Note: This paper is dedicated to Aaron and Elizabeth Waters on the occasion of Dr. Waters' retirement.
New isotopic analyses of lead are reported for 40 volcanic rocks from the Cascade Mountains of Washington, Oregon, and California. Strontium isotopic compositions were also determined in 33 volcanic rocks from the same area. In addition, lead and strontium isotopic data are given for feldspar and whole-rock samples of prominent varieties of crystalline basement rocks from northern Washington. The Sr87/Sr86 values of the volcanic rocks average 0.7037, with no significant difference between andesite and high-alumina basalt. The ratios exhibit no measurable correlation with strontium concentrations over a range of 200 to 1,500 ppm. Strontium in the Cascade rocks is slightly more radiogenic than that in oceanic ridge basalts, but less radiogenic than in most continental basalt, including the Columbia River basalt. Comparisons with published data for strontium in Pacific Ocean sediments and coastal graywackes in Oregon indicate much higher Sr87/Sr86 values in the sedimentary rocks when compared to the Cascade volcanic rocks, thus ruling out anatectic models involving large sediment contributions.
Lead isotopic compositions from the volcanic rocks are variable but tend to be rather constant at a single volcanic center. A noteworthy feature of the lead data are the higher Pb207/Pb204 values (or larger µ values in a model age diagram) for the Cascade volcanic rocks compared to published Pb207/Pb204 values and µ values for Pacific Ocean tholeiitic basalt. The Pb206/Pb204 values from the Cascade volcanic rocks are higher than in oceanic sediments from the northeast Pacific Ocean but are lower than the average ratio for three samples of Oregon coastal sedimentary rocks. There is no demonstrable difference between lead isotopes in high-alumina basalt and andesite at particular volcanos, although the lead data are more variable than the strontium isotope data. Strontium in crystalline basement rocks from northern Washington is substantially more radiogenic than that in the calcic and calc-alkalic volcanic rocks of the Cascade Mountains. The same is true of lead in some, though not all, of the basement rocks. The isotopic data are utilized, along with other trace-element data to test models of andesite genesis. The isotopic compositions of lead and strontium in andesite and high-alumina basalt from the Cascade Mountains are consistent with derivation of the two lava types from a common source, or with derivation of andesite by differentiation of high-alumina basalt magma. The data do not support models in which the Cascade calcic and calc-alkalic magmas were formed by melting of Oregon coastal eugeosynclinal sedimentary rocks, or by anatexis of basement rocks similar to those now exposed near volcanic centers in northern Washington. Various subduction zone melting models are also considered. Mixing models, whereby radiogenic strontium and lead are added to magmas containing lead and strontium with isotopic compositions similar to those in oceanic ridge tholeiite, do not fit the observed data satisfactorily because values of Pb206/Pb204 in Pacific Ocean sediments are too low to account for the Pb206/Pb204 values in the Cascade lavas. Mixtures of eugeosynclinal sedimentary rock and oceanic ridge tholeiite can probably account for the Cascade volcanic rock isotopic data in a general way, but the uniform isotopic composition of strontium in the lavas over a wide range of concentrations is hard to understand in terms of such a model. The same is true for strontium in the oceanic sediment mixing model. The uniform isotopic composition of strontium in the Cascade lavas, irrespective of strontium concentration and geologic setting, argues against contamination of magma with radiogenic strontium during ascent. A similar argument can be made for lead.
The best explanation of the isotopic and trace-element data appears to be a multistage model in which orogenic andesite is derived from three or more stages of partial melting of mantle material in which crustal materials play an insignificant role. This model best explains the uniform Sr87/Sr86 values in the lavas over a wide range of strontium concentrations. It can also explain why strontium in lavas extruded at continental margins, such as in the Cascade Mountains and in Japan, has the same isotopic composition as strontium in lavas from the Mariana arc system in the Pacific Ocean.
Pb206/Pb204 and Pb208/Pb204 values in Cascade lavas may show a correlation with crustal thickness, but this does not prove that crustal contamination is responsible for the trend. The lead and strontium data exhibit no correlation with the Quartz Diorite Line of Moore.