UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Effects of habitat and food on demographic classes and population dynamics of a habitat specialist, the… Galindo Leal, Carlos 1991

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-UBC_1991_A1 G33.pdf [ 6.65MB ]
Metadata
JSON: 831-1.0100480.json
JSON-LD: 831-1.0100480-ld.json
RDF/XML (Pretty): 831-1.0100480-rdf.xml
RDF/JSON: 831-1.0100480-rdf.json
Turtle: 831-1.0100480-turtle.txt
N-Triples: 831-1.0100480-rdf-ntriples.txt
Original Record: 831-1.0100480-source.json
Full Text
831-1.0100480-fulltext.txt
Citation
831-1.0100480.ris

Full Text

E F F E C T S O F H A B I T A T A N D F O O D O N D E M O G R A P H I C C L A S S E S A N D P O P U L A T I O N D Y N A M I C S O F A H A B I T A T S P E C I A L I S T , T H E R O C K M O U S E . by CARLOS GALINDO-LEAL B.Sc, Universidad Autonoma Metropolitana, Mexico, 1979 M.Sc, University of British Columbia, 1984 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE STUDIES DEPARTMENT OF ZOOLOGY We accept this thesis as conforming to the required standard THE UNIVERSITY OF BRITISH COLUMBIA October, 1991 C) Carlos Galindo-Leal, 1991 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of The University of British Columbia Vancouver, Canada Date DE-6 (2/88) A B S T R A C T I examined the relations between habitat structure, population var iabi l i ty and patterns of microhabitat use by demographic classes (sex, age, residency) i n the rock mouse i n southern Durango, Mexico. I also examined indiv idual and population responses to experimental changes i n food abundance, and the consequences for the prevalence of botfly parasit ism. I used a gradient of habitats i n manzanita-oak shrubland to analyze demographic var iabi l i ty and microhabitat use patterns. If demographic parameters of habitat specialists are closely associated wi th habitat structure, I expected populations to be more s imi lar i n habitats w i th matching habitat structure. Demographic characteristics were more dissimi lar i n those grids wi th the greatest differences i n vegetation characteristics. I also tested the hypothesis that populations w i th higher breeding densities have higher adult survival , lower recruitment and higher stabil ity. The areas wi th highest breeding densities had low juvenile and subadult recruitment i n the breeding season. Both were relatively stable dur ing the first year of study, but one declined to extinction dur ing the second year. I examined microhabitat use by individuals to test the hypothesis that demographic classes differ i n their spatial distr ibution, part icular ly dur ing the breeding season. There were differences i n microhabitat use among sexes and ages, as well as among resident and transient indiv iduals, part icularly i n the breeding season. ii I conducted two short-term (4 mo) supplementary food addition experiments during breeding and non-breeding seasons to test the hypothesis that females are more responsive to food resources than males, especially during reproduction. In two grids I added food on relatively widely spaced point sources to compare the effects within grids on fed and unfed subpopulations. I also compared the populations on treatment grids and on a control grid. Among fed subpopulations females responded more intensely to food additions than males. They gained weight and had improved reproduction and improved survival, during both wet and dry seasons. Males responded less consistently than females. There were also substantial effects at the population level. Populations on both experimental grids increased as a result of the food addition. Reproduction and adult recruitment also improved. Lastly, I tested the hypothesis that botfly infestation is male-biased and related to increased movement. Sexes differed seasonally in infestation rates. During both years adult females were heavily infested in the fall, whereas adult males were mainly infested in winter. Infestation rates were negatively related to motility. Females had higher infestation in the breeding season when they move less, whereas males had higher infestation in the non-breeding season. In the experimental grids higher infestation rates occurred among individuals, particularly females, who used traps near food stations. iii T A B L E OF C O N T E N T S ABSTRACT ii T A B L E O F CONTENTS iv LIST O F T A B L E S vi LIST OF FIGURES vii A C K N O W L E D G E M E N T S ax Chapter 1 GENERAL INTRODUCTION The rock mouse 3 Organization of the thesis 3 Chapter 2 HABITAT STRUCTURE AND DEMOGRAPHIC VARIABILITY INTRODUCTION 5 STUDY A R E A 7 M E T H O D S 8 Live-trapping 9 Trappability 10 Habitat sampling 12 R E S U L T S 13 Habitat structure 13 Overstory 13 Understory 15 Heterogeneity and similarity 15 Demography 18 Density and population trends 18 Seasonality of reproduction 20 Differences in recruitment 24 Changes in body weight 28 Survival and residence time 29 Spatial distribution 29 DISCUSSION 32 Habitat structure and habitat specialists 36 Breeding densities and population stability . . 37 Habitat structure and demographic stability 38 Chapter 3 MICROHABITAT DIFFERENTIATION AMONG DEMOGRAPHIC CLASSES INTRODUCTION 42 M E T H O D S 44 R E S U L T S 47 Relations between vegetation components 47 Vegetation components and demographic classes 48 Microhabitat use by demographic classes 54 Microhabitat differentiation among grids 57 iv Survival in different microhabitats 58 Microhabitat breadth ,60 Microhabitat overlap 60 DISCUSSION 63 Variability in microhabitat partitioning 65 Microhabitat breadth and overlap 67 Chapter 4 POPULATION RESPONSES TO CLUMPED POOD INTRODUCTION 71 M E T H O D S 73 Experimental design 75 Wet season 76 Dry season 76 R E S U L T S .77 Wet season (June to December) 77 Density and seasonal patterns 77 Changes in body weight 81 Differences in recruitment 81 Changes in survival 84 Effects on reproduction 86 Dry season (December to May) 88 Density and seasonal patterns 88 Changes in body weight 90 Differences in recruitment 50 Changes in survival 91 Effects on reproduction 94 DISCUSSION 94 Chapter 5 ECOLOGICAL CONSEQUENCES OF SEX DIFFERENCES: BOTFLY PARASITISM INTRODUCTION 102 M E T H O D S 104 R E S U L T S 106 Seasonality of infestation 106 Infestation rates 106 Effect of food addition on infestation rates 108 Temporal patterns in sex and age related infestation 110 Infestation and breeding condition 112 Movements and infestation 112 Effects on survival 114 Spatial variability in infestation rate 116 DISCUSSION 118 Chapter 6 GENERAL CONCLUSIONS Methodological problems 126 L I T E R A T U R E CITED 127 v L I S T O F T A B L E S Table 2.1. Minimum and maximum trappability estimates 11 Table 2.2. Habitat characteristics of all grids 14 Table 2.3. Summary of habitat structure characteristics 16 Table 2.4. Recruitment of resident and transients 25 Table 2.5. Seasonal recruitment 27 Table 2.6. Seasonal survival of residents 30 Table 2.7. Summary of demographic parameters 38 Table 3.1. Vegetation characteristics of grids 49 Table 3.2. Overstory cover and demographic classes 51 Table 3.3. Microhabitat use from pooled data 55 Table 3.4. Microhabitat use by demographic classes 56 vi LIST OF FIGURES Figure 2.1. Classification of grids by habitat structure 17 Figure 2.2. Population trends of resident individuals 19 Figure 2.3. Annual changes in sex ratio 21 Figure 2.4. Length of breeding season 22 Figure 2.5. Mean residence time 31 Figure 2.6. Spatial distribution of residents 33 Figure 2.7. Seasonal changes in diet 40 Figure 3.1. Microhabitat availability 53 Figure 3.2. Residence time in different microhabitats 59 Figure 3.3. Microhabitat breadth for demographic classes 61 Figure 3.4. Microhabitat overlap for demographic classes 62 Figure 4.1. Experimental design 74 Figure 4.2. Population trends 78 Figure 4.3. Sex ratio of fed and total population .80 Figure 4.4. Changes in body weight 82 Figure 4.5. Recruits and transients in control and experimental grids 83 Figure 4.6. Survival in control and experimental grids in the wet season 85 Figure 4.7. Residence time of fed and unfed individuals in the wet season 87 Figure 4.8. Survival in control and experimental grids in the dry season 92 Figure 4.9. Residence time of fed and unfed individuals in the dry season 93 Figure 5.1. Seasonality of bot fly infestation 107 vii Figure 5.2. Utilization of food stations 109 Figure 5.3. Temporal patterns of sex and age biases in infestation I l l Figure 5.4. Seasonal changes in movements 113 Figure 5.5. Residence time of infested and non-infested individuals 115 Figure 5.6. Spatial variability in infestation rate 117 viii ACKNOWLEDGEMENTS I thank my advisor Charles J. Krebs for his support throughout this work. His dedication to research has always been a source of inspiration. I also thank the members of my commitee C. Lee Gass, A.R.E. Sinclair, Judith H. Myers and William E. Neill, for their advice throughout the development of my work and for their editorial help in the final stages of the manuscript. My external examinors Beatrice Van Home and Jack S. Millar made thorough reviews and excellet comments that improved this thesis. In particular, I thank C. Lee Gass for his active support as a scientist, as a teacher, and as a friend. Many hours of his time and several revisions of the thesis greatly improved the final outcome. During 30 months of fieldwork in the Sierra Madre of Durango, several people voluntarily helped me with different aspects of the study. I was very fortunate to meet Arturo Hernandez who took a break from school to continue my study single-handed for three months while I was back in Vancouver. He did a superb job. Gabriel Moreno, Manuel Weber and Angeles Morales also helped me trapping and provided excellent company. From the Ejido San Juan Michis I also thank Enrique and Belen Contreras and their family and Jose Angel and Maria del Refugio Moreno and their family for friendship and warm hospitality. Jorge Servin made the fieldwork considerably less pleasant. ix I thank Walter Klenner, Finbarr Horgan, and Durell Kapan, colleagues from the Ecology Group of UBC, who revised part or all the thesis and offered excellent suggestions for improvement. Richard Repasky often listened to my methodological concerns and gave me many insightful suggestions. Paolo Domenici and Guillermo Giannico offered their warm support. Alistair Blachford and Joerg Messer from the Biosciences Data Center continuously assisted me with the task of data management. Charlene Higgins helped me to meet the deadline. Financial support for this project was provided by Consejo National de Ciencia y Tecnologia (CONACYT) Mexico, Institute de Ecologia, A.C., and Ancient Cultures. I dedicate this thesis to my wife Laura Jamieson, who participated actively in all stages of this thesis; moving to Mexico, finding a study area, trapping mice. She also listened to too many mouse monologues and made many final editorial improvements. I also dedicate this thesis to my family, which in spite of the distance continues to support my whims. X C h a p t e r 1. G E N E R A L I N T R O D U C T I O N An ima l populations vary continuously both temporally and spatially. They may increase and decrease erratically, remain relatively constant, or cycle. Demographic changes are the collective outcome of the use of resources by indiv iduals, whose requirements and abilities differ according to age, sex, and social and reproductive status (Sutherland and Parker 1985), as well as susceptibility to predation, starvation, parasit ism, injuries and diseases (Clutton-Brock et al., 1982, Sadleir 1984). Therefore, patterns of resource and habitat use by demographic classes may range from complete segregation to complete overlap, considerably affecting the structure and dynamics of populations (Werner and G i l l i am 1984, G i l l i am and Fraser 1988). In tu rn , spatial and temporal patterns of avai labi l i ty of resources alter the social interactions and the distr ibution of individuals (Clutton-Brock and Harvey 1978, Davies and Lundberg 1984, Ostfeld 1985). Habitat structure strongly influences the abundance and distribution of resources such as food, nest sites, protective cover and mates (Cody 1985, Be l l et al., 1990), and therefore i t affects distr ibution and movements of individuals, demographic parameters, population abundances and ult imately, species distr ibution. 1 Insectivorous and granivorous rodents usually have relatively stable populations compared to their herbivorous counterparts of similar size. Their densities range from 5 to 50 individuals per hectare, unlike herbivores which may reach from 100 to 800 individuals per hectare (Taitt and Krebs 1985). Field mice of the genus Peromyscus are among the most well known small mammals. Their populations often exhibit relatively low annual fluctuations in density, and social factors strongly influence maintenance of their stability (Sadleir 1965, Healey 1967, Terman 1968, Harland et al, 1979, Fairbairn 1977, Mihok 1979, Metzgar 1971, Hansen and Batzli 1978, Harland et al., 1979, Van Home 1981, 1982, Taitt 1981, Halpin 1981, Nadeau et al., 1981, Galindo and Krebs 1987). However, population characteristics vary in relatively small geographical areas in relation to differences in habitat quality (Van Home 1981, Krohne 1989). The vast majority of the research on Peromyscus populations comes from two species (P. maniculatus and P. leucopus). Because these have the widest geographical distribution of the 53 members of the genus (Kaufman and Kaufman 1989, Carleton 1989), many aspects of their ecology might not represent the genus. Most Peromyscus species have more restricted distributional ranges (Carleton 1989) and many are probably habitat specialists. One of these species, the rock mouse (Peromyscus difficilis), is the focal point df this study. 2 The rock mouse. The rock mouse (Peromyscus difficilis) is a habitat specialist of montane conifer forests (Hoffineister 1986). It is a relatively large, semi-arboreal species with a distributional range along both the western and eastern Sierra Madre in Mexico and the southwestern U.S.A. It inhabits xeric areas with rocky outcrops and abundant shrubs and trees such as oak, sagebrush, mountain mahogany, pinyon and juniper (Wilson 1968, Holbrook 1978). Like other species of Peromyscus in xeric areas, it is also highly specific in microhabitat use (Holbrook 1978, Hoffineister 1986) and has small litters of 3 to 4 young (Alvarez and Polaco 1984, Hoffineister 1986). Some authors divide rock mice into two species: P. nasutus in U.S.A., and P. difficilis in Mexico (Carleton 1989). Little is known of the demography and other aspects of the ecology of this species. Organization of the thesis. I used an observational approach to examine the relations between habitat structure, demographic variability and microhabitat use patterns by demographic classes. I also used an experimental approach, manipulating the abundance and spatial distribution of food to examine individual and population responses. Finally, I used both natural variation and experimental manipulation to analyze the prevalence of botfly parasitism among sex classes. I first described the population dynamics of the rock mouse in relation to habitat characteristics. I tested the hypothesis that demographic parameters of habitat specialists are closely associated with habitat 3 structure. I investigated the relation between breeding densities, adult survival, recruitment and stability, and their habitat structure correlates (Chapter 2). I examined microhabitat use by individuals to test the hypothesis that demographic classes (sex, age, residency) differ in their spatial distribution, particularly in the breeding season. I also analyzed microhabitat use by individuals as related to habitat heterogeneity and population density (Chapter 3). I tested the hypothesis that females are more responsive to food resources than males, especially in the breeding season, using supplementary food on individual home ranges. I investigated the effects of food on the supplemented individuals and on the demography of the local populations (Chapter 4). Finally, I analyzed temporal patterns of botfly infestation to sex and age classes of rock mice to test the hypothesis that infestation is male-biased and related to increased movement (Chapter 5). In summary, I examined the effects of habitat and food on demographic classes and population dynamics of rock mice, as well as their susceptibility to parasitism. 