Southern xenoliths: First glimpse of Jurassic (ca. 160 Ma) crust

Robert J. Stern1*, Elizabeth Y. Anthony2, Minghua Ren2, Brian E. Lock3, Ian Norton4, Jun-Ichi Kimura5, Takashi Miyazaki5, Takeshi Hanyu5, Qing Chang5, and Yuka Hirahara5 1Geosciences Department, University of at Dallas, Richardson, Texas 75083-0688, USA 2Department of Geological Sciences, University of Texas at El Paso, El Paso, Texas 79968, USA 3Department of , Box 44530, University of Louisiana, Lafayette, Louisiana 70504, USA 4Institute for , Jackson School of Geosciences, University of Texas at Austin, Austin, Texas 78758, USA 5Institute for Research on Earth Evolution (IFREE), Japan Agency for Marine-Earth Science and Technology (JAMSTEC), Yokosuka 237-0061, Japan

ABSTRACT Salt that was deposited on top of the Late Juras- No direct information about the age and composition of rift-related igneous activity associ- sic crust, occasionally bring up samples of igne- ated with the Late Jurassic opening of the Gulf of Mexico exists because the igneous rocks are ous rocks (Lock and Duex, 1996; Ren et al., deeply buried beneath sediments. Three salt diapirs from southern Louisiana exhume samples 2009). The Five Islands of southern Louisiana of alkalic igneous rocks; these salt domes rise from the base of the sedimentary pile and overlie are part of seven uniformly spaced salt diapirs an isolated magnetic high, which may mark the position of an ancient volcano. Three samples that defi ne a linear northwest trend (Fig. 1). The from two domes were studied; they are altered but preserve relict igneous minerals including Five Islands trend overlies transitional crust strongly zoned clinopyroxene (diopside to Ti-augite) and Cr-rich spinel rimmed with titanite. thought to have formed during Gulf of Mexico 40Ar/39Ar ages of 158.6 ± 0.2 Ma and 160.1 ± 0.7 Ma for Ti-rich biotite and kaersutite from two opening (Dobson and Buffl er, 1997; Harry and different salt domes are interpreted to represent the time the igneous rock solidifi ed. Trace Londono, 2004). The diapirs containing mafi c element compositions are strongly enriched in incompatible trace elements, indicating that the igneous rocks (Jefferson, Avery, and Weeks) igneous rocks are low-degree melts of metasomatized upper mantle. Isotopic compositions of rise over a magnetic high (Fig. 1), as might be Nd and Hf indicate derivation from depleted mantle. This information supports the idea that expected from a signifi cant volume of buried crust beneath southern Louisiana formed as a magma-starved rifted margin on the northern mafi c igneous rocks. Salt mine exposures reveal fl ank of the Gulf of Mexico ca. 160 Ma. These results also confi rm that some magnetic highs highly deformed bedding defi ned by interlay- mark accumulations of mafi c igneous rocks buried beneath thick sediments around the Gulf ered halite and anhydrite with inclusions of of Mexico margins. Oligocene sandstone, shale, and igneous rocks, as well as pockets of water, oil, and gas (Lock INTRODUCTION lack of correlatable, spreading-related magnetic and Duex, 1996). Structure in the salt domes The Gulf of Mexico opened as the western- anomalies. Indirect evidence indicates opening is essentially vertical, with multiphase isocli- most arm of Tethys, related to breakup of Pangea between ca. 165 and 139 Ma (Late Jurassic) and nal folding. The original stratigraphic position and synchronous with opening of the Central that the transitional crust ranges from a narrow, of the igneous samples is not known, but must Atlantic (Pindell, 1985). In spite of this general magma-rich volcanic rifted margin beneath the have been immediately beneath, interbedded understanding about when and how it opened, Texas coast to a broader, magma-poor rifted with, and/or intruded into or above the salt the Gulf of Mexico is a rare example where the passive margin beneath Louisiana (Mickus et (Fig. 2). The age of the Louann Salt is bracketed origin of a sizable oceanic basin at low latitudes al., 2009; Stern and Dickinson, 2010). We have on stratigraphic grounds (Salvador, 1991) as is unclear, due largely to thick blanketing sedi- no direct way to sample and study this Jurassic post–Early Jurassic to pre–late Oxfordian, prob- ments (to 16 km; Muehlberger, 1992) and the seafl oor, but salt diapirs, sourced from Louann ably mostly Callovian (165–161 Ma; Walker and Geissman, 2009). We report here mineral