are so similar in all other respects? The geographicpositions of these features reveal the answer.

Minna Bluff (elev. 1,060 meters) juts into the RossIce Shelf like a giant snow fence and either traps ordiverts the snow that prevailing southerly windssweep off the ice shelf. Black Island is directly northof Minna Bluff and so does not get much blownsnow. White Island has nothing to the south of itbut open ice shelf, so snow can be blown over anddeposited on this island. The difference in snow ac-cumulation on the two islands is perhaps accentuatedby the wedge shape of White Island with its wideside to the south, which favors the collection of snow.Black Island is shaped with a point to the south thatdiverts the winds around it.

The difference in snow accumulation on Black Is-land and White Island thus seems to be best explainedby the geographic and meteorologic conditions de-scribed, although additional factors may be involved.

Density of thestratiform Dufek intrusion,

Pensacola Mountains, AntarcticaA. B. FORD

U.S. Geological SurveyMenlo Park, California

S. W. NELSON

Department of GeologyUniversity of Nevada

An immense layered gabhroic complex, the Dufekintrusion, makes up the entire northern one-third ofthe Pensacola Mountains near the head of the Wed-dell Sea. Discovered only in 1957 on an IGY traversefrom Ellsworth Station (Aughenbaugh, 1961; Walker,1961), the complex was mapped in entirety, geo-physically surveyed, and extensively sampled by ateam of U.S. Geological Survey geologists, geophysi-cists, and topographic engineers in the austral sum-mer of 1965-1966 (Schmidt and Ford, 1966, 1969;Ford and Boyd, 1968; Behrendt et al., 1966). Com-pilation and analysis of the field data and laboratorystudies of samples have continued; this report brieflysummarizes some of this work.

Parts of the intrusive body are excellently exposedin enormous escarpments that provide two completesections for study: a lower one, 2 kilometers thick, inthe Dufek Massif, and an upper one, also about 2kilometers thick, in the Forrestal Range. Although

Publication authorized by the Director, U.S. GeologicalSurvey.

feldspathic pyroxenite

90- (pyroxene cumulate)

gabbroicpyroxenite'(plagioclase—pyroxenecumulate)

TO--

-pyroxene-plogioclos

Y.

06. ..-. -----

mafic gabbro •:

-(plogioclose—pyroxenet/iCL cumulate)50-------------...1 ---------%•1S7

gobbro(

• .-cumulate)30- - --------3•.....------------

onorthositic gabbro

- (pyroxene—plagioclasecumulate)

10-VY•Z ...............onorthositeA:(plogloclose cumulate)

-

*I IIIIII2502.803.003.203A0

density

Figure 1. Variation of density, in grams per tubic centimeter,with pyroxene content, in volume percent, and rock type for the

Dufek Massif section.

the basal zone, an intermediate zone estimated to beon the order of 2 kilometers thick, and the roof arenot exposed, indirect evidence suggests that the totalthickness is at least 7 kilometers (Ford, 1970) andthat the layered mafic rocks extend for great distancesbeneath adjoining continental ice sheets, probablyover an area of at least 34,000 square kilometers inall (Behrendt, 1971). These estimates indicate thatthe body is comparable in size to some of the largestlayered mafic complexes in the world. Although suchbodies typically occur in a Precambrian craton set-ting, the Dufek body lies in a recurrently activeorogenic belt marginal to the antarctic craton. Thelatest major deformation in the belt took place inprobable Triassic time (Ford, in press), as indicatedby the presence of Permian fossils in folded beds andby Middle Jurassic potassium-argon dates (R. W.Kistler, written communication, 1969) for the post-orogenic Dufek intrusion. Radiometric dating andchemical characteristics suggest that the body is re-lated to Ferrar diabase intrusive activity (Compstonet al., 1968) elsewhere in the Transantarctic Moun-tains.

