Introduction

Of Domestic and Wild

Guinea Pigs: Studies in
Sociophysiology,
Domestication, and Social
Evolution
Norbert Sachser
Westflische Wilhelms-Universitt Mnster,
Institut fr Neuro- und Verhaltensbiologie,
Abteilung fr Verhaltensbiologie,
Badestrasse 9, D-48149 Mnster, Germany
Among mammals a majority of each individuals daily expectations, motivations, and behaviors are directed to encounters
with conspecifics. Therefore the knowledge of the genesis, control, and consequences of social interactions is crucial for understanding their social life. We present here our research on
the sociophysiology, domestication, and social evolution of wild
(Cavia aperea and Galea musteloides) and domestic (Cavia
aperea f. porcellus) guinea pigs, which summarizes general rules
for many group-living mammals. It is shown that social interactions have consequences not only for the individuals reproductive success but also for their degrees of stress and welfare. The
way in which individuals interact is controlled not only by the
present environment but also by the previous social experiences which they have gathered during their behavioral development. Furthermore, the study of ontogeny does not begin at
birth, because prenatal social factors acting on pregnant females can also affect the way in which the offspring will interact
when adult. In addition, to understand the genesis of interactions between domesticated animals implies knowledge of the
behavioral and physiological changes which occurred during
the process of domestication. Finally, understanding the social
interactions among individuals of the wild ancestor of the domesticated form requires knowledge of how their behavior patterns were brought about by natural selection during the process of social evolution.

Naturwissenschaften 85, 307317 (1998)

Q Springer-Verlag 1998

In higher vertebrates a majority of each individuals

daily expectations, motivations, and behaviors are
directed to encounters with conspecifics (Hendrichs
1978). Therefore knowledge of the genesis, control,
and consequences of social interactions is crucial for
understanding the social life of mammals. The study
of the genesis involves the effects of evolution, domestication, culture, and ontogeny. Control refers to
short-term regulation by external stimuli impinging
on the animals and by internal physiological and
mental events. The consequences concern the effects on the organism itsself, on its social and physical environment and on its fitness. [This classification follows Dewsbury (1992) in most points.] Until
now mammalian social interactions have rarely been
analyzed in a way that gives equal attention to these
different levels of analysis. Thus at present behavioral biology is far from understanding the social life
of mammals in a comprehensive way.
About 20 years ago we began studying domestic guinea pigs using a descriptive ethological approach
and were fascinated by the unexpected complexity
of these animals social life (Sachser 1986, 1994a;
Sachser and Hendrichs 1982). Soon, however, it became clear that questions concerning the immediate
consequences of social interactions for the organism
in terms of stress and welfare could not be answered
in a satisfactory way by purely behavioral observations. Did subdominants, for example, really suffer a
higher degree of stress than dominants? Was life at
high population densities indeed more stressful than
at low densities? (The answer to both questions is
no, as is shown below.) At the beginning of the
1980s we developed a technique for taking blood
samples from guinea pigs ear vessels in a nonstressful way (Sachser and Prve 1984), from which the
concentrations of several hormones (cortisol, nore307

pinephrine, testosterone) were determined (e.g.,

Sachser 1987). Thereafter the simultaneous recording of ethological and endocrinological parameters
provided a reliable method for studying the physiological consequences of social interactions (e.g.,
Sachser 1994a; Sachser et al. 1994). In addition, it
became clear that adult guinea pigs interactive behavior is significantly influenced by their pre- and
postnatal social environment (Sachser and Kaiser
1996; Sachser and Lick 1991). The first part of this
contribution summarizes our sociophysiological
findings, which by now have emerged as general
rules for many group-living mammals.
Guinea pigs (Cavia aperea f. porcellus) are domesticated animals which are adapted to man-made housing conditions, but which are not found in natural
habitats. In contrast, their ancestor, the wild cavy
(Cavia aperea), is among the most common and
widespread rodents of South America (Stahnke and
Hendrichs 1988). Understanding the genesis of social interactions in the domesticated form requires
consideration not only of the individuals ontogeny
but also of the effects deriving from the process of
domestication. The second part then summarizes the
ethological and physiological differences between
the domestic guinea pig and its wild ancestors. As is
shown below, the process of domestication has led
to typical behavioral developments which are also
found in comparisons between wild and domestic
forms of other species. In addition, some new findings on endocrinological changes are presented
which have clearly helped domesticated animals to
adapt to man-made housing conditions.
Understanding the social interactions among individuals of the wild ancestor of the domesticated
form requires knowing the way in which their behavior patterns have come about by natural selection during the process of social evolution. A most
promising way to analyze the social evolution of the
wild cavy is to compare its patterns of social interactions with those of a closely related species and to
determine the consequences for differential reproduction (the so-called comparative approach). The
third part of this contribution summarizes the initial
results of our comparison between two species of
wild guinea pigs: the wild cavy (Cavia aperea) and
the yellow-toothed cavy (Galea musteloides). This
comparison provides new general insights into the
evolution of mammalian mating systems.

