Do all Archaean TTG rock compositions represent former melts

This paper seeks to argue one simple point. That the diversity in chemical composition in Archaean granitoids of the tonalite-trondhjemite-granodiorite suite (TTG) is primarily the product of crystal fractionation, specifically plagioclase and hornblende fractionation. If this is the case TTG compositions from a given suite can be reduced to the least fractionated samples which will represent the primary melt composition(s).
There are two important implications of this proposal. (1) The subdivision of TTGs into high-, medium- and low-pressure types on the basis of their trace element concentrations may be an artefact of crystal fractionation, and (2) if hornblende fractionation is important in TTG genesis then this argues for an amphibolitic source rather than an eclogitic source and greatly simplifies the number of geodynamic options available in the genesis of TTG magmas.
The argument is supported by two case studies from TTGs in the lower and middle crust of the Lewisian Complex, NW Scotland and the lower crust of the Northern Marginal Zone of the Limpopo Belt, Zimbabwe. In both case studies there is strong evidence to show that Archaean TTGs can experience significant hornblende and plagioclase fractionation.
As a working rule therefore all TTG suites should be examined for the possible effects of fractional crystallisation and geochemical models for their origin should be based upon the least fractionated samples.