Pictures appeared one at a time for five seconds each: eight spiders, eight flowers, shuffled together. Every spider had been matched to a flower for colour, for size on screen, and for brightness. Pupils shrink in bright light too, so the researchers levelled brightness across every image before measuring anything else. Sample size, sixteen infants per experiment, was fixed in advance by a power calculation, before any data came in. Half the babies met the spiders first and half met the snakes, so plain fatigue cannot explain the pattern.