Title: Aggression and familiarity as factors in mate selection in Peromyscus polionotus and Peromyscus maniculatus
Full Citation
Permanent Link: http://ufdc.ufl.edu/UF00102855/00001
 Material Information
Title: Aggression and familiarity as factors in mate selection in Peromyscus polionotus and Peromyscus maniculatus
Physical Description: Book
Language: English
Creator: Webster, Daniel George, 1948-
Copyright Date: 1983
 Record Information
Bibliographic ID: UF00102855
Volume ID: VID00001
Source Institution: University of Florida
Holding Location: University of Florida
Rights Management: All rights reserved by the source institution and holding location.
Resource Identifier: oclc - 09818074
ltuf - ABZ0641

Full Text

MATE SELECTION IN Peromyscus polionotus
AND Peromyscus maniculatus






To my wife Carole and my daughter Danielle
and to my parents


Many individuals have generously contributed to this work, and

while these individuals deserve much of the credit for its completion,

any faults which may remain are my own. I thank Dr. D. A. Dewsbury for

his support and suggestions, for providing a critical scientific

atmosphere in which to conduct research, and for his example as a

scientist. I would also like to thank Drs. H. J. Brockmann, C.

VanHartesveldt, and W. B. Webb for their encouragement and support, and

for contributing of their time and knowledge. I particularly thank

Dr. M. E. Meyer whose consistent encouragement and support were a major

factor in my completion of the Graduate Program and the present work.

Several of my fellow students also contributed to this work. My

appreciation to D. Kim Sawrey, Bruce Ferguson, Larry Shapiro, Allen

Hodges, and Denis Baumgardner, for their support, encouragement and

critisms. I would especially like to thank Dean Williams for his time,

and for his patience in explaining some of the mysteries of electronics

to a neophyte. Our animal caretakers, T. C. Fryer and I. Washington,

also deserve my appreciation for the excellent care they have given my

animals. Last, but most of all, I would like to thank my wife Carole

who has supported me through the bad times and the good, and my daughter

Danielle whose patience and understanding exceed her years.

This research was funded in part through NSF Grant BNS78-05173 to

Dr. D. A. Dewsbury.



ACKNOWLEDGEMENTS ................................................. i

ABSTRACT ..................................... ...... .............. vi


I INTRODUCTION .............................................. 1

Aggression as a Factor in Mate Selection .................. 4
Familiarity as a Factor In Mate Selection ................11
Kin Familiarity ...................................... 12
Familiar Others .. .................................... 14


Selection of Species ....................**............... 18
Approaches to the Study of Social Preference ............. 19
General Experimental Information .............*........... 22
General Methods ........................... ** ......... *25
Subjects ........... -..............................- **25
Apparatus ......................................... ** ..26
Seminatural apparatus ................................ 26
Preference apparatus ................................. 29
Aggression apparatus ................................. 32

I I I EXPERIMENTS ................ ..................* ........ 35

Seminatural Experiments ............................. ... 35
Introduction ........................................ 35
Subjects ................... ......................... *36
Procedure ............................... *.......... ** 36
Results .............................. ............... *40
Discussion ........................................ **61
Aggression and Familiarity Preference Tests ............. 63
Introduction ....... ............... ................ .. 63
Subjects ............ ............................. ***63
Procedure ................. .......................... *64
Results .................. ..................... . .... 67
Discussion ......... .............................. ***78

Sibling Preference Tests ............................... 81
Introduction ........................................... 81
Subjects ....... .................................... 82
Procedure ........................................... 82
Results .............................................. 83
Discussion ........................................... 90

IV GENERAL DISCUSSION .................................... 92

Aggressive Ability as a Factor in Preference ............. 93
Ecology, Mating System, and Aggressive Abillty ......... 94
Peromyscus maniculatus ............................... 95
Peromyscus pollonotus ............................... 101
Aggression and Mate Selection ......................... 106
Intraspeclfic aggression ............................ 106
Interspecific aggression ........................... 108
Early breeding ..................................**** 109
Bruce effect .......... ............................ *109
Heritable aggressive ability ........................ 111
Familiarity as a Factor In Preference ................... 113
Prior History ......................................... 114
Ecology and Social System -........................... 115
Peromyscus maniculatus .........**.................. 115
Peromyscus polionotus ........*..................... 117
Kinship as a Factor in Preference ......................* 119
Inbreeding In Peromyscus maniculatus .................* 120
Evidence for and against inbreeding ................. 120
Ecological and social factors ..................... 121
Inbreeding In Peromyscus pollonotus ................... 125
Evidence for and against inbreeding ................. 125
Ecological and social factors ..................... 126
Evolution of Monogamy in Peromyscus Pollonotus ......... 129
Summary ..... ....................... .................... 133

REFERENCES .......... .............. .............** ..... 137

BIOGRAPHICAL SKETCH ....................................* 162

Abstract of Dissertation Presented to the Graduate Council
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy

MATE SELECTION IN Peromyscus polionotus
AND Peromyscus maniculatus


Daniel George Webster

April 1983

Chairman: Donald A. Dewsbury
Major Department: Psychology

Aggressive ability and familiarity were examined as factors in

the social preference and mate selection of males and females of the

monogamous species polionotus, and the polygamous species P.

maniculatus. The aggressive behaviors and nesting behavior of P.

polionotus were observed in a seminatural apparatus; factors assessed

were 1) familiarity based on cohabitation, and 2) aggressive ability

as determined through aggressive interactions. Groups observed were

composed of either two pairs of familiar opposite-sexed individuals,

or two unfamiliar animals of each sex. Preferences of both species

were assessed in an automated preference apparatus. In addition to

the two factors assessed in the seminatural apparatus, the effects of

familiarity based on relatedness were assessed in the preference

apparatus. Measures recorded were the number and duration of visits

to stimulus animals. In the seminatural setting E. pollonotus of

both sexes displayed aggression, and nested more frequently with the

more aggressive of two opposite-sexed individuals. Males of this

species also exhibited a behavior, aggressive digging, that may

signal their aggressive status. Peromyscus polionotus and P.

maniculatus of both sexes exhibited evidence of preference for more

assertive opposite-sexed individuals (high rather than low tendency

to interact) in the preference apparatus. Peromyscus maniculatus of

both sexes also displayed preference for individuals of the opposite

sex that they had previously been housed with, but such familiarity

did not affect preference in P. polionotus. The lack of a

significant effect of familiarity on preference in P. polionotus was

consistent with the nesting behavior of this species in the

seminatural apparatus. Differences in the responses of P.

maniculatus and P. polionotus to familiar individuals may be based on

differences in the opportunities individuals of these species have to

use this factor in mate selection.

Peromyscus polionotus females demonstrated significant

preference for siblings over nonsiblings, and males tended to display

higher sibling than nonsibling scores. Inbreeding in EP polionotus

may permit individuals of this species to found populations in

isolated patches of favorable habitat. Lack of significant

preference by P. maniculatus for siblings or nonsiblings was

interpreted as due to competing preference responses in this species.


The success of an organism in leaving a numerous
posterity is not measured only by the number of
its surviving offspring, but also by the quality
or probable success of these offspring. It is
therefore a matter of importance which particular
Individual of those available is to be their other
parent.(Fisher, 1958, p. 143)

Few factors are as important to an individual's reproductive

success as is the selection of a mate. Mate selection involves more

than simply the identification of potential partners as to species and

sex. In order for an individual to maximize Its future representation

in the gene pool it must also select the "best" possible mates on the

basis of a variety of other considerations.

There has been much theoretical speculation on the proximate and

ultimate bases of mate selection and on the relative Importance of

mate selection to different mating systems. In the last decade a

considerable effort has been made to collect empirical data on mate

selection, but large gaps still remain in the existing data which must

be filled before many important theoretical problems can be resolved.

Particularly lacking are data on male choice and data on mate

selection in mammalian species, especially those considered to be

monogamous. The lack of data on male choice probably stems In large

part from a traditional emphasis on female choice and male-male

competition (Bateman, 1948; Daly & Wilson, 1978; Trivers, 1972;

Williams, G. C., 1966) and the belief that male choice was either

nonexistent or negligible In the face of this competition. More

recently, however, It has been suggested that mate selection should be

of some consequence to males as well as females (e.g., Dewsbury,

1982c). The lack of data on monogamous mammalian species may be due In

part to two factors that make it somewhat difficult to observe these

species: 1) monogamy appears to be uncommon in mammals (Alexander,

1974; Crook, 1977; Kleiman, 1977; Orians, 1969) and 2) most mammals,

especially the smaller species, are largely nocturnal (Vaughan, 1978).

Study of the proximate factors Involved in mate selection is

extremely important to the resolution of issues about the evolution and

relative importance of mate selection in various mammalian taxa.

Factors that have been proposed to be of major importance to mate

selection are of two general types: those related to the genotype of a

potential mate, such as genetic quality (Trivers, 1972, 1976; Zahavi,

1975), and relatedness (Maynard Smith, 1956); and those related to

resources, such as the ability to accrue resources (Trivers, 1976) and

parental investment (Bateman, 1948; Trivers, 1972). The relevance of

any particular factors as criteria In mate selection will vary among

species, between sexes, and across mating systems, as a function of

differences in a group of interrelated variables Including ecological

factors (Borgia, 1979; Emlen & Orlng, 1977; Halliday, 1978).

Differences in the degree to which individuals of various species

utilize particular factors as criteria in mate selection would be

expected to reflect differences in the adaptiveness of those criteria

to mate selection in those species. Comparative studies of mate

selection or social preference in different species, therefore, can

provide an empirical basis for evaluation of hypotheses on the

importance of various factors to mate selection under different social

or mating systems. Such comparisons would be most effective when the

species compared were closely related (King, 1970) so that the species

do not differ in so many respects as to obscure the relationships of


Two factors of the sort that might be expected to be of broad

importance as criteria for mate selection across most species, but may

be expected to vary in importance among species, are the aggressive

ability of a potential mate, and that animal's degree of familiarity

with the individual expressing choice. An individual of high

aggressive ability could be defined as one that is highly competent in

the performance of aggressive behavior (i.e., displays, threats, and

fights). Individuals of high aggressive ability would be expected to

perform well in competition for contested resources and defense of

mates and/or offspring. Familiarity can be of two types, these are 1)

familiarity gained through exposure to other unrelated individuals, and

2) familiarity with kin. In the first sense, familiarity may provide a

basis for evaluation of former mates or a means of discriminating

between two potential mates. In the second sense, familiarity may

provide a basis for avoidance of inbreeding, or as a yardstick for

comparison of potential mates (Bateson, 1978, 1980).

This study was designed to provide data on the importance of

aggression and familiarity as factors affecting mate preference in

males and females of two closely related species of muroid rodents with

different mating systems: the monogamous species Peromyscus polionotus

and the polygamous species Peromyscus maniculatus. In light of the

relative lack of data on mate selection in monogamous species the major

focus of this study is on P. polionotus. The discussions that follow

provide a brief review of relevant theoretical factors in mate

selection, and a demonstration of the importance of aggression and

familiarity as factors in mate selection, with an emphasis on mammalian


Aagression as a Factor in Mate Selection

Darwin (1859) recognized that some males could gain a reproductive

advantage over other males by defeating them in fights for females. It

has been suggested (Bateman, 1948) that the evolution of male-male

competition had its basis in differences in male-female strategies of

investment in gametes. Males are generally considered to invest little

energy in the production of gametes, whereas females invest a great

deal (Bateman, 1948; Orians, 1969; Stacey, 1982; Trivers, 1972).

Because of their larger investment females are valuable to males, and

the probability that females will obtain mates is high, but their

reproductive success will be limited by the number of gametes they can

produce. Males, however, because they invest little energy in the

production of individual gametes, can afford to produce large numbers

of gametes--with which they could potentially fertilize large numbers

of females. A male's reproductive success, therefore, may be greatly

influenced by the number of females he mates with--and males may be

expected to compete to fertilize females (see however, Dewsbury, 1982c;

Nakatsuru & Kramer, 1982). This argument was extended by Trivers

(1972), who restated it in a more general form based on the overall

relative level of parental investment of the two sexes, which he

hypothesized to determine the intensity of male-male competition in

species as well as the form of the mating system (see however, Kleiman

& Malcolm, 1981, p. 371; Wickler & Seibt, 1981). The level of

success achieved by a male in competitive mating may often depend upon

his ability to dominate other males. This notion has been examined in

a wealth of studies, recently reviewed by Dewsbury (1982b), on the

relationship between "dominance" and various aspects of reproduction.

A male's ability to conquer other males, however, also provides

females witn a basis by which to judge him against other males--a

basis for "female choice" (Darwin, 1874). Borgia (1979) has suggested

that a female's best indication of the relative overall genetic quality

of a male is provided through his aggressive interactions with other

males. Females choosing aggressive males, or allowing such males to

mate with them, may in effect be selecting "good genes" for their

offspring (Maynard Smith, 1956; Trivers, 1972). Selection for good

genes has also been suggested as a basis for the evolution of extra-

vagant sexually dimorphic characteristics (Fisher, 1930), as a basis

for lek behavior (Borgia, 1979), and as a basis for female choice in

Drosophila (Partridge, 1980). The major obstacle to the use of

heritable factors as a basis for choice is the problem of "using up"

the genetic variance for a trait (Krebs & Davies, 1981; Maynard Smith,

1978). Several factors, however, have been suggested to act to

maintain genetic variance; these include 1) advantages for

heterozygotes (Borgia, 1979), 2) variation in the optimal genotype in

space and time (Krebs & Davies, 1981), 3) factors such as chronic

parasitism, that may result in cyclic changes in the optimal genotype

(Hamilton & Zuk, 1982), and 4) rate of mutation in polygenic characters

(Lande, 1976).