4 C h a p t e r 2. H A B I T A T S T R U C T U R E A N D D E M O G R A P H I C V A R I A B I L I T Y O F A H A B I T A T S P E C I A L I S T . I N T R O D U C T I O N Habitat structure may influence habitat suitabi l i ty both directly and indirectly, through its effects on microclimate, food abundance (fruits, seeds, insects), the avai labi l i ty and spatial distr ibution of nest sites, and the architecture of protective cover (Cody 1985, Be l l et al., 1990). Consequently, an imal distr ibution and abundance are often associated wi th structural features of the habitat such as the amount of cover, foliage height diversity, depth of perennial grasses, and density of woody understory (Wilson 1968, B rown and L ieberman 1973, Rosenzweig 1973, M'Closkey 1975, Holbrook 1978, Thompson 1982, Verner et al., 1986, Kaufman and Kaufman 1989). Spat ia l differences in habitat structure should be expected to affect the distr ibution of individuals, demographic parameters, population abundance and ult imately, species distribution. Opt imal habitats and distributional boundaries represent the extremes of demographic var iabi l i ty for populations. Between these l imi ts , habitats vary widely i n their capacity to sustain populations. Several models 5 dist inguish two extreme types of populations i n contrasting habitats: Populations i n primary, central, survival , or source habitats are characterized by higher stability, higher survival rates, higher reproductive rates, more stable age distributions, and lower extinction rates. Populations i n secondary, marginal , colonizing, or sink habitats have lower survival rates, lower reproductive rates, less stable age distributions, and higher extinction rates (Anderson 1970, Smi th et al., 1978, Soule 1973, Pu l l i am 1988). Thus, habitat suitabi l i ty is best characterized by h igh reproduction and survival rates which often result i n greater population stabil ity (Van H o m e 1981, 1986). Therefore, to characterize demographic parameters and investigate population regulation i t is necessary to closely examine the relation between habitat and demography (Van Home 1986, Pu l l i am 1988). The demographic characteristics of some populations of small mammals are related to habitat structure (Van H o m e 1981, Bondrup-Nielsen 1987, Ostfeld and Klosterman 1986, Ostfeld et al., 1985, Krohne and Baccus 1985), yet others exhibit very s imi lar demographic patterns i n different habitats (Petticrew and Sadleir 1974, Sadleir 1974, Su l l i van 1979, Parmenter and MacMahon 1983, Adler and Wi lson 1987). Most studies on the effect of habitat characteristics on demography of smal l mammals have compared generalist species i n highly contrasting habitats. Generalists are often selected because their wide distribution and relatively h igh abundance make them easy to study. However, generalists can withstand a wide variety of conditions by definition, and their indiv idual distribution and demographic 6 responses may not strongly reflect habitat characteristics. In contrast, the demography of habitat specialists should be more closely associated to habitat structure since their abundance often declines sharply outside suitable habitats (Adler and Wilson 1987). The rock mouse (Peromyscus difficilis) is a habitat specialist of montane conifer forests. It is a relatively large, semi-arboreal species distributed along the western and eastern Sierra Madre in Mexico and southwestern U.S.A. In this region, it inhabits rocky outcrops in pinyon pine, juniper and oak forests. In this chapter I describe the population dynamics of the rock mouse in relation to habitat characteristics. I tested two hypotheses: First, demographic parameters of habitat specialists should be closely associated with habitat structure. Second, populations with higher breeding densities should have higher adult survival, less recruitment and more stable densities than populations with lower breeding densities. In addition, I analyzed the habitat structure correlates of demographic stability. STUDY AREA The study took place on the eastern slopes of the western Sierra Madre (23° 25' N; 104° 15', W) in southern Durango, Mexico, from February 1986 to July 1988. Annual precipitation averages from 50 to 70 cm and is concentrated in five months between June and October. Mean monthly 7 temperatures fluctuate from 17.4° to 20.4° C. The study area is located in a dry-temperate oak-pine forest at 2400 m altitude. Dominant tree species include several species of oaks (Quercus durifolia, Q. sideroxyla, Q. eduardii, Q. chihuahuensis, Q. convalata, Q. potosina, Q. rugosa), and pines (Pinus arizonica, P. chihuahuana, P. engelmani, P. leiophylla, P. teocote). Point-leaf manzanita (Arctostaphylos pungens) and guazapol (Ceanothus buxifolius) are the main components of the shrub stratum. The study area is within the buffer zone of the Michilia Biosphere Reserve (MAB-UNESCO). METHODS Six grids (HC1-4, LC, MC) were established in oak-manzanita shrubland in March 1986, after a previous survey to locate rock mouse populations. The names of the grids reflect their similarity in terms of the relative composition of cover. Grid LC (Low Cover) had the lowest cover and grids HC1-4 (High Cover) had moderate to high cover. Grid MC had mixed cover, both in terms of species and plant strata (see heterogeneity section). Distances between grids ranged from 500 m to 4 km. Each grid comprised 64 Longworth traps (8 x 8) set at 20 m intervals, covering an area of 2.6 ha. Traps were baited with whole oats. Traps were placed on the ground initially, but most were moved later to 1-2 m on the closest tree to avoid disturbance by grey foxes (Urocyon cinereoargenteus) and peccaries (Dicotyles tajacu). Trapping sessions lasted two nights and were scheduled at two to four week intervals. Traps were open in the afternoon, checked for the next 8 two consecutive mornings and were left locked open between trapping sessions. Trapping took place from March 1986 to May 1987 (except January 1987) on a l l six grids. In M a y 1987 two grids (HC3, MC ) were enlarged to 8 X 17 trap stations covering 5.4 ha, to increase sample sizes for experimental manipulations discussed elsewhere (Chapter 4). These large grids were trapped unt i l Ju ly 1988. I only discuss here the demography of the population on gr id M C , since gr id H C 3 was used for experimental manipulations (Chapter 4). A t least four species of Peromyscus have been recorded i n the study area (Alvarez and Polaco 1984). Most individuals I trapped were P. difficilis following morphological criteria from H a l l (1981) and Comely et al., (1981). Other species present were P. maniculatus, Reithrodontomys megalotis, Sigmodon leucotis, Neotoma mexicana, and Tamias bulleri, but their numbers were extremely low and they were caught only occasionally. Besides rock mice, a persistently low number of P. maniculatus were caught on two grids (HC4, MC) . L i v e - t r a p p i n g . Trapped individuals were ear-tagged, and their sex, reproductive condition, weight (nearest gram) and location of capture recorded. A few individuals lost their tags, but most of these were identified from their trapping history. Breeding condition i n males was determined by recording visible cauda epydidimis (Jameson 1950) and females were 9 recorded as having perforated or non-perforated vaginas. Size of nipples was scored as large, medium or small and evident pregnancies were recorded. Individuals were assigned to one of three age categories: Adults (sexual maturity or > 22 g), Subadults (molting, brown pelage and/or >19 and < 23 g) and juveniles (grey pelage and < 20 g). I used only adult males to analyze differences in body weight among grids and seasons to avoid the confounding effect of pregnancy in females. Sex ratios are shown as the proportion of females in the total population. Trappability. I used the complete enumeration method to estimate population size, since Jolly-Seber inflates estimates when immigration is high relative to number of residents (Adler and Wilson 1987). Furthermore, both methods provide similar results when densities are low as in this case (Galindo-Leal 1991). Enumeration gave accurate population estimates since maximum and minimum trappabilities were higher than 50% for all grids (Hilborn et al., 1976, Krebs and Boonstra 1984). There were no consistent differences in trappability between males and females on any grids (Table 2.1). Individuals trapped on two or more sessions were considered residents and recruits; others were considered transients. I used contingency tables and log-likelihood tests (G tests) to compare differences in the recruitment of resident and transient individuals as well as temporal patterns of recruitment. Because sample sizes were small, I made no distinction between age classes in comparing residence time, and grouped seasons into 10 Table 2.1. Minimum and maximum trappability estimates. Table 2.1. Minimum and maximum trappability estimates for P. difficilis. Sample sizes in parentheses. Trappability Grid Minimum Maximum Unweighted Males Females Males Females HC1 0.77 (26) 0.62 (20) 0.88 (47) 0.81 (37) HC2 0.74 (20) 0.76 (15) 0.75 (42) 0.85 (33) HC3 0.53 (21) 0.62 (26) 0.71 (53) 0.79 (46) HC4 0.60 (30) 0.65 (30) 0.72 (54) 0.78 (48) L C 0.67 (13) 0.58 (19) 0.85 (34) 0.78 (32) MC 0.58 (20) 0.67 (30) 0.69 (47) 0.82 (55) mean 0.65 (130) 0.65 (140) 0.77 (277) 0.81 (251) 11 two periods: spring-summer when most individuals were overwintered adults, and fall-winter, when most individuals were new recruits. Habitat sampling. Habitat sampling was carried out during September 1986. Density and cover of perennial vegetation was estimated in each trapping grid using 10 x 10 m (100 m2) quadrats centered on each trap station (64 quadrats per grid). Thus sampling units covered 40% of each grid. Oaks, pines, junipers and madrones (Arbutus spp.) were pooled into generic categories, and plants were assigned to overstory (>1.50 m) or understory (<1.50 m). Percentage cover was estimated in 25% increments by placing two parallel line transects 5 m apart in every quadrat. Because cover and density were highly correlated for all plant taxa (Chapter 3), only cover was used in several analyses. I used Morisita's Index of Dispersion (Lj) to describe the degree of aggregation of overstory perennial plants (Krebs 1989) and the reciprocal of Simpson's Diversity Index (1/D) to compare heterogeneity of overstory and understory vegetation cover among grids (Krebs 1989). In this study, this index ranges from 1 to 9 (number of categories). Herbaceous plants were almost absent from this habitat, as is common in other chaparral shrub communities (Swank and Oechel 1991), and hence, were not included in the analysis. To test for differences in total plant density and cover among grids I used an analysis of variance (ANOVA) on log-transformed data for densities and arcsin-transformed data for cover (Zar 1984). If results of the ANOVA were significant, I used multiple contrasts (Scheffes test), to determine 12 differences among grids or groups of grids (Wilkinson 1988). The level of significance was set at =0.05. When sample sizes were unequal I used Weighted Means Analysis. I compared habitat composition among grids using Morisita-Horn's Index of Similarity for cover, which ranges from 0 (no similarity) to 1 (complete similarity) (Krebs 1989). Grids were then classified using average linkage clustering (Wilkinson 1988). R E S U L T S Habitat structure Because prehminary trapping showed rock mouse populations were absent in areas with low shrub density and low cover, all grids were established in oak-manzanita shrubland; density, cover and species composition of perennial plants varied among grids (Tables 2.2 and 2.3). Overstory. Total overstory varied among grids (ANOVA F=18.74, d.f.=5,378, p<0.001). Two had low overstory density (LC, MC), three had medium density (HC1, HC3, HC4), and one had high density (HC2), these differences were significant (Multiple contrasts, p<0.05), mostly due to differences in the density of oaks and manzanita. In some areas manzanita and oaks had similar densities (HC3, LC, MC), whereas in others manzanita was twice as dense as oaks (HC1, HC2, HC4). Pine density was high on grid MC and very low in the other grids. 13 Table 2.2. Habitat characteristics of all grids. T a b l e 2 . 2 . Habitat characteristics of all grids. Mean density (individuals per 100 m2), mean percentage cover and dispersion (Id) of perennial plants. Cover values may be over 1.0 since crowns overlap. Differences between individual contributions and totals are due to exclusion of madrones and junipers. Sample sizes are all 64. H C 1 H C 2 H C 3 HC4 LC M C D e n s i t y 1 2 . 8 O v e r s t o r y T o t a l 1 7 . 2 2 2 . 4 1 9 . 1 1 9 . 9 1 3 . 6 Manzanita 11.3 13.8 9.2 12.4 5.3 3.3 Oaks 5.1 7.9 9.5 6.6 6.6 4.8 Pines 0.7 0.6 0.3 0.5 0.7 4.7 U n d e r s t o r y T o t a l 5 . 2 5 . 3 5 . 8 5 . 2 5 . 8 8 . 1 Manzanita 3.1 3.3 2.8 2.7 2.5 2.5 Oaks 1.3 1.7 2.6 2.1 2.2 2.9 Pines 0.2 0.1 0.1 0.3 0.1 1.3 C o v e r O v e r s t o r y T o t a l 1 . 0 1 . 2 3 1 . 0 1 1 . 0 7 0 . 9 0 1 . 0 Manzanita 0.52 0.66 0.51 0.6 0.32 0.19 Oaks 0.39 0.48 0.45 0.39 0.51 0.42 Pines 0.06 0.08 0.03 0.03 0.05 0.35 U n d e r s t o r y T o t a l 0 . 1 6 0 . 1 6 0 . 2 0 0 . 1 6 0 . 1 3 0 . 2 0 Manzanita 0.09 0.08 0.08 0.08 0.06 0.05 Oaks 0.05 0.07 0.10 0.07 0.05 0.07 Pines 0.01 0.01 0.01 0.01 0.01 0.05 D i s p e r s i o n O v e r s t o r y Manzanita 1.4 1.2 1.4 1.4 1.7 2.7 Oaks 1.4 1.3 1.4 1.6 1.6 1.3 Pines 1.1 1.1 3.8 3.7 1.2 1.4 H e t e r o g e n e i t y 3.1 2.9 3.0 2.8 2.9 4.2 14 Grids also differed in overstory cover (ANOVA F=8.63, d.f.=5,378, p<0.001), and could be separated into three groups. However, this time the moderate cover grids were MC, HC1, HC3, and HC4 (Multiple contrasts, p>0.05). As with density, grid HC2 had the highest cover and grid LC the lowest (Multiple contrasts, p<0.05). Mean percentage of manzanita cover varied substantially from 32% to 66%, whereas oak cover was very similar among grids, ranging from 39% to 51%. Pine cover reached 35% on grid MC, but was negligible in the rest ranging from 3 to 8%. Manzanita was highly aggregated on grid MC where it had the lowest density, but not on the other grids (Table 2.2). Oaks had similar dispersion on all grids, whereas pines were highly clumped on two grids (HC3, HC4). U n d e r s t o r y . Both understory density and cover were signficantly higher only on grid MC (density: ANOVA F=3.24, d.f.=5,378, p<0.05; Multiple contrasts, p<0.05; cover: ANOVA F=2.89, d.f.=5,378, p<0.05; Multiple contrasts, p<0.05). In the other five grids, the density of understory was very low (Multiple contrasts, all p>0.05). The differences were mostly due to the higher pine density in the understory of grid MC. H e t e r o g e n e i t y a n d s i m i l a r i t y . Grid MC was most heterogeneous, others were similar (Table 2.2, 2.3). All grids had similar cover composition, ranging from .69 to 1.0, but four grids were particularly similar (HC1-4) (Morisita-Horn indices above 0.99). Grids LC and MC were each in their own group, but the former was closer to the other four grids than MC (Fig. 2.1). 15 Table 2.3. Summary of habitat structure characteristics. Table 2.3. Summary of habitat structure characteristics in all grids. + indicates relatively high; - relatively low. No sign indicates moderate. HC1 HC2 HC3 HC4 L C MC Habitat Structure Overstory density + - -cover + Understory density _ _ _ - _ + cover - - - - - + Manzanita aggregation - + Pine aggregation - - + + - -Heterogeneity + Cover similarity a a a a b c 16 Figure 2.1. Classif ication of grids by habitat structure. . 6 . 7 .8 .9 COVER 1.0 HC2 HC4 HC1 HC3 LC MC Figure 2.1. Classification of grids using relative composition of both under and overstory cover of perennial plants, using Morisita's Similarity Index and average linkage clustering. 17 Demography Density and population trends. Rock mouse populations on al l grids were low and relatively stable throughout the study. Densities fluctuated seasonally from 1.6 to 13.3 individuals per hectare. Population numbers declined or were already low i n the spring and continued this way dur ing the first ha l f of the wet season (June-Aug.), then increased through the rest of the wet season (Sept.-Oct.) and the first ha l f of the dry season (Nov.-Jan.). Populations either declined at the end of the dry season (Apr-Jun) or remained stable. There was a sharp pulse of recruitment on 4 grids during A p r i l or May . On average, the m in imum number of individuals per gr id was 8.0 (s.e.=l. l , n=6) and the max imum was 29.3 (s.e.=2.3, n=6). The grids could be divided into three types according to within-year population trends: f luctuating, stable h igh density and stable low density (Fig. 2.2). F luctuat ing populations (HC1, H C 2 , HC3) had low spring, summer and fal l density (minima of 4-9/grid) and high winter density (maxima of 27-34/grid). They increased from 3 to 8.3-fold. Stable h igh density populations (HC4, M C ) had higher summer densities (minima of 11-13/grid), and lower winter densities (maxima 23-25/grid) than fluctuating populations. They increased from 1.9 to 2.1-fold. However, the "stable h igh density" population on gr id H C 4 went extinct i n the spring of 1987. Stable low density populations had both lower summer (8 individuals) and winter densities (19 individuals) than the other types (LC). This population increased 1.9-fold. 18 Figure 2.2. Population trends of resident individuals. Figure 2.2. Fluctuating (HC1-3), stable low density (LC) and stable high density (HC4, MC) populations. Minimum number alive of resident individuals in all grids. Hatched areas indicate wet season in 1986-1987. 19 Sex ratio (proportion of females) varied seasonally and among grids. At the beginning of the breeding season (May-August) it was even or slightly but consistently biased towards females on most grids (HCl, HC3, HC4, LC, MC). From September to April it was even (LC), female-biased (HC3, MC), or male-biased (HCl, HC2, HC4) (Fig. 2.3). In 1987 and 1988 the sex ratio was relatively even on two grids and male-biased on one grid (Chapter 4). Seasonality of reproduction. Breeding activity was strongly synchronized with the wet season and was therefore highly seasonal, lasting from 6 to 7 months. Most females were in breeding condition from June to December each year, but there was some variability between years and among grids. Breeding females were caught from late May to December in 1986, from February to December in 1987, and again in June and July in 1988. Early breeding in 1987 was likely due to an unusually heavy snowfall in February 1987. In June 1986 breeding began with the wet season on most grids (HCl, HC2, HC4 and MC), or slightly earlier (HC3, LC), and declined during November and December. By February (1987) there was a complete absence of breeding on all grids except grid HC4 (Fig. 2.4). In 1987, up to 40% of females were breeding before the wet season began. Pregnant females were recorded on most grids during February (HC4), March (HCl, HC2, HC3) and April (MC), but not on grid LC (Fig. 2.4). Breeding finished in November and early December. In 1988 breeding began in June in grid MC. 20 Figure 2.3. Annual changes in sex ratio. Figure 2.3. Proportion of females in fluctuating (HC1-3), stable low density (LC), and stable high density (HC4, MC). populations. Stippled areas indicate wet season in 1986-1987. 21 Figure 2.4. Length of breeding season. FEMALES 1986 1987 HC1 I • — ^ M I I I I HC3 i i i mt I M — I I J LC | I I — — I I I MC I I — — I i m 1 9 8 ? M C 1 9 8 8 M A M J J A S O N D J F M A M J J MALES 1986 1987 HC1 I I — I I \ HC3 I — I I I I I ' HC4 i i • i t n n m l » l MC i i — % i M — i i • mm 1 9 8 7M C i • » i 1 9 8 8 i = i M A M J J A S O N D J F M A M J J Figure 2.4. Breeding season of females and males in 1986 and 1987. Solid bars indicate more than 50% individuals breeding; stippled bars indicate less than 50%; empty bars indicate no breeding. 