x i [I ‘I' I II ulflr‘ - I I II? qfiESLS mmumrmn MtcmeAN STATE UN EAST LANSING, MICH. 48824 This is to certify that the dissertation entitled DISPERSAL IN SEMI-NATURAL POPULATIONS OF PEROMYSCUS MANICULATUS: SEASONAL AND HOR- MONAL RELATIONSHIPS presented by Kevin Labran Murphy has been accepted towards fulfillment ofthe requirements for Ph . D. degree in Zoology I r i k” - é/LMI 0‘ L \(‘j‘A/‘L G Major professor Date O/Wz 4, Wm MSU is an Affirmative Action/Equal Opportunity Institution 0-12771 MSU ’ LIBRARIES “ RETURNING MATERIALS: Place in book drop to remove this checkout from your record. FINES will be charged if book is returned after the date stamped below. Vi? uh 'L i DISPERSAL IN SEMI-NATURAL POPULATIONS OF PEROMYSCUS MANICULATUS: SEASONAL AND HORMONAL RELATIONSHIPS By Kevin Labran Murphy A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Zoology 1983 ABSTRACT DISPERSAL IN SEMI-NATURAL POPULATIONS OF PEROMYSCUS MANICULATUS: SEASONAL AND HORMONAL RELATIONSHIPS BY Kevin Labran Murphy In prairie deermice (PeromLscus maniculatus bairdi), juvenile dispersers are often pubertal. Furthermore, the rate of departure from the natal area by young & 1': bairdi is usually greatest during the spring and summer, a period which coincides with the season of greatest reproductive recruitment of juveniles into the population. Causes for this post-natal dispersal may vary, and several hypotheses that address these in small mammals were reviewed. One hypothesis, the sexual search hypothesis, predicts that the: 1) departure rate of juveniles from their natal area will be positively associated with levels of sexually attractive stimuli present in the environment, 2) pubertal mice will be more likely to disperse than non-pubertal individuals, and 3) age at dispersal by juveniles can be experimentally delayed by suppressing their sexual maturation, or advanced by sexual stimulation. These predictions were tested in a series of controlled field experiments with E; & bairdi in south-central Michigan. Populations of known familial composition were established by introducing mated pairs of deermice into an old-field artificially provided with favorable microhabitats. These microhabitats consisted of mouse-accessible enclosures with. nest-boxes, surplus food, and water. This semi-natural arrangement provided at least partial control of variables (e.g. population density, predation pressure, nutrition, and intraspecific competition) which are suspected to be causally related to dispersal. Juveniles were experimentally treated with melatonin implants in one experiment to delay gonadal maturation; while, in another experiment, they were treated with gonadal steroids in an attempt to stimulate sexual activity. Positive seasonal and hormonal relationships between age at puberty and age at dispersal were found. The results indicated that: 1) pubertal mice were more likely to disperse than non-pubertal individuals, 2) age at puberty was a good predictor of age at disappearance for all mice from {all seasons, 3) summer-born, and melatonin-treated mice attained puberty and dispersed at an older age than their spring and control counterparts, 4) in the fall, dispersal rates of steroid-treated mice did not differ from controls. The close association between age at post-natal dispersal and ontogenetic maturation of the gonads lends credence to the argument that gonadal hormones may be important in activating mate-seeking behaviors in mice. To my parents, Bernard and Madelene Hulmes Murphy who made it all possible ii ACKNOWLEDGEMENTS I extend my sincerest thanks to John A. King, my advisor, for his guidance and enthusiastic support throughout this study» 11 am also grateful to Richard W. Hill, Lawrence 1. O'Kelly, and Donald 0. Straney for their constructive comments and encouragement. Dr. Straney generously provided computing facilities and programming assistance in the preliminary stages of data analysis. I also thank John L. Gill and Clyde R. Anderson for their assistance with statistical analyses. I thank Susan Hill for providing facilities and materials for the histological analyses. My special thanks go to the following students who provided dedicated and unselfish assistance in collecting data during the field work: Lori Aldrich, Susan Comeau, Jerry Gidner, Karen Kolbasa, Jay Kralik, Dana Lindemann, and Peggy Vogt. To my fellow graduate students and friends who provided encouragement and support, I acknowledge my sincerest appreciation. This research was supported in part by a U.S. Public Health Service Research Grant, 5 R01 MB 29414, and by a teaching assistantship provided by the Department of Zoology. iii TABLE OF CONTENTS Page LIST OF TABLES O O O O O O O O O O O O O O O O I O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O I O Vii LIST OF FIGURES O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O Xi INTRODUCTION. 0 C O O O O O I O O O O C C O O O O O O O I O O O O O O I O O O O O O O O O O O O O O I O O O O O O O O O O 1 CHAPTER ONE - LITERATURE REVIEW: DISPERSAL IN SMALL MAMMALS........ 6 DefinitionSOOOOOOOO00.0.00...OOOOOOOOOOOOOOOOOOOOOOOOOIOOOOOOO MeasurementSOOCOCOOOOOCOO000......00......OOOOOOOOOOOOIOOOOOOC Attributes Of Dispersers O O C O O O O O O O O O O O O O O O O O O O O O I O O O O O O O O O O O O O 1 O‘CDO‘ SEXOOOCCCOOOOOOOOOOOOOOO00....0.0.0.0000...OOOOOOOOOOOOOOOO 17 “a and weight 0 O O O O O O O O O I O O O I O O O O O O O O O O O O O O O O O O O O I O O O O O O O O O 17 ReprOductive Status 0 O O O O O O O O O O O O O O O O O O O O O O O O O O I I O O O I O O O O O O O 19 GenotYPeOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO 21 BehaViorOOOOOOOOOOOOOOO00.0.0.0...0......OOOOOOOOOOOOOOOOOO 22 Field StudieSOOOOOOOOOOOOOOOOOOOOOCOOOOOOOOOOOOOOOOOIOOO 22 laboratory StUdieSOOOOOOOOOOOOO00.0.00...OOOOOOOOOOOOOI. 24 Behavioral Hypotheses......................................... 28 Social Subordination Hypothesis............................ 29 Genetic-Behavioral Polymorphism Hypothesis................. 31 Presaturation-Saturation Dispersal Hypothesis.............. 32 Social Cohesion Hypothesis................................. 32 Sexual Search Hypothesis................................... 33 Behavioral Hypotheses: Do They Explain General Attributes of Dispersers?................................................ 34 Demographic Generalizations................................ 35 Population Density...................................... 35 Sex..................................................... 36 Age/Height.............................................. 37 Reproductive Condition.................................. 37 Behavioral Generalizations................................. 38 Aggression.............................................. 38 Other Behaviors......................................... 39 Hypothesis Testing............................................ 42 iv CHAPTER TWO - EFFECTS OF MELATONIN ON THE REPRODUCTIVE SYSTEM OF P. M. BAIRDI MAINTAINED UNDER DIFFERENT PHOTOPERIODSOOOOOOO0.00.00.00.00000000000000000000... 45 Introduction.................................................. 45 Background.................................................... 46 Induction of Sexual Behavior............................... 47 Delay of Sexual Maturation................................. 47 Materials and Methods......................................... 50 Melatonin Implants......................................... 51 Control Implants........................................... 53 Implantation Technique..................................... 53 Analysis................................................... 56 Results....................................................... 57 MaleSOO00....OOOOOOOOCOOOOOOOOOOO...OOOOOOOCOCOOOOOCOOOOOOO 57 Testes Weight........................................... 57 Testicular Index........................................ 58 Histological Assessment of Gonadal Condition............ 61 Body Weight............................................. 64 Females.................................................... 64 Vaginal Patency......................................... 64 Reproductive Tract Weight............................... 66 Histological Assessment of Follicular DevelOpment....... 68 Body Weight............................................. 70 Discussion.................................................... 72 CHAPTER THREE - EXPERIMENTAL TEST OF THE SEXUAL SEARCH HYPOTHESIS OF DISPERSAL IN A SEMI-NATURAL POPUI-‘ATION 0FikBAIRDICOOCCCOCOOOOOOOOOOOOOOCOO 77 Introduction.................................................. 77 General Materials and Methods................................. 77 Experiment 3.1 - Spring and Summer Dispersal in Juveniles: Effect of Delaying Gonadal Maturation......................... 87 Materials and Methods...................................... 87 Analysis................................................ 88 Mmmusn.n.u.u.u.u.u.n.u.n.u.u.u.u.u.u.u..91 Experiment 3.2 - Fall Dispersal in Juveniles: Effect of Stimulating Sexual Behavior.........................123 Materials and Methods......................................123 Analysis................................................126 ResultBOOOOOOOI...0.0.000IOOOOOOOOOOOI.0.0.0.00000000000000128 D18¢u881°n0000000000000OOOOOOOOOOOOOOOOOOO0.00.00.00.000000000141 CHAPTER FOUR - PATTERNS OF NOCTURNAL ACTIVITY, SPACE-USE, AND NEST COHABITATION IN FREE-RANGING MALE g; & BAIRDI AS REVEALED BY RADIOTELEMETRY..........163 IntIOdUCtionoooooso.ooooooooooooooooooooooo00000000000000.0000163 Materials and Methods.........................................l64 Analysis...................................................l68 Results.......................................................169 Social Interactions........................................l70 Activity Patterns..........................................172 GeneraIOOOOOOOOOOOOIO...0.00.0000...00.0.000000000000000172 During Parturition Periods O O O O O O I O O O O O O O O O O O O O O O O O O O O O O .174 D18CU881ODOOOOOOOOO0......OOOOOOOOOOOOO0.0.0.00.00000000000000178 SOCial Interactions O I O I O I O O O O O O O O O O O O O O O O O O O O O O O O O I O C I O O O I .178 ACtiVity Patterns. 0 O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O O 186 GENERAL DISCUSSION...0.0.0.000... 0... 0.0.0.0.... O. 0...... .00... 00.0191 BIBLIOGRAPHYOOOOOOOOOOOO...O....0.0.0.0.0...O...00.000.00.000000000207 vi TITLE 1.1 2.1 2.2 2.3 2.4 2.5 2.6 2.7 3.1 3.2a 3.2b 3.2c 3.2d 3.2e 3.3 LIST OF TABLES Predictions of five behavioral hypotheses for the cause of dispersal.................................. Analysis of variance of testes weights...................... Analysis of variance of testes index scores................. Analysis of variance of male body weights................... Frequencies with which females were perforate or imperforate at the end of the 2 week treatment periodOCOOOCOOOOC.0.0.0....OOIOOOOOOOOOOOOOOO0.0000000000... Analysis of variance of female reproductive traCt weightSOOOOOOOOOOOOOOOOOOOOOOOOOOOOO...IOOOOOOOOOOO... Frequencies with which females possessed corpora lutea at the end of the 2 week treatment period............. Analysis of variance of female body weights................. Introduction dates of the original bisexual pairs........... Frequencies with which melatonin-and control- treated juveniles (at least 30 days old) became pubertal or remained non-pubertal.)......................... Frequencies with which males and females became pubertal or remained non-DUbertalO0.0.0.0.0...OOOOOOOOOOOOOO Frequencies with which individuals in each season became pubertal or remained non-pubertal.................... Frequencies with which males and females (in each season) became pubertal or remained non-pubertal............ Frequencies with which melatonin-and control- treated juveniles (in each season) became pubertal or remained non-pubertal........................... Analysis of variance of age at puberty...................... vii PAGE 30 58 61 64 66 7O 7O 85 92 93 93 93 95 96 94 3 '»l.‘ .V‘J ‘AJ LA) 3.4 Preliminary mixed-effect analysis of variance Of age at pubertYOOOOOOOOOOOIOOOOOOOOOOOOOOOOOOOOOOOOO0.0... 98 3.5 Analysis of variance of weight at puberty................... 99 3.6 Preliminary mixed-effect analysis of variance Of weight at DUbertYOOOOOOOOOOOOOOOOOOIOOOOIOOOOOOOOOIOOOOOI 99 3.7 Frequencies with which melatonin-and control- treated females (at least 50 days old) became pregnant...O...00......OOOOOOOOOOOOOOOOOOO0.00...0.0.0.0....101 3.8a Frequencies with which melatonin- and control- treated juveniles disappeared from the population orstayedOOOOOOOOOOOOOOOOOOOOOOOOCOOOOOOOOOOOOOOOOOOOOOOO...102 3.8b Frequencies with which individuals in each season disappeared from the population or stayed................... 102 3.8c Frequencies with which males and females disappeared from the population or stayed................... 103 3.8d Frequencies with which males and females (for each treatment) disappeared from the population orstayedOOOOOOOOO0.0000000000000000000000000000000000000000103 3.8e Frequencies with which individuals in each season (for each treatment) disappeared from the pOpulation orstayedCOOOOOOCOOOOOO0.00.00.00.00...OOOOOOCOOOOOOOOOOO...104 3.8f Frequencies with which males and females (in each season) disappeared from the population or stayed........... 105 3.8g Frequencies with which melatonin- and control- treated juveniles (in each season) disappeared from the population or stayed............................... 106 3.9 Analysis of variance of age at disappearance................ 104 3.10 Preliminary mixed-effect analysis of variance Of age at disappearance...OOOOIOOOOOOOOOOOOOOOOOOOOOOOOOOOOO 108 3.11 Analysis of variance of weight at disappearance............. 109 3.12 Preliminary mixed-effect analysis of variance Of waight at disappearance.................................. 109 3.13s Numbers of melatonin- and control-treated stayers that made at least one inter-couplet movement. Stayers are those individuals which remained in the population until at least the median duration of occupancy for all treated mice (42 days).............................................. 111 viii 3.13b 3.13c 3.13d 3.13e 3.14e 3.14b 3.14c 3.14d 3.17 3.18 3.19 3.20 Frequencies with which stayers in each season made at least one inter-couplet movement.................... Frequencies with which male and female stayers made at least one inter-couplet movement.................... Frequencies with which male and female stayers (for each treatment) made at least one inter- couplet movementOOOOOOOOOOOOOOOO0.00......00.000000000000000 Frequencies with which stayers in each season (for each treatment) made at least one inter- couplet movementCOOOOOOOOOOOOOOOOOOOOQOOOOOO00.0.0.0....0... Frequency with which individuals were pubertal at disappearanceOO0.00000000000000.00....OOOOOOOOOOOOOOOO... Frequencies with which melatonin - and control- treated juveniles were pubertal at disappearance (regardless Of age)OOIOOOOOOOOOOOOOOO...OOOOOOOOOOOOOOOOOOOO Frequencies with which individuals in each season were pubertal at disappearance (regardless of age).......... Frequencies with which males and females were pubertal at disappearance (regardless of age)............... Latency to disappear following attainment of pubertYOOOOOIOOOOOOOOOOOOOOOOOOIOOOOOIOOOOOOOO0.00.00.00.00. Frequencies with which males and females (for each treatment) became pubertal or remained non-pUbertaIOOOOOOOOOOOOOOO...OCOCOOOOOCOOOOOOOOOO0.00...... Frequencies with which males and females (for each treatment) disappeared from the population or stayed. Stayers are individuals which remained in the population until at least the median age of occupancy for all treated mice (33 days)................................................... Frequencies with which steroid - (EB and TP pooled for males and females) and control-treated stayers made at least one inter-couplet movementOOOOOOOOOOOOOOOOOOOOIOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO Frequence with which individuals were pubertal at disappearanceOOOOOOOOOC.0.00.00.00.00...OOOOOIOOOOOOOOOOO... Frequencies with which males and females (for each treatment) were pubertal at disappearance................... ix 112 112 113 113 114 115 115 115 116 117 129 132 136 137 3.21 Latency to disappear following attainment of puberty among mice that became pubertal before disappearing...0.0.0.0....OOOOOOOOOOOOOO0......0.00.00.00.00 138 3.22 Age at disappearance for melatonin-treated mice and untreated mice from six previous seasons........... 147 3.23 Age at disappearance for control mice and untreated mice from six previous seasons.............................. 148 3.24 Age at disappearance for testosterone-treated mice and untreated mice from 6 previous seasons............. 158 3.25 Age at disappearance for estradiol-treated mice and untreated mice from six previous seasons................ 158 3.26 Age at disappearance for control mice and untreated mice from six previous seasons.................... 159 4.1 Percent of total observations in which collared males occupied a couplet or a nest (if in a couplet)OOOOCOOCOOOOOOOOOOOOOOOOOOOOOOOOOOO0.000.000.0000... 170 4.2 Percent of total observations in which collared males were alone or cohabiting with other mice in a couplet or in a nest (if in a couplet)................. 171 4.3 Mean activity measures of radiotracked males................ 173 4.4 Relationship between time of parturition by closely neighboring females and activity of COIIared “1318800000000000.000000000000009.coco-0000000000000 181 TITLE 2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8 2.9 3.1 3.2 3.3 3.4 3.5 3.6 LIST OF FIGURES Beeswax pellet implant tOOIOOOOOOOOOOOOO0.0.0.0....0.0.0.... Mean paired testes weights. Vertical lines indicate :1 SOEOOOOCOOOOOOOOOOOC0.0.0.0000...OOOOOOOOOOOOOOOOO0.0... Mean testis index scores. Vertical lines indicate : 1 SOEOOIOO00....OOOOOOOOOOOOOOOOOOOOOOOOOOOOOIOOOOOOOOOOOO Mean spermatogenic index scores. Vertical lines indicate : SOEOOOOOOOOOOOOO0.00.00.00.00.0.00.0000...0...... Mean interstitial cell index scores for males. Vertical lines indicate 1:1 S.E............................. Mean body weights of males. Vertical lines indicate :1 SOEOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOIOOOOOOOOOOOOOCOOOOO Mean reproductive tract weights of females. Vertical lines indicateil SOEOOOOOOOOOOOOOOOOOOOOOIOOOOOOOOOOOOOOOO Mean follicular index scores. Vertical lines indicate 1 l. SOEOOOOOOOOOOOOOOOOOOOOOO.IOOOIOOOOOOOOOOOOOO. Mean body weights of females. Vertical lines indicate :- l SOEOOOOOOOOOOOOOOOOOOIOOOOOOOOOOOOOOOOOOO00.... MapOf themuseC1tie8 StUdyareaoooooooooo000.000.000.000. Schematic diagrams of an arena and a nest-box............... Pattern of introduction of breeding pairs to couplets at the mouse cities. Darkened circles represent introductory couplets............................. Mean age at puberty. Vertical lines indicate : 1 SOEO0.0...OOOOOOCOOOOOOOOOOOOOOO00......OOOOOOOOOOOOOOOOOO Mean weight at puberty. Vertical lines indicate : l 80E...00......OOOOOCOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO Mean age at disappearance. Vertical lines indicate : l 80E...I.0..OOOOOOOOOOOOOOOO0.00......0000...... xi PAGE 52 59 60 62 63 65 67 69 71 78 82 84 97 100 107 3.7 3.8 3.9 3.10 3.11 3.12 3.14 3.15 Mean weight at disappearance. Vertical lines indicateil SOEOOOOOOOOOOOOOOOOCOOOOOOOIOOOOOOOOOOOOOOIOOO. Relationship between age at disappearance and age at puberty for mice in the Summer 1981. The following regression equations are for melatonin-treated and control-treated mic? respectively: Y - 0.99X + 1.94 (n 8211, r = 0.81); Y = 0.70X.+ 13.81 (n = 13, r - 0.45) ............... Relationship between age at disappearance and age at puberty for mice in the Spring 1982. The following regression equations are for melatonin-treated and control-treated mice, rEspectively: Y = 0.80X + 10.16 (n = 32, r2 = 0.69); Y - 0.94X + 5.72 (n = 37, r .0.13)....COOO0.00...OOOOOOOOOOOOOOCOOOOOOOOOOIOOOOOOOOO Relationship between age at disappearance and age at puberty for mice in the Summer 1982. The following regression equations are for melatonin-treated and control-treated mice, rfispectively: Y - 0.97X + 7.65 (n = 14, r2 - 0.74); Y - 1.12X + 3.96 (n 8 14, r .0.30)....OOOOOOOOOCOO...OOOOOOOOOOOOOOOOOO...0.0.0.0... Relationship between weight at disappearance and weight at puberty for mice in the Summer 1981. The following regression equations are for melatonin-treated and control-treated mice, rgspectively: Y a 0.84X + 2.36 (n 8 112 r - 0.70; Y - 0.39X + 7.47 (n = 13, r = 0.21)............. Relationship between weight at disappearance and weight at puberty for mice in the Spring 1982. The following regression equations are for melatonin-treated and control-treated micE, respectively: Y = 0.90X + 1277 (n = 31, r 8 0.55); Y - 1.37X - 3.11 (n = 36, r - 0.35)....................... Relationship between weight at disappearance and weight at puberty for mice in the Summer 1982. The following regression equations are for melatonin-treated and control-treated mice, rsspectively: Y - 0.71X + 5.21 (n - 14,2 r - 0.23); Y - 0.43X + 7.78 (n - 14; r - 0.42)........... Mean age at puberty. Vertical lines indicate 1 1 SOEOOOOOOOOOOOOOOOOO0.0.0.0....OOOOOOOOOCOCOOOOOOOOOOOO Mean weight at puberty. Vertical lines indicate :1 SOECOOOOOOOOOOOOOOOOOOIOOOOOOOOOOOOOOOOCOOOOO0.0.0.0... xii 110 118 119 120 122 124 125 130 131 3.17 3.18 4.1 4.2a 4.2b 4.2c Mean age at disappearance. Vertical lines indicateil SOEOOIOOOOOOOOOOOOOOOOOOI.OOOOOOOOOOOOOOOOOOOO 134 Mean weight at disappearance. Vertical lines indicateil SOEOOOOOOOOOOOOOOOIOOOOOOOOOOOOOOOOOOOOOOOOOO. 135 Relationship between age at disappearance and age at puberty for mice in the Fall 1981. The following regression equations are for control-treated, TP-treated, and EB-treated mice, respectively. Y = 4.00X - 64.00 (2 - 3, r = 0.84); Y = 1.30X - 6.60 (n r - 0.56); Y = 1.81X - 15.11 (n a 3, r 2" 5’ 8 0.98)............ 139 Relationship between weight at disappearance and weight at puberty for micce in the Fall 1981. The following regression equations are for TP- treated and EB-treated mice, respectively. A regression analysis was not possible for control- treaEed mice (see text): Y = 1.32X - 2.522(n = 5, r - 0.82); Y = 1.07X + 1.11 (n = 3, r = 0.89).......... 140 Schematic diagram of a radiocollar. The diagram on the right shows an unpotted radiocollar with underlying electronic components. The one on the left shows a radiocollar that is potted with dental acrylic and fitted with a protective sleeve................ 165 Positions of 3 collars males obtained during 2 weeks of radiotracking in the Fall 1981. The positions are superimposed on a map of the mouse cities and indicate the relative frequency of observation at any given location as well as the position of males relative to the position of neighboring parturient females (see text for explanation)............................................... 175 Positions of 5 collared males obtained during 3 weeks of radiotracking in the Fall 1981. The positions are superimposed on a map of the mouse cities and indicate the relative frequency of observation at any given location as well as the position of males relative to the position of neighboring parturient female (see text for explanation)...................................... 176 Positions of 7 collared males obtained during 2 weeks of radiotracking in the Spring 1982. The positions are superimposed on a map of the mouse cities and indicate the relative frequency of observation at any given location as well as the position of males relative to parturient females (see text for explanation.)................................ 177 xiii 4.3a 4.3b Total nightly distances traveled by radiocollared males in the Fall 1981. The graphs illustrate the distances traveled by males during the parturition and non-parturition periods of neighboring females......... 179 Total nightly distances traveled by radiocollared males in the Spring 1982. The graphs illustrate the distances traveled by males during the parturition and non-parturition periods of neighboring females........ 180 xiv INTRODUCTION Dispersal can be viewed as the movements made by an animal as it departs its natal area (Gaines and McClenaghan, 1980, after Endler, 1977). These movements are "continual rather than periodic" and may occur "within or between generations.” This rather simple and general definition belies the difficulty of studying dispersal, a life-history characteristic that is, perhaps, fundamental to all animals (Howard, 1960). The study of these movements in natural populations is particularly difficult for secretive, nocturnal small mammals because dispersal cannot be observed directly. lhstead, dispersal movements are monitored indirectly with the aid of live-trapping, nest-boxes, radiotelemetry, and other tracking techniques. Some general demographic attributes of rodent dispersers are gradually emerging from field studies. For ‘most species studied, dispersal is a density-independent phenomenon. Dispersers are commonly young, lightweight individuals of either sex who move or disappear from their natal area at, or near, the age of puberty (e.g. Bekoff, 1977) though older individuals of certain species also make dispersal movements (see Tardif, 1979). For a more thorough review of the attributes of dispersers, including behavioral and genetic characteris- tics, see the Literature Review. Despite knowledge of certain demographic, behavioral, and genetic attributes of dispersers, little is known about the underlying causes or mechanisms that prompt dispersal in small mammals. In population studies of the cricetine rodent Peromyscus, the season of active breeding is often characterized by an increase in the number of dispersing juveniles which are often pubertal (e.g. Healey, 1967; King, 1983; Saldier, 1965). The coincidence of juvenile dispersal, juvenile sexual maturation, and breeding by resident adults has led to the hypothesis that juveniles move in response to social pressure or aggression from resident males (Gaines and McClenaghan, 1980; Krebs, 1978). Nevertheless, the evidence for the occurrence of social pressure or aggression in natural populations is weak and is derived mainly from questionable laboratory tests or from the assessment of skin wounds. Despite the inferences of aggression in populations, aggressive behaviors of small, nocturnal mammals have rarely 'been observed in the field (e.g. Anderson, 1980). Since dispersal occurs predominantly during the breeding season, an alternative explanation for dispersal is that it results from the search for mates by the sexually active individuals, which most often are the pubertal recruits (King, 1983). One prediction of the sexual search hypothesis is that the dispersal rate of all individuals in the population will be positively associated with the level of sexually attractive stimuli present in the environment. These stimuli may be those provided by the sexually active individuals present in the population, but they also include the sexually' attractive signals produced by these individuals. Although there is considerable anecdotal reference to the existence of these signals in the environ- ment (e.g. Moore, 1965), little is known about the nature of these stimuli or the modality of their perception. In many rodents, however, sexual attraction or sexual recognition stimuli are commonly presumed to be odorous substances produced in the preputial gland in both sexes (e.g. Vandenbergh, 1975). If sexually attractive stimuli promote dispersal in sexually active individuals, I predicted that the initia- tion and rate of juvenile dispersal could be affected by 1) altering the levels of sexually attractive stimuli in the environment and/or 2) manipulating the sexual maturation or sexual behavior of juveniles. Most P; m;_bairdi disperse in the spring and summer when juveniles mature sexually. Dispersal rates in the fall are low, and only adults continue to ‘breed (Howard, 1949; King, 1983). These. two seasonal patterns permit an experimental test of the two complementary predictions of the sexual search hypothesis. For spring and summer seasons, I predicted that a reduction in the level of sexually attractive stimuli and/or the prevention of sexual maturation in juveniles would result in lower than normal dispersal rates. Conversely, in the fall, an increase in the level of sexually attractive stimuli and/or number of sexually active juveniles would result in greater than normal dispersal rates. A field experiment test of the sexual search hypothesis poses several methodological problems. The first concerns the control of other independent variables (e.g. population density, predation pressure, nutrition, and intraspecific competition) which are suspected to be causally related to dispersal. To provide some control of these variables, I introduced structured populations of §;_m; bairdi into a study site that was provided with favorable microhabitats consisting of mouse-accessible enclosures with nest-boxes and surplus food and water. A second technical problem involves measurement of dispersal in the c1 it CK EX as ti field. I used inter-enclosure movement data and the disappearance of marked individuals from the study area to measure dispersal. Dispersal measured as disappearance, however, was irreconcilably confounded with mortality. Finally, a third methodological problem concerns the selection of techniques for manipulating either the sexually attractive stimuli or the sexual maturation of juveniles. Given that little is known about the nature of sexually attractive stimuli, I chose to manipulate the sexual maturation of juveniles by treatment with exogenous hormones. Chapter 2 deals with the selection and testing of a hormonal treatment for delaying sexual maturation of juvenile 2;.2; bairdi. I chose a technique which involved treating mice with melatonin, a hormone with antigonadal properties. It was necessary to learn whether melatonin would be capable of suppressing gonadal development of mice reared under the stimulating influence of a long-day photoperiod. In the laboratory, I performed standard histological examinations of testes and ovaries to assess the antigonadal effects of melatonin treatment on mice reared under the influence of different photoperiods. Specific hypotheses for these treatments are described in this chapter. Melatonin treatments were used to delay sexual maturation of free-ranging juveniles during the spring and summer. In this experiment (Chapter 3; Experiment 3.1), field-born mice were randomly assigned to either a melatonin or a control treatment group. Stimula- tion of sexual behavior in free-ranging juveniles was attempted in the fall experiment (Experiment 3.2) by treating mice with exogenous implants of gonadal steriods. The specific hypotheses tested in these experiments are described in Chapter 3. In the proposed test of the sexual search hypothesis (Chapter 3), I attempted to control intraspecific competition in the pOpulation and thus eliminate aggression as a probable cause for juvenile dispersal. Nonetheless, the possibility remained that adult-juvenile aggression (e.g. Fairbairn, 1977a, 1978a) might explain the results predicted for Experiments 3.1 and 3.2. I therefore: made direct radiotelemetric observations of the activity patterns of free-ranging adult P_._ m: bairdi in an attempt to estimate the likelihood of adult-juvenile aggression in the field. The Specific hypotheses tested are described in the final chapter (Chapter 4). CHAPTER ONE LITERATURE REVIEW DISPERSAL IN SMALL MAMMALS Definitions Any review of small mammal dispersal is beset with problems of terminology. It is difficult to reach a consensus for a definition of dispersal because most interpretations are specifically related to a particular organism or method of study. A commonly used definition is that of Howard (1960, p. 159): Dispersal of an individual vertebrate is the movement the animal makes from its point of origin to the place where it reproduces or would have reproduced if it had survived and found a mate. Howard's definition is certainly consistent with methods of study and results from earlier research on the population biology and social behavior of small rodents, particularly Peromyscus spp. (Burt, 1940; Howard, 1949; Dice and Howard, 1951; Nicholson, 1941). Using nest-boxes in their studies of P; maniculatus bairdi and P; leucoEus, these authors found that young mice typically leave the natal nest when they reach sexual maturity, establish a home range of their own away from the place of birth, and remain somewhat sedentary thereafter. Dispersal has characteristically been reserved to describe the first movements that a young animal makes as it permanently leaves its natal area. Recently, authors have espoused a more dynamic interpretation of dispersal because animals move or change their home range at other times for a variety of reasons (Bekoff, 1977; Gaines and McClenaghan, 1980; Smith, 1978; Tardif, 1979). Brown (1975:49) viewed dispersal simply as the ”movements of animals from a source...” Another widely used and general definition is that of Lidicker (1975:104) who described dispersal as ...any movements of individual organisms or other propagules in which they leave their home area, sometimes establishing a new home area...