ARTICLE

Geological significance of 40Ar/39Ar mica dates across amid-crustal continental plate margin, Connemara (Grampianorogeny, Irish Caledonides), and implications for the evolutionof lithospheric collisions1

Anke M. Friedrich and Kip V. Hodges

Abstract: The Connemara region is a world-class example of a regional-scale, high-temperature metamorphic terrain. Its rockrecord documents formation of a bi-vergent orogenic wedge and associated calkalkaline magmatism in a an arc–continentcollisional setting (Grampian orogeny), for which a protracted evolution was inferred based on a >75 Ma spread in U–Pb, Rb–Sr,and K–Ar mineral ages. In contrast, geological field observations imply a simple relationship between syntectonic magmatism,bi-vergent deformation, and Barrovian-type metamorphism. We explore the significance of the spread in apparent cooling agesusing 40Ar/39Ar mica thermochronometers of varying grain sizes and composition, collected across metamorphic grades rangingfrom staurolite to upper sillimanite. We integrated geological and previously published geochronological evidence to identify a32 Ma range (ca. 475–443 Ma) of permissible cooling ages and distinguished them from those dates not related to cooling afterhigh-temperature metamorphism. Variations in 40Ar/39Ar dates at a single locality are ≤10 Ma, implying rapid cooling (≥6–26 °C/Ma)following metamorphism and deformation. A distinct cooling age variation (≥15 Ma) occurs on the regional scale, consistent withspatial differences in the metamorphic, magmatic, and deformational evolution across Connemara. This cooling record relates to alateral thermal gradient (30 °C/km) in an evolving arc–continent collision, rather than to differential unroofing of the orogen. Ourresults imply that the large (≥50 Ma) spread in thermochronometers commonly observed in orogens does not automatically translateinto a protracted cooling history, but that only a small number of thermochronometers supply permissible cooling ages.

Résumé : La région de Connemara des Calédonides irlandaises est un exemple de classe mondiale de terrain métamorphique de hautetempératured’envergurerégionale.Saformationestreliéeaumagmatismecalcoalcalindanslecontexted’unecollisionarc–continent,pourlaquelle une évolution prolongée a été inférée a la lumière de la distribution sur plus de 75 Ma d’âges U–Pb, Rb–Sr, et K–Ar sur différentsminéraux. Une telle évolution ne correspond pas aux observations géologiques sur le terrain, qui indiqueraient plutôt une relation simpleentre un magmatisme, une déformation et un métamorphisme de type barrovien syntectoniques. Nous avons examiné la signification dela distribution des âges apparents de refroidissement a l’aide de thermochronomètres 40Ar/39Ar sur micas de granulométries et composi-tions variées, prélevés a différents degrés de métamorphisme allant de la zone a staurotide a la zone a sillimanite supérieure. Nous avonsintégré les données géologiques et de l’information géochronologique déja publiée pour délimiter une fenêtre de 32 Ma (de 475 a 443 Maenviron) pour les âges de refroidissement possibles et avons distingué ces derniers des âges non reliés au refroidissement ayant suivi lemétamorphisme de haute température. La variation des âges 40Ar/39Ar en un même endroit est ≤ 10 Ma, ce qui indiquerait un refroidisse-ment rapide (≥ de 6 a 26 °C/Ma) suivant le métamorphisme et la déformation. Des variations claires des âges de refroidissement (≥ 15 Ma)existent a l’échelle régionale, correspondant aux variations spatiales de l’évolution métamorphique, magmatique et de la déformation al’échelle de la région de Connemara. Cet historique de refroidissement est relié a un gradient latéral de température et de vitesse dedéformation dans une collision arc–continent dynamique, plutôt qu’a la dénudation tectonique différentielle de l’orogène. Nosrésultats indiquent que la large distribution (≥ 50 Ma) des thermochronomètres couramment observée dans les ceintures orogéniquesne reflète pas nécessairement une histoire de refroidissement prolongé, mais que seul un petit nombre de thermochronomètresproduisent des âges de refroidissement admissibles pour un contexte donné. [Traduit par la Rédaction]

Introduction

Studies of thermal histories of ancient metamorphic terrainsusing thermochronologic techniques frequently yield large ranges ofapparent ages, sometimes spanning hundreds of millions of years(Onstott and Peacock 1987; Mezger et al. 1991; Hodges and Bowring

1995). Such ranges are often interpreted in terms of a protractedcooling history, assuming reasonable closure temperatures formineral-isotopic systems (e.g., Cliff 1985; Heaman and Parrish1991; Hodges 1991, 2014; Baxter 2010). However, studies of themechanisms by which radiogenic isotopes are lost from mineralsystems (e.g., Lee 1995) suggest that assigning an exact closure

Received 4 January 2016. Accepted 26 August 2016.

Paper handled by Editor Ali Polat.

A.M. Friedrich* and K.V. Hodges.† Department of Earth, Atmospheric, and Planetary Sciences, Massachusetts Institute of Technology, Cambridge,MA 02139, USA.Corresponding author: Anke M. Friedrich (email: friedrich@lmu.de).*Present address: Department of Earth and Environmental Sciences, Ludwig Maximilian University of Munich, 80333 Munich, Germany.†Present address: School of Earth and Space Exploration, Arizona State University, Phoenix, AZ, USA.1This paper is part of a special issue that honors the careers of Kevin C. Burke and John F. Dewey.Copyright remains with the author(s) or their institution(s). Permission for reuse (free in most cases) can be obtained from RightsLink.

1258

Can. J. Earth Sci. 53: 1258–1278 (2016) dx.doi.org/10.1139/cjes-2016-0001 Published at www.nrcresearchpress.com/cjes on 6 October 2016.

Can

. J. E

arth

Sci

. Dow

nloa

ded

from

ww

w.n

rcre

sear

chpr

ess.

com

by

51.1

71.9

6.78

on

04/2

8/17

For

pers

onal

use

onl

y.

temperature to any particular mineral sample is not entirelystraightforward (Watson and Baxter 2007). For example, ranges ofdates provided by a single mineral-isotopic system may reflectlocal differences in deformational history or mineral chemistryrather than protracted cooling (e.g., Mulch et al. 2005; Bröckeret al. 2013). One approach to assessing the importance of thisproblem is to compare thermochronologic data for minerals ofknown composition with independent constraints on the thermaland deformational history of the metamorphic terrain fromwhich they were collected.

In this paper, we present thermochronologic data for the Neo-proterozoic metamorphic complex of Connemara, one of theworld’s best exposed sections of middle crust—a type-example ofan island-arc continent collision (Grampian orogeny, e.g., Dewey2005; Dewey and Ryan 2016). At Connemara, a protracted coolinghistory spanning at least 75 million years has been inferred fromearlier Rb–Sr and K–Ar results (e.g., Elias et al. 1988, Cliff et al.1996). The significance of this “Grampian tail” (Dewey and Mange1999) continues to play an important role for tectonic models ofisland arc – continent collisions, involving subduction polarity-reversal and establishment of Andean-type margins, such as theone documented in western Ireland (e.g., Dewey 2005). Conne-mara provides a special opportunity to compare the thermo-chronologic record as documented in metamorphic rocks (e.g.,Yardley et al. 1987) with very tight constraints on the thermal anddeformational history of the area provided by the U–Pb geo-chronology of synmetamorphic intrusive igneous rocks of theConnemara complex (Friedrich et al. 1999a, 1999b). We employedthe 40Ar/39Ar method on muscovite, biotite, and phlogopite frommetapelitic and calcsilicate rocks sampled at different metamorphicgrades, ranging from staurolite to upper sillimanite. The results werethen evaluated against independent constraints. We found a sub-stantial spread in 40Ar/39Ar dates, largely consistent with the resultsof previous studies, but the distribution of dates seems less a simplefunction of protracted slow cooling than a function of spatially com-plex variations in the thermal history of the region during metamor-phism, variable incorporation of excess 40Ar in mineral systems, andlocal, post-orogenic resetting of 40Ar/39Ar systematics.

We integrated our results with published structural, paleomag-netic, and geochronologic data over three scales to derive internallyconsistent evolutionary diagrams of the arc–continent collision atConnemara. We provide an explanation for the thermal evolutionof the orogen in context with geodynamic processes involving thelithosphere and the sublithospheric upper mantle.

Geologic frameworkThe Connemara region of the Irish Caledonides (Fig. 1) exhibits the

thermal and deformational characteristics of an arc–continent colli-sional environment (e.g., Leake and Tanner 1994; Dewey 2005;Dewey and Ryan 2016; fig. 17.3 in Brown et al. 2011). The formationof double-vergent nappe structures and regional-scale metamor-phism is closely related to emplacement of mafic to felsic plutonsof the Connemara complex (Leake 1989), a calc-alkaline magmaticprovince that formed during the middle Ordovician Grampian–Taconic orogeny (ca. 474–463 Ma) along the eastern Laurentianmargin (e.g., Dewey and Ryan 2016; Friedrich et al. 1999a; van Staalet al. 1998, 2013). The country rocks for the Connemara complexare metasedimentary units of the Neoproterozoic Dalradian Su-pergroup, which represents the rift-to-drift transition related tobreakup of Rodinia and the formation of the Iapetus ocean (e.g.,Soper et al. 1999). The outcrop pattern of abundant marker hori-zons in this stratigraphy delineates a complicated three-stage de-formational history (e.g., Leake and Tanner 1994). The overallstructure is dominated by the gently east-plunging, range-scaleConnemara antiform, which exposes a spectacular three-dimensional cross section through the double-vergent Dalradianfold-and-thrust belt and the Connemara complex (Figs. 1 and 2; e.g.,

Leake and Tanner 1994). The deepest structure in the subverticalsouthern limb (“the steep belt”) of the Connemara antiform is theMannin thrust (Fig. 1; Leake et al. 1983); it carries the Dalradianmetasedimentary package and the Connemara complex in itshanging wall, and juxtaposes these units southwards againstgreenschist facies metarhyolite of the Delaney Dome Formationin the footwall (e.g., Draut and Clift 2002). The shallowest struc-tures are normal faults (e.g., the Renvyle–Bofin Slide, Boyle andDawes 1991; Clift et al. 2004), which are exposed north of theantiformal crest.

At Connemara, the Dalradian units preserve two macroscopicphases of metamorphism: an early, low- to high-pressure amphi-bolite facies event (M2) and a late, high-temperature, medium- tolow-pressure event (M3; e.g., Barber and Yardley 1985). The M2event was similar to that recorded by metamorphic assemblagesin age-equivalent rocks elsewhere in the Irish and British Cale-donides (e.g., Leake and Tanner 1994; Yardley et al. 1987; Cliff et al.1996; Dewey et al. 2015). However, the M3 event is unique to Con-nemara and the distribution of M3 isograds suggests that thisevent was related to intrusion of the Connemara magmatic com-plex (e.g., Yardley et al. 1987). Despite the structural complexity ofthe region, these isograds strike generally east–west, transectingall major structures, suggesting a simple thermal evolution dur-ing the late stages of M3. However, the relative age relationshipsbetween deformational structures, intrusive phases of the Conne-mara complex, and M2 and M3 metamorphic assemblages provideevidence for a more complicated thermal history overall.

Deformation historyThe oldest preserved fabric (S1) occurs only as inclusion trails in

garnet porphyroblasts with so far unknown tectonic significance(Figs. 2 and 3A; e.g., Leake and Tanner 1994). The earliest macro-scopic folds in Connemara (F2) are east–west-trending open toisoclinal. They are associated with an axial planar schistosity (S2),defined by muscovite and biotite, which is mainly preserved innorthern Connemara (Fig. 3B). There the S2 schistosity was de-formed by small-scale folds during a D3 event, which was accom-panied by retrograde muscovite + chlorite growth (Fig. 3C; e.g.,Boyle and Dawes 1991). Farther south, the S2 schistosity was de-formed by small F3 folds (Fig. 3D) and large eastward plunging F3fold-and-thrust nappes. Especially in pelitic lithologies, S2 wasoverprinted by a penetrative S3 schistosity defined by biotite in-tergrown with fibrolite or prismatic sillimanite (MS3; Fig. 3E). Insouthern Connemara, where M3 temperatures were high enoughto cause partial melting in the Dalradian sequence, evidence forsyn-anatectic deformation demonstrates a clear temporal rela-tionship between D3 and M3. In this region, F3 folds were dis-rupted by syndeformational anatectic melting, indicating thatanatexis followed the main D3 deformation (Fig. 3F). The Manninthrust, which places the Connemara complex and its countryrocks over metarhyolites of the Delaney Dome Formation, is re-garded as a late D3 structure (e.g., Tanner et al. 1989).

Fabrics and structures of D3 age were deformed on a regionalscale by the F4 Connemara antiform, described above. The factthat M3 metamorphic isograds cut across D4 structural trendsrequires that the M3 metamorphic peak stretched over the D3–D4deformational interval. The youngest structures in the region maybe extensional shear zones exposed in northern Connemara. One,described by Williams and Rice (1989) and Clift et al. (2004), occursdirectly below the Silurian unconformity (Fig. 1); its hanging wallis largely concealed by post-deformational Silurian strata. The second—referred to as the Renvyle–Bofin Slide—corresponds to the northernoutcrop boundary of the Dalradian Supergroup at Connemara andhas a demonstrated top-to-the-north normal shear sense (Boyle andDawes 1991; Clift et al. 2004; Wellings 1998). However, the Renvyle–Bofin Slide probably is a smaller structure relative to the postulateddetachment below the Silurian unconformity (Clift et al. 2004).

Friedrich and Hodges 1259

Published by NRC Research Press

Can

. J. E

arth

Sci

. Dow

nloa

ded

from

ww

w.n

rcre

sear

chpr

ess.

com

by

51.1

71.9

6.78

on

04/2

8/17

For

pers

onal

use

onl

y.

Intrusive historyIntrusions of the Connemara complex were either broadly syn-

chronous with or postdate crustal deformation. The oldest plu-tons, consisting of ultramafic and gabbroic rocks, occur in bothnorthern (Dawros–Currywongaun–Doughruagh complex, Fig. 1;Kanaris-Sotiriou and Angus 1976) and southern Connemara (e.g.,Cashel–Lough Wheelaun intrusion, Leake 1989). A younger intru-sive suite, which consists mainly of quartz diorite and granite,intruded older mafic plutons only in southern Connemara. Thegabbros of northern Connemara were emplaced syntectonicallyinto the same tectonostratigraphic level as the southern Conne-mara gabbros (Leake and Tanner 1994). Wellings (1998) suggestedthat the Dawros–Currywongaun–Doughruagh complex intrudedduring D2 deformation. If he is correct, a recent U–Pb zircon agefor a gabbro from the complex constrains the age of D2 in north-ern Connemara as 474 Ma (Friedrich et al. 1999a). The Cashel–Lough Wheelaun intrusion of southern Connemara intruded justprior to or early during D3 deformation (Tanner 1990) at 470 Ma(U–Pb zircon age, Friedrich et al. 1999a). A minimum age for bothD3 and D4 deformation is provided by the undeformed Oughte-

rard granite (e.g., Tanner et al. 1997), which has a U–Pb xenotimecrystallization age of 463 Ma (Friedrich et al. 1999a).

