One of the primary goals of the Apollo missions was to provide information about the crater production rate on the moon, and hence on the Earth. Before the Apollo missions relative ages of different surfaces could be calculated by comparing the number of craters of a given size on the different surfaces. By returning samples from different lunar surfaces for radiometric dating, it became possible to attach absolute ages to different surfaces and to re-construct the cratering history of the early solar system, as in figure 1.

Cratering rates on the moon.
Figure 1. Lunar crater production rates through geologic time as reconstructed from the measurement of crater densities on the lunar surface and from absolute age dating of returned lunar rocks. Firm correlations can only be reconstructed for (1) the well-characterized basalt surfaces (3.8-3.2 Gyr) and (2) the contemporary meteorite flux based on current astronomical observations (t = 0). The ages of Tycho and Copernicus are inferred from indirect evidence. From F. Horz et al. p.84.

Upon study of figure 1 it becomes apparent that, while the cratering history of the solar system between 4.5 and 3.0 billion years (Gyr) is well known, that little is known afterwards. Copernicus, Tycho and a handful of other craters are the only radiometrically dated impacts younger than 3 Gyr, and even these dates are subject to interpertation. Impacts on the Earth can also be dated to further constrain the cratering history, but erosion and plate tectonics makes this a difficult task, and the results are, at best, incomplete.

Our most fundamental goal is to provide data about the crater production rates within the last 3 Gyr, although the method we shall propose will determine relative rates only, not absolute rates. We will do this by 40Ar/39Ar dating (a variant on potassium-argon dating) of glass spherules from random craters. Initially we will measure the ages of 30 to 40 lunar spherules; ultimately we would like to obtain the ages of several hundred to a thousand or more.

Interest in the recent terrestrial cratering record was piqued in 1980, when Luis and Walter Alvarez, Frank Asaro and Helen Michel published the hypothesis that the impact of a comet or asteroid on the Earth was the cause of the Cretaceous catastrophe, the mass extinction that killed over 50% of the Earthís life, including the dinosaurs. The "smoking gun" for this event has now been found, with the identification of the Chicxulub crater in the Yucatan as the site of the impact and its dating to precisely the time of the extinctions, to within 0.1 million years (Myr). The age was determined by 40Ar/39Ar dating of individual glass spherules thrown out by the impact. It was done at the Berkeley Geochronology Center.

Furthermore, it is not known if this impact was an isolated event, or part of an intense period of impacts. Comet showers are brief periods of intense comet impacts predicted by J. Hills in 1981. He showed that such a shower, bringing 109 or more comets into the inner solar system, would result if a star passed close to the Oort comet cloud. In 1984 we showed that the effect of a comet storm might be to give an illusion of an extended extinction period, perhaps accounting for the widely-held belief that the mass extinctions were extended in time rather than abrupt. A series of impacts taking place in a 1-million year period would give a series of extinctions, "stepwise," according to a later terminology, which if not recognized would look like a gradual extinction over the same period. Soon afterwards several extinctions were analyzed according to this model, and the stepwise extinction pattern was seen as consistent with the data.

In 1984, paleontologists Raup and Sepkoski analyzed the marine extinction data over the past 250 Myr and found a distinct periodicity of about 26 Myr. Alvarez and Muller analyzed the cratering record on the Earth and found a periodicity of about 30 Myr. However, the small number of well-dated terrestrial impacts means that this result is somewhat ambiguous. If these periodicities exist and if they are linked, our views on the evolution of life would be profoundly affected.

As mentioned above, the terrestrial cratering record is not well preserved. The Moon, however, preserves craters much better than the Earth. This led to Horz's suggestion that we sample the impact glasses from several hundred craters on the moon during another series of lunar exploration. Our proposal is to analyze the glass spherules formed in impacts on the moon, which, while providing similar data, has the substantial advantage of not requiring further lunar missions. The second goal of our project, therefore, is to search for the presence of structure, such as peaks created by comet showers. One of the most interesting results that we can anticipate would be the discovery of clusters of impacts, perhaps one centered about 65 Ma, or 250 Ma at the time of the Permian-Triassic extinctions. Perhaps we could eventually see a series of such events, separated in time by the 26 Myr that Raup and Sepkoski have found in the mass extinctions on the Earth.

