kilometers to the east, must have risen at least 2000 and2100 meters since their formation, if they were formedentirely beneath ice level (as evidenced by the palagonitebreccias). Mount Siple, 400 kilometers northwest ofToney Mountain, may have risen as much as 3000meters. These displacements appear to have been differ-ential, and related to movements along late Cenozoicblock faults (LeMasurier, 1972a, 1972b), rather than toglacio-isostatic rebound.

Hollin (1962, 1968) considered the maximum in-creases in Quaternary ice level that may be used as asecond limiting case. In his 1962 paper, Hollin pre-sented a profile suggesting that the ice level may havebeen about 1000 meters higher than today at MountMurphy (located very near the present grounding line),and roughly 250 meters higher than today at MountTakahe (about 175 kilometers inland), during the Würmmaximum glaciation. In Hollin's 1968 paper, however,it was noted that these increases in ice level probablywere overestimated because the effects of isostatic load-ing (which would depress the bed of the glacier) wereunderestimated. Thus, if an ice level increase of 250meters is considered an overestimate at Mount Takahe,it seems reasonable to estimate that approximately 2000meters of uplift occurred since the formation of MountTakahe. It therefore appears that Mount Takahe hasbeen rising at an average rate of about 1 centimeter peryear (or 36 feet per 1000 years), if one assumes essen-tially a maximum age for the volcano (200,000 years).This rate favorably compares with uplift rates in activeorogenic belts (Schumm, 1963). Similar rates may havecharacterized the uplift of Mount Murphy and MountSiple, because glacial erosion in Marie Byrd Land appearsto depend on the exposure of rock above ice level, andthe glacial dissection of these two mountains is muchmore advanced than that of Mount Takahe (Andrewsand LeMasurier, 1973).

Marie Byrd Land appears to be part of a large tec-tonic province characterized by late Cenozoic block fault-ing and basalt—trachyte volcanism that extends fromthe Ross Sea region, through Marie Byrd Land to theAntarctic Peninsula. Webb (1972) reviews evidence forblock faulting in the McMurdo Sound region and pre-sents evidence for a minimum of 262 meters of upliftin Wright Valley within the last 3.4 million years.Evidence from Mount Takahe suggests that tectonicactivity in this province may be more recent thanformerly thought and, in some places, may be continu-ing today. If this is true, then the relative aseismicityof Antarctica is more puzzling than ever. The explana-tion that Antarctica is aseismic and tectonically dormantbecause it lies entirely within a lithospheric plate is notvery satisfying from a geologic point of view. This workis supported by National Science Foundation grant GV-

2 5 328.

References

Andrew's, J . T., and W. E. LeMasurier. 1973. Rates of Quarter-nary erosion and corrie formation, Marie Byrd Land, Ant-arctica. Geology, 1: 75-80.

Hollin, J . T. 1962. On the glacial history of Antarctica.Journal of Glaciology, 4: 173-195.

Hollin, J . T. 1968. The antarctic ice sheet and the Quaternaryhistory of Antarctica. In: Pilaeoecology of Africa and Ant-arctica (E. M. van Zinderen Bakker, ed.). Capetown, A. A.Balkema, 5: 109-138.

LeMasurier, W. E. 1972a. Volcanic record of Cenozoic glacialhistory of Marie Byrd Land. In: Antarctic Geology andGeophysics (R. J . Adie, ed.). Oslo, Universitetsforlaget.251-260.

LeMasurier, W. E. 1972b. Volcanic record of antarctic glacialhistory: implications with regard to Cenozoic sea levels. In:Polar Geomorphology (R. J . Price and D. E. Sugden, com-pilers). Institute of British Geographers. Special Publication,4:59-74

LeMasurier, W. E. 1972c. Marie Byrd Land Quaternary vol-canism: Byrd ice core correlations and possible climatic in-fluences. Antarctic Journal of the U.S., Vu(S) : 139-141.

Schurnrn, S. A. 1963. The disparity between present rates ofdenudation and orogeny. U.S. Geological Suriey. Profes-sional Paper, 454-H.

