The differential allocation of information processing resources over time, here termed "temporal attention," may be achieved by relatively automatic "exogenous" or controlled "endogenous" mechanisms. Over 100 years of research has confounded these theoretically distinct dimensions of temporal attention. The current report seeks to ameliorate this oversight by novel application of 2 experimental methodologies. A scheme imported from the animal learning literature (Rescorla's "truly random control" procedure) was used to eliminate any temporal contingency between signals and targets. An auditory stimulus imported from the psychophysical literature (correlated vs. uncorrelated noise) was used to provide a salient signal that entailed no local or global change in intensity. Purely endogenous temporal attention (generated by a reliable signal-target contingency in the absence of a change in intensity) is characterized by robust improvements in speed and accuracy of responding. Purely exogenous temporal attention (generated by an intensity increase in the absence of contingency) is characterized by a brief period of faster responding. When exogenous temporal attention is elicited in the context of endogenous temporal attention, the decrease in response time that follows an intense signal is accompanied by a decrease in response accuracy. (Contains 1 table, 5 figures, and 17 footnotes.)