chemical, whole-rock chemical, Nd and Hf iso- Figure 1. Magnetic anom- topic compositions, and radiometric ages for Mississippi aly map of Louisiana Ala (United States) and en- Avery to Weeks ~12 km Louisiana virons. Red is magnetic 32°N high, blue is magnetic Texas low. Seven black dots are Cenozoic sediments Av salt domes of Five Islands Fl trend. Note that three salt J W domes containing mafi c 30°N xenoliths (J—Jefferson, Late Jurassic and Cretaceous 2. Alkalic mafics Av—Avery, W—Weeks) sediments intruded into salt? 3 . Lava lt? are above magnetic high s plucked above sa interpreted to mark pres- Louann Salt (176–158 Ma) ence of buried mafi c lavas. Magnetic anoma- 1. Alkalic mafics plucked from beneath salt? 28°N Gulf of Mexico lies are from Maus et al. (2007). Fl—, Ala— Figure 2. Possible relationship between salt Alabama. 94°W 92° 90° 88° and alkalic mafi c rocks. Mafi c rocks could be plucked by rising salt from underneath or above, or could have intruded. Note distance *E-mail: [email protected]. between two salt domes containing xenoliths.

© 2011 Geological Society of America. For permission to copy, contact Copyright Permissions, GSA, or [email protected]. GEOLOGY,Geology, April April 2011; 2011 v. 39; no. 4; p. 315–318; doi: 10.1130/G31635.1; 4 fi gures; 1 table; Data Repository item 2011108. 315 three samples of igneous rocks entrained in two (Fig. DR3). Diopside also exhibits core-over- 1 of these salt bodies. A third salt dome, Jeffer- growth textures with two distinct core composi- A 2EO kaersutitic amphibole son Island (Fig. 1), also contains altered mafi c tions: one with a high Cr, Si and a second with K/Ca I xenoliths (Balk, 1953), but samples of this were low Cr, Si (Fig. DR1). Rims for both core types 200 0.1 not available for study. This is the fi rst time that exhibit titanopyroxene (MgSi2 = TiAl2) and Ca- such information has been presented for igneous tschermaks (MgSi = AlviAliv) substitution and 180 Q rocks that formed when the Gulf of Mexico is are identical to matrix diopside, indicating that 160.1 ± 0.7 Ma (MSWD = 2.37) thought to have opened, in Jurassic time. We use the rim compositions were in equilibrium with 160 N P this information to further our understanding of the fi nal melt. The diopside cores and chromite LM O

Gulf of Mexico formation. are probably xenocrysts and have compositions Apparent age (Ma) 140 similar to minerals in mantle rocks, e.g., abys- Integrated Age = 167.1 ± 0.8 Ma RESULTS sal peridotites from the Vulcan Fracture Zone 120 Methods are summarized in the GSA Data (Dick, 1989) or Samoan xenoliths (Wright, 100 Repository1 (Ar results and chemical and iso- 1987; Hauri and Hart, 1994). Kaersutitic amphi-

10 K/Ca topic analytical procedures). Three samples of bole occurs as small grains in W26 and as large B W26 biotitte porphyritic igneous rocks from two salt domes phenocrysts in 2EO (Fig. DR4). Similar min- 200 1 were studied. Sample 2EO is from the 1000′ eral associations have been reported for alkalic (~304 m) level of the Weeks Island dome, and magmas from the Jasper (Gee et al., 1991) and 180 W26 is from the 1300′ (~395 m) level and W25 Line Island (Natland, 1976). Ti- 158.6 ± 0.2 Ma (MSWD = 1.49) is from the 1600′ (~486 m) level of the Avery rich (~5%–7% TiO ) biotite is intergrown with 160 2 F K C E G H I J Island dome. The samples are altered, as is magnetite and quartz. Magnetite has substantial D

obvious from petrographic examination, which jacobsite (Mn) and ulvospinel (Ti) components, Apparent age (Ma) 140 B reveals that primary igneous minerals such as typically (Mn0.5Fe1.1)(Fe0.9Ti0.5)O4. Relict igne- Integrated Age = 158.1 ± 0.3 Ma clinopyroxene are replaced by secondary quartz, ous feldspar was not found. This mineral assem- 120 vermiculite, calcite, hematite, and K-feldspar blage, i.e., mantle-like diopside and Cr-spinel 0 20 40 60 80 100 39 (Figs. DR2, DR8, and DR9 in the Data Reposi- with Ti-rich rims and Ti-rich hydrous phases of Cumulative % Ar Released tory). The K-feldspar assemblage is similar to biotite and amphibole, indicates involvement Figure 3. 