(for constant etching time); blank values, but not necessarily efficiencies, tend to rise with increasing temperature, These results are consistent with the observations of Qaqish and Besant (1976) and Costa- Ribeiro and Labdo (1975). The observed variations among generally comparable exposures are probably due to etch bath temperature fluctuations, the relatively uncontrolled introduction of two new operations (etch bath stirring and ultrasonic rinse), and failure to achieve sampleto-sample consistency in setting the track acceptance thresholds for the image analyzer. Better control over all of these parameters can be achieved without prohibitive cost or effort; however, the practice of treating each piece of film as an individual "counter" with separate but equivalent sample, standard, and background areas reduces the. number of assumptions required to quantify the film data. The sediment sample exposures (Table 2) show good agreement between repeat counts, and indicate that a value of about 5 x 10°74 tracks/mm™ 2 (pCi/g)-! hr-! is a good estimate of film sensitivity to reasonably homogeneous solid sample alpha activity. Although 0.5 mil (12.8 um) Mylar is an optimal energy degrader for surface-plated Pu activity, the self-absorption in a three-dimensional sample reduces or eliminates the need for additional energy degraders. There appears to be little difference between direct contact sensitivity and sensitivity with 0.25 mil (6.4 um) Mylar interposed, so it should be possible to use such a film when it is desirable to contain the sample or to protect the film from direct physical or chemical interaction with the sample. The Enewetak soil samples were intended as a test of the potential of short exposure studies of contaminated environments using the combined sample-standard~background approach to single film analysis. However, the extreme heterogeneity of the samples frustrated a rigorous comparison. The following points may nonetheless be made on the basis of Table 3: (a) the track densities for the standard exposures on the different films were in satisfactory agreement; (b) blanks were reasonably consistent, although apparently somewhat higher than the value of about .06 mm ~2 suggested by Eastman Kodak; (c) the calculated sensitivities were generally comparable in magnitude to the values obtained in Table 2, although the sample heterogeneity makes it doubtful that the published bulk activity values can be directly applied to this type of calculation; and (d) the relative standard deviation provides a crude but useful measure of sample homogeneity. The effectiveness of sample homogenization procedures (e.g., ball-milling), the distribution of environmental activity in relation to subcomponents of a sample, and the implications of inhomogeneous activity distribution in terms of sampling and analytical design and procedures are all subjects which have received less than adequate attention, due in large part to the lack of an appropriate research method. Alpha autoradiography permits addressing these subjects on a variety of levels. The standard deviation of the average track density is one useful index, as is subjective visual observation. However, use of image analyzing systems will permit a much more powerful approach to the question of partitioning 604