Citation
Copulatory behavior of Rattus rattus

Material Information

Title:
Copulatory behavior of Rattus rattus
Creator:
Estep, Daniel Quen, 1949-
Copyright Date:
1975
Language:
English
Physical Description:
vii, 136 leaves : ill. ; 28cm.

Subjects

Subjects / Keywords:
Ejaculation ( jstor )
Female animals ( jstor )
Mating behavior ( jstor )
Pregnancy ( jstor )
Rats ( jstor )
Reproduction ( jstor )
Roofs ( jstor )
Satiety ( jstor )
Species ( jstor )
Ultrasonics ( jstor )
Dissertations, Academic -- Psychology -- UF ( lcsh )
Psychology thesis Ph. D ( lcsh )
Rats -- Behavior ( lcsh )
Genre:
bibliography ( marcgt )
non-fiction ( marcgt )

Notes

Thesis:
Thesis--University of Florida.
Bibliography:
Bibliography: leaves 112-117.
General Note:
Typescript.
General Note:
Vita.
Statement of Responsibility:
by Daniel Quen Estep.

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright [name of dissertation author]. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Resource Identifier:
025293515 ( AlephBibNum )
02787279 ( OCLC )
AAT0766 ( NOTIS )

Downloads

This item has the following downloads:


Full Text











COPULATORY BEHAVIOR OF Rattus rattus










By

DANIEL QUEN ESTEP


A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF
THE UNIVERSITY OF FLORIDA
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE
DEGREE OF DOCTOR OF PHILOSOPHY















UNIVERSITY OF FLORIDA

1975





















































UNIVERSITY OF FLORIDA


3 1282 08552 8437



















ACKNOWLEDGEMENTS


I would like to express my gratitude to Professor

Donald A. Dewsbury for his continued help, advice and

criticism of all aspects of this study. I would also like

to thank Profs. J. H. Kaufmann, E. F. Malagodi, M. E, Meyer

and W. B. Webb for their careful reading of the manuscript.

Dr. Steven Humphrey, Dr. L. E. Harms and Dr. T. Kelly all

provided valuable aid in various aspects of the research.

Their assistance was appreciated. I would also like to

thank Theodore Fryer and Isaiah Washington for caring for

the animals used in this research. Finally, I would like to

thank my wife, Barbara, for typing the manuscript and for

her love, patience, understanding and moral support

throughout my years of graduate study.














TABLE OF CONTENTS


Page

ACKNOWLEDGEMENTS...................................... iii

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

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

The Study of Copulatory Behavior................... 1

The Description of Copulatory Behavior.............. 4

The Adaptive Significance of Variations
in Copulatory Behavior................................ 7

Methods in the Study of Adaptive Significance........ 8

Variations in Copulatory Behavior and
Reproductive Success................................. 11

The Natural History of Roof Rats.................... 19

The Present Research.................................. 21

EXPERIMENT 1.................................... ....... 24

Materials and Methods ....................... ......... 24

Subjects.......................................... 24

Apparatus .... ..... .................... ......... 25

Procedures............................... ........ 26

Behavioral Measures.............................. 27

Results and Discussion................................ 32

Basic Motor Patterns of Copulation............... 32

-Quantitative Measures of Copulatory Behavior.... 35

Categorization of Behavior...... .. ............. 57

Ultrasonic Vocalizations... ...................... 71










TABLE OF CONTENTS (continued)


Page

EXPERIMENT 2............................................ 84

Materials and Methods ............................. 85

Subjects........................................ 85

Apparatus....................................... 85

Procedures ................................... ... . 85

Behavioral Measures.............................. 88

Results and Discussion..... ........................ 88

GENERAL DISCUSSION.................................. ... 102

REFERENCES.............................................. 112

APPENDIX A. ANALYSES OF VARIANCE FOR EXPERIMENT 1.... 118

APPENDIX B. ANALYSES OF VARIANCE FOR EXPERIMENT 2.... .127

BIOGRAPHICAL SKETCH .................................... 136















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



COPULATORY BEHAVIOR OF Rattus rattus

By

Daniel Quen Estep

August, 1975

Chairman: Donald A. Dewsbury
Major Department: Psychology

The copulatory behavior of a wild population of roof

rats (Rattus rattus) was investigated in two experiments.

In the first experiment the copulatory behavior of 12 pairs

of laboratory-reared wild roof rats was observed on a total

of 71.tests, each carried to a 30 minute satiety criterion.

Females were brought into behavioral estrus with the aid of

exogenous hormones. The basic motor patterns of copulation

were described and the standard measures of copulatory

behavior were taken. Categorizations of the behaviors

accompanying copulation were also made on one test each for

each male and each female of 11 pairs of animals. The

occurrence and duration of bouts of ultrasonic vocalizations

were also recorded on each of these 22 categorization tests.

With regard to most of the aspects of their copulatory

behavior, roof rats appeared essentially identical to

laboratory or Norway rats (Rattus norvegicus). Roof rats








were found to display a pattern of copulatory behavior

characterized by no copulatory lock, no intravaginal

thrusting, multiple ejaculations and usually multiple

intromissions prior to ejaculation. Roof rats were also

similar to Norway rats with regard to most of the quantita-

tive measures of copulation, the behaviors accompanying

copulation and patterns of ultrasonic vocalization. Both

male and female roof rats were observed to emit ultrasonic

vocalizations at 28 kHz during copulation.

In the second experiment the role of multiple

ejaculations in the initiation of pregnancy was investigated.

Sixteen young females were mated twice, once to males mating

to only one ejaculation and once to males mating to a 30

minute satiety criterion. Eighty-seven percent of the females

mated to satiety became pregnant while only 56% of the

females mated to one ejaculation became pregnant. It was

concluded that like Norway rats, roof rats may reach the

maximal probabilities of pregnancy with only one ejaculation,

although it appears that roof rats may require more intro-

missions than Norway rats to become progestational.

The results from these two experiments were compared to

similar data from other muroid rodents and the results were

discussed in terms of the adaptive significance of variations

in copulatory behavior. Finally, the differences and

similarities in copulatory behavior of roof rats and labora-

tory rats were discussed in terms of possible behavioral

isolating mechanisms between the two species.














INTRODUCTION


The research reported in this dissertation provides

the first quantitative description of the copulatory

behavior of wild roof rats (Rattus rattus). Included in

this description is an analysis of the behaviors

accompanying copulation, including ultrasonic vocaliza-

tions. In addition there is an experimental analysis of

the role of multiple ejaculations in the initiation of

pregnancy in this species. To provide a rationale for this

research and to review the appropriate background literature

the introduction has been divided into several sections.

The first section provides the rationale for studies of

copulatory behavior and of adaptive significance. The

latter sections review the relevant literature in these

areas, provide some background information about roof rats

and outline the research to be reported.


The Study of Copulatory Behavior


Mammalian copulatory behavior provides an excellent

locus for comparative behavioral analysis. Among the

reasons for this are the following. First, it has been

found that the copulatory behavior of mammals is highly

stereotyped within species but can show considerable

variability between different species. Second, copulatory









behavior is of great biological significance, being near the

heart of biological fitness. Finally, copulatory behavior

can be studied in the laboratory making it an excellent

preparation for experimental analysis.

The variability seen in patterns of copulation among

mammals provides the very essence of comparative behavior

analysis. If there were little variability in copulatory

behavior, comparative analysis would be impossible. The

fact that the variability in copulatory behavior occurs

primarily between species rather than within suggests that

patterns of copulatory behavior may be under significant

genetic control and subject to evolutionary pressures.

This in turn suggests that species differences in copulatory

behavior may be due to different environmental pressures

selecting for certain patterns over others. As pointed

out by Dewsbury (1972b), comparative analyses of these

species differences and the evolutionary pressures

responsible for them may contribute significantly to our

general understanding of the evolution of both behavior and

of reproductive systems in general. The ways in which

mammalian copulatory behavior can vary will be summarized

below.

The reproductive success of a particular mammalian

organism is in large part dependent upon its success in

copulatory behavior. Except in cases of human intervention,

organisms that do not engage in appropriate copulatory

behavior do not leave offspring. The biological fitness








of an organism is expressed in terms of the relative numbers

of viable, fertile offspring left by that individual, in

sexually reproducing organisms. Efficient, successful

copulatory behavior is absolutely essential in order to

maximize reproductive success and hence, biological fitness.

Successful copulatory behavior involves a number of factors.

Mating must occur when there are maximal probabilities that

viable, fertile offspring result. Similarly, the mating

itself must not only transfer adequate numbers of sperm from

male to female, but must also provide the necessary and

sufficient stimulation to induce successful pregnancy.

Studies by Adler (1969), Dewsbury and Estep (1975), and

McGill (1970), have shown that the simple transfer of sperm

is not sufficient to induce normal pregnancy in some

mammalian females and that specific patterns of copulatory

stimuli from the male may be necessary. Small variations in

these normal patterns of behavior can disrupt normal

reproduction and hence lower biological fitness (Adler, 1969;

Adler & Zoloth, 1970). It is clear, therefore, that

copulatory behavior is near the heart of the reproductive

success of an individual.

For a number of reasons copulatory behavior is an

excellent behavior pattern for laboratory study. It is

first of all highly stereotyped in its appearance and can

be reliably elicited under a variety of conditions. Few

mammalian behavior patterns show such stability. In

addition, the occurrence of copulatory behavior does not









depend upon prolonged training or extended practice among

subjects. All of this means that the investigator can

generate stable baselines in behavior both within and

between subjects of a given population. Such stable

baselines are ideal for experimental analyses and detailed

comparisons across populations and species.

The Description of Copulatory Behavior


Comparative analysis begins with the basic description

of the events to be compared. Before comparisons become

meaningful, however, it is necessary to develop a system

of classifying the behaviors observed. Dewsbury (1972b)

has developed a taxonomy of mammalian copulatory behavior

which suggests the ways in which copulatory behavior can

vary and provides a framework for summarizing and comparing

patterns of copulatory behavior.

As Dewsbury (1972b) points out, various aspects of

the entire copulatory situation may-vary from species to

species. First, the antecedent conditions for copulation

may differ. Some species, for example, breed seasonally

while others breed the year round. The time of day,

location of copulation and social context may also be

significant and vary between species. Second, the nature

of courtship or precopulatory behaviors may vary. Some

Species may have elaborate courtship behaviors while others

do not. Third, the postures assumed during copulation may

vary. Some species such as humans have ventral-to-ventral








copulation (Ford & Beach, 1951) while others, such as rats,

have dorsal-ventral copulation (Beach & Jordan, 1956).

Fourth, the behaviors accompanying copulation may vary. In

some species such as laboratory rats, (i.e., domesticated

Norway rats, Rattus norvegicus), copulation itself may

account for less than one percent of the total time of a

copulatory sequence (Dewsbury, 1967). The behaviors

occurring during the rest of the time may vary considerably

among species. Finally, the copulatory behavior itself may

vary in a qualitative or quanitative fashion.

After careful review of the available data on copulatory

behavior in mammals, Dewsbury (1972b) proposed that copulatory

behavior may vary in four basic ways. First, there may or

may not be a lock or physical tie between male and female

during copulation. Second, intravaginal thrusting may or

may not occur after the male has gained penetration. Third,

multiple insertions without sperm transfer, may or may not

be necessary prior to ejaculation. Finally, multiple

ejaculations may or may not occur during the copulatory

sequence. As there are two alternatives to each of these
4
four attributes of copulatory behavior, there are 2 or 16

different patterns of behavior possible. Dewsbury has

classified the copulatory behavior of laboratory rats for

example, as pattern #13 because laboratory rats (Rattus

norvegicus) show no copulatory lock, have no intravaginal

thrusting, have multiple intromissions prior to ejaculation

and have multiple ejaculations within a normal copulatory









sequence. House mice (Mus musculus) by contrast, have been

assigned pattern #9 on the basis of the lack of a copulatory

lock, and the presence of intravaginal thrusting, multiple

intromissions and multiple ejaculations.

The various attributes of copulatory behavior can vary

quantitatively as well as qualitatively. The frequencies

and durations of copulatory events may vary as well as the

latencies and intervals between events. Laboratory rats,

Mongolian gerbils (Meriones unguiculatus), and rice rats

(Oryzomys palustris), for example, all display pattern #13

copulatory behavior but show substantial variability with

regard to the quantitative measures of copulation (Beach &

Jordan, 1956; Davis, Estep & Dewsbury, 1974; Dewsbury,

1970). Any complete description of the copulatory

behavior of a given species must include an account of the

quantitative variations as well as the qualitative

variations.

Descriptions of the copulatory behavior of mammals

can be found scattered throughout the biological and

behavioral literature. However, as Dewsbury (1972b) notes,

most of these descriptions are inadequate for most comparative

purposes. In recent years Dewsbury and his colleagues have

been collecting systematic quantitative descriptions of the

copulatory behavior of a variety of muroid rodents (See for

example, Dewsbury,-1970; 1971; 1972a; 1973b; 1974a; 1974b;

Estep, Lanier & Dewsbury, in press; Gray & Dewsbury, 1973).

These studies provide the kind of complete descriptions









discussed above (including analyses of motor patterns,

quantitative and qualitative variations in pattern and

behaviors accompanying copulation) that are necessary to

facilitate interspecific comparisons. Other good

descriptions are to be found scattered among other groups

of mammals such as primates (See Dewsbury, 1972b).

Among the behavioral patterns found to accompany

copulation, one of the most interesting is the occurrence

of ultrasonic vocalization. These high frequency

vocalizations have been reported to occur in a variety of

species and in a variety of social situations (Sewell, 1970).

With regard to the ultrasonic vocalizations occurring during

copulatory sequences the work of Barfield and Geyer (1972;

1975) on laboratory rats is most important. They have found

that male laboratory rats have regular patterns of ultrasonic

vocalization that follow most ejaculations. They have

described and quantified this pattern in detail and have

discussed its possible biological functions. Descriptions

of such patterns in other species would provide added detail

to the description of the behaviors accompanying copulation

and perhaps allow some assessment of the functions of such

behaviors in different species.

The Adaptive Significance of Variations

in Copulatory Behavior


While the description of copulatory behavior is a

worthwhile and necessary first step in the understanding of









mammalian copulatory patterns, it is not an end unto itself.

Such descriptions achieve significance only when related to

some meaningful question concerning copulatory behavior.

One such question concerns the biological function or

adaptive significance of variations in copulatory behavior.

Students of animal behavior have often wondered why different

species of mammals display such diversity in their copulatory

behaviors. However, few systematic studies have been

generated to investigate this question. Dewsbury (1972b)

and his colleagues are attempting one such systematic

investigation within a taxonomically restricted group of

mammals muroid rodents. By studying a behaviorally

diverse but closely related group such as the muroids,

probabilities of discovering the biological significance of

differences in copulatory behavior are greatly enhanced.

How then, might one investigate the adaptive significance

of variations in copulatory behavior?

Methods in the Study of Adaptive Significance


Dewsbury (1973a) lists three methods for investigating

questions of adaptive significance: The method of adaptive

correlation, the behavior-genetic method and the experimental

method. The first, the method of adaptive correlation,

involves correlating variations in the behavior of interest

with variations in other behaviors, morphology and ecologi-

cal variables. By sampling a variety of organisms with

different behaviors and looking for such correlations,









commonalities and consistent differences should become

apparent that would suggest the functions of such

differences and the environmental pressures responsible for

them. A classic example of this method is presented by

Cullen (1957) in the study of cliff nesting Kittiwakes. She

was able to relate specific reproductive adaptations of

these birds to their cliff nesting habits when comparisons

were made to related non-cliff nesting gulls. The functions

of these special adaptions to the nesting habitat became

clear only when compared to other species who differed in

their nesting habits.

The behavior-genetic method involves the use of inbred

lines of animals and therefore is of restricted utility.

This method involves comparing the behavior of F1 hybrids

with the behavior of the inbred lines that gave rise to

them. By comparing the scores of the F1s to those of the

parents, conclusions can be drawn concerning the strength

of the selective pressures operating on the parental types

and therefore the adaptedness of the behavior. Bruell

(1964, 1967) discusses the utility of this method in some

detail. A recent example of this method is the work of

Dewsbury (1975b). He has utilized this method to investigate

the adaptive significance of variations in the copulatory

behavior of laboratory rats.

The final method discussed by Dewsbury is the

experimental method. By directly manipulating the behavior

or the results of the behavior and looking for effects on









reproductive success or fitness, one can suggest possible

adaptive functions for the behavior. A good example of

this method concerns the egg shell removal behavior of gulls

studied by Tinbergen (1963). Many gull species engage in

egg shell removal behaviors soon after chicks are hatched.

By systematically removing and replacing egg shells of

various colors and looking at chick mortality, Tinbergen

was able to show that egg shell removal helped prevent

detection of the nest by predators and reduced chick

mortality. The behavior was clearly adaptive and increased

the reproductive success of those engaging in such behavior.

Similar analyses have been carried out by Adler and his

colleagues (See Adler, 1974 for a review), Dewsbury and

Estep (1975) and Lanier, Estep and Dewsbury (in press) in an

attempt to discover the adaptive significance of variations

in copulatory behavior. The results of these analyses will

be summarized below.

Each of the three methods discussed above has its own

set of advantages and disadvantages. The methods differ

with regard to the kinds of information they can generate

and the requirements for their utilization in terms of

subjects and procedures. It should be noted that the

methods are not contradictory in nature but may, in fact,

be used to complement each other to provide a more thorough

analysis of the adaptive significance of a particular

behavior.








For a number of reasons, the experimental method may

be particularly attractive to those engaged in laboratory

investigations of copulatory behavior. First, it does not

necessarily require that a number of species be studied and

compared directly. Second, investigations such as those of

Adler, Dewsbury and Estep and others might be conducted

along with a basic description of the copulatory behavior of

a species thus allowing immediate access to information

concerning the possible adaptive significance of the pattern

described. Finally, experimental analyses of adaptive

significance can provide data for a level of comparative

analysis beyond simple description.


Variations in Copulatory Behavior and Reproductive Success


What, then, is the adaptive significance of variations

in copulatory behavior? How might variations in copulatory

behavior affect reproductive success and fitness? Adler

(1969) and others have found that apparently small

variations in male copulatory behavior can directly affect

the female's reproductive physiology and thereby affect

reproductive success. In a comprehensive review of the

mammals, Adler (1974) notes that variations in male behavior

can affect a female's reproductive physiology at many points

in her reproductive life. Perhaps the most intensively

studied of these phenomena concerns the effects of male

behaviors on various aspects of the normal reproductive

cycle of adult females.









As Adler notes (1974) the normal mammalian ovarian cycle

has three sequential phases: 1) the follicular phase when

ova and their follicles grow and mature, 2) the ovulatory

phase when the mature follicle releases the egg and 3) the

luteal phase when the ruptured follicle becomes a corpus

luteum which secretes progesterone which aids in the

maintenance of pregnancy. Conaway (1971) notes that there

is considerable variability among mammalian species in the

lengths of the various phases of the ovarian cycle, their

occurrence and the kinds of stimuli necessary to elicit

or modify them. In some species, for example, the

ovulatory phase only occurs when the female receives

copulatory stimulation. These animals are referred to as

induced ovulators. Prairie voles (Microtus ochrogaster) and

montane voles (Microtus montanus) are induced ovulators

(Gray, Davis, Zerylnick & Dewsbury, 1974; Richmond a Conaway,

1969). In other species, such as laboratory rats, ovulation

occurs spontaneously. Similarly, the luteal phase may

be spontaneous, as in humans, or induced, as in laboratory

rats. While it has been known for quite some time that

copulatory behavior was instrumental in initiating some

of these physiological changes, the exact nature of the

stimuli remained obscure. A series of experiments by Adler

and his colleagues was designed to analyze the effects

of variations in male laboratory rat behavior on successful

pregnancy in the female by affecting her reproductive

physiology.









Male laboratory rats typically have a number of mounts

and mounts with vaginal penetration, called intromissions,

prior to ejaculation. An organized group of such mounts and

intromissions, terminating in ejaculation is termed an

ejaculatory series. Male laboratory rats normally display

several such ejaculatory series prior to sexual satiation

(that is, a period of 30 minutes with the same female

without an intromission or ejaculation).

Wilson, Adler and Le Boeuf (1965) found a correlation

between the amount of copulatory stimulation delivered by

sexually vigorous males and the probability of pregnancy

in the females receiving -this stimulation. One group of

females was allowed to mate undisturbed until the males

attained their first ejaculation, usually after about nine

intromissions. The other group was mated with males which

only had three or fewer intromissions prior to their first

ejaculation. Ninety percent of the females allowed a full

complement of intromissions became pregnant while only

22% of the females receiving three or fewer intromissions

became pregnant. Wilson et al. thus demonstrated that

multiple intromissions prior to ejaculation (not just the

ejaculation itself) were necessary to initiate normal

pregnancy in young virgin female laboratory rats. Adler

(1969) later replicated these results using a different

strain of rats.

Apparently, the male's behavior functioned in two ways

to induce normal pregnancy. First Adler (1969) found that









a sufficiently high number of preejaculatory intromissions

was necessary for normal sperm transport. Those females

having only one intromission prior to ejaculation had no

sperm in their uteri 1-3 hours after copulation while those

that had 2 or more intromissions had high numbers of sperm

in their uteri. Chester and Zucker (1970) later extended

these results and found that only 50% of the females

receiving one ejaculation on the male's first insertion had

sperm in their uteri 1/2 4 hours later. The biological

importance of such rapid sperm transport vwas demonstrated

by Adler (1968, 1969). He found that those females

receiving high numbers of intromissions prior to ejaculation

had normally developing embryos 3 days after copulation

while those receiving low numbers of preejaculatory

intromissions had unfertilized and degenerating ova in their

fallopian tubes. Thus multiple intromissions were necessary

for normal sperm transport and fertilization. It should

also be noted that the ejaculatory reflex and the hard

coagulated copulatory plug delivered by the male at

ejaculation were also found to be necessary for normal sperm

transport (Blandau, 1945).

The second way in which the male's behavior was found

to influence the initiation of pregnancy in the female was

through the induction of the luteal phase of the cycle. As

noted earlier, the luteal phase is induced in laboratory rats

and is essential for the maintenance of normal pregnancy

in the female. The progesterone secreted by the corpus









luteum is essential for preparing the uterus to receive the

fertilized embryo and allowing implantation. Adler (1969)

showed that 100% of the females given a high number of

intromissions showed a cessation of behavioral receptivity

while only 18% of the females given three or fewer

intromissions showed a cessation of behavioral receptivity.

This implied a release of gestational hormones in response

to the high levels of stimulation which in turn caused

the cessation of behavioral receptivity. In a further

elaboration of this work, Adler (1969) gave varying numbers

of intromissions to females but did not allow them to

receive an ejaculation. His data show that the proportion

of females ceasing behavioral receptivity increased as a

function of increased copulatory stimulation. Adler (1968)

further showed an increased incidence of pregnancy among

females inseminated following reduced numbers of intro-

missions if an injection of exogenous progesterone were

given following copulation. Finally, Adler, Resko and Goy

(1970) measured circulatory levels of plasma progestins

directly in females given varying amounts of copulatory

stimulation and found a direct correlation between the

amount of copulatory stimulation and levels of progestin.

All of these data confirm the importance of multiple

intromissions in the induction of the luteal phase and the

subsequent production of progestins, which is in turn

essential for the maintenance of normal pregnancy. Although

the mechanism is not yet fully understood, experimental









evidence reviewed in Adler (1974) suggests that the induction

of the luteal phase occurs as a result of a neuroendocrine

reflex which stimulates corpora lutea growth and thus the

secretion of progestins.

All of the above evidence suggests that the frequency

of intromissions is important in induction of normal

pregnancy in laboratory rats. Work by Adler and Zoloth

(1970) further suggests that the timing of copulatory

events in laboratory rats may also.be important. They

found that if ejaculations were followed too closely by

additional intromissions or by manual cervical stimulation,

sperm transport and presumably, pregnancy could be disrupted.

Thus the patterning of post-ejaculatory events is also

important in induction of normal pregnancy.

The work of Adler and his coworkers shows that a single

normal ejaculatory series provides the necessary and

sufficient stimulation for induction of normal pregnancy.

However, laboratory rats typically have multiple ejaculatory

series. If one ejaculation normally provides all the

stimulation necessary for normal pregnancy, what then is the

biological function of multiple ejaculatory series? Recent

work by Davis (1974) provides one possible answer. Davis

found that while one ejaculatory series was indeed

sufficient to induce normal pregnancy in young, virgin

female laboratory rats, this was not the case in older

multiparious females. Davis found that none of the older

females receiving one complete ejaculatory series became









pregnant while 66% of those receiving 3 series and 92% of

those receiving five series became pregnant. Moreover,

those receiving five series had significantly more

developing embryos in their uteri than those receiving only

3 series. Because none of the older females receiving

just one series showed a cessation of regular estrous

cycles, it was suggested that the reproductive failure of

these females was due to inadequate induction of luteal

function. Subsequent measurements of plasma progestins

in older females receiving varying numbers of ejaculations

supported this suggestion. It appears therefore, that

multiple series delivered by a male could extend the

reproductive life of aging females and increase the number

of offspring produced, thus increasing the fitness of those

males delivering multiple ejaculations.

Copulations beyond the first ejaculation also appear

to be important in the induction of pregnancy in young

females of other species. Dewsbury and Estep (1975) have

demonstrated that only 5% of the cactus mice (Peromyscus

eremicus) females given just one ejaculatory series become

pregnant. Forty-seven percent and 60% respectively of the

females allowed to mate to satiety and to satiety plus

pairing overnight became pregnant. Thus, copulation beyond

the first ejaculation in this species is necessary to

maximize the probabilities of pregnancy. Work by Lanier et al.

(in press), has shown a similar effect in Syrian

golden hamsters (Mesocricetus auratus). In two experiments








they found that 20% and 40% respectively, of the females

receiving just one copulatory series became pregnant. In

both experiments 100% of the females mated to a 15 min

satiety criterion became pregnant. They also demonstrated a

direct relationship between the number of ejaculatory series

received and the proportion of females becoming pregnant.

