OUP user menu

Diversity of social and mating systems in cavies: a review

Oliver Adrian, Norbert Sachser
DOI: http://dx.doi.org/10.1644/09-MAMM-S-405.1 39-53 First published online: 16 February 2011


Cavies (family Caviidae) are a group of rodents distributed over much of South America. The high diversity of habitats inhabited by different species is paralleled by a high diversity of social organizations. Comparisons of behavioral and reproductive strategies between species can provide valuable insights into mammalian social evolution. In this review we 1st summarize the literature to give an idea of the diversity of social and mating systems within the genera Cavia, Galea, Microcavia, and Kerodon. Social systems range from solitary animals through pairs and harems to multimale-multifemale groups. Mating systems of cavies include monogamy and promiscuity and different forms of polygyny. We then review behavioral strategies of females and males that account for this social diversity. Particularly, the role of females and the potential of males to monopolize mates are examined. In some species, for example, estrous females actively solicit copulations with several males and thereby prevent their monopolization by single males. We also discuss adaptations of reproductive physiology to different mating systems. Finally, environmental factors influencing aspects of the social and mating systems of cavies are considered. Particularly, differences in resource distribution, predation risk, and climatic conditions might explain the great variation in species social organization.

Key words
  • behavioral strategies
  • cavies
  • Caviidae
  • communal burrowing
  • mating system
  • paternities
  • predation risk
  • sexual size dimorphism
  • social system
  • spacing behavior

Cavies or guinea pigs (Caviidae) are a small group of caviomorph (New World hystricognath) rodents distributed over much of South America. The domestic guinea pig, 1 of only 3 ancient New World domesticated mammals (Lavallée 1990), has experienced an extraordinary worldwide distribution as a laboratory animal and a pet. In pre-Columbian times guinea pigs were sacrificed in religious ceremonies to the gods of the ancient cultures of the central andes. They also represented an important source of food for humans, as they do today (Sandweiss and Wing 1997).

The guinea pig was domesticated from a wild form still abundant in Peru (Hückinghaus 1961; Spotorno et al. 2004, 2007), but more than a dozen wild species exist on the continent. Four genera of cavies—Cavia (common or wild cavies), Galea (yellow-toothed cavies), Microcavia (mountain or desert cavies), and Kerodon (rock cavies or mokos)—can be distinguished easily. All have a similar appearance on 1st view and will be identified as guinea pigs even by a layperson. Upon closer look, Kerodon represents the genus whose morphology differs the most. Recent molecular phylogenetic studies revealed that Kerodon is related more closely to the capybara (Hydrochoerus hydrochaeris) than to the other cavies (Rowe and Honeycutt 2002; Trillmich et al. 2004).

The capybara traditionally has been placed in a family of its own (Hydrochoeridae) but now is included in the family Caviidae (Woods and Kilpatrick 2005), forming a subfamily with Kerodon (Hydrochoerinae). The remaining 3 genera of cavies constitute the subfamily Caviinae. A 3rd subfamily (Dolichotinae) contains 2 species of maras (Dolichotis patagonum and D. salinicola), which resemble small ungulates or short-eared lagomorphs more than other cavies. Within the genera Cavia and Galea several species have been described that later were synonymized with other species. Knowledge of the number of and relationships between different species and subspecies and their geographic distribution remains limited. An overview of cavy species and subspecies currently accepted is given by Woods and Kilpatrick (2005). The species G. monasteriensis (Solmsdorff et al. 2004) has not yet been included there. According to a recent phylogenetic study G. monasteriensis might be a junior synonym of G. musteloides boliviensis, whereas the lowland forms of G. musteloides would not be conspecific to the highland forms and should be named G. leucoblephara (Dunnum and Salazar-Bravo 2010). All field studies on G. musteloides cited in this review (with the exception of Mares et al. 1981) and the laboratory studies for which the origin of the study animals is reported are based on the lowland forms.

Caviids inhabit almost all of South America with the exception of western Chile and Tierra del Fuego. The northwestern edge of the continent and parts of the Amazon Basin are occupied only by the capybara but not the cavies. Distribution maps of the different species are given by Mares and Ojeda (1982), Eisenberg (1989), Redford and Eisenberg (1992), and Eisenberg and Redford (1999). The wide distribution of the family results in 1 or several species present in all major habitat types of South America (Mares and Ojeda 1982). Capybaras live in the tropical rain forests of the Amazon Basin and surrounding savannas and swamps. Maras inhabit the deserts, grasslands, scrubs, and dry forests of the southern part of South America. Habitats of cavies include grasslands, savannas, marshes, moist and dry forests, scrubs, deserts, montane regions, and cultivated areas. A variety of different ecotypes has been recognized in this group and include semiaquatic forms with webbed feet (H. hydrochaeris and C. magna) and semiscansorial (M. australis and K. rupestris), burrowing (Dolichotis spp. and Microcavia spp.), cursorial (Dolichotis spp.), and ground-dwelling forms (Cavia spp. and Galea spp.).

This diversity of habitats and ecotypes is reflected in a high diversity of social and mating systems. The male(s) of an established social unit (e.g., a monogamous pair, a harem, or a multimale-multifemale group) is (are) often not the exclusive mating partner(s) of the female(s) of that unit. Extrapair or extragroup copulations and paternities occur in many species, particularly in birds (Birkhead and M0ller 1992, 1995), but also in mammals (Digby 1999; Martin et al. 2007; Munshi-South 2007). Thus, both the mating behavior (social mating system) and the genetic outcome of matings (genetic mating system) are not necessarily determined by the composition of a social group.

A classical view of social organization of animals is that resource distribution determines the distribution of females in space and time. Female distribution then explains male distribution, and these patterns together with male-male competition for mates effect different forms of social organization (including mating systems—Clutton-Brock and Harvey 1978; Emlen and Oring 1977). However, the role of females is much more active than expressed in this view. Female, just as well as male, behavioral strategies for survival and successful reproduction represent the proximate factors generating the diversity of social and mating systems in different species. Behavioral strategies find their expression also in morphological (e.g., sexual size dimorphism) and reproductive (e.g., sperm characteristics and estrous type) adaptations and have consequences for male and female reproductive success. Ultimately, behavioral strategies evolve as adaptations to specific environmental conditions. Thus, the distribution of food resources, predation risk, climatic conditions, or costs of burrow construction and maintenance, among other factors, must be considered when assessing determinants of social and mating systems.

Beginning with Rood (1969), (1970), (1972), multiple detailed studies on the social organization in the natural habitat of different cavy species have been conducted, complemented by investigations of captive colonies in seminatural enclosures and laboratory studies, to answer specific questions on the social and mating system through experimental research under controlled and standardized conditions. In this review we 1st summarize the available literature to note the remarkable diversity of social and mating systems within this group of South American rodents and present data on spacing behavior, social interactions, sexual size dimorphism, mating, and paternity. Where it seemed relevant to us, we also included unpublished field data from our group. We focus here on the cavies (Fig. 1). The maras and capybaras are excluded from this review. Information on their social and mating systems can be found in Herrera et al. (2011) [this issue]), Lord (2009), Macdonald et al. (2007), Taber and Macdonald (1992a), (1992b), and references therein. Because it represents a special case due to artificial selection during the long process of domestication (Künzl et al. 2003; Künzl and Sachser 1999), we also do not consider the domestic guinea pig. Its social organization is the central theme of papers by Fuchs (1980), Sachser (1986), Sachser et al. (2004), and references therein.