22 The breeding season, as I defined it was shorter for males than for females. It also began with the wet season but lasted from 4 to 6 months. Breeding males were caught from late May to early December in 1986, from February to November in 1987, and in June and July in 1988. In 1986, most males in reproductive condition were caught from June to October (86%) but a few were caught in May (5%) (HC4, LC) and in November and early December (9%) (HC3, HC4). Peak numbers of reproductive males occurred during July and August when most grids had 50% or more breeding males. The breeding season also ended at different times, from August (LC) to October (HC1, HC2, MC) to November (HC3) to December (HC4). In 1987, breeding males were recorded two or three months earlier, in February (HC1) and March (HC3). Breeding continued until early November, but most males were breeding from July to October. In 1988 males began breeding in June (Fig. 2.4). Among a total of 56 females that were evidently pregnant, 39 (70%) were overwintered adults. The rest, 17 (30%) were young of the same year. Overwintered adults had one (51%), two (31%) or three (18%) pregnancies during the breeding season. Most young of the year were pregnant only once during that year (94%), but one was pregnant twice (6%). The average numbers of pregnancies of overwintered adults varied among grids from 1.1 (HC4), 1.6 (HC1, HC3), 1.8 (MC) to 2.0 (LC) and 2.4 (HC2). Small sample sizes preclude statistical analysis. 23 Excluding evident pregnancies, breeding females were significantly heavier (x=33.4, s.e.=1.29) than non-breeding adult females (x=26.8, s.e.=0.2; Weighted Means Analysis F=53.5, d.f.=l,132 p<0.001). No female less than 25 g showed signs of breeding. Breeding males (x=30.9, s.e.=0.85) were also significantly heavier than non-breeding adult males (x=27.19, s.e.=0.27; Weighted Means Analysis F=17.7, d.f.=l,165 p<0.001). No male under 27 g showed signs of breeding. Differences in recruitment. There were substantial differences in total recruitment among grids. Recruitment varied from low (LC), and moderate (HC1, HC2) to high (HC3, HC4, MC; Table 2.4). The age composition of recruits differed significantly among grids (G=30.34, d.f.=10, p<0.001). The largest contributions to total G values were from the low proportion of subadult recruits on grid MC. Differences among grids other than MC were not significant (G=13.92, d.f.=8, p>0.05; Table 2.4). The ratio of adult to subadult and juvenile recruits was similar on all grids except grid LC. The proportions of juveniles and subadult recruits also varied among grids. Two grids had mostly subadult recruits (HCl, HC3), two grids had even numbers (LC, HC4), and two grids had mostly juvenile recruits (HC2, MC), particularly grid MC. Recruitment also varied seasonally. In 1986 many adults recruited during spring. Mean monthly recruitment for these months was 9 individuals per grid (s.e.=1.8, n=18). From June 1986 to May 1987, monthly recruitment averaged only 2.1 (s.e.=0.2, n=72). Among adults, there was 24 Table 2.4. Recruitment of resident and transients. Table 2.4. Percentages of recruits and transient individuals of different age classes. J = juveniles; SA = subadults; A = adults. Sample sizes (n) are shown in parentheses. Recruits Transients Grid J SA A n J SA A n HC1 12 25 63 (59) 8 21 71 (24) HC2 22 13 64 (45) 23 17 60 (30) HC3 11 23 66 (64) 28 13 59 (32) HC4 18 18 64 (71) 26 19 55 (31) L C 29 29 41 (44) 50 5 45 (22) MC 31 3 66 (58) 36 12 52 (42) 25 a small bias towards males, particularly in two grids (HC1, HC3). Pooled results from all age classes indicate that most recruitment occurred in spring, autumn and winter. Recruitment was very low (7-11%) on all grids in summer, and also low in the spring of 1987 (3-7%; Table 2.5). Grids differed significantly in seasonal patterns of recruitment (G=40.0, d.f.=20, p<0.005). Largest contributions to total G values indicated that the greatest differences were due to high recruitment during winter on grid HC2 and during autumn on grids HC1 and LC, but the test was not significant when grids HC1 and HC2 were excluded from the contingency table (G=13.8, d.f.=12, p>0.05). Juvenile recruitment varied substantially among grids, from late July (HC4), late August (HC3, LC, MC), and late September (HC1) to October (HC2). Populations differed by as much as two months in the date of first juvenile recruitment. Juvenile recruitment was highest in September and late November and December (LC, MC), in December (HC2, HC3) or was relatively constant (HC1, HC4). Juveniles and subadults mostly recruited in fall and winter. Recruitment was highly seasonal in some areas with 83 to 94% of all juveniles and 74 to 86% of all subadults recruited in this period (HC1, HC2, HC3, LC). In other areas it was more prolonged with 67 to 71% of all juveniles and 53 to 57% of all subadults recruited in fall and winter (MC 26 Table 2.5. Seasonal recruitment. T a b l e 2.5. Percentages of recruits and transient individuals i n different seasons. Sp = spring; Su = summer; Au= autumn; W i = winter. Sample sizes are shown i n parentheses. 1986 1987 R e c r u i t s S p S u A u W i S p n H C l 31 7 47 10 5 (59) H C 2 22 11 27 36 4 (45) H C 3 29 11 27 28 5 (64) H C 4 51 8 23 14 4 (71) L C 27 7 48 11 7 (44) M C 52 7 21 17 3 (58) T r a n s i e n t s H C l 54 8 16 21 0 (24) H C 2 30 7 27 20 17 (30) H C 3 31 6 19 34 9 (32) H C 4 19 13 26 19 23 (31) L C 36 14 41 5 5 (22) M C 43 12 24 14 7 (42) 27 and HC4). Very few juveniles recruited in summer or spring. Pooling juveniles and subadults, the mean monthly recruitment was 5.3 individuals (s.e.=0.7, n=30) in fall and winter. Outside this period, monthly recruitment averaged only 0.98 individuals (s.e.=0.2, n=51). Recruitment was even among sexes both for juveniles and subadults. Transient individuals constituted between 29 and 33% of all the recruits on most grids (HC1, HC3, HC4, LC), but grids HC2 and MC had higher proportions (40 and 42 % respectively). The age composition of transients was similar among grids (G= 13.81, d.f.=10, p>0.05) (Table 2.4). Most transients were also adult individuals, but the percentages varied widely. Grid LC had the fewest adult transients (Table 2.4). Juvenile transients outnumbered subadults on all grids except grid HC1. On grid LC, 50% of all transients were juveniles. Transients followed a similar temporal pattern to that of recruits. In 1986, they were most abundant in the spring, autumn and winter and less abundant in the summer. In 1987 they were less abundant in spring. Seasonal differences among grids were not significant (Table 2.5). Changes in body weight. Seasonal changes in body weight were marked on all grids, increasing, from a spring minimum of 27-28 g to a summer peak of 32-33 g, then decreasing steadily to the spring minimum. In the breeding season average body weight was significantly higher only in grid HC3 (x=33.8, s.e.=1.2; Weighted Means Analysis, F=2.39, d.f.=5,72, P<0.05). 28 S u r v i v a l a n d r e s i d e n c e t i m e . Survival was relatively high and similar for both adult males and adult females. Spring and summer adult females survived slightly but not significantly better than adult males. Survival was very similar between the sexes in autumn, winter and spring of 1987. This pattern was particularly evident on four grids (HCl, HC2, HC3, LC). On grids MC and HC4, adult survival was very similar between the sexes or even higher for males (Table 2.6). Small sample sizes precluded analysis of juvenile survival. Sexes did not differ in residence time (Three-way ANOVA, F=0.754, d.f.=l,281, p>0.05). There was a significant interaction between grid and season (Three-way ANOVA, F=2.70, d.f.=5,281, p<0.05). Residence time was longer in the fall-winter period on grids H C l and HC3, whereas the opposite occurred on grid HC2. There was no difference in the rest of the grids (Fig. 2.5). S p a t i a l d i s t r i b u t i o n . During the breeding season individuals were widely dispersed on the grids. Male home ranges were slightly larger than female home ranges, particularly during breeding (Chapter 5). There was little overlap either between or within sexes, particularly on grids with low breeding densities (Fig. 2.6). 29 Table 2.6. Seasonal survival of residents. Table 2.6. Average two week survival of resident adults of both sexes in different seasons. Sp = spring; Su = summer; Au = autumn; Wi = winter. Sp Su Au Wi Sp Mean H C l females .84 .84 .79 .93 .89 .86 males .80 .62 .91 .92 .86 .82 HC2 females 1.0 .89 .89 .89 .84 .90 males 1.0 .91 .81 .88 .71 .86 HC3 females .78 .92 .92 .84 .97 .89 males .67 .94 .85 .83 .91 .84 HC4 females .88 .81 .88 .86 .79 .84 males .82 .90 .91 .85 1.0 .89 L C females .96 .92 .83 .73 .87 .86 males .80 .84 .85 .90 .94 .87 MC females .81 .90 .86 .85 .91 .87 males .78 .83 .90 .89 .95 .87 mean females .88 .88 .86 .85 .88 mean males .81 .84 .87 .88 .89 30 Figure 2.5. Mean residence time. Figure 2.5. Mean residence time (days) during spring-summer (•) and fall-winter (O) in different grids. Bars represent one standard error. 31 D I S C U S S I O N I will first compare the population dynamics of the rock mice with those of other species of Peromyscus and then discuss their demographic variability in relation to habitat structure. Unlike herbivorous rodents which reach densities of 100 to 800 individuals per hectare with amplitudes of numerical change over 10-fold (Taitt and Krebs 1985), insectivorous and granivorous rodents usually range from 5 to 50 individuals per hectare, and increase less than 5-fold in their annual fluctuations. In this study, rock mice fluctuated annually in density, similar to northern species of Peromyscus (Montgomery 1989). Numbers were lowest during the summer, increased in the fall and winter, and declined in the spring. Densities were relatively low and the amplitudes of population fluctuation were small (from 2 to 8-fold) as in other Peromyscus species living in lower latitudes (P. eremicus, P. boylii, P. californicus, P. polionotus, P. gossypinus) (Kaufman and Kaufman 1989, Montgomery 1989). Reproduction was usually restricted to the wet season from June to November or December. Earlier breeding in 1987 may have occurred because of increased moisture caused by an unusually heavy snowfall in February. Consequently, demographic parameters, like low numbers of recruits and transients during spring 1987, resembled those of the previous summer. Rock mouse populations have a relatively even sex ratio (Chapter 4). Adults are similar in size (males 27.5 g, s.e.=0.21, n=297; females 27.6 g, 32 Figure 2.6. Spatial distr ibution of residents. GRID HC1 GRID HC2 GRID MC GRID HC4 Figure 2.6. Spatial distribution of resident males (filled dots) and females (open dots) in two low density grids and two high density grids during the first three months of breeding season. In most grids there was little overlap between males and females. 33 s.e.=0.26, n=240), and both sexes apparently have exclusive and non-overlapping home ranges in the breeding season. This is in contrast to most reports of Peromyscus where sex ratios are male-biased and male home ranges overlap with those of females (Wolff 1989, Kaufman and Kaufman 1989). Exclusive use of space by males and females is exhibited by mammals which make food caches, such as red squirrels, kangaroo rats, pikas and pack rats (Smith and Reichman 1984). Several other species of Peromyscus, particularly those living in semi-arid areas, cache seeds (Barry 1976). Rock mice cache acorns, but it is not known to what extent they rely on them (Alvarez and Polaco 1984). Rock mice also show a seasonal cycle in average body weight and a threshold body weight seems necessary to attain breeding condition. The decline in average body weight was highly influenced by the recruitment of light weight adults in the fall and winter. The increase in average body weight, however, most likely represents a response to increased food availability at the start of the wet season. c Habitat structure and habitat specialists. Throughout its range, distribution of rock mice is highly restricted to habitats with particular structural features such as protective cover, characteristic plant species composition and rocky outcrops (Wilson 1968, Holbrook 1978, Hoffmeister 1986). In fact, this species has the most restricted habitat distribution among 8 Peromyscus species in Arizona (Hoffmeister 1986). 34 By definition, habitat specialists are absent or in extremely low abundances in other than their.preferred habitat. Thus, comparison of demographic characteristics is limited to smaller variations in habitat characteristics. At a gross scale, all grids in this study have the same habitat: oak-manzanita shrubland. Nevertheless, rock mouse distribution was sensitive to slight modifications of habitat structure, and the demographic characteristics were more extreme on grids with greatest differences in vegetation characteristics (MC, LC; Table 2.3). The results support the hypothesis that demographic parameters of habitat specialists are closely associated with habitat structure. Grid MC had the highest heterogeneity, highest understory cover, highest pine density and highly clumped manzanita (Table 2.3), and also differed demographically from the other grids (Table 2.6). Its sex ratio was consistently female-biased, had higher breeding densities, higher stability and a shorter breeding season. Recruitment was higher than on other grids in the spring of 1986, but was lower in the fall (Table 2.7). That this grid had lower juvenile and subadult recruitment in fall and winter and more transients may indicate a higher turnover due to repressed immigration. Grid HC4 shared several of the demographic characteristics of grid MC (Table 2.7), but there were also differences. Grid HC4 had few transients, moderate subadult recruitment and a longer breeding season. In spite of high breeding densities during summer and relative stability during fall and winter 1986, the population on grid HC4 went extinct during the spring of 35 Table 2.7. Summary of demographic parameters Table 2.7. Summary of results on demographic parameters in all grids. + indicates relatively high; - relatively low. No sign indicates moderate. Demographic characteristics Summer to winter increase Proportion of females Breeding density Reproduction Recruitment Adults Subadults/Juveniles Transients Adults Subadults/Juveniles Residence time (1st) (2nd) Body weight H C l HC2 HC3 HC4 L C MC 8.3 3.0 5.7 1.9 2.4 2.1 0.48 0.45 0.53 0.43 0.56 0.57 - + - + - - + + + -+ + - + + + + + - + + - + -- + - - - + + + + + - + + -- + + - + + - + 36 1987. Habitat structure in this grid was most similar to that in grids H C l , HC2, and HC3. Habitat structure was also substantially different on grid LC. This grid had the least overstory and understory cover and also differed demographically from the rest of the grids in several parameters: consistently lower density, adult recruitment, low numbers of adult and subadult transients and low to medium survival. The other grids (HCl, HC2, HC3), had similar habitat structure among them. Grid HC3 had slightly lower total plant density than the other two, but very similar cover. Demographically, these grids had low to medium breeding densities, lower stability, high adult recruitment and many transient adults (Table 2.7). B r e e d i n g d e n s i t i e s a n d p o p u l a t i o n s t a b i l i t y . Van Home (1981) described two general types of summer demography for Peromyscus maniculatus which were related to serai stages of coastal coniferous forest. First, relatively stable and high density populations, with mostly adult individuals, were found in intermediate serai stages. These had a high perennial shrub cover and some canopy closure. Second, less stable, high density populations, with mostly juvenile individuals, were found in earlier and later serai stages with less understory cover. In my study, the population on grid MC was somewhat equivalent to Van Home's first type, because high breeding densities were accompanied by high stability, and juvenile and subadult recruitment were low in the breeding season. This supports the hypothesis that populations with higher breeding densities, 37 particularly of females, have higher survival, lower recruitment and have more stable densities. Populations on grids HC4 and LC were also relatively stable. However, the population on grid HC4 in spite of higher breeding density and its resemblance with that on grid MC, declined to extinction during the spring of 1987. Krohne (1989) also reported that populations of P. leucopus with similar demography during part of the year may differ during others. The low densities and few recruits and transients of grid LC with sparse overstory and understory cover suggest that its stability resulted from unsuitable habitat. The populations on grids H C l , HC2, and HC3 were more similar to Van Home's second type. They all had high percentage of juvenile and subadult recruits and population changes from summer to winter were more dramatic than on the other grids. Habitat structure and demographic stability Ostfeld et al. (1985) underscored cover as a good index of habitat quality. Their populations of California voles (Microtus californicus) had higher peak densities, female-biased sex ratios, higher juvenile recruitment, and longer persistence in habitats with dense cover. In this study, high within-year stability was not unrelated to plant cover or density per se, but occurred in the grid with highest heterogeneity in composition of plant cover. Heterogeneous habitats might provide a combination of protective cover with more diverse and stable food supplies, 38 since different plant components provide alternative resources at different times. The branching configuration of manzanita shrubs offers dense, protective cover, while hollowed oaks and junipers provide nest sites. Rock mice are typical Peromyscus in that they are opportunistic feeders whose diet changes widely in relation to seasonal changes in food availability (Alvarez and Polaco 1984) (Fig. 2.7). In terms of food, manzanita provides flowers and fruits during spring and summer, whereas oaks, junipers and pines provide seeds in late fall and throughout winter. Insects are consumed mostly in winter and spring. While my results support the idea that differences in demographic parameters of habitat specialists are closely related to subtle changes in habitat structure, other studies of small mammals have reported populations in very distinct habitat types to be very similar in demography (Petticrew and Sadleir 1974, Sadleir 1974, Sullivan 1979, Adler and Wilson 1987). There are at least three plausible reasons for demographic similarities in the face of habitat differences in those studies. First, most of the studied species were extreme habitat generalists with wide geographic distributions and including diverse habitats. For example, the most well-studied species of Peromyscus (P. maniculatus and P. leucopus) are the only ones among 42 mainland species that have distributional ranges covering more than 30° of latitude. The rest cover less than 25° (15 species) or less than 10° of latitude (25 species; Carleton 1989). Second, many studies have been restricted to only a part of the annual cycle, usually summer and fall. Populations in 39 Figure 2.7. Seasonal changes in diet. Figure 2.7. Seasonal changes in the diet of the rock mice in the study area. Sample size in parenthesis. Modified from Alvarez and Polaco (1984). 