I compared the effects on plasma corticosterone levels of two methods of collecting blood samples during standardized capture and handling stress protocols. In one method, individual nestling American Kestrels (Falco sparverius) were bled at three time periods: when initially removed from the nest, and 15 and 30 min later. In the other method, siblings removed from a nest were bled once each, either at the time of removal, or 15 or 30 min later. I found that there was no difference between the two groups in plasma corticosterone levels at the first sampling period, but 15 and 30 min after capture the singly-bled birds had significantly higher plasma corticosterone levels than the multiply-bled nestlings. The results suggest that data from multiply-bled birds underestimate actual circulating hormone levels. The underlying mechanism for this phenomenon is unknown, although it may involve hemodilution.