Fig. 1

Taxonomy of Caviidae (following Woods and Kilpatrick 2005) at the generic level and above, genera of the cavies (light gray background color), and species described in this review.

Subsequently, we review behavioral strategies of females and males responsible for the diversity of social and mating systems. Particularly, the role of females and the potential of males to monopolize mates are examined. After this we describe adaptations of reproductive physiology in cavies with different mating systems. Finally, environmental factors influencing certain aspects of the social and mating systems of cavies are discussed.

Social and Mating Systems: Spacing Behavior, Social Interactions, Sexual Size Dimorphism, Mating, and Paternity

Cavia aperea.—Distribution of C. aperea is associated with areas of dense ground vegetation and surrounding zones of short grass where the animals forage (Asher et al. 2004, 2008; Bonaventura et al. 2003; Cassini and Galante 1992). Male and female cavies have stable home ranges (Asher et al. 2004, 2008; Rood 1969, 1970, 1972), those of males being larger than those of females (Table 1). Male home ranges overlap those of 1–3 females almost completely, whereas little or no overlap is found among neighboring males; neighboring females overlap about 40% on average (Asher et al. 2004, 2008). Although little home-range overlap exists among males, they do not behave territorially; home-range borders are neither scent-marked nor defended against intruding males. However, females are scent-marked, and other males are chased by a resident male when they approach a resident female (Asher et al. 2004, 2008). Distribution and overlap of home ranges allows the differentiation of distinct spatial units: pairs (1 adult male and 1 adult female) and harems (1 adult male and 2 or 3 adult females). In harems home-range overlap is greater between the male and his females than between the females, and females are found at greater distances from each other than from the male. The number of females belonging to a male unit can depend on the home-range sizes of the females. At a site in Uruguay (Asher et al. 2008) with a 2.4-fold greater population density than at a site in Brazil, female home ranges were smaller (Asher et al. 2004; Table 1), and in Uruguay harem groups dominated, whereas in Brazil the predominant type of social unit was the pair.

View this table:
Table 1

Male and female individual home-range sizes (m2) in cavies. Note that absolute home-range sizes must be compared with caution among studies because of different estimation methods and sample sizes. ∆ signif. = difference between males and females statistically significant (P < 0.05). MCP = minimum convex polygon.

Home-range size (mean ± SD)Δ signif.MethodSource
Males (n)Females (n)
Cavia aperea1,387 ± 657 (9)1,173 ± 464 (8)Exclusive boundary stripRood (1972)
880 ± 217 (5)549 ± 218 (7)YesMCP 100%Asher et al. (2004)
604 ± 324 (4)320 ± 249 (12)YesMCP 100%Asher et al. (2008)a
165 ± 22 (2)MCP 100%Asher et al. (2008)b
543 ± 503 (7)302 ± 368 (15)YesMCP 100%Asher et al. (2008)a
Cavia magna11,830 ± 6,210 (24)7,670 ± 3,970 (25)YesMCP 100%Kraus et al. (2003)c
Cavia intermedia1,700 ± 1,000 (7)1,700 ± 1,300 (5)NoMCP 100%Salvador and Fernandez (2008a)
Galea musteloides4,275 (1)Exclusive boundary stripRood (1972)
12,083 ± 5,416 (6)2,250 ± 1,041 (4)YesMCP 100%M. Asher, University of
Münster, pers. comm.
Galea spixii872 ± 497d632 ± 978dNoIntertrap distancesLacher (1981)
Microcavia australis3,942 ± 2,420 (22)2,187 ± 940 (16)Exclusive boundary stripRood (1972)e
7,720 ± 1,160 (5)3,525 ± 382 (3)Exclusive boundary stripRood (1970, 1972)e
6,737 ± 4,443 (6)3,517 ± 2,449 (12)NoMCP 95%Ebensperger et al. (2006)
3,738 ± 1,322 (9)1,963 ± 1,143 (14)YesMCP 100%Lüttmann (2006); Ranft (2005)
  • a Data of the same population during early summer and late summer, respectively.

  • b Small satellite males in contrast to large resident males above.

  • c Drifting home ranges.

  • d Sample size not reported in original source.

  • e Data of the same population including all animals with ≥6 records and animals with ≥50 records, respectively.

In Uruguay all harem males were large, with a body mass > 500 g (Asher et al. 2008). Within 2 of the spatial units a small but sexually mature male of <350-g body mass (satellite male) was present whose home range was much smaller than those of the large resident male and the females. A 3rd category of mature males was termed roaming males. These were of intermediate size (350–500 g), had no stable home ranges but moved across areas much larger than the home-range areas of resident males, and were found for 1 or a few days within each resident male's home range. The following scenario might explain the relationship between male size and home ranges (Asher et al. 2008). Young males disperse from their natal area and become satellite males at neighboring sites. They assume the dominant position if the resident male dies. When they grow too large, the resident male perceives them as potential rivals and evicts them. The young males then start a roaming life searching for a vacant resident male position.

Cavia aperea is crepuscular, leaving the dense vegetation in the morning and late afternoon for short foraging bouts on more open sites. The animals forage alone or with a conspecific (Asher et al. 2004, 2008; Rood 1972). In most cases 2 animals foraging together are the male and a female of a harem unit; joint foraging of 2 females from the same harem is much rarer (Asher et al. 2008). Aggression is rare in these feeding groups as are social interactions in general (Rood 1972). Males maintain proximity to females, whereas the latter seem to ignore them.

Social interactions are difficult to observe in C. aperea because most occur in dense vegetation. Hence, most information on social interactions stems from observations of captive groups. C aperea is characterized by a low threshold for aggression and possesses a greater variety of aggressive behavioral patterns compared to the cavies G. musteloides and M. australis (Rood 1972). Aggressive interactions usually occur between 2 males or 2 females or between females and subadults. In the captive groups of Sachser et al. (1999) and Stahnke (1987) adult males were incompatible; not more than 1 adult male could be kept in an enclosure because fighting between 2 males always escalated, leading to severe injury. In the groups of C. aperea of Rood (1972) the males formed linear dominance hierarchies. Dominance depended on body mass and former residency. When a new male was introduced into a group he was attacked severely by the resident male, which often caused the death of the intruder. Loss of dominance also resulted in the death of former alpha males within a month after they withdrew from any social interactions. Frequent alpha male aggression toward subordinate males was a primary cause of high male mortality in the captive colonies.