[excludingl short-term exploratory movements. Even Lidicker's definition may be limiting because it attempts to distinguish between ”short-term exploratory movements" from others that are, presumably, dispersal movements. Is it valid to make a distinction between short-term exploratory movements and other, longer movements? I believe not. In some situations, this distinction may have more to do with the methods used to record animal movements and less with the actual behavior of the animals being measured. For example, in live-trap studies, movement distances made by animals can be inferred from straight line measurements made between two trap locations (e.g. Dice and Howard, 1951). These measurements apply to the distance between the two points concerned but not necessarily to the actual distances traveled. Thus, some animals that appear to be making short-term exploratory movements may, in fact, be traveling as far as animals that are classified as dispersers (see Gottfried, 1982). To define dispersal according to the length and direction of movements is not recommended because these characteristics may also vary from species to species and from individual to individual and, in themselves, do not serve to distinguish dispersal from other movement types (Bekoff, 1977; see also Madison, 1980c). In this review I favor a general definition of dispersal adopted by Gaines and McClenaghan (1980:164, after Endler, 1977) who viewed dispersal as "movements of only a short distance made by individuals away from a natal site." According to these authors, dispersal movements are "continual rather than periodic and can occur within or between generations." This view of dispersal allows me to consider animal movements without restrictions in their timing or magnitude. Measurements Dispersal of individuals or propagules has been demonstrated in some stage of the life cycle of almost every species of plant and animal (Howard, 1960). Although dispersal in small mammals has recently been the focal point of much ecological and behavioral research (Gaines and McClenaghan, 1980), broad generalities of the causes and consequences of this phenomenon have been slow to emerge. This is partly due to inconsistencies in conceptual definitions described above to vague and Operational definitions of dispersal. Therefore, before I describe attributes of dispersers and hypotheses prOposed to explain the underlying proximal mechanisms of dispersal, I will critically review the methodologies that apply to the study of dispersal. One fundamental problem with studying dispersal in small mammal populations is a logistic one. Although dispersal is often characterized as involving a series of movements away from a natal site, the processes are difficult to observe directly, particularly among small, secretive, nocturnal mammals. Direct visual observation of these movements is possible in studies of large, diurnal species like ground squirrels (e.g. Armitage, 1973; Carl, 1971; Downhower and Armitage, 1981; King, 1955; Slade and Balph, 1974; Yeaton, 1972). In contrast, small, nocturnal species must be studied by using indirect methods, such as live-trap sampling‘ of ‘undisturbed. or' depopulated grids, resource structuring experiments, fenced enclosure experiments, nest-box; monitoring, radiotelemetry, radioactive tagging, and. other tracking methods (e.g. marked feces, chemically-treated tracking papers). By using live-trap methods, dispersal can be estimated by sampling individuals that immigrate into a study area. The study area to be monitored can either be left undisturbed as a control, or depopulated (residents removed). In control study areas, newly-recruited (captured) animals can be young animals born in the areas, previously uncaptured older residents, or immigrants. Since it is difficult, if not impossible, to distinguish the source of newly captured animals, dispersal rates can only be estimated. Fairbairn (1977a, 1978a) estimated dispersal rates for h m; austerus by comparing rates of survival and recruitment of mice within a study area. Fairbairn assumed that high rates of movement by animals within the general papulation should be reflected in high rates of recruitment and low rates of survival on the control areas. In spite of the difficulties in distinguishing the source of newly-recruited animals, this method has provided insight into the mechanisms of recruitment (Boonstra, 1978; Fairbairn, 1977b, 1978a; Hansen and Batzli, 1978; Sullivan, 1977). In removal-trapping studies, dispersal rates are estimated in two ways. One assessment of dispersal is made by determining the number of former marked residents of control grids which appear in traps on removal grids. Alternatively, the rate of dispersal can be measured by comparing the number of dispersers which colonize a removal area to the 10 population density of a control area. Implicit in removal studies are two assumptions: 1) that animals trapped on the removal area are indeed immigrants and not individuals making shorter, exploratory movements from closely surrounding areas (e.g. Stickel, 1946; Joule and Cameron, 1975), and 2) that animals taken from removal grids will not influence the population dynamics of animals from adjacent control grids (Gaines e_t a_l., 1979b). Removal-trapping has been the most commonly used method for studying dispersal in small mammals (e.g. Fairbairn, 1978a; Gaines e_t fl” 1979b; Joule and Cameron, 1975; Krebs gt a_l., 1978; Myers and Krebs, 1971; Stickel, 1946; Sullivan, 1977; Tamarin, 1977). In most removal experiments, the authors assume that removal-trapping measures dispersal patterns that would occur in unmanipulated populations. This method was criticized by Dobson (1981) who argued that removal areas constitute artificial dispersal sinks that may produce a "vacuum effect" (after Gaines st 21., 1979a). By comparing dispersal data from removal and non-removal (unmanipulated) study areas, Dobson demonstrated that data collected from. removal studies ”cannot be assumed to represent an accurate measure of...dispersal...occurring in unmanipulated papulations..." Dispersal has also been studied by restructuring the environment. Gaines g a_l. (1979b) compared a field burning method with two other experimental techniques (removal-trapping, fenced enclosures) ix: an analysis of dispersal in fluctuating Microtus ochrogaster populations. The authors found remarkable consistency in the data obtained by the three experimental methods. In this study, the vegetation on two experimental grids was burned, enabling the authors to examine the dispersal of individuals into this suboptimal habitat. A major 11 assumption of this experiment was that all animals subsequently live-trapped on the burned grids were dispersers. The authors admitted, however, that this assumption was not completely valid because several animals captured on the burned grid were, in fact, residents that survived the burning. Several investigators have added supplemental food. to natural populations of voles and mice and examined the effects of this manipulation on immigration rates and other demographic characteristics (Desy and Thompson, 1983; Flowerdew, 1972; Gilbert and Krebs, 1981; Smith, 1971; Taitt, 1981; Taitt and Krebs, 1981; Taitt st 31., 1981). Taitt and Krebs (1981) suggested that experimental structuring of other resources (e.g. estrous females, nest sites) in future studies could lead to a greater understanding of the influences of resource availability on spacing behavior (see also Hypothesis Testing below). Fenced enclosures have enabled workers to control crucial variables (e.g. immigration rates, predation risk) and identify all individuals that attempt to "disperse" by leaving the enclosure via exit holes or pitfall traps (Gaines 35.21,, 1979b; Riggs, 1979, from Gaines and McClenaghan, 1980). Implicit in this technique is the assumption that all individuals have equal opportunities to "find" the exit holes or pitfall traps. Gaines _e_t_:_ 11;. (1979b) conceded that individuals whose normal home ranges were close to the fence lines and exit holes were more likely than other voles to enter the exit holes. Fenced enclosures have also provided a method for investigating the role of dispersal as a population regulating mechanism. The enclosure used by Gipps and Jewell (1979) was vole-proof and prevented any iuigration or emigration. During the course of their study, 12 population densities of bank voles (Clethrionomys glareolus) within the enclosure reached extremely high .1evels, an observation repeatedly observed in other enclosure studies in which dispersal was prevented (Boonstra and Krebs, 1977; Crowcroft and Rowe, 1957; Lidicker, 1976; Krebs 35 al., 1969). Drift fences with pitfall traps have been used to study dispersal movements in several species of small mammals, including the old-field mouse P; polionotus (Briese and Smith, 1974; Garten and Smith, 1974). Implicit in all the methods described above which utilize either pitfall traps or exit holes, is the assumption that mice "captured" are actually attempting to exit or enter the study grid. Recent technological advances in radiotelemetry methods (for reviews see Amlaner and MacDonald, 1979; Cheesman and Mitson, 1982) have permitted detailed investigations of movement and space use by small voles and mice. Radiotelemetry studies have revealed patterns of activity and space-use among free-ranging brown lemmings, Lemmus trimucronatus (Banks gt al., 1974), red-backed voles, C; gapperi (Chute §£_ al., 1974; Herman, 1977), collared lemmings, Dicrostonyx ggoenlandicus (Brooks and Banks, 1971), meadow voles, M; ochrogaster (L.L. Getz, C.S. Carter, pers. comm.), white-footed mice,_§; leucopus (Madison, 1977; Mineau and Madison, 1977; B. Ormiston, pers. comm.; Holff and Hhrlbutt, 1982), and deer mice, P; maniculatus (Murphy and Gidner, 1982; Wolff and Hurlbutt, 1982). Compared with live-trapping methods, radiotelemetry techniques provide more precise measurements of an animal's activity and movement. However, radiotelemetric methods have their own limitations. Implicit in these methods is the assumption that they do not interfere with normal behavior and movement 13 capability. The use of radiocollars could result in reduced activity of monitored animals (Hamley and Falls, 1975; Webster and Brooks, 1980). In spite of recent improvements in radiotelemetric techniques, few quantitative data are available on the behavior of free-ranging small mammals at the time of dispersal. The lack of information is conspicuous for juveniles, in particular, because of their small size (< 15 g for many mouse-sized species). Most reliable radiotransmitters available on the market today are still too large and heavy to be used for tagging animals that weigh less than 20 g. Given that the predominant dispersers in most small mammals are juveniles, and that radiotelemetric monitoring is not yet feasible for these small-sized animals, it is not surprising that little is known regarding the behavioral attributes of dispersers in the field. Consequently, most radiotelemetry studies are concerned with describing patterns of activity and space-use among free-ranging adults. In addition to radiotelemetry, remote censusing with radioisotope labels has also contributed to our understanding of movement patterns among free-ranging small mammals. 'Wolff and Holleman (1978) recommended in 25552 labelings of progeny as a technique for studying kin groups and dispersal of juvenile Microtus spp. and house mice, Mus musculus from their natal home range. Radioisotope labeling was also used in studies of movement in the field vole, M; agrestis (Godfrey, 1954; Myllymaki 35 al., 1971), Townsend's vole, M; townsendii (Hilborn and Krebs, 1976), and coast mole, Scapanus orarius (Schaefer, 1982). A limiting consideration for the use of radioisotope labeling in the study of movements in small mammals is the potential risk that the d5! C01 a: pa: rac‘ 11: Fem \ rel Pop! “SH -" If“ Sim 110: and HQSt 14 associated radiation exposure may produce detrimental effects which could be manifested in behavioral changes (Wolff and Holleman, 1978). Compared with radiotelemetric methods, the monitoring of radioisotope- 1abeled animals provides only limited information. This is due, in part, to limitations in the distance that Geiger counters can detect radioactively labeled animals. Another drawback is that monitoring is limited to only a few individuals at a time (e.g. Godfrey, 1954). Patterns of movement and space-use have also been inferred from tracking toe-clipped individuals in a population with the use of chemically treated tracking papers. Space-use by individual .M; musculus was determined in a natural population by using a combination of live-trapping and tracking paper methods (Fitzgerald g£_§l;, 1981). Spatial relationships among wood mice, Apodemus sylvaticus were inferred from data collected from tracking marked feces (Randolph, 1977). Finally, the use of nest-boxes in studies of free-ranging Peromyscus spp. has allowed interpretations of animal movements relative to the social and familial history of individuals in a population. Howard (1949), Dice and Howard (1951), and King (1983) used nest-boxes to study dispersal in populations of prairie deer mice (P; m;_bairdi). Nicholson (1941) and Trudeau 35 El; (1980) similarly studied populations of P; leucopus. Most microtines (Microtus spp.) do not regularly use nest-boxes (M. Gaines, pers. comm.), although Gipps and Jewell (1979) reported nest-box use by E; glareolus. The use of nest-boxes is most feasible for Peromyscus spp. An advantage of nest-box monitoring over live-trapping methods is that nest-boxes do not impose restrictions on movements of mice which be ad af 118: bi] bi: onl 15 choose to use them. On the other hand, a mouse which enters a trap and is captured is not only restricted from further movement, but is behaviorally removed from the population until its release. In addition, little is known about how trap spacing or trappability affects behavioral measures (Madison, 1980b). Another advantage of a nest-box study is that it provides an opportunity for determining the birth dates of animals born in the population. With a knowledge of birth dates, the age at dispersal or other events can be determined with reasonable certainty. In live-trapping studies, however, ages can only be estimated (see Attributes of Dispersers below). The monitoring of animals from birth provides an opportunity for comparing the timing of maturational events (e.g. puberty) with dispersal. For example, an interesting question often asked is whether dispersal precedes the onset of sexual maturation, or vice-versa (see also Attributes of Dispersers below). Thus far, this question has proven difficult to answer in removal studies that utilize live-trapping methods (e.g. Beacham, 1981). If a disperser is trapped in a depopulated grid and found to be in breeding condition, we have no way of determining whether the individual became sexually mature prior to dispersing, or after dispersing but prior to being trapped. Nest-box monitoring is at a disadvantage when it comes to measuring extensive dispersal movements of individuals which leave the study area (Howard, 1949; King, 1983). Dobson (1981) criticized the use of nest-boxes because he believed that they may influence social pressures that might influence dispersal tendencies. One could also argue that studies which attempt to describe dispersal movements of small mammals with the use of live traps can also be biased. For 16 example, trappability of deermice has been shown to be a function of the olfactory cues in the traps themselves (e.g. Mazdzer, 35 El' 1976) and of trap spacing (e.g. Howard, 1949). Despite these criticisms, the use of nest-boxes is recommended in dispersal studies which are planned to allow manipulation or control of resource variables such as nest sites, food, refuges, and protection from predators (see Hypothesis Testing below). Another fundamental problem with studying dispersal in natural populations lies in our inability to distinguish between mortality or dispersal as the cause of disappearance of animals from a study site. Although Hilborn (1975) experimentally separated disappearance of M; townsendii into dispersal and survivorship components, in most studies investigators do not separate these two sources of disappearance. Attributes of Dispersers A fundamental characteristic of all dispersers is their movement. We know about dispersal and associated movements from mapping the pattern of dispersion of individuals in space and time. It is often difficult to determine, from the map alone, either the causes for the observed dispersion patterns, or the perceptual mechanisms which enable the animals to move and to space themselves. Fortunately, we can collect demographic, behavioral, and genetic data which may provide clues to help explain the underlying mechanisms of dispersal. Even with these data, it still seems that much of what we know about the causes of dispersal and spacing patterns in small mammals is like a naive observer's understanding of a game of chess: the players are seen but the rules of movement and the underlying strategies and mechanisms remain obscure . 17 The study of dispersal is becoming increasingly common and much new data concerning the characteristics of dispersers are being collected. in: spite of the variety of life history patterns of the animals studied, some broad generalizations regarding the demographic and behavioral attributes of dispersers are slowly emerging (for other views see Gaines and McClenaghan, 1980; Dobson, 1982). Sex In many multiannually cycling species (e.g. Microtus spp.) studied, males were the predominant dispersing sex (e.g. Beacham, 1981; Boag and Murie, 1981; Dobson, 1981; Gaines _e£ 3L, 1979b; Keith and Tamarin, 1981; Krebs gt a]._._, 1978; McClean, 1982; Riggs, 1979, in Gaines and McClenaghan, 1980). Tamarin (1977) reported that M; pennsylvanicus exhibited shifts in the sex ratios of dispersers, with males predominating in the winter and females predominating in the sunnner. Seasonal patterns were also described for M; pennsylvanicus (Dueser gt al., 1981), and cotton rats, Sigmodon hispidus (Stout and Demmer, 1982). There were no significant differences reported in sex ratios between dispersers and non-dispersers in .3; maniculatus (Fairbairn, 1978a; King, 1983) or _P_._ leucogus (Nadeau it; al., 1981; Tardif, 1979). On the other hand, Gottfried (1982) found male P; leucopus to be the predominant dispersing sex. Age and Weight Perhaps the most prevalent finding from dispersal studies is the observed movement or disappearance of juveniles from their natal area at, or near, the age of puberty in P; maniculatus (Dice and Howard, 1951; Howard, 1949, 1960, King, 1983; Petticrew and Sadlier, 1974), P; 18 leucopus (Hansen and Batzli, 1978; Nadeau 33 al., 1981), Mus musculus (Anderson, 1970; Myers, 1974), Microtus spp. (Beacham, 1979; Keith and Tamarin, 1981; Webster and Brooks, 1981b), Clethrionomys Aglareolus (Mazurkiewicz and Rajska, 1975; Wiger, 1982), Sigmodon hispidus (Cameron, 1977), Reithrodontomys fulvescens (Cameron, 1977), Spermophilus spp. (Boag and Murie, 1981; Dunford, 1977; MCClean, 1982; Michener and Euchener, 1977; Rongstadt, 1965; Slade and Balph, 1974; Yeaton, 1972) and Marmota_§pp. (Armitage, 1981; Bronson, 1964; Davis gt al., 1964). Although dispersal movements are commonly seen in juveniles, they can also be observed in older animals in P; 1eucopus (Gottfried, 1982; Stickel, 1946), P; polionotus (Briese and Smith, 1974; Gentry, 1966), Mus musculus (Delong, 1967; Rowe st 31., 1964; Strecker, 1954), Microtus spp (Myers and Krebs, 1971; Pucek and Olszewski, 1971; Tamarin, 1977; Van ‘Vleck, 1968), Dipodomys sp. (Fitch, 1948), and Marmota flaviventris (Armitage, 1962, 1977). In some studies, adult dispersers predominated in certain seasons. This has been observed in 22. leucopus (Gottfried, 1979), Microtus spp. (Lidicker, 1976), Mus musculus (Newsome, 1969), Tamiasciurus hudsonicus (Kemp and Keith, 1970; Rusch and Reeder, 1978) and Gerbillus allenby; (Abramsky and Sellah, 1982). In comparing dispersers with residents, dispersers were found to be younger and lighter in weight in §;_maniculatus (Fairbairn, 1978a), Microtus sp. (Beacham, 1979; Gaines gt al., 1979a; Krebs 35 al., 1978; Myers and Krebs, 1971; Riggs, 1979, in Gaines and McClenaghan, 1980), Clethrionomys glareolus (Kozakiewicz, 1976), and Sigmodon spp. (Joule and Cameron, 1975; Stout and Demmer, 1982). In other studies of 385' DESI 19 Microtus, however, the ages of dispersers and residents did not differ (Tamarin, 1977; Verner, 1979, in Gaines and McClenaghan, 1980). In most of these studies, age could only be assessed from measures of body weights or inspection of developmental molts. Age is probably the most difficult character to measure accurately, particularly in animals from live-trapping studies. Dispersers caught in snap traps and killed can be examined more closely than live animals. Postmortem examinations can provide relatively accurate measurements of age, which can be determined by examining eye lens weights, closure of epiphyseal plates, and dental annulations (Pucek and Lowe, 1975). On the other hand, age determination of live-trapped animals must be based on morphological changes associated with age. Typically, body weight and pelage characteristics are used to define broad age categories (e.g. adult, subadult, juvenile, nestling). Known ages can be assigned to individuals sampled from nest-boxes if the date of birth was previously recorded. Reproductive Status According to Gaines and McClenaghan (1980), the reproductive condition of dispersers (regardless of age) showed considerable variation among the species studied. In many species, the reproductive status of dispersing males was a random subset of the resident male population. However, dispersing subadult male M; townsendii (Beacham, 1979, 1981; Krebs e_t al., 1978), _M_._ ochrogaster (Myers and Krebs, 1971), M; pennsylvanicus (Dueser 3.5 al., 1981; Keith and Tamarin, 1981), and M; breweri (Keith and Tamarin, 1981) were more likely to be reproductively active than subadult male residents. Dispersing males of all ages were more likely than residents to be in breeding condition 20 in E; glareolus (Kozakiewicz, 1976), M; californicus (Riggs, 1979, from Gaines and McClenaghan, 1980), M; breweri and M; pennsylvanicus (Tamarin, 1977). In reviewing the literature, Gaines and McClenaghan (1980) found no discernible pattern in the reproductive condition of dispersing females. In certain species (M; pennsylvanicus and Mh_breweri), dispersing females were more likely than resident females to be breeding when trapped (Tamarin, 1977). Nevertheless, in some studies, there was a greater proportion of resident females in breeding condition compared to dispersing females (Myers and Krebs, 1971; Gaines et al., 1979a,b; Keith and Tamarin, 1981). In certain seasons, female 3; leucopus dispersers were more likely to be reproductively active than residents (Tardif, 1979). Dispersing female M; townsendii (Krebs .25 al., 1976), and P; maniculatus (Sullivan, 1977) reached puberty at an earlier age than residents. Measurement of reproductive condition may also be problematical. For males, the reproductive status can be inferred by measuring the position of the testes. Scrotal testes imply reproductive competence, whereas abdominal testes are indicative of sexually immature status. A critical review of this procedure was provided by Jameson (1950). A more quantitative technique involves laparotomy and the use of a testicular index (Johnston and Zucker, 1980c). This method allows for repeated quantitative assessment of gonadal status but has been reserved mainly for laboratory studies. Repeated and frequent surgical gonadal assessment of males in the field is not recommended. For females, vaginal patency is used as an indicator of puberty (Clark, 1938; Whitsett and Miller, 1982; for a critical review see 21 Rogers and Beauchamp, 1974). Pregnancies cannot be verified either visualLy or by palpation until relatively late in gestation for most small mammals. Parity is difficult to assess for females sampled in live-traps, and is possible for females monitored in nest-boxes only if the female uses the nest-boxes to raise her litters. Genotype Electrophoretic analyses of blood plasma proteins revealed genetic differences between residents and dispersers in some species. Myers and Krebs (1971) reported finding differences in the genetic make-up of residents and dispersers at the leucine aminopeptidase (LAP) and transferrin (TF) loci in ‘M; pennsylvanicus and ‘M; ochrogaster populations. Similar results were described for an esterase locus in ‘M; ochrogaster (Pickering 35 al., 1974), M; pennsylvanicus (Keith and Tamarin, 1981), and M; breweri (Keith and Tamarin, 1981). In contrast to the above studies, Verner (1979, in Gaines and McClenaghan, 1980), and Riggs (1979, in Gaines and McClenaghan, 1980) found no evidence for differential genetic composition in residents or dispersers. Although much of the work done on genetic analysis of dispersal has been performed on cycling vole species, some analyses have also been done on non-cycling small mammal species. Michener and Michener (1977) found no evidence for a differential loss of genotypes from a population of Richardson's ground squirrels (SpermOphilus richardsonii). Although genetic differences have been found at particular loci, it remains to be shown that these genetic differences between dispersers and non-dispersers are causally related to dispersal, or any other behavior. 22 Behavior Field Studies In natural papulations, the behavioral attributes of dispersers can be inferred from direct, visual observations (limited mostly to larger, diurnal species), radiotracking, monitoring of radioisotope- labeled animals, and measurement of skin wounds. Most behavioral measures recorded in the field are related to aggression. The social subordination and genetic behavioral polymorphism 'hypotheses for a proximal mechanism of dispersal predict that aggressive behavior increases with population density, and interactions between dominant and subordinate animals results in the dispersal of the latter (see Behavioral Hypotheses below). From direct observations of yellow-bellied marmots Marmota flaviventris, Armitage (1973) attributed the dispersal of juvenile males to aggression. A similar pattern was described for male arctic ground squirrels, Spermophilus undulatus by Carl (1971). Pfeiffer (1982) credited aggression as the cause for the disappearance and dispersal of juvenile female Wyoming ground squirrels (§E_elegans). On the other hand, observed aggression in other ground squirrels did not explain juvenile dispersal. In round-tailed ground squirrels, -§; tereticaudus, mothers were aggressive only towards non-kin but this did not explain the dispersal of juveniles from their mother's home range (Dunford, 1977). No relationship between aggression and dispersal was detected in populations of Uinta ground squirrels, _S_:_ armatus (Slade and Balph, 1974), S; richardsonii (Yeaton, 1972), or M; flaviventris (Svendsen, 1974). 23 The skins of secretive, small mammals can be inspected, postmortem, for evidence of wounds caused by attacks from other animals (e.g. Rose and Gaines, 1976; Ziesenis £5 31;, 1975). Higher incidences of wounding were recorded as density increased in populations of M: musculus (Rowe 35 al., 1964; Southwick, 1958), and & pennsylvanicus (Christian, 1970). Krebs (1964) found higher levels of wounding in lemmings during peak population densities, but when population density declined, levels of wounding remained high. In other microtine species, no relationship between population density and wounding could be found (Batzli and Edtelka, 1971; Rose and Gaines, 1976; see also Laboratory Measures below). If aggression is indeed an important mechanism for eliciting dispersal in subordinate individuals, and if aggression between dominant and subordinate individuals resulted in wounding and dispersal of the latter, then one should expect to find more incidences of wounding in dispersers than in non-dispersers. I found no evidence from the literature that supports this prediction. Christian (1970, 1971) suggested that subordinate (lightweight?) voles disperse as a result of increased aggression arising from increasing levels of density, but he found no evidence of scarring in dispersers. Furthermore, if lightweight, subordinate individuals are the predominant disperser class (e.g. Baird and Birney, 1982), one would expect to see higher levels of wounding in lighter weight individuals. Rose and Gaines (1976) found no difference, however, in wounding levels of lightweight versus heavyweight ME_ ochrogaster. The authors suggested that lighter weight voles may not be the predominant dispersers. CC 53' sta SCI for the rGSid \ E resid, 24 Krebs and. Myers (1974) discussed the problems associated. with collecting and interpreting skin wound data (see also Behavioral Hypotheses: Do They Explain General Attributes of Dispersers? below). In larger, diurnal species like ground squirrels, agonistic interactions can sometimes be visually recorded (e.g. Armitage, 1977; Carl, 1971). Laboratory Studies In paired encounters in neutral arena tests in the laboratory, the aggressive behavior of residents and dispersers have been compared. In staging paired encounters, two individuals (usually two males who are strangers) are placed into a ”fighting arena" and allowed to interact for a limited period of time. Broad categories of agonistic behavior (Clarke, 1956; Eisenberg, 1963) are recorded for each individual during the test. Myers and Krebs (1971) found no consistent patterns in aggressive behavior for the two Microtus species they studied. Krebs g£_al. (1978) found dispersing M; townsendii to be more submissive than residents. These authors concluded that dispersing M; townsendii were subordinate individuals, despite finding more wounding in residents than in dispersers. Resident §h_mt_austerus males were more likely to exhibit aggression during pairwise encounters with a laboratory-trained fighter than dispersing males (Fairbairn, 1978b). Reich 35 31; (1982) found aggression to be independent of the dispersal rate for male M; breweri. Laboratory studies have also compared exploratory behaviors of residents and dispersers. Myers and Krebs (1971) found dispersers in .M; pennsylvanicus and M; ochrogaster to be less exploratory than residents in a vertical maze test. Fairbairn (1978b) reported similar 25 results for dispersers in _P_._ m; austerus. On the other hand, in a study of & leucopus, Tardif (1979) found dispersers to be more likely to cross a water barrier and drop into an unfamiliar area than residents (a measure of exploration). In a test of general activity, dispersing & m; austerus showed a tendency to be more active than residents (Fairbairn, 1978b), but Myers and Krebs (1971) failed to detect a difference in activity between residents and dispersers in Microtus spp. Other behaviors have also been examined. Tardif (1979) compared the behavioral responses of residents and dispersers in tests designed to measure the animal's willingness to sample novel fluids and eat a diversity of foods. He hypothesized that mice dispersing across unfamiliar habitat were more likely than residents to encounter unpredictable and novel fluids and foods. Tardif also reported that more field-caught dispersers than residents: 1) sampled novel fluids, and 2) exhibited higher feeding diversity. Savidge (1974) examined social factors which influence the rate at which juvenile f; m; bairdi leave their natal nest site. In his experiments, parents and siblings played important roles in determining the rate of natal dispersal of the juveniles. Juveniles were attracted to their father and typically departed from the natal site with him. Lactating mothers, who were classified as aggressive, increased the rate of departure of their previous litter. A serious problem with all the studies described above is whether the behavioral tests are valid assays of dispersal-related behavior. The behaviors measured in the laboratory may not be related to the behavior of dispersers in the field (see also Gaines and McClenaghan, 26 1980). Anderson (1980) criticized