Does infall end before the class I stage?

We have observed HCO+ J = 3-2 toward 16 class I sources and 18 class 0 sour ces, many of which were selected from Mardones et al.'s recent observations . Eight sources have profiles significantly skewed to the blue relative to optically thin lines. We suggest six sources as new infall candidates. We f ind an equal "blue excess" among class 0 and class I sources after combinin g this sample with that of Gregersen et al. We used a Monte Carlo code to s imulate the temporal evolution of line profiles of optically thick lines of HCO+, CS, and H2CO in a collapsing cloud and found that HCO+ had the stron gest asymmetry at late times. If a blue-peaked line profile implies infall, then the dividing line between the two classes does not trace the end of t he infall stage.