Among females intrasexual aggression is much less pronounced, but it also occurs frequently (Rood 1972; Sachser et al. 1999; Stahnke 1987). In the field females seem to avoid each other (Asher et al. 2008). In captive groups they establish linear dominance hierarchies (Rood 1972) that are age-dependent (Sachser et al. 1999). Only high-ranking females successfully breed in such colonies. An adult male might intervene in a fight between 2 females by placing himself between the adversaries and starting courtshiplike behavior (Stahnke 1987). Unfamiliar adult males and females frequently exhibit heightened aggression (Rood 1972; Stahnke 1987). Between males and females familiar with each other, interactions are predominantly amicable, but rare. Social grooming and close bodily contact almost never occur (Asher et al. 2004; Rood 1972; Sachser et al. 1999). In mate-choice experiments, where females had access to 4 (Hohoff 2002) or 2 (Adrian et al. 2008a) different males, each female had a clear social preference for 1 of the males, measured as the relative time spent with each male.

Because females seem to avoid each other and interact only aggressively, it is not surprising that nursing of other females' pups is extremely rare (Rood 1972; Stahnke 1987). Females seem to distinguish between their own and alien young and may aggressively repel the latter (Rood 1972). Males of C. aperea show paternal behaviors in the form of social play and grooming and only rarely are aggressive toward their pups (Adrian et al. 2005).

In conclusion, although large mixed-sex groups were described (Rood 1972), the social system of C. aperea seems better characterized by 1-male units (Table 2). Although a stable social relationship between the male of a pair or harem and his female(s) exists, no such relationship exists among females of the same harem.

View this table:
Table 2

Social and mating systems and associated traits in the cavies. For references to the data see the main text. Note that different terms have been used here for the description of social and mating systems to emphasize their different character. In the literature the term monogamy, for example, is used for both pair-living and exclusive mating with 1 partner by both male and female.

Social systemMating systemSexual size dimorphismRelative testes sizeMultiple paternityPaternal behavior
Cavia apereaPair/haremFemale-defense polygynyMales > femalesLow compared to G. musteloidesInfrequentModerate
Cavia magnaSolitaryPromiscuity?Males > females???
Cavia intermedia??Males = females???
Galea musteloidesSolitary/multimale-multifemale groupsPromiscuityFemales > malesHighFrequentNo
Galea spixii?Lek-type of male-domi nance polygyny (?)???No
Galea monasteriensisPairMonogamyFemales > malesLow compared to G. musteloidesAbsentFrequent
Microcavia australisFemale-centered multimale-multifemale groupsPromiscuityMales = femalesHigh??
Microcavia niata??????
Kerodon rupestrisHaremResource-defense polygynyMales = females??Frequent

In captive colonies with several males usually only the alpha male (and sometimes the beta male) exhibits frequent courtship behavior toward the adult females (Rood 1972). Courtship behavior of other males seems to be inhibited by the alpha male. However, the other males show courtship behavior to subadults and juveniles. When pregnant females near term, the alpha male starts to guard them, maintaining close proximity and chasing other approaching males away. The alpha male is usually the 1st to copulate, but subordinate males also can copulate successfully with the female in the course of a mating chase. The alpha male is highly aggressive toward rivals but can prevent other males from copulating only in about one-half of all cases (Rood 1972). The other males usually ejaculate rapidly on the 1st intromission in contrast to alpha males, which mount the female several times before ejaculation.

In the field courtship and sexual behavior were observed exclusively between members of the same social group of C. aperea (Asher et al. 2004). Microsatellite analyses of 15 litters with >2 pups born to females captured in the field while gravid revealed that multiple paternity occurred in 13–27% of litters, a value much lower than in G. musteloides, the only other cavy species in which paternity has been examined (Asher et al. 2008). All individual males identified as sires of the offspring had a body mass of >500 g, thus belonging to the category of large resident, harem males. No multiple paternities were observed in litters from mate-choice tests in which females could mate with 2 or 4 different males, although several females had copulated with > 1 male (6 litters each—Adrian et al. 2008a; Hohoff 2002).

As is common when males compete for access to females, males of C. aperea are significantly heavier than nongravid females (Asher et al. 2004, 2008; Bonaventura et al. 2003; Sachser et al. 1999). In the study population of Asher et al. (2008) males were, on average, 17% heavier. Large resident males were about 32% heavier than females, whereas no significant difference in body mass was found between roaming or satellite males and adult females. The exclusivity of reproductive success of large males, strong sexual size dimorphism, and rarity of multiple paternities point to male contest competition for females. The mating system can be described as female-defense polygyny with limited promiscuous mating (Table 2).

Cavia magna.—The social organization of C. magna is very different from that of C. aperea. C. magna lives in wetlands that periodically are flooded. Adaptations of this species to these environmental conditions are webbed feet and good swimming ability. Individuals of both sexes of C. magna were distributed randomly at a site in Uruguay (Kraus et al. 2003). Home ranges (Table 1) were significantly larger in males than in females and depended on water levels. When water levels were high, male home-range size increased, whereas female home-range size decreased. Home ranges were not stable but shifted from month to month across the study area independently of the water level. Male home-range fidelity was lower than that of females. Home ranges overlapped strongly within and between sexes. At high water levels overlap was higher because the animals congregated in the drier areas. Male-male and male-female overlap tended to be higher than female-female overlap. Male home ranges did not completely encompass single female home ranges, as is the case in C. aperea (Asher et al. 2004, 2008).

Analysis of static and dynamic interactions of animals within the overlap zones suggested that use of these zones was random. Group memberships were not apparent, and no stable social relationships between males and females or among females existed (Kraus et al. 2003). Intersexual interactions were more frequent than intrasexual ones. Animals neither avoided nor were attracted by conspecifics sharing parts of their home ranges. During foraging several animals sometimes aggregated. Territorial behavior was not observed. In accordance with this, an increasing population density did not lead to smaller home ranges, as would be expected in territorial species, but rather to an increase in home-range overlap. Nonstationary use of space is unusual for rodents. In this species it might have evolved as an adaptation to the regular flooding and varying distribution of food resources. The social system can be described as solitary (Table 2).

The larger male home ranges suggest that males might exhibit a roaming strategy and mate with every estrous female they encounter. The unpredictable location of females and their solitary lifestyle makes the monopolization of several females impossible. Rather, temporary consortships between a male and an estrous female might occur. Because reproduction is not strongly synchronized among females, a male also might form consortships with different females in succession. The mating system is described as overlap promiscuity (Kraus et al. 2003); however, no data exist on multiple mating by females. The sexual size dimorphism found with males heavier than females (Ximenez 1980) points to some degree of contest competition, and an estrous female might be monopolized by a guarding male.

Cavia intermedia.—This species was described recently (Cherem et al. 1999). C. intermedia is restricted to Moleques do Sul Island off southern Brazil. Information relating to its social and mating system is largely lacking. C. intermedia has a high and stable population density and a balanced sex ratio (Salvador and Fernandez 2008a). In contrast to C. aperea and C. magna, home-range size does not differ between males and females (Table 1), and sexual size dimorphism is absent (Salvador and Fernandez 2008b). These findings clearly point to differences from the other Cavia species in both social organization and mating system.

