Journal of A rchaeological S cience (1996) 23, 283–287 The Effect of Firing Temperature on the Elemental Characterization of Pottery James W. Cogswell, Hector Neff and Michael D. Glascock Research Reactor Center, University of M issouri, Columbia, M O 65211, U.S.A. (R eceived 13 October 1994, revised manuscript accepted 7 February 1995) Neutron-activation analysis (NAA) of clay test tiles fired to temperatures from 100 to 1100C shows that with the possible exception of bromine, no volatilization of 34 elements occurs. Thus, there is little reason to consider firing temperature a potential confounding factor in ceramic compositional studies.  1996 Academic Press Limited Keywords: NEUTRON-ACTIVATION ANALYSIS, CERAMICS, FIRING TEMPERATURE, ELEMENT VOLATILIZATION. Introduction T his study addresses the effects of pottery firing temperature on bulk elemental concentrations. Several researchers have provided information on this question, but their results were mixed. Kilikoglou, Maniatis and Grimanis (1988) found that firing had no effect on 20 elements in fired clay samples. Attas, Ya ffe and Fossey (1977) reached the same conclusion for 19 elements. On the other hand, Perlman and Asaro (1969) noted that bromine was volatilized from their standard pottery clay and from a natural clay. Rye and Duerden (1982: 63–64) reported that firing up to 815Ca ffected bromine, chlorine, and sulphur concentrations. Poole and Finch (1972: 84) found that sulphur and possibly strontium are a ffected either by firing or burial. Storey (1988) detected an unexplained loss of yttrium in fired clay samples. With the exception of Perlman and Asaro’s (1969) work on standard pottery clay, acceptance of these previous studies is limited primarily by their use of natural clays. Sample heterogeneity was reduced to some extent in all of these studies, but volatilization of significant amounts of carbon compounds and water may have obscured detection of minor- and trace- element volatilization. Samples were analysed in their unfired state and at a limited number of firing tempera- tures, so analytical error combined with sample in- homogeneity may have masked subtle changes induced by volatilization. The study reported here was designed to eliminate these potential problems. Methods A 2kg sample of Ohio Redart clay from a single bag was used for this study. This commercial clay was selected for use because of its low organic and water content (Table 1), and because its chemical composi- tion is already well known as a result of its use as a check standard at the Missouri University Research Reactor (MURR). Deionized water was added to the clay sample until a uniformly workable consistency was attained. (Plastic gloves were worn through all phases of sample preparation to prevent contamina- tion.) Test tiles were formed by pressing wet clay into plastic petri dishes measuring 7 mm deep by 60 mm diameter. The test tiles were air dried while covered by paper towels, removed from the plastic forms, and then further dried in an oven at 100C for 24h. Batches of three test tiles were fired to temperatures ranging from 150C to 1100C in increments of 50C using a Fisher Isotemp model 495A electric furnace. Heating rate for each batch was 15C min 1 to the maximum tempera- ture, a 60 min ‘‘soaking’’ at the desired temperature, and an average of 2h of cooling to ambient tempera- ture. Five 2 g samples from each three-tile batch were powdered with an agate mortar and pestle, placed in glass vials, and held at 100C until NAA samples were prepared. Sample irradiation and analysis were conducted at MURR utilizing procedures described by Glascock (1992). In brief, 200 mg of each sample was weighed into polyethylene vials, irradiated for 5 s, allowed to decay for 25 min, then counted on a high-resolution T able 1. M ineralogical analysis of Ohio R edart clay* Illite 50% Kaolinite 10% Montromillionite/illite 30% Quartz 10% Total water 5% Organics 0·5% *From Resco Products, Inc. product information. 283 0305-4403/96/020283+ 05 $12.00/0  1996 Academic Press Limited