SOCIAL FACTORS CONTRIBUTING TO THE DEPARTURE 0F PEROMYSCUS MANICULATUS BAIRDI FROM THEIR NATAL SITE Thesis for the Degree of Ph. D. MICHIGAN STATE UNIVERSITY IRVIN RAY SAVIDGE 1970 This is to certify that the thesis entitled Social Factors Contributing To The Departure of Peromyscus Maniculatus Bairdi From Their Natal Site presented by Irvin Ray Savidge has been accepted towards fulfillment of the requirements for Ph.D. degree in ZQQng! N It We, /] Major professor I \ . ‘ Date August 10, 1970 0-169 tr unmi- a»... ~ LIB RA R Y Fidel-.1355» “tau: ’1“ Y BINDING BY IIIIAG 8 SUNS' W54 ABSTRACT SOCIAL FACTORS CONTRIBUTING TO THE DEPARTURE OF PEROMYSCUS MANICULATUS BAIRDI FROM THEIR NATAL SITE BY Irvin Ray Savidge Parental factors and individual differences contributing to the rate of natal site departure by young Peromyscus maniculatus bairdi were studied in the laboratory by using an electric shock barrier of 0.2 milli-amperes between a home cage and another cage. The number of juveniles crossing the barrier each day was recorded from 21 to 48 or 55 days of age. The mice that crossed were returned to the home cage each morning and the shock was turned off one day per week. The rate of departure increased with age. There was no significant sex difference. The rate of crossing by juveniles was correlated with the father's movements across the shock grid. When the father was restricted to the home cage or to the opposite cage, the rate of crossing was significantly higher in juveniles moving toward their father than in juve- niles moving away from him. Restraining the mother decreased the rate of crossing on non-shock days, whereas the presence of a subsequent litter J Irvin Ray Savidge increased the rate of crossing on non-shock days. The off- spring of aggressive mothers with a subsequent litter crossed at a higher rate than the offspring of non-aggressive mothers with a subsequent litter. Differences between litters were found in juveniles tested as isolates from 21 to 48 days of age. Litters tested together frequently crossed as groups rather than independent- 1y. Parental factors and individual differences contribute to the rate of natal site departure of young deermice. The presence of an aggressive mother with a subsequent litter increases the rate of departure and the presence of sibs or a non-aggressive parent decreases the rate of departure. SOCIAL FACTORS CONTRIBUTING TO THE DEPARTURE OF PEROMYSCUS MANICULATUS BAIRDI FROM THEIR NATAL SITE BY Irvin Ray Savidge A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Zoology 1970 ACKNOWLEDGMENTS I wish to thank John A. King for his guidance during the development and execution of this research and Martin Balaban, William Cooper and Dean Haynes for their contribu- tions to my doctoral program. Gary R. Connor designed and aided in the construction of the electrical components of the experimental apparatus. During the first year of my graduate studies at Michigan State University I was supported by an NDEA Title IV Fellowship and subsequently by a Traineeship on NIH Animal Behavior Training Grant GM 1751. Additional support was provided by NIH Research Grant No. 5 R01 EYOO447. ii TABLE OF CONTENTS LIST OF TABLES . . . . . . . . . . . . LIST OF FIGURES. . . . . . . . . . . . INTRODUCTION . . . . . . . . . . . . . LITERATURE REVIEW. . . . . . . . . . . METHODS AND RESULTS. . . . . . . . . . Experiment I--Group Composition . Experiment II--Adult Male . . . . Experiment III--Subsequent.Litter of Adult Female. . . . . . . Experiment IV--Aggressiveness of Adult Experiment V--Isolated Juveniles. DISCUSSION 0 O O O O O O O O O O O O 0 Adult Male. . . . . . . . . . . . Adult Female. . . . . . . . . . . Individual Differences. . . . . . Overview. . . . . . . . . . . . . SUMMRY O O O O O O O O 0 O O O O O O 0 LITERATURE CITED . . . . . . . . . . . iii and Restraint Female . Page iv vi 11 21 24 27 52 46 46 48 49 50 53 55 10. LIST OF TABLES Sex differences in rate of crossing barrier. . Distribution of crossings of juveniles within the family group compared to binomial expecta- tion (weeks 5 to 6)..-Significance.suggests juveniles may be crossing together rather than independently . . . . . . . . . . . . . . Differences between families within treatments from analysis of variance. . . . . . . . . . . Experiment I--Group Composition. Correlation of total crossings for the five week period (day 21 to day 55) . . . . . . . . . . . . . . .Experiment IV--Aggressiveness of the adult female. Test of aggressiveness of the adult females. . . . . . . . . . . . . . . . . . . . Experiment V--Isolated Juveniles. Number of crossings of the individual animals from day 21 to day 48 O O O O o O O O 0 O O O O O O 0 0 Experiment I--Group Composition. Mean number of crossings per juvenile and standard errors by week (8 replicate families per group). Shock days only. . . . . . . . . . . . . . . . Analysis of variance tables for Experiment I-- Group Composition. . . ... . . . . . . . . . . Analysis of variance tables for Experiment 11-- Adu 1t Ma 1e 0 o o o o o o o o o o o o o o o 0 0 Experiment II--Adult Male. Means and standard' errors of the numbers of crossings per juve- nile.by week (6 replicate families per group). Shock days only. . . . . . . . . . . . . . . . iv Page 17 18 19 20 29 54 57 58 59 4O LIST OF TABLES-~continued TABLE 11. 12. 13. 14. 15. Analysis of variance tables for Experiment III-—Subsequent Litter and Restriction of Adult Female. . . . . . . . . . . . . . . . . Experiment III--Subsequent Litter and Restric- tion of Adult Eemale. Mean number of cross- ings per juvenile and standard errors by week Analysis of variance table for Experiment IV-- Aggressiveness of the Adult Female. . . . . . Experiment IV--Aggressiveness of the Adult Female. Means and standard errors of the numbers of crossings per juvenile by week (6 replicate families per group). Shock days only. . . . . . . . . . . . . . . . . . . . . Analysis of variance table for Experiment V-- Isolated Juveniles. . . . . . . . . . . . . . Page 41 42 45 44 45 LIST OF FIGURES FIGURE Page 1. Experimental Apparatus. . . . . . . . . . . . . 