Pyroxene compositional trends in theDufek Intrusion, Pensacola Mountains

G. R. HIMMELBERG and A. B. FORD

Department of Geology, University of Missouri,Columbia

U.S. Geological Survey, Menlo Pat-k, California

The Dufek intrusion in the northern Pensacola Moun-tains is a stratiform mafic complex with many similari-ties in primary structure and stratigraphy to the Still-water and Bushveld complexes of Montana and SouthAfrica, as well as to the much smaller Skaergaard in-trusion of east Greenland. The layered gabbros are ex-posed in two partial, non-overlapping stratigraphic sec-tions, one in the Dufek Massif and the other in theForrestal Range. Each section is nearly 2 kilometersthick. The intrusion, described in Ford and Boyd(1968) and Ford (1970), consists dominantly of gab-bro interlayed with minor anorthosite and pyroxeniteand capped concordantly by granophyre.

Ford (1970) demonstrated that the most pronouncedwhole-rock chemical trend, stratigraphically ascendingin the layered series of gabbros, is progressive iron-enrichment. We recently initiated mineralogical andchemical studies of coexisting Ca-rich and Ca-poor py-roxenes, the principal mafic minerals in the complex,

260 ANTARCTIC JOURNAL

P__-

I Ca

Figure 1 Trend of crystalli-zation of Ca-rich and Ca-poor pyroxenes from thelower part of the Dufek in-trusion. Open circles are inter-cumulus pyroxenes, closedcircles are cumulus pyroxenes.Because of crowding, tie lineshave not been shown for allpairs. The Skaergaard (solidlines) and Palisade (brokenlines) trends are shown for

comparison. Mg

in order to further elucidate the crystallization historyand differentiation trends of the intrusion. This progressreport summarizes compositional trends of the pyrox-enes thus far analyzed. Fifteen of the samples are fromthe lower partial section, Dufek Massif, and one isfrom the lowermost exposed gabbro in the upper sectionin the Forrestal Range. Analyses were made with anAR L E MX-SM electron microprobe.

Both the Ca-rich and Ca-poor pyroxenes have abun-dant exsolved Ca-poor and Ca-rich pyroxene, respectively.In most of the Ca-poor pyroxene hosts the exsolvedphase occurs as lamellae and irregularly distributed blebs.In the Ca-rich pyr?xene hosts the exsolved phase occursas lamellae. In order to determine as close as possiblethe bulk composition for the pyroxenes, two grainseach of Ca-poor and Ca-rich pyroxene in each samplewere traversed with adjacent analytical areas by using a20 to 25 micron diameter beam. The data presentedin this report are based on averages of the two grains.Differences in major oxide content between the twoanalyzed grains rarely exceed ± 3 percent of the amountpresent. The standards were minerals and syntheticglasses of known composition. Corrections were madefor the background, mass absorption, secondary fluores-cence, and atomic number. Iron values are total ironcomputed as Fe2+.

The variations, with fractionation, of some of the ma-jor cations (iron, magnesium, and calcium) are shown inthe conventional pyroxene quadrilateral (fig. 1). Thosepyroxenes plotted with open circles are intercumulusand occur in anorthosite from the lowest exposed rocksof the Dufek Massif. All the other plotted pyroxenesare cumulus in origin (they accumulated on the cham-ber floor by settling from the melt). The most Mg-rich and Ca-poor pyroxene analyzed probably crystal-lized as a primary hypersthene. The twinning, characterof exsolution textures and compositions of all the otherCa-poor pyroxenes, lead us to interpret them as hyper-

atomic % Fe

sthene inverted from pigeonite (Hess, 1941; Polder-vaart and Hess, 1951; Brown, 1957). The Ca-rich py-roxenes are augite.