Galea musteloides.—This species is a habitat generalist. In Salta Province, Argentina, it occurs in 4 of the 5 major habitat types. It is most common in moist microhabitats (Mares et al. 1981). In a capture-mark-recapture study on a wild population in Nacunan, Argentina (M. Asher, University of Münster, pers. comm.), the sex ratio was biased in favor of males. Telemetry data revealed that home ranges of males were >5 times as large as those of females (Table 1). Female home ranges usually did not overlap, whereas the large male home ranges overlapped strongly with those of several females and several other males. This situation differs from that in C. aperea but is similar to that in C magna, although intersexual home-range size differences are not as large in the latter species. The large aggregation of about 25 animals found at a 48-m-long stone wall in Buenos Aires Province, Argentina, also points to extensive home-range overlap (Rood 1969, 1972).

In captive colonies males and females form dominance hierarchies (Rood 1972; Schwarz-Weig and Sachser 1996). Male-male aggressive interactions are frequent despite stable linear hierarchies. Some authors could keep only 1 adult male in a cage due to high intermale aggression (Rood and Weir 1970; Weir 1970). Male-male aggressive interactions occur more than twice as often as those among females (Schwarz-Weig and Sachser 1996). Females frequently direct socio-positive behaviors toward each other. Dominance hierarchies among them are less stable and not strictly linear (Rood 1972). Reproductive status and age are factors affecting female dominance relationships. Lactating females occasionally can improve their dominance status temporarily, and older animals are usually dominant over younger ones. Adult females are dominant over low-ranking (in the male hierarchy) males but subordinate to the alpha male. Intersexual aggression is rare compared to intrasexual aggression. All individuals of a mixed-sex group can be observed regularly sitting together in close bodily contact in rest periods that alternate with activity periods throughout the day (Schwarz-Weig and Sachser 1996). Social tolerance is higher in G. musteloides than in C. aperea. However, subadult animals experience much aggression from adults, mainly from those of the same sex. Adult males are aggressive toward young males but less so toward young females (Rood 1972). Males of G. musteloides never engage in direct parental activities. They treat even their own offspring rather aggressively, beginning at birth (Adrian et al. 2005).

Allonursing is common in captive groups of G. musteloides (Künkele and Hoeck 1995) and also was observed in the field (Rood 1972). All lactating females in a group nursed offspring of other females, and the majority of pups received milk from females other than their mother (Kiinkele and Hoeck 1995). However, females can discriminate between their own and alien pups and favor the former. Alien pups older than 8 days are treated aggressively (Künkele and Hoeck 1989).

In conclusion, field data are ambiguous, some pointing to a solitary social system (M. Asher, University of Münster, pers. comm.) and others to multimale-multifemale groups (Rood 1972), which is possibly due to intraspecific variation in the social organization of G. musteloides (Table 2). Examination of laboratory data argues against a solitary life of this species, because communal huddling and frequent allonursing point to closer relationships, at least among females.

Female estrus is the only situation where the position of the dominant male is challenged (Rood 1972). Increased male interest in females begins 1 or a few days before parturition and thus just before the next estrus (all cavies exhibit a postpartum estrus). Often, several males gather near the female. The alpha male shows mate-guarding behavior and occasionally attempts to chase other males away. As soon as the female leaves the birth site after parturition all males usually follow her in a mating chase.

Although often mate-guarding the receptive female and attacking his rivals, the alpha male does not successfully prevent other males from copulating (Rood 1972; Schwarz-Weig and Sachser 1996). Moreover, Rood (1972) found that the alpha male was seldom the 1st male to copulate. The occurrence of multiple paternities in litters from females caught gravid in the field was determined using microsatellites (M. Asher, University of Münster, pers. comm.). The percentage of multiple paternities amounted to at least 50–80% of litters. In captive groups consisting of 4 male and 6 or 7 female G. musteloides that could mate freely, multiple paternities were detected in 83% of litters with >1 pup (Keil et al. 1999). In contrast to C. aperea and C. magna, a reversed sexual size dimorphism is present in G. musteloides, which points to little effect of body size in male-male competition. In a captive population nongravid females were, on average, 15% heavier than males (Sachser et al. 1999). In a feral population in Argentina females were about 8% heavier than males (M. Asher, University of Münster, pers. comm.). G. musteloides clearly has a promiscuous mating system (Table 2).

Galea spixii.—Like G. musteloides, G. spixii is a habitat generalist that occurs in a wide variety of open habitats. Male and female home ranges overlap strongly (Lacher 1981). The home ranges of G. spixii are strikingly smaller than those of G musteloides (Table 1). Males scent-mark at certain locations of their home ranges, and tolerance between males is very low. In captive groups both sexes form linear dominance hierarchies. Agonistic behavior varies, depending on whether a dominance hierarchy is already established (Lacher 1981). Unfamiliar animals, both males and females, are overtly aggressive toward each other. Even in established groups social grooming is rare. The general aggressiveness of adults toward juveniles, even toward their own progeny, is remarkable in this species. Paternal care has never been observed. The social system, although not fully described, might be similar to that of G. musteloides.

Aggression increases as a female approaches estrus. Males try to deter competing males from mating through overt aggression. According to Lacher (1981), however, females ultimately choose with which males to copulate. He states that the mating system of G. spixii can be described best as a lek-type male dominance polygyny (Table 2). It would be interesting to gather genetic data on the mating system to specify differences from G. musteloides.

Galea monasteriensis.—This species was described in recent years (Solmsdorff et al. 2004—but named Galea sp. by Hohoff et al. [2002] and Galea sp. no v. by Trillmich et al. [2004]). All data stem from captive animals, and virtually nothing is known from field studies. Unfamiliar individuals of G. monasteriensis are extremely aggressive toward each other, much more than are individuals of G. musteloides. When 2 males or 2 females confronted each other, fighting escalated in all encounters without exception and thus had to be terminated by the experimenters to prevent serious injury to the animals (Hohoff et al. 2002). As well, 60% of male-female encounters led to escalated aggression. Aggression was initiated more often by the female than by the male.

In a series of mate-choice tests where a female had unrestricted access to 2 males housed singly, she usually exhibited a clear social preference for one of the males, measured as the relative time spent with the males in their respective enclosures (Adrian et al. 2008b). In mate-choice experiments where a female could choose between 4 different males, females not only spent more time with a single male but also directed more contact and sociopositive behaviors toward this preferred male (Hohoff 2002). Adrian et al. (2008b) measured the effects of separation and subsequent reunion in established pairs of G. monasteriensis on stress hormones and behavior of males. If the female was removed from the home enclosure, blood Cortisol levels of the male increased significantly. Upon reunion, a decline in Cortisol levels was noticeable. Reunion also led to increased socio-positive and courtship and sexual behaviors of the male toward the female and to more spatial proximity. Agonistic behaviors never were observed during reunion. These effects of separation from and reunion with a partner point to a social bond between the male and the female of a pair of G. monasteriensis.