Metamorphic historyFrom north to south across Connemara, Grampian metamor-

phic conditions increase progressively from upper greenschist toupper amphibolite facies (Fig. 1). Four major zones, trending east–west, have been identified based on the distribution of progradepelitic mineral assemblages (Barber and Yardley 1985; Yardleyet al. 1987; Boyle and Dawes 1991): (1) the garnet–staurolite zone,(2) the staurolite–sillimanite transition zone, (3) the sillimanite–muscovite zone, and (4) the sillimanite–K-feldspar zone. In southernConnemara, near the principal outcrop region of quartz-dioritesof the Connemara igneous complex, muscovite is no longer stableand the diagnostic metapelitic assemblage becomes sillimanite +biotite + cordierite + K-feldspar + quartz + plagioclase. Many me-tasedimentary outcrops are migmatitic in the southernmost partsof this zone, suggesting that temperatures were high enough topromote widespread anatexis.

Fig. 1. Generalized geologic map of Connemara, western Ireland, showing the approximate location of metamorphic isograds, theConnemara intrusive complex, the Silurian unconformity, late orogenic extensional faults (RBS, Renvyle–Bofin Slide), major F4 folds (such asthe Connemara antiform), and the Mannin thrust, which emplaces the Connemara complex and the Dalradian country rocks over low-grademetamorphic rocks of the Delaney Dome Formation (DDF). DCDC, Dawros–Currywongaun–Douruagh complex of northern Connemara;C-LWG, Cashel–Lough Wheelaun gabbro (Leake and Tanner 1994). Cross section A–A= is shown in Fig. 2.

1260 Can. J. Earth Sci. Vol. 53, 2016

Published by NRC Research Press

Can

. J. E

arth

Sci

. Dow

nloa

ded

from

ww

w.n

rcre

sear

chpr

ess.

com

by

51.1

71.9

6.78

on

04/2

8/17

For

pers

onal

use

onl

y.

Despite the simplicity of the metamorphic pattern at Conne-mara, textural relationships suggest that the observed assem-blages grew during two prograde events. Minerals diagnostic of thegarnet–staurolite assemblage help define the S2 schistosity in north-ern Connemara, whereas sillimanite–muscovite and sillimanite–K-feldspar assemblages help define the S3 schistosity in centraland southern Connemara (Barber and Yardley 1985; Boyle andDawes 1991; Yardley et al. 1987, and references therein). This hasled previous workers to define an M2 event on the basis of thenorthern Connemara assemblages and an M3 event on the basis ofcentral and southern Connemara assemblages. Estimated peakM2 conditions in northern Connemara were 550 ± 50 °C and6–8 kb (Boyle and Dawes 1991). Peak M3 temperatures were atleast 650 °C in the middle of the sillimanite–K-feldspar zone, in-creasing to �750 °C in southernmost Dalradian outcrops (Barberand Yardley 1985). Near the main outcrop area of the quartz dior-ites, M3 pressures during anatexis reached �5.5 kb, but may havedeclined to <3 kb during crystallization of the leucosome.

Indirect constraints on the age of M2 and M3 are provided by theU–Pb ages of synmetamorphic intrusive rocks. Based on the work ofWellings (1998), S2 and the M2 assemblages that define it are at leastpartly coeval with intrusion of the Currywongaun gabbro at 474 Ma(Friedrich et al. 1999a). The M3 event overlapped in time with em-placement of the bulk of the Connemara complex (Yardley et al.1987), which occurred during the 470–463 Ma interval (Friedrich etal.1999a, 1999b). While textural evidence suggests that amphibolite-facies metamorphism should be attributed to two separate events atConnemara, the two were separated in time by no more than a fewmillion years and the prograde M2–M3 transition is probably bestregarded simply as a temperature increase during progressive orogeny.

Post-Grampian extensional faulting, erosion,sedimentation, igneous activity, and late-stage fluid flow

The high-grade Dalradian metasedimentary rocks and the Con-nemara intrusions are overlain unconformably by unmetamor-phosed shallow marine sedimentary rocks of lower Silurian age

Fig. 2. Summary of the deformational features and structural setting of the Connemara region. See Fig. 1 for location of cross section A–A=and abbreviations. The dominant schistosities and fold axes are from Tanner and Shackleton (1979).

Friedrich and Hodges 1261

Published by NRC Research Press

Can

. J. E

arth

Sci

. Dow

nloa

ded

from

ww

w.n

rcre

sear

chpr

ess.

com

by

51.1

71.9

6.78

on

04/2

8/17

For

pers

onal

use

onl

y.

(Fig. 1; e.g., Leake and Tanner 1994; Ryan and Dewey 2011; Cliftet al. 2004). This unconformity marks the end of the Grampianorogeny in the geologic record at Connemara. The unconformityis estimated to have formed at ca. 443 Ma, but no later than 435 Ma(Clift et al. 2004; Tucker and McKerrow 1995), such that the asso-ciated hiatus may be up to 29 Ma long. The real hiatus duration isshorter, because Dalradian rocks are overlain locally by a Late

Ordovician alluvial cobble-conglomerate (Lough Mask Formation;Derryveeney Formation; Graham et al. 1989; Clift et al. 2004; Ryanand Dewey 2011). Direct evidence for erosional unroofing of theDalradian sequence is documented in northeastern Connemaraand South Mayo by the first arrival of metamorphic clasts in thesedimentary record of the Rosroe and Mweelrea Formations be-tween ca. 468 Ma and at least 462 Ma (cf. fig. 13.5 in Ryan and

Fig. 3. Representative field photographs and photo micrographs of the deformation sequence in Connemara. (A) S1, schistosity is onlypreserved as inclusion trails within garnet. (B) S2, schistosity is the dominating penetrative fabrics at the low-metamorphic zones of northernConnemara, but a weak F3 overprint is usually observed. Exposure from the shore at Tully Mountain. (C) Sample AF17, a muscovite-garnetmetapelite in which muscovite + chlorite define a retrograde mineral assemblage, probably equivalent to the M3 assemblages farther south.(D) Impure marble of the Lakes Marble Formation. This picture is from Cur Hill, at the staurolite–sillimanite transition zone, and shows thetypical F3 folds that occur at a variety of scales throughout Connemara. F3 folds refold minor F2 fold axes. (E) The main S3 foliation plane in astaurolite–sillimanite–garnet schist from the staurolite–sillimanite transition zone at Cur Hill (sample AF79). (F) Outcrop “migmatitic”metasedimentary rocks, ca. 50 m northeast of the parking lot at the Alcock and Brown landing site memorial. The field relationships showthe S2 schistosity folded around a F3 fold-axis in a metasedimentary rock. This block has been disrupted subsequent to F3 folding, asindicated by its random orientation relative to the strong anatectic foliation defined within the anatectic metapelite rock. The foliation (S3=)is defined by alignment of paleosome and leucosome. (G) Outcrop F2–F4 fold generations. F2 and F3 fold axial traces are subparallel, whereasF4 folds are almost orthogonal to the older folds. F4 folds are accompanied by granitic pegmatite. Outcrop at the head of Errislanaanpeninsula, ca. 150 m southwest of the lighthouse.

1262 Can. J. Earth Sci. Vol. 53, 2016

Published by NRC Research Press

Can

. J. E

arth

Sci

. Dow

nloa

ded

from

ww

w.n

rcre

sear

chpr

ess.

com

by

51.1

71.9

6.78

on

04/2

8/17

For

pers

onal

use

onl

y.

Dewey 2011). By Upper Silurian to Devonian time, the Connemarametamorphic complex was buried again under �3–4 km of ma-rine sedimentary rocks (Graham et al. 1989). At this time, a majorpost-Grampian granitic intrusion—the Galway batholith—intrudedolder plutons of the Connemara igneous complex (e.g., Burke1957; Leake and Tanner 1994). The thermal pulse associated withbatholith intrusion led to the development of a regionally extensivehydrothermal system (Jenkin et al. 1992), which may be responsiblefor local retrogression of M2–M3 metamorphic assemblagesthroughout Connemara.

Previous thermochronologyK–Ar and Rb–Sr mineral dates ranging from 490 to 415 Ma were

interpreted as representing a prolonged period of unroofing (e.g.,Giletti et al. 1961, Dewey et al. 1970; Elias et al. 1988; Dewey 2005)or local resetting related to hydrothermal alteration (e.g., Leggoet al. 1966; Miller et al. 1991). A discordant U–Pb zircon date ofca. 490 Ma for the Cashel–Lough Wheelaun gabbro (Jagger et al.1988), intruded just before or during the early stages of D3, hasbeen used to suggest that M3 metamorphism began at about thattime (Cliff et al. 1996). This interpretation gained support fromslightly younger K–Ar hornblende dates from some outcrops(Elias et al. 1988). Mica K–Ar and Rb–Sr cooling dates clusteredbetween 460 and 450 Ma in southern Connemara, leading previ-ous workers to suggest two important episodes of cooling in theregion—one between 490 and 480 Ma, and the other between 460and 450 Ma, separated by an interval of relatively slow cooling(e.g., Elias et al. 1988). However, not all early thermochronologicdata fit into this simple model. One K–Ar hornblende date of420 Ma was reported from northern Connemara, and severalsouthern Connemara samples yielded mica dates in the 420–410 Ma age range (e.g., Elias et al. 1988). Such anomalies led to thedevelopment of complex models for regional cooling involvingdifferential uplift along high-angle faults and localized reheatingevents related to hydrothermal activity (Elias et al. 1988). In par-ticular, Elias et al. (1988) found a general southward increase ofhornblende K–Ar dates (of up to 480 ± 10 Ma) with increasingmetamorphic grade and concluded that they cooled before thenorthern lower-grade rocks.

An alternative model was proposed by Miller et al. (1991). Theseauthors determined K–Ar dates for 21 hornblendes and severalmicas from across Connemara, obtaining a large range of datessimilar to that reported by Elias et al. (1988). In one 30 m widequarry in southern Connemara, hornblende dates for amphibo-lites containing abundant quartz + epidote veins ranged over70 million years. Miller et al. (1991) found a weak positive correla-tion between K–Ar hornblende dates and �D values for the sameamphiboles, suggesting that the range of hornblende dates repre-sented variable amounts of Ar loss related to hydrothermal activity.In this scenario, K–Ar hornblende dates provide little constraint onthe thermal evolution of the Connemara region.

U–Pb geochronologic analysies of abraded single-zircon crystalsfrom intrusive rocks across Connemara (Friedrich et al. 1999a,1999b) helped to place constraints on the range of viable interpre-tations of K–Ar and Rb–Sr thermochronologic data. For example,a more precise redetermination of the age of the Cashel–LoughWheelaun gabbro and a new determination of the age of theCurrywogaun gabbro of southern Connemara demonstrate thatpeak M2–M3 metamorphism is no older than 475–470 Ma(Friedrich et al. 1999a), such that older published K–Ar horn-blende dates must represent excess 40Ar contamination. More-over, given the very brief duration of Grampian orogenesis atConnemara, we would expect that all amphibole and mica coolingdates would cluster within a few million years, between perhaps 460

and 450 Ma (Friedrich et al. 1999a). The Ordovician–Silurian bound-ary provides a lower bound on the post-Grampian cooling history. Inthis paper, we report new 40Ar/39Ar geochronologic data obtained inan attempt to better understand the extent and cause of the largerange of K–Ar and Rb–Sr ages obtained by previous workers.

MethodsWe collected muscovite, biotite, and phlogopite from metapelitic

and calcsilicate rocks from small subareas within each metamor-phic zone (Fig. 4). We assumed that the effective closure of amineral-isotopic system is a volume diffusion process, governedby effective diffusion dimension and chemical composition(McDougall and Harrison 1988; Watson and Baxter 2007; Baxter2010; Hodges 2014). Although studies suggest that fast diffusionpathways can complicate the geometry and rate of diffusion inminerals (Lee 1995), the effective diffusion dimension of micacrystals empirically appears to be similar to the physical grain size(Hames and Bowring 1994). In the hope of recovering a largerportion of the cooling history at each locality, we concentrated onanalyzing micas with a range of composition and grain size. Weseparated muscovite, biotite, and phlogopite from crushed or un-crushed rock and purified following standard procedures (Appen-dix A; Hodges et al. 1994).

We performed both high-precision incremental heating analysesof large samples (several milligrams) with a resistance furnace, andmicroanalysis of between 1 and 10 individual crystals with an Ar-ionlaser, which yields most of its energy in the blue–green visible spec-trum (Appendix A; supplementary data, Table ST42; Fig. SF22). Tocontrol the effects of sample composition and grain size, we deter-mined the compositions of between three and five representativecrystals of each analyzed sample (Tables ST1, ST2, and ST32), andmeasured the grain size of each crystal analyzed (Tables 1–3). Thegrain sizes of biotite, muscovite, and phlogopite from northern Con-nemara are relatively uniform and small (typically 300–500 �m indiameter). More significant grain size variations are found in south-ern Connemara, with the largest grains (>6000 �m diameter) occur-ring in some muscovite-bearing pegmatites.

All age calculations were based on assuming an initial 40Ar/36Arratio of modern atmosphere (295.5). Although previous work sug-gests that excess 40Ar contamination does exist in Connemara inminerals—particularly in amphibole as noted above—attempts toisolate such components and correct for them using inverse iso-tope correlation diagrams were hampered by uniformly high ra-diogenic Ar yields for all samples. Fortunately, this characteristicmeans that calculated dates are relatively insensitive to our choiceof initial 40Ar/39Ar. We interpret the dates reported here as cool-ing ages, but the presence of small amounts of excess 40Ar maymake some of the dates close overestimates of the time of coolingbelow the effective closure temperature interval.

