The evolution of social dominance is thought to require phenotypic traits that convey honest information regarding an individual's status1. These ‘badges of status’ could be communicated by any means, but only avian plumage has been examined and provides equivocal support for the hypothesis2. By chemical manipulation of the odour of socially naive male cockroaches (Nauphoeta cinerea) we show that the sex pheromone of males determines rather than reflects status. The individual chemical constituents of the pheromone have contrasting effects on status, but the evolution of the blend that confers dominance may be constrained by contrasting selection pressures resulting from female mate choice.

The male-produced pheromone of N. cinerea that attracts females is composed of three compounds: 3-hydroxy-2-butanone, 2-methylthiazolidine and 4-ethyl-2-methoxyphenol3,4. The quantities of these are heritable and co-vary with male status5,6. Female mate choice is based on pheromonal differences among males7,8. We manipulated the quantities of these three components individually and in combinations to investigate their effect on interactions between males.

Adult males were isolated on emergence, and a filter-paper circle (6 mm diameter) was glued to the pronotum. The pheromonal profile of males was manipulated by adding individual pheromone components: a combination of the two developmentally and genetically correlated components5,6, a blend of all three components, or acetone (control) onto the filter paper. The quantity applied was based on the mean quantity of pheromone isolated from the sternal pheromone glands of dominant males5,6. After a 15-min recovery period, a control male and a manipulated male were placed in an open arena (17×12×6.5 cm) for observations. The social status of each male was determined on the basis of patterns of agonistic behaviour5,6,7,8,9.

Status depended on the specific pheromone manipulation (Fig. 1). When 3-hydroxy-2-butanone alone was increased in quantity, males were likely to be treated and to react as subordinates. When either 2-methylthiazolidine or 4-ethyl-2-methoxyphenol was increased, there was an increased likelihood of being dominant. This effect was even greater when the quantities of both compounds were increased together. The influence of 3-hydroxy-2-butanone is cancelled by the combined effects of 2-methylthiazolidine and 4-ethyl-2-methoxyphenol; increasing all three compounds simultaneously did not affect status. Thus, the compounds have individual, additive and contrasting effects on status.

Figure 1: Pairs of males were treated, one with the compound (or mixture) and one with acetone (10 μl placed on a filter paper on the pronotum).
figure 1

The experiment was performed blind, with respect to which chemicals were being applied. Males with increased 3-hydroxy-2-butanone were more likely to be treated and act as subordinate (binomial test, P=0.005). In contrast, males with increased 2-methylthiazolidine (P=0.054) or 4-ethyl-2-methoxyphenol (P=0.115) were slightly more likely to be treated as and to act dominant. This effect was stronger for males treated with increased 2-methylthiazolidine and 4-ethyl-2-methoxyphenol (P=0.005). Increasing the amount of all three compounds in ratios corresponding to the individual manipulations had no effect on status (P=0.345).

Full size image

Our manipulations did not result in abnormal behaviour. The direction and intensity of interactions were stable within a one-minute observation period. Manipulated males behaved the same as control males of the same status and there were no obvious differences in the way that males were treated. There were no significant differences in the frequency of aggressive acts by males scored as dominant or the frequency of submissive acts by subordinate males. Highly aggressive acts associated with uncertain initial status, such as grappling, biting and kicking, were no more frequentin those interactions resulting in the manipulated male being scored as dominantas in the interactions where the control male was scored as dominant.

A dual function allows badges of status to be ‘honest’ signals1,10. The sex pheromone of N. cinerea may have a dual function, acting in both male-male competition and female mate choice5,6,7,8,9. Male-male competition influences mating success in N. cinerea5,9 so selection could be expected to increase the quantities of both 2-methylthiazolidine and 4-ethyl-2-methoxyphenol and decrease the quantity of 3-hydroxy-2-butanone. There is no genetic constraint to this pattern of evolution: the quantities of 2-methylthiazolidine and 4-ethyl-2-methoxyphenol are genetically correlated but the quantity of 3-hydroxy-2-butanone is genetically independent of these and is presumably free to evolve6. However, female N. cinerea also use all three components of the sex pheromone to discriminate among males5,6,10. Sex differences in either the signal perception or the functions of the signal components can thus maintain variations in this pheromone and its multicomponent structure.