The Guinea Pig: A Study in

Sociophysiology
Social Interactions, Social Relationships,
Social Organization
Various forms of social interactions are observed in
guinea pigs. On the one hand, animals compete with
each other in agonistic encounters. These establish
dominance relationships and stratify the individuals
into relative social positions. On the other hand, social interactions can also proceed in a sociopositive
and/or sexual way, which may result in the establishment of social bonding. The overall structure of
dominance relationships and social bondings constitute the animals social organization.
In our first experiment we studied the effects of increasing population density on the patterns of social
interactions and reproductive success. A small number of guinea pigs (four males and two females)
were placed in a 16 m 2 enclosure. The animals were
allowed to reproduce freely, and after 20 months
there were about 50 individuals in the colony. However, the astonishing point was that even at such
high population numbers the reproductive success
of the females (that is, the number of surviving
offspring/time) did not decline, and the number of
fights between individuals did not increase distinctly
(Sachser 1986, 1994a). The question thus arose:
which mechanisms allow guinea pigs in contrast to
many other mammals (Christian 1975) to cope so
effectively with high population numbers? The
answer to this question is shown in Fig. 1 : guinea
pigs change their patterns of social interactions and
their social relationships, that is, their social organization, when population numbers increase (Sachser
1986).

Fig. 1. Various forms of social organization in guinea pigs. Low density: arrows among males indicate direction of aggressive behaviors.
High density: lines between males and females indicate individual social bondings. Alphas (circled males) dominate nonalphas (noncircled
males). Broken lines, the borders of territories
308

At low population numbers (for example, three

males and three females) the social organization is
characterized mainly by a linear dominance hierarchy among the adult males. Subordinate males retreat whenever a higher ranking conspecific approaches; this largely precludes threat displays and
fights. Individuals of identical rank are never found.
The highest ranking male shows much more courtship behavior towards each of the females than any
other male, and he is probably the father of the
offspring. Social bondings do not exist between
males and females. Among the females there is also
a linear rank order (Thyen and Hendrichs 1990).
However, their agonistic interactions are less pronounced than those among males. Between the
sexes fighting and threat displays do not occur.
When population numbers increase, guinea pigs
change their social organization. Groups of 1015 or
more split into subunits, each consisting of one to
four males and one to seven females. The highest
ranking male of each subunit, the alpha, establishes
long-lasting social bondings toward all females of his
subunit. The alphas guard and defend their females
around estrus, and more than 85% of offspring are
sired by them, as shown by DNA fingerprinting (Fisbeck and Sachser, unpublished data). The lower
ranking males also have bondings with the females
of their subunits, that is, they interact predominantly
with these animals. Alphas of different subunits respect each others bondings, that is, they do not
court other alphas females even if these are receptive. In general, individuals belonging to a given
subunit live in an area that does not overlap with the
area of other subunits. It is in these areas that most
of the social interactions are displayed, and where
the individuals have their resting and sleeping
places. The alphas defend the borders of these areas
around their females estrus (see Sachser 1986,
1994a).
Social organization at high population numbers is
therefore characterized by the following three
points: (a) splitting of the whole group into subunits
provides all individuals with social and spatial orientation; (b) escalated fighting is rare because alphas
respect the male-female bondings of other alphas;
(c) the individuals different social positions are stable over months, and the basic patterns of social organization are independent of individual animals.
Thus the change in social organization from a strictly dominance-structured system at low population
numbers to a system in which long-lasting bondings
are predominant at high numbers can be regarded
as a mechanism for facilitating adjustment to increasing population density.
What are the causes for this change? At low density

the highest ranking male monopolizes all females.

However, the costs, for example, in time and energy
spent in agonistic encounters and in maintaining the
exclusive access to all females within the whole area,
increases with the population of competitors and females. It is economical to defend all females only as
long as the net benefits (in terms of reproductive
success) exceed net costs. When the relationship between benefits and costs becomes unprofitable, an
alternative behavior yielding a higher reproductive
success is preferable, that is, controlling a certain
number of females within a certain area and respecting the same behavior pattern displayed by other
males. Thus the cause for change in the social organization is considered to be the highest ranking
males change in reproductive strategy for maximizing his fitness (Sachser 1986).
At low population numbers the highest ranking
male probably sires most offspring. This reproductive advantage results from his dominance. At high
population numbers the alphas reproduce with
their females. This reproductive advantage results
from two mechanisms: (a) alphas respect the ownership of other alphas, and (b) they dominate nonalphas in agonistic interactions. Thus at low and
high population numbers a polygynous mating system is typical for domestic guinea pigs.

Physiological Consequences of Social

Stratification
Social interactions have a profound effect on the pituitary adrenocortical (PAC) and the sympathetic
adrenomedullary (SAM) systems (e.g., Henry and
Stephens 1977; Sachser 1994a,b; von Holst 1990).
The activation of each of these systems plays a major role in adjusting an individual to social and nonsocial challenges by providing the organism with energy and shifting it into a state of heightened reactivity. Although the short-term or moderate activation of both systems represents an adaptive mechanism to cope with conflict situations, the long-term
hyperactivation of both the SAM and the PAC system is related to the etiology of irreversible injury
and even death (Henry 1982; Henry and Stephens
1977; von Holst 1990). Blood glucocorticoid concentrations cortisol being the major component in guinea pigs and man can be used as a measure of
PAC activity; serum catecholamine concentrations
epinephrine, norepinephrine and adrenal tyrosine
hydroxylase activities are reliable indicators of SAM
activity.
When male guinea pigs living at different population
numbers are compared, they show an increased
309