39Ar weighted mean ages and their 2� errors were calculated forall total fusion and incremental release analyses. These ages rep-resent the 39Ar volume-averaged total gas age of a sample. Forincremental heating analyses, a plateau date is reported if at leastthree consecutive steps overlap within 2� error and contain over50% of the 39Ar released. For total fusion analyses, results frommultiple analyses of an individual sample were combined statis-tically to give an apparent age distribution for the sample. Theseare shown as normalized probability distribution plots (e.g.,Fig. SF22). Interpretation of the total fusion analyses was guided bythe shape of the probability distribution. If the distributionshowed a single mode, without secondary peaks or significantskewness, we assumed that the 39Ar weighted mean date of allanalyses represents the cooling age of the sample. If the apparentage distribution is multi-modal or skewed, the 39Ar weighted

2Supplementary data are available with the article through the journal Web site at http://nrcresearchpress.com/doi/suppl/10.1139/cjes-2016-0001.

Friedrich and Hodges 1263

Published by NRC Research Press

Can

. J. E

arth

Sci

. Dow

nloa

ded

from

ww

w.n

rcre

sear

chpr

ess.

com

by

51.1

71.9

6.78

on

04/2

8/17

For

pers

onal

use

onl

y.

mean date may be geologically meaningless, and we interpretdistinct peaks as the best indicator of age components repre-sented by the fusion analyses. Skewness of peaks toward olderdates can often suggest contamination by excess 40Ar, assumingthat the excess component is distributed inhomogeneously suchthat multiple total fusion analyses of a sample will yield ages withvariable excess. In this case, we regard the sample mode as a closeminimum estimate of a samples’ cooling age (e.g., Fig. SF22).

Following Dodson (1973), we calculated closure temperaturesfor micas from existing constraints for Ar diffusivity and grainsize by assuming a cooling rate (Table 4). Estimated closure tem-peratures used in this study are shown in Table 5.

Results of 40Ar/39Ar and microprobe analysesAll sample locations are shown in Fig. 4, their coordinates are

given in Tables 1–3. Results of geochemical microprobe and 40Ar/

Fig. 4. Generalized geological map of Connemara showing the locations of samples analyzed and discussed in this study. Refer to Fig. 2B forthe internal structure of the fold belt within the Dalradian Supergroup, which enabled sampling of three different lithologies (metapelite,calcsilicate, and amphibolite) at each metamorphic grade. Numbers refer to samples collected for 40Ar/39Ar analyses. Additional informationabout the samples is given in the supplementary data2. Refer to Tables 1–3 for location coordinates of the samples.

1264 Can. J. Earth Sci. Vol. 53, 2016

Published by NRC Research Press

Can

. J. E

arth

Sci

. Dow

nloa

ded

from

ww

w.n

rcre

sear

chpr

ess.

com

by

51.1

71.9

6.78

on

04/2

8/17

For

pers

onal

use

onl

y.

39Ar analyses are presented in Tables ST1–ST3, and ST42, respec-tively, and the Argon mineral ages are summarized in Tables 1–3.Figures SF1 and SF22 show the plateau and total fusion results. Thespatial distribution of the 40Ar/39Ar dates are shown in Fig. 5. Ingeneral, 40Ar/39Ar dates of biotite and muscovite are the oldest innorthern Connemara (Table 1), younger in central Connemara(Table 2), and the youngest in the migmatitic rocks of southernConnemara (Table 3) where they also show the largest range in dates.

Interpretation of 40Ar/39Ar mica dates acrossConnemara

Because the interpretation of these results depends on theirregional geological context, the data collected in this study aresummarized and interpreted individually and in geographic or-der, from the garnet–staurolite zone in northern Connemara tothe sillimanite–K-feldspar zone in southern Connemara (Figs. 5and 6). Below we also provide a justification for our samplingstrategy in the context of earlier published work.

Northernmost garnet–staurolite zoneTo test the interpretation of Boyle and Dawes (1991), based on

the Elias et al. (1988) K–Ar data, that cooling from �500–300 °C innorthern Connemara occurred over the 470–430 Ma interval, weanalyzed coexisting muscovite and biotite from two garnetiferousrocks from northernmost Connemara.

Furnace incremental heating experiments for muscovite (XKMs =0.73) and biotite (XAnn = 0.47) from sample AF16 yield essentially

flat spectra but no statistically defined plateaus (Table 1; Fig. SF12).The weighted mean 40Ar/39Ar dates for the flat segments of thesespectra are 471.0 ± 1.0 and 462.5 ± 1.1 Ma for muscovite and biotite,respectively. Total fusion results for larger muscovite (XKMs = 0.67)and biotite crystals from metapelite AF7 show symmetric apparent-age distributions with weighted mean dates of 476.4 ± 3.7 and 468.4 ±3.1 Ma for these minerals, respectively (Table 1; Fig. 5; Fig. SF12).

We interpret the 40Ar/39Ar dates of all four analyses as coolingages. The age differences among the four minerals are consistentwith differences in closure temperatures related to differences ingrain size. Collectively, the data are consistent with cooling at amoderate rate over the ca. 476–463 Ma interval, and we found noevidence to support protracted cooling in northernmost Conne-mara.

Garnet–staurolite and staurolite–sillimanite transition zonesPreviously published K–Ar ages from these zones range be-

tween 450 and 400 Ma for biotite and between 470 and 430 Ma formuscovite (Elias et al. 1988). To evaluate the significance of thisspread in dates, we analyzed muscovites, biotites, and phlogopitesfrom a variety of metasedimentary rocks (Tables 2 and 3; Fig. 5;Figs. SF1 and SF22).

Most 40Ar/39Ar analyses yield 39Ar weighted mean dates olderthan 470 Ma, but a few muscovites, one biotite, and one of thephlogopites yield 40Ar/39Ar dates that are significantly younger.We first report the >470 Ma dates, and then discuss the signifi-cance of the younger results.

Table 1. Summary of 40Ar/39Ar analyses for muscovite, biotite, and phlogopite from northern Connemara.

Sample RocktypeLocality(Nat’l. Grid Ref.) Mineral

Grain radius(range) (�m)

No. of grainsper analysisor weight

39Ar weightedmean date (Ma)

2�

error(Ma)

Mode(range)(Ma)

Plateau or flatsegment (Ma)

Garnet–staurolite zoneAF7 Metapelite 693.636 Muscovite 250–355 1–4 476.4 3.7 471–476 —

Biotite 250–355 1–5 468.4 3.1 469–476 —AF16 Metapelite 670.637 Muscovite 125–150 3.3 mg 469.2 1.5 — 471.0±1.0

Biotite 125–150 4.0 mg 452.4 1.5 — 462.5±1.1AF12 Metapelite 663.607 Biotite 125–250 1–5 481.5 4.3 471–483 —AF15 Mylonite — Muscovite — — 452.0 1.1 — 452.5±1.4AF40 Calcsilicate 662.583 Phlogopite (w/altered) 125–250 5.3 mg 472.9 2.4 455, 495 —AF49 Metapelite 804.574 Muscovite (crappy) 125–150 3.0 mg 450.7 1.2 — 452.3±1.0

Biotite 125–150 0.4 mg 454.7 3.3 — 459.8±2.5Staurolite–sillimanite zoneAF35 Metapelite 849.538 Biotite 150–250 1–7 472.2 4.3 466–474 —AF37 Calcsilicate 842.540 Phlogopite 100–150 4.7 mg 476.3 4.5 483 —AF66 Calcsilicate 760.554 Phlogopite (w/altered) 125–150 2.1 mg 399.0 3.0 — 418.5±1.6 (432)AF67 Semipelite — Muscovite — — 452.1 2.4 — 448.7±1.8AF70 Metapelite 914.508 Biotite 125–250 28.2 mg 459.0 0.7 0.7 466.8±1.0

Biotite 125–250 — 474.6 5.4 470–481 (477) —AF73 Metapelite 915.508 Biotite 200–350 1–2 441.2 443, 473 —AF79 Metapelite — Biotite 150–300 — 472.8 1.0 — 473±1

Note: Preferred values are noted by bold type.

Table 2. Summary of 40Ar/39Ar analyses for biotite and phlogopite from central Connemara.

Sample Rocktype

Locality(Nat’l.Grid Ref.) Mineral

Grain radius(range) (�m)

No. ofgrains peranalysis orweight

No. ofanalyses

39Ar weightedmean date (Ma)

2� error(Ma)

Mode (range)(Ma)

Plateau or flatsegment (Ma)

AF27 Calcsilicate 835.490 Phlogopite 250–350 3–6 9 468.3 2.5 443, 479 —AF28 Calcsilicate 836.492 Biotite 225–450 4 10 453.9 1.7 442, 447, 472 —AF32 Calcsilicate 811.500 Phlogopite 300 1–4 10 473 2 460–465 —AF34 Metapelite 806.501 Biotite 300–400 Furnace — 444.9 2.9 — 448.5±2.0

Biotite 150–350 1–5 14 459.3 6.9 460–465 —AF38 Calcsilicate 718.531 Phlogopite 400–800 Furnace — 445.8 0.8 — 455.8±1.3

Phlogopite 500–1250 1 10 466.9 2.2 472 —AF52 Calcsilicate 856.505 Phlogopite 125–250 9.4 mg — 454.8 1.0 — 456.7±2.6

Note: Preferred values are noted by bold type.

Friedrich and Hodges 1265

Published by NRC Research Press

Can

. J. E

arth

Sci

. Dow

nloa

ded

from

ww

w.n

rcre

sear

chpr

ess.

com

by

51.1

71.9

6.78

on

04/2

8/17

For

pers

onal

use

onl

y.

Total fusion analyses of biotite (XAnn = 0.41) from one staurolite–garnet schist (AF12), biotites from three staurolite–garnet–sillimaniteschists (AF35, XAnn = 0.37; AF70, XAnn = 0.43; AF79, XAnn = 0.45),and phlogopites from two calcsilicate rocks (AF37, XAnn = 0.06;AF40, XAnn = 0.01) yielded similar dates (Table 1; Fig. SF22). BiotiteAF12 has a 39Ar weighted mean date of 478.0 ± 5.1 Ma, whichbroadly overlaps with the sample mode of 471–483 Ma. Total fu-sion analyses of the other micas yield 39Ar weighted mean dates of472.2 ± 4.3 (AF35), 474.6 ± 5.4 (AF70), 476.3 ± 4.5 (AF37), and 472.9 ±2.4 Ma (AF40). More precise incremental release analyses ofbiotites AF70 and AF79 yielded 39Ar weighted mean dates of 466.8 ±1.0 and 473.1 ± 1.1 Ma for �70% of the released gas. In all casesexcept the AF40 phlogopite, these dates broadly overlap with thesample modes of the respective probability distributions, so weinterpret them to yield approximate cooling ages (cf. Table 1). Incontrast, the probability distribution of phlogopite AF40 showsdistinctive modes at 455 and 495 Ma, inconsistent with the 39Arweighted mean age, indicating that the sample is probably con-taminated by heterogeneously distributed excess 40Ar.

Younger 40Ar/39Ar dates were obtained for micas in samplescollected in or near fault zones. Muscovite AF15 from a myloniticmetasedimentary rock collected within the Renvyle–Bofin Slide(Fig. 5). Conventional furnace analysis of this sample yielded aplateau date of 452.5 ± 1.4 Ma. Given the probability that themylonitic fabric developed at low temperature (for example,there is no evidence for plastic deformation of feldspars in thinsection) and the fact that the muscovite helps define the myloniticfabric, we interpret 452 Ma as a close estimate of the time ofmylonitization. Two additional samples were collected near smallerfaults that offset the Renvyle–Bofin Slide. The incremental releasespectra of muscovite (XKMs = 0.6) and biotite (XAnn = 0.39) from astrongly foliated garnet schist rock (AF49) displayed flat segmentswith mean dates of 452.3 ± 1.0 and 458.1 ± 2.3 Ma, respectively(Table 1; Fig. SF12). Muscovite from another garnet schist (AF67,XKMs = 0.64) has a plateau date of 448.7 ± 1.8 Ma. The similarT

able

3.Su

mm

ary

of4

0A

r/39A

ran

alys

esfo

rm

usc

ovit

ean

dbi

otit

efr

omso

uth

ern

Con

nem

ara.

Sam

ple

Roc

kty

pe

Loca

lity

(Nat

’l.G

rid

Ref

.)M

iner

alLa

belo

fan

alys

isG

rain

radi

us

(ran

ge)(

�m

)

No.

ofgr

ain

sp

eran

alys

isor

wei

ght

No.

ofan

alys

es

39A

rw

eigh

ted

mea

nda

te(M

a)

2� erro

r(M

a)

Mod

e(r

ange

)(M

a)Pl

atea

uor

flat

segm

ent

(Ma)

Mig

mat

itic

sill

iman

ite–

K-f

eld

spar

zon

eA

F18

Met

apel

ite

833.

415

Bio

tite

b.a

>10

00M

app

ing

1127

6.5

5.1

270,

320

—B

ioti

teb1

150–

350

1–3

1048

9.4

3.3

453,

485,

515

—B

ioti

teb.

5137

51

Lase

rst

ep44

5.3

2.1

445±

2In

v.sa

ddle

(370

–507

)bi

otit

eb.

furn

250–

750

10.3

mg

Furn

ace

446.

82.

344

7±2

Inv.

sadd

le(3

63–5

07)

AF1

8M

etap

elit

e83

3.41

5M

usc

ovit

em

120

0–25

01

1045

7.4

2.7

459

—M

usc

ovit

e8.

2.1

400–

500

3La

ser

step

456.

52.

2—

456

±2M

usc

ovit

em

f.38

200–

250

2La

ser

step

455

2—

453±

1A

F22

Met

apel

ite

823.

443

Bio

tite

b81-

9025

0-40

01–

210

458.

03.

645

9—

Bio

tite

b15-

2075

–125

15–2

010

Cal

cse

p/4

47.6

6.6

440,

446

—B

ioti

teb3

6-40

200–

400

1–3

5C

alc

sep

.45

3—

AF4

4Pe

gmat

ite

814.

410

Bio

tite

2.10

0�

300

3La

ser

step

387.

65.

9—

418±

6A

F45

Met

apel

ite

814.

410

Bio

tite

b112

5–15

01–

710

434.

53.

143

7,46

3—

Bio

tite

b8.1

640

0–50

01

Lase

rst

ep40

0.8

2.9

—St

air

step

(327

–420

)A

F45

Met

apel

ite

814.

410

Mu

scov

ite

m11

.13

700

3La

ser

step

442.

72.

2—

439

±3M

usc

ovit

em

.b25

00M

app

ing

844

0.2

2.7

440–

450

—M

usc

ovit

em

8.8

400–

500

4La

ser

step

432.

72.

6—

430

±3M

usc

ovit

em

f.4

�20

08

Lase

rst

ep43

3.1

5.4

—42

9±5

No

te:P

refe

rred

valu

esar

en

oted

bybo

ldty

pe.

Table 4. 40Ar diffusion data for mica.

