A comparison between male-induced abortions and male-induced estrus in prairie voles (Microtus ochrogaster)


Material Information

A comparison between male-induced abortions and male-induced estrus in prairie voles (Microtus ochrogaster)
Physical Description:
vi, 173 leaves : ; 29 cm.
Bryan, Judith C., 1944-
Publication Date:


Subjects / Keywords:
Prairie vole   ( lcsh )
Sexual behavior in animals   ( lcsh )
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )


Thesis (Ph. D.)--University of Florida, 1993.
Includes bibliographical references (leaves 158-172).
Statement of Responsibility:
by Judith C. Bryan.
General Note:
General Note:

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001925699
notis - AKA1618
oclc - 30608360
System ID:

Full Text








Many people helped me in this research and in my

graduate career, but two individuals were particularly

helpful: the chairman of my committee, Dr. Donald A.

Dewsbury, and Dr. Jane Brockmann. Both of them taught

me to think carefully about various issues in animal

behavior and to write precisely and logically. I also

would like to thank the other members of my committee,

Drs. Marc N. Branch, Louis J. Guillette, and Neil E.

Rowland for sharing their expertise in their particular


My peers in the Departments of Psychology and

Zoology have also made my graduate experience easier

and more fun than it would have otherwise been. I

especially thank Jo Manning who was always ready to

sympathize with my difficulties and to discuss the

issues. Bruce Ferguson, John Pierce, Steve Taylor, and

Allen Salo were also very helpful.

Staff members of the Department of Psychology were

also always ready to assist me. In particular Cheryl

Phillips, Chris Wilcox, Theodore Fryer and Isaiah

Washington all contributed.




ABSTRACT ................................................v


1 INTRODUCTION......................................1


Male-induced Abortions in House Mice.............14
Discovery of Male-induced Abortions........... 14
Circumstances of Male-induced Abortions
in House Mice................................17
Mechanisms of Male-induced Abortions in
House Mice...................................27
Wild House Mice...............................38
Summary of House Mouse Studies................ 39
Male-induced Abortions in Microtine Rodents.... 40
Prairie Voles (Microtus ochrogaster)..........43
Field Voles (Microtus agrestis)................46
Meadow Voles (Microtus pennsylvanicus)........ 51
Other Microtines..............................55
Male-induced Abortions in Other Species
of Rodents.....................................58
Syrian Golden Hamsters (Mesocricetus auratus).58
Mongolian Gerbils (Meriones unquiculatus).....59
Deer Mice and White-footed Mice (Peromyscus)..62
Djungarian Hamsters (Phodopus campbelli)......67
Laboratory Rats (Rattus norvegicus) ...........68
Male-induced Estrus in House Mice and Voles.....69
House Mice.....................................69
Prairie Voles..................................76


General Goals, Design, and Methods...............78
General Methods................................83
Experiment 1.....................................88
Methods....................................... 88


Experiment 2 ....................................95
Experiment 3....................................101

4 GENERAL DISCUSSION.............................107

Summary and Discussion of Present
Experimental Results.........................107
Other Evidence of Two Separate Mechanisms......114
Evidence For and Against the
Existence of a Single Sensory Mechanism.... 115
Evidence For and Against the
Existence of a Single Hormonal Mechanism...121
Conclusions about the Mechanism.............. 132
Discussion of Other Nonadaptive Hypotheses
and Supporting Evidence......................132
Probability of Occurence in the Wild......... 132
The Excessive Stress Hypothesis for Male-
induced Abortions.........................141
Adaptive Significance of Male-induced
Abortions ................................... .142
Advantages for Males........................ 143
Advantages for Females.......................145


REFERENCES.......................................... 158

BIOGRAPHICAL SKETCH................................ 173

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



Judith C. Bryan

August, 1993

Chairman: Dr. Donald A Dewsbury
Major Department: Psychology

A nonadaptive explanation has been proposed to

account for the observation that some female rodents

abort when exposed to unfamiliar males early in

pregnancy. According to this explanation, male-induced

abortions are a side-effect of the generally adaptive

mechanism by which males induce estrus in females. If

this is the correct explanation, then there must be a

single mechanism which is responsible for both male-

induced estrus and male-induced abortions. The goal of

this research was to test the hypothesis that a single

mechanism is responsible for both phenomena.

One way to demonstrate that two different

mechanisms, rather than one common mechanism, are

involved in male-induced estrus and male-induced

abortions would be to show that different responses are

given to the same set of treatment conditions,

depending on the state of the female. In this work I

tested the single-mechanism hypothesis by exposing

pregnant and anestrous female prairie voles to

heterospecific males of two congeneric species, montane

voles (Microtus montanus) and meadow voles (M.

pennsylvanicus). If there is a single mechanism, one

would predict that congeneric males should induce both

abortions and estrus, or they should do neither.

The results show that male montane and meadow

voles induce abortions in female prairie voles at about

the same rate as do unfamiliar male prairie voles.

However, neither male montane nor meadow voles induce

estrus in female prairie voles, under conditions in

which there is a high rate of estrus induction by male

prairie voles. I concluded that two different

mechanisms are involved in male-induced estrus and

male-induced abortions.



Within the last thirty years, a major focus in the

field of animal behavior has been to explain the

behavior of animals in terms of the contributions of

that behavior to the individual animal's lifetime

reproductive success (Dawkins, 1976, 1982; Maynard

Smith, 1964; Williams, 1966). Specific behavioral

traits can be thought of as adaptations, and are often

assumed to have been shaped by natural selection

working at the level of the individual or gene.

The "adaptationist" assumption, that any

particular inherited trait has been specifically shaped

by natural selection for its current function in

maximizing the animal's inclusive fitness, has been

criticized by some (e.g. Gould & Lewontin, 1979; Gould

& Vrba, 1982), and defended by others (e.g. Mayr, 1983;

Parker & Maynard Smith, 1990). The latter claim that,

among other things, the assumption leads to a better

understanding of the selective pressures operating on

the trait, and of the constraints on an animal's

response to the pressures.

I agree with the adaptationist's view that

adaptive explanations can generally increase

understanding and generate further research, and that

many, if not most, of the behavioral characteristics of

animals will, when thoroughly investigated, turn out to

be adaptive. However, it is apparent that even in

specific cases where plausible nonadaptive explanations

have been proposed, there has been a tendency to

dismiss these explanations without further

investigation, and to test only adaptive hypotheses for

the function of the behavior.

Traits that appear to decrease the inclusive

fitness of the individuals that display them are of

particular relevance to the arguments between

adaptationists and their critics, because in these

cases it is more likely that nonadaptive explanations

will be proposed to account for the traits. Male-

induced abortions are a case in point. In some species

of rodents, if a newly mated or pregnant female is

exposed to an unfamiliar male (defined as a male other

than the one with which she mated), she may fail to

implant the embryos or she may resorb or expell already

implanted embryos, depending on her species and timing

of the exposure.

Some terms used in this dissertation need

definition. First, I will use the term "abortion" to

include both loss of implanted embryos and implantation

failure. In most of the early literature describing

the effect of unfamiliar males on pregnancy, the term

"pregnancy block" was used to designate the failure of

implantation that is associated with female contact

with unfamiliar males during the first few days after

mating. The use of the word "block" implies that

something prevents pregnancy from being established.

Later research indicated that post-implantation loss of

embryos is also associated with female contact with

unfamiliar males. The term "pregnancy block" does not

seem to be an appropriate term to use to describe a

loss of pregnancy that can occur as late as

midgestation. The term "male-induced abortions" was

used by Stehn and Jannett (1981). Other authors have

used terms such as "pregnancy disruption," and

"pregnancy failure." It is not certain that pre-

implantation and post-implantation loss of embryos are

due to the same process, however, I believe it is

useful for now to use a single term when talking

generally about these phenomena. Webster's New World

Dictionary of American English, 3rd College Edition

(1988), defines "abortion" as "any spontaneous

expulsion of an embryo or a fetus before it is

sufficiently developed to survive;" therefore,

"abortion" appears to be an appropriate general term.

Second, I will use the term "male-induced" in

reference to estrus and abortions to describe a

situation in which the presence of a male increases the

probability that the female will abort or show evidence

of estrus. The term "induced" in reference to estrous

cycles was used by Richmond and Conaway (1969). Early

use of the term "induced" in the field of mammalian

reproduction occurred in connection with estrus that

followed the injection of estrogens or gonadotropins.

Injection of hormone "induced" estrus in susceptible

females. Later, the term was often used to refer to a

type of ovulation that occurs in association with

copulation (see discussion in Greenwald, 1956). In

some females ovulation rarely or never occurs in the

absence of copulation (induced ovulation). The term

"induced" was applied, with a similar meaning, to the

estrous cycles of female mammals that rarely show

evidence of estrus unless they are in contact with

males. Females may have been selected to use cues

provided by males, along with other cues, to regulate

their reproductive cycles.

Finally, I will use the term "mechanism" in a very

broad sense to refer to the entire process of estrus

induction or abortion including reception of the

stimulus, conveyance of the signal to the pituitary,

response of the pituitary, and response of the

reproductive tract.

If a female rodent aborts when exposed to an

unfamiliar male, successful reproduction is delayed.

On first consideration, this delay would seem to have

the effect of reducing lifetime reproductive success

for that female, especially since rodents have short

lifespans, which should favor immediate reproduction

(Williams, 1966; also, see Keverne & Rosser, 1986). An

adaptationist approach to this problem would be to

suppose that these abortions, in spite of delaying

reproduction, may actually lead to higher lifetime

reproductive success for females that abort in certain

circumstances, compared with those that do not.

Several explanations have been proposed to explain

how abortions might be adaptive for females, including

infanticide avoidance and inbreeding avoidance (Huck,

1984; Labov, 1981b; Schwagmeyer, 1979; Storey, 1986).

However, there is at least one plausible explanation

that does not assume that male-induced abortions, per

se, are adaptive. This explanation is that male-

induced abortions are an epiphenomenon, that is, that

they are a side-effect of a generally adaptive

mechanism by which males induce estrus in females.

According to this explanation, there is only one

mechanism which has both effects, only one of which,

estrus induction, is adaptive. While specific

hypotheses based on adaptive explanations have been

tested (e.g. Elwood & Kennedy, 1990), I know of no

attempts to test the single mechanism hypothesis.

The nonadaptive, single mechanism hypothesis was

proposed by Bronson and Coquelin (1980) for house mice,

and by Stehn and Richmond (1975) for prairie voles.

According to these authors, male-induced abortions are

simply a side-effect of an estrus-induction system in

which exposure to unfamiliar males induces estrus in

females. If male-induced abortions are a side effect,

they may not have been specifically selected for any

function and therefore it cannot be assumed that they

are adaptive. In proposing that male-induced abortions

are a side effect, Bronson and Coquelin (1980)

suggested that there is strong circumstantial evidence

that the physiological mechanism associated with male-

induced abortions is identical to the mechanism that is

associated with male-induced estrus.

Keverne and Rosser (1986) sharply criticized

adaptive explanations for male-induced abortions in

house mice and proposed that male-induced abortions are

a cost of male-induced estrus. They also pointed out

the advantages to both males and females of male-

induced estrus. Stehn and Richmond (1975) stated their

opinion that male-induced abortions in prairie voles

are "an unavoidable side effect of the endocrine events

associated with estrus induction" (p. 1212). Thus

these authors believe that male-induced abortions are

simply estrus induction that occurs during pregnancy,

resulting in abortion as a side effect. Bronson and

Coquelin (1980) maintained that in house mice, contact

between pregnant females and unfamiliar males

sufficient to induce abortions happens so rarely in the

wild that this response has not been selected against.

In this dissertation I will describe the results

of a series of experiments that were performed to test

the specific hypothesis that the two phenomena, male-

induced estrus and male-induced abortions, are actually

the result of a single mechanism and are essentially

the same response to the same stimulus. One way to

test whether two phenomena are caused by the same

mechanism is to manipulate a single variable in both

situations and compare the responses to that variable

(Dewsbury, 1985). If the responses are different, then

there must be some difference between the controlling

mechanisms. I compared male-induced abortions and

male-induced estrus in female prairie voles, Microtus

ochrogaster, by exposing pregnant or anestrous females

to heterospecific males. If the mechanisms controlling

male-induced abortions and male-induced estrus are the

same, then heterospecific males should either both

induce estrus and cause abortions in female prairie

voles, or they should do neither. If they do neither,

it could be simply because female prairie voles are not

sensitive to the pheromones of heterospecific males.

In order to increase the chances that female

prairie voles would be sensitive to heterospecific male

pheromones, I used congeneric males of two species:

montane voles (Microtus montanus) and meadow voles

(Microtus pennsylvanicus). If female prairie voles are

sensitive to pheromones from these two closely related

species, there are two possibilities for their

responses. First, if anestrous females come into

estrus and pregnant females abort when exposed to

heterospecific males, this would be evidence for a

common mechanism. Second, if females do one but not

the other, this would be evidence that the mechanisms

are different.