They concluded that multiple.ejaculations are essential in

Syrian golden hamsters to maximize the probabilities of

pregnancy.

It should be clear from this review of the literature

that variations in male copulatory behavior can dramatically

effect the reproductive physiology of the female, the

induction of pregnancy and ultimately the fitness of

organisms involved. Experimental investigations of the

biological function of various aspects of the male's

copulatory pattern have suggested possible functions for

various aspects of the patterns such as multiple intro-

missions and multiple ejaculations. In addition species

differences have been found in the function of similar

aspects of the behavior. Laboratory rats and Syrian golden

hamsters have multiple intromissions and multiple

ejaculations; however, multiple ejaculations maximize the

probability of pregnancy in young female hamsters while

they do not in young laboratory rats. As mentioned

previously, experimental investigations of the biological

function of patterns of copulation carried out in conjunction

with basic descriptive analyses provide an added level of









analysis to comparative investigations of copulatory

behavior. They also facilitate the study of the adaptive

significance of variations in copulatory behavior and

contribute to our general understanding of the evolution

of behavior and reproductive systems. Few such descriptions

and experimental analyses exist, however, even for species

of muroid rodents. Such a detailed description of the

copulatory behavior and an experimental analysis of the

biological functions of the behavior seems particularly

warranted for one species of muroid rodent the roof rat

(Rattus rattus). Roof rats are closely related to the

common laboratory rat and are very common, living in close

association with man. However, very little is known of

their behavior including their copulatory behavior.

The Natural History of Roof Rats


According to Simpson (1945), the roof rat or black rat

(Rattus rattus L.) is a member of the order Rodentia, super-

family Muroidea, family Muridae and subfamily Murinae.

Simpson recognizes 68 different genera within the Murinae,

one of which is the genus Rattus. According to Ellerman

(1941), there are over 554 forms in the genus, many of which

are subspecies. Walker (1964) notes that this makes the

genus Rattus the largest genus of all mammals. Most of the

284 distinct species recognized by Ellerman are tropical in

distribution and are found chiefly in southeast Asia and

Africa (Walker, 1964). Two species have gained world wide








distribution: the Norway rat, Rattus norvegicus, and the

roof rat, Rattus rattus. Norway rats tend to be a more

temperate species and are thought to have originated in the

area around the Caspian Sea, whereas roof rats are

considered to be a more tropical species originating in

Indo-Asia (Robinson, 1965).

According to Walker, Robinson and others, both roof

rats and Norway rats gained their global distributions

through their associations with men. Roof rats supposedly

were first brought to Europe during the crusades by

Europeans on their way back from the Middle East. Roof rats

brought with them the plagues which ravaged Europe during

the 12th, 13th and 14th centuries. The Norway rat

supposedly did not get to Europe until approximately the

16th century. In northern climates the newly introduced

Norway rat quickly displaced the smaller, less vicious roof

rat. In more tropical climates the displacement of roof

rats by Norway rats has been much slower. According to

Walker (1964) roof rats reached North America with the first

explorers, while Norway rats did not reach North America

until the 18th Century.

Roof rats and Norway rats are very similar morphologically

but can be distinguished by several reliable keys.

According to Hall and Kelson (1959) Norway rats are larger

in body size than-roof rats. Norway rats have 12 teats

while roof rats have only 10. Roof rats have tails longer

than their entire bodies while Norway rats have tails equal
to or shorter than their body lengths.








Roof rats and Norway rats often exist as commensals of

man, living in barns, grain storage areas and human

habitations. Roof rats tend to be arboreal, nesting in the

roofs of houses and barns, while Norway rats tend to nest

in burrows or in protected nest sites on the ground.

According to Telle (19661 and Ewer (1971) roof rats form

complex social groups with a linear male hierarchy or

dominant male type of social organization. Both Telle and Ewer

report that group territories are actively defended against

intruders. Barnett (1963), Ewer (1971) and Telle (1966)

provide some information on other aspects of the behavior

of this species such as foraging, aggression, parental care,

maintenance activities and other social interactions.

Barnett and Telle also discuss interspecific interactions

between roof rats and Norway rats. According to Barnett

(1963) roof rats and Norway rats do not hybridize in the

laboratory or the field.

The Present Research


Ewer (1971) has observed the copulatory behavior of both

captive and free ranging roof rats. She describes their

behavior as conforming "...in general with what has been

described by Barnett in Rattus norvegicus... [p.155]". Her

description suggests that roof rats have a pattern

characterized by multiple intromissions prior to ejaculation,

no lock and probably multiple ejaculations prior to sexual

satiety. This is indeed similar to the pattern shown by








Norway rats. While Ewer provides data on the events

preceding copulation, postures and motor patterns assumed

during copulation and a rough description of the copulatory

pattern itself, much remains to be done to provide a

detailed quantitative description of the copulatory

behavior of this species. If the maximal information is to

be gained from interspecific comparisons the copulatory

pattern should be described in as much detail as possible.

Without this added detail, conclusions about the adaptive

significance cf variations in copulatory behavior must

remain tenuous and preliminary.

The research reported here consists of two experiments.

The first provides a quantitative description of the

copulatory behavior of roof rats. The second provides an

analysis of the stimulation essential for initiation of

successful pregnancy in this species. In addition to

providing data that will facilitate interspecific comparisons

and aid in our understanding of the adaptive significance of

variations in patterns of copulation, these data might also

provide some insight into possible behavioral isolating

mechanisms between roof rats and Norway rats. As noted

earlier the two species are occasionally sympatric, at

times living in the same building, but never hybridizing.

It is possible that one of the mechanisms preventing

hybridization involves differences between the species in

their copulatory behavior. If the differences were

substantial enough, they might prevent the induction of


~ _






23


pregnancy when heterospecifics did happen to mate thus

preventing successful hybridization.














EXPERIMENT 1


While detailed descriptions of copulatory behavior

exist for a variety of muroid rodents including laboratory

rats (Dewsbury, 1972b), no such detailed quantitative

description exists for the common roof rat. This experiment

was designed to provide such a description. Provided are an

analysis of the basic motor patterns, a qualitative and

quantitative, description of the pattern of copulation, and a

description of the behaviors accompanying copulation

including ultrasonic vocalizations. Such detail should

facilitate interspecific comparisons and provide a framework

for experimental analyses of the adaptive significance of

various aspects of the copulatory behavior of this species.


Materials and Methods


Subjects

All of the subjects of Experiments 1'and 2 were the

first generation laboratory reared offspring of wild trapped

animals. The parents of these subjects were 11 males and 8

females all live-trapped as adults at the University of

Florida Poultry Farm, southwest of the main campus'of the

University of Florida. Several of these original specimens

were positively identified as Rattus rattus L. after close








morphological examination by Dr. Steven Humphrey of the

Florida State Museum.

The subjects for Experiment 1 were 12 male and 12

female laboratory reared offspring of these wild trapped

animals. All the subjects were weaned at 30 days of age and

housed in isolation in Wahman suspension cages measuring

18 cm x 17.5 cm x 24.5 cm. All animals received Purina

Lab Chow and water ad lib. The colony was housed in an air

conditioned room maintained at approximately 21 C. A

reversed light-dark cycle of 16 hr light, 8 hr dark was

enforced throughout the experiment with light onset occurring

at 1800 hr daily. One 25 watt red light bulb illuminated

the room at all times.


Apparatus

Animals were tested in clear plastic cages measuring

38 cm x 20 cm x 48.5 cm covered with stainless steel cage

tops made of 2 mm diameter rods. San-i-Cel brand litter

material covered the bottom of the cages and served as a

substratum for testing. Behavioral events were recorded

on a 20 channel Easterline-Angus event recorder and by hand

with written descriptions when necessary. Ultrasonic

vocalizations were detected with the aid of a Holgate

Mark V ultrasonic receiver and a Telequipment Model S54A

Oscilloscope. The oscilloscope was calibrated at '50 ,sec/

cm and 0.5 volts/cm. The ultrasonic receiver was calibrated

with the aid of a frequency generator and was found to be

accurate within + 5 kHz in a range from 20 kHz to 100 kHz.








The ultrasonic receiver is a superheterodyne instrument

that transduced the ultrasonic input into an audible output.

This output was monitored directly with the use of head-

phones and on occasion the audible output was also fed into

the oscilloscope and visual representations of the signal

were observed.


Procedures

All of the males used in this experiment had been

pretested for copulatory behavior at least once prior to the

start of formal testing and proved to be reliable copulators.

Pretesting consisted of pairing a male with a female in

natural estrus and allowing him to mate uninterrupted for

at least one hr. Males had not been allowed to explore

the apparatus prior to protesting. Females were not pre-

tested for copulatory behavior prior to formal testing but

all had been permitted to explore the test apparatus on at

least one occasion of 20-30 min. All of the animals

were between 130 and 180 days of age at the start of the

experiment and between 210 and 260 days of age at the time

of their final test.

Animals were tested in the colony room approximately

3 hrs after the beginning of the dark phase of the diurnal

cycle. During the experiment four 25 watt red light bulbs

illuminated the testing cages. Males were introduced into

the cages 5 to 10 min before the start of testing and

allowed to habituate to the situation.

Tests were started with the introduction of the female
into the testing cage. Males were given 30 min from the








start of the test to initiate copulation. On tests where

the male did not achieve an intromission within this time

period, the test was terminated and scored as a "negative".

On tests where the male did initiate copulation, the test

was allowed to continue until 30 min had elapsed since

the male achieved his last intromission or ejaculation.

At this time the test was terminated and scored as a

"positive". Each male was tested with the same female at

two wk intervals until each pair had achieved six positive

tests or had received four consecutive negative tests.

On the fifth and sixth positive tests for each pair a

categorization was made of the behaviors accompanying

copulation, and the occurrence of ultrasonic vocalizations

in the 22-33 kHz frequency range was noted. On those tests

where ultrasonic vocalizations were to be observed, the

Holgate ultrasonic receiver was utilized, tuned to 28 kHz

and the microphone was placed in the center of the food tray

of the cage top.

All females were brought into behavioral estrus with

intramuscular injections of 0.1 mg estradiol benzoate

followed by 1.0 mg of progesterone given approximately 75 hrs

and 5-6 hrs before testing, respectively.


Behavioral Measures

Pretesting had indicated that the copulatory pattern of

roof rats was essentially the same as that of Norway rats.

That is copulations were found to be organized into groups

or series of. mounts, mounts with vaginal penetration









(intromissions) and intromissions terminating in ejaculation

ejaculationss). Each, such organized series of copulatory

behaviors usually ended with the occurrence of an ejaculation.

Males have a number of such ejaculatory series before

reaching sexual satiety. Accordingly, the standard measures

of copulation for this species were adopted from those

developed by Beach and Jordan (1956) for laboratory rats.

These standard measures of copulation were:

ML Mount latency. The time from the start of testing

until the first mount with pelvic thrusting (with or without

intromission).

IL Intromission latency. The time from the start of

testing until the first vaginal intromission by the male.

MF Mount frequency. The number of mounts with pelvic

thrusting but without intromission within an ejaculatory

series.

IF Intromission frequency. The number of intro-

missions within an ejaculatory series.

EL Ejaculation latency. The time from the first

intromissions of a series to the ejaculation of the series.

MIII Mean interintromission interval. The mean

interval separating the intromissions of a series. This

is calculated by dividing EL by IF.

PEI Postejaculatory interval. The time from the

ejaculation of one-series until the resumption of

copulation as indicated by the first intromission of the

next series.








It should be noted that the measures IF, EL and MIII

are not completely independent of each other for a given

ejaculatory series (EL = IF x MIII).

On those tests where ultrasonic vocalizations were to

be observed, several measures of vocalizations were taken.

The postejaculatory ultrasonic vocalizations of roof rats

were found to be similar in pattern to those of Norway rats

and therefore the parameters of postejaculatory vocalizations

developed by Barfield and Geyer (1975) were utilized for roof

rats. These measures were:

VL Vocalizations latency. The time from the

ejaculation until the beginning of the ultrasonic

vocalizations.

VT Vocalization termination. The time from the

ejaculation until the end of the ultrasonic vocalizations.

In addition two derived measures were taken:

PEI-VT An estimate of the time spent in ultrasonic

vocalization. Calculated by subtracting the VT for a given

copulatory series from the PEI in which it occurred.

VT/PEI The proportion of the PEI spent in

ultrasonic vocalization. Calculated by dividing the VT for

a given series by the PEI in which it occurred.

After preliminary observations of the copulatory

behavior of roof rats, a system was developed for the

categorization of the behaviors accompanying copulation

into several mutually exclusive categories. The occurrence

and duration of each of these categories of behavior were









recorded on an Esterline-Angus event recorder and later

decoded to generate a profile of the behaviors occurring

during the various phases of the copulatory sequence. Using

this system the behaviors accompanying copulation were

categorized on the fifth positive test for each male and on

the sixth positive test for each female. The categorization

was initiated at the start of the test and terminated at the

end of the 30 min satiety criterion. The behavioral

categories were adapted from those of Dewsbury (1967; 1970;

1971; 1972a; 1973b; 1974a; 1974b) and are listed below. All

behaviors are common to both males and females unless

otherwise noted.

Mount (Males Only) Mounting, intromitting or

ejaculating by males.

Lordosis (Females only) the posture taken by females

when males mount. It involves a stance in which the female

plants her feet firmly, straightens her hind legs and lifts

her head forming an inverted arch in her back.

Chase (Males only) Following or chasing the female

orienting towards her at all times and usually culminating

in a mount of the female.

Run From Male (Females only) Involves the female's

darting away from the male or running from a pursuing male.

Simultaneous ear wiggling and tail rattling behaviors may

also occur.

Sniff Involves an approach or orientation toward the

partner in which the animal appears to sniff the body of

the partner.









Upright Defense Occurs when the animal stands upright

on its hind legs orienting the ventrum towards the partner.

Sometimes involves pushing or slashing of the partner with

the forepaws and/or audible vocalizations.

Genital Groom Self Scratching, licking or manipulation

of the animal's own genital area.

General Groom Self Scratching, licking or manipulation

of the animal's own body with the exception of the genital

area.

Genital Groom Partner Scratching, licking or

manipulation of the partner's genital region.

General Groom Partner Scratching, licking or

manipulation of the partner's body with the exception of the

genital region.

Groomed A sitting posture adopted by an animal while

being groomed by its partner. Ears are usually back and

eyes are closed.

Climb Climbing or hanging upside down on the cage

top of the testing cage.

Pull Female (Males Only) Slashing at or pulling on a

female which is climbing or hanging on the cage top.

Feet on Walls A posture in which the animal stands

upright and leans against the sides of the cage with its

forepaws.

Dig Digging-or pushing the substratum with.the paws.

Locomotor-Explore Walking movements sometimes

accompanied by nosing of the substratum, walls or cage top,








mouthing or manipulation of the substratum or feces, and

upright postures other than those listed elsewhere.

Sit Adoption of a sitting posture by the animals

involving little movement other than slight mouth and head

movements.

Lie Lying down. Self explanatory.


Results and Discussion


Basic Motor Patterns of Copulation

The basic motor patterns of copulation of roof rats

were found to be almost identical to those of laboratory rats.

Male roof rats mounted females several times prior to

ejaculation. The male typically palpated the female's sides

vigorously while making numerous rapid shallow pelvic

thrusts. On some of these mounts the male gained vaginal

penetration intromissionn), characterized by a deep pelvic

thrust followed by a rapid dismount; on other mounts the

male dismounted without achieving intromission (termed

mounts). Usually after several mounts and intromissions

the male mounted, achieved intromission and ejaculated. The

ejaculatory intromission was characterized by a much deeper

pelvic thrust and a much slower dismount. The ejaculation

was always followed by a period of sexual inactivity known

as the postejaculatory interval (PEI). An organized chain

of copulatory events culminating in an ejaculation (and

including the following PEI) is known as an ejaculatory series.

Male roof rats were found to achieve an average of 4.3 such

ejaculatory series before reaching sexual satiety.








On three of the 302 ejaculatory series observed, three

different males were observed to ejaculate on a single

insertion, that is with no intromissions preceding the

ejaculation. This pattern, although rare in roof rats, has

never been observed in normal Norway rats. It is possible

therefore that there is a fundamental difference between the

two species with regard to this ability to ejaculate on a

single insertion.

In a very strict sense roof rats are categorized as

pattern #15 in Dewsbury's taxonomy of copulatory behavior

(Dewsbury, 1972b) because they show no copulatory lock, no

intravaginal thrusting, ejaculations can be achieved on a

single insertion and because they typically attain multiple

ejaculations. In that the phenomena of ejaculation on a

single insertion was observed so rarely in roof rats, it

might also be reasonable to classify roof rats as pattern #

13 animals. This pattern characterizes species such as

Norway rats (Beach & Jordan, 1956), Syrian golden hamsters

(Beach & Rabedeau, 1959), and Mongolian gerbils (Kuehn 6

Zucker, 1968) which have no lock, no intravaginal thrusting,

multiple ejaculations and multiple intromissions always

preceding ejaculation. The final resolution of this

problem of classification awaits further data on other

populations of roof rats. If the:phenomenon of ejaculation

on a single insertion proves to be very rare, then it would

probably be more reasonable to classify roof rats as pattern

#13.








Ewer's (1971) description of the copulatory behavior of

a laboratory colony of roof rats and of a free-living

population indicates that the animals she observed also

displayed pattern #13 copulatory behavior. It is clear from

her'description of the behavior of the laboratory colony that

the animals she observed had multiple intromissions and

probably multiple ejaculations.' No mention is made of the

presence of a lock or of intravaginal thrusting. She also

notes that while complete "sequences" of copulatory behavior

were not observed in the free-living population "All stages

were, however, seen a number of times on different

occasions" [p. 156].

When mounted by a male, the female roof rats demonstrated

a lordotic posture similar to that displayed by female

laboratory rats. The lordotic posture was characterized by

an extension of the hind legs and elevation of the head

accompanied by a spinal flexure resulting in an inverted arch

in the.female's back. The tail was deflected-laterally.

Prior to mounting, receptive females often displayed a

series of behaviors including rapid ear wiggling, tail

rattling and approaches to the male followed by rapid darting

away and stopping several cm away. These behaviors were

observed on every test in which the females were judged

behaviorally receptive and were not observed in the few

tests where the females were not behaviorally receptive.

Females were judged behaviorally receptive solely on the

basis of their responses to mounting by males: a behaviorally








receptive female adopted a lordotic posture in response to

male mounting. It is of interest to note that Ewer (1971),

observed the patterns of darting and tail rattling in free

ranging wild Rattus rattus females during copulatory

sequences. Ewer makes no mention of the pattern of ear

wiggling observed in the present population of animals.

Darting and ear wiggling have been reported numerous times

for receptive female laboratory rats (Beach, 1942; 1943;

1956) but tail rattling has never been reported for this

species.


Quantitative Measures of Copulatory Behavior

Quantitative data for the copulatory behavior of roof

rats were collected on 71 tests from twelve pairs of animals.

Each pair was to receive six tests each carried to a 30

minute satiety criterion; however one male died prior to

his last test.

Table 1 presents the means and standard errors of the

mean for the standard measures of copulatory behavior for

the first three and the last three ejaculatory series.

Standard errors were computed using a formula derived by

Marks (1947). This formula was developed for those

situations in which there are unequal numbers of tests for

different subjects. As can be seen from Table 1, it took

male roof rats an average of approximately 1.5 min to

achieve their first mount and a little over two minutes to

achieve their first intromission from the start of the test.




























V)
41




U) M




rd -








C

2 0)


4) U)


Uf)






W Q


cc




5-,











0 0-4





0
uti
er
ni!-
*T3 0
Ci-H v










F1-1









u~c


o1


--


t-4







A S
H tn



mda


'0 0 -0

o Co M
-4 N ,


+1 +1 +1





i-4




c'3 cd
l 1- p.
-4 )-


N 0







Ln
00






+1 +l +1



,04
+1 +1 +1




\0
N





























C D
0 0 Co


+1 +1 +1

i) *Gj* l-


N
N







C

+1 +1 +1

0 0 I)

1M \0
\0



r--




+I +1 +1



N 00








r-
-i

+1 +1 +1

,-. '0 l -


U


4.




,CM
I M
-4


r-l
+il












\ O
- N


+1 +1

V) N

N M

\*O

















































t0 L0
+1 +i





I)









+1 +I














+1 +i



C> CO












'-4
+i +i

M N


'-4 -4 n

~ H 0-








Once started, roof rats typically attained a mean of 4.3

ejaculations before reaching sexual satiety. Sexual satiety

was attained between series on 46 of 71 tests or 64.8% of

the time. This is to say, satiety was reached after the

completion of an ejaculation and without the animal

initiating another series, on 64.8% of the tests. On 35.2%

of the tests, males initiated but did not complete an

ejaculatory series before reaching sexual satiety. During

their first copulatory series, male roof rats had a mean of

5.1 mounts and a mean of 7.6 introrissions prior to

ejaculation. The mean interval between intromissions was

found to be a little over one min while the mean

ejaculation latency was somewhat over nine min in length.

The mean postejaculatory interval following the first series

lasted approximately four min.

Analyses of variance were performed for all eight of

the standard measures of copulatory behavior to assess the

effects of repeated testing on these measures. Those

measures that occurred more than once per test (i.e. MF, IF,

EL, MIII and PEI) were subjected to different analyses than

those that did not (ML, IL, EF) in order to assess any

changes that might occur across series as well as across

tests.

Male roof rats were found to be remarkably variable in

the number of ejaculations that they might attain on any

given test. Males were found to have as many as eight

ejaculations or as few as one ejaculation prior to reaching










sexual satiety. This variability in the number of series

made analysis of data from all series for all tests on all

subjects impossible. Therefore, only those animals having

three or more series on their first four tests were included

in the statistical analysis. Eight pairs of animals met

this criterion and it is on the data from these animals that

the following statistical analyses were performed.

The results of these analyses of variance are

summarized in Table 2. Complete ANOVA tables for each

measure appear in Appendix A. One way analyses of variance

for repeated measures were performed on the measures ML, IL

and EF to assess the effects of repeated testing on these

measures. As the results in Table 2 show, there were no

significant test effects for any of these measures. It can

therefore be concluded that no significant changes occurred

with regard to these measures over the first four tests.

For the measures MF, IF, EL, MIII and PEI, two way

analyses of variance for repeated measures were performed.

The results of these analyses also appear in Table 2. On

no measure was a significant test effect found. Significant

changes across series were noted for the measures IF, EL and

PEI but not for the measures MIII and MF. None of the tests

by series interactions proved to be significant.

The Newman-Keuls method (Winer, 1974) was applied to

the measures IF, EL and PEI to determine which series were


CIC~









Table 2


Results of Analyses of Variance of
Eight Measures of Copulatory Behavior


Behavioral
Measure Source df F p

a
ML tests 3,21 2.30 NS

IL tests 3,21 1.33 NS

EF tests 3,21 0.96 NS

MF tests 3,21 1.45 NS

series 2,56 1.42 NS

tests'x
series 6,56 0.73 NS

IF tests 3,21 1.32 NS

series 2,56 58.40 <.001

tests x
series 6,56 0.87 NS

EL tests 3,21 0.19 NS

series 2,56 30.08 <.001

tests x
series 6,56 0.64 NS

MIII tests 3,21 0.21 NS

series 2,56 2.46 NS

tests x
series 6,56 0.76 NS

PEI tests 3,21 1.81 NS

series 2,56 68.42 < .001

tests x
series 6,56 1.11 NS


a
NS = Not significant






40







Table 3


Results of Newman-Keuls Analysis for Three
Measures of Copulatory Behavior


Series


Mean and
Standard Error


406.9

84.9

154.8




249.6

283.1

341.6


Comparison


45.6

6.5

33.4


10.6

11.6

15.6


vs 2

vs 3

vs 3


r q r


13.95

12.40

1.55




10.42

8.16

2.26




0.60

16.34

10.41


Note. df =

a


56 for all comparisons


Time Measures in .Seconds


p < .01


Behavioral
Measure


1

IF 2

3


1
a
EL 2

3


1
a
[ 2

3








significantly different from the others for each of these

measures. The results of these analyses appear in Table 3.

For IF and EL it was found that the means for the first

series were significantly larger than the means for the

second series and the third series. Means for series two

were not significantly different from means for series

three for either measure. For PEI, the mean for the third

series was significantly greater than the means for the

first and significantly greater than the means for the

second. The mean PEI of series one was not significantly

different from the mean PEI of series two.

It can be concluded from these results that no

significant changes as a function of test occurred over

the first four tests in any of the eight standard measures

of copulatory behavior. Significant changes across the

first three series can be seen only for the measures IF, EL

and PEI. For the measures IF and EL significant decreases

are seen from first to later series. For PEI the third

series shows a significant increase over the first two.

The changes in the standard measures of copulatory

behavior across series are of considerable interest. Several

studies have dealt with the changes in copulatory behavior

of laboratory rats as sexual satiety is approached (Beach &

Jordan, 1956; Dewsbury, 1968; Karen & Barfield, 1975). This

research was generated primarily by an interest in the

nature of the underlying mechanisms responsible for such

changes. It is therefore of some theoretical interest to









compare the changes in the copulatory behavior of laboratory

rats and roof rats as satiety is approached. The

statistical procedures performed previously do not allow for

an adequate analysis of changes in copulatory behavior as

satiety is approached because only the first three series

were analyzed and on many tests the animals completed more

than three copulatory series. The means and standard errors

presented in Table 1 provide some indication of changes

that occurred in various measures as male roof rats

approached sexual satiety. With this method of data

presentation means and standard errors were determined for

the first three series for all subjects. Then the means and

standard errors were calculated for the last complete

copulatory series (or Nth series) regardless of whether it

was the eighth or the first series. In a similar fashion

the next to last series (N-1) and the second from the last

series (N-2) were computed. This system does not accurately

portray the changes across series for those tests where

considerably more or fewer than six series are needed to

reach satiety. For example, on a test where only two series

were completed before satiety was reached, data from those

two series were represented twice as the first and second

series respectively and as the N-1 and N series, respectively.