SAM activity at high densities. PAC activity, in contrast, is not at all affected by population numbers,
that is, a male living in a large colony does not show
higher cortisol concentrations than a male living in a
small group or together with one female only. These
endocrinological data support our behavioral findings: a change in density does not necessarily mean
an increase in social stress for the individuals as long
as a stable social environment is maintained by social mechanisms (Sachser 1990, 1994a).
At high and low population numbers males take different social positions, which are stable over
months. Alphas, for example, always clearly dominate nonalphas of the same subunit. These dominance relationships are independent of place and
time. Alphas bite more often and are less often bitten than nonalphas, and they display far more courtship and sexual behavior than the lower ranking
males (Sachser 1990). Surprisingly, despite these
clear differences in behavior and status, alphas and
nonalphas do not differ significantly in indices of
PAC and SAM activities (Sachser 1987, 1994a;
Sachser et al. 1998), that is, having low social status
does not necessarily entail a higher degree of social
stress than having high social status. Established social relationships resulting in predictable behavior
are seen as a main reason for this, since all individuals live in a stable social organization.
Male-female bonding is very strong in colonies of
guinea pigs (see above; see also Jacobs 1976; Sachser 1986). There are three categories of females for
an individual colony-living male: (a) his bonded females with whom most amicable interactions take
place, (b) females which live in the same colony, and
with whom he is familiar but has no social ties, and
(c) unfamiliar females which live in a different colony, and which he has never before encountered. Interestingly, the males endocrine stress response
when placed in an unfamiliar cage, that is, his increase in PAC activity, is sharply reduced when the
bonded female is present. In contrast, the presence
of a strange female or of one with whom he is merely acquainted has little effect. Thus the effect of various types of relationships differ remarkably, and
substantial social support is given only by the bonding partner (Sachser et al. 1998).
To summarize, guinea pigs establish complex and
long-lasting stable social structures in which the individuals take different social positions. Established
social relationships resulting in predictable behavior
provide all members of the social system with high
security. Bonding partners give social support in
challenging situations. As a consequence, changes in
population parameters such as a rise in population
numbers and stratification into different social po310

sitions do not seem adversely to affect the animals

welfare and health. The question arises, which factors enable guinea pigs to arrange in such a nonstressful and nonaggressive way even at high densities? The answer is: (a) a great tolerance toward
conspecifics, which has been acquired during the
process of domestication (see below); (b) the ability
to establish and to respect dominance relationships;
(c) the ability to establish and to respect social
bondings. As the following section shows, however,
whether these abilities are realized depends on the
social conditions under which the individuals were
reared.

Effects of Social Experiences on Physiology

and Behavior
When two adult males that have grown up in different large colonies are placed into an unfamiliar enclosure in the presence of an unfamiliar female, they
quickly establish stable dominance relationships
without displaying overt aggression. No significant
changes in PAC and SAM activities are found,
either in the dominant or in the subdominant male
(Sachser and Lick 1991). However, such a peaceful stratification into different social positions requires that the opponents have been engaged in
agonistic interactions with older dominant males
around puberty, as is the case in individuals reared
in colonies. In such encounters they experience the
role of a subdominant individual, whereby they acquire the social skills needed to adapt to conspecifics in a nonaggressive and nonstressful way (Sachser
1993; Sachser and Lick 1991; the evidence that the
time around puberty is decisive for social development is summarized in Sachser et al. 1994).
In contrast, a male which grows up singly, or with a
female, is prevented from agonistic interactions
around puberty (since in this species no fighting and
threat displays are found between the sexes). Thus
these social skills cannot be learned. When two
males, both reared in such a way, confront one another in the presence of an unfamiliar female in an
unfamiliar enclosure, high levels of aggressive behavior are displayed, and escalated fighting is frequent. During the first days of the confrontation no
stable dominance relationships are established. We
had to stop about half of the experiments to avoid
irreversible injuries and even death in the losers.
Distinct and persistent increases in PAC activities
were found mainly in the subdominant males
(Sachser and Lick 1991; Sachser et al. 1994). There
is evidence that fighting ability is not what determines the outcome of such contests, but that the

Fig. 2. Serum glucocorticoid concentrations in

colony (solid lines; np6) and individually
(dotted lines; np6) reared males before and
after transfer into an unfamiliar colony for
20 days. **P~0.01, ***P~0.001. (With permission from Sachser and Renninger 1993)

winners are the males that succeed in establishing a

bond toward the female.
The crucial role of social experiences has also been
shown in study taking a different approach (Sachser
and Renninger 1993). Colony- and individually
reared males were introduced singly into unfamiliar
colonies of conspecifics for a period of 20 days. Colony-reared males easily adjusted to the new social
situation. On the first day they explored the new environment but did not court any female, thereby
preventing attacks from the male residents which
had established bondings toward the females. In the
course of the following days they gradually integrated into the social network of the established colonies and could even gain a higher ranking social
position than that which they had had in their native
colonies. In the new colonies changes could not be
determined in either their body weights or in their
PAC and SAM activities on the 1st, 3rd, 6th, 10th,
and 20th days (see Fig. 2). In contrast, individually
reared males were frequently involved in threat displays and fighting. As a consequence they responded to the new situation with substantial decreases in body weight and with extreme increases in
PAC activity (Sachser and Renninger 1993).

