Do number counts tell us anything about the distribution of distant sources

The C − method for finding the spatial distribution of radio sources requires knowledge of both the flux and redshift for each source, while source counts yield little information on the detailed distribution but do not require time consuming redshift determinations. The possibility of obtaining an approximate density evolution using a log N –log S related method with redshift information from a subsample is examined in some artificial cases with the hope that something better than a (1 + z ) n type law may be found. It appears that for realistic models there is little information on the evolutionary behaviour at large redshifts, and a rough criterion for when this is true is given.