Effects on retinue and rescue behavior
Upon examining the queen in an ant colony, one notices that she is surrounded by “retinue” workers that often lick her cuticle or otherwise groom her, a scenario commonly observed in other social insects such as honeybees and termites (Wilson 1971; Bortolotti and Costa 2014). The most immediate function of the queen pheromone is attraction, enabling workers to locate the queen, so as to care for or rescue her. In a way, the queen acts as the “gravity center” that holds the colony together. This notion is evident in the rock ant Temnothorax rugatulus, a frequently emigrating species; when a colony was experimentally split into two nests under equal conditions, all members almost always reunited with the queen in one of the nests (Doering and Pratt 2016).
Early investigations pointed to the existence of chemical pheromones by showing that queens were attractive to their workers, and that this attraction capacity could be transferred in chemical extracts. Pheidole workers would adopt corpses of Lasius queens treated with Pheidole queen extracts (Stumper 1956). Queenless workers of Myrmica will generally accept queens of a closely related species (Brian 1986a, 1988a, b), which is likely due to the similarity of the chemical makeup of queen recognition signals among recently diverged taxa, a feature that ultimately may pave the way for the rise of social parasites (Lenoir et al. 2001). The chemical basis of retinue behavior pheromones was verified from experiments showing the same effect by queen corpses or queen cuticular extracts in a wide range of taxa (Watkins and Cole 1966; Jouvenaz et al. 1974; Fowler and Roberts 1982; Hölldobler and Wilson 1983).
Fire ant workers (S. invicta) exhibit emphatic and swift responses to a queen exposed outside of the colony: workers will (1) quickly be attracted to her, (2) cluster around her, (3) move brood items to or around her, (4) form a pheromonal trail that the queen can follow back to the nest, and/or (5) pull the queen towards the nest, should she not move voluntarily (Glancey et al. 1983). Workers exhibit the same series of stereotyped behaviors toward a paper dummy dosed with reproductive queen hexane extract as they do toward live queens or fresh queen corpses, i.e., collectively retrieving the treated dummy into the nest and keeping it there for hours (Trible and Ross 2016; Zeng et al. 2022). Additionally, queenless workers infrequently exhibit vibrant body shaking upon retrieving a queen (Zeng, personal observation), which resembles the jerking response to a queen or king in termites (Funaro et al. 2018, 2019).
A general correlation between the intensity of attraction and the fecundity or weight of a queen was noted by many studies (Sommer and Hölldobler 1995; Hannonen et al. 2002). Such a trend is pertinent to other queen pheromone effects in the discussion that follows, and is most parsimoniously explained by a higher level of pheromone production occurring in more fertile queens (Fletcher and Blum 1983a). This collective, accurate assessment of the fertility condition of an individual is fundamental to the signaling function of queen pheromones (Keller and Nonacs 1993).
Effects on colony maintenance
In many ant species, a newly mated queen performs all necessary tasks to initiate the growth of the colony by herself. She prepares a nest site (e.g., digs a burrow in the soil) and raises the first cohort of workers from the eggs she then lays. But when worker adults emerge, the queen stops engaging in colony maintenance tasks and transitions to acting strictly as an egg layer (Cassill et al. 2002; Majidifar et al. 2022). The continued presence of the queen not only ensures the steady production of fertilized eggs and, subsequently, additional brood and adults, but a few studies suggest that the queen also boosts the level of worker activity in brood care and acts to maintain the cohesion of the colony. In other words, the queen (or queen pheromones) may function as a “catalyst” to boost colonial development and maintenance, on top of being a “gravity center” to maintain colony cohesion, as shown in the studies below.
In broodless colonies of Cataglyphis cursor, overall worker activities were reduced after queen removal (Berton et al. 1992). Queenright workers of Myrmica sp. and Manica sp. antennated the brood more often and stayed longer with the brood than queenless workers (Vienne et al. 1998). In contrast, queenless workers tend to leave the nest, interacting more often with adult workers instead of the brood. Similarly, in Atta sexdens, workers departed more frequently from the nest, exhibited higher mortality and lowered refuse accumulation, but showed no change in foraging efficiency, when the queen was removed (Della Lucia et al. 2003; Sousa-Souto and Souza 2006). In Temnothorax curvispinosus, queenright sub-colonies outperformed queenless counterparts in various task efficiencies and were more resistant to fungal pathogens (Keiser et al. 2018).