The northern elephant seal (Mirounaa angustirostris) provides an

example of the importance of male aggressive ability in female choice

in a natural setting. During the breeding season males of this species

establish dominance hierachies which are maintained through threat and

combat. Dominant males guard groups of females from other males, and

account for the majority of first copulations (LeBeouf, 1974; LeBoeuf &

Peterson, 1969). Females help insure that they will be inseminated by

an aggressive dominant male by vocalizing loudly if a subordinate male

attempts to mount. The dominant male, alerted by the female, drives

off the subordinate male and copulates with the female (Cox & LeBoeuf,

1977). The authors note that females of many polygynous species might

be expected to incite male-male competition in this manner. Cox (1981)

has observed that female elephant seals are less likely to vocalize If

the male mounting them has just displayed dominance over another male

and suggested that females may thus select for males that frequently

display their aggressive ability. Cox hypothesized that in general "in

species where social status of males Is correlated with their genetic

fitness, female choice is likely to be based on social signals which

are used in competition between males" (p. 197). Similar hypotheses

have been proposed by Borgia (1979) and Alexander (1975). Cox (1981)

has provided examples of several species in which females appear to use

male's aggressive signals toward each other as a basis for choice. The

list consists of a fairly diverse array of species including

territorial birds (Armstrong, 1973; Thorpe, 1961), tree frogs (Whitney

& Krebs, 1975a, 1975b), sticklebacks (Tinbergen, 1951), and a

lek-forming bird, the ruff Philomachus pugnax (Hogan-Warburg, 1966).

Male aggressive ability need not be heritable or manifest at the

time of mating to be an important factor in female choice. Male

differences in aggressive ability, for example, will be related to

differences in their ability to acquire and hold territories (Brown, J.

L., 1964). A female choosing a male with a superior territory is in

effect also choosing a male that has been able to first obtain that

territory in competition with other males, and further to maintain it

in the face of threats from other males.

Female preference for aggressive males has also been demonstrated

in laboratory choice tests. Sexually experienced female brown lemmings

(Lemmus trimucronatus) in estrus were found, in olfactory choice tests,

to prefer dominant over subordinate males (Huck, Banks, & Wang, 1981).

Female preference was also predictive of male performance in later

dominance tests; estrous females again preferred dominant males while

diestrous females preferred subordinate males. The authors found that

dominant males had heavier testes and higher testosterone levels than

subordinates, and they hypothesized that female choice might be based

on differences in androgen dependent male odors. Estrous females were

also found to exhibit more copulation with dominant males in

tether-choice tests (Huck & Banks, 1982).

Costanzo and Renfrew (1977) studied the preference of sexually

experienced and naive female rats for dominant and subordinate male

odors. Ovariectomized sexually naive females displayed no preference

in hormonally induced estrus. Ovariectomized sexually experienced

females preferred dominant males when not injected with hormones, but

subordinate males when in hormonally induced estrus.

As discussed earlier a female may often invest relatively more than

a male in offspring, and a female's reproductive success may often be

more limited than a male's. Because of these factors females are

generally considered to be more choosey than males when selecting mates

(Bateman, 1948; Burley, 1977, 1981; Daly & Wilson, 1978; Trivers,

1972; Williams, G. C., 1966) and the majority of studies of social or

mate preference have been studies of female choice. Male reproductive

success, however, depends on not only the number of mates males acquire

but the quality of these mates as well (Wade, 1979). Rails (1976) has,

for example, suggested that larger females may often be better mothers.

Factors such as this might be considered trivial if males were capable

of inseminating an unlimited number of females. However, although

males appear to produce an enormous number of sperm and therefore to

be capable of inseminating an enormous number of females, these sperm

are emitted as ejaculates--of which a male may produce only limited

numbers (Dewsbury, 1982c; Nakatsuru & Kramer, 1982). Mate selection

may, therefore, often be of some consequence to males as well as

females (Dewsbury, 1982c).

The aggressive ability of females may be an important criterion for

male choice. Females with high aggressive ability may be more capable

than females with low ability in such behaviors as defense of young or

nest sites, or in competition for food items. Males of some species

have also demonstrated an ability to discriminate the aggressive status

of other individuals, and therefore the potential to use this factor in

mate selection. Male rats, for example, spend more time investigating

odor from a dominant male than from a subordinate (Krames, Carr, &

Bergman, 1969). Male mice also discriminate between the odors of

dominant and subordinate males (Carr, Martorano, & Krames, 1970) and

investigate areas marked by dominant males less than those marked by

subordinate males (Jones & Nowell, 1973). Both male and female

saddle-backed tamarins investigate dominant male scent marks more than

those of subordinate males (Epple, 1973, 1974).

Although the majority of preference studies to date have been

conducted with polygamous species, there is no reason to believe that

members of monogamous species should be any less adept at

discrimination based on differences in aggressive status, or that the

ability to make such discrimination should be any less useful to

members of monogamous species. Aggression serves many of the same

functions in monogamous species as in polygamous species. Although

male-male competition and aggression near the time of copulation may be

less important in monogamous than in many non-monogamous species,

aggression serves many functions at other times in an organism's

lifetime, and success in these other aggressive encounters will be just

as important to monogamous as to polygamous males. For example,

defense of resources critical to raising young, and in many cases to

initially attracting a mate, is hypothesized by some authors to be

universally displayed by males of monogamous species. Kleiman (1977)

states that "The male's territorial defense, which prevents the over-

utilization of necessary resources, is practiced by males of all

monogamous species" (p. 54). Kleiman (1977) has also indicated that

females of monogamous mammalian species may commonly be as involved as

males in territorial defense and contrasts this behavior with a lack

of territorial defense by non-monogamous females. This hypothesis is

similar to one previously proposed by Wilson (1975) as one of the

biasing ecological conditions for monogamy that two adults are

required to defend valuable resources contained in the territory. If

monogamous females help to acquire and defend resources then monogamous

males may do well to choose aggressive females as mates.

An additional consideration, related to the possible function of

aggression as a factor in mate selection in monogamous species, is the

"early breeding" hypothesis proposed by Darwin (1874). According to

this theory the most healthy, more dominant members of a species will

come into breeding condition earlier in the season, and establish

territories earlier, than the less vigorous subordinate individuals.

The more dominant individuals of each sex should then choose each other

as the "better" mates in preference to subordinate individuals. Two

species whose behavior may support the occurrence of this form of

selection are the artic skua (Stercorarius parasiticus) and the

mourning dove (Zenaldura macroura). Individuals generally pair for

life In the arctic skua (O'Donald, 1959). Darker males of this species

tend to breed earlier than birds of light or intermediate phenotype,

and females breeding for the first time are more successful with dark

males. Early breeding and large territory size are both correlated

with early hatching, and success of new pairs with dark males may be

related to male ability to hold large territories (Davis & O'Donald,

1976). Dominant male mourning doves prefer dominant females in

experiments with penned populations (Goforth & Baskett, 1971).

Dominant pairs breed earlier than subordinate pairs, and more offspring

of dominant pairs survive.

Based on theoretical considerations, and on the results of

research conducted to date, predictions can be made as to the relative

importance of aggression as a factor In mate selection in species with

polygamous and monogamous mating systems. Males that are above average

in their aggressive abilities should generally be preferred as mates by

females of monogamous and polygamous species. Females of monogamous

species, because they may form long-term pair bonds and consequences of

their choice may therefore be more long-term, might be expected to

display stronger preference than polygamous females. Males of

monogamous species might be expected to display preference for females

with higher aggressive ability because, among other considerations,

such females may be in breeding condition earlier and might be expected

to better share responsibility for territorial defense. Polygamous

males might also prefer females with higher aggressive ability because

such females may be better able to defend young or resources necessary

for raising young, but preference should not be as critical for these

males as for monogamous males.

Familiarity as a Factor in Mate Selection

Familiarity could be an important factor in mate selection In

several ways. It would be expected, for example, that early

familiarity with conspecifics should aid Individuals In discriminating

between members of their own and other species. Familiarity could

additionally be an important factor in recognition of kin, and

therefore a factor of importance in kin-selection and in avoidance of

inbreeding. It may also be of importance to animals of many species to

be capable of discriminating between unrelated strange and familiar

individuals. Two aspects of familiarity, recognition of kin and of

familiar others, will be discussed further.

Kin Familiarity

Bateson (1978, 1980) has hypothesized that early experience with

kin Imprints individuals to aspects of both kin and species, and that

such kin familiarity is important to selection of mates of the

appropriate species and to avoidance of inbreeding at sexual maturity.

Avoidance of inbreeding Is an important consideration in mate selection

because inbreeding often leads to inbreeding depression--a reduction

in the viability of inbred offspring and/or of their ability to

reproduce. Inbreeding depression has been documented in a variety of

species from ungulates (Rails, Brugger, & Ballou, 1979) to rodents

(Hill, 1974), birds (Bulmer, 1973), and Drosophila (Maynard Smith,

1956). In many mammalian species inbreeding is avoided in part because

individuals of one sex emigrate from the natal group before sexual

maturity. This type of emigration has been observed in chimpanzees

(Pusey, 1980), lions (Bertram 1975, 1976), olive baboons (Packer,

1979), black-tailed prairie dogs (Hoogland, 1982), and a variety of

other species. Intergroup transfer of individuals in mammalian species

and the significance of this behavior to avoidance of inbreeding have

recently been discussed by Packer (1979). Emigration from natal groups

may be mediated by adult aggression in many species (e.g., langurs:

Suglyama, 1965; elephant-shrews: Rathbun, 1979). However, female

(and male?) choice has also been suggested as a factor (Hoogland, 1982;

Wittenberger, 1981) since Individuals should be selected to emigrate

from natal groups If 1) relatives refuse to mate with them and 2)

alternative mates are not readily available. In at least some rodent

species reproductive maturity may be inhibited by pheromones produced

by adults (Bedlz & Whitsett, 1979; Drickamer, 1979; Lawton &

Whitsett, 1979; Lombardi & Whitsett, 1980). For young of such species

emigration from the natal group may provide the major opportunity for


Although it would appear that avoidance of inbreeding should

generally be the rule, certain circumstances may favor inbreeding. If

Inbreeding were not detrimental, individuals could increase their

inclusive fitness through mating with relatives (Bengtsson, 1978;

Maynard Smith, 1978). Female control of the sex of offspring in the

wasp Euodynerus foraminatus may be used to counterbalance the

detrimental effects of inbreeding and allow individuals of this

species to take advantage of the increase in relatedness resulting from

inbreeding (Cowan, 1979). Bengtsson (1978) hypothesized that it would

be adaptive for individuals to inbreed if the costs of inbreeding were

lower than the costs that would be incurred in dispersal or in

competition for mates in the natal group. Costs incurred through

dispersal would include such factors as increased exposure to predators

and to unfavorable environmental conditions. Inbreeding may also at

times be suited to specialized environmental situations; as noted by

Mayer (1970) "An outbreeder may also be so well buffered that it

stagnates evolutionarily. At the other end is the extreme inbreeder

which has found a lucky genotypic combination that permits it to

flourish in a specialized environmental situation . (p. 245).

Shields (1982) compared the advantages and disadvantages of

outbreeding, inbreeding, and asexual reproduction and concluded that

inbreeding is often more advantageous than commonly assumed and should

be "expected to be common in organisms produced by stable lineage-

environment associations" (p. 274).

Although inbreeding may be adaptive under particular circumstances

individuals of most species should, given a choice, prefer to breed

with nonsiblings rather than siblings. Some support is lent to this

statement by the observation that the initiation of breeding in sibling

pairs is often delayed in comparison to the initiation of breeding in

nonsibling pairs (Batzli, Getz, & Hurley, 1977; Dewsbury, 1982a;

Hill, 1974; McGuire & Getz, 1981). The contribution of preference per

se to these findings is, however, difficult to assess. Animals in

these studies were not allowed a choice of mates and delayed

reproduction, or lack of reproduction, may result from several factors

(Dewsbury, 1982a) in addition to preference. It would be expected that

reproductively mature males and females of most species should prefer

to associate with nonsiblings rather than siblings, and that this

preference should be apparent in choice tests.

Familiar Others

Differences in familiarity need not only be defined in terms of

differences in relatedness, differences in familiarity may also be

defined in terms of differences in the amount and type of contact an

individual has had with others. Familiarity in this sense is an

important factor in many aspects of an animal's behavior, including

mate choice. Individuals should improve their reproductive success

by retaining mates that they have previously bred successfully with

and choosing unfamiliar others over mates with which breeding has

previously failed. Coulson (1966), for example, found kittiwakes

(Rissa tridactyla) were much more likely to change mates if breeding

had been unsuccessful; and red-billed gulls (Larus novaehollandiae

scopulinus) exhibit similar behavior (Mills, 1973). Although some

authors (e.g., Halliday, 1978) have indicated that mate choice based

on previous reproductive performance should only be of importance to

species which pair bond for more than one season, all that is

actually required is an ability to recognize previous mates. This

ability has been demonstrated in several species including rats

(Carr, Demesquita-Wander, Sachs, & Maconi, 1979; Carr, Hirsch, &

Balazs, 1980; Krames, Costanzo, & Carr, 1967), lemmings (Huck &

Banks, 1979) and prairie voles (Ward, Baumgardner, & Dewsbury, 1981).

The ability to recognize previous mates may also enhance reproduction

through allowing earlier breeding. This function of familiarity

has generally been stressed for monogamous species (e.g., Daly &

Wilson, 1978; Wilson, 1975).

The advantages of familiarity to mate selection, based on

reproductive performance or early breeding, are related to an

individuals previous breeding experience. Familiarity may, however,

also bias selection of mates by sexually Inexperienced individuals.