40Ar/39Ar age and K/Ca spectra. A: Deep Sea Drilling Project Site 453 samples of moderately depleted mantle, overprinted by Weeks Island kaersutitic amphibole (sample described by Natland (1982). Intergrown with water-rich alkaline melts, probably low-degree 2EO). B: Avery Island biotite (sample W26). calcite is mcgillite, a manganous hydroxychol- partial melts. These melts were strongly alka- MSWD—mean square of weighted deviates. orosilicate (Stevenson et al., 1984). Because a line, most likely undersaturated basanite or oliv- signifi cant component of this alteration involves ine nephelinite (Anthony et al., 1989; Panina carbonate and water-rich vermiculite, a simple and Usoltseva, 2008). strength elements (HFSE, e.g., Ti, Zr, Hf, Nb, way to quantify alteration is with measured loss Biotite (W26) and kaersutitic amphibole Y; Fig. 4A). This interpretation is confi rmed on ignition (Table DR1 in the Data Repository) (2EO) with primary, igneous morphologies by elevated abundances of immobile incom- for whole-rock samples. By this criterion, and were dated at the New Mexico Institute of Min- patible trace elements (Fig. 4A), the most consistent with assessment based on thin-sec- ing and Technology using 40Ar/39Ar techniques incompatible of which are one to two orders tion examination, W26 is the least altered, fol- (Fig. 3; for Ar results, see the Data Repository). of magnitude enriched relative to normal mid- lowed by 2EO. W25 is the most altered. Both yield well-defi ned Ar plateau ages: 158.6 oceanic ridge (N-MORB). This implies The rocks contain fresh igneous minerals of ± 0.2 Ma for W26 biotite and 160.1 ± 0.7 Ma for that the alkalic mafi c magma was derived from diopside, Mg-Al chromite, titanite, kaersutite, 2EO kaersutite. These minerals are inferred to extremely low degree of partial melts (~2%– and Ti-rich biotite; these are set in a more-altered be original igneous phases, so we interpret these 3%) in contrast to MORB melts produced by matrix comprising diopside, titanite, biotite, ages as approximating when the magma cooled. ~10% melting (Fig. 4A). The mantle source apatite, and vermiculite (representative mineral These are the fi rst radiometric dates for igneous was more enriched than expected for MORB- compositions and textures are in Table DR2, and rocks that can be directly related to opening of type asthenosphere considering extremely Figs. DR1, DR3, DR4, and DR5). Primary min- the Gulf of Mexico in Jurassic time. These ages high highly incompatible elements. Positive eral assemblages reveal signifi cant disequilibria are older than previously reported ca. 146 Ma anomalies for Zr and Hf suggest that melting between cores with compositions characteristic 40Ar/39Ar dates from biotite diorite xenoliths of amphibole in the source was involved. of depleted mantle (diopside, Cr-rich spinel) found in salt diapirs in the La Popa Basin, Nd and Hf are much less mobile than Sr and and rims that are strongly enriched in Ti (tita- northeastern Mexico, interpreted as metamor- Pb during alteration, so their isotopic composi- naugite, titanite), indicating that early refractory phic ages by Garrison and McMillan (1999). tions are most likely to refl ect those of the unal- minerals reacted with an alkaline mafi c melt. The ages reported here are, however, consistent tered magmas and mantle source. This inference Spinels have Cr# (100 Cr/Cr + Al) from 36 to 42 with zircon dates obtained from laser ablation− is supported by the fact that Nd and Hf isoto- (Table DR2) and are typically rimmed by titanite inductively coupled plasma−mass spectrometry pic compositions are indistinguishable for the from La Popa xenoliths (J. Amato, 2010, per- three samples (Table 1), in spite of the fact that 1GSA Data Repository item 2011108, Fig- sonal commun.), consistent with the hypothesis these show different extents of alteration. These ures DR1−DR9, Ar results, chemical and isoto- that rifting related to the opening of the Gulf of results indicate derivation of the magma from pic analytical procedures, Table DR1 (whole-rock Mexico also affected the interior of northeastern mantle with a long-term depletion history (high chemical data), and Table DR2 (representative Mexico (Stern and Dickinson, 2010). Lu/Hf and Sm/Nd). The mantle source was not mineral compositions), is available online at www .geosociety.org/pubs/ft2011.htm, or on request from Alteration disturbed primary igneous com- as depleted as MORB-type mantle, but similar to [email protected] or Documents Secretary, positions, especially Si, alkali metals, alka- the depleted mantle source of Hawaiian ε ε GSA, P.O. Box 9140, Boulder, CO 80301, USA. line earths, Pb and H2O, but not high fi eld ( Hf ~+9, Nd ~+7 at 160 Ma; Table 1; Fig. 4B).