The coexisting pyroxenes show a general iron-enrich-ment, with fractionation. Details of this trend are illus-trated by the plot of the cation ratio Fe/(Fe + Mg)versus stratigraphic height (fig. 2). Through a cumula-tive thickness of approximately 1000 meters, betweenabout the 500-meter and 1500-meter levels, the ironenrichment is small and accompanied by minor reversalsin the cation ratio of Fe/(Fe + Mg). Ford (1970)demonstrated a similar trend in the gabbros for thevariation in whole-rock mafic index, (FeO + Fe203/FeO + Fe 2O + MgO) X 100, which he interpretedas being related to chemical changes accompanying con-vective activity. The mafic-index trend line is approxi-mately parallel to the pyroxenes' trend line below thetop 500 meters of the Dufek Massif section—that is,below the level at which cumulus magnetite first appears.Minor reversals in pyroxene composition in the StillwaterComplex also were interpreted as being related to con-vective overturn (Hess, 1960). Jackson (1970)and Page et al. (1972) have shown that even withinindividual cyclic units of the Stillwater Complex cumulusminerals can vary widely in composition. Page et al(1972) ascribe the variations to oscillatory or cyclicprocesses within a single. magma batch.

Comparison with pyroxene trends of other largestratiform intrusions suggests that the pyroxenes fromthe Dufek Massif are more like those from the Skaer-gaard intrusion (Brown and Vincent, 1963) than likethose from the Bushveld complex (Atkins, 1969), withrespect to the miscibility gap's extent. Striking simi-larities also exist when compared to the trend estab-lished for the Palisades Sill pyroxenes (Walker et al.,1973). More Mg-rich pyroxenes are expected to occurin the unexposed basal layers of the Dufek intrusion.Similarily, the slight compositional gap between the

September-October 1973 261

"t.,71•

/ 704/slayeredgabbro

/ AAS5AAssAs

IAsAS

50 -' AA

anorthosi te0A

DufekMassifsect/on

augiteo hypersthene

6- granophyre/

'C 90o'c

80ForrestolRange layeredsect/on gabbro

5

E A

)

.

(concealed)-c

G)-C

C.)-cci0CP

05-

.4-

(I)

Figure 2. Variation of pyrox-ene cation ratio Fe! (Fe +Mg) with stratigraphic height.Open symbols are intercumu-lus pyroxenes, closed symbolsare cumulus pyroxenes. Thebroken line shows the approx-imate trend of mafic index ingabbroic rocks and granophy-

02 04 0.6 0.8 re (from Ford, 970). Num-bers indicate numerical value

Cation ratio Fe/(Fe+Mg) of maflc index.

pyroxenes from the uppermost part of the Dufek Massifstudy, with assistance from National Science Founda-and those from the lowermost exposed gabbros of thetion grant GA-18445.Forrestal Range probably is because of concealment ofan intermediate section of the complex. We expect thatthe iron-enrichment trend of the pyroxenes will be ex-tended when analyses of higher level pyroxenes arecompleted. References

This work is supported by National Science Founda-tion grant AG-238. The University of Missouri pur-Atkins, F. B. 1969. Pyroxenes of the Bushveld Intrusionchased the electron microprobe that was used for thisSouth Africa. Journal of Petrology, 10: 222-249.

262 ANTARCTIC JOURNAL

Brown, G. M. 1957. Pyroxenes from the early and middlestages of fractionation of the Skaergaard intrusion, EastGreenland. Mineralogical Magazine, 31: 511-543.

Brown, G. M., and E. A. Vincent. 1963. Pyroxenes from thelate stages of fractionation of the Skaergaard intrusion, EastGreenland. Journal of Petrology, 4: 175-197.

Ford, A. B. 1970. Development of the layered series andcapping granophyre of the Dufek intrusion of Antarctica.In: Symposium on the Bushveld Igneous Complex and OtherLayered Intrusions (D. J . L. Visser and G. von Gruenewaldt,eds.). Geological Society of South Africa. Special Publica-tion, 1: 494-510.

Ford, A. B., and W. W. Boyd, Jr. 1968. The Dufek intru-sion, a major stratiform gabbroic body in the PensacolaMountains, Antarctica. Proceedings of the 23rd Interna-tional Geological Congress, 2: 213-228.

Hess, H. H. 1941. Pyroxenes of common mafic magmas.American Mineralogist, 26: 515-535, 573594.

Hess, H. H. 1960. Stillwater igneous complex, Montana: aquantitative mineralogical study. Geological Society of Amer-ica. Memoir, 80. p. 121.

Jackson, E. D. 1970. The cyclic unit in layered intrusions-acomparison of repetitive stratigraphy in the ultraniafic partsof the Stillwater, Great Dyke, and Bushveld Complexes. In:Symposium on the Bushveld Igneous Complex and OtherLayered Intrusions (D. J . L. Visser and G. von Gruenewaldt,eds.). Geological Society of South Africa. Special Publi-cation, 1: 391-424.