Females may, for example, require a male to exhibit some evidence of

"commitment" to forming a pair-bond prior to copulation. The male

may fill this requirement by investing a large amount of his time in

the relationship, and thus preclude his finding another female

(Maynard Smith, 1977), and/or by demonstrating his ability to provide

resources (e.g., Nisbet, 1973). Evidence of commitment is likely to

be most important to members of monogamous species, especially those

in which individuals form prolonged or lifelong bonds, because many

individuals of these species may only choose a mate once in their

lifetime. Individuals of species that form prolonged pair bonds

might also be expected to be more "prepared" (Seligman, 1970) to

recognize differences in familiarity, than would be individuals of

species that do not pair bond, if familiarity were important to the

maintenance of pair bonds.

Many authors have indicated that males of polygamous species

should mate with as many different individuals as possible (Adler,

1978; Bateman, 1948; Dawkins, 1976; Williams, G. C, 1966; Zucker &

Wade, 1968). Although as noted previously males may have a limited

capacity to mate and should therefore be somewhat selective when

allowed a choice of partners (Dewsbury, 1982c; Nakatsuru & Kramer,

1982), it may still often be to a males advantage to obtain

additional matings if the opportunity is presented. Familiarity may

therefore be of importance (at least to polygamous males) in

identification of females a male has already mated with. However,

because mate infidelity could have serious consequences for

monogamously mated individuals (Grafen & Sibly, 1978; Trivers, 1972),

it is likely that individuals of monogamous species have been

selected to detect potential philanderers, and to select against such

individuals as mates (e.g., Erickson & Zenone, 1976). It may be

expected therefore that polygamous males would be more likely than

monogamous males to prefer novel over familiar partners. This

prediction is consistent with previous suggestions that males of

monogamous species should be less likely than males of polygamous

species to exhibit a "Coolidge effect" (Thomas & Birney, 1979;

Wilson, Kuehn, & Beach, 1963; but also see Dewsbury, 1981a,b).

As a general set of predictions it might be expected that

individuals of monogamous species would display greater preference for

familiar individuals than would individuals of polygamous species.

Because of differences in the consequences of choice it might also be

expected that females would display stronger preference than males

(e.g., Burley, 1981). Monogamous females would be expected to display

the strongest preference for familiar individuals. Monogamous males

should be expected to display some preference for familiar

individuals, as might also polygamous females (unless greater benefits

result from producing multiply sired litters). Polygamous males,

however, due to a greater possibility of increasing their reproductive

success through mating with more than one female, may be expected to

display some preference for novel individuals of the opposite sex.


The discussions in this section provide a brief rationale for the

choice of the particular species and experimental procedures that were

followed in this study. This section provides general methodological

Information common to all experiments in this study, and a description

of the apparatus used in these experiments.

Selection of Species

One of the methods that may be particularly suited to exposing and

interpreting differences In social behavior among species Is the

comparative approach (Dewsbury & Rethlingshafer, 1973; King, 1970).

Murold rodents are a group that is particularly suited to the use of

the comparative method (Dewsbury, 1974, 1978). The two murold rodent

species that were chosen for comparison in this study, the monogamous

Peromyscus polionotus (oldfleld mouse) and polygamous Peromyscug

manlculatus (deer mouse), are both members of the maniculatus species

group of the subgenus Peromyscus. Because the majority of mammalian

species are considered to be non-monogamous (Alexander, 1974; Crook,

1977; Kleiman, 1977; Orlans, 1969) and relatively little information

is available about mate choice In monogamous mammalian species, It was

considered to be particularly Important that one of the species

selected for comparison in the present study be a monogamous species.

Although monogamy has been suggested for several rodent species,

"except for polionotus, the data are circumstantial" (Foltz, 1981a,

p. 665) and are open to more than one interpretation. Data supportive

of monogamy in P. polionotus include the consistent finding that the

majority of reproductively mature individuals are captured as

heterosexual pairs (Blair, 1951; Foltz, 1981a; Rand & Host, 1942;

Smith, 1966), and behavioral (Blair, 1951) and electrophoretic (Foltz,

1981a) evidence that pairs form long-term reproductive associations.

Information about mate selection in this species was also considered to

be of importance, in addition to comparative considerations, because

one subspecies, the beach mouse (PE polionotus leucocephalus), is

presently considered endangered.

In contrast to P, polionoQus, electrophoretic evidence indicates

polygamy for P. maniculatus (Birdsall & Nash, 1973; Merritt & Wu,

1975) and females may even, on occasion, raise young communally

(Hansen, 1957). Dewsbury (1981c) suggests that maniculatus may be even

more promiscuous than indicated by electrophoretic studies because

these studies do not consider the effect of factors such as

"differential fertilizing capacity" (Lanier, Estep, & Dewsbury, 1979)

that may affect estimates of the number of matings that have occurred.

Approaches to the Study of Social Preference

Ideally social behavior and social preference should be studied in

natural settings. Although this approach may be utilized successfully

with diurnal and highly visible species, it is often an impractical, or

nearly impossible, approach for many species. As an alternative

investigators have often turned to the study of populations in outdoor

(Agren, 1976; Boice, 1977; Boice & Adams, 1980; Gipps & Jewell,

1979; Jannett, 1980; Lidicker, 1980) or indoor (Bowen & Brooks, 1978;

Crowcroft & Rowe, 1963; Getz & Carter, 1980; Hill, 1977; Poole &

Morgan, 1976; Reimar & Petras, 1967; Thiessen & Maxwell, 1979;

Thomas & Birney, 1979) seminatural enclosures. While these enclosures

do not replicate natural conditions in many respects, they do allow the

investigator to approximate some aspects of the natural setting, and

allow a degree of control over experimental variables that is generally

not available in nature.

Even in a seminatural apparatus, however, interactions may often

be so complex that it is difficult to evaluate the effects of any

single variable on a particular behavior such as social preference.

This problem has led investigators to the use of even more controlled

situations, such as preference apparatus of various types, to evaluate

the role of various factors in social preference. In the typical

preference paradigm an animal (the "choice" animal) is allowed to

express preference by "choosing" between two or more alternative

stimuli. Behavioral measures of preference may include factors such as

the number of approaches, number of visits, duration of visits, time

spent huddling together, mating activity, and a variety of other

measures. Use of preference apparatus, in addition to allowing more

controlled investigation of particular factors (e.g., familiarity) than

may be available in seminatural apparatus, allows the experimenter to

control the degree of contact between choice animals and stimulus

animals. In a tether preference apparatus, for example, stimulus

animals are tethered in a fixed area while the choice animal Is allowed

free access to the apparatus and may express preference through

proximity or contact behaviors, or under appropriate conditions, mating

behavior (Ward et al., 1981; Huck & Banks, 1982; Webster, Williams, &

Dewsbury, 1982). An alternative method used in preference tests is to

place collars on the stimulus animals, and then place these animals in

compartments with doorways of a size large enough to allow access by

choice animals, but too small for the collared animals to pass through

(Mainardi, Marsan, & Pasquali, 1965; McDonald & Forslund, 1978).

Direct contact between choice animals and stimulus animals may also be

prevented by simply constructing stimulus compartments or containers so

that they are not accessible by the choice animal (Agren & Meyerson,

1977; Carmichael, 1980; Carr, Wylie, & Loeb, 1970; Murphy, 1977;

Webster, Sawrey, Williams, & Dewsbury, 1982). Experimenters have also

opted at times to test preference for odors from stimulus animals

rather than using the animals themselves (Carr et al., 1980; Fass,

Guterman, & Stevens, 1978; Gilder & Slater, 1978; Huck & Banks, 1979,

1980; Krames et al., 1967; Ruddy, 1980), or to restrict choice cues

to olfactory cues by using anesthetized stimulus animals (Landauer,

Banks, & Carter, 1977; Landauer, Seidenberg, & Santos, 1978; Murphy,


While preference apparatus offer the opportunity for greater

control over variables than do seminatural apparatus, the conditions

under which preference is assessed do not approximate natural

conditions as closely as do conditions in seminatural apparatus. With

preference apparatus, therefore, one may run a greater risk of

obtaining results that are misleading in respect to behavior under more

natural conditions. Social preferences may, for example, sometimes be

expressed less strongly in preference apparatus than they would be

In a more natural context; one might therefore be more likely to

falsely reject a factor as unimportant to social preference in these

tests. One way to minimize this problem is to first assess a species'

social behavior in more natural settings, such as in seminatural

apparatus, and select factors for preference experiments on the basis

of those results. Alternatively one might use results from seminatural

experiments in part as a guide in interpretation of results from

preference experiments.

General Experimental Information

This study was designed to provide data on aggression and

familiarity as factors in the social preference of monogamous and

polygamous species. Partial data are available about the function of

these factors in the social preference of the representative species

chosen for this study, P. polionotus and P. maniculatus.

Available evidence indicates that aggression may be an important

factor in the social behavior of P. maniculatus, that more aggressive

males may sire more offspring than less aggressive males, and that

differences in male aggressive ability may be important in female

choice in this species. In P. maniculatus blandus Blair and Howard

(1944) found that, in experimental populations consisting of two

individuals of each sex, one male would generally establish dominance

over the other. The dominant male generally nested with both females

more frequently than did the subordinate, and the authors were able to

establish (through coat-color markers) that dominant males sired the

majority (19 of 21) of litters in their study. Dewsbury (1979, 1981c)

found that male dominance in manlculatus balrdi was positively

related to copulatory behavior, and that dominant males not only

copulated more than subordinates, but that they also sired a larger

number of offspring (Dewsbury, 1981c). Elsenberg (1962) observed that

after the formation of dominance relationships between male P.

maniculatus gambelil, females of this species that had been paired with

subordinate males for two weeks prior to aggression tests generally

failed to remain with their subordinate male partners and nested

Instead with the dominant male.

Blair and Howard (1944) studied two subspecies of P. pollonotus,

P. pollonotus albifrons and P. pollonotus leucocephalus, and found

little evidence of aggression against conspeciflcs by individuals of

either sex. In addition all four individuals (two males and two

females) in a group were frequently found nesting together. From these

observations the authors concluded that PE pollonotus were a very

social species. Field observations, however, do not support the notion

that adult EL polionotus are highly social. Although E. pollonQtus

are commonly found in family groups composed of a male, a female, and

young (Blair, 1951; Foltz, 1979; Rand & Host, 1942; Smith, 1966;

personal observations), sexually mature Individuals of the same sex are

never (Smith, 1966), or very infrequently (Blair, 1951; Rand & Host,

1942) found together in the same nest. In addition Blair (1951)

observed wounding In some transient and immature Individuals and also

observed, in trap and release experiments, that adult females often

chased other females from nests. Smith (1967) has suggested that

females of this species "are normally dominant over their mates and

play a major role in the process of pair formation and maintenance of

the pair bond" (p. 236). P. polionotus have also been observed to

exhibit aggression in some laboratory tests (Garten, 1976; Smith,

Garten, & Ramesy, 1975), but the conditions for these tests do not

allow evaluation of the function of aggression in a social context, or

as a factor in social preference.

Few data are available on the function of aggression and

familiarity in the social behavior of P. polionotus, or in the

social behavior of monogamous species in general. The first set of

experiments in this study was designed to provide such data. These

experiments were conducted in a seminatural apparatus that was designed

with artificial burrows. This design takes into consideration the

semifossorial habits of P. polionotus, and thereby allows an

approximation of natural conditions in this species.

The seminatural experiments with P. polionotus were followed by

preference experiments on both P. polionotus and P. maniculatus.

These experiments allowed preference based on aggressive ability and

familiarity to be assessed under the same conditions for both species,

and thus allowed a direct comparison of the relative value of these

factors in the social preference of these two species. The first set

of these experiments examines aggressive ability and familiarity with

unrelated individuals (based on previous contact) as factors in social

preference; the second experiment examines preference for siblings.

General Methods


Subjects for this study were 45 to 65 day old Individuals of two

species of muroid rodent, Peromyscus polionotus subgriseus and

Peromyscus maniculatus bairdi. The P. polionotus were laboratory-bred

animals one to four generations removed from the wild. The parental

stock was obtained from two different subpopulations in the Ocala

National Forest in Florida. The first group of these animals was

trapped in 1978 from road shoulders along State Road 316 between Salt

Springs and Eureka. Additional animals for breeding stock were trapped

in 1980 from road shoulders along U.S. Highway 19. These two

populations are from the same general area as that listed as population

25 by Selander, Smith, Suh, Johnson, and Gentry (1971). The method of

capture was similar to that detailed by Foltz (1979).

It is not possible to determine how many generations removed from

the wild the EP maniculatus were. This colony was founded at the

University of Florida with animals obtained from near East Lansing,

Michigan, In 1970, and additional wild stock has been added on several

occasions since.

Animals were housed In clear plastic cages measuring 48 x 27 x 13

cm or 29 x 19 x 13 cm with wood shavings as bedding. Purina laboratory

animal chow and water were provided ad lib. Prior to serving as

subjects all animals were maintained as litters. Peromyscus pollonotus

litters were weaned at 22 or 23 days of age, P. maniculatus were

weaned at 21 days of age. Animals that exhibited obvious physical

defects, such as extensive tail wounds or missing tails, were not

selected for study.

Animals of both species were maintained on a reversed 16L:8D

photoperiod. All adaptation and testing were conducted during the dark

portion of the photoperiod. With the exception of observations

conducted in the seminatural apparatus, which was in a separate room,

all studies and adaptation periods were conducted in the P. pollonotus

colony room. Procedural details specific to particular studies are

described in the methods sections of those studies.