Prairie voles are excellent animals to use to

demonstrate differences between male-induced estrus and

abortions using the method described above, because

they usually remain in an anestrous (noncycling, often

referred to as diestrous) condition when housed

individually, but often come into estrus within 2-6

days of being placed in contact with a male (Richmond &

Conaway 1969; see further discussion in chapter 2).

Therefore estrus induction can be measured relatively

easily in terms of presence versus absence, rather than

by more subtle differences in the timing of estrus.

Prairie voles also have a high incidence of male-

induced abortions, even late in pregnancy (Richmond &

Stehn, 1975). Also, as shown by previous experience in

this laboratory, both estrus and abortions occur with a

high probability in female prairie voles separated from

males by wire mesh barriers. The use of this technique

avoids some of the procedural problems, discussed

below, that have been encountered in other studies of

male-induced abortions.

In the second chapter of this dissertation,

evidence for the occurrence of male-induced abortions

and male-induced estrus in a number of rodent species

under various conditions will be reviewed (see summary

in Table 1). This background information is essential

to understanding the arguments regarding the identity

of the two mechanisms presented in the discussion

(chapter 4). I will begin the review with the

literature on house mice, because male-induced

abortions were first discovered in house mice and much

research has been done using this species. I will then

review the literature covering male-induced abortions

in other species and describe some of the differences

between these species and house mice. Lastly, I will

review studies of male-induced estrus in house mice and

other species and compare and contrast these two

phenomena, with emphasis on the nature of the

pheromones that are associated with their occurrence

and the hormonal mechanisms involved.

In the third chapter of this dissertation,

experimental work which was designed to dissociate, if

possible, the mechanism involved in male-induced

abortions from that involved in male-induced estrus


will be presented. In the fourth chapter I will relate

the findings of these experiments to other evidence

that the two processes are separable, and review the

hypotheses that have been proposed regarding the

adaptive significance of male-induced abortions.



0 CO

4-) 4-
U) U

o 0 N
0 C aC


t U
01 a)

- c(

4P 0
U) a

-1 r-l


4-) 4-)
a) a-
M co co
(U co N- W

4J a -) 4-
U)~ U) U) U

* 0 -*
o 0 0 0' N O Pe
O~r~ a~glcO






























H I I o

>4 -H

4 a
U) 0d 0 ( H
o -1 rl (

H- H
co cG
a\I o

4Ji 4-
4J 4J
a) a)

(U (U

b b
-) h)

41i 4.)
to U)
C C:

m3 X

%D co L
m H

4J 4JV 4J 4J 4- 4J 4J J 4J 4J 4J
0 u u 0u 0 0u 0 0u a u 0
4.) 4J V- 4- 4. 4.- 4 .) V-I 4 4-I -H' V 4- .
0 0 0 0 0 0 0 0 0 0 0 0 0 0

0 0

r-i 4i
PH (U 0
H r4 U .
3 (U (U U
c: a rH

CO i-l O
00 r o W

4P 0 0(
41 r-H H 4-4 9
0) CO CO 0) $4
9 to1 ro 0 r0
co H H H
40 -1 QH )

U to to u u
A (0 0rz

U) U U) C)


0 0) 0 C)(
i-q r- 4 r-i A4 p

r.4 r 0 .
H H H H .-I ,!

-H *- *r4 -4 0 0
g g 2 fi aC

0 C0 N
m qw cc 0 c 0 co CO qw a% N ( n
.0 '0 II I O CDO r- I H 0 CO O0
In rN r-H
Cj co rN

4) 4 4J 4 a 04 J 4 0 J 0 0 CO
0 0 ro0 0 0 *i 0 0 A
4- 4-3 434- Vrd 4J 4-3 4J 4 V 4.- 4. )
C C C C -4 C C )U ) C C r-
0 0 0 0 0-I 0 0 0 0 r- 0 0 10
Ou u u -O q u o A -4 o O
0 0 0
0 t 44-
(/) q4o~a ~












U) 0
0 -O

4 9



N C0


4.3 4-3
W Ua











( H H *
Hl o M 4
co 0h 0 0 0)
ON (1 rql ri -ri H_ C
HI H H C H -
ON cn n 03
SCO CO 4 4) 0
-P A 0 C (G 0 ) 4. to 1 q0U
4 0 c il rH a 4a ) (0 .
0 0 co u i ) 4) 0 0 0w
g H $4 ON H H H H
(U H H (U ( (H t
4) 0 co C a 0 ( I at a 4a -H
U4 ) e O O AUCO

03 ,a H H $ H I C4 4C )
44 H 0 1 0 0 0 0
0) 4J H it 0 > 4 > >1 m4 m $ 4:
4 w U z % 4. 4.
*A 4J 0

0 OH
D o0 $4 U 04

to -r- 0
o0 0

-H 0 M -H
4 40 0m
0 0 (t>,
4. 4J 4. 4)
t .r i k 9$ O O 44 V
0 0 0 0 0 0 0 0 0 0 0 0 y 9
t o 0 (d -H (a -HI t o -iH t .H q itU
S4 4 V 4 $ 0) 04) 0 $ UQ)
V C 9 9 p p f V4 9 k r P4)r
S O H O 4 4V 0d
0 0 0 4U 0 m (U rI 0 ( 0 ( *u
0 En 0 0 0 p r-0 0 .Q 0) to
r. (00 0 -H q 0) 3
4 w 0 4 0 (0 ( 0 44) -

0 p V 9 V 00 0 0
H > 0 4-) H .(

0 r r. ) 0 (U + -HOo
C H 4) -H H0 v.
035 H W H p A pi. I
V ( $ H 0 0 (o
0 0 w 0 H p 4 .J
4 $ (U 13 ( U$ (0

d 0 ri 0 o ) -H 0 U 0 .
0 0 0i V (r h-HH0

3 4 O $3 04 4
01 H $4 O 0 0 0 ( 0 0 ( *
H0 0 4 $ -H 0 U) C V 0 03
0 0 0 $4 0 M 0 $
(U ( H -H 0 0 M0 0U
S 0 Q g U



Male-induced Abortions in House Mice

Discovery of Male-induced Abortions

Bruce (1959) described a phenomenon she called

"pregnancy block" that was caused by exposure of mated

female house mice, Mus musculus, of an inbred

laboratory strain to unfamiliar males of another strain

24 h after the original mating. More details plus

additional data from the original experiments were

presented in a follow-up paper (Bruce 1960a). In most

of the studies involving house mice, unfamiliar males

of the same strain as the female and the stud male are

called "strange males", while unfamiliar males of a

different strain are called "alien males." The male

that does the original mating is usually called the

"stud male." In studies involving species other than

house mice, unfamiliar males are often called "strange


The phenomenon was further explored in subsequent

experiments by Bruce and her colleagues (e.g. Bruce,

1963, 1965; Parkes & Bruce, 1962). In these

experiments, female mice were housed with males and

examined each morning for evidence of copulation

(presence of vaginal plugs). When a plug was found,

the female was removed from the stud male's cage. The

mated females were housed overnight, alone or with one

or two other females, and the next day (24-36 h after

mating) each mated female was exposed to one of a

number of test conditions for the next 7-10 days.

Conditions included exposure to (a) an unfamiliar male

of the same strain as the female, (b) an unfamiliar

male of a different inbred strain, (c) a testosterone-

injected male, (d) a castrated male, (e) an adult

female, (f) an ovariectomized female or (g) the

original stud male. Vaginal smears were taken every

day to check for a return to estrus (judged by the

appearance of large numbers of cornified cells in the

smear). A return to estrus was considered to be

evidence of termination of a pregnant or pseudopregnant


The major finding of these experiments was that a

significantly higher proportion of females exposed to

unfamiliar males returned to estrus within seven days,

than did females exposed to other females or to the

stud males. Bruce did not report the results of

statistical tests for many of the data presented in her

1959, 1960a, and 1965 papers. However, if the data

presented in the 1960a paper are analyzed using a chi-

square test, there is a highly significant difference

in the responses of those females re-exposed to the

stud male versus those exposed to an unfamiliar male of

a different strain (N=113, 1= 47.66, p<.0001).

Exposure to castrated males of the same strain as

the female was also associated with a significantly

higher proportion of females aborting (26%) than

exposure to stud males (all pregnant or pseudopregnant)

according to Bruce's (1960a) paper (data analyzed by

present author; N=96, X=13.83, R<.0002). Alien males

induced abortions at a higher rate than intact strange

males (i.e. unfamiliar males of the same strain as the

female) (63% versus 28%), showing the difference

between strange and alien males. No castrated alien

males were used in this study, however, there was no

difference in the percentage of females returning to

estrus after exposure to intact strange males and

castrated strange males (28 versus 26%) (N=119, X=.03,

p>.85). However, in a later paper Bruce (1965)

provided evidence that a higher percentage of mated

females returned to estrus when exposed to unfamiliar

intact males than when exposed to males castrated at 8

weeks of age and tested at least 2 months later (N=253,

"=45.96, p<.0001). The original finding was

attributed to unknown factors. However, it appears

that when the unfamiliar male is the same strain as the

female, the same relatively low rate of abortions is

associated with either intact or castrated males. This

may mean that these abortions are associated with a

difference process, such as disturbance, than the

process that produces the abortions associated with

alien males.

Bruce (1960a) reported that proximity, but not

direct contact between the mated female and the

unfamiliar male, was required to produce the effect;

however her conditions did not exclude contact with

male urine (see below). In addition, it was shown that

increasing the number of males from 1 through 16 did

not significantly increase the incidence of abortions.

A number of follow-up experiments were done by

several different investigators, and these were of

three basic types: (a) those aimed at delineating the

circumstances under which male-induced abortions occur

in laboratory strains of house mice, (b) those aimed at

discovering the mechanism, and (c) those directed

toward determining the generality of the phenomenon. I

will review each of these in turn.

Circumstances of Male-induced Abortions in House Mice

Strain and parity effects

Bruce (1960a) discovered in her first experiments

that when the unfamiliar male was of a different inbred

strain than the stud male, pregnancy was more likely to

fail than if the unfamiliar male was of the same strain

as the stud male. Bruce (1963) found that in order to

demonstrate maximal blocking in inbred strains, it was

very important that the stud male and alien male be of

different strains, but the female's strain was not

important. Bruce attributed this to the fact that the

female could more easily discriminate between the two

males if they were of different strains (see chapter 4

for a different interpretation).

There is great variability among mouse strains

both in the ability of the males to block pregnancies,

and in the tendency of females to abort their

pregnancies when exposed to strange or alien males

(Chapman & Whitten, 1969; reviewed in Marchlewska-Koj,

1983). Females of only a few strains, including the

Parkes strain used by Bruce, and also the Balb/c, CBA,

and SJL/J strains, are susceptible (Whitten, 1973).

These differences appear to be due to genetic

differences among strains; in males the differences may

be related to differences in the ability to produce the

chemical involved, which is androgen-dependent (see

below). Injection of androgen into male mice of a

strain that normally does not induce abortions enhances

their ability to do so (Hoppe, 1975).

Chipman and Fox (1966a) found that young females

(2 months) were more susceptible to male-induced

abortions than older females (6 months), but there was

no difference between nulliparous and parous females of

the same age.

Timing effects

In house mice, the ability of an alien male to

induce abortions depends on the timing of the exposure

of the female to the male or his odors. Based on a

study in which strange males were introduced to mated

females at various times after mating, Bruce (1961)

claimed that females were most vulnerable within 48 h

of mating. However, although chi-square tests of

significance were done, the values were not given.

Furthermore, the numbers of females tested or the

numbers returning to estrus were not given for all of

the days that males were introduced; only a graph of

percentages of females returning to estrus was given.

Some numbers of females in various categories were

given in the text; comparing the proportion of females

that returned to estrus when the male was introduced on

day 1, 24 h post-mating (81%) with the proportion when

the male is introduced on day 5 (39%), using a chi-

square test, shows a significant difference (N=124, )

=20.70, E<.0001). Therefore, I must conclude, based on

the numbers that were given and on examination of the

graph, that there was no significant decrease in

abortions until day 5. When males were introduced

after day 5, very few females returned to estrus,

however the exact numbers were not given. Bruce (1961)

claimed that females had to be exposed to the alien

males for at least 12 h, and that durations of 48 and

72 h gave the maximum effect and were not significantly

different from each other.

It has been shown (Chipman, Holt, & Fox, 1966),

that exposure to the alien male does not have to be

continuous. Females that were exposed to alien males

for three 15-min periods per day for 4 days had

pregnancy rates (determined by autopsy on day 7 after

mating) that were not significantly different from the

rates for females that were housed continuously with

alien males for 4 days; both were significantly

different from unexposed controls.