1O 2. Experiment I--Group Composition. Crossings per Juvenile by week. Shock days only. . . . . . . 14 3. Experiment I--Group Composition. Correlation of crossings of juveniles with crossings of the adults. 0 O O O O O O O O O O O O O O 0 O O O O 16 4. Experiment II--Adult Male. Means and standard errors of crossings per juvenile by week. . . . 25 5. Experiment III-—Subsequent Litter and Restric- tion of Adult Female. Mean number of crossings per juvenile by week. Shock days only. . . . . 26 6. Aggressiveness of the Adult Female. Means and standard errors of the crossings per juvenile by week. Shock days only . . . . . . . . . . . 51 7. Comparison of juveniles tested individually with juveniles tested as litters. . . . . . . . 56 vi I NT RODUCTI ON The biology of dispersal includes generalities regard— ing gene flow and population regulation that vary little across widely divergent taxa. Simultaneously it includes details of behavior and ecology that may differ between species or subspecies. Dispersal may be defined as the movement of an animal from its natal point of origin to its permanent homesite. The dispersal movement may be either a long or a short dis- tance (such as those to an adjacent home range) (Howard, 1960). Whether the dispersal movement is long or short, it consists of three phases: 1) leaving the natal site, 2) crossing a barrier (which may be only distance or may include physical and biological obstacles), and 3) settling in a new area. Social behavior probably mediates many of the factors influencing the initiation of dispersal, specifically the leaving of the natal site. The social interactions contrib- uting to the departure of juveniles from their natal site have only been postulated. Individual differences observed among juveniles leaving their natal sites may reflect dif- ferences in social stimuli or differences in their sensitivi- ties to these stimuli. Peromyscus maniculatus bairdi was chosen as the experi- mental species because 1) it is organized as family groups, 2) it is adaptable to the laboratory, and 5) a relatively large amount of information is available on its movements in the field (Dice and Howard, 1951; Stickel, 1968) and on its population dynamics (Terman, 1968). The hypothesis tested is: Social factors and individual differences deter- mine the rate of natal site departure of young Peromyscus maniculatus bairdi. To control for environmental variables such as weather, light cycle, habitat, and physical barriers, a laboratory situation was used. A shock grid served as a barrier between two identical cages, thus maintaining con- stancy in the resistance to the juveniles' leaving their natal site. LITERATURE REVIEW Social behavior probably mediates many of the factors influencing small mammals to leave their natal site. .For example, in house mice (Mus musculus) limited food resources presumably increased dispersal via the social system before the food supply was depleted (Strecker, 1954). The dis— persal movements of deermice (g, m, bairdi) just prior to sexual maturity also suggest a social factor (Howard, 1949); as does the emigration of different age classes of muskrats (Ondatra zibethica) during drought (Errington, 1963). The stimuli from other members of the family may be either attractive or repulsive as in the occasional dispersal of littermates of E. polionotus together (Smith, 1968). The action of the social hierarchy tends to disperse flag musculus upon the attainment of sexual maturity (Brown, .1955). In an expanding colony of Rattus norvegicus, con- flicts split it into family subgroups (Barnett, 1958). Resident adults of g, m, austerus are antagonistic toward intruding juveniles in a laboratory maze and removal of adu1ts in the field improved juvenile survival (Sadlier, 1965). One of the factors involved in the settling of dis- persant rodents (Mus musculus (Delong, 1967), Peromyscus maniculatus austerus (Healey, 1967), and Ondatra zibethica (Errington, 1965)) is the presence of residents in the new area. The removal of residents, however, did not increase the rate of settling by migrant Appdemus (Andrzejewski and Wroclawek, 1962). The social behavior (antagonism toward the immigrants) and not the absolute population density is probably responsible for the failure of immigrants to estab- lish residence in favorable habitat. The aggressiveness of female Peromyscus in defense of their nest against conspecifics varies with the species (reviewed by Layne, 1968). In some species the young may continue to associate with the mother after weaning. In a few cases both litters may continue nursing for a few days. Brown (1966) views the social organization of small mammals as consisting of a dominant male who travels freely throughout a neighborhood consisting of the home ranges of the subordinate males and females. He (Brown, op. cit.) Suggests replacing the concept of home ranges of individual mice with the concept of each individual fitting into a social pattern. Information on social interactions within natural populations is essential for the understanding of their dynamics (Terman, 1968). Lidicker (1962) argues that emigrants may be those individuals most sensitive to density bUt not less poorly adapted than non-emigrants. The young animals are most affected by population pressures (Terman, 1968). In addition to the environmental factors which can initiate dispersal, Howard (1960) postulated an "innate" dispersal mechanism. "Environmental dispersal is a density dependent factor, whereas innate dispersal is independent of density, but both are presumed to be inherited traits" (Howard, 1960, p. 152). Blair (1955) also postulated "an inherent tendency to