SeminaturaJ apparatus

The seminatural apparatus was a large square Plexiglas arena 125

cm on a side and 46 cm deep. The sides of this arena were constructed

with 1/4 inch Plexiglas and the floor was constructed with 1/2 Inch

plywood and painted grey. The arena was partitioned, with four 85 cm

lengths of 1/4 inch Plexiglas, into a square central area that measured

85 cm on a side and four right angle triangular corner compartments

with sides of 85 cm, 61 cm, and 61 cm (See Figure 1).

Two nest boxes were attached to each corner compartment. Nest

boxes were constructed with sides of 1/4 inch Plexiglas and 1/4 inch

plywood backs. They measured 10 x 9 x 8 cm and had hinged Plexiglas

lids to provide access for removal of animals and cleaning. A 3.2 cm

diameter hole cut in the front of each nest box provided access for the

animals. In each corner compartment two matching 3.2 cm diameter

holes, cut in the sides of the apparatus, 45.5 cm from the corner and

1.5 cm from the floor, provided access to the nest boxes. The front of







one of the nest boxes for each corner compartment was connected with

silicon cement directly to the side of the apparatus in line with one

of these holes. The other nest box in each corner compartment was

connected to the second opening by means of a 48 cm length of Tygon

polyethylene tubing with an internal diameter of 2.5 cm and an external

diameter of 3.2 cm, and thus formed an artificial burrow. Silicon

cement was used to attach one side of the nest box to the apparatus and

to connect the tubing to the openings for the nest box and the


Animals could gain access from the corner compartments to the

central area of the apparatus through 3.2 cm diameter holes centered on

and 1.5 cm from the bottom of the partition which formed the

compartment. A 2.5 cm hole 3 cm from the right angle corner and 2 cm

from the floor of the apparatus provided access for the drinking tube

of a water bottle.

The seminatural apparatus was in a room separate from the colony

room but maintained on a 16L:8D photopericd identical to that

maintained in the colony room. The apparatus was illuminated in the

light phase of the photoperiod by four 75-watt incandescent-white bulbs

and two 60-watt red bulbs, each suspended three feet above the floor of

the apparatus, and during the dark phase of the photoperlcd by the two

60-watt red bulbs alone.

Behavioral measures were recorded by means of a 20-channel

Esterline-Angus event recorder. The behaviors exhibited by each group

of animals, during their four days in the semlnatural apparatus, were

also recorded on videotape using a Hitachi CCTV low light television

camera, and a Panasonic time lapse VTR video tape recorder set on a 72

hour record mode.

Preference apparatus

The preference apparatus was a three chambered rectangular box

with a hinged lid; It was constructed of 1/4 Inch Plexiglas and

measured 44 x 21.5 x 20 cm (Figure 2). The inside measurements of the

two end chambers were 10 x 21.5 x 20 cm. These chambers were open to

the central area through a 7 x 7 cm opening. The end chambers were

designed to accommodate small removable "choice chambers" which

measured 10 x 8 x 6 cm. A "stimulus box" with an inside measurement of

8 x 7 x 8 cm was attached to each end chamber, and was open to It

through a 6 x 5 cm opening. When choice chambers were placed in the

end chambers, therefore, one end of the choice chamber was accessible

from the central area, while the other end was open to the stimulus

box. A hardware cloth screen installed In the opening to the stimulus

box and a second three-sided piece of hardware cloth which fit the

inside of the stimulus box provided a "double screen" between the

choice chamber and the stimulus box.

A bank of three red-sensitive photocells (peak response at 735

nm), wired in a series behind each choice chamber, registered entries

to the chamber. The light sources for the photocells were 60-watt red

light bulbs placed 27 cm in front of the apparatus and directly in

front of the bank of photocells. Each photocell was attached to the

end of a tubular 4.3 cm piece cut from a 12 x 75 cm disposable plastic

culture tube. The outsides of these tubes were painted black; this in

effect columnated the light to the photocells. The photocells were

uo O---- \


> L





situated such that in order to register a visit to the choice chamber,

an animal had to be completely Inside the chamber, and at least

partially in the half of the chamber closest to the stimulus animal.

Each bank of photocells was wired In a series with the coil circuit of

a 10,000 ohm, 24 VDC DPDT relay (see Figure 3). These relays were

powered through output from a variable power supply set for a

continuous output of 26 volts. One of the normally closed circuits of

each of these relays was wired into the pen circuit of an Esterline-

Angus event recorder; this provided a permanent record of entries into

each choice chamber. Output from another normally closed circuit of

these relays was used to control a second set of relays on a relay

rack. Two banks of Sodeco counters received input through the normally

open contacts of these relays. The first bank of counters was wired in

a series to pulse former and recorded the number of visits to each

chamber regardless of visit duration. Input to the second bank of

counters was regulated by means of a recycling timer set to produce

pulses at 1/3 of a second. These counters recorded the total duration

of visits to the nearest 1/3 of a second. Session duration was

automatically controlled via another timer (not displayed) which

controlled the input to the recycling timer and counters.

Aggression apparatus

The "aggression arena" was constructed from a large 48.5 x 38

x 20 cm plastic cage. Two 3.2 cm diameter holes were cut in the two

longer sides of the cage centered 25.5 cm apart and 3.5 cm from the

bottom. Silicon cement was used to form a gasket around each hole.

Matching holes and gaskets were placed on one side of two 48 x 27 x 13

cm plastic cages. The larger cage and two smaller cages could then be

connected by 6.5 cm lengths of Tygon polyethylene tubing (Internal

diameter of 2.5 cm and external diameter of 3.2 cm) to form the

"aggression apparatus" (See Figure 4). During adaptation procedures

the aggression arena and smaller cages were connected with unobstructed

lengths of tubing; a piece of metal screen in the center of each

length of tubing prevented animals from traveling through the tubes

during tests. A 1/4 Inch 53.5 x 43 cm Plexiglas lid was placed over

the large cage, and wire cage lids over the smaller cages, while

testing was conducted.

FIGURE 4 Aggression Apparatus


Seminatural Experiments

This section is divided into three major subsections; the three

sets of experiments that comprise this study are each described within

separate subsections under the headings of Seminatural Experiments,

Aggression and Familiarity Preference Tests, and Sibling Preference

Tests. Each of these subsections begins with a brief introduction to

the experiments in that subsection, and specific information on

subjects and procedures for these experiments. This information is

followed by the results of these experiments and a brief discussion of

the results. The results of all three sets of experiments in this

study are discussed together, in the context of the ecology of EP

polionotus and P. maniculatus and theoretical considerations, in the

General Discussion section.


This experiment was designed to provide information on aggression

and familiarity as factors in the social behavior and mate selection of

E. polionotus. (Similiar types of data already exist for P.

maniculatus: Blair & Howard, 1944; Dewsbury, 1979, 1981c: Elsenberg,

1962.) Although it is difficult to establish the relevance of factors

in social preference per se with seminatural observations, such data

can provide an indication of the functions a factor may serve in

nature, and can provide Indications of whether a factor may be of

importance in social preference.


Subjects were 40 male and 40 female LP. pollonotus. Prior to

serving as subjects, animals were maintained as previously described in

the section on general methods. Subjects were selected using the

criteria described in the general methods, and the additional criterion

that the animals within any group had no common grandparents.


In order to better separate and evaluate the roles of aggression

and familiarity in the social behavior of P. polIonotus animals were

observed under two different experimental conditions, the "paired"

condition and the "single" condition. Twenty animals of each sex were

assigned to either the single condition or the paired condition.

Animals In each condition were divided into 10 groups; two animals of

each sex were assigned to each group. Each of the 10 groups of animals

in each condition, single or paired, were treated separately. Animals

within each group were lightly anesthetized with ether and shaved In

one of the following four patterns: (1) band shaved around the neck;

(2) band shaved around the middle; (3) band shaved at the rear; (4)

no shaved area. Shaving was performed one day prior to beginning the

first experimental manipulation. Approximately equal numbers of males

and females received each shave pattern. Subjects under both the

single and paired conditions were exposed to a series of three

different experimental manipulations. These manipulations, in order,

and their durations were "nest building," 4 days; seminaturall

isolation," 3 days; and seminaturall interaction," 4 days.

During the four-day nest building period animals were housed in 48

x 27 x 13 cm plastic cages on San-i-cel bedding. Animals in the single

groups were housed individually; animals in the paired groups were

housed as two separate pairs of opposite-sexed animals. Three 2-inch

square "Nestlets" (Ancare Corp.) were provided as nesting material in

each cage. The type of nest built was assessed just prior to the

beginning of the dark period on the next 4 consecutive days. Nests

were rated as one of three types: (0) no nest; (1) platform nest;

(2) covered nest.

The seminatural isolation period began in the first dark phase

which followed the nest building period. Animals were transferred from

the colony room to the room containing the seminatural apparatus

approximately 15 min after the beginning of the dark phase. Animals

from single groups were each placed individually in corner

compartments, with animals of the same sex in compartments diagonally

opposite each other. Animals from paired groups, which had been

maintained as pairs during nest building, were transferred to the

seminatural apparatus in the same paired relationship. Pairs were

placed in corner compartments of the apparatus diagonally opposite each

other. The opening from each corner compartment to the central area

was closed with a solid black rubber stopper so that animals were

restricted to the compartment in which they had been placed. The floor

of the central area and of the corner compartments had been covered

with San-i-cel to a depth of approximately 1.5 cm; each corner

compartment also contained three Nestlets, and food and water was

available ad lib.

Animals were maintained in the corner compartments for 3 days.

Activity during this entire period was videotaped. Each pair of

animals in the paired groups was also observed for three alternate

10-minute periods during the dark phase on the day the animals were

Introduced, and on the following two days. On the first day,

observation was begun as soon as all animals had been placed in the

apparatus. On each of the following 2 days one of the pairs was

designated as the first pair to be observed, and the first period of

observation was begun when the members of that pair had emerged from

their burrow. The behavioral and aggressive measures that were

recorded were similar to categories described by Allin and Banks (1968)

and Colvin (1973). Measures were recorded by means of a 20-channel

Esterllne-Angus event recorder. The measures, and definitions of each,

were as follows:

Approach Scored when an animal came within one and
one-half body lengths of another while
oriented toward it.

Attack Scored when one animal lunged at or charged
another but did not pursue the other or
initiate vigorous biting behavior. This
behavior could be accompanied by a single
bite or attempts to bite.

Scored when one animal pursued another.







Scored when one animal's attack on another
escalated to vigorous biting behavior by
both individuals. Generally the initiator
would knock the other animal over, or roll
to one side with the other animal clenched
in its jaws while shaking its head, often
simultaneously clawing with the rear claws.
Fighting often resulted after one animal did
not retreat when attacked, but rather attempted
to defend itself or at the end of a chase if
the pursuing animal caught the other.

A very vigorous form of fighting; rough-and-
tumble fights were only scored when both
animals were tumbling end over end while
attempting to bite and claw each other.

Scored when one animal retreated upon
another animals approach.

Scored when one animal, upon approach or
attack by another, either rolled over on its
back, or reared back upon its hind legs with
its nose pointed up, and made no attempt to
defend Itself. Both approach and submission
or attack and submission were scored for each

This was a very vigorous form of digging
behavior much more Intense than the type of
digging these animals have been observed to
perform in an isolated test (Webster, Williams,
Owens, Geiger, and Dewsbury; 1981). Although
this behavior was not generally directed toward
an opponent, it was very similar to that described
by Allin and Banks (1968).

Under both experimental conditions, single and paired, the

isolation period was followed by the seminatural interaction period.

Ten min prior to the first dark phase in this period the rubber

stoppers were removed from the partitions between each corner

compartment and the central area. Behavioral and aggressive

interactions between animals, as defined above, were recorded during

the first hour of the dark phase for 4 consecutive days. Nesting

relationships were recorded each day, for the last 3 of these 4 days,

20 min prior to the beginning of the dark phase. An animal was defined

as having nested with another if it was found in the same nest with the

other, and videotape records verified that it had not switched nests

between the period extending from after the first 1/2 hr of the

preceding light phase to 1/2 hr prior to the nest check. Activity

during the entire isolation and interaction stages was videotaped.

The seminatural apparatus, including the artificial burrows

and nest boxes, was thoroughly cleaned with a solution of Sterigent (a

deodorant and disinfectant soap) before each group of animals was

introduced to the apparatus. Water bottles were also cleaned and

refilled, and fresh San-i-cel, food, and Nestlets were placed in the



Peromyscus polionotus appeared to adapt very quickly to the

seminatural apparatus in general, and to the artificial burrows in

particular. All but five of the 40 individuals in the paired condition

entered and explored the artificial burrows within the first hour of

observation, and all except two pairs of the 20 pairs observed in the

paired condition had constructed at leat a rudimentary nest in the

burrow by the end of their first day in the apparatus. Individuals in

the paired or single groups were only infrequently observed nesting in

the alternate nest box, although this box was frequently used for

feeding and as an escape when individuals were attacked. Aggressive

relationships between Individuals in each group appeared to remain

fairly stable over their four days together In the seminatural

apparatus. In all 10 single groups, and in seven of the 10 paired

groups, the individual with the highest total frequency of aggressive

behavior (sum of all attacks, chases, fights, and rough-and-tumble

fights) on day one, still exhibited the highest frequency of aggressive

behavior on day four. In the other three paired groups the individual

that exhibited the highest frequency of aggressive behavior on day two

also exhibited the highest frequency on day four.

Each animal in a group could potentially interact with twice as

many opposite-sexed individuals as same-sexed individuals. To adjust

for this bias, the mean value of any measure of an animal's interaction

with both opposite-sexed individuals, rather than the total, was used

in analyses Involving opposite-sexed Individuals.