Proposed techniques

We believe that we can obtain these dates without going back to the moon by giving up the knowledge of which craters we have dated, and by measuring only relative cratering rates, not absolute ones. Material from large impacts is thrown great distances; the rays from craters such as Copernicus and Tycho reach far across the face of the moon this is reasonable, since typical impact velocities for comets are 30-45 km/sec, much greater than the lunar escape velocity of 2.38 km/sec. In our work with lunar soil, we have discovered 195 spherules in about 0.9 gm of lunar soil. The lunar spherules are expected to come from the cores of the impacts, and have some of the highest velocities of all the ejecta. Thus every gram of lunar soil could contain fragments from many, perhaps hundreds, of different craters.

A relatively recent improvement in radioisotope dating allows the time of the impacts to be measured with remarkable precision. For terrestrial samples younger than 0.1 Gyr, Paul Renne has obtained accuracies better than ± 0.2 million years from single spherules. For the success of this project it is not necessary to duplicate such precision; indeed, accuracies of several million years will give important information on cratering rates. However one of the goals of this pilot project is to see what precision can be obtained with present equipment.

As we shall discuss below, we have received from NASA two grams of lunar soil; we have already separated 195 spherules from 0.9 gm of this material. In this initial phase of the project we are trying to learn as much about the spherules as possible. We are taking scanning electron microscope pictures and electron microprobing all of the spherules recovered from the soil sample. This summer we plan to date the first 18 spherules we have finished analyzing (with the SEM and electron microprobe) to determine the quality of the dates that we can obtain using 40Ar/39Ar and 37Ar/38Ar cosmic ray exposure age dating. In the second phase of this project we will choose approximately 20 spherules to date, based on their dissimilarity from each other. We would like to find 20 spherules from 20 separate craters. Ultimately we may have to reject (as likely from the same crater) any two spherules with similar ages that do not have sufficiently different chemical compositions. From physical appearance and microprobe analysis we will select spherules that are relatively young, different from each other, and originate from impacts rather than volcanic events. Several elements, including titanium, appear to vary significantly across the lunar surface. We hope to be able to identify spherules formed by different impacts by comparing the concentrations of elements (like titanium) in the spherules. We are attempting to distinguish between impact and volcanic spherules in several ways. The absence of schlieren and rock fragments, high volatile concentrations and high Mg/Al ratios are all indicative of volcanic glasses. We have also chosen samples from the maria to de-emphasize the many early craters that represent the heavy bombardment that took place prior to the flooding of the lunar plains. In addition, if we chose spherules with highland composition, we are more likely to find spherules from non-local sources (and therefore we are more likely to find spherules from different craters). Furthermore, highlands spherules found on the maria are likely to be younger than the maria they are emplaced on unless they were transported to their location on the maria without having their radiometric clock reset. Other chemical and geophysical methods for distinguishing recent spherules will be considered.

Preliminary work

We have already made some progress. In May 1993 we applied to NASA for two grams of lunar soil, from Apollo-11 and Apollo-12, in order to separate, analyze, and date spherules. The samples arrived in November 1993. As we had requested, they were samples collected near the surface of the lunar soil, rather than buried deeply. From about 0.4 g of the Apollo-11 sample, we separated 95 spherules and fragments of spherules. From about 0.5 g of the Apollo-12 sample we separated 100 spherules and fragments of spherules. Fifty-two of these spherules (as of April 1995) have been polished and examined with a scanning electron microscope and twenty-three have been chemically analyzed by an electron microprobe operated by the University of California.

In the individual pages for spherules 9 and 25 we show examples of scanning electron microscope and electron microprobe data gathered from each of the spherules. Visible are rock clasts, partially-melted clasts, and the glass matrix. This is a standard procedure that has been perfected for terrestrial spherules by John Donovan and Tim Teague. The spherules that we have measured so far all have statistically-significant chemical differences. Although we can not yet claim that they came from different craters, we are encouraged.

