I conducted a vegetation removal experiment using American Goldfinch (Carduelis tristis) nests to test the hypothesis that predation rates vary with concealment in old field habitats in eastcentral Illinois. Daily predation rates were 0.05 for manipulated nests and 0.04 for control nests. Although manipulated nests were much less concealed than control nests, the probability of predation did not differ significantly between treatments or years. Logistic regression models indicated that nests initiated earlier in the breeding season had a greater probability of predation than nests initiated later in the breeding season. These results indicate that time of breeding season may be more important than concealment in explaining probability of predation of American Goldfinch nests in this old field system.