As noted in the Introduction, consistency has yet to be attained in published terminology for plutonic units in the study area. Ludman and Hill (1986) used Moosehorn Intrusive Complex as a collective term for plutonic units in the Calais area; however, Jurinski (1987) and Hogan and Sinha (1989) used the term Moosehorn Igneous Complex. Ludman and Hill (1986) excluded Staples Mountain Gabbro from their Moosehorn Intrusive Complex, although its close spatial association and petrographic features (Fig. 2, Table 1) appear to justify its inclusion.

Hence, as used here, the MPS consists of five plutons termed the Staples Mountain Gabbro, St. Stephen Gabbro, Calais Quartz Diorite, Baring Granite, and Elliott Mountain Diorite (Fig. 2), each named according to its dominant and most characteristic rock type, although each contains a variety of components (Table 1). Smaller bodies of gabbro, diorite, and granite are present throughout these main plutons. The individual pluton names were used by previous workers, except for Elliott Mountain, a new name introduced by McLaughlin (2003) for part of the Calais gabbro/diorite of Hogan and Sinha (1989). Hogan and Sinha (1989) included all of the mafic-intermediate plutonic rocks between the Baring and Red Beach granite bodies in the Calais gabbro/diorite; however, we use the term Calais Quartz Diorite only for the gabbro-diorite intrusive complex of Ludman and Hill (1986, 1990), which also contains a substantial component of granite and metasedimen-tary xenolithic material. We use the name Elliott Mountain Diorite to refer to the gabbro unit of Ludman and Hill (1986, 1990). That unit extends into the adjacent map area of Abbott (1986), who described it as complex mixture of mainly gabbro, diabase, and granodiorite, intruded by younger granite. We consider the Calais and Elliott Mountain units to be different from one another overall, although both contain a wide variety of rock types and distinction between them at the outcrop scale is ambiguous (see below). Similar and probably co-magmatic mafic rock types also occur throughout the Baring Granite unit, as well as in the Meddybemps Granite to the south. Like previous workers, we exclude clearly younger (Devonian?) granitic plutons in the area, such as Meddybemps, Charlotte, and Red Beach (Fig. 2), from the MPS.

Distinction between gabbro and diorite is problematic in the MPS, especially in outcrop and hand specimen. We use the term gabbro when we consider that the rock is, or was originally before alteration, dominated by pyroxene as the ferromagnesian mineral component, and diorite when amphibole is the dominant ferromagnesian mineral. Hence we term the hornblende gabbro of some earlier workers (e.g., Abbott 1986) as diorite.

The MPS is characterized by complex zones in which diorite forms enclaves in granite, suggesting that the diorite magma was only partially crystallized at the time of emplacement of the granite magma, and that the two magmas commingled (Hill and Abbott 1989; Jurinski 1990; Ludman and Hill 1990; Hill 1991). These enclaves vary in shape from amoeboid to angular, and in size from cm-scale to m-scale, and were described in detail by previous workers (e.g., Hill and Abbott 1989; Jurinski 1990). The dioritic component dominates in the areas included in this study in the Calais Quartz Diorite and Elliott Mountain Diorite, whereas the granitic component dominates in the area shown as Baring Granite (Fig. 2). A complex relationship also appears to exist between dioritic and gabbroic magmas in the MPS, as locally diorite and quartz diorite appears to cross-cut gabbro, and xenoliths of gabbro are present in diorite and quartz diorite, but overall their petrographic and chemical features are gradational. In this study, the boundary between the St. Stephen Gabbro and Calais Quartz Diorite units is drawn so as to exclude, in so far as possible, gabbroic rocks from the latter pluton. Dykes of granite occur in gabbro in both the Staples Mountain and St. Stephen plutons, indicating that the granite is at least slightly younger than (and not mingled with) the gabbro magma. Because they are similar in appearance, these granite dykes are assumed to be co-magmatic with the granite that dominates the Baring Granite unit of the MPS.