Several significant differences In aggression were apparent in

comparisons between males and females in both the paired and single

conditions. Males were more aggressive than females in a statistically

significant larger number of groups by all measures except the number

of rough-and-tumble fights (designated in tables as r&t fights; see

Table 1). Within the single condition males were the more aggressive

sex in a significantly larger number of groups by all measures except

the number of rough-and-tumble fights and the duration of

rough-and-tumble fights. Within the paired condition males were the

more aggressive sex In a significantly larger number of groups for the

measures of number of attacks, number of fights, duration of fights and

number of approaches. There was no measure, for either the paired or

Table 1

Comparison of the Number of Groups In Which
Males or Females Were More Aggressive

Total (N=20) Paired (N=10) Single (N=10)

Measure Male Female Male Female Male Female

No. of attacks 19 1*** 9 1* 10 0**
No. of fights 17 2*** 9 1* 8 1*
Duration of fights 18 2*** 9 1* 9 1*
No. of chases 17 3** 8 2 8 1*
Duration of chases 17 3** 8 2 9 1*
No. of r&t fights 13 5 6 2 7 3
Duration of r&t fights 15 3** 8 2 7 1
No. of approaches 19 1*** 10 0** 9 1*
No. of displacements 15 4* 7 3 8 1*

All durations are in seconds.
Sign test 2-tail *-<.05 **p<.01 ***q<.001

single condition, for which females were more aggressive than males in

a larger number of groups.

Males in general also exhibited higher total levels of aggression

than females by all measures (see Table 2). Males, in both single and

paired groups, exhibited a higher frequency of attacks, fights, chases,

displacements and approaches than females. They also exhibited longer

durations of chases and fights than females in both types of groups

(see Table 3). Although differences in the total amount of submissive

behavior exhibited by males and females across both groups were not

statistically significant (1=1.28, df=19, p>.05),males had more

submissive behavior directed toward them than did females

(paired-=2.71, df=19, p<.05).

Because aggressive digging was not immediately recognized as a

possible correlate of aggression, it was not recorded for the three

Initial paired groups. It was recorded for all subsequent paired and

all single groups. No significant differences in the frequency,

duration, or mean duration of aggressive digging were apparent in

comparisons between paired and single animals. Males displayed higher

levels on all of these measures than did females, with the exception of

the comparison of the average duration of aggressive digging for single

animals. High-aggression males (those in each group with the highest

total frequency of aggressive behavior) displayed a higher frequency of

aggressive digging than low-aggression males (paired-=2.30, df=16,

p<.05). The difference between high and low-aggression animals in the

level of aggressive digging they displayed may be due In part to

differences in the response of these two classes of individuals to

Table 2

Comparison of Total Aggression Within Groups by Males
and by Females In Both Paired and Single Conditions

Male Female

Measure Mean (SE) Mean (SE) t

No. of attacks 46.30 5.95 10.20 2.41 5.23****
No. of fights 14.55 2.24 3.05 1.07 4.23****
Duration of fights 23.00 3.96 4.60 1.46 4.31****
No. of chases 116.20 12.18 21.85 5.51 5.98****
Duration of chase 679.10 79.01 128.40 32.79 5.59****
No. of r&t fights 5.00 1.44 1.25 .42 2.60*
Duration of r&t fights 10.30 3.04 2.80 1.39 2.35*
No. of approaches 64.10 7.91 23.40 3.61 5.62****
No. of displacements 12.80 2.59 3.40 .90 3.43***

All durations are In seconds.
Paired -test df=19 2-tail *P<.05 ***,<.01

se a
Oh * *5 *
Ul.CNl' r mN~Imrtrh-- *

C~~for-tO hooO h-^'M'~r-.aoh.
.- o -CP itno'on
0 4

E n
0 C4 In 4

o z -1 0 a 4 W o

C -. '^*^or.0 Mr~0 vr -0)
uW ON|0N.0
( i *1 *a 0 1 0 w I Z,
W- W N nI 10 .5T nn )0 r- wt o

z ooocoomp0Oh00-0ow
C r-ON.0Or-uOlfs.Or.K5.TOO.
L" I a'0^ 'aa""'0 'a s "

5 *: N '-N

N &U asC140 a 50

D S/) N; 4 *h 4- N
-y o
0 co r ww '00 0

0 tm a v -nn0a .00
c. 10 N 0 n C- w N6"Wr4 CO t
0)O .-o 0tr-.5 .O0N 0 0 ; 0 n
(A 00A 'o 0 '-0'o

o a
S_ Wa).. S N 3O In 4-oInP0C)^io
SN 0 -
0 C %0a0
a) c ccirmioK ri s*

1 u Es as a 0 aOr
.0 10 *55 Se a e* LL.
U01 VI (0m0~10N C

C-4 N Wnn~~-
s ^cn ^n no -:: ?our

see' Ce easemsrri *c on r-Unn a..0'.r..5-

fQ H. N-^ -,- -

0 I N r- CN 1

4- .- 4


0 r
a~ l cc
0x A 0 c 0 0 )0 10 r0 10

^.^~~ ~ 0, 2 0-**** *-
^4.L Na V) A500N .*
Q) f V

o1 5 S N -N 5-N U 4
t10 a a 0o

I- a*,

CC 0-0 LA O)I I ) 0 fl

0 .S N L" N0 00

o c: n^csm-- irsir- m-OO^t N I I

0 M^.--ow^Nf\- (

a S
Q+- -f > XN |
a=+- 0 C
a- a o c -a
(A a 4- 0)0) l
+- wo < i-a- ,
a Un L. C

a'* U C~5 0
o*- o..-+-0o- e >o'-
o ox: oCS- o0 *
iOcaOc:- c ioo c
U+--0OS0O-O00- 0- c
(Qr:--- 1f +- ^Ql-- 4- L ~- 0.*-
+i- aa.- 'o -a 04- -5- aw- 0)0- o -0
** oI.

00- 0- 0- 000-
as as SC C|
.- ta .-a. .-aC)
:z-1zr ac H ~o
i-IZI- ZZCb-

noises within the apparatus. High-aggression animals often

Investigated noises made by other Individuals In the apparatus.

Low-aggression Individuals generally Ignored these noises or retreated

from them. Low-aggression individuals that attempted to dig,

therefore, In effect increased the probability of attack by

high-aggression Individuals, whereas high aggression Individuals dug

without Interference.

Individuals directed more aggression toward same-sex than

opposite-sex Individuals In both the paired and single conditions (see

Tables 4 and 5). Males in both single and paired groups directed more

fights and rough-and-tumble fights, and longer durations of these

behaviors, against other males than against females. Females in both

single and paired groups directed more attacks toward same-sexed than

opposite-sexed Individuals. Males and females in both types of groups

directed more chases toward same-sex than opposite-sex individuals.

Although many of the differences in aggression between sexes were

significant, and many significant differences were also found in the

level of aggression directed at same versus opposite-sexed Individuals,

the overall levels of aggression displayed by animals in the single and

paired conditions were very similar (means and standard errors were

presented in Table 3). Animals In the two conditions displayed

significant differences on only two of the aggressive measures: paired

animals displayed longer average durations of fights with same-sexed

Individuals and a greater number of approaches to same-sexed

Individuals than did animals in the single condition (=2.12, df=78,

p<.05; and =2.10, dL=78, .<.05 respectively).


19 0 )CM M- n U9 91 00% r

Ln mr C4- nn L- 1

0 f 0tooo %Cm D 4m0a

a C w- C3 %0
VN 00 0

e o

cE ooCmOin i O o
a P '0 (4o(o0 r

0 P n X '

4 S

x NM MCCDo Cnnon n
L 0
o a

0 IC N 4- 4
0 -

-- E k CO-N CD aLn n n r-C9x v
CI) co c m o o
)* * *

o cnn w N o t re C
nE .n fn MCI
Sa m o o o o o o o o ir f
-) N1 N --5

m cn
a v -N m N -o co An0 mI 0 U)

0n 0

IA. 0
V) 0 E In n4DO.D W'n *o

i- 4O a O e
0 -30 (

SC 0 0in

us U L c IA
(0 10 J-_ 4*)

(n O* O o
u --4000 a-

to) c- o

-- -2' l0s-h
o 0 0%
m z Ez%' M07 3a
0 -. '0 ( Nr.%la

o o.

U) (U C L0 C 0O Iu '

S00 0 ) 0 3 O

3-l Z O '.0-Z <00


rn 0m w' 0N N 0Mtn 0 on in

N-o NN

(0n- U'oPmmoUoPo
4- o

to c
V 0 ifa W! Ci -W! ci Ci
S0 0 . .. ..

0 *

0): 0 C PO C4I WN Wo o'
l) X 0 O O6- C4T4 in
Z | O C O CD CDWM %w C3M

0 %( aW % C o Nn O


0 0 +U 0- *% U
Ln 10 a tn a. C) Fn t v Cl- 0
a ) V


0 _

06 4 -W* N

QO -- l 06

-* 01 0'0mI0%-in o 'I ri n

X ** x

ou &C o-- o0
| *r *0 C

4 0 c 0. 0-
2_ U C L CM C Lo(


L- m o cL!

0l >0 z I M
0 in+-0 000

In u 00 El-

in I- C c uincLC (A

Comparisons of aggressive measures between the two conditions

within each sex yielded significant differences for females only (means

and standard errors were presented in Tables 4 and 5). Paired females

exhibited a greater number of approaches to same-sex Individuals and a

greater number of displacements of same-sex individuals than did single

females (=2.57, df=38, p<.02; and =2.08, d=38, p<.05,


The various measures of aggression recorded tended to be

correlated with each other. The total frequency of attacks, chases,

fights, and rough-and-tumble fights were correlated within each sex for

both paired and single groups, as were the duration of chases, fights,

and rough-and-tumble fights. The frequency of approaches was

correlated with the number of chases for males and females in both

conditions and with the number of attacks for paired males and females

and single males (see Table 6). The amount of submissive behavior

directed toward single females was correlated with the frequency and

duration of fights (Pearson correlation, r=.724 and =.798,

respectively, p<.001) and the frequency and duration of rough-and-

tumble fights (Pearson correlation, =.704 and .800, respectively,

p<.001). Submission was also correlated with the total frequency of

approaches for paired females and single males (Pearson correlation,

=.444 and =.515, respectively, p<.05). Frequency of aggressive

digging was correlated with the total frequency of aggressive behavior

(combined frequencies of attacks, fights, chases, and rough-and-tumble

fights) for paired males and females (L=.638, .<.05, and =.845, p<.001

respectively) and single males (=.697, p<.001). Duration of

-- *KI

(- %.


(N co
* *

* **

K *2
* *

r- r-



*** *

r- rC CO o0
r~ ~ o c
* * *

* *

2Kr *2T CM

to l co



*-k0** <
N3 0 N N

*% *
r~ OK o

e>M 0 oenee


\o n- co


* ***
2K **2K
* ***
* *2*

coCM T M I %0 "r co r- r-



or- r- N)
* e *

0% m IM
\) 00 00
o co c


NM r r- 0o
* * *




CM co

K ***
IK *e
K **2**
7' co co 0 oro 0
)N %0 N 'o r- n Ln





- O

0 L


o c



0 0

L (0
0 )0


+- l

U)a) U)

4- U) U)0)EE E c

u C)nD) C CUUUW--
-0(D E U) n (D ) U) ) CO(O (O4-'4-4-
-C (D -- ( E 4-) EO--
)00 U CC= a=)CD C = D L L C7)+- 4-
U) -- (0 (0 0)U U 0)u u a.U)n U- x3od
(1)4- 4- 0 U 0 (u (U (0 a.- L L
ui)C I- L.0)C- 0 4- 0 MDOvo
C O0)4- L.U U) La. L "a0v 0V
S- o4a .- C 4- .n U) 4-0 U)-n o C C c
U S- L (C 0 .C U (oD i 1 0 (0 ( 4c

CO O C CO c C c C Co C 4-4- c0 Un nu)

( U0 U 4-4-4-4- ) ) )-- 0

4-4-4-4-4- 0 00 0) 4-4- 0.0 0 0
4-4-4-4-4-- --_-__-C .dod Q.
M CO C0 C0 C4 4- '4- U i- LL C C C C

............... L L L
;0000200000000000 n


+- v

U 0

-0 *


- O

i- oC
0) 4-z
c 0

0 (C-


-0 II
< ozt



aggressive digging was correlated with total frequency of aggressive

behavior for single males (r=.527, .<.05) and paired females (r=.845,

.<.001). An animal's weight did not appear to be an important factor

in aggressive Interactions. Only the correlation between weight and

the frequency of rough-and-tumble fights In single females was

statistically significant (r=.378, p<.05).

For paired groups comparisons were also made between the level of

aggressive behavior that occurred while pairs were confined to the

corner compartments and the level after access to the entire apparatus

was allowed. Very few aggressive interactions of any type (chases,

fights, attacks, rough-and-tumble fights) were observed between pair

members during the period pairs were confined to the corner

compartments (Mean total frequency of aggressive interaction per paired

Individual=.20, range = 0 4.0). Therefore only the total frequency

of aggressive behavior, rather than the frequency of behavior In each

category, was used for comparison.

The total frequency of all aggressive Interactions (attacks,

fights, rough-and-tumble fights, and chases) between pair members while

confined to the corner compartments was not useful In predicting

frequency of later aggressive interactions with pair members

(males:r=.034, .>.05; females: r=.087, p>.05) or total frequency of

aggressive Interactions (males: =.015, p>.05; females: r=.118,


Pairing did, however, have some effects on later levels of

aggression. Males in paired groups were less aggressive to the females

with which they had previously been paired than to females with which

they had not been paired by the measures of frequency and duration of

chases (one-tall palred-=2.18 and 2.22, respectively, df=19, P<.05).

