A critical part of the exploration for mineral deposits or of quantitative mineral resource assessments is the estimation of how large undiscoveredeposits might be. Typically, this problem is addressed using grade and tonnage models in which a major source of variation in possible sizes is accounted for by the differences in types of deposits (Cox and Singer, 1986; Mosier and Page, 1988; Bliss, 1992). It is clear from studies of petroleum exploration that larger oil fields tend to be found early in the process (Arps and Roberts, 1958). If the same behavior exists in mineral exploration, then tonnage models constructed from local data may be biased estimators of the tonnages of any remaining undiscoveredeposits in the area. Although Singer and Mosier (1981) showed that larger porphyry copper deposits should be found earlier than smaller deposits in a given geologic and exploration environment, there are no definitive studies that we could find which actually test the hypothesis that larger mineral deposits are discovered early in the exploration of a region.

In this paper the hypothesis that larger mineral deposits are discovered early in the exploration of a region is tested by examining the relationship between discovery order and size of known mercury deposits in the California Coast Ranges. We then present a new maximum likelihood approach to modeling the size distribution of undiscovered mineral deposits by examining the sizes of the mercury deposits discovered early in the exploration process.