The concentrations of uranium, yttrium, sodium, iron, zirconium, manganese, calcium, and nickel in 75 mill-pulp samples of uranium deposits in the Salt Wash member of the Morrison formation on the Colorado Plateau have been found, by statistical tests, to be significantly related to the size of the deposits represented by the samples.

The elements mentioned above are related to the formation of the deposits in a variety of ways. Zirconium is an intrinsic element, contained principally in the detrital syngenetic fraction of the host sandstone. Calcium, manganese, and sodium are intrinsic elements contained principally in epigenetic (diagenetic) carbonate in the host sandstone. Uranium, yttrium, and nickel are principally extrinsic elements, introduced into the host sandstone by uranium mineralization or related processes. Somewhat more than half of the iron is probably intrinsic and the remainder is extrinsic.

Three methods can be used to estimate the size of uranium deposits in the Salt Wash, member within broad limits. Method 1 is based on simple linear-regression theory; method 2 is based on multiple-regression theory (long method); and method 3, on multiple-regression theory (short method).

For methods 1 and 2 the estimated log size of each deposit can be computed from tables showing the known concentration of uranium, yttrium, sodium, iron, zirconium, manganese, calcium, and nickel in the deposits, as determined by semiquantitative spectrographic analysis. For method 3 the estimated size or log size can be read directly from a table showing known concentration of uranium and yttrium only.

About 80 percent of the tonnage-size estimates from method 1 will be within a factor of 13 (12-14) of the true sizes. The precision of the size estimates from method 2, the long multiple-regression method, is highly variable. Some estimates from method 2 will be within a factor of 12 of the true size at the 80-percent confidence level; others, within a factor of 40 at the 80-percent confidence level. About 80 percent of the tonnage-size estimates from method 3, the short multiple-regression method, will be within a factor of 15 (13-16) of the true size.

A group of 40 deposits of known size was used to test the theoretical derivation of the confidence intervals given above. It was concluded from the test that the confidence intervals describe the precision of the methods correctly.

The methods for estimating the size of uranium deposits are useful where the ore is poorly exposed or where an independent estimate is desired. The error of the estimates may be quite large, as indicated by the confidence limits given above; but the estimates can be used to, at least, distinguish very large from very small deposits. They also may serve to encourage or discourage further expenditures in the development and exploration of ore bodies. The methods of size estimation can be particularly useful in attempts to appraise or compare groups of deposits or mining districts, inasmuch as the average estimate of size of deposits in a group is more precise than any single estimate.

The methods for estimating size are established only for deposits in the Salt Wash member of the Morrison formation. Tests indicate that the equations calculated for deposits in the Salt Wash fail completely if applied to deposits in other stratigraphic units, such as the Moss Back and Shinarump members of the Chinle formation. A further restriction, not completely evaluated at present, is that semiquantitative spectrographic analyses of mill-pulp samples are required. No tests have been made to determine the precision and accuracy of the methods when other types of samples, such as drill core, are used.