Mineral XAnn Do (cm2/s) EA (kcal/mol K) Reference

Biotite 0.54 0.077 47.1±1.5 Harrison et al. 19850.71 0.4 50.5±2.2 Grove and Harrison 1996

Muscovite — 0.00039 43 Hames and Bowring 1994Phlogopite 0.04 0.75 57.9±2.6 Giletti 1974

Table 5. Closure temperatures for Argon diffusion in mica fromConnemara.

Sample MineralGrainradius (mm)

Coolingrate (°C/Ma)

Closuretemperature (°C)

AF7 Biotite 300 6–20 322–340Muscovite 300 6–20 372–394

AF12 Biotite 200 6–20 311–328AF16 Biotite 150 3–20 294–320

Muscovite 150 3–20 318–349AF35 Biotite 200 6–26 311–332AF37 Phlogopite 125 6–26 392–414AF70 Biotite 200 6–26 311–332AF79 Biotite 200 6–26 320–348AF38 Phlogopite 600 6–26 376–402AF52 Phlogopite 200 6–26 406–429AF18 Muscovite 225 7–14 344–356

Muscovite 450 7–14 368–381AF22 Biotite 100 7–14 294–302

Biotite 325 7–14 327–336

Note: Closure temperatures for a mineral pair from the same sample(muscovite–biotite) or from the same locality (phlogopite and biotite) were cal-culated assuming internally consistent cooling rates. In all calculations, weassumed that a cylinder best describes the mica diffusivities. The closure tem-perature equation is from Dodson (1973). General uncertainites are ±50 °C.

1266 Can. J. Earth Sci. Vol. 53, 2016

Published by NRC Research Press

Can

. J. E

arth

Sci

. Dow

nloa

ded

from

ww

w.n

rcre

sear

chpr

ess.

com

by

51.1

71.9

6.78

on

04/2

8/17

For

pers

onal

use

onl

y.

muscovite dates suggest an important phase of post-M3 deforma-tion and dynamic recrystallization at ca. 450 Ma. The older AF49biotite date probably reflects excess 40Ar.

Another anomalously young date was provided by phlogopitefrom a calc-silicate rock (AF66). Its spectrum displayed a flat seg-ment with a 39Ar weighted date of 418.5 ± 1.6 Ma. This date issimilar to many K–Ar dates obtained by Elias et al. (1988) from thesame region. Most likely, it reflects local resetting by hydrothermalfluids related to intrusion of the post-orogenic Galway batholith.

Sillimanite–K-feldspar zoneIn this zone, we concentrated on biotites and phlogopites from

metasedimentary samples collected near Derryclare Lough areain central Connemara (Fig. 5; Figs. SF1 and SF22; Table 2). BiotiteAF34 (XAnn = 0.29) from a sillimanite–K-feldspar–biotite gneiss hasa plateau date of 448.5 ± 2.0 Ma. Phlogopite AF38 (XAnn = 0.04)from a metasomatic diopside rock yields a plateau date of 455.8 ±1.3 Ma, consistent with a U–Pb date of ca. 462 Ma for titanite from

the same sample (Friedrich 1998; Friedrich et al. 1999b). Phlogo-pite AF52 (XAnn = 0.02) from another metacarbonate rock has amore complicated release spectrum with a flat segment corre-sponding to a date of about 457 Ma. We interpret the plateau andflat segment dates of these three samples as cooling ages.

These cooling ages are significantly younger than those of tri-octahedral micas with similar compositions and grain sizes fromthe garnet–staurolite and staurolite–sillimanite transition zonesof northern Connemara (Tables 1 and 2; Fig. 5). The boundarybetween the cooling age provinces broadly coincides with themuscovite-breakdown isograd and the crest of the Connemaraantiform.

Several additional samples from the sillimanite–K-feldspar zoneyielded 40Ar/39Ar dates that were complicated and difficult to in-terpret unambiguously. The frequency distribution of laser fusiondates for phlogopite AF27 (Xann = 0.05) shows distinctive modes at479 and 443 Ma. Petrographically, this sample contains phlogo-

Fig. 5. Generalized geological map of Connemara showing the 40Ar/39Ar mica dates as a function of metamorphic grade and distance frommajor intrusions. Only those results are shown that yield a single date. Refer to Tables 1–3 and the supplementary data (Files S1, S6–S12)2 foradditional results. Ages and their 2� uncertainties are reported in Ma.

Friedrich and Hodges 1267

Published by NRC Research Press

Can

. J. E

arth

Sci

. Dow

nloa

ded

from

ww

w.n

rcre

sear

chpr

ess.

com

by

51.1

71.9

6.78

on

04/2

8/17

For

pers

onal

use

onl

y.

pites of two different grain sizes, one somewhat larger than500 �m and one significantly smaller. It may be that the twomodes in the release spectrum correspond to different coolingages for the two phlogopites. However, the ca. 479 Ma date seemsinconsistent with other data from the area, particularly U–Pb ti-

tanite dates that may record cooling through a much higher clo-sure temperature (≥650 °C; Friedrich 1998; Hodges 2014; Schoene2014) roughly 15 million years later. This older mode may reflectsome degree of excess 40Ar contamination, but Friedrich (1998)provided textural evidence that the titanites crystallized hy-

Fig. 6. Time–space diagram of 40Ar/39Ar dates for the Connemara region. The data are grouped into four categories, based on theirrelationship to the geological context. Contraints on the timing of major regional heating significantly above the Argon–mica closuretemperatures were provided by U–Pb zircon dating of the Cashel–Lough Wheelaun and the Currywongaun mafic-ultramafic plutons (Friedrichet al. 1999a). Older 40Ar/39Ar dates are interpreted to yield excess Argon (Category 4). The shaded areas (Category 1) contain all permissiblecooling ages as constrained by the age of the magmatic complex, peak metamorphism (Friedrich et al. 1999b), fluid infiltration (Category 3),and the Silurian hiatus. Category 2 dates may represent extensional unroofing of the Connemara orogenic wedge. The north–south axis is notdrawn to scale.

1268 Can. J. Earth Sci. Vol. 53, 2016

Published by NRC Research Press

Can

. J. E

arth

Sci

. Dow

nloa

ded

from

ww

w.n

rcre

sear

chpr

ess.

com

by

51.1

71.9

6.78

on

04/2

8/17

For

pers

onal

use

onl

y.

drothermally at low ambient temperatures, perhaps related tofar-reaching fluid circulation related to emplacement of theOughterard granite.

A similar interpretation can be proposed for data from anotherphlogopite separate (AF32, XAnn = 0.03; Fig. SF22); although itslaser fusion dates have a 39Ar weighted mean of 473.7 ± 2.3 Ma, thefrequency distribution is positively skewed away from a singlemode between 460 and 465 Ma. Biotite AF28 (XAnn = 0.42) from abiotite–sillimanite–garnet schist has a 39Ar weighted mean dateof 453.9 ± 21.7 Ma, but its population distribution displays threemodes. The youngest, at 442–447 Ma, is similar to the young modefor the AF27 phlogopite, and this range may reflect localized re-setting of Ar systematics by hydrothermal fluid interactions ordynamic recrystallization during deformation.

Migmatitic portions of the sillimanite–k-feldspar zoneOur 40Ar/39Ar research on migmatitic rocks of the sillimanite–

K-feldspar zone was designed to augment data presented else-where by Friedrich et al. (1999a, 1999b) pertinent to the detailedthermal evolution of southern Connemara. Our earlier work sug-gested relatively rapid cooling over the 460–377 °C temperatureinterval between 460 and 454 Ma near the northern distributionlimit of anatectic melt products in the sillimanite–K-feldsparzone. As described in the following paragraphs, the 40Ar/39Ardates for samples collected closer to the main outcrop region ofthe Connemara igneous complex are notably younger than themica ages reported by Friedrich et al. (1999b) and much youngerthan the crystallization ages of igneous phases in the Connemaracomplex.

Previous workers found abundant evidence that the K–Ar horn-blende chronometer cannot be applied with confidence to amphi-bolites from southern Connemara (Miller et al. 1991). The mostcommonly proposed explanation for the tremendous range inhornblende apparent ages was the disturbance of isotopic system-atics by hydrothermal fluids related to intrusion of the ca. 400 MaGalway batholith and related “post-orogenic” plutons (Leggo et al.1966). To seek additional support for this hypothesis, we analyzedmicas from metapelitic schists and older granitic dikes collectedat various distances from late plutons. In general, the muscovitesyielded relatively simple release spectra with plateaus or flat seg-ments, whereas the biotite results were much more complex.

Biotite AF22 is from a schist collected 10 km away from thepost-orogenic Roundstone granite (Fig. 5). Petrographic analysisprovided no evidence of hydrothermal alteration (Fig. SF32). Laserfusion analyses yielded symmetric frequency distributions with39Ar weighted mean dates of 453 and 440 Ma for the �400 and�150 �m size fractions, respectively (Table 2), consistent with alower closure temperature for the smaller size fraction. Substan-tially older than the 420 Ma crystallization age of the granite,these dates probably represent cooling after M3 metamorphism.

Collected 5 km south of the AF22 outcrop, AF18 is a garnet–K-feldspar schist containing biotite (XAnn = 0.53), secondary musco-vite (XKmu = 0.74) that replaces K-feldspar, and chlorite that fillsfractures in garnet (Fig. 4). Incremental heating analysis of coarse(900 �m) muscovite yielded a plateau date of 456.5 ± 2.2 Ma,whereas a finer-grained fraction (450 �m) had a plateau date of453.6 ± 0.9 Ma (Fig. SF22). Laser fusion analyses of fragments of a450 �m single crystal of this muscovite had a 39Ar weighted meandate of 457.4 ± 3 Ma, which lies within the uncertainty of the twoother analyses. The similarity of these ages with those obtainedfor M3 micas farther north in the “migmatite” zone of Connemarasuggests that retrogression in AF18 occurred at a relatively hightemperature. Laser total fusion analyses for two aliquots of biotite(Xann = 0.53) from this sample have 39Ar weighted mean ages of489.1 ± 3.3 and 275.6 ± 5.1 Ma (Table 3; Fig. SF12). To better under-stand this complex behavior, we also performed a furnace incre-mental heating experiment on a 10 mg biotite aliquot with theresistance furnace, and a laser incremental heating experiment

on a single 750 �m crystal (Table 3; Fig. SF12). Both experimentsyielded similar 39Ar weighted mean dates of 446.8 ± 2.3 and445.3 ± 2.1 Ma, respectively. In both cases, the spectra displayedsimilar stair-step shapes with age components of ca. 508 andca. 365 Ma (Fig. SF12). Given U–Pb constraints that M3 occurredover the 460–474 Ma interval, the 508 Ma component must reflectcontamination with excess 40Ar. The ca. 365 Ma component ismuch younger than other biotite cooling ages from this area anddoes not correspond to a previously identified thermal event. This40Ar/39Ar date, however, may correspond to one of several low-temperature fluid alteration events that affected the Connemararegion after emplacement of the Galway granites (cf. Jenkin et al.1992).

We also analyzed micas from two samples collected less than5 km east of the Roundstone granite (Fig. 4). Sample AF45 is fromthe paleosome of an anatectic pelitic gneiss that surrounds peg-matitic leucosome AF44. Furnace incremental heating analysisof the biotite (XAnn = 0.49) from the leucosome yields a plateaudate of 418.4 ± 6.5 Ma. Given the crystallization age of this pegma-tite (467 ± 2 Ma, U–Pb zircon; Friedrich et al. 1999b), this 40Ar/39Ardate probably reflects complete resetting during intrusion of theRoundstone granite, and possibly limited resetting at ca. 350 Maor younger.

The 40Ar/39Ar systematics of the biotite (XAnn = 0.49) in AF45 aremore complex (Figs. SF1 and SF22). Laser fusion analyses of ali-quots of up to seven grains have a 39Ar weighted mean date of434.5 ± 3.1 Ma. However, frequency distribution plots show threedistinct modes at ca. 420, 437, and 463 Ma (Fig. SF12). Laser incre-mental heating of another aliquot yielded a stair-step profile witha flat segment at 420.1 ± 6.5 Ma and a minimum age of 327 ± 20 Ma(Fig. SF12). The simplest interpretation of this spectrum is thatthe sample experienced a major Ar re-equilibration event atca. 420 Ma followed by a second, less profound event at ≤330 Ma.Muscovite (XKMs = 0.63) in AF45 occurs as fine-grained pinite andas large (up to 5000 �m), randomly oriented, secondary muscovitebooks (Fig. SF32). Laser incremental heating analyses of differentsize fractions yielded simple plateaus or flat segments corre-sponding to dates between 439 and 429 Ma (Fig. SF12; Table 3).Large (700 �m) crystals gave the oldest dates, whereas the smallestcrystals (200 �m) gave the youngest dates. Despite the positivecorrelation between grain size and apparent age, these cannot beclosure ages related to simple cooling after M3 metamorphismbecause AF45 is from a structural level below the older Silurianunconformity (Fig. 4). Conceivably, the young ages imply partialresetting associated with intrusion of the Roundstone granite.

Significance of the 40Ar/39Ar mica age patternThe 40Ar/39Ar mica dates for Connemara fall into four interpre-

tive groups (Fig. 6), based on the U–Pb geochronological con-straints of the main intrusive phases (Friedrich et al. 1999a), thetiming of peak metamorphism (Friedrich et al. 1999b), and otherpreviously published constraints from the geological record (e.g.,Cliff et al. 1993, 1996; Ryan and Dewey 2011; Chew and Strachan2013, Wellings 1998; Boyle and Dawes 1991; Yardley et al. 1987).The last Category can be defined because the ca. 400–420 Mapost-orogenic granites are the youngest, well-dated, potential causesfor resetting, and because dates much older than ca. 475 Ma areinconsistent with U–Pb zircon constraints on the age of the earlyGrampian Cashel–Lough Wheelaun gabbro (Friedrich et al. 1999a).

Category 1 agesMica ages that record the thermal relaxation of the Grampian

orogen show a spatial pattern opposite that proposed by Eliaset al. (1988) (Figs. 5 and 6), but our study does not include any argonresults for hornblende. The oldest reliable Category 1 ages—ranging from 478 to 467 Ma—were obtained for garnet–staurolitezone and staurolite–sillimanite transition zone samples collectedin northern Connemara. To the south, toward the Connemara

Friedrich and Hodges 1269

Published by NRC Research Press

Can

. J. E

arth

Sci

. Dow

nloa

ded

from

ww

w.n

rcre

sear

chpr

ess.

com

by

51.1

71.9

6.78

on

04/2

8/17

For

pers

onal

use

onl

y.

igneous complex, Category 1 biotite ages decrease to as youngas 453–440 Ma in the highest temperature portions of thesillimanite–K-feldspar zone. The only complication in this patternoccurs along the extreme northern margins of the Dalradian out-crop region, where biotite ages from staurolite-zone rocks areslightly younger (468–463 Ma) than those a few kilometers tothe southeast in the staurolite–sillimanite transition zone (478–467 Ma). One possible explanation for the relatively young ages innorthernmost Connemara is limited argon degassing related tolate extensional deformation (Boyle and Dawes 1991).