The Dufek body, which is considerably more dif-ferentiated than any known Ferrar diabase sheet, iscomposed of a highly varied suite of layered rocksranging from anorthosite and granophyre to pyroxe-nite and magnetite. The great bulk, however, isgabbro with variable amounts of the principal min -eral phases, plagioclase, pyroxene—both clinopyrox-ene and orthopyroxene—and magnetite or other iron-and titanium-rich oxides. Bulk-rock densities closelyreflect the varying major mineral content (fig. 1), as

September-October 1972 147

FORRESTAL RANGE

N

DUFEK MASSIF SECTION

N •N N••'0; .• a • NI. •:IL

SECTION

LA,N

J

stratigrophicheight(kilometers)UI a,)

Figure 2. Variation of density with stratigraphic height in Dufek intrusion. x pyroxenite layers; y magnetitite layers; z anorthosite layers.

In

C

Ea,I-

U)0a,E00C

40— FORRESTALRANGE

n: 22320- A B

IF-I-fl-.;-.t F]2702.903.10330 i504

80 -DUFEK

— MASSIFE60-

U—CB

40-0

20-

2IOdensity

Figure 3. Frequency distribution of rock densities in the Dufekintrusion.

well as chemistry, particularly total iron oxides in therocks. Magnetic susceptibility (K) shows a generalpositive correlation with density in each of the ex-posed sections (Griffin, 1969). Using the curve offig. 1, density measurement can provide a rapidmeans for preliminary classification of many Dufekrocks.

Many aspects of magmatic history are clearly re-flected in the vertical distribution of rock-density

variants in the complex (fig. 2). Physical processesthat operated during consolidation of the immensemagma reservoir, including lateral current activityand gravity settling, were analogous to processes thatoperate during accumulation of some types of water-lain clastic materials that form sedimentary rocks.The igneous Dufek "sediments" accumulated on thechamber floor in a crystal rain, interrupted episodi-cally by turbidity-current-like floodings across thefloor by crystal-rich magma currents presumably gen-erated by convection. Evidence of scour along thefloor, which rose as crystals accumulated, is clear atseveral levels. Early currents carried mainly pyrox-enes, later ones mainly plagioclases. The many sharpfluctuations in the density curve of fig. 2 correspondto thin pyroxenitic and anorthositic layers formedthereby, as well as to gravity-accumulated magnetiteconcentrations that occur mainly in the ForrestalRange section.

The continual separation of crystals led to progres-sive changes in melt composition, to accompanyingchemical changes in later formed crystals, and to theappearance of new phases in successively higher cumu-lates. Early separation of magnesium-rich pyroxene,and presumably olivine in the unexposed basal layers,resulted in increasing iron content of the originallytholeiitic melt and eventually in the crystallization ofmagnetite, locally in great amounts. The melt be-came enriched in alkalies and silica by early, andcontinual, fractionation of calcium-rich plagioclaseand enriched in water owing to the anhydrous char-acter of all early phases. Such changes led, in thefinal stage of consolidation, to the development of a300-meter-thick capping layer of alkaline granitic

364

F

148 ANTARCTIC JOURNAL

residue of granophyre containing iron-rich clino-pyroxene, hornblende, and biotite.

Densities measured on approximately 600 samplesrange widely from as low as 2.65 grams per cubiccentimeter for granophyre and 2.70 g/cc for an-orthosite to as much as 3.30 g/cc for pyroxenite and3.50 g/cc or more for magnetitite. Most gabbros(pyroxene-plagioclase cumulates) lie in the range2.80-3.20 g/cc (figs. 1 to 3). Weighted according tolayer thicknesses, the average density of the DufekMassif section is about 2.95 g/cc; of the ForrestalRange section, about 3.03 g/cc. The estimated aver-age for the entire body, taking into consideration theprobable densities of unexposed sections, approximatesthat of R. A. Daly's average gahbro or norite (Dalyet al., 1966), about 2.98 g/cc, and only slightly ex-ceeds that of about 2.95 g/cc measured on rocks fromlittle differentiated diabase sills in the southern Pensa-cola Mountains. The upward increase in averagedensity, contrasting with general upward decreasecommon in thin diabase sills elsewhere (Jaeger, 1964),obviously reflects the strong trend of iron enrichmentduring fractionational crystallization of the Dufekmagma. The Dufek body is a highly inhomogeneousmass, and such differences in density for differentparts of the total stratigraphic section as indicatedhere should be considered in future more detailedgravity studies when sub-ice terrain maps becomeavailable.

