Separate research streams have identified synchrony and arousal as two factors that might contribute to the effects of human rituals on social cohesion and cooperation. But no research has manipulated these variables in the field to investigate their causal - and potentially interactive - effects on prosocial behaviour. Across four experimental sessions involving large samples of strangers, we manipulated the synchronous and physiologically arousing affordances of a group marching task within a sports stadium. We observed participants' subsequent movement, grouping, and cooperation via a camera hidden in the stadium's roof. Synchrony and arousal both showed main effects, predicting larger groups, tighter clustering, and more cooperative behaviour in a free-rider dilemma. Synchrony and arousal also interacted on measures of clustering and cooperation such that synchrony only encouraged closer clustering-and encouraged greater cooperation-when paired with physiological arousal. The research helps us understand why synchrony and arousal often co-occur in rituals around the world. It also represents the first use of real-time spatial tracking as a precise and naturalistic method of simulating collective rituals.