Category 1 ages are defined as related to cooling subsequent tothe M3 event in northern Connemara and to cooling followinganatexis in southern Connemara (Figs. 6 and 7). No significantdiscontinuity occurs in the pattern across the sillimanite isograd(Figs. 5 and 6), which has been suggested by some authors ascorresponding to the effective northern distribution limit of pro-grade M3-metamorphic assemblages (Yardley 1976; Tanner andShackleton 1979). Even if the northern limit of prograde M3-metamorphism extends only as far north as sillimanite-in isograd,as has been proposed elsewhere (Yardley et al. 1987), the thermaleffects of M3 extended throughout Connemara (cf. Boyle andDawes 1991).

Three explanations could satisfy the general north-to-southyounging of Category 1 ages. The first, which is consistent withsome U–Pb titanite and monazite dates of less than 470 Ma incentral and southern Connemara (Cliff et al. 1996; Friedrich et al.1999b; Friedrich 1998) is that the older 40Ar/39Ar dates of northern

Connemara represent contamination by excess 40Ar. We cannotdisprove this interpretation with the data at hand; our sampleswere so radiogenic that the use of isotope correlation diagrams toevaluate this problem was impractical. However, the consistencyof most northern Connemara 40Ar/39Ar ages from a single subareais not a pattern typical of pervasive excess 40Ar contamination.Most well-documented examples of sample suites contaminatedwith excess 40Ar are characterized by irreproducible results formultiple aliquots of the same sample, wide variations in datesfrom single outcrops, and (or) complex release spectra (Harrisonand McDougall 1981; von Blanckenburg and Villa 1988; Arnaudand Kelley 1995). Because many of the young titanites found inConnemara are thought to have grown during late- or post-M3metasomatic events (Friedrich 1998; Friedrich et al. 1999b) andbecause we cannot preclude a similar origin for the few youngmonazites dated in the region, we see no reason to reject the40Ar/39Ar evidence for the M3 metamorphism in northern Conne-mara at ca. 474–470 Ma.

A second possible explanation for the north–south age trend isthat the region was unroofed diachronously—first in the north—about the east–west-trending Connemara antiform (e.g., Eliaset al. 1988). This interpretation is consistent with the coincidenceof the abrupt break in cooling ages with the axial trace of thisantiform, but inconsistent with evidence from (i) paleomagneticstudies that northern Connemara had already cooled to below�510–580 °C before folding (Morris and Tanner 1977; Robertson1988), and (ii) structural studies which show that the antiform

Fig. 7. Temperature–time (Tt) diagram showing 40Ar/39Ar muscovite, biotite, and phlogopite cooling ages and other dates for northern andsouthern Connemara. Refer to Fig. 6 and Friedrich et al. 1999b for additional constraints on the timing of deformation (e.g., D3) and peakmetamorphism (e.g., metapelite AF45). The thick grey curved lines mark a likely Tt-path for northern and southern Connemara, respectively.The thin vertical lines mark magmatic or hydrothermal resetting events at times when the ambient temperature was significantly below theclosure temperature intervals for mica at Connemara (Category 3 dates). The 452 Ma 40Ar/39Ar age of a metasedimentary mylonite (sample AF15) from northern Connemara is interpreted to yield the time of extensional mylonitization at an ambient temperature below the closure ofArgon diffusion in muscovite. The timing of Grampian compressional and extensional deformation are best determined in northernConnemara (cf. Fig 6), while in southern Connemara thermochronometers recorded cooling following the dominant Barrovian metamorphicevent. The youngest Category I ages in southern Connemara coincide with the timing of extension in northern Connemara, consistent withcooling during extensional collapse. The Silurian nonconformity provides a hard constraint on the end of post-Barrovian cooling andextensional collapse at Connemara. [Colour online.]

1270 Can. J. Earth Sci. Vol. 53, 2016

Published by NRC Research Press

Can

. J. E

arth

Sci

. Dow

nloa

ded

from

ww

w.n

rcre

sear

chpr

ess.

com

by

51.1

71.9

6.78

on

04/2

8/17

For

pers

onal

use

onl

y.

formed as a late (D4) feature (Leake et al. 1983). Furthermore,structural, petrologic, and thermobarometric constraints on pa-leodepth trends during M3 suggest synchroneity across the Con-nemara region.

We favor a third explanation consistent with all available geo-chronologic and structural data: that intrusion of intermediate tofelsic phases of the Connemara igneous complex over the ca. 468–463 Ma interval maintained M3 temperatures high in southernConnemara while regional cooling occurred in northern Conne-mara. Textural relations in dated metamorphic rocks further sug-gest that structures and fabrics commonly attributed to D3 areolder in the north than in the south; for example, micas thatdefine the S3 schistosity in northern Connemara had cooled wellbelow �400 °C while synkinematic micas were still growing atupper amphibolite facies in southern Connemara. While this modelrequires high lateral temperature gradients of up to 20 °C/km, wesuggest that such a thermal regime would be consistent with a arc-collisional setting like that of Connemara.

Results of thermal modeling studies further support our expla-nation. Reverdatto and Polyansky (2004) found that a lateral ther-mal gradient of �14 °C/km may be sustained over 5–6 Ma if abasaltic-type heat source is assumed to have occurred structurallyabove the Connemara complex. Dewey and Ryan (2016) concludedthat Barrovian-type PT conditions, as observed in Connemara,may be achieved in under 5 Ma by an actively thrusting hot hang-ing wall over colder footwall. In their model, significant heatingof the footwall is restricted to within 7 km of the thrust contact.

Category 2 agesCategory 2 ages may provide close minimum estimates on the

age of the Renvyle–Bofin Slide and related late structures in cen-tral and northern Connemara (Figs. 5–7). If this interpretation iscorrect, extensional deformation in this part of the Grampianorogen was taking place no more than about 15 million years afterpeak metamorphism associated with large-scale crustal shorteningand Barrovian-type metamorphism in the island arc – continent col-lisional environment.

Category 3 agesMost Category 3 ages record resetting of the Ar systematics of

micas throughout much of Connemara at ca. 420 Ma (Figs. 6 and7). This event cannot be regarded as part of the Grampian orogeny,because, both, north- and south-vergent structures are mucholder and Dalradian rocks had been covered unconformably bymarine Silurian sedimentary deposits by 435 Ma (Leake andTanner 1994). Instead, the spread of Category 3 ages may be re-lated to intrusion of granites that resulted in context of north-ward subduction beneath the Laurentian margin from LateOrdovician to Devonian time (Mayoian phase of Dewey and Ryan2016). If this interpretation is correct, the thermal effects of thepost-orogenic magmatic event extended far beyond southern Con-nemara, perhaps through a regional metasomatic mechanismsimilar to that invoked for older ca. 462 Ma metasomatic depositsin Connemara (Yardley et al. 1991; Cliff et al. 1993). Such a largeregional extent of hydrothermal activity related to emplacementof the Galway batholith has been suggested by Jenkin et al. (1992)for retrograde alteration in southern Connemara.

Two other anomalous dates may be related to similar alterationevents. The frequency distribution plots for AF27 phlogopite andAF28 biotite display modes between 450 and 440 Ma. These datesare significantly younger than those obtained for compositionallysimilar minerals at the same structural level in the sillimanite–K-feldspar zone. Neither AF27 or AF28 show evidence of post-metamorphic dynamic recrystallization, but irregularly spaced�1–3 mm quartz-filled fractures occur orthogonal to the S3 folia-tion defined by biotite and phlogopite in these samples. There-fore, we tentatively attribute the 450–440 Ma dates to retrograderesetting. An extremely young near-plateau segment on the incre-

mental heating spectrum for AF45 biotite (ca. 330 Ma) suggests thatsome alteration may have occurred long after intrusion of the Gal-way batholith and related plutons. However, the AF45 biotite incre-mental heating analysis was poor, with very large uncertainties, andlaser analyses of the same sample yielded discrepant results.

Thermal and structural evolution of ConnemaraWe integrate the new 40Ar/39Ar data along with previously pub-

lished geochronologic, structural, metamorphic, and paleomag-netic data and suggest a refined thermal and tectonic model of theGrampian orogeny at Connemara. We refer to three simplifieddevelopmental north–south cross sections (Fig. 8), which wereconstructed from youngest to oldest (see Appendix B), while dis-cussion of the cross section follows geological convention, fromoldest to youngest. The oldest section (Fig. 8A), drawn to representthe structural pattern at ca. 474–470 Ma, the time of syntectonic(D2) magmatism (see fig. 6 in Friedrich et al. 1999b and referencestherein), is dominated by a regional-scale recumbent fold of D2age (Yardley et al. 1987). The D2 event, responsible for the foldnappe and the S2 schistosity throughout Connemara, was at leastin part synchronous with M2 metamorphism. Although directconstraints on the initiation age and duration of D2 and M2 arelimited, Wellings (1998) suggested that the ca. 474 Ma Currywon-gaun gabbro intruded during the regional D2 and M2.

Numerous 40Ar/39Ar cooling ages for M3 sillimanite-grade micasin northern Connemara are within uncertainty of this age or onlyslightly younger at 475 Ma (Figs. 5, 6, and 7), so we conclude thatthe transition from D2/M2 to D3/M3 in northern Connemaraoccurred within a few million years. On a regional scale, thistransition may simply represent a deformational continuum atprogressively higher temperatures.

Both D3 and M3, as defined by F3 folds and S3 axial planarschistosity in Dalradian rocks, affected all of Connemara (Fig. 9;e.g., Barber and Yardley 1985; Treloar 1985; Yardley et al. 1987).Both D3 and M3 appear to have been very short-lived phenomenain northern Connemara (ca. 470 Ma), but they may have lasted7–8 million years in southern Connemara as a result of moreprolonged and more voluminous magmatic activity (470–463 Ma;Leake 1989; Friedrich et al. 1999a, 1999b).

M3= and D3 reached their peak intensities at 468–466 Ma insouthern Connemara, coincident with widespread quartz dioritemagmatism and anatexis of metasedimentary rocks (Fig. 8B;Tanner et al. 1997; Friedrich et al. 1999b). At that time, a very highthermal gradient persisted across Connemara, with temperaturesof ≥750 °C in the southernmost sillimanite–K-feldspar zoneand <300 °C in northern Connemara (Figs. 7, 8B, and 9). D3 struc-tures of this age include an intense foliation in anatecticmetapelite rocks (Fig. 3F) and wide ductile zones, especially insouthwestern Connemara above the Mannin thrust (Leake et al.1983), both of which contrast significantly with the older F3 foldsand tectonite fabrics that developed at lower temperaturesthroughout Connemara.

Figure 8C represents the regional structure after developmentof the D4 Connemara antiform, but before late extensional defor-mation. In combination with paleomagnetic data, the 40Ar/39Arcooling ages provide tight constraints on the age of the antiform.Paleomagnetic analyses of mafic rocks indicate that the northernConnemara gabbros cooled below �510–580 °C, the nominalrange for the Curie temperature of magnetite, before folding ofthe Connemara antiform, whereas the southern Connemara gab-bros cooled below this temperature range after development ofthe antiform (Morris and Tanner 1977; Robertson 1988). Thenorth–south variation in Category I 40Ar/39Ar mica ages (Fig. 7)thus brackets formation of the Connemara antiform to betweenca. 468 and ca. 463–462 Ma. Additional confirmation of the lowerage bracket comes from the U–Pb crystallization age of 462.5 Mafor the post-F4 Oughterard granite (Tanner et al. 1997; Friedrich

Friedrich and Hodges 1271

Published by NRC Research Press

Can

. J. E

arth

Sci

. Dow

nloa

ded

from

ww

w.n

rcre

sear

chpr

ess.

com

by

51.1

71.9

6.78

on

04/2

8/17

For

pers

onal

use

onl

y.

Fig. 8. Schematic evolutionary cross-sections summarizing the major thermal and structural evolution of the Connemara region in threetime slices. Refer to Appendix B for details regarding the restorations (Friedrich 1998). The three panels emphasize the tight constraints ontiming of deformation, metamorphism, and cooling. (A) Restoration starts at the time of syn-kinematic emplacement of mafic and ultramaficplutons (Wellings 1998). The main penetrative metamorphic fabrics (D2 and D3) formed at this time interval. The orientation of the regional-scale isograde implies more voluminous mafic magmatism in southern Connemara. (B) The interval between 468 and 467 Ma marks the timeof peak Barrovian metamorphism with anatexis, a switch from northerly to southerly vergence of orogenic fabrics and structures, and a shiftin magmatism from mafic to felsic (e.g., Leake and Tanner 1994). Radiometric dates are from Friedrich et al. (1999a, 1999b). [Colour online.]

1272 Can. J. Earth Sci. Vol. 53, 2016

Published by NRC Research Press

Can

. J. E

arth

Sci

. Dow

nloa

ded

from

ww

w.n

rcre

sear

chpr

ess.

com

by

51.1

71.9

6.78

on

04/2

8/17

For

pers

onal

use

onl

y.

et al. 1999a). The persistence of “late” M3 high-temperatures insouthern Connemara during development of the F4 Connemaraantiform may help explain the asymmetry in deformational styleacross the crest of the antiform, with more intense deformationand shearing in the southern limb (locally referred to as “the steepbelt”, e.g., Leake and Tanner 1994). Formation of the antiform waslinked kinematically with ductile deformation of the quartz dior-ites and gabbros of southern Connemara along an incipient Man-nin thrust-shear zone that eventually resulted in formation of thebrittle Mannin thrust.

At ca. 462 Ma, the waning stages of Connemara igneous com-plex activity maintained temperatures in the southern part of theregion relatively high, perhaps more than 150 °C higher thantemperatures in northern Connemara. However, extensive hy-drothermal systems extended northward from the igneous com-plex, leading to local resetting of mica chronometers and thecrystallization of metasomatic assemblages such as the titanite-bearing diopside rocks described by Yardley et al. (1991) and datedby Friedrich (1998). Similar episodes of metasomatic or hy-drothermal alteration events appear to have occurred more re-cently in the history of the region, consistent with the occurrenceof Category 3 dates (Fig. 6).