the practice of extrapolating conclusions of laboratory’ arena encounters to natural populations. According to Anderson, the practice of assigning either dominant or subordinate status to animals in the field based on observations of aggression in neutral arenas was "on the par with inferring through observation of boxing matches that there are two kinds of human beings." Furthermore, the relationship between aggression and dominance, and that between aggression and dispersal, has been criticized (Bekoff, 1977; see also review by Bernstein, 1981). There is evidence from studies with rats and other rodents which indicates that aggressive individuals may not necessarily be dominant, and vice-versa (Archer, 1970; Barnett, 1975). Another problem with laboratory tests is that the conditions of the test are not relevant to conditions that prevail in the field at the time of dispersal. For instance, consider the validity of using only strangers in neutral arena tests (e.g. Krebs, 1970). If individuals who disperse do so as a result of encounters with only strangers, then a test for strangers in a neutral arena is valid. But, suppose would-be dispersers are influenced differently by encounters with familiar individuals than they are by strangers? If would-be dispersers in the field are as likely to interact with familiar animals as with strangers, then how valid are interpretations of data derived from neutral arena tests that only pair strangers? All individuals in the papulation are not likely to be perceived equally by would-be dispersers. Indeed, Vestal and Hellack (1978) demonstrated that wild-caught Peromyscus spp. are capable of recognizing and distinguishing neighbors from strangers. In this study, strangers were 27 more likely than neighbors to be aggressive in paired encounter tests (see also Grau, 1982; Halpin, 1981; Kareem and Barnard, 1982; Rajska-Surgiel, 1976). Tests which ignore this behavioral asymmetry are likely to lead to interpretations that have dubious value. The same can be said for behavioral tests which ignore the experiential history of the tested individuals. Behavioral differences observed between dispersers and residents are likely to be explained by differences in the experiences of these two groups of individuals. For instance, in Tardif's (1979) water barrier test, wild-caught .3; leucopus were more likely than laboratory-reared g:_leucopus to cross the barriers. In his tests, 42% of the wild-caught mice (n=33) crossed the barriers, compared with only 182 of the laboratory-reared mice (n-62). In an examination of behavioral development, Tardif and Gray (1978) manipulated the early feeding experiences of two groups of laboratory-reared £;| leucopus and later compared their feeding diversity scores with those of wild-caught residents and dispersers. Both laboratory-reared individuals with experience of an unpredictable diet, and wild-caught dispersers, were classified as generalized feeders. Similarly, residents, and laboratory-reared mice exposed to a stable diet were more specialized feeders. From the results of this experiment, we learn that experiences unrelated to actual dispersal may confound the results of tests used to compare the behavior of dispersers and residents. Despite the limitations described above, several tests did distinguish the behavior of dispersers from that of non-dispersers. Could any of these tests be used to predict, a priori, whether a given 28 individual is likely to disperse? If a laboratory test were capable of this, it would suggest that some individuals might have a genetic predisposition to disperse as has been suggested by Howard (1960). In a pilot experiment I chose to test the predictability of Tardif's (1979) water barrier test by first testing young field-caught h a: bairdi in the barrier apparatus and then reintroducing them to their home nest in a field study site (see King, 1983). I tested the hypothesis that mice which cross the barriers in the laboratory would be more likely than non-crossers to disperse after being returned to the field. The results of this experiment were equivocal; barrier-crossing mice were no more likely than non-crossers to disperse (disappear). I observed similar results for a general activity test and Tardif's novel fluid test. The results of these experiments suggest two interpretations. First, animals may have a predisposition to disperse, but the tests chosen were not valid assays of dispersal-related behaviors. Secondly, individuals disperse as a result of unique social and non-social experiences, and not as a result of some predisposition factor. Behavioral Hypotheses Three of the five hypotheses for dispersal reviewed below are extensions of behavioral hypotheses prOposed to explain the mechanisms underlying population regulation of multiannually cycling species (see Gaines and McClenaghan, 1980). These are the social subordination, genetic-behavioral polymorphism, and presaturation-saturation dispersal hypotheses. Gaines and McClenaghan are commended for their review of the literature and attempt to outline testable and alternative hypotheses that may account for proximal causes of dispersal in small re th in. in: The suE Fur '35: dis; than and 13C?! Suba; resid 29 mammals. Unfortunately, it is difficult to discriminate three of their outlined hypotheses on the basis of predictive value. Compared to the three hypotheses referred to above, the social cohesion and sexual search hypotheses described below offer alternative predictions (see also Table 1). Social Subordination Hypothesis In 1970, Christian proposed that intraspecific competition resulting from increases in population density is a "major component of the force to disperse." (p. 85). According to this hypothesis, increasing population density leads to an increase in intraspecific interactions which, in turn, result in increased levels of aggression. The increased levels of aggressive interactions between dominant and subordinate animals may lead to a forced dispersal of the latter. Furthermore, Christian (1971) predicted that levels of aggression would be influenced by the proportion of fertile males in the population. These premises were based on laboratory data showing that agonistic behavior increases with population density (Chitty, 1967; Christian and Davis, 1964). The social subordination hypothesis results in the following predictions (Gaines and McClenaghan, 1980): 1) levels of aggression will be positively associated with population density, 2) dispersal will be more common during phases of peak population density than during phases of increasing or declining density, 3) aggression and dispersal will be influenced by physiological responses to increasing population density, 4) the predominant dispersers will be subadult males that have just reached puberty, and 5) compared with residents, dispersers will be social subordinates. 30 sumo «Iowan-«e acuueuau-nnoum hams amonommuv Esau-nau-m asauuvuus mu mus-coaumuou seduces-u I m I n unuuuemum mu au3090nu-uou Ibuucuauuuuh I h I u I a wouuuqoum mu sq£maouulumu unquumnuz zuzcmm acaxmm 29—musou at—ocw BL a<=mn10~bm2uu U a 20~H~Gz~ maze: A<—ocm u<fl3 m¢O~><=mn OZ~U<=un nonhuzub ¢c~><=fln nuuhm=<¢ ° + TRAP 8 @} —=50metem . p ‘ , . COUPLET w s. TRAP LINE :50". * 0 g 0 o '°°~§m.wda<:§ .~+ + NAGADOIN ROAD Figure 3.1 Map of the mouse cities study area. 79 predominated, with interspersed patches of Solidago spp., Daucus carota, Cirsium vulagare, Aster spp., and Rubus spp. occurring secondarily. Small stands of Acer negundo and Populus deltoides were located to the north of the study grid and several small trees of these species were interspersed throughout the grid. Two fence rows of mature trees and shrubs were located to the south and north of the study grid. An ephemeral pond, varying seasonally and annually in size and water depth, was located to the south of the study grid and fence row. A 2 m high chain link fence paralleled Hagadorn and Sandhill roads, and a gate to the south of the grid provided privileged access to the study site. During the course of my research, areas adjacent to the study site were used in other experiments by members of Michigan State University's Departments of Botany and Zoology. Abundant small mammals at the study site included Microtus (pennsylvanicus, Zapus hudsonicus, and Blarina brevicauda. Less abundant species included Peromyscus leucopus, Scalopus .aquaticus, Sorex cinereus, Marmota monax, Procyon lotor, Mustela nivalis, and Mustela frenata. Occurring rarely were Felis domestica and vulpes vulpe . Four species of snakes were found at the study site; these included Thamnophis sirtalis sirtalis, Thamnophis butleri, Lampropeltis triangulum triangulum, and Storeria dekayi dekayi. Trapping records from a study preceding this one at the same site (King, 1983) revealed that 3;.2; bairdi were probably not originally present. This study site does not offer the most favorable _P_._ P; bairdi habitat (King, 1983). In the mid-Michigan area this subspecies is most commonly found in or near disturbed agricultural fields of 80 alfalfa, corn, oats, wheat or clover (J. King, pers. comm.; Linduska, 1950). The nearest agricultural field to the study site was an isolated corn field located approximately 2 km to the northeast. Nevertheless, trapping records kept since the completion of King's study and prior to the beginning of the present research revealed the presence of some indigenous g; m;_bairdi. These mice, however, were probably the wild-born descendants of experimental mice introduced in earlier experiments. As part of the general scheme to provide experimental control of available nest sites, food, water, and refuges from predators, the study site was provided with microhabitats that were proven to be favorable to the mice (King, 1903). The microhabitat enclosures consisted of 32 cylinders arranged in pairs, with 16 pairs forming a square grid, 120 m on a side (Figure 3.1). Each pair of cylinders was spaced 40 m apart. The cylinders (hereafter called arenas) were constructed from 3.2 mm sheet aluminum. Each.arena was 2 m in diameter and 100 cm high. ‘The bottom of each arena was buried 15 cm below the ground surface (Figure 3.2) to discourage burrowing animals from gaining entry to the arenas. Two arenas were spaced 2 m apart to form a pair (hereafter called a couplet). Each arena was provided with four shelters constructed from concrete building blocks (20 cm x 20 cm x 40 cm) and concrete patio blocks (5 cm x 20 cm x 40 cm). Access passageways into and.between the three chambers of each building block (hereafter called a nest) were provided by cutting 2.5 cm x 2.5 cm square notches (Figure 3.2). A concrete patio block was placed at the top and bottom of each nest. A 30 cm x 50 cm plywood roof, waterproofed with marine varnish, was placed between the 81 top of the nest and the patio block to provide extra insulation and waterproofing for the nest below. The four nests were placed strategically within each arena to provide mice with a diversity of available microhabitats. Two of the nests were placed at ground level on opposite sides within the circular arena, one was buried so that its roof was at ground level, and the fourth was placed upon a building block laying on its side in the center of the arena (Figure 3.2). Thus, each arena had one subterranean, one elevated, and two surface-level nests. A small 2.5 cm diameter exit hole, located 20 cm high on the arena wall, allowed passage of mice into and out of the arena. One of the surface nests was placed directly beneath the exit hole and an additional surface nest was placed outside of the arena wall beneath the hole (Figure 3.2). Peromyscus m; bairdi readily climb the surface nests and easily locate the exit hole. The height of the exit hole served as a deterrent to arena entry by more- terrestrial vertebrates such as M;_pennsylvanicus, B; brevicauda, g; cinereus, and snakes. Each arena was provisioned with supplemental food, water, and nesting material. Unshelled sunflower seeds were added on a regular basis to one of the surface nests in each arena (Figure 3.2) and to additional surface nests placed halfway (20 m) between couplets. A second surface nest; and the subterranean. nest received. compressed cotton nesting material (Nestlets, Ancare Corp.). Fresh Nestlets were added to these nests on a regular basis as needed during the course of the experiments. Supplemental water was provided during seasonal dry periodsm Periodic maintenance at the arenas included trimming weeds and grasses when they obstructed the nests, realigning displaced 82 .xonlumoa a can «sous an wo «assuage oauoaonom N.m ouswfim hmwz ouEDOOO 02.301m om>0§m¢ .343 NEW—.30 It; .34; >> 5:5: ._. zm3m m is o p30 .191 E: my 263 me u 0.275 (C. Anderson; pers. comm.)), these factors were pooled with experimental error to gain degrees of freedom for testing the remaining effects. Due to the unbalanced data, reduction in sums of squares were calculated to provide the most conservative estimate of the sums of squares due to any main effect (Speed and Hocking, 1976). Factorial and one-way ANOVAS were conducted with the aid of SPSS computer programs (ONE-WAY and MANOVA). Linear regression plots were made with the assistance of an SPSS plotting program called PLOT. T-tests were performed with the aid of the SPSS program called T-TEST. Results Effects of Treatments on Attainment of Puberty The null hypothesis that attainment of puberty is independent of treatment could not be rejected (X2 - 0.61; df - l; p > 0.60; Table 3.2a). 92 Table 3.2a Frequencies with which melatonin- and control-treated juveniles (at least 30 days old) became pubertal or remained non-pubertal. Treatment (Sexes, Seasons Pooled) Response Melatonin Control Pubertal 68 68 Non-Pubertal 34 26 Totals N=102 N=94 2 x = 0.61, df = 1, p > 0.60 Melatonin-treated mice, present in the population for at least 30 days, were as likely to become pubertal as controls that were present in the population for a comparable length of time. Further tests of the independence of the primary criterion of interest, attainment of puberty, and the secondary criteria of season (Summer 1981, Spring 1982, and Summer 1982) and sex were performed. The acceptance or rejection of the null hypotheses is based on two-tailed chi-square tests. Puberty was found to be independent of sex (X2 - 0.001; df . 1; p > 0.06; Table 3.2b), but not season (X2 - 11.13; df =- 2; p < 0.01; Table 3.2c). A significantly greater percentage of mice became 93 Table 3.2b Frequencies with which males and females became pubertal or remained non-pubertal. Sex (Treatments, Seasons Pooled) - Response Male Female Pubertal 66 7O Non-Pubertal 29 31 Totals N=95 N=101 2 x = 0.001, df =1, p > 0.60 Table 3.2c Frequencies with which individuals in each season became pubertal or remained non-pubertal. Season (Treatments, Sexes Pooled) Response Summer 1981 Spring 1982 Summer 1982 Pubertal 29 71 36 Non-Pubertal 17 16 27 Totals N=46 N=87 N=63 x2 = 11.13, df = 2, p < 0.01 94 pubertal in the spring season than summer. Following procedures outlined by'(HJJ.(l978), I conducted further tests involving two-way combinations of treatment, season, and sex in an attempt to discover any dependencies among the secondary criteria as they affected the attainment of puberty. The results of these tests (Table 3.2d-e) reveal that, regardless of treatment, male mice in the spring season were more likely to become pubertal than males in either of the summer seasons. Although melatonin treatment had no effect on the number of mice which attained puberty relative to control-treated mice, melatonin-treated individuals reached puberty at an older age than controls as predicted (F - 34.400; df 8 1/113; p < 0.001; Table 3.3, Figure 3.4). Table 3.3 Analysis of variance of age at puberty. Analysis of Variance Source of Variance DF MS F PROB Treatment 1 2033.746 34.401 p < 0.001 Sex 1 167.966 2.840 NS Season X Treatment 2 146.392 2.476 NS Season X Treatment X Litter 2 202.694 3.428 p < 0.05 Error 113 59.118 Total 119 There was also, however, a significant three-way interaction of treatment by season by litter order (F = 3.428; df - 2/113; p < 0.05; Table 3.3). Melatonin-treated individuals from the first-born litters in all seasons reached puberty later than controls. In both summer seasons, however, the age that second-litter mice reached puberty was 95 >5 mo.o A a .m n a .H a we .No.o u NQ¢~ ngfim .m> HwGfi HUSH—m N no.0 v a .m u a .H u we .ow.m n ma sass umeaam .m> Nwoa maauam no.0 v a .m u e .H u as .ma.m n wx Nmsa umaasm .m> was“ maauam Amoamzv comwumaaoo commom cmozuom o~.o .A a .N u we .mo.s n Nx mo.o my a .N u we .mo.m u we mmuz osuz omuz wan asuz omnz masses ms w m Na m m Hmuumnsmuaoz oN Nm ma as am as Hmsumaam Nmma umaaam ~m¢z maauam swam usessm Nmma amassm mass waauam anal smassm mmaoammm mmHmEmm mafia: Atopooa mucosummmHv Mom .kuuonsalco: vocwmaou no Hmuuonsn season Asommow sumo Lowv monEow can mmama Loans sags quucosvmum vN.m sandy 96 86 A a .N n :0 .8; u New 26 A a .N n .3 .26 u Nx amnz Hquz Hmuz Hmnz ocuz mmuz mamuoe q~ o 0 ma OH dfi Hmuuonsmlaoz ma mm ma ma cm «A Hmuuonsm Nwmfi seesaw Nwmfi wcauam fiwmfi seesaw Nwmfi umsasm mmmg magnum Hmmfi seesaw mmsoamom Houucoo CHflOUme—L Avoaoom mmxomv ucoEummuH .Hmuumpaa memos: Acommom comm powv mmafico>=h voumwuulfiouucoo was Icfi=0umaoa gowns saws mowocoavoum w~.m magma 97 .m.m u H mumoqvcw moawa Hmouuuw> .huuoasa um owm coo: ¢.m shaman >Pmmm3a Nwm_ mus—5:5 Nam. ozEam _mm. «Us—23m N amt... _ «mi... N 5:... _ amt... . «EC... .553 on: on: m7: 9.: _Nucomuc t... or: Nu: m»: 2": __nc .928 _H— 5.3.66.2 g on (S AVG) 39V 98 only slightly greater for melatonin-treated individuals than controls. The preliminary, mixed effect ANOVA revealed no significant effect of mother on the age that mice attained puberty (Table 3.4). Table 3.4 Preliminary mixed-effect analysis of variance of age at puberty. Analysis of Variance Source of Variance DF MS F PROB Treatment 1 32.18 1.25 0.267 Sex 1 37.32 1.45 0.232 Litter 1 13.13 0.51 0.476 Season 2 10.04 0.39 0.676 Treatment X Sex 1 6.18 0.24 0.629 Treatment X Season 2 59.46 2.31 0.106 Treatment X Litter 1 0.51 0.02 0.881 Sex X Litter 1 3.35 0.13 0.720 Sex X Season 2 7.72 0.30 0.739 Litter X Season 2 9.27 0.36 0.701 Treatment X Litter X Season 2 71.81 2.79 0.068 Treatment X Sex X Season 2 0.02 0.00 0.999 Treatment X Sex X Litter 1 1.80 0.07 0.796 Sex X Litter X Season 2 0.26 0.01 0.989 Sea X Trt X Lit X Sex 2 0.77 0.03 0.967 Mother Within Season 18 4.89 0.19 p>>0.50 Error 75 25.74 As expected, melatonin-treated mice weighed more at puberty than controls (F - 17.809; df 8 1/118; p < 0.001; Table 3.5, Figure 3.5). Furthermore, there was a significant interaction of season by sex (F - 3.556; df - 2J118; p < 0.05; Table 3.5), with males from the spring season consistently weighing more at puberty than females. As was the case for age at puberty, there was no significant effect of mother on the weights of mice at puberty (Table 3.6). 99 Table 3.5 Analysis of variance of weight at puberty. Analysis of Variance Source of Variance DF MS F PROB Treatment 1 47.801 17.809 p < 0.001 Treatment X Sex 1 2.692 1.002 NS Season X Sex 2 9.545 3.556 p < 0.05 Season X Treatment X Litter 2- 6.952 2.590 NS Error 112 2.684 Total 118 Table 3.6 Preliminary mixed-effect analysis of variance of weight at puberty. Analysis of Variance Source of Variance DF MS F PROB Treatment 1 5.36 2.33 0.104 Sex 1 1.10 0.48 0.490 Litter 1 0.55 0.24 0.627 Season 2 0.76 0.33_ 0.719 Treatment X Sex 1 3.73 1.62 0.206 Treatment X Season 2 0.01 0.00 0.998 Treatment X Litter 1 0.01 0.00 0.998 Sex X Litter 1 0.21 0.09 0.766 Sex X Season 2 14.49 6.30 0.003 Litter X Season 2 0.44 0.19 0.824 Treatment X Litter X Season 2 5.27 2.29 0.108 Treatment X Sex X Season 2 0.32 0.14 0.870 Treatment X Sex X Litter 1 0.07 0.03 0.859 Sex X Litter X Season 2 1.17 0.51 0.603 Sea X Trt X Lit X Sex 2 0.51 0.22 0.800 Mother Within Season 18 0.05 0.02 p2>0.50 Error 75 2.30 As predicted, melatonin-treated females were less likely to 2 become pregnant than control females (one-tailed X - 5.73; df - 1; p < 0.025; Table 3.7). Fifty-seven percent of the control females 100 .m.m a .I... mucus—J was: Hmowuum> .huuonaq um unmask new: n.n ouswwm >mem3a Nwm_ mmZEDm Nmm_ 02QO _mm. KNEE—am 3.25“. 3.2: «342mm $.22 $25.. 8.2: or: a... a... mu: 9.: an: enema": once": Tc .3: O _ 93 : _ - ('6)1H0I3M .2200 D 5:20.02 § m. 101 Table 3.7 Frequencies with which melatonin- and control-treated females (at least 50 days old) became pregnant. Treatment (Seasons Pooled) Responses Melatonin Control Pregnant 3 8 Not Pregnant 15 6 Totals N=18 N=14 2 *X = 5.73. df = 1. n < 0.025 *one-tailed (at least 50 days old, an age at which first pregnancies can be visibly detected) became pregnant compared with only 172 of the melatonin-treated females. Seasons were pooled in the above analysis due to the small number of females that became pregnant in each season 0 Effects of Treatment on Dispersal In: all of the populations studied there was a continuum in the length of residency for the mice.. I chose to distinguish between mice that disappeared from the population and those that stayed. These two categories were differentiated on the basis of whether residency was less than or greater than the median length of occupancy (in days) for all treated mice. In other words, individuals which remained in the population for at least the median length of occupancy (42 days) were considered to be stayers. Mice that disappeared from the population prior to this time were classified as mice that disappeared. As predicted, melatonin-treated individuals were less likely to disappear than controls (one-tailed X2 - 39.03; df - 1; p < 0.001; Table 3.8a). 102 Table 3.8a Frequencies with which melatonin- and control-treated juveniles disappeared from the population or stayed. Treatment (Sexes, Seasons Pooled) Response Melatonin Control Stay* « 72 29 Disappear - 33 82 Totals N=105 N=111 2 **X = 39.03, df = l, p ‘< 0.001 *Individuals who remained in population until at least the median age of occupancy for all treated mice (42 days) **One-tailed I performed additional tests to determine the degree of independence of residency (stay or disappear) from the secondary criteria of interests - season and sex. The acceptance or rejection of the null hypotheses is based on two-tailed chi-square tests. Length of residence was independent of sex (X 2 a 0.72; df - 1; p > 0.60; Table 3.8c) 5.. not season (x2 = 21.20; df . 2; p < 0.001; Table 3.8b). Table 3.8b Frequencies with which individuals in each season disappeared from the population or stayed Season (Treatments, Sexes, Pooled) Response Summer 1981 Spring 1982 Summer 1982 Stay 27 30 44 Disappear 19 7O 26 Totals N=46 N=100 N=70 x2 = 21.2, df = 2, p < 0.001 103 Table 3.8c Frequencies with which males and females disappeared from the population or stayed. Sex (Treatments, Seasons Pooled) Response Male Female Stay 46 55 Disappear 59 56 Totals 105 111 2 x = 0.72, df = 1, p > 0.60 Significantly more mice disappeared from the spring population than expected. Tests involving two-way combinations of treatment, season, and sex revealed that control mice (for each sex and in each season) were more likely to disappear than melatonin-treated individuals (Table 3.8d-e). Similarly, spring mice (for each sex and treatment) Table 3.8d Frequencies with which males and females (for ech treatment) disappeared from the population or stayed. Sex (Seasons Pooled) Male Female Response Melatonin Control Melatonin Control Stay 35 ll 34 20 Disappear 18 41 15 41 Totals N=53 N=52 N=49 N=62 2 x . 21.54, df = 1, x2 = 14.54, df = 1, p < 0.001 p < 0-001 104 Table 3.8e Frequencies with which individuals in each season (for each treatment) disappeared from the population or stayed. Season (Sexes Pooled) Summer 1981 Spring 1982 Summer 1982 Response Melatonin Control Melatonin Control Melatonin Control Stay 19 8 25 5 28 16 Disappear 4 15 23 47 6 20 Totals N=23 N=23 N=48 N=52 N=34 N836 X2 = 10.02, df = 1, X2 = 19.54, df = 1, X2 = 11.46, df = 1, p < 0.01 p < 0.001 p < 0.005 were more likely to disappear than mice from either summer season (Table 3.8f-g). In addition to disappearing from the population with less frequency than controls, melatonin-treated. mice disappeared at an older age (F a 34.759; df - 1/161; p < 0.001; Table 3.9, Figure 3.6). Table 3.9 Analysis of variance of age at disappearance. Analysis of Variance Source of Variance DF MS F PROB Treatment 1 3824.511 34.759 p‘<0.001 Sex 1 309.363 2.812 NS Litter 1 0.148 0.001 NS Season 2 698. 767 6.351 p < 0.005 Sex X Treatment 1 4.920 0.045 NS Error 161 110.029 Total 167 105 mo.o A a .m a a .a n he .sm.o a mx Neal “maasm .m> lama smaasm mo.o A a .m u a .a u we .om.~ u ex lama hoaaam .m> Neal wcauam N mo.o v a .m u a .a u as .ms.w a ax NmaH umaaam .6> ~wa~ waauam N "Andamammv mcomfiummaou commmm somaumm mo.o A a .m u a .z u we .mo.o mo.o v a .m u a .H u we .om.m mo.o v a .m n a .a as .mm.m N mx Nmmfi assasm .m> ~wm~ seesaw mx Nwm~ umaaam .m> Nwmfi wcauam mx Hmmfi umeasm .m> Nwmfi magnum "Ammamzv mcomfiummsoo commwm comauom mo.o v a .N u we .6N.m n Nx so.o v a .N u we .No.ms u Nx osuz msuz oNuz emuz mmuz omnz masses as on Na ma as a Hawaamman 6N As «a «a ma Ma swam Neal umaasm Nmma waauam lama sagasm mass umaaam Nmma waauam amaH umaasm smaoamum mHmEmm mam: Ammaoom muchummqu xmm .vohmum mo coaumasaoa ecu Eoum cwumoqmmmfiv Acommom some uowv monEow was moans noun: :uas mofiucoscoum mw.m manmh 106 mo.o .A a .m u a .s u we .mm.o a wx Nmms smeaam .m> sass umaasm Ho.o ”v a .m u a .a u as .sa.sa u mx Nmaa “maasm .m> Nma_ mcauam mo.o “v a .m n a .a he .mo.a n mx sass umaasm .m> News wcausm "Aaouucouv msomwummaoo mcommom comaumm mo.o .A a .m n e .H u as .Ho.o n mx Nmsa amassm .m> Hess umaaam mo.o my a .m u a .a n as .aa.~ n mx mass umeasm .6» Nmsa magnum no.0 V. a .m u e .H u as .oa.s n wx Amos umsasm .m> mama waauam "Acflaoumamzv maomaummEou mcommom cowsumm moo.o v. a .N n as .Hm.sa u Nx Ho.o v a .N u we .NH.HH n «x omuz Nmuz mmnz . smuz msnz mmuz masses .WMII . MMII MMII. Will .MMII will. ummaammaa as m m wN mm as swam Nwma amassm Nmsa waspam Amos umeasm «mos usesam mama wcauam lama smeasm mmaoammm HOHUGOU _ GHGOUMHQZ Awoaoom mmxmmv.ucmEummuH .co>Mum no cofiumasqoa may scum kumoqamwmav Acomwom some nowv moawco>sn kuwmuulaouucoo uo Icacouoaos nouns spas moaocosaoum mw.m «Hawk 30 SPRING '82 SUMMER'BZ n=24 45 n=59 .- n Meloionin Dc 1 I ..... 2222222222222 5444443$3310u 4 0 $>0.50 Error 109 66.94 Melatonin-treated individuals weighed more at disappearance than controls (F 9 20.648; df - 1/167; p < 0.001; Table 3.11, Figure 3.7). There was no significant effect of mother on the weights of disappearing mice (Table 3.12). 109 Table 3.11 Analysis of variance of weight at disappearance. Analysis of Variance Source of Variance DF MS F PROB Treatment 1 130. 788 20. 648 p < 0.001 Sex 1 22.544 3.559 NS Season X Treatment 2 2.024 0.320 NS Error 167 6.334 Total 171 Table 3.12 Preliminary mixed-effect analysis of variance of weight at disappearance. Analysis of Variance Source of Variance DF MS F PROB Treatment 1 14.82 2.56 0.113 Sex 1 12.68 2.19 0.142 Litter 1 2.26 0.39 0.532 Season 2 4.34 0.75 0.473 Treatment X Sex 1 3.18 0.55 0.461 Treatment X Season 2 10.71 1.85 0.162 Treatment X Litter 1 1.74 0.30 0.586 Sex X Litter 1 0.46 0.08 0.784 Sex X Season 2 0.29 0.05 0.948 Litter X Season 2 0.98 0.17 0.841 Treatment X Litter X Season 2 5.04 0.87 0.420 Treatment X Sex X Season 2 4.75 0.82 0.441 Treatment X Sex X Litter 1 0.06 0.01 0.907 Sex X Litter X Season 2 2.08 0.36 0.696 Sea X Trt X Lit X Sex 2 2.55 0.44 0.643 Mother Within Season 18 0.46 0.08 p:>0.50 Error 107 5.79 - a _ b 3 2 II SEED; O/A // W\\\\\\\\\\\\\\\\\\\\\= n=44 FEMALES 42 DISAPPEARANCE 35 n=50 MALES ..I O 111 As an alternative measure of dispersal-related movement I assessed the rate of inter-couplet movements made by mice that did not disappear (i.e. resided in the p0pulation at least 42 days) from the population. Resident, melatonin-treated mice were less likely to make inter-couplet movements (at least one) than resident controls (one-tailed X2 - 8.48; df = l; p < 0.005; Table 3.13a). Again, I Table 3.13a Numbers of melatonin- and control-treated stayers that made at least one inter-couplet movement. Treatment (Sexed, Seasons Pooled) Response Melatonin Control Move 20 17 No Move 52 12 Totals N=72 N=29 2 **x = 8.48, df = 1, p < 0.005 *Individuals who remained in population until at least the median age of occupancy for all treated mice (42 days). **One-tailed. tested further to determine the degree of independence of inter- couplet movement from the secondary criteria of interest - season and sex. The acceptance or rejection of the null hypotheses is based on two-tailed chi-square tests. Inter-couplet movement was independent of both season (X2 - 4.23; df - 2; p > 0.20; Table 3.13b) and sex (X2 - 0.49, df - 1; p > 0.60; Table 3.13c). Although control mice were more likely to make inter-couplet movements than. melatonin- treated individuals, this was only true for’ males (regardless of 112 Table 3.13b Frequencies with which stayers in each season made at least one inter-couplet movement. Season (Treatments, Sexed Pooled) Response Summer 1981 Spring 1982 Summer 1982 Move 7 9 21 No Move 20 21 23 Totals N=27 N=30 N=44 2 x = 4.23, df = 2, p > 0.40 Table 3.13c Frequencies with which male and female stayers made at least one inter-couplet movement. Sex (Treatments, Seasons Pooled) Response Male Female Move 17 20 No Move 29 35 Totals N=46 N=55 x2 = 0.49, df = 1, p > 0.60 season, Table 3.13d) and mice in the Summer 1982 season (regardless of sex, Table 3.13e). 113 Table 3.13d Frequencies with which male and female stayers (for each treatment) made at least one inter-couplet movement. Sex (Seasons Pooled) Male Female Response Melatonin Control Melatonin Control Move 8 9 12 8 No Move 24 5 28 7 Totals N=32 N=14 N=40 N=15 x2 = 6.46, df = 1, p < 0.05 x2 = 2.57, df = 1, p > 0.20 Table 3.13e Frequencies with which stayers in each season (for each treatment) made at least one inter-couplet movement. Season (Sexes Pooled) Summer 1981 Spring 1982 Summer 1982 Response Melatonin Control Melatonin Control Melatonin Control Move 5 2 7 2 8 13 No Move 14 6 20 1 18 5 Totals N=19 N=8 N=27 N=3 N=26 N=18 x2 - 0.005, df = 1, x2 = 1.97, df = 1, x2 = 7.33, df = 1, p > 0.60 p > 0.20 p < 0.02 Few juveniles in any season became trapped on either the perimeter or peripheral traplines (Figure 3.1). On the peripheral traplines (100 m away from the mouse cities), three juveniles were 114 trapped in the Spring 1982 compared with only one in each of the summer seasons. Most trap captures of juveniles occurred on the perimeter trapline immediately' surrounding (20 m. away) the mouse cities. On this trapline, three juveniles were trapped in the Spring 1982 compared with seven in the Summer 1982 and none in the Summer 1981. For all seasons combined, the trap-captured mice were a random subset of the trappable population with regard to their sex, reproductive condition and hormone treatment. Relationships Between Attainment of Puberty and Dispersal I found that mice were more likely than not to be pubertal at their last known residency in the population (one-tailed X2 - 18.36; df = 1; p < 0.001; Table 3.14a). Mice that disappeared were nearly Table 3.14s Frequency with which individuals were pubertal at disappearance. Responses Observed Expected Pubertal 117 88.5 Non-Pubertal 60 88.5 Totals 177 177 2 X = 18.36, df = 1, p < 0.001 two times more likely to be pubertal than non-pubertal. Furthermore, pubertal status was independent of treatment (X2 a 0.004; df I 1; p > 0.60; Table 3.14b), season (X2 - 4.01; df - 1; p > 0.05; Table 3.14c), and sex (X2 -_0.14; df = 1; p > 0.60; Table 3.14d). 