disperse, stimulated by physiological changes as the animal becomes sexually active." Howard (1949, 1960) considers the animals dispersing short distances (such as to a nearby home range) to be “environmental dis- persants" and those dispersing long distances to be "innate dispersants." This view confounds the factors determining whether or not an animal will leave its natal site with the factors determining how far it will travel before settling permanently. The spread of an introduced allele through a popula- tion of house mice has been studied by Anderson, Dunn, and Beasley (1964). They introduced a teallele onto Gull Island by releasing male mice that were heterozygous to this locus. The slow spread of the allele was attributed to the closed social system of this species. Many species probably have a social system intermediate between the closed structure examplified by Mg§_and the open system envisioned by most genetic models. Such intermediate systems will be difficult to distinguish from open systems in which the demes are isolated by distance. The discovery by Rasmussen (1964) of a shortage of heterozygotes for the blood group poly- morphisms of g. m, gracilis within a large continuous popu— lation in northern Michigan suggests such an intermediate system for Peromyscus. Although territoriality has not been demonstrated in this genus, the large volume of literature on spatial distributions in the field (Stickel, 1968) indi- cates a behavioral mechanism is preventing panmixia. METHODS AND RESULTS Several decades of field work on Peromyscus have con- tributed almost nothing to our knowledge of the behavioral interaction between family members. The utilization of a laboratory design permitted controlling environmental vari- ables such as weather, food supply, and light cycle. Preliminary studies indicated that the size of the cages was not a significant factor in determining if the juvenile leaves the home cage. Several types of barriers (water, maze, and shock) were considered. Shock permitted the best control of the intensity of the barrier and was most effec- tive. The shock level to be used was determined by placing mice on the shock grid and observing their reaction to the shock. It was also found that mice would not cross the shock barrier if no opportunity for exploration of the ap- 'paratus was provided. Subjects. The mice used in these experiments were descendents of Peromyscus maniculatus bairdi trapped in Central Michigan and had been in the colony for less than four generations. Bisexual pairs of adults were housed in 5" x 11" x 6" deep plastic cages and maintained in the laboratory colony prior to the experiment. Each cage contained wood shavings, cotton, and ad libitum food (Purina Mouse Chow), and water. The shavings and cotton were changed on alternate weeks. .Large litters were re- duced to five mice shortly after birth and litters with less than four mice were not used. Apparatus. The apparatus was designed to reduce the frequency of crossing a barrier between the home cage and another identical cage. A 5" x 2.5" high passageway with an 18" grid electrified with a 0.2 milliamperes shock con- nected two 16" x 20" x 8.5" deep plastic cages with wire mesh lids (Figure 1). The electric grid acted as a barrier to the free passage of mice from one cage to another. -Each cage contained wood shavings, cotton, and ad libitum food and water. The shavings and cotton were not changed during the experiment. In some experiments, adults were restricted to one cage by an additional barrier of 1/2" wire mesh through which juveniles could pass onto the grid and to the other cage. The adults were too large to squeeze through the 1/ " wire mesh. The light cycle was 8 hours dark and 16 hours light. General Procedure. Parents with their litter were placed in the test apparatus before the litter reached 14 days of age, and the locations of the mice were checked and recorded each morning. Those found in the opposite cage were returned to the home cage. The mean numbers of cross- ings per litter were used to test for treatment effects. .msumnmmmd Hmucmfiwummxm .a ousmflm 10 \ . LINN/u V "J, Afiw 11 Weeks are numbered by the age of the juvenile at the begin- ning of the week. One day per week the shock was discon- nected to allow the mice to explore the entire apparatus. Analysis. Analysis of variance was used to determine the effects of the treatments. Juvenile males and females within families were paired and sex differences of the juveniles were examined with paired t-tests within treatment groups. The distribution of crossings was compared to an expected calculated from the binomial distribution with a Chi square test. The expected was based on the sum across families of the binomial expansion where number of crossings number of mice x number of days Other comparisons appliCable to a specific experiment are discussed in the results of the respective experiments. The above analyses were applied to the data for the six days per week when the shock was turned on. Total crossings on non-shock days were also compared among treatments when the number of crossings justified analysis. .Experiment I—-Group Composition Methods. This experiment was designed to test the in- fluence of parents on the rate of dispersal of the juveniles. An adult pair with a litter was placed in the apparatus when the litter was less than 14 days of age. When the juveniles were 20 days old, four types of group combinations were 12 produced: 1) parents with litter (no adults removed), 2) adult male with litter (female removed), 5) adult female with litter (male removed), and 4) litter only (both adults removed). Eight replicates of each group were tested until the juveniles were 55 days old. The adults were not re- strained and could move across the grid. Results. The presence of one or both parents did not affect rate of leaving the home cage of the