Males also exhibited a higher frequency of approach to females with

which they had been paired (one-tail paired-=1.78, df=19, p<.05).

None of these comparisons were significant for females (Number and

duration of chases, =.38 and .76 respectively, approach; =1.08,

df=19, P>.05).

A problem arises when one attempts to compare different classes of

animals as to their levels of aggressive interactions with one another.

The problem is that the total amount of contact between different

classes of animals may vary. For example, If females tend to avoid

other animals, but males do not, Individuals would have more

opportunities to be aggressive to males than to females. A difference,

therefore, in the level of aggression an individual expresses toward

Individuals of one class versus another may reflect a true difference

in the frequency of aggression, or a difference in the frequency of

access to Individuals of the two classes. One method of gaining a

clearer understanding of the level of aggression, and differences in it

between classes of individuals, Is to construct a scale or index for

comparisons which accounts for differences In frequency of contact.

An "aggressive Index" was calculated for this study by dividing

the total frequency of all aggressive encounters initiated by an animal

(frequency of attacks, chases, fights, and rough-and-tumble fights)

toward any other class of individuals by the total number of contacts

initiated by that animal toward that class of Individuals (total

aggressive encounters plus the frequency of approaches and

displacements). Although there may be qualitative differences in

approaches which elicit displacement or submission and those which do

not (for example an aggressive individual may signal its status through

adopting a particular posture), no means was available in the present

study to detect these cues. Therefore displacements or approaches with

submission were classed as contacts without aggression for purposes of

constructing the index.

No significant differences were apparent between paired and single

groups, or in comparisons between males or females of these groups, in

the total frequency of contacts or the frequency of contact with

same-sexed or opposite-sexed individuals (see Talbe 7). Single females

did, however, have a significantly higher overall aggressive index

(frequency of all aggressive behaviors divided by frequency of all

contacts) and a higher aggressive index against opposite-sexed

individuals than paired females.

The total frequency of contact was significantly higher for males

than for females in the paired and single conditions. Although the

overall aggressive index was significantly higher for paired males than

paired females, the difference between single males and females was not

significant (see Table 8).

Paired males and females and single males all displayed a

significantly higher frequency of contact with same-sexed than

opposite-sexed individuals. Whereas both paired males and females

displayed a significantly higher aggressive index against same-sexed

than against opposite-sexed individuals, this comparison was not

significant for single males or females (see Table 9). Paired males,

o4- C o

!D Go %0 04 r
(N (N -

S0 C O W % CM e l
E -W 1l 0C -T

00 000 cn


S* *0-0

C% 4
Q)% wo WN


a> a -
U O 0O'.c
rt-c 000 o o"o

00 c 1 1!


t C+ C C O

4-) t/c . 9,4 ^ -

C 0 U) -C
0 a
E0 irn r o o oo

4- ll4-

X0 C in%-(N-
I 0i)
c 0 3 4 )
0) V) C C a .oc.4C

0 o-

4- a: _0 4- S

00 0- %0 10

a) I) * % -
> 0I

a)0 V 3 ,
S U) 00 0 0

4- -D ~ 4E -0
0 0)

O 0) L N. co

0- 0^
1- 0



N 0





* *

K> 0
N0 -



Nt 0









a) 4-
















a) V

u- a












a) -










u ) c





- Un
Ch co

* *
* S
* *
* *
-n o

o co

o I



O <
OX -

> o|

to -
o ro






0- -0

r= o


,I- 0
-o o




r r-

t a_


0 -
U)1 U.)
I U)

0 C
0. (0
0 S

I Q )
0 -

0 C

x w

0 C

0% o
%0 0

1o r-
o o


oo r-
%0 0A



l n

'0 u0


o r-~


1 0)
10 U)

- 0%

N -

N -

NO -
LA n


c c

D 00

o o
0 C4


r- \

0 0!




N -

N en
N 0
o o


L 4n








10 -

( 0

0 0%







0Q -

but not paired females, exhibited a lower aggressive index against pair

members than against opposite-sexed non-pair member (two-tail I-test,

df=19, males and females respectively: =2.25, p<.05; t=1.41, p>.05).

The frequency of nesting arrangements, for days on which complete

nest data were available, is presented in Table 10. It is of interest

that two animals of the same sex nested together only once without an

opposite-sexed animal also present. Animals nested as two opposite-

sexed pairs on over 1/3 of the days for which data were available and

almost another 1/4 of the nesting arrangements observed included one

opposite-sexed pair.

Only data on nesting behavior for days on which nesting

relationships were known for all individuals in a group were analyzed

statistically. Nest data were available for all but one paired and one

single group. Three days of nest data (the total possible) were

available for two of these nine paired, and seven of these nine single

groups. Two days of data were available for six paired and two single

groups, and only one day of data was available for one of the paired

groups. The mean number of days of data available for nine paired and

nine single groups were 2.11 days and 2.78 days, respectively.

The analyses of nesting behavior presented in the tables are based

on the "Wilcoxon-test" (Siegel, 1956). Analysis of nesting behavior by

this test in the present study may give more "weight" to observations

from groups for which more days of data are available. Significant

comparisons in the tables that were not also significant by the "sign-

test" (Siegel, 1956) are noted in text. Animals generally nested with

opposite-sexed rather than same-sexed animals. Differences in nesting

Table 10

Frequency of all Possible Nesting Arrangements

Paired Single
Total Groups Groups

All four together 4 1 3
Two opposite-sex pairs 15 7 8
Two males, one female 3 3 0
Two females, one male 6 4 2
Two males 0 0 0
Two females 1 1 0
One opposite-sex pair 12 2 10
None together 3 1 2

Only days on which all animals could be accounted for are included
(Total number of days=44).


z r--

N W -
0 C O -r O

0 a) cCOM-

IW -C'
a u - -

L m

-rC EX 4
-a 1 1 I *

*x a

0.0) c o x
() a 0rV (

0 0 )
i~o- o -

U) a)- -I C o
oi 4- U 0
> t 0

0E X C-

kn t
0 N C nar- X v
0. L* Il *

w 0 0 U
EU 0

U U Ut *0

-- s -s s
ON I) 0-

zz an N e

-CD t

go ODU 4
C" Q

LL- x r O)0
00L C0 -o - C
0)r +- ID C.
0 5 -D
CO *

) U E 4 *I

,A 0 -
EX C r- C

inU 0
oC Ic I -

)4- Q) C 00

1w wl 01 X *

0 LO) 1 D = -
-EX cv-0

cn = 0.11
l..U) U)0o Zu
c ~ C_< /







o -


0 3

c) w

I -

31 t )

0 -


0 -


X L-

* *


(M '*O IO

r- 0 0

- ?N 0

' (s 0

r- N
(N -

\o in

K% 0

- CM

1 0 '0
* *

L. U

- 0- g
M -
* *

N- -1

t w- (


an c


C -0 0)
L 0)

< 0 (I)

frequency with same and opposite-sexed animals were significant for all

comparisons except paired animals in general, and paired females (see Table

11). (By the sign test the comparison for all females was also non-

significant, x=10). In general animals also nested with high-aggression

rather than low-aggression animals of the opposite sex (see Table 12). This

was true for animals in paired groups and single groups, and for males and

females. The finding of non-significance for comparisons within paired and

single groups, except single females, may be due to the small number of non-

tied observations available for comparison. (By the sign test comparisons

for paired animals and males in general were also non-significant, x=2 for

both comparisons). Total frequency of aggression was correlated with

frequency of nesting with opposite-sexed individuals for paired males and

single males (Pearson correlation, r=.710, p<.001 and r=.519, p_<.05

respectively) but not for paired or single females (r=.066 and r=.219

respectively, p>.05).

Animals did not nest more frequently with familiar individuals.

Comparisons based on all paired animals, males, or females were all

non-significant (Wilcoxon, N=number of non-tied observations: all

animals, N=24, z=1.24; males, N=10, z=1.22; females,

N=14, z=.60). Neither the number of days pairs had built nests

together during the nest building stage, nor the average type of nest

built, was correlated with nesting frequency with pair members (Pearson

correlation, nest days, r=.395; nest type, r=.250).


These seminatural observations provide evidence that aggression is

likely to be an important factor in social interactions in P.

polionotus. As predicted of the behavior of monogamous species

(Klelman, 1977) frequent aggression was displayed by both males and

females, and the majority of aggressive behavior was directed against

same-sexed individuals. The suggestion that females of this species

are normally dominant over males (Smith, 1967) was not supported by the

results of the present study. Males displayed much higher levels of

aggression than did females, and females generally exhibited very low

levels of aggression toward males. In addition, high-aggression males

more frequently performed a behavior, aggressive digging, that could

function to display their aggressive status.

Although the frequencies of aggressive behaviors In paired and

single groups were very similar, the aggressive index indicated that

individuals In paired and single groups behaved differently.

Individuals in single groups did not appear to discriminate between the

targets of their aggression as well as paired individuals did.

Overall, the differences In aggressive behavior between paired and

single groups may Indicate a tendency for reduced aggression toward

opposite-sexed Individuals, especially pair members, in paired

Individuals. This is particularly true of paired males, which

displayed both reduced total frequencies of aggression and a lower

aggressive index, against females with which they had been paired than

against those with which they had not been paired.

Superior aggressive ability would appear to provide some social

benefits for individuals, as both males and females nested more

frequently with high-aggression rather than low-aggression individuals

of the opposite sex. On the other hand familiarity, although It

appeared to reduce aggression between pair members, did not have a

significant effect on nesting behavior. It would appear from these

observations that, in P. pollonotus, aggressive ability may be a more

potent factor in social preference than in familiarity. However,

because It is unlikely that Individuals under the present conditions

were always able to control who nested with them, It is probably best

to use caution in interpreting these results.

Aggression and Familiarity Preference Tests


The results of the seminatural experiments Indicated that

aggression may be an important factor in social preference in P.

pollonotus, but cast some doubt on the importance of familiarity in the

social preference of this species. This experiment was designed to

test preference based on aggression and familiarity in P. pollonotus

and EL maniculatus in a more controlled manner, through the use of a

preference apparatus.


Subjects were Individuals of two species of muroid rodents, fE

pollonotus and E. manculatus. A total of 40 animals of each species

served as subjects for aggressive tests and aggression preference

tests. Prior to serving as subjects, animals were maintained as

described in the general methods section. Within each species these

animals were each assigned to one of 10 groups, with two animals of

each sex per group. In addition to the criteria described in the

general methods section, no Individual within each group could be

related by more than two common grandparents to any other animal in the


Following the aggression tests and aggression preference tests the

animals described above also served as "stimulus" animals for

familiarity preference tests, while an additional 40 animals, 20 of

each sex of each species, served as "choice" animals for these tests.


Procedures were identical for each experimental group. Animals

for each group were separated from litter mates and individually housed

in 48 x 27 x 13 cm clear plastic cages. Peromyscus maniculatus were

moved from their colony room to the P. polionotus colony room. On the

following day, within the first 1/2 of the dark phase of the

photoperiod, animals were lightly anesthetized with ether and marked

for identification by shaving them in one of two patterns: either (1)

a band was shaved from around the neck area, or (2) a band was shaved

from around the middle of the animals. One animal of each sex was

shaved In each pattern. Animals were placed in 48 x 27 x 13 cm plastic

cages modified (as previously described under aggression apparatus) for

aggression testing.

All adaptation and testing were conducted during the dark portion

of the photoperlod. Animals were adapted to the preference apparatus

on the two days following the marking procedure. On the first of these

two days animals were adapted to the procedure that would be used when

they served as "choice" animals.

Adaptation to "choice" procedures was as follows: the animal was

placed In the start box for 5 min, followed by 1 hour free in the

apparatus without other animals present. Animals that did not exit the

start box within 1 1/2 min after the door was lifted were gently

prodded with the eraser end of a pencil, often simply lifting the lid

of the start box slightly provided sufficient stimulus for the animal

to leave the box. The same procedure was followed during tests.

The day following adaptation to choice procedures, animals were

adapted to "stimulus" conditions. Adaptation for stimulus animals

consisted of being placed In the stimulus boxes at either end of the

preference apparatus for 1 hour. Animals which were tested together as

stimulus animals for experimental tests were also adapted together.

During adaptation of stimulus animals for the aggression preference

tests an opposite-sexed pretestt" animal was allowed free in the

apparatus and its visits to either chamber recorded. This period was

designated as the pretest period. Each pretest choice animal had been

adapted to the apparatus previously, and served In several protests

with animals of the opposite sex.

Following adaptation to the preference apparatus animals were

adapted to the "aggression apparatus" for 2 hours on each of the next

three consecutive days. For these adaptation periods the animals home

cage was connected to the aggression arena by means of lengths of Tygon

tubing inserted into the holes cut in the sides of the home cage and

into the sides of the aggression arena. Animals were restricted to the

arena for the first 40 min of the 2 hr period by means of #6 black

rubber stoppers inserted in the ends of the connecting tubes. The

stoppers were removed for the remainder of the period so that the

animal had access to both the arena and its home cage.

Aggression tests were conducted on 3 consecutive days following

adaptation to the aggression apparatus. Males and females were

observed. Tests were conducted by placing the two shaved animals of

the same sex Into the aggression arena and observing them for 40 min.

Behavior during this period was categorized as approach, displacement,

aggressive, or submissive. Frequency of all aggressive behaviors

(total of all attacks, chases, and fights) was Included under one

category because aggressive behaviors other than attacks were extremely


Aggressive preference tests were conducted on the 2 days following

aggression tests. Two animals of one sex, that had been tested

together In the aggression tests, each served once as choice animals.