Strength of stimulus effects

Experiments designed to test the susceptibity of

females exposed to different numbers of males have had

mixed results. Bruce (1963) found no effect of

increasing the number of males. In another experiment,

Chipman and Fox (1966a) found a lower pregnancy rate

among females exposed to six alien males than to one or

three alien males. The pregnancy rate for females that

were not exposed was significantly higher than that in

any of the exposure conditions. The major difference

between these two studies was the technique used to

expose mated females to unfamiliar males. In the Bruce

(1963) study, females were placed with the males in a

20 x 15 x 5 inch box. In the Chipman and Fox (1966a)

study unfamiliar males were placed above the females in

runways with wire mesh bottoms so that females were

exposed to male urine, but there was limited bodily

contact with the males. As will be described below, in

some studies full contact with males appears to be more

effective in inducing abortions than contact with urine

only or contact through screens. Thus there may be a

greater possibility for increasing rates of abortion by

making the stimulus stronger in non-contact experiments

than in contact experiments.

Effect of the presence of the stud male

When recently mated females were individually

confined in a cage and placed in a larger cage

containing a group of alien males, a high rate of

abortions was obtained. However, Parkes and Bruce

(1961) reported that if the stud male was confined with

the female in the same situation, the rate of pregnancy

blocking was much less. These results are difficult to

evaluate, however, as no actual data were presented.

In another study, Thomas and Dominic (1987) reported

that full contact with the stud male gave complete

protection against male-induced abortions caused by

exposure to alien male excreta. These results were

obtained when alien males were confined in a corral

placed on top of the mesh-topped cage containing the

pair. Also, if both males were confined in separate

corrals on top of the female's cage there was a

significantly lower abortion rate than when the female

was exposed to alien male excreta only.

Effect of the presence of other females

Bruce (1963) also reported that if recently mated

females were placed in a group and the whole group was

exposed to alien males, the proportion of females that

returned to estrus (42 of 146 for groups of eight

females) was much less than when females were exposed

individually (119 of 209). However, the proportion

that returned to estrus when grouped was still greater

than that of females not exposed to alien males (9 of

129). Thus grouping females appears to make it

somewhat less likely that they will abort when exposed

to alien males.

Effect of familiarity and copulation

How the female mouse recognizes the stud male so

that she does not react to his urine has been

investigated in house mice and in other species (see

below). Experiments in house mice have led different

investigators to two opposing conclusions.

In the first set of experiments, it was shown that

females that had very short (3 h) exposures to stud

males were less likely to abort when exposed to alien

males than females with at least 24 h of exposure (Lott

& Hopwood, 1972). Short-exposure females also did not

abort when re-exposed to the stud male. Females with

short exposures to stud males were obtained by using

only estrous females that mated within a short time

after stud male introduction. In previous studies

(e.g. all of those done by Bruce and her colleagues)

males were placed with cycling females, which means

that there was a longer time on average (usually longer

than 48 h) between the introduction of the male and his

removal after copulation had been detected. Lott and

Hopwood concluded that when no male is recognized as

the stud male, because females require longer exposure

to the stud male for memory formation to occur, then

male-induced abortions are less likely. They suggested

that exposure to the stud male sensitized females, so

that they aborted when exposed to the odors of all

other males. The term "sensitization" is used to

describe a situation in which an animal increases the

amplitude or frequency of a response as a result of

repeated exposure to a stimulus. The use of the term

in the current context implies that a stimulus

repeatedly given by the stud male increases the

probability that a female will remain pregnant.

However, in house mice removal of the stud male after

copulation does not result in a decreased probability

of pregnancy (see chapter 2).

Different results were obtained by other

investigators. Bloch (1974) tested the reaction of

females to stud males which had mated with the female

in postpartum estrus and then had been removed the next

morning and reintroduced two days later. In this

situation the females were exposed to the stud male for

less than 24 h because they were already in estrus when

the males were introduced ("unfamiliar" stud male).

When stud males were unfamiliar, they induced abortions

of pregnancies that they had caused in 51% of the

females tested, compared to 27% of control females that

failed to deliver litters.

Rosser and Keverne (1985) performed a similar

experiment, except that females in cycling estrus were

exposed to a stud male only long enough for mating to

occur (about 20 min). Each female was then placed into

a soiled cage previously occupied by a male belonging

to another strain. After 48 h the female was re-

exposed to the stud male. A high proportion of male-

induced abortions was obtained (75-80% returned to

estrus, depending on strain); in this experiment, as in

the one by Bloch (1974), males were inducing abortions

of pregnancies that they had caused.

Rosser and Keverne (1985) also tested the duration

of exposure to the stud male that would be necessary if

the female were to maintain pregnancy when re-exposed

to the stud male after 48 h. If females were separated

from the stud male after only 30 min, no females

remained pregnant when re-exposed to him later. If

females were separated after 4.5 h, most remained

pregnant, and all remained pregnant if the stud males

were present for 6 h.

Bloch (1974) also found that if a male was allowed

to mate with a female and to remain with her through

gestation, but was removed just before parturition and

another male was allowed to mate with the female, the

original stud did not block the pregnancy that was

caused by the other male.

The experiments by Bloch (1974) and by Rosser and

Keverne (1985) show that a female must spend at least a

few hours with a male, as well as copulating with him,

before she recognizes him as the stud male, and that

this recognition persists even if the female

subsequently mates with another male. The persistence

of the memory for the stud male was confirmed by Kaba,

Rosser, and Keverne (1988); they also found that an

intervening pregnancy will cause the memory to fade

more quickly (but it does not fade within the time

period used by Bloch).

Also related to the above findings is the report

of Parkes and Bruce (1961) that exposing the female to

an alien male before her mating with the stud male

reduced the incidence of male-induced abortions when

the same alien male is used as the stimulus male.

Familiarity, then, tends to lower the probability that

a female will react to a specific male, and this

applies to both stud and alien males. Copulation is a

powerful means of creating familiarity, but it is not

the only factor.

Rosser and Keverne (1985) injected 6-hydroxy-

dopamine (6-OHDA) into the accessory olfactory bulbs of

female mice; this procedure temporarily depletes the

neurotransmitter noradrenaline. According to Keverne

and de la Riva (1982), 6-OHDA interferes with the

formation of the "memory" of the stud male, making

treated females unable to discriminate him from other

males. Females injected with 6-OHDA and vehicle-

injected females were mated with stud males; the males

were removed when a plug was found and reintroduced 24

h later. Eight 6-OHDA injected females remained alone

to control for the effects of the procedure alone; all

of these remained pregnant. None of the 6-OHDA-

injected females had implantation sites on autopsy,

while all of the vehicle-injected females re-exposed to

stud males were pregnant. Rosser and Keverne concluded

that females abort when exposed to any male that is not

recognized by the female, and that the memory of a male

is normally formed during copulation.

As will be described below, experiments using

field voles showed that when there is no stud male,

pseudopregnancies are blocked. The preponderance of

the available evidence seems to support the hypothesis

that a female habituates to the odor of the stud male

during exposure to him, and this removes the potential

for reaction to his odor. There is another alternative

explanation: Bronson and Coquelin (1980) suggested that

it is the male that habituates to the female's odor.

The result of this habituation is that the male stops

producing the pheromone responsible for abortions and

estrus induction. As will be shown below, this

explanation is unlikely to be correct for some species

of voles that have male-induced estrus.

Mechanisms of Male-induced Abortions in House Mice

The stimulus

Channel. Since the estrous cycles of mice had

already been shown to be sensitive to stimulation from

conspecifics via the olfactory channel (e.g. Van der

Lee & Boot, 1956b; Whitten, 1956), it was suspected

that male-induced abortions might also be triggered by

stimuli acting on the olfactory system. Bruce and

Parrott (1960) tested this hypothesis by removing the

olfactory bulbs from females before mating them and

then exposed them to alien males. Most mated

bulbectomized females exposed to alien males remained

pregnant, whereas most intact females aborted.

Source of the stimulus. Bruce and her colleagues

then turned to the question of the nature of the

olfactory stimulus that triggers abortions. In earlier

experiments (Bruce, 1960a), when females were placed in

small cages and then these cages were placed inside

larger cages containing males, the females had high

rates of return to estrus. She concluded that direct

contact with the strange male was not necessary.

Therefore, it was hypothesized that some volatile

excretion of the males was responsible for male-induced

abortions. To test this hypothesis Bruce (1960b)

investigated the ability of male-soiled bedding to

elicit the reaction. She found that as long as the

soiled bedding was replaced twice each day (with new

soiled bedding), and the animals were tested in glass

jars with reduced ventilation, the proportion of

females that returned to estrus after exposure to the

bedding was similar to that when the females were

exposed to the males themselves.

In another experiment (Parkes & Bruce, 1962) it

was shown that the important factor was frequent

replacement of the soiled bedding, not reduced

ventilation. Twice a day for three days, they placed

mated females into a different standard mouse cage,

immediately after five males had been removed. These

females had a higher rate of return to estrus compared

with females placed into fresh cages on the same

schedule. Changing cages only once a day for three

days was not nearly as effective. Furthermore, they

tested several exposure durations, from 12 h to 3 days,

and found that time course of effectiveness of the

stimulus was the same as in the experiments described

above (Bruce, 1960a) in which females were exposed

directly to males. It was concluded that the stimulus

is present in male excreta, and that it dissipates

fairly rapidly (within hours).

The above experiments involving exposure to soiled

bedding demonstrated that the stimulus is something

left behind by males that affects the olfactory system.

The following experiments were designed to determine if

the substance is volatile and whether it is found in

urine, feces, or both.

Nature of the stimulus. Dominic (1966a) exposed

females to male excreta without the possibility of

direct body contact between males and females; he found

that exposure to male urine alone was associated with a

similar high rate of abortions compared with females

exposed to both urine and feces. However, there was no

direct comparison of females exposed to feces alone

versus females exposed to urine alone.

In another experiment Dominic (1966a) showed that

females would abort if a small drop of fresh male urine

was placed on their noses three times a day for three

days. A lower, but still significant, proportion of

females aborted when 3 ml of male urine was placed on

their bedding twice a day. Urine used in this

experiment was preserved and stored, rather than fresh,

which may account for the smaller proportion of females

aborting. Urine that is stored without a preservative

does not induce abortions (Parkes & Bruce, 1961).

Bruce (1960a) assumed that the substance was

volatile since females separated from males by wire

mesh cage walls still aborted. However, Rajendren and

Dominic (1984) showed that direct contact between a

female and male or his urine was necessary to produce

abortions. They placed the alien male and mated female

in separate corrals, 4-5 cm apart, within a larger box,

and compared the incidence of abortions in this group

(only 8 of 49 returned to estrus) versus a group in

which only the male was corralled (35 of 45 returned to

estrus). In a control group in which the female was

corralled 4-5 cm from an empty corral, 5 of 47 returned

to estrus. They concluded that in previous experiments

using single wire mesh barriers (e.g. Bruce, 1960a)

females could contact confined males or their urine

sufficiently to produce abortions.

Further evidence that the substance is non-

volatile was obtained by Marchlewska-Koj (1981), who

confirmed that urine that had been evaporated under

vacuum and then diluted with water could induce

abortions when applied to the nares of mated females.

Any volatile substance should have been eliminated by

this procedure.

Effect of androgens. Production by males of the

substance that elicits the abortion response in females

is apparently androgen dependent, since it is not

produced by castrated males and it is produced by

androgenized females (Bruce, 1965; Dominic, 1965).

Further evidence was provided by Bloch and Wyss (1973),

who injected alien males with an anti-androgen and

showed that these males did not block pregnancies.

Hoppe (1975) also injected testosterone into inbred

males of a strain that normally does not elicit

abortions; females exposed to treated males aborted

more frequently than females exposed to uninjected

males of this strain. Therefore strain differences in

the ability of male mice to trigger the abortion

response may be related to differences in the levels of

androgens (or their metabolites) in the urine.

Effect of social dominance. Labov (1981) showed

that dominant males did not have a greater tendency to

induce abortions than subordinate males; the percentage

of females not pregnant was essentially equal in both

conditions. However, this result was directly

contradicted by the results of Huck (1982) who found

that a higher percentage of females exposed to strange

dominant males aborted than females exposed to strange

subordinate males. In both studies the females and the

strange males were of the Swiss albino strain. Huck

used slightly older females, but females in both groups

were nulliparous and at least 45 days old. Subordinate

male mice have been shown to have depressed

testosterone levels in some studies (e.g. McKinney &

Desjardins, 1973), but neither Huck nor Labov measured

testosterone levels.

The reason for the difference between these two

studies is not clear. The percentage of females

aborting after exposure to dominant males in the Huck

study was the same as for dominant males in the Labov

study, but the subordinate males in the Huck study were

much less effective in eliciting the abortion response.