Effects of the Prenatal Social Environment

on Physiology and Behavior
Recently we discovered that even prenatal social
factors have a profound effect on animals physiology and behavior when adult (Sachser and Kaiser
1996; Kaiser and Sachser 1998). This experiment
compared daughters whose mothers had either lived

in a stable social environment (SSE) during pregnancy and lactation (SE mothers) or in an unstable
social environment (USE) during this period (UE
mothers). The SSE was made by keeping the group
composition (one male, five females) constant; in
the USE situation every third day two females from
different groups were exchanged. After weaning,
groups of daughters were established from UE
mothers (UE daughters) and groups of daughters
from SE mothers (SE daughters) consisting of four
females each. When adult, the spontaneous behavior of the daughters was recorded in their home
cages. Surprisingly, the UE daughters were characterized by a distinct behavioral masculinization: they
displayed behavioral patterns intensive naso-anal
licking, rumba which are essential parts of the
male courtship behavior that in mixed-sex groups of
guinea pigs are never shown by females (Fig. 3).
This behavioral masculinization corresponds to significantly higher serum testosterone concentrations
in UE than SE daughters. PAC activity does not differ between the two categories of females. Significantly higher adrenal tyrosine hydroxylase activities,
as an indication of SAM activity (Fig. 3), and adrenal weights in UE than SE daughters, however, indicate higher degrees of stress in UE daughters (Kaiser and Sachser 1998). Interestingly, the behavioral
differences between UE and SE daughters are due
to the social instability during pregnancy while the
period of lactation does not seem to be of importance for this phenomenon (Sachser and Kaiser
1996).
What physiological mechanism caused the UE
daughters behavioral masculinization? We favor
the following hypothesis. The experimentally in311

Fig. 3. Relative frequencies of male typical courtship behavior (rumba) and adrenal tyrosine hydroxylase activities of SE and UE daughters. Left,
medians, ranges and 25% and 75% quartiles (each
box np12). Right, meanscSEM (each column
np8). **P~0.01, ***P~0.001. (With permission
from Sachser and Kaiser 1997)

duced change in group composition (i.e., social instability) probably constituted a stressor for the
pregnant guinea pigs. As a consequence, an activation of the PAC system can be expected. This would
result in increased secretion of glucocorticoids and
also androgens originating from the zona reticularis
of the adrenal cortex (Heap 1979). The androgens
would cross the placenta and during a critical phase,
which in the precocial guinea pig is during pregnancy (MacLusky and Naftolin 1981), would cause a
masculinization of the female embryonic hypothalamus (Breedlove 1992; Suchecki and Neto 1991).
This could cause typically male behavior.

Conclusions
What general conclusions can be drawn from our
studies in sociophysiology? When we look at the relevant literature in a comparative way four points
emerge which obviously apply to many mammalian
species:
1. Social stratification is a general phenomenon in
all species studied so far, both in their natural habitats and in captivity. Even in species in which close
cooperation, social tolerance, and amicable relationships predominate, distinct differences in social status are observed among the adult individuals of a
group (e.g., Frame et al. 1978). Taken together, the
findings from guinea pigs and other nonhuman
mammalian species suggest that in stable social systems established dominance relationships result in
predictable behavior, and as a consequence differences in social status do not lead to differences in
stress and health. Thus both high- and low-ranking
312

individuals can live in a nonstressful way (Sachser

1994b). Under conditions of social instability, however, significant increases are found in disease susceptibility, mediated by hormonal responses. Lower
ranking individuals which lose control of their social
position show an extreme increase in PAC activity
that ultimately leads to deficiencies of the immune
system. In contrast, animals that attempt to cope actively with recurrent threats to their social positions
are characterized by a heightened SAM activation,
ultimately resulting in cardiovascular disease. Since
this behavioral pattern is often found in dominants
of species in unstable conditions, it is by no means
always the lower ranking animal which is affected
negatively by its social status (Bradley et al. 1980;
Christian 1975; Henry and Stephens 1977; Kaplan et
al. 1982; Sachser 1994b; Sapolsky 1983; von Holst
1990).
2. In group-living mammals the ability to establish
and respect dominance relationships is a prerequisite to establishing stable social systems. Whether
this ability is realized, however, depends on social
experiences during behavioral development. Social
rearing conditions significantly affect behavioral
patterns in adulthood (e.g., Harlow and Harlow
1962; Immelmann et al. 1982). Furthermore, specific
behavioral patterns in adulthood are related to the
cause of stress and disease (Henry and Stephens
1977; von Holst 1990). It is generally believed that
early experiences in mother-offspring and/or siblingsibling interactions are most important for the development of behavioral strategies in later life (Immelmann et al. 1982). Our present data in guinea
pigs, however, point to the time around puberty as
crucial for the acquisition of those social skills

needed to fit into stable social structures and to adapt to unfamiliar conspecifics in a nonstressful and
nonaggressive way (Sachser 1993; Sachser and Lick
1991; review: Sachser et al. 1994). Thus a causal relationship between social rearing conditions around
puberty, aggressive behavior as adults, and the degree of social stress during the establishment of
dominance relationships is likely, but has not yet
been verified as a general rule.
3. Stress responses can be ameliorated by the presence of members of the same species (e.g., Gust et
al. 1994; Hennessy and Ritchey 1987; Henry 1993;
Stanton et al. 1985). Frequently, however, such social support cannot be provided by any conspecific
(Mendoza et al. 1991; von Holst 1986, 1987), but the
ability to give social support is restricted to bonding
partners, as is seen very clearly in our present data
from adult guinea pigs (Sachser et al. 1998).
4. The study of behavioral ontogeny does not begin
only after birth, because prenatal factors can affect
the offsprings physiological and behavioral development (Gandelmann 1992; Hines 1995). Not only
do artifical physical stressors act on the pregnant female but also mild natural social stressors, such as
instability of the social environment during pregnancy, and these factors can have conspicuous effects,
such as a distinct masculinization of the females behavior (Sachser and Kaiser 1996; Kaiser and Sachser
1998).