These studies showed that the presence of the queen induced a higher level of brood care, stronger cohesion, and better overall performance of the colony. These results are consistent with findings in honeybees where queen mandibular pheromones are shown to stimulate a wide range of worker task performance, including foraging, defense, combing building, and brood rearing (Bortolotti and Costa 2014). However, unlike honey bees, these effects in ants remained to be associated with a chemical basis (Box 1, Q1).
At the proximate level, these shifts in worker behavior may be explained as secondary effects of the attractiveness of the queen causing, for instance, a more structured spatial distribution of the brood and, as a result, more efficient brood care. At the ultimate level, the behavioral shift of workers may be associated with the alignment of their evolutionary interests with the queen’s from a kin selection perspective (Hamilton 1964; Foster et al. 2006).
In hymenopteran societies, workers gain a higher inclusive fitness by rearing sister nestmates (related by 0.75) rather than producing their own sons (related by 0.5) due to the relatedness asymmetry stemming from haplodiploidy and other ecological factors from colonial living. Workers thus “work for the colony”, but only when a queen is present to dominate reproduction. When the kinship structure changes in the absence of the queen, individual workers are expected to prioritize their own fitness by reproducing instead of performing colony maintenance duties (Bourke 1988; Wenseleers et al. 2020b). This argument may apply to other effects discussed below, such that adult workers and gynes behave to maximize their inclusive fitness as they stay informed about the presence of the queen through queen pheromones.
Effects on nestmate discrimination
Similar to other eusocial insects, the nest of an ant colony comprises the physical structure, food storage, and brood items. It requires substantial investment to build, and is subject to stealing and raiding, often from sympatric conspecific colonies (Holldobler and Michener 1980; Tschinkel 1992; Sturgis and Gordon 2012). To safeguard the nest from intruders and ensure sustainable growth, a colony must deploy effective nestmate discrimination. Each ant colony carries a set of cuticular chemical odor labels, which workers use to distinguish nestmate from non-nestmate conspecifics (Ozaki et al. 2005; Sturgis and Gordon 2012).
Many studies indicated that the presence of one or more queens has a significant impact on the odor label of a colony, such that queenless colonies exhibit reduced territoriality, acting as if they have lost some component of their distinct colony identity. Compared to queenless workers, workers from queenright colonies are more subject to aggression by non-nestmate conspecific workers, as well as being more aggressive themselves towards such workers.
Queenless S. invicta workers of the monogyne social form (single queen per colony) received little aggression from their original nestmates when returned to the natal queenright colony. However, if they tended a foreign queen for only 15 minutes, they were attacked by their original nestmates (Obin and Vander Meer 1989). Queenless workers also became less aggressive themselves after queen removal, becoming completely docile after about two weeks (Vander Meer and Alonso 2002). Thus, in monogyne fire ants, a colony’s unique chemical identity is attributable, at least in part, to its sole reproductive queen.
Similar phenomena were demonstrated in other species. Queenless workers of Cataglyphis niger did not show aggression towards and did not receive aggression from original nestmate workers from the queenright parent colony (Lahav et al. 1998). In Camponotus species, worker aggression towards non-nestmates largely disappeared in queenless colonies and reappeared after an unrelated queen was adopted into the colony (Carlin and Hölldobler 1983, 1986, 1987). Additionally, queens with less developed ovaries or that were incompletely inseminated had a weaker such effect than normal queens (Carlin and Hölldobler 1987).
However, as a counterexample, queen presence had little impact on nestmate discrimination in Camponotus aethiops, a discrepancy that was attributed to the presence of heritable chemical cues on workers (van Zweden et al. 2009). Another example came from Rhytidoponera confusa, where nestmate discrimination also was not affected by the presence of the queen (Crosland 1990). However, in R. confusa, queen-worker dimorphism is not pronounced and the colonies can reproduce without the queen (Ward 1981, 1983). Because the queen is not an indispensable component of the colony identity, it is plausible that her absence does not impact worker nestmate recognition abilities (Carlin and Hölldobler 1991).
