Recent advances in sociogenomics, namely the ability to characterize molecular pathways in non-model organisms, provide novel opportunities to address questions concerning the evolution of social behavior in an ecological context by comparing taxonomically relevant species [1, 2]. This evo-devo approach to behavior promises to be particularly successful in understanding the most extreme form of social organization, eusociality . Eusocial species exhibit cooperative brood care, overlapping generations of individuals living together, and a division of labor in which one or few individuals monopolize reproduction. The Hymenoptera, including the ants, bees, and wasps, are ideally suited for comparative studies of social behavior as species in this group display a diversity of forms of social organization and represent multiple independent evolutionary origins of eusociality . Research in the behavioral genetics of social insects reveals that molecular pathways associated with behaviors in social species include behavioral genes expressed in solitary species [reviewed in [5–9]]. To date, we have little information on whether genetic mechanisms are conserved across eusocial species or whether derived social phenotypes are a product of convergent evolution driven by alternative molecular pathways.
In a eusocial insect colony, behaviors must be coordinated or synchronized among potentially thousands of individuals at any point in time. In addition to a reproductive division of labor, non-reproductive workers in highly derived eusocial species also display a division of labor among adult workers who perform specialized colony tasks . The success and survival of the colony depends on a variety of tasks; some tasks occur inside the nest, such as brood care and nest construction, while others occur primarily outside the nest, such as nest maintenance, nest defense, and foraging for food.
Task allocation, the coordination of behaviors in eusocial insect colonies, often occurs without any central or hierarchical command . Individuals can display plasticity in their propensity to perform these behaviors throughout their lifetime . In some highly derived eusocial species, individuals display age polyethism, i.e. they progress through sequences of several tasks in a development-dependent manner [1, 4, 12]. A recent review article by Toth and Robinson  highlighted comparisons between honeybee and harvester ants to illustrate potential scenarios explaining the independent evolution of age polyethism in eusocial insects. Age-related polyethism in ants and bees could either arise as a result of convergent evolution of behavioral phenotypes  or as a result of parallel evolution of behaviors via a conserved molecular mechanism shared by a common ancestral phenotype [1, 13].
Efforts to determine the physiological and genetic bases for behavioral task development in non-reproductive eusocial workers have thus far focused on pathways involved in the development of nurse bees and foragers in honeybees [6, 14–18]. One promising candidate pathway involves the role of the circadian clock in regulating age-dependent rhythmicity [14, 16, 19, 20]. Honeybee foragers, which likely benefit from coordinating an internal clock with timing of nectar production and availability or sun-compass navigation, exhibit behavioral rhythmicity and a daily oscillation in the clock gene, period . In contrast, nurse bees, which provide brood care inside the nest, exhibit arrhythmicity in locomotor behavior, and lower overall levels and weaker cyclical expression of period mRNA [14, 16]. The results from honeybees suggest a link between foraging behavior, up-regulation of the period gene, and onset of circadian rhythms [14, 16, 20]. Given that age polyethism is a derived state in eusocial insect colonies [1, 4], we ask whether foraging behavior is associated with the developmental regulation of circadian clock in the harvester ant, a species that evolved eusociality and age-related polyethism independently of bees [4, 21].
The seed-eating harvester ants, genus Pogonomyrmex, are the behavioral model for task allocation in ants [1, 10, 12, 22]. Large colonies of harvester ants (10,000–12,000 workers) inhabit the desert floors of the southwestern United States and the organization of these colonies runs, well, like clockwork. Young harvester ant workers tend to remain deep inside the nest with the single queen and perform tasks related to brood care, while older workers perform nest maintenance and patrolling tasks at the nest entrance and foraging tasks outside of the nest [12, 23]. In P. barbatus, the species best characterized for behavior, foragers have specific behavioral patterns that relate to daily temperatures and environmental conditions . In summer months, they emerge from the nest early in the morning and the majority of foraging occurs during the morning hours until the sun rises and temperatures increases. By late morning, most foraging activity has ceased, although some activity occurs again in the late afternoon. During the winter months, harvester ant colonies show little activity outside of the nest. Like honeybees, harvester ant foraging activity is dependent on light and temperature, but the developmental timeline of workers is vastly different between bees (0–20 days) and ants (months – year). Thus, we are interested in whether the association of foraging behavior and circadian rhythms is conserved across diverse developmental time scales.
We cloned a harvester ant ortholog to the period gene, PoPer, from a closely related species that is easily maintained in the laboratory, Pogonomyrmex occidentalis. We analyzed the expression of PoPer in brains isolated from individual workers in field-collected colonies housed under controlled light and temperature conditions in the laboratory. We show that P. occidentalis workers display a daily oscillation in period mRNA expressed in the brain and that this oscillation is endogenous. We compare the expression patterns of nest workers with foragers and show that differences in daily expression patterns correspond to differences in locomotor activity. In addition, we show that patterns of daily oscillations in period mRNA in foragers differ across seasons.