For a test where eight series were completed the first

three series were represented in series one, two and three,

respectively and the last three series were represented in

the series labeled N-2, N-l and N. The middle two series


__








would not be represented at all. Given the considerable

variability shown in ejaculation frequency across tests for

roof rats another method of data presentation was employed

for the analysis of changes in behavior as satiety was

approached. This method was originally developed by

Larsson (1956) and later refined by Karen and Barfield (1975)

for use with laboratory rats. The essence of the system

involves plotting the changes in behavior across series as

a function of the number of completed series obtained prior

to sexual satiety. Means were calculated for all tests from

all animals that had only two series prior to satiety, all

tests where the animals had only three series prior to

satiety and so on until curves were generated for all tests

where two, three, four, five and six ejaculations occurred

prior to satiety. No curves were generated for those

tests in.which one, seven or eight ejaculations were

obtained, there being too few such tests to construct

meaningful curves. Finally, these separate curves were

plotted on a variable axis, thus allowing direct comparison

of all initial and terminal points for each measure regardless

of the number of ejaculations received. Figures 1 through

4 present these data for the measures IF, EL, MIII and PEI,

respectively.

Figure 1 reveals that there was a clear decline in IF

from first to later series. In addition, those tests where

three, four, five and six ejaculations occurred prior to

satiety showed a small but consistent rise in IF during the

terminal series.

























Figure 1. Intromission frequency across successive series
for those tests in which 2, 3, 4, 5 or 6 ejaculations
preceded sexual satiety. For the group having 2 ejaculations,
n = 10, for the group having 3 ejaculations, n = 14, for the
group having 4 ejaculations, n = 16, for the group having
5 ejaculations, n = 14, and for the group having 6
ejaculations, n = 8.





45












I0-
-I IF
'"..



--*



n 4-




2 3 f4
2 \32





"2 3 4 5
2 3 4
I 2 3
I 2
Series



























Figure 2. Ejaculation latency across successive series for
those tests in which 2, 3, 4, 5 or 6 ejaculations preceded
sexual satiety. For the group having 2 ejaculations, n = 10,
for the group having 3 ejaculations, n = 14, for the group
having..4 ejaculations, n = 16, for the group having 5
ejaculations, n = 14, and for the group having 6
ejaculations, n = 8.











(1050)"-..


'. EL (Sec)
'*
'OO ,
.


I I


3
3
2
2


Series


I I


4I '5 6
4 5
3 4
3
2


600-



500-

(,D
^C-

) 400


E
- 00



200-



100-


I I I _



























Figure 3. Mean interintromission interval across successive
series for those tests in which 2, 3, 4, 5 or 6 ejaculations
preceded sexual satiety. For the group having 2 ejaculations
n = 10, for the group having 3 ejaculations, n = 14, for the
group having 4 ejaculations, n = 16, for the group having 5
ejaculations, n = 14, and for the group having 6
ejaculations, n = 8.








(.. 200)


MIII (Sec) .
p3




/ *5




6 /
.\.\\i.// jl


I 2 3 4 5 6
I2 3 4 5
2 3 4


Series


AN


140


120


100


80-


60-


40-


20-




























Figure 4. Postejaculatory interval across successive series
for those tests in which 2, 3, 4, 5 or 6 ejaculations
preceded sexual satiety. For the group having 2 ejaculations
n = 10, for the group having 3 ejaculations, n = 14, for the
group having 4 ejaculations, n = 16, for the group having 5
ejaculations, n = 14, and for the group having 6
ejaculations, n = 8.
















PEI (Sec)










,o6

(4 5
02 -
C3 -----
a,- -A


I 2' 3 4 '5 6
I 2 3 4 5
S2 3 4
S2 3
I 2
Series


800

U)
600
E

400








The graphs for EL in Figure 2 and for MIII in Figure 3

reveal very similar U-shaped functions for most tests. For

the tests in which three, four, five and six ejaculations

occurred the EL showed a decline from first to middle series

and then an increase in the terminal series. For those tests

in which only two ejaculations took place, EL declined from

first to last series. The measure MIII is seen in Figure 3

to have declined from first to middle series and then

increased rapidly in the later series for those tests in

which three, four, five and six ejaculations occurred. For

those tests in which only two ejaculations occurred the MIII

increased from first to last series.

While the measures IF, EL and MIII appeared to

display more or less U-shaped functions the measure PEI

seen in Figure 4 appeared to have an entirely different

function. As satiety was approached, the PEI tended to

increase across successive series, rather than decrease and

then increase.

It had been noted earlier that with regard to the basic

motor patterns of copulation, roof rats were found to be

very similar to laboratory rats. Roof rats also appear to

be quite similar to laboratory rats with regard to the

quantitative measures of copulatory behavior. Table 4

compares the present quantitative data for roof rats with

those from three different strains of laboratory rats in two

different studies by Dewsbury (1968; 1975b). The measures

compared are ML, IL, MF, IF, EL, MIII and PEI from the first














Table 4


Comparisons of the Standard Measures of Copulatory Behavior
Between Roof Rats (Rattus rattus) and 3 Strains of
Laboratory Rats (Rattus norvegicus)

Behavioral a b
Measure Roof Rats Laboratory Rats

c c d
LEW F344 Long-Evans


ML 98.5 + 18.6 85 239 -------


IL 146.3 + 28.0 191 359 -------


EF 4.3 + 0.4 ------ ----- .7.4


MF 5.1 + 1.2 13.9 :7.4 6.7


IF .7.6 + 0.7 11.9 9.8 10.8


EL 563.4 + 118.5 1057 607 687


MIII 70.3 + 9.3 95 65 62.6


PEI 269.2 + 15.5 533 314 372


a
Data expressed as means and standarderrors
b
Data expressed as means
c
Data from Dewsbury, 1975
d
Data from pewsbury, 1968


__







copulatory series and EF for those strains tested to satiety.

These data were chosen for comparison because they were

collected in the same laboratory in which the roof rat data

were collected and because procedures were much the same

for all three studies. The ML and IL for the LEW strain

were very close to those observed for roof rats while the

other measures differed considerably between the two species.

For the measures MF and MIII both the F344 strain and the

Long-Evans strain appear close to the range of scores for

roof rats. For the measures EL and PEI the F344 strain more

closely approximates the scores shown by roof rats. The

intromission frequency in the first series appears somewhat

lower for roof rats than any of the three laboratory rat

strains presented here.

The ejaculation frequency also appears somewhat lower

for roof rats than for the Long-Evans strain. Comparisons

of EF's from three other studies show that laboratory rats

tested in different laboratories with varying amounts of

prior sexual experience do show some variability in this

measure. Beach and Jordan (1956) tested both albino and

hooded strains and found the average EF to be 6.2 and 6.4

in two different experiments. Dewsbury tested Long-Evans

rats with varying amounts of prior sexual experience and

found the EF to vary between 5.5 and 7.6. Males with very

little sexual experience had lower EF's than those who had

considerable experience. These scores are still somewhat

higher however, than the EF of 4.3 observed for roof rats.







As the data in Table 4 and the above mentioned data for

EF illustrate, there is some variability in the copulatory

behavior of laboratory rats tested in different laboratories.

Some of this variability is due, no doubt, to differences in

housing, maintenance and testing procedures of the animals,

prior sexual experience and other environmental variables.

Some of the variability is also due to genetic differences

between strains of laboratory rats. Several studies have

shown there to be quantitative differences in the copulatory

behavior of different strains of laboratory rats (Dewsbury,

1975b; McLean, Dupeire & Elder, 1972; Whalen, 1961). This

variability within laboratory rats makes it difficult to

assess possible species differences between roof rats and

laboratory rats with regard to the quantitative measures of

copulation. To adequately compare the behavior of these two

species it would be necessary to rear and test representatives

of both species under identical conditions. Given the

strain differences in behavior it would also be necessary to

sample a wide range of domesticated and wild populations of

Rattus norvegicus and to also sample a range of wild

populations of Rattus rattus. Under these conditions it

would be possible to draw firm conclusions about species

differences in the quantitative aspects of copulatory

behavior.

Despite the limitations of comparing the behavior of

different species from different experiments, several points

can be made from the data presented in Table 4. In that the







mean ML and MIII for roof rats fall within the range of

means presented for laboratory rats, it is probable that

the two species do not differ with regard to these measures.

The measures MF, IF and EL show considerable individual

variability for the roof rats tests as can be seen by the

high standard errors. In fact 24% of the IF's, 26% of

the MF's and 30% of the EL's collected for roof rats fall

within the range of scores presented for laboratory rats.

These data suggest that with regard to these three measures

roof rats and laboratory rats may not differ. There is

less individual variability within the measures IL, PEI and

EF. It is possible that the differences seen between the

two species with regard to these measures may be due to

real species differences. It is also possible that the

differences, should they prove to be reliable, are not due to

species differences but to the effects of domestication on

the behavior of laboratory rats. There are at present no

data on the copulatory behavior of wild Rattus norvegicus

and it is possible that wild Norway rats are even more

similar to wild roof rats than are domesticated Norway rats.

Comparisons among wild and domesticated stocks of Norway

rats and wild stocks of roof rats would provide evidence for

this suggestion.

It can be concluded from this comparison of the

quantitative measures of copulatory behavior of roof rats and

laboratory rats that in five of the eight measures compared

there appears to be very little difference between the two








species. With regard to the other three measures, roof rats

and laboratory rats do not appear radically different, but

small reliable species differences might exist.

With regard to changes that occur in the copulatory

behavior of roof rats as exhaustion is approached, a

similar conclusion can be drawn. The curves generated for

IF, EL, MIII and PEI in Figures 1-4 for roof rats are almost

identical in form to those presented by Karen and Barfield

(1975) for laboratory rats. They are also similar in form

to those functions described by Dewsbury (1968) for

laboratory rats using a slightly different method of

presentation and to those presented by Larsson (1956). It

should be noted that Larsson did not in fact conduct

standard satiety tests with his animals but only tested the

animals for a one hour period. Nevertheless Larsson's

functions for the measures IF, EL and PEI are strikingly

similar to the present ones for roof rats. These facts

suggest that similar mechanisms underly the sexual arousal

and exhaustion of roof rats and laboratory rats. Theoretical

speculations concerning the nature of these mechanisms can be

found in papers by Beach (1956), Beach and Jordan (1956),

Beach and Whalen (1959), Dewsbury (1968), Karen and Barfield

(1975), and McGill (1965).


Categorization of Behavior

A categorization of the behaviors accompanying

copulation was made on one test each for 11 males and 11

females. A twelfth pair of animals was not categorized due
to the death of the male prior to categorization.







Figures 5 and 6 present the categorization data for

males and females, respectively. Data from 15 of the

possible 16 mutually exclusive categories of male behavior

are presented in Figure 5. The sixteenth category, Dig, was

never found to occur more than 1% of the time and was

dropped from further analysis. The data in Figure 5 are

presented as per cent of time spent in each category during

the IL, first, second and last EL, first and second PEI

and the 30 minute satiety criterion. All of the periods

are further divided into quarters with the exception of the

IL. The data are expressed as means and standard errors.

During the intromission latency period males spent

about 85% of their time in activities directed at the

female: sniffing, genital and general grooming and chasing.

Particularly frequent were the episodes of sniffing and

chasing. The EL periods were predominated by copulation

related activities such as chasing, mounting, pulling on the

female, sniffing and genital grooming of the female. Males

also spent extended periods of time grooming their own

genitals; an activity highly associated with copulatory

activity. Males spent from 13-23% of their time in locomotor

exploratory behavior and 6-17% of their time sitting. The

majority of the postejaculatory intervals were spent in

activities such as locomotor-exploratory behavior, sitting

and general grooming. Fourteen percent of the time was spent

in genital grooming and only 2-4% of the time was spent

chasing and mounting females. Grooming and sniffing females













0)
u i
0)











m
cd
aC
a






cu
cH )

rd

.0
o *H



-H




P.H
*dC

Sed
0(0






m0 C

o *o


N 0 C)



*rld2











LO cQ
UN N


00
nd 0 N


3 0 (n


l ct





bF NaN




























































311 O .IN33903d















V *
CO







0
3 0

















0
nao






u )
(IS












Um
-C4k





00
0 a0
~H






E .*qO




0c0
.,Q


0 0
t@ 9




0 O-
0C1

cC'



















u) 0







to a
*H 4)1

ou





* U)







CL C cC


















U i


I- 1 1
v, i o_ i <






SI -I I
.-i-, t







SI I ., -



I I
i = ,' i i l i i.
I-, -- I -
I. I -s ">.----'-----





2 ) -.i

- II 0 iNI.



3V^11 dO INJRa~d


b

-

:'-'.


.^-*w




a..


^t.-

..ur







j a.





63


occupied 7-13% of the time. Mounting and chasing both

tended to occur in the final quarter of the PEI just prior

to the next.series. Very little time Was devoted to

copulation-related activities such as chasing and mounting

during the satiety period as compared to the EL periods.

What copulation-related activity did occur. was restricted

to the first two quarters only. About 65% of the period was

taken up with locomotor-exploratory behaviors and sitting.

Figure 6 presents the percent of time females spent in

13 of the 15 possible categories of female behavior. The

data are presented as means and standard errors. The

categories of Dig and Lie were excluded because during no

period did either category occupy more than 1% of the time.

During the IL period females behavior appeared to be directed

at the male. Sniffing, genital and general grooming, being

groomed, upright defense, running from the male and lordosis

accounted for 58% of the period. The remainder of the

period was spent in locomotor-exploratory behaviors, feet

on walls and climbing. 'Females spent from 16-35% of their

time during the EL periods in active copulation or activities

involving the male-including running, lordosis, being

groomed, upright defense and grooming the male. Sitting and

locomotor-exploratory activities (not including climbing)

accounted for 25-53% of the females' time in this period.

Climbing behavior was observed to occur interspersed in

bouts of both locomotor-exploratory and copulatory behaviors.

Often, especially in later series, females would climb on

the cage top in an apparent attempt to avoid the chasing








male. Climbing was observed to occur from 22-38% of the

time during EL periods. The majority of the two PEI periods

were devoted to locomotor exploratory behaviors and climbing.

Only 1-5% of the time was spent running from males or in

lordosis; these behaviors occurred during the fourth quarter

of the periods prior to the next series. During the

satiety period females spent the majority of their time in

locomotor-exploratory behaviors and climbing. Thirteen

percent of the time was spent in sitting. Only 4% of the

time was devoted to running from the male and lordosis;

this occurred primarily during the first two quarters.

Figures 7 and 8 present the categorization data for

males and females in a slightly different fashion but one

that facilitates interspecific comparisons. For the males,

10 of the 15 categories presented in Figure 5 have been

collapsed into six mutually exclusive categories and overall

means calculated for each time period. The category

locomotor-explore includes locomotor-explore, climb and feet

on walls. Chase-mounts includes chase, mount and pull

female. The categories general groom self, genital groom

self, sniff and sit are the same as in Figure 5. For females

10 of the 13 categories in Figure 6 have been collapsed into

nine categories in Figure 8. Locomotor-explore includes

locomotor-explore and feet on walls. The categories run

from male, lordosis, sniff, genital groom self, general.

groom self, climb and sit are the same as in Figure 6.















Cd
O
Fd-



nd

0 co
*H






-4C

4-
S*-






00

uC


- a)

04
0
CD
O





0 H


*H 44
dH
c e


0 O-
0 4-' 0
bO F-.

04Q

r-l $


0 cd
U 4-4VT3






C cd





0 n
Cd



Q.U) n)




oxo

U)

r- c)
i^ (U +->
t-. 3 (

&0 ni *H U






















XI


E






F











U
*I t t I





s u~
1-^









I 1- I




4
0I
















I-)
1 r I I 1-

000000
s_ ) ___ -


mi








I

I I t 1 f


0I

ol


Ln

E
,
-- l.?-:- :4?5.? ??;i











cI~
t i t !


r4









ct


000000
a me-u


auwil 10 lU3OJd


z



_I-
VI



Pd N



u-
UJ

*-j
CL







z
Ud
ci
1-1





SN




Id .
CL
d-
w
_i
r-


i


1
















u
CO









a,
*H 4J
> cd






Ok
05.
0
rd
'-o


o -H
4-J


4 -U
a0
Od U


0
4-4 L-4
0


0 -H
*H 4-
4-4

c C- *-
*H 4- In
4 0
04-1 0



U 4+j4
Cd





cd aQm
' 0C CC
0 x )


O0
rn n

M Cd 4)


C) a)
cP i



04-S
t an a







Lcd aC



< -1

















i I











i f '- f / I / /I I I
E- 1








1 r !
U)
0 0 *
.^0 ^
*O: V*
i < | __ 1 -- - -- -
s ^ ^ZZMQ


n ^ ^ ^^
i^, *if& ^////////^
Y C ^
^ i p r
i l l L l l - l - i -


0 0 0 0 0 0
v0 m- 10 N 10


0 0 0 0 C
awlo 10 u -a:ad
siWij_ in uo-a


) 0 0 0 0 0 0
o o ooo o
uroN-








The collection of categorization data for laboratory

rats by Dewsbury (1967) allows detailed comparisons to be

made with the present data for roof rats. A few qualitative

and quantitative differences in behavior between the two

species are apparent. However, as with the other aspects

of copulatory behavior discussed thus far, the most

striking features of these comparisons are the great

similarities in behavior. The copulation-related behaviors

of chase-mount in males and lordosis in females were almost

identical in quantitative as well as qualitative comparisons

between the two species. Similarly the behaviors of

sniffing, locomotor-explore and non-genital grooming, are

quite similar for males of both species. The patterns of

genital grooming and general grooming, sitting, upright

defense and lying down were also similar for females of

both species.

The patterns of change in behaviors across time periods

(i.e., from IL to EL to PEI) within the first series were

very similar for males of both species.with regard to the

behaviors genital grooming and sitting. Both species showed

more genital grooming during the EL period and more sitting

during the PEI period. Patterns of change across time

periods were also very similar for laboratory rats and roof

rat females for the behaviors sniffing, run from male and

locomotor-explore.

Quantitatively, laboratory rat males tended to spend

more time genital grooming during EL and PEI periods than did

roof rat males. However, roof rat males tended to spend








more time sitting during these two periods than did

laboratory rat males. Female roof rats tended to spend

much more time running from males during the IL and EL

periods than did female laboratory rats. However, female

laboratory rats tended to spend about twice as much time in

locomotor-exploratory activities during all time periods

than did female roof rats.

Qualitatively, the only major differences between the

two species appears to be with regard to the occurrence of

allo-grooming in roof rats. Male and female laboratory rats

are not reported to engage in genital and general grooming

of the partner while male and female roof rats engage in low

but consistent levels of these behaviors. Climbing behavior

is also not reported for laboratory rats; however, Dewsbury's

testing situation provided no opportunity for his subjects

to climb. It is doubtful whether they would climb even if

given the opportunity. On several occasions laboratory rats

have been tested in the roof rat testing cages and at no

time have laboratory rats been observed to climb on the cage

top.

Quantitatively and qualitatively roof rats appear very

similar to laboratory rats with regard to the behaviors

accompanying copulation. The differences between laboratory

rat and roof rat females in running from the male may be

related to the occurrence of the darting, ear wiggling and

tail rattling pattern observed in female roof rats that was

discussed earlier. The occurrence of this presumed

solicitation pattern may indicate a real species difference








between roof rats and Norway rats. It may be that roof rat

males require more stimulation from the female in order to

initiate copulation than do laboratory rats. It is also

possible that the same behavior patterns occur in wild

Norway rats but that through domestication, these patterns

have been lost. There are no clear explanations for the

other differences observed between the two species.


Ultrasonic Vocalizations

Ultrasonic vocalizations were observed on a total of

22 tests for 11 pairs of animals, each pair being observed

on two tests. These were the same tests during which the

categorizations of behavior were done: the fifth and sixth

positive tests for each pair. Identification of individual

animalsproducing ultrasonic vocalizations in social

situations is sometimes difficult because such vocalizations

can be produced by more than one animal at one time in a

given test. Using only the ultrasonic receiver it is

impossible to localize the sounds well enough to positively

identify the source of the vocalizations. In most cases,

however, it was possible to correlate the occurrence of.

the ultrasonic vocalization with rhythmic chest movements

in one but not the other animal. These chest movements,

probably rhythmic exhalations producing the vocalizations,

are also seen in isolated animals during vocalizations.

Singing was never observed to occur in the absence of such

chest movements by one or the other animal.








During protesting, scans of the frequency range from

20 kHz to 60 kHz during various phases of the copulatory

sequence revealed the only observable vocalizations to

occur at approximately 28 kHz. As a result during actual

testing the frequency dial of the ultrasonic receiver

was set at about 28 kHz and the occurrence and duration of

all vocalizations occurring in this range were recorded on

one channel of the event recorder. This procedure minimized

the probabilities of detecting ultrasonic vocalizations

occurring at other frequencies and it is therefore possible

that vocalizations were made at other frequencies but went

undetected. Sewell (1967) for example reports male

vocalizations in the 50 kHz range during.mounting activity

in laboratory rats.

Both males and females were observed to emit ultrasonic

vocalizations during copulatory sequences. For both males

and females, vocalizations were observed at 28 + 5 kHz in

frequency. These vocalizations occurred in pulses of 1-3

sec in length with pulses occurring in trains of variable

length; some up to 3 min in length on occasion.

Some sex differences were noted in the occurrence of

ultrasonic vocalizations. Males were observed to vocalize

only during PEIs and the satiety criterion period.

Ultrasonic vocalizations followed 91% of the 83 total

ejaculations observed on these 22 tests. On 8 tests male

vocalizations were observed that were not directly

associated with ejaculations or other copulatory activities.








All of these vocalizations occurred during the satiety

criterion period. Twenty-two percent of these vocalizations

appeared to be correlated with sudden noises occurring in

the testing room.

Twenty-five instances of female vocalization were

observed on 14 of the 22 tests. Female vocalizations,

unlike male vocalizations, were observed during EL periods

as well as PEI periods and the satiety periods. However,

only 24% of the females' vocalizations occurred during EL

periods; the other 75% czcurred during PEIs and the satiety

period. Only 20% of all female vocalizations appeared to be

correlated with sudden noises.

Table 5 presents quantitative data (means and standard

errors) on male postejaculatory vocalizations from eight

subjects on 16 tests. Only those subjects having at least

three series on both tests were included in order to

facilitate analysis of changes in behavior across series.

The measures of vocalization and other copulatory measures

have been defined previously. The numbers in parentheses

are means and standard errors for data collected by Barfield

and Geyer (1975) on laboratory rats and have been included to

facilitate cross-species comparisons.

As can be seen from the table, the measures IF and EL

show significant decreases from first to second series. The

measures.VT and PEI-show significant increases from first

to second series. The measures VL, PEI-VT and VT/PEI do

not change significantly from first to second series. It










Table 5


Means and Standard Errors of the Quantitative Measures of Male
Postejaculatory Vocalization and Other Measures of Copulatory
Behavior and Results of T-tests


Series


t-test for
related pairs
(df = 7)


(174.6 + 45.6) -------------
IL 45.0 + 19.8 -------------


( 11.1 + 1.5)
IF 8.8 t 1.0


(859.0 121.0)
EL 595.9 + 152.5


( 36.0 + 4.0)
VL 36.4 6.8


(328.0 + 14.9)
VT 161.6 + 18.9


(429.0 + 18.9)
>EI 256.8 + 19.0


PEI-VT



VT-PEI


(101.0 + 15.1)
95.2 + 17.6


( 0.77 + 0.03)
0.63 + 0.07


( 5.2 + 0.5)
3.4 0.3


(337.0 + 33.8)
106.9 + 12.5


( 42.8 + 10.2)
41.4 5.7


(415.0 + 19.0)
210.5 + 25.5


(528.0 + 19.9)
294.9 t 20.2


(117.0 + 18.5)
85.0 + 15.5


( 0.79 + 0.3)
0.71 0.06


t = 5.35 **



t = 3.05 *


t = 0.65


t = 3.16 *



t = 4.58 **


t 0.89



t = 1.84


Note: N = 8, 2 tests per animal.


Measures in parentheses are
and Geyer, 1975.


for Norway rats from Barfield


* p < .02
** p < .01


Behavioral
Measure








takes male roof rats from 36-41 sec to initiate ultrasonic

vocalizations following an ejaculation and such vocalizations

continue for 2.5 to 3 min. Roof rat males spend from 63-71%

of their total PEI in vocalization activity.

When the present data are compared to those of

Barfield and Geyer (1975) for laboratory rats, striking

similarities are seen between the two species. The measures

VL and PEI/VT are almost identical in the two species.

Laboratory rats tend to spend more absolute time in

vocalization than do roof rats, as measured by VT. However,

the total PEIs for laboratory rats are considerably longer

than those of roof rats. When the percent of the PEI spent

in vocalization is examined it reveals a much closer

similarity between the two species. Roof rats spend 63%

of their first PEI in vocalization; laboratory rats spend

77%. Roof rats spend 71% of their second PEI in vocalization,

laboratory rats 79%.

It has been suggested that in laboratory rats the

measures VT and PEI/VT are fair estimates of the absolute

and relative postejaculatory refractory periods, respectively

(Barfield & Geyer, 1975; Karen & Barfield, 1975). Similarly

VT/PEI has been suggested to be a reasonable estimate

of the percentage of the PEI taken up by the absolute

refractory period. Again, the close correspondence

between these two species with regard to these measures

suggests that similar mechanisms underlie the arousal,

maintenance and exhaustion of their copulatory behavior.






76

Barfield and Geyer (1972, 1975) report that male

laboratory rats are most often inactive during ultrasonic

vocalizations, lying quietly or occasionally grooming

themselves or moving slowly about. Similar behavior for

female laboratory rats during male ultrasound is also

reported by these authors with females reported "...also

often quiescent"(Barfield & Geyer, 1975, p. 726).

The collection of categorization data concurrently with

ultrasonic vocalization data provides an excellent

opportunity to quantify the behaviors of male and female

roof rats during ultrasonic vocalizations.

The behavior of 11 males and 11 females during male

postejaculatory vocalizations was categorized on one test

for each male and each female. The per cent of time males

spent in six categories of behavior during the first, second

and terminal PEI vocalizations is presented in Figure 9.