115 Table 3.14b Frequencies with which melatonin- and control-treated juveniles were pubertal at disappearance (regardless of age). Treatment (Sexes, Persons Pooled) Response Melatonin Control Pubertal 57 ' 60 Non-Pubertal 29 31 Totals N=86 N=91 x2 =- 0.004, df =1, p > 0.60 Table 3.14c Frequencies with which individuals in each season were pubertal at disappearance (regardless of age). Seasons (Treatments, Sexes Pooled) Response Summer 1981 Spring 1982 Summer 1982 Pubertal 23 67 27 Non-Pubertal 11 27 22 Totals N=34 N=94 N=49 2 X 8 4.01, df = 1, p > 0.05 116 Table 3.14d Frequencies with which males and females were pubertal at disappearance (regardless of age). Sex (Treatments, Seasons Pooled) Response Male Female Pubertal 58 59 Non-Pubertal , 28 32 Totals N=86 N=91 x2 = 0.14, df = 1, p > 0.60 Regression analyses of age and weight at disappearance on age and weight at puberty were performed on melatonin-treated mice and controls in each season. I conducted separate regression analyses for each season because seasonal effects on age at disappearance, age at puberty, (as in season x treatment x litter order effect), and weight at puberty (as in season x sex effect) were discovered in earlier analyses of variance (above). Only mice which became pubertal prior to disappearance could be included in the analyses. To determine whether the responses of mice used in the regression analyses were representative of all other treated mice (including mice which disappeared prior to reaching puberty, and mice which failed to disappear after reaching puberty but prior to the completion of the experiment), I used t-tests to compare the ages and weights at puberty and disappearance for the different groups of mice. For instance, for each treatment-season combination, the ages and weights at puberty were compared for both disappearing and non-disappearing mice. 117 Similarly, the ages and weights at disappearance were compared for both pubertal and non-pubertal mice. Age at Disappearance vs Age at Puberty In all three seasons (regardless of treatment), mice which disappeared from the pOpulation did so very soon after (< eight days) reaching puberty (Table 3.15; see also Figures 3.8-10). The latency Table 3.15 Latency to disappear following attainment of puberty. Summer 1981 Spring 1982 Smer 1982 Treatment N X days i (8.8.) N X days I (3.3.) N X days : (S.E.) Melatonin 11 1.9 (1.4) 32 2.7 (0.8) 14 6.5 (2.0) Control 13 3.7 (1.5) 37 4.0 (1.5) 14 7.7 (2.6) Difference Between Treatments (t) 1.71 0.86 0.70 Probability 0.10>p>0.05 p > 0.30 p > 0.40 to disappear following puberty was not significantly different for melatonin-treated individuals vs controls. Latency to disappear was greater, however, in the Summer 1982 season compared to the other two seasons. The overall simple linear regression equations for age at disappearance vs age at puberty were significant for melatonin-treated mice in the Summer 1981 (r2 - 0.81; F - 36.71; df - 1/9; p << 0.001, Figure 3.8), Spring 1982 (r2 - 0.69, F - 67.20; df - 1/30; p << 0.001, Figure 3.9), and Summer 1982 (r2 - 0.74; F - 35.12; df . 1/12; p << 0.001, Figure 3.10). Similarly, for control-treated mice, the regression equations for age at disappearance vs age at puberty was significant for all three seasons (Summer 1981: r2 - 0.45; F - 8.82, 118 SUMMER 1981 74+ 1)- 8&* WP H 821' U.) >. c D m 581 L) 0 z 0 a 1P 0: '0 (I (9 DJ 50* 3: a: 1,. - 0') 0—0 0 J c’ 4+» u E ' l I t I. 3 m (I 38* HI I. umnm 0 MELATONIN - CONTROL 2 i i % Jr i 9 i ¢ 4 1L # J! ‘ : ‘f 20 28 32 3B 44 SO 58 82 88 RGE HT PUBERTY (DRYS) Figure 3.8 Relationship between age at disappearance and age at puberty for mice in the Summer 1981. The following regres- sion equations are for melatonin-treated and contrsl- treated mice respectively: Y - 0.9?X + 1.94 (n - 11, r - 0.81); Y - 0.70X + 13.81 (n - 13, r - 0.45). 119 SPRING 1982 741* II ar- 68*" 17- —~ 62; (.0 E o D 4» . m 661' U 2 c b n: a: U 501* l O. u D. u % Jr 0 0 0—0 G) c’ 4+; 0 4 o p— 0 (I +» 0' 0 Lu 0 a / m (I 30 0 MI II . m 0 MELATONIN - CONTROL ZGrf vL 3 i # ¢ 3 § 3 ‘r § 3 5 1L i ¢ 20 28 32 38 44 50 58 82 88 HGE RT PUBERTY (DHYS) Figure 3.9 Relationship between age at disappearance and age at puberty for mice in the Spring 1982. The following regres- sion equations are for melatonin-treated and control- tfeated mice, respectively: Y - 0. 8 + 10.16 (n - 32, - 0.69); Y - 0. 94X +,5. 72 (n - 37, r - 0.13). 120 SUMMER 1982 741* 88+ 621'- 561P 4a- RGE HT DISHPPEHRRNCE (DRYS) CI 4% 321* L I LEM 1/ I o MELATONIN <9 2- I CONTROL L L 1 a A g a a a g a a a L 2 . w . v 20 28 '32 '38 f '44 ' '60 58 82 88 ROE RT PUBERTY (DHYS) Figure 3.10 Relationship between age at disappearance and age at puberty for mice in the Summer 1982. The following regression equations are for melatonin-treated and contrel-treated mice, respectively: Y - 0.2?X + 7.65 (n - 14, r - 0.74); Y - 1.12X + 3.96 (n - 14, r - 0.30). 121 df = 1/11, p a 0.013, Figure 3.8; Spring 1982: r2 = 0.30; F - 5.32; df = 1/12; p = 0.040, Figure 3.10). I compared the regression equations of melatonin-treated mice and controls in each season and found no significant differences in their slopes (Summer 1981: t = 0.74; df = 20; p > 0.40; Spring 1982: t' = 0.27; df = 63; p > 0.50; and Summer 1982: t = 0.23; df - 24; p > 0.50). The assumption of equal variances about the regression equations necessary for this analysis was not violated. When I compared the age at disappearance for mice that did or did not become pubertal, significant differences were found only for control-treated mice in the Summer 1981 and Spring 1982 seasons. In both of these seasons, mice used in the regression analyses (i.e. those which became pubertal prior to disappearing) disappeared at a later age than those mice which were not included in the regression (Summer: t' - 2.90; df :18; p = 0.009; Spring 1982: t = 4.55; df 49; p < 0.001). The comparison of ages at puberty for mice that did or did not subsequently disappear revealed a significant difference only for melatonin-treated mice in the Summer 1981. In this season, mice used in the regression analysis (i.e. those which attained puberty and subsequently disappeared) reached puberty at a later age than‘those mice which were not included in the regression (t' = 3.24; df=12; p < 0.05). Weight at Disappearance vs Weight at Puberty The overall simple linear regression equations for’ weight at disappearance vs weight at puberty were significant for melatonin- 2 treated mice in the Summer 1981 (r a 0.70; F - 22.17; df - 1/9; p < 0.005, Figure 3.11) and Spring 1982 (r2 - 0.55; F - 35.81; df = 1/29; 122 SUMMER 1981 2 {b 2211- C) H 1P liJ L) D Z 8‘: m 184’ 11.1 0. 0.. C: (D H ‘1' D ’— C: .1 F— IM- C) (D u 5 I t—O U L; I 1- I II I 0 1L II LEQEMD 0 MELATONIN .. . CONTROL 1 6 T '10 T '14 ' ‘18 WEIGHT RT PUBERTY (G) Figure 3.11 Relationship between weight at disappearance and weight at puberty for mice in the Summer 1981. The following regres- sion equations are for melatonin-treated and contrel- treated mice, respectively: Y - 0284X.+ 2.36 (n - 11, r - 0.70; Y - 0.39X + 7.47 (n - 13, r - 0.21). 123 p << 0.001, Figure 3.12), but not Summer 1982 (r2 a 0.42; F - 3.58; df - 1/12; p - 0.083, Figure 3.13). For control-treated mice, the regression equation for weight at disappearance vs weight at puberty was significant in the Spring 1982 (r2 - 0.35; F - 8.87; df - 1/34; p << 0.012, Figure 3.12), but not Summer 1981 (r2 - 0.21; F - 2.92; df - 1/11; p - 0.115; Figure 3.13). I compared the regression equations of melatonin-treated mice and controls in each season and found no differences in their slapes for Summer 1981 (t - 1.10; df - 20; p > 0.40), Spring 1982 (t - 1.00; df ' 63; p > 0.40), or Summer 1982 (t = 0.55; df - 24: p > 0.50). When the weight at disappearance for mice that did or did not become pubertal was compared, significant differences were found only for melatonin- and control-tested mice in the Spring 1982. In this season, mice used in the regression analysis (i.e. those which became pubertal prior to disappearing) weighed more at disappearance than those mice which were not included in the regression (Controls: t' - 3.33; df=45; p < 0.005; Melatonin-treated mice: t = 2.34; df . 39; p = 0.024). I found no significant differences when I compared the weight at puberty for mice that did or did not subsequently disappear. EXPERIMENT 3.2: FALL DISPERSAL IN JUVENILES: EFFECTS OF STIMULATING SEXUAL BEHAVIOR The general hypothesis tested in this experiment is that dispersal is more likely to occur in juvenile 3;.2; bairdi that are sexually active than non-sexually active. Materials and Methods On the day of tagging, juveniles were temporarily removed from their nest-boxes and brought into the laboratory to receive an implant 124 SPRING 1982 25 1% HEIGHT RT DISFIPPERRRNCE (G) 101 22T- 1&- Figure 3.12 I o I I o o o m 0 ., m 0 II a El 0 m I m I / I I I u a) IJEEMD U I o MELATONIN I CONTROL T 110 I 11‘ V '18 HEIGHT RT PUBERTY (G) Relationship between weight at disappearance and weight at puberty for mice in the Spring 1982. The following regres- sion equations are for melatonin-treated and contrgl- treated mice, respectively: Y - 0.30X + 1.77 (n - 31, r - 0.55); Y - 1.37X - 3.11 (n - 36, r - 0.35). 125 SUMMER 1982 2Q fir- 224 o ... ,_ o LaJ u o 2 SE a: 18» UJ & a: o 2 .. ' D r- I. 0 a: o I— w» 9 (33 u a. IIII II ‘ U-I o 3 1r- ’0 l - II 1| 0 mm I O MELATONIN .. I CONTROL 8 ' '10 ' '14 "3 HEIGHT RT PUBERTY (G) Figure 3.13 Relationship between weight at disappearance and weight at puberty for mice in the Summer 1982. The following regres- sion equations are for melatonin-treated and contrgl- treated mice, respectively: Y - 0.31X + 5.21 (n - 14, r - 0.23); Y - 0.43X + 7.78 (n - 14; r - 0.42). 126 treatment. Juvenile males from each litter randomly received either a testosterone propionate (TP) or control implant. Females from each litter received either an estradiol benzoate (EB) or control implant. Following the procedures of Clemens and Pomerantz (1982), males were subcutaneously implanted with either an empty (control) 10 mm length of Silastic capsule (Dow Corning Corp., 1.47 mm i.d. x 1.95 mm o.d.) or one containing finely' ground ‘TP. Similarly, females received either an empty capsule or one filled with finely ground EB. These implant dosages are known to induce copulatory behavior among male EE_ 1‘; bairdi and receptivity among female &_ m_. bairdi (L. Clemens, S. Pomerantz; pers. comm.). Following implantation (within four hours), mice were returned to their appropriate nest-boxes. (fix all subsequent nest-box: checks, implant-treated mice were checked by palpation to confirm the presence of the implant. One individual lost its implant and was therefore discarded from any of the analyses. Analysis The following three hypotheses concern the effect of gonadal steroid treatment on puberty. Although it was predicted that steroid treatments would stimulate sexual activity, gonadal development should not be affected. 1. The probability of ultimate attainment of puberty by juveniles in the population is independent of treatment (TP, EB, or control). 2. Age at puberty is independent of treatment. 3. Weight at puberty is independent of treatment. 127 The remaining hypotheses pertain to the general prediction that dispersal will be more likely to occur, and at an earlier age, among steroid-treated individuals than controls. Dispersal here is defined as it is in Experiment 3.1. 1. Steroid-treated individuals are more likely to disappear from the pOpulation than control-treated individuals. 2. Steroid-treated individuals will disappear from the population at an earlier age than control-treated individuals. 3. Steroid-treated individuals will weigh less at disappearance than will control-treated individuals. 4. Resident, steroid-treated individuals are more likely to make inter-couplet movements than resident, control-treated individuals. 5. Steroid-treated mice are more likely to be captured in peripheral or perimeter traplines than control-treated mice. I also predicted that disappearing mice are more likely to be pubertal than non-pubertal. Furthermore, I predicted that a positive relationship exists between age at puberty and age at disappearance. I also anticipated that the weight of individuals at puberty will be positively associated with weight of individuals at disappearance. Various parametric and non-parametric tests were used to analyze the data: Chi-square tests of idependence (X2), t-tests (t, t' ), analysis of variance (one-way), simple linear regression (r2), and Mann-Whitney U test of independence. T-tests, one-way ANOVAS, and linear regression were conducted with the aid of SPSS programs (T-TEST, ONE-WAY, REGRESSION, and PLOT). Mann-Whitney U tests were 128 used when the more strict assumptions of the parametric tests were violated. Results (hm October 30, November 3, and 4, I recorded five weasel attacks on families in the mouse cities. In these five predations by two known M; nivalis, at least ten adults and 20 juveniles were known to have been killed. In most of the attacks the adult females were nursing young; thus, four litters were either killed altogether or reduced in size. To eliminate the possibility of including lost individuals into the category of disappeared instead of killed, I chose to eliminate from the analyses all juveniles that survived the attempted predation. Two weasels were caught; one was in the act of killing nestlings when it was discovered and removed; the other was trapped in a live-trap adjacent to one of the arenas. The four affected families were replaced (except for any survivors) with five new' breeding pairs from the laboratory ‘mouse colony' (see General Materials and Methods above). Effects of Treatments on Attainment of Puberty The null hypothesis that ultimate attainment of puberty is independent of hormone treatment was rejected for males (TP vs Control: x2 - 12.08, df = 1, p < 0.002, Table 3.16), but accepted for females (EB vs Control: 2:2 = 0.90, df - 1, p > 0.60, Table 3.16). Significantly more TP-treated males (in the population at least 21 days) became pubertal than control-treated males. In fact, none of the control males attained puberty during the course of the 129 Table 3.16 Frequencies with which males and females (for each treatment) became pubertal or remained non-pubertal. Sex Male Female Response TP Control EB Control Pubertal 10 0 7 4 Non-Pubertal 1 5 1 2 Totals N=11 N=5 N=8 N=6 2 2 X = 12.08, df = 1, X 8 0.90, df 8 1, p < 0.002 p > 0.60 experiment. Females in the EB treatment group were no more likely to become pubertal than females in the control group. Due to the small sample sizes and unequal variances, I used Mann-Whitney U tests to compare the age at puberty of controls and steroid-treated mice. Since no control males attained puberty, no comparisons with TP-treated males was possible (see Figure 3.14). For females, where a comparison of controls and EB-treated females was possible, I found no significant different in the age at puberty (U - 7, p - 0.115; Figure 3.14), although EB-treated females were slightly younger compared to controls. Although the age at puberty for EB-treated females was not significantly less than the age at puberty for controls, their weights were significantly different (U - 4, p a 0.036, Figure 3.15). EB-treated females weighed less at puberty than controls. Again, no W//////////mm IMALES PUBERTY at puberty. Vertical lines indicate 1 1 S.E. 3:109) L A M E F LES * PUBERTY eight at puberty. Vertical lines indicate 1 1 S.E. m m e r m. .1 F 132 comparisons of weight at puberty for TP-treated males vs controls was possible (see Figure 3.15). No females from either treatment group became pregnant during the duration of the experiment. Effects of Treatment on Dispersal Following the procedures described above in Experiment 3.1 for distinguishing stayers from mice which disappear, I tested the hypothesis that steroid-treated mice were more likely to disappear than controls. For males and females, occupancy (stay vs disappear) was found to be independent of hormone treatment (Males: one-tailed X2 - 0.90, df - 1, p > 0.30; Females: one-tailed X2 - 0.57, df - 1, p > 0.30; Table 3.17). Table 3.17 Frequences with which males and females (for each treatment) disappeared from the population or stayed. Stayers are individuals which remained in the population until at least the median age of occupancy for all treated mice (33 days). Sex Male Female Response TP Control EB Control Stay* 4 5 5 3 Disappear 6 3 3 4 Totals N=10 N=8 N=8 N=7 **x2 = 0.90, df = 1, **x2 - 0.57, df = 1, p > 0.30 p > 0.30 *Individuals that remained in population until at least the median age of occupancy for all treated mice (33 days) **One-tailed. 133 An unexpected result was the finding that control males disappeared from the pOpulation at an earlier age than did TP-treated males (U - 1, p = 0.024, Figure 3.16). Control females, on the other hand, did not differ from EB-treated females with regard to the age of the their disappearance (U = 5, p = 0.429; Figure 3.16). Furthermore, the weights at disappearance did not differ for steroid-treated mice vs controls for either males or females (Males: U = 8, p = 0.452; Females: U = 5, p = 0.429, Figure 3.17). As an alternative measure of dispersal-related movement, I assessed the rate of inter-couplet movement for mice that did not disappear (in this case, mice which resided in the population for at least 33 days, the median length of occupancy). Due to the small sample of responses for mice from either the TP or EB treatment groups, all steroid-treated mice were pooled for the analysis. Inter-couplet movement was independent of treatment (one-tailed X2 = 0.02, df . 1, p > 0.30, Tab.e 3.18). Very few resident mice from either treatment group made inter-couplet movements. The live-trapping data from this experiment revealed a pattern similar to that found in Experiment 3.1; that is, few juveniles were ever found in traps on either the peripheral or perimeter traplines. Two mice were captured on the peripheral trapline and one was caught on the perimeter trapline. TwO (one male, one female) of the three captured mice were pubertal at the time of capture. 404 38 " Steroid 36- :JCOnTrOl 34F ' 32‘ A30 - >- < 28 , 326 +- Lu 0 24 n=6 n=3 ' n=3 n=4 MALES FEMALES DISAPPEARANCE T .4 . E .. . o wwwwww 3:132, 136 Table 3.18 Frequencies with which steroid- (EB and TP pooled for males and females) and control-treated stayers made at least one inter-couplet movement. Treatment (Sexes Pooled) RespOnse' . Steroid (EB, TP) Control Move 2 2 No MOve 7 6 Totals N=9 N=8 2 **x = 0.02, df = 1, p > 0.30 *Individuals that remained in population until at least the median age of occupancy for all mice (33 days). **One-tailed. Relationships Between Attainment of Puberty and Dispersal Mice which disappeared from the population were equally likely to be pubertal as non-pubertal (one-tailed X2 - 1.0, df - 1, p) 0.10; Table 3.19). Further examination of these data revealed, however, that TP-treated males were more likely to be pubertal at disappearance than control males (X2 - 5.65; df - 1, p < 0.05; Table 3.20). Steroid-treated females, on the other hand, did not differ from control females (X2 - 0.06, df - 1, p > 0.60, Table 3.20). Repression. analyses of age and weight at disappearance on age and weight at puberty were attempted on TP-treated males, EB-treated' females and controls (sexes combined). The rationale behind utilizing only those animals which became pubertal prior to their disappearance is explained above in Experiment 3.1. The regresSion of weight at disappearance on weight at puberty was not possible for control males 137 Table 3.19 Frequency with which individuals were pubertal at disappearance. Responses Observed Expected Pubertal 10 8 Non—Pubertal 6 ,. 8 Totals N;T6 N:T6 x2 = 1.00, df = 1, p > 0.10 Table 3.20 Frequencies with which males and females (for each treatment) were pubertal at disappearance. Sex Male Female Response TP Control EB Control Pubertal 5 0 2 3 Non-Pubertal 1 3 1 1 Totals N=6 N=3 N=3 N=4 x2 a 5.65, df = 1, x2 . 0.06, df = 1, p < 0.05 p > 0.60 because all of these animals weighed the same at puberty (the independent variable). Age at Disappearance vs Age at Puberty Regardless of treatment (steroid or control), mice which disappeared from the population did so very soon (x - 3 days; Table 138 3.21, see also Figures 3.18-19) after showing signs of puberty. The latency to disappear from the population following the attainment of puberty'was not significantly different for steroid-treted mice vs controls. Table 3.21 Latency to disappear following attainment of puberty among mice that became pubertal before disappearing. Treatment N 5.: days : (S.E.) TP 5 1.2 (1.20) EB 3 4.0 (0.87) Control 3 5.0 (2.00) The simple linear regression equations for age at disappearance vs age at puberty were not significant for the steroid-treated mice (TP-males; r2 = 0.56; F - 3.76; df - 1/1; p - 0.148; EB-females: r2 - 0.84; F - 5.33; df - 1/1; p - 0.260). The slope of the regression line for each steroid treatment group was not significantly different from the slope of the control group regression line. Weight at Disappearance vs Weight at Puberty The overall simple linear regression equation for weight at disappearance vs weight at puberty was significant for TP-treated males (r2 8 0.88; F - 21.21; df - 1/3; p - 0.019), but not for EB-treated females (r2 - 0.89, F - 8.33; df - 1/1; p - 0.212). As explained above, a regression analysis was not possible for the control group. 139 FHLL 1981 74* 681* 1b 1" 624. 00 >. C: ., c: m 5611’ L) 1: cr 1* c: c: uJ 501 a. n. c: v a: D 441* p— c: 1* Lu c: c: as: 11- 32* mm 4» . OCONTROL ||TP 28v +EB 4,. 2- ' ¢_, ¢ :4 ¢ :7 c : .1 :5 r : .11 t : 4 20 28 32 38 44 60 88 82 88 ROE HT PUBERTY (DHYS) Figure 3.18 Relationship between age at disappearance and age at puberty for mice in the Fall 1981. The following regression equations are for control-treated, TP-treated, and EP-treated mice, respectively. Y I 4.0 - 64.00 (n I 3, r I 0.84); Y I 1.30X - 6.60 (n I 5, r I 0.56); I I 1.81X - 15.11 (n I 3, r I 0.98). 140 FRLL 1981 2 221* CD [1.1 U 2 E c 18'" UJ L L ¢ (.0 H 1’ D .— C: P— lh* I (D H U 3 «1- 10'1- 1r- 1 Figure 3.19 10 14 LEDEMP 0 CONTROL .TP +EB 4 V HEIGHT RT PUBERTY (G) db 18 Relationship between weight at disappearance and weight at puberty for mice in the Fall 1981. The following regres- sion equations are for TP-treated and EB-treated mice, respectively. A regression analysis was not possible for .32X - 2.52 (n I 5, - 0089)e cpntrol-treated mice (see text): Y I r I 0.82); Y I 1.07X + 1.11 (n I 3, r 141 Discussion This study aimed to investigate the role of sexual maturation in the postnatal dispersal of juvenile P; & bairdi. Although many authors have noted that dispersers are commonly young, pubertal individuals, few have argued that the postnatal dispersal of juveniles may be causally related to their sexual maturation. Instead, many have implied that dispersal is a density dependent phenomenon and that individuals, particularly the very young or very old, are forced to disperse from an area in response to limitations in such critical resources as food and living space (Christian, 1970; Krebs, 1978a; Lidicker, 1975). King (1983) questioned the importance of density in controlling dispersal and found, instead, by controlling several density related variables, that dispersal in juveniles was associated primarily with seasonal patterns in their sexual recruitment to the population. He proposed that juveniles of either sex disperse from their natal site in search of mates. Furthermore, it was proposed that this sexual search behavior by juveniles commences with the onset of their sexual maturity, a time when pubertal mice are attracted to relevant sexual stimuli in their environment. In partially controlled experiments with semi-natural populations °f.£; E; bairdi, I tested predictions of the sexual search hypothesis by manipulating the expression of sexual maturity in juveniles. Juvenile dispersal (i.e. disappearance) is characteristically high in the spring and summer, seasons in which juveniles are known to be sexually recruited into the breeding population (King, 1983). In the first experiment I attempted to delay the dispersal of spring and summer-born juveniles by’ delaying their sexual maturation through 142 exogenous treatment with the antigonadal hormone, melatonin. In the fall of the year, however, juveniles typically fail to breed and are less likely to disperse. Thus, in the second experiment I attempted to stimulate dispersal in [juveniles via their treatment ‘with. the gonadal steroids, testosterone propionate and estradiol benzoate. Critical to both of the experiments described here was the assumption that gonadal steroids can facilitate the expression of mate-searching behavior. For instance, in the first experiment, melatonin was expected to indirectly delay sexual searching in mice by retarding their gonadal maturation and subsequent production of gonadal steroids. In the second experiment I expected that direct exogenous treatment of juveniles with gonadal steroids would stimulate sexual search behavior. Turning to the literature on hormones and behavior for support, one finds that the expression of sexual behavior in P;_m:_bairdi and other species is clearly dependent on the presence of gonadal steroid hormones (Pomerantz g£_sl;, 1983; see also reviews by Larsson, 1979; Morali and Beyer, 1979). As evidence for this relationship, one finds that in prepubertal mice and in adult mice that have been castrated, exogenous treatment with gonadal steroids can either initiate or restore normal sexual behavior (Larsson, 1979). In most studies though, gonadal steroids are known to exert control of the copulatory elements of normal sexual behavior, i.e. lordosis in females, and mounting, intromission, and ejaculation in males. Little is known, however, about how other, pre-copulatory behaviors (e.g. mate-seeking behavior) are affected by these hormones. One recent study, however, has shown that gonadal steroids can activate courtship behavior as well as copulatory behavior in male 3; m; bairdi 143 (Pomerantz, it}. &, 1983). Gonadal steroids can also stimulate locomotory and exploratory behaviors, which are surely associated with mate-seeking activity (Blizard, 1983; Holekamp, 1982; Mather, 1981, Moore, 1965). For example, implants of testosterone in males, and estradiol in females, are known to increase certain components of locomotor activity in hamsters (Ellis and Turek, 1979; Morin, £31., 1977). Holekamp (1982) suggested that gonadal steroids can stimulate natal dispersal in ground squirrels through their mediating effects on juveniles' exploratory behavior. If sexually motivated mice are, indeed, predisposed to exhibit increased locomotory and exploratory behaviors as a result of the influence of gonadal steroids, one should expect these mice to move from a home nest at puberty in search of a mate. Evidence from an earlier field study supports this contention. Howard (1949) found it unlikely that juvenile & 1': bairdi breed within close vicinity of their natal nest. He demonstrated that more than half of the young mice he studied moved away from their natal home range before breeding. If, as King (1983) and I have proposed, pubertal juveniles disperse in search of potential mates, one should see sexual search behavior manifested in measurable movements away from their area of birth. In the present study, I selected inter-couplet movement as one of three possible indicators of sexual search behavior. During both experiments, at least half of the couplets at the mouse cities were occupied by other mice at any given census. It stands to reason that if sexually mature mice were dispersing in search of mates, the search may very well extend to other couplets. In the first experiment, mice 144 which were treated with melatonin to delay sexual maturation should, according to the sexual search hypothesis, be less likely to make inter-couplet movements than controls. The results indicated that, of all the mice that resided in the nest-boxes for at least 42 days (median age of residency), melatonin-treated mice were, indeed, unlikely to move from their natal couplet (Table 3.13a). Although there was no direct dependence of these movements on sex or season, closer analysis revealed that, when controlled for season, control males were more likely to move from their natal couplet than melatonin-treated males. The contingency test ( Table 3.13d) suggests that melatonin treatment is probably not affecting the males' tendency to move differently than it affects females because, in both sexes, melatonin-treated individuals were at least two times more likely to remain in their natal couplet than move. Control mice of either sex, (n1 the other hand, were more likely to move from the natal couplet than not move. When sex was controlled, a significant treatment effect was found only in the summer 1982 season. Again, the factor most likely to be responsible for this significant effect was the tendency of control mice to be more inclined to move from their natal couplet. In the second experiment, steroid-treated mice were just as likely to move from their natal couplet as controls (Table 3.18). The proportion of all mice that moved was lower in the fall experiment than in the spring and summer. In spite of the small sample sizes, this apparent low rate of movement by mice born in the fall seemed to be fairly typical of fall-born mice from previous years (King, 1983). 145 Evidence of behavior related to sexual searching need not be restricted to inter-couplet ‘movements. Mate seeking behavior is likely to involve an exploration of the individual's surroundings in general. Any widespread exploration of the environment by an individual is certain to increase its chances of encountering a potential mate (Mather, 1981). Unfortunately, nest-box censuses only revealed where mice rested during the day, and these checks were made only every third day. Nocturnal movements could be made by mice that appeared, from the census of nest-boxes, to be ensconced in one arena or couplet. I hoped, therefore, that by monitoring traplines that surrounded the mouse cities I would be able to identify mice that were moving away during the night from their natal couplet and into the surrounding fields. If the sexual search hypothesis is correct, one should expect to find that sexually mature mice, or mice that were treated with gonadal steroids, would be more likely to explore the surrounding fields and become trapped than sexually immature mice. Unfortunately, too few juveniles were trapped in any season to allow a reliable interpretation of the data. Of those mice that were trapped, neither sex, hormone treatment, nor reproductive condition were distinguishing characteristics. It was surprising, in light of the relatively large juvenile component to the population in most seasons, that so few juveniles were caught in the traps. Nevertheless, this low capture rate corresponds well with the trapping results of an earlier study at the mouse cities (King, 1983). The results of the present study and that of King (1983) show that most mice either remained in the mouse cities and occupied nest-boxes, or disappeared from the population, although some mice 146 only occupied nests periodically. For purposes of this discussion, disappearance is defined as a permanent, discontinued daytime occupation by a mouse of any nest-box. In the first experiment, half of all tagged juveniles disappeared from the population in each season by the time they reached 42 days of age. Of those mice which disappeared by this age 71% (82/115) were controls (Table. 3.8a). In all three seasons, controls were more likely to disappear than melatonin-treated mice. In the fall experiment, steroid-treated mice of either sex were just as likely to disappear as controls of either sex (Table 3.17). In the fall, nearly one out of every two mice disappeared prior to reaching 33 days of age. It is interesting to compare these rates of disappearance with those from earlier studied populations at the mouse cities. In his study of one summer and two spring seasons, King (1983) reported that 75.4% (129/171) of all tagged juveniles (from first and second litters only) disappeared from the pOpulation by the end of each ten week experiment. This value is remarkably similar to that obtained for control mice in the first experiment of the present study, where 73.9% (82/111) disappeared prior to reaching 42 days of age (Table 3.8a). Thus the high rate of loss for controls in the present study appears to be fairly typical of other spring and summer populations. Compared with this high rate of juvenile disappearance, the loss among melatonin-treated mice was significantly lower (Table 3.8a: 31.42 - 33/115). This value for the melatonin-treatment group is more typical of mice in the fall, a season when juveniles are more likely to be sedentary. In two previous fall experiments, King (1983) reported that, on the average, 33.9% (40/118) of the juveniles were lost from 147 the population. Compared with this value, mice in the present fall experiment were more likely to disappear, but this rate of disappearance was still below the loss rate recorded for control mice in the spring and summer seasons. The value for disappearance described above for melatonin-treated mice is somewhat misleading, however, because it is only a measure of those mice which disappeared prior to reaching 42 days of age. Many more melatonin-treated individuals actually disappeared, but did so sfggg 42 days of age. Mice that were treated with this hormone disappeared, on the average, 9 days later than controls (Figure 3.6). In fact, this difference in age at disappearance can be seen even in comparisons with untreated mice from 4 previous spring and summer seasons at the mouse cities (from King, 1983; Table 3.22). The ages Table 3.22 Age at disappearance for melatonin-treated mice and untreated mice from 6 previous seasons. Age at Disappearance (Days) Season x (t S.E.) Summer 1979 30.11 (0.84)a Spring 1979 30.44 (1.68)ab Spring 1980 33.41 (1.64)abc Summer 1980 36.10 (1.04)bd Spring 1982* 38.84 (1.35)Cde Fall 1979 39.27 (2.80)abef Fall 1978 39.59 (3.09)abef Summer 1981* 44.80 (1.87)df Summer 1982* 47.79 (2.83)f *Experimental seasons (all others are baseline seasons; see text for explanation). Means sharing a common letter are not significantly different. 