A combined field and laboratory study of plutonic rocks from the San Jacinto Mountains of southern California was conducted in order to investigate the nature and origins of strontium and oxygen isotope heterogeneities within batholithic rocks.
Geological mapping has allowed differentiation of three major and many minor masses of plutonic rock. Contacts between units are offset up to 6 km by faults of the Neogene San Jacinto fault system, which has a total right-lateral displacement of 29±1 km. The early small intrusives range from olivine gabbro through granite. They were intruded by three larger plutons of relatively homogeneous biotite-hornblende-titanite tonalite. The oldest major intrusive unit, Unit I, is an elongate body of dimensions 40x8 km. Before complete solidification it was intruded by Unit II, an irregular tabular mass 25 km long and a few kilometers wide. Unit III, in turn, intruded Unit II before it was completely solidified, producing a roughly rectangular mass 20x12 km that appears to funnel in downwards.
Mineral foliations and banding, schlieren, and xenolith orientations within each unit usually parallel the nearest contact. Alignment of foliations and apparent flow-sorting and scour features seem to reflect flow patterns within each chamber. Mlfic synplutonic dikes (quartz diorite to tonalite) intruded into the tonalites, and were commonly broken up and redistributed as linearly extensive xenolith trains. From these relationships it is interpreted that: 1) magma adjacent to pluton walls had considerable yield strength, as it could fracture to allow dike emplacement; 2) magmatic flow adjacent to pluton walls was capable of moving material some distances (up to km) to create the xenolith trains; and 3), that the dikes are potentially the feeders through which material was added to the inflating magma chambers. Each major tonalite unit spans a limited compositional range of from mafic tonalite (Colour Index > 15) to low-K granodiorite (Colour Index < 10). Volumetrically minor felsic differentiates extend the compositional range through to granodiorite. Units I and II average slightly more mafic overall compositions than does Unit III. All units are comprised of plagioclase (An30-40) [50-55%], quartz[20-30%], K-feldspar[1-8%], biotite (10-15%], hornblende[0-5%], titanite[0-2%] and accessory zircon, apatite, allanite and ilmenite. Variations in mineral abundances are geographically systematic only within Unit III, which grades from marginal mafic tonalite to central low-K granodiorite.
Mineral compositions throughout the major tonalites are remarkably uniform. The An content of the bulk plagioclase falls from An40 ± 1 in the most mafic tonalites to An30 in low-K granodiorites; Mg/(Mg + Fe) of biotite and hornblende drop similarly from 0.44 to 0.36. The entire observed range of plagioclase compositions within the major tonalites is An44 to An25 (and to An47 in mafic xenoliths). The sole opaque mineral is almost pure ilmenite. This homogeneity of mineral compositions implies remarkable stability of physico-chemical conditions throughout crystallization of each unit.
Major and trace element abundances reflect the general homogeneity of these rocks. Most have SiO2 in the range 63-68 wt.%; minor felsic differentiates extend to 71 wt.%. 60% of analyzed samples from Unit III fall in the restricted compositional range 66-68 wt.% SiO2; the majority of samples from Units I and II are more mafic than this. Major elements (excepting K2O) define excellent linear arrays on Harker diagrams. K2O shows a diffuse curvilinear pattern. Trace elements generally considered "compatible" (including the transition metals), and Sr also define linear arrays on Harker diagrams. Other trace elements, especially Ba, Rb, Pb, Th, U and REE show more complex behaviour. "Mafic" tonalites, (<65.5 wt.% SiO2) have simply covarying trace element endowments. Minor felsic differentiates, collected on the basis of field evidence for in situ fractionation, have higher Si, K, Rb, Ba, U and Th. "Normal" tonalites (66.5 < SiO2 < 70.0) have trace element and K endowments intermediate between the mafic tonalites and the felsic differentiates.
Mafic tonalites, comprising about half the exposed rocks, crystallized from liquids which derived their geochemical characteristics before injection into the high-level magma chambers. The minor felsic differentiates are considered end-products of fractional crystallization within the magma chamber; the "normal" tonalites are interpreted as crystallizing from liquids of intermediate character, i.e., mixtures of "primitive" and fractionated liquids.
