The treatment of radiation in contemporary climate models is considera
bly more complex than that used by Arrhenius. While this offers many m
ore possibilities for investigations than were available a century ago
, the very complexity of the models now in use demands that they be su
bjected to wide-ranging evaluation. This paper discusses some of the w
ays in which the radiation codes and models may now be compared with o
bservations and some of the issues which those comparisons have raised
. The opportunities for studies of the clear-sky greenhouse effect off
ered by data from the re-analysis project performed at the European Ce
ntre for Medium-range Weather Forecasts are highlighted, and some prel
iminary results from comparisons with the Hadley Centre Climate Model
are shown.