Domestic Guinea Pigs and Wild Cavies: A

Study in Domestication

ture emerges: The daughters integrate into the linear dominance hierarchy of the females, which is
age dependent, the oldest female filling the highest
and the youngest female the lowest position. In contrast, the father and his sons become rather incompatible when the sons attain sexual maturity. Then
in most cases they must be taken out of the groups
because otherwise the father will injure and even
kill them (Kuczius and Sachser 1992). Thus one of
the major differences between wild cavies and
domestic guinea pigs is the considerably higher aggressiveness of the wild form, which makes it nearly
impossible to keep adult males together in the presence of females. Correspondingly, domestic male
and female guinea pigs display more sociopositve
behaviors than their wild ancestors. In addition,
overt courtship behavior is more frequently expressed, and a lower threshold for vocalization is
found in the domesticated form (Knzl and Sachser
1997; Rood 1972; Stahnke 1987).
Recently we assessed PAC and SAM activities in
wild cavies and domestic guinea pigs to test for possible endocrinological changes during the process of
domestication (Knzl and Sachser 1997). The animals were housed in groups of one adult male and
two adult females under standardized conditions.
When the males were caught, and a blood sample
was taken from their ear vessels, cortisol titers did
not differ significantly between males of the wild
and domesticated forms. In contrast, serum epinephrine and norepinephrine concentrations were five to
sevenfold higher in the wild cavies than in the
domestic guinea pigs (Fig. 4).

The Behavior and Physiology of Cavia

aperea f. porcellus and Cavia aperea
Guinea pigs were domesticated in South America
30006000 years ago (Herre and Rhrs 1990). Interestingly, the behavior patterns are similar in the
domesticated and the wild animals; distinct differences, however, occur in behavioral frequencies and
thresholds (Knzl and Sachser 1997; Rood 1972;
Stahnke 1987). Social interactions between members
of the domesticated form therefore proceed in a
completely different way from those between wild
cavies. When a male domestic guinea pig is kept together with one or several females, his mature sons
and daughters will integrate rather peacefully into
the social system of the group, and all animals will
cohabitate in a nonaggressive and nonstressful way
(as shown above). When wild cavies are kept in the
same group composition, a completely different pic-

Fig. 4. Serum catecholamine concentrations in males of wild cavies

(np5) and domestic guinea pigs (np7). MeanscSEM. **P~0.01.
(With permission from Knzl and Sachser 1997)
313

Conclusions
It appears that in the guinea pig the process of domestication has led to typical traits reduced aggressiveness, increased tolerance of conspecifics, larger
size that have also been found in comparisons between wild and domestic forms of other species
(Boice 1971; Fox 1978; Herre and Rhrs 1990).
These results are what one would predict if those
who originally domesticated the guinea pig for food
had selected the largest animals for breeding and
eliminated aggressive troublemakers (Rood 1972).
Our physiological finding the decreased reactivity
of the SAM system might explain a further conspicious trait of domesticated animals compared to
their wild ancestors: the distinctly reduced degree of
nervousness and timidy (Hemmer 1982; Price 1973;
Smith 1972) which obviously helps them to adjust to
man-made housing conditions (Knzl and Sachser
1997).
It nevertheless remains surprising, that such an
agreeable animal as the domestic guinea pig should
have been derived from highly aggressive ancestors.
This obvious paradox might be explained not only
by changes that occurred during the process of domestication but also in the following way: the wild
cavy has a wide distribution, from Venezuela and
Colombia to northern Argentina. The process of
domestication obviously took place in the Andes
(Herre and Rhrs 1990). In contrast, the wild cavies
which have been studied so far did not originate
from a mountain region but from the grasslands of
Buenos Aires province. It is well conceivable that
owing to the different ecological conditions under
which the two populations live, distinct differences
in social tolerance may have evolved by natural selection. The guinea pig may therefore have been
domesticated from a population of more tolerant
animals than those which were studied by Rood
(1972), Stahnke (1987), and our group (Knzl and
Sachser 1997; Kuczius and Sachser 1992).

Wild and Yellow-Toothed Cavies: A

Study in Social Evolution
Mating System, Testis Size, and Reproductive
Success
The wild cavy (C. aperea) is one of about ten species
of wild guinea pigs which belong to four genera
(Cavia, Galea, Microcavia, Kerodon) of a subfamily
(Caviinae). In wild cavies adult males are incompati314

ble, whereas females organize themselves into a linear dominance hierarchy (as shown above). Escalated agonistic interactions between the sexes never
occur. As a consequence, a polygynous mating system exists: whenever a female comes into estrus,
only one male is present. This male thus mates with
several females, whereas every female mates with a
single male. The closely related yellow-toothed cavy
(G. musteloides) behaves in a completely different
way: these animals can be kept in large mixed-sex
colonies. Brief phases of high activity alternate several times daily with phases of inactivity, during
which all or nearly all members of the group huddle
together with close bodily contact. The mating system is promiscuous; although the highest ranking
male guards the female during estrous, he is not successful in preventing the lower ranking males from
copulating (Rood 1972; Schwarz-Weig and Sachser
1996). Promiscuous mating is achieved by the females behavior, which attracts the attention of all
the males and makes it impossible for a single male
to monopolize her. The females thus are actively involved in bringing about promiscuity (Schwarz-Weig
and Sachser 1996).
Variation in mating systems is frequently related to
relative testis size (Harcourt et al. 1981; Harvey and
Harcourt 1984; Kenagy and Trombulak 1986). In
species in which only one male mates (polygynous
and monogamous species), low testis weights in relation to body weights are found, whereas males of
promiscuous species are characterized by high relative testis weights. Presuming that testis size is positively correlated with the volume of ejaculate, sperm
counts, and sperm motility (Harvey and May 1989;
Mller 1988), it is reasonable to assume that in
promiscuous species males do not compete for access to females via agonistic encounters but that behavioral competition is replaced by sperm competition.
With this theory in mind, we determined the testis
size in both species (Fig. 5). Indeed, the promiscuous G. musteloides males had distinctly higher absolute and relative testes masses than the polygynous C. aperea although the latter are characterized
by a 25% higher body weight (Schwarz et al. 1994).
The relative testis masses, which were calculated according to the formula of Kenagy and Trombulak
(1986), can be compared directly with those of other
mammals. The values found in C. aperea are within
the range that is typical for polygynous species. In
contrast, the relative testis size of G. musteloides
males is among the highest ever recorded in a terrestrial mammalian species with a known promiscuous
mating system (Kenagy and Trombulak 1986;
Schwarz-Weig and Sachser 1996). Interestingly, the

