In a randomised crossover trial of 29 healthy adults with normal glucose tolerance, a single 30-minute bout of moderate exercise was performed once in the morning (07:00-10:00) and once in the evening (16:00-19:00), with continuous glucose monitoring over an 8-day free-living protocol. Evening exercise produced a significantly larger acute drop in blood glucose (P < 0.001), yet mean daily glucose exposure did not differ between conditions. Evening sessions were followed by greater within-day glycaemic variability (SD P = 0.009; CV P = 0.008), and appetite rose after exercise regardless of timing.
The practical message is reassuring: for metabolically healthy people, the best time to exercise is the time they will actually do it. The bigger evening dip is a short-lived shape change, not a gain in total glycaemic control. Larger variability after evening exercise is not clearly harmful in people with normal glucose tolerance, but its meaning is uncertain. The real question is whether timing matters for people with prediabetes or type 2 diabetes, where earlier trials hint at afternoon or evening benefits. This cohort cannot answer that. Single bouts, a short monitoring window and healthy participants also limit how far the findings extend to habitual training over months. This is a preprint that has not been peer reviewed, so the statistics and analysis choices, including the circadian-aligned windows, may be revised. As evidence, it is incremental and confirmatory: it adds resolution to the timing debate rather than overturning it.