TEPHRACHRONOLOGY

Mass spectrometric data from tephra samples are summarized in Table T4 and presented graphically in Figure F5 Complete experimental data and plots are available in electronic files from the journal or on request from the authors. Fitted Ar isotopic ratios from step measurements are used in two ways. Assuming that initial sample Ar compositions were atmospheric (initial ⁴⁰Ar/³⁶Ar = 295.5), step ages are plotted against cumulative percent ³⁹Ar released, as age spectrum, or plateau diagrams. In addition, isotope correlation diagrams (⁴⁰Ar/³⁶Ar vs. ³⁹Ar/³⁶Ar) are examined for collinear step compositions; the slope is equivalent to age since closure, and the ⁴⁰Ar/³⁶Ar intercept reveals the initial Ar composition of the system (rock or mineral). We accept an apparent age as an accurate estimate of the sample crystallization age if several statistically testable conditions are met (Dalrymple et al., 1980; Pringle and Duncan, 1995), namely (1) a well-defined, mid- to high- temperature plateau is formed by at least three concordant, contiguous steps representing >50% of the ³⁹Ar released; (2) a well-defined isochron exists for the plateau step Ar compositions; (3) the plateau and isochron ages are concordant; and (4) the isochron ⁴⁰Ar/³⁶Ar intercept is atmospheric composition.

Most of the sample ages presented in Table T4 meet the criteria listed above. Total gas ages are calculated by recombining all steps from each sample and are roughly equivalent to conventional K-Ar ages. Plateau ages (2 uncertainties) are the mean of from three to eight step ages, representing >50% of the total sample ³⁹Ar, weighted by the inverse of variance. Sample 183-1138A-12R-3, 5-7 cm, glass (A1) produced a saddle-shaped age spectrum (Fig. F5A) with a three-step plateau and older ages at the lowest temperature and highest temperature steps, suggestive of small amounts of excess (nonatmospheric) ⁴⁰Ar trapped in the quenched glass at the time of crystallization. The corresponding three-step isochron produced a significantly younger age (5.99 ± 0.71 Ma) and an initial ⁴⁰Ar/³⁶Ar intercept greater than the atmospheric value (295.5). The isochron age is preferred because of the evidence for excess ⁴⁰Ar and because it fits stratigraphically with the much better resolved biotite age from Sample 183-1138A-13R-1, 22-24 cm (A2), in the underlying core. This second sample produced a tight plateau age (six of seven steps; 99% of the total gas released) and concordant isochron age (Fig. F5B). We accept this age as a reliable estimate of the eruption and immediate deposition events. Sample 183-1138A-15R-1, 73-75 cm, glass (A3) also showed evidence of undegassed, excess ⁴⁰Ar, in the form of old ages at high-temperature steps (Fig. F5C). The lowest three temperature steps produced a poorly resolved plateau, but the isochron based on all steps indicated a younger age and a nonatmospheric initial Ar composition. The isochron age, although imprecise, appears to be the better estimate of the age of crystallization and deposition of this ash layer. Two samples were analyzed from interval 183-1138A-19R-1, 110-112 cm, with compositions of glass (A5) (Fig. F5D) and biotite (A4) (Fig. F5E). These produced concordant plateau and isochron ages with atmospheric initial Ar compositions. We use the weighted mean of the isochron ages (11.45 ± 0.22 Ma) as the best estimate of the age of this horizon.