STATISTICAL ANALYSIS APPLIED TO THE PROBLEM OF ROCK ORDERING IN THE LITHOSPHERE. Yu.L. Voitekhovsky, Institute of Geology, Komi Science Centre, Ural Division of RAS, Centre de Geostatistique, Ecole Nationale Superieure des Mines de Paris

Indicator functions are used to test spatial correlations between mineral grains of different species in a rock. The biggest range obtained from all the covariograms is regarded to be the radius of minimal mineral aggregates representing a rock texture. These aggregates are defined to be those mineral clusters which build a rock by their superposition on one another. In accordance with statistical theory, translation of such clusters in a rock space is possible with controlled variation. In this sense, the principle of rock ordering looks like intermediate between the strict order and absolute chaos. The developed theory has been applied to gabbronorite from the Fedorova-Pansky Tundra intrusive, the Kola Peninsula.