In reality, about the only circumstance I can imagine in which I
might assign a facies to a single sample would be in a teaching lab
with students where one might ask (many actually have asked) "what
facies would this rock belong to?". In any kind of research setting,
we all work with multiple samples for which we have detailed
information from thin sections and maybe even probe analyses that
allow us to be much more precise. Discussions of "facies" are
typically in tectonics rather than petrology papers, where the
broadness of scale generally blurs single-sample detail, and in my
opinion this is the important point.
Thanks, Bob, this is what I meant in my previous mail. There is always a research side and a teaching side to these things. When you teach introductory courses or go on field trips discussing basic petrography and interpretation of metamorphic rocks, facies can be used as a quick-and-dirty approach to P and T. In this context, I would expect a student to tell me that a staurolite-andalusite schist indicates low-pressure amphibolite facies conditions ("most likely" could be added, to exercise the necessary caution). In a field trip situation, there is, of course, a better chance to look at a variety of rocks.
Re- the fossil analogy of Bruce: I don't think that modern palaeontology would be content with having a detailed stratigraphic table with fossil zones etc. entirely detached from actual (quantitative) ages. I had the impression that these guys were quite busy attaching radiometric ages to all their boundaries. Clearly, the zones and boundaries are still defined by fossils, but the numbers are not meaningless, even though they will be refined as time goes on. We can also turn the whole thing around and say: an age of 40 Ma means Eocene (even though the Eocene is defined by fossils for all I know), just as 800 C and 6 kbars means granulite facies conditions.
School of Geological & Computer Sciences
University of Natal