In contrast to G. musteloides and G. spixii, where males never engage in direct paternal activities and treat even their own offspring rather aggressively, males of G. monasteriensis exhibit social play and grooming behavior with their offspring (Adrian et al. 2005). Offspring-directed aggression is rare. When transferred with her pups to an unfamiliar environment, a mother of G. monasteriensis has an attenuating influence on her pups' increased blood Cortisol levels, and mother-offspring interactions are characterized by physical closeness and sociopositive behavior (Hennessy et al. 2006). Such an influence is not found when the pups are transferred to an unfamiliar female. In contrast, interactions between pups of C. aperea and their mother and interactions between the pups and an unfamiliar female do not differ markedly.

The high incompatibility between adults, the existence of a social bond with physiological correlates, and the frequency of paternal behaviors are strong evidence for a pair-based social system in G. monasteriensis (Table 2). Other data also argue for a monogamous mating system. In 4-male choice tests conducted by Hohoff (2002) the socially preferred male was always a copulation partner of the female—although several females copulated with 2 males—and he gained a higher number of copulations than the 2nd male. In contrast to females of G. musteloides tested in identical experiments, copulations with 3 or all 4 males never occurred. In the 2-male choice tests multiple paternities did not occur in litters resulting from these tests, and in 7 of 8 litters the socially preferred male sired the offspring (Adrian et al. 2008b). Multiple paternities also did not occur in other choice tests (Hohoff 2002; Hohoff et al. 2002). All of 15 litters of >2 pups resulting from these experiments were sired by a single male. A reversed sexual size dimorphism in G. monasteriensis, with females being 9% heavier than males (Hohoff et al. 2002), points to little direct male competition for access to females.

Microcavia australis.—This species occurs throughout most of western and southern Argentina and some parts of southern Chile in arid and semiarid habitats (Tognelli et al. 2001). It is sympatric with G. musteloides at some locations but has much higher population densities (Contreras 1966; Rood 1969, 1970). Relationships between the 2 species are amicable; animals use the same runways and burrow systems, and interspecific grooming occurs (Rood 1970).

Females of M. australis can be found most of the time in the vicinity of a bush or a group of bushes, whereas males wander around visiting different female sites (Rood 1970, 1972). A particular plant structure, weeping branches of certain plant species almost reaching the ground, are used preferentially by M. australis (Tognelli et al. 1995). Underneath these bushes burrow systems are located in which the animals spend the night. Several cavies share such a burrow system with a number of entrance holes and nests and form a social unit, also called nesting association (Contreras and Roig 1978; Ebensperger et al. 2006; Lüttmann 2006; Ranft 2005). No switching of females between nesting associations over a study period of about 1–2 weeks suggested some stability of associations (Ebensperger et al. 2006). Newly established social units consist exclusively of young animals, which indicates dispersal from the native unit at adolescence (Contreras and Roig 1978). Social units can comprise between 2 and 38 members with various sex ratios (Contreras and Roig 1978; Ebensperger et al. 2006). Adult males do not associate with a specific nesting association but spend successive nights in nests of different social units (Lüttmann 2006; Ranft 2005). Some animals of both sexes do not share nests with conspecifics (Ebensperger et al. 2006).

Animals show high fidelity to their burrow systems, fleeing to the nearest entrance of their system in the event of danger, even when many entrance holes to burrow systems of other groups are closer. Members of a social unit defend their burrow system against nonmembers (Contreras and Roig 1978). Above the ground no territorial behavior is obvious, and space is shared among members of different social units that interact during aboveground activities (Contreras and Roig 1978). However, home ranges of both males and females of the same social unit overlap extensively (>60% on average—Ebensperger et al. 2006; Lüttmann 2006; Ranft 2005), whereas overlap of animals from different social units is much lower, usually <10%. Male home ranges are about twice the size of those of females (Table 1; Ebensperger et al. 2006; Lüttmann 2006; Ranft 2005; Rood 1970, 1972).

Social tolerance is higher in M. australis than in C. aperea or G. musteloides (Rood 1972). Among males relationships are hostile, but aggressive interactions are avoided and physical injuries are rare. Data on dominance relationships are ambiguous. Both a stable, linear hierarchy (Rood 1970, 1972) among males with overlapping home ranges and a lack of a rigid linear hierarchy (Contreras and Roig 1978) have been reported. The character of female-female interactions varies. Aggressive and amicable interactions can be observed (Rood 1970, 1972). This variation might depend on the membership of females in social units. In a discrete social unit of 4 females and 4 males, female-female interactions were predominantly sociopositive, and agonistic interactions occurred rarely. Females initiated social interactions with other females much more often than they did with males (Lüttmann 2006). Such amicable social interactions among females contrast strikingly with female-female relationships in other cavy species, particularly with those of G. monasteriensis, which exhibits complete intrasexual incompatibility. Male-female interactions are almost always amicable (Rood 1970, 1972) and usually initiated by the males. Individuals of M. australis can be seen regularly resting together in bodily contact. These groups contain several females and juveniles but usually only 1 adult male. Huddling during severe weather conditions could serve to conserve body heat (Rood 1970).

Allonursing is common (Contreras and Roig 1978; Rood 1970, 1972), particularly when females inhabit the same bush; however, females can discriminate between their own pups and those of other females. They occasionally are aggressive toward other pups attempting to suckle (Rood 1970).

Originally, social groups of M. australis were described as noncohesive, mere aggregations at resources (Rood 1970). Subsequent studies emphasized that the animals do not merely aggregate but form stable social units (Contreras and Roig 1978; Ebensperger et al. 2006; Lüttmann 2006; Ranft 2005) that are female-centered, with males associating more loosely with female units (Table 2). Contradictory results possibly can be ascribed to intraspecific variation of the social organization.

During the reproductive season Contreras and Roig (1978) observed an increased tendency of males of M. australis to move to the areas of neighboring social units. Males sometimes form temporary associations with gravid and lactating females and spend several hours each day with them (Rood 1970, 1972). At the time of female estrus, intermale aggression is enhanced. Several males congregate in the vicinity of the female's home bush, and the dominant male frequently attacks and chases away his rivals when they approach the female. However, this mate-guarding is not very successful. As in G. musteloides, mating chases involving several males occur, and a female usually mates with several males. In contrast to G. musteloides, copulation by a male during mating chases often is impeded by frequent attacks by other males. Successful copulations occur more often when a single male encounters a receptive female (Rood 1970).

Although body size in this species varies according to climatic differences within the total distribution area (Taraborelli et al. 2007), no sexual dimorphism in body mass exists in M. australis (Lüttmann 2006; Ranft 2005; Taraborelli and Moreno 2009). All available data point to a promiscuous mating system in this species (Table 2); however, no genetic data exist on the incidence of multiple paternity.

