The durations of animals' brief vocalizations provide conspecifics with imp
ortant recognition cues. In the present experiments, zebra finches and huma
ns (trained musicians) were rewarded for responding after S+ (standard) aud
itory signals from 56 to 663 ms and not for responding after shorter or lon
ger S- (comparison) durations from 10 to 3684 ms. With either a single stan
dard (Experiment 1) or multiple standards (Experiment 2), both zebra finche
s and humans timed brief signals to about the same level of accuracy. The r
esults were in qualitative agreement with predictions from scalar timing th
eory and its connectionist implementation in both experiments. The connecti
onist model provides a good quantitative account of temporal gradients with
a single standard (Experiment 1) but not with multiple standards (Experime
nt 2).