Abstract
We have used theG-mode central method to classify sets of major oxide chemical data of lunar rocks (163 averages) and lunar glasses (921 separate analyses). These data were selected from the Lunar Data Base using the following criteria: (1) the amount of SiO2, Al2O3, TiO2, FeO, MgO, CaO, Na2O, and K2O were measured by the same group of investigators and (2) the sum of these 8 major oxides was in the range 97.0–103.0 wt.% for the rocks and 99.0–101.0 for the glasses.
TheG-mode central method attempts to recognize homogeneous sets or groups of samples within a raw data matrix. The original multivariate distribution is reduced to a set ofG values which follow a ‘quasi-Gaussian univariate’ distribution. Each of the homogenous groups consists of those samples that can be described by a specific normal distribution of the computedG values. We have followed previous suggestions and have not yet experimented with the effects of variations in estimates of precision, the nature of the distribution(s), or the influence of percentage formation on the recognition of homogenous groups.
Fifteen groups of lunar rocks have been recognized but three of these groups are very small and certainly not homogeneous. All eight variables contribute to the recognition of the 12 retained groups (151 averages) with TiO2−Al2O3−FeO being the most effective ternary subset for the recognition and definition of these groups. A combination ofQ-mode cluster analysis (using all 8 major oxides and cos θ as the measure of similarity) and spatial position in the TiO2−Al2O3−FeO ternary allowed recognition of five families of the 151 retained lunar rock averages. Oxide wt.% and cation normative mean vectors are given for each of the 12 groups. We have assigned names to each of the families on the basis of comparisons with published information but these names are to be considered as descriptive andnot genetic.
A total of 8 families and 16 groups of lunar glasses have been recognized and vector means for the oxide wt.% and cation percentage normative components are given along with the number of samples in each group and the percentage of the total number of retained samples (885) accounted for by each group. Again, names used to describe each group should not be considered as having genetic significance.
Article PDF
Similar content being viewed by others
Avoid common mistakes on your manuscript.
References
Butler, J. C.: 1975, inOrigins of Mare Basalts and Their Implications for Lunar Evolution, Lunar Science Institute, Houston, Texas, U.S.A., pp. 15–19.
Butler, J. C.: 1977,Phys. Earth Planet. Int 15, 275–286.
Butler, J. C.: 1978,Math. Geol. (in press).
Chayes, F.: 1971,Ratio Correlation, University of Chicago Press, Chicago, Illinois, U.S.A.
Coradini, A., Fulchignoni, M., and Gavrishin, A. I.: 1976,The Moon 16, 175–190.
Coradini, A., Fulchignoni, M., Fanucci, O., and Gavrishin, A. I.: 1977,Computers and Geosciences 3, 85–105.
Coradini A., Gavrishin, A. I., and Bianchi, R.: (in preparation).
Glass, B. P.: 1976,Proc. 7th Lunar Sci. Conf. 679–693.
Kork, J. O.: 1977,Math. Geol. 9, 543–563.
Lindqvist, L.: 1976,Computers and Geosciences 1, 129–146.
Reid, A. M., Warner, J., Ridley, W. I., Johnston, D. A., Harmon, R. S., Jakeš, P., and Brown, R. W.: 1972,Proc. 3rd Lunar Sci. Conf. Geochim. Cosmochim. Acta 1, 363–378.
Taylor, S. R.: 1975,Lunar Science, A Post-Apollo View, Pergamon Press, New York, U.S.A. 372 pp.
Vamman, D. T. and Papike, J.: 1977,Proc. 8th Lunar Sci. Conf. 3161–3193.
Warner, J. L.: 1976, NASA, JSC, Houston, Texas, U.S.A.
Author information
Authors and Affiliations
Rights and permissions
About this article
Cite this article
Bianchi, R., Butler, J.C., Coradini, A. et al. A classification of lunar rock and glass samples using theG-mode central method. The Moon and the Planets 22, 305–322 (1980). https://doi.org/10.1007/BF01259287
Received:
Issue Date:
DOI: https://doi.org/10.1007/BF01259287