The per cent of time spent by females in eight categories

of behavior during the male's first, second and terminal

PEI vocalizations are presented in Figure 10. Only six of

the possible 16 categories of male behavior and only eight

of the possible 15 categories of female behavior are

presented because in both cases the remaining behaviors

occurred no more than 1% of the time.

Figure 9 reveals that male roof rats spent the

majority of their vocalization time sitting. The percentage

of vocalization time spent sitting increased over successive

series. Less than 12% of the vocalization time was spent in


__
















*O

00
00
0



o
0



UOH






rn 0




fl)
0





80 4

*r-4 0
4- .1




0(0
-H 0



.C C)



4-)
-H O













V) 0

Cd tO 4
0 0e
tdo C 0



O co
0 a)
0 4-) !,



- C
CC cn
*H) 0





r0 0
C) CC







-H QeC)
a0 U
C-l 'd
5-4-)
4c1 i

0 nC
C TU
C X>
0 n
*r-i i

nj o C


































U) a )!l

C" 0
.0 0 4OU, '-4





> ,
O -d3.3---1 LI
630 .
63 v ~s 0









Sa5i-- wD touD i
Ini5 CL







I- I I 1 I I ,
0 0 0 0 0
C ID ,t c

aui!jL JO ;uajad


~I_












o


*0
rn n




B '0
0 +j








n.
0
OF-



r.4)
In 0




00

OF-.
0








0 +-I)
0
Or*













4-1









00
U0




C)H
O Q
on


Cr4l 0













a )


OF-






-H Q L
4-1 0 )


m 101

0 0)

- 0




-41 c)
0 4LA0

H u 0


0 0








*Hd H
n, a 0































rs
illga uo r.a



Ja 'a a uD |'oua'
JlaS =9 Imoua9
(llS "0 I41u49.







1ausey suD IrraunD
("S





a4s 'uD le)auD








saussed u~ Igrlung
il'S W" ItlJal3
j1's riilua )



;iti uos t la

pa'Y'I


C)

c Oc
C/)



O






-, -


o
Sz


I I I I I I
0 0 0 0 0 0
0
auj!l jo luaJad


F E


Ls








locomotor-exploratory behavior, the remainder was spent in

quiescent activities such as grooming, sitting and lying.

This is very similar to the reports of Barfield and Geyer

(1972, 1975) with regard to laboratory rat activities during

vocalization. One point of difference is worthy of note.

Barfield and Geyer report that laboratory rats spend much

of their vocalization time lying down. Roof rats spend very

little of their vocalization time lying down. During the

majority of their vocalization time, male roof rats sit in a

relaxed but alert posture. Indeed, on several occasions

vocalizations were observed to be interrupted for short

periods of time by sudden noises in the testing room. Male

roof rats appeared to be immobile during periods of

vocalization, but are not totally unresponsive to environ-

mental stimuli.

The activities of female roof rats during male

vocalizations appear to be more variable. Figure 10 shows

that female roof rats spent the majority of the vocalization

time in such non-quiescent activities as locomotor-

exploratory behavior, feet on walls and climbing during the

first two vocalizations. However, during the last

vocalization, females spent much more of their time in such

quiescent activities as sitting and self-grooming. Female

roof rats appear to be much more active than female

laboratory rats during the early postejaculatory vocalizations

of the .male. However, during the final postejaculatory

vocalizations, female roof rats appear more similar to female

laboratory rats spending long periods of time in quiescent
behaviors.








Overall, the postejaculatory vocalizations of roof rats

appear remarkably similar to those reported for laboratory

rats in their quality and their temporal patterning.

The biological function of male postejaculatory

vocalizations is not clear. Barfield and Geyer (1972) note

that in laboratory rats 22 kHz vocalizations can be observed

in a number of different social situations. They suggest

that the 22 kHz frequency may be "a ...'carrier frequency'

for signals denoting states of contact avoidance" (p. 1350).

Specifically in sexual situations they suggest it may

function to inhibit female sexual behavior during periods

when the male is incapable of sexual activity. In a later

paper (Barfield & Geyer, 1975) they suggested that the

postejaculatory vocalizations might have a facilitative

effect on reproduction by affecting the reproductive

physiology of the female. Barfield and Geyer also noted that

sounds have been shown to be significant in initiating specific

reproductive responses in the females of some species of birds

(Barfield, 1971). It was suggested that postejaculatory

vocalizations by male rats may have a similar function.

Another possibility not mentioned by Barfield and Geyer concerns

the reduction of competition between conspecific males.

If 22 kHz vocalizations really do communicate a state of

contact avoidance in laboratory rats it is quite possible

that this might keep other conspecific males away from the

vocalizing male and the female with which he is mating.

This might prevent a take-over of the female by a strange






83


male when the vocalizing male is incapable of sexual

activity, thereby reducing competition between the vocalizing

male and others. Further research should clarify the

adaptive significance of posejaculatbry vocalizations in the

reproductive process.














EXPERIMENT 2


The results of Experiment 1 demonstrated that the

copulatory behavior of roof rats is fundamentally the same

as that of laboratory rats. However, the first experiment

did not provide any indication as to the importance of

various aspects of the male's behavior in the initiation

of successful pregnancy in female roof rats. As was pointed

out earlier, Adler (1969), Chester and Zucker (1970) and

others have shown that for laboratory rats, a single

ejaculatory series provides sufficient stimulation to induce

successful pregnancy in 80-90% of the females. In contrast,

Dewsbury and Estep (1975) and Lanier et al. (in press) have

demonstrated that in cactus mice (Peromyscus eremicus) and

Syrian golden hamsters (Mesocricetus auratus), copulations

beyond the first ejaculation are essential to maximize the

probability of successful pregnancy. Although roof rats

appear quite similar to laboratory rats in their copulatory

behavior, it is possible that quite different aspects of the

copulatory pattern are important for successful pregnancy in

roof rats. The present experiment was designed to assess the

significance of multiple ejaculations in the initiation of

pregnancy in roof rats. Comparisons are made between females

given just one ejaculation and those given a complete satiety

test with regard to the initiation of pregnancy.









Materials and Methods


Subjects

The subjects of Experiment 2 were nine male and sixteen

female laboratory-reared offspring of wild trapped roof rats.

All of these animals were bred, weaned, and maintained in

the same fashion as those of Experiment 1.


Apparatus

Behavioral tests were conducted in the same clear,

plastic cages as used in Experiment 1. Behavioral events

were also recorded in the same fashion as Experiment 1 using

the event recorder.


Procedures

Daily vaginal smears were taken on all females starting

when the females were 60 days of age and continuing throughout

the study. Smears were taken between 0800 and 1100 hrs using

a thin wire loop. These vaginal smears were examined micro-

scopically and a determination was made of the stage of the

female's estrous cycle on the basis of the cell types present

in the smear. The stages of the estrous cycle were classified

according to the criteria used by Hardy (1972) for labora-

tory rats. These were: proestrus, predominance of nucleated

epithelial cells; estrus, predominance of cornified epi-

thelial cells; metestrus, cornified and nucleated epithelial

cells and some leucocytes; diestrus, preponderance of

leucocytes with some nucleated epithelial cells.


_ __








After the females began to show regularity in their

estrous cycles, they were pretested for copulatory behavior

and fertility. On the afternoon of vaginal proestrus, the

proestrus female was placed in a testing cage with a male

and, if receptive, was allowed to mate for at least three

copulatory series. If the female was not behaviorally

receptive, she was returned to her home cage and retested

at the next occurrence of vaginal proestrus. Those females

showing no regular estrous cycles, having three consecutive

negative tests, or mating but not becoming pregnant and

delivering pups after two mating tests were eliminated from

the study. This system of protesting allowed positive

verification of fertility for both males and females and

allowed both males and females to gain experience in the

standard testing situation.

Females judged fertile by this method and which

demonstrated two consecutive estrous cycles of 3-5 days

duration were tested twice, once in each of two conditions.

Half of the females were randomly assigned to first serve

in the one ejaculation condition (1 Ejac), the other half

served first in the satiety condition (Satiety). The design

was counterbalanced so that those animals tested in the 1

Ejac condition on their first test served in the Satiety

condition on their second test and vice versa. All of the

animals were between 123 and 192 days of age at the time of

their first test and between 179 and 240 days of age at the

time of their second test.


I








Mating tests were conducted on the afternoon of

vaginal proestrus approximately three hours into the dark

phase of the light cycle. The testing procedure was similar

to that of Experiment 1 with males introduced into the test

cages 5-10 minutes before the females. Testing started with

the introduction of the female. If the male failed to gain

an intromission within 30 minutes after the start of the

test, the test was terminated and the female was retested at

the next occurrence of vaginal proestrus.

Each female assigned to the 1 Ejac condition was allowed

to mate with a proven fertile male until he had achieved one

ejaculation. At this point the test was terminated and the

female was returned to her home cage. Each female assigned

to the Satiety condition was allowed to mate with a proven

fertile male until he had reached an arbitrary satiety

criterion of 30 minutes without an intromission. The test

was terminated at this time and the female was returned to

her home cage.

Following a positive test a female continued to receive

daily vaginal smears and the presence or absence of regular

estrous cycles was noted. Females not becoming pregnant were

allowed to display at least 2 complete 3-5 days cycles before

being retested. Those becoming pregnant were allowed to

deliver their litter, then retested after the resumption of

regular 3-5 day estrous cycle activity. Females displaying

at least 12 consecutive days of diestrus following a positive

mating test but not delivering pups were judged to be









pseudopregnant. Those displaying diestrous smears for 20

consecutive days and delivering pups between days 21 and 22

were judged to have become pregnant as a result of the mating.

Those females continuing to show regular 3-5 day cycles

following mating were judged to be neither pregnant nor

pseudopregnant.


Behavioral Measures

The standard measures of copulatory behavior were

taken on all tests.


Results and Discussion


A total of 16 females completed both tests and are

included in the final data analysis. Table 6 presents the

percentages of tests in which females became pregnant,

pseudopregnant or continued to cycle as a result of

receiving one ejaculation or satiety. A binomial test

(Siegel, 1956) for the proportions of females becoming

pregnant in each of the two test conditions, revealed that

there was no significant difference between the 1 Ejac and

Satiety conditions in the proportions of females becoming

pregnant (X = 1, N = 7; p = .06, 1-tailed). A binomial

test for the proportions of females becoming pregnant in

each testing order also revealed no significant differences

(X = 2, N = 7; p = .45, 2-tailed).

Although there were no statistically significant

differences between the 1 Ejac condition and the Satiety

condition in the proportion of females becoming pregnant,


~1~11~





89








Table 6


Percent of Tests in Which Females Became Pregnant,
Pseudopregnant or Continued to Cycle as a Result of
Differing Amounts of Copulatory Stimulation


Outcome


1 Ejaculation


% Pregnant


% Pseudopregnant


% Cycling


Mean Litter Size


8.0 + 0.6
(n = 9)


Note. n = 16 for each condition except as noted
size data


b
Satiety


7.1 + 0.7
(n = 14)




for litter


IF = 9.5


EF = 5.4


_ __








it is interesting to note that there appeared to be a trend

for a higher proportion of females to become pregnant after

receiving multiple ejaculations. Also worthy of note is

the fact that all of the females allowed to mate to satiety

showed a progestational response (pseudopregnancy or

pregnancy) while only 68% of those females receiving one

ejaculatory series showed a progestational response. This

response is an essential prerequisite for normal pregnancy

in the female laboratory rat and, it can be assumed, for all

female mammals not having a functional luteal phase in their

estrous cycles, such as roof rats.

A t-test for unrelated samples comparing litter sizes

for those females that had litters in each condition,

revealed no significant differences in litter size as a

function of different numbers of ejaculations t(21 = 0.96;

p ( .05, 1-tailed).

Table 7 reveals that within the 1 Ejac condition there

was a trend for females receiving high numbers of intro-

missions to become pregnant or pseudopregnant while those

receiving few intromissions prior to ejaculation tended to

continue cycling. A t-test for unrelated samples indicated

that those females becoming progestational had significantly

more preejaculatory intromissions than those females

continuing to cycle (Mean IF for progestational females

11.7, Mean IF for cycling females 5.4, t(14) = 5.06, p < .001).

Thus, as in laboratory rats, there appears to be a

direct relationship between the amount of copulatory


__





91








Table 7



Mean IF and Resultant Female Response
in 1 Ejaculation Condition


Outcome n IF




Pregnant 9 11.4 + 0.8



Pseudopregnant 2 13.0 + 1.0



Cycling 5 5.4 1.0


Data presented as means and standard errors








stimulation received by a female and the probability that

she will become pregnant. Although roof rat females

can become pregnant after a single ejaculation provided that

they receive a sufficiently high number of preejaculatory

intromissions, in the average first series females only

receive 7.4 10.1 intromissions. In the present experiment

the average first ejaculatory series provided only enough

stimulation to induce pregnancy in 56% of the roof rats so

stimulated. This percentage was increased to 87% when the

copulatory stimulation was increased to satiety, an average

of 5.4 ejaculations. Because there was no statistically

significant difference between the Satiety and 1 Ejac

conditions with regard to the proportion of females that

became pregnant, it appears that like Norway rats, roof rats

reach maximal probabilities of pregnancy with a single

ejaculatory series. However, it is difficult to believe that

the 56% pregnancy rate in the 1 Ejac condition constitutes

the maximal probabilities of pregnancy. It is quite possible

that given a larger sample size the proportion of females

pregnant in the 1 Ejac condition would be significantly

different from the proportion of females pregnant in the

Satiety condition.

It should also be noted that all of the work done thus

far by Adler, Chester and Zucker and others on the functions

of multiple ejaculations in Rattus norvegicus has been done

on domesticated forms. To date no one has examined the

functions of multiple ejaculations in wild Rattus norvegicus.






93

It is possible that the lack of a function for multiple

series in this species, with regard to the induction of

pregnancy, is an effect of domestication. That is, multiple

ejaculations may be necessary for the induction of pregnancy

in wild Norway rats but through the process of domestication

this function has been lost. Domestication has been found to

affect a wide variety of behaviors in both Norway rats and

house mice (Mus musculus)(Barnett, 1963; Boice, 1972,1973;

Lockard, 1968; Smith, 1972). There is even some evidence

to suggest that domestication has affected the copulatory

behavior of house mice (Estep, Lanier & Dewsbury, in press).

Research on the induction of pregnancy in wild Norway rats

should-provide evidence to-confirm-or deny the possibility

that domestication has affected the stimulus requirements

for the normal initiation of pregnancy in this species.

The 32 tests of copulatory behavior conducted in this

experiment provide an opportunity to compare the behavior of

male roof rats paired with females in natural estrus to those

paired with females in an artificially induced estrus. The

females tested in this experiment were of necessity in

natural estrus while those in Experiment 1 were brought into

estrus with the aid of exogenous hormones.

Tables 8 and 9 present data from males mated with

hormone-induced estrous females from Experiment 1 and data

from males mated with females in natural estrus from the 1

Ejac and Satiety conditions, of the present experiment. In

Table 8 means and standard errors for the latency measures




Full Text

PAGE 1

COPULATORY BEHAVIOR OF Raff us' fattus By DANIEL QUEN ESTEP A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 1975

PAGE 2

UNIVERSITY OF FLORIDA niiiiiiiiiii 3 1262 08552 8437

PAGE 3

ACKNOWLEDGEMENTS I would like to express my gratitude to Professor Donald A. Dewsbury for his continued help, advice and criticism of all aspects of this study. I would also like to thank Profs. J. H. Kaufmann, E. F. Malagodi, M. E. Meyer and W. B. Webb for their careful reading of the manuscript. Dr. Steven Humphrey, Dr. L. E. Harms and Dr. T. Kelly all provided valuable aid in various aspects of the research. Their assistance was appreciated. I would also like to thank Theodore Fryer and Isaiah Washington for caring for the animals used in this research. Finally, I would like to thank my wife, Barbara, for typing the manuscript and for her love, patience, understanding and moral support throughout my years of graduate study. 111

PAGE 4

TABLE OF CONTENTS Page ACKNOWLEDGEMENTS. . • • • iii ABSTRACT • vi INTRODUCTION .......=.. 1 The Study of Copulatory Behavior 1 The Description of Copulatory Behavior 4 The Adaptive Significance of Variations in Copulatory Behavior ........ v 7 Methods in the Study of Adaptive Significance....... 8 Variations in Copulatory Behavior and Reproductive Success •• •••• 11 The Natural History of Roof Rats... 19 The Present Research 21 EXPERIMENT 1 24 Materials and Methods 24 Subjects 24 Apparatus . . . ^-> Procedures 26 Behavioral Measures . . 27 Results and Discussion . . 32 Basic Motor Patterns of Copulation 32 • Quantitative Measures of Copulatory Behavior.;.... 35 Categorization of Behavior • 57 Ultrasonic Vocalizations... 71 iv

PAGE 5

TABLE OF CONTENTS (continued) Page EXPERIMENT 2 . . . * 84 Materials and Methods • • 85 Subjects '... 8 5 Apparatus 85 Procedures. 85 Behavioral Measures.... . ... . 88 Results and Discussion.......... • 88 GENERAL DISCUSSION. .-.'.. . 102 REFERENCES. • • • • H 2 APPENDIX A. ANALYSES OF VARIANCE FOR EXPERIMENT 1... . 118 APPENDIX B. ANALYSES OF VARIANCE FOR EXPERIMENT 2 .127 BIOGRAPHICAL SKETCH .................... .'.: ............. 136

PAGE 6

Abstract of Dissertation Presented to the Graduate Council of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy COPULATORY BEHAVIOR OF Rattus rattus By Daniel Quen Estep August, 1975 Chairman: Donald A. Dewsbury Major Department: Psychology The copulatory behavior of a wild population of roof rats ( Rattus rattus ) was investigated in two experiments. In the first experiment the copulatory behavior of 12 pairs of laboratory-reared wild roof rats was observed on a total of 71. tests, each carried to a 30 minute satiety criterion. Females were brought into behavioral estrus with the aid of exogenous hormones. The basic motor patterns of copulation were described and the standard measures of copulatory behavior were taken. Categorizations of the. behaviors accompanying copulation were also made on one test each for each male and each female of 11 pairs of animals. The occurrence and duration of bouts of ultrasonic vocalizations were also recorded on each of these 22 categorization tests. With regard to most of the aspects of their copulatory behavior, roof rats appeared essentially identical to laboratory or Norway rats ( Rattus norvegicus ) . Roof rats

PAGE 7

were found to display a pattern of copulatory behavior characterized by no copulatory lock, no intravaginal thrusting, multiple ejaculations and usually multiple intromissions prior to ejaculation. Roof rats were also similar to Norway rats with regard to most of the quantitative measures of copulation, the behaviors accompanying copulation and patterns of ultrasonic vocalization. Both male and female roof rats were observed to emit ultrasonic vocalizations at 28 kHz during copulation. In the second experiment the role of multiple ejaculations in the initiation of pregnancy was investigated. Sixteen young females were mated twice, once to males mating to only one ejaculation and once to males mating to a 30 minute satiety criterion. Eighty-seven percent of the females mated to satiety became pregnant while only 56% of the females mated to one ejaculation became pregnant. It was concluded that like Norway rats, roof rats may reach the maximal probabilities of pregnancy with only one ejaculation, although it appears that roof rats may require more intromissions than Norway rats to become progestational. The results from these' two experiments were compared to similar data from other muroid rodents and the results were discussed in terms of the adaptive significance of variations in copulatory behavior. Finally, the differences and similarities in copulatory behavior of roof rats and laboratory rats were discussed in terms of possible behavioral isolating mechanisms between the two species. vii

PAGE 8

INTRODUCTION The research reported in this dissertation provides the first quantitative description of the copulatory behavior of wild roof rats ( Rattus rattus ) . Included in this description is an analysis of the behaviors accompanying copulation, including ultrasonic vocalizations. In addition there is an experimental analysis of the role of multiple ejaculations in the initiation of pregnancy in this species. To provide a rationale for this research and to review the appropriate background literature the introduction has been divided into several sections. The first section provides the rationale for studies of copulatory behavior and of adaptive significance. The latter sections review the relevant literature in these areas, provide some background information about roof rats and outline the research to be reported. The Study of Copulatory Behavior Mammalian copulatory behavior provides an excellent locus for comparative behavioral analysis. Among the reasons for this are the following. First, it has' been found that the copulatory behavior of mammals is highly stereotyped within species but can show considerable variability between different species. Second, copulatory

PAGE 9

behavior is of great biological significance, being near the heart of biological fitness. Finally, copulatory behavior can be studied in the laboratory making it an excellent preparation for experimental analysis. The variability seen in patterns of copulation among mammals provides the very essence of comparative behavior analysis. If there were little variability in copulatory behavior, comparative analysis would be impossible. The fact that the variability in copulatory behavior occurs primarily between species rather than within ruggests that patterns of copulatory behavior may be under significant genetic control and subject to evolutionary pressures. This in turn suggests that species differences in copulatory behavior may be due to different environmental pressures selecting for certain patterns over others. . As pointed out by Dewsbury (1972b), comparative analyses of these species differences and the evolutionary pressures responsible for them may contribute significantly to our general understanding of the evolution of both behavior and of reproductive systems in general. The ways in which mammalian copulatory behavior can vary will be summarized below. The reproductive success of a particular mammalian organism is in large part dependent upon its success in copulatory behavior. Except in cases of human intervention, organisms that do not engage in appropriate copulatory behavior do not leave offspring. The biological fitness

PAGE 10

of an organism is expressed in terms of the relative numbers of viable, fertile offspring left by that individual, in sexually reproducing organisms. Efficient, successful copulatory behavior is absolutely essential in order to maximize reproductive success and hence, biological fitness. Successful copulatory behavior involves a number of factors. Mating must occur when there are maximal probabilities that viable, fertile offspring result. Similarly, the mating itself must not only transfer adequate numbers of sperm from male to female, but must also provide the necessary and sufficient stimulation to induce successful pregnancy. Studies by Adler (1969), Dewsbury and Estep (1975), and McGill (1970), have shown that the simple transfer of sperm is not sufficient to induce normal pregnancy in some mammalian females and that specific patterns of copulatory stimuli from the male may be necessary. Small variations in these normal patterns of behavior can disrupt normal reproduction and hence lower biological fitness (Adler, 1969; Adler $ Zoloth, 1970). It is clear, therefore, that copulatory behavior is near the heart of the reproductive success of an individual. For a number of reasons copulatory behavior is an excellent behavior pattern for laboratory study. It is first of all highly stereotyped in its appearance and can be reliably elicited under a variety of conditions. Few, mammalian behavior patterns show such stability. In addition, the occurrence of copulatory behavior does not

PAGE 11

depend upon prolonged training or extended practice among subjects. All of this means that the investigator can generate stable baselines in behavior both within and between subjects of a given population. Such stable baselines are ideal for experimental analyses and detailed comparisons across populations and species. The Description of Copula tory Behavior Comparative analysis begins with the basic description of the events to be compared. Before comparisons become meaningful, however, it is necessary to develop a system of classifying the behaviors observed. Dewsbury (1972b) has developed a taxonomy of mammalian copulatory behavior which suggests the ways in which copulatory behavior can vary and provides a framework for summarizing and comparing patterns of copulatory behavior. As Dewsbury (1972b) points out, various aspects of the entire copulatory situation may vary from species to species. . First, the antecedent conditions for copulation may differ. Some species, for example, breed seasonally while others breed the year round. The time of day, location of copulation and social context may also be significant and vary between species. Second, the nature of courtship or precopulatory behaviors may vary. Some species may have elaborate courtship behaviors while others do not. Third, the postures assumed during copulation may vary. Some species such. as humans have ventral-to-ventral

PAGE 12

copulation (Ford § Beach, 1951) while others, such as rats, have dorsal-ventral copulation (Beach § Jordan, 1956). Fourth, the behaviors accompanying copulation may vary. In some species such as laboratory rats, (i.e., domesticated Norway rats, Rattus norvegicus) , copulation itself may account for less than one percent of the total time of a copulatory sequence (Dewsbury, 1967) . The behaviors occurring during the rest of the time may vary considerably among species. Finally, the copulatory behavior itself may vary in a qualitative or quanitative fashion. After careful review of the available data on copulatory behavior in mammals, Dewsbury (1972b) proposed that copulatory behavior may vary in four basic ways. First, there may or may not be a lock or physical tie between male and female during copulation. Second, intravaginal thrusting may or may not occur after the male has gained penetration. Third, multiple insertions without sperm transfer, may or may not be necessary prior to ejaculation. Finally, multiple ejaculations may or may not occur during the copulatory sequence. As there are two alternatives to each of these 4 four attributes of copulatory behavior, there are 2 or 16 different patterns of behavior possible. Dewsbury has classified the copulatory behavior of laboratory rats for example, as pattern #13 because laboratory rats ( Rattus norvegicus) show no copulatory lock, have* no intravaginal thrusting, have multiple intromissions prior to ejaculation and have multiple ejaculations within a normal copulatory

PAGE 13

sequence. House mice ( Mus musculus) by contrast, have been assigned pattern #9 on the basis of the lack of a copulatory lock, and the presence of intravaginal thrusting, multiple intromissions and multiple ejaculations. The various attributes of copulatory behavior can vary quantitatively as well as qualitatively. The frequencies and durations of copulatory events may vary as well as the latencies and intervals between events. Laboratory rats, Mongolian gerbils ( Me ri ones unguiculatus ) , and rice rats ( Oryzomys pnlustris ) , for example, all display pattern #13 copulatory behavior but show substantial variability with regard to the quantitative measures of copulation (Beach § Jordan, 1956; Davis, Estep § Dewsbury, 1974; Dewsbury, 1970) . Any complete description of the copulatory behavior of a given species must include an account of the quantitative variations as well as the qualitative variations. Descriptions of the copulatory behavior of mammals can be found scattered throughout the biological and behavioral literature. However, as Dewsbury (1972b) notes, most of these descriptions are inadequate for most comparative purposes. In recent years Dewsbury and his colleagues have been collecting systematic quantitative descriptions of the copulatory behavior of a variety of muroid rodents (See for example, Dewsbury, 1970; 1971; 1972a; 1973b; 1974a; 1974b; Estep, Lanier § Dewsbury, in press; Gray § Dewsbury, 1973). These studies provide the kind of complete descriptions

