INTRODUCTION.
Meteoriticists have long known that the populations of various meteorite types are different for falls and finds (1). This is attributed to the preferrential collection of iron meteorites as finds because they are dense and easier to distinguish from terrestrial rocks. It was hoped that collection of Antarctic meteorites on ice would be less biased and that Antarctic meteorites would be more representative than other meteorite finds. The last few years have seen a major debate about whether Antarctic meteorites represent different populations than meteorite falls (2,3). We present a review of the populations of meteorites in our collection, pairing of Antarctic meteorites, and the abundances of rare meteorites. We show that our best estimate of Antarctic meteorite populations is very similar to that of falls, except that rare meteorites are more abundant in Antarctic meteorite collections, presumably because they are small and hard to find in other environments.

POPULATIONS.
The populations of major meteorite types for non-Antarctic falls, finds, and Antarctic finds are given in Table 1. The first three populations are based simply on the total numbers of meteorites. This overall distribution of Antarctic meteorites is more similar to that of non-Antarctic falls than to finds, however the Antarctic population is enriched in ordinary chondrites and depleted in achondrites, stony-irons and irons relative to that of falls. This comparison is flawed by the uncertainty in the number of individual meteorites that the Antarctic meteorites represent. The Antarctic data do not subtract for the possible pairing of two or more meteorite fragments (see below). To avoid this problem, some investigators (6,7) have chosen to compare mass distributions like those in columns 4 and 5. However the simple mass distributions listed here are scewed by the inclusion of large Antarctic irons not found on ice as part of the thorough meteorite search (6,8). Analysis and modelling of the meteorite mass distributions (6,7) concluded that there was no significant difference between falls and Antarctic finds. However, Lipschutz and colleagues (2, 9) have argued based on the numbers of meteorites, that differences in populations are significant. As part of this debate, we offer a simple method of estimating the numbers of Antarctic meteorites that includes pairing corrections for meteorite showers.

PAIRING.
Meteorites often fall as showers of a few to many fragments. When this happens most places in the world, all fragments are grouped together as the same meteorite. In the Antarctic meteorites from many different falls are concentrated together (8) and it is not simple to determine which ones are from the same fall, or paired specimens. The problem of pairing in Antarctic meteorites has been discussed by several authors (10-12) who have concluded pairing would reduce the number of Antarctic meteorites by factors of 2-10, but most likely 2-6. It is relatively simple to evaluate pairing among the less common meteorite types. Meteorites that are similar to each other and collected in close proximity are most likely from the same fall. Our database of US Antarctic meteorites (4) includes pairing estimates which are known with varying certainty (11). There are 150 pairing groups among the 5700 classified meteorites, with 2-678 meteorites per group. There are, however, only 16 groups with over 10 meteorites and of these only 5 groups have over 25 meteorites. Among the less common meteorites, the number of meteorites per group exceeds 10 in only a few cases: the 20 ALH aubrites, 14 ALH eucrites, 34 ALH CM2, 16 EET CR2, 47 EET CK5 carbonaceous chondrites, and 21 PCA EH3 enstatite chondrites. The average number of specimens per group for the 67 pairing groups of the less common meteorites is 5, in the range of previous estimates of pairing ratios. It is clear that the number of meteorites in a group increases with the the total mass of the group. Among ordinary chondrites pairing is much less certain, but there are some large pairing groups for the ALH L3, QUE L5 and EET L6 chondrites. It is much harder to estimate pairing among H compared to L chondrites, and fewer H pairing groups. We do not believe that this is accurate, but that there are unidentified pairings among H chondrites. Huss (7) proposed the existence of a large ALH H5 pairing group based on mass distributions. In the absence of accurate pairing data for ordinary chondrites, we feel that it is best to use the average paring ratio of 5 from the less common meteorites. When the actual pairing-corrected numbers are used for the five less common meteorite types, and a pairing ratio of 5 is used for ordinary chondrites, the populations of Antarctic meteorites are as shown in the last column of Table 1. This population is remarkably similar to that of non-Antarctic falls as given in the first column. We see no evidence that the overall population of meteorite types has changed over time.

Figure 1. Meteorite populations by number. This diagram compares populations for falls, finds and Antarctic meteorites and shows that Antarctic meteorites are more like falls than finds.

Figure 2. Pairing of Antarctic meteorites. This histogram of the number of meteorites per pairing group shows that the average number of Antarctic meteorites per pairing group is 5.

RARE METEORITES.
There is, however, strong evidence that the abundance of rare or anomalous meteorites is significantly higher among Antaractic meteorites (3,13-15). Table 2 lists the Antarctic meteorite populations of each of the meteorite subtypes. The last subtype in each major type is that for rare or unusual meteorites. These generally include several different types of meteorites, as for example, the achondrites include acapulcoites, angrites, brachinites, lunar and martian meteorites. Among non-Antarctic meteorites this subtype would include only a few specimens and not amount to more than 1-2%. For achondrites, stony-irons, irons, and carbonaceous chondrites the abundance of rare meteorites is 10-20%! This surprising abundance of rare meteorites in Antarctica is attributed to the fact that it is easier to find small meteorites on ice than on land. Most rare Antarctic meteorites are small specimens ( < 30g). An exception to this is the planetary meteorites. Although lunar meteorites are found almost exclusively in Antarctica, and most lunar meteorites are < 30g, 3 of the 11 are ~500g. Martian meteorites are generally large. Only one of the 4 Antarctic meteorites is small, the others being, 500g, 1900g and 7900g. It would appear that a difference in impact dynamics makes these rare planetary meteorites larger specimens.

Figure 3. Meteorite populations by number, pairing corrected. The populations of Antarctic meteorites, when corrected for pairing, is identical to that of falls.

REFERENCES.

1. Mason B. (1962) Meteorites. John Wiley and Sons.
   
2. Dennison J., Lingner D., and Lipschutz M. (1986) Nature 319:390-393.
   
3. Koeberl C. and Cassidy W. (1991) Geochim. Cosmochim. Acta 55:3-18.
   
4. Grossman J. (1994) Meteoritics 29, 100-143 and Ant.Met.News. (1994), 17(1).
   
5. Graham A.L., Bevan A.W.R., and Hutchison R. (1985) Catalogue of Meteorites. 4th ed. British Museum (Natural History).
   
6. Cassidy W.A. and Harvey R.P. (1991) Geochim. Cosmochim. Acta 55, 99-104.
   
7. Huss G. (1991) Geochim. Cosmochim. Acta 55:105-111.
   
8. Cassidy W., Harvey, R., Schutt J., Delisle G., and Yanai K. (1992) Meteoritics 27:490-525.
   
9. Lipschutz M. (1989) Workshop on Differnences between Antarctic and non-Antarctic Meteorites, LPI, 59-61.
   
10. Graham A.L. and Annexstad J.O.(1989) Antarctic Sci. 1: 3-14.
    
11. Scott E.R.D. (1989) Smithsonian Contrib. Earth Sci. 28: 103-111.
    
12. Ikeda Y. and Kimura K. (1992) Meteoritics 27:435-441.
    
13. Marvin U.B. (1983) New Scientist 17, 710-715.
    
14. Clarke R.S. (1986) International Workshop on Antarctic Meteorites, LPI, 28-29.
    
15. Wasson J. (1990) Science 249:900-902.