40 different habitats might be similar demographically while they were being studied but different in winter (Krohne 1989; this study). Third, habitat structure and demography may have been evaluated on different spatial scales. For example, Adler and Wilson (1987) reported similar summer demography of P. leucopus in different habitat types, but that conclusion was based on very small trapping grids (0.28 ha). Home range sizes of Peromyscus range from 0.02 ha to over 0.30 ha (Wolff 1989). Therefore, while the characterization of the habitat might truly represent that area sampled, the characterization of the demography of Peromyscus does not; most individuals captured in that study probably lived elsewhere. General aspects of the demography of rock mice are similar to those of northern species, but other features are very different. Most Peromyscus species have restricted distributional ranges (Carleton 1989) and many are probably habitat specialists. In fact, many of the large sized species seem to have specialized nest site requirements (P. californicus, P. truei, P. boylii; Merritt 1974). However, most of our knowledge (Kirkland and Layne 1989) comes from only 2 of the 53 recognized Peromyscus species (Carleton 1989). These two have the widest geographical distribution of all and are therefore atypical (Montgomery 1989). Further research on the other 51 little-studied species should provide fruitful insights into the relations of habitat structure, demographic variability and geographical distribution. 41 C h a p t e r 3 M I C R O H A B I T A T D I F F E R E N T I A T I O N A M O N G D E M O G R A P H I C C L A S S E S I N T R O D U C T I O N Most studies of habitat use have been concerned wi th the coexistence of species (Cody 1985, Kaufman and Kaufman 1989, Montgomery 1989). A common feature of studies showing differences i n habitat use is the assumption that a l l individuals of a population behave similar ly; no distinction is made between demographic classes (i.e. sex and age groups). However, patterns of resource and habitat use of demographic classes wi th in species may vary from complete segregation to complete overlap, and such differences have important consequences for the structure and dynamics of the population (Werner and G i l l i am 1984, G i l l i am and Fraser 1988). Sex, age, and social and reproductive status affect the requirements of individuals and the capacity to ful f i l l ecological demands (Sutherland and Parker 1985). Ecological differences between the sexes are usual ly associated w i th sexual dimorphism, but they may also be present i n sexually monomorphic species. For example, h igh energetic demands of lactat ing 42 female mammals (Millar 1989) are likely to produce patterns of resource use different from those of other individuals in the population. Ecological differences between age classes are associated with size differences and are extreme among holometabolous insects and amphibians (Werner and Gilliam 1984). Social status may also influence resource use since dominant individuals may monopolize resources through territorial behavior. Consequently, subordinate individuals, may be displaced to less favorable microhabitats (Fretwell 1972). Since large size is often a prerequisite to attaining dominant status (Morse 1974), juveniles are often subordinate to adults (Healey 1967, Van Home 1982, De Laet 1985), and the larger sex, dominant to the smaller one (Jones 1985, Temeles 1986). While few studies have analyzed intraspecific differences in resource use, those few have revealed differences in microhabitat use among sex or age classes in a variety of taxa, including plants (Bierzychudek and Eckhart 1988), invertebrates (Werner and Gilliam 1984), lizards (Schoener 1977), birds (Morse 1985, Jones 1985, Cody 1985) and mammals (Clutton-Brock et al., 1982, Bowers and Smith 1979, Van Home 1982, Kincaid and Cameron 1985, but see M'Closkey 1985). Age and sex differences in habitat use are well documented among many large mammals, particularly in ungulates (Clutton-Brock et al., 1982). Among small mammals, some studies have documented differences in habitat use among age (Merkt 1981, Van Home 1982, Kincaid and Cameron 1985) and sex classes (Bowers and Smith 1979, Morris 1984, Seagle 1985, Kincaid 43 and Cameron 1985), but others have not found differences (Morris 1984, M'Closkey 1985, Kaufman et al., 1985). Age related and sex-specific differences i n microhabitat use due to behavioral interactions have been documented for some birds (Peters and Grubbs 1983, Morse 1985, Jones 1985, Cody 1985, Desrochers 1989). In this chapter I analyzed microhabitat use of rock mice by sex, age, residence status and season. This species is sexually monomorphic (Chapter 2), and therefore, ecological differences between sexes cannot be related to size differences. I tested the hypotheses that demographic classes differ i n microhabitat use, and that these differences would increase dur ing the breeding season, since agonistic behavior increases dur ing reproduction (Sadleir 1965). I also predicted that subordinates would show smaller microhabitat breadths and less overlap than dominant individuals (Morse 1974). F ina l ly , I studied microhabitat use by individuals as related to habitat heterogeneity and population density. METHODS Six grids separated by 500 m to 4 k m , were set i n s imi lar habitat (Oak-manzanita shrubland) after a previous survey to locate rock mice populations. The study area has been described elsewhere (Chapter 2). Each gr id (2.6 ha) had 64 Longworth traps (8 x 8) at 20 m intervals. Trapping sessions lasted two nights and were scheduled at two to four week intervals 44 in every month with the exception of January. Traps were baited with whole oats, left in place, and locked open between trapping sessions. All individuals caught were ear-tagged, and their sex, reproductive condition, weight (nearest gram) and trap station were recorded. They were assigned to two age categories: Adults (sexually mature or > 22g) and juveniles (grey pelage and < 22g). Individuals trapped for two or more sessions were considered residents whereas those trapped only once were considered as transients. Van Home (1982) suggested the need to correct capture variables obtained in grid trapping because of a likely edge effect of the trapping grid. To determine the existence of edge effect I compared the mean number of captures in the 4 concentric perimeters of the grids, pooling data from all grids. No significant differences were found among the 4 means (Weighted Mean Analysis F=0.76, d.f.=3, p>0.05), so, there was no need for correction. The importance of the boundary strip decreases as the size of the grid increases (White et al., 1982), so the lack of edge effect in this study might be due to the larger grid size (2.6 ha) used and the low densities encountered. Microhabitat sampling was carried out during September 1986. I selected perennial vegetation as the main habitat variable since previous studies have shown strong correlations between structure provided by this habitat component and utilization by Peromyscus species in similar habitats 45 (Wilson 1968, M'Closkey 1975, Holbrook 1978). This should be particularly true for the highly arboreal Peromyscus difficilis. I classified perennial vegetation on 10 x 10 m (100 m2) quadrats centered on each of the 384 trap stations into two categories: understory with plants below 150 cm and overstory with plants above 150 cm. The number of individuals of each genus was counted: oaks (Quercus spp.), pines (Pinus spp.), junipers (Juniperus spp.), and madrones (Arbutus spp.). Only one species of manzanita (Arctostaphylos pungens) was present in the areas. Madrones were not recorded in the understory. Canopy cover of individual genera in both strata was estimated in 25% increments by placing two parallel line transects 5 m apart. Measurements of aggregation (Morisita's Index of Dispersion = I<j) and heterogeneity (reciprocal of Simpson's Diversity Index = 1/D) are described elsewhere (Chapter 2). I determined hierarchical microhabitat categories using TWINSPAN (Two Way Indicator Species Analysis) (Gauch 1982, Jongman et al., 1987). Twinspan uses reciprocal averaging to separate groups. Unlike other classification programs, Twinspan arranges clusters placing similar samples close together (Gauch 1982). To analyze microhabitat use by demographic classes of rock mice in relation to availability I used Replicated Goodness of fit tests (Sokal and Rohlf 1981) and partitioned the total into contributions due to individual classes. I distinguished four demographic classes (adult males and females, juveniles and transients) and two seasons (breeding/non-breeding). The breeding season corresponds to summer and fall. 46 To evaluate niche breadth, I used Hurlbert's standardized measure (B'A), which includes differences in resource state availability and is sensitive to the selectivity of rare resources. It ranges from 0 to 1 (Krebs 1989). To analyze habitat overlap, I used Hurlbert's Index (L), which also takes into account resource availability. This index is 0 when two species share no resources, 1.0 when both species use resources in proportion to availability, and more than 1.0 when the two species use certain resources more intensively than others and their use coincide (Krebs 1989). Confidence intervals for both breadth and overlap were calculated using replicated values from the 6 areas. Only the first capture of every session was included for the analyses since the location of the second capture was not independent (Williams adjusted Log-likelihood ratio Gadj=22.67, d.f.=l, p<0.001). I used ANOVAs to analyze differences in residence time by sex and age classes in different microhabitats, and to analyze differences in microhabitat breadth by demographic classes (Sokal and Rohlf 1981). RESULTS Relations between vegetation components All 5 vegetation components (manzanita, oak, pine, juniper and madrone) were represented on all grids, with the exception of grid HC3 which lacked madrones. Manzanita and oaks were the dominant groups followed by pines. The two former components constituted 98% of the total density of perennial vegetation on most grids. 47 On grid MC the density of pines was much higher; pines along with oaks made up the dominant group. Junipers and madrones were relatively rare on all grids, usually < 0.3/100 m 2. Only on grid MC did junipers attain densities equivalent to those of pines (Table 3.1). On all grids cover and density were significantly correlated for all vegetation groups in both overstory and understory (all Spearman rs>0.3, n=64, p<0.05). Therefore, I carried on the analysis using only cover. Both manzanita and oaks were highly clumped on all grids, although their degree of aggregation varied. Oaks and manzanita were least clumped on grid HC2. Pines had a low degree of aggregation on most grids, although they were highly clumped on grids HC3 and HC4. Vegetation groups were highly correlated. Manzanita and oak cover were negatively correlated on all grids but MC. Manzanita and pine cover were also negatively correlated on 3 grids (HCl, HC2, HC4). Finally, pine and oak cover were negatively correlated only on grid MC, where pines were most abundant (Table 3.1). According to the index of diversity (1/D), grids can be grouped into low (HC2, HC4, LC), medium (HCl, HC3) and high heterogeneity (MC) (Chapter 2). V e g e t a t i o n c o m p o n e n t s a n d d e m o g r a p h i c c l a s s e s O v e r s t o r y : Age and sex classes, particularly adult males, were positively correlated with 48 Table 3.1. Vegetation characteristics of grids. T a b l e 3 . 1 . Vegetation characteristics o f grids. Mean Percentage cover ( C ) ; Mean number of individuals/100 m 2 (D); Morisita's index of dispersion; (Id). Heterogeneity (1/D). Spearman correlations (rv) between cover of overstory components. * p<0.5, ** p<0.01, *** p<0.001. Signs indicate trends of non-significant correlations. M a n z a n i t a C D I d H C l 52.3 11.1 1.4 H C 2 65.8 13.8 1.2 H C 3 51.2 9.2 1.4 H C 4 59.9 12.2 1.4 L C 32.4 5.3 1.7 M C 19.1 3.3 2.7 C o r r e l a t i o n s M a n z a n i t a - o a k s H C l -0.40 ** H C 2 -0.30 * H C 3 -0.58 * * * H C 4 -0.60 * * * L C -0.45 * * * M C + O a k s c D I d C 38.7 5.1 1.4 5.5 47.7 7.9 1.3 8.2 44.5 9.5 1.4 2.7 39.8 6.6 1.6 3.1 50.9 6.6 1.6 5.1 41.9 4.8 1.3 34.8 M a n z a n i t a - p i -0.28 ** -0.33 ** -0.32 ** P i n e s D I d 1 / D 0.7 1.1 3.07 0.6 1.1 2.87 0.3 3.8 3.02 0.5 3.7 2.84 0.7 1.2 2.87 4.7 1.4 4.17 P i n e - o a k s + + -0.38 ** 49 manzanita cover, as indicated by analysis of pooled captures from all 6 grids. In contrast, all groups, except adult females, were negatively correlated with pine cover. Correlations with oak, juniper and madrone cover were not significant (Table 3.2). The distribution of age and sex classes in relation to cover varied among grids. Captures of adult females were positively correlated with manzanita cover in 3 grids (HCl, HC4, LC) and with oak cover in two grids (HC2, MC) and not significantly correlated with any vegetation component on grid HC3. Captures of adult males were positively correlated with manzanita cover on all grids except HC2 and were negatively correlated with oak cover on grid HC4. Captures of juvenile females were positively correlated with manzanita cover in two grids (HC4, LC) and negatively correlated with pine cover on grid LC. Captures of juvenile males were positively correlated with manzanita cover on grid H C l and with oak cover on grid HC2 (Table 3.2). There were no significant correlations with juniper cover. Similarly, captures of both adult males (Spearman rs=-0.27, p<0.05) and adult females (Spearman rs=-0.27 p<0.05) were negatively correlated with madrone cover only on grid HC4. Understory: The few correlations between demographic classes and different components of the understory, resembled those with the overstory. Analysis of all 6 grids pooled indicated that only captures of adult males 50 Table 3.2. Demographic classes and overstorey cover Table 3.2. Spearman rank correlations of captures by sex and age classes and overstory cover of dominant vegetation components. All grids have n=64. Trends of non-significant correlations are indicated with signs. * p<0.5, ** p<0.01, *** p<0.001. Adults Juveniles Females Males Females Males All Grids n=384 Manzanita 0.13 * 0.40 *** 0.17 ** 0.15 ** Oaks + - - -Pines + -0.24 *** -0.17 ** -0.11 * H C l Manzanita 0.27 * 0.28* + 0.27 * Oaks - - + + Pines - + - -HC2 Manzanita + + Oaks 0.34 ** + + 0.32 ** Pines + + - -HC3 Manzanita + 0.29* _ + Oaks + - + -Pines + - - -HC4 Manzanita 0.26* 0.43 *** 0.37 ** + Oaks - -0.27 * - -Pines + - - -L C Manzanita 0.35 ** 0.31* 0.27 * + Oaks + - - -Pines - - -0.25 * + MC Manzanita + 0.29 * _ + Oaks 0.28* + - -Pines - - - + 51 were positively correlated with manzanita cover (Spearman rs=0.18, n=384, p<0.001).Both adult males (Spearman rs=-0.17, n=384, p<0.01) and juvenile females (Spearman rs=-0.12, n=384, p<0.05) were negatively correlated with pine cover. Again, variability among grids was high. Adult females were positively correlated with oak cover on grid HC2 (Spearman rs=0.31, n=64, p<0.05). Adult males were positively correlated with manzanita cover in two grids (HC3: Spearman rs=0.25, n=64, p<0.05; MC: Spearman rs=0.28, n=64, p<0.05). There were no significant correlations between juveniles of either sex and any vegetation component of the understory. TWINSPAN divided the 384 trap stations into two microhabitats. Their description is as follows: M i c r o h a b i t a t I : Overstory dominated by medium to high manzanita cover (40 to 80%) and low to high oak cover (20 to 80%), with a few junipers, madrones and pines interspersed. Understory mostly with both low manzanita cover and low oak cover (< 20%). This microhabitat was further divided in l a , with high manzanita cover and low pine cover, and l b , with lower manzanita cover and medium oak cover. M i c r o h a b i t a t I I : Overstory dominated by medium to high pine cover (> 60%), high oak cover (> 80%) and medium to low manzanita cover (< 50%). Understory with low cover of manzanita, pines and oaks (< 20%). 52 Figure 3.1. Microhabitat availability. Figure 3.1. Microhabitat availability (number of traps) on the 6 grids. Microhabitat I with overstorey dominated mostly by manzanita. I a has high manzanita cover and low pine cover; I b has low manzanita cover and medium oak cover. Microhabitat II with overstorey dominated mostly by oaks and pines. 53 The relative abundance of microhabitats varied among grids. Microhabitat I was more abundant (70%) than microhabitat I I on four grids (HC1-4), similar to on grid LC, or less abundant (<20%) on grid MC (Fig. 3.1). Microhabitat use by demographic classes Microhabitats were not used randomly by rock mice, with the exception of grid HC3. Contingency table analysis showed significant heterogeneity (Gh) among demographic classes in their microhabitat use, and significant deviations (Gp) from the expected frequencies (Table 3.3). Most significant contributions occurred during the breeding season. Adult males and juveniles had significantly more captures in microhabitat I , whereas transients used significantly more microhabitat I I . During the non-breeding season only adult males had a significantly non-random microhabitat use. They were captured more often in microhabitat I (Table 3.4). Results from individual grids were heterogeneous but some trends were evident. Adult females had more significant contributions than other groups. Adult females were captured more often in microhabitat I on most grids (HCl, LC, MC). However, they used microhabitat I I significantly more on grid HC2. Numbers of adult male captures were proportional to habitat availability in the breeding season, but males consistently used microhabitat I. Males had a significantly higher use of microhabitat I on two grids (LC, HC4) during the non-breeding season. 