The youngest structures for which we have age constraints arethe extensional faults of northern Connemara. Last movementalong the Renvyle–Bofin Slide occurred at ca. 452 Ma based on the40Ar/39Ar muscovite crystallization age of mylonite AF15. This ageis consistent with the recent reinterpretation of the structural age ofthis fault to be post-D3 (Wellings 1998). It seems likely that exten-sional faulting entirely postdates Grampian deformation at Conne-mara, but it was concurrent with rapid exhumation and cooling.

DiscussionWe assimilated thermochronological (Fig. 7), structural (Fig. 8),

and thermobarometric data in a refined PTt-path (Fig. 9) and estab-lished relationships between the regional features and their geody-

namic context (Fig. 10). We then discuss the arc–continent collisionmodel proposed by Dewey (2005) and Dewey and Ryan (2016).

Constraints on the PTt evolution of deformationAt Connemara, Dalradian metasedimentary strata of Lauren-

tian affinity document progressive deformation (D1–D4) andmetamorphism, which captured the transition from north- tosouth-vergent orogenic deformation under clockwise PT condi-tions (Fig. 9). First, multi-phase north-vergent deformation oc-curred under prograde- and peak-pressure conditions and coevalwith emplacement of calcalkaline mafic to ultramafic plutons.The extra heat source lead to Barrovian metamorphism, crustalanatexis, and formation of migmatite under peak-temperature,but decreasing pressure conditions. Next, progressively localizedsouth-vergent deformation accommodated crustal-scale shorteningunder decreasing PT conditions. Last, exhumation, extensional de-formation, granitic intrusion, and hydrothermal circulation docu-ment the waning stages of the orogeny.

The thermal peak outlasted north-vergent fold-nappe forma-tion such that cooling ages (Category I, Fig. 6) provide a lower limiton the timing of this deformation (D3). But because southernConnemara also experienced Barrovian metamorphism at highertemperatures, only the Category 1 cooling ages (475–470 Ma) ofnorthern Connemara are useful to define the timing of D3 (Figs. 7and 9), overlapping with ages of the mafic syn-D2 and syn-D3plutons dated by Friedrich et al. (1999a). Exhumation with decom-pression of about 4 kbar must have occurred relatively rapid tolimit cooling between ca. 474 and 472 Ma to <150 °C (Figs. 7 and 9).This yields an exhumation rate of �7 km/Ma, which is comparableto exhumation of modern orogens such as Taiwan (Liu et al. 2000;Willett et al. 2003; Byrne et al. 2011).

In southern Connemara, U–Pb geochronological constraintsplace the anatectic Barrovian event at 468–467 Ma, coeval withintermediate quartz-diorite intrusions (Friedrich et al. 1999b).Therefore, significant cooling from peak temperatures (�750 °C,Yardley et al. 1987) to the highest closure temperature inferred for

Fig. 9. Pressure–temperature–time (PTt) diagram for Connemara. The PT data are from Yardley et al. (1987) and Boyle and Dawes (1991). Thepath has been refined by integration with geochronological results from Friedrich et al. (1999a, 1999b), and this study. Numbers are given inMa and relate to event ages discussed in text. The blue line represents the clockwise PT-path representative for the garnet zone of northernConnemara. The red line represents the PT-path for the migmatite zone in southern Connemara. Filled circles mark well-known time markerson the PT diagram (peak pressure, peak temperature). Small open circles represent geochronological constraints from Friedrich et al. (1999a,1999b), while large open circles represent the best estimate for position of Category I cooling ages. [Colour online.]

Friedrich and Hodges 1273

Published by NRC Research Press

Can

. J. E

arth

Sci

. Dow

nloa

ded

from

ww

w.n

rcre

sear

chpr

ess.

com

by

51.1

71.9

6.78

on

04/2

8/17

For

pers

onal

use

onl

y.

mica in this study (�430 °C) occurred between ca. 468 and 460 Maat a rate of �40 °C/Ma. Category I 40Ar/39Ar mica thermochronom-eters record cooling from �430 °C to <320 °C between ca. 460 and452 Ma, at �10 °C/Ma. A lateral thermal gradient of up to 30 °C/kmpersisted between ca. 472 and 460 Ma (Figs. 7, 8B, and 9).

Formation of the major south-verging thrust and range-scale anti-formpostdatecrustalanatexis,but itwasfinishedbythe timeofgraniteintrusion in eastern Connemara at ca. 463 Ma (Friedrich et al. 1999a,1999b). Therefore, D4 structures formed between 467 and 463 Ma. Theintervalbetweenca.468and467Mabestdefinesthereversal inpolarity,i.e., the time at which activity on north-vergent structures diminished,while it increased along south-vergent structures.

Geometric and kinematic boundary conditionsTo reconstruct the geometric context in which D2 and D3 fold

nappes formed, we combined structural and PTt information(Figs. 8 and 9). The overall geometry of the Grampian structures atConnemara define the mid-crustal core of a bi-vergent orogenicwedge. A basal north-verging decollement (not exposed) may beinferred based on the existence of north-vergent D2 and D3 foldnappes (Fig. 8A). The basal south-verging structure is representedby the Mannin thrust and related features, such as the Clifdensteep belt and the Connemara antiform (Fig. 8C). Decompressionand exhumation started along basal north-vergent structures, asD2 and D3 nappes began to detach at ca. 474 Ma to a depth equiv-alent of �5 kbar at ca. 470 Ma, and continued facilitated by south-vergent motion between ca. 467 and 463 Ma (Figs. 9 and 10). Insouthern Connemara, significant decompression occurred at ele-vated temperatures (above 400 °C), while in northern Connemaraat least some decompression occurred at temperatures below300 °C (cf. Boyle and Dawes 1991), resulting in an environmentthat favors brittle fracturing and hydrothermal fluid circulation.

To examine the relationship between the regional-scale struc-tures and their geodynamic context, we rescaled the Connemaraorogenic wedge to plate-boundary dimensions (cf. Fig. 8A and10A). In Fig. 10, we suggest a permissible sectional configurationthat would result from collision between the Laurentian marginand an intraoceanic island arc, as proposed by Dewey (2005). Animportant modification to this model was recently proposed byRyan and Dewey (2011) and tested by Dewey and Ryan (2016). Theseauthors no longer consider Connemara as an “outlier terrane”,which was emplaced to its present position following the Gram-pian orogeny, but that the Connemara crust should be consideredin-situ relative to the Laurentian margin of western Ireland; itrepresents the thinned Laurentian margin which was overriddenby the postulated Lough Nafooey island arc system, sedimentaryremnants of which are preserved in the South Mayo basin to thenorth (Fig. 1; e.g., Ryan and Dewey 2011). This suggestion solves along-standing problem in Connemara geology and allowed us torelate field observations and our analytical results to refine theproposed model (Fig. 10).

Island arc – continent collision at Connemara (Grampianorogeny)

The key elements of the arc–continent collision model pro-posed by Dewey (2005) are as follows: (i) northward-directed ob-duction and collision of an intraoceanic island-arc system (LoughNafooey arc and South Mayo forearc region, Dewey 2005) as thepassive continental Laurentian plate margin with its Dalradiansedimentary cover rocks entered an oceanward-facing subductionzone, followed by (ii) subduction polarity reversal, (iii) extensionalcollapse, roll back (cf. Clift et al. 2004), and (iv) formation of anactive Laurentian margin. Dewey and Ryan (2016)’s suggestionthat hot island-arc mantle of the overriding plate acted as a heatsource to calcalkaline plutonism and Barrovian high-temperaturemetamorphism in the lower plate, is a key to connect local- andregional-scale observables. Below we critically address the evolu-tion of Connemara in this context, referring to Fig. 10.

By 490 Ma, an intraoceanic island arc had formed (LoughNafooey arc; Dewey and Mange 1999) and no later than 478 Ma theLaurentian continental margin and its Dalradian strata enteredthe subduction zone. By 474 Ma, a major low-angle shear zoneformed at the plate interface (Fig. 10B), successively incorporatingDalradian cover strata as they detached from their basement. Thesequence deformed internally as it moved north- and upwards, de-fining an incipient orogenic wedge. Exhumation may have occurredby some combination of thrusting and erosion of the oceanic arc inthe hanging wall, which is consistent with thick deposits of volcanicarc debris in the South Mayo basin (e.g., Sheefrey Formation, Ryanand Dewey 2011). Based on geometric arguments illustrated inFig. 10B, we postulate that peak pressures should be higher in ex-posed southern Connemara rocks, although only limited evidencefor this has been found to date (cf. Leake and Tanner 1994).

As Dalradian rocks were thrust below the oceanic arc – mantle,a transient inverted thermal gradient established. The hot oceanicarc – mantle acted as a heat source, which resulted in formation ofcalcalkaline mafic and ultramafic intrusions and—within a fewmillion years—to Barrovian-type metamorphism (Dewey andRyan 2016). These conditions were established by 468 Ma (Fig. 8;Friedrich et al. 1999b). Thermal modeling of this scenario byReverdatto and Polyansky (2004) and Dewey and Ryan (2016) im-plies that a hot mantle heat source must have been adjacent to, orobliquely on top of the Connemara metamorphic and magmaticcomplex from a southerly direction (Figs. 10B and 10C). At 468–467 Ma, a lateral thermal gradient of �20–30 °C/km developedacross Connemara (Figs. 7, 8B, and 10C). This gradient is a keysupportive element of the model that a hot upper plate, whichwas continuously moving northwards, provided the heat for high-temperature metamorphism and anatexis in southern Conne-mara, while northern Connemara—being farther from the major

Fig. 10. Schematic profiles of the island–arc continent collision (Grampian orogeny) in Connemara at five time intervals. (A) Plate boundaryconfiguration across an intra-oceanic subduction zone prior to collision of the inferred Lough Nafooey arc with the Laurentian continent(Dewey 2005). (B–D) Reconstructed within the frame of the island arc – continent collision model (Dewey 2005; Dewey and Ryan 2016) andrefined by the constraints provided by our data (e.g., Figs. 6, 8, and 9). We assume that the subducting Laurentian lithosphere steepened as itentered the sub-asthenospheric mantle, which acted as a geometric quasi-pin point to further reconstruction. The Laurentian slab and itsextended margin (Dalradian) thereby moves under the oceanic island–arc lithosphere, equivalent to ophiolite obduction in the geologicalrecord as suggested by Dewey and Ryan (2016). (B) The sharp bend in the downgoing Laurentian plate facilitated detachment of crustal nappes(D2 and D3) and their northward emplacement over Laurentia, marking the onset of collision. (C) Continuing ocean-ward motion ofLaurentia, driven by asthenospheric flow, resulted in south-vergent lithospheric necking and formation of upward-propagating lithosphericmantle shear zones, in both plates, followed by (D) thrusting of Laurentian mantle–lithosphere over oceanic mantle–lithosphere, demarkingdeep-seated collision and formation of the Grampian orogenic wedge. Continuing plate convergence results in further southward thrusting ofLaurentia, slab detachment, upward propagation of mantle shear-zones, and final emplacement of the orogenic wedge over the formervolcanic island arc. As the bi-vergent wedge shortens at depth, its rocks move upwards and are subject to erosion, horizontal extension, andcollapse. (E) An Andean-type margin may have established in western Ireland (e.g., Dewey 2005) by ca. 450 Ma. By ca. 440 Ma, marinesedimentary rocks were deposited onto extended and exhumed mid-crustal rocks terminating the possibility of any protracted cooling atConnemara following the Grampian orogeny. Note: arrows indicate relative plate motion, schematically, not to scale.

1274 Can. J. Earth Sci. Vol. 53, 2016

Published by NRC Research Press

Can

. J. E

arth

Sci

. Dow

nloa

ded

from

ww

w.n

rcre

sear

chpr

ess.

com

by

51.1

71.9

6.78

on

04/2

8/17

For

pers

onal

use

onl

y.

Friedrich and Hodges 1275

Published by NRC Research Press

Can

. J. E

arth

Sci

. Dow

nloa

ded

from

ww

w.n

rcre

sear

chpr

ess.

com

by

51.1

71.9

6.78

on

04/2

8/17

For

pers

onal

use

onl

y.

heat source—cooled rapidly to below 300 °C. Figure 10C providesan internally consistent geometric solution for this scenario. Theelevated temperatures also facilitate localization of deformationand disruption of earlier D3 fabrics by formation of incipientsouth-vergent deformation within the steep belt of southern Con-nemara (Fig. 8B).

To meet the inferred PT conditions requires actively obductingupper plate no thicker than �20–25 km (Dewey and Ryan 2016).Therefore, the upper portion of the downgoing plate must be inlow-angle contact with the overriding arc lithosphere, which lim-its any asthenospheric mantle that may be trapped at the inter-face (Fig. 10B). As collision proceeds, the oceanic arc – lithospheredetaches from its sublithospheric mantle along a postulated intra-mantle shear zone. This zone occurs near the contact with themantle and is facilitated by a fairly sharp bend in the downgoingLaurentian plate, which could form by slab-steeping above a slowsub-asthenospheric mantle. This lithospheric bend is the geomet-ric basis for the upcoming collision and formation of the secondvergence direction, which enabled the reversal of subduction po-larity. A similar geometric configuration has been observed underTaiwan (Carena et al. 2013).

The switch from north- to south-vergent structures in Conne-mara is one of the most fundamental features upon which thearc–continent collision model is built (Dewey 2005). Integration ofstructural, metamorphic, and geochronological data bracket thetime of this polarity reversal to a short interval between 468 and467 Ma (Figs. 9 and 10B; Friedrich et al. 1999b). In Fig. 10C, weillustrated a geometrically permissible solution, in which litho-spheric necking in the downgoing plate occurs as a response tocontinuing plate convergence and incipient continent-ward sub-duction of the oceanic mantle – lithosphere.

Reversal of subduction polarity may be completed only once theformerly subducting plate tears away (Fig. 10D). This change invergence of compressional structures is driven by the deep-seatedcollision between subducted Laurentian mantle – lithospherewith oceanic mantle – lithosphere in a mobile asthenosphericenvironment.

Timing of extension and slab roll-backDewey (2005) and Ryan and Dewey (2011) argued that exten-

sional collapse of the orogen occurred at 466 Ma, based on thesudden appearance of staurolite-grade detritus in the sedimen-tary record of South Mayo. Because these authors cannot reconcilethis observation in any other way, they argue that extensionalfaulting is an effective way to unroof high-grade metamorphicrocks, suddenly, without having first eroded lower-grade rocks. Ifcorrect, the first appearance of staurolite would mark the onset ofextensional collapse, and in turn, slab roll-back as the underlyingcause of extension (upper portion of the Rosroe Formation;466 Ma; figs. 3 and 4C in Dewey 2005; fig. 13.2 in Ryan and Dewey2011; fig. 3 in Dewey et al. 2015; fig. 3 in Dewey and Ryan 2016). Thisscenario implies that extensional collapse and slab roll-backwould have occurred prior to formation of the south-vergent D4structures, such as the Mannin thrust, which we find difficult toreconcile with the kinematic requirements that the deeper por-tions of the orogenic wedge were under compression at that time(Fig. 10C).