One difference between these two studies is that Labov

continued daily male-male encounters throughout the

experiments, while Huck apparently tested the males for

dominance first and then did the experiments. It is

difficult to see how this difference in procedures

could have produced the observed differences in

results. Continued encounters would be expected to

produce lower testosterone levels in subordinate males

and reduce their ability to induce abortions. Instead,

subordinate males continuously exposed to dominant

males in the Labov experiments, in which no effect of

dominance was found, appear to have been more likely to

induce abortions than would be expected based on this


Chemical analyses. The substance excreted by male

mice that induces abortions was found to be a non-

volatile substance that is found in the peptide

fraction of urine (Marchlewska-Koj, 1981). The active

fraction was isolated from both bladder urine and

excreted urine using G-75 Sephadex chromatography and

thin-layer electrophoresis. These results indicate

that the substance is not made in the preputial glands;

further evidence of this is that there are strains of

house mice that lack preputial glands, yet can induce

abortions. Deer mice also lack preputial glands but

have both male-induced estrus and male-induced

abortions (Bronson & Marsden, 1984). Hoppe (1975)

suggested that the substance may be made in the


The sensory receptors

Bellringer, Pratt, and Keverne (1980) identified

the location of the receptors of the abortion-inducing

pheromone as the vomeronasal organ (VNO) by destroying

the VNO of females and then mating them and exposing

them to alien males. Females with VNO lesions remained

pregnant when exposed to alien males, whereas sham-

operated females showed high rates of abortion. The

VNO sends projections to the accessory olfactory bulb

(Barber & Raisman, 1974), and from there projections go

to the amygdaloid nucleus in rabbits and rats (Scalia &

Winans, 1975).

Lloyd-Thomas and Keverne (1982) compared the

effects of VNO lesions with the effects of main

olfactory receptor lesions (peripheral anosmia) that

were caused by zinc sulfate infusions into the nasal

cavities. All females were exposed to alien male

bedding, and only those females with VNO lesions had

pregnancy rates similar to those of the control

(nonoperated and sham-operated) females. They

concluded that the abortion-inducing pheromone acts

only through the VNO.

The hormonal mechanism

It was suspected that the immediate cause of male-

induced abortions in females exposed to alien males is

a failure of the corpora lutea to become fully

functional and secrete sufficient progesterone to

support pregnancy (Bruce & Parkes, 1960). This in turn

was thought to be due to failure of the pituitary

luteotrophic activity that is normally triggered by


In many rodents (guinea pigs are an exception),

the corpus luteum is short-lived and does not become

fully functional unless a luteotrophic signal is sent

by the pituitary, and this signal is sent only if the

female has received sufficient cervical stimulation

from copulation (induced luteal phase; Van Tienhoven,

1983). It was shown by Milligan, Charlton, and Versi

(1979) in field voles, Microtus aqrestis, that the

stimulation given by the male at the time of mating is

sufficient to maintain some luteal function for about

10 days, even if the original set of corpora lutea is

destroyed and a new set is created by LHRH injection.

Effects of LH. The identity of the luteotrophic

signal has been explored in several experiments. It

was first thought that LH might be the signal, since

injections of an antiserum to LH was associated with

abortions (Munshi, Purandare, & Rao, 1972). Further

experiments showed that administration of the LH

antiserum on days 2-4 after mating did not block

implantation, but that females sacrificed on days 8 and

10 after mating showed evidence of abortions of

implanted embryos (Munshi & Nilsson, 1973). Therefore,

LH was shown to be necessary to maintain pregnancy, but

not at the time that male-induced abortions occur in

house mice.

Effects of prolactin. Bruce and Parkes (1960),

suspected that the pheromone from alien males

interfered with the secretion of prolactin, which was

known to be luteotropic. Therefore they tested the

effect of prolactin by injecting newly mated females

with prolactin before exposing them to alien males.

This procedure prevented male-induced abortions.

Mednick, Barkley, and Geschwind (1980) found that in

mated house mice corpus luteum function depends on

prolactin secretion until about 5 days after mating,

and between then and about midpregnancy the corpus

luteum depends on LH. Hypophysectomy at midpregnancy

has no effect on pregnancy maintainance, and this is

now known to be due to the presence of a placental

luteotropic hormone (Cerruti & Lyons, 1960).

The involvement of prolactin in the mechanism of

male-induced pregnancy disruptions was further

confirmed by testing the responses of females mated in

postpartum estrus. Females that mated while lactating

were less likely to have pregnancies blocked after

exposure to alien males, and the proportion of females

that aborted was related to the number of pups being

nursed and thus to the amount of prolactin present

(Bruce & Parkes, 1961).

Dominic (1966b) also injected newly mated females

with prolactin and found that prolactin-injected

females exposed to alien males had the same low rate of

return to estrus as undisturbed females. He also got

similar results in females with ectopic homografts of

anterior pituitary tissue, which are known to produce

prolactin. Dominic (1966b) confirmed that females

exposed to alien males had degenerating corpora lutea

by examining the ovaries histologically.

Injections of reserpine produced the same effect

as injections of prolactin (Dominic, 1966b). Reserpine

is a dopamine antagonist, and dopamine inhibits the

release of prolactin. Prolactin secretion is necessary

for early pregnancy maintainance because it stimulates

the corpus luteum to secrete progesterone, so Dominic

(1966b) also injected newly mated females with

progesterone and again the females were immune to the

effects of urine from alien males.

Effects of gonadotrophin. Injections of

gonadotrophin sufficient to cause ovulation in mice

have been shown to lead to a failure of implantation,

and this is believed to be due either to the direct

effect of the gonadotrophin, or to release of LH, or to

secondary effects on estrogen levels (Hoppe & Whitten,

1972). In rats, ovulatory doses of gonadotrophin given

before implantation are luteolytic.

In a different study, when gonadotrophin (PMSG)

injections were given to female house mice prior to

mating and then these females were exposed to male odor

alone, all but one female failed to implant before day

9, but only 6 of 15 females returned to estrus

(Bellringer et al., 1980). What is more remarkable is

that diapausing embryos (embryos that had not implanted

but were still viable) were recovered from most of the

females on day 9 after mating. These females included

both those that had returned to estrus, and those that

had not. In most previous experiments, pregnancy block

was judged to have occurred if females returned to

estrus within 7 days; therefore, if there had been

females with diapausing embryos in these studies, they

would have been counted as remaining pregnant.

The relevance of this finding to the

interpretation of previous studies is difficult to

assess. In none of the other studies was gonadotrophin

injected before mating; Bellringer et al. used this

technique to induce superovulation, and it is possible

that the occurrence of diapausing embryos is restricted

to this situation. Furthermore, as will be shown

below, PMSG has some unusual effects with regard to


Bellringer et al. (1980) showed that the

diapausing embryos would implant if the females

received progesterone injections. They claimed that

the occurence of diapausing embryos may be another

effect of alien male pheromones. Furthermore, they

stated that if some females in other studies had

diapausing embryos, and thus did not return to estrus

within the usual 7-day time limit (and were scored as

remaining pregnant), then the incidence of male-induced

abortions may have been underestimated, because females

would have been counted as pregnant even though their

embryos were eventually lost.

Wild House Mice

Some of the first attempts to establish the

generality of the findings in laboratory mice involved

testing animals recently derived from wild-trapped

house mice for the occurrence of male-induced abortions

(Chipman and Fox, 1966b). A significant difference was

found between females exposed to strange males and

those not so exposed (16% versus 76% remained

pregnant). In addition, it was found that other kinds

of disturbance, including handling and cage changes,

also resulted in reduced pregnancy rates. This is

contrary to the findings of Bruce (1960a) with

laboratory mice, which were not affected by handling.

In fact, in a later experiment (Bruce, Land, &

Falconer, 1968), it was found that a group of females

of an outbred strain that had been handled more than

usual just prior to being tested for the first time

were less likely to abort when exposed to alien males

than other laboratory mice, but after becoming

accustomed to handling they showed the same high rate

of male-induced abortions as other strains. Thus the

effect of disturbance has not been consistent over

different studies, but may be a factor in experiments

involving species that have not been bred in captivity

for generations if care is not taken to see that the

exposed and unexposed groups are subjected to the same

amounts of disturbance.

Summary of House Mouse Studies

Evidence based on studies using house mice shows

that male-induced abortions occur in early pregnancy

when females are exposed to a peptide or peptide-linked

component of male urine. This substance interferes

with prolactin secretion by pregnant females, resulting

in the failure of the progesterone secretion that is

necessary to maintain pregnancy. Females apparently

habituate to the urine of the stud male, and do not

react to it. As will be shown below, some of the

conclusions based on studies using house mice have been

modified by results obtained from other species.

Male-induced Abortions in Microtine Rodents

When male-induced abortions were investigated in

rodents other than house mice, some new complications

arose. In house mice, abortions seem to be due

entirely to failure of embryos to implant at the proper

time. Embryos that do not implant are eventually

expelled. As will be shown below, in some other

species abortions occur after implantation when the

female is exposed to a strange male. I have used the

term "abortion" to describe both of these situations,

although different processes may be involved, one in

which the process of implantation is interfered with,

and another in which an already implanted embryo is

resorbed or expelled. There is at least some evidence,

described below, that a stronger stimulus, or a

combination of stimuli, is necessary to trigger post-

implantation abortions.

A second complication, which will be described

below, is that in some species, simple removal of the

stud male (in order to replace him with a strange male)

has an effect on pregnancy maintainance.

Researchers studying male-induced abortions in

microtine rodents frequently employ procedures that are

different than those described for house mice. Many

microtines have an induced behavioral estrus, as well

as induced ovulations. In the experiments described

below, the time of mating was often not known, since

pairs are simply put together and it was assumed that

they mated within 2-3 days of pairing. Often, no

attempt was made to determine the day of mating, or

even whether mating actually occurred, by looking for

plugs or sperm in the vagina. However, the degree to

which various microtine females require male stimuli to

induce receptivity is highly variable, even within

species (Sawrey & Dewsbury, 1985; Taylor, 1990). This

means that some females may have mated immediately

after pairing, while others did not mate until 2-3 days

after pairing, or did not mate with the original male

at all. Strange males were introduced at various times

after the original pairing, but it is not precisely

known on what day of the pregnancy of each individual

female that this was done.

Another problem with many of the studies described

below is that the strange male is allowed direct access

to the female, by being placed in the same cage with

her. This is the same procedure that is employed in

studies using house mice; however, when this procedure

is used in an experimental design that does not allow

determination of the exact time of mating, paternity of

the litter must be assigned based on time to

parturition. Failure to correctly assign paternity to

litters can result in identifying as non-aborting

females that are pregnant as a result of mating with

the strange male. In previous studies of microtines,

often litters delivered past an assigned time limit

were assumed to have been sired by the strange male,

and therefore that female was scored as having aborted

the first litter, even though she may not have been

pregnant. This procedure could also lead to errors in

identifying aborting females if gestation lengths are

affected by any of the manipulations. Furthermore,

when paternity is based on time to parturition, animals

that do not have litters must be discarded from the

analysis (e.g. Storey, 1986, 1990) which can distort

the results if there are more of these in experimental

than control groups. This might happen if there was a

lower conception rate from matings of females that had

recently undergone abortions, than in the original

matings. In only one experiment (Smale, 1988) were

vasectomized males used as strange males, avoiding the

problem of assigning paternity to litters.

Exposing the female directly to the strange male

may also cause confounding of effects due to

generalized stress (from fighting) and effects due

specifically to the presence of a strange male. This

confounding is due to the fact that abortions may also

be caused by mechanisms that involve the pituitary-

adrenal axis (see discussion in chapter 4).

Prairie Voles (Microtus ochrogaster)

Male-induced abortions in prairie voles were first

reported by Stehn and Richmond (1975), who showed that

abortions can occur very late in pregnancy in this

species when females are caged with strange males. The

rate of pre-implantation abortions was quite high,

often 90% or more (see Table 1). In the Stehn and

Richmond study there was no significant decrease in the

abortion rate for animals exposed at different times

during pregnancy, until days 14-15. These authors

suggested that there seems to be an association between

a greater dependence on male pheromones for estrus

induction and a greater susceptibility to male-induced

abortions late in pregnancy, and that this may be due

to the identity of the two processes.

The occurrence of midgestation abortions due to

the presence of strange males was investigated by Stehn

(1978), and further confirmed by Stehn and Jannett

(1981) using a colony of prairie voles more recently

derived from the wild than those used in the earlier

experiments. Midgestation abortions were also found in

Microtus montanus and M. pinetorum (Stehn & Jannett,

1981). In Microtus ochrogaster, midgestation abortions

occurred in lactating females, although to a lesser

extent than in females that were pregnant for the first

time. In a different study using multiparous females,

Kenny, Evans, and Dewsbury (1977) found midgestation

abortions in 4 of 13 females. This appears to indicate

a lower rate of midgestation abortions in multiparous

females, however, direct comparison is difficult due to

procedural differences between the two studies.