juveniles (Figure 2, Tables 7 and 8) but all treatments exhibited a highly significant increase in rate of leaving the home cage with age of the litter. Interaction between group composi- tion and age was not significant. No effect of the treat- ments was found on the non-shock days. Sex of the juveniles did not affect the rate of crossing (Table 1). The distribu- tion of crossings for the first four weeks was significantly different from the expected binomial distribution with too few nights having one crossing and too many having none or more than two crossings (Table 2). This unified action among litter mates indicates that individuals of a litter may not have acted independently and there may have been an at- traction between them. The number of times juveniles were found in the opposite cage was correlated with the crossings of the adult male but not with the adult female (Figure 5, Table 4). 15 .maco mmmp Moocm .xmmz an waecm>sm mom mmcemmouo .cofluflmomEoo msonwllH ucoEHHomxm .vausmflm 14 q! N onsmwm mxwmz I mm¢ “#1 OHM—A HHDU< ll mflgwh HHHuu‘ 00 o o 0 :3 flag I.I l atrueAnp/sfiurssoxo Figure 5. .15 Experiment I--Group Composition. Correlation of crossings of juveniles with crossings of the adults. A (J:A ): A (J:A ): A Pr (J:A ): A Pr (J:A ): Adult male present; crossing of juveniles correlated with cross- ings of father. Adult female present; crossings of juveniles correlated with 'crossings of mother. Adult pair present; crossings of juveniles correlated with cross- ings of father. Adult pair present; crossings of juveniles correlated with cross- ‘ings of mother. _O-5 1.0 -O.5 16 CORRELATION OF CROSS ING Ao(J:Ao) ** APr( Jon) Age of Juveniles Figure 5 _ 0.01 _ 0.05 no crossing by adult * II'U'U A9(J:A9r) A Pr(J:A 3 ) 17 Table 1. Sex differences in rate of crossing barrier. Crossings per female juvenile subtracted from crossings per male juvenile within each family. d Paired t df P Experiment I Adult Pair -0.86 1.70 7 N.S. Adult Male 0.81 1.45 7 .8. Adult Female 0.58 0.60 7 .S. Litter Only 0.55 0.64 7 N.S. Experiment II Male Across 0.24 0.55 5 N.S. Male Home 0.70 1.02 4 N.S. Experiment III Restrained female with subsequent litter —0.58 0.67 4 N.S. Non-restrained female with ‘subsequent litter —0.55 0.98 4 N.S. Restrained female without subsequent litter -0.10 0.60 4 N.S. Non-restrained female without - subsequent litter 0.07 0.16 4 N.S. Experiment IV Aggressive Female -0.01 0.02 5 N.S. Non-aggressive Female -0.65 2.57 4 N.S. 18 Table 2. Distribution of crossings of juveniles within the family group compared to binomial expectation (weeks 5 to 6). Significance suggests juveniles may be crossing together rather than independently. x2 df 1? Experiment I Adult Pair 45.48 5 0.001 Adult Male 10.85 2 0.01 Adult Female 20.54 2 0.001 Litter Only 10.12 2 0.01 Experiment II Male Across 8.97 5 0.05 Male Home 1.14 1 N.S. Experiment III Restrained female with subsequent litter 5.57 2 N.S. Non-restrained female with subsequent litter 12.25 2 0.01 Restrained female without subsequent litter 0.01 1 N.S. Non-restrained female without subsequent litter 5.66 2 N.S. Experiment IV Aggressive Female 18.45 2 0.001 Non-aggressive Female 2.54 1 N.S. 19 Table 5. Differences between families within treatments from analysis of variance. F df P Experiment I Adult Pair 2.42 7,51 0.05 Adult Male 2.20 7,51 N.S. Adult Female 7.55 7,27 0.005 Litter Only 4.69 7,51 0.005 Experiment II Male Across 4.18 5,22 0.01 Male Home 5.59 5.19 0.025 -Experiment III Restrained female with subsequent litter 5.45 5.22 0.025 Non-restrained female with subsequent litter 8.42 5.22 0.005 Restrained female without ‘ subsequent litter 7.57 5.21 0.005 Non-restrained female without subsequent litter 14.51 5.25 0.005 Experiment IV .Aggressive Female (1+x transformation) 1.51 5,20 N.S. Non-aggressive Female 1.28 5,21 N.S. 20 Table 4. Experiment I--Group Composition. Correlation of Total Crossings for the Five Week Period (day 21 to day 55). N=8 Treatment Combination Shock r P Adult Pair JzAd' on 0.45 N.S. Adult Pair JzAd' off 0.52 N.S. Adult Pair J:A9 on 0.15 N.S. Adult Pair J:A9 off 0.51 N.S. Adult Pair Ad':AQ on 0.78 0.05 Adult Pair AdgAQ off 0.52 N.S. Adult Male JzAd‘ on 0.76 0.05 Adult Male JzAd' off 0.62 N.S. Adult Female JzAQ on -0.40 N.S. Adult Female J:A9r off 0.56 N.S. 21 Experiment II--Adult Male Methods. Since the number of crossings of the juveniles in Experiment I was correlated with the number of crossings of the adult male, an experiment was designed to test whether the adult male attracted the juveniles. An adult pair with their litter were placed in the apparatus with the adults restricted to the home cage. When the juveniles were 20 days old, the adult female was removed. In group 1 the adult male was restrained in the opposite cage, while in group 2 the adult male was restrained in the home cage with the litter. Each group had six replicates. The number of crossings of the juveniles was recorded from 21 to 48 days of age. The data of weeks 5 and 4 and weeks 5 and 6 were combined to reduce the proportion of zero scores and the data were then transformed by adding 1.0 and taking the square root to attain homogeniety of variance before testing for main effects. Results. The juveniles crossed the grid at a signifi- cantly higher rate to move toward the adult male than to move away from the adult male on both shock and no shock days (Figure 4, Tables 9 and 10). No effect of age was found and no interaction between age and treatment. .No sex difference was found (Table 1). The distribution of crossings toward the father indicated the juveniles may have crossed as groups. The number of crossings in the group with the father at home was too small to test with a Chi square. .The families within 22 .xmwB >9 0HHC0>5n mom nonfimmono mo mnouum Unmpcmvm Ucm mono: .mamz pasuEIIHH ucmsfluwmxm .e musmflm 25 e ousmem mxooz I 00¢ d uunttnti mm0H0< mam: ”230$ lll meow ”Hm: “Hand I 0.0 N.0 0.0 0.0 0.0 m.fi atruaAnp/sfiutssoxo 24 each treatment group appear to be different from each other, but the individual juveniles may not have acted independent- ly which would invalidate this comparison. Experiment III-~Subsequent Litter and Restriction of Adult Female Methods. Since family differences were found in Experi- ment I when the adult female was present, the following experiment was designed to test two features of the female which may influence the rate of crossing of the juveniles. The effect of a second litter and the restraint of the mother on grid crossing by juveniles was examined in a cross- classified design with six replicates in each of the follow- ing groups: A pregnant female with a litter was placed 1) in a test apparatus with the restraining barrier present and 2) without the restraining barrier. A non-pregnant female with a litter was placed, 5) in a test apparatus with 'the restraining barrier and 4) without the restraining ibarrier. The number of crossings made by the juveniles were :recorded from 21 days of age to 48 days of age. Weeks 5 and 4 and weeks 5 and 6 were combined to reduce the proportion of zero scores before analysis. Results. Neither a subsequent litter nor restraint of 'the adult female had a significant effect on crossings during the shock days. Increased age of the juveniles increased line rate of leaving the home cage, but none of the possible interactions were significant (Figure 5, Tables 11 and 12) . 25 .maco mmmp xoonm .xmm? ma waflc0>sfl mom mmcflmmouo mo H0985: one: .mamfimm ua904 mo cOADUAHummm 0cm Houuflq ucozvmmnsmllHHH Damaguomxm .m musmflm 26 m ousmam 0x003 I 004 pocwmuummm uoz .acmcmoum pocwmuumwm uoz .ucmcmmum uoz Umcflmnummm .ucmcmmum pocwmnummm .usmcmwum uoz 33mm 0.32 I...| 32mm “:53 Ill . «é madam...— uasec l mamewm paspd ..... w.fi atruaAnp/sfiurssoxo 27 On non-shock days, however, a subsequent litter increased dispersal and restraining of the mother decreased dispersal, but there was no interaction between the subsequent litter and the restraining. No sex differences were found (Table 1). Groups 1 and 4 had a distribution of crossings not signifi- cantly different from the expected binomial. Group 2 had a distribution of crossings that was significantly different from the binomial expectation with too many days with no crossings, too many days with two or more crossings and too few with one crossing. Group 5 had too few crossings to allow comparison (Table 2). (The expected value of days with two or more crossings was too small to validly use Chi square.) Since the distribution of crossings in group 1 and 4 did not differ from random, the individual juveniles of these groups can be assumed to be acting independently. The individuals were then treated as samples to compare the families within a treatment. All four treatment groups had significant differences between families indicating the Ixopulation of adult females, from which the sample was drawn, ‘was not homogeneous with respect to an unknown trait influ- encing dispersal (Table 5). EXEriment IV--Aggressiveness of the Adult Female Methods. Incidental observations in the previous experi— ments suggested that some parent females attacked their Young when the young were returned from the opposite cage 28 whereas others did not. To test whether this difference in aggressiveness could explain the heterogenity of results obtained within previous groups, female mice with litters were divided into two groups according to whether or not they attacked a strange weanling mouse introduced into their cage. The test for aggressiveness consisted of: 1) probing the female with a forceps, 2) removing the litter for one minute and returning it to the female, and 5) introducing a strange juvenile to the female's cage for 1 minute. Since the responses of the female to the first two tests were not distinct, the behavior toward the strange juveniles was used to separate the aggressive and non-aggressive females (Table 5). .Two groups of six replicates each were established: 1) aggressive females and 2) non-aggressive females. A preg- nant female with her litter was placed in each apparatus with the restraining barrier present. The number of cross- ings of the first litter were recorded from 21 to 48 days of age. Weeks 5 and 4 and weeks 5 and 6 were combined be- fore analysis to reduce the proportion of zero scores. Results. The juveniles in the "aggressive" group crossed the grid at a significantly higher rate than those in the "non-aggressive" group. Age of the juveniles was not significant (0.10sn madco>zn uncommon Amwmvv odmfimm mmm mo xom ou ammonom Houudd madc0>sb oncommmm mo 00¢ mo mm0c0>dmmonmmw mo umoe .modmfiom udsvm 0:» .on50m wasps mo mmoco>dmmmumm¢II>H ucoeduomxm. .m OHQMB 50 .maco mwmw xooom .x003 >9 madcm>sn Mom mmcdmmouo 030 m0 muouuo pnmvcmum pom memo: .mamfiom waned may mo mmoco>dmmonmm¢ .0 whomdm 51 0 0nsmdm 0x003 I 00¢ 0HmE0m 0>dmm0ummmIcoz 0H0§0m 0>dmm0ummd II I 0.0 0.0 atrueAnp/sfiurssoxo 52 crossings in the “aggressive" treatment was significantly different from the expected binomial (Table 2). The "non- aggressive" grouping had too few crossings to allow com- parison. Although the individuals within a family may not have crossed independently of each other, the families within a treatment group were not significantly different from each other. This consistency among families within a treatment was in contrast to the families of the previous experiments, which failed to control for the aggressiveness of the females. There was only one crossing on a non-shock day in the "aggressive" treatment and none in the "non- aggressive" treatment. Experiment V--Isolated Juveniles Methods. Since Experiment I indicated significant dif- ferences between litters of juveniles with no adult present, this experiment was designed to determine if these results were real or merely an artifact of the