The Elliott Mountain Diorite is dominated by dioritic rocks (gabbro of Abbott 1986 and Ludman and Hill 1990), although granodiorite and granite are also present, especially in the eastern part of the unit (Abbott 1986). Where observed, the contact between the Calais Quartz Diorite and the Elliott Mountain Diorite appears to be sharp and marked by a reduction in grain size in the diorite as the contact is approached, suggesting the presence of a chilled margin. Based on the lack of observed mingling relationships during this study, we suggest that the Elliott Mountain Diorite may be at least somewhat younger that the Baring Granite, but more detailed mapping is needed along their contact in order to confirm this suggestion. The Elliott Mountain Diorite is older than the Charlotte, Magurrewock Lakes, and Red Beach granites that intrude it on the south and east, and which are excluded from the MPS.

In addition to units included in the five main plutons, small bodies of dioritic and gabbroic rocks (e.g., Woodland Dump gabbro of Ludman and Hill 1990) also are present in the MPS and adjacent units. A few samples were collected from these units for comparison with the main dioritic and gabbroic units of the MPS.

Petrographic features of the main units of the MPS are summarized in Table 1, and commented on briefly below.

Dated sample SS00-124 is a white, medium- to coarsegrained monzogranite, typical of the Baring Granite unit. Zircon grains were separated from a 20 kg sample in the geochronology laboratory at the Massachusetts Institute of Technology by standard techniques, abraded to obtain clean, unaltered grains, and dated using methodology as described by Schmitz and Bowring (2003). Two pale yellow, slender prisms (length:width = 2.8:1) with blunt terminations, as well as two fragments of such prisms, were analyzed (Table 2). All four fractions are concordant, yielding a precise crystallization age of 421.1 ± 0.8 Ma (Fig. 4), at 2 ó error.

This age is the same, within error, as the age of 423 ± 3 Ma obtained by the U-Pb (zircon) method for the Utopia Granite (M.L. Bevier, 2001, personal communication to S.M. Barr). The age is similar also to ages of various other plutons of the coastal Maine magmatic province (Fig. 1), such as the South Penobscot Pluton (419 ± 2 Ma, U-Pb, zircon; Stewart et al. 2001), Spruce Head Pluton (421 ± 1 Ma, U-Pb, zircon; Tucker et al. 2001), Cadillac Mountain intrusive complex (424 ± 2 and 419 ± 2 Ma, U-Pb, zircon; Seaman et al. 1995), and Sedgwick Pluton (419.5 ±1 Ma, U-Pb, zircon; Stewart et al. 2001).

Sample SS98-36 from the olivine gabbro unit of the St. Stephen Gabbro (location shown on Fig. 2) contained relatively large and abundant interstitial phlogopite suitable for ⁴⁰Ar/³⁹Ar dating. Phlogopite grains were picked from the crushed sample using tweezers under a binocular microscope. For irradiation in the McMaster University nuclear reactor (in Hamilton, Ontario), the selected grains were placed individually into holes machined in aluminum disks. The flux monitor was the hornblende standard, MMhb-1 (assumed age = 520 Ma; Samson and Alexander 1987). Laser analyses were made with a Nd-YAG system operated in continuous mode with the beam expanded to approximately cover the grain. Power was then increased in a series of steps until complete fusion was achieved. Isotopic analyses were made using a VG 3600 mass spectrometer.

The single grain ages are plotted against ³⁹Ar abundance for each analysed grain (Fig. 5). With one exception (Spot 5), no significant differences in apparent age were observed among the fourteen grains. The mean age is 421 ± 4 Ma (2ó uncertainty, which includes the estimated error in the irradiation parameter, J). This age is interpreted to be the time of cooling of phlogopite through its argon closure temperature (ca. 300– 400°C, McDougall and Harrison 1999), and likely approximates the time of crystallization of the gabbro. It is similar within error to a previously reported amphibole K-Ar age of 418 Ma (Wanless et al. 1973) from the St. Stephen Gabbro. However, it is significantly older than the biotite (presumably phlogopite) K-Ar age from the same pluton (Wanless et al. 1973). It is also similar to the previously reported ⁴⁰Ar/ ³⁹Ar age of 422.7 ± 3 Ma reported for hornblende from the Pocomoonshine Gabbro-Diorite (West et al., 1992).

