More than 150 years ago, R.W. Bunsen (1851) first proposed that magmas of contrasting composition could mix to produce a suite of intermediate hybrids. Subsequent research, summarized by Pitcher (1997) and Wilcox (1999), has elevated the status of magma mixing and demonstrated its importance in the evolution of many igneous suites from a variety of magmatic provinces. It has been shown to be a fundamental petrogenetic process in many sub-volcanic environments, often triggering eruption (Murphy et al. 1998; Bachmann et al. 2002), and at deeper crustal levels forming dynamic, open-system magma chambers (Barnes et al. 1986).

The thermal, physical, and compositional gradients that result when mafic and felsic magmas are juxtaposed produce an array of spectacular field relations and a variety of mineral-scale disequilibrium textures. Examples are widespread and have been described from the British Tertiary Igneous Province (Harker 1904; Elwell 1958; Wager and Bailey 1953), the Irish Caledonides (Baxter and Feely 2002), and the Sierra Nevada batholith (Reid et al. 1983, 1993; Frost and Mahood 1987). Hibbard (1991, 1995) described a number of disequilibrium textures produced during magma mixing/hybridization that can be utilized to indicate that coeval mafic and felsic magmas have interacted to varying degrees. In addition, geochemical and isotopic data (Holden et al. 1987; Reidet al. 1983, 1993; Bateman 1995) also have been used to discriminate between magma mixing and other competing magma chamber processes such as fractional crystallization or assimilation.

The coastal Maine magmatic province (CMMP) hosts more than 100 mafic and felsic plutons (Fig. 1). Many of these preserve abundant field and geochemical evidence to indicate that mafic and felsic magmas have interacted extensively (Hogan and Sinha 1989). In these "fossil", high-level magma chambers, such as the Pleasant Bay Complex (Wiebe 1993a), field relations indicate that mafic magma was episodically injected into the crystallizing felsic magma chamber. Evidence for mixing and mingling of coeval magmas has also been preserved in the volcanic sequences of the CMMP such as the Cranberry Island series (Seaman et al. 1995). However, many other plutons of the CMMP, particularly the "younger" Devonian granites (e.g. the Lucerne, Deer Isle, and Mount Waldo plutons, see Fig. 1), lack obvious field evidence that their crystallization was interrupted by episodic injections of mafic magma. In this paper we describe the textural assemblage and crystallization history of the Mount Waldo granite, of which only a ~ 350 m thick section (most likely toward the top of this intrusion) is currently exposed. We suggest that, even though direct field evidence is lacking, magma mixing played an important role in the evolution of the Mount Waldo pluton. In addition, we discuss the origin of these mingling-related textures in other CMMP intrusions where there is unequivocal evidence for magma mixing/mingling at the present level of erosion. Comparisons between known magma mixing systems and suspected or "cryptic" systems within the same magmatic province could enable us to elucidate magma chamber processes within the "younger" granites of the CMMP, even though only their upper structural levels can be examined directly.

The coastal Maine lithotectonic block is composed of a series of fault-bounded, NE-trending terranes (Osberg et al. 1985). These peri-Gondwanan terranes were accreted to the Laurentian margin of the North American continent during the Middle Devonian, at the peak of the Acadian orogeny (Stewart et al. 1995; Osberget al. 1995; Bradley et al. 2000). Emplaced into the coastal block are over 100 post-tectonic intrusions that "stitch" these terranes together. Chapman (1962) first recognized that the coastal Maine plutons form a distinct magmatic province, which he termed the Bays-of-Maine igneous complex. He subdivided this complex into an older phase of layered gabbro and granite and a younger phase of coarse-grained granite, a subdivision that even with new radiometric ages still has some relevance. Hogan and Sinha (1989) re-named the province the coastal Maine magmatic province (CMMP), recognizing that these two phases are petro-tectonically related.

The CMMP is composed of a number of intrusions that are diverse in composition, complexity, and age. They range from gabbro to diorite, and granodiorite to granite, with two-mica peraluminous granite also present. Many are composite, zoned intrusions whereas others have crystallized from a single magma batch. The CMMP intrusions range in size from small stocks (< 1 km²) to large sub-batholithic bodies (> 1500 km²). Hogan and Sinha (1989) subdivided the plutons into four types based on the degree of interaction between mafic and felsic magmas observed within each intrusion as indicated by the abundance of enclaves and composite dikes. The plutons vary in age from Early Silurian (Wenlockian) to Late Devonian (Frasnian), i.e. 425 – 370 Ma (Bradley et al. 2000; Stewart et al. 1995; West et al. 1995). As Chapman (1962) noted, the younger phase is dominated by coarse-grained granite plutons some of which (e.g. the Lucerne and Deblois plutons) form the largest intrusions of the CMMP. Therefore the assembly of the CMMP represented some 50 Ma of crustal recycling, supply of mantle-derived mafic magmas, and a significant volumetric addition to this part of the North American continent.

