Implications are drawn for probable issues that might arise in any discussion of the desirability or the constitutionality of capital punishment. In particular, the issues raised in Miller and Cockriell (1976) are interpreted in light of the empirical evidence. The conflicting evidence on deterrence manifested in the Canadian time series as presented in papers by Avio and Layson is discussed. It is argued that the extension of the time-series past 1960 requires statistical allowance for structural changes, and that these changes, and not the reduction in execution rates, explain the post-1960 variation in homicide rates. Bias arising from measurement errors in the use or ratio risk variables also is commented upon. The discussion on arbitrariness and discrimination reviews research on capital case decisions made by Cabinet. Cabinet did not make its decisions randomly, but it did rely upon certain unprincipled standards. The disparities in execution risk for different ethnic and racial groups are noted. The constitutionality of capital punishment is considered in light of this empirical evidence. Comparisons are drawn with the statistical evidence presented in the recent McClesky case (U.S.). It is argued that Cabinet's erroneous reliance on a deterrent effect in making its commutation decisions may provide legitimate grounds to test the constitutionality of capital punishment, even if the courts, following Miller, determine that retribution is a sufficient social value to preserve the constitutionality of capital punishment per se. (Author abstract)