There is some evidence that midgestation

abortions in prairie voles may require that the female

be caged with the unfamiliar male; that pheromones

alone are not sufficient (Stehn, 1978). This may also

be the case in meadow voles (see below). Thus, females

in midpregnancy may require stronger stimulation

including direct physical contact with the unfamiliar

male, before undergoing abortions. Smale (1988)

demonstrated a significantly greater number of

abortions in 7-day pregnant female prairie voles that

had the urine of unfamiliar males placed on the

external nares twice a day on days 7-12 after mating,

than females treated with stud male urine or water.

Midgestation abortions are problematical because they

cannot be caused by exactly the same mechanism as

abortions occurring before day 6-7, since in

midgestation the luteotropic hormone is a placental

lactogen. Exactly what the differences are between

pregnancy in house mice and voles that makes the latter

but not the former subject to midgestation abortions,

has not, to my knowledge, been investigated.

Chipman, Holt, and Fox (1966) showed that

exposure of house mice to strange males did not have to

be continuous. Hofmann, Getz, and Gavish (1987) tested

the effect of exposing female prairie voles to strange

males for three 15 min periods per day for 2 days at

different times during pregnancy. They found that if

the stud male was present except during the short

periods when the females were exposed to strange males,

there was no increase in abortions over control groups.

However, if the stud male was permanently removed and

the female was exposed to brief contacts with strange

males, then there was a significantly higher abortion

rate than in females housed alone after stud male

removal. Hofmann et al. used a different strange male

for each "visit."

The effect of castration on the ability of strange

males to trigger abortions was tested by Smale (1988).

She found that castrated males or their urine were as

effective stimuli for male-induced abortions as intact

males or their urine, which is very different from

previous findings in house mice and field voles (see

below). Smale's experiments were performed on day 7

after pairing, thus these may be post-implantation

abortions. Interpretation of her data is somewhat

complicated by low pregnancy rates in control groups.

Since the original males and females were simply placed

together with no check on when or if they actually

mated, the low pregnancy rates may reflect failure of

the first males to mate.

Information on the possible cost to females of

male-induced abortions was obtained by Stehn (1978),

who found that the next litter of a female that had

undergone an abortion showed a reduction in the number

of nursing young, in spite of equal implantation rates,

when compared with females whose first litter was not


Field Voles (Microtus agrestis)

The first report of male-induced abortions in a

microtine rodent was by Clulow and Clarke (1968), in

field voles. Female voles were placed with males for

two days; evidence for mating was obtained from daily

vaginal smears. After two days of cohabitation with

the stud male, mated females were transferred to a

clean cage with either the stud male or a strange male

for 24 h. The females were then isolated for the

remaining gestation period. Of the 20 females re-

exposed to the stud male, 16 delivered litters, while

only five of the females exposed to strange males

littered. Use of this particular technique could have

led to an underestimation of the rate of male-induced

abortions, because the female may have still been in

estrus when transferred, and mating by the second male

during the 24 h period the female was with him was not

prevented. This means that some of the five litters

delivered in the strange-male condition could have been

sired by the strange male. Even so, the result was

statistically significant.

Milligan (1976a) obtained some results in Microtus

agrestis that differ from those in experiments on other

species. High rates of abortion were obtained only

when there was direct contact between the strange male

and the female. Males placed on the other side of

single or double wire mesh barriers had no effect.

Urine and feces placed in the female's cage, or housing

the female in male-soiled cages had no effect. The

only stimulus not involving direct body contact that

had a significant effect was the placement of a cage

containing three strange males directly above the

female's cage. All of these conditions, except the the

presence of double barriers, are effective in

triggering abortions in female house mice. The reason

that soiled cages had no effect may be that females

were placed in cages soiled by males, but the soiled

bedding (type not specified) was not renewed at all

during the time of exposure. As in the case of mice

(Bruce 1960a, see above), this may result in

insufficient stimulation. The reason for the lack of

abortions in females placed across single barriers from

strange males is uncertain. Milligan suggested that,

in addition to olfactory stimuli, field voles may

require other stimuli from males to induce abortions.

I have not found any studies in which urine was applied

directly to the nares of field voles; therefore, there

is no direct evidence that olfactory stimuli from male

urine are a primary factor in male-induced abortions in

field voles.

Another difference found by Milligan (1979) was

that, unlike the house mice in the Rosser and Keverne

(1985) study, but like the house mice in the Lott and

Hopwood (1972) study, very short (1 hour) exposures to

the stud male did not result in abortions in females

when they were re-exposed to the same stud male (16 of

19 remained pregnant). However, both females exposed

to stud males for very short times and females exposed

to stud males for 2 days aborted when exposed to

strange males. Milligan concluded that the female

field vole's memory of the stud male is formed very

quickly, and that once this memory is formed, the

female reacts to the pheromones of all males except the

stud. This is the same conclusion reached by Rosser

and Keverne (1985) for house mice, except for the time

course of memory formation.

If females are made pseudopregnant by LHRH

injection plus vaginal stimulation (therefore no stud

male), there is a higher rate of return to estrus for

females exposed to strange males than for females left

alone. Thus pseudopregnancies, as well as pregnancies,

are terminated by exposure to strange males. This has

also been found in house mice made pseudopregnant by

mating them to vasectomized males (Dominic, 1966a).

Again, the conclusion is that when there is no memory

formed (because there is no stud male), the female

reacts to all males. Females do not react to the stud

male because they become habituated to his odor.

As in previous studies with house mice, castrated

strange males did not induce abortions, but castrated

strange males with testosterone implants did (Milligan,


Milligan (1976b) described changes in the ovaries

and uterus that accompanied male-induced abortions.

Mated females were exposed to strange males between 48

and 72 h after the original mating. Corpus luteum

degeneration began after 72 h and was well advanced at

96 h (48 h after the strange male was introduced).

Uteri examined 96 h after the original mating resembled

non-pregnant uteri.

In experiments similar to those done in house

mice, Charlton, Milligan, and Versi (1978) found that

prolactin injections make male-induced abortions less

likely to occur, and that abortions are less likely to

occur in lactating females. Prolactin injections

prolong the life of the corpus luteum in animals

induced to ovulate by LH-RH injections; without

prolactin, corpora lutea formed in this manner

degenerate rapidly. An effect of injections of the

prolactin inhibitor bromocriptine was also found. If

females were injected with bromocriptine 12 h before

mating, the concentrations of plasma prolactin measured

15 min after mating were significantly reduced. It had

previously been found that there is normally a

prolactin surge immediately after mating (Milligan &

MacKinnon, 1976). In addition, Charleton et al. (1978)

discovered that bromocriptine injections during the

first 5 days of pregnancy were associated with

degenerating corpora lutea 3 days after the injection;

after day 5 bromocriptine was no longer effective in

disrupting corpus luteum function. They concluded

that, as in house mice, prolactin is luteotrophic for

the first 6 days of pregnancy in field voles, but

afterwards a placental luteotropin is responsible for

maintaining progesterone secretion. A placental

lactogen has been identified in field voles and bank

voles by Forsyth and Blake (1976); this hormone has

prolactin-like activity and is presumably luteotrophic.

All of these results in field voles strongly suggest a

similar mechanism for male-induced abortions in voles

as was discovered in house mice.

Meadow Voles (Microtus pennsylvanicus)

Male-induced abortions in meadow voles were first

described by Clulow and Langford (1971), using the same

techniques as were used in field voles (Clulow &

Clarke, 1968). A significant effect was obtained in

females exposed to strange males (12 of 20 females

produced litters in the group exposed only to stud

males, while only 4 of 20 delivered litters in the

group exposed to strange males). This study in meadow

voles is subject to the same limitations as the

previously described experiments in field voles; a

lessening of the magnitude of the effect because of

mating by the second male. Mating by the strange male

with 3 of the 4 females that delivered litters in the

strange male group was observed; these females were

counted as non-aborting. Furthermore, the authors

attribute the high rate of pregnancy failure in the

control group to disturbance caused by experimental

procedures, including the twice-daily vaginal smear


Evidence indicating exactly when male-induced

abortions occur in meadow voles was obtained by Mallory

and Clulow (1977). They exposed nulliparous mated

female meadow voles to strange males on day 5 after

mating, and then killed groups of three each on days 6,

10, 15, and 20. Viable embryos were found in the day 6

group, but not in any of the later groups, indicating

that abortions occurred between 2 and 4 days after

exposure to strange males. Abortions did not occur in

females mated in postpartum estrus and exposed to

strange males on the second day after mating.

When experiments were conducted to determine

whether or not direct contact with the strange male was

necessary, it was found that meadow voles are similar

to house mice and not like field voles, in that

abortions occurred in females exposed to strange males

across a barrier, and in females exposed to bedding

soiled by strange males (Watson, Clulow, & Mariotti,

1983). An unusual finding is that the highest rate of

abortions was found in females separated from the

strange male by a double wire mesh barrier with a

2.5 cm gap between the male's side and the female's

side. This apparatus would seem to exclude female

contact with a non-volatile pheromone in the urine,

which is thought to be the stimulus for male-induced

abortions in other species. A similar high rate of

abortions was obtained when the female was separated

from male-soiled bedding by a the same kind of double

barrier. The authors appear to have assumed that the

abortion-inducing substance is volatile, and they did

not offer any reason why their results differed from

those of other double-barrier studies. A sawdust

substrate and cotton bedding was used in all

conditions. It is possible, but unlikely, that enough

urine-soaked substrate could have been transferred from

male to female compartments to cause the consistently

high rate of abortions observed in the groups in which

double barriers were used. Urine transfer between

compartments beneath the double barriers might have

occurred if males frequently urinated against the

barrier, but there is no evidence that this is the


Evidence that abortions can occur after

implantation in meadow voles was obtained by Storey

(1986). There was no difference between the rate of

male-induced abortions in parous females caged with

strange males on day 4 after pairing with the stud

male, and the rate of abortions in those exposed on day

12. It appears that in both groups there was a

moderate proportion of females that aborted, however,

interpretation is made difficult by the fact that all

the females that failed to become pregnant in the

experiment were discarded from the analysis unless they

passed a post-experiment fertility test, and it is not

clear from which groups these discarded females came.

As noted above, if females that have aborted are more

likely to fail to mate or conceive with the strange

male than control females are with the stud, discarding

nonpregnant females will give misleading results.

Storey (1986) also showed that among all females

exposed on day 4, there was a high rate of abortions in

nulliparous females, but only a moderate amount in

parous females. The rates of abortion were nearly

equal for parous and nulliparous females when exposure

to the strange male was on day 12. Because the ages of

the females were not given, it is not clear whether the

difference in sensitivity of the females to stimuli

from strange males was due to age or parity.

In another experiment (Storey, 1986) in which

males were separated from the females by wire mesh

barriers, it was found that exposing a female

simultaneously to both the stud male and a strange male

resulted in pregnancy rates that were not significantly

different from the rates obtained when the female was

across from a strange male only, or the stud male only.

However, in another experiment, the stud male was

confined with the female across from the new male on

day 12; there was a very high proportion of females

remaining pregnant (85%), whereas if the strange male

was confined with the female and the stud male was on

the other side of the barrier, the pregnancy rate was

low (13%). This was true only when the transfer of

males was done on day 12. Early in pregnancy (day 4),

there were similar rates whether it was the stud male

or the strange male that was across the barrier (60-


In summary, according to this study, in early

pregnancy, females have a high rate of abortion when in

direct contact with the strange male, unless the stud

male is nearby but not in direct contact. In late

pregnancy, however, females have only a moderate rate

of abortion when in direct contact with strange males,

unless the stud male is nearby; if he is nearby the

abortion rate is very high. Storey (1986) suggested

that females abort late in pregnancy when the strange

male is in contact and the stud male is close but not

in contact because there is a high potential for

infanticide when a new male has displaced the old male

but the old male remains nearby. The old male provides

an additional cue to the new male that the latter is

not the sire of the litter, although why this cue

should be important late in pregnancy but not early is


Other Microtines

California voles (Microtus californicus)

Strange male California voles induced abortions in

females in large outdoor enclosures (Heske, 1987).

Abortions occurred even if the stud male was still

present and had killed or injured the strange male.

