Five sampling methods were used for sampling lady beetles in grain sor
ghum fields in the Texas High Plains. Removal sampling provided accura
te estimates of absolute population for adult coccinellids but not for
larvae. Sweep net sampling was not correlated with absolute populatio
n for adults or larvae. Timed counts of coccinellids were correlated w
ith absolute population for adults, but results for larvae were inconc
lusive. A regression model was developed to convert relative estimates
of adult populations obtained by limed counts to estimates of absolut
e population. The model accounted for 80% of the variability in the re
lationship between the relative density estimated by timed counts and
absolute density estimated by removal sampling.