juveniles not cross- ing the grid independently within families. An adult female with a litter consisting of 2 males and 2 females was placed in each of five apparatuses without the restraining barrier. When the juveniles were twenty days old they were placed individually in other apparatuses and their crossings re- corded until they were 48 days old. The shock was turned off on the fourth day of each week and any juveniles that crossed the barrier were returned each day to the cage to 55 which they were originally introduced. The analysis was done on the total number of crossings of a mouse over the four week test period. Results. The differences between litters and the interaction between sex and litter were significant but the sex difference was not (Tables 6 and 15). These results indicate that individuals within a litter act more alike than individuals from different litters when not given the opportunity to respond to each other. The juveniles in this experiment crossed the grid more frequently than the juveniles tested as litters in experi— ment I-4 indicating a social attractiveness of littermates (Figure 7). 54 Table 6. Experiment V--Isolated juveniles. Number of crossings of the individual animals from day 21 to day 48. Family 1 2 3 4 5 I 2 4 5 4 8 5.5 2 5 0 0 9 5 9 3 1 6 5.0 1 16 3 2 6 I 2.0 8.0 2.2 1.8 7.2 55 .mu0uuda mm U0u00u m0adc0>sn Qud3_>aamsed>dccd ©0000“ m0adc0>sfl mo condummfioo .n 0Hsmdm 56 b 0H50dm 0x003 ad 00¢ mn0uuea 00 m0ddc0>sh 0COH< m0ddc0>sb L. 0.0 N.0 ¢.0 0.0 #.d 0.d 0.d satruaAnp/sfiurssora 57 mm.odmm.N dd.0800.0 00.0H0¢.0 Nd.0dd>.0 md.odmm.0 00.0de.0 .m om.oAmm.a mm.ouom.o om.oama.o oa.oamm.o mo.oAda.o am.oAHa.o HHco umuqu 00.0H¢¢.m 0N.0Hm0.d 0d.0dmm.0 0N.0HNO.d em.0H0>.0 0N.0HON.0 0H050m Haswd mm.0H0>.N ed.odmm.0 0d.0dmm.0 0N.0d00.0 0N.odmm.0 dd.0d0m.0 0H0: uaowfi $0.0Hmm.m 0N.0Hmm.0 NN.0H0>.0 NN.0H00.0 0m.0d00.0 0d.0de.0 Hdmm pasve H0009 m e m N d uc0Eum0HB x003 .waco 0000 x0050 .AQDOH0 H00 m0HHH80w 0u00HHm0H 00 #003 0Q mHOHH0 UHmvcmum 0:0 0HHc0>9n H00 m0chmOHo mo H0985: c002 .coHuHmomEoo QDOHGIIH uc0EHH0mxm .5 0HQMB 58 Table 8. Analysis of variance for Experiment I--Group Composition. df MS F P Shock Days Source: Group Composition 5 0.922 1.40 N.S. Error between ‘28 0.659 Total Between 51 Weeks 4 1.59 5.57 0.01 Composition x Weeks 12 0.558 0.76 N.S. Error within 112_ 0.446 Total within 128 Total 159 Non-shock Days Source: Group Composition 5 0.27 0.79 N.S. Replicates 28 0.54 Total 51 59 Table 9. Analysis of variance for Experiment II--Adult Male df MS F P Shock Days (1+x transformation), by two week period Source: Location of Male 1 1.428 6.15 0.025 Error between .10 0.252 Total between .11 Total week period 1 0.077 1.0 N.S. Location of male x weeks 1 0.000 1.0 N.S. Error within 19_ 0.447 Total within .12 Total g; Non-shock Days, Source: Location of male 1 2.757 8.04 0.025 Replicates 10 0.541 Total 11 4O Nm.ddmm.m 0¢.OHmo.d 00.000d.d dd.odmm.0 mm.0H0>.0 mmOHU< 0H02 00.0dm>.0 0N.0H0¢.0 00.0d00.0 Nd.0dMN.0 00.0H00.0 0Eom 0d02 H0008 0 m w m 0005000HB x003 .maco 0000 xoonm .A050H0 H00 00HHH500 0H00HH00H 00 £003 0Q 0HHC0>sm H00 m0cH000H0 mo 0H0QEsc 00» mo 0H0HH0 0H00cmum 000 00002 .0H02 uds0m.0 mN.odm¢.0 0d.oth.0 H0HHHH 5005000Q50 H5onud3 0H0E0m 005H0HH00HI502 >N.000¢.0 0N.00¢N.0 00.00¢0.0 m0.0HNd.0 00.0000.0 H0HHHH 5505000950 u50£uH3 0H0E0w.005d0H000m m>.odm>.d 00.00m>.0 md.odm0.0 wd.odmN.0 0N.odmm.0 HOHHHH 5005000A50 SHH3 0H080m 000H0HH00HI502 mm.odmm.d 0N.odmd.0 00.00mm.0 md.odmm.0 0N.odm¢.0 H0HHHH HG05000Q50 LHH3 0H0E0m 005H0H500m H0509 , d m N d uc0Eum0HB x003 .waco 0>00 x0000 .A050H0 H00 00HHHE0M 0u0oHH00H 00 #003 mn 0H0HH0 0H00c000 050 0HHC0>5n H00 005H000H0 mo H0QE5c c002 .0H0E0m HH500 mo mcoHHUHH500H 050 H0uudd us05000£50IIHHH 550EHH00xm .Nd 0HQ0B Table 15. Analysis of variance for Experiment IV--Aggressive- Shock days only by 2 ' ness of the Adult Female. week period. Total within df MS F P Source: Aggressiveness 1 1.245 6.11 0.05 Error between 10 0.204 Total between 11_ Weeks 1 1.554 4.61 0.1x0.05 Weeks x aggressiveness 1 0.625 1.87 N.S. Error within 19_ 0.555 12 25 Total 44 0d.odm0.0 0N.0d0N.0 00.00Nd.0 00.0H00.0 dd.0dNN.0 0>H000H000Icoz mm.0dmm.d 0d.00M0.0 0N.000¢.0 00.0H0d.0 0d.0HmN.0 0>d000H00< H0008 m m w m 050E500HB x003 .0a50 0000 #0000 .A050H0 H00 00HHHE0M 0H00HH00H 00 £003 09 0HH50>50 H00 005H000Ho mo H0QE5c 005 m0 0H0HH0 0H005050 050 05002 .000800 HH500 000 no 00050>H000H00H 050EHH00xm .ed 0HQ09 45 Table 15. Analysis of variance for Experiment V--Isolated Juveniles. Shock days only. df MS F P Source: Litters 4 58.51 9.46 0.005 Sex 1 11.25 1.0 N.S. Litters x Sex 4 19.69 4.86 0.025 Error 10 4.05 Total 19 DISCUSSION The hypothesis tested in these experiments was: social factors and the individual differences determine the rate that young Peromyscus maniculatus bairdi leave their natal site. The following are considered likely factors in the dispersal of P, m, bairdi. Both parents and sibs play a role in determining the rate of dispersal of the juveniles. The father and the mother influence the juveniles differently. Individual differences occur both in the behavior of the mother toward her weaned offspring and between members of different litters tested in similar social environments. Adplp_Male In Experiments I and II the juvenile Peromyscus maniculatus bairdi were attracted to the father. Several field observations indicate that the father is also attrac- tive in field condition and may aid the juveniles in their initial explorations. A father and his four offspring (E, m, bairdi) were captured in the same trap three hundred feet from their home by Howard (1949). Rainey (1955) ob- served three g, leucopus removing chopped grain from a live trap with no indication of competition or hostility. 