54 Table 3.3. Microhabitat use from pooled data. T a b l e 3 . 3 . Microhabitat use by demographic classes from all grids. Replicated Goodness of fit tests. d.f.= degrees of freedom. * p<0.5, ** p<0.01, *** p<0.001. P o o l e d H e t e r o g e n e i t y T o t a l d.f. 1 7 8 A l l 20.07 *** 19.26 ** 39.33 *** H C l 7.36 ** 8.02 15.38 * H C 2 6.98 ** 9.8 16.78 * H C 3 2.98 1.15 4.48 H C 4 9.13 ** 8.7 17.83 * L C 20.38 *** 37.1 *** 57.48 *** M C 13.34 ** 11.6 24.94 ** 55 Table 3.4. Microhabitat use by demographic classes. Table 3.4. Microhabitat use by demographic classes, of individual classes. Tendency to use microhabitat I microhabitat I I = -. *p<0.5, ** p<0.01, ** p<0.001. Breeding Season Residents Adult females Adult males Juveniles Transients Non-breeding Season Residents Adult females Adult males Juveniles Transients All H C l HC2 HC3 + + -5.2** + +4.0* + + + +4.4* + + + -6.9** - -4.9* + + +5.2** + +18.0** + - + + + - + + + + + Log-likelihood values = +. Tendency to use HC4 L C MC + +37.7*** -+ + + +9.6** + + -5.3** + + - +14.5*** +4.2* +6.5** + + + + +4.1* 56 Juveniles captures were distributed in proportion to microhabitat availability on most grids, except on grid HC4, where they had a significantly higher use of microhabitat I during the breeding season. Microhabitat use of transients changed seasonally. In the breeding season they had a significantly higher use of microhabitat I I in two grids (HC2, LC), whereas in the non-breeding season they were caught mostly on microhabitat I. The contribution was only significant on grid MC (Table 3.4). M i c r o h a b i t a t d i f f e r e n t i a t i o n a m o n g g r i d s . Although the replicated goodness of fit tests indicated significant heterogeneity only on grid LC, the analysis of contributions from individual demographic classes indicated that differences in microhabitat use occurred on most grids. According to the heterogeneity values, which represent the number of demographic classes with significantly non-random microhabitat use, grids could be separated into three groups. The highest differentiation occurred on grid LC. On grids H C l , HC2, HC4, and MC, one and two demographic classes had significant non-random distributions. Microhabitat differences were completely absent on grid HC3 (Table 3.4). I investigated whether the spatial variability in microhabitat use was 0 related to habitat heterogeneity (Bowers and Smith 1979), or population density (Morris 1984). As a measure of the degree of microhabitat differentiation, I used the value of the Heterogeneity G (Table 3.3). Habitat heterogeneity and the degree of microhabitat partitioning were not correlated 57 (Spearman r s=-0.09, d.f.=6, p>0.05). In fact, the highest differentiation i n microhabitat use occurred on grid L C , a site wi th very low habitat heterogeneity. S imi lar ly , there was no relation between average mouse density and the degree of microhabitat part it ioning i n either the breeding (Spearman r s=-0.09, d.f.=6, p>0.05) or the non-breeding season (Spearman r s=0.29, d.f.=6, p>0.05). S u r v i v a l i n d i f f e r e n t m i c r o h a b i t a t s . According to the analysis of microhabitat use based on capture intensity, adult males were associated wi th manzani ta cover and consistently used microhabitat I more than microhabitat I I . In turn, adult females used both manzanita cover and oak cover and had no definite microhabitat affinity. Since abundance may not indicate best habitat suitabil i ty (Van H o m e 1982, 1986), I investigated differences i n survival among individuals l i v ing i n different microhabitats. I classified individuals i n three classes according to which microhabitat they used; individuals that used either microhabitat I or I I exclusively, and individuals that used both microhabitats. Residence time differed significantly, among sexes and microhabitats (Two-way A N O V A , Sex F=7.19, d.f.=l,28, p<0.05; microhabitat F= 12.99, d.f.=2,28, p<0.001). Furthermore, there was a strong interaction between sex and microhabitat. (Two-way A N O V A , F=4.92, d.f.=2,28, p<0.015). Residence time was s imi lar for both males and females i n microhabitat I (Fig. 3.2). In contrast, females resided longer than males i n microhabitat I I . Both males and females that used both microhabitats had longer residence times than 58 Figure 3.2. Residence time i n different microhabitats. 160 0 ' ' ONLY I ONLY II BOTH I AND II MICROHABITAT USE Figure 3.2. Mean residence time (days) of adult males (o). adult females (#), and of juveniles Q). living in microhabitat I or II exclusively and using both microhabitats. Bars represent 95% confidence intervals. 59 those that used either microhabitat exclusively. Residence time of juveniles was similar among the three categories (ANOVA F=0.76, d.f.=2,13, p>0.05). M i c r o h a b i t a t b r e a d t h . To analyze both microhabitat breadth and overlap, I used the next hierarchical subdivision of TWINSPAN for microhabitat I. Microhabitat I I could not be subdivided because the resulting sample size was too small. Microhabitat breadth (B'^ ) was relatively high for all demographic classes of mice. There were no significant differences among any of the classes (ANOVA F=1.91, d.f.=7,40, p>0.05), but there were some trends. Resident females had smaller microhabitat breadth than resident males during both breeding and non-breeding seasons. Microhabitat breadth differences between these two groups were significant when seasons were pooled, (ANOVA, F=8.65, d.f.=l,22, p<0.01). Juveniles had similar microhabitat breadth to adults but less variability in the non-breeding season. Transient individuals had lower microhabitat breadth in the breeding season than other classes, but during the non-breeding season their microhabitat breadth increased and was similar to that of other classes. Both juveniles and transients had a tendency for larger microhabitat breadth in the non-breeding season (Fig. 3.3). M i c r o h a b i t a t o v e r l a p . Microhabitat overlap between demographic classes was relatively high for all groups, and for most groups overlap (L) was equal or larger than 1. Adult males and females had similar overlap in both seasons. Transients and both adult males and females had higher overlap 60 Figure 3.3. Microhabitat breadth for demographic classes. FEMALES MALES JUVENILES TRANSIENTS Figure 3.3. Hurlbert's microhabitat breadth for demographic classes in the breeding (#) and non-breeding seasons (o)- Squares represent means for pooled seasons. Bars represent 9 5 % confidence intervals. 61 Figure 3.4. Microhabitat overlap for demographic classes. FEMALES FEMALES MALES VS VS VS MALES JUVS JUVS FEMALES MALES JUVS VS VS VS TRANS TRANS TRANS Figure 3.4. Hurlbert's microhabitat overlap for demographic classes in the breeding (•) and non-breeding seasons (O). Bars represent 95% confidence intervals. 62 during the non-breeding season than during the breeding season. However, adult females and transients had a higher overlap in the non-breeding season than adult males and transients. In contrast, juveniles and both adult males and females had lower overlap in the non-breeding season. Juveniles and transients had the lowest overlap of all groups and it was similar in both seasons (Fig. 3.4). DISCUSSION If differences in microhabitat use among sex, age and resident classes are the result of behavioral interactions, then resident adults should occupy the most suitable habitats, while juveniles and transients should occupy less suitable habitats. In turn, since behavioral interactions and spacing behavior are most important during the breeding season (Sadleir 1965, Healey 1967), differences in microhabitat use should be most evident during this season. In this study differences in microhabitat use among sexes and ages were evident as were differences among resident and transient individuals. Furthermore, such differences were most evident during the breeding season. All demographic classes were more strongly correlated with overstory cover than with understory cover, as might be expected with an arboreal species. Microhabitat analysis also indicated seasonal changes in use, particularly among adult females and transients. 63 Differences in male survival coincided with microhabitat use. More males used microhabitat I than I I , and had higher survival in this microhabitat. Females, on the other hand, had good survival in both microhabitats, but were less consistent than males in their habitat use. Both sexes survived best when they used both microhabitats. This is probably due to the fact that manzanita shrubs and oaks are both important components in the habitat requirements of both males and females, particularly females. While manzanita shrubs provide mostly food and cover, oaks also provide nesting sites. Since manzanita and oak are negatively correlated, habitats with high cover might lack nest sites. Thus, areas which include both microhabitats might be most favorable. Differential microhabitat use by the sexes has been reported in Peromyscus maniculatus (Bowers and Smith 1979), P. leucopus (Morris 1984, Seagle 1985) and Sigmodon hispidus (Kincaid and Cameron 1985). Differential microhabitat use by age classes has been reported in Peromyscus maniculatus (Van Home 1982, Merkt 1981) and Sigmodon hispidus (Kincaid and Cameron 1985). On the other hand, some studies found no differential microhabitat use among sexes in Microtus pennsylvanicus (Morris 1984), and Peromyscus leucopus (Kaufman et al., 1985) or among sex, age and resident classes in four species of Heteromyid rodents (M'Closkey 1985). Unfortunately, methodological differences in the above studies make comparisons (iifficult. First, some researchers use both between and within-session live-trapping captures (Bowers and Smith 1979, Merkt 1981, Morris 64 1984, M'Closkey 1985). However, since multiple captures of the same individuals might not be independent some studies discarded non-independent captures in order to make statistical tests valid (Van Home 1982, Seagle 1985, Kincaid and Cameron 1985, this study). Second, the level of discrimination of demographic classes differs between studies. For example, demographic classes have been pooled in: transient and resident individuals (Bowers and Smith 1979, Van Home 1982, Morris 1984, Kincaid and Cameron 1985), sexes (Van Home 1982), or ages (Bowers and Smith 1979). Other researchers have analyzed only one season. In this study, I pooled sexes for juveniles and both sexes and ages for transient individuals. This practice increases sample size but likely obscures some of the patterns which we are attempting to investigate. V a r i a b i l i t y i n m i c r o h a b i t a t p a r t i t i o n i n g . Differences in microhabitat use are often documented in some habitats but not in others. For example, Bowers and Smith (1979) found that female Peromyscus maniculatus used moist areas, whereas males used more xeric areas in both Pinus ponderosa and Artemisia-Sarcobatus communities, but not in the most xeric Atriplex-Eurotia community. Morris (1984) found microhabitat differences among sexes of P. leucopus in both grassland and sumac habitats but not in forest or old-field habitats. Further, microhabitat differences in grassland habitat were evident only one out of two years. Seagle (1985) found intrasexual microhabitat differences also in P. leucopus in deciduous forest but not in cedar glade habitat. Van Home (1982) reported that differences in habitat 65 use between adult and juvenile P. maniculatus were more apparent in some grids than in others. Furthermore, some studies that found sex differences in microhabitat partitioning, reported higher selectivity for females (Bowers and Smith 1979, Seagle 1985), whereas other found males to be more selective (Merkt 1981, Morris 1984). Even within the same species, different studies have reported completely opposite results. Morris (1984) found higher use by males of microhabitats with greater protective cover, whereas Seagle (1985) found higher use of these microhabitats by females. In this study most classes showed consistent patterns, however, on grid LC adult females used microhabitat I and on grid HC2 they used microhabitat I I . Two hypotheses have been proposed to account for such spatial variability. First, microhabitat partitioning can only occur in heterogeneous habitats (Bowers and Smith 1979). Second, microhabitat partitioning is density dependent and occurs only at high densities (Van Home 1982, Morris 1984). In this study I found no relation between the degree of microhabitat partitioning and either habitat heterogeneity or density. Other authors have also found microhabitat differences in the least heterogeneous habitats (Morris 1984, Seagle 1985). In my study, the absence of these relations might be due to lower variability between sites. Since this species is a habitat specialist, differences between habitats and densities on different grids (Chapter 2) are very small compared to other studies which have 66 analyzed distinct habitats (Bowers and Smith 1979, Morris 1984, Seagle 1985). Microhabitat breadth and overlap. Published predictions regarding microhabitat breadth and overlap of dominant and subordinate individuals are conflicting. In a review of interspecific niche relations, Morse (1974) suggested that subordinate individuals have narrower niches and overlap less with dominants. On the other hand, Van Home (1982) predicted subordinate individuals (juveniles) to have broader niches, since forced emigration might disperse them into a wider range of habitats. It seems likely that the breadth of habitat used by subordinates will depend on relative habitat suitability. If the habitats available are at all suitable, the subordinates will increase their niche breadth in their attempt to avoid dominant individuals (decrease overlap). On the other hand, if suitable habitats are not available their niche will be smaller. In turn, habitat suitability might depend on species-specific requirements. Subordinates in generalist species might be able to use a wider range of habitats than those in specialist species. Predictions about temporal changes in overlap are also ambiguous. Less overlap is predicted for seasons when resources are scarce (winter). This prediction, however, does not take into account the behavior of the animals in relation to activities other than food acquisition, such as reproduction. For example, LLewellyn and Jenkins (1987) predicted that if interference competition is potentially important, both high population 67 density and reduced resources should result in reduced niche overlap. This pattern, however, would not be expected if seasonal changes in the behavior of the animals were taken into account. During the breeding season energetic demands of females increase because of the physiological costs of pregnancy and lactation (Millar 1989), and the cost of defending nest sites to prevent infanticide (Wolff 1989). At the same time, physiological demands of males may also increase due to the cost of maintaining access and exclusivity to females. However, information on energetic costs of reproduction in males is lacking at present. If the above is true, then during the breeding season while population density is the lowest and resources are presumably highest, the intensity of intraspecific competition might also be very high. Inversely, in the non-breeding season, spacing behavior might be less intense in spite of higher densities and potentially lower resources. The predictions would therefore be opposite to those of LleweUyn and Jenkins (1987). In spite of lower densities during the "good" breeding season, there will be small overlap, since agonistic interactions are more common. Aggression within and between species varies seasonally with reproduction (Sadleir 1965, Healey 1967, Turner and Iverson 1973, Turner et al., 1975), and it has been suggested that decreased aggression between species during the non-breeding season results in winter coexistence (Turner et al., 1975). In this study there was a tendency for larger microhabitat breadth among juveniles and transients in the non-breeding season. Also, overlap between resident adults and transients was higher. Juveniles, had overlaps 68 in the opposite direction to that predicted: they were higher in the breeding season and lower in the non-breeding season. Adult females had smaller breadths than males. Since this pattern was consistent between seasons it might be the result of some basic differences in selectivity by the sexes. Bowers and Smith (1979) suggested that differences in microhabitat use by sexes in Peromyscus maniculatus were due to their different requirements and social interactions. Dominant females may exclude males from using the best microhabitats. There is increasing evidence among other rodent species that females might have exclusive areas during the breeding season (Galindo and Krebs 1987, Kaufman and Kaufman 1989), which would affect the spatial distribution of other demographic classes. Alternatively, Morris (1984) suggested that differential microhabitat use might just be the result of intrinsic differences in habitat requirements by the sexes. Females have very high energy requirements in the breeding season and are likely to spend more time than males, hoarding or foraging, increasing their vulnerability to predation. Females should therefore, select safe nest sites. My results indicate that in this population of rock mice, behavioral interactions may influence microhabitat differentiation. First, differences were most apparent during the breeding season. Second, overlap between transients and adults of both sexes, but particularly females, was higher in the non-breeding season than in the breeding season. This was the result of seasonal changes in microhabitat use by transients. In turn, some differences seem to be related to sex-specific requirements. Manzanita 69 shrubs provide mostly protection, whereas oak trees provide nest sites as well. Higher affinity of females for oaks may be the result of their more stringent requirements for adequate nest sites; many nests were inside hollow oaks. Merritt (1974) suggested that large sized Peromyscus species, seem to have specialized nest site requirements. At present it is difficult to distinguish among the hypotheses, since to date all studies on small mammals have only described patterns. Future research should be directed toward testing these hypotheses experimentally. This might be done by removing demographic classes and monitoring shifts in microhabitat use by the remaining classes. Although removal experiments to investigate interspecific competition are common, the analogous experiments to investigate intraspecific competition, while equally feasible, are almost lacking. Studies such as Peters and Grubbs (1983) with downy woodpeckers and Desrochers (1989) with black-capped chickadees could serve as models. To my knowledge these are the only two examples in which the role of behavioral interactions has been shown to determine differences in the use of microhabitats by the sexes. These might be extended to demographic classes other than sex and carried over different seasons. If kinship were known, dividing demographic classes according to relatedness might clarify some of the patterns even further. 