All 35 samples were analyzed on a fully automated Rigaku 3370 spectrometer in the GeoAnalytical Laboratory at Washington State University (Table 1). The laboratory uses a rock powder Li₂B₄O₇ (1:2) mixture fused in a carbon crucible in a muffle furnace at 1000ºC to form a single glass bead that is used for the analysis of both major and trace elements. The detailed procedures, including the tabulated precision and estimates of accuracy (including comparison with data from other laboratories and techniques) are given in Hooper et al. (1993) and in Johnson et al. (1997).

The same 35 samples were also analyzed on a Sciex Elan (model 250) ICP/MS for the 14 REEs and 11 other trace elements (Table 1). Ba, Nb, Y, Pb, Rb, and Sr duplicate the XRF analyses. Of these, the values for Ba, Nb, and Pb by ICP/MS are preferred at the relatively low abundances present in these rocks, while the XRF values for Y, Rb, and Sr are the more precise. The analytical techniques employed for the ICP/MS data are described in detail by Knaack et al. (1994), which includes precision determinations for each element and comparisons with data from other laboratories as a measure of absolute accuracy.

Seven samples were analyzed at the University of Alberta for Sr and Nd isotopes under the direction of Dr. R.A. Creaser (Table 2). Sample preparation followed standard procedures, modified from Richard et al. (1976). Samples were dissolved in concentrated HF at ~120ºC for at least 48 hr. This was followed by HNO₃ treatment and conversion to chloride salts using concentrated HCl. Sr and REEs were separated from the whole sample by passing the sample through calibrated ion exchange columns. Sr separates were passed through a second calibrated ion exchange column before analysis. Nd and Sm were separated from the other REEs by passing through di-2-ethylhexyle-coated teflon powder columns, and were then analyzed.

Sr samples were analyzed on a VG Micromass 30 single-collector mass spectrometer. Results were standardized using National Bureau Standard 987 with an ⁸⁷Sr/⁸⁶Sr value of 0.71025. Nd and Sm samples were analyzed on a VG Micromass 354 fully automated multiple collector mass spectrometer, and samples were standardized to an internal standard, which ran true to La Jolla standard value of 0.511848, so no correction was made.

Oxygen isotope data were determined on plagioclase separates from 19 samples. The isotope ratios are reported in the familiar ? notation, relative to V-SMOW (Vienna-standard mean ocean water). Conventional fluorination techniques (Clayton and Mayeda, 1963) were used to extract O₂ from the powdered samples. ClF₃ was used as an oxidizing agent (Borthwick and Harmon, 1982). O₂ was converted to CO₂ on a resistance-heated carbon rod. The isotope ratio of this gas was measured on a gas source, isotope ratio, mass spectrometer in the geochemical laboratories at Washington State University. Raw data were corrected to an NBS-28 (African glass sand) value of 9.58, using the in-house standard MM-1 (Mica Mountain pegmatite quartz) which has a value of 12.9 on the NBS-28 scale. Replicate analyses of MM-1 show a standard deviation of <0.2.

A precursory survey of the compositions of the main mineral phases, plagioclase, pyroxene, and oxides, is presented in Table 3 (see "ASCII Tables" in the Table of Contents). Olivine pseudomorphs are also present in many samples but no fresh olivine, from which reliable compositions could be obtained, was found.

The mineral compositions were determined on an automated Cameca Camebax Microelectron microprobe with four wave-length dispersive spectrometers. The standards used were supplied by Charles Taylor (Na = albite #4, Mn = spessartine garnet, K = MAD-10 orthoclase, Mg = olivine #1, Ni = olivine #1, Cr = USNM 746, Johnston meteorite hypersthene, Ca = diopside #1, Al = kyanite #1, Ti = sphene #1, Si = diopside #1, Ba = NBS glass, K373.), using 10-s counts with a 4µ-beam at 20 KV and 13 nA. The probe's analytical data were corrected for atomic number, absorption, and fluorescence effects.