These results point to the intricate nature of nestmate discrimination in ant colonies, which is governed by both environmentally derived cues and genetic factors (Carlin and Hölldobler 1983; Helanterä et al. 2011; Sturgis and Gordon 2012; Caliari Oliveira et al. 2022). Unraveling the specific mechanisms by which queen pheromones affect nestmate discrimination remains a challenge (Box 1, Q2). One possibility is that workers obtain queen cuticular odor cues in small colonies through common contact with the queen (Lahav et al. 1998). Another possibility is that queenless workers retain the ability to detect non-nestmates but lack the incentives to act aggressively toward them.
Maintenance of reproductive dominance
Inhibition of larval sexual development
As the hallmark of eusociality, reproductive division of labor necessitates that members of the queen caste maintain their reproductive dominance. In small colonies, the queen may reasonably manipulate larval caste fate and suppress worker reproduction through physical actions, which is frequently observed in wasp societies where caste dimorphisms are not distinct (Ross and Matthews 1991). However, in populous colonies, the effect is often achieved via an essential and well-studied class of queen pheromones that exert their effects onto almost all life stages of female colony members (Smith and Liebig 2017; Holman 2018).
To begin with, queen pheromones have been shown to inhibit sexualization (development as queens) of female larvae, thus biasing female development toward worker production over gyne (virgin winged queen) production.
Early studies on the subject came from Myrmica species, where the queen induces improved larval survival, earlier pupation, and lower larval and pupal weights, in apparent accord with the notion of the queen stimulating general colony function and cohesion (Brian 1957, 1986b; Brian and Carr 1960). When the queen was present, large gyne-destined larvae received less care from workers than small worker-destined larvae (Brian and Hibble 1963); moreover, the workers lethally bite gyne-destined larvae (Brian and Carr 1960; Brian 1973). The presumed pheromone was not volatile while, notably, structural and topological features of the queen played a role in the pheromonal effect (Brian 1970, 1973).
In the Pharaoh’s ant, Monomorium pharaonis, the presence of a fertile queen inhibited development of sexual brood (Petersen-Braun 1975, 1977). Unlike the case in Myrmica, this effect was disseminated specifically by the queen-laid eggs, but not queen corpses or solvent extracts (Berndt and Nitschmann 1979; Edwards 1987; Boonen and Billen 2017). Queenright workers always accept worker broods from a foreign colony but would cannibalize any introduced sexual broods (Edwards 1991). Monocyclic diterpene neocembrene, a compound produced only by egg-laying queens, was found to be a pheromonal component in M. pharaonis as it elicited weak “queen retinue” attraction as well as inhibited production of sexuals (Edwards and Chambers 1984; Oliveira et al. 2020).
Evidence from the two above examples and many other species strongly suggested that queen pheromones affect caste development through the behavior of workers, including biting the larvae to suppress growth evident by bite marks on larvae, or simply killing sexualized larvae (see Supplementary Information for more details; Table 1; Box 1, Q3). This is also the case in honey bees and stingless bees where workers control the caste fate of larvae, besides genetic and maternal factors (Bueno et al. 2023).
Adult workers probably sense queen pheromones mainly through antennal chemoreceptors, which was supported by strong electrophysiological responses of worker antennae towards queen extracts and candidate queen pheromones (D’Ettorre et al. 2004b; Holman et al. 2010; de Narbonne et al. 2016). Following perception, workers respond to queen pheromones by actively suppressing the sexual development of some larvae via the behaviors above (Box 1, Q4).
Inhibition of reproductive physiology of adult females
As the best studied effect in many groups of social insects, queen pheromones suppress physiological changes tied to the onset of reproduction in adult females in the colony, which can be measured in diverse behavioral and physiological changes in the trajectory of reproductive development, including dealation, ovarian activation, weight gain, and finally egg laying.
Dealation, or wing shedding, is the first observable indication of the onset of reproduction development in adult gynes in many ants. Take the example of S. invicta, virgin gynes in queenright colonies typically remain winged until a mating flight event, after which the newly mated gynes kick off their wings to initiate colony funding underground (Tschinkel 2013). However, these gynes can dealate as soon as 12 hours after separation from fertile queens, with their alary muscles beginning to histolyze simultaneously followed by ovarian development, and oviposition starts in another two to three days (Fletcher and Blum 1981, 1983b; Vargo and Laurel 1994). By this time, the gyne begins to exhibit attractiveness in the formation of a queen retinue and to produce the inhibitory pheromones (Vargo 1999).