70 C h a p t e r 4 P O P U L A T I O N R E S P O N S E S T O C L U M P E D F O O D I N T R O D U C T I O N The dispersion, abundance, renewability, and seasonality of resources can influence the social interactions and spatial distr ibution of individuals (Clutton-Brock and Harvey 1978, Davies and Lundberg 1984, Ostfeld 1985). Effects of resource distr ibution on social systems have been proposed for several groups of animals ( Jarman 1974, Clutton-Brock and Harvey 1978, Rubenstein 1981, Ostfeld 1985). Social systems and social interactions, i n turn , may affect the demographic characteristics of a population (Dunbar 1985, Clutton-Brock and Albon 1985). Populations of Peromyscus species, l ike many other insectivorous and granivorous rodents (French et al., 1975), are relatively more stable than herbivorous rodents (Kaufman and Kaufman 1989, Montgomery 1989), that often fluctuate many-fold. Population stabil ity seems to be greatly influenced by social interactions (Sadleir 1965, Healey 1967, Terman 1968, Har land et al., 1979, Tai t t 1981, Lusk and M i l l a r 1989). For example, spacing behavior of adult males and females affects the distr ibution of individuals (Chapter 3), l imi ts recruitment of new individuals into the population (Sadleir 1965, Healey 1967, Fa i rba i rn 1977, M ihok 1979, Metzgar 1971, Tai t t 1981, Hansen 71 and Batzli 1978, Harland et al., 1979, Halpin 1981, Nadeau et al., 1981, Galindo and Krebs 1987) and inhibits maturation of young females (Lusk and Millar 1989). Female Peromyscus usually overlap less and have smaller home ranges than males (Wolff 1989). Dispersion and individual interactions vary seasonally (Sadleir 1965, Lusk and Millar 1989). In the breeding season resources are most abundant, but the behavioral interactions increase with reproductive activity. In the non-breeding season, resources decrease and breeding stops. Several authors have suggested that breeding females should be highly responsive to environmental conditions (Boonstra 1977, Taitt 1981, Fordham 1971, Ostfeld 1985, Anderson 1989). Lactation requires an increase from 74 to 147% over the energy requirements of non-lactating individuals for Peromyscus (Glazier 1985, Millar 1989). The behavior of reproductive females should reflect their attempt to insure access to sufficient food resources. I investigated the effect of supplementary food on individuals and populations, to test the hypothesis that females are more responsive to food availability than males and that these differences are more pronounced during the breeding season. Food was supplemented on a sample of home ranges of individual male and female Peromyscus during both breeding and non-breeding seasons. The following predictions were tested; 1) Clumped food should most strongly affect the body weight, residence time and reproduction of resident adult females, compared with adult males; and 2) 72 these changes should be particularly evident in the wet season. I also investigated whether reproduction, recruitment, survival and population trends were affected by the highly clumped food addition. METHODS The study took place from May 1987 to July 1988 in southwestern Durango, Mexico (23° 25' N; 104° 15' W) in a dry temperate oak-pine forest. See Chapter 2 for detailed habitat descriptions. Experimental design. I used two different experimental designs to analyze individual and population responses to food provisioning. To analyze individual responses I used two grids (I, II). Unlike most food addition experiments which provide food near every other trap station, I placed food stations near the most-used trap station in the home range of half of the individuals in each grid. In this design individuals are considered experimental units. Individuals were chosen randomly with the restriction that food stations should not be in contiguous home ranges. Therefore, most food stations were widely spaced (Fig. 4.1). I considered experimental individuals (fed) to be those that used the closest trap to each food station, control individuals (non-fed) to be those never caught in these traps. Only one of a total of 51 fed individuals survived from the wet to the dry season and was considered experimental in both seasons. 73 Figure 4.1. Exper imental design. GRID I 0 o O 0 0 • o o O 0 0 o 0 o o o CONTROL O 0 O 0 o o O 0 0 o 0 o o o 0 0 GRID II 0 o 0 0 0 0 o o O 0 O 0 ° • o o 100 m 0 0 O 0 0 o o o • 00 0 0 0 • 0 o o o o o o 0 o o o 0 o o o 0 o 0 0 O 0 O 0 O 0 o o o o 0 o O 0 O 9 0 0 0 0 0 0 0 o o o 0 0 0 0 0 0 0 0 0 • o 0 0 0 0 o o 0 0 0 o o o o o 0 • o 0 o o 0 0 0 O 0 0 o o o o • ° o o o 0 o • • o O 0 O 0 o o 0 o O 0 O 0 m o 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 o o o o o o o o o o o o o 0 0 0 0 0 0 0 0 0 0 0 0 0 0 * 0 0 0 0 0 ^ 0 # 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 Figure 4.1. Experimental design. Grids were more than 600 m apart. Small circles represent trap stations, Full circles indicate position of food stations during the wet season. Stippled area in control grid represents 1 hectare. 74 The objective of this design was to eliminate pseudoreplication by interspersing experimental and control units. Furthermore, this design allows more replication than when grids are used as experimental units (Hurlbert 1984). Grids were analyzed individually, since the duration of experimental manipulations differed and so did the population dynamics of rock mice.. To analyze population responses I considered grids as experimental units. The control grid was situated between the two experimental grids, I and II, 600 m and 1 km from them. No movements between grids were documented. This design achieved neither replication nor interspersion, and suffered from pseudoreplication. Consequently I opted for a conservative approach and used no inferential statistics in the comparisons (Hurlbert 1984). Instead I searched for consistent differences between both pretreatment and treatment periods within and between experimental grids. Accurate population estimates were provided by the complete enumeration technique since both maximum and minimum trappabilities (Hilborn et al. i976, Krebs and Boonstra 1984) were above 64% for all areas. The three large grids (I, II, control) were in oak-manzanita shrubland in May 1987 and were trapped until July 1988. Each grid had 136 Longworth traps (8 x 17) set at 20 m intervals and cover an area of 5.4 ha. Some traps were placed on the ground and some were fixed with wide elastic bands between 1 - 2 m high on the closest trees and were baited with whole oats. Live-trapping methods have been described elsewhere (Chapter 2). 75 Individuals trapped for two or more sessions were considered residents and those trapped only once were considered as transients. I analyzed these two categories separately. Recruitment included resident individuals only. Food stations consisted of two 21 cans containing a mixture of sunflower seeds and whole-grained oats, nailed close to each other into the closest tree to trap station between 0.5 and 2 m high. Each can had a small entrance, to prevent birds from taking food. Two cans were provided to insure a constant food supply. They were refilled every two weeks but were rarely emptied before refilling. The experiments were carried out during both wet and dry seasons on both grids. Wet season. The rains begin in June and continue until November or December, corresponding with summer and fall. Pretreatment period lasted from May to July 27, 1987. Food stations were filled from then until December 5th on grid I and November 20th on grid II due to heavy trap disturbance. There were 13 food stations on grid I and 8 food stations on grid II. Dry season. The dry season spans from December to May corresponding to winter and spring. Pretreatment period lasted from January to April 7th 1988. Food stations were filled from then until July 13th on both grids. To avoid familiarity with locations from previous treatment, I moved all food stations one trap station away. In this period there were 13 food stations on grid I and 11 on grid II. 76 Sex ratios are shown as the proportion of females to the total population. I used contingency tables and log-likelihood tests (G) to compare sex, age and residency of control and experimental individuals. I used ANOVAs to analyze differences in residence time of experimental and control individuals (Sokal and Rohlf 1981). R E S U L T S In the following section I present first population responses to food addition, followed by responses from fed and unfed individuals, for wet and dry seasons. Wet season (June to December) Density and seasonal patterns. Rock mouse populations are usually very low at the end of the dry season (Apr-Jun) and during the first half of the wet season (Jun-Aug) (Chapter 2). They increase during the second half of the wet season (Sept-Oct) and continue increasing steadily throughout the first half of the dry season (Nov-Jan). Food addition began at the end of July, before populations began to increase, and lasted until December, when they began to decline (Chapter 2). Pretreatment numbers were similar on grid II and the control grid, and were 1.6 times higher on grid I. After treatment began, numbers on both experimental grids increased at a higher rate, reaching twice the numbers of the control during September, and began declining in October (Fig. 4.2). In 77 Figure 4.2. Population trends. Figure 4.2. Population trends of experimental (1:0; ll:fu) and control (•) areas. Stippled rectangles indicate periods of food addition. 78 contrast, the population on the control gr id did not begin increasing unt i l September and reached h igh numbers at the end of October and November. A t this point a l l three populations had simi lar numbers. The annual f luctuation on the control population was s imi lar to that of unmanipulated grids i n the previous year (Chapter 2). On grid II an excessive number of traps was disturbed by predators i n November causing a temporal decline i n the population and the experiment was ended on this gr id. Sex ratios on both experimental grids were even both during the pretreatment period (I, II, x=0.5) and after food addition (I x=0.47; II x=0.52). Sex ratio on the control gr id was more biased towards males both dur ing pretreatment (x=0.41) and treatment (x=0.44). The number of fed individuals was 35 on gr id I and 16 on grid II, out of a total of 196 and 97 individuals caught i n each (17.9% and 16.5% respectively) dur ing the treatment period. Age and sex composition of fed indiv iduals on both grids did not differ from the whole population (grid I: G=0.58, d.f.=3, p>0.05; gr id II: G=3.72, d.f.=3, p>0.05). However, there were clear sex differences i n the temporal response. Whereas pretreatment sex ratios on both grids were close to 1:1, sex ratio of fed individuals became highly skewed toward females soon after food was provided. Males began to use food stations more readily on gr id I by the end of September and the sex ratio fluctuated from close to 1 to 2. O n grid II, sample sizes of fed individuals were too smal l to obtain reliable sex ratios but the trend was s imi lar (Fig. 4.3). Most fed individuals on both grids were residents 79 Figure 4.3. Sex ratio of fed and total population. i 1.0 0.21 • 1 1 • i • i • — — i — — ' 1 •— J A S O N D Figure 4.3. Proportion of females in the population (empty symbols) and among fed individuals (full symbols), in grid I (circles), and grid II (square) in the wet season. Stippled area indicates treatment period. Sessions with less than five individuals were eliminated. 80 (1=77%; 11=81%). On both grids there were slightly more residents using food than expected by the composition of the population but the differences were not significant (G's < 2.5, d.f.=l, p>0.05). Changes in body weight. Fed and unfed females had similar weights at the beginning of the treatment on grid I. By mid-September fed females were heavier than unfed females and continued to be heavier throughout the treatment period. On grid II, fed females were heavier right from the start of treatment, but by October the difference between fed and unfed females diminished (Fig. 4.4). On both grids unfed males were slightly heavier than fed males at the beginning of the experiment. The trend reversed after mid-September, and fed males were heavier than unfed males throughout the rest of the treatment period. The difference in weight was particularly evident in October and November (Fig. 4.4). Differences in recruitment. Pretreatment recruitment was similar on grids II and control, and slightly lower on grid I. After food addition recruitment increased on both experimental grids. However, this response was evident only during the first half of the treatment period (Fig. 4.5). In August and September, there were 24 recruits in each experimental grid, but only 13 on the control grid. The sex composition of recruits was very similar among grids but the age composition differed substantially. Experimental grids recruited 7 (I) and 2.4 (II) times more adults than subadults and juveniles together. In contrast, the control grid recruited 2.3 times more juveniles and subadults than adults. During October and November there 81 Figure 4.4. Changes in body weight. Figure 4.4. Changes in body weight of fed (•) and un-fed (O) males and females during the wet season. Bars represent 1 standard error. 82 Figure 4.5. Recruits and transients in control and experimental grids. Figure 4.5. Number of recruits and transients in control (o) and experimental grids (I •;!!•). Stippled areas indicate treatment periods. 83 was little difference in recruitment between the three grids (15 and 14 recruits on experimental grids and 18 on the control grid). Before food addition there were very few transients on any of the three grids. After food addition the number of transients increased on all grids, including the control. However, more transients were caught on both experimental grids and their age composition differed from those on the control. On the experimental grids, 72% (I) and 41% (II) of transients were adults, but only 15% of the transient individuals were adults on the control grid. C h a n g e s i n s u r v i v a l . Females on all 3 grids had very high survival during the pretreatment period. During the treatment period, female survival was slightly lower on all grids due to a slowly declining general trend through the fall (Fig. 4.6). Considerable reductions in survival occurred in September and October on grid I, and after November on both experimental grids. Average two-week female survival during the treatment period was slightly higher on the control (x=0.87) than in the experimental grids (I, x=0.85; II, x=0.79). Males also had high survival on all 3 grids in the pretreatment period. There was also a slowly declining trend in survival in the fall. During the treatment period male survival also declined in September and October on all three grids. It continued to decline after November on both experimental grids, but not on the control grid. Average two-week male survival during the treatment period was higher on the control grid (x=0.93) than on experimental grids (I x=0.86, II x=0.74). In summary, food addition had no 84 Figure 4.6. Survival in control and experimental grids in the wet season. Figure 4.6. Two week survival of females and males on experimental grids (I •; II •) and control grid (O) during the wet season. Stippled areas indicate treatment periods. 85 apparent effect on survival of either males or females during the wet season (Fig. 4.6). Average residence times of both fed males and fed females were 1.5 times longer than those of unfed individuals on grid I (Two-way ANOVA F=4.95, d.f.=l,59, p<0.05). On grid II, fed females stayed 1.6 times longer than unfed females, but the difference was not statistically significant (ANOVA F=1.82, d.f.=l,24, p>0.05). Average residence time of fed males on grid II was slightly lower than that of unfed males. No statistical test was carried out since only 3 resident males used food stations (Fig. 4.7). Effects on reproduction. During pretreatment, more than 50% of the females were breeding on all three grids. After food addition, a high percentage of females continued breeding on both experimental grids but not on the control grid. In this period fewer than half of the females were in breeding condition in only 1 and 2 of the 8 trapping sessions on grids I and II respectively. In contrast, fewer than half of the females were breeding in 5 sessions on the control grid. During the pretreatment period, 5 females were breeding on the control grid, whereas only 2 were breeding on both experimental grids. Most females were recorded as pregnant only once during this period. Only one female on the control grid was pregnant two times. After food addition, there was a reversal in the trend. The number of females breeding on the experimental grids (1=20; 11=23) was twice as high as on the control grid (n=10). During this time some females had from one to 86 Figure 4.7. Residence time of fed and unfed individuals in the wet season. 