Microcavia niata.—Two field colonies studied in northern Chile consisted of 15 and 17 individuals, respectively, living in small areas of 100 m2 each (Marquet et al. 1993). The smaller colony consisted of 2 adult males, 5 adult females, and 8 juveniles. Neither sexual nor agonistic behaviors were observed, but sociopositive interactions, mainly grooming, occurred among colony members. Cavies reacted highly aggressively toward intrusions of noncolony members. They retreated into their burrow system when a colony member emitted alarm calls (Marquet et al. 1993).

Kerodon rupestris.—This species is a habitat specialist restricted to the Caatinga, a semiarid region in northeastern Brazil. It is associated with rock piles, where it takes refuge from predators and unpleasant weather conditions. K. rupestris climbs in trees and shrubs to feed on leaves (Lacher 1981).

Males defend a rock pile and aggressively exclude other males from this resource, and several females inhabit the rock pile of a dominant male. Thus, a 1-male-multifemale (harem) group is formed (Table 2). In captive groups dominance hierarchies among females are linear, whereas male hierarchies are not (Lacher 1981). Male-male behavior is usually aggressive, but sociopositive behavior can occur in groups with an established dominance hierarchy. In the captive colony studied by Lacher (1981) the alpha male was groomed frequently by subordinate males. The beta male also was groomed but never groomed other males. Social grooming never was observed among females, and it was infrequent between males and females. Other social behaviors (crawl-over and huddling) did occur between females. However, aggression among females is pronounced in this species. High stress levels of females, due to female-female aggression and female aggression toward juveniles other than their own, might be responsible for the high juvenile mortality (Lacher 1981). Females seem to belong to the same harem due to their reliance on the shelter site only. The harem is an internally competitive group; frequent aggression toward offspring of other females of the same harem points to competition among females (Lacher 1981).

Males are very tolerant toward their offspring. In captive pairs of K. rupestris males spent much time with the young and exhibited paternal behaviors such as grooming them and huddling with them (Tasse 1986).

In the captive colony observed by Lacher (1981) the alpha male and some subadult males, but not the beta male, were involved in mating chases with a female in estrus. The mating system of K. rupestris has been named resource-defense polygyny (Lacher 1981); however, no genetic data on the mating system are available. Contrary to the pattern observed in other polygynous mammals, sexual size dimorphism is not found in K. rupestris.

Role of Females and Male Potential to Monopolize Them

The active role of females in shaping the social and mating system of a species has long been underestimated, particularly in mammals (Clutton-Brock and McAuliffe 2009). In G. musteloides, however, female behavior is obviously a decisive factor that prevents monopolization by males. Receptive females suddenly start racing around, often changing direction and then stopping abruptly. The females thus attract the attention of all nearby males, which then participate in a mating chase (Rood 1972; Schwarz-Weig and Sachser 1996). The male closest to the female tries to position its chin on the rump of the female while both animals are moving. Other males follow in a row. The alpha male shows this courtship behavior of chin-rump-follows with the highest frequency. When the female stops, the male behind her mounts to copulate. Although a copulating male is sometimes pushed away by another male, in most cases the queuing males wait until copulation has finished. They then resume the mating chase until the female stops again and another male mounts to copulate. A female often does not terminate her soliciting behavior until all males present have copulated with her (Schwarz-Weig and Sachser 1996).

In choice tests with G. musteloides, in which a female had access to 4 males while direct male-male competition and harassment of the female by the males were prevented, 7 of 10 females copulated with 2 males and 1 each with 3 and all 4 males, respectively (Hohoff et al. 2003). Males gained a median number of 12 copulations. These were not in direct succession, however; females switched partners several times, thus mixing up sperm extensively. First-mating males did not gain significantly higher numbers of copulations than 2nd-mating males. However, males with higher within-trial body-mass rank gained a higher number of copulations. A significant correlation was found between the number of male courtship behavior acts elicited and the number of copulations gained. Females thus exercised mate choice by favoring heavier, more-courting males, and simultaneously enabled sperm competition by switching between multiple partners. Because reproductive monopolization of a female by a male is circumvented by the behavior of the female, no selective pressure for the evolution of a large male body mass (as a proxy of fighting strength) acted upon males. The high percentage of multiple paternities found both in captive groups and in the field (M. Asher, University of Münster, pers. comm.; Keil et al. 1999) reflects female promiscuity and inability of males to monopolize females.

Females gain fitness benefits through their polyandrous mating behavior (Keil and Sachser 1998). Proportions of females of G musteloides that became gravid, litter sizes, and birth mass of offspring did not differ between females paired with either 4 males or a single male; however, females paired with a single male had significantly more stillborn offspring and offspring that died before weaning. Thus, females paired with 4 males weaned significantly more young. Although we can only speculate about the mechanisms responsible for this difference, this is a rare example of direct fitness benefits of polyandrous mating in a mammal (Alcock 2005).

Males of G. musteloides, however, do not lack counter-strategies to gain fitness benefits. Keil et al. (1999) determined male dominance ranks in mixed-sex groups and subsequently analysed paternities. Males of all ranks sired offspring. Yet, higher ranking males sired significantly more offspring than lower ranking individuals. The authors suggest that the high level of aggression among males of G. musteloides might suppress the gonadal activity in lower ranking males, giving dominant males an advantage in sperm competition (see also Kruczek and Styrna 2009).

In M. australis mating behavior also is promiscuous. From what is known from the field, females likely play the same role in the mating system as in G. musteloides. Males cannot successfully monopolize females. Competition for parentage of the offspring operates via sperm competition rather than directly (see below). Detailed studies on paternities and possible fitness benefits of female polyandrous mating, as demonstrated in G. musteloides, are needed for M. australis.

The promiscuous behavior of G. musteloides and M. australis contrasts sharply with that of G. monasteriensis. In the latter species females mate with 1 or 2 males, even when they could mate with up to 4 males in a mate-choice apparatus (Adrian 2001; Hohoff 2002). Mate switching is rare and so is extensive sperm mixing. Females of G. monasteriensis obviously exhibit a different behavioral strategy than females of G. musteloides. G. monasteriensis is the only cavy species where the existence of a pair-bond has been demonstrated experimentally (Adrian et al. 2008b). In addition, paternal behaviors (grooming and playing with the young) are performed frequently compared to other species (Adrian et al. 2005). Because G. monasteriensis lives in high-altitude habitats in the Bolivian Andes where climatic conditions can be harsh, a female might need a male partner to raise offspring successfully. For a male it might pay to stay, where the probability of paternity is high, rather than roaming in search of polygynous matings. In the cavies a general link between probability of paternity and paternal behavior exists. Males of G. monasteriensis and C. aperea, both with 1-male systems and no or a low frequency of multiple paternities, frequently show paternal behaviors, whereas no paternal behavior is observed in G. musteloides, which mates promiscuously and has a high frequency of litters with multiple paternity (Adrian et al. 2005).