males which mate with more than one male. Indeed,

females which in a mating experiment were paired
with four males and became pregnant, weaned significantly more surviving offspring than females
which were paired with a single male. Litter sizes did
not differ between the groups (Keil and Sachser
1998). The data support the hypothesis that promiscuous females copulate with several males to induce
sperm competition and/or zygote selection and
thereby increase the viability of their offspring
(Eberhard 1996). The physiological mechanism behind this phenomenon remains to be clarified.

Conclusions

Fig. 5. Relative testis size in mammals (according to Kenagy and

Trombulak 1986). Left, mammalian species with single-male (monogamous and polygynous) and multimale (promiscuous) mating systems. (Data from Kanagy and Trombulak; values shown are only
from those species whose type of mating system is known). Each dot,
the value of a species; columns, medians. Right, individual values of C.
aperea and G. musteloides. Each triangle, the value of a single individual; columns, medians. (With permission from Schwarz-Weig and
Sachser 1996, extended by unpublished data from Schwarz-Weig and
Sachser)

relative testis size of the domestic guinea pig is quite

similar to that of the C. aperea, its wild ancestor
(Sachser, unpublished data). This finding corresponds well to a polygynous mating system which is
typical not only for the wild but also for the domesticated form (see above).
Recent findings from DNA fingerprinting have confirmed what behavioral observations and testis size
suggest for G. musteloides: the existence of a promiscuous mating system. In three mixed-sex groups of
this species (four males and six/seven females) multiple paternity was found on average in more than
80% of the litters, that is, more than one and up to
three males were represented as fathers (Keil et al.,
submitted). To our knowledge, a comparably high
percentage of multiple paternities has been described only in Beldings ground squirrels (Hanken
and Sherman 1981).
Traditionally it was thought that multiple matings
are without any benefits for females. Their role in
achieving promiscuous matings should therefore be
rather passive. From our behavioral observations it
appears, however, that female G. musteloides are actively involved in bringing about promiscuity
(Schwarz-Weig and Sachser 1996). We therefore
studied whether a reproductive benefit exists for fe-

Many questions concerning the proximate mechanisms which help to establish and maintain different
social and mating systems can only be answered under controlled laboratory conditions. The same may
apply when specific hypotheses concerning the evolution of mating systems are tested. In our case such
a laboratory approach revealed that C. aperea and
G. musteloides are two closely related species with
divergent mating systems and functional variations
in testis size. The polygynous mating system of C.
aperea is brought about mainly by the incompatibility of the adult males. The promiscuous mating system of G. musteloides results from the higher compatibility of the males and the females soliciting behavior when receptive. Moreover, it is shown for the
first time that a female mammal has a reproductive
advantage from promiscuous mating. However, a laboratory approach cannot elucidate the ultimate factors which have resulted in different individual behaviors, in different interaction patterns, and thus in
different social and mating systems in these two species of wild guinea pigs. For this, differences in behavior must be related to differences in ecological
conditions (e.g., distribution of food and/or predators) in the natural habitats of C. aperea and G. musteloides. Until now these data are largely lacking.
We have therefore recently begun to carry out such
field studies in cooperation with colleagues from
South America to understand how each species behavior is adapted to the specific environmental conditions under which it lives.

Conclusions
In 1963 Niko Tinbergen, one of the founders of
ethology, formulated the famous four problems of
behavioral biology: the problem of causation, the
315

problem of survival value, the problem of ontogeny,

and the problem of evolution. He insisted that a
comprehensive, coherent science of behavioral biology must give equal attention to each of these and
to their integration. Recently Dewsbury (1992) suggested an up-to-date reformulation of Tinbergens
four problems, that is, to consider them in relation
to the genesis, control, and consequences of behavior. One might differ over whether he was able to
succeed in this. (I personally think that he did, but
for a criticism see Alcock and Sherman 1994). In any
case agreement has existed for decades that a comprehensive understanding of behavioral phenomena
requires research at quite different levels of analysis.
However, in practice such multilevel approaches are
rare. This is a pity, for we have also had to learn in
our research that a comprehensive understanding of
the social life of guinea pigs is possible only if their
behavior is analyzed from different points of view.
In doing this a complex picture emerged: social interactions have consequences for the individuals reproductive success not only but also for their degrees of stress and welfare. The way in which individuals interact is controlled not only by their sex,
age, physiological state, and present environment
but also by the social experiences which they have
undergone. Furthermore, the study of ontogeny
does not begin at birth, because prenatal social factors can also affect the animals behavior when adult. In addition, to understand the genesis of interactions between domesticated animals requires knowledge of the behavioral and physiological changes
which have occurred during the process of domestication. Finally, understanding the social interactions
between the wild ancestors of the domesticated
form includes knowing how their behavior patterns
have evolved by natural selection.
Acknowledgements. I thank Claudia Bger, Trevor G. Cooper, Gerti
Dcker, and Sylvia Kaiser for critical comments on the manuscript.
This work was supported by grants form the Deutsche Forschungsgemeinschaft (Sa389/1-3). All experiments comply with current German
laws.