Wiebe (1993a, b) described a specific category of CMMP intrusion that he named mafic and silicic layered intrusions (MASLI). Examples such as the Pleasant Bay, Cadillac Mountain, and Vinalhaven complexes display compelling field evidence indicating that mafic magmas were injected into, and flowed along the aggrading floors of these felsic magma chambers. Fault block rotation during Mesozoic extension, combined with Cenozoic glaciation, provides almost complete cross sections within these complexes, from their lower replenishment zones to their upper granitic/volcanic carapaces. At lower structural levels there is ample evidence for the injection of mafic magmas with accompanying magma mixing and mingling. Mafic pillow mounds (comparable to pillow basalts), granitic pipes intruding upwards into overlying mafic material, development of hybrid magmas of intermediate composition (that gave rise to enclaves), abundant disequilibrium textures (quartz ocelli and rapakivi feldspar), and large-scale compositional layering are common in the MASLI. Such features suggest that these magma chambers were dynamic, and were fuelled, both thermally and compositionally, by mafic injections.

The Mount Waldo granite (MWG) crops out over an area of ~ 150 km² at the northern end of Penobscot Bay (Fig. 2). Its contact with the surrounding country rock has been well mapped, along with a 2 – 3 km wide andalusite-bearing contact aureole (King 1977; Wones 1991; Stewart 1998). Hornblende geobarometry calculations (Lux et al. 2000) yielded a range of pressures depending on the calibration used: 3.1– 3.8 kbars (Hammarstrom and Zen 1986); 3.2– 3.9 kbars (Hollister et al. 1987); 3.7– 4.3 kbars (Schmidt 1992). Although the results vary, they suggest that the MWG was most likely emplaced at around 3.5 kbars ( ~12 kms). Gravity studies (Sweeney 1976) indicate that the MWG is < 7 km thick and is a steep-sided, funnel-shaped intrusion. U-Pb zircon data (Stewartet al. 1998) yielded an age of 371± 2 Ma, which is in close agreement with the Rb-Sr age of 374 ± 3 Ma (J.P. Hogan, unpublished data).

The MWG is typically a medium grey, coarse-grained, seriate to porphyritic, biotite ± hornblende granite with colour index < 15. It contains abundant large K-feldspar crystals (up to 6 cm) many of which, but not all, are mantled with plagioclase to form rapakivi texture. Accessory present minerals are titanite, allanite, zircon, apatite, and magnetite. Minor amounts of molybdenite occur along some joint surfaces (M. Yates, personal communication, 2000). The granite commonly displays a foliation, defined primarily by the feldspar, and to a lesser extent, by biotite. Apart from some minor undulatory extinction in quartz there is little evidence of any solid-state deformation, and we agree with Trefethen (1944) that the foliation is principally magmatic in origin. However, the anisotropy of magnetic susceptibility study of Callahan and Markley (2003) indicated that a small increment of crustal strain occurred during the late stages of crystallization of the MWG.

The petrography of the MWG varies significantly within the pluton and at the outcrop scale. Many parts of the granite are distinctly heterogeneous with variable textures and mineral-ogy. Two main units of the MWG can be mapped in the field – one, the more typical MWG described above, forms the major part of the intrusion. The other, here termed the Mt. Ephraim unit, consists of granite that is more equigranular in texture and generally lacks the large feldspar crystals that are abundant in the main unit. It has a lower colour index and crops out in the W and NW parts of the pluton. Field relationships between the two units are unclear; however, the Mt. Ephraim equigranular granite is also present as dikes that intrude the main MWG unit, suggesting that it may have been a slightly later pulse.

Smaller scale textural variants, aplites, magmatic enclaves, coarse-grained feldspar domains, and mafic layering or schlieren are also observed in the MWG and contribute to its overall heterogeneous appearance. The enclave suite ranges in composition from quartz diorite, to tonalite and granodiorite, with more leucocratic varieties also common (Gibson et al. 2001). The enclaves have magmatic textures, are typically fine- to medium-grained with equigranular and porphyritic textures, and range in size and morphology from micro-enclaves and mafic clots (< 2 cm in diameter) to leucocratic enclaves which reach ~ 1 m in diameter. The coarse-grained feldspar domains (e.g. Fig. 4b) are composed of many different feldspar varieties (see section 4a below) along with interstitial quartz, and minor amounts of biotite. They commonly enclose enclaves of all compositions and are ubiquitous in the MWG. Similarly, the mafic layering or schlieren occur throughout the pluton. They are mineralogically similar to the granite but have higher concentrations of biotite and hornblende, and up to 12% accessory minerals. They form planar layers, "ladder dikes", trough, and ring structures (Lux and Gibson 2000), and are comparable to layering described in other granitic bodies by Reid et al. (1993), Wilshire (1969), and Weinberg et al. (2001). These authors attributed the formation of mafic layering to magmatic processes, in particular crystal settling and/or flow segregation. The petrography and field relationships of these textural variants are summarized in Table 1, and modal data (Fig. 3) indicate the wide range of "granitic" compositions evident within the MWG and demonstrate its overall heterogeneity.