148 at disappearance for melatonin-treated mice were consistently higher in all seasons except for 2 previous fall seasons. Table 3.23 indicates that the ages at disappearance for controls and untreated mice from the previous seasons were not significantly different. In addition to a treatment effect on disappearance, there was also a seasonal effect. The proportion of disappearing mice was highest in the spring season (Table 3.8b), and spring mice disappeared at an earlier age than summer mice (Figure 3.4). Table 3.23. Age at disappearance for control mice and untreated mice from 6 previous seasons. Age at Disappearance (Days) Season x (i S.E.) Summer 1979 30.11 (0.84)8 Spring 1979 30.44 (1.68)ab Spring 1982* 31.10 (1.26)ab Spring 1980 33.41 (1.64)ab Summer 1981* 33.85 (1.87)ab Summer 1980 36.10 (1.04)b Summer 1982* 36.96 (2.83)ab Fall 1979 39.27 (2.80)ab Fall 1978 39.59 (3.09)ab *Experimental seasons (all others are baseline seasons; see text for explanation). Means sharing a common letter are not significantly different. At this point in the discussion I would like to address the question of why control mice in the first experiment were more likely to disappear, and disappear earlier, than melatonin-treated mice. This question might be easier to answer if I knew whether these two groups of mice disappeared for the same reason. Unfortunately, it is impossible to know for certain why any mouse would disappear. We can 149 be certain, however, that mice which disappeared from a nest-box either left the nest-box permanently, or died there. The results of this experiment suggest that nest mortality does not explain the observed differences in the rates and ages of disappearance. There is little indication to suggest that control individuals were more likely to die at the nest, or die earlier there, than melatonin-treated mice. With the exception of the deaths of some young nestlings due to abandonment by their mother, few mice were found dead in the nest-boxes. In the three seasons of this experiment only 4 dead juveniles were recovered from the nests. The cause of death was determined for only one of these mice, a control juvenile that was killed and partially eaten in the spring 1982 by a weasel, presumably M; nivalis. Incidentally, in the same season, one entire litter of preweanling mice was found dead in a nest-box. Of each mouse killed by a weasel there was evidence of the characteristic puncture wound at the base of the skull. In order for nest predation to explain the earlier disappearance of control mice, predators would be expected to prey selectively on control mice that were younger than melatonin-treated individuals. Although it is possible that more mice could have been preyed upon at the nest, particularly by weasels, it is unlikely that a predator would behave in the manner alluded to above. For instance, in weasel attacks, the weasel typically kills most or all of the occupants of a nest-box or even an arena (Murphy, pers. obs.). Because occupied arenas were typically inhabited by relatively equal proportions of controls and melatonin-treated mice (usually littermates), weasel predation would not explain why control 150 mice occupants disappeared at an earlier age than their melatonin-treated counterparts. Other predators, too, were unlikely to be the cause of the differential disappearance among treatment groups. Although the arenas and concrete nest-boxes were accessible to the agile, small-bodied weasels, they probably afforded the mice considerable protection from other potential predators, particularly snakes, shrews, raccoons, and owls. Snakes were sometimes common in nest-boxes that lie outside of the arenas (O nest, Figure 3.2). The arena wall was effective in preventing most snakes from gaining entry to the arenas although they were sometimes found in nest-boxes inside of the arena. Another potential predator that was rarely seen inside of an arena was the short-tailed shrew. Shrews and snakes were both probably excluded from the arenas due to the elevated entrance hole on the arena wall. Raccoons, on the other hand, were sometimes persistent in gaining entry to the arenas by climbing over the arena wall. This problem was severest in the summer 1982 during which several raccoons developed a theretofore rarely observed ability to Open the lids of several nest-boxes. During the course of approximately one week in this season, raccoons entered several arenas formerly occupied by tagged mice and removed the lids to all of the nest-boxes inside. I suspected that in a few of these arenas, the raccoons may have caught and killed the occupants. The removal of the troublesome raccoons through. intensive live-trapping, and the deve10pment of a nest-box security lid, prevented any further nest disturbances by raccoons in that season. The arguments presented above suggest that predation at the nest can not adequately explain 151 the observed patterns of disappearance in the present study. Therefore, in all likelihood, mice that disappeared probably moved permanently away from the nest-boxes. If this is true, why, then, did control mice leave their nest at an earlier age than melatonin-treated individuals? It is interesting to note here that melatonin treatment alone could affect the activity of treated mice. In experiments with rats, melatonin treatment can depress wheel-running activity (WOng and Whiteside, 1968). Chronic administration of melatonin was also effective in suppressing locomotor activity in house sparrows (Hendel and Turek, 1978). When the melatonin implant was removed from a bird, locomotor activity was restored to pre-treatment levels. Whether melatonin in the present study delayed the departure of treated mice from their nest because of its antigonadal effects or due to the suppression of locomotor activity (or both), is unknown. Although I have argued that the movement of mice from their natal area results from their initiation of mate-searching behavior, the possibility remains that these movements represent a motivated search for some other resource that is not available in the vicinity of the natal nest. Some potentially limited resources that could influence these movements include food, water, refugia and/or nesting sites. If, for the sake of argument, control mice were assumed less tolerant of a resource limitation than melatonin-treated individuals, and more likely to move in search of it, then a shortage of a critical resource at the natal nest could explain their earlier disappearance. For example, perhaps controls and melatonin-treated mice differed in their nutritional requirements. The results show that controls were, 152 indeed, lighter in weight than melatonin-treated mice when they disappeared. In spite of ample food supplies at the nests, one could argue that controls were forced to leave the nest to locate some deficient element to their diet. This argument can be disputed, however, on two grounds. Firstly, although controls weighed less at disappearance (Figure 3.7), the observed weight difference was more likely to be due to the chronological differences in age at disappearance between controls and melatonin-treated mice than to a deficiency in growth in the former. Secondly, there is laboratory evidence which indicates that melatonin-treatment does not significantly affect body weight in_§; leucOpus (Johnston and Zucker, 1980a; Lynch st 51;, 1978; see also Results, Chapter 2). Thus, it is unlikely that nutrition contributed much to the observed treatment differences in disappearance. The same can probably be said of water. During dry periods I was careful to provide extra water at each arena. Although controls were more likely to move from their natal couplet or disappear than melatonin-treated mice, they probably did not do so for a lack of available nests or refugia. There were always surplus nest-boxes available throughout the duration of each experiment. For instance, over half of the available arenas (x I 18.03), and nearly one-third of the available couplets (x I 5.31), were unoccupied at any given census. Nearly 4 arenas (x I 3.67) were 23135 occupied at all during each ten week experiment. King (1983) reported similar values for unoccupied arenas (x I 19.2; x I 5.0). It is also unlikely that parasitism or disease could explain any treatment effect on disappearance. One could argue, though, that controls may be more inclined to leave the nest-boxes than 153 melatonin-treated mice because the former were more susceptible to parasitism or disease at the nest. For instance, many common mouse parasites (e.g. fleas and mites) are sometimes more common in a nest than on a mouse (Whitaker, 1968). Perhaps, then, controls leave the nest to avoid infestation. This is unlikely, however, because almost all mice in the present study, regardless of treatment, already had some parasites (fleas) before they disappeared. Of the few observed incidences of parasitism by botfly larvae (Cuterebra sp.), just as many control mice were afflicted as melatonin-treated individuals (5 cases each). To summarize the arguments presented so far, it appears that nestbox-related mortality or parasitism cannot adequately explain why melatonin-treated mice disappeared from their natal couplet at an older age than controls. It also seems improbable that melatonin-treated individuals and controls would respond differently to potential deficiencies in food or nesting sites. Furthermore, because predation, food supply, and nesting sites were at least partially controlled in each season, it is doubtful that these factors could explain the seasonal effects on disappearance. Why, then, did spring mice disappear at an earlier age than summer mice, and why did melatonin-treated mice disappear at an earlier age than controls? An important clue to an understanding of the observed treatment and seasonal effects may be found in the association between disappearance and the attainment of puberty. Irrespective of treatment, the majority of mice which disappeared did so very soon (2-8 days) after reaching puberty (Table 3.15). Indeed, regression analyses of the mice that disappeared following puberty revealed that 154 the age at puberty is a good predictor of age at disappearance. This relationship held for controls as well as melatonin-treated mice, regardless of season (Figures 3.8-10). This relationship serves to explain why melatonin-treated individuals disappeared at a later age than controls. Disappearance and puberty were also dependent on season. For example, the prOportions of mice that reached puberty and disappeared were both highest in the spring (Tables 3.2c, 3.8b). Furthermore, spring mice reached puberty at an earlier age than summer mice (Figure 3.4). One possible explanation for the observed difference in the age at puberty and in the proportion of pubertal mice could be the concomitant seasonal change in the photoperiod. Photoperiods changed substantially from one season to the nest. Daylength increased from about 13.5 to 15.3 (1.8) hours in the spring experiment and decreased from 15.1 to 12.6 (2.5) hours in the summer and from 12.2 to 9.5 (2.7) hours in the fall. Although the total length of daylight available to the mice in the summer and spring seasons was comparable, the direction of change in the daylength differed. Perhaps it is this decline in daylength, independent of treatment effect, which is responsible for summer mice reaching puberty in fewer numbers and at an older age than spring mice. Seasonal changes in photoperiod are known to exert a strong influence on the reproductive development in ‘2; maniculatus and other temperate-climate small mammals (e.g. Johnston and Zucker, 1980b,c; Petterborg and Reiter, 1980; Whitsett gt al., 1983). 155 Another notable difference between the spring and summer seasons is the lower density of mice found in the two summers. Compared to the large spring season population, there were 381 fewer treated juveniles (at least 30 days old) present in the summer 1982 population and 892 fewer in the summer 1981 population (Table 3.2c). These differences can be attributed primarily to the variation in the size of litters born to the introduced adult females. It is not intuitively clear if the difference in densities could explain the observed pattern. of pubertal. deve10pment .and. disappearance in the different seasons. Because the spring season was not replicated, the possibility remains that the seasonal difference is due to a sampling error. Nevertheless, it appears that the seasonal effect on puberty probably accounts for the observation that spring mice disappeared at an earlier age and in greater proportions than summer-born mice. This close association between age at disappearance and ontogenetic maturation of the gonads lends credence to the argument that gonadal hormones may be important in activating mate-seeking activity in mice. If gonadal steroids are important in this regard, then mice that are treated with these hormones should be more inclined to depart their natal area, and leave at an earlier age, than mice which received no treatment. The results of the second experiment revealed, however, that steroid-treated mice were just as likely to disappear as controls (Table 3.17). The expected treatment effect on age at disappearance was also not realized. In fact, control males disappeared at a younger age than steroid-treated males (Figure 3.16). In spite of the small sample sizes, I will offer several possible explanations for these results. Firstly, the gonadal steroids chosen 156 for the exogenous treatments simply may not have stimulated mate-seeking behaviors, though they may be involved in activating other sexual behaviors. Perhaps other gonadal hormones, or even different metabolites of the ones used in the present study (TP, EB) may be more important in facilitating the search component of sexual behavior. Indeed, other investigators have shown that the degree to which sexual behavior in P; m; bairdi is affected by gonadal steroids is influenced by the particular sexual behavior under consideration (Pomerantz, g£_§l;, 1983). Another reason why gonadal steroids were possibly ineffective could be that they were administered to mice at an age at which behavioral development was relatively insensitive to the hormones. In a recent unpublished study, Holekamp (1982) tested the hypothesis that concurrent, circulating gonadal steroids cause natal dispersal in juvenile Belding's ground squirrel (Spermophilus beldingi). Her data contradicted this prediction and indicated, instead, that Erinatal exposure to androgens (in males) is required to activate dispersal later on. Although perinatal exposure to gonadal steroids may be important for the dispersal of juveniles in the present study, I would argue that circulating gonadal hormones associated with puberty are also critical for the expression of mate-seeking dispersal. I believe this is true because in mice treated with melatonin, puberty and dispersal were both delayed. Finally, I wish to consider the possibility that by treating mice in the fall with gonadal steroids I was actually able to stimulate natal dispersal in not only the steroid-treated mice but also in the 157 controls. I consider this as a possibility because of the following observation. The results of this study indicated that the ages at disappearance for controls as well as steroid-treated mice were significantly younger than that of untreated mice from 2 previous fall seasons (King, 1983). In fact, mice in the present fall experiment disappeared, on the average, at an earlier age than mice from m previously studied season at the mouse cities (Tables 3.24-26). I offer the following explanation for why controls and treated mice both disappeared at such an early age. Following their treatment with gonadal steroids, mice may increase their exploratory activity and soon move from their natal couplet. At this time in the year, the breeding season is nearing its end and fewer individuals, particularly juveniles, will be in breeding condition. Thus, sexually stimulated juveniles may be expected to leave their natal area in search of the remaining individuals in breeding condition (see Nadeau 35 31;, 1981). Control individuals, in turn, may then disperse due to their perception of other active individuals in the population. Unfortunately, any of the above scenarios could serve to explain the disappointing results of the steroid experiment. Clearly, more research is required to determine the roles that sexual maturation and gonadal hormones have in mediating postnatal dispersal. Although I have demonstrated in the first experiment that a delay in sexual maturation in juveniles can result in a delay in their natal dispersal, I can only assume that gonadal hormones may be involved. These field experiments should be repeated, but an attempt should be made to determine if a causal relationship exists between particular gonadal steroids and dispersal. One way of accomplishing this would 158 Table 3.24 Age at disappearance for testosterone-treated mice and untreated mice from 6 previous seasons. Age at Disappearance (Days) Season x (i S.E.) Fall 1981* 26.75 (1.26)b Summer 1979 30.11 (0.84)bc Spring 1979 30.44 (1.68)abc Spring 1980 33.41 (1.64)abc Summer 1980 36.10 (1.04)a Fall 1979 39.27 (2.80)ac Fall 1978 39.59 (3.09)ac *Experimental season (all others are baseline seasons: see text for explanation). different. Meaning sharing a common letter are not significantly Table 3.25 Age at disappearance for estradiol-treated mice and untreated mice from 6 previous seasons. Age at Disappearance (Days) Season x (i 3&E-) Fall 1981* 27.40 (1.54)b Summer 1979 30.11 (0.84)bc Spring 1979 30.44 (1.68)abc Spring 1980 33.41 (1.64)abc Summer 1980 36.10 (1.04)8 Fall 1979 39.27 (2.80)ac Fall 1978 39.59 (3.09)abc *Experimental season (all others are baseline seasons; Means sharing a common letter are not see text for explanation. significantly different. 159 Table 3.26. Age at disappearance for control mice and untreated mice from 6 previous seasons. Age at Disappearance (Days) Season x (j S.E.) Fall 1981* 27.40 (3.44)b Summer 1979 30.11 (0.84)bc Spring 1979 30.44 (1.68)abc Spring 1980 33.41 (1.64)abc Summer 1980 36.10 (1.04)8 Fall 1979 39.27 (2.80)ac Fall 1978 39.59 (3.09)ac *Experimental season (all others are baseline seasons; see text for explanation). Means sharing a common letter are not significantly different. be to monitor blood steroid levels in the juveniles before they disperse. This approach is being attempted by I. McDonald and M. Taitt (in Taitt and Krebs, 1982) in a study designed to examine the role of gonadal steroids in controlling certain female behaviors in field populations of voles. Although hormonal manipulation of animals in the field is becoming more common, the hormonal treatments are almost exclusively restricted to adult individuals (e.g. Krebs 55 31,, 1977; Taitt and Krebs, 1982). Only in the present study and one by Ho1ekamp (1982) were juveniles treated with hormones. Furthermore, most studies were designed to investigate aggressive behaviors. In these field studies on aggression, and in their laboratory counterparts (e.g. Whitsett 25 ‘21:, 1979), some degree of causal relationship was observed between gonadal hormones and aggression, particularly for adult males (see review by Bronson and Desjardins, 1971). Although the authors of 160 these studies did not examine juvenile dispersal in the field, many invariably suggested that, in the field, hormonally-mediated aggression in adults is important in controlling, among other things, mortality and dispersal of juveniles (Ayer and Whitsett, 1980), and recruitment of juveniles to the pOpulation (Whitsett g a_l_._, 1979). Some of these investigators, as well as others (Fairbairn, 1977a,b, 1978a; Healey, 1967; Sadlier, 1965), might argue, for example, that in the present study, pubertal juveniles were driven from their natal site by aggressive adult conspecifics. If sexual maturity in juveniles was an aggression-eliciting cue for adults, it is possible that juveniles were forced to disperse from their natal area only after attaining puberty; If this were true, adult aggression could explain the results of the present experiment where juveniles, irrespective of treatment, disappeared shortly after becoming pubertal. This aggression-related argmnent, however, can be disputed on several grounds. Firstly, overt aggression among conspecifics has never been directly observed in natural populations (see Chapter 4). Even indirect evidence of aggression in the field is rare. In the present experiment only 5 juveniles in a total of 4 experimental seasons exhibited any signs of scars that may, or may not, have been the result of wounding. Secondly, the widely held notion that juvenile dispersers are forced out of their natal area by their parents needs to be re-examined in light of the discovery by Halpin (1981) that parents do not show aggression towards their young in £g_ & austerus, and _P; 1': saturatus. Adult aggression is usually only documented under laboratory conditions in which adults are allowed to 161 interact with other adults, or with unrelated juveniles (e.g. Ayer and Whitsett, 1980; Rowley and Christian, 1976; Whitsett 25 21;, 1979). In the present study, owing to the cohesive nature of the family groups (see also King, 1983), the adults most likely to be near the juveniles prior to their dispersal from their natal area are the parents. If, as suggested by Halpin, parents do not exhibit aggression towards their offspring, the juveniles must disperse on their own initiative. Furthermore, the fact that juveniles which remained in the population. exhibited familial cohesion.‘with. their parents argues against aggression (see also King, 1983). Alternatively, aggression could be involved in provoking dispersal if adults other than the parents attack and drive out juveniles. This, too, is unlikely to have been important in the present study because adults from. different family' groups rarely' intermingled (see also King, 1983). Halpin (1981:341) suggested that is possible that EEEEE a juvenile left its natal site, ”the aggressive behavior of unrelated adults may have an effect on its survival and recruitment into the population." Even this type of aggression is probably rare in the present study because there was always an ample supply of space, food, and nesting sites available to support a growing population. Although there is evidence from the daytime nest-box observations which indicates that juveniles are unlikely to be found with unrelated adults, I cannot be certain that this spacing pattern exists during the night, when mice are active and likely to be gone from the nest-boxes. For instance, could adults be overtly aggressive towards unrelated juveniles during nighttime encounters? If juvenile dispersal was the result of nighttime aggression by unrelated adults, 162 one should expect to find a high degree of interaction between unrelated adults and juveniles during the night. I tested this prediction in the next chapter by using radiotelemetric methods to study the patterns of nocturnal activity, space use, and nest-cohabitation among selected adult males. CHAPTER FOUR PATTERNS OF NOCTURNAL ACTIVITY, SPACE-USE AND NEST COHABITATION IN FREE-RANGING MALE g; M; BAIRDI AS REVEALED BY RADIOTELEMETRY Introduction Aggressive behavior of resident breeding adults towards juveniles has frequently been implicated as the primary cause for juvenile dispersal in Peromyscus spp. (e.g. Sadlier, 1965; Healey, 1967; Fairbairn, 1977a,b, 1978a). In fact, aggression has been invoked as an explanation for juvenile dispersal in many small mammals, including microtines (Chitty, 1955; Christian, 1971; Krebs, 1970; Myllymaki, 1977; Reich Egmal;, 1982). As far as I am aware, no one has yet directly observed aggression in unconfined mice in the field, although Terman (1961) described nocturnal observations of aggression in D; m; bairdi between residents and aliens tethered outside of the residents' nest-boxes. In addition, there is indirect evidence of aggression from wounding scars in udcrotines (Christian, 1971; Iidicker, 1973, 1980; Madison, 1980a; Rose and Gaines, 1976; Turner' and Iverson, 1973), though Anderson (1980; see also Literature Review) questioned the validity of this measure. Most of the evidence for aggression between adults and juveniles comes primarily from observations of mice in arena encounters or enclosed populations. Results from these laboratory situations are often extrapolated to unconfined natural populations. Few investigators question the validity of these extrapolations; however, Anderson (1980) warned that "extrapolations from these 163 164 confined situations to nature can be justified only when supported by direct Observations of undisturbed animals in the field.” I took his recommendation and planned a direct, radiotelemetry study of activity patterns of free-ranging & m_. bairdi in an attempt to estimate the likelihood of adult-juvenile aggression in the field. Healey (1967) examined the role of aggression in population regulation in deer mice and admitted that practically nothing is known about the way aggression occurs in the natural habitat. Healey hypothesized that aggressive male mice were more active and therefore more capable of occupying large home ranges than docile males. He tested the first part of the hypothesis and found that aggressive mice were more active than docile mice. He went on to suggest (1967:383) that "the number of social contacts a mouse makes must depend, in part, on how active it is. In this case, the aggressive mice would be more likely to encounter and threaten a strange juvenile.” It was unfortunate that he could not test the second part of his hypothesis. In this experiment I used radiotelemetric methods to examine the likelihood of encounter between resident breeding adult males and other mice in a population with known familial composition. Materials and Methods The study site for the radiotelemetric analysis was the same as described for Experiments 3.1 and 3.2 (General Materials and Methods, Chapter 3). The animals used in this experiment consisted of 16 of the 17 mated adult male Peromyscus maniculatus bairdi introduced as members of the original bisexual pairs in the Fall 1981 and Spring 1982 seasons of Experiments 3.1 and 3.2 Nine mice were radiocollared in the Fall 1981 season and 7 mice were studied in the Spring 1982 season. 165 On the ninth day of confinement to the introduction couplet (see General Materials and Methods, Chapter 3), each adult male chosen for radiotracking was fitted with a miniature radiotransmitter which followed the design of Shields (1976). Each transmitter package consisted of transmitter elements, a battery, and an rintegral tuned-loop antenna that also served as the collar (Figure 4.1). Figure 4.1 Schematic diagram of a radiocollar. The diagram on the right shows only the electronic components of the radiocollar. The one on the left shows a radiocollar that is potted with dental acrylic and fitted with a protective sleeve. Prior to the attachment of a radiocollar to a mouse, the transmitter package was potted with dental acrylic to protect the electrical components from potential damage that could result from gnawing (Madison, 1977) or moisture. Finished radiotransmitter packages weighed from 1.9 to 2.4 g. No mouse was collared with a transmitter that weighed more than 152 of the animal's body weight. The radiocollar was fitted to the mouse by' customizing the tuned-100p antenna to fit snugly around the neck. The collar and its circuitry were completed by soldering the ends of the antenna leads together with a low-wattage portable soldering gun. Observations of mice fitted with radiocollars of this design in a preliminary test of the transmitter's 166 efficacy indicated some skin irritation occurred from exposure to the bare wires of the antenna-collar. To solve this problem each collar was fitted with a small sleeve of plastic heat-shrink tubing. This sleeve prevented the wire antenna leads from causing undue irritation to the skin. Early attempts at attaching collars on non-experimental mice in the field resulted in less than satisfactory results. Therefore, all mice tested in this experiment were collared in the laboratory. Due to the incomplete synchrony in the introduction dates of the bisexual pairs (Table 3.1, Chapter 3) not all males in each season were collared on the same day, although each male was collared on the ninth day of its confinement to its couplet. Mice were removed from their respective couplets during the daytime and brought back to the laboratory for the attachment of a radiocollar. An attempt was made to return the collared mouse to its home couplet within four hours after being removed. All mice were returned to the study site before sunset. During the males' absence, the females and their nursing young were left confined to their respective couplets. On the tenth day of confinement, the arena hole was opened and the radiocollared mice and their respective families were allowed the opportunity to move freely about the site for the remainder of the 10-week experiment (Experiments 3.1 and 3.2). The mean age of firstborn litters in the population at the beginning and end of radiotracking for both seasons was 14.2 i 0.9 (S.E.) and 42.0 i 3'8 (S.E.) days, respectively. Radiotelemetric monitoring of the tagged mice began on the first night following the opening of the arena hole. This allowed one day for the mouse to adjust to wearing the radiocollar before it was radiotracked. Hamley and Falls (1975) and others recomended that 167 radiocollared mice be monitored for any signs of abnormal behavior that might be a result of either the collaring procedure or of the collar itself. From preliminary observations of collared mice in the laboratory, I determined that a one-day adjustment period was sufficient. Herman, (1977) also .accepted. a 24-hour familiarization period. During my preliminary observations, I noted that, for most mice, locomotory, grooming, and feeding behaviors were normal within 30 minutes after the radiocollar was attached. Monitoring involved surveying the site with a frequency modulated (FM) radio receiver once every two hours beginning shortly after sunset and continuing until shortly before sunrise. Due to differences in night length between the fall and spring (fall: R I 13.5 hrs; spring: 5 I 9.75 hrs), six radiolocations were possible during nights in the fall compared with only five nightly observations during the spring. Locations of tagged mice were determined by a single observer who slowly searched the areas near the couplets and in the surrounding habitat at the study site. During the day, the collared mice and their families often occupied nests within particular couplets. Knowledge of these daytime nest positions offered a clue, at the start of each tracking period, of where to begin radio surveillance. These areas were the first ones scanned at the beginning of each tracking period. If the radio signal detected indicated that a collared mouse was present in a couplet, the nests located within the couplet (both arenas) were quietly checked. The use of a flashlight confirmed the presence of a collared mouse. The location of the collared mouse and, whenever possible, the location and identity of all other mice found in the nest-boxes were recorded. If a radiocollared mouse was not found 168 in a couplet and its radio signal was not detected in the area immediately surrounding its last known nest location, surveillance was carefully and quietly extended to the study site at-large. Searching continued until maximum signal strength was obtained. By reducing antenna gain, a successively smaller area of signal strength could be localized. At this point, scanning with a flashlight occasionally revealed the mouse moving slowly about in the grass. A position fix was recorded on a prepared map of the study site that was partitioned into 2 m squares. Any additional information regarding the collared mouse was recorded as notes on the site-maps. If the signal of a collared mouse was localized but the mouse was not directly observed, a description of the animal's likely refuge (if not in a nest) was noted. Also noted were descriptions of any encounters between a collared mouse and any other mice. In addition, at the beginning of each night of radiotracking, the air, temperature and sky conditions were noted. Minimum temperatures for each night were obtained from the National Weather Service office at Michigan State university. At the end of every two-hour interval, the following data were collected and recorded: 1) location and identity of each collared male, and 2) the identity of each mouse found cohabiting with the collared male (if the collared male was found in a couplet). Analysis To answer the question concerned with male activity, several activity’ measures were examinedt These included: 1) frequencyr of movement - the number of two-hour intervals during which a male changed locations/total two-hour intervals observed; 2) distance moved - distance moved/interval; and 3) average step length - distance moved/ number of intervals during which a male changed locations. 