Alcock J, Sherman P (1994) The utility of the proximate-ultimate dichotomy in ethology. Ethology 96 : 5862
Boice R (1971) Some behavioral tests of domestication in norway
rats. Behav 42 : 198231
Bradley AJ, McDonald IR, Lee AK (1980) Stress and mortality in a
small marsupial (Antechinus stuartii, Mcleay). Gen Comp Endocrinol 40 : 188200
Breedlove SM (1992) Sexual differantiation of brain and behavior. In:
Becker JB, Breedlove SM, Crews D (eds) Behavioral endocrinology. MIT Press, Cambridge, pp 3971
316

Christian JJ (1975) Hormonal control of population growth. In: Eleftheriou BE, Sprott RLS (eds) Hormonal correlates of behavior, vol.
1. Plenum, New York, pp 205274
Dewsbury D (1992) On the problems studied in ethology, comparative psychology, and animal behaviour. Ethology 92 : 89107
Eberhardt WG (1996) Female control: sexual selection by cryptic female choice. Princeton University Press, Princeton
Fox MW (1978) The dog: its domestication and behavior. Garland
STPM, New York, pp 319
Frame LH, Malcolm JR, Frame GW, van Lawick H (1978) Social organization of African wild dogs (Lycaon pictus) on the Serengeti
plains, Tanzania 19671978. Z Tierpsychol 50 : 225249
Gandelmann R (1992) Psychobiology of behavioral development. Oxford University Press, New York
Gust DA, Gordon TP, Brodie AR, McLurie HD (1994) Effect of a
preferred companion in modulating stress in adult female rhesus
monkeys. Physiol Behav 44 : 681684
Hanken J, Sherman PW (1981) Multiple paternity in Beldings ground
squirrel litters. Science 212 : 351353
Harcourt AH, Harvey PH, Larson SG, Short RV (1981) Testis weight,
body weight and breeding systems in primates. Nature 293 : 5557
Harlow HF, Harlow MK (1962) Social deprivation in monkeys. Sci
Am 207 : 136146
Harvey PH, Harcourt AH (1984) Sperm competition, testes size, and
breeding systems in primates. In: Smith RL (ed) Sperm competition
and the evolution of animal mating systems. Academic, Orlando,
pp 589600
Harvey PH, May RM (1989) Out for the sperm count. Nature
337 : 508509
Heap RB (1979) Hormone und Graviditt. In: Austin CR, Short RV
(eds) Fortpflanzungsbiologie der Sugetiere, vol 3. Parey, Berlin,
pp 6792
Hemmer H (1982) Domestikation Verarmung der Merkwelt. Vieweg, Braunschweig Wiesbaden
Hendrichs H (1978) Die soziale Organisation von Sugetierpopulationen. Sugetierkdl Mitt 26 : 81116
Hennessy MB, Ritchey RL (1987) Hormonal and behavioral attachment responses in infant guinea pigs. Dev Psychobiol 20 : 613625
Henry JP (1982) The relation of social to biological processes in disease. Soc Sci Med 16 : 369380
Henry JP (1993) Psychological and physiological responses to stress:
the right hemisphere and the hypothalamo-pituitary-adrenal axis,
an inquiry into problems of human bonding. Int Physiol Behav Sci
28 : 368386
Henry JP, Stephens PM (1977) Stress, health, and the social environment. A sociobiologic approach to medicine. Springer, New York
Heidelberg Berlin
Herre W, Rhrs M (1990) Haustiere zoologisch gesehen. Fischer,
Stuttgart
Hines M (1995) Human behavioral sex differences: a role for gonadal
hormones during early development? Psychol Bull 118.1 : 55107
Immelmann K, Barlow GW, Petrinovich L, Main M (eds) (1982) Verhaltensentwicklung bei Mensch und Tier. Das Bielefeld-Projekt.
Parey, Berlin
Jacobs WW (1976) Male-female associations in the domestic guinea
pig. Anim Learn Behav 4 : 7783
Kaiser S, Sachser N (1998) The social environment during pregnancy
and lactation affects the female offsprings endocrine status and behaviour in guinea pigs. Physiol Behav 63 : 361366
Kaplan JR, Manuck SB, Clarkson TB, Lusso FM, Taub DM (1982)
Social status, environment, and atheriosclerosis in cynomolgus
monkeys. Arteriosclerosis 2 : 359368
Keil A, Sachser N (1998) Reproductive benefits from female promiscuous mating in a small mammal. Ethology (in press)
Kenagy GJ, Trombulak SC (1986) Size and function of mammalian
testes in relation to body size. J Mammal 67 : 122

