BRUCE D. MARSH and MICHAEL J. ZIEG, M.K. Blaustein Department of Earth and Planetary Sciences, Johns Hopkins University, Baltimore, Maryland 12218
Layered intrusions are of nearly mystical proportions in igneous petrology. Although they are sometimes unusually large, reaching approximately 500,000 cubic kilometers (km3) for the Bushveld of South Africa, they are especially noted for their beautiful layers of minerals (Parsons 1987; Cawthorn 1996). The minerals normally found in common basaltic rocks have been somehow sorted to produce a high degree of order. What exact physical or chemical processes lead to this layering has long been debated. Do these layers simply reflect "snow fall" events where the crystallizing magma suddenly churned and sent down a shower of crystals? Could these layers reflect a chemical condition in the magma leading to locally preferred nucleation and growth of a certain mineral? Or could it be a combination of several chemical and physical processes? Regardless of the actual processes involved, however, layering surely records the temporal evolution of the magma much as tree rings do for trees. The main difficulty in reading layering is that the final rock texture, which is so critical to interpretation, generally results from a long and slow process of cooling that may erase, through annealing, sensitive indicators of the conditions of initial formation of the layer. And because cooling time increases with the square of body size, large layered intrusions have cooled so slowly that it is often difficult to recognize the original textures. Small magmatic bodies [i.e., those less than about 500 meters (m) thick] have short cooling times, but they generally show little or poorly developed layering. Thus, our understanding of how layering originates and evolves in magmatic systems contains a gap.
This gap may be closed by our recent discovery of a small layered intrusion associated with the Basement Sill in the McMurdo Dry Valleys. The Basement Sill has long been known to contain a tongue of large orthopyroxene crystals (i.e., phenocrysts) (Gunn 1966), but its tremendous aerial extent and locally massive character have only become appreciated over the past few years (Marsh 1996; Marsh and Philipp 1996).
The Basement Sill is a sheet of basaltic rock suddenly injected during the breakup of Antarctica and Africa about 180 million years ago. The sill crops out over an area of about 5,000 km3 from Cathedral Rocks in the south to the Debenham Glacier area in the north and also perhaps well beyond this region. Northeast of Lake Vida, the Basement Sill reaches a thickness of about 700 m. We have traced the tongue of phenocrysts throughout the dry valleys.
A distinctive feature of the phenocryst tongue is the presence of slight but pervasive layering. The layers, often more aptly described as stringers, are exclusively due to natural sorting of the two principal constituent minerals; orthopyroxene (brown-green) and plagioclase feldspar (white). The layers or stringers are found only in the tongue, usually reach thicknesses of only a few centimeters, and are laterally continuous up to about 10 m.
At the east end of the Dais in western Wright Valley is a well-developed layered intrusion about 200 m thick; the lower half is covered ( see figure 1). The layering is evident on scales from centimeters to tens of meters. From as far away as 15 km down Wright Valley, the coarsest layering is evident as alternating light and dark bands. And close up, these coarse (10-30 m) bands consist of many smaller sublayers of from 1 centimeter (cm) to 3-4 m in thickness. On the outcrops, the coarse layering is unrecognizable unless its presence is already known. The alternating light and dark colors of these coarse layers suggest a slight, but distinctive, alternating dominance of orthopyroxene and plagioclase. Individual distinct layers generally do not exceed 20-30 cm in thickness and are often laterally continuous for only a few meters. But even where the rock appears unlayered, there is everywhere, however slight, modal banding. Plagioclase-rich layers are particularly distinct and two near the base of the exposed section (northeast Dais) are essentially anorthosites (15 and 35 cm thick, see figure 2). These layers are laterally continuous for about 50-70 m; they then fade and reappear laterally along strike at several more locations. The style of layering is generally maintained at each horizon even though specific layers may be discontinuous.
The loss of lateral continuity of these layers in several places may be related to depositional scouring as is often seen in particle debris flows or slurry flows. Some 20 m upward in this section is a large trough feature; it is perhaps 50 m wide and 3-5 m deep and consists of a well-sorted assemblage of small [1-3-millimeter (mm)] grains of orthopyroxene and plagioclase. The overall appearance is similar to a large block of sandstone. The trough appears to have been filled by settling of a vast assemblage of fine-grained minerals under quiet magmatic conditions.
Individual layers, regardless of size, show gradational bottoms and sharp tops; often there are local concentrations of long (3-4-cm) rodlike orthopyroxenes lying flat in the horizon of the layer itself. Grain size generally fines upward. Of the many types of layers, thin (approximately 10-cm) layers of massive orthopyroxene and plagioclase are distinctive. In hand sample, the plagioclase is not fine grained, as is generally seen, but is massive with interspersed massive clots of orthopyroxene. Local vertical pods of coarse, dark orthopyroxene are also evident, and they show an apparent cross-cutting relationship to the horizontal layers. Irregular in form, these pods are 2-3 m tall and up to 0.5 m wide.
The overall appearance of this layered sequence clearly represents a depositional sequence (as opposed to in situ crystallization) in a dynamic environment. The crystals involved in sorting, pyroxene and plagioclase, were carried by the invading magma. They did not grow after emplacement. That this layered body is near the inferred filling point of the Basement Sill itself ( see Marsh and Philipp 1996), coupled with the juxtaposition of delicately sorted layers and scour and fill troughs, suggests local ponding of the infilling sill magma perhaps near the outer reaches of a periodically avalanching thick pile of crystals.
The mechanism and efficiency of crystal sorting are clearly due to the great difference in size between the pyroxene (3-30-mm) and plagioclase (0.1-3-mm) crystals. Any slight concentration of pyroxene crystals functions as a sieve to the plagioclase crystals, allowing them to settle freely through the pyroxenes to form a layer of plagioclase. And pyroxene crystals isolated within a concentration of plagioclase settle due to their weight and size until encountering more pyroxenes. The natural tendency is to form layers of each mineral with the overall form of the layer reflecting its mode of deposition. The textures so formed are remarkably similar to those of the orthopyroxene/plagioclase rocks of the huge Stillwater layered intrusion of Montana. The principal differences are in the slightly larger size of the Stillwater plagioclase and the overall annealed form of the Stillwater textures. Both of these features are clearly due to the enormously longer cooling time (hundreds of thousands of years) of Stillwater, which allows for significant postdepositional recrystallization.
In summary, the presence of large concentrations of two minerals of disparate size and density in the Dais magma has unavoidably led to layering. The relatively short cooling time (approximately 1,000 years) of the Dais intrusion has preserved the original textures unusually well, making this body of perhaps singular importance to the study of layered intrusions.
This research was supported by National Science Foundation grant OPP 94-18513.
Cawthorn, R.G. (Ed.). 1996. Layered intrusions. Amsterdam: Elsevier Science Publishers.
Gunn, B.M. 1966. Modal and element variation in antarctic tholeiites. Geochima et Cosmochima Acta , 30, 881-920.
Marsh, B.D. 1996. Sill sequences as piecemeal layered intrusions and links to volcanism: Evidence from the Ferrar dolerites of Antarctica. EOS, Transactions of the American Geophysical Union , 77, S292.
Marsh, B.D., and J.R. Philipp. 1996. Three-dimensional magmatic filling of Basement Sill revealed by unusual crystal concentrations. Antarctic Journal of the U.S. , 31(2), 39-40.
Parsons, I. (Ed.). 1987. Origins of igneous layering . Dordrecht: D. Reidel.