316 GEOLOGY, April 2011 A istics of depleted mantle (Cr-rich spinel and 100 W26 diopsidic clinopyroxene) and rims showing 2EO Figure 4. Plots of geochemi- strong alkaline affi nities (titanite and Ti-augite; cal and Nd-Hf isotopic data. 40 W25 see Figs. DR1 and DR3). Ti-rich amphibole and 30 A: Relatively immobile trace biotite are related to the late alkaline overprint, element abundances nor- 20 malized to normal mid-oce- and demonstrate the hydrous nature of the infi l- anic ridge basalt (N-MORB; trating alkaline melt. These petrographic fea- 10 Sun and McDonough, 1989). tures can be explained if depleted mantle was Elevated abundances of infi ltrated by a volatile-rich alkaline metaso- high fi eld strength elements 4 matic melt, perhaps accompanying early stages and rare earth elements are 3 of extension and mantle upwelling as the Gulf

Sample/N-MORB consistent with original lam- 2 prophyric magma. B: Hf-Nd of Mexico began to open. Because the meta- isotope diagram comparing somatic agent transported Ti, it must have car- 1 initial isotopic composi- ried HFSEs as well as large ion lithophile ele- More Incompatibility Less tions of Nd and Hf in 160 Ma alkalic mafi cs with those ments. Thus, the isotopic compositions of Nd U Ta Ce Nd Zr Eu Gd Dy Ho Tm Lu of present mantle sources and Hf are almost certainly dominated by the Th Nb La P Sm Hf Ti Tb Y Er Yb (EMI, EMII [enriched man- metasomatizing melt. These isotopic composi- tle types], DMM [depleted tions indicate derivation from a mantle source 28 Atlantic MORB mantle MORB], HIMU [high with a time-integrated history of light rare earth 24 B Iceland 238U/204Pb ratio]). Also shown DDMMMM element and Hf/Lu depletion, with no evidence Indian MORB is Indian-Pacifi c mantle do- 20 main boundary in south- for participation of old continental lithosphere. Indian 16 Hawaii western Pacifi c. X-Y axes Rarotonga Pacific Instead, these samples have an oceanic isotopic are in ε units to allow us to 12 character, as expected for the igneous rocks of Society correlate present-day man- EEMIIMII an embryonic oceanic basin. Hf 8 tle source and salt dome ε Pitcairn Xenoliths samples. References for The correspondence of salt diapirs containing 4 mantle sources and Indian- alkalic mafi c xenoliths and a magnetic high in HHIMUIMU Rurutu 0 Pacifi c boundary are from coastal Louisiana supports the hypothesis that Tubuai -4 Salters and White (1998) and magnetic fabrics along the northern Gulf of EEMIMI St. Helena Pearce et al. (2007). -8 Mexico refl ect rift-related crustal structure, with Kerguelen -12 magnetic highs marking accumulations of mafi c -8 -6 -4 -2 0 2 4 6 8 10 12 14 igneous rocks and magnetic lows corresponding εNd to intervals of stretched continental crust (e.g., Mickus et al., 2009). This study is, to our knowl- edge, the fi rst time that magnetic highs in the northwestern Gulf of Mexico have been directly TABLE 1. Nd AND Hf ISOTOPIC RESULTS correlated with mafi c igneous rocks in the crust. Sample 143Nd / 144Nd 147Sm / 144Nd ε (160)* 176Hf / 177Hf 176Lu / 177Hf ε (160)† Nd Hf This is consistent with geophysical cross sec- ±2σ ±2σ tions depicting the Louisiana margin as lack- W26 0.512941 ± 7 0.1409 7.05 0.282944 ± 6 0.00377 9.20 ing extensive magmatism (Harry and Londono, W25 0.512944 ± 8 0.1511 6.90 0.282960 ± 5 0.00354 9.81 2004). Extrapolating from our observations, we W25(r) 0.512943 ± 6 0.1522 6.86 0.282948 ± 6 0.00361 9.39 2EO 