ing effect. The counterflow and drift are verified and quantitated to some degree by wind measurements at the different levels on the KMOX-TV tower. They are reported in Volume V, Meteorology and Topography. The results of the correlation between meteorological stability classes and the AISI sampler data for each site are given in Table 24. At 6 of the 12 sites the results have the same pattern as the high-volume sampler results, a decrease in suspended particulates from class 2 to class 4 and an increase for class 5. Four of the six sites not having this pattern had an increase from class 2 to 3, but a decrease from class 3 to 4, and an increase from class 4 to 5. The other two sites showed increases from class 3 to 4. These results indicate that although wind speed is the more influential factor, thermal mixing has a greater influence on particulates collected by the AISI sampler than on that collected by the high-volume air sampler. The difference reflects the influence of the smaller particles collected by the AISI sampler. The class 1 condition did not occur during February 1964. 5,6,7 slightly stable, stable, and extremely stable. b = Feb 1964 only. c Jan and Feb 1964 only. SULFATION 2 Lead peroxide candles were located at 42 sites to measure sulfation rates. The results are reported in milligrams of sulfur trioxide per 100 square centimeters per day and are given in Tables 25 to 30. The results ranged from 0.10 to 8.86 mg 2 S03/100 cm3/day for the period March 1963 through February 1964. The arithmetic mean for all sites during this period was 1.228 mg S03/100 cm/day, and the geometric mean was 0.886 mg S03/100 cm /day. Site 433-565, which is a low-populationdensity residential-to-rural site, had the lowest yearly geometric mean of 0.220 mg S03/100 cm3/day, and site 472-680, which is heavily industrialized, had the highest of 5.264 mg S03/100 cm3/day. 2 2 2 The monthly and seasonal trends are shown in Figure 86. With the exception of the increase in July and August, these trends follow those found in Nashville? The July and August anomaly was examined by plotting individual June-to-July increases for each station on a map. This was done in two ways, with both showing essentially the same results. The first was based on the hypothesis that the pollutant would be uniformly distributed over the area and that it would cause an increase 3' of at least 0.15 mg S03/100 cm2/day. The second assumed a pollutant distribution that would cause a 20 percent or more increase from the June value. The map, Figure |