Although complex mingling, mixing, and gradational contacts characterize relations in the MPS, as described above, heterogeneous rocks were avoided in sampling for geochemistry in this study. We focused on homogeneous-looking samples that appeared characteristic of the main units of the MPS (Table 4). Some of the samples analysed as part of this study (MH designation in Table 4) were obtained from the collection of M. Hill and analyzed as part of the present project. The geochemical data of Gaskill (1999) were integrated into our study after a re-examination of his thin sections and samples. However, we have not included the large chemical data base of Paktunc (1989) because of uncertainty in locations and petro-graphic features of the analyzed samples. As demonstrated by McLaughlin (2003), they do not significantly change the conclusions of this study.

Also included in Table 4 are data for the Bocabec Pluton. The major element data for these samples are from Fyffe (1971), and trace element analyses were obtained from the same sample powders as part of the present study. In addition, four other granitic samples from the thesis collection of Fyffe (1971) were analyzed by K. Thorne and D. Lentz, and those unpublished data are included, with permission, in Table 3. Additional data from Ludman and Hill (1990), Coughlin (1986), McLeod (1990) (including data listed in his Appendix 4), and Thorne and Lentz (2001) are included on the diagrams, but not listed in Table 4.

Plots against SiO2 are used to illustrate chemical characteristics of the units of the MPS (Fig. 6). Dioritic and gabbroic samples from the Staples Mountain, St. Stephen, and Elliott Mountain plutons all have less than about 50% SiO₂, but show differences in other major element oxides. The Staples Mountain samples show a wide range in most elements, consistent with their wide range in rock types and cumulate features as described by Coughlin (1986). Their most distinctive characteristics compared to the dioritic and gabbroic samples are their high Fe and Ti contents (Fig. 6a, c). The Elliott Mountain samples tend to overlap in composition with gabbro from the St. Stephen Gabbro, but have higher TiO2, Na2O, and P2O5 and lower MgO contents (Fig. 6a, f, h, d). Compared to samples from the gabbro unit in the St. Stephen Gabbro, the olivine gabbro and troctolite samples have lower SiO2, but also a trend toward lower TiO2, Na2O, and P2O5 contents. Variations in Al2O3, Fe2O3^t, MgO, and CaO are consistent with the occurrence of cumulate textures in most of the analysed samples, and can be accounted for by olivine, pyroxene, and plagioclase separation or accumulation. In general, the trend from the more mafic samples of the St. Stephen Gabbro to the higher SiO2 samples of the Calais Quartz Diorite can be explained by fractionation of plagioclase, olivine, and pyroxene, and it is likely that these plutons are co-magmatic. In comparison, both the Staples Mountain Gabbro and Elliott Mountain Diorite units show differences in major element compositions which suggest that they are not genetically related to the St. Stephen and Calais plutons. Sample SS00-108 from a gabbroic dyke(?) in the Baring Granite has higher SiO2 than any samples analyzed from the Staples Mountain Gabbro, but also shows high Fe2O3 and low MgO, as well as high Na2O and P2O5 contents (Fig. 6f, h). In contrast, sample SS00-129 from the Woodland Dump gabbro of Ludman and Hill (1990) is similar to samples of similar SiO2 content from the St. Stephen Gabbro.

With one exception (granodiorite SS00-142), analyzed samples from the Baring Granite are all of granite composition, and have high SiO2 contents (more than 68%), and correspondingly low contents of other major elements except Na2O and K2O (Fig. 6). These chemical features are consistent with the high contents of sodic plagioclase, K-feldspar, and quartz and low abundance of ferromagnesian minerals in the samples. Chemical similarity between the Baring Granite and the Utopia Granite (data from McLeod 1990) is apparent, although the compositional range reported for the Utopia Granite extends to higher SiO2 content.