The two animals of the opposite sex, that had been tested together on

the aggression test, served as stimulus animals for both tests. Tests

were arranged by placing one of the same-sexed pair of stimulus animals

in each of the boxes at the ends of the preference apparatus 15 min

prior to the beginning of the test, and the choice animal In the start

box 10 min later. Tests were initiated by raising the door to the

start box, and thereby allowing the choice animal access to the

apparatus. Test duration was 1 hr, timed from when the choice animal

exited the start box. Order of testing was counterbalanced for sexes

across days. Although the stage of estrus was not controlled for in

these tests, smears were taken for each female after she had served as

a stimulus animal and on the following day.

On the day after the conclusion of aggression tests for a group,

the members of the group were each housed in a 29 x 19 x 13 cm plastic

cage with an Individual of the opposite sex. These "new" opposite-

sexed individuals had been adapted on the previous day to the apparatus

under the procedures described for choice animals. These animals each

served once as choice animals In the familiarity preference tests that

followed. Animals that had been paired for the aggression preference

tests also served together as stimulus animals for the familiarity

preference tests (The two same-sexed stimulus animals for each of these

tests were two Individuals that had been partners In tests for

aggression). This arrangement produced four tests for each group, two

tests for each same-sexed set of stimulus animals.

Smears were obtained from the females of each group prior to the

onset of the dark period on the seventh day after animals had been

paired. Tests were conducted with stimulus females only when both of

these females exhibited smears consisting of at least 75% leucocytes.

This type of smear would normally Indicate a nonreceptive state.

Females that displayed sperm on the smear were tested after they

displayed this type of smear. Choice females were tested individually

if they displayed smears with at least 75% leucocytes. Test procedures

for choice and stimulus animals were as described previously.


Aggression Tests. The level of aggression (attacks, chases,

fights, and rough-and-tumble fights) displayed in the aggression tests

were very low. No aggressive behaviors were displayed in four of the

ten groups by females of either species. Peromyscus polionotus males

did not display aggression In two groups, while P. maniculatus males

did not display aggression In three groups. The comparison of total

values (for all three tests) for the four behavioral categories

recorded are displayed by sex for each species in Table 13. Only the

comparison of P. pollonotus males and females on frequency of

aggressive behavior was significant, with males displaying a higher

frequency of aggressive behaviors.

Because the frequency of aggression in these tests was very low

it was difficult to determine which animal of a pair was actually the

"most aggressive". Instead, the total frequency of approaches and

aggressive behaviors was used to provide an indication of which animal

of a pair might be more aggressive. This total frequency score,

although not an aggression score per se, does allow a comparison of the

tendency to initiate interactions or "assertiveness" of the two

Individuals in a pair. It would seem reasonable to expect that the

less timid of two individuals under these test conditions might also be

more likely to exhibit more aggression under other conditions. The

individual of a pair of animals in aggression tests that displayed the

greatest tendency to initiate Interactions will be termed the

"high-interaction" individual, while the individual that displayed the

lower tendency to interact will be termed the "low-interaction"


Aggression Preference. Table 14 presents data on the preferences

of animals of each sex for high or low-interaction Individuals of the

opposite-sex. No preference for high-interaction animals of the

opposite sex was displayed by P, pollonotus of either sex. Peromyscus

maniculatus males, however, did display significantly longer durations

of visits with high-interaction females than low-interaction females

%o i co oC

CO C% U11%
a- in '0 0


C-4 I- r- o
N Op

O' r- n 0


.' N UN ..

4 -

N- c-


C4 C4 ni

no tnLn o
o4 0 W 0
i mn -t
ac i o

O 4-

0 -

L. L 0- E


m =
0 u


co w























0 1
- l
Z 0>

0 I')
00 v


.c u

1 0 in
-I-~ 0

- w c
K> 01 0%

C 0%
-- N r

In 0 PI

rr %

*o IN 0%


'0in 0%

%tI0% 00
400 %


0n ocM

0% 0
ino K


0A 0 (D
ino co
-- 0



mo r

L 0
4- 0

o I- -Z

- rN 0%

o) 0% -

A CM 0%
N 0

o i-1 -
0C 00
0% in -
IA 0% N

N W% N

mIn -I
a> m f


Ln Co C4

(A 04

o -

CO r- tU
0 CM CC)

o N N

C) in (A
In 0%k
in IA r-



F' 0%
*t '0
^'0 0%

w0 00 Ia
0 0'0


4- In
In -

4- (
0> -
F5 *5


-4+- -
In 0 -
L- x

4- 0o
4- <0
z \I-


u L/)


- 0)

S.- 4-)
4- 4-


E3- E


"- -0
E a

c> c

C 0
.) U


and 16 of 20 males visited high-interaction more frequently than

low-interaction females (sign test, N=20, x=4, p<.01).

High-interaction P. maniculatus females and low-interaction P.

maniculatus males both displayed significantly longer durations of

visits with high rather than low-interaction animals of the opposite

sex (means for high-interaction females with high and low-interaction

animals = 1155 sec and 159 sec, one-tall paired-I: df=8, =2.37, means

for low-interaction males with high-interaction and low-interaction

animals = 1559 sec and 468 sec, one-tail paired-: df=9, =2.13;

p<.05 for both comparisons).

Although the stage of estrus for females in aggression tests was

not controlled, data were available for this factor. Comparisons of

male preference for high-interaction and low-interaction females were

made for these tests In which both stimulus individuals were In

diestrus (see Table 15). The only significant finding was a preference

by P. maniculatus males for high-interaction females by the measure of

duration of visits. This is also the only comparison which had been

significant when data from all females was included. Comparisons could

not be performed within the non-diestrous condition as there were no

cases for P. polionotus in which both stimulus females were

non-dietrous, and only two such cases for P. maniculatus females.

Comparisons of male preference for diestrous versus non-diestrous

high-interaction females and diestrous versus non-diestrous low-

interaction females are presented in Table 16. Peromyscus polionotus

males displayed more visits to non-dlestrous high-interaction females

than to diestrous high-interaction females, and P. maniculatus males

Table 15

Comparison of Male Preference for
High-Interaction and Low-Interaction
Females in Diestrus

High Low
Interaction Interaction

Measure Mean (SE) Mean (SE) t

F. polionotus

No. of visits
Total duration of visits
Mean duration of visits

e. maniculatus

No. of visits
Total duration of visits
Mean duration of visits











All durations are in seconds.
Paired- EP. polionotus df=13
1-tall *D<.05

P, manlculatus df=5


n *

m 4 0 %Q


'tF co ot r-W W

E w

4,- Cr -
uo ro c o : K% r

0 C CDM i0n

0 0
0 40 A nz 90 0 0
g r.- i n

-o 0 -
0 0
v0 SM M C *q-

rO C
I M W! W!- 9 9 '


03 X 000 0
v nc

'o 0 l CD -I 0 W%
E C Cll V C4

>>- >u
w Jw0 *D0
-- c0N 0
CD E %0 % Coo i J %-

QO vl c cIfolOflO0

+S- D- 0

Cn ri- U- CL C; 00"0

0 0 0-- g c

O 0000400

= >z >x 0
An nW u I

( c -0 0 > L-

Z - 0- V (

(D 0 00 Q 1
=--z- 2-- 0

1i a^~ to E


S% 0 co r co N

u 0% 10 -0 0% Im
4- -

o 2 % r. r

4-- o Mo>4
LA Z. lT K-N


C3 O

a 0 a w%

0- E er14 0 co C M

at LU a
; *o
(ULL +

0 a a a

0- Im A o0

o -

ro c ---L

0w 0 o -'O- '

S0 1

(r 9.s c- c K to>(Moo
4- 0 0 c* 4

"D. 0CM ~0-O

'D A
03- -a g

c 4- -u O
'+D; --0- 0

xzo- x
4- 4-



X 2-1X '*

visited non-dlestrous low-Interaction females more frequently than

diestrous low-interaction females. All other comparisons of male

preference based on stage of estrus were non-significant. The stage of

estrus did not significantly affect female preference for high or

low-interaction males of either species (see Table 17).

Familiarity Preference. Male and female P. pollonotus did not

display preference for familiar individuals over unfamiliar individuals

of the opposite sex by any measure, although the comparison of the

total duration of visits did approach significance for females

(one-tall paired-, df=19, =1.47, p=.054). Peromyscus maniculatus

males and females displayed significantly more visits to familiar than

to unfamiliar individuals of the opposite sex. Females also displayed

longer durations of visits to familiar males than to unfamiliar males

(see Table 18), and a greater number of females exhibited more visits

to familiar than to unfamiliar males (one-tall sign test, N=18, 2=5,


Because stimulus animals In familiarity preference tests had

previously been evaluated in aggression tests, the "attractiveness" of

these animals in familiarity preference tests could also be evaluated

on the basis of their tendency to Interact with other animals. These

comparisons are presented in Table 19. Peromyscus Dollonotus females

displayed significantly more visits to high-interaction than to

low-interaction males, and the difference in the number of females that

spent longer durations with high-interaction than low-interaction males

was also significant (one-tail sign test, N=20, x=5, p<.05).


tu 0.

ra U

E 4-

- (E

Sr- 2

*.- t
E -

rt no
LL 0
S- X

na c>

4- CO

4 *-

0- 0




C. c


L (

c c




a- W% Uo

0% In 0

in N>

0>- -

z- x

in -

- I

-- 0

..- -O

z -

* N

CM 00
*- 0 f
M N -

- 10 0

0% N

o0 r-


CM 01 0

U* .% *


o r.


co -

trn 0

0M a




4-- 0



nc ~ a -. co $


Sc W% 4 a, n

E a
S1 4- .

V) :n o0


0o) 0
4 r -- o 0%

4-- 50 04 NO in %
I 0- C. C4 r- o

0 .)
4-L/ 4) In
w oL- 0 !

'-4 ( 0 0 a oC'
.) ( --
W r-D .4-- .l..

0 0 c_% in N 0 co

4- E "
toi- ID o

4 Fn C- in %a n
1ro 4-
0) 4 0%I

ro r 0 C- 4 Ln (N y

4- a

cu a O S ^

S- -


0 0 C
>J >
.4- I.-


a ; \IiC
01 *I 0-


Whereas individuals observed in the seminatural apparatus had

exhibited appreciable levels of aggressive behavior, levels of

aggression displayed by individuals in the aggression apparatus were

very low. This finding was somewhat unexpected. The object of the

procedures followed In the present study had been to condition

individuals to treat the aggression arena as an extension of the home

cage. In previous observations Individuals housed in cages identical

to those used to form the aggression arena had displayed fairly high

levels of aggression to Intruders after one to two weeks of residency

(unpublished observations). In the present study individuals were

tested for aggressivity in the arena on the seventh through ninth days

of residency in the home cage. Although individuals did not have

access to the home cage during aggression tests, they had been allowed

to travel freely between the home cage and aggression arena during

adaptation. In addition, during aggression tests the home cages and

the aggression arena were attached in a manner that allowed a fairly

free flow of air between them. Many Individuals were observed to spend

long periods sniffing and gnawing at the entrances to their home cages.

Individuals therefore appeared to recognize their home cage as opposed

to a strange cage (unfortunately these observations were not

quantified). During aggression testing, however, Individuals did not

behave as residents in the aggression arena, nor did they defend the

entrances to their home cage. Animals In these tests were

simultaneously exposed to olfactory cues from both their home cage and

that of their opponent. These test conditions may have led to

conflicting "fight" and "flight" responses (See Hinde, 1966), and

thereby resulted in low aggression scores.

It Is of interest that Individuals In the seminatural apparatus

displayed high levels of aggression toward one another immediately upon

being allowed access to the central arena, even though none of these

Individuals had previous exposure to this area, and had been in

residence In the connected home areas only three days. The observation

of higher levels of aggression under the seminatural conditions than

under aggression-test conditions may be due In part to the fact that

Individuals in the seminatural apparatus were exposed to opposite-sexed

Individuals during tests while individuals in the aggression apparatus

were not (Barnett, Evans & Stoddart, 1968; Brain, Benton, & Bolton,

1978; deCantazaro, 1981; Flannelly & Lore, 1977; O'Donnell,

Blanchard, & Blanchard, 1981). Exposure to females has been

demonstrated to Increase male-male aggression In P maniculatus balrdl

(Terman, 1982; Dewsbury, Personal communication). Exposure to

opposite-sexed Individuals was, however, not required to elicit

aggression In the previously mentioned tests of resident B. polionotus

in aggression-arena-sized cages; and males in the seminatural

single-housed condition were observed in several tests to Initiate high

levels of attacks and chases prior to exposure to females.

Preferences for high-interaction (and presumably more aggressive)

individuals were not as strong as might have been predicted from the

results of the seminatural experiments in the present study, or from

the results of previous studies (e.g., Blair & Howard, 1944;

Eisenburg, 1962). Although in the majority of comparisons (over 80%)

scores on the measures of preference for high-interaction individuals

were higher than those for low-interaction individuals, most of the

differences displayed between high and low-interaction individuals were

non-significant. This may be a reflection of the low levels of

aggression displayed In these tests. Previous investigators (e.g.