Red-backed voles (Clethrionomys gapperi)

Clulow, Franchetto, and Langford (1982) reported

that exposure to strange males triggered abortions in

nulliparous female red-backed voles. They also

reported a significant difference in the proportion of

females aborting for two groups of females that were

not exposed to strange males, but one group was exposed

to a cage change. However, there was also a difference

in the timing of separation of the female from the stud

male, which could account for the observed difference

(see discussion in chapter 3). There was no difference

between parous females subjected to a cage change, and

parous females subjected to both a cage change and

presence of a strange male. Parous females subjected

to both cage change and male change were less likely to

abort than were nulliparous females subjected to both

changes. However, using data from this study to

compare the responses of parous and nulliparous females

is risky because all changes were made at 24 h post-

coitum for the nulliparous females, but immediately

post-coitum for the parous females. Because of the

lack of proper controls, data from this study cannot be

used to differentiate responses due to disturbance

(cage changes) and a specific reaction to strange males

or removal of the stud male.

Collared lemmings (Dicrostonyx groenlandicus)

A high rate of infanticide by male collared

lemmings on infants sired by other males was noted in

an experiment by Mallory and Brooks (1978), and they

believed that this should be associated with a high

rate of abortions caused by strange males. This

hypothesis was tested (Mallory & Brooks, 1980) by

exposing mated females to strange males for 24 h on day

4 after mating. A high rate of male-induced abortions

was obtained (87%), in spite of the use of an exposure

time that was shorter than those used in studies of

other species. A high rate of abortions was also

obtained in a group of females that was handled on

three separate occasions during pregnancy, but not

exposed to strange males.

Sagebrush voles (Lagurus curtatus)

The only microtine rodent that has failed to show

male-induced abortions when tested under conditions

similar to those used for other species is Laqurus

curtatus (Stehn & Jannett, 1981). Females were exposed

to strange males at about day 12 of pregnancy; only one

female failed to deliver in the experimental group, as

did one female in the control group. In the same paper

results were given for prairie and montane voles, both

of which showed male-induced abortions under the same

testing conditions as were used for the sagebrush


Male-induced Abortions in Other Species of Rodents

Syrian Golden Hamsters (Mesocricetus auratus)

Huck (1984) tested female Syrian golden hamsters

for the tendency to abort pregnancies when exposed to

strange males the day after stud male mating. All

females delivered litters, indicating a lack of the

male-induced abortion response. It is difficult to

compare this study with those by other researchers

involving other species. Female golden hamsters are

larger than males and are very aggressive. They chase

males away after mating and remain hostile to male

approach during pregnancy. Thus, they may seldom be

exposed to pheromones from strange males, and they may

not recognize stud males or treat them differently than

strange males. To my knowledge, no experiments have

been done in which male urine is placed on the noses of

female golden hamsters. Such an experiment would be

necessary to determine if the mechanism is present but

not usually triggered because of female avoidance of

contact with males during pregnancy.

In a subsequent study it was found that female

golden hamsters may abort when exposed to other females

(Huck, Bracken, & Lisk (1983). In this situation,

direct contact between the females is necessary, and

subordinate females but not dominant females, are

affected by exposure to strange females. Abortions in

subordinate females consisted of both implantation

failure and resorbtion of implanted embryos (Huck,

Lisk, Miller, & Bethel, 1988). The occurrence of

abortions in subordinate females was positively

correlated with the amount of fighting between the

subject female and the strange female (Huck et al.,

1988). There was also a negative correlation between

fighting and progesterone levels in pregnant

subordinate females. These findings seem to indicate

that in golden hamsters female-induced abortions may

not be the result of the same mechanism as male-induced

abortions in other species. If the mechanism is the

same it is probably not an estrus induction mechanism,

as the usual effect of a dominant female on a

subordinate is to suppress estrus.

Mongolian Gerbils (Meriones unguiculatus)

The occurrence of male-induced abortions has also

been described in Mongolian gerbils (Norris & Adams,

1979; Rohrbach, 1982). Norris and Adams (1979) found

that the rates of abortion were similar for females

exposed to strange males or exposed to other

environmental disturbance, as was noted in wild house

mice above. They also found that strange females

induced abortions, as in golden hamsters. Rohrbach

(1982) found that females exposed to either strange

males or strange females terminated pregnancies at a

much higher rate than females subjected to

environmental disturbance only, although the latter

females had a higher rate than females exposed to

neither strange males nor environmental disturbance.

In contrast to the findings in house mice, the

presence of the stud male did not reduce the incidence

of abortions in female gerbils. Also, re-exposure of

the female to her regular partner caused abortions of

pregnancies caused by a strange male, even though the

female's exposure to the strange male was very brief

(only long enough for copulation to take place). In

one experiment the female had been housed with two

males, but had mated with only one. The other male did

not induce an abortion of the pregnancy caused by the

first male. This is also similar to the results with

house mice in which prior familiarity with the alien

male reduced the incidence of abortions.

In most of the tests in the Rohrbach (1982) study,

the female was separated from her regular partner for

the 10 days of the test. In a different study of

Mongolian gerbils, Norris (1985) claimed that abortions

were due to the separation of the female from her

regular partner, rather than being due to exposure to

the strange male. None of 20 females delivered litters

when removed to a clean cage without the stud male on

the day after mating, but 7 of 20 littered when moved

to a clean cage with a strange male. This would seem

to indicate that removal of the stud male had as strong

an influence on pregnancy rates as the presence of a

strange male.

Unlike house mice, females of some species of

rodents have higher successful pregnancy rates if the

stud male is present at least during the first day or

two of gestation (see discussion in chapter 3). Norris

did not introduce a strange male without removing the

stud, to see if there was any effect attributable to

the presence of the strange male per se. Rohrbach

(1982) performed this test, placing the female and her

partner on one side of a barrier, and the strange male

on the other. In this situation 65% of the females

returned to estrus, indicating that the presence of a

strange male induced abortions even when the stud male

was not removed.

Both Rohrbach (1982) and Norris (1985) used

females in postpartum estrus in their experiments; it

is not clear exactly when the pups were removed, but in

both studies removal apparently occurred just before

testing. Thus these females were probably lactating

for some of the duration of the test, and in spite of

this abortion rates were high. In house mice and some

other species, lactating females do not abort when

exposed to strange males.

Deer Mice and White-footed Mice (Peromyscus)

The first investigation of male-induced abortions

in a species other than house mice was done by

Eleftheriou, Bronson, and Zarrow (1962), who subjected

newly mated female Peromyscus maniculatus bairdii to

various changes in their environments, including

exposure to strange males. Changes in the physical

environment, including placing the females in larger

cages than they had previously occupied, was associated

with pregnancy loss. There was also an effect of

exposure to strange males. When females were housed in

small cages, there was a much lower implantation rate

when strange males were present than when the females

were alone. However, when the females were housed in

large cages there was an equally high failure rate for

isolated females as for females housed with strange

males. This last effect was attributed to the

unfamiliar type of housing.

In further experiments (Bronson & Eleftheriou,

1963), it was shown that, as in house mice, direct

contact with the strange male was necessary, and

increasing the number of males did not increase the

incidence of abortions. Also, exposure of mated female

deer mice to male house mice resulted in pregnancy

rates that were lower, but not significantly lower,

than rates for females housed either with stud males or

alone. In another experiment using a larger number of

subjects, Bronson, Eleftheriou, and Garick (1964),

found significantly decreased pregnancy rates in female

Peromyscus maniculatus bairdii exposed to strange male

deer mice (both P. m. bairdii and P. m. gracilis), male

house mice, and female P. m. gracilis. Male house mice

were less effective than strange male P. m. bairdii.

In both of the deermouse studies, there were large

numbers (about 16%) of females that were apparently

pseudopregnant, even in control groups. The reason for

this was not clear.

Bronson, Eleftheriou, and Dezell (1969) found

that, as in house mice, recently mated female deer mice

that received injections of prolactin at the same time

they were exposed to strange males had significantly

higher implantation rates than exposed females not

receiving prolactin. In addition, because it was

thought that abortions might be the result of

generalized disturbance acting through the pituitary-

adrenal axis rather than a specific response to strange

males, Bronson et al. (1969) tested the effect of

various doses of ACTH on mated female deer mice. They

found no decrease in implantation rates in injected

females as compared with those in uninjected or

vehicle-injected females. Exposure of pregnant females

to strange males was followed by an increase in plasma

corticosterone over the levels in control females

exposed to stud males or no males, but this increase

apparently was not the cause of the higher rates of

implantation failure in females exposed to strange

males. There was no difference between mated

adrenalectomized females and mated sham-operated

females in their responses to strange males. This is

contrary to findings in house mice by Snyder and

Taggert (1967) that adrenalectomized females were less

likely to abort than intact females when both were

exposed to strange males. The latter finding was true

of both inbred albino females and females of a brown

strain that was descended from wild mice.

The effect of parity and the presence of the stud

male on male-induced abortions was tested in deer mice

by Terman (1969). He found that parous females were

less likely than nulliparous females to abort when

exposed to strange males, and that the presence of the

stud male in addition to the strange male makes

abortion significantly less likely than when the

strange male, but not the stud male, is present. At

the same time, however, the pregnancy rate was

significantly less when a strange male was present than

when he was not, even if the stud male was also


In house mice the rate of male-induced abortions

is not significantly different in females exposed to

strange males after implantation than in females not so

exposed. In some voles, however, male-induced

abortions may take place after the time of implantation

(Kenney, Evans, & Dewsbury, 1977; Stehn & Richmond,

1975). In deer mice, there have been conflicting

results from experiments conducted after implantation.

Kenney et al. (1977) reported that some parous female

deer mice that were exposed to strange males on day 14

after insemination aborted. However, the proportion of

exposed females returning to estrus was low (no control

females returned to estrus), and O'Keefe, Pinkston, and

Terman (1985) claimed that there was no significant

difference when the Kenney et al. data were analyzed

using the Fisher exact probability test. In their own

experiment, in which they used nulliparous female deer

mice, O'Keefe et al. (1985) found no significant

difference in pregnancy rates of females exposed to

strange males at various times after implantation, when

compared with unexposed controls.

Dewsbury (1982) reported that female deer mice

that mated sequentially with two different males had

lower pregnancy rates than females that mated with just

one male, when the number of ejaculations, disturbance,

and pacing of copulation were controlled. In another

experiment it was found that females that mated with

one male and then were placed in the cage of a strange

male (separated from him by a wire mesh barrier)

immediately afterward also had lower pregnancy rates.

Females mated with one male in a testing cage and then

placed alone in his home cage had a significantly

higher pregnancy rate than females placed alone in the

home cage of a strange male. This differs from the

kind of male-induced abortions described by Bruce, in

that exposure to the strange male immediately follows

mating to the stud male, and exposure to both males is

relatively brief (Dewsbury, 1982).

Dewsbury (1985), found that when females were

lactating, exposure to more than one male at the time

of mating did not result in lowered pregnancy rates

when compared with females exposed to just one male.

He suggested that there may be commonality in mechanism

between the pericopulatory pregnancy block in deer mice

and the implantation failure that occurs when exposure

to the strange male is 24-48 h after mating.

In white-footed mice (Peromyscus leucopus) female-

induced abortions were reported by Haigh, Cushing, and

Bronson (1988). Young females paired with males at

weaning failed to produce young if another adult female

(or her odor) was also present. The effect was very

long-lasting (as long as the older female was present)

and it was not due to puberty delay, as there was

evidence of cycling and mating by the younger females.

It was concluded that failure of the embryos to implant

in these females was the reason for lack of successful

pregnancies. A similar effect has been found in P.

eremicus by Skryja (1978), thus female-induced

abortions may be more common than currently thought.

Djungarian Hamsters (Phodopus campbelli)

Females of this species also undergo

pericopulatory pregnancy blocks similar to those in

Peromyscus maniculatus (see above). The pericopulatory

block was reported by Wynne-Edwards and Lisk (1984).

They exposed estrous female Djungarian hamsters to two

males. If both males were free to interact with the

female she did not become pregnant, but returned to

estrus and mated again four days later. This sequence

of mating and then returning to estrus continued for as

long as 17 days. Some of these females were sacrificed

and examination of the ovaries showed that ovulations

had occurred. If one of the males was confined, the

first mating of the female with the free male did not

result in pregnancy, but subsequent matings were often


Responses of female Djungarian hamsters to

exposure to strange males that occurred 24-48 hours

after mating were tested by Wynne-Edwards, Huck, and

Lisk (1987). In this study there was an unusual

pattern of effects of removal of the stud male. There

was a high percentage of successful pregnancies if the

stud male was either removed 2 h after mating began, or

was allowed to stay with the female for 48 h or more.

If he was removed 24 h after mating there was a

significant decrease in pregnancy rate. There was no

difference in pregnancy rates for females left alone

after 24 h, females whose mates were replaced with a

strange male after 24 h, and females whose mates were

confined in smaller enclosures within the female's cage

after 24 h. Like Norris (1985) for Mongolian gerbils,

Wynne-Edwards et al. concluded that the primary factor

in pregnancy loss in Djungarian hamsters is removal of

the stud male, and that the female is especially

sensitive to stud male removal at 24 h after mating.

