Atmospheric density fluctuations which affect 'seeing' have been studied to examine details of their characteristics and ways of at least partial compensation. Work involved two 60 ft erectable masts, each with five fast temperature sensors, and also measuring equipment set up on a pole at a height of 160 feet on the top of a rather narrow 160 ft building. Wind speed measuring equipment was also installed. Extensive data has been taken as a function of height, horizontal separation, wind speed, and time of day. It has been found that the Kolmogorov-Taylor approximation has validity, but that there are important deviations from it. The power law of fluctuations as a function of frequency devices somewhat from this KT model. The Taylor approximation for movement of fluctuations by wind is a reasonable approximation for modest distances but the time length of its validity is no longer than about 13 seconds. Major turbulences are shown to have similar vertical and horizontal sizes, and to vary in size approximately linearly with wind speed. Correction of approximately one half of the path length fluctuations of light passing through the earth's atmosphere appear to be possible from measurements monitoring fluctuations near the earth's surface.