The presently exposed mid-crustal section in Connemara ap-pears to have undergone the transition from shortening to exten-sion at the scale of rock samples no earlier than 463 Ma. In thefield, this transition may be marked by the post-D4 intrusion ofgranites and related evidence of fracturing and hydrothermalfluid circulation. Thus, extensional collapse of the orogenicwedge, and thus its cause, postdate formation of D4 south-vergentstructures. A lower bound to significant late Grampian extensionwould be provided by Category I cooling ages for southern Con-nemara, which may reflect extensional unroofing of the orogenuntil ca. 452 Ma (Fig. 9).

We therefore suggest that the sharp contact between ophioliticand staurolite-grade metamorphic detritus in the stratigraphicsection of South Mayo at 466 Ma marks the time at which ero-sional unroofing of the orogen exhumed the major thrust thatjuxtaposed the hot ophiolitic material and arc mantle againsthigh-grade metamorphic rocks of the lower plate. The hypothesisof Dewey and Ryan (2016) that hot oceanic mantle was thrust overthe Dalradian rocks and heated them from above is consistentwith this conclusion. After 463 Ma, slab roll-back may have en-hanced extension and morphologic decay of the orogenic lid, re-quiring only one major phase of extension at ca. 450 Ma. TheSilurian nonconformity may be a late surface expression of thisprocess.

Timing of post-Grampian processesOur geochronological results confirm that Grampian events at

Connemara took place progressively and within a few millionyears (Friedrich et al. 1999a, 1999b), while our thermochronologi-cal results imply that cooling following the Grampian orogenywas locally rapid and finished long before deposition of marineSilurian rocks, as originally proposed by Fitch et al. (1964), butopposed by Dewey (2005). We postulate that the Grampian“tail”—our Category 3 (Fig. 6)—is not related to slow cooling re-lated to denudation and erosion as proposed by Dewey (2005), andconclude that Silurian thermochronological dates bear no rele-vance to models of Grampian orogeny; they may relate to post-Grampian magmatism and hydrothermal circulation related toactive continental margin processes, which were established bythat time along the Laurentian margin (Fig. 10E).

ConclusionsThe spatial pattern of the 40Ar/39Ar data presented here sup-

ports the previously reported large spread in mineral dates fromthe Connemara region. However, consideration of our data in thecontext of other petrographic, geochronologic, and stratigraphicconstraints, suggests that some dates represent regional coolingsubsequent to metamorphism, others record episodes of dynamicrecrystallization, still others reflect alteration episodes, and a feware so contaminated by excess 40Ar as to be geologically meaning-less. The regional cooling ages display patterns that can be linkeddirectly to temporal and spatial variations in deformation, meta-morphism, and magmatism. Northern Connemara cooled below�300 °C by ca. 470 Ma, whereas southern Connemara was not atsuch low temperatures until ca. 450 Ma. This phenomenon prob-ably was related to large lateral thermal gradients in an evolvingarc–continent collisional setting. One important implication ofthe Connemara study is that regional trends in cooling ages mayreflect regional variations in the orogenic temperature structure(such as distance to the heat source) rather than regional varia-tions in the unroofing history, as is commonly assumed.

The significance of the geologic record to proper interpretationof mineral cooling ages cannot be overemphasized. If we knewless about the stratigraphic, intrusive, and structural history ofConnemara, it would be a simple matter to interpret the broadspread in mineral ages from this region in terms of protractedcooling over nearly one hundred million years. This realization isa reminder that geochronology is of limited value in regions thathave not been characterized adequately through careful fieldmapping and structural analysis.

AcknowledgementsThis paper is dedicated to Kevin Burke, John Dewey, and Samuel

Bowring. Financial support was provided by an NSF grant awardedto K.V.H. and by Samuel Bowring to A.M.F. Bill Oleszewski isthanked for help with the argon sample preparation and analysis.The authors thank reviewer Paul Ryan and editor Ali Polat forconstructive comments.

1276 Can. J. Earth Sci. Vol. 53, 2016

Published by NRC Research Press

Can

. J. E

arth

Sci

. Dow

nloa

ded

from

ww

w.n

rcre

sear

chpr

ess.

com

by

51.1

71.9

6.78

on

04/2

8/17

For

pers

onal

use

onl

y.

ReferencesArnaud, N.O., and Kelley, S.P. 1995. Evidence for excess argon during high-

pressure metamorphism in the Dora Maira massif (western Alps, Italy), usingan ultra-violet laser ablation microprobe 40Ar-39Ar technique. Contributionsto Mineralogy and Petrology, 121: 1–11. doi:10.1007/s004100050086.

Barber, J.P., and Yardley, B.W.D. 1985. Conditions of high-grade metamorphismin the Dalradian of Connemara, Ireland. Journal of the Geological Society,142: 87–96. doi:10.1144/gsjgs.142.1.0087.

Baxter, E.F. 2010. Diffusion of noble gases in minerals. Reviews in Mineralogyand Geochemistry, 72(1): 509–557. doi:10.2138/rmg.2010.72.11.

Boyle, A.P., and Dawes, I.P. 1991. Contrasted metamorphic and structural evolu-tions across a major ductile/brittle displacement zone in NW Connemara,western Ireland. Geologische Rundschau, 80: 459–480. doi:10.1007/BF01829377.

Bröcker, M., Baldwin, S., and Arkudas, R. 2013. The geological significance of40Ar/39Ar and Rb–Sr white mica ages from Syros and Sifnos, Greece: a recordof continuous (re) crystallization during exhumation? Journal of Metamor-phic Geology, 31(6): 629–646. doi:10.1111/jmg.12037.

Brown, D., Ryan, P.D., Afonso, J.C., Boutelier, D., Burg, J.P., Byrne, T., et al. 2011.Arc–continent collision: the making of an orogen. In Arc–Continent Colli-sion. Edited by D. Brown and P.D. Ryan. Frontiers in Earth Sciences. Springer-Verlag Berlin Heidelberg. pp. 477–493. doi:10.1007/978-3-540-88558-0_17.

Burke, K. 1957. An outline of the structure of the Galway granite. GeologicalMagazine, 94(6): 452–464. doi:10.1017/S001675680007014X.

Byrne, T., Chan, Y.C., Rau, R.J., Lu, C.Y., Lee, Y.H., and Wang, Y.J. 2011. Thearc–continent collision in Taiwan. In Arc–Continent Collision. Edited byD. Brown and P.D. Ryan. Frontiers in Earth Sciences. Springer-Verlag BerlinHeidelberg. pp. 213–245. doi:10.1007/978-3-540-88558-0_8.

Carena, S., Suppe, J., and Wu, Y.-M. 2013. Contrasting lithospheric-scale andcrustal-scale architecture and kinematics of the Taiwan arc–continent colli-sion. GSA Denver Annual Meeting, 27–30 October 2013, Paper #379–8.

Chew, D.M., and Strachan, R.A. 2013. The Laurentian Caledonides of Scotlandand Ireland. In New Perspectives on the Caledonides of Scandinavia andRelated Areas. Edited by F. Corfu, D. Gasser, and D.M. Chew. Geological Soci-ety, London, Special Publications, 390. doi:10.1144/SP390.16.

Cliff, R.A. 1985. Isotopic dating in metamorphic belts. Journal of the GeologicalSociety, 142: 97–110. doi:10.1144/gsjgs.142.1.0097.

Cliff, R.A., Yardley, B.W.D., and Bussy, F. 1993. U–Pb isotopic dating of fluidinfiltration and metasomatism during Dalradian regional metamorphism inConnemara, western Ireland. Journal of Metamorphic Geology, 11: 185–191.doi:10.1111/j.1525-1314.1993.tb00141.x.

Cliff, R.A., Yardley, B.W.D., and Bussy, F. 1996. U–Pb and Rb–Sr geochronology ofmagmatism and metamorphism in the Dalradian of Connemara, westernIreland. Journal of the Geological Society, 153: 109–120. doi:10.1144/gsjgs.153.1.0109.

Clift, P.D., Dewey, J.F., Draut, A.E., Chew, D.M., Mange, M., and Ryan, P.D. 2004.Rapid tectonic exhumation, detachment faulting and orogenic collapse inthe Caledonides of western Ireland. Tectonophysics, 384(1–4): 91–113. doi:10.1016/j.tecto.2004.03.009.

Dewey, J.F. 2005. Orogeny can be very short. Proceedings of the National Acad-emy of Sciences, 102(43): 15286–15293. doi:10.1073/pnas.0505516102.

Dewey, J.F., and Mange, M. 1999. Petrography of Ordovician and Silurian sedi-ments in the western Irish Caledonides: tracers of a short-lived Ordoviciancontinent–arc collision orogeny and the evolution of the Laurentian Appalachian–Caledonian margin. Geological Society, London, Special Publications, 164(1):55–107. doi:10.1144/GSL.SP.1999.164.01.05.

Dewey, J.F., and Ryan, P.D. 2016. Connemara: its position and role in the Gram-pian Orogeny. Canadian Journal of Earth Sciences, 53. This issue. doi:10.1139/cjes-2015-0125.

Dewey, J.F., McKerrow, W.S., and Moorbath, S. 1970. The relationship betweenisotopic ages, uplift and sedimentation during Ordovician times in westernIreland. Scottish Journal of Geology, 6(2): 133–145. doi:10.1144/sjg06020133.

Dewey, J.F., Dalziel, I.W., Reavy, R.J., and Strachan, R.A. 2015. The Neoprotero-zoic to Mid-Devonian evolution of Scotland: a review and unresolved issues.Scottish Journal of Geology, 51(1): 5–30. doi:10.1144/sjg2014-007.

Dodson, M.H. 1973. Closure temperature in cooling geochronological and petro-logical systems. Contributions to Mineralogy and Petrology, 40: 259–274.doi:10.1007/BF00373790.

Draut, A.E., and Clift, P.D. 2002. The origin and significance of the Delaney DomeFormation, Connemara, Ireland. Journal of the Geological Society, 159(1):95–103. doi:10.1144/0016-764901034.

Elias, E.M., Macintyre, R.M., and Leake, B.E. 1988. The cooling history of Conne-mara, western Ireland, from K–Ar and Rb/Sr age studies. Journal of the Geo-logical Society, 145: 649–660. doi:10.1144/gsjgs.145.4.0649.

Fitch, F.J., Miller, J.A., and Brown, P.E. 1964. Age of Caledonian orogeny andmetamorphism in Britain. Nature, 203: 275–278. doi:10.1038/203275a0.

Friedrich, A.M. 1998. 40Ar/39Ar and U–Pb geochronological constraints on thethermal and tectonic evolution of the Connemara Caledonides, western Ire-land. Ph.D. thesis, MIT, Cambridge, MA, 227 pp.

Friedrich, A.M., Bowring, S.A., Martin, M.W., and Hodges, K.V. 1999a. Short-livedcontinental magmatic arc at Connemara, western Irish Caledonides: Impli-cations for the age of the Grampian orogeny. Geology, 27(1): 27–31. doi:10.1130/0091-7613(1999)027<0027:SLCMAA>2.3.CO;2.

Friedrich, A.M., Hodges, K.V., Bowring, S.A., and Martin, M.W. 1999b. Geochrono-

logical constraints on the magmatic, metamorphic and thermal evolution ofthe Connemara Caledonides, western Ireland. Journal of the Geological So-ciety, 156: 1217–1230. doi:10.1144/gsjgs.156.6.1217.

Giletti, B.J. 1974. Studies in diffusion I: Argon in phlogopite mic. In GeochemicalTransport and Kinetics. Edited by A.W. Hofmann, B.J. Giletti, H.S. Yoder, andR.A. Yund. Washington, Carnegie Institute of Washington Publication,pp. 107–115.

Giletti, B.J., Moorbath, S., and Lambert, R.S.J. 1961. A geochronological study ofthe metamorphic complexes of the Scottish Highlands. Quarterly Journal ofthe Geological Society, 117(1–4): 233–264. doi:10.1144/gsjgs.117.1.0233.

Graham, J.R., Leake, B.E., and Ryan, P.D. 1989. The geology of South Mayo,western Ireland including 1:63 360 geological map with cross-sections. Uni-versity of Glasgow, pp. 1–75.

Grove, M., and Harrison, T.M. 1996. 40Ar* diffusion in Fe-rich biotite. AmericanMineralogist, 81, 940–951. doi:10.2138/am-1996-7-816.

Hames, W.E., and Bowring, S.A. 1994. An empirical evaluation of the argondiffusion geometry in muscovite. Earth and Planetary Science Letters, 124:161–169. doi:10.1016/0012-821X(94)00079-4.

Harrison, T.M., and McDougall, I. 1981. Excess 40Ar in metamorphic rocks fromBroken Hill, NSW. Implications for 40Ar/39Ar age spectra and the thermalhistory of the region. Earth and Planetary Science Letters, 55: 123–149. doi:10.1016/0012-821X(81)90092-3.

Harrison, T.M., Duncan, I., and McDougall, I. 1985. Diffusion of 40Ar in biotite:temperature, pressure, and compositional effects. Geochimica et CosmochimicaActa, 49: 2461–2468. doi:10.1016/0016-7037(85)90246-7.

Heaman, L., and Parrish, R. 1991. U–Pb geochronology of accessory minerals. InApplications of Radiogenic Isotope Systems to Problems in Geology. Edited byL. Heaman and J.N. Ludden. Toronto, Mineralogical Association of Canada,pp. 59–102.

Hodges, K.V. 1991. Pressure–temperature–time paths. Annual Review of Earthand Planetary Sciences, 19: 207–236. doi:10.1146/annurev.ea.19.050191.001231.

Hodges, K.V. 2014. Thermochronology in orogenic systems. In Treatise on Geo-chemistry, 2nd ed. Elsevier, pp. 281–308.

Hodges, K.V., and Bowring, S.A. 1995. 40Ar/39Ar thermochronology of isotopi-cally zoned micas: insights from the southwestern USA Proterozoic orogen.Geochimica et Cosmochimica Acta, 59: 3205–3220. doi:10.1016/0016-7037(95)00209-I.