In contrast to many other plutons of the CMMP, the MWG lacks appreciable amounts of mafic material at the present level of erosion. In fact, only one mafic dike has been mapped (Trefethen 1944) in the entire intrusion, unlike the other granites of the CMMP such as the Deer Isle Complex, where composite dikes are present. However, the textural/mineralogical assemblage observed in the MWG suggests that mafic magma may have played a more important role in the evolution of this pluton than its absence would seem to imply.

A variety of textures, typical of the mixing and mingling of coeval mafic and felsic magmas, are evident at the field and microscopic scales in the MWG including its petrographic variants described above. These textures can be grouped into two categories: a) those pertaining to the feldspar population, and b) those observed in the enclave suite. When considered collectively, these textures suggest that strong thermal, compositional, and physical gradients were operating in the Mount Waldo magma chamber.

The assemblage of textures described above, along with field relations observed in the MWG, indicate a definite role for the mixing and mingling of disparate magmas (Hibbard 1981) and the formation of hybrids as represented by the enclave suite. Many of the feldspar textures indicate they were formed in two stages. The dissolution textures suggest that initially, re-heating resorbed the grains. Subsequently, re-growth occurred where there was a supply of material of differing composition to form the overgrowth rims. However, comparable textures have been ascribed also to decompression associated with magma ascent and eruption (Cherry and Trembath 1978; Nekvasil 1991). Importantly, in the MWG feldspars with variable textures are observed commonly juxtaposed in the same hand sample or thin section (see Figs. 4c, 6b, 6c).

The feldspar population of the MWG consists of K-feldspar, rapakivi feldspar, and plagioclase phenocrysts alongside boxy/spongy cellular plagioclase and variably zoned plagioclase grains. The latter, typically with resorbed, high An cores, are observed alongside weakly zoned crystals. No systematic variation of this feldspar population has been recognized with structural level or position in the Mount Waldo intrusion. However, if these textures had originated by a decompression event, some systematic variation should be evident. A decompression origin would also require two populations of crystals – those that were present before decompression and those that crystallized afterwards. It is also difficult to explain the mixture of inherited phenocrysts within the magmatic enclaves and across the host/enclave contact (Fig. 6a) by a decompression event. If the T and P conditions (or ∆T and ∆P) were similar, then plagioclase zonation should be consistent for each grain.

Evidence from other MASLI-style intrusions of the CMMP suggests that the mixing/mingling textures were formed deeper in the magma chamber, adjacent to the mafic input zones. In the Vinalhaven intrusion, hybrid magmas of intermediate composition developed and exhibit a wide range of disequilibrium textures in the zone of mixing and mingling between the mafic and felsic layers. Rapakivi feldspars are much more abundant (nearly every grain has a plagioclase mantle), as are resorbed quartz and titanite occelli, suggesting that they originated there. The hybrid rocks developed are texturally, mineralogically, and compositionally similar to enclaves found in the higher, granitic portions of this intrusion (Luxet al. 2001). The distribution of these textures within this pluton suggests that once mafic input occurred, mixing textures were formed and subsequently distributed around the magma chamber by convection. The latter was probably initiated (and possibly maintained) by the thermal shock of mafic magmas intruding the crystallizing felsic chamber. Mafic magmas are much hotter than felsic magmas and this heat is dispersed as cooling and crystallization progresses. In the MWG mixing/mingling textures were formed at the interface between these disparate magmas and then circulated throughout the chamber resulting in the variable feldspar population and mixture of plagioclase types within the same sample. This process was also responsible for the distribution of enclaves. Feldspar megacrysts may have been incorporated in the enclaves at the time of their formation or later as they both circulated through the chamber.

Once mixing in the system was initiated, it became an auto-catalytic process due to the thermal, chemical, and physical gradients set up in the magma chamber and their subsequent effects on the body of the magma chamber by convection. These gradients perpetuated themselves throughout the chamber and were responsible for the highly variable nature of the MWG. Textural variants such as the coarse feldspar domains (Table 1 and Fig. 4b), possibly initially cumulate layers, were subsequently disrupted and transported through the chamber, collecting enclaves of variable composition. As a result, the higher level of the pluton could have been a complex mix of crystallizing framework and passageways of magma flow creating the conditions, i.e., flow segregation and/or crystal settling, conducive for the formation of the mafic schlieren.

In conclusion, MASLI-style magma chambers were common in the early stages of the development of the CMMP. However, the younger Devonian plutons, here exemplified by the MWG, may also have crystallized in dynamic, hybrid magma chambers as implied from complex the feldspar population and enclave textures described here. These plutons are in effect "cryptic" MASLI, where the body of evidence points to mixing as an important process in their evolution, although the actual mixing zone cannot be observed at the current erosional levels. Additional whole-rock and mineral geochemistry along with isotopic data will enable us to critically test this model, and to evaluate the extent of the involvement of mafic magmas in the generation and crystallization of granitic rocks in the CMMP.