The fecundity of a queen is correlated to her weight and the ability to suppress reproductive development in nestmate queens. The correlation is likely due to a link between weight and the level of pheromone production (Fig. 3). Evidence comes from the fact that queenless monogyne workers consistently recognized and adopted the heavier queen of two presented as their new queen (Fletcher and Blum 1983a). Corpses of heavier queens suppressed dealation for longer than light-weight queen corpses (Fletcher and Blum 1983b; Willer and Fletcher 1986). Such an inhibitory effect acts on other egg-laying reproductive queens as well: the addition of live queens or queen corpses reduced fecundity of all nestmate queens in polygyne colonies (Vargo and Laurel 1994).
Perception of queen pheromones via the antennal sensilla leads to a downregulation of dopamine production, which in turn suppresses the production of juvenile hormones (JH) and inhibits reproductive development (Robinson and Vargo 1997; Boulay et al. 2001). JHs are critical regulators not only of reproduction, but of development and behavior, throughout the lifecycle of insects (Jindra et al. 2013). Topical treatment of alate gynes with JH or JH analogue induced dealation in S. invicta even in the presence of the queen, overriding the inhibitory effect of the queen pheromone (Vargo and Laurel 1994). Notably, JH treatment can yield opposing effects on reproductive development depending on the species and the size of the treatment doses used, suggesting a condition-dependent cost of JH and calling for more studies on the endocrinological regulation of reproduction (Robinson and Vargo 1997; Cuvillier-Hot et al. 2004; Penick et al. 2011; Holman 2012).
The inhibitory effects on worker reproduction received detailed studies in Camponotus, where eggs are again the dissemination agents of the pheromones. The addition of queen-laid eggs prohibited workers from laying eggs in queenless colonies; in addition, worker-laid eggs were less prone to destruction when applied with queen cuticular hydrocarbons (CHCs) (Endler et al. 2004). As a further support, surface chemical profiles of eggs corresponded to the cuticular chemical profiles of the respective egg-laying queen or worker (Endler et al. 2006).
In Lasius species, the compound 3-methylhentriacontane (3-MeC31) suppressed egg-laying of workers in Lasius species, making it the first identified ant queen pheromone with inhibitory effects on worker reproduction (Holman et al. 2010, 2013; Holman 2012). Further analysis indicated a slower evolution of this compound compared to other CHCs in Lasius, hinting at potential evolutionary constraints on queen signals, but did not agree with the findings in Temnothorax species (Brunner et al. 2011; Holman et al. 2013). Other studies have shown inhibitory effects on worker reproductive physiology by live queens, queen corpses, or queen-laid eggs across various ant taxa (Table 1; Supplementary Information; Box 1, Q5)
Another general effect of such inhibitory pheromones is a shortening of longevity. Reproduction and longevity are typically a trade-off in animals (De Loof 2011; Blacher et al. 2017). However, in social insects, reproduction and longevity are instead positively linked (Blacher et al. 2017), perhaps due to a reproductive division of labor where queens are liberated from costly daily tasks. When worker ants become reproductively active after queen removal, they also showed extended lifespans, as documented a some ant species as well as other social insects (Kohlmeier et al. 2017; Vollet-Neto et al. 2018; Majoe et al. 2021; Negroni et al. 2021).
Harpegnathos saltator workers are fully capable of reproduction (Peeters et al. 2000). Removal of the queen (and her pheromones) prompted workers to engage in antennal duels, a ritualistic competition to re-establish hierarchy that occurs in many ponerine species (Powell and Tschinkel 1999; Peeters et al. 2000; Penick et al. 2014). Winners of these duels transition into gamergates (Sasaki et al. 2016), which had about five times the lifespan of normal workers (Yan et al. 2022). These gamergates displayed queenlike physiology, with decreased brain and optic lobe volumes, and decreased venom production. They also behaved more like queens, remaining inside the nest and hiding from intruders (Penick et al. 2021). Nevertheless, these queen-like traits can revert back to a worker-like state if a gamergate is exposed to a strong source of queen pheromones (Penick et al. 2021), such is the case in other social insects (Van Oystaeyen et al. 2014).
