I observed kissing couples in public places (international airports, large railway stations, beaches and parks) in the United States, Germany and Turkey. The head-turning behaviour of each couple was recorded for a single kiss, with only the first being counted in instances of multiple kissing. The following criteria had to be met to qualify: lip contact, face-to-face positioning, no hand-held objects (as these might induce a side preference), and an obvious head-turning direction during kissing. Subjects' ages ranged from about 13–70 years.

Of 124 kissing pairs, 80 (64.5%) turned their heads to the right and 44 (35.5%) turned to the left (Fig. 1). This roughly 2:1 ratio is significantly different from 50% (χ2 = 5.34, d.f. = 1, P < 0.05). As the couples come from a biologically adult age range, this result indicates that adults have a head-turning bias towards the right side, just like embryos and newborns.

Figure 1
figure 1

The number of couples who turn their heads to the right rather than to the left when kissing predominates by almost 2:1 (64.5%: 35.5%; n = 124 couples).

Preferential use of the right foot, ear or eye is also evident as a 2:1 ratio7, which raises the possibility that these biases may be decreed by the observed head-turning preference. The incidence of right-handedness, however, is about 8:1 (ref. 8), so this particular asymmetry cannot be the result of a simple bias attributable to a right-sided head-turning tendency — the genetic origins of this trait may be different8, or cultural factors may have modified an original 2:1 pattern.

It takes two people to kiss (Fig. 2), so the bias of individuals cannot be judged from observations made on pairs. For example, what happens when a right-turner kisses a left-turner? If we assume that kissers with opposite biases could go either way with equal probability, then the individual biases should match those of the couples. For example, if the individual bias is also 2:1 to the right and if couplings are random, then four of nine pairs would be right kissers, one of nine would be left-kissing, and four of nine would be mixed; if choice is random in the last group (that is, two of these four pairs are right-turning and two are left-turning), the result for the nine couples would be a right-turn kissing bias in six of them. This presumed 2:1 distribution for individuals thus predicts the observed 2:1 ratio for couples, indicating that the group bias may also reflect individual asymmetry.

Figure 2
figure 2


The Kiss: the couple in Auguste Rodin's masterpiece are turning their heads to the right to kiss each other.

In birds, a preference for turning the head to the right before hatching induces motor9, visual10 and cognitive11 asymmetries. If a similar effect occurs in humans and an initial asymmetry in the direction of head turning stimulates various functional left–right differences12, then the head-turning bias of the newborn would have to overlap in time with the establishment of adult asymmetries. My finding that turning of the head to the right is preferred throughout adulthood suggests that, by fostering a constant bias towards the right, this mechanism may be able to induce or enhance right-sided asymmetries of perception and action.