Laboratory rats (Rattus norvegicus)

In a study of laboratory rats, in which the

females were all of the Long-Evans strain, it was shown

that there was no effect of exposure of the female to a

strange male of her strain or of a different strain on

pregnancy rates (Davis & deGroot, 1964). The number of

animals used in this study (4-6 per group) was much

lower than in all of the studies reported above where

male-induced abortions were found and only 2 of 22

females were not pregnant at autopsy.

Bruce (1966) reported that she was also unable to

demonstrate male-induced abortions in laboratory rats,

however, no details were given.

Male-induced Estrus in House Mice and Voles

House Mice

Estrous cycles in isolated females

Female mice housed in isolation have estrous

cycles of about 5-6 days (Whitten, 1958). Behavioral

receptivity and ovulation are spontaneous in isolated

females; that is, they occur regularly without mating

or the presence of a male. The cycle can be divided

into 4 stages, based on the those described for rats

with 4-day cycles (Feder, 1981). These are: proestrus,

during which the eggs go through the final preparation

for ovulation; estrus, the time when the female

ovulates and is receptive to the male; diestrus I

(=metestrus), when the corpora lutea secrete

progesterone (little or none in house mice); diestrus

II, in which the corpora lutea regress and the

reproductive tract returns to the proestrus state.

These stages can be detected in females by

examination of vaginal smears (Snell, 1941). Estrus in

rats and house mice is triggered by a rise in estrogen

during proestrus, and is accompanied by vaginal

cornification (Feder, 1981). When estrogen levels

reach a threshold, an LH surge is triggered and

ovulation occurs. In late proestrus there are also

peaks in FSH, prolactin and progesterone.

Alteration of estrous cycles by exposure to males

Although estrus is spontaneous in house mice,

contact with an intact male house mouse has an effect

on the timing of the female estrous cycle; the exact

effect depends on the previous housing conditions of

the female. As noted above, female house mice kept in

isolation have 5-6 day estrous cycles; however, females

housed in contact with males often have 4 day cycles

(Whitten, 1958). The effect of the male in speeding up

estrous cycles in individually housed females was

assumed to be related to the effect of a male on

grouped females, described next.

Grouping of sexually mature female house mice into

a single cage is associated with a prolongation or

cessation of their estrous cycles (Lamond, 1959; Van

der Lee & Boot, 1955, 1956; Whitten, 1959). This

effect appears to be due to a pheromone produced by

grouped females, since cages soiled by grouped females

caused prolonged cycles in singly housed females

(Champlin, 1971). Exposure of group-housed adult

females to males or their urine causes synchronization

of estrous cycles in house mice (Marsden & Bronson,

1964; Whitten, 1956), and a number of other animals

(see review in McClintock, 1983). Synchronization in

this context means that when previously group-housed

females are exposed to a male, there is a higher

proportion of females mating on the third night after

introduction of the male, than would be expected if the

timing of estrus were independent of male presence.

Therefore, the male has "induced" estrus in the

females, and the effect of the male has been referred

to as an estrus-inducing effect (Bronson & Dezell,


It should be noted that two different procedures

have been used in estrus-induction experiments with

grouped females: one method is to introduce a male into

a group of females (e.g. Whitten, 1959), and the other

method is to remove each female from the group and pair

her individually with a male (e.g. Whitten, 1956). If

only the latter method is used, it is very difficult to

dissociate the effects of removal of the female from

the group from the effects of male exposure. Marsden

and Bronson (1965) showed that mere removal of a

previously group housed female to isolation can induce

estrus with the same timing as removing her from the

group and pairing her with a male. It appears that

there are two separate effects; however, the two

effects are apparently not additive.

Whitten (1959) and Gangrade and Dominic (1984)

showed that introducing a male into a group of females

induced estrus even though the females remained

grouped. This has often been interpreted to mean that

the male's estrus-inducing pheromone can overcome the

estrus-suppressing pheromone of the other females.

However, it is not certain whether the male actually

induced estrus or merely interfered with the production

of the estrus-suppressing pheromone, thereby permitting

estrous cyclicity to resume. Better evidence for a

direct role for males in estrus induction comes from a

study by Whitten (1958), which showed that individually

housed females that had males confined in small wire

baskets placed in their cages had shorter cycles than

females that remained alone.

Problems in the definition of estrus induction

I have used the term "estrus induction" to refer

to the process by which the presence of males or

bedding soiled by males is associated with the presence

of estrus in females. Researchers working with house

mice and other spontaneous ovulators often use the term

"induction of ovulation", since estrus and ovulation

are closely linked in these species, however,

copulation is often what is measured. One problem in

discussing the similarity between male-induced estrus

and male-induced abortions is that the terms "estrus

induction" and "induction of ovulation" have been

applied to several processes which may be different.

For instance, there may be two different estrus-

induction processes occurring in grouped females

exposed to males. Whitten (1959) claimed that under

some conditions prolonged cycles in female house mice

housed in larger groups (e.g., 30) are due to anestrus,

judged by the failure of the uteri to show a decidual

reaction and by reduced ovarian weights and lack of

corpora lutea. It had previously been shown that if

female house mice are housed in fairly small groups

(e.g., 4) in most of the animals the prolonged estrous

cycles are due to pseudopregnancy (Van der Lee & Boot,

1955). Ryan and Schwartz (1977) showed that females

housed in groups of 20 (but with the same space per

animal as in the Whitten, 1959 study) have elevated

progesterone levels, which they claimed are not

different from those of females made pseudopregnant by

infertile matings. These grouped females also have

active corpora lutea. Ryan and Schwartz asserted that

Whitten's technique for assessing the decidual reaction

was not as accurate as the one they used, and that

group-housed females show cycles of a length compatible

with pseudopregnancy; they are not acyclic.

The evidence presented by Ryan and Schwartz (1977)

is convincing; however in the literature there is often

reference to two separate processes in grouped females,

depending on the size of the group. Some authors

(e.g., McClintock, 1983) refer to the prolonged cycles

as "estrus suppression," which avoids the controversial

questions about their exact nature, i.e., whether there

is corpus luteum activity (pseudopregnancy) or not

anestruss). The effect of the male is called "estrus

synchrony" or "estrus induction."

Pubertal female mice that are going from a state

of anestrus to their first estrus come into estrus

sooner if the females are exposed to strange males.

This process is usually termed "puberty acceleration,"

but it is also referred to as "estrus induction." It

seems to have often been assumed that all these forms

of estrus induction are the same process, triggered by

the same pheromone. However, if the prolonged cycles

of grouped-housed females are pseudopregnancies with

elevated levels of progesterone, as suggested by Ryan

and Schwartz (1977), then this "estrus induction" is

different from the "estrus induction" that results from

the exposure of anestrous, isolated, or prepubertal

female house mice to conspecific males.

Pregnant females undergoing male-induced abortions

start from a condition in which the levels of many

reproductive hormones, especially progesterone, are

higher than the levels during anestrus. Many studies

involving estrus induction in house mice involve

pseudopregnant females (group housed in groups of

5-10). These females would have hormone levels similar

to those in pregnant females. Estrus induction in

pseudopregnant female house mice and male-induced

abortions in pregnant female mice may indeed be the

result of the same process (see below). However,

acceleration of estrus in singly housed females or in

pubertal females may be a different process. It is

these latter kinds of estrus induction that Bronson and

Coquelin (1980) claimed have been selected for as part

of the opportunistic reproductive strategies of house

mice, and it is this system that they believe produces

male-induced abortions as a side-effect. Therefore,

when considering the question whether male-induced

abortions are a side effect, the contrast that needs to

made is that between male-induced abortions and male-

induced estrus other than that in pseudopregnant

grouped females. As Bronson and Coquelin suggested,

induction of estrus in grouped pseudopregnant females

is probably irrelevant in the wild. If anything is a

side effect, it is just as likely that estrus induction

in pseudopregnant females is a side-effect of male-

induced abortions.

None of the evidence so far, however, rules out

the possibility that in pregnant and pseudopregnant

mice, "estrus induction" is really a two step process:

termination of the pregnancy or pseudopregnancy

followed by an immediate return to estrous cyclicity.

Estrus induction in pubertal females may include a

different "first step", followed, as in pseudopregnancy

and abortions, by an establishment of estrous


Prairie Voles

Female prairie voles rarely become sexually

receptive unless exposed to males (Richmond & Conaway,

1969; for a comparative review, see Sawrey & Dewsbury,

1985). Richmond and Conaway found that only 1 of 36

isolated females, smeared daily for 4 weeks, showed

signs of vaginal cornification. Shapiro and Dewsbury

(1990) found that over a 3 day period of daily vaginal

smears, 6 of 30 female prairie voles that were housed

in isolation showed at least one estrous smear, defined

as a smear containing at least 50% cornified cells.

Richmond and Conaway found that a number of treatments,

including moving isolated females to small

compartments, as well as exposing females to males or

male-soiled cages, induced estrus; however, exposing

females to males, either directly or across a mesh

barrier had by far the greatest effect.

The ability of castrated males to induce estrus in

adult female prairie voles apparently has not been

tested, however, castrated males are much less

effective than intact males in inducing reproductive

activation (measured by uterine weight gain) in

pubertal females (Carter, Getz, Gavish, McDermott, &

Arnold, 1980).

Like pubertal female house mice, female prairie

voles exposed to males (after having been isolated from

them) begin from a condition of anestrus (noncycling,

often referred to as diestrus) in which levels of all

reproductive hormones are probably low. Female montane

voles (Microtus montanus) that had diestrous smears for

2 consecutive days had low plasma levels of LH and

progesterone compared with recently mated females

(Gray, Davis, Kenney, & Dewsbury, 1976). It is likely

that adult female prairie voles that are not paired

with males have the same low levels of hormones.

If this is true, then male-induced estrus in

previously isolated female prairie voles is due to a

similar process as male-induced estrus in anestrous

(but not pseudopregnant) female house mice. It has

been suggested (Taylor, 1990) that male-induced estrus

is not an all-or-none phenomenon, but rather that

females of some species are more "male-dependent" for

estrus induction than females of other species. Using

Taylor's terminology, it can be said that female house

mice are essentially male-independent, while prairie

voles are highly dependent on male stimuli for estrus




General Goals, Design, and Methods

Three experiments were performed to test the

hypothesis that male-induced abortions and male-induced

estrus are the result of a single mechanism. As

mentioned previously, the experimental strategy was to

expose pregnant or anestrous female prairie voles to

heterospecific males to determine if the latter induced

estrus and/or abortions. If heterospecific males

induce both estrus and abortions, this would be

evidence for a common mechanism. If, however,

heterospecific males induced estrus but not abortions,

or vice versa, this would be evidence for two separate


Before testing the effect of heterospecific males,

it was necessary to establish that the abortions

observed could be attributed to the presence of a

strange conspecific or heterospecific male, rather than

to the removal of the stud male. In some species of

rodents, higher pregnancy rates are obtained if the

stud male is allowed to remain with the female for at

least the first one or two days of gestation, than if

he is removed. This seems to be true of prairie voles


(Richmond & Stehn, 1976), montane voles (Berger &

Negus, 1982), Djungarian hamsters (Wynne-Edwards, Huck,

& Lisk, 1987), and Mongolian gerbils (Norris, 1985).

In these species, if male exchanges are done within

24-48 h of mating, the effect of the presence of the

strange male may be confounded with an effect of the

absence of the stud male. However, it has also been

found that removal of the stud male on day 4 or day 12

after mating had no effect in prairie voles (Richmond &

Stehn, 1976). Similar results were obtained for female

montane voles by Berger and Negus (1982) and by Taylor

(1990), that is, removal of the stud male after 2 h of

mating significantly reduced pregnancy rates when

compared with a condition in which a wire mesh

partition was placed between the male and female to

prevent further mating, but the male was not removed.

It was not known whether, in the population of

prairie voles used in this experiment, there might be

an effect on pregnancy of removal of the original male

on day 2. Therefore, in Experiment 1 the percentage of

females delivering litters was determined for females

that were: a) left with the original male; b) left

alone starting on day 2 after mating; or c) separated

from the stud male and exposed to an unfamiliar male on

day 2 after mating. The last group was included

because testing of the ability of heterospecific males

versus unfamiliar conspecific males to induce abortions

could not be done concurrently with Experiment 1. The

inclusion of the stud male and unfamiliar conspecific

male groups in both experiments was done to provide

assurance that conditions had not changed between


In Experiment 2, the effect of replacement of the

original male prairie voles with male montane voles,

male meadow voles, and unfamiliar male prairie voles on

the delivery of litters by mated female prairie voles

was compared with the effect of retaining the original


In Experiment 3, the ability of unfamiliar male

prairie voles, meadow voles, and montane voles to

induce estrus in previously isolated female prairie

voles was compared to the rate of spontaneous estrus in

females that remained alone. Estrus was determined by

examining vaginal smears. Previous work in this

laboratory (e.g. Taylor, 1990) has suggested that while

behavioral receptivity is not always associated with a

high level of vaginal cornification in female prairie

voles, females that show cornified smears are much more

likely to be receptive than females with predominantly

leucocytic smears; the latter are almost never

receptive. Female prairie voles rarely show complete

cornification, as is often described in house mice.