PAGE 14

discussed above (including analyses of motor patterns, quantitative and qualitative variations in pattern and behaviors accompanying copulation) that are necessary to facilitate interspecific comparisons. Other good descriptions are to be found scattered among other groups of mammals such as primates (See Dewsbury, 1972b). Among the behavioral patterns found to accompany copulation, one of the most interesting is the occurrence of ultrasonic vocalization. These high frequency vocalizations have been reported to occur in a variety of species and in a variety of social situations (Sewell, 1970). With regard to the ultrasonic vocalizations occurring during copulatory sequences the work of Barfield and Geyer (1972; 1975) on laboratory rats is most important. They have found that male laboratory rats have regular patterns of ultrasonic vocalization that follow most ejaculations. They have described and quantified this pattern in detail and have discussed its possible biological functions. Descriptions of such patterns in other species would provide added detail to. the description of the behaviors accompanying copulation and perhaps allow some assessment of the functions of such behaviors in different species. The Adaptive Significance of Variations in Copulatory Behavior While the description of copulatory behavior is a worthwhile and necessary first step in the understanding of

PAGE 15

mammalian copulatory patterns, it is not an end unto itself. Such descriptions achieve significance only when related to some meaningful question concerning copulatory behavior. One such question concerns the biological function or adaptive significance of variations in copulatory behavior. Students of animal behavior have often wondered why different species of mammals display such diversity in their copulatory behaviors. However, few systematic studies have been generated to investigate this question. Dewsbury (1972b) and his colleagues are attempting one such systematic investigation within a taxonomically restricted group of mammals muroid rodents. By studying a behaviorally • diverse but closely related group such as the muroids, probabilities of discovering the biological significance of differences in copulatory behavior are greatly enhanced. How then, might one investigate the adaptive significance of variations in copulatory behavior? Methods in the Study of Adaptive Significance Dewsbury (1973a) lists three methods for investigating questions of adaptive significance: The method of adaptive correlation, the behavior-genetic method and the experimental method. The first, the method of adaptive correlation, involves correlating variations in the behavior of interest with variations in other behaviors, morphology and ecological variables. By sampling a variety of organisms with different behaviors and looking for such correlations,

PAGE 16

commonalities and consistent differences should become apparent that would suggest the functions of such differences and the environmental pressures responsible for them. A classic example of this method is presented by Cullen C1957) in the study of cliff nesting Kittiwakes. She was able to relate specific reproductive adaptations of these birds to their cliff nesting habits when comparisons were made to related non-cliff nesting gulls. The functions of these special adaptions to the nesting habitat became clear only when compared to other species who differed in their nesting habits. The behavior-genetic method involves the use of inbred lines of animals and therefore is of restricted utility. This method involves comparing the behavior of 1? 1 hybrids with the behavior of the inbred lines that gave rise to them. By comparing the scores of the F-^s to those of the parents, conclusions can be drawn concerning the strength of the selective pressures operating on the parental types and therefore the adaptedness of the behavior. Bruell (1964, 1967) discusses the utility of this method in some detail. A recent example of this method is the work of Dewsbury Cl975b). He has utilized this method to investigate the adaptive significance of variations in the copulatory behavior of laboratory rats. The final method discussed by Dewsbury is the experimental method. By directly manipulating the behavior, or the results of the behavior and looking for effects on

PAGE 17

10 reproductive success or fitness, one can suggest possible adaptive functions for the behavior. A good example of this method concerns the egg shell removal behavior of gulls studied by Tinbergen (1963) . Many gull species engage in egg shell removal behaviors soon after chicks are hatched. By systematically removing and replacing egg shells of various colors and looking at chick mortality, Tinbergen was able to show that egg shell removal helped prevent detection of the nest by predators and reduced chick mortality. The behavior was clearly adaptive and increased the reproductive success of those engaging in such behavior. Similar analyses have been carried out by Adler and his colleagues (See Adler, 1974 for a review), Dewsbury and Estep (1975) and Lanier, Estep and Dewsbury (in press) in an attempt to discover the adaptive significance of variations in copulatory behavior. The results of these analyses will be summarized below. Each of the three methods discussed above has its own set of advantages and disadvantages. The methods differ with regard to the kinds of information they can generate and the requirements for their utilization in terms of subjects and procedures. It should be noted that the methods are not contradictory in nature but may, in fact, be used to complement each other to provide a more thorough analysis of the adaptive significance of a particular behavior.

PAGE 18

11 For a number of reasons, the experimental method may be particularly attractive to those engaged in laboratory investigations of copulatory behavior. First, it does not necessarily require that a number of species be studied and compared directly. Second, investigations such as those of Adler, Dewsbury and Estep and others might be conducted along with a basic description of the copulatory behavior of a species thus allowing immediate access to information concerning the possible adaptive significance of the pattern described. Finally, experimental analyses of adaptive significance can provide data for a level of comparative analysis beyond simple description. Variations in Copulatory Behavior and Reproductive Success What, then, is the adaptive significance of variations in copulatory behavior? How might variations in copulatory behavior affect reproductive success and fitness? Adler (1969) and others have found that apparently small variations in male copulatory behavior can directly affect the female's reproductive physiology and. thereby affect reproductive success. In a comprehensive review of the mammals, Adler (1974) notes that variations in male behavior can affect a female's reproductive physiology at many points in her reproductive life. Perhaps the most intensively studied of these phenomena concerns the effects of male behaviors on various aspects of the normal reproductive cycle of adult females.

PAGE 19

12 As Adler notes (19 74) tne normal mammalian ovarian cycle has three sequential phases: 1) the follicular phase when ova and their follicles grow and mature, 2) the ovulatory phase when the mature follicle releases the egg and 3) the luteal phase when the ruptured follicle becomes a corpus luteum which secretes progesterone which aids in the maintenance of pregnancy." Conaway (1971) notes that there is considerable variability among mammalian species in the lengths of the various phases of the ovarian cycle, their occurrence and the kinds of stimuli necessary to elicit or modify them. In some species, for example, the ovulatory phase only occurs when the female receives copulatory stimulation. These animals are referred to as induced ovulators . Prairie voles ( Microtus ochrogas te r) and montane voles ( Micro tus montanus) are induced ovulators (Gray, Davis, Zerylnick § Dewsbury, 1974; Richmond § Conaway, 1969). In other species ,such as laboratory rats, ovulation occurs spontaneously. Similarly, the luteal phase may be spontaneous, as in humans, or induced, as in laboratory rats. While it has been known for quite some time that copulatory behavior was instrumental in initiating some of these physiological changes, the exact nature of the stimuli remained obscure. A series of experiments by Adler and his colleagues was designed to analyze the effects of variations in male laboratory rat behavior On successful pregnancy in the female by affecting her reproductive physiology.

PAGE 20

13 Male laboratory rats typically have a number of mounts and mounts with vaginal penetration, called intromissions, prior to ejaculation. An organized group of such mounts and intromissions, terminating in ejaculation is termed an ejaculatory series. Male laboratory rats normally display several such ejaculatory series prior to sexual satiation (that is, a period of 30 minutes with the same female without an intromission or ejaculation). Wilson, Adler andLe Boeuf (1965) found a correlation between the amount of copulatory stimulation delivered by sexually vigorous males and the probability of pregnancy in the females receiving this stimulation. One group of females was allowed to mate undisturbed until the males attained their first ejaculation, usually after about nine intromissions. The other group was mated with males which only had three or fewer intromissions prior to their first ejaculation. Ninety percent of the females allowed a full complement of intromissions became pregnant while, only 22% of the females receiving three or fewer intromissions became pregnant. Wilson et al. thus demonstrated that multiple intromissions prior to ejaculation (not just the ejaculation itself) were necessary to initiate normal pregnancy in young virgin female laboratory rats. Adler (1969) later replicated these results using a different strain of rats. Apparently, the male's behavior functioned in two ways to induce normal pregnancy. First Adler (1969) found that

PAGE 21

14 a sufficiently high number of preejaculatory intromissions was necessary for normal sperm transport. Those females having only one intromission prior to ejaculation had no sperm in their uteri 1-3 hours after copulation while those that had 2 or more intromissions had high numbers of sperm in their uteri. Chester and Zucker (1970) later extended these results and found that only 501 of the females receiving one ejaculation on the male's first insertion had sperm in their uteri 1/2 4 hours later. The biological importance of such rapid sperm transport war, demonstrated by Adler (1968, 1969). He found that those females receiving high numbers of intromissions prior to ejaculation had normally developing embryos 3 days after copulation while those receiving low numbers of preejaculatory intromissions had unfertilized and degenerating ova in their fallopian tubes. Thus multiple intromissions were necessary for normal sperm transport and fertilization. It should also be noted that the ejaculatory reflex and the hard coagulated copulatory plug delivered by the male at ejaculation were also found to be necessary for normal sperm transport (Blandau, 1945) . The second way in which the male's behavior was found to influence the initiation of pregnancy in the female was through the induction of the luteal phase of the cycle. As noted earlier,* the luteal phase is induced in laboratory rats and is essential for the maintenance of normal pregnancy in the female. The progesterone secreted by the corpus

PAGE 22

15 luteum is essential for preparing the uterus to receive the fertilized embryo and allowing implantation. Adler (1969) showed that 100% of the females given a high number of intromissions showed a cessation of behavioral receptivity while only 18% of the females given three or fewer intromissions showed a cessation of behavioral receptivity. This implied a release of gestational hormones in response to the high levels of stimulation which in turn caused the cessation of behavioral receptivity. In a further elaboration of this work, Adler (1969) gave varying numbers of intromissions to females but did not allow them to receive an ejaculation. His data show that the proportion of females ceasing behavioral receptivity increased as a function of increased copulatory stimulation. Adler (1968) further showed an increased incidence of pregnancy among females inseminated following, reduced numbers of intromissions if an injection of exogenous progesterone were given following copulation. Finally, Adler, Resko and Goy (1970) measured circulatory levels of plasma progestins directly in females given varying amounts of copulatory stimulation and found a direct correlation between the amount of copulatory stimulation and levels of progestin. All of these data confirm the importance of multiple intromissions in the induction of the luteal phase and the subsequent production of progestins, which is in turn essential for the maintenance of normal pregnancy. Although the mechanism is not yet fully understood, experimental

PAGE 23

16 evidence reviewed in Adler (1974) suggests that the induction of the luteal phase occurs as a result of a neuroendocrine reflex which stimulates corpora lutea growth and thus the secretion of progestins. All of the above evidence suggests that the frequency of intromissions is important in induction of normal pregnancy in laboratory rats. Work by Adler and Zoloth (1970) further suggests that the timing of copulatory events in laboratory rats may also.be important. They found that if ejaculations were followed too closely by additional intromissions or by manual cervical stimulation, sperm transport and presumably, pregnancy could be disrupted. Thus the patterning of post-ejaculatory events is also important in induction of normal pregnancy. The work of Adler and his coworkers shows that a single normal ejaculatory series provides the necessary and sufficient stimulation for induction of normal pregnancy. However, laboratory rats typically have multiple ejaculatory series. If one ejaculation normally provides all the stimulation necessary for normal pregnancy, what then is the biological function of multiple ejaculatory series? Recent work by Davis (1974) provides one possible answer. Davis found that while one ejaculatory series was indeed sufficient to induce normal pregnancy in young, virgin female laboratory rats, this was not the case in older multiparious females. Davis found that none of the older females receiving one complete ejaculatory series became

PAGE 24

17 pregnant while 66% of those receiving 3 series and 921 of those receiving five series became pregnant. Moreover, those receiving five series had significantly more developing embryos in their uteri than those receiving only 3 series. Because none of the older females receiving just one series showed a cessation of regular estrous cycles, it was suggested that the reproductive failure of these females was due to inadequate induction of luteal function. Subsequent measurements of plasma progestins in older females receiving varying numbers of ejaculations supported this suggestion. It appears therefore, that multiple series delivered by a male could extend the reproductive life of aging females and increase the number of offspring produced, thus increasing the fitness of those males delivering multiple ejaculations. Copulations beyond the first ejaculation also appear to be important in the induction of pregnancy in young females of other species. Dewsbury and Estep (1975) have demonstrated that only 5% of the cactus mice ( Peromyscus eremicus ) females given just one ejaculatory series become pregnant. Forty-seven percent and 60% respectively of the females allowed to mate to satiety and to satiety plus pairing overnight became pregnant. Thus, copulation beyond the first ejaculation in this species is necessary to maximize the probabilities of pregnancy. Work by Lanier et al, (in press)., has shown a similar effect in Syrian golden hamsters C Mesocricetus auratus ) . In two experiments

PAGE 25

they found that 201 and 401 respectively, of the females receiving just one copulatory series became pregnant. In both experiments 100% of the females mated to a 15 min satiety criterion became pregnant. They also demonstrated a direct relationship between the number of ejaculatory series received and the proportion of females becoming pregnant. They concluded that multiple ejaculations are essential in Syrian golden hamsters to maximize the probabilities of pregnancy. It should be clear from this review of the literature that variations in male copulatory behavior can dramatically effect the reproductive physiology of the female, the induction of pregnancy and ultimately the fitness of organisms involved. Experimental investigations of the biological function of various aspects of the male's copulatory pattern have suggested possible functions for various aspects of the patterns such as multiple intromissions and multiple ejaculations. In addition species differences have been found in the function of similar aspects of the behavior. Laboratory rats and Syrian golden hamsters have multiple intromissions and multiple ejaculations; however, multiple ejaculations maximize the probability of pregnancy in young female hamsters while they do not in young laboratory rats. As mentioned previously, experimental investigations of the biological function of patterns of copulation carried out in conjunction with basic descriptive analyses provide an added level of

PAGE 26

19 analysis to comparative investigations of copulatory behavior. They also facilitate the study of the adaptive significance of variations in copulatory behavior and contribute to our general understanding of the evolution of behavior and reproductive systems. Few such descriptions and experimental analyses exist, however, even for species of muroid rodents. Such a detailed description of the copulatory behavior and an experimental analysis of the biological functions of the behavior seems particularly warranted for one species of muroid rodent the roof rat (Rattus rattus ) . Roof rats are closely related to the common laboratory rat and are very common, living in close association with man. However, very little is known of their behavior including their copulatory behavior. The Natural History of Roof Rats According to Simpson (1945), the roof rat or black rat (Rattus rattus L.) is a member of the order Rodentia, superfamily Muroidea, family Muridae and subfamily Murinae. Simpson recognizes 68 different genera within the Murinae, one of which is the genus Rattus . According to Ellerman (1941), there are over 554 forms in the genus, many of which are subspecies. Walker (1964) notes that this makes the genus Rattus the largest genus of all mammals. Most of the 284 distinct species recognized by Ellerman are tropical in distribution and are found chiefly in southeast Asia and Africa (Walker , 1964) . Two species have gained world wide

PAGE 27

20 distribution: the Norway rat, Rattus norvegicus , and the roof rat, Rattus rattus . Norway rats tend to be a more temperate species and are thought to have originated in the area around the Caspian Sea, whereas roof rats are considered to be a more tropical species originating in Indo-Asia (Robinson, 1965). According to Walker, Robinson and others, both roof rats and Norway rats gained their global distributions through their associations with men. Roof rats supposedly were first brought to Europe during the crusades by Europeans on their way back from the Middle East. Roof rats brought with them the plagues which ravaged Europe during the 12th, 13th and 14th centuries. The Norway rat supposedly did not get to Europe until approximately the 16th century. In northern climates the newly introduced Norway rat quickly displaced the smaller, less vicious roof rat. In more tropical climates the displacement of roof rats by Norway rats has been much slower. According to Walker (1964) roof rats reached North America with the first explorers, while Norway rats did not reach North America until the 18th Century. Roof rats and Norway rats are very similar morphologically but can be distinguished by several reliable keys. According to Hall and Kelson (1959> Norway rats are larger in body size than" roof rats. Norway rats have 12 teats while roof rats have only 10. Roof rats have tails longer than their entire bodies while Norway rats have tails equal to or shorter than their body lengths.

PAGE 28

21 Roof rats and Noruay rats often exist as commensals of man, living in barns, grain storage areas and human habitations. Roof rats tend to be arboreal, nesting in the roofs of houses and barms, while Norway rats tend to nest in burrows or in protected nest sites on the ground. According to Telle (19661 and Ewer (1971) roof rats form complex social groups with a linear male hierarchy or dominant male type of social organization. Both Telle and Ewer report that group territories are actively defended against intruders. Barnott (1963), Ewer (1971) and Telle (1966) provide some information on other aspects of the behavior of this species such as foraging, aggression, parental care, maintenance activities and other social interactions. Barnett and Telle also discuss interspecific interactions between roof rats and Norway rats. According to Barnett (1963) roof rats and Norway rats do not hybridize in the laboratory or the field. The Present Research Ewer (1971) has observed the copulatory behavior of both captive and free ranging roof rats. She describes their behavior as conforming "...in general with what has been described by Barnett in ffattus norvegicus . . . [p. 155]". Her description suggests that roof rats have a pattern characterized by multiple intromissions prior to ejaculation, no lock and probably multiple ejaculations prior to sexual satiety. This is indeed similar to the pattern shown by

PAGE 29

22 Norway rats. While Ewer provides data on the events preceding copulation, postures and motor patterns assumed during copulation and a rough description of the copulatory pattern itself, much remains to be done to provide a detailed quantitative description of the copulatory behavior of this species. If the maximal information is to be gained from interspecific comparisons the copulatory pattern should be described in as much detail as possible. Without this added detail, conclusions about the adaptive significance c.f variations in copulatory behavior must remain tenuous and preliminary. The research reported here consists of two experiments. The first provides a quantitative description of the copulatory behavior of roof rats. The second provides an analysis of the stimulation essential for initiation of successful pregnancy in this species. In addition to providing data that will facilitate interspecific comparisons and aid in our understanding of the adaptive significance of variations in patterns of copulation, these data might also provide some insight into possible behavioral isolating mechanisms between roof rats and Norway rats. As noted earlier the two species are occasionally sympatric, at times living in the same building, but never hybridizing. It is possible that one of the mechanisms preventing hybridization involves differences between the species in their copulatory behavior. If the differences were substantial enough, they might prevent the induction of

PAGE 30

23 pregnancy when heterospecifics did happen to mate thus preventing successful hybridization.

PAGE 31

EXPERIMENT 1 While detailed descriptions of copulatory behavior exist for a variety. of muroid rodents including laboratory rats (Dewsbury, 1972b), no such detailed quantitative description exists for the common roof rat. This experiment was designed to provide such a description. Provided are an analysis of the basic motor patterns, a qualitative and quantitative, description of the pattern of copulation, and a description of the behaviors accompanying copulation including ultrasonic vocalizations. Such detail should facilitate interspecific comparisons and provide a framework for experimental analyses of the adaptive significance of various aspects of the copulatory behavior of this species. Materials and Methods Sub j ec t s All of the subjects of Experiments l'and 2 were the first generation laboratory reared offspring of wild trapped animals. The parents of these subjects were 11 males and 8 females all live-trapped as adults at the University of Florida Poultry Farm, southwest of the main campus ' of the University of Florida. Several of these original specimens were positively identified as Rattus rattus L. after close 24

PAGE 32

25 morphological examination by Dr. Steven Humphrey of the Florida State Museum. The subjects for Experiment 1 were 12 male and 12 female laboratory reared offspring of these wild trapped animals. All the subjects were weaned at 30 days of age and housed in isolation in Wahman suspension cages measuring 18 cm x 17.5 cm x 24.5 cm. All animals received Purina Lab Chow and water ad lib. The colony was housed in an air conditioned room maintained at approximately 21 C. A reversed light-dark cycle of 16 hr light, 8 hr dark was enforced throughout the experiment with light onset occurring at 1800 hr daily. One 25 watt red light bulb illuminated the room at, all times. Apparatus Animals were tested in clear plastic cages measuring 38 cm x 2.0 cm x 48.5 cm covered with stainless steel cage tops made of 2 mm diameter rods. San-i-Cel brand litter material covered the bottom of the cages and served as a substratum for testing. Behavioral events were recorded on a 20 channel Easterline-Angus event recorder and by hand with written descriptions when necessary. Ultrasonic vocalizations were detected wi'th the aid of a Holgate Mark V ultrasonic receiver and a Telequipment Model S54A Oscilloscope. The oscilloscope was calibrated at '50 y^ sec/ cm and 0.5 volts/cm. The ultrasonic receiver was calibrated with the aid of a frequency generator and was found to be accurate within + 5 kHz in a range from 20 kHz to 100 kHz.

PAGE 33

26 The ultrasonic receiver is a superheterodyne instrument that transduced the ultrasonic input into an audible output. This output was monitored directly with the use of headphones and on occasion the audible output was also fed into the oscilloscope and visual representations of the signal were observed. Procedures All of the males used in this experiment had been pretested for copulatory behavior at least once prior to the start of formal testing and proved to be reliable copulators Pretesting consisted of pairing a male with a female in natural estrus and allowing him to mate uninterrupted for at least one hr. Males had not been allowed to explore the apparatus prior to pretesting. Females were not pretested for copulatory behavior prior to formal testing but all had been permitted to explore the test apparatus on at least one occasion of 20-30 min. All of the animals were between 130 and 180 days of age at the start of the experiment and between 210 and 260 days of age at the time of their final test. Animals were tested in the colony room approximately 3 hrs after the beginning of the dark phase of the diurnal cycle. During, the experiment four 25 watt red light bulbs illuminated the testing cages. Males were introduced into the cages 5 to 10 min before the start of testing and allowed to habituate to the situation. Tests were started with the introduction of the female into the testing cage. Males were given 30 min from the

PAGE 34

.27 start of the test to initiate copulation. On tests where the male did not achieve an intromission within this time period, the test was terminated and scored as a "negative". On tests where the male did initiate copulation, the test was allowed to continue until 30 min had elapsed since the male achieved his last intromission or ejaculation. At this time the test was terminated and scored as a "positive". Each male was tested with the same female at two wk intervals until each pair had achieved six positive tests or had received four consecutive negative tests. On the fifth and sixth positive tests for each pair a categorization was made of the behaviors accompanying copulation, and the occurrence of ultrasonic vocalizations in the 22-33 kHz frequency range was noted. On those tests where ultrasonic vocalizations were to be observed, the Holgate ultrasonic receiver was utilized, tuned to 28 kHz and the microphone was placed in the center of the food tray of the cage top. All females were brought into behavioral estrus with intramuscular injections of 0.1 mg estradiol benzoate followed by 1.0 mg of progesterone given approximately 75 hrs and 5-6 hrs before testing, respectively. Behavioral Measures Pretesting had indicated that the copulatory pattern of roof rats was essentially the same as that of Norway rats. That is copulations were found to be organized into groups or series of mounts, mounts with vaginal penetration

PAGE 35

28 (intromissions) and intromissions terminating in ejaculation (ejaculations).. Each, such organized series of copulatory behaviors usually ended with the occurrence of an ejaculation, Males have a number of such ejaculatory series before reaching sexual satiety. Accordingly, the standard measures of copulation for this species were adopted from those developed by. Beach and Jordan (1956) for laboratory rats. These standard measures of copulation were: ML Mount latency. The time from the start of testing until the first mount with pelvic thrusting (with or without intromission) . IL Intromission latency. The time from the start of testing until the first vaginal intromission by the male. MF Mount frequency. The number of mounts with pelvic thrusting but without intromission within an ejaculatory series. IF Intromission frequency. The number of intromissions within an ejaculatory series. EL Ejaculation latency. The time from the first intromissions of a series to the ejaculation of the series. Mill Mean interintromission interval. The, mean interval separating the intromissions of a series. This is calculated by dividing EL by IF. PEI Postejaculatory interval. The time from the ejaculation of oneseries until the resumption of copulation as indicated by the first intromission of the next series.

PAGE 36

29 It should be noted that the measures IF, EL and Mill are not completely independent of each other for a given ejaculatory series (EL = IF x Mill). On those tests where ultrasonic vocalizations were to be observed, several measures of vocalizations were taken. The postejaculatory ultrasonic vocalizations of roof rats were found to be similar in pattern to those of Norway rats and therefore the parameters of postejaculatory vocalizations developed by Barfield and Geyer (1975) were utilized for roof rats. These measures were: VL Vocalizations latency. The time from the ejaculation until the beginning of the ultrasonic vocalizations . VT Vocalization termination. The time from the ejaculation until the end of the ultrasonic vocalizations. In addition two derived measures were taken: PEI-VT An estimate of the time spent in ultrasonic vocalization. Calculated by subtracting the VT for a given copulatory series from the PEI in which it occurred. VT/PEI The proportion of the PEI spent in ultrasonic vocalization. Calculated by dividing the VT for a given series by the PEI in which it occurred. After preliminary observations of the copulatory behavior of roof rats, a system was developed for the categorization of the behaviors accompanying copulation into several mutually exclusive categories. The occurrence and duration of each of these categories of behavior were

PAGE 37

30 recorded on an Esterline-Angus event recorder and later decoded to generate a profile of the behaviors occurring during the various phases of the copulatory sequence. Using this system the behaviors accompanying copulation were categorized on the fifth positive test for each male and on the sixth positive test for each female. The categorization was initiated at the start of the test and terminated at the end of the 30 min satiety criterion. The behavioral categories were adapted from those of Dewsbury (1967; 1970; 1971; 1972a; 1973b; 1974a; 1974b) and are listed below. All behaviors are common to both males and females unless otherwise noted. Mount (Males Only) Mounting, intromitting or ejaculating by males. Lordosis (Females only) the posture taken by females when males mount. It involves a stance in which the female plants her feet firmly, straightens her hind legs and lifts her head forming an inverted arch in her back. Chase (Males only) Following or chasing the female orienting towards her at all times and usually culminating in a mount of the female. Run From Male (Females only) Involves the female's darting away from the male or running from a pursuing male. Simultaneous ear wiggling and tail rattling behaviors may also occur. Sniff Involves an approach or orientation toward the partner in which the animal appears to sniff the body of the partner.