Induction of worker policing
Although the queen is the dominant reproductive member, workers in many ant species can potentially produce males by laying unfertilized eggs. These egg-laying workers pose a source of conflict over male parentage within the colony, as well as a cost to the colony productivity (Helanterä and Sundström 2007; Bourke and Franks 2019). The queen(s) repress worker reproduction through pheromonal inhibition as discussed in the above section, or by dominance behavior and destruction of worker-laid eggs (Heinze and Smith 1990; Bourke 1991). Workers themselves also police reproduction of nestmate workers, a behavior documented in many social hymenopterans that might have evolved concurrently with eusociality (Ratnieks 1988; Frank 1995, 2003; Wenseleers et al. 2020a).
Typical acts of policing in social insect colonies include direct aggression toward adults or destruction of their eggs (Ratnieks and Visscher 1989; Beekman and Oldroyd 2005). Mechanistically, workers must i) recognize queen presence through queen pheromones and ii) correctly assess the fertility status of colony members, as well as recognize the origin of offspring through pheromones present on the egg surface or post-embryonic cuticle (Ratnieks 1995; Oi et al. 2015b). In some species, queens actively mark suspect individuals with pheromones to “command” worker policing.
Pachycondyla workers would lay viable embryonated eggs when the workers were physically separated from the queen (Dietemann and Peeters 2000). Worker-laid eggs were eaten more frequently by nestmate workers than queen-laid eggs, and such policing was more prominent when the queen was present (D’Ettorre et al. 2004a). Notably, the potential pheromonal cues by which workers distinguish egg origin were persistent and non-transferable through mutual contact between the eggs (D’Ettorre et al. 2006).
Workers in the genus Formica could distinguish nestmate eggs from non-nestmate eggs, and worker-laid eggs from queen-laid eggs, but the latter ability was displayed only when an adult queen was present (Helanterä and Sundström 2005, 2007; Helanterä and Ratnieks 2009a, b; Chernenko et al. 2013). Corresponding to the above finding, hydrocarbon profiles of eggs displayed robust and consistent differences among species, colonies, and even among matrilines within a colony, demonstrating a link between genetic variation and potential pheromonal variation (Helanterä et al. 2014; Helanterä and d’Ettorre 2015).
In some cases, the egg-laying workers themselves, but not their eggs, were subject to policing. In Temnothorax unifasciatus, that reproductive workers were attacked, not by random nestmates, but only by a select few workers who would become dominant reproductives upon queen removal (Stroeymeyt et al. 2007). Likewise, in Novomessor cockerelli (previously Aphaenogaster cockerelli), worke-laid eggs did not differ from queen-laid eggs in their surface chemical profiles and were not policed (Smith et al. 2008a). Instead, egg-laying workers were attacked by nestmate workers (Smith et al. 2011). Reproductive status is signaled by unbranched alkanes, as the application of these compounds on non-reproductive workers induced nestmate aggression, but only in the presence of a queen (Smith et al. 2009). The queen also attacked and marked reproductive workers for aggression by discharging compounds from her Dufour's gland onto the target worker (Smith et al. 2012a). This is similar to the finding in Dinoponera quadriceps, where high-ranking gamergates mark challengers with Dufour’s gland secretion to direct aggression by low-ranking workers (Monnin et al. 2002).
In Odontomachus brunneus, a hydrocarbon, (Z)-9-nonacosene, was identified as a fertility signal, based on three lines of evidences: i) its higher abundance in reproductive individuals, ii) the typical submissive gesture of nestmate workers towards workers treated with the compound (Fig. 2), and iii) the nestmate policing (biting and pulling) of treated workers in queenright colonies (Medeiros et al. 1992; Smith et al. 2012b, 2013). The role of this compound was conserved across geographic populations, but it must function synergistically with other pheromonal chemicals (Smith et al. 2013, 2015).
Induction of execution of superfluous reproductive adults
A queenright monogyne colony is generally not expected to accept additional reproductive queens as this would decrease indirect fitness benefits to workers despite the apparent benefit of larger social groups (Hamilton 1964; Gardner et al. 2011). In species with strong caste dimorphism, superfluous queens are eliminated by workers, which may be considered as an extreme form of policing, as these queens cannot transition back to a worker-like state and contribute to colony tasks. As a requirement, workers rely on queen pheromones that signal the presence and identity of the true queen.
