In order to remove impurities, natural illite is often subjected to chemical pretreatments before analysis, the assumption being that such pretreatments have little or no effect on the composition of illite. Bulk analyses are sometimes reported before and after pretreatment (Aja et al. 1991) but, because the contribution of the impurities to the bulk composition is unknown, the possible effects of chemical pretreatments on the composition of illite cannot be assessed. Recently, Rosenberg and Hooper (1996) have shown that analytical transmission electron microscopy (ATEM) techniques, which exclude impurities, can be used to determine the bulk compositions of relatively pure illite by averaging the analyses of at least 50 crystals per sample. This approach can be applied to impure illite in order to determine the effect of chemical pretreatments on the composition of natural illite.

We sampled individual growth rings from three ancient remnant bald cypress (Taxodium distichum) trees from a massive buried deposit at the mouth of the Altamaha River on the Georgia Coast to determine the best technique for radiocarbon (14C) dating pretreatment. The results of our comparison of traditional ABA pretreatment and holocellulose and α-cellulose fractions show no significant differences among the pretreatments (<1 sigma) thereby suggesting that ABA pretreatment will prove sufficient for the development of a high-resolution 14C tree-ring chronology based on these ancient bald cypresses which will indicate whether the U.S. Southeast is subject to a regional radiocarbon offset.