Male boreal toads (Bufo boreas) are thought to return to the breeding site every year but, if absent in a particular year, will be more likely to return the following year. Using Pollock's robust design we estimated temporary emigration (the probability a male toad is absent from a breeding site in a given year) at three locations in Colorado, USA: two in Rocky Mountain National Park and one in Chaffee County. We present data that suggest that not all male toads return to the breeding site every year. Our analyses indicate that temporary emigration varies by site and time (for example, from 1992 to 1998, the probability of temporary emigration ranged from 10% to 29% and from 3% to 95% at Lost Lake and Kettle Tarn, respectively). Although the results provide weak evidence that males are more likely to return after a year's hiatus, a general pattern of state-dependent temporary emigration was not supported. We also hypothesized relationships between temporary emigration and a number of weather variables. While some competitive models included weather covariates, imprecise and variable estimates of the effects of these covariates precluded fully defining their impact on temporary emigration.