169 To answer the question concerning the likelihood of encounter between collared males and other mice, the following measures of social interaction were used. From records of nest checks at couplets where a collared mouse was found, the mice found cohabiting with a collared mouse were classified as either a family member or stranger. The percent of total observations in which a collared male was either alone, or cohabiting with a family member or stranger was determined. In addition, home range estimates of the collared mice were plotted from locations recorded on the site-maps. The specific hypotheses tested here are: 1. Activity measures for fall and spring males are not expected to differ. :2. In both seasons, males will be more likely to move, and move farther, on nights when females are exhibiting post-parous estrus than on other nights. 3. The home ranges for fall and spring males are not expected to differ. Comparisons of activity between seasons were made using t-tests. Within-male comparisons of activity during parturition periods and non- parturition periods were made using paired t-tests. Tests of the likelihood of males moving on nights during the parturition period versus non-parturition periods were made using a Wilcoxon matched-pairs signed-ranks test. Results The nine males in the fall 1981 (hereafter called Fall) were radiotracked for an average of ten nights each (i I 51 observational 170 male) during the period from 20 October to 4 December. In the spring 1982 season (hereafter called Spring), seven males were radiotracked for an average of 12 nights each (1.: I 51 observations/male) from 27 April to 13 May. No collars were lost in either season. One collared male in the Fall was killed by a least weasel but the body and collar were recovered. Radiotracking of another Fall male was stOpped prematurely due to an injury caused by a malfunctioning collar. There were no injuries or deaths directly attributable to the collars during the Spring. All of the collars were removed from the mice at the end of each radiotracking season. Social Interactions In both seasons, the males were most often found nesting or resting within a nest in a couplet (Table 4.1). Spring males were more Table 4.1 Percent of total observations in which collared males occupied a couplet or a nest (if in a couplet). Occupation Z (N obs) Nest- Couplet (If in Couplet) Fall 1981 Spring 1982 Fall 1981 Spring 1982 Present 87 (403) 72 (258) 93 (375) 98 (254) Absent 13 ( 58) 28 ( 98) 7 ( 28) 2 ( 4) Totals 100 (461) 100 (356) 100 (403) 100 (258) likely than Fall males to be radiolocated outside of the couplet. Only rarely was a male observed between nests within an arena. 0n the few occasions when a male was visually observed in the field away from an 171 arena, the mouse was usually alone. However, on one occasion in the Spring, I was able to identify another mouse that was nearby; it was a 21-day-old son of the collared male. Both mice happened to be taking refuge in a small hole located approximately 15 m from the "home" couplet. Although juveniles were not radiotracked, some indication of their activity patterns was provided by checking nests in arenas occupied by collared males. Juveniles 16 days old or younger were almost always found in a nest, while Older juveniles were found in nests less frequently. When present in couplets, both collared males and other mice were frequently found in the nest containing sunflower seeds. In both seasons, males were more likely to be found sharing a couplet with family member than with a stranger, although the male was usually alone in a nest (Table 4.2). Family members consisted of the Table 4.2 Percent of total observations in which collared males were alone or cohabiting with other mice in a couplet or in a nest (if in a couplet). Cohabitation Condition Z (N obs) Male in Nest Male in Couplet (If in Couplet) Fall 1981 Spring 1982 Fall 1981 Spring 1982 With Family 54 (217) 63 (161) 19 ( 72) 30 ( 75) With Stranger 6 ( 24) 1 ( 3) 3 ( 11) 0 ( 1) With Family & Stranger 3 ( 14) 0 ( 0) 0 ( O) 0 ( 0) Alone 37 (148) 36 ( 94) 78 (292) 70 (178) Totals 100 (403) 100 (285) 100 (375) 100 (292) 172 male's mate and her offspring. A male was found with a stranger in the same couplet on 27 observations (Table 4.2: Fall I 24; Spring I 3) with either a lone adult, or an adult with offspring present. Collared males were never found with only a strange juvenile, and only six observations were with other adult males. The six male-male encounters occurred on two nights between two males in the Fall (Mouse D and I; Figure 4.2). On these two nights, a strange collared male (I) was found in an arena shared by a formerly-collared male (D) and D's mate and offspring. There was evidence of severe wounding on the head and belly of the resident male (D). On several observations during these two nights, the intruder male was found in the same nest with D's mate and juvenile offspring. On these observations, D was found alone in a separate nest. On the third day following the initial observed encounter between the two males, D moved 40 111 south to an adjacent couplet. The intruder male maintained D's original mate as a consort for an additional two weeks until the end of the experiment. Twenty-one observations (78%) of male-stranger encounters were between a collared male and an adult female. On 13 of these observa- tions, the female had either just given birth on that day or the previous night. Males were frequently found in the same couplet with a postparous female (either his original mate or a stranger) even though, in some cases, the male had never before been observed in that couplet. Activity Patterns General In the Fall, collared males changed locations, on the average, once out of every four observations, moving nearly 26 m at a time (Table 4.3). Spring males had a higher frequency of movement than Fall 173 AEV summon mz mo.c om.nm ou em.m~ om.o + no.oN ON.mm ou mm.w~ mn.N + Nn.mN nmum mwmuoe< I I Aam>umunfi\av mz mm.o on.mfi on 55.0 mo.H + em.m om.NH Ou mw.~ om.N + em.w oo>oz monmuMHn Aam>uoucfi mz o~.~ «0.0 cu oc.o wane + mm.o oo.o ou oo.o wo.o H mN.c \mo>oav unmao>oa mo honosumum Auv mnommom mwnmm .mvm m ownmm .m.m x hua>wuo< ammsumm I oucouomman area mnauem arms Harm nommmm .mmame ewxomuuoaomu no nonsense >ua>uuom coo: m.e manna 174 males but the difference was not significant (t I 1.10; df I 14; p > 0.10). There was also no difference in the average step length between Spring and Fall males (t I 0505; df I 14; p > 0.25). When distance traveled is measured as meters moved/Z-hr interval (to correct for differences in night length between Fall and Spring nights), one finds that, again, Fall males moved no further than Spring males (t I 0.37; df I 14; p > 0.25). Figures 4.2a-c illustrate the space use by the nine Fall and seven Spring males. Males typically restricted their ‘movements between neighboring couplets. Males Often centered their activity around a couplet that served as a diurnal refuge. Activity areas are estimated by joining peripheral positions by a minimum number of straight lines and measuring the enclosed area (Madison, 1978). The average activity area for mice that moved was 0.077 1 0.03 hectares for Fall males and 0.112 : 0.02 hectares for the seven Spring males. Two males in the Fall and one male in the Spring were never observed outside of their home couplet. I detected no clear pattern of weather influence on activity patterns. Mice appeared to be just as active on cold, snowy nights in the Fall as they were on mild, dry nights in the Spring. The average low temperatures on nights during the Fall and Spring radiotracking periods were 39.0 : 8.1°C and 45.4 : 2.30C, respectively. During,parturitionpperiods Males from both seasons were more likely to move on nights during periods when neighboring females were in parturition than on nights when no parturition occurred (Figures 4.2a-b; Wilcoxon matched-pairs signed-ranks test, p < 0.05). Males in the Fall moved significantly 175 FROM TO 0AIO/29 ”/7 I8 IO/23 "/3 CO 10126 ”/4 .I. COUPLET I! > 20 It or TIMES 9 POSITION OFdWRING PARTURITION I 5-20 OBSERVED AT f POSITION OFQDIRING PARTURITION I <5 LOCATION Figure 4.2a Positions of 3 collars males obtained during 2 weeks of radiotracking in the Fall 1981. The positions are super- imposed on a map of the mouse cities and indicate the relative frequency of observation at any given location as well as the position of males relative to the position of neighboring parturient females (see text for explanation). 176 FROM To FROM TO DO II/Ia ~II/2I AG II/Is II/zo OE II/I3 II/2I II/zs 12/4 IF II/Is II/2I IH II/Is II/ZI IU29 e/4 IV29 Ivso 01 was 12/4 _— COUPLET I >zo'# OF TIMES p POSITION OFd'OURINC PARTURITION I s-zo OSSERVEOAT f POSITION OFQOLBING PARTURITION I <5 LOCATION Figure 4.2b Positions of 5 collared males obtained during 3 weeks of radiotracking in the Fall 1981. The positions are super- imposed on a map of the mouse cities and indicate the relative frequency of observation at any given location as well as the position of males relative to the position of neighboring parturient female (see text for explanation). + COUPLET p POSITION OPdwRINs PARTURITION f POSITION OPQOLRINC PARTURITION Figure 4.2c f6 \ \ D3 \4 . \ fps}, ATE p41‘ 93.4.7 03 177 ./ \ ./°‘6| ./ n). . . l p) L. + -+TT\.\. .pl\. + 4-‘1 94 p4 FROM TO PROM To OJ 4121 514 ON 4128 519 UK 4128 5H3 A0 51I 51m IL 4121 an: AP 4130 51s I. M 4130 5/13 I > 20 it OF TIMES I 5-20 OBSERVED AT I < 5 LOCATION Positions of 7 collared males obtained during 2 weeks of radiotracking in the Spring 1982. The positions are superimposed on a map of the mouse cities and indicate the relative frequency of observation at any given location as well as the position of males relative to parturient females (see text for explanation.) 178 greater distances on. nights during parturition. periods than. during nights of non-parturition periods (Figures 3a, b; Table 4.4; t 8 3.32; df = 3; p < 0.05). Males in the Spring also moved greater distances during nights of parturition periods but not significantly so (t = 1.86; df = 5; 0.10 > p > 0.05). Discussion Social Interactions The results of this experiment indicate that a considerable degree of familial cohesion exists among free-ranging _P_. & bairdi during their nocturnal activities. These findings are in close agreement with results of diurnal nest-box censusing reported in Experiments 3.1 and 3.2 (Chapter 3), and in King (1983). Males from both seasons spent a majority of time each night in a couplet. However, spring males were more likely than fall males to be absent from a couplet (Table 4.1). On nearly 602 of all Spring observations, males were found sharing a couplet with at least one family member (Table 4.3). The presence of family members in a couplet could have attracted a male to the couplet and maintained his residency there. A male that shares a couplet with his family may be also assist his mate in parental care of their offspring. One fall male nested with his lZ-day Old offspring for two nights following the permanent disappearance of his mate. Paternal care has been observed in California mice, Peromyscus californicus parasiticus (Dudley, 1974) and in a laboratory-reared deer mice (Horner, 1947); however, observations of direct attentiveness by males towards their offspring in the present study were rare. I typically found a male and his family in separate nests or even separate arenas within a couplet. This was true especially when females were nursing 179 .moamaom wouuonswwoo mo ovouuoq sous IwusquAIoos can sowuuusuuoa onu mcausv mouse as voao>ouu moocouowo osu ououuusfinu magnum one .fiwafi dash oau :« mono: wouoaaooowvou an voao>muu moosoumwv maunwac HmuOH mn.< Ousmum >oz >02 uua >oz vn~.38 83,582....3. cn~.3a~.uo~eseeo o D I ‘1 2338.25. o~ L ON 0‘ i 0' HM .3.§.z* o. . 8 u. go: . O. a I 8 w .23 Suited 8: . 1 8. I. 8:... 52.28.. E 8: . A . 8. w. 3...: so: .... Mn“. 1 m Mn" M. 2% eaq -23 98 >2. >9. >2. o vn~.83.~8o.senn 58222292». acaemennxn. 0.>¢(20..—(hm. v ..ON A , a. m. .. r 100 T... .. Y 60 fl . . Too..l ,4 ”i r o“. W . A . ‘0'. M m a 1 .wgm . . ;z. 1 m_. n— : m a S 0 +5. m .;;:. ........ Ra >oz poo o n a ...non 3.158838 oifizsfiz. ou . «w . an o. a I 5.3.“ A 3 .m- 00 4 w “:2. . 8 m...“ o. A r o 9 oo. L 4 OS on. . . . u. m oc. L . .3. M O.— .. v . 8. «I on. w .m We» I v m an. ..m m o. M 0°" ., ...................... “mint...manna.“ N 180 200 O a: K > I50 «1 . (I . F. — Ioo- - (n - a: _ E I u 5°: ' 5 : (untimrymr nuawlzsa aawI234ssvssmnma m NAT APR MAY 200 8 r -’ Z I; ISO: < _ a: - '- I00: a, _ 05 : 5 so} 5 : (Itmimy)0._ : : ' . , : . .; 27282930! 2345 8789l0lll2|3 30! 23 45 c 10910" I2I3 APR MAY man a 200 II] .J '5’ I50 < _ E : 4) Ice-_- x .- .l-I_J I In 5°: 3 : (static-aymr ' ' 5? 7 , ‘ zszswlzsass‘las I23456799I0III2I3 APR MAY MAT zoo_ no; ; ? um um .00: .PWIflon Potiod Z Pormm 50: OM“ : *Noighbor (stofloncyWI— ;' Z ; : ' ’ 30| 2 3 4 5 e 7 a 9 APR MAY Figure 4.3b Total nightly distances traveled by radiocollared males in the Spring 1982. The graphs illustrate the distances traveled by males during the parturition and non-parturi- tion periods of neighboring females. 181 .AsmHHmOImcov mo.o v mg .ummqu vwoammo .cOfiufiuauuma mo oumw mnu nouwm was .wcfiusv .muommn unmwn wso mo wcwumfimaoo vowuma %vam a .mOOfiumn nuon wafiusv vo>oa umfiu moama haao movafioch mz om.~ Aom.~v afi.m o AoS.Nv mH.m~ e mama wasnam « ~m.m Aofi.mV om.o S AH~.¢V om.ma S Hams Hams on A.m.m we m moaaz A.m.m we m mosaz commmm mooauwm :mm3umm vaHHmm coaufiusuummIaoz nmvowumm :oauaunuumm mo:OHOHMHn wcfiusa munwaz weapon mucwfiz Aam>nmucw\av mmoamz vmuoaaoo hm voHo>muH mucoumfin was .moaoa voumHHoo mo aufi>wuom awamawm wcfiuonnwfioc hammoao he cofiufiusuuma mo mafia monsoon awnmcowumHOM <.¢ manna 182 young pups. The apparent avoidance between adult males and adult females with Offspring is a pattern observed during both diurnal nest-box censusing (Experiment 3.1, 3.2, Chapter 3; King, 1983; Howard, 1949) and nocturnal monitoring. Hill (1977) suggested that dispersion patterns Observed in natural populations of Peromyscus result more from mutual avoidance than from aggression. If males are providing parental care for their offspring during nocturnal periods, it probably occurs during brief encounters within the nests. MOst encounters within nests were between an adult male and his mate. Rarely was an adult male found in a nest with only his offspring. No aggression. was detected during observed encounters between family members. II am not aware of any published reports of observed aggression between family members in natural populations of Peromyscus, although aggression between mothers and their offspring has been observed in a laboratory study by Savidge (1974). Under the conditions of this study, Savidge suggested that some mothers attacked their offspring after he introduced the young to the mothers' cage, while other mothers did not. Savidge also suggested that differences in aggressiveness between mothers might explain Observed differences in the rates of departure of their offspring from the natal cage. He hypothesized that offspring of ”aggressive" mothers would depart the natal cage at a higher rate than offspring of "non-aggressive” mothers. He based his classification of aggressiveness on whether or not females attacked strange juveniles. A test of his hypothesis revealed that juveniles with "aggressive" mothers departed from the natal cage at a higher rate than those with ”non-aggressive” mothers, although he did 183 not report any observations of aggression between mothers and their offspring. Most demonstrations of aggression under laboratory conditions have been observed between non-family members (Rowley and Christian, 1976; Ayer and Whitsett, 1980; Gleason, gt 3.1;, 1981; Whitsett, E Fit: 1979). Halpin (1981) reached an opposite conclusion in a study of interactions between parents and their own offspring. She reported ”an almost complete lack of aggression between adults and their own young." Halpin went on to suggest that parental aggression may be more prevalent under field conditions after the birth of a second litter, however, I found no field evidence of parental aggression towards firstborn juveniles either before or after the birth of a second litter. Males were found in a couplet with only a stranger on just 72 of all observations (Table 4.2). In the majority of these observations, the male was found with a lone adult female, or an adult female with her offspring present. I never observed a male in a couplet with only a strange juvenile. Encounters with a strange adult female usually occurred on nights during her period of parturition and post-partum estrus. Interactions of this type were also revealed by radiotelemetry for meadow voles, Microtus 4pennsylvanicus (Madison, 1978, 1980b). Madison reported that radiocollared males were significantly' more likely to overlap in space with neighboring males and females when females were in estrus than when they were not in estrus. In the present study, some observations of couplet sharing between a male and an adult female revealed that the male was visiting the couplet for the first time. This suggests interesting questions regarding the 184 mechanisms by which males detect the presence of estrous or pregnant females. It is likely that males, in their wanderings between couplets, are attracted to cues provided by an estrous or pregnant female. Chemical signals produced by female rodents are probably the most important cues for attracting males to females (see reviews by Aron, 1979; Brown, 1979; Gosling, 1982). Moore (1962, in Bronson, 1979) demonstrated that female 3; maniculatus produce substances that attract males. In mice, the attractiveness of female urine increases during estrus or pregnancy (Bellamy and Davies, 1971; Davies and Bellamy, 1972). Whether the male is responding primarily to cues localized at the females' home couplet, or distributed generally in the field surrounding the couplet, is not known. Although adult females were not radiotracked, their habits could sometimes be inferred from checking nests in couplets occupied by collared males. Females spend some portion of each night wandering in the field away from the couplets as indicated by occasional trap captures. I don't know the amount of space used by an average adult female, but if it is similar to that of a male, she has the opportunity to release olfactory signals (in her urine or other secretions) over a wide area surrounding her home couplet. The attractiveness of urine from estrous or pregnant females to males could explain the observed cohabitation of males with females during their parturition periods. Only six male-stranger encounters were between two males. These encounters provided the only evidence (p. 172 ) of intraspecific aggression between two mice. Although there have been reports of Observed male-male aggression among laboratory reared Peromyscus (Healey, 1967; Ayer and Whitsett, 1980), no one has reported aggressive 185 encounters in a natural, unconfined pOpulation. The likelihood of male-male aggression is probably rare in natural populations because male-male interactions are similarly infrequent. Data reported from nest-box studies indicate that, during breeding seasons, same-sex cohabitations are rare (Nicholson, 1941; Terman, 1961; King, 1983). Although males were frequently found in a couplet shared by family members, males were Often alone in couplets (Table 4.2). Therefore, social attraction among family members is not necessary to explain the attractiveness of couplets to adult males. Males Observed in otherwise unoccupied couplets were frequently found in the nest containing sunflower seeds, and I assumed the mice were feeding. It appears that even an, unoccupied couplet, with its supplemental food.Iand 'refuge sites, is attractive to free-ranging males. In fact, three males (two Fall, one Spring) were never Observed outside of their home couplet during the radiotracking period. Although males were radiolocated at locations away from a couplet on 156 observations (Table 4.1), I rarely visually observed the mice. On one occasion, however, I observed a male and his weanling son enter a refuge together. Other field Observations indicate that juveniles may be attracted to the father. Howard (1949) reported finding a father and his four offspring (g; m; bairdi) in a live trap 300 ft. frmm their home nest-box. Howard suggested that juvenile §;_!:_bairdi follow their parents about in the process of becoming familiar with the parents' home range. In a laboratory investigation of possible dispersal mechanisms in 2;.24 bairdi, Savidge (1974) found that weanling juveniles are attracted to the father. 186 Activity Patterns Although radiotelemetry studies have contributed greatly to our understanding of the activity patterns of free-ranging microtines (e.g. Banks, 2£_al;, 1974; Brooks and Banks, 1971; Herman, 1977; Madison, 1978, 1980b; Webster and Brooks, 1981a) and E; leucopus (e.g. Madison, 1977; Mineau and Madison, 1977; B. Ormiston, unpub. obs.), little is known about the movements of free-ranging §:_gt_bairdi. In this study, males from both seasons typically made just over one move per night. The average distance traveled per move was the same for males from both seasons (Table 4.3). No mouse moved further than 145 m on any given night. Based on live-trap records and a knowledge of nest-box homesites, Howard (1949) reported maximum. nightly linear movements ranging from 15 to 107 m. Dice and Howard (1951 reported that g; E; bairdi seldom travel more than 123 m in a straight line from their homesites. These distances are comparable to the nightly values for cumulative distances traveled in this study (Figures 4.3a, b). Most mice moved less than 100 m on any given night. Males were just as likely to move on nights at the beginning of the tracking period as they were on any other nights (Figures 3a, b). Similarly, there was no discernible pattern of hourly movement on any given night. Males could be found moving at any hour of the night, from shortly after sunset until shortly before sunrise. Males from both seasons were more likely to move and to move farther on nights during periods when neighboring adult females were in parturition than on nights when no parturition occurred (Table 4.4; Figures 4.3a, b). These results suggest that males actively search for females, and this search intensifies on nights during parturition periods of neighboring 187 females. Female 2; maniculatus in estrus (and presumably, post-partum estrus) are known to produce signalling pheromones that are attractive to adult males (MOore, 1962; from Bronson, 1979; Moore, 1965). A male probably detects these signalling odors during his wanderings away from his nest and, upon detecting the odor, increases his level of activity in search of the estrous female. Curiously, some males that shared a couplet with a mate on the night of her parturition also moved far on that night (mice B, H, and L: Figure 4.2a, b). I would expect a male to remain in close proximity to his mate during the time she was in post-partum estrus to insure his chance of acquiring an exclusive mating with her. Dewsbury (1979, 1982) reported that female 3; £1 bairdi in cycling estrus require more than one ejaculatory series from a male in order to maximize probability of pregnancy. However, if the female is in post-partum estrus, a single ejaculatory series is sufficient for maximum probability of pregnancy. Thus, if a male wanted to insure a fertile copulation with a post-partum estrous female, only one ejaculation series is required. Remaining close by the female would not necessarily represent any additional reproductive benefit for either the male or female. This could explain the behavior of males B and H (Figures 2a, b) who were found only once in the same arena with their mates on the night of parturition and were absent from them during the remainder of that night's observation. From Dewsbury's findings, I predict that males should spend more time in close vicinity of cycling estrous females than with post-partum estrous females. In addition to searching behavior on the part of males, estrous females could also be active in search for males. Doty (l972a,b, 1973) reported that male urine can act as an attractant to estrous female 2; 188 PL bairdi. Post-partum lactating females are known to spend considerable time away from the nest (_13_._ leucOpus: Hill, 1972; Harland and Millar, 1980) and would, therefore, be likely to encounter other mice. Although all observed encounters between post-partum females and males took place in her home couplet, it is possible for a male and female to encounter each other while both were wandering away from the nest. Further work with telemetry systems under field conditions is required to learn more about the dynamics of mating behavior of £1 1; bairdi. Reserach is currently underway to develop a sensitive electronic tagging system that can be used to monitor the activity of individual mice at a nest (R. Hill, pers. com.). This recording system could be used to gain considerable information on patterns of nest visitation among reproductively active mice. In the present study, radiotracking methods revealed information on patterns of activity, space utilization, and nest cohabitation among free-ranging adult male L & bairdi that would have been difficult or impossible to collect with conventional live-trapping methods. However, all studies which utilize radiotelemetry can be criticized on several grounds (see also Literature Review). The most significant problem pertains to the assessment of the effect of the radiotrans- mitters upon the animal's behavior (Brooks and Banks, 1971 ; see also Hamley and Falls, 1975; Webster and Brooks, 1980). In this study, I relied mostly on subjective criteria to evaluate possible transmitter related interference of normal behavior patterns. I detected no noticeable effects of radiocollars on grooming or feeding behaviors in mice confined in the laboratory. Although a radiocollar fit snugly around the mouse's neck, feeding behavior did not appear to be 189 disturbed. In the food supplement nest in the field, collared mice were frequently seen feeding normally on sunflower seeds. Neverthe- less, I was concerned that the stress of carrying a 2 g collar around might result in an energy deficit and a loss of weight in the mouse. I compared body weights of mice before and after radiotracking and discovered that post-tracking weights were within 1 g of pre-tracking weights. One exception was a mouse which lost 2 g in body weight (19.0 g to 17.0 g). If wearing a radiocollar caused an energy deficit in the mice, body weights could have been maintained in one of two ways. First, a collared mouse could have responded to the burden of carrying around extra weight by reducing his activity and restricting his use of space to the close vicinity of his home couplet. Secondly, a collared mouse could feed more, in which case he could stay in the couplet and feed on the supplemental sunflower seeds, or he could forage widely in search of other food. In this study and others (Howard, 1949; Howard and Evans, 1961), mice stored large caches of wild weed seeds, suggesting that mice were spending time foraging away from the nest and not restricting their intake of food to sunflower seeds. Although mice may have been foraging for food, a comparison of size of activity areas used by males in the present study with home ranges measured by Blair (1940) suggests that mice in the present study were actually restricting their foraging to areas near their home couplet (Figures 2a-c). Blair's estimate of male home range size is 2.5 to 4 times the average area of activity recorded for mice radiotracked in the spring and fall, respectively (see Results). Although radiocollars could be restricting normal activity, I believe the disparity in estimates of space use can be explained by Blair's use of live traps over a longer 190 period, and by recognizing that mice in the present study were newly introduced to the study site. Blair's mice were natural residents in a completely natural field. The study site of the present study was also not considered natural habitat for g; E; bairdi in south-central Michigan (King, 1983). Despite potential radiocollar interference and restrictions in the sampling regime (e.g. sampling only once every two hours), I believe that radiotracking, in conjunction with nest-box censusing, was a valid technique for measuring the likelihood of encounter between adult males and other mice in the population. The results of this study indicate that: 1) adult males restricted their movements to areas between their home couplet and neighboring couplets; 2) males were more likely to move, and move farther, on nights during parturition periods than on nights during non-parturition periods; 3) males spent a majority of each night in the same couplet with family members and were more likely to be found sharing a nest with a relative than with a stranger; and 4) aggression between adult males and other mice was rare; a single observation of aggression occurred between two adult males. 