were recorded; for this, both an average track density per mm* and a standard deviation (expressed as percent of track density/mm) were calculated. Finally, a sensitivity factor (tracks mm~2 (pci/g)~! hr-!) was calculated for each exposure, based on the estimated bulk activity of the sample. The tracks produced by this sample showed an extremely uniform distribution, with no evidence of particulate clusters. Table 3 gives the results of short-term exposures to soil samples from Runit I., Enewetak Atoll; these samples were selected from among those analyzed for the 1972 Enewetak radiological survey (U.S. Atomic Energy Commission, 1973) on the basis of their known high Pu concentrations. Sample A is 70-5205- 24A, Sample B is 74-5221-24A, and Sample C is 76-5223-24A; all samples had been ball-milled to a powder as part of the original analytical procedures. Exposures were carried out as with the Bravo Crater sediment, except that portions of the film were shielded for blank measurement and for subsequent exposure to a commercial 24lam standard to provide for efficiency comparisons between exposures. Track counts were recorded for each square mm observed in order to calculate both average track density and its percent standard deviation for each exposure. These samples were found to be extremely heterogeneous, with most of the activity concentrated in particles represented by track clusters of various sizes (0.01 to 0.5 mm diameter); this feature is reflected in the high and erratic standard deviations calculated for the sample R values. The sensitivities, based on the published bulk Pu activities of the samples, were calculated as above and then normalized to the lowest observed standard track density value in the set. DISCUSS ION Qaqish and Besant (1976) have demonstrated that, with careful optimiza- tion and control of procedures, both LR-115 and CA 80-15 have detection efficiencies of 98-99%, a linear dependence of etch rate on temperature, and track diameter, which is a function of particle energy. They conclude that it is possible to control combined etching, track identification, and statistical uncertainties at the 1% level. The level of effort required to achieve this, however, tends to negate the track detector's inherent advantages of speed, simplicity, and economy. This study, by contrast, has attempted to address the question of how easily the commercially available cellulose nitrate films may be made to yield useful results. The studies of standard source exposures (see Table 1) show that, with the techniques adopted, efficiencies range from 66 to 91% of the incident alphas for etching conditions close to the manufacturer's recommendations. The blank values and the efficiencies both drop as temperature is reduced 602