Garnet is one of the most important indicator minerals used in the search for kimberlites, and the presence of purple Cr-rich pyrope is an almost-certain indication that one is on the right track. Other garnets, however, are less reliable indicators of kimberlite, and some can be downright misleading.
In the Journal of Geology in 1975, J.B. Dawson and W.E. Stephens published a classification scheme for garnets from kimberlites, sorting the garnets into 12 groups based on their chemical composition. John Gurney later modified this scheme (drawing on earlier work by Gurney and Switzer in 1973) to distinguish, within the Cr-rich peridotite garnet population, between those from “ordinary” garnet peridotites (“G-9” garnets) and those with close similarities to the unique population of Ca-poor, Cr-pyrope garnets occurring inside diamonds (termed “G-10” garnets). Gurney’s method is useful in determining “good” kimberlite targets in terms of diamond potential. In the explosion of recent kimberlite/diamond exploration activity, especially in Canada, the Dawson-Stephens classification scheme has been applied to garnets acquired in regional exploration programs — typically garnets from tills, soils and alluvial occurrences. Identification of garnets that correspond to various Dawson-Stephens categories (“G-3″,”G-4”, or “G-5” garnets, for example) has been cited as evidence of a kimberlite source for minerals in the heavy mineral samples.
To apply this method to garnets not known to be from a kimberlite, however, ignores the facts that (1) the method was proposed to classify “garnets from kimberlite” and (2) garnets from ordinary igneous and metamorphic crustal rocks will fall into most of the 12 categories when classified using the flow-chart provided by Dawson and Stephens. A non-exhaustive survey of published compositions of garnets revealed that garnets corresponding to groups 1, 3, 4, 5, 6, 7, 8 and 9 occur in crustal rocks.
The most abundant crustal garnets (those from amphibolite-grade pelitic and basic schists and gneisses) typically correspond to group 5. This group is particularly misleading because most of the original data used to define this group are now known to be either garnets from crustal granulites (Somerset Island Fe-rich garnet xenocrysts) or unusually Fe-rich, low-T eclogites from non-kimberlite pipes in Arizona and Utah that are more similar to crustal eclogites than they are to mantle eclogites. Mantle examples of group 5 garnets do exist, but they are exceedingly rare and unlikely to be encountered in a kimberlite exploration program.
So where does this leave the usefulness of garnets in kimberlite exploration? First of all, one must examine the garnet compositions with more care than is involved in simply determining their Dawson-Stephens group number. Mantle eclogite garnets can usually, though not always, be readily distinguished from crustal almandines in groups 3 to 6 by the more Fe-rich and TiO2-poor character of the latter. Overlap between crustal grossulars and uvarovites and those from the mantle (typically groups 7 and 8) is extensive, but the mantle examples are usually more magnesian and these garnets are not common. The combination of high MgO and TiO2 in groups 1 and 2 are good hints of a mantle source, as is high MgO coupled with high Cr2O3 in the peridotite-derived garnets of groups 9-12. Crustal examples of group 9 garnets are usually fairly Cr-poor (< 2-3 wt% Cr2O3). Purple Cr-rich pyropes remain the best kimberlite indicator garnet. Few kimberlites in the world have more eclogitic (orange) garnets than peridotitic (purple) garnets. The Dawson-Stephens garnet classification method was meant to be used for garnets from kimberlite so as to understand better their origins. If it is used for garnets from other geologic environments, its limitations must be recognized or its application can lead to unfounded conclusions, and perhaps costly decisions.
— Daniel Schulze is with the department of geology at the University of Toronto’s Erindale College, Mississauga, Ont.
Be the first to comment on "GUEST COLUMN — Caution advised in interpreting garnet data"