191 GENERAL DISCUSSION Very few topics in field biology offer more challenge to the investigator than dispersal. Dispersal is variously described in the literature on small mamals. It has been regarded as immigration or emigration (Lidicker, 1962, 1975), disappearance (Beacham, 1979; Pfeiffer, 1982), searching behavior (King, 1983), exploratory migration (Mather, 1981), long-distance movement or travel (Myllymaki, 1977; Watts, 1970), "leaving the area" (Fairbairn, 1978a), or simply, movement (Christian and Invis, 1964; Grant, 1978, Fairbairn, 1978a). Frequently, some of these words are used interchangeably even in the same study. Furthermore, dispersal can be defined as, among other things, the movements of an animal from a source (Brown, 1975), or the movement of an animal from its place of birth to a place where it breeds (Howard, 1960). Adding to this inconsistency is the problem of measuring the movements of animals in the field. Typically, dispersal movements among small, secretive animals are inferred from data collected with live-trapping methods. In many live-trap studies, a disperser is regarded as a marked individual that is captured in a location some distance away from its former area of residence. unfortunately, the criteria used to define a disperser vary from one study to the next; therefore, it is hardly surprising to find so many descriptions of dispersal (see Literature Review, Chapter 1). A second challenge to a dispersal biologist is the determination of a cause or causes for dispersal. In live-trapping studies, besides knowing the length and the direction of movement made by a disperser, very little is known regarding the process of the dispersal event. NOnetheless, a knowledge of the dispersers' sex, age, weight, genotype, 192 and reproductive condition at the time of capture provides some clue to understanding the causes and mechanisms of dispersal. For example, in an annually cycling species, such as Peromyscus, dispersers are usually, but not always, young, lightweight individuals of either sex which may, or may not, be sexually mature. One explanation for the cause of dispersal of these individuals is the response of young individuals to aggression and social pressure by resident adults (e.g. Fairbairn, 1978a; Petticrew and Sadlier, 1974). Frequently, only one, or at most, a few causes are ascribed to dispersal in any particular study because many investigators disregard the considerable variation in the characteristics of dispersers. For instance, there are many examples in the literature of attempts made by authors to identify from a varied sample of dispersers, a predominant or "mean disperser" - for example, a young, sexually mature juvenile - and then propose a cause for its dispersal. As an example, it may be suggested that because this disperser is sexually active, it is dispersing to search for a mate, or to avoid mate competition with older resident adults. While these explanations may apply to the sexually mature disperser, what about those dispersers that are sexually inactive? Perhaps these other dispersers are not moving to find a mate but, instead, are moving to secure some other resource such as food. Often, these other dispersers and the reasons behind these different movements are ignored. Certainly, the variation in disperser characteristics suggests that individuals move or disperse for many reasons. Why, then, is there an apparent reluctance on the part of many investigators to propose multiple explanations for dispersal? Part of the reason may be that it is intuitively appealing to expect a universal cause for the 193 phenomenon. Secondly, and more importantly, few experiments are designed to identify and experimentally examine, the ecological variables, such as resource availability, that could influence dispersal. An individual's decision to move or disperse depends on its unique perception of the spatial and temporal availability of important resources such as mates, food, shelter, etc. Although few investi- gators would deny the important influence that these ecological variables may have on dispersal, these factors have been, nevertheless, difficult to assess and control in the field. Often, this problem can lead to fortuitous interpretations of the causes of dispersal. For example, in her study of §._maniculatus, Fairbairn (1978a) argued that the cause of dispersal in lightweight subordinate males varied seasonally. She believed that in one season these individuals dis- persed as a result Of competition with larger, dominant males over the availability of receptive mates. In another season, she believed that this same type of individual dispersed due to local limitations in the supply of food and nesting sites. Because these variables were not controlled in each season, it is diffuclt to be certain, for instance, that food supply was limited in one season and not the other. Obviously, if the supply of mates and food were both limited in the same season, dispersal could be attributed to either one or both. In addition to a weak knowledge of how resource availability affects dispersal, in many studies little is also known about the social processes involved. Although Anderson (1980) has said that "dispersal is a social, rather than an individual, process,” the social context under which dispersal occurs is often ignored or misinter- 194 preted. Some investigators have argued that juvenile dispersers may be forced out of their natal area by their parents. Although behavioral interactions between parents and their offspring are certain to influence the offspring's decision to disperse (Savidge, 1974), parental aggression is probably rare under natural field conditions (Terman, 1961). I doubt that many of the investigators who support this aggression argument understand the true nature of the behavioral interactions that occur between adults and juveniles in the field. MOst of the evidence for juvenile-directed adult aggression is derived from laboratory studies in which adults are exclusively paired with unrelated juveniles. This may be the wrong social context for judging whether aggression in adults can explain the dispersal of juveniles from their natal area. For instance, field evidence suggests that prior to natal dispersal, juvenile g; m; bairdi are likely to be found with their parents and not with unrelated adults (Howard, 1949; King, 1983; Chapter 4). Furthermore, the likelihood of adult aggression toward their own offspring is low (Halpin, 1981). These results, therefore, challenge the importance and relevance of adult aggression as a catalyst for juvenile dispersal, at least for _P_._ maniculatus. This example illustrates the value of knowing something about the early social history of the disperser (Bekoff, 1977, 1981), which is easier to obtain in nest-box studies because the familial composition of the population can be determined. For live-trap studies, on the other hand, the social history of the dispersers is a ”black box.” To summarize, I see four limitations in the way that dispersal has been studied in the past. The first is that operational definitions of dispersal are so varied that it becomes difficult to interpret what 195 kinds of movements are being measured. Animals move for many reasons. Secondly, very little is known about the spatial and temporal distribution of resources that may affect an individual's movement patterns. Third, an understanding of dispersal requires that we attempt to take into account the social context under which animals disperse, something that has been difficult to do. Fourth, in the past, many investigators have relied almost exclusively on the visible characteristics of dispersers caught in traps to infer the causes of dispersal. Looking for the causes of dispersal in this manner can be likened to someone searching for lost keys under a streetlight, not because they were lost there, but because that is where the light is. Investigators have looked at the circumstantial evidence of age, weight, sex, reproductive condition of the dispersers and suggested probable causes for the dispersal. What is needed though are more attempts to identify and experimentally control features of the ecological and social stage from which individuals disperse. In the past 5 years descriptive studies have been replaced by experimentation. In these studies particular attention has been placed on controlling certain variables known to be important in dispersal, and manipulating others. Manipulations have included the provision of supplemental food (Desy and Thompson, 1983; Gilbert and Krebs, 1981; Taitt, 1981), alteration of sex ratios (Redfield g a_l., 1978), and provisions of extra refuges (Taitt st 31;, 1981). The importance of aggressive behavior as a mechanism for promoting dispersal has been examined by using exogenous hormone treatments (Krebs g 11;, 1977; Taitt and Krebs, 1982). The endocrinological cause of natal dispersal 196 in ground squirrels was also examined with the aid of gonadal steroid treatments (Holekamp, 1982). In many studies, the potential influences of environmental features and individual differences in resource requirement on dispersal were considered. Nonetheless, limitations in the experimental designs need to be overcome before we can understand what effect the manipulations in behavior or resource availability have on dispersal. For example, Taitt and Krebs (1982) compared the rates of immigration of voles onto a grid containing ”aggressive-females" (Testosterone-treated) and several control grids. Immigration of females onto the testosterone grid was higher than it was on either control grid. These results, however, are difficult to interpret because the testosterone and control grids also differed quantitatively with respect to vegetation and cover, which are also known to influence spacing behavior (Taitt £5 31;, 1981). Many ‘variables have been. associated ‘with. dispersal, including cover or shelter, nesting sites, nutrition, population density, sex, age, reproductive condition, competition, and predation. In most live-trap studies, few‘ of these independent variables are experi- mentally controlled in contrast to some success in enclosure studies (e.g. Gipps and Jewell, 1979). Unfortunately, dispersal in fenced enclosures is restricted (cf. Riggs, 1979). Although I criticized the practice of hypothesizing one cause of dispersal, I felt justified to test only the sexual search hypothesis because I partially controlled resource availability and other variables known to influence dispersal. Peromyscus m; bairdi readily uses nest-boxes (Howard, 1949; Dice and Howard, 1951; King, 1983) which 197 allowed me to control the availability of shelter and nest sites and exert some control on predation. Furthermore, by introducing a known number of established breeding pairs into the arenas I was able to control initial demographic features of the pOpulations studied. Because 3: a: bairdi readily ate sunflower seeds, I was also able to control the quantity of food available to the mice. The lack of mortality at the nest suggests that disappearing juveniles dispersed from their natal area. Natal dispersal was most likely to occur among juveniles that had recently become pubertal. By treating mice with the anti-gonadal hormone melatonin, I was able to retard pubertal development and delay natal dispersal in the treated mice. If mice disperse in search of a resource that is limited in their area of birth (e.g. Dice and Howard, 1951), then many of these resources were at least partially controlled in the present study: nesting sites, quantities of food, and shelter from predators. One resource that I did not control was mate availability. In the small family groups of mice that are characteristic of_I_’_._ & bairdi (Howard, 1949; King, 1983; present study), juveniles rarely mate with siblings or their parents. For example, in Howard's study, only 4 - 101 of the litters observed could have been produced by matings between siblings or between parents and offspring. Inbreeding avoidance among siblings has also been observed in the laboratory. Hill (1974) reported that sibling pairs of g_._ & bairdi had lower reproduction and reached reproductive age later than unrelated pairs. Parent-offspring matings are also uncommon but probably occur more Often than matings between siblings (Haigh, 1983, Howard, 1949, Spevak, 1981). 198 The infrequency of consanguineous matings could be attributed to reproductive inhibition among young deermice by their parents. Deermice mothers can completely inhibit reproduction in their daughters if they are housed together in the same cage (Haigh, 1983). Pubertal development in the present study was delayed by melatonin treatment but not by the presence of the mother. Perhaps parental influences may inhibit reproductive success of juveniles even though parents do not appear to delay gonadal development in their offspring. Clearly, more research is needed to determine how parent-offspring social inter- actions in the field influence reproductive competence in young mice. Juveniles may have little Opportunity to mate if they remain in their family group. Therefore, it behooves sexually maturing juveniles to leave their natal area in order to reproduce. In the present study, departure from the natal couplet for other resources was probably not the cause of dispersal. Nesting sites, refugia, food, and water were in ample supply, which leaves potential mates as the most limited resource at the natal couplet. Juvenile dispersal may be based upon the individual's gonadal maturation and perception of sexually attractice stimuli in the environment surrounding the natal area. Various odors, particularly those found in urine, are known to be attractive to the Opposite sex (e.g. see review by Aron, 1979), and may explain why mice leave their natal nest site. It is certainly plausible that these and other social signals may be found in excretory' markings .located. throughout the individual's environment. These signals could be detected when 2-week-old juveniles begin to explore the area immediately surrounding their natal nest (Chapter 4; Howard, 1949; Layne, 1968). General 199 activity increases a few days after eye-opening (R. Hill, pers. comm.), and juveniles may make excursions from their nest by following their parents (Howard, 1949; Chapter 4). By the age of puberty, they have acquired considerable information regarding their environment, including sexually attractive stimuli and other social signals (Mather, 1981). Perhaps the absence of social signals from the environment could explain why steroid-treated mice failed to exhibit any more movement than controls in the fall experiment (Experiment 3.2). In the absence of sexually attractive stimuli, perhaps no level of gonadal steroids is sufficient to stimulate sexual searching in treated mice. Indeed, sexually attractive stimuli may have been reduced in the fall experiment because fewer juveniles were recruited into the reproductively active portion of the population than were spring and summer-born mice. Unfortunately, it was impossible to assess or directly control the level of sexually attractive stimuli. The effect that the levels of sexual stimuli have on the rate of dispersal in sexualy active individuals could be tested by manipulating the prOportion of sexually mature males and cycling females in the population. In the fall, an increase in the number of sexually mature mice should increase the rate of dispersal among steroid-treated juveniles. .A decrease in the level of sexually active mice, on the other hand, should reduce the rate of movement. Because odors may provide the most information to an individual regarding the location and reproductive condition of potential mates (e.g. Eisenberg and Kleiman, 1972), further tests of the sexual search hypothesis could involve manipulations of the olfactory capabilities in 200 pre-pubertal mice. Spring and summer-born mice that are made anosmic should be less likely to disperse from their natal site at puberty than untreated individuals. Sexually attractive signals may not have been the only stimuli important for the departure of pubertal juveniles from the natal area. Juveniles may also leave the nest to avoid the potential reproductive inhibition brought about by the social influence of their parents or other adults. This reproductive inhibition by parents toward their offspring may be what some authors regard as "adult aggression" (e.g. Fairbairn, 1978a). The nature of this adult influence, however, is apparently more subtle than overt aggression. In one 6 month study, for example, although reproduction was inhibited in juvenile female _P; 11 bairdi housed in cages with either their mother or an unrelated adult female, there was no sign of wounding in any juvenile, and only 2 females (22) died (Haigh, 1983). In the present study, evidence of wounding in juveniles prior to their dispersal was absent. Although adults exhibit aggression towards unrelated juveniles in the laboratory (Whitsett g _ai._, 1979), unrelated juveniles are infrequently encountered at nests in the field (Chapter 4). Familial cohesion was strong in the present study during the night as well as during the day. Juveniles appear to escape the natal situation on their own accord rather than be forced out by parents or rarely encountered unrelated adults. Possible explanations for the disappearance of mice after their natal dispersal include: 1) traveling completely off the site, 2) suffering mortality during their travels, or 3) living undetected within the study area. Even though mice that disappeared may have 201 moved completely off the study site, the results of the perimeter live-trapping suggested otherwise. Only 5 juveniles were captured on the perimeter traplines located 100 m away from the mouse cities, and only 10 were caught in the closer, peripheral lines located 20 m away (Figure 3.1). One possible explanation for the few trap captures at these distances is that mice respond less to traps which lie outside of their home range (Kozel and Fleharty, 1979; Robinson and Falls, 1965). Thus, juveniles could very well have moved off-site without being captured. Juveniles were capable of moving at least 300 m from their natal couplet. One summer-born juvenile from a pilot experiment in 1980 moved approximately 350 m from the mouse cities to a grassy area that was monitored with live traps in another study. This individual was trapped several times within a small area, suggesting that it had established a new home range. One boundary of the mouse cities was not monitored with traps at the 100 m distance due to the presence of a road on the east side of the study site (Figure 3.1). Since roads inhibit movements of rodents (Kozel and Fleharty, 1979), it is unlikely that disappearing mice dispersed to habitat east of this road. Mice that disappeared following their departure from the natal couplet may have died. Once juveniles commence movement away from the safety of the arenas at their natal couplet, the risk of predation is likely to increase. Because juveniles would be expected to have less information regarding their environment than an adult, juveniles might spend more time exploring their surrounds (Mather, 1981) and therefore be more susceptible to predation than adults. This would explain why proportionately’ more juveniles disappeared during each. season than 202 adults. Mortality could also explain why so few mice were caught in the traplines surrounding the study area. How many individuals shunned the concrete nest-boxes and arenas in favor of natural nests and dwelled undetected on the study site is unknown. While most individuals either occupied the nest-boxes continuously or disappeared entirely, some mice used the nest-boxes on a periodic basis. It was common for these mice to be found in a particular nest-box for 2 or 3 censuses followed by an absence of comparable length. Obviously, the periodic nest-box users moved about the study area and used natural refuges and nests when they did not occupy a nest-box. Although many disappearing juveniles could have lived undetected in the study area, there was no way to confirm it. Radiotelemetry was considered at one point as a method to measure dispersal movements and determine the fate of mice that left the nest-boxes but was abandoned due to the technical limitations of radio-collating small juveniles (Chapter 4). Determining the fate of individuals who leave their natal site is a difficult problem surpassed only by determining whether a disperser is successful in securing matings. This question is of particular interest in the present study because it was assumed that mice dispersed in search of mates. Unfortunately, there was no way to determine if disappearing males and females were more reproductively successful than mice which remained in the mouse cities (see Table 3.7). For mice that disperse in search of mates, success depends primarily on the likelihood of encountering other sexually active individuals. All things being equal, a mouse that searches far and 203 wide is more likely to acquire a mate than one that does not. Dispersers which spend a great deal of time in transit or searching, however, also face a greater change of dying, particularly if their movements take individuals into unfamiliar habitat (Metzgar, 1967). In light of the risks involved in dispersal it is sometimes difficult to visualize any benefits to the disperser. Indeed, many researchers have at least implied that dispersal is more likely to benefit the pOpulation or species than the individual (Howard, 1960; Van Valen, 1971; Wynne-Edwards, 1962). These advantages could include 1) the promotion of outbreeding, 2) the spread of new genetic material, 3) the reinvasion of habitats depopulated by local extinction, 4) range extension of the species, and 5) population regulation. Opposing the above premise of group selection are those who have argued that, in spite of the risks involved, the potential benefits belong to the individual who disperses (Bekoff, 1977; Fairbairn, 1978b; Myers and Krebs, 1971; Tardif, 1979). A commonly cited benefit is that the disperser is likely to encounter more potential mates (e.g. Lidicker, 1962). Natal dispersal of juveniles may even be advantageous to the parents as well as their offspring. Mice that disperse from their natal site at puberty may increase their chances of reproducing as soon as they are physiologically able by avoiding competition with their parents or siblings (Bekoff, 1977). In an interesting paper on microtine behavior, Anderson (1980) argued that instead of favoring the juvenile that disperses, natal dispersal aids the parents whose behavior induces it. To support his argument, Anderson considered the breeding strategy of a resident adult male in relation to the breeding opportunities and dispersal of his 204 offspring. He argued that ”if the cost of inbreeding is not too high to permit a net gain,” an adult male would realize a significant reproductive gain by mating with his daughters. His sons would be capable of providing a ”jackpot" contribution to his reproductive fitness if they were to leave their natal area and breed successfully elsewhere. Therefore, rather than permitting his sons to compete for copulations in the natal area, a father would benefit by expelling his pubertal sons and mating with his daughters. According to Anderson, an adult female would be expected to benefit by expelling her offspring only if their continued presence in the natal area would threaten her resource base and her future reproductive capability. Specific predictions of Anderson's hypothesis that parents cause the dispersal of their offspring to the parents' benefit include 1) individuals will disperse at the onset of puberty, 2) males will be more likely to disperse and move farther than females, 3) aggressive combat will be rare in unconfined situations, 4) adults will expel their offspring without overt aggression, and 5) proportionately fewer offspring will disperse towards the end of the breeding season. All of these predictions have some tentative support in the literature, as well as by data from the present study. For instance, juveniles dispersed at the onset of sexual maturation. There was also no sign of physically damaging aggressive behavior between juveniles and adults. Furthermore, my data and those of King (1983) both support the prediction that the rate of juvenile dispersal towards the end of the breeding season in the fall will be less than it is in the spring and summer 0 205 Although these data support Anderson's predictions, I disagree with his contention that the major benefactor of juvenile dispersal is the parent instead of the individual that disperses. Critical to the distinction Of these two hypotheses regarding the benefactor of dispersal is whether adults actually expel offspring from the natal area, and whether they do this by overt aggression towards their offspring. Anderson's arguments address only the question of expulsion, not the mechanism. Although he argues effectively against serious combat among unconfined voles, he fails to provide any evidence that parents actually expel their offspring, albeit non-aggressively. I am not aware of anyone who has described the non-aggressive expulsion of young mice by their parents. Parent-offspring behavioral interactions of this nature are surely complex and would be difficult to measure in unconfined situations. Bekoff (1977) recommended collecting data on the patterns of social interactions among littermates and their parents throughout their early development and during dispersal to determine why certain individuals disperse. This recommendation is more suitable to the study of larger, diurnal social mammals such as canids and marmots which are amenable to the study of individual behavioral phenotypes than to the small, nocturnal cricetine rodents. In spite of these limitations, the use of nest-boxes in the present study allowed me to examine indirectly the patterns of parent-offspring interactions. Juveniles and their parents certainly did not appear to avoid interaction because they were commonly found together in the same arena or couplet and sometimes in the same nest (see also King, 1983). Although extended cohabitation between parents 206 and some of their offspring is not proof that parents did not expel others, this familial cohesion would suggest it is unlikely. One way to examine parent-offspring interactions in a nocturnal species such as Peromyscus would be to monitor the activity of parents and their offspring at a nest with the aid of radiotelemetry. A continuously recording, passive identification device (CRPID) that will be capable of recording the movements, to and from the nest, of all tagged occupants, including very small individuals, is currently being developed and tested (R. Hill, pers. comm.). By comparing the behavioral ontogeny of dispersers with 'non-dispersers, as well as comparing the nest activity patterns of their parents, we might be able to identify factors that are necessary for juvenile dispersal. Ultimately, questions concerned with whether dispersal benefits one individual or anotehr can only be answered by comparing the effect that dispersal has on the fitness Of both types of individuals. Related to the above example, for instance, does juvenile dispersal contribute more to the fitness of the disperser or to the parent which might promote it? Another question is if dispersal maximizes the fitness of an individual disperser (Gaines and McClenaghan, 1980). All too often the success of a disperser, relative to a non-disperser, is measured in terms of its survivability. Far more relevant to the fitness question are studies that are designed to explore the "fitness of dispersing individuals relative to their fitness had they not dispersed (Gaines and McClenaghan, 1980:189).” BIBLIOGRAPHY Abramsky, Z., and C. Sellah, 1982. Competition and the role of habitat selection in Gerbillus allenbyi and Meriones tristami: a removal experiment. Ecology, 63:1242-1247. Amlaner, C.J. and D.W. MacDonald, eds. 1979. A Handbook on Bio- telemetry and Radiotracking. Pergamon Press, Oxford, New York, xix + 804 pp. Anderson, P.K. 1970. Ecological structure and gene flow in small mammals. Symp. Zool. Soc. London, 16:299-325. Anderson, P.K. 1980. Evolutionary implications of microtine behavioral systems on the ecological stage. Biologist, 62:70-88. Archer, J. 1970. Effects of population density on behavior in rodents. pp. 169-210 in Social Behavior in Birds and Mammals (Crook, J.H., ed.). Academic Press, New York, 492 pp. Armitage, K.B. 1962. Social behavior of a colony of the yellow- bellied marmot (Marmota flaviventris). Anim. Behav., 10:319-331. Armitage, K.B. 1973. Population changes and social behavior following colonization by the yellow-bellied marmot. J. Mamm., 54:842-864. Armitage, K.B. 1981. Sociality as a life-history tactic of ground squirrels. Oecologia, 48:36-49. Aron, C. 1979. Mechanisms of control of the reproductive function by Olfactory stimuli in female mammals. Physiol. Rev., 59:229- 284. Ayer, M.L. and J. Mal Whitsett. 1980. Aggressive behavior of female prairie deer mice in laboratory populations. Anim. Behav., 28:763-771. Baird, D.D. and E.C. Birney. 1982. Pattern of colonization in Microtus pennsylvanicus. J. Mamm., 63:290-293. Banks, E., R. Brooks, and J. Schnell. 1974. A radiotracking study of home range and actvity of the brown lemming (Lemmus trimucronatus). J. Mamm., 56:888-901. 207 208 Barnett, S.A. 1975. The Rat: A Study in Behavior. University of Chicago Press, Chicago, xiv + 318 pp. Batzli, G.O. and F.A. Pitelka. 1971. Condition and diet of cycling populations of the California vole, Microtus californicus. J. Mamm., 52:141-163. Beacham, T.D. 1979. Size and growth characteristics of dispersing voles, Microtus townsendii. Oecologia, 42:1-10. Beacham, T.D. 1981. Some demographic aspects of dispersers in fluctuating populations of the vole Microtus townsendii. Oikos, 36:273-280. Bekoff, M. 1977. Mammalian dispersal and the ontogeny of individual behavioral phenotypes. Am. Nat., 111:715-732. Bekoff, M. 1981. Development of agonistic behavior: Ethological and ecological aspects. pp. 161-176 in Multidisciplinary Approaches to Aggression Research. Elsevier7NOrth Holland Biomedical Press. Bellamy, D. and V.J. Davies. 1971. Sexual attractants in rodent urine. J. Endocrinology, 51:xix. Bernstein, 1.8. 1981. Dominance: The baby and the bathwater. Behav. Brain Sci., 4:419-457. Blair, W.F. 1940. A study of prairie deer-mouse populations in southern Michigan. Amer. Midl. Nat., 24:273-305. Blizard, D.A. 1983. Sex differences in running-wheel behaviour in the rat: The inductive and activational effects of gonadal hormones. Anim. Behav., 31:378-384. Boag, D.A. and J.O. Murie. 1981. POpulation ecology of Columbian ground squirrels in southwestern Alberta. Can. J. 2001., 59:2230-2240. Boonstra, R. 1976. Experimental studies of the population processes in the vole Microtus townsendii. Unpubl. Ph.D. dissertation, Univ. of British Columbia. Boonstra, R. 1978. Effect of adult Townsend voles (Microtus townsendii) on survival of young. Ecology, 59:242-248. Boonstra, R. and C.J. Krebs. 1977. A fencing experiment on a high- density population of Microtus townsendii. Can. J. 2001., 55:1166-1175. Briese, L.A. and M.H. Smith. 1974. Seasonal abundance and movement of nine species of small mammals. J. Mamm., 55:615-629. Bronson, F.H. 1964. Agonstic behavior in woodchucks. Anim. Behav., 12:470-478. 209 Bronson, F.H. and C. Desjardins. 1971. Steroid hormones and aggressive behavior in mammals. pp. 43-63 in_Physiology of Aggression and Defeat (Eleftheriou, B.E. and J.P. Scott, eds.). Pergamon Press, New York. Brooks, R.J. and E.M. Banks. 1971. Radiotracking study of lemming home range. Comm. Behav. Biol., 6:1-5. Brown, J.L. 1975. The Evolution of Behavior. W.W. Norton and Co., New York, xix + 761 pp. Brown, R.E. 1979. Mammalian social odors: A critical review. pp. 104-162 in Advances in the Study of Behavior, vol. 10 (Rosenblatt, J.S., gt_§l,, eds.). Academic Press, New York. Burt, W.H. 1940. Territorial behavior and populations of some small mammals in southern Michigan. Misc. Publ. Mus. Zool., Univ. Michigan, 45:1-58. Cameron, G.N. 1977. Experimental species removal: Demographic responses by Sigmodon hispidus and Reithrodontomys fulvescens. Jo Mam. , 58 : 488-506 0 Cardinali, D.P. and M.I. Vacas. 1978. Feedback control of pineal function by reproductive hormones-a neuroendocrine paradigm. J. Neural Transm., Suppl., 13:175-201. Carl, E. 1971. Population control in arctic ground squirrels. Ecology, 52:395-413. Cheesman, C.L. and R.B. Mitson. 1982. Telemetric studies of vertebrates. Symp. Zool. Soc. London, vol. 49, Academic Press, London, 368 pp. Chitty. D. 1955. Adverse effects of population density upon the viability of later generations. pp. 57-67 i2_The Numbers of Man and Animals (Cragy, J.B. and N.W. Pirie, eds.). Oliver and Boyd, Edinburgh, vi + 152 pp. Chitty, 1960. POpulation processes in the vole and their relevance to general population theory. Canad. J. Zool., 38:99-113. Chitty, D. 1967. The natural selection of self-regulatory behaviour in animal pOpulations. Proc. Ecol. Soc. Aust., 2:51-78. Christian, J.J. 1970. Social subordination, population density, and mammalian evolution. Science, 168:84-90. Christian. J.J. 1971. Fighting, maturity, and population density in Microtus pennsylvanicus. J. Mamm., 52:556-567. Christian, J.J. and D.E. Davis. 1964. Endocrines, behavior, and population. Science, 146:1550-1560. 210 Chute, F.S., W.A. Fuller, P.R.J. Harding, and R.B. Herman. 1974. Radiotracking of small mammals using a grid of overhead wire antennas. Can. J. 2001., 52:1481-1488. Clark, F.H. 1938. Age of sexual maturity in mice of the genus Peromyscus. J. Mamm., 19:230-234. Clarke, J.R. 1956. The aggressive behavior of the vole. Behaviour, 9:1—230 Clemens, L.G. and B.A. Gladue. 1979. Neuroendocrine control of adult sexual behavior. pp. 73-103 in Review of Neuroscience, vol. 4 (Schneider, D.M., ed.). Raven Press, New York. Clemens, L.G. and S. M. Pomerantz. 1981. Male sexual behavior in deer mice (Peromyscus maniculatus) following castration and hormone replacement. Horm. Behav., 15:183-196. Clemens, L.G. and S.M. Pomerantz. 1982. Testosterone acts as a prohormone to stimulate male copulatory behavior in male deer mice (Peromyscus maniculatus bairdi). J. Comp. Physiol. Psych., 96:114-122. Crowcroft, P. and F. Rowe. 1957. The growth of confined colonies of the wild house mouse (Mus musculus LL). Proc. Royal Zool. Soc. London, 129:359-370. Dark, J., P.G. Johnston, M. Healy and I. Zucker, 1983. Latitude of origin influences photOperiodic control of reproduction of deermice (Peromyscus maniculatus). Biol. Reprod., 23:213-220. Davies, V.J. and D. Bellamy. 1972. The olfactory response of mice to urine and effects of gonadectomy. J. Endocrinology, 55:11-20. Davis, D.E., J.J. Christian, and F. Bronson. 1964. Effect of exploitation on birth, mortality and movement rates in a woodchuck population. J. Wildlf. Mgmt., 28:1-9. DeLong, K.T. 1967. Population ecology of feral house mice. Ecology, 48:611-634. Desjardins, C. 1981. Latitudinal gradients in the responsiveness of the rodent reproductive system to photic stimuli. Biol. Reprod., 24:23 Abstract. Desjardins, C. and M.J. Lopez. 1983. Environmental cues evoke differential responses in pituitary testicular function in deer mice. Endocrinology, 112:1398-1406. Desy, B.A. and C.F. Thompson. 