For a six week period in 1952, a network of twenty high volume air samplers was used to measure the concentration of particulate pollution in Detroit and Windsor. Using a graphical multiple regression technique, regression curves are obtained which relate particulate pollution directly with temperature anomaly, and 24 hour temperature change, and inversely with wind speed, and duration of precipitation.
The relationships were modified for application to independent data covering each of the four seasons. The results appeared useful, although rigorous claims about significance cannot be made where graphical methods are employed.
* Paper presented at 131st National Meeting of the American Meteorological Society, Columbus, September, 1954. Published by permission of the Controller of Meteorological Services, Department of Transport, Canada.
** Present Affiliation—University of Michigan, Engineering Research Institute.