PAGE 38

31 Upright Defense Occurs when the animal stands upright on its hind legs orienting the ventrum towards the partner. Sometimes involves pushing or slashing of the partner with the forepaws and/or audible vocalizations. Genital Groom Self Scratching, licking or manipulation of the animal's own genital area. General Groom Self Scratching, licking or manipulation of the animal's own body with the exception of the genital area. Genital Groom Partner Scratching, licking or manipulation of the partner's genital region. General Groom Partner Scratching, licking or; manipulation of the partner's body with the exception of the genital region. Groomed A sitting posture adopted by an animal while being groomed by its partner. Ears are usually back and eyes are closed. Climb Climbing or hanging upside down on the cage top of the testing cage. Pull Female (Males Only) Slashing at or pulling on a female which is climbing or hanging on the cage top. Feet on Walls A posture in which the animal stands upright and leans against the sides of the cage with its forepaws. Dig Diggingor pushing the substratum with, the paws.Locomotor-Explore Walking movements sometimes accompanied by nosing of the substratum, walls or cage top,

PAGE 39

32 mouthing or manipulation of the substratum or feces, and upright postures other than those listed elsewhere. Sit Adoption of a sitting posture by the animals involving little movement other than slight mouth and head movements. Lie Lying down. Self explanatory. Results and Discussion Basic Motor Patterns of Copulation The basic motor patterns of copulation of roof rats were found to be almost identical to those of laboratory rats. Male roof rats mounted females several times prior to ejaculation. The male typically palpated the female's sides vigorously while making numerous rapid shallow pelvic thrusts. On some of these mounts the male gained vaginal penetration (intromission) , characterized by a deep pelvic thrust followed by a rapid dismount; on other mounts the male dismounted without achieving intromission (termed mounts). Usually after several mounts and intromissions the male mounted, achieved intromission and ejaculated. . The ejaculatory intromission was characterized by a much deeper pelvic thrust and a much slower dismount. The ejaculation was always followed by a period of sexual inactivity known as the postejaculatory interval (PEI) . An organized chain of copulatory events culminating in an ejaculation (and including the following PEI) is known as an ejaculatory series. Male roof rats were found to achieve an average of 4.3 such ejaculatory series before reaching sexual satiety.

PAGE 40

33 ' On three of the 302 ejaculatory series observed, three different males were observed to ejaculate on a single insertion, that is with no intromissions preceding the ejaculation. This pattern, although rare in roof rats, has never been observed in normal Norway rats. It is possible therefore that there is a fundamental difference between the two species with regard to this ability to ejaculate on a single insertion. In a very strict sense roof rats are categorized as pattern #15 in Dewsbury's taxonomy of copulatory behavior (Dewsbury, 1972b) because they show no copulatory lock, no intr avaginal thrusting, ejaculations can be achieved on a single insertion and because they typically attain multiple ejaculations. In that the phenomena of ejaculation on a single insertion was observed so rarely in roof rats, it might also be reasonable to classify roof rats as pattern # 13 animals. This pattern characterizes species such as Norway rats (Beach § Jordan, 1956), Syrian golden hamsters (Beach $ Rabedeau, 1959), and Mongolian gerbils (Kuehn $ Zucker, 1968) which have no lock, no intravaginal thrusting, multiple ejaculations and multiple intromissions always preceding ejaculation. The final resolution of this problem of classification awaits further data on other populations of roof rats. If the phenomenon of ejaculation on a single insertion proves to -be very rare, then it would probably be more reasonable to classify roof rats as pattern #13.

PAGE 41

34. Ewer's (1971) description of the copulatory behavior of a laboratory colony of roof rats and of a free-living population indicates that the animals she observed also displayed pattern #13 copulatory behavior. It is clear from her description of the behavior of the laboratory colony that the animals she observed had multiple intromissions and probably multiple ejaculations.' No mention is made of the presence of a lock or of intravaginal thrusting. She also notes that while complete "sequences" of copulatory behavior were not observed in the free-living population "All stages were, however, seen a number of times on different occasions" [p. 156]. When mounted by a male, the female roof rats demonstrated a lordotic posture similar to that displayed by female laboratory rats. The lordotic posture was characterized by an extension of the hind legs and elevation of the head accompanied by a spinal flexure resulting in an inverted arch in the female's back. The tail was deflected laterally. Prior to mounting, receptive females often displayed a series of behaviors including rapid ear wiggling, tail rattling and approaches to the male followed by rapid darting away and stopping several cm away. These behaviors were observed on every test in which the females were judged behaviorally receptive and were not observed in the few tests where the females were not behaviorally receptive. Females were judged behaviorally receptive solely on the basis of their responses to mounting by males: a behaviorally

PAGE 42

35 receptive female adopted a lordotic posture in response to male mounting. It is of interest to note that Ewer (1971). observed the patterns of darting and tail rattling in free ranging wild Rattus rattus females during copulatory sequences. Ewer makes no mention of the pattern of ear wiggling observed in the present population of animals. Darting and ear wiggling have been reported numerous times for receptive female laboratory rats (Beach, 1942; 1943; 1956) but tail rattling has never been reported for this species. Quantitative Measures of Copulatory Behavior Quantitative data for the copulatory behavior of roof rats were collected on 71 tests from twelve pairs of animals Each pair was to receive six tests each carried to a 30 minute satiety criterion; however one male died prior to his last test. Table 1 presents the means and standard errors of the mean for the standard measures of copulatory behavior for the first three arid the last three ejaculatory series. Standard errors were computed using a formula derived by Marks (1947) . This formula was developed for those situations in which there are unequal numbers of tests for different subjects. As can be seen from Table 1, it took male roof rats an average of approximately 1.5 miri to achieve their first mount, and a little over two minutes to achieve their first intromission from the start of the test.

PAGE 43

36 CO

PAGE 44

37 Once started, roof rats typically attained a mean of 4.3 ejaculations before reaching sexual satiety. Sexual satiety was attained between series on 46 of 71 tests or 64.81 of the time. This is to say, satiety was reached after the completion of an ejaculation and without the animal initiating another series, on 64.8% of the tests. On 35.2% of the tests, males initiated but did: not complete an ejaculatory series before reaching sexual satiety. During their first copulatory series, male roof rats had a mean of 5.1 mounts and a mean of 7.6 intromissions prior to ejaculation. The mean interval between intromissions was found to be a little over one min while the mean ejaculation latency was somewhat over nine min in length. The mean postejaculatory interval following the first series lasted approximately four min. Analyses of variance were performed for all eight of the standard measures of copulatory behavior to assess the effects of repeated testing on these measures. Those measures that occurred more than once per test (i.e. MF, IF, EL, Mill and PEI) were subjected to different analyses than those that did not (ML, IL, EF) in order to assess any changes that might occur across series as well as across tests. Male roof rats were found to be remarkably variable in the number of ejaculations that they might attain on any given test. Males were found to have as many as eight ejaculations or as few as one ejaculation prior to reaching

PAGE 45

38 sexual satiety. This variability in the number of series made analysis of data from all series for all tests on all subjects impossible. Therefore, only those animals having three or more series on their first four tests were included in the statistical analysis. Eight pairs of animals met this criterion and it is on the data from these animals that the following statistical analyses were performed. The results of these analyses of variance are summarized in Table 2. Complete ANOVA tables for each measure appear in Appendix A. One way analyses of variance for repeated measures were performed on the measures ML, IL and EF to assess the effects of repeated testing on these measures. As the results in Table 2 show, there were no significant test effects for any of these measures. It can therefore be concluded that no significant changes occurred with regard to these measures over the first four tests. For the measures MF, IF, EL, Mill and PEI, two way analyses of variance for repeated measures were performed. The results of these analyses also appear in Table. 2. On no measure was a significant test effect found. Significant changes across series were noted for the measures IF; EL and PEI but not for the measures Mill and MF. None of the tests by series interactions proved to be significant. The Newman -Keuls method (Winer, 19 74) was applied to the measures IF, EL and PEI to determine which series were

PAGE 46

39 Table 2 Results of Analyses of Variance of Eight Measures of Copulatory Behavior Behavioral Measure

PAGE 47

40 Table 3 Results of Newman-Keuls Analysis for Three Measures of Copulatory Behavior Behavioral Mean and Measure Series Standard Error Comparison r q ] IF 1

PAGE 48

41 significantly different from the others for each of these measures. The results of these analyses appear in Table 3. For IF and EL it was found that the means for the first series were significantly larger than the means for the second series and the third series. Means for series two were not significantly different from means for series three for either measure. For PEI, the mean for the third series was significantly greater than the means for the first and significantly greater than the means for the second. The mean PEI of series one was not significantly different from the mean PEI of series two. It can be concluded from these results that no significant changes as a function of test occurred over the first four tests in any of the eight standard measures of copulatory behavior. Significant changes across the first three series can be seen only for the measures IF, EL and PEI. For the measures IF and EL significant decreases are seen from first to later series. For PEI the third series shows a significant increase over the first two. The changes in the standard measures of copulatory behavior across series are of considerable interest. Several studies have dealt with the changes in copulatory behavior of laboratory rats as sexual satiety is approached (Beach $ Jordan, 1956; Dewsbury, 1968; Karen $ Barfield, 1975). This research was generated primarily by an interest in the nature of the underlying mechanisms responsible for such changes. It is therefore of some theoretical interest to

PAGE 49

42 compare the changes in the copulatory behavior of laboratory rats and roof rats as satiety is approached. The statistical procedures performed previously do not allow for an adequate analysis of changes in copulatory behavior as satiety is approached because only the first three series were analyzed and on many tests the animals completed more than three copulatory series. The means and standard errors presented in Table 1 provide some indication of changes that occurred in various measures as male roof rats approached sexual satiety. With this method of data presentation means and standard errors were determined for the first three series for all subjects. Then the means and standard errors were calculated for the last complete copulatory series (or Nth series) regardless of whether it was the eighth or the first series. In a similar fashion the next to last series (N-l) and the second from the last series (N-2) were computed. This system does not accurately portray the changes across series for those tests where considerably more or fewer than six series are needed to reach satiety. For example, on a test where only two series were completed before satiety was reached, data from those two series were represented twice as the first and second series respectively . and as the N-l and N series, respectively. For a test where eight series were completed the first three series were represented in series one, two and three, respectively and the last three series were represented in the series labeled N-2, N-l and N. The middle two series

PAGE 50

43 would not be represented at all. Given the considerable variability shown in ejaculation frequency across tests for roof rats another method of data presentation was employed for the analysis of changes in behavior as satiety was approached. This method was originally developed by Larsson (1956) and later refined by Karen and Barfield (1975) for use with laboratory rats. The essence of the system involves plotting the changes in behavior across series as a function of the number of completed series obtained prior to sexual satiety. Means were calculated for all tests from all animals that had only two series prior to satiety, all tests where the animals had only three series prior to satiety and so on until curves were generated for all tests where two, three, four, five and six ejaculations occurred prior to satiety. No curves were generated for those tests in which one, seven or eight ejaculations were obtained, there being too few such tests to construct meaningful curves. Finally, these separate curves were plotted on a variable axis, thus allowing direct comparison of all initial and terminal points for each measure regardless of the number of ejaculations received. Figures 1 through 4 present these data for the measures IF, EL, Mill and PEI , respectively. Figure 1 reveals that there was a clear decline in IF from first to later series. In addition, those tests where three, four, five and six ejaculations occurred prior to satiety showed: a small but consistent rise in IF during the terminal series.

PAGE 51

Figure 1. Intromission frequency across successive series for those tests in which 2, 3, 4, 5 or 6 ejaculations preceded sexual satiety. For the. group having 2 ejaculations, n = 10, for the group having 3 ejaculations, n = 14, for the group having 4 ejaculations, n = 16, for the group having 5 ejaculations, n = 14, and for the group having 6 ejaculations, n = 8.

PAGE 52

lO-i 45 IF &8 c
PAGE 53

Figure 2. Ejaculation latency across successive series for those tests in which 2 , 3, 4, 5 or 6 ejaculations preceded sexual satiety. For the group having 2 ejaculations, n = 10, for the group having 3 ejaculations, n ='14, for the group haying .4 ejaculations , n .= 16, for the group having 5 ejaculations, n = 14, and for the group having 6 ejaculations, n = 8.

PAGE 54

47 (1050)*-. \ EL (Sec) o CD CD E 600 5004003002001004 ' 3 Series 6 5 4 3 2

PAGE 55

Figure 3. Mean interintromission interval across successive series for those tests in which 2, 3, 4, 5 or 6 ejaculations preceded sexual satiety. For the group having 2 ejaculations n = 10, for the group having 3 ejaculations, n = 14, for the group having 4 ejaculations, n = 16, for the group having 5 ejaculations, n = 14, and for the group having 6 ejaculations , n = 8 .

PAGE 56

2049 ..,|(200) Mill (Sec) ..•* 140120100o CD CO CD e 8060402 ' I 3 4 . « 2 3 4 2 3 2 6 5 4 3 2 Series

PAGE 57

Figure 4. Postejaculatory interval across successive series for those tests in which 2, 3, 4, 5 or 6 ejaculations preceded sexual satiety. For the group having 2 ejaculations n = 10, for the group having 3 ejaculations, n = 14, for the group having 4 ejaculations, n = 16, for the group having 5 ejaculations, n = 14, and for the group having 6 ejaculations, n = 8.

PAGE 58

800 } 100051 P'EI (Sec) o in £ i800600400200^6 Series

PAGE 59

; 52' The graphs for EL in Figure 2 and for Mill in Figure 3 reveal very similar U-shaped functions for most tests. For the tests in which three, four, five and six ejaculations occurred the EL showed a decline from first to middle series and then an increase in the terminal series. For those tests in which only two ejaculations took place, EL declined from first to last series. The measure Mill is seen in Figure 3 to have declined from first to middle series and then increased rapidly in the later series for those tests in which three, four, five and six ejaculations occurred. For those tests in which only two ejaculations occurred the Mill increased from first to last series. While the measures IF, EL and Mill appeared to display more or less U-shaped functions the measure PEI seen in Figure 4 appeared to have an entirely different , function. As satiety was approached, the PEI tended to increase across successive series, rather than decrease and then increase. It. had been noted earlier that with regard to the basic motor patterns of copulation, roof rats were found to be very similar to laboratory rats. Roof rats also appear to be quite similar to laboratory rats with regard to the quantitative measures of copulatory behavior. Table 4 compares the present quantitative data for roof rats with those from three different strains of laboratory rats in two different studies by Dewsbury (1968; 1975b). The measures compared are ML, _IL, MF, IF, EL, Mill and PEI from the first

PAGE 60

53 Table 4 Comparisons of the Standard Measures of Copulatory Behavior Between Roof Rats (RattUs rattus) and 3 Strains of Behavioral Measure

PAGE 61

54 copulatory series and EF for those strains tested to satiety, These data were chosen for comparison because they were collected in the same laboratory in which the roof rat data were collected and because procedures were much the same for all three studies. The ML and IL for the LEW strain were very close to those observed for roof rats while the other measures differed considerably between the two species For the measures MF and Mill both the F344 strain and the Long-Evans strain appear close to the range of scores for roof rats. For the measures EL and PEI the F344 strain more closely approximates the scores shown by roof rats. The intromission frequency in the first series appears somewhat lower for roof rats than any of the three laboratory rat strains presented here. The ejaculation frequency also appears somewhat lower for roof rats than for the Long-Evans strain. Comparisons of EF's from three other studies show that laboratory rats tested in different laboratories with varying amounts of prior sexual experience do show some variability in this measure. Beach and Jordan (1956) tested both albino and hooded strains and found the average EF to be 6.2 and 6.4 in two different experiments. Dewsbury tested Long-Evans rats with varying amounts of prior sexual experience and found the EF to vary between 5.5 and 7.6. Males with very little sexual experience had lower EF's than those who had considerable experience. These scores are still somewhat higher however, than the EF of 4.3 observed for roof rats.

PAGE 62

55 As the data in Table 4 and the above mentioned data for EF illustrate, there is some variability in the copulatory behavior of laboratory rats tested in different laboratories. Some of this variability is due, no doubt, to differences in housing, maintenance and testing procedures of the animals, prior sexual experience and other environmental variables. Some of the variability is also due to genetic differences between strains of laboratory rats. Several studies have shown there to be quantitative differences in the copulatory behavior of different strains of lr.bjratory rats (Dewsbury, 1975b; McLean, Dupeire $ Elder, 1972; Khalen, 1961). This variability within laboratory rats makes it difficult to assess possible species differences between roof rats and laboratory rats with regard to the quantitative measures of copulation. To adequately compare the behavior of these two species it would be necessary to rear and test representatives of both species under identical conditions. Given the strain differences in behavior it would also be necessary to sample a wide range of domesticated and wild populations of Rattus norvegicus and to also sample a range of wild populations of Rattus rattus . Under these conditions it would be possible to draw firm conclusions about species differences in the quantitative aspects of copulatory behavior . Despite the limitations of comparing the behavior of different species from different experiments, several points can be made from the data presented in Table 4. In that the

PAGE 63

56 mean ML and Mill for roof rats fall within the range of means presented for laboratory rats, it is probable that the two species do not differ with regard to these measures. The measures MF, IF and EL show considerable individual variability for the roof rats tests as can be seen by the high standard errors. In fact 24 ? o of the IF's, 26% of the MF's and 30% of the EL's collected for roof rats fall within the range of scores presented for laboratory rats. These data suggest that with regard to these three measures roof rats and laboratory rats may not differ. There is less individual variability within the measures IL, PEI and EF. It is possible that the differences seen between the two species with regard to these measures may be due to real species differences. It is also possible that the differences, should they prove to be reliable, are not due to species differences but to the effects of domestication on the behavior of laboratory rats. There are at present no data on the copulatory behavior of wild Rattus norvegicus and it is possible that wild Norway rats are even more similar to wild roof rats than are domesticated Norway rats. Comparisons among wild and domesticated stocks of Norway rats and wild stocks of roof rats would provide evidence for this suggestion. It can be concluded from this comparison of the quantitative measures of copulatory behavior of roof rats and laboratory rats that in five of the eight measures compared there appears to be very little difference between the two

PAGE 64

57 species. With regard to the other three measures, roof rats and laboratory rats do not appear radically different, but small reliable species differences might exist. With regard to changes that occur in the copulatory behavior of roof rats as exhaustion is approached, a similar conclusion can be drawn. The curves generated for IF, EL, Mill and PEI in Figures 1-4 for roof rats are almost identical in form to those presented by Karen and Barfield (1975) for laboratory rats. They are also similar in form to those functions described by Dewsbury (1968) for laboratory rats using a slightly different method of presentation and to those presented by Larsson (1956) . It should be noted that Larsson did not in fact conduct standard satiety tests with his animals but only tested the animals for a one hour period. Nevertheless Larsson's functions for the measures IF, EL and PEI are strikingly similar to the present ones for roof rats. These facts suggest that similar mechanisms underly the sexual arousal and exhaustion of roof rats and laboratory rats. Theoretical speculations concerning the nature of these mechanisms can be found in papers by Beach (1956) , Beach and Jordan (1956) , Reach and V.'halen (1959), Dewsbury (1968), Karen and Barfield (1975) , and McGill (1965) . Categorization of Behavior A categorization of the behaviors accompanying copulation was made on one test each for 11 males and 11 females. A twelfth pair of animals was not categorized due to the death of the male prior to categorization.

PAGE 65

58 Figures 5 and 6 present the categorization data for males and females, respectively. Data from 15 of the possible 16 mutually exclusive categories of male behavior are presented in Figure 5. The sixteenth category, Dig, was never found to occur more than 1% of the time and was dropped from further analysis. The data in Figure 5 are presented as per cent of time spent in each category during the IL, first, second and last EL, first and second PEI and the 30 minute satiety criterion. All of the periods are further divided into quarters with the exception of the IL. The data are expressed as means and standard errors. During the intromission latency period males spent about 85% of their time in activities directed at the female: sniffing, genital and general grooming and chasing. Particularly frequent were the episodes of sniffing and chasing. The EL periods were predominated by copulation related activities such as chasing, mounting, pulling on the female, sniffing and genital grooming of the female. Males also spent extended periods of time grooming their own genitals; an activity highly associated with copulatory activity. Males spent from 13-23% of their time in locomotor exploratory behavior and 6-171 of their time sitting. The majority of the postejaculatory intervals were spent in activities such as locomotor-exploratory behavior, sitting and general grooming. Fourteen percent of the time was spent in genital grooming and only 2-4% of the time was spent chasing and mounting females. Grooming and sniffing females

PAGE 66


PAGE 67

60 3WIJL JO lN33H3d

PAGE 68

+-> Q •H . d) o U -H 3 U U) O rt ft d> S u S J! -H ' U +> o . o rt a> p u cr o px • rt u w r-t C• 4-lrt tf) O 4-> OTJ CPCI O rt •h m. 4-> O W rt pi t-3 +-> rt •H C 0) p e/> (D t/r f-i tn (D ft oo. ft o> H X p. u<. a) rt

PAGE 69

62 3WI1 JO ±N30H3d

PAGE 70

63 occupied 7-13% of the time. Mounting and chasing both tended to occur in the final quarter of the PEI just prior to the next series. Very little time was devoted to copulation-related activities such as chasing and mounting during the satiety period as compared to the EL periods. What copulation-related acitivity did occur, was restricted to the first two quarters only. About 65$ of the period was taken up with locomotor-exploratory behaviors and sitting. Figure 6 presents the percent of time females spent in 13 of the 15 possible categories of female behavior. The data are presented as means and standard errors. The categories of Dig and Lie were excluded because during no period did either category occupy more than 1% of the time. During the IL period females behavior appeared to be directed at the male. Sniffing, genital and general grooming, being groomed, upright defense, running from the male and lordosis accounted for 58% of the period. The remainder of the period was spent in locomotor-exploratory behaviors, feet on walls and climbing. Females spent from 16-35% of their time during the EL periods in active copulation or activities involving the male-including running, lordosis, being groomed, upright defense and grooming the male. Sitting and locomotor-exploratory activities (not including climbing) accounted for 25-531 of the females' time in this period. Climbing behavior was observed to occur interspersed in bouts of both locomotor-exploratory and copulatory behaviors. Often, especially in later series, females would climb on the cage top in an apparent attempt to avoid the chasing

PAGE 71

64 male. Climbing was observed to occur from 22-381 of the time during EL periods. The majority of the two PEI periods were devoted to locomotor exploratory behaviors and climbing. Only 1-51 of the time was spent running from males or in lordosis; these behaviors occurred during the fourth quarter of the periods prior to the next series. During the satiety period females spent the majority of their time in locomotor-exploratory behaviors and climbing. Thirteen percent of the time was spent in sitting. Only 4% of the time was devoted to running from the male and lordosis; this occurred primarily during the first two quarters. Figures 7 and 8 present the categorization data for males and females in a slightly different fashion but one that facilitates interspecific comparisons. For the males, 10 of the 15 categories presented in Figure 5 have been collapsed into six mutually exclusive categories and overall means calculated for each time period. The category locomotor-explore includes locomotor-explore, climb and feet on walls. Chase-mounts includes chase, mount and pull female. The categories general groom self, genital groom self, sniff and sit are the same as in Figure 5. For females 10 of the 13 categories in Figure 6 have been collapsed into nine categories in Figure 8. Locomotor-explore includes locomotor-explore and feet on walls. The categories run from male, lordosis, sniff, genital groom self, generalgroom self, climb and sit are the same as in Figure 6.

PAGE 72

XI

PAGE 73

66 *-&/////?///. mm?i i i F *w> >w&«m £ -C j i i i \ mzzfa m& t o o o o o o IO ^jI*} GJ — o o o o o o iO ^ro cvl — <

PAGE 74

rC

PAGE 75

68 1

PAGE 76

69 The collection of categorization data for laboratory rats by Dewsbury (1967) allows detailed comparisons to be made with the present data for roof rats. A few qualitative and quantitative differences in behavior between the two species are apparent. However, as with the other aspects of copulatory behavior discussed thus far, the most striking features of these comparisons are the great similarities in behavior. The copulation-related behaviors of chase-mount in males and lordosis in females were almost identical in quantitative as well as qualitative comparisons between the two species. Similarly the behaviors of sniffing, locomotor-explore and non-genital grooming, are quite similar for males of both species. The patterns of genital grooming and general grooming, sitting, upright defense and lying -down were also similar for females of both species. The patterns of change in behaviors across time periods (i.e., from IL to EL to PEI) within the first series were very similar for males of both species with regard to the behaviors genital grooming and sitting. Both species showed '. ' ' ' f -.'.. , more genital grooming during the EL period and more sitting during the PEI period. Patterns of change across time periods were also very similar for laboratory rats. and roof rat females for the behaviors sniffing, run from male and locomotor-explore. Quantitatively, laboratory rat males tended to spend more time genital grooming during EL and PEI periods than did roof rat males. However, roof rat males tended to spend

PAGE 77

70 more time sitting during these two periods than did laboratory rat males. Female roof rats tended to spend much more time running from males during the IL and EL periods than did female laboratory rats. However, female laboratory rats tended to spend about twice as much time in locomotor-exploratory activities during all time periods than did female roof rats. Qualitatively, the only major differences between the two species appears to be with regard to the occurrence of allo-grooming in roof rats. Male and female laboratory rats are not reported to engage in genital and general grooming of the partner while male and female roof rats engage in low but consistent levels of these behaviors. Climbing behavior is also not reported for laboratory rats; however, Dewsbury's testing situation provided no opportunity for his subjects to climb. It is doubtful whether they would climb even if given the opportunity. On several occasions laboratory rats have been tested in the roof rat testing cages and at no time have laboratory rats been observed to climb on the cage top. Quantitatively and qualitatively roof rats appear very similar to laboratory rats with regard to the behaviors accompanying copulation. The differences between laboratory rat and roof rat females in running from the male may be related to the occurrence of the darting, ear wiggling and tail rattling pattern observed in female roof rats that was discussed earlier. The occurrence of this presumed solicitation pattern may indicate a real species difference

PAGE 78

71 between roof rats and Norway rats. It may be that roof rat males require more stimulation from the female in order to initiate copulation than do laboratory rats. It is also possible that the same behavior patterns occur in wild Norway rats but that through domestication, these patterns have been lost. There are no clear explanations for the other differences observed between the two species. Ultrasonic Vocalizations Ultrasonic vocalizations were observed on a total of 22 tests for 11 pairs of animals, each pair being observed on two tests. These were the same tests during which the categorizations of behavior were done: the fifth and sixth positive tests for each pair; Identification of individual animals producing ultrasonic vocalizations in social situations is sometimes difficult because such vocalizations can be produced by more than one animal at one time in a given test. Using only the ultrasonic receiver it is impossible to localize the sounds well enough to positively identify the source of the vocalizations. In most cases, however, it was possible to. correlate the occurrence of the ultrasonic vocalization with rhythmic chest movements in one but not the other animal. These chest movements, probably rhythmic exhalations producing the vocalizations, are also seen in isolated animals during vocalizations. Singing was never observed to occur in the absence of such chest movements by one or the other animal.