All animals used in these experiments were born in

colonies maintained at the University of Florida. Each

species occupied a separate room or rooms in the animal

care facility of the Department of the Psychology.

There were two rooms housing prairie voles, one housing

the breeding colony and the experimental animals to be

used in this study and the other housing experimental

animals used in other studies. Male and female prairie

voles were maintained in the main prairie vole colony

room from birth until they were entered into an

experiment. Montane and meadow voles were born and

maintained in their respective colony rooms, except

that fertility testing of male meadow voles was done in

a different room, occupied only by meadow voles.

The prairie vole colony was derived from animals

obtained from the University of Illinois in 1983.

Except for a period lasting about a year and ending 4-5

years ago, in which mating of relatives was done to

obtain animals homozygous for transferring alleles, an

effort has been made to minimize inbreeding. The

montane vole colony was derived from animals obtained

in 1985 from a colony at the University of Utah (Berger

& Negus, 1982). The meadow vole colony originated from

animals obtained from the University of Massachusetts

in 1986.

All colony rooms are windowless and air-

conditioned; the temperature was maintained at 60-65 F.

All animals were maintained on 16:8 light-dark cycles

with lights off at 1200 h. Cages of all colony animals

were changed weekly except when the experimental

protocols described below dictated an alternative

schedule. Pine shavings were used for bedding; water

was continuously available. Prairie voles were given

Purina Rabbit Chow ad libidum; montane voles also

received Purina Rat Chow in addition to Rabbit Chow.

Meadow voles received Purina Rat Chow only.

All animals used in the experiments described

below were fertility tested by being placed with a non-

experimental animal of the opposite sex long enough to

produce at least 1 but not more than 2 litters. Female

prairie voles were allowed at least 10 days to recover

after removal of young before being used in


Animals of all three species were housed after

weaning in litter groups in 48 x 27 x 13 cm clear

polycarbonate cages. Fertility testing was done in the

same size cages. After fertility testing, animals were

housed individually in 29 x 19 x 13 cm polycarbonate

cages except when being used in experiments. Males of

all species were housed singly for at least 10 days

before being used. Fertility-tested male prairie voles

were assigned to two different pools, one to be used as

studs and the other to be used as unfamiliar males;

they were assigned by drawing colored discs from a cup.

Female prairie voles were between 90 and 205 days

old when assigned to a group (see below for means for

each group). Male prairie voles were 134 to 277 days

old, male meadow voles were 110 to 282 days old, and

male montane voles were 103 to 336 days old when used.

Male and female prairie voles were used only once.

However, because of a shortage of fertility-tested male

meadow and montane voles, some had to be used twice,

once in Experiment 2 and once in Experiment 3.

Experiment 1 was run between February 12, 1991 and

August 30, 1991. Experiment 2 was run between August

28, 1991 and December 7, 1991. Experiment 3 was run

between December 7, 1991 and February 10, 1992.

General Methods

All experiments were carried out in 48 x 27 x 13

cm polycarbonate cages that were divided into two equal

sized (504 cma) compartments by a removable wire mesh

screen. Animals could see, hear, smell, and touch each

other through the mesh, but extensive physical contact

was not possible. Urine deposited on or near the

barrier could easily pass through to the other side and

be directly contacted by the other animal. Animals

occasionally transferred bedding material from one

compartment to the other by pulling it underneath the


The experimental protocols for Experiments 1 and 2

were the same and are described here; the protocol for

Experiment 3 is described in the section devoted to

that experiment. In both Experiments 1 and 2 mated

pairs were required; these were obtained in the

following manner. Each female was started in the

experiment as soon as possible after fertility testing.

Logistical considerations prevented starting more than

4 females per day, and animals could not be run every

day; therefore, short delays (10 days to 3 weeks)

sometimes occurred between fertility testing and the

beginning of an experiment. At the start of the

experiment for each female, a fertility-tested male was

selected for her from the stud male pool by drawing

numbered plastic discs from a cup, with the constraint

that he could not be a sibling of the female or of the

male used in fertility testing the female.

Each pair was then placed into a divided cage,

with the male in the front compartment and the female

in the rear compartment. Pairing was done between 0900

and 1200 h. Approximately 48 h after pairing, at the

beginning of the dark period, the cage was placed on a

testing table, and the barrier was removed. Copulatory

behavior was observed during a 2-h period, and the

measures of copulatory behavior described below were

recorded on a computer using software developed by

R. Kumar of the University of Florida. Light for

observation was provided by a 15 watt red light placed

on top of the cage. If a pair failed to copulate

within 30 min, the barrier was replaced and the pair

was tested again the next day. Pairs that had not

copulated after 3 days were separated for at least 10

days before being tested again; in this case the female

was assigned a new male as above. About 15% of pairs

failed to copulate during the 3-day period in both

experiments 1 and 2. Whether such pairs were tested

again depended on how many animals were available for

use at the time; only 9 females that failed to mate the

first time were tested again, and only 3 of those

copulated on the second test.

Behavioral measures recorded for each pair

included mount latency (ML), the time in seconds

between removal of the barrier and the first mount or

intromission by the male; intromission latency (IL),

the time in seconds between removal of the barrier and

the first intromission; and ejaculation latency (EL),

the time in seconds from the first intromission of a

series until ejaculation. Also recorded were the

number of intromissions in a series (IF), number of

thrusts (TF), and total number of ejaculations (EF).

At the end of the 2-h mating session, each mated

pair was assigned to an experimental group by drawing

numbered plastic discs from a cup, with the constraint

that littermate sisters could not be assigned to the

same group, and nonlittermate sisters could not be

assigned to the same group unless applying that rule

would lead to no assignment for that pair. No more

than two full sisters were allowed in each experimental

group. This was done so that all of the genetic

variability present in the colony would be represented

in each of the groups.

The barrier was replaced and the pair was left

undisturbed for approximately 48 h. After 48 h, the

male was removed from his compartment, and all of the

bedding material from his compartment was replaced with

fresh bedding.

Depending on the experimental protocol, described

under each experiment, the original male was then

reintroduced, or an unfamiliar conspecific or

heterospecific male was introduced. The cage

containing the pair was then moved from the prairie

vole colony room to the colony room housing the species

of the unfamiliar male. This was done so that the

females that were being exposed to meadow or montane

vole males would not simultaneously be exposed to odors

from male prairie voles in the prairie vole colony

room. In the case of groups with familiar or

unfamiliar prairie vole males as stimulus animals, the

cage was removed to the second prairie vole room. Thus

in all groups, the cage was removed from the main

prairie vole room, where the mating had occurred, to

another room that contained only males and females of

the same species as the stimulus male.

Pairs were left undisturbed, except that I looked

into each cage daily to check food and water, which

were replaced as needed. Cage lids were not removed at

these times. On day 12 after mating, cages were moved

to a cart or table, and the lids were removed in order

to change several handfuls of bedding from each side.

There was no effort made to remove equal amounts of

bedding, rather, an effort was made to leave the cages

equally soiled.

On day 18 or 19 after mating, females were gently

picked up and palpated to assess pregnancy. This was

done to guard against the possibility that litters

would be born and destroyed before being recorded. In

no case was there a discrepancy between the results of

palpation at this time and subsequent delivery or non-

delivery of a litter. On the mornings of days 21 and

22 after mating, females were examined for litters; if

a litter was found the number of pups was recorded.

Females not delivering litters by day 22 were checked

again the following morning. Only one female gave

birth on day 23; the remainder that had not given birth

by day 23 were confirmed to be nonpregnant by thorough


Experiment 1

The purpose of this experiment was to compare the

responses of females to exposure to an unfamiliar male

in the absence of the stud male from day 2 of

pregnancy, with the responses of females that were

separated from the stud male but not exposed to

unfamiliar males.


Mated pairs were obtained and assigned to

experimental groups as described above so that at the

end of testing each group contained twenty mated pairs.

A total of 62 female, 62 stud male, and 20 unfamiliar

male prairie voles were used in Experiment 1. Two

females (one in Group 1 and one in Group 2; see below

for group definitions) died before completion of the

experiment; the pairs involved were replaced with new

pairs. Age distributions of the females were as

follows: Group 1 mean = 152 days, range = 120-205 days;

Group 2 mean = 155 days, range = 126-201 days; Group 3

mean = 145 days, range = 125-195 days. Age

distributions of males were: Group 1 mean = 202 days,

range = 121-275 days; Group 2 mean = 198 days, range =

139-252 days; Group 3 (studs) mean = 192 days, range =

149-281 days; Group 3 (unfamiliar) mean = 203 days,

range = 145-281 days.

On day 2 after mating, the following manipulations

were made after the original stud male was removed and

the bedding changed (as described above): for Group 1,

the stud male was returned to his compartment and

remained across from the female for the remainder of

gestation; for Group 2, the female was left across from

an empty compartment; for Group 3, an unfamiliar male

prairie vole was placed in the vacant compartment.

Unfamiliar males were selected from the unfamiliar male

pool by drawing numbered discs from a cup, with the

constraint that the unfamiliar male could not be a

sibling of either the female, the stud male, or the

male used to fertility test the female. This was done

to avoid the possibility that the female would have had

previous contact with that male's odor.

Animals were maintained and pups counted at the

end of gestation as described above.


The results of Experiment 1 are summarized in

Table 2. Results of a chi-square test show a

significant effect of treatment on pregnancy among the

three groups ( 1 = 18.59, d.f = 2, N = 60, p < .001).

There was no statistically significant difference

between the percentage of females that delivered

litters when the original stud male stays throughout

gestation and the percentage delivering when there is

no male across from the female after day 2 (Group 1

versus Group 2; Fisher's exact probability test, N =

40, p > .17). Therefore, removal of the original male

is not by itself associated with abortions under these

experimental conditions. Females that were housed

across from unfamiliar males starting on day 2 after

mating were less likely to deliver litters than females

that remained across from the stud male (Group 1 versus

Group 3; N = 40, p < .0002), or across from an empty

cage (Group 2 versus Group 3; N = 40, p < .005).

Analysis of the copulatory behavior data using a

parametric ANOVA test showed that there were no

significant differences in any of the measures between

mated pairs assigned to the three different groups (see

Table 3). These results show that the differences in

percentage of females delivering litters in the

different groups were not due to any systematic

differences in the mating behavior of the pairs

assigned to each of the groups.

The mean litter size for Group 2 females was

higher than for the other two groups (see Table 4).

Because of the low N in some of the groups, a Kruskal-

Wallis ANOVA by ranks test was chosen to compare the

mean number of pups per litter (for females that

delivered litters). This test revealed a significant

difference over all groups (p = .03; see Table 4). The

statistical significance shown by the ANOVA is due to a

high mean (5.0) for Group 2 (no male), and this high

mean is due to two females, each of which had litters

of 8 pups, an unusually large number. There is no

difference in litter size between Group 1 (original

male) and Group 3 (unfamiliar male).

Mean age of the males did not differ significantly

between stud and unfamiliar males t (38) =.91, p >.05,

or between unfamiliar males paired with pregnant or

nonpregnant females, t (18) = .59, E >.05. In 10 of 20

cases the females remained pregnant when the stud male

was older than the unfamiliar male, and in the other 10

cases the female aborted when the stud male was older

(see discussion in chapter 4).

Table 2
Numbers and Percentages of Females Delivering Litters
in Experiment 1.

Type of Exposure Number Number

from day 2 tested of litters %

Group 1 Stud Male

Group 2 No Male

Group 3 Unfam. Male

Note. Fisher's exact probability test
Group 1 versus Group 2, E > .17
Group 1 versus Group 3, E < .0001

Table 3
Measures of Copulatory Behavior in Animals Assigned to Stud
Male, No Male, and Unfamiliar Male Groups in Experiment 1

Type of Male

Measure Stud None Unfamiliar F P

ML 164.2 112.6 155.3 .40 .68

IL 235.5 149.8 165.2 .61 .55

EL 333.2 236.1 271.1 .55 .58

IF 9.4 9.0 8.1 .19 .82

TF 39.6 42.2 40.6 .13 .87

EF 3.7 4.2 4.4 1.74 .18

Note: All measures except EF are for the first series only.

Table 4
Comparisons of the Number of Pups per Litter Among the
Groups in Experiments 1 and 2.

Type of Male

P/lit. Stud None Unfam. Meadow Montane H E

Exp 1 3.6 5.0 3.9 -- -- 6.98 .03

Exp 2 4.2 -- 3.7 3.4 4.7 4.38 .22