Police reports of Index crimes, arrests, and population estimates for 1960 through 1977 for 49 U.S. cities are used to compare trends in arrest rates, crime rates, and the arrests per crime (A/C) ratio - a measure sometimes correlated with crime rates to assess the deterrent effect of arrest. Since the correlation of the A/C ratio with the crime rate can produce tautological results in a trend analysis, comparisons of trends for crime rates and arrest rates for specific cities are examined. The results do not support a deterrence hypothesis, but indicate that the A/C ratios for some of the nation's largest cities are unstable over time and sensitive to changes in reporting practices. (Publisher abstract)