In a controlled heat study published this year, participants worked moderate-intensity shifts in hot conditions, took an hour-long break in a cool room, then returned to work. Their core temperature hit the occupational safety threshold faster in the afternoon. The break hadn't restored morning-level capacity. The heat was still in them.
A separate analysis of American time-use diaries (2003–2019) found the behavioral trace: on days reaching 100°F, workers logged 36% more non-work activity while at the workplace. The effect was sharpest among hourly workers and women, the people with least control over when they slow down and most scrutiny when they do.
The common adaptation advice is to start earlier, work before peak heat. But a morning hour claimed for outdoor labor is subtracted from sleep, from recovery, from the drive to drop children at school. A body that woke at 4 a.m. carries that deficit into the afternoon the same way it carries unresolved heat. The cool window shrinks, and everything that used to spread across the full day crowds into what remains.
What determines whether a cool hour actually restores the body
Three attributes shape whether a given hour helps or just looks cooler on a thermometer:
Its temperature where you are — 75°F outdoors doesn't mean 75°F inside a poorly insulated apartment or a building that absorbed heat all afternoon.
What preceded it — the body carries prior heat forward. A 2023 review of cooling during passive rest found that in 8 of 50 datasets, workers' core temperature showed no net cooling at all during breaks. A cool room is not the same as a cooled body.
What it displaces — if early morning goes to work, it's no longer available for sleep. If a cool evening hour goes to errands postponed from the heat of the day, it competes with the body's recovery window. Childcare, transit, and school schedules don't shift on the same clock as the temperature.
Sources: Kenny et al., Am. J. Ind. Med. 2026 (core-temperature study, 60 participants, chamber conditions); Econ. & Human Biology 2026 (ATUS diary data, 2003–2019)