Data from the Bocabec Pluton are also shown on Fig. 6. As described in detail by Fyffe (1971) and McLeod (1990), these samples represent the mafic and felsic components in the suite, and also the intermediate units that were inferred to have formed by mixing and hybridization between the mafic and felsic magmas. The range of compositions encompasses most of the samples from the Calais Quartz Diorite, as well as the gabbro unit of the St. Stephen Gabbro, thus supporting their co-genetic relationship. In contrast, the Staples Mountain and Elliott Mountain samples generally lie outside the Bocabec trend, implying that they are not closely linked to the other mafic magmas. The data also show that the granitic component of the Bocabec Pluton is chemically similar to the Baring Granite, and support the interpretation (Fyffe 1971; McLeod 1990) that the intermediate samples formed by a mixing process. Intermediate sample (SS00-142) from the MPS is chemically similar to intermediate samples from the Bocabec Pluton, suggesting that it may have had a similar origin, as a result of mixing between Baring Granite and Calais Quartz Diorite magmas.

Trace element data are more limited than major element data, and are mainly for samples from the present study (Table 4). Trace elements are available for only one sample from Staples Mountain Gabbro and variable numbers of samples from the Bocabec Pluton and Utopia Granite, depending on the element. Trends in Rb, Sr, and Ba within the mafic and felsic sample sets are generally consistent with feldspar fractionation (Figs. 7a-c). Gabbroic samples have mainly low Rb and Ba, and high Sr, whereas the Calais Quartz Diorite generally has less Sr and more Rb and Ba. Some Utopia Granite samples and one Baring Granite sample have high Rb and low Ba, indicative of a highly evolved composition in which Ba has been partitioned into and removed with K-feldspar (Fig. 6g). However, overall, the Baring Granite is characterized by high Ba content (up to almost 1100 ppm).

Elements Zr, Nb, Y, and V display wide variation and few trends on silica variation diagrams (Figs. 8a-d), related to the control of these elements by the abundance of mainly pyroxene and magnetite (e.g., Miyashiro and Shido 1975; Pearce and Norry 1979) which vary widely in the dioritic and gabbroic samples. The Utopia Granite overlaps the composition of the Baring Granite, but generally has higher Nb and Y, and lower Zr (Figs. 8a– c). Some of the intermediate to granitic Bocabec Pluton samples have elevated Zr, Nb, and Y, and low V (Fig. 8). High Zr content in intermediate samples may represent the build-up of that element in the magma prior to onset of zircon crystallization.

Ni and Cr show wide spread in the gabbroic and dioritic samples (Figs. 9a, b) related to variations in olivine and pyroxene contents. The Staples Mountain Gabbro sample (NB00-133) is low in both elements compared to most St. Stephen Gabbro samples, further support for their lack of genetic relationship. Cu and Co also display a wide range in gabbroic to intermediate samples, with a few high values up to 150 and nearly 100 ppm, respectively (Figs. 9c, d). Ni, Cr, and Cu are low in all the granitic samples, but Co contents are high in the Baring Granite compared to most granitic Bocabec Pluton and Utopia Granite samples. Pb and Zn also differ between the Baring and Utopia granites, being generally higher in the Baring samples (Figs. 9e, f). Zn contents are highest in the intermediate rocks.

Interpretation of chemical affinity and tectonic setting for units of the MPS is somewhat ambiguous, as illustrated by representative, commonly used diagrams in Fig. 10. With the exception of the Staples Mountain and Woodland Dump samples, which appear tholeiitic, gabbroic and dioritic samples show little iron enrichment, and calc-alkaline affinity is suggested (Fig. 10a). Most samples have low Nb/Y ratios indicative of subalkalic affinity, although some St. Stephen and Bocabec samples plot in alkalic fields (Fig. 10b). Alkalic affinity is not supported by the V-Ti diagram (Fig. 10c), which suggests with-in-plate tholeiitic character for mafic samples. On the other hand, the Ti-Zr-Y diagram suggests calc-alkaline (arc) setting (Fig. 10d).