Measured primary δ18O values vary from +9.0 to +10.6. Metasedimentary country rocks have δ18O values of +11.5 to +13.5. Exchange of oxygen between plutons and country rock is minor and limited to narrow border zones. Within Unit III primary δ18O correlates with position. A marginal zone of variable values (+9.0-+10.0) gives way to regularly increasing values (+10.0-+10.5) inwards. Primary δ18O correlates with Colour Index. Within the central part of Unit III the observed range in δ18O values can be explained by crystallization of modally variable rocks from a liquid of constant 18O/16O (to±0.2 per mil). The δ18O values of the more mafic marginal rocks (and of mafic rocks from Units I and II) also correlate with Colour Index; lower SiO2 rocks have lower δ180. This correlation cannot simply result from varying mineral abundances, but must reflect variations in δ180 values of the liquids from which these rocks crystallized.
Calculated initial 87Sr/86Sr (Sri) varies substantially among rocks from each major tonalite unit (Unit I: 0.7060-0.7076; Unit II: 0.7060-0.7074; Unit III: 0.7058-0.7073). These variations appear geographically regular at the kilometer scale within each pluton. The complex patterns, however, differ fundamentally from the general regular west-to-east increase in Sri reported for the batholith (Early and Silver, 1973), and observed in the small early intrusives from the San Jacinto Mountains (0.7057-0.7077). Sri within these rocks shows no identified correlation with other geochemical and petrological parameters.
The Sr isotope data indicate that melt production, transport, and crystallization processes combined were not capable of completely homogenizing initial variations in Sri within the liquids from which these rocks crystallized. This further implies that either the time scale for convection was large compared to that for crystallization, or that the length scale for convection was small compared to the size of the plutons. Field evidence suggests considerable flow within the magma chamber; estimation of rheological parameters suggest that flow was within a laminar flow convective regime.
The combined observations are compatible with crystallization from an intermittently recharged, continuously fractionating system. Recharge tended to buffer both the thermal and chemical properties of liquids within the magma chamber; it gave a mechanism for introducing isotopic variations that are incorporated into this continuously crystallizing system. The mafic dikes are suggested to be conduits through which some of these liquids were injected into the various magma chambers. Chemical buffering by continued recharge is also compatible with the observation that the majority of these rocks have geochemical features interpreted as resulting from the action of processes prior to injection of liquids into the high-level magma chambers.
Rocks with low Sri (0.7058-0.7068) generally have intermediate δ18O values (+9.7-+10.3), and fall near the low-δ18O side of the batholithic trend defined by Taylor and Silver (1978). Rocks with high Sri (>0.7072) cover the entire observed range in δ18O values (+9.0-+10.6), and overlap the field defined for the San Jacinto - Santa Rosa Mountains block by Taylor and Silver. These data require involvement of material from three isotopically distinct source materials in the generation of these rocks. Two of these components (one with low Sri, low δ18O; one with high Sri, high δ18O) are common to the bulk of the batholith to the south and west. The third (high Sri, low δ18O) seems unique to the San Jacinto - Santa Rosa Mountains block; its relative importance within the San Jacinto rocks appears to correlate negatively with SiO2, suggesting that it was associated with relatively mafic liquids. The oxygen isotopic data imply that as much as 35% of this component may be present in some rocks. This component has isotopic and inferred geochemical characteristics compatible with old, slightly enriched (in Rb relative to Sr) subcontinental lithosphere. The low-Sri, low-δ18O component appears to be either (or both) normal depleted mantle or (subducted) oceanic crust. Tile oxygen data imply that the third (high-Sri, high-δ18O) component has had a prior history at the Earth's surface; it could be either sediment, or igneous material altered at low temperatures such as hydrothermally altered oceanic crust. Geochemical features (K, Rb, LREE abundances) appear more compatible with sedimentary material.
These data are compatible with, but do not prove, a model for this source region as being a mixture of normal depleted mantle, oceanic crust, old slightly enriched "subcontinental lithosphere", and subducted sediment. This model source contains variously 0-35% (oxygen atom basis) subcontinental lithosphere, up to 25% sedimentary component, and apparently requires material of both basaltic (oceanic crust) and depleted mantle composition to balance isotope systematic systematics.
Tile combined data show 1) that the source volumes for the batholithic rocks were heterogeneous at the scale of hundreds of meters or greater, 2) that the effects of these source heterogeneities were at least partially preserved throughout melt production, transport, and crystallization, and 3) that the net effect of a persistent recharge-fractional crystallization process within the magma chambers was to buffer the composition of the bulk of the rock near that of the early-crystallizing solids.