Monogyne Solenopsis invicta workers imprint on the pheromonal signature of their mother queen, killing any other dealate (wingless reproductive) queens presented to the colony (Fletcher and Blum 1983a; Gotzek and Ross 2007). Only when a colony is rendered queenless for a week or more will it accept an unrelated queen, and the longer the colony stays queenless, the more accepting of a foreign queen it becomes (Fletcher 1986; Vander Meer and Alonso 2002). Additionally, when presented with multiple reproductive queens, such hopelessly queenless workers usually select the most physogastric one, which might be due to a higher amount of fertility signal produced by such queens (Fletcher and Blum 1983a).
In Aphaenogaster senilis, workers attack supernumerary gynes and only the oldest gyne ascends to become the sole reproductive queen (Chéron et al. 2009). In Argentine ants, Linepithema humile, queenless colonies show lower aggression towards intruder queens compared to queenright colonies, which usually kill intruder queens within 24 hours (Vásquez and Silverman 2008). Adoption decisions were not influenced by fecundity, but by similarity of CHC profile to the nestmate queens (Vásquez and Silverman 2008; Vásquez et al. 2008). Thus, there appears to be a tight but variable linkage between queen pheromones involved in nestmate recognition, fertility signaling, and regulation of reproduction in ants.
Regulation of colony social structure
An archetypical colony of social insects is composed of a single family, headed by a single queen and her offspring workers. It is less known that multiple-queen colonies occur in many ant taxa and such variation of colony social structure, being either monogyne (single-queen) or polygyne (polygyne), exist both within species and across species. Phylogenetic analysis suggested that eusociality of ants evolved under the monogyne condition, while polygyne forms subsequently evolved independently in many ant taxa (Ross and Carpenter 1991; Huges et al 2018). In a general sense, the evolution of social structure (from monogyny to polygyny) and the evolution of eusociality in ants, raised similar problems as to why individuals are willing to forgo personal reproductive success for the benefit of group reproductive output.
Queen pheromones are also involved in the regulation of such variation in colony social structure, an important but often overlooked class of function. Although this function of queen pheromones may be ubiquitous in diverse ant taxa (Hölldobler and Carlin 1985; Evison et al. 2012; Abril and Gómez 2019), the only such case that has received careful study to date is the regulation of colony social form in S. invicta (Box 1, Q6).
In stark contrast to the single-queen, monogyne form, the polygyne form houses multiple reproductive queens, as many as a few hundred, in a colony. The two social forms are distinct from each other in many other natural history traits, such as nest density in the wild, the average weights of alate gynes and reproductive queens, colony founding mode, and worker size distributions (Keller and Ross 1995; Gotzek and Ross 2007; Tschinkel 2013; Huang and Wang 2014). The genetic underpinning of this social form polymorphism in S. invicta and several congeners is an inversion-based selfish genetic element termed the Social b (Sb) supergene. The element spans a large portion of chromosome 16, comprising three adjacent inversions and encompassing over 500 described genes (Yan et al. 2020; Stolle et al. 2022; Helleu et al. 2022).
In monogyne colonies, all female members are homozygous for alternate, wild-type haplotype (SB), and only one SB/SB reproductive queen is tolerated. In polygyne colonies, all reproductive queens and over half of the worker population are heterozygous at the supergene locus. Polygyne workers enforce this striking genotype composition of their queens in a green-beard fashion, accepting additional Sb-carrying queens but executing SB/SB queens, including nestmate SB/SB gynes shortly after they emerge as adults (Ross and Keller 1998; Keller and Ross 1998).
The pheromonal basis of the supergene genotype signal was first demonstrated by the findings that polygyne workers rubbed against SB/SB queens were attacked by their nestmate workers (Keller and Ross 1998). Specific cuticular hydrocarbons were found to be uniquely present on the cuticle of SB/Sb queens, the abundance of which increased as the fertility of the queen increased (Eliyahu et al. 2011). Trible and Ross (2016) showed that polygyne workers showed strong preferences toward polygyne queen extracts over monogyne queen extracts, confirming the presence of a supergene pheromone. Zeng et al. (2022) then showed that a complex blend of unsaturated CHCs functioned as this signal of queen supergene status to workers (Box 1, Q7). However, beyond the recognition of supergene status, the precise mechanisms by which multiple Sb-carrying queens are permitted in polygyne colonies remain elusive. A potential general explanation may be that Sb queens are perceived as identical individuals, despite substantial variation in their fertility status and material apportionment (Ross 1988).