140 UN-FED FED UN-FED FED GRID I GRID II Figure 4.7. Average residence time (days) of un-fed and fed individuals in grids I and II in the wet season. Circles represent females and squares represent males. Bars indicate 1 standard error. Sample sizes in parenthesis. 87 three pregnancies. The average number of pregnancies in 4 months of treatment was slightly larger on grid I (x=1.7), than on grid II (x=1.3) and on the control grid (x=1.4). During the pretreatment period more than 50% of the males were breeding on all three grids. More males were breeding on experimental grids during most sessions of the treatment period. During this period fewer than half of the males were in breeding condition in only 2 and 3 of the 8 trapping sessions on grids I and II respectively. In contrast, fewer than half of the males were breeding in 5 sessions on the control grid. More fed (1=86%; 11=89%) than unfed females (1=71%; 11=79%) on both grids were pregnant. The average number of pregnancies per female was also higher among fed females on both grids. On grid I, fed females had an average of 1.86 pregnancies in 4 months (s.e.=0.312, n=14) and on grid II, the average was 1.3 in 4 months (s.e.=0.289, n=9). Unfed females had an average of 1.1 pregnancies on grid I (s.e.=0.2, n=17) and 1.2 on grid II (s.e.=0.175, n=19). The difference was statistically significant only on grid I (F=4.95, d.f.=l,29, P<0.05). The proportion of males breeding was similar among fed and unfed individuals on grid I (G=0.02, d.f.=l, p>0.05). There were too few fed males on grid II to assess differences. D r y s e a s o n ( D e c e m b e r t o M a y ) D e n s i t y a n d s e a s o n a l p a t t e r n s . Rock mouse populations usually decline from December to March and have low densities during late spring and early 88 summer (Chapter 2). When the treatment began population trends and densities differed markedly among grids. Populations on experimental grid I and on the control grid were low during the pretreatment period (Fig. 4.2). Numbers did not change after food addition, but followed a similar pattern on both grids, slightly increasing in the summer. In contrast, the population on experimental grid II was 4.5 times higher than control densities and was gradually declining from peak numbers in February. After food addition numbers continued declining slowly throughout April and stabilized in May at about 3 times the other two grids. On both experimental grids, average sex ratios were slightly biased toward females before food addition (I, x=0.6, II x=0.63). After food was added, average sex ratios became more biased toward males (I, x=0.38) or were more even (II, x=0.55). Sex ratio on the control grid was more biased towards males both during pretreatment (x=0.34), and treatment (x=0.33). There were 9 fed individuals on grid I and 12 on grid II, out of a total of 37 and 72 individuals caught in each grid respectively during the treatment period. No juveniles were caught on grid I during the dry season. In terms of age and sex there was no difference between the composition of fed individuals and that of the whole population on grid I (G=0.01, d.f.=l, p>0.05). On grid II both juveniles and adults were caught. The age and sex composition of fed individuals did not differ from that of the population (G=0.61, d.f.=13, p>0.05). Most fed individuals on grid I (78%) and all on grid 89 II were residents. These proportions were similar to those in the whole population (I G=3.15, d.f.=l, p>0.05; II G=1.49, d.f.=l, p>0.05). Changes in body weight. There were too few individuals on grid I to assess reliably differences in body weight between fed and unfed individuals. On grid II, unfed and fed females were very similar, but fed males were slightly heavier than unfed males after 1.5 months. Sample sizes were very small and variability was high. Differences in recruitment. Pretreatment recruitment was similar on grid I and the control grid, and very high on grid II. Recruitment during the treatment period was higher on both experimental grids than on the control grid. The response, however was evident only during the first two months of the experiment. In April and May, there were 12 and 9 recruits on grids I and II respectively, but only 4 on the control grid. In June and July, there were 3 and 8 recruits on grids I and II, but 11 on the control grid (Fig. 4.5). No juveniles and very few subadults recruited during April and May, and therefore only differences in sex composition were analyzed. Most individuals recruited in the experimental grids were males (proportion of females: I, x=0.25, II, x=0.33). In contrast, female recruits outnumbered males on the control grid (x=0.75). The number of transients during pretreatment was low on grids II and control grid, and higher on grid I. At this time most transient individuals were adults. During treatment the number of transients decreased on the 90 control grid, and on grid II, but increased on grid I. During this period all transients were adults. There was no evident sex-bias among adults. Changes in survival. Female survival during the pretreatment period was lower on both experimental grids than on the control grid. During the treatment period, average two-week female survival increased on both experimental grids, but was slightly higher on the control grid (x=0.92) than on experimental grids (I, x=0.71; II, x=0.88). This difference was due, however, to a decline in June. Males also had low two-week survival during pretreatment on grids I and control grid. After the addition of food, male survival improved on grid II. Male survival on grid I was already very high and continued to be so. Both experimental grids had higher survival than the control until June, when survival decreased to the level of the control. Average male survival during the treatment period was slightly higher on both experimental grids (I, II x=0.9) than on the control grid (x=0.84) (Fig. 4.8) . On grid I, average residence times of fed and unfed males were very similar, but fed females stayed 2.1 times longer than unfed females. Sample sizes, however, were too small for statistical analysis. On grid II fed females stayed slightly longer (1.2 times) than unfed females. Fed males on this grid had longer residence time (1.5 times) than control males, but the difference was not statistically significant (ANOVA F=1.823, d.f.=l,24, p>0.05) (Fig. 4.9) . 91 Figure 4.8. Survival in control and experimental grids in the dry season. Figure 4.8. Two-week survival of females and males in experimental grid (l#; II •) and control grid (O) during the dry season. Stippled areas indicate treatment periods. 92 Figure 4.9. Residence time of fed and unfed individuals in the dry season. HI 120 100 (3)(8) (4)(3) (25)(20) (7)(3) LU o z LU 9 cn LU CC 80  60 < LU 2 40 20 _L J i L t UN-FED FED GRID I UN-FED FED GRID II Figure 4.9. Average residence time (days) of un-fed and fed individuals in grids I and II in the dry season. Circles represent females and squares represent males. Bars indicate 1 standard error. Sample sizes in parentheses. 93 Effects on reproduction. No females were breeding during the pretreatment period on any of the 3 grids. Breeding began in early June on grid II and in the second half of June on grid I and control grid. On all three grids fewer than half of the females were breeding in June and July, and only one pregnancy occurred in any grid. Males in breeding condition were recorded in late May on experimental grid I and in June on the other two grids (II, control). During July more than half of the males were breeding on all three grids. The treatment period began at the end of the dry season, before breeding usually occurs (Chapter 2). Only one female on each grid, both of them fed individuals, showed signs of pregnancy (June 21 and 22). No unfed females showed signs of pregnancy until middle July. There were no obvious differences in the breeding condition of fed and unfed males, however the low numbers of fed males on both precluded analysis. D I S C U S S I O N Several researchers studies have proposed that females should be more responsive to food resources than males, particularly during the breeding season, when behavioral strategies among sexes may diverge the most (Boonstra 1977, Taitt 1981, Fordham 1971, Ostfeld 1987, Ims 1987a,b, Anderson 1989, Galindo and Krebs 1987). The results from this study support this idea. Females responded most intensely to food additions. 94 Females who used food stations had improved reproduction and improved survival, during both wet and dry seasons. Males used food stations less than females, and they gained weight and had improved survival, but less consistently than females. Although male reproductive activity seemed less affected, reproductive performance was difficult to quantify in males. Most previous food addition studies of populations (Fordham 1971, Boonstra 1977, Hansen and Batzli 1978, Taitt 1981), and more detailed studies of individuals (Ostfeld 1986, Ims 1987a,b, this study) have shown that the distribution and abundance of artificial food might differentially affect the spatial distribution, reproduction and survival of the sexes. Very few have used natural food distributed naturally. Recently, Glendinning and Brower (1990) monitored populations of Peromyscus melanotis both inside and outside overwintering aggregations of monarch butterflies. Monarch butterflies represent a natural high protein food source for this rodent which is adapted to their high cardenolide content. Mouse populations inside aggregations were denser, had a higher proportion of adult breeding females than those outside. These results substantiate previous studies which have provided artificial supplementary food. Reproductive requirements induce females to be more responsive to increased food. The high energetic demands of breeding females should increase both their total time spent foraging and the distance from the nest during this period. Both will increase the susceptibility of pups to infanticide (Wolff and Cicirello 1989). These opposite forces will be particularly strong 95 during lactation when energetic requirements are higher and pups are very vulnerable. Exclusive access to food through territoriality might allow females to increase their reproductive effort reducing both their foraging time and distance, and reducing the risk or losing offspring to potential infanticide (Wolff and Cicirello 1989). This link between food and female reproduction would have consequences for the distribution of males (Boonstra 1977, Ostfeld 1985) and for processes affecting recruitment, survival and maturation of young (Taitt 1981, Hansen and Batzli 1978, Galindo and Krebs 1987, Lusk and Millar 1989). In this study the effects of food addition during the breeding season were highly biased toward adults both for recruits and transients. Glenndining and Brower (1989) also found that juvenile residency was lower in the butterfly sites where populations were denser and female-biased. Resource levels might alter female strategies to weaning young. When food is plentiful and consistently available in time and space, females may increase reproductive fitness by increasing the number of litters and prompting juvenile dispersal. Under these conditions more young find places to settle. This would occur in years, seasons or locations of high productivity, during food bonanzas or at the beginning of the breeding season when conditions are most favorable. In contrast, females might tolerate their offspring to remain in the area, perhaps long enough to inherit it from them under less favorable conditions, such as in low productivity areas or late in the breeding season. There is some evidence that females with subsequent 96 litters are more likely to induce dispersal of their previous litters (Savidge 1974), and that later litters of a breeding season are more likely to recruit into the population (Hansen and Batzli 1978). Boutin (1990) underscored two consistent outcomes of populations to experimental food additions: there is often an increase from 1.5 to 2.5 fold, but they also decline in the presence of food. In this study, I expected negligible effects at the population level. Since food was provided in widely spaced and highly clumped food sources. Food stations were located at only 10% of the traps on grids I and at 8% of traps on grid II in areas of 5.4 ha. Most food addition studies provide many food stations, often one at every other trap station (Gilbert and Krebs 1981, Taitt 1981, Wolff 1985, Briggs 1986). The food addition in the wet season surprisingly produced strong effects in the population dynamics of rock mice, consistent with the first general outcome. Populations on both experimental grids increased as a result of the food addition. Both reproduction and adult recruitment improved. However, higher recruitment of adults on experimental grids than control grids suggest that the response on experimental grids was due more to immigration than to local recruitment of juveniles. The wet season results also support Boutin's second general outcome, because both experimental grids reached higher densities than the control and then declined to control levels despite the added food. The dry season experiments are more difficult to interpret. Food was added when the populations on both the control grid and experimental grid I 97 were already low after a gradual decline throughout winter. In contrast, the population on grid II had undergone an unusual increase during the dry season (winter) prior to food addition and was very dense. This population also declined somewhat during food addition, but remained approximately 3-fold higher than the other two populations and also 3-fold higher than its previous early summer density. It is difficult to know if the moderate decline and the high summer density were due to the food addition, since the control population was already low. However, both increased recruitment and survival on this grid after food addition indicate a plausible effect of the supplementary food. The puzzling question is: why did the population increase during the dry season when no food was added? All three grids had similar trends during the previous dry season and early wet season. All previously documented populations of rock mice in my study area declined over the dry season (winter-spring) (Chapter 2). The increase throughout the dry season in this study resulted from very high recruitment. Recruits may have been the juveniles which resulted from higher breeding intensity during the food addition experiment in the wet season. In turn, they might have been individuals that immigrated to the grid as a consequence of the accidental removal of a substantial fraction of the resident population. In mid-November the population declined sharply after traps were found disturbed. This observation supports the idea that factors other than food prevent 98 further increases (Boutin 1990), and thus, food addition experiments coupled with removal experiments might be particularly useful (Klenner 1991). Most food supplementation experiments with terrestrial vertebrates have assessed the effects of food addition at a population level (Boutin 1990). Supplemental food provided evenly over large areas often results in home range reductions by both sexes (Taitt and Krebs 1981, Taitt 1981, Mares et al, 1982, Sullivan et al, 1983) or by females only (Taitt et ah, 1981, Ostfeld 1986). However, individuals may monopolize food resources (Clutton-Brock and Albon 1985, Boutin 1990), especially when these are clumped, and therefore the benefits of the food provided may not be distributed evenly in the population (Boutin 1990). In this study I analyzed both general population responses and responses of individuals who were likely to have received the benefits of the food addition. As in other food supplementation studies on relatively stable rodent populations (Peromyscus and Apodemus), in this study I documented increased densities through immigration, earlier breeding, and increased reproduction after food addition (Boutin 1990). Furthermore, as in other studies I detected no effect on general summer survival (Boutin 1990), although fed individuals did show longer residence times during this period. This suggests that the effects of food provisioning might not be obvious at the population level even when some individuals responded. This is likely to occur if the food is not shared equally by the population, but monopolized by some individuals (Clutton-Brock and Albon 1985, Boutin 1990). Since individuals are most likely to respond to the local 99 abundance and distribution of food resources by altering their movements and other behavior (Clutton-Brock and Harvey 1978, Davies and Lundberg 1984, Ostfeld 1985), the spatial distribution of food addition is of major interest. Most studies, however, have largely ignored the distribution of supplementary food. Recently, two studies have closely examined behavioral responses of small mammals to spatial patterns of food distribution (Ostfeld 1986, Ims 1987a,b). Ostfeld (1986) monitored movements of California voles before and after adding fresh carrots evenly to one 100 m 2 patch during 10 days on each of two years. Although he reported territoriality in males but not in females, the food was probably added over too great an area to detect territoriality in females, since their home ranges were of similar size or smaller than the food patch. Using smaller patches of natural food in different spatial patterns and over longer time might be more appropriate. Ims (1987a,b) investigated the effects of the distribution of food on dispersion and dominance in Clethrionomys rufocanus. He added a mixture of whole-grained oats, corn and sunflower seeds at point sources on a small island. Females with additional food had reduced home ranges and increased overlap. In spite of the highly clumped food distribution, females did not behave territorially as predicted by Ostfeld (1985). Both of the above studies examined responses to only one type of food distribution. Further experiments should compare food distributions, preferably using foods similar to the natural food, and over longer periods, to 100 document the population consequences. The study of Noyes et al., (1982) on behavioral responses by feral house mice to food distribution in small enclosures could be used as a model for manipulations in more natural conditions. In that study, clumped food had a pronounced effect on the social structure of house mice. Dominant individuals monopolized food stations with strong consequences on the overall demography. Enclosures with clumped food had lower population densities and lower proportion of breeding individuals than those with dispersed food (Stueck and Barrett 1978). The majority of food addition studies have not incorporated replicate treatments (Boutin 1990). Appropriate replication is particularly difficult for small mammal studies which are highly dependent on live trapping grids to obtain demographic information. Experimental manipulations of small mammals using highly spaced individuals as experimental units, as in the present study, overcome the problem of pseudoreplication and increase the limited number of replicates provided by live trapping grids. 