PAGE 79

72 During pretesting, scans of the frequency range from 20 kHz to 60 kHz during various phases of the copulatory sequence revealed the only observable vocalizations to occur at approximately 28 kHz. As a result during actual testing the frequency dial of the ultrasonic receiver was set at about 28 kHz and the occurrence and duration of all vocalizations occurring in this range were recorded on one channel of the event recorder. This procedure minimized the probabilities of detecting ultrasonic vocalizations occurring at other frequencies and it is therefore possible „, that vocalizations were made at other frequencies but went undetected. Sewell (1967) for example reports male vocalizations in the 50 kHz range during mounting activity in laboratory rats. Both males and females were observed to emit ultrasonic vocalizations during copulatory sequences. For both males and females, vocalizations were observed at 28 + 5 kHz in frequency. These vocalizations occurred in pulses of 1-3 sec in length with pulses occurring in trains of variable length; some up to 3 min in length on occasion. Some sex differences were noted in the occurrence of ultrasonic vocalizations. Maleswere observed to vocalize only during PEIs and the satiety criterion period. Ultrasonic vocalizations followed 911 of the 83 total ejaculations observed on these 22 tests. On 8 tests male vocalizations were observed that were not directly associated with ejaculations or other copulatory activities.

PAGE 80

73 All of these vocalizations occurred during the satiety criterion period. Twenty-two percent of these vocalizations appeared to be correlated with sudden noises occurring in the testing room. Twenty-five instances of female vocalization were observed on 14 of the 22 tests. Female vocalizations, unlike male vocalizations, were observed during EL periods as well as PEI periods and the satiety periods. However, only 24% of the females' vocalizations occurred during EL periods; the other 751 occurred during PEIs and the satiety: period. Only 20% of all female vocalizations appeared to be correlated with sudden noises. Table 5 presents quantitative data (means and standard errors) on male postejaculatory vocalizations from eight subjects on 16 tests. Only those subjects having at least three series on both tests were included in order to facilitate analysis of changes in behavior across series. The measures of vocalization and other copulatory measures have been defined previously. The numbers in parentheses are means and standard errors for data collected by Barfield and Geyer (1975) on laboratory rats and have been included to facilitate cross-species comparisons. As can be seen from the table, the measures IF and EL show significant decreases from first to second series. The measures VT and PEI show significant increases from first' to second series. The measures VL, PEI-VT and VT/PEI do not change significantly from first to second series. It

PAGE 81

74 Table 5 Means and Standard Errors of the Quantitative Measures of Male Postejaculatory Vocalization and Other Measures of Copulatory Behavior and Results of T-tests Behavioral Series t^-test for Measure 1 2 related pairs (df > 7) (174.6 + 45.6) IL 45.0 + 19.8 ( 11.1 + 1.5) ( 5.2 + 0.5) IF 8.8 ± 1.0 3.4+0.3 t = 5.35 (859.0 + 121.0) (337.0+33.8) EL 595.9 + 152.5 106.9+12.5 t =3.05 ( 36.0 + 4.0) (42.8 + 10.2) VL 36.4 + 6.8 41.4 +5.7 t = 0.65 (328.0 + 14.9) (415.0 + 19.0) VT 161.6 + 18.9 210.5 + 25.5 t = 3.i6 * (429.0 + 18.9) (528.0 + 19.9) PEI 256.8 + 1.9.0 294.9 + 120.2 t = 4.51 (101.0 + 15.1) (117.0 + 18.5) PEI-VT 95.2 + 17.6 85.0 + 15.5 t = 0.89 (0.77+0.03) ( 0.79 + 0.3) VT-PEI 0.63 + 0.07 0.71+0.06 t = 1.84 Note: N = 8, 2 tests per animal. Measures in parentheses are for Norway rats from Barfield and Geyer, 1975. * p < .02 ** p < .01

PAGE 82

75 takes male roof rats from 36-41 sec to initiate, ultrasonic vocalizations following an ejaculation and such vocalizations continue for 2.5 to 3 min. Roof rat males spend from 63-71% of their total PEI in vocalization activity. When the present data are compared to those of Barfield and Geyer (1975) for laboratory rats, striking similarities are seen between the two species. The measures VL and PEI/VT are almost identical in the two species. Laboratory rats tend to spend more absolute time in vocalization than do roof rats, as measured by VT. However, the total PEIs for laboratory rats are considerably longer than those of roof rats. When the percent of the PEI spent in vocalization is examined it reveals a much closer similarity between the two species. Roof rats spend 631 of their first PEI in vocalization; laboratory rats spend ; 771. Roof rats spend 71% of their second PEI in vocalization, laboratory rats 79% . It has been suggested that in laboratory rats the measures VT and PEI/VT are fair estimates of the absolute and relative postejaculatory refractory periods, respectively (Barfield $ Geyer, 1975; Karen $ Barfield, 1975). Similarly VT/PEI has been suggested to be a reasonable estimate of the percentage of the PEI taken up by the absolute refractory period. Again, the close correspondence between these two species with regard to these measures suggests that similar mechanisms underlie the arousal, maintenance and exhaustion of their copulatory behavior.

PAGE 83

76 Barfield and Geyer (1972, 1975) report that male laboratory rats are most often inactive during ultrasonic vocalizations, lying quietly or occasionally grooming themselves or moving slowly about. Similar behavior for female laboratory rats during male ultrasound is also reported by these authors with females reported "...also often quiescent"(Barfield § Geyer, 1975, p. 726). The collection of categorization data concurrently with ultrasonic vocalization data provides an excellent opportunity to quantify the behaviors of male and female roof rats during ultrasonic vocalizations. The behavior of 11 males and 11 females during male postejaculatory vocalizations was categorized on one test for each male and each female.. The per cent of time males spent in six categories of behavior during the first, second and terminal PEI vocalizations is presented in Figure 9. The per cent of time spent by females in eight categories of behavior during the male's first, second and terminal PEI vocalizations are presented in Figure 10. Only six of the possible 16 categories of male behavior and only eight of the possible 15 categories of female behavior are presented because in both cases the remaining behaviors occurred no more than II of the time. Figure 9 reveals that male roof rats spent the majority of their vocalization time sitting. The percentage of vocalization time spent sitting increased over successive series. Less than 121 of the vocalization time was spent in

PAGE 84

.

PAGE 85

78 in c o> o C. =3 o o JC o CD J^ o to o CL o -e: 5= e: fc: ,,35 1115 l«l'»'3 1!S •d«3--=oT JI'S »0 l«!«»3 O O UJ 00 9UUJ1 p iU30J9d

PAGE 87

80

PAGE 88

locomotor-exploratory behavior, the remainder was spent in quiescent activities such as grooming, sitting and lying. This is very similar to the reports of Barfield and Geyer (1972, 1975) with regard to laboratory rat activities during vocalization. One point of difference is worthy of note. Barfield and Geyer report that laboratory rats spend much of their vocalization time lying down. Roof rats spend very little of their vocalization time lying down. During the majority of their vocalization time, male roof rats sit in a relaxed but alert posture. Indeed, on several occasions vocalizations were observed to be interrupted for short periods of time by sudden noises in the testing room. Male roof rats appeared to be immobile during periods of vocalization, but are not totally unresponsive to environmental stimuli. The activities of female roof rats during male vocalizations appear to be more variable. Figure 10 shows that female roof rats spent the majority of the vocalization time in such non-quiescent activities as locomotorexploratory behavior, feet on walls and climbing during the first two vocalizations. However, during the last vocalization, females spent much more of their time in such quiescent activities as sitting and self-grooming. Female roof rats appear to be much more active than female laboratory rats during the early postejaculatory vocalizations of the male. However, during the final postejaculatory vocalizations, female roof rats appear more similar to female laboratory rats spending long periods of time in quiescent behaviors.

PAGE 89

Overall, the postejaculatory vocalizations of roof rats appear remarkably similar to those reported for laboratory rats in their quality and their temporal patterning. The biological function of male postejaculatory vocalizations is not clear. Barfield and Geyer (1972) note that in laboratory rats 22 kHz vocalizations can be observed in a number of different social situations. They suggest that the 22 kHz frequency may be "a ...'carrier frequency' for signals denoting states of contact avoidance" (p. 1350). Specifically in sexual situations they suggest it may function to inhibit female sexual behavior during periods when the male is incapable of sexual activity. In a later paper (Barfield § Geyer, 1975) they suggested that the postejaculatory vocalizations might have a facilitative effect on reproduction by affecting the reproductive physiology of the female. Barfield and Geyer also noted that sounds have been shown to be significant in initiating specific reproductive responses in the females of some species of birds (Barfield, 1971). It was suggested that postejaculatory vocalizations by male rats may have a similar function. Another possibility not mentioned by Barfield and Geyer concerns the reduction of competition between conspecific males. If 22 kHz vocalizations really do communicate a state of contact avoidance in laboratory rats it is quite possible that this might keep other conspecific males away from the vocalizing male and the female with which he is mating. This might prevent a take-over of the female by a strange

PAGE 90

83 male when the vocalizing male is incapable of sexual activity, thereby reducing competition between the vocalizing male and others. Further research should clarify the adaptive significance of posejaculatory vocalizations in the reproductive process.

PAGE 91

EXPERIMENT 2 The results of Experiment 1 demonstrated that the copulatory behavior of roof rats is fundamentally the same as that of laboratory rats. However, the first experiment did not provide any indication as to the importance of various aspects of the male's behavior in the initiation of successful pregnancy in female roof rats. As was pointed out earlier, Adler (1969) , Chester and Zucker (1970) and others have shown that for laboratory rats, a single ejaculatory series provides sufficient stimulation to induce successful pregnancy in 80-901 of the females. In contrast, Dewsbury and Estep (1975) and Lanier et'al. (in press) have demonstrated that in cactus mice ( Peromyscus eremicus ) and Syrian golden hamsters ( Mesocricetus auratus ) , copulations beyond the first ejaculation are essential to maximize the probability of successful pregnancy. Although roof rats appear quite similar to laboratory rats in their copulatory behavior, it is possible that quite different aspects of the copulatory pattern are important for successful pregnancy in roof rats. The present experiment was designed to assess the significance of multiple ejaculations in the initiation of pregnancy in roof rats. Comparisons are made between females given just one ejaculation and those given a complete satiety test with regard to the initiation of pregnancy. 84

PAGE 92

85 Materials and Methods Subjects The subjects of Experiment 2 were nine male and sixteen female laboratory-reared offspring of wild trapped roof rats. All of these animals were bred, weaned, and maintained in the same fashion as those of Experiment 1. Apparatus Behavioral tests were conducted in the same clear, plastic cages as used in Experiment 1. Behavioral events were also recorded in the same, fashion as Experiment 1 using the event recorder . Procedures Daily vaginal smears were taken on all females starting when the females were 60 days of age and continuing throughout the study. Smears were taken between 0800 and 1100 hrs using a thin wire loop. These vaginal smears were examined microscopically and a determination was made of the stage of the female's estrous cycle on the basis of the cell types present in the smear. The stages of the estrous cycle were classified according to the criteria used by Hardy (1972) for laboratory rats. These were: proestrus, predominance of nucleated epithelial cells; estrus, predominance of cornified epithelial cells; metestrus, cornified and nucleated epithelial cells and some leucocytes; diestrus, preponderance of leucocytes with some nucleated epithelial cells.

PAGE 93

86 After the females began to show regularity in their estrous cycles, they were pretested for copulatory behavior and fertility. On the afternoon of vaginal proestrus, the proestrus female was placed in a testing cage with a male and, if receptive, was allowed to mate for at least three copulatory series. If the female was not behaviorally receptive, she was returned to her home cage and retested at the next occurrence of vaginal proestrus. Those females showing no regular estrous cycles, having three consecutive negative tests, or mating but not becoming pregnant and delivering pups after two mating tests were eliminated from the study. This system of pretesting allowed positive verification of fertility for both males and females and allowed both males and females to gain experience in the standard testing situation. Females judged fertile by this method and which demonstrated two consecutive estrous cycles of 3-5 days duration were tested twice, once in each of two conditions. Half of the females were randomly assigned to first serve in the one ejaculation condition (1 Ejac), the other half served first in the satiety condition (Satiety) . The design was counterbalanced so that those animals tested in the 1 Ejac condition on their first test served in the Satiety condition on their second test and vice versa. All of the animals were between 123 and 192 days of age at the time of their first test and between 179 and 240 days of age at the time of their second test.

PAGE 94

Mating tests were conducted on the afternoon of vaginal proestrus approximately three hours into the dark phase of the light cycle. The testing procedure was similar to that of Experiment 1 with males introduced into the test cages 5-10 minutes before the females. Testing started with the introduction of the female. If the male failed to gain an intromission within 30 minutes after the start of the test, the test was terminated and the female was retested at the next occurrence of vaginal proestrus. Each female assigned to the 1 Ejac condition was allowed to mate with a proven fertile male until he had achieved one ejaculation. At this point the test was terminated and the female was returned to her home cage. Each female assigned to the Satiety condition was allowed to mate with a proven fertile male until he had reached an arbitrary satiety criterion of 30 minutes without an intromission. The test was terminated at this time and the female was returned to her home cage. . Following a positive test a female continued to receive daily vaginal smears and the presence or absence of regular estrous cycles was noted. Females not becoming pregnant were allowed to display at least 2 complete 3-5 days cycles before being retested. Those becoming pregnant were allowed to deliver their litter, then retested after the resumption of regular 3-5 day estrous cycle activity. Females displaying at least 12 consecutive days of diestrus following a positive mating test but not delivering pups were judged to be

PAGE 95

pseudopregnant . Those displaying diestrous smears for 20 consecutive days and delivering pups between days 21 and 22 were judged to have become pregnant as a result of the mating. Those females continuing to show regular 3-5. day cycles following mating were judged to be neither pregnant nor pseudopregnant. Behavioral Measures The standard measures of copulatory behavior were taken on all tests. Results and Discussion A total of 16 females completed both tests and are included in the final data analysis. Table 6 presents the percentages of tests in which females became pregnant, pseudopregnant or continued to cycle as a result of receiving one ejaculation or satiety. A binomial test (Siegel, 1956) for the proportions of females becoming pregnant in each of the two test conditions, revealed that there was no significant difference between the 1 Ejac and Satiety conditions in the proportions of females becoming pregnant (X = 1, N = 7; p = .06, 1-tailed) . A binomial test for the proportions of females becoming pregnant in each testing order also revealed no significant differences (X = 2, N = 7; p = .45, 2tailed) . Although there were no statistically significant differences between the 1 Ejac condition and the Satiety condition in the proportion of females becoming pregnant,

PAGE 96

89 Table 6 Percent of Tests in Which Females Became Pregnant, Pseudopregnant or Continued to Cycle as a Result of Differing Amounts of Copulatory Stimulation a b Outcome 1 Ejaculation Satiety Pregnant 56 : 87 % Pseudopregnant '< . , '; 13 13 % Cycling 31 Mean Litter Size .8.0 + 0.6 7.1 + 0.7 (n = 9) (n = 14) Note, n =16 for each condition except as noted for litter size data a IF = 9.5 b EF = 5.4

PAGE 97

90 it is interesting to note that there appeared to be a trend for a higher proportion of females to become pregnant after receiving multiple ejaculations. Also worthy of note is the fact that all of the females allowed to mate to satiety showed a progestational response (pseudopregnancy or pregnancy) while only 681 of those females receiving one ejaculatory series showed a progestational response. This response is an essential prerequisite for normal pregnancy in the female laboratory rat and, it can be assumed, for all female mammals not having a functional luteal phase in their estrous cycles, such as roof rats. A t-test for unrelated samples comparing litter sizes for those females that had litters in each condition, revealed no significant differences in litter size as a function of different numbers of ejaculations t(21 = 0.96; p < .05, 1 -tailed) . Table 7 reveals that within the 1 Ejac condition there was a trend for females receiving high numbers of intromissions to become pregnant or pseudopregnant while those receiving few intromissions prior to ejaculation tended to continue cycling. A t-test for unrelated samples indicated that those females becoming progestational had significantly more preejaculatory intromissions than those females continuing to cycle (Mean if for progestational females 11.7, Mean IF for cycling females 5.4, t(14) = 5.06, p< .001) Thus, as in laboratory rats, there appears to be a direct relationship between the amount of copulatory

PAGE 98

91 Table 7 Mean IF and Resultant Female Response in 1 Ejaculation Condition Outcome a IF Pregnant Pseudopregnant Cycling 11.4 + 0. 13.0 + 1.0 5.4 ± 1.0 a Data presented as means and standard errors

PAGE 99

92 stimulation received by a female and the probability that she will become pregnant. Although roof rat females can become pregnant after a single ejaculation provided that they receive a sufficiently high number of preejaculatory intromissions, in the average first series females only receive 7.4 10.1 intromissions. In the present experiment the average first ejaculatory series provided only enough stimulation to induce pregnancy in 561 of the roof rats so stimulated. This percentage was increased to 871 when the . copulatory stimulation was increased to satiety, an average of 5.4 ejaculations. Because there was no statistically significant difference between the Satiety and 1 Ejac conditions with regard to the proportion of females that became pregnant, it appears that like Norway rats, roof rats reach maximal probabilities of pregnancy with a single ejaculatory series. However, it is difficult to believe that the 56% pregnancy rate in the 1 Ejac condition constitutes the maximal probabilities of pregnancy. It is quite possible that given a larger sample size the proportion of females pregnant in the 1 Ejac condition would be significantly different from the proportion of females pregnant in the Satiety condition. It should also be noted that all of the work done thus far by Adler, Chester and Zucker and others on the functions of multiple ejaculations in Rattus norvegicus has been done on domesticated forms. To date no one has examined the functions of multiple ejaculations in wild Rattus norvegicus .

PAGE 100

93. It is possible that the lack of a function for multiple series in this species, with regard to the induction of pregnancy, is an effect of domestication. That is, multiple ejaculations may be necessary for the induction of pregnancy in wild Norway rats but through the process of domestication this function has been lost. Domestication has been found to affect a wide variety of behaviors in both Norway rats and house mice ( Mus mus cuius ) (Barnett, 1963; Boice, 1972,1973; Lockard, 1968; Smith, 1972). There is even some evidence to suggest that domestication has affected the copulatory behavior of house mice (Estep, Lanier § Dewsbury, in press). Research on the induction of pregnancy in wild Norway rats should-provide evidence to confirm -or deny the possibility that domestication has affected the stimulus requirements for the normal initiation of pregnancy in this species. The 32 tests of copulatory behavior conducted in this experiment provide an opportunity to compare the behavior of male roof rats paired with females in natural estrus to those paired with females in an artificially induced estrus. The females tested in this experiment were of necessity in natural estrus while those in Experiment 1 were brought into estrus with the aid of exogenous hormones. Tables 8 and 9 present data from males mated with hormone -induced estrous females, from Experiment 1 and data from males mated with females in natural estrus from the 1 Ejac and Satiety conditions, of the present experiment. In Table 8 means and standard errors for the latency measures

PAGE 101

94 Table 8 Comparison of the Behavior of Males Tested with Females in Artificial vs Natural Estrus for Measures in the First Series Natural Estrus Behavioral Artificial . , . Measure Estrus Satiety 1 Ejac a ML 98.5 + 18.6 104.8 + 21.1 85.6 + 19.9 a : '".'...-.>•-'.. IL 146.3 + 28.0 124.6 +30.2 96.9 + 20.6 EF 4.3 + 0.4 5.4+0.4 MF 5.1 + 1.2 3.7 + 0.7 8.1 + 3.1 IF 7.6 + 0.7 10.1 + 0.8 9.5 * 0.9 EL 563.4 + 118.5 621.8 + 86.1 863.7 + 176.1 a Mill 70.3 +9.3 64.0 + 7.6 82.7 + 12.6 PEI 269.2 + 15.5. 281.1 + 18.9 ------Note. All data presented as means and standard errors a All time measures expressed in seconds

PAGE 102

95 H

PAGE 103

96 and the measures of the first ejaculatory series are presented. The data on artificial estrous behavior came from Table 1 of the first experiment and are based on 71 tests from twelve pairs of animals. The data on natural estrous behavior came from the present experiment and are based on sixteen pairs of animals, each tested once in both conditions Analyses of variance were performed to determine whether estrous state of the female did in fact affect the standard measures of copulatory behavior. The data for artificial estrous tests were from Experiment 1. Data from the first two series of the first test from all twelve pairs of animals were used. In Experiment 2, ttests for related samples indicated that there were no significant differences between the 1 Ejac condition and the Satiety condition for the eight standard measures of copulatory behavior. As a result, the data from the first two series on the 16 tests of the Satiety condition constituted the data for the natural estrous condition. One-way analyses of variance were applied to the measures occurring only once per test (i.e., ML, IL and EF) . Two way analyses of variance were used on the remaining measures that occurred more than once per test, the results of these analyses of variance can be seen in Table 10. Complete ANOVA tables appear in Appendix B. As can be seen from Table 10, significant changes from first to second series can been seen for the measures MF, IF and Mill. Estrous condition of the female apparently

PAGE 104

97 Table 10 Results of Analyses of Variance Comparing Different Estrus Conditions Behavioral

PAGE 105

98 affected only two of the eight measures: IL and IF. For IL, males mated to natural estrous females initiated copulation much sooner than those males mated to females in artificial estrus (Mean IL, natural estrus = 124.5, Mean IL artificial estrus = 322.6). For the measure IF a significant interaction of estrous condition with series was found. Analyses of variance for simple effects of changes across series indicated that in both conditions there was a significant decrease in IF from first to second series (natural estrus F(l,27) = 91.98, p<.001; artificial estrus F(l,27) = 18.67, 2 <.001). However as multiple t_-tests revealed, there was no significant difference in the second series IF between conditions (t_(27) = 0.1), while the first series IF for the natural estrous condition was significantly larger than the first series IF for the artificial estrous condition (t_(27) = 3.8, jd<.001). This indicates that the interaction was due to a much greater drop in IF from first to second series in the natural estrous condition than in the artificial estrous condition. As there was no difference between estrous conditions with regard to the second series IF, estrous condition appears only to affect the first series IF. It is not clear why estrous condition of the female should affect only the measures IL and the first series IF. If estrous condition had a general facilitory or inhibitory effect on the copulatory behavior of a given pair of animals it might be expected that several measures of

PAGE 106

99 copulatory behavior would be affected, and that those measures would be affected in a similar fashion. This does not appear to be the case. While males who copulated with natural estrous females initiated copulation sooner than those mated with artificial estrous females, the males mated with natural estrous females required more intromissions in the first series to achieve ejaculation. It is of interest to note that the ^irst series IF of the animals in natural estrus in this experiment was much closer to the IF reported for laboratory rats than the first series IF for animals tested in artificial estrus. Adler (1969) and Wilson, Adler and I e Boeuf (1965) reported first series IFs for laboratory rats mated with females in natural estrus to be 10.3 and 9.4, respectively. In these two experiments such average IFs provided enough stimulation to induce pregnancy in 83-901 of the females. However, in roof rats one ejaculatory series with an average of 9.5 intromissions only induces pregnancy in 56% of the females. This discrepancy is most easily explained in terms of differential sensitivity on the part of the females of the two species. While an average of 10 intromissions maximizes the probabilities of pregnancy in laboratory rat females, an average of 10 intromissions does not appear to provide enough stimulation to maximize the probabilities of pregnancy in roof rat females. This observed difference between species in female sensitivity to comparable copulatory stimulation may be accounted for by one of two

PAGE 107

ino possible reasons. First, the difference may reflect true species differences in female sensitivity. Alternatively, the difference may be due to an increased sensitivity on the part of laboratory rat females to male copulatory stimulation as a result of domestication. The data in Table 8 are also of interest with regard to species differences between roof rats and laboratory rats in the quantitative measures of copulation. Comparisons of the quantitative data of Table 8 for natural estrous tests for roof rats with the data in Table 4 for laboratory rats reinforces the conclusions drawn earlier about possible species differences. It had been noted earlier that there was overlap between species in the means for ML and Mill when comparing laboratory rats in artificial estrus with roof rats in artificial estrus. This overlap becomes even more pronounced when the data for natural estrous copulations in roof rats are compared to the data for laboratory rats. Where no overlap was noted between species with regard to the measures MF, IF and EL in Table 4, the natural estrous data for roof rats presented in Table 8 shows an overlap with the laboratory rat data for these three measures. No overlap had been found previously with regard to the measures IL, PEI and EF and it was hypothesized that perhaps there might be real species differences with regard to these measures. When the data for natural estrous roof rats are compared to those for laboratory rats in artificial estrus there appears to be no evidence for

PAGE 108

101 overlap with regard to the measures IL and PEI. The EF for natural estrous roof rats is higher than that for those tested in artificial estrus but it is still somewhat lower than that reported for sexually experienced laboratory rats. It should be obvious to the reader that the appropriate species comparisons are not of roof rats in natural estrus versus laboratory rats in artificial estrus, but rather both species in natural estrus. There are, however, no comparable data for laboratory rats in natural estrus, hence these interspecific comparisons must be interpreted with caution. It is possible that laboratory rats tested in natural estrus would show changes in their behavior that would make laboratory rats and roof rats appear less similar with regard to the quantitative measures of copulation. However, the intromission frequency data from Adler (1969) and V.'ilson et al. (1965) suggest that this may not be the case .