This work is supported by National Science Founda-tion grant AG-238.

References

Aughenbaugh, N. B. 1961. Preliminary report on the geologyof the Dufek Massif. International Geophysical Year WorldData Center A Glaciology. Glaciology Report, 4: 155-193.

Behrendt, J . C. 1971. Interpretation of geophysical data inthe Pensacola Mountains, Antarctica. Antarctic Journalof the U.S., VI(5): 196-197.

Behrendt, J . C., L. Meister, and J . R. Henderson. 1966. Air-borne geophysical study in the Pensacola Mountains,Antarctica. Science, 153 (3742) : 1373-1376.

Compston, W., I. McDougall, and K. S. Heier. 1968. Geo-chemical comparison of the Mesozoic basaltic rocks ofAntarctica, South Africa, South America, and Tasmania.Geochemica et Cosmochimica Acta, 32(2): 129-149.

Daly, R. A., G. E. Menger, and S. P. Clark, Jr. 1966. Den-sity of rocks. In: Handbook of Physical Constants (S. P.Clark, Jr., ed.). Geological Society of America. Memoir,97: 19-26.

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

Ford, A. B. In press. The Weddell orogeny-latest Permianto early Mesozoic deformation at the Weddell Sea marginof the Transantarctic Mountains. In: Antarctic Geology

and Geophysics (R. J . Adie, ed.). Oslo, Universitets-forlaget.

Ford, A. B., and W. W. Boyd, Jr. 1968. The Dufek intrusion,a major stratiform gabbroic body in the Pensacola Moun-tains, Antarctica. Proceedings of the 23rd InternationalGeological Congress, vol. 2: 213-228.

Griffin, N. L. 1969. Paleomagnetic properties of the Dufekintrusion, Pensacola Mountains, Antarctica. MS Thesis.University of California, Riverside. 93 p.

Jaeger, J . C. 1964. The value of measurements of densityin the study of dolerites. Journal of the Geological Societyof Australia, 11. 133-140.

Schmidt, D. L., and A. B. Ford. 1966. Geology of the north-ern Pensacola Mountains and adjacent areas. AntarcticJournal of the U.S., 1(4): 125.

Schmidt, D. L., and A. B. Ford. 1969. Geologic Map ofAntarctica (Pensacola and Thiel Mountains) (Sheet 5).Antarctic Map Folio Series, 12.

Walker, P. T. 1961. Study of some rocks and minerals fromthe Dufek Massif, Antarctica. International GeophysicalYear World Data Center A Glaciology. Glaciology Report,4: 195-213.

Rb-Sr and K-Ar dating of rocksfrom southern Chileand West Antarctica

MARTIN HALPERN

Geosciences DivisionUniversity of Texas at Dallas

Geological and geophysical field programs in thesouth of Chile (Halpern, 1970) and in West Antarc-tica have provided the opportunity for collecting sam-ples of igneous and metamorphic rocks for radiometricdating. The aim of this program was to establish thechronology of principal rock units so that the geologichistory of these remote regions of the earth's crustcould be understood. Rubidium-strontium isotopicage analyses were carried out at the University ofTexas at Dallas and potassium-argon isotopic datingat the University of Leeds, England.

In southern Chile, metamorphic rocks constitutethe oldest known rocks. Gneiss from the 'basement'of the Magellan Basin at the Atlantic entrance to theStrait of Magellan have been rubidium-strontiumtotal rock dated at 306 ± 156 million years (Xf3 =1.47 x 10 per year) with an initial strontium-87to strontium-86 ratio of 0.7112 ± 0.0033. Biotitefrom a sample of the gneiss has been rubidium-stron-tium and potassium-argon dated as Permian, implyingthat the 'basement' of the Magellan Basin has beeninvolved in one or more Paleozoic geologic events.Paraschists from the 'basement' complex along Chile's

September-October 1972 149