In C. aperea the tendency of females to copulate polyandrously is much lower than in G. musteloides. However, in a 4-male mate choice experiment, half of the 12 females tested copulated with >1 male (Hohoff 2002). Polyandrously mating females had significantly more stillborn offspring than did monandrously mating females (Hohoff 2002). This analysis, however, was based on the total number of pups. If the number of females with and without stillborn offspring was compared, the difference between polyandrously and monandrously mating females was not significant. In contrast to G. musteloides, in the choice tests with C. aperea 1 male was clearly preferred as social partner, and this male gained the highest number of copulations. Mate switching was much rarer; 4 of the 6 females mating with several males changed their copulation partner only once, and the other 2 females changed them twice and 5 times, respectively. DNA fingerprinting of litters resulting from the mate-choice tests revealed that all pups were sired by preferred partners of females (Hohoff 2002). Mating order did not seem to affect probability of successful fertilization.

In females of C. aperea mate choice is much more pronounced than in G. musteloides. Adrian et al. (2008a) found that male body mass predicted female preferences. In contrast, male body mass was not a significant factor in determining copulation success in another study (Hohoff 2002). In the natural habitat the significance of female mate choice in C. aperea is difficult to evaluate. This species usually lives in habitats with abundant food resources throughout the year (Asher et al. 2004), leading to stable and relatively small home ranges (Table 1); thus several females are found close to each other in small areas. Such a situation generally generates a high environmental potential for male polygamy (Emlen and Oring 1977), because defense of several females against other males becomes feasible and females can be monopolized by powerful males. However, because female home ranges overlap only partially due to the uniform distribution of abundant food plants, males can establish only small harems (Asher et al. 2004). Selection pressure of direct male-male competition has led to the evolution of a sexual size dimorphism in C. aperea. Because the females of the same harem avoid each other, harem males cannot guard and defend all of their females simultaneously, creating opportunities for female mate choice and extragroup copulations.

For K. rupestris, the rock piles where animals take shelter are critical, highly clumped resources. Females rely on these resources for reproduction. Thus, several females gather at single rock piles and create the potential for males to monopolize them. However, in contrast to C. aperea, where female-defense polygyny is present, K. rupestris exhibits resource-defense polygyny. A female can mate outside the rock piles, and whether females frequently perform extragroup copulations is not known.

Adaptations of Reproductive Physiology

Both male and female cavies exhibit adaptations in reproductive physiology to the mating system of their species. In promiscuous mating systems sperm competition can replace contest competition in males. Hence, males are selected for the number and quality of their sperm and—because sperm number depends on gonad size—for high testes and epididymides mass. In single-male mating systems (monogamy and polygyny with female monandrous mating) gonad size is comparatively small (Kenagy and Trombulak 1986; Munshi-South 2007; Ramm et al. 2005) because few sperm suffice to fertilize the females.

The relationship described above is obvious in the cavies (Fig. 2; Table 2). Male relative testes mass in the promiscuous G. musteloides is among the highest recorded for mammals (Cooper et al. 2000; Sachser et al. 1999). Relative testes mass is almost 3 times higher than in the polygynous C. aperea, and relative epididymides mass is more than twice as high (Cooper et al. 2000). Relative testes and epididymides masses of males of G musteloides also are significantly higher than those of males of the monogamous G. monasteriensis (Hohoff et al. 2002). During the reproductive season males of the promiscuous M. australis show very large scrotal testes (K. Lüttmann, University of Münster, pers. comm.), similar to those of G. musteloides. Gonad size is not maintained throughout the year. In G. musteloides (Gutiérrez et al. 1995) and C. aperea (Barlow 1969) gonads regress in winter when reproduction is minimal.

Fig. 2

Males of Microcavia australis, Cavia aperea, and Galea musteloides (left to right) during the breeding season. Whereas the polygynous C. aperea has relatively small testes, those of the 2 promiscuous species are extremely large (photographs by K. Lüttmann, s. Ranft, and M. Asher).

Gonad size is not the only adaptation of male reproductive physiology. Sperm characteristics also differ significantly among cavy species with different mating systems (Cooper et al. 2000). Sperm of C. aperea are larger than those of G. musteloides, swim faster, and are more prone to acrosome damage. Smaller size of sperm of G. musteloides allows the production of even higher sperm numbers (see also Holt 1977). In the monogamous G. monasteriensis the percentages of motile sperm and sperm with intact acrosomes in the cauda epididymidis are distinctly lower than in G. musteloides.

Different sperm characteristics likely also evolved as adaptations to female estrous types. As all cavies, female G. musteloides exhibit postpartum estrus (Weir 1974). If they are isolated at this time or fail to conceive after mating, the next estrus is induced by the presence of a male (Touma et al. 2001). Although Weir (1971), (1973) argued that direct physical contact was necessary to induce estrus, Touma et al. (2001) found that contact through a wire mesh was sufficient to induce vaginal estrus within 2–3 days. Ovulation then occurred spontaneously without stimuli from copulation. Because ovulation takes place several hours after sperm deposition (Weir 1971), sperm of G musteloides will have been selected for longevity rather than velocity.

The induction of estrus is atypical for caviomorph rodents (Weir 1974). It is an advantage if population density is very low and males are not constantly present at a female's living site, which might be the case in some populations of G. musteloides. This species is a habitat generalist and also dwells in deserts and other areas not supporting large groups of animals. Population densities of G. musteloides generally have been reported to be lower than those of M. australis (Contreras and Roig 1978; Rood 1969, 1970, 1972). Induced estrus allows a solitary nongravid female to resume reproduction immediately when she encounters a fertile male.

Microcavia australis has an estrous cycle of approximately 15 days (Rood 1972). Because males are always present in this group-living species, no advantage for induced estrus exists. The same is true for C. aperea. Estrus occurs spontaneously and is followed by spontaneous ovulation and corpus luteum activity. Touma et al. (2001) found a median cycle length of 15 days, whereas other authors reported a mean cycle length of 20.6 days (Rood and Weir 1970; Weir 1970). The cyclicity of female estrus, in combination with a short receptive phase around ovulation and a polygynous-monandrous (1-male) mating system, might have favored the evolution of larger, faster sperm in lower numbers in C. aperea than in G. musteloides.

Environmental Factors

The social organization of animals is dependent in part on the environment. For only some cavy species are data available concerning both the social and mating systems and the ecological conditions under which they live. These will be exemplified here.

Resource distribution is an important factor affecting social organization. K. rupestris, for example, relies on rock piles. These are clumped, and hence females aggregate at this resource, providing the basis for the harem-resource-defense polygyny system of the species (Lacher 1981). In C. magna frequent flooding of the habitat leads to frequent changes of the distribution of areas with food and shelter prompting the cavies to shift their home ranges regularly (Kraus et al. 2003). The unpredictable location of females prevents males from monopolizing them and is a definitive factor for the solitary-promiscuous system.