Hodges, K.V., Harries, W.E., and Bowring, S.A. 1994. 40Ar/39Ar age gradients inmicas from a high-temperature – low-pressure metamorphic terrain: evi-dence for very slow cooling and implications for the interpretation of agespectra. Geology, 22: 55–58. doi:10.1130/0091-7613(1994)022<0055:AAAGIM>2.3.CO;2.

Jagger, M.D., Max, M.D., Aftalion, M., and Leake, B.E. 1988. U–Pb zircon ages ofbasic rocks and gneisses intruded into the Dalradian rocks of Cashel, Conne-mara, western Ireland. Journal of the Geological Society, 145: 645–648. doi:10.1144/gsjgs.145.4.0645.

Jenkin, G.R.T., Fallik, A.E., and Leake, B.E. 1992. A stable isotope study of retro-grade alteration in SW Connemara, Ireland. Contributions to Mineralogy andPetrology, 110: 269–288. doi:10.1007/BF00310743.

Kanaris-Sotiriou, R., and Angus, N.S. 1976. The Currywongaun–Doughruagh syn-tectonic intrusion, Connemara, Ireland. Journal of the Geological Society,132(5): 485–508. doi:10.1144/gsjgs.132.5.0485.

Leake, B.E. 1989. The metagabbros, orthogneisses and paragneisses of the Con-nemara complex, western Ireland. Journal of the Geological Society, 146:575–596. doi:10.1144/gsjgs.146.4.0575.

Leake, B.E., and Tanner, P.W.G. 1994. The geology of the Dalradian and associ-ated rocks of Connemara, western Ireland: a report to accompany the1:63 360 geological map and cross sections. Royal Academy, Dublin.

Leake, B.E., Tanner, P.W.G., Singh, D., and Halliday, A.N. 1983. Major southwardthrusting of the Dalradian rocks of Connemara, western Ireland. Nature,305: 210–213. doi:10.1038/305210a0.

Lee, J.K.W. 1995. Multipath diffusion in geochronology. Contributions to Miner-alogy and Petrology, 120: 60–82. doi:10.1007/BF00311008.

Leggo, P.J., Compston, W., and Leake, B.E. 1966. The geochronology of the Con-nemara granites and its bearing on the antiquity of the Dalradian series.Quarterly Journal of the Geological Society, 122: 91–116. doi:10.1144/gsjgs.122.1.0091.

Liu, T.K., Chen, Y.G., Chen, W.S., and Jiang, S.H. 2000. Rates of cooling anddenudation of the Early Penglai Orogeny, Taiwan, as assessed by fission-trackconstraints. Tectonophysics, 320(1): 69–82. doi:10.1016/S0040-1951(00)00028-7.

McDougall, I., and Harrison, T.M. 1988. Geochronology and Thermochronologyby the 40Ar/39Ar Method. New York, Oxford University Press, 212 p.

Mezger, K., Rawnsley, C.M., Bohlen, S.R., and Hanson, G.N. 1991. U–Pb garnet,sphene, monazite, and rutile ages: implications for the duration of high-grade metamorphism and cooling histories, Adirondack Mtns., New York.Journal of Geology, 99: 415–428. doi:10.1086/629503.

Miller, W.M., Fallick, A.E., Leake, B.E., Macintyre, R.M., and Jenkin, G.R.T. 1991.Fluid disturbed hornblende K–Ar ages from the Dalradian rocks of Conne-mara, western Ireland. Journal of the Geological Society, 148: 985–992. doi:10.1144/gsjgs.148.6.0985.

Morris, W.A., and Tanner, P.W.G. 1977. The use of paleomagnetic data to delin-eate the history of the development of the Connemara Antiform. CanadianJournal of Earth Sciences, 14: 2601–2613. doi:10.1139/e77-225.

Friedrich and Hodges 1277

Published by NRC Research Press

Can

. J. E

arth

Sci

. Dow

nloa

ded

from

ww

w.n

rcre

sear

chpr

ess.

com

by

51.1

71.9

6.78

on

04/2

8/17

For

pers

onal

use

onl

y.

Mulch, A., Cosca, M.A., Andresen, A., and Fiebig, J. 2005. Time scales of defor-mation and exhumation in extensional detachment systems determined byhigh-spatial resolution in situ UV-laser 40Ar/39Ar dating. Earth and PlanetaryScience Letters, 233(3–4): 375–390. doi:10.1016/j.epsl.2005.01.042.

Onstott, T.C., and Peacock, M.W. 1987. Argon retentivity of hornblendes: a fieldexperiment in a slowly cooled metamorphic terrane. Geochimica et Cos-mochimica Acta, 51: 2891–2903. doi:10.1016/0016-7037(87)90365-6.

Reverdatto, V.V., and Polyansky, O.P. 2004. Modelling of the thermal history ofmetamorphic zoning in the Connemara region (western Ireland). Tectono-physics, 379(1–4): 77–91. doi:10.1016/j.tecto.2003.10.005.

Robertson, D.J. 1988. Paleomagnetism of the Connemara gabbros, western Ire-land. Geophysical Journal of the Royal Astronomical Society, 94: 51–64.

Ryan, P., and Dewey, J.F. 2011. Arc–continent collision in the Ordovician ofwestern Ireland: stratigraphic, structural and metamorphic evolution. InArc–Continent Collision. Edited by D. Brown and P.D. Ryan. Springer. Fron-tiers in Earth Sciences. Springer-Verlag Berlin Heidelberg. pp. 373–401. doi:10.1007/978-3-540-88558-0_13.

Samson, S.D., and Alexander, E.C. 1987. Calibration of the interlaboratory 40Ar/39Ar dating standard, MMhb-1. Chemical Geology, 66: 27–34. doi:10.1016/0168-9622(87)90025-X.

Schoene, B. 2014. U–Th–Pb geochronology. In Treatise on Geochemistry, SecondEdition, Volume 4. Edited by H.D. Holland and K.K. Turekian. Elsevier Science,Oxford. pp. 341–378.

Soper, N.J., Ryan, P.D., and Dewey, J.F. 1999. Age of the Grampian orogeny inScotland and Ireland. Journal of the Geological Society, 156: 1231–1236. doi:10.1144/gsjgs.156.6.1231.

Tanner, P.W.G. 1990. Structural age of the Connemara gabbros, western Ireland.Journal of the Geological Society, 147: 599–602. doi:10.1144/gsjgs.147.4.0599.

Tanner, P.W.G., and Shackleton, R.M. 1979. Structure and stratigraphy of theDalradian rocks of the Bennabeola area, Connemara, Eire. In The Cale-donides of the British Isles – Reviewed. Edited by A.L. Harris, C.H. Holland, andB.E. Leake. Geological Society, London, Special Publications, 8: 243–256.

Tanner, P.W.G., Dempster, P.J., and Dickin, A.P. 1989. Time of docking of theConnemara terrane with the Delany Dome Formation, western Ireland. Jour-nal of the Geological Society, 146: 389–392. doi:10.1144/gsjgs.146.3.0389.

Tanner, P.W.G., Dempster, P.J., and Rogers, G. 1997. New constraints upon thestructural and isotopic age of the Oughterard Granite, and on the timing ofevents in the Dalradian rocks of Connemara, western Ireland. GeologicalJournal, 32: 247–263. doi:10.1002/(SICI)1099-1034(199709)32:3<247::AID-GJ744>3.0.CO;2-Y.

Treloar, P.J. 1985. Metamorphic conditions in central Connemara, Ireland. Jour-nal of the Geological Society, 142: 77–86. doi:10.1144/gsjgs.142.1.0077.

Tucker, R.D., and McKerrow, W.S. 1995. Early Paleozoic chronology: a review inlight of new U–Pb zircon ages from Newfoundland and Britain. CanadianJournal of Earth Sciences, 32: 368–379. doi:10.1139/e95-032.

Van Staal, C.R., Dewey, J.F., Mac Niocall, C., and McKerrow, W.S. 1998. TheCambrian–Silurian tectonic evolution of the northern Appalachians and Brit-ish Caledonides: history of a complex west, and south-west Pacific-type seg-ment of Iapetus. In Lyell: the Past is the Key to the Present. Edited byD.J. Blundell and A.C. Scott. Geological Society, London, Special Publications,143: 199–242. doi:10.1144/GSL.SP.1998.143.01.17.

Van Staal, C.R., Chew, D.M., Zagorevski, A., McNicoll, V., Hibbard, J., Skulski, T.,and Sylvester, P.J. 2013. Evidence of Late Ediacaran hyperextension of theLaurentian Iapetan margin in the Birchy complex, Baie Verte Peninsula,northwest Newfoundland: implications for the opening of Iapetus, forma-tion of peri-Laurentian microcontinents and Taconic–Grampian orogenesis.Geoscience Canada, 40(2): 94–117. doi:10.12789/geocanj.2013.40.006.

von Blanckenburg, F.V., and Villa, I.M. 1988. Argon retentivity and argon excessin amphiboles from the garbenschists of the western Tauern Window, east-ern Alps. Contributions In Mineralogy and Petrology, 100: 1–11. doi:10.1007/BF00399435.

Watson, E.B., and Baxter, E.F. 2007. Diffusion in solid-Earth systems. Earth andPlanetary Science Letters, 253(3–4): 307–327. doi:10.1016/j.epsl.2006.11.015.

Wellings, S.A. 1998. Timing of deformation associated with the syn-tectonicDawros–Currywongaun–Doughruagh complex, NW Connemara, westernIreland. Journal of the Geological Society, London, 155: 25–37. doi:10.1144/gsjgs.155.1.0025.

Willett, S.D., Fisher, D., Fuller, C., En-Chao, Y., and Chia-Yu, L. 2003. Erosion ratesand orogenic-wedge kinematics in Taiwan inferred from fission-track ther-mochronometry. Geology, 31(11): 945–948. doi:10.1130/G19702.1.

Williams, D.M., and Rice, A.H.N. 1989. Low-angle extensional faulting and theemplacement of the Connemara Dalradian, Ireland. Tectonics, 8: 417–428.doi:10.1029/TC008i002p00417.

Yardley, B. 1976. Deformation and metamorphism of Dalradian rocks and theevolution of the Connemara cordillera. Journal of the Geological Society,132: 521–542. doi:10.1144/gsjgs.132.5.0521.

Yardley, B., Barber, J.P., and Gray, J.R. 1987. The metamorphism of the Dalradianrocks of western Ireland and its relation to tectonic setting. PhilosophicalTransactions of the Royal Society, London, A, 321: 243–270. doi:10.1098/rsta.1987.0013.

Yardley, B., Bottrell, S.H., and Cliff, R.A. 1991. Evidence for a regional-scale fluidloss event during mid-crustal metamorphism. Nature, 349: 151–154. doi:10.1038/349151a0.

Appendix A. Analytical methodsWe obtained muscovite, biotite, and phlogopite separates by stan-

dard crushing, sieving, Wilfley, magnetic, and heavy liquid separa-tion. Two representative crystals of each sample were used todetermine the major element compositions using a JEOL 733 electronmicroprobe operating with a beam current of 10 nA and an acceler-ating voltage of 15 kV. The remaining separates were cleaned ultra-sonically with distilled water and wrapped in aluminium foil beforeirradiation at the McMaster University Reactor in Hamilton, Ontario.Fast neutron flux was monitored using MMhb hornblende (520.4 Ma,Samson and Alexander 1987). Corrections for interfering reactionswere accomplished by including K2SO4 and CaF2, and KCl salts in theirradiation packages. All samples were analyzed by a MAP 215-50 gassource mass spectrometer at MIT. Gas was extracted from the samplein three ways: furnace incremental heating, laser incremental heat-ing, and laser fusion. The first method involved experiments on5–50 mg samples in a double-vacuum resistance furnace for a periodof 10 min per increment. Laser incremental heating was performedwith a defocused, Coherent Ar-ion laser on 1–10 crystal aliquots. Pro-gressive heating was accomplished by firing the laser for 2 min in-tervals at successively higher power levels until each sample melted.Laser total fusion of 1–30 crystals was achieved by firing a defocusedlaser beam for 30 s at high power (typically >20 W). Representativesystems blanks for laser microanalysis for the M/e 40, 39, 38, 37,36 (moles) were 9 × 10−16, 2 × 10−17, 8 × 10−18, 1 × 10−17, and 8 × 10−18.Blanks varied in the furnace as a function of temperature and load-ing cycle, but were typically at least an order of magnitude higherthan those for laser analyses. All analyses were corrected for blanks,interfering isotopes, and mass discrimination.

For incremental heating analyses, plateaus were defined asthree or more consecutive release steps with a total of >50% of the39Ar released that are overlapping within 2� uncertainties irre-spective of the contribution of the J-value uncertainty. For sam-ples not yielding plateaus, total gas ages or 39Ar weighted meanages are reported. To determine the best age and assess the uncer-tainties of total fusion analyses, we used the Monte Carlo ap-proach described by Hodges and Bowring (1995).

Appendix B. Construction of schematic serial crosssections

The Connemara antiform plunges about 10° eastward over adistance of 80 km, exposing a roughly 10 km thick crustal section(Leake and Tanner 1994). This down-plunge view allows great in-sight into the three-dimensional structure of the Connemara an-tiform and the gabbro and gneiss complex, which facilitated asimple cross section construction. We started with the present-day configuration and added all metamorphic, paleomagnetitc,and geochronologic constraints. Next, we used published knowl-edge about the timing and sense of motion along the main struc-tures. Restoration of both older time slices, prior to 460 Ma,involved several assumptions. The Silurian unconformity is par-allel to the structurally highest unit, rather than cutting acrosspreviously folded layers. This implies that the unconformity wastilted together with the Dalradian block sometime after the Gram-pian orogeny. We assumed that the vertical offset on the Renvyle–Bofin Slide of northern Connemara was minor, because there isno significant break in metamorphic grade across the fault. Thegeometry of unfolding of the F4 Connemara antiform was basedon paleomagnetic data that required a 20° southward tilt of the F3fold axes prior to F4 folding (Robertson 1988; Morris and Tanner1977). Paleomagnetic constraints on the pre-F4 orientation of re-gional F3 fold axes and geometric considerations require restora-tion of the Mannin thrust late during the development of theConnemara antiform, a relationship shown schematically by asteep incipient Mannin thrust (Fig. 8). The present-day surfacetrace is shown on the cross sections as a reference line.

1278 Can. J. Earth Sci. Vol. 53, 2016

Published by NRC Research Press

Can

. J. E

arth

Sci

. Dow

nloa

ded

from

ww

w.n

rcre

sear

chpr

ess.

com

by

51.1

71.9

6.78

on

04/2

8/17

For

pers

onal

use

onl

y.