0.512940 ± 5 0.1489 6.87 0.282950 ± 7 0.00352 9.45 predict that if and when mafi c Jurassic xenoliths JB-2 0.513079 ± 6 0.283241 ± 7 are recovered from salt domes along the Texas JNdi-1 0.512091 ± 5§ margin, they will be higher degree tholeiitic JMC-475 0.282141 ± 11** basalts, inasmuch as these overlie a much more Note: (r) indicates duplicate dissolution and analysis. continuous and broad magnetic high (Mickus et *Relative to CHUR (chondritic uniform reservoir) 143Nd/144Nd = 0.512638. al., 2009). †Relative to CHUR 176Hf/177Hf = 0.282772 (Blichert-Toft and Albaréde, 1997). §Mean of 3 analyses. Deposition of the Louann Salt marked a criti- **Mean of 7 analyses used for normalization to 176Hf/177Hf = 0.282160. cal time between rifting and seafl oor spreading in the nascent Gulf of Mexico. Salt was depos- ited on transitional crust on either side of the Gulf of Mexico as the gulf began to open. Seawater The isotopic and trace element data consid- DISCUSSION was able to fl ow into the basin, but this narrow ered together indicate that the mantle source Three points are discussed in this section: (1) connection was repeatedly closed, allowing sea- was recently enriched as a result of metaso- the signifi cance of ca. 160 Ma low-degree melts water in the basin to evaporate. Untold cycles matism; this enrichment may have been asso- of metasomatized mantle beneath rifted crust of of seawater fl ooding and evaporation occurred ciated with Gulf of Mexico rifting. Alkalic the Louisiana margin; (2) the implications of salt before the basin widened suffi ciently that com- mafi c composition is typical of rift-initiation dome mafi c xenoliths for understanding regional munication with the world ocean was perma- magmas from strongly metasomatized source magnetic fabric; and (3) the signifi cance of xeno- nently established. Salt deposition thus ended regions (Maria and Luhr, 2008; Elkins-Tanton lith ages for understanding the age of the Louann about the time that seafl oor spreading began to et al., 2007), a conclusion consistent with Salt and when Gulf of Mexico rifting occurred. form true oceanic crust, or shortly thereafter. petrographic evidence that the salt dome xeno- Relict igneous minerals in the xenoliths are The age of salt deposition thus constrains when liths formed from volatile-rich melts. strongly zoned, with cores showing character- Gulf of Mexico rifting occurred, but has been

GEOLOGY, April 2011 317 diffi cult to determine. Salvador (1987) correlated Siberian fl ood basalts: Results from experi- in Hussong, D.M., et al., Initial reports of the the Louann Salt with the Callovian Huehue- mental petrology: Contributions to Mineral- Deep Sea Drilling Project, Volume 60: Wash- tepec Formation evaporite sequence in Mexico ogy and Petrology, v. 153, p. 191–209, doi: ington, D.C., U.S. Government Printing Offi ce, 10.1007/s00410-006-0140-1. p. 579–599. and other stratigraphic relationships to bracket Garrison, J.M., and McMillan, N.J., 1999, Evidence Panina, L.I., and Usoltseva, L.M., 2008, Alkaline-ul- deposition as post–Early Jurassic to pre–late for Jurassic continental-rift magmatism in trabasic mantle-derived magmas, their sources, Oxfordian (176 Ma to 158 Ma). Allowing for NE Mexico: Allogenic metaigneous blocks in and crystallization features: Data of melt inclu- the imprecision of the stratigraphic data, the El Papalote evaporite diapir, La Popa Basin, sion studies: Lithos, v. 