Kuczius K, Sachser N (1992) Soziale Organisation und Verhalten von

Wildmeerschweinchen (Cavia aperea). Z Sugetierkd 57
[Suppl]:3637
Knzl C, Sachser N (1997) Verhaltensbiologische Untersuchungen
zur Domestikation des Meerschweinchens. In: Aktuelle Arbeiten
zur artgemen Tierhaltung 1996. KTBL no 376, Darmstadt, pp
99109
MacLusky NJ, Naftolin F (1981) Sexual differentiation of the central
nervous system. Science 211 : 12941303
Mendoza SP, Lyons DM, Saltzman W (1991) Sociophysiology of
squirrel monkeys. Am J Primatol 23 : 3754
Mller AP (1988) Ejaculate quality, testis size and sperm competition
in primates. J Hum Evolut 17 : 479488
Price EO (1984) Behavioral aspects of animal domestication. Q Rev
Biol 59 : 132
Rood JP (1972) Ecological and behavioural comparisons of three genera of Argentine cavies. Anim Behav Monogr 5 : 183
Sachser N (1986) Different forms of social organization at high and
low population densities in guinea pigs. Behaviour 97 : 253272
Sachser N (1987) Short-term responses of plasma norepinephrine,
epinephrine, glucocorticoid and testosterone titers to social and
non-social stressors in male guinea pigs of different social status.
Physiol Behav 39 : 1120
Sachser N (1990) Social organization, social status, behavioural strategies and endocrine responses in male guinea pigs. In: Balthazart J
(ed) Hormones, brain and behaviour in vertebrates. Comparative
Physiology, vol. 9. Karger, Basel, pp 176187
Sachser N (1993) The ability to arrange with conspecifics depends on
social experiences around puberty. Physiol Behav 53 : 539544
Sachser N (1994a) Sozialphysiologische Untersuchungen an Hausmeerschweinchen. Gruppenstrukturen, soziale Situation und Endokrinium, Wohlergehen. Parey, Berlin
Sachser N (1994b) Social stratification and health in non-human
mammals a case study in guinea pigs. In: Ellis L (ed) Social stratification and socioeconomic inequality, vol 2. Praeger, Westport,
pp 113121
Sachser N, Hendrichs H (1982) A longitudinal study on the social
structure and its dynamics in a group of guinea pigs (Cavia aperea f.
porcellus). Sugetierkdl Mitt 30 : 227240
Sachser N, Kaiser S (1996) Prenatal social stress masculinizes the females behaviour in guinea pigs. Physiol Behav 60 : 589594
Sachser N, Kaiser S (1997) The social environment, behaviour and
stress a case study in guinea pigs. Acta Physiol Scan 161 [Suppl
640]:8387
Sachser N, Lick C (1991) Social experience, behaviour and stress in
guinea pigs. Physiol Behav 50 : 8390

Sachser N, Prve E (1984) Short-term effects of residence on the testosterone responses to fighting in alpha male guinea pigs. Aggr Behav 10 : 285292
Sachser N, Renninger SV (1993) Coping with social conflict: the role
of rearing conditions in guinea pigs. Ethol Ecol Evol 5 : 6574
Sachser N, Drschlag M, Hirzel D (1998) Social relationships and the
management of stress. Psychoneuroendocrinol (in press)
Sachser N, Lick C, Stanzel K (1994) The environment, hormones and
aggressive behaviour a five-year-study in guinea pigs. Psychoneuroendocrinology 19 : 697707
Sapolsky R (1983) Endocrine aspects of social instability in the olive
baboon (Papio anubis). Amim J Primat 5 : 365379
Schwarz E, Kuczius K, Sachser N (1994) Paarungssystem, Geschlechtsdimorphismus und Hodengewichte: Ein Vergleich von
Wild- und Wieselmeerschweinchen. Verh Dtsch Zool Ges 87 : 57
Schwarz-Weig E, Sachser N (1996) Social behaviour, mating system
and testes size in cuis (Galea musteloides). Z Sugetierkd
61 : 2538
Smith RH (1972) Wildness and domestication in Mus musculus. A
behavioral analysis. J Comp Physiol Psychol 1 : 2229
Stahnke A (1987) Verhaltensunterschiede zwischen Wild- und Hausmeerschweinchen. Z Sugetierkd 52 : 294307
Stahnke A, Hendrichs H (1988) Meerschweinchenverwandte Nagetiere. In: Grzimek B (ed) Grzimeks Enzyklopdie Sugetiere.
Kindler, Munich, pp 314357
Stanton ME, Patterson JM, Levine S (1985) Social influences on conditioned cortisol secretion in the squirrel monkey. Psychoneuroendocrinology 10 : 125134
Suchecki D, Neto JP (1991) Prenatal stress and emotional response of
adult offspring. Physiol Behav 49 : 423426
Tinbergen N (1963) On aims and methods of ethology. Z Tierpsychol
20 : 411433
Thyen Y, Hendrichs H (1990) Differences in behaviour and social organization of female guinea pigs as a function of the presence of a
male. Ethology 85 : 2534
Von Holst D (1986) Psychosocial stress and its pathophysiological effects in tree shrews (Tupaia belangeri). In: Schmidt TH, Dembroski
TM, Blmchen G (eds) Biological and psychological factors in cardiovascular disease. Springer, Berlin Heidelberg New York, pp
476489
Von Holst D (1987) Physiologie sozialer Interaktionen Sozialkontakte und ihre Auswirkungen auf Verhalten sowie Fertilitt und Vitalitt von Tupajas. Physiologie Aktuell 3 : 189208
Von Holst D (1990) Stre ein Problem sozialer Gesellschaften.
Mnch Med Wochenschr 132 : 811819

317