The Dufour’s gland is found in all female hymenopterans (bees, ants and wasps) and it is believed to secrete compounds that are utilized in defense by workers and reproduction in queens (Martin et al. 2005). In the worker and queen honey bee, the gland is associated with the sting apparatus and is found in the same position in both castes. The glandular secretion passes through the membrane between the sting lancets. This exit is very narrow and indistinct, and is close to the setosa membrane, a hairy region of cuticle, which surrounds the entire sting bulb and acts as a platform for pheromone release (Lensky et al. 1995, Martin et al. 2005). Previously, it was called the basic or alkaline gland and was thought to be associated with the venom, as it is in ants (Martin et al. 2005). In solitary bees, which build underground nests, the Dufour’s gland produces hydrophobic (water hating) substances used to line the nest and protect eggs against extreme fluctuations in humidity and against microorganisms (Abdalla et al. 2004). Some speculation still surrounds the precise role of the Dufour’s gland in the honey bee.

There is great variability in the size and chemical composition of the Dufour’s gland among the female castes and different aged honey bee workers. The gland of a three-day-old worker bee is tubular and markedly curved near its distal end and less so at the proximal end, unlike other species of bees (Ikenga and Chapman 1989). The Dufour’s gland in queens is approximately 10 times greater in volume than a normal worker gland. The glands of queenless workers are slightly larger than those of queen-right workers, but these differences are not statistically different (Katzav-Gozansky et al. 2000). The queen’s gland is hypertrophied (enlarged due to productivity) and contains up to 20 times more material than that of workers. Among the workers, the amount of secretion found in the glands of queenless egg-laying workers was at least four times higher than that of queen-right nurses or foragers (Katzav-Gozansky et al. 2002). Katzav-Gozansky et al. (2000) demonstrated that the Dufour’s gland, both in workers and queens, is metabolically active throughout the individual’s life.

Chemical analyses of the gland’s secretion revealed that this gland exhibits caste-specific differences (Katzav-Gozansky et al. 1997b). In honey bees, the gland possesses only two classes of compounds, hydrocarbons and esters. The glandular secretions of workers are composed of a series of odd n-alkanes, and the secretions of queens are additionally fortified with wax-type esters. The main esters found in queen glandular secretion are: tetradecyl tetradecanoate, tetradecyl hexadecenoate, hexadecyl tetradecanoate, tetradecyl octadecenoate and hexadecyl hexadecenoate (Katzav-Gozansky et al. 2002). Queens and queenless egg-laying workers possess a bouquet of esters and hydrocarbons, whereas queen-right workers (nurses or foragers) and queenless foragers produce exclusively hydrocarbons (Katzav-Gozansky et al. 2000). While glands of egg-laying workers obtained under queenless conditions produced both hydrocarbons and esters, this was not true for queenless foragers. The reason for these differences between queenless workers is not known, but may be linked to either their age or task. It is possible that some physiological changes that occur in old foragers are irreversible, as was shown in hypopharyngeal glands (Huang and Robinson 1996).

The gland is known to secrete but not produce hydrocarbons (Katzav-Gozansky et al. 1997a) which are used for waterproofing the cuticle and possibly involved in chemical communication (Abdalla and Cruz-Landim 2001). In both queen and worker glands, no in vitro incorporation of radioactivity into hydrocarbons was observed. This suggests that the copious amounts of hydrocarbons that are present in the gland result from sequestration rather than de novo (from the beginning) synthesis (Katzav-Gozansky et al. 2000). In addition to the hydrocarbons, the queen’s gland secretes a series of esters (Katzav-Gozansky et al. 1997b, Martin et al. 2002) which function as part of the multi-sourced queen signal (Katzav-Gozansky et al. 2003), while the workers’ Dufour’s glands secrete eicosenol (Martin et al. 2004) a low volatile alarm pheromone (Pickett et al. 1982). This may explain why workers are attracted by virgin queen Dufour’s gland extracts and repelled by worker Dufour’s gland extracts (Abdalla and Cruz-Landim 2001).

Both in vivo (within the organism) and in vitro (done in a cell free system) studies have demonstrated that ester biosynthesis in the Dufour’s gland is not a caste-fixed phenomenon. Queenless workers that start to develop ovaries also biosynthesize the queen-type esters. Moreover, glands from queen-right nurse bees incubated in vitro also produce these esters, after a certain delay (Katzav-Gozansky et al. 1997b).