103, p. 431–444, doi: Nuevo Leon, Mexico, in Bartolini, C., et al., 10.1016/j.lithos.2007.10.009. ca. 160 Ma igneous activity represented by the eds., Mesozoic sedimentary and tectonic his- Pearce, J.A., Kempton, P.D., and Gill, J.B., 2007, xenoliths is essentially contemporaneous with tory of north-central Mexico: Geological Soci- Nd-Hf evidence for the origin and distribution the end of salt deposition. The radiometric ages ety of America Special Paper 340, p. 319–332, of mantle domains in the S.W. Pacifi c: Earth reported here thus constitute an important con- doi: 10.1130/0-8137-2340-X.319. and Planetary Science Letters, v. 260, p. 98– straint, that rift-related mafi c igneous activity Gee, J., Staudigal, H., and Natland, J.H., 1991, Geol- 114, doi: 10.1016/j.epsl.2007.05.023. ogy and petrology of Jasper : Journal Pindell, J.L., 1985, Alleghanian reconstruction and occurred ca. 160 Ma, about the same time that of Geophysical Research, v. 96, p. 4083–4106, the subsequent evolution of the Gulf of Mexico, the Louann Salt deposition ended. doi: 10.1029/90JB02364. Bahamas, and Proto-Caribbean Sea: Tectonics, These results fi rmly support the interpreta- Harry, D.L., and Londono, J., 2004, Structure and evo- v. 4, p. 1–39, doi: 10.1029/TC004i001p00001. tion that the Gulf of Mexico opened in Late lution of the central Gulf of Mexico continen- Ren, M., Stern, R., Lock, B., Griffi n, R., Anthony, tal margin and coastal plain, southeast United E., and Norton, I., 2009, Origin of igneous rock Jurassic time and encourage geoscientists to States: Geological Society of America Bulletin, fragments from South Louisiana salt domes: search for more evidence of early igneous activ- v. 116, p. 188–199, doi: 10.1130/B25237.1. Gulf Coast Association of Geological Societies ity associated with its opening in the many salt Hauri, E.H., and Hart, S.R., 1994, Constraints on Transactions, v. 59, p. 641–651. diapirs around the fl anks of the Gulf of Mexico. melt migration from mantle plume: A trace Salters, V.J.M., and White, W.M., 1998, Hf Isotope element study of peridotite xenoliths from constraints on mantle evolution: Chemical Ge- Savai’I, western Samoa: Journal of Geophysi- ology, v. 145, p. 447–460, doi: 10.1016/S0009 ACKNOWLEDGMENTS cal Research, v. 99, p. 24,201–24,321, doi: -2541(97)00154-X. This work benefi tted from conversations with Tim 10.1029/94JB01553. Salvador, A., 1987, Late Triassic–Jurassic paleoge- Lawton, Kate Giles, and Jeff Amato, and thought- Lock, B.E., and Duex, T.W., 1996, Xenolithic inclu- ography and origin of Gulf of Mexico Basin: ful comments by Dennis Harry, Jim Natland, and an sions within the salt at Weeks Island, Louisi- American Association of Petroleum Geologists anonymous reviewer. The research was supported by ana, and their signifi cance: Gulf Coast Asso- Bulletin, v. 71, p. 419–451. the Texas Advanced Research Program (grant 003661- ciation of Geological Societies Transactions, Salvador, A., 1991, Triassic–Jurassic, in Salvador, 0003-2006) to Stern and Anthony. This is University v. 46, p. 229–234, doi: 10.1306/2DC40B23 A., ed., The Gulf of Mexico Basin: Boulder, of Texas at Dallas Geosciences contribution 1210. -0E47-11D7-8643000102C1865D. Colorado, Geological Society of America, Ge- Maria, A.H., and Luhr, J.F., 2008, Lamprophyres, ba- ology of North America, v. J., p. 131–180. REFERENCES CITED sanites, and basalts of the Western Mexican Vol- Stern, R.J., and