46 47 Adult male g, I, noveboracensis are occasionally found in the nest boxes with females and their litters when the litters were twenty-five days old or older (Nicholson, 1941). On five occasions he found single adult males living with litters after the mother left the nest box, but on nine occasions the adult male did not remain with the litter after weaning by the mother. Young 2. m, bairdi follow their parents about in the process of becoming familiar with the parental home range (Howard, 1949). Survival (disappearance in the field is considered as mortality) of juvenile g, m, austerus is negatively corre- lated to the aggressiveness of the adult males (Sadlier, 1965; Healey, 1967). In their laboratory studies they used alien juveniles introduced into their apparatus with resident adults and observed aggression. The behavior of an adult male toward strange juveniles is therefore different from his behavior toward his familiar offspring. The attractiveness of another mouse is, however, not restricted to the adult male. The distribution of crossings within families frequently was non-random (Table 2). Litters crossed the grid in groups more frequently than expected and alone less frequently than expected indicating a social attractiveness among the littermates. Singly tested individuals of a litter also crossed more frequently than littermates tested in groups (Figure 7). In the field littermates of g, polionotus occasionally disperse together 48 (Smith, 1968). Multiple captures of Peromyscus in single live traps have also been reported (Burt, 1940; Blair, 1942). The tendency of the juvenile to cross as groups was reduced when an adult was restrained to the home cage. This sug- gests that the attractiveness of the juveniles leaving may be less than the attractiveness of the adults. Adult Female The influence of the mother on the juveniles leaving the natal site varies with the circumstances. Except in the case of an aggressive female with a sub- sequent litter, the mother attracts the juveniles. In con- trast to the father, however, the attraction of the mother decreases as the juveniles become older as indicated by the increased rate of grid crossing of the juveniles with age. Although no correlation of crossings of the juveniles with the mother was found in Experiment I (Figure 5, Table 4), the rate of crossing of juveniles in Experiment III was greater on non-shock days if the mother was not restricted (Table 12). This suggests that the juveniles may have crossed the grid with their mother. A subsequent litter also increased the rate of grid crossing of the previous litter on non-shock days. The difference is not significant on shock days probably because of the heterogeniety of the females with respect to aggres- siveness. Restrained females without a subsequent litter 49 of Experiment II-5 compared with the aggressive and non- aggressive females of Experiment IV suggest that mothers without subsequent litters have the same effect on their juveniles as non-aggressive mothers with a subsequent litter. The effect of a subsequent litter is, therefore, dependent upon the aggressiveness of the female. Since the difference between aggressive and nonmaggressive females is seen only in the presence of a subsequent litter, they would all behave as non-aggressive mothers toward the last litter of the season. This, in conjunction with delayed puberty (Howard, 1949), may explain the failure of the last litter of the season to disperse until the following spring. In the field many females abandon the previous litter or force it out of the nest when the next litter is born. It is not known if there is a correlation between female aggressiveness and whether a female abandons her previous litter or evicts them from the nest in the field situation. Even if a female abandons her litter, her aggressiveness toward the juveniles in the home range may be a factor in the initiation of their dispersal. Burt (1940) reported observing an adult female g, leucopus chasing a young female. He considers old males to be more tolerant than old females toward both young and adults of the same sex. Individual Differences In addition to the differences in rate of departure resulting from the individual differences in aggressiveness 50 of the mother, differences between families were found that could not be attributed to the effect of a subsequent litter on the adult female. For example, family differences in the treatment of litters were found in I-4, male across (II-1), male home (II-2), restrained female without subse— quent litter (III-5) and non-restrained female without sub- sequent litter (III-4) (Table 5). The family differences of two of these treatments, litter only (I-4) and male across (II-1) could be explained by the tendency of the juveniles to disperse together (Table 2) as Smith (1968) observed in the field for P. polionotus. The results of Experiment V (Isolated Juveniles) suggest an inherent difference between the juveniles of the different families. Maternal influences prior to weaning have not been ruled out since no cross fostering was done. Inherent differences between individuals in the tendency to disperse is strongly championed by Howard (1960). Overview The observations of this study viewed in the context of the results of the various field studies allow us to specu- late on the dynamics of dispersal in field populations of Peromyscus maniculatus bairdi, which is probably similar to other subspecies and species of Peromyscus with minor modifi- cation. The initiation of dispersal in widely divergent genera of rodents may also be similar in some aspects. 