PAGE 109

GENERAL DISCUSSION The data presented in Experiments 1 and 2 provide a reasonably complete description of the various aspects of the copulatory behavior of wild roof rats. Before discussing the implications of these results it is perhaps wise to review briefly the findings of these two experiments. The copulatory behavior of wild roof rats could be classified as pattern #13 or pattern #15 in Dewsbury's taxonomy of copulatory behavior. Roof rats have no copulatory lock and no intravaginal thrusting and they usually, but not always, have multiple intromissions prior to ejaculation. They typically have multiple ejaculations prior to sexual satiety. This is the same pattern seen in a number of muroid rodents including Norway rats. The basic motor patterns and postures of copulation are identical to those seen in Norway rats . Quantitatively the patterns of copulation of Norway and roof rats also appear quite similar. Roof rats and Norway rats are probably not different with regard to five of the eight quantitative measures taken. There are possibly real species differences with regard to the measures IL, EF and PEI. Roof rats and laboratory rats show similar functional changes in the quantitative measures of copulation as they approach sexual satiety. This implies that the underlying 102

PAGE 110

103 mechanisms controlling copulatory behavior are the same or at least very similar in the two species. With regard to the behaviors accompanying copulation roof rats and laboratory rats again appear very similar. Roof rats engage in allogrooming and climbing behaviors not reported for Norway rats and female roof rats spend much more time running from males than do Norway rats. Female roof rats also display a pattern of darting and tail-rattling not seen in laboratory rats. The occurrence of the allogrooming and increased runring on the part of females may indicate that more mutual stimulation is required to initiate copulation in this species than in Norway rats. Wild roof rats emit ultrasonic vocalizations in sexual contexts with the majority of vocalizations occurring in males following ejaculations. This is similar to patterns of ultrasonic vocalization reported for laboratory rats. Roof rats vocalizations occur at 28 kHz, while Norway rat vocalizations tend to occur at 22 kHz. The quantitative measures of ultrasonic vocalizations are also quite similar between the two species. Like male Norway rats, male roof rats tend to be quiescent during vocalizations while female roof rats tend to be active during most male vocalizations. With regard to the stimulation necessary for the initiation of pregnancy, roof rats and Norway rats again appear similar. In domesticated Norway rats one normal ejaculatory series provides sufficient stimulation to maximize the probabilities of pregnancy (Adler, 1969).

PAGE 111

104 Results of statistical analyses indicate that in roof rats, one ejaculatory series may also maximize the probabilities of pregnancy. However, it appears that roof rat females may need more preejaculatory intromissions than Norway rat females in order to become pregnant and/or progestational. Further research is needed to clarify the exact nature of the relationship between copulatory stimulation and the initiation of pregnancy in wild roof rats. Finally, it was observed that there may be a few differences between the copulatory behavior of male roof rats mated with natural estrous females and male roof rats mated with females in artificial estrus. The implications of these results were not clear. Many of the results found in these two experiments were discussed in terms of comparisons with laboratory (Norway) rats. This was done for two reasons: first, many of the measures of copulatory behavior and most of the initial descriptions of copulatory behavior in muroid rodents were first done on Norway rats, making Norway rats the ''standard" by which most other descriptions are compared. Second, there is an intrinsic interest in how the behavior of these two sympatric, closely related species compare. However, it is also of some interest to compare the behaviors of wild roof rats to the behaviors of other muroid rodents. Such comparisons can be made along a number of parameters such as the basic pattern of copulatory behavior, quantitative aspects of the pattern, et cetera. In this

PAGE 112

105 discussion three such parameters will be considered: the basic pattern of copulation, the quantitative aspects of the copulatory pattern and the induction of pregnancy. As had been noted earlier roof rats display pattern #13 copulatory behavior according to the taxonomy of copulatory behavior of Pewsbury (1972b). The data presented in that paper show that pattern #13 copulatory behavior is quite common among many species of muroid rodents for which the copula tory pattern is known. Thus roof rats do not appear to be radically different from many muroid species with regard to the basic pattern of copulatory behavior. Comparisons of the quantitative aspects of the copulatory pattern make sense only when species are compared that have similar copulatory patterns (that is, when pattern #13 animals are compared to other pattern #13 animals). Table 11 presents quantitative data for the standard measures of copulatory behavior for nine muroid species that all display pattern #13 copulatory behavior. It should be clear from the data presented in this table that wild roof rats fall well within the range of scores of other muroid rodents on all of the standard measures of copulatory behavior. As with the qualitative aspects of copulatory behavior, roof rats do not appear to be very different from other muroid species with regard to the quantitative aspects of copulatory behavior. Finally, we can compare roof rats with other muroid rodents with regard to the stimulation necessary for the initiation of pregnancy . Such data have been reported

PAGE 113

106

PAGE 114

107 thus far for only seven muroid species other than roof rats: Norway rats (Adler, 1969), Cactus mice (Dewsbury $ Estep, 1975), Syrian golden hamsters (Lanier et al., in press), Montane voles (Davis, Gray, Zerylnick § Dewsbury, 1974), prairie voles (Gray, Zerylnick, Davis 5 Dewsbury, 1974), meadow voles (Gray, 1974) and house mice (McGill, 1970). In six of these eight species, extensive copulations occur after the first ejaculation. Only house mice and prairie voles tend to reach sexual satiety after completing just one or two ejaculatory series. In five of the six species that have extensive copulations beyond the first ejaculation, it appears that these extra copulations are essential to maximize the probabilities of pregnancy. In cactus mice, Syrian golden hamsters, montane voles and meadow voles, copulations beyond the first ejaculation appear to be necessary to maximize the probabilities of pregnancy in young virgin females (Davis et al., 1974; Dewsbury $ Estep, 1975; Gray, 1974; Lanier et al . , in press). In Norway rats, multiple ejaculations appear necessary to maximize the probabilities of pregnancy in old multiparious females but not young virgin females (Davis, 1974). It is possible that multiple ejaculations maximize the probabilities of pregnancy in roof rats as well. In prairie voles and house mice, species that do not usually have more than one or two ejaculatory series, maximal probabilities of pregnancy appear to be reached with a single ejaculatory series. This suggests that there is a fine co-adaptation within muroid species between male behavior and female reproductive physiology. In those species where stimulation beyond the first ejaculation is necessary to

PAGE 115

108 maximize the probabilities of pregnancy, males usually provide that extra stimulation. In those species that do not need the extra stimulation to maximize the probabilities of pregnancy such stimulation is usually not given. The differences in the copulatory behavior of roof rats and Norway rats alluded to throughout this paper do not appear to correspond in any meaningful way with the known differences in morphology,, life history or ecology of these two species. Perhaps the most striking thing about the copulatory behavior of roof rats is the great similarity between roof rats and Norway rats. They are after all distinct species adapted to different ways of life and slightly different adaptive zones. Other closely related species are known to show more dramatic differences in their copulatory behavior than those observed here. Cactus mice ( Peromyscus eremicus) , California mice ( Peromyscus calif ornicus) and old field mice ( Peromyscus polionotus) are all classified in the same genus, yet each shows a different pattern of copulatory behavior. Old field mice show pattern #13. characterized by no intravaginal thrusting, multiple intromissions and multiple ejaculations (Dewsbury, 1971) . California mice on the other hand, show pattern #11 characterized by intravaginal thrusting, no multiple intromissions but multiple ejaculations (Dewsbury, 1974a). Cactus mice show yet another pattern of copulation, pattern #9, characterized by intravaginal thrusting, multiple intromissions and multiple ejaculations (Dewsbury, 1974b) .

PAGE 116

109 In light of such data the similarities between roof rats and Norway rats are most striking. It should also be repeated that the differences and similarities seen between roof rats and Norway rats may not represent true species differences and similarities at all but rather reflect changes that have occurred in Norway rats as a result of domestication. There are at present no comparable quantitative data on the copulanory behavior of wild populations of Norway rats. Until these data are collected the differences and similarities between Norway rats and other species must be interpreted with caution. It was noted in the Introduction that one of the reasons for the present study of the copulatory behavior of wild roof rats was to assess the possibility that a behavioral isolating mechanism might be responsible for the lack of hybridization between sympatric populations of roof rats and Norway rats. The present data show the copulatory behavior of roof rats and Norway rats to be very similar and provide no support for the notion that differences in copulatory behavior might act as behavioral isolating mechanisms between these two species. This does not deny the possibility that other behavioral isolating mechanisms might exist that keep the species from hybridizing. One such mechanism might be pheromonal in nature such that the members of one species would not be sexually attracted to members of the opposite sex of the other species.

PAGE 117

110 It is also possible that copulation itself provides no isolating mechanism at all between the two species. Hybridization might be prevented by infertility or inviability of offspring produced by heterospecific matings. There is in fact some evidence to suggest that heterospecific matings between the two species produce no viable young. Gray (1972) cites several studies of attempted hybridization between roof rats and Norway rats. The better of these is a study by Hiraiwa and Yoshida (195 6) . Using artificial insemination techniques these authors report normal pregnancytype weight gains in 26 of 48 Rat t us rattus females inseminated with R. norvegicus sperm. However, half way through pregnancy all of these females showed bloody vaginal discharges and weight loss. No viable young were delivered of these inseminations. The eggs of six other inseminated females were examined soon after fertilization and found to be developing normally up to the second cleavage of the eggs. Hiraiwa and Yoshida interpreted these results as indicating that the females had become pregnant as a result of the heterospecific inseminations but that the embryos had died around the midpoint of gestation and had been reabsorbed. Further study of the behavior of heterospecific pairs of Norway and roof rats might provide some indication of other possible isolating mechanisms. Pairings of estrous female Norway or roof rats with heterospecific males might provide some evidence for possible pheromonal, visual or auditory precopulatory isolating mechanisms. Such studies

PAGE 118

Ill could also provide further evidence for or against postcopulatory isolating mechanisms such as hybrid inviability,

PAGE 119

REFERENCES Adler, N. T. Effects of the male's copulatory behavior in the initiation of pregnancy in the female rat. Anatomical Record , 1968, 160 , 305. Adler, N. T. Effects of the male's copulatory behavior on successful pregnancy of the female rat. Journal of Comparative and Physiological Psychology , 1969, 69, 613-622 Adler, N. T. The behavioral control of reproductive physiology. In W. Montagna § W. A. Sadler (Eds.), Reproductive Behavior . New York: Plenum Publishing Corporation, 1974. Adler, N. T., Resko, J. A., $ Goy, R. W. The effect of copulatory behavior on hormonal change in the female rat prior to implantation. Physiology and Behavior , 1970, 5_, 1003-1007. ~" Adler, N. T., § Zoloth, S. R. Copulatory behavior can inhibit pregnancy in female rats. Science , 1970, 168 , 1481-1483. : Barfield, R. J. Gonadotrophic hormone secretion in the female ring dove in response to visual and auditory stimulation by the male. Journal of Endocrinology , 1971, ; 49, 305-310. : Barfield, R. J., § Geyer, L. A. Sexual behavior: Ultrasonic postejaculatory song of the male rat. Science , 1972, 176 , 1349-1350. Barfield, R. J. § Geyer, L. A. The ultrasonic postejaculatory vocalization and the postejaculatory refractory period of the male rat. Journal of Comparative and Physiological , Psychology , 197b, «8_, /Z5-/54. : ~ : Barnett, S. A. The rat: A study in behavior . Chicago: Aldine Publishing Company, 19 b5. r ~~. Beach, F. A. Importance of progesterone to induction of sexual receptivity in spayed female rats. Proceedings of the Society of Ex perimental Biology and Medicine , 1942, bl, 369-371. ~ : 112

PAGE 120

113 Beach, F. A. Effects of injury to the cerebral cortex upon the display of masculine arid feminine mating behavior by female rats. Journal of Compara tive Psychology, 1943, 3_6, 169-198. ' : Beach, F. A. Characteristics of masculine "sex drive". In M. R. Jones (Ed.), Nebraska Symposium on Motivation : 1956. Lincoln: University of Nebraska Press, 1956. Beach, F. A. § Jordan, L. Sexual exhaustion and recovery in the male rat. Quarterly Journal of Experimental Psychology , 1956, 8_, 121-133": Beach, F. A., § Rabedeau, R. G. Sexual exhaustion and recovery in the male hamster. Journal of Comparative and Physiological Psychology , 1959, 32, 56-61. Beach, F. A., § Whalen, R. E. Effects of ejaculation on sexual behavior in the male rat. Journal of Comparative and Physiological Psychology , 1959, bZ, Z49-Z54. Blandau, R. J. On factors involved in sperm transport through the cervix uteri of the albino rat. American Journal of Anatomy , 1945, 77, 253-272. ~ Boice, R. Some behavioral tests of domestication in Norway rats. Behaviour , 1972, 42, 198-231. Boice, R. Domestication. Psychological Bulletin , 1973, 80 , 215-230. Bruell, J. H. Inheritance of behavioral and physiological characters of mice and the problem of heterosis. American Zoologist , 1964, 4_, 125-138. Bruell, J.-H. Behavioral heterosis. In J. Hirsch (Ed.)., Behavior-genetic analysis . New York: McGraw-Hill, 1967. Chester, R. V., $ Zucker, I. Influence of male copulatory behavior on sperm transport, pregnancy and pseudopregnancy in female rats. Physiology and Behavior , 1970, 5, 35-43. Conaway, C. H. Ecological adaptation and mammalian reproduction. Biology of Reproduction , 1971, 4_, 239-247. Cullen, E. Adaptations in the kittiwake to cliff -nesting. Ibis , 1957, 9_9, 275-302. Davis, H. N. The importance of age and male behavior in successful reproduction by female rats (Rattus norvegicus ) . Doctoral dissertation, University of Florida, 1974. ""****"

PAGE 121

114 Davis, H. N., Estep, D. Q. , § Dewsbury, D. A. Copulatory behavior of Mongolian gerbils ( Me ri ones un g u i c ul a t U s ) . Animal Learning and Behavior , ly/4, £, 69-/3. Davis, H. N., Gray, G. D.,, Zerylnick, M., § Dewsbury, D. A. Ovulation and implantation in montane voles ( Microtus mo nt anus) as a function of varying amounts of copulatory stimulation. Hormones and Behavior , 1974, 5, 383-388. Dewsbury, D. A. A quantitative description of the behavior of rats during copulation. Behaviour, 1967, 29, 154-178. Dewsbury, D. A. Copulatory behavior of rats: Changes as satiety is approached. Psych ological Reports , 1968, 22, 937-943. ~ Dewsbury, D. A. Copulatory behavior of rice rats ( Oryzomys palustris) . Animal Behaviour , 1970, 18, 266-2757 Dewsbury, D. A. Copulatory behavior of old-field mice (Peromyscus polionotus subgriseus) . Animal Behaviour, 1971, 19, 192-204. Dewsbury, D. A. Copulatory behavior of cotton rats (Sigmodon hispidus) . Zeitschrift fur Tierpsychologie, 1972, 3U, 477-487. Ca) " ~~ Dewsbury, D. A. Patterns of copulatory behavior in male mammals. Quarterly Review of Biology, 1972, 4 7, 1-33 Cb) : " [ ~~ Dewsbury, D. A. Comparative psychologists and their quest for uniformity. Annals of the New York Academy of Science , 19 73, 22_3, 147-167. (.aj . : Dewsbury, D. A. Copulatory behavior of montane voles ( Microtus montanus ) . Behaviour , 1973, 4_4, 186-202. (b) Dewsbury, D. A. Copulatory behavior of California mice (Peromyscus californicus). Brain, Behavior and Evolution , 19T47 b) , 9b-i06.(al Dewsbury, D. A. Copulatory behavior of wild-trapped and laboratory-reared cactus mice ( Peromyscus eremicus) from two natural populations. BehavioraT hioiogy , 1974, n, 315-326. (b) .; ~ Dewsbury, D. A. Copulatory behavior of Peromyscus crinitus and Peromyscus f lor id anus . American Midland Maturatist , 1975, 93, 468-471. (aj . Dewsbury, D. A. A diallel cross analysis of the genetic determinants of copulatory behavior in rats. Journal of. Comparative and Phy s iolo gical Psychology , 1975, 88, 713-722 Ibi : ; —

PAGE 122

115 Dewsbury, D. A. Copulatory behavior of white-footed mice ( Peromyscus leucopus ) . Journal of Mammalogy , in press. Dewsbury, D. A., § Estep, D. Q. Pregnancy in cactus mice: Effects of prolonged copulation. Sci ence , 1975, 187, 552-553. ELlerman, J. R. The families and genera of living rodents (2 vols.). London: British Museum of Natural History, 1941. Estep, D. Q., Lanier, D. L. , § Dewsbury, D. A. Copulatory behavior and nest building behavior of wild house mice ( Mus mus cuius ) . Animal Learning and Behavior , in press. Ewer, R. F. The biology and behaviour of a free-living population of black rats ( Rattus rattus ) . Animal Behaviour Monographs , 1971, 4C3J , 127-173. Ford, C. S., § Beach, F. A. Patterns of sexual behavior . New York: Harper § Row, Publishers, 1951. Gray, A. P. Mammalian hybrids . Slough, England: Commonwealth Agricultural Bureau, 1972. Gray, G. D. Ovulation and implantation in Microtus as a function of copulatory behavior . Doctoral dissertation University of Florida, 1974. Gray, G. D., Davis, H. N., Zerylnick, M., $ Dewsbury, D. A. Oestrus and induced ovulation in montane voles. Journal of Reproduction and Fertility , 1974, 38_, 193-196. Gray, G. D., § Dewsbury, D. A. A quantitative description of copulatory behavior in prairie voles ( Microtus ochrogaster ) . Brain, Behavior and Evolution , 1973 , 8 , 437-4 52. Gray, G. D., Zerylnick, M. , David, H. N. , § Dewsbury, D. A. Effects of variations in male copulatory behavior on ovulation and implantation in prairie voles, Microtus ochrogaster . Hormones and Behavior , 1974, 5_, 389-396. Hall, E. R. , 5 Kelson, K. R. The mammals of North America (2 vols.). New York: Ronald, 1959. Hardy, D. F. Sexual behavior in continuously cycling rats. Behaviour , 1972, 41_, 288-297. Hiraiwa, Y. K. , § Yoshida, H. Conception in a cross between Rattus norvegicus and Rattus rattus II. Breeding experiments using artificial insemination. Animal J Breeding Abstracts , 1956, 2_4, 285. (Abstract] —

PAGE 123

116 Karen, L. M. , § Barfield, R. J. Differential rates of exhaustion and recovery of several parameters of male rat sexual behavior. Journal of Compar a t i ve and Phy s iolo gicaT Psycho logy , 197 5,88,6957~QT. Kuehn, R. E., § Zucker, I. Reproductive behavior of the Mongolian gerbil ( Merlones UngUicuTatus) . Journal of Comparative and Physio logical Psycho logy , 1968, 6b"7~T47-752 Lanier, D. L., Estep, D. Q. , § Dewsbury, D. A. Copulatory behavior of golden hamsters: Effects on pregnancy. Physiology and Behavior , in press. Larsson, K. Conditioning and sexual behavior in the male albino ratT bteckhoim: Almqvist tj wiksell, lybb.. Lockard, R. B. The albino rat: A defensible choice or a bad habit. American Psychologist , 1968, 23_, 734-742. Marks, E. S. Sampling in revision of the Stanford-Binet scale. Psychological Bulletin , 1947, 44, 413-444. McGill, T. E. Studies of the sexual behavior of male laboratory mice: Effects of genotype, recovery of sex drive and theory. In F. A. Beach (Ed.) Sex and behavior . New York: John Wiley § Sons, inc. , T965. McGill, T. E. Induction of luteal activity in female house mice. Hormones and Behavior , 1970, 1, 211-222. McLean, J. H., Dupeire, W. A., Ill, $ Elder, S. T. Strain differences in the mating behavior of Sprague-Dawley, Long-Evans and Wistar male rats. Psychonomic Science, 1972, 29_, 175-176. " Richmond, M. , § Conaway, C. H. Induced ovulation and oestrus in Microtus ochrogaster . In Biology of reproduction in mammals . Oxford: Blackwell Press , 1969 . Robinson, R. The genetics of the Norway rat . International Series of Monographs in Pure and Applied Biology (,vol Z4J . New York: Pergamon Press, 1965 . Sewell, G. D. Ultrasound in adult rodents. Nature, 1967, 216, 512. _ Sewell, G. D. Ultrasonic communication in rodents. Nature , 1970, 227_, 410. ' ~~~ Siegel, S. Nonparametric statistics for the behavioral sciences" New York: McGraw-Hill, I9bb. Simpson, G. G. The principles of classification and a classification of mammals. Bulletin of the American Museum of Natural Hi story, ly4b, 8b, 1-55U.

PAGE 124

117 Smith, R. H. Wildness and domestication in Mus musculus : A behavioral analysis. Journal of Comparative and Phy s ToTogicaT Psycho logy , 1972, 79_, 22-29. Telle, H.-J. Contribution to the knowledge of behavioral patterns in two species of rats, Rattus norvegicus and Raft us' rattus . Ze i t schr if t 'fur' angewordt e zoo logy , 1966, 53, 129-196 (Trans. V. N. Nekrassoff-National Research" Council of Canada, 1972) . Tinbergen, N. The shell menace. 'Nafu'r'aT "History , 1963, 72, 28-35. Walker, E. P. Mammals of the world (2 vols.). Baltimore: Johns Hopkins Press, 1964. Whalen, R. E. Strain differences in sexual behavior of the male rat. Behaviour , 1961, 18^, 199-204. Wilson, J. R., Adler, N. T., § Le Boeuf, B. The effects of intromission frequency on successful pregnancy in the female rat. "Proceedings' of the National Academy of Sciences , 1965, 53, 1392-1395. Winer, B. J. Statistical princ iples in experimental design (2nd ed.). New York: McGraw-Hill, 1974'. " ~~

PAGE 125

APPENDIX A . ANALYSES OF VARIANCE FOR EXPERIMENT 1

PAGE 126

119 Table A-l Analysis of Variance for ML Source df £S MS F Total 31 68,505.87 ---Subjects 7 15,720.37 Tests 3 16,579.62 5,526.54 2.30 Error 21 ;50, 353.01 2,397.76

PAGE 127

120 Table A2 Analysis of Variance for IL Source df §S . MS/ L Total 31 868,843.5 Subjects 7 231,754.5 . "--Tests 3 114,010.75 38,003.58 1.53 Error 21 523,078.25 24,908.49

PAGE 128

121 Table A-3 Analysis of Variance for EF Source ' M ' SS_ MS F Total 31 76.72 2.47 Subjects 7 32.47 4.64 Tests 3 5.35 1.78 0.96 Error 21 38.9 1.85

PAGE 129

122 Table A-4 Analysis of Variance for MF Source df SS ' MS

PAGE 130

123 Table A-5 Analysis of Variance for IF Source df SS MS Total

PAGE 131

124 Table A6 Analysis of Variance for EL Source df SS MS Total

PAGE 132

125 Table A7 Analysis of Variance for Mill Source df SS MS

PAGE 133

126 Table A8 Analysis of Variance for PEI Source df SS MS

PAGE 134

APPENDIX B ANALYSES OF VARIANCE FOR EXPERIMENT 2

PAGE 135

128 Table B-l Analysis of Variance Comparing Estrus Condition for ML Source df SS MS F Between 1 . 32,587.74 32,587.74 2.15 Within 26 392,798.69 15,107.64

PAGE 136

129 Table B-2 Analysis o£ Variance Comparing Estrus Condition for IL Source df SS MS F Between 1 268,884.00 268,884.00 4.52 Within 26 1,546,616.85 59,485.26 * p < .05

PAGE 137

130 Table B-3 > Analysis of Variance Comparing Estrus Condition for EF Source df SS MS F Between 1 7.03 7.03 2.42 Within 26 75.60 2.91

PAGE 138

.131 Table B-4 Analysis of Variance Comparing Estrus Condition for MF Source df . SS MS Total 55 992.84 Between

PAGE 139

132 Table B5 Analysis of Variance Comparing Estrus Condition for IF Source df SS MS Total 915.55 Between

PAGE 140

133 Table B-6 Analysis of Variance Comparing Estrus Condition for EL Source df SS MS Total 55 355,263,033.1 Between

PAGE 141

134 Table B-7 Analysis of Variance Comparing Estrus Condition for Mill Source df SS MS ] Total

PAGE 142

13 5 Table B-8 Analysis of Variance Comparing Estrus Condition for PEI Source

PAGE 143

BIOGRAPHICAL SKETCH Daniel Quen Estep was born March 19, 1949 in San Antonio, Texas. In May, 1967 he graduated from Robert E. Lee High School in Baytown, Texas and in the fall of 1967 enrolled in the University of Texas at Austin. He graduated from the University of Texas at Austin in May, 1971, receiving a Bachelor of Arts degree with a major in Psychology. In September, 1971 he enrolled in the Department of Psychology of the Graduate School of the University of Florida. In December, 1973 he received a Master of Arts degree in psychology from the University of Florida. Since January, 1974 he has been working towards the Doctor of Philosophy degree under the direction of Professor Donald A. Dewsbury. Daniel Quen Estep is married to the former Barbara Melton of Dallas, Texas. 136

PAGE 144

I certify that I have read this study and that in my opinion it conforms to acceptable' standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree .ofx Doctor of Philosophy,. <&,.^~/ Donald A. Dewsbury, Cha Professor of Psychology man I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. I certify that I have read this study and. that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. TH F. Malagodi Associate Professor of Psychology

PAGE 145

I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. / r ~W. E . Meyer ~ Professor of Psychology I certify that I have read this study and that in my opinion it conforms to acceptable standards of scholarly presentation and is fully adequate, in scope and quality, as a dissertation for the degree of Doctor of Philosophy. W. B. Mtebb Professor of Psychology This dissertation was submitted to the Graduate Faculty of the Department of Psychology in the College of Arts and Sciences and to the Graduate Council, and was accepted as partial fulfillment of the requirements for the degree of Doctor of Philosophy. August, 1975. Dean, Graduate School