Huck et al., 1981) have hypothesized that the differences in odors

displayed by dominant and subordinate animals are mediated by

differences in the physiological changes Induced in these Individuals

through aggressive encounters. The levels of aggression in the present

tests may not have been high enough to fully induce the physiological

changes necessary for clear-cut discrimination on the part of choice


Although levels of aggression were low in aggression tests, females

of both species, and E. maniculatus males, displayed significant

preferences for the more "assertive" individual of the opposite sex

under some preference test conditions. Significant preferences were

not displayed for low-interaction individuals of either sex in either

species. Under the assumption that Individuals that are more assertive

would also normally be more aggressive, these findings are consistent

with the hypothesis proposed earlier that Individuals of these species

should prefer aggressive opposite-sexed Individuals, and also with the

nesting behavior exhibited by fP pollonotus In the seminatural


In familiarity preference tests P, maniculatus of both sexes

displayed preference for familiar Individuals of the opposite sex,

while P. polionotus did not display such preference. Although the

lack of preference displayed by E. pollonotus In the familiarity

preference tests Is consistent with observations made in the

semlnatural study, results of the familiarity preference tests are

contrary to predictions made earlier as to the behavior of these two


Sibling Preference Tests


Individuals of most rodent species will be exposed to siblings

during development. Such exposure may, of itself or In conjunction

with genetic factors, influence mate selection (Grau, 1982; Halpin &

Hoffman, 1982; Smith, 1966). The general consensus holds that, except

under special circumstances, Individuals should prefer to breed with

non-relatives (Daly & Wilson, 1978; Dewsbury, 1982a; Krebs & Davies,

1981; Wittenberger, 1981). Rasmussen (1970), however, has suggested

that Inbreeding may be fairly extensive in some species of Peromyscus,

particularly P. maniculatus (Rasmussen, 1964); and Howard (1949) has

proposed that inbreeding may account for as much as 10 percent of

breeding in P. maniculatus. Smith (1966) has suggested that "a

considerable amount of Inbreeding" (p. 50) also occurs in En


Not all Investigators agree with these proposals however.

Selander (1970) has questioned the genetic basis for Rasmussen's (1964)

conclusions regarding E. maniculatus, and other Investigators

(Dewsbury, 1982a; Hill, 1974) have demonstrated suppressed

reproduction for sibling matings in this species. Foltz (1981b) has

also questioned Smith's (1966) proposal of high levels of inbreeding In

P. pollonotus. The present study was designed to evaluate preference

for siblings in P. pollonotus mand. P maniculatus.


Both P. pollonotus and ,. maniculatus served as subjects for

sibling preference tests. Subjects were selected from 12 litters of

each species that contained at least two animals of each sex. Two

animals of each sex were selected from these litters, under criteria in

the general methods section, and maintained together throughout the

experimental procedure. Litters selected for groups were unrelated by

more than one common grandparent. Two litters of each species

comprised an experimental group.


Individuals were ear-punched for identification, and .

maniculatus litters were moved to the P. pollonotus colony room. On

the following day animals were adapted to the preference apparatus,

without other Individuals present, as both stimulus animals and choice


Preference tests were of two types, same-sex tests and opposite-

sex tests. In same-sex tests choice animals had a choice between a

same-sexed sibling stimulus animal and a same-sexed nonslbling. In

opposite-sex tests animals had a choice between two opposite-sexed

individuals, one a sibling and one a nonslbling. Each animal in a

group served twice as a stimulus animal and once as a choice animal for

each type of test. In an opposite-sex test, for example, males 2 and 4

would serve as stimulus animals for choice females 2 (one of the 2 male

siblings) and 4 (a 4 male sibling). They would each also serve as

choice animals with stimulus females 1 (a 2 male sibling) and 3 (a 4

male sibling).

Preference tests were conducted on the two days following

adaptation. All tests of one type (i.e. opposite-sex tests) and half

of the tests of the other type were conducted on the same day. Testing

was completed on the second day. The order of test type across days

was counterbalanced. The number and duration of visits to each chamber

were recorded on each test and for the adaptation period. Vaginal

smears were obtained for each female before the beginning of the dark

cycle on each test day.


Peromyscus polionotus females demonstrated a preference for

siblings over nonsiblings (see Table 20). Sibling males were visited

significantly more frequently than nonsibling males, and a

significantly larger number of females spent longer durations with

sibling rather than nonsibling males (sign test, N=24, x=6, 2-tall

p<.05). They also spen+ significantly longer durations with sibling

rather than nonsibling females. None of the comparisons for siblings

versus nonsiblings were significant for P. pollonotus males or for EP

maniculatus of either sex (see Tables 20 and 21). It is, however, of

Interest that scores for P. pollonotus males on tests with opposite-

sexed animals mirror those of P. polIonotus females (higher scores for

siblings), while for all but one measure (average duration of visits

for females) E. maniculatus display higher scores for opposite-sexed



N ~ N

-0 Nc N

-N N-O

0 t
Z T C coc
Ei Z r e-

0 CM i
E Q O co in c

tA aN %0i

0) CQi, co e

o +

C) 0C

cn Ln c r

0x C -0 (NM

w) - w In
-E 2
co O- C-ON

(1 1C *- *4*-
00 N a

O c-4or- -n

Ia 0 O pp C40
- -1N
* 0 M
rQ s C(4 *

CA% 0 o %0
r-S +N WI

4- 4-

do 00 ;c
(UJ - aa %-c- n
(j 00 co N.
x a% C> mmo N OM
ao Wc x-
W) li 0 -i a

U C -cl MO-N N V

Sc 0T M -&-l


+- 2N a 0,

0- o- 10 0
0 0 C0
C4 C 00( 0 0)

u c = 4- I - x
0*- 0*- L4-
0 O 00

2: 2<
.4-0 --



- L



n I

.- -


*- I4

ro c

o -o



0 cU

z )


C 0

c c
0 (0

z C
Z "

r- 01 p-
SCo -
KC 0

a) ( F4

in C4 -



N c-
r-O -
.O. V%
N -
- o -

%0 Lco
- %0 co

00 0 0

co o

F- (N -


+- 4-

- CO 0

10 4 -

o> co

0 0 Q)

CO -


4-- 0
-0 3


2 1-


N -

--10 %

(N r- c

co !n
O W-0

C in



-n N

r- (n

0 mF


0> -~


0 (0


-- 0(D
z 2:

S- m

' T


0 0
2 -

Table 22

Comparison of Male Preference for
Sibling and Nonslbling Females in Diestrus

Sibling Nonsibling

Measure Mean (SE) Mean (SE) t

P. polionotus

No. of visits
Total duration of visits
Mean duration of visits

L maniculatus

No. of visits
Total duration of visits
Mean duration of visits










All durations are in seconds.
-test P. polionotus dL=11
p>.05 for all comparisons

P. maniculatus df=14


% % co CM U"% 3
Go 0 0 r- 'T

u, U) Nq,)-N'o
W n 03 NNC
in 1n M T (>

ID Z %O'0 %0%0%0

0 ca
o) a (M 94 M ~c -
SCD r- % -0w

o C-

4- Z 0 w C 0 r-r-N

C- 0C
t( Qe co arcoor-


VI0 N I .. .

4-) 1
f C1O C 00 CO

*in V C)
C2 C1 C14

(o re a D z 'o 'D O 'Doio
o E o)

Cd 1 0 0 I WI) Nw -

a) r Q1 0 a a I 5oa%
0) 0 ogoo o a
o co- i
co W CUh ON 4 ao

*Or0 M Ccla

m w 0 co C n
(Ai - - - -
% N N CO% C

0 3o ou

a) X CA-KnN CN -
U- r 10

Q'0) a

me >
S* -U 00 a

4- OOo o a m


Q)+-- 4- o

0 0 0- a
.c >
S0 0COC -0

0 0Z 0 01 a0
5:- Z +- C -:
--ut rn n -

II I-fl 4 4 0 0-(


in C r- 0% r-
St *N s tcN

LU I- oT a' N-
I Ao0 %0 i

4- a -,ce
S Z vo

I a M


co omu*noomo
(A4 4 -

N a LU Nme
cn E rU !
2 CM -

0 4t- Z -ocoo rc n

0 =II I 23 ooor a

U) C %0
a ro

-0 4-
a 0 0 CCanm ar-

C O i N hN-IOr-

S. .

/) a: CM 4- -
& C-

U) 4-)

0O m 1 o CAi

SC --W
0> 0

(a C C C i V
a a

0 0 C 01 0

,.5 o4 0 a r_
+-- 0 *
cL i CI

0 I C 0 0

01 o o 2
S - i -
-0 -

S0) 0 0 -

x -- - m z x

3 0 0 0

--C) O B D"2

S zzt-c-. <1


%0!? 0% C4 ,

co 'n N Nin I'

W o

S ra. r- goo .n No-W
( to \or


e z on nnm-U

So o n
mi 0 1 D. .
O . 0 >0r


-n W- c .

0 i- co

4"-- .0 -- 0

r- *rrG a
CULL ) Z ur-p--i

P fo %0 -NW M O- a

A -
0V) N2" 0

c ---r

S-t- -. .

c-r-1 0 0- cncm 0

+ 43) rzN iN 9

S s% 04
En '0 LN%% a0n



' z- 0 4- o SOrI A

a) = 0 oo or

O l c mC +OO o
- U,

a) O I On=
o o
1-c -.Z L

0 N QO> N> -0
V 00

0- 0 0
gj C iri~mmmin >

0 N S MOO- N- 0
0n '0 oO

in c l *o 0
a.-- O cN

Sin c C > -

0 0 0 m -
In >

cl I c

0 c 4- 00 4-

_0 0 0 C 0 )

0)00 03
C 4- 4- 0 0


4- 4z l--- 0

In 00CC
3c 4 4- 0-
0 00000
z zz-,-ri

As with the aggression preference tests comparisons were made of

male preference for siblings and nonslblings In diestrus and of male

preference for diestrous versus non-dlestrous siblings and nonslblings.

None of these comparisons were significant for either EP pollonotus or

EP maniculatus males (see Tables 22 and 23). Comparisons of the

preferences of females in different stages of estrus did, however,

yield some significant differences. Although P. pollonotus females

exhibited no significant differences in preferences based on stage of

estrus, P. maniculatus females did display significantly greater

preference for nonsibling males, by the measures of total and average

durations of visits, when in a non-diestrous condition over a diestrous

condition (see Tables 24 and 25).

The possibility of litter effects on preference scores was also

evaluated by conducting a one-way analysis of variance on the

"difference scores" between preference scores for siblings and

nonsiblings. Significant litter effects were found for the average

duration of visits by E pollonotus males to females (F(11,12)=3.77,

-<.05). and for the total duration of visits by EP pollonotus females

to males (F(11,12)=3.08, p<.05).


While Peromyscus pollonotus males generally displayed higher

scores on preference measures for siblings than for nonslbllngs, these

comparisons were not significant. PE polionotus females, however, did

display significant preferences for siblings over nonsiblings. The

results of this study, therefore, may offer some support for proposals

of high levels of Inbreeding in P. polionotus (Smith, 1966). The

observation that scores for preference by EP maniculatus males and

females were generally In the opposite direction from those for P.

polionotus Is of Interest. This difference may indicate a greater

preference for siblings by F. polionotus than by P. maniculatus.

It is also of interest that non-dlestrous P. maniculatus females

displayed a greater preference than diestrous females for nonsibling

males, since females in a non-dlestrous condition would be more likely

to be receptive. These results are consistent with previous studies

demonstrating suppression of reproduction in sibling matings of P.

maniculatus (Dewsbury, 1982a; Hill, 1974), and the implicit conclusion

from these studies that inbreeding should be avoided by members of this



In the general discussion the observations of the present study

are examined in light of previous research, and interpretations are

suggested for these observations that are consistent with the ecology

and mating systems of f. maniculatus and E. polionotus. The general

discussion section is divided into six subsections. The first of these

subsections examines aggressive ability as a factor in preference.

This subsection begins with a fairly extensive overview of the ecology

of P. maniculatus and P. polionotus, and also presents evidence for

the role of aggression in the ecology of these two species. The

ecological information presented in this subsection serves as a

background for discussions which follow in all subsequent subsections.

The discussion of the ecology of these species is followed by a

discussion of factors that may provide an adaptive basis for selection

of aggressive mates in these species.

The second subsection deals with familiarity as a factor in

preference and with the effect of prior breeding experience on

preference for familiar individuals, and includes a discussion of the

opportunities that may be available for P. maniculatus and P.

polionotus to utilize familiarity as a basis for mate selection. This

subsection is followed by a subsection on the related topic of kinship

as a factor in preference. The subsection on kinship discusses

evidence for and against inbreeding in maniculatus and P.

pollonotus, and the ecological and social factors that may affect

preference for siblings as mates in these species. This subsection is

followed by a subsection on the evolution of monogamy in P.

pollonotus, and the summary for this study.

Aggressive Ability as a Factor in Preference

Many of the observations In the present study were consistent with

the general hypothesis that aggressive ability may be Important in the

social behavior and mate selection of monogamous and polygamous

species. They were also consistent with the specific predictions that

aggression should be an important component of the social behavior of

P. pollonotus, and that P. pollonotus and P. maniculatus of both

sexes should prefer mates with high aggressive ability. In the

seminatural experiments aggression was routinely displayed by both

male and female P. pollonotus, and both males and females of this

species nested more frequently with the more aggressive of the two

opposite-sexed individuals. Female P. pollonotus also displayed

preference for the more assertive of two males in familiarity

preference tests; and preference for more assertive individuals was

displayed by P. maniculatus males and high-interaction P. manlculatus

females in aggression preference tests.

The preferences exhibited by these species may be mediated through

advantages in reproduction gained by individuals that choose mates with

good aggressive ability over those that choose mates with poor

aggressive ability. Some of the possible advantages accrued by

Individuals that choose mates with high aggressive ability have been

University of Florida Home Page
© 2004 - 2010 University of Florida George A. Smathers Libraries.
All rights reserved.

Acceptable Use, Copyright, and Disclaimer Statement
Last updated October 10, 2010 - Version 2.9.8 - mvs