51 For example, muskrat mothers also appear to vary in their aggressiveness toward their offspring (Errington, 1965). The behavioral mechanisms underlying the dispersal of juveniles in the breeding season are more comparable to those studied here than during the non-breeding season. Shortly before weaning, the father often joins the mother and litter (Nicholson, 1941). At that time, or slightly before, the young begin exploring the home range of their parents probably both alone and with the father. The mother may then move to another nest site in the same home range to give birth to her next litter and the juveniles extend their explorations. Some juveniles apparently explore more widely than others. During this time the mother, if she is of the aggressive type may drive the juveniles from her home range. At the onset of sexual maturity, if the young have not previously been driven from their natal home range by their mother, some will make extensive moves to suitable vacant areas perhaps discovered earlier during their explora- tions. Those driven from their home range prior to puberty probably do not settle down until the onset of sexual matur- ity and may be driven widely if the neighboring residents are aggressive. Several aspects of the influence of social behavior on dispersal of mice remain to be studied. For example, the interactions of various family members, such as, the inter- action of an aggressive mother in the presence of the father, 52 may be different from either parent alone. Behavioral modi- fications induced by environmental change may explain seasonal changes and yearly differences in dispersal. The influence of other individuals outside the family is prob— ably different in the different phases of dispersal and the elucidation of these differences will increase our under— standing of behavioral population regulatory mechanisms. For example, an adult female may behave differently toward strange juveniles than toward his offspring. Past experiences of the dispersants also undoubtedly influence the observed responses. A description of the interactions and relative influences of social stimuli, previous experience, and indi- vidual differences could provide a theoretical framework for interpreting Peromyscus population dynamics. SUMMARY Social interactions and individual differences in Peromyscus maniculatus bairdi influence the rate at which juveniles leave their natal site. In a family group the father is attractive to the juveniles and does not expell them. The social influence of the mother depends upon the presence of a subsequent litter and her aggressiveness in defending her litters. Two types of females were found with respect to aggressiveness. An aggressive mother with a subsequent litter will increase the rate of departure of her previous litter. There is a tendency for littermates to leave together and no sex difference was found. Differ- ences were found between litters when the litter members were tested separately. The social behaviors within family groups of P, m, bairdi determine the rate at which juveniles leave their natal site. 55 LI TERATURE CI TED LITERATURE CITED Anderson, P. K., L. C. Dunn, and A. J. Beasley. 1964. Introduction of a lethal allele into a feral house mouse population. Amer. Natur. 98:57-64. Andrzejewski, R., and H. Wroclawek. 1962. Settling by small rodents a terrain in which catching out had been per- formed. Acta Theriol. 62257-274. Barnett, S. A. 1958. An analysis of social behavior in wild rats. Proc. 2001. Soc. London 150:107-152. Blair, W. F. 1942. Size of home range and notes on the life history of the woodland deer-mouse and eastern chipmunk in northern Michigan. J. Mammal. 25:27-56. Blair, W. F. 1955. Population dynamics of rodents and other small mammals. Advances in Genetics 5:1-41. Brown, L. E. Home range and movement of small mammals. Symp. 2001. Soc. London 18:111-142. Brown, R. Z. 1955. Social behavior, reproduction and popu- change in the house mouse (Mus musculus L.). Ecol. Monogr. 25:217-240. Burt. W. H. 1940. Territorial behavior and populations of some small mammals in southern Michigan. Misc. Publ. Mus. Zool., Univ. Mich. 4531-58. DeLong, K. T. 1967. Population ecology of feral house mice. Ecology 48:611-654. Dice, L. R., and W. E. Howard. 1951. Distance of dispersal by prairie deermice from birthplace to breeding sites. Contrib. Lab. Vert. Biol., Univ. Mich. 50:1-15. Errington, P. L. 1965. Muskrat populations. Iowa State Univ. Press, Ames. ‘665 pp. Healey, M. C. 1967. Aggression and self-regulation of popu- lation size in deermice. Ecology 48:577-592. 55 56 Howard, W. E. 1949. Dispersal, amount of inbreeding, and longevity in a local population of prairie deermice on the George Reserve Southern Michigan. Contrib. Lab. Vert. Biol., Univ. Mich. 45:1-52. Howard, W. E. 1960. Innate and environmental dispersal of individual vertebrates. Amer. Midl. Natur. 65:257-289. Layne, J. N. 1968. Ontogeny, pp. 148-255. l2 J. A. King. Biology of Peromyscus (Rodentia). Amer. Soc. Mammal. Spec. Publ. No. 2. Lidicker, W. 2., Jr. 1962. Emigration as a possible mechanr ism permitting the regulation of population density below carrying capacity. Amer. Natur. 96:29-55. Nicholson, A. J. 1941. The homes and social habits of the wood-mouse (Peromy8cus leucopus noveborascensis) in southern Michigan. Amer. Midl. Natur. 25:196-225. Rainey, D. G. 1955. Observations on the white-footed mouse in eastern Kansas. Trans. Kansas Acad. Sci. 58:225-228. Rasmussen, D. I. 1964. Bood group polymorphism and inbreed- ing in natural populations of the deer mouse Peromyscus maniculatus. Evolution 18:219-229. Sadlier, R. M. F. S. 1965. The relationship between agonistic behaviour and population changes in the deermouse Peromyscus maniculatus (Wagner). J. Anim. Ecol. 54:551- 552. Smith, M. H. 1968. Dispersal of the old-field mouse. Peromyscus polionotus. Bull. Georgia Acad. Sci. 26:45-51. Stickel, L. F. 1968. Home range and travels, pp. 575-411. Ig_J. A. King. Biology of Peromyscus (Rodentia). Amer. Soc. Mammal. Spec. Publ. No. 2. Strecker, R. L. 1954. Regulatory mechanisms in house-mouse populations: the effect of limited food supply on an unconfined population. Ecology 55:249-255. Terman, C. R. 1968. Population dynamics, pp. 412-450. £2 J. A. King. Biology of Peromyscus (Rodentia). Amer. Soc. Mammal. Spec. Publ. No. 2.