1960s, maps of the magnetic anomalies on the ocean floor showed symmetrical patterns at mid ocean ridges, cut by discontinuities. With simple paper models, starting from Vine and Matthews’ work, Wilson suggested that transform faults were not later faults affecting ridges, but rather essential parts of the ridge system. Lynn Sykes at Lamont showed that earthquake motions along the faults matched Wilson’s model, and that transform faults away from the ridges were inactive. suggested. He also suggested that more complex models combined with geochemical modelling could account for variability of some trace elements in rocks formed by partial melting.

READ THE PAPER — McKenzie, D. 1984. The generation and compaction of partially molten rock. Journal of Petrology. 25, 713-65. 


Draft of figure 4 on the pore geometry of mineral grains, from the working file for “The generation and compaction of partially molten rock” (1984).