At any given time, temperatures could differ by more than 15 °C. There is a relationship between the reduction in Tabtop and diving behaviour, with slight decreases during periods of shallow diving (c) and larger decreases during deep diving bouts (a, b, d). Although Tabtop may be several degrees lower than Tstom (1 is for Tstom, a is for Tabtop) the source of cooling could still be the ingested food because large thermal differences exist in the stomach (see text). Note that, during the resting phase preceding shallow dives, Tstom is still higher than Tabtop, which might indicate local heat production during digestion. When temperature data are available for the bottom part of the abdomen (when Tabbot is below 22.7 °C) as shown here in two diving bouts, these events roughly parallel Tabtop, with a slight delay, even during the rewarming phase. B, Representative changes in body temperature of a king penguin at sea 5 h after leaving the chick (bird K departed on 18 February at 06:49 h for 8.0 d of foraging). At this stage, the stomach is empty and any ingestion of food or water is detected by a typical precipitous fall in Tstom (ingestion events 1 to 7; ref. 13). Although there was no ingestion during the 8 h preceding the first deep-diving bout, Tabtop progressively decreased by more than 1 °C, suggesting that this decline is due to diving per se. The further decrease of Tabtop during a deep diving bout (event b) could be partly due to ingestion (events 4–6). However, the transient drop in both Tstom and Tabtop preceding the rewarming phase (event c), which routinely occurs during a surface phase, argues for a mechanism independent of feeding.