Dickinson, W.R., 2010, Gulf of Mex- Anthony, E.Y., Segalstad, T.V., and Neumann, E.R., canic Belt: Volatile contents and a vein-wallrock ico is a Jurassic backarc basin: Geosphere, v. 6, 1989, An unusual mantle source region for melting relationship: Journal of Petrology, v. 49, p. 739–754, doi: 10.1130/GES00585.1. nephelinites from the Oslo Rift, Norway: Geo- p. 2123–2156, doi: 10.1093/petrology/egn060. Stevenson, J.S., Stevenson, L.S., and Donnay, J.D.H., chimica et Cosmochimica Acta, v. 53, p. 1067– Maus, S., and 22 others, 2007, EMAG2: A 2-arc min 1984, A postscript to the mcGillite story: Cana- 1076, doi: 10.1016/0016-7037(89)90211-1. resolution Earth Magnetic Anomaly Grid com- dian Mineralogist, v. 22, p. 265–268. Balk, R., 1953, Jefferson Island salt dome, Louisi- piled from satellite, airborne, and marine mag- Sun, S.-S., and McDonough, W.F., 1989, Chemical ana: American Association of Petroleum Ge- netic measurements: Geophys- and isotopic systematics of oceanic basalts: Im- ologists Bulletin, v. 37, p. 2455–2474. ics Geosystems, v. 8, Q08005, doi: 10.1029/ plications for mantle composition and processes, Blichert-Toft, J., and Albaréde, F., 1997, The Lu-Hf 2006GC001521. in Saunders, A.D., and Norry, M.J., eds., Mag- isotope geochemistry of chondrites and the Mickus, K., Stern, R.J., Keller, G.R., and Anthony, matism in the ocean basins: Geological Society evolution of the mantle-crust system: Earth and E.Y., 2009, Potential fi eld evidence for a vol- of London Special Publication 42, p. 313–345, Planetary Science Letters, v. 148, p. 243–258, canic rifted margin along the Texas Gulf Coast: doi: 10.1144/GSL.SP.1989.042.01.19. doi: 10.1016/S0012-821X(97)00040-X. Geology, v. 37, p. 387–390, doi: 10.1130/ Walker, J.D., and Geissman, J.W., compilers, 2009, Dick, H.J.B., 1989, Abyssal peridotites, very slow G25465A.1. : Geological Society of spreading ridges and ocean ridge magmatism, Muehlberger, W.R., 1992, Tectonic map of North America, doi: 10.1130/2009.CTS004R2C. in Saunders, A.D., and Norry, M.J., eds., Mag- America, Southeast and Southwest sheets: Wright, E., 1987, Mineralogical studies of Samoan matism in the ocean basins: Geological Society Tulsa, Oklahoma, American Association of Pe- ultramafi c xenoliths: Implications for upper of London Special Publication 42, p. 71–105, troleum Geologists, scale 1:5,000,000. mantle processes, in Keating, B., et al., eds., doi: 10.1144/GSL.SP.1989.042.01.06. Natland, J.H., 1976, Petrology of volcanic rocks Seamounts, islands, and atolls: American Geo- Dobson, L.M., and Buffl er, R.T., 1997, Seismic stra- dredged from seamounts in the Line Islands, in physical Union Geophysical Monograph 43, tigraphy and geologic history of Jurassic rocks, Heirtzler, J., et al., Initial reports of the Deep p. 221–234. northeastern Gulf of Mexico: American Asso- Sea Drilling Project, Volume 33: Washing- ciation of Petroleum Geologists Bulletin, v. 81, ton, D.C., U.S. Government Printing Offi ce, Manuscript received 31 July 2010 p. 100–120. p. 759–777. Revised manuscript received 28 October 2010 Elkins-Tanton, L.T., Draper, D.S., Agee, C.B., Jew- Natland, J.H., 1982, Petrography and mineral com- Manuscript accepted 7 November 2010 ell, J., Thorpe, A., and Hess, P.C., 2007, The positions of gabbros recovered in DSDP Hole last lavas erupted during the main phase of the 452 on the western side of the Mariana Trough, Printed in USA

318 GEOLOGY, April 2011