Page, N. J . , R. Shimek, and R. Huffman, Jr. 1972. Grain-sizevariations within an olivine cumulate, Stillwater Complex,Montana. U.S. Geological Survey. Professional Paper, 800-C,C29-C37.

Poldervaart, A., and H. H. Hess. 1951. Pyroxenes in thecrystallization of basaltic magma. Journal of Geology, 59,472-489.

Walker, K. R., N. G. Ware, and J . R. Lovering. 1973. Com -positional variations in the pyroxenes of the differentiatedPalisades Sill, New Jersey. Geological Society of America.Bulletin, 84, 89-110.

Analysis of antarctic geophysical data

C. R. BENTLEY, H. K. ACHARYA, J . L. CLAPP,

J . W. CLOUGH, H. KOHNEN, and J . D. ROBERTSON

Geophysical and Polar Research CenterDepartment of Geology and Geophysics

University of Wisconsin, Madison

Continuing analysis of antarctic geophysical data fol-lows several lines, including studies of ice properties (asrevealed by seismic and electromagnetic wave propaga-tion experiments near Byrd Station), west antarcticgravity maps, Roosevelt Island strain data, and theo-retical studies of seismic wave propagation. Appendedis a bibliography of papers on these subjects (since

Contribution number 299, Geophysical and Polar ResearchCenter, Department of Geology and Geophysics, University ofWisconsin, Madison.

Bentley et al., 1969). What follows is a summary ofrecent results not yet published.

1. Seismic velocities obtained from short refractionprofiles can be used to predict density at depths betweeno and 10 meters, with a standard error of about 0.01gm/cm3.

2. A newly recognized and extensive horizon at depthsof 25 to 30 meters, marking an apparent change in thedensification rate, has been found in West Antarctica.The horizon's existence suggests that two distinct mech-anisms successively dominate the metamorphic processbetween the depth of closest packing of snow grains andthe urn-ice boundary.

3. Measurement of P-wave attenuation in ice nearByrd Station led to the determination of a very low valuefor the internal friction. From comparison with labora-tory measurements (Kuroiwa, 1964), it appears that aslight but significant contamination of the antarctic iceby ionic impurities (Gow, 1968) and the ambient icetemperature (-28 0 C.) result in the falling of seismicfrequencies at a dissipation minimum between spectralregions dominated respectively by grain-boundary phe-nomena, and by the fundamental relaxation spectrum.

4. Analysis of electromagnetic wide-angle reflectionmeasurements shows that correction for refraction in theupper portion of an ice sheet need not be made whencalculating mean velocities. Even for a reflector as shal-low in depth as 100 meters, the error introduced byassuming straight line geometry is only about 10 nanosec,well below the time resolution of the measurements. Foraccurate measurements, however, the length of wide-angle profiles must be limited to distances correspondingto the reflection path for a ray at grazing incidence onthe surface. The amplitude of the reflected wave, whichchanges markedly along the length of a profile, can playan important role in the measurements.

5. Previously existing maps of gravity anomalies inWest Antarctica have been supplemented by new dataand have been contoured by a computer. Three-dimensional modeling has been used to prepare an Airy isostaticgravity anomaly map. This map reveals several imbal-ances, which may be caused by: up-warping of the M-discontinuity in Ellsworth Land and beneath the Hollick-Kenyon Plateau; dense, lower-crustal material unusuallynear the surface southwestward from the WhitmoreMountains; the extension beneath the Rockefeller Plateauof a pre-Cretaceous geosyncline known to exist in theEdsel Ford Ranges; the southern boundary of the Ceno-zoic volcanic province in Marie Byrd Land. A deepnegative anomaly of unknown cause exists along theBakutis Coast. On the hypothesis of a recent retreat ofthe ice sheet, no more than 40 percent (and probablymuch less) of the anomaly can be attributed to incom-plete isostatic rebound.

6. A Rayleigh wave group-veloiity curve, applicableto the known velocity-depth and density-depth curves in

September-October 1973 263