Recent reports on patterns and occurrence of torpor and other natural hypothermic states in birds have prompted a revision of many longstanding opinions. For instance, a larger assortment of birds covering a wider range of body mass than previously recognized uses energy-conserving mechanisms in the face of abundant food supplies. Thus, although acute energetic stress triggers the occurrence of hypometabolic states in many birds, energy deficits can no longer be regarded as the sole stimulus for avian torpor. Additionally, the phenology of this phenomenon (phases, duration, depth) shows more interspecific variation than previously appreciated, and traditional concepts of the phases of torpor are not always adequate. Hence, inclusive definitions of torpor based on physiological and/or behavioral criteria have become more difficult to formulate. However, average nighttime body temperature, which is highly consistent throughout the class Aves (38.54 degrees +/- 0.96 degrees C, n = 202), provides a convenient reference for detecting natural hypothermic states. Of the putative ecological factors associated with avian torpor, food specialization seems a prime candidate as an ultimate factor in the occurrence of this state in birds. With few exceptions, all of these animals capable of torpor are either frugivorous, nectarivorous, or insectivorous, suggesting a correlation between thermoregulatory pattern and predictability of food supply. To date, no clear answer exists as to whether the variety of thermoregulatory patterns evident in birds involves discrete mechanisms or merely steps in a physiological continuum. However, I suggest that the occurrence of differences in torpor patterns among closely related species (e.g., within families) favors the latter interpretation.