1983. Effects of supplemental food on a Microtus pennsylvanicus population in central Illinois. J. Mime ECO]. o , 52:127-140. 211 Dewsbury, D.A. 1979. Copulatory behavior of deer mice (Peromyscus maniculatus): III. Effects on pregnancy initiation. J. Comp. Physio. Psych., 93:178-188. Dewsbury, D.A. 1982. Ejaculate cost and male choice. Am. Nat., 119:601-610. Dice, L.R. and W.H. Howard. 1951. Distance of dispersal by prairie deer mice from birth places to breeding sites. Contrib. Lab. Vertebr. Biol. Univ. Michigan, 50:1-15. Dobson, S.F. 1981. An experimental examination of an artificial dispersal sink. J. Mamm., 62:74-81. Dobson, S.F. 1982. Competition for mates and predominant juvenile male dispersal in mammals. Anim. Behav., 30:1183-1192. Doty, R.L. 1972a. Odor preferences of female (Peromyscus maniculatus) and white-footed mice (Peromyscus leucopus) to homospecific and heterospecific urine odors. J. Comp. Physiol. Psych., 84:296-303. Doty, R.L. 1972b. Odor preferences of female Peromyscus maniculatus bairdi for male mouse odors of h & bairdi and & leucopus noveboracensis as a function of estrous state. J. Comp. Physiol. Psych., 81:191-197. Doty, R.L. 1973. Reactions of deermice (Peromyscus maniculatus) and white-footed mice (Peromyscus leucopus) to homospecific and heterospecific urine Odors. J. Comp. Physiol. Psych., 84:296-303. Downhower, J.F. and K.B. Armitage. 1981. Dispersal of yearling yellow-bellied marmots (Marmota flaviventris). Anim. Behav., 29:1064-1069. Drickamer, L.C. and B.M. Vestal. 1973. Patterns of reproduction in a laboratory colony of Peromyscus. J. Mamm., 54:523-528. Dudley, D. 1974. Contributions of paternal care to the growth and development of the young in Peromyscus californicus. Behav. Biol., 11:155-166. Dueser, R.D., M.L. Wilson, and R.K. Rose. 1981. Attributes of dis- persing meadow voles in open-grid populations. Acta Theriol., 26:139-162. Dunford, C. 1977. Behavioral limitation Of round-tailed ground squirrel density. Ecology, 58:1254-1268. Edwards, D.A. 1970. Induction of estrus in female mice: Estrogen- progesterone interactions. Horm. Behav., 1:299-304. Eisenberg, J.F. 1963. The intraspecific social behavior of some cricetine rodents of the genus Peromyscus. Amer. Midl. Nat., 69:240-246. 212 Eisenberg, J.F. and D.G. Kleiman. 1972. Olfactory communication in mammals. Ann. Rev. Ecol. Syst., 3:1-32. Ellis, G.B. and F.W. Turek. 1979. Changes in locomotor activity associated with the photOperiodic response of the testes in male golden hamsters. J. Comp. Physiol., 132:277-284. Ellis, L.C. 1969. The direct action of melatonin and serotonin on testicular androgen production in_vitro. J. Reprod. Fert., 18:159 Abstract. Endler, J. 1977. Geographic Variation, Speciation, and Clines. Princeton University Press, Princeton, 246 pp. Fairbairn. D.J. 1977a. The spring decline in deer mice: Death or dispersal? Can. J. Zool., 55:84-92. Fairbairn, D.J. 1977b. Why breed early? A study of reproductive tactics in Peromyscus. Can. J. Zool., 55:862-871. Fairbairn, D.J. 1978a. Dispersal of deer mice, Peromyscus manicu- latus: Proximal causes and effects on fitness. Oecologia, 32:171-193. Fairbairn, D.J. 1978b. Behaviour of dispersing deer mice (Peromyscus maniculatus). Behav. Ecol. Sociobiol., 3:265-282. Fitch, H.S. 1948. Habits and economic relationships of the tulare kangaroo rat. J. Mamm., 29:5-35. Fitzgerald, B.M., B.J. Karl, and H. Meller. 1981. Spatial organiza- tion and ecology of a sparse population of house mice (Mus musculus) in a New Zealand forest. J. Anim. Ecol., 50:489-518. Flowerdew, J.R. 1972. The effect of supplementary food on a popu- lation of woodmice (Apodemus sylvaticus). J. Anim. Ecol., 41:537-552. Gaines, M.S., C.L. Baker, and A.M. Vivas. 1979a. Demographic attributes of dispersing southern bog lemmings (Synaptomys cooperi). Oecologia, 40:91-101. Gaines, M.S., A.M. Vivas, and C.L. Baker. 1979b. An experimental analysis of dispersal in fluctuating vole populations: Demographic parameters. Ecology, 60:814-828. Gaines, M.S. and L.R. McClenaghan, Jr. 1980. Dispersal in small mammals. Ann. Rev. Ecol. Syst., 11:163-196. Garten, C.T. and M.H. Smith. 1974. MOvement by Oldfield mice and population regulation. Acta Theriol., 19:513-514. 213 Gentry, J.B. 1966. Invasion of a one-year abandoned field by Peromyscus polionotus and Mus musculus. J. Mamm., 47:431-439. Gilbert, B.S. and C.J. Krebs. 1981. Effects of extra food on Peromyscus and Clethrionomys populations in the southern Yukon. Oecologia, 51:326-331. Gill, J.L. 1978. Design and Analysis of Experiments in the Animal and Medical Sciences. Vol. 1. Iowa State University Press, Ames, xiii + 409 pp. Gipps, J.H.W. 1982. The effects of testosterone and scopolamine HBr on the aggressive behavior of male voles, Microtus townsendii. Canad. J. Zool., 60:946-950. Gipps, J.H.W. and P.A. Jewell. 1979. Maintaining pOpulations of bank voles, Clethrionomys glareolus, in large outdoor enclosures, and measuring the response of population variables to the castration of males. J. Anim. Ecol., 48:535-556. Gipps, J.H.W., M.J. Taitt, C.J. Krebs, and Z. Dundjerski. 1981. Male aggression and the population dynamics of the vole, Microtus townsendii. Canad. J. Zool., 59:147-158. Glass, J.D. and G.R. Lynch. 1982. Evidence for a brain site of melatonin action in the white-footed mouse, Peromyscus leucopus. Neuroendocrinology, 34:1-6. Gleason, P.E., S.D. Michael, and J.J. Christian. 1981. Prolactin- induced aggression in female Peromyscus leucopus. Behav. Neural Biol., 33:243-248. Godfrey, G.R. 1954. Tracing field voles (Microtus agrestis) with a Geiger-Muller counter. Ecology, 35:5-10. Goldman, E., V. Hall, C. Hollister, P. Roychoudhury, L. Tamarkin, and W. Westrom. 1979. Effects of melatonin on the reproduc- tive system in intact and pinealectomized male hamsters maintained under various photoperiods. Endocrinology, 104:82-88. Gosling, L.M. 1982. A reassessment of the function of scent marking in territories. Z. Tierpsychol., 60:89-118. Gottfried, B.M. 1979. Small mammal populations in woodlot islands. Am. Midl. Nate, 102:105-1120 Gottfried, B.M. 1982. A seasonal analysis of small mammal populations in woodlot islands. Canad. J. Zool., 60:1660-1664. Goulet, L.A. 1979. Aspects of population dynamics and social behaviour in the Richardson's ground squirrel as modified by a chemosterilant (mestranol). Unpubl. Ph.D. dissertation, Simon Fraser University, British Columbia. 214 Grant, P.R. 1978. Dispersal in relation to carrying capacity. Proc. Natlo Acado SCio, 75:2854-28580 Grau, H.J. 1982. Kin recognition in white-footed deermice (Peromyscus leucopus). Anim. Behav., 30:497-505. Grocock, C.A. and J.R. Clarke. 1974. Photoperiodic control of testis activity in the vole, Microtus agrestis. J. Reprod. Fert., 39:337-347. Haigh, G.R. 1983. Effects of inbreeding and social factors on the reproduction of young female Peromyscus maniculatus bairdii. Jo Ham. , 64:48-54. Halpin, Z.T. 1981. Adult-young interactions in island and mainland populations of the deermouse Peromyscus maniculatus. Oecologia, 51:419-425. Hamley, J.M. and J.B. Falls. 1975. Reduced activity in transmitter- carrying voles. Canad. J. Zool., 53:1476-1478. Hansen, L. and G.O. Batzli. 1978. The influence of food availability on the white-footed mouse: Populations in isolated woodlots. Harland, R.M. and J.S. Miller. 1980. Activity of breeding Peromyscus leucopus. Canad. J. Zool., 58:313-316. Hayne, D.W. 1978. Experimental design and statistical analyses. pp. 3-13 in POpulations of Small Mammals Under Natural Conditions (Snyder, D.P., ed.). Univ. Pittsburgh Pymatuning Lab. Ecol. Spec. Publ. Ser., vol. 5. Healey, M.C. 1967. Aggression and self-regulation of population size in deer mice. Ecology, 48:377-392. Hendel, R.C. and F.W. Turek, 1978. Suppression of locomotor activity in sparrows by treatment with melatonin. Physiol. Behav., 21:275-278. Herman, T.B. 1977. Activity patterns and movements of subarctic voles. Oikos, 29:434-444. Hilborn, R. 1975. Similarities in dispersal tendency among siblings in four species of voles (Microtus). Ecology, 56:1221-125. Hilborn, R. and C.J. Krebs. 1976. Fates of disappearing individuals in fluctuating populations of Microtus townsendii. Canad. J. Zool., 54:1507-1518. Hill, J.L. 1974. Peromyscus: effect of early pairing on reproduction. Science, 186:1042-1044. 215 Hill, J.L. 1977. Space utilization of Peromyscus: Social and spatial factors. Anim. Behav., 25:373-389. Hill, R.W. 1972. The amount of maternal care in Peromyscus leucopus and its thermal significance for the young. J. Mamm., 53:774-790. Hoffmann, K. 1973. The influence of photOperiod and melatonin on testis size, body weight and pelage colour in the Djungarian hamster (Phodopus sungorus). J. Comp. Physiol., 85:267-282. Holekamp, K.E. 1982. Natal dispersal in Belding's ground squirrels: Its endocrinological causes. Paper presented at a conference on sociality in ground squirrels. 30 September-2 October 1982, Banff, Alberta, Canada. Horner, E. 1947. Parental cart of young mice of the genus Peromyscus. J. Mamm., 28:31. Howard, W.E. 1949. Dispersal, amount of inbreeding, and longevity in a local population of deer mice on the George Reserve, southern Michigan. Contrib. Lab. Vertebr. Biol. Univ. Michigan, 43:1-50. Howard, W.E. 1960. Innate and environmental dispersal of individual vertebrates. Am. Midl. Nat., 63:152-161. Howard, W.E. and F.C. Evans. 1961. Seeds stored by prairie deer mice. Jo MaMo , 42:260‘263o Howard, W.E. and R.E. Marsh. 1969. Mestranol as a reproductive inhibitor in rats and voles. J. Wildlf. Mgmt., 33:403-408. Jameson, E.W., Jr. 1950. Determining fecundity in male small mammals. J. Mamm., 31:433-436. Johnston, P.G. and I. Zucker. 1980a. Antigonadal effects of mela- tonin in white-footed mice (Peromyscus leucopus). Biol. Reprod., 23:1069-1074. Johnston, P.G. and I. Zucker. 1980b. Photoperiodic regulation of the testes of adult white-footed mice (Peromyscus leucopus). Biol. Reprod., 23:859-866. Johnston, P.G. and I. Zucker. 1980c. Photoperiodic regulation of reproductive development in white-footed mice (Peromyscus leucopus). Biol. Reprod., 22:983-989. Joule, J. and G.N. Cameron. 1975. Species removal studies. I. Dispersal strategies of sympatric Sigmodon hispidus and Reithrodontomys fulvescens populations. J. Mamm., 56:378-396. Kao, L.W.L. and J. Weisz. 1977. Release of gonadotrophin-releasing hormone (GnRH) from isolated, perfused medial basal hypothalamus by melatonin. Endocrinology, 100:1723-1726. 216 Kareem, A.M. and C.J. Barnard. 1982. The importance of kinship and familiarity in social interactions between mice. Anim. Behav., 30:594-601. Keith, T.P. and R.E. Tamarin. 1981. Genetic and demographic differences between dispersers and residents in cycling and non-cycling vole populations. J. Mamm., 62:713-725. Kemp, C.A. and L.B. Keith. 1970. Dynamics and regulation of red squirrel (Tamiasciurus hudsonicus) populations. Ecology, 51:763-779. King, J.A. 1955. Social behavior, social organization and population dynamics in a black-tailed prairie dog town in the Black Hills of South Dakota. Contrib. Lab. Vertebr. Biol. Univ. Michigan, 48:1-123o . King, J.A. 1973. The ecology of aggressive behavior. Ann. Rev. Ecol. Syst., 4:117-137. King, J.A. 1983. Seasonal dispersal in a semi-natural population of Peromyscus maniculatus. Canad. J. Zool., 61:in press. Kozakiewicz, M. 1976. Migratory tendencies in a population of bank voles and a description of migrants. Acta Theriol., 21:321-338. Kozel, R.M. and E.D. Fleharty. 1979. Movements of rodents across roads. Southwest. Nat., 24:239-248. Krebs, C.J. 1964. The lemming cycle at Baker Lake, Northwest Territories, during 1959-1962. Arct. Inst. North Am. Tech. Pap., 15. 104 pp. Krebs, C.J. 1966. Demographic changes in fluctuating populations of Microtus californicus. Ecol. Mbnogr., 36:239-273. Krebs, C.J. 1970. Microtus population biology: Behavioral changes associated with the population cycle in M; ochrogaster and 'M; pennsylvanicus. Ecology, 51:34-52. Krebs, C.J. 1978a. A review of the Chitty Hypothesis of population regUIationo cano Jo ZOOIo, 56:2463-2480o Krebs, C.J. 1978b. Aggression, dispersal, and cyclic changes in populations of small mammals. pp. 49-60 in_Aggression, Dominance, and Individual Spacing. (Krames, L., P. Pliner, and T. Alloway, eds.). Plenum, New York. Krebs, C.J., Z.T. Halpin, and J.N.M. Smith. 1977. Aggression, test- osterone, and the spring decline in populations of the vole, Microtus townsendii. Canad. J. Zool., 55:430-437. 217 Krebs, C.J., B. L. Keller, and R.B. Tamarin. 1969. Microtus population biology: Demographic changes in fluctuating populations of M. ochrogaster and M;_pennsylvanicus in southern Indiana. Ecology, 50:587-607. Krebs, C.J. and J.H. Myers. 1974. Population cycles in small mammals. Adv. Ecol. Res., 8:267-399. Krebs, C.J., J. A. Redfield, and M.J. Taitt. 1978. A pulsed-removal experiment on the vole Microtus townsendii. Canad. J. Zool., 56:2253-2262. Krebs, C.J., I. Wingate, J. LeDuc, J.A. Redfield, M. Taitt, and R. Kilborn. 1976. Microtus population biology: Dispersal in fluctuating populations of M; townsendii. Canad. J. Zool., 54:79-95. Larsson, K. 1979. Features of the neuroendocrine regulation of male sexual behavior. pp. 77-163'in_Endocrine Control of Sexual Behavior (Beyer, 0., ed.). Raven Press, New York. Layne, J.N. 1968. Ontogeny. pp. 148-253 i2_Biology of Peromyscus (Rodentia) (King, J.A., ed.). Spec. Publ., Amer. Soc. Mamm., Vol. 2. xiii + 593 pp. Lidicker, W.Z. Jr. 1962. Emigration as a possible mechanism permitting the regulation of population density below carrying capacity. Amer. Nat., 96:23-29. Lidicker, W.Z. Jr. 1973. Regulation of numbers in an island population of the California vole, a problem in community dynamics. Ecol. Monogr., 43:271-302. Lidicker. W.E. Jr. 1975. The role of dispersal in the demography of small mammal populations. pp. 103-128 in Small Mammals: Their Productivity and Population Dynamics (Petruscewicz, K., F.B. Golley, and L. Ryszkowski, eds.). Cambridge Univ. Press, New York, xxv + 451 pg. Lidicker, W. Z. Jr. 1976. Experimental manipulation of the timing of reproduction in the California vole. Res. Pop. Ecol., 18:14-27. Lidicker. W.Z. Jr. 1980. The social biology of the California vole. Biologist, 62:46-55. Linduska, J.P. 1950. Ecology and land-use relationships of small mammals on a Michigan farm. Mich. Dept. Cons., Fed. Aid Project 2-R, ix + 144 pp. Lloyd, J.A. 1980. Interaction of social structure and reproduction in populations of mice. pp. 3-22 in Biosocial Mechanisms of Population Reguoation (Cohen, M.D., R.S. Malpass, and H.G. Klein, eds.). Yale university Press, New Haven, London, xxiii + 406 pp. 218 Lloyd, J.A. and J.J. Christian. 1967. Relationship of activity and aggression to density in two confined pOpulations of house mice (Mus musculus). J. Mamm., 48:262-269. Lombardi, J.R. and J.M. Whitsett. 1980. Effects of urine from con- specifics on sexual maturation in female prairie deer mice, Peromyscus maniculatus bairdi. J. Mamm., 61:766-768. Lorenz, K. 1973. On Aggression. Methuen, London, 273 pp. Lynch, G.R. 1973. Seasonal changes in thermogenesis, organ weights, and body composition in the white-footed mouse, Peromyscus leucopus. Physiol. Zool., 51:280-299. Lynch, G.R. and A.L. Epstein. 1976. Melatonin induced changes in gonads, pelage, and thermogenic characters in the white-footed mouse (Peromyscus leucopus). comp. Biochem. Physiol., 530:67-69. Lynch, G.R., F.D. Vogt, and H.R. Smith. 1978. Seasonal study of spontaneous daily torpor in the white-footed mouse, Peromyscus leucopus. Physiol. Zool., 51:289-299. Madison, D.M. 1977. Movements and habitat use among interacting Peromyscus leucopus as revealed by radiotelemetry. Canad. Field-Nat., 91:273-281. Madison, D.M. 1978. Behavioral and sociochemical susceptibility of meadow voles (Microtus pennsylvanicus) to snake predators. Amer. Midlo Nato, 100323-28o Madison, D.M. 1980a. An integrated view of the sociobiology of Microtus pennsylvanicus. Biologist, 62:20-33. Madison, D.M. 1980b. Space use and social structure in meadow voles, Microtus pennsylvanicus. Behav. Ecol. Sociobiol., 7:65-71. Madison, D.M. 1980c. Movement types and weather correlates in free-ranging meadow voles. pp. 34-42 in Proceedings of the Fourth Eastern Pine and Meadow VOle Symposium (Beyers, R.E., ed.). Hendersonville, N.C. Mares, M.A., T.E. Lacher Jr., M.R. Willig, N.A. Bitar, R. Adams, A. Klinger, and D. Tazik. 1982. An experimental analysis of social spacing in Tamias striatus. Ecology, 63:267-278. Margolis, D.J. and G.R. Lynch. 1981. Effects of daily melatonin injections on female reproduction in the white-footed mouse, Peromyscus leucopus. Gen. Comp. Endocrinol., 44:530-537. Mather, J.G. 1981. Wheel-running activity: A new interpretation. Mammal Rev., 11:41-51. Mazdzer, E., M.R. Capone, and L.C. Drickamer, 1976. Conspecific odors and trappability of deer mice (Peromyscus leucopus noveboracensis). J. Mamm., 57:607-609. 219 Mazurkiewicz, M. and E. Rajska. 1975. Dispersion of young bank voles from their place of birth. Acta Theriol., 20:71-81. McLean, I.G. 1982. The association of female kin in the arctic ground squirrel Spermophilus parryi. Behav. Ecol. Sociobiol., 10:91-99. Metzger, L.H. 1967. An experimental comparison of screech owl predation on resident and transient white-footed mice (Peromyscus leucopus). J. Mamm., 48:387-391. Michener, G.R. and D.R. Michener. 1977. Population structure and dispersal in Richardson's ground squirrels. Ecology, 58:359-368. Mihok, S. 1981. Chitty's hypothesis and behavior in subarctic red- backed voles Clethrionomys gapperi. Oikos, 36:281-295. Millar, J.S., F.B. Wille, and S.L. Iverson. 1979. Breeding by Peromyscus in seasonal environments. Canad. J. Zool., 57:719-727. Mineau, P. and D. Madison. 1977. Radio-tracking of Peromyscus leucopus. Canad. J. Zool., 55:465-468. Moore, R.E. 1962. Olfactory discrimination as an isolating mechanism between.§§_gh_fuginus and P; polionotus leucocephalus. Unpubl. Ph.D. dissertation, Uhiv. of Texas, Austin. Moore, R.E. 1965. Olfactory discrimination as an isolating mechanism between Peromyscus maniculatus and Peromyscus polionotus. Amer. Midlo Nato , 73:85-100o Morali, G. and C. Beyer. 1979. Neuroendocrine control of mammalian estrous behavior. pp. 33-60 ig_Endocrine Control of Sexual Behavior (Beyer, C., ed.). Raven Press, New York. Morin, L.P., K.M. Fitzgerald, and I. Zucker. 1977. Estradiol shortens the period of hamster circadian rhythms. Science, 196:305-307. Murphy, K.L. and J.L. Gidner, 1982. Patterns of activity, space use, and nest cohabitation within a semi-natural population of deer mice. Paper presented at the 62nd Annual Meeting of the American Society of Mammologists, 20-24 June, Snowbird, Utah. Myers, J.H. 1974. Genetic and social structure of feral house mouse populations on Grizzly Island, California. Ecology, 55:747-759. Myers, J.H. and C.J. Krebs. 1971. Genetic, behavioral and reproductive attributes of dispersing field voles Microtus pennsylvanicus and Microtus ochrogaster. Ecol. Monogr., 41:53-78. Myers, P. and L.L. Master. 1983. Reproduction of Peromyscus maniculatus: Size and compromise. J. Mamm., 64:1-18. 220 Myllymaki, A. 1977. Intraspecific competition and home range dynamics in the field vole Microtus agrestis. Oikos, 29:553-569. Myllymaki, A., A. Paasikallio, and U. Hakkinen. 1971. Analysis of a ”standard trapping” of Microtus agrestis (L.) with triple isotope marking outside the quadrat. Ann. Zool. Fenn., 8:22-34. Nadeau, J.H., R.T. Lombardi, and R.H. Tamarin. 1981. Population structure and dispersal of Peromyscus leucopus on Muskeget Island. Newsome, A.E. 1969. A population study of house mice temporarily inhabiting a south Australian wheatfield. J. Anim., Ecol., 38:341-359. Nicholson, A.J. 1941. The homes and social habits of the wood mouse (Peromyscus leuc0pus noveboracensis) in southern Michigan. Amer. Midlo Nato, 25:196—223o Pedersen, T. and H. Peters. 1968. Proposal for a classification of ocytes and follicles in the mouse ovary. J. Reprod. Fert., 17:555-557. Petterborg, L.S. and R.J. Reiter. 1980. Effect of photoperiod and melatonin on testicular deve10pment in the white-footed mouse, Peromyscus leucopus. J. Reprod. Fert., 60:209-212. Petticrew, B.G. and R.M.F.S. Sadlier. 1974. The ecology of the deer mouse Peromyscus maniculatus in a coastal coniferous forest. I. population dynamics. Canad. J. Zool., 52:107-118. Pfeiffer, S. 1982. Disappearance and dispersal of Spermophilus elegans juveniles in relation to behavior. Behav. Ecol. Sociobiol., 10:237-243. Pickering, J., L.L. Getz, and C.S. Whitt. 1974. An esterase phenotype correlated with dispersal in Microtus. Trans. 111. State Acad. Sci., 67:471-475. Pomerantz, S.M., E. Fox, and L.G. Clemens. 1983. Gonadal hormone activation of male courtship ultrasonic vocalizations and male coulatory behavior in castrated male deer mice (Peromyscus maniculatus bairdi). Behav. Neurosci., 97:462-469. Potvin, N., J.M. Bergeron, M. Norman, and A. Cyr. 1982. Evaluating the sterile male method on red-winged blackbirds: clinical evaluation of thiotepa as a sterilant. Can. J. Zool., 60:460-465. Pucek, A. and J. Olszewski. 1971. Results of extended removal catches of rodents. Ann. 2001. Fenn., 8:37-44. Pucek, Z. and W.P.W. Lowe. 1975. Age criteria in small mammals. pp. 52-72 in Small Mammals: Their Productivity and Population Dynamics (Golley, F.B., K. Petrusewicz, and L. Ryszkowski, eds.). Cambridge University Press, New York, xxv + 451 pp. 221 Rajska-Surgiel, E. 1976. Interactions between individuals of a pOpulation of the bank vole, Clethrionomys glareolus (Schreber, 1980)o EKGlo POlo, 24:3-35o Randolph, S.E. 1977. Changing spatial relationships in a population of Apodemus sylvaticus with the onset of breeding. J. Anim. ECOlo, 46:653-676o Redfield, J.A., M.J. Taitt, and C.J. Krebs. 1978. Experimental alteration of sex ratios in populations of Microtus townsendii, a field vole. Canad. J. Zool., 56:17-27. Reich, L.M., K.M. Wood, B.E. Rothstein, and R.H. Tamarin. 1982. Aggressive behavior of male Microtus breweri and its demographic implications. Anim. Behav., 30:117-122. Reiter, R.J. 1974. Circannual reproductive rhythms in mammals related to photoperiod and pineal function: A review. Chronobiologia, 1:365-395. Reiter, R.J., M.D. Rollag, E.S. Panke, and A.F. Banks. 1978. Melatonin: Reproductive effects. J. Neural Trans., Supple., 13:209-223. Riggs, L.A. 1979. Experimental studies of dispersal in the California vole, Microtus californicus. Unpubl. Ph.D. dissertation, Univ. of California, Berkeley, 236 pp. Rintamaa, D.L., P.A. Mazur, and S.H. Vessey. 1976. Reproduction during two annual cycles in a population of Peromyscus leucopus noveboracensis. J. Mamm., 57:593-595. Robinson, W.L. and J.B. Falls. 1965. A study of homing of meadow miceo Amo Midlo Nato, 73:188-224o Rogers, J.G. Jr. and G.K. Beauchamp. 1974. Relationships among three criteria of puberty in Peromyscus leucopus noveboracensis. Jo MaMo , 55:461‘462o Rongstadt, O.J. 1965. A life history study of thirteen-lined ground squirrels in southern Wisconsin. J. Mamm., 46:76-87. Rose, R.K. and M.S. Gaines. 1976. Levels of aggression in fluctuating populations of the prairie vole, Microtus ochrogaster in eastern Kansas. J. Mamm., 57:43-57. Rowe, F.P., B.J. Taylor, and A.H.J. Chudley. 1964. The numbers and movements of house mice (Mus musculus L.) in the vicinity of four corn ricks. J. anim. Ecol., 32:87-97. Rowley, M.H. and J.J. Christian. 1976. Intraspecific aggression of Peromyscus leucopus. Behav. Biol., 17:249-253. 222 Rudel, H.W. and F.A. Kincl, 1966. The biology of antifertility steroids. Acta Endocrinologica, 51 (Suppl. 105):1-45. Rusak, B. and L.P. MOrin. 1976. Testicular responses to photoperiod are blocked by lesions of the suprachiasmatic nuclei in golden hamsters. Biol. Reprod., 15:366-374. Rusch, D.A. and W.G. Reeder. 1978. Population ecology of Alberta red squirrels. Ecology, 59:400-420. Rust, C.C. and R.K. Meyer. 1969. Hair color, moult, and testis size in male short-tailed weasels treated with melatonin. Science, 165:921-922. Sadlier, R.M.F.S. 1965. The relationship between agonistic behavior and population changes in the deer mouse, Peromyscus maniculatus (Wagner). J. Anim. Ecol., 34:331-352. Sadlier, R.M.F.S. 1969. The Ecology and Reproduction in Wild and Domestic Mammals. Methuen, London, xii + 321 pp. Satterthwaite, F.E. 1941. Synthesis of variance. Psychometrika, 6:309-316. Savidge, I.R. 1974. Social factors in dispersal of deer mice (Peromyscus maniculatus). Amer. Midl. Nat., 91:395-405. Schaefer, V.H. 1982. Movement and diet activity of the coast mole Scapanus orarius True. Cad. J. Zool., 60:480-482. Schafer, E.W. Jr., R.B. Brunton, and N.F. Lockyer. 1976. Evaluation of 45 chemicals as chemosterilants in adult male quail (Coturnix coturnix). J. Reprod. Fert., 48:371-375. Shields, L.J. 1976. Telemetric determination of the activity of free- ranging rodents: The fine structure of Microtus californicus activity patterns. Unpubl. Ph.D. dissertation, Uhiv. of California, Los Angeles, California. Slade, N.A. and D.P. Balph. 1974. Population ecology of Uinta ground squirrels. Ecology, 55:989-1003. Smith, M.H. 1971. Food as a limiting factor in the population ecology of Peromyscus polionotus (Wagner). Ann. 2001. Fenn., 8:109-112. Sokal, R.R. and F.J. Rohlf. 1969. Biometry. W.H. Freeman, San Francisco, xxi + 776 pp. Southwick, C.H. 1958. Population characteristics of house mice living in English corn ricks: Density relationships. Proc. 2001. Soc. London, 131:163-175. Speed, F.M. and R.R. Hocking. 1976. The use of the R( )- notation with unbalanced data. Amer. Stat., 30:30-33. 223 Spevak, T.A. 1981. Breeding structure in semi-natural populations of Peromyscus maniculatus bairdii. Paper presented at the 6lst Annual Meeting of the American Society of Mammalogists, 7-11 June, Oxford, Ohio. Stenseth, N.C. 1978. Demographic strategies in fluctuating populations of small rodents. Oecologia, 33:149-172. Stenseth, N.C. 1981. On Chitty's theory for fluctuating populations: Importance of genetic polymorphism in the generation of regular density cycles. J. Theoret. Biol., 90:9-36. Stickel, L. 1946. The source of animals moving into a depopulated area. J. Mamm., 27:301-307. Stout, I.J. and R.J. Demmer. 1982. Cotton rat invasion of sand pine scrub habitat. J. Mamm., 63:236-242. Strecker, R.C. 1954. Regulatory mechanisms in house mouse popula- tions: The effect of limited food supply on an unconfined population. Ecology, 35:249-253. Sullivan, T.P. 1977. Demography and dispersal in island and mainland populations of the deer mouse, Peromyscus maniculatus. Ecology, 58:964-978. Sullivan, T.P. and D.S. Sullivan. 1982. Population dynamics and regulation of the Douglas squirrel (Tamiasciurus douglasii) with supplemental food. Oecologia, 53:264-270. Svendsen, G.E. 1974. Behavioral and environmental factors in the spatial distribution and population dynamics of a yellow-bellied marmot population. Ecology, 55:760-771. Taitt, M.J. 1981. The effect of extra food on small rodent popula- tions: 1. Deer mice (Peromyscus maniculatus). J. Anim. Ecol., 50:111-124. Taitt, M.J., J.B. Gipps, C.J. Krebs, and Z. Dundjerski. 1981. The effect of extra food and cover on declining populations of Microtus townsendii. Canad. J. Zool., 59:1593-1599. Taitt, M.J. and C.J. Krebs. 1981. The effect of extra food on small rodent populations: II. Voles (Microtus townsendii). J. Anim. Ecol., 50:125-137. Taitt, M.J. and C.J. Krebs. 1982. Manipulation of female behavior in field populations of Microtus townsendii. J. Anim. Ecol., 51:681-690. Tamarin, R.H. 1977. Dispersal in island and mainland voles. Ecology, 58:1044-1054. 224 Tamarin, R.H. 1980. Dispersal and population regulation in rodents. pp. 117-134 in Biosocial Mechanisms of POpulation Regulation (Cohen, M.N., R.S. Malpass, and H.G. Klein, eds.). Yale University Press, New Haven, London, xxiii + 406 pp. Tamarkin, L., W.K. Westrom, A.I. Hamill, and B.D. Goldman. 1976. Effect of melatonin on the reproductive systems of male and female Syrian hamsters: A diurnal rhythm in sensitivity to melatonin. Endocrinology, 99:1534-1541. Tardif, R.R. 1979. Dispersal of Peromyscus leucopus within and between woodlots. Unpubl. Ph.D. dissertation, Michigan State Univ., East Lansing, Michigan, 112 pp. Tardif, R.R. and L. Gray. 1978. Feeding diversity of resident and immigrant Peromyscus leucopus. J. Mamm., 59:559-562. Terman, G.R. 1961. Some dynamics of spatial distribution within semi-natural populations of prairie deer mice. Ecology, 42:288-302. Trobec, R.J. and L.W. Oring. 1972. Effects of testosterone proppri- onate implantation on lek behavior of sharp-tailed grouse. Amer. Midlo Nato, 87:531-536o Trudeau, A.M., G.R. Haigh, and S.H. Vessey. 1980. The use of nest boxes to study the behavioral ecology of Peromyscus leucopus. Ohio J. Sci., 80:91 Abstract. Turek, F.W., C. Desjardins, and M. Menaker. 1976. Differential effects of melatonin on the testes of photOperiodic and non- photoperiodic rodents. Biol. Reprod., 15:94-97. Turner, B.N. and S.L. Iverson. 1973. The annual cycle of aggression in male Microtus pennsylvanicus and its relation to population parameters. Ecology, 54:967-981. Turner, B.N., S.L. Iverson, and K. L. Severson. 1980. Effects of castration on Open-field behavior and aggression in male meadow voles (Microtus pennsylvanicus). Canad. J. Zool., 58:1927-1932. Turner, C.D. and J.T. Bagnara. 1976. General Endocrinology, W.B. Saunders, Philadelphia, x + 596 pp. Vandenbergh, J.G. 1975. Hormones, pheromones, and behavior. pp. 551-584 in Hormonal Correlates of Behavior, vol. 2 (Eleftheroiu, B.E. and R.L. Sprott, eds.). Plenum Press, New York, xvii + 806 pp. Van Valen, L. 1971. Group selection and the evolution of dispersal. Evolution, 25:591-598. Van Vleck, D.B. 1968. Movements of Microtus pennsylvanicus in relation to depopulated areas. J. Mamm., 49:92-103. 225 Verner, L. 1979. The significance of dispersal in fluctuating populations of Microtus ochgrogaster and M; pennsylvanicus. Unpubl. Ph.D. dissertation, Univ. of Illinois, Urbana, 86 pp. Vestal, B.M. and J.J. Hellack. 1977. Effects of available space on social interactions in male white-footed mice (Peromyscus leucopus). Behav. Biol., 19:289-299. Watson, A. 1970. Territorial and reproductive behavior of red grouse. J. Reprod. Fert., 11 (Suppl.):3—l4. Watts, C.H.S. 1970. Long distance movement of bank voles and wood mice. J. Zool. Lond., 161:247-256. Webster, A.E. and R.J. Brooks. 1980. Effects of radiotransmitters on the meadow vole, Microtus pennsylvanicus. Canad. J. Zool., 58:997-1001. Webster, A.E. and R.J. Brooks. 1981a. Daily movements and short activity periods of free-ranging meadow voles, Microtus pennsylvanicus. Oikos, 37:80-87. Webster, A.B. and R.J. Brooks. 1981b. Social behavior of Microtus pennsylvanicus in relation to seasonal changes in demography. Jo Mammo , 62:738-751o Whitaker, J.O. Jr. 1968. Parasites. pp. 254-311 in Biology of Peromyscus (Rodentia) (King, J.A., ed.). Spec. Publ., Amer. soc. Mamm., Vol. 2. xiii + 593 pp. Whitsett, J.M. and A.D. Lawton. 1982. Social stimulation of reproductive development in male deer mice housed on a short- day photoperiod. J. Comp. Physiol. Psych., 96:416-422. Whitsett, J.M., A.D. Lawton, and L.L. Miller. 1983. Daylength and pubertal development in male deer mice: Photosensitive stages. Submitted for publication. Whitsett, J.M., A.D. Lawton, and L.L. Miller. 198 . Daylength and pubertal development in male deer mice: Photosensitive stages. Whitsett, J.M., L.E. Gray, Jr., and G.M. Bediz. 1979. Gonadal hormones and aggression toward juvenile conspecifics in prairie deer mice. Behav. Ecol. Sociobiol., 6:165-168. Whitsett, J.M. and L.L. Miller. 1982. Photoperiod and repro- duction in female deer mice. Bio. Reprod. 26:296-304. Wiger, R. 1982. Roles of self-regulatory mechanisms in cyclic populations of Clethrionomys with special reference to .9; glareolus: A hypothesis. Oikos, 38:60-71. 226 Wolff, J.O. and B. Hurlbutt. 1982. Day refuges of Peromyscus leucopus and Peromyscus maniculatus. J. Mamm., 63:666-668. Wolff, J.O. and D.F. Holleman. 1978. Use of radiosotope labels to establish genetic relationships in free-ranging small mammals. Jo Mano , 59:859-860o Wong, R. and 0.3.0. Whiteside. 1968. The effect of melatonin on the wheel-running activity of rats deprived of food. J. Endocrinol., 40:383-384. Wynne-Edwards, V.C. 1962. Animal Dispersion in Relation to Social Behaviour. Oliver and Boyd, Edinburgh, 653 pp. Yeaton, R.I. 1972. Social behavior and social organizations in Richardson's ground squirrel (Spermophilus richardsonii) in Saskatchewan. J. Mamm., 53:139-147. Ziensis, J.S., D.E. Davis, and D.E. Smith. 1975. Diel variations in the aggressive behaviour of the mouse, Mus musculus. Anim. Behav., 23:941-948. Zucker, I., P.G. Johnston, and D. Frost. 1980. Comparative, physiol- ogical and biochronometric analyses of rodent seasonal reproductive cycles. Prog. Reprod. Biol., 5:102-133.