Granitic samples from the Baring Granite have lower Y and Nb contents, consistent with origin in a volcanic arc setting, whereas granitic samples from the Bocabec Pluton and Utopia Granite are generally more evolved and plot mainly in the within-plate granite field (Fig. 10e). Elevated Zr and Ga/Al ratio suggest A-type characteristics for these samples (Fig. 10f ).

Rare-earth element (REE) data were obtained for 15 samples from the MPS during the present study (Table 5). Additional REE data for the MPS were taken from Ludman and Hill (1990) and data for the Bocabec pluton and Utopia Granite are from McLeod (1990). Some of the data from McLeod (1990) show erratic features that suggest analytical error, but the overall patterns are consistent enough to enable comparison with the MPS.

In comparison with the other units, the St. Stephen gabbro has low REE abundance and a relatively flat REE pattern (Fig. 11a). The positive Eu anomalies suggest plagioclase accumulation. The sample with highest REE abundance shows some enrichment in light REE relative to heavy REE, and the pattern is very similar to that of the Calais Quartz Diorite sample with lowest REE abundance. The samples from the Calais Quartz Diorite show parallel chondrite-normalized patterns, with a change from a slight positive to a slight negative Eu anomaly with increasing REE abundance and increasing SiO₂ content.

Mafic samples from the Bocabec Pluton display patterns similar to the Calais Quartz Diorite (Fig. 11b). However, an intermediate sample with over 58% SiO₂ has higher REE abundances and a pattern more similar to those of the Baring Granite samples (Fig. 11c). Three samples from the Elliot Mountain Diorite (Fig. 11c), all from Ludman and Hill (1990) are similar to the Calais Quartz Diorite, but do not show the small negative Eu anomalies that characterize the Calais samples.

Samples from the Baring Granite, Utopia Granite, and granite in the Bocabec Pluton all show strong negative Eu anomalies that could be explained by feldspar fractionation (Fig. 11c, e). They are most pronounced in the Utopia Granite, consistent with its more evolved character (e.g., Fig. 7e). However, in general the REE patterns are indicative of similar and related origins for these units.

Gabbro dyke(?) sample SS00-108 in the Baring Granite has high total REE abundance and a positive Eu anomaly (Fig. 11d). The high REE content, light REE enrichment compared to heavy REE, and strong positive Eu anomaly set it apart from other samples. It may be related to the Staples Mountain Gabbro, as suggested by some of its other chemical features, but no REE data are yet available from Staples Mountain for comparison.

Five samples were analyzed for Sm-Nd isotopes (Table 6). The calculated åNd at 421 Ma for gabbro sample SS98-9 from the St. Stephen Gabbro is 0.4, with a depleted mantle model age of 1567 Ma, whereas two dioritic samples from the Calais Quartz Diorite have higher åNd values of 1.03 and 3.4 at 421 Ma, with depleted mantle ages of 1184 and 881 Ma, respectively. For the Baring Granite, the åNd is only slightly lower at – 0.40, with a similar depleted mantle age of 1250 Ma. The gabbro dyke(?) in the Baring Granite has a positive åNd of 1.51. All of these values are well below depleted Mantle values (ca. +7 to 8 at ca. 421 Ma), showing that significant crustal material was melted to produce even the mafic units in the MPS. On the other hand, the åNd values, even for the Baring Granite, are too high for the magmas to have had an ancient crustal source (likely to be – 7 or less at ca. 421 Ma; DePaolo 1988).

These results are consistent with those reported by Whalen et al. (1994), which were in the order of +2 to +3 for the Utopia Granite and granite in the Bocabec Pluton. Gabbro in the Bocabec Pluton yielded a higher value of +5 (all values recalculated to 421 Ma).