The queen’s Dufour’s gland secretion constitutes a component of the multi-sourced queen pheromone signal. It is attractive to workers, which form a retinue around the scented source in concentrations as low as 1/8 queen-equivalents. Bioassays reveal the ester fraction, not the hydrocarbons, to be the active constituents (Katzav-Gozansky et al. 2003). The function of the esters was corroborated by assays with the synthetic queen-esters mixture, which successfully mimicked the queen’s secretion. When the queen’s ester mixture was applied either on a glass slide or on another worker, a retinue formed around the 'surrogate queen' (Katzav-Gozansky et al. 2001). As predicted from the queen-like secretion exhibited by egg-laying workers, their glandular secretion was also attractive to nestmates, however, to a lesser degree than that of the queen; while that of non-egg-laying workers was totally inactive (Katzav-Gozansky et al. 2003). The amount of secretion found in the glands of egg-laying workers is about 1/7-1/8 queen equivalents. Worker-attraction towards the glandular secretion of virgin queens was also demonstrated by Abdalla and Cruz-Landim (2001). Occasionally, under queenless conditions some workers attract a small but recognizable retinue, presumably by exuding some components of the queen pheromone. Such workers are designated as false queens (Crewe and Velthuis 1980). Since Dufour’s esters seem to differ significantly between queen-right workers and egg-laying workers, Dufour’s gland secretion may be involved in marking these workers as false queens.

Katzav-Gozansky et al. (2000) demonstrated that the synthesis of the gland’s secretion is regulated. In the presence of a queen, ester production is inhibited. Once the queen is removed, the physiologically unconstrained gland starts to biosynthesize the queen-specific esters after a certain lag-time needed for the buildup of precursors and the enzymatic machinery. Since there seems to be a link between ovarian development and the occurrence of the queen specific esters in the gland, it is possibly a marker of queen fecundity, a fundamental parameter affecting colony fitness.

Abdalla and Da Cruz-Landim (2001) completed a series of experiments to assess possible roles of the honey bee Dufour’s gland secretion. Bioassays with extracts of queen and worker glands from two colonies were made under artificial conditions, in which nestmate and non-nestmate forager workers were tested. The results demonstrated that forager workers display behavioral responses when exposed to Dufour’s gland extracts of nestmates, but remain indifferent when exposed to non-nestmate extracts. They were also able to demonstrate that forager workers are attracted by virgin queen gland extracts and repelled by forager worker extracts. These data demonstrate that the Dufour’s gland secretion is colony- and caste- specific. The attractant remains in the nest. The repellent effect of forager worker extracts is interpreted as an alarm-like pheromone. The attractant effect of virgin queen extracts could be useful in the swarming process to attract scout bees to the eventual founder virgin queen or to attract old foragers to the virgin queen that remains in the nest.

Ratnieks (1993, 1995) suggested that Dufour’s gland secretion serves as an egg-marking pheromone helping workers to discriminate between queen-laid and worker-laid eggs. Several features of Dufour’s gland in honey bees support its possible function as the source of an egg-discriminating pheromone. This abdominal gland opens into the dorsal vaginal wall (Billen 1987) and its secretion may therefore be controllably applied onto the egg before deposition. Katzav-Gozansky et al. (2001) extensively evaluated the role of Dufour’s gland secretion as a egg-discriminator pheromone and was able to refute this hypothesis. Workers honey bees were exposed to two combs simultaneously, one containing queen- or worker-laid eggs and the second containing treated or untreated worker-laid eggs. Treatments included extracts of Dufour’s gland secretion as well as the synthetic esters that were identified in the secretion. Policing was clearly detected both in queenright and queenless colonies by the swift removal of worker, but not of queen eggs. However, neither the glandular secretion nor its synthetic ester constituents were able to protect worker-born-eggs from policing. Treated worker eggs were removed significantly faster than queen eggs, and at the same rate as non-treated worker eggs. Chemical analyses of the queen abdominal tips revealed the presence of Dufour’s esters, indicating that the glandular secretion oozes out and spreads over the cuticle around the genital chamber. However, contamination while ovipositing may also explain the minute amounts of these esters that were detected on the egg surface.