Our samples are low in potassium so an important milestone will be to determine the accuracy that we can obtain. Based on the prior experience Paul Renne has had with dating spherules, we could obtain an accuracy of ± 1 million years on some of the spherules; however even much less accurate results will be interesting. Lunar samples require corrections for spallation and the possible implantation of argon from solar wind. We will measure all argon isotope masses from mass 36 through 40 and plan to use isotope correlation diagrams to help analyze solar wind, spallation, and other trapped components.

The ages of the spherules will give us important new information on lunar cratering rates; in particular, we hope to determine whether the cratering rate has been gradually declining over the last 3 billion years, and whether impact pulses have occurred.

Other questions

We are proposing this pilot study, in part, because there are some questions that are difficult to address until we do experimental measurements. Comet showers that affect the Earth could miss the moon; since the number of impacts will be smaller (the moon has only 25% of the Earth's cross-sectional area). However the record is much better preserved since, except for impacts, the surface of the Moon has been virtually undisturbed for over four billion years. If the comet storms are accompanied by large numbers of small objects (in addition to the 1 km comets) then a record of them could have been left on the moon.

Prior work

The ability to date individual spherules by 40Ar/39Ar dating became available for the first time in the late 1970s, when Huneke at Cal Tech measured five glass spherules. His interest was in finding spherules from a local fire fountaining event, and they were chosen from the lunar soil because they matched other material known the pyroclastic event. Several other papers in which the ages of spherules were measured followed this same pattern. Apparently because of the difficulty of measuring many samples, very special spherules were always chosen. In the late 1980s and early 1990s interest in measuring spherules waned, just as the ability to measure large numbers of individual spherules became practical. We have found no example in the literature of 40Ar/39Ar dating of chemically-dissimilar spherules, those that presumably come from different craters.


E.C. Alexander, M.R. Coscio, J.C. Dragon, K. Saito, 11th Lunar and Planetary Science Conference, (New York, Pergamon Press, p. 1663-1677, 1980).

L. Alvarez, W. Alvarez, F. Asaro, H. Michel, Science 208, 1095-1108 (1980).

W. Alvarez and R.A. Muller, Nature, 311 635-636 (1984).

Basaltic Volcanism Study Project, Basaltic Volcanism on the Terrestrial Planets, (New York, Pergamon Press, 1981).

J. Hills, Astron. J. 86, 1730-40 (1981).

F. Horz, Mass Extinctions and Cosmic Collisions: A Lunar Test, in Lunar Bases and Space Activities of the 21st Century, (Houston, TX, Lunar and Planetary Institute, p. 349-358, 1985).

F. Horz, R. Grieve, G. Heiken, P. Spudis, and A. Binder, Lunar Surface Processes, in Lunar Sourcebook, (Houston, TX, Lunar and Planetary Institute, p. 61-120, 1991).

J.C. Huneke, 9th Lunar and Planetary Science Conference, (New York, Pergamon Press, p. 2345-2362, 1979).

P. Hut et al., Nature 329, 118-126 (1987).

R. A. Muller, Evidence for a Solar Companion Star, LBL-18271 (August 1984), published in The Search for Extraterrestrial Life: Recent Developments (M.D. Papagiannis, editor, D. Riedel Pub. Co., Dordrecht, Holland, p. 233-243, 1985).

D. M. Raup and J.J. Sepkoski, Proc. Nat. Acad. Sci. USA 81, 801-805 (1984).

R. R. Spangler, R. Warasila, J. W. Delano, Proc. of 14th Lunar and Planetary Sci. Conf., part 2, JGR 89 Suppl, p. B487-B497 (1984).

C. Swisher, J.M. Grajales-Nishimura, A. Montanari, S.V. Margolis, P. Claeys, W. Alvarez, P. Renne, E. Cedillo-Pardo, F. Maurrasse, G.H. Curtis, J. Smit and M.O. McWilliams, Science 257, 954 (Aug 14, 1992).