The distribution of food resources is a codetermining factor for differences in the social organization of C. aperea and G. musteloides. Although both species might be sympatric at some locations (Contreras 1966), the typical habitat of C. aperea is more humid, with abundant food that supports high numbers of cavies with small home ranges. The even distribution of food does not lead to a clumping of females as in K. rupestris, and males of C. aperea can monopolize only 1 or a few females, thus forming pairs or small harems (Asher et al. 2004, 2008). In contrast, the typical habitat of G. musteloides is dryer, and vegetation might be sparse. Under such conditions competition for food is high, and the cavies use larger areas to find enough food leading to the solitary lifestyle of this species and to nonoverlapping home ranges of the females (M. Asher, University of Münster, pers. comm.). Males, in turn, roam even more widely in search of receptive females (Table 1). Because G. musteloides is a habitat generalist, the species might exhibit a different social organization in habitats with more abundant food. Observations by Rood (1972) hint at such a conclusion, because he described multimale-multifemale groups rather than solitary animals. M. australis also occurs sympatrically with G. musteloides at some locations (Contreras 1966; Lüttmann 2006; Rood 1969, 1972). Whereas the latter primarily feeds on grasses, M. australis has exploited another food resource by climbing into shrubs and trees to feed on leaves and fruits (Campos et al. 2001; Monge et al. 1994). This habit of 3-dimensional foraging supports higher population densities compared to those found in G. musteloides, which might have influenced the evolution of a higher social tolerance in M. australis.

In addition to resource distribution, other ecological factors such as predation pressure, climatic conditions, and, in burrowing species, soil characteristics might be important to the development of a given social system. Predation pressure is high in both C. aperea and M. australis (Asher et al. 2004, 2008; Taraborelli et al. 2008). Comparing both species, we can see that different strategies might have arisen to solve the same problem. M. australis was studied at 2 sites in Argentina that differed in predation risk (Taraborelli et al. 2008). At El Leoncito, where plant cover is lower, cavies responded earlier when predator models were presented than at Ñacuñán. Group size of cavies affected vigilance behavior. Although cavies did not forage farther from their burrow when a higher number of individuals was present, individual vigilance decreased with larger group size while total group vigilance increased (Taraborelli 2008). Increased group vigilance can enhance survival in the presence of multiple predators. Vigilance behavior of M. australis is thus in accordance with the hypothesis that the degree of sociality is related to predation risk in caviomorph rodents (Ebensperger 1998; Ebensperger and Blumstein 2006—but compare the situation for C. aperea described below). In M. niata enhanced vigilance for predator detection appears to represent a benefit of group living as well. Marquet et al. (1993) observed cavies to flee into their burrow system upon alarm calling from group members.

Cavia aperea uses a very different antipredator strategy compared to M. australis. This species has no burrows in which to seek shelter from predators. When threatened the animals freeze or dash into dense vegetation where they freeze and allow a predator or a human to approach within a few meters (Asher et al. 2008; Rood 1972). Such a cryptic predator-avoidance strategy precludes the evolution of large groups, because the presence of many active animals in a small area increases the risk of detection by a predator. Foraging groups of C. aperea are small and often only comprise 2 animals (Asher et al. 2004, 2008). Cassini (1991) suggested that foraging in groups is an antipredator strategy of C. aperea similar to that of M. australis. Cavies had longer foraging bouts and scanned the environment marginally less frequently when foraging in a group. However, the observation that usually a harem male and 1 of his females form such foraging groups, whereas 2 females very rarely do so (Asher et al. 2008), might indicate that in addition to increased vigilance mate-guarding is another cause of group foraging in C. aperea.

Potential benefits of group living, other than group vigilance, were tested in M. australis by comparing populations from the 2 sites in Argentina differing in climatic conditions and food resource availability (Taraborelli 2009; Taraborelli and Moreno 2009). At El Leoncito population density was almost 4 times higher than at Ñacuñán (Taraborelli and Moreno 2009). Social groups were formed of 5 individuals on average at El Leoncito and of 3.3 individuals at Ñacuñán. At El Leoncito the index of association among individuals was significantly higher and the rate and proportion of agonistic behavior were lower than at Ñacuñán. El Leoncito is the site with stronger climatic harshness and lower winter temperatures (down to − 4°C— Taraborelli 2009); having larger groups here could serve for more effective social thermoregulation. Stronger association and less agonistic behavior among members of a social group point to higher social tolerance, which might be an adaptation to reduce water and energy loss in this harsh environment.

Sharing costs of burrow construction and maintenance also is often quoted as an advantage of group living in burrowing species. Communal burrowing was identified as a general factor in the evolution of group living in caviomorph rodents (Ebensperger and Blumstein 2006; Ebensperger and Cofré 2001). Comparison of burrow systems of M. australis at El Leoncito and Nacunan showed that they were significantly smaller and less complex at El Leoncito (Taraborelli 2009) where soils are softer and hence energetic costs of burrow digging should be lower. At El Leoncito social groups contained a higher number of animals than at Ñacuñán. No linear relationship was found between group size and areas encompassed by burrow systems. Thus, in contrast to benefits from group vigilance and more effective thermoregulation, the hypothesis that larger groups form to share costs of burrow construction and maintenance is not supported in M. australis.

Concluding Remarks

A high diversity of social and mating systems exists within the cavies, a group with relatively few species. Earlier generalizations that they would exhibit promiscuous mating systems, form no permanent social bonds, and live in noncohesive groups that are nothing more than aggregations associated with natural resources (Rood 1972) have turned out to be inadequate. Phylogenetic constraints, although suggested by Rowe and Honeycutt (2002), appear not to play a fundamental role in the evolution of social organizations in cavies. Consider the remarkable differences that can be found even within the same genus as, for example, between G. musteloides and G. monasteriensis or C. aperea and C. magna (Trillmich et al. 2004). These differences provide strong evidence that ecological factors are more important (Lacher 1982) and have shaped the morphology, physiology, and behavior of individuals. Male and female reproductive and behavioral strategies under given environmental conditions bring about the different social and mating systems whereby females are significantly involved in this process.

Studies on the social life of cavies already have provided valuable insights into the social evolution of this group and of mammals in general. However, further studies could increase our understanding of the mechanisms and functions underlying the formation of different social and mating systems. Research on intraspecific variation in social organization deserves more attention. The wide distribution of some cavy species and their occurrence in various habitats are promising prerequisites. Also, genetic studies on paternities and degrees of kinship within groups could reveal fitness benefits of different behavioral and reproductive strategies that determine what we observe as social and mating systems.


We thank M. Asher for providing unpublished field data on G. musteloides. K. Lüttmann, S. Ranft, and M. Asher took the photographs of the cavies. We are grateful to the guest editors of this special issue, L. Hayes and L. Ebensperger, for inviting us to contribute a paper. Comments by E. Herrera, J. Koprowski, and 2 anonymous reviewers were very helpful in improving earlier drafts of the manuscript. Our studies on cavies have been supported continuously by grants of the German Research Foundation (Deutsche Forschungsgemeinschaft) to NS. Current funding: FOR1232; Sa389/11-1.


  • Special Feature Editor was Barbara H. Blake.

Literature Cited

View Abstract