The no observed effect concentration (NOEC) is currently a major summary st
atistic in ecotoxicity testing. Its use is heavily criticized, partly becau
se it is a poor estimator of "safe" chemical concentrations. In this short
communication, we review the limited information available on the percentag
e effect that corresponds with the NOEC, a value designated the ECNOEC, and
calculate ECNOEC values for fish growth data. The average ECNOEC for diffe
rent test protocols was found to vary widely, with values for chronic studi
es ranging from 10 to 34%. Individual results varied even more widely. This
analysis suggests that the NOEC is neither a consistent summary statistic
nor an indicator of safe concentrations of toxic chemicals.