101 C h a p t e r 5 E C O L O G I C A L C O N S E Q U E N C E S O F S E X D I F F E R E N C E S : B O T F L Y I N F E S T A T I O N O F R O C K M I C E . I N T R O D U C T I O N Asymmetr ical parental investment between males and females results in sex-related differences i n behavior (Krebs and Davies 1981). Female mammals i n particular, continue their investment long after the young are born. Differences i n behavior of males and females may influence the r isks they are exposed to such as increased susceptibility to predation, starvation, parasit ism, injuries and diseases. L i t t le is known about the ecological consequences of different behaviors at the indiv idual and population levels (Clutton-Brock et al, 1982, Sadleir 1984). Recently, some studies have documented microhabitat differences between male and female small mammals (references i n Chapter 3), and also differences i n their spatial distr ibution (Montgomery 1989). Males have often larger home ranges than females and male-biased dispersal is the common pattern among mammals. These differences may result i n sex-related r isks. 102 For example, differential susceptibility to predation between the sexes may be a consequence of differences in their movements (Daly et al., 1990). Botfly parasitism might be a sex-related risk due to differences in the use of space and dispersal by the sexes. Botflies spend their larval stages as subcutaneous parasites of North American small mammals, feeding on living tissue to obtain energy and protein for the adult stage. Each species is highly specific, infesting only one or two host species even when others are available. Infestations are highly seasonal (July to December; Dunaway et al., 1967) and highly variable both temporally and spatially, in their prevalence and intensity. While physiological changes in infected individuals are documented (Sealander 1961, Payne et al., 1965, Bennett 1973), their effect on populations is not clear (Wecker 1962, Clough 1965, Goertz 1966, Miller and Getz 1969, Iverson and Turner 1969, Getz 1970, Hunter et al., 1972, Timm and Cook 1979, Boonstra et al., 1980, Munger and Karasov 1991). A number of authors have proposed that botfly parasitism of small mammals is significantly biased towards males (Catts 1982, but see Xuhua and Millar 1990). It has been suggested that since infestation requires that mice contact botfly eggs, greater movements by males increase the probability of encounter (Catts 1982, Xuhua and Millar 1990). However, since botflies lay their eggs in, or close to their host nests (Bennett 1973, Catts 1967, 1982, Capelle 1970, Timm and Cook 1979), individuals that confine their activities near the nest might be more exposed to infestation. In 103 this case, females will be more exposed since they have to remain near the nest during the breeding season. In this chapter I describe the seasonal patterns of botfly infestation of Peromyscus difficilis. I tested the hypotheses that there is a male bias in botfly infestation, and that infestation is related to increased movement. In turn, I analyzed sex and age temporal patterns in infestation. Finally, I assessed the effect of botfly parasitism on survival of infested mice. METHODS The study took place from February 1986 to July 1988 in southwestern Durango, Mexico (23° 25' N; 104° 15' W) in a dry temperate oak-pine forest. Detailed descriptions of habitats and grid characteristics are presented elsewhere (Chapter 2). I sampled six 2.6 ha live-trapping grids (HCl, HC2, HC3, HC4, LC, MC) from March 1986 to May 1987 and three 5.6 ha grids (control, I, and II) from June 1987 to July 1988. On two grids (I, II) I experimentally investigated sex-specific responses to highly clumped food, using food stations near 21 of 198 traps from August to December 1987 (Chapter 3). Food stations consisted of two bins containing a mixture of sunflower seeds and oats ad libitum, nailed into the closest tree between 0.5 and 2 m high. Each had a small entrance to prevent birds from taking food. Traps were set every two or three weeks and were checked every morning for two days. They were locked open after every trap session. 104 Individuals trapped were ear-tagged, and released, and in every trapping session I recorded the following information: sex, reproductive condition, weight (nearest gram) and trap station. Reproductive condition of males was assessed by the position of testes (scrotal vs. abdominal) and of females by nipple size, perforated vagina, and evident pregnancies. Each individual was assigned to one of two age categories: Adults (sexual maturity or > 22 g) and juveniles (grey pelage and < 22 g). I regarded those individuals trapped for two or more sessions as residents and those trapped only during one session as transients. I also recorded the number and position of any warbles or scars. Botflies were only identified up to genus Cuterebra sp. I obtained an index of changes in movements using the mean distance between the traps where mice were caught in consecutive trapping sessions. Each individual contributed with only one distance per season. For this analysis, I used the movements of individuals before I detected any signs of botfly infection, excluding from analysis any individuals already infested at first capture. I estimated sex and age biases of infestation by log-likelihood tests (G tests). ANOVAs were used to analyze differences in movements (log(x+l) transformation) and standardized residence times (arcsine transformation). (Sokal and Rohlf 1981). For most analyses the data were pooled from all grids within years, since sample sizes were very small. 105 R E S U L T S During both years all warbles occurred on the back and flanks of the mice. Unlike most reports on warble location for Peromyscus leucopus (references in Xuhua and Millar 1990), I found no warbles near the inguinal region. S e a s o n a l i t y o f i n f e s t a t i o n . In 1986 botfly warbles first appeared during the second week of September, then increased through November, declined slightly in December, and disappeared after February (Fig. 5.1). In 1987 infestation began earlier in a few individuals in May and August, but the infestation followed the same pattern as the year before during fall and winter (Fig. 5.1). I n f e s t a t i o n r a t e s . Number of warbles per infested host ranged from 1 to 4. In 1986, 78% the of infested individuals had one warble, 15% had two, 5% had three. Only one individual (2%) had 4 warbles. A total of 28 females and 27 males had botfly warbles. Most infested individuals were adults, both among females (79%) and among males (74%). In 1987, 68% of the infested individuals had one warble, 23% had two and 9% had three. Individuals with more than three warbles were not encountered. A total of 22 females and 25 males were infested with botfly larvae. Again, most infected individuals were adults: 91% of the females and 84% of the males. Infestation rates were very similar between the sexes. In 1986 females had a mean infestation rate of 1.29 larvae per infected individual (s.e.=0.101, 106 Figure 5.1. Seasonality of bot fly infestation. 1986 I J A I A U J . A -1987 N M CO _J < o > Q Z Q UJ h-W HI U. 20 - 10 u_ O z o r-rr O Q. o DC Q. 20 10 A M J J Figure 5.1. Annual pattern of bot fly infestation in Peromyscus difficilis during 1986, and 1987. Stippled bars indicate females. Empty bars indicate males. 107 n=28), whereas males had a rate of 1.37 (s.e.=0.152, n=27). In 1987 females had a slightly higher mean rate of 1.41 larvae (s.e.=0.107, n=22) whereas males had a slightly lower mean of 1.24 (s.e.=0.087, n=25). These differences in infestation rate between sexes (Two-way ANOVA, F=0.13, d.f.=l, p>0.05) or years (F=0.001, d.f.=l, p>0.05) were not significant. E f f e c t o f f o o d a d d i t i o n o n i n f e s t a t i o n r a t e s . During 1987,1 recorded 47 infested individuals, most (85%) on experimental grids (I, II). The low number (7) of infested individuals on the control grid was similar in the previous year, which probably reflects habitat characteristics of this site (Chapter 2, 3). All infested individuals on the control grid had only one warble. During the infestation period (August to December), 21 of 198 traps on experimental grids were near food stations (Chapter 4). Infested individuals were caught preferentially at food stations (G=12.6, d.f.=l, p<0.001) (Fig. 5.2). The 15 of 40 individuals on grids HC3 and II with more than one larva were caught preferentially at traps near food stations (G=10.5, d.f.=l, p<0.001) (Fig. 5.2). Infested individuals that used traps near food stations had an average of 1.62 larvae (s.e.=0.180, n=13), whereas infested individuals that did not use trap near food stations had a mean of 1.37 larvae (s.e.=0.143, n=27). Among infested individuals on experimental areas, females (x=1.63, s.e.=0.175, n=19) were slightly more infested than males (x=1.29, s.e.=0.101, 108 Figure 5.2. Utilization of food stations. 4 0 WITHOUT FOOD Figure 5.2. Utilization of traps with and without nearby food stations by infested individuals, and by infested individuals with more than one warble. Stippled bars indicate observed values. Open bars indicate expected values. 109 n=21). The difference was nearly significant (Weighted Means Analysis, F=3.08, d.f.=l,38, p=0.087). T e m p o r a l p a t t e r n s i n s e x a n d a g e r e l a t e d i n f e s t a t i o n . Males and females had similar infestation rates over the whole infestation season, but the temporal pattern of their infestation was different. In 1986, 82% of females with botflies were recorded during autumn and 18% in winter. Conversely, males were infested during both autumn (52%) and winter (48%). Infestation was not randomly related to sex or age. In the fall adult females were more heavily infested (G=12.2, d.f.=3, p<0.01), and juveniles of both sexes were less infested than expected by their proportions in the population (Fig. 5.3). In winter there was also a significant sex bias in botfly infestation. However, this time adult males were more heavily infested (G=9.8, d.f.=3, p<0.025), while adult females and juveniles were less infested than expected by their relative proportions in the population (Fig. 5.3). The pattern was similar in 1987. All infested females were recorded during autumn and none in winter. Males were infested earlier than the previous year, with a few infested individuals during May and August (8%). Most males, however, were infested in autumn (76%). The remainder (16%), were infested in winter. Similarly to the previous year, in 1987 botfly infestation was not proportional to sex and age classes. During autumn, adult females were most heavily infested, but the bias was not significant (G=7.3, d.f.=3, p>0.05) (Fig. 5.3). During winter, only adult males were found infested (Fig. 5.3). 110 Figure 5.3. Temporal patterns of sex and age biases in infestation. 20 15 10 5 S 0 15 10 5 •\"S II AUTUMN 1986 I f AUTUMN 1987 WINTER 1986 WINTER 1987 | m i l l JZL 12 FEMALES MALES FEMALES MALES FEMALES MALES FEMALES MALES ADULTS JUVENILES ADULTS JUVENILES Figure 5.3. Temporal patterns of sex and age biases in infestation during autumn and winter 1986 and 1987. Stippled bars indicate observed values. Empty bars indicate expected values. i n I n f e s t a t i o n a n d b r e e d i n g c o n d i t i o n . Rock mice reproduced from June to December (Chapter 2). Botfly warbles appeared during the second half of the breeding season, coinciding with recruitment of juvenile and subadult individuals. In fact, most infested individuals were fall recruits (79% of female fall recruits; 63% of male fall recruits). The rest were either summer recruits (11% females; 11% males) or winter recruits (10% females; 26% males). Infestation was unrelated to breeding condition of adults of both sexes in either year (Log-likelihood tests, all p > 0.05). M o v e m e n t s a n d i n f e s t a t i o n . Previous studies have suggested that individuals that move more are more susceptible to botfly infestations (review in Catts 1982, but see Xuhua and Millar 1990). I therefore investigated the relation between movements and infestation rates. Movements differed seasonally between males and females (Repeated Measures ANOVA, F=15.01, d.f.=l,10, p<0.005). Male movements increased in the breeding seasons (summer 1986, spring 1987), whereas female movements decrease at this time (Fig. 5.4). Comparison of season by season indicated significantly smaller movements for adult females in fall (t-test t=2.43, d.f.=9, p<0.05; fig. 5.4), when they were more infested. Likewise, males had significantly smaller movements during winter (t-test t=2.82, d.f.=7, p<0.05; fig. 5.4) and were also more infested. However, comparison of movements of infested and healthy individuals before the appearance of warbles did not indicate differences. Movements of 112 Figure 5.4. Seasonal changes in movements. x 0 I i ' 1 • 1 • i ' i SPRING SUMMER AUTUMN WINTER SPRING Figure 5.4. Home range index of males (O) and females (•) throughout the seasons. Smalle female ranges occurred during breeding in summer 1986, and spring 1987. 113 infested individuals of either sex and healthy individuals were similar during autumn (two-way ANOVA F's <1.6, d.f.=l,132, p>0.05). During winter, no infested females fulfilled the requirements for analysis, and there was no difference in the movements of infested and non-infested males (ANOVA F=0.81, d.f.=l,26, p>0.05). Effects on survival. I used residence time on grids as an index of survival to evaluate the effect of botfly larvae on mice. I included in this analysis individuals recruited during fall and winter, comprising most infested individuals in the population. I excluded infested adult individuals recruited during summer to minimize possibly spurious variability in residence time. Since I stopped trapping most areas during May 1987, residence time was standardized, dividing it among the potential residence times. Both infested females and males had significantly longer residence times than non-infested individuals (two-way ANOVA F=5.85, d.f.=l, p<0.05). However, the results changed drastically when transient individuals were excluded. There was a significant interaction between infestation, sex and year (three-way ANOVA, F=10.91, d.f.=l,116, p<0.001). In 1986, infested adult and juvenile females resided slightly less time on grids than non-infested females, whereas infested adult and juvenile males had longer residence times than those of non-infested males. The pattern was reversed during 1987 (Fig. 5.5). 114 Figure 5.5. Residence time of infested and non-infested individuals. FEMALES MALES 1.0 0.8 -0.6 -UJ I— 0.4 -UJ O Z 0.2 UJ Q (75 UJ 0.8 0.6 0.4 0.2 ADULTS JUVENILES FEMALES -> 1 1 1 ADULTS JUVENILES MALES 1986 1987 Figure 5.5. Standardized residence time of infested and non-infested individuals in 1986, and 1987. Infested individuals (•). Non-infested individuals (O). Bars represent 2 standard errors. 115 Spatial variability in infestation rate. Did grids differ in infestation rates? In 1986 from 4% to 27% of individuals were infested. On grids HC2 (17%), HC3 (23%), HC4 (13%), and LC (24%); the differences were not significant (G=7.1, d.f.=3, p>0.05). Grid H C l had more infested individuals than expected (27%) by the relative abundance of mice in this area, whereas grid MC had a smaller proportion (4%) (G=11.6, d.f.=5, p<0.05) (Fig. 5.6). In 1987 the proportion of individuals infested ranged from 10% to 29%, and grids also differed (G=26.9, d.f.=2, p<0.001). Grid HC3 (29%) had more whereas grid MC (10%) and grid II (11%) had fewer infested individuals than expected by the numbers of mice present in these areas (Fig. 5.6). On grids studied both years, there were consistent trends of infestation. Grid HC3 had a relatively high proportion of infested mice during both years. In turn, grid MC had the lowest proportion of infested individuals during both years. Furthermore, there were also differences in infestation between microhabitats (Chapter 3). More infested females lived exclusively in microhabitat I than in I I (G=16.1, d.f.=l, p<0.001). Similarly, most infested males lived exclusively in microhabitat I (G=4.3, d.f.=l, p<0.05). Microhabitat I had higher manzanita (Arctostaphylos pungens) cover than microhabitat I I , where pine cover was higher but the understory cover was more open (Chapter 3). 116 Figure 5.6. Spat ia l var iabi l i ty i n infestation rate. LU I-< 20 15 10 1986 I-< I-00 LU LL 30 25 20 15 10 5 1987 HC1 HC2 HC3 HC4 LC MC Figure 5.6. Spatial variability in infestation rate in 1986, and 1987. Stippled bars represent observed proportion. Empty bars represent expected proportions. 117 DISCUSSION As in other studies, botfly infestation was highly seasonal. At higher latitudes infestation usually begins in July and continues until November or December (Dunaway et al., 1967); here it began in September and continued until January and February, with peaks in October and November. Since high temperature and low humidity inhibit larval development (Catts 1982), seasonality of botfly infestation is partially determined by rainfall. However, rains begin in June in the study area, so there is a delay period of several months until botflies are present. There is no information on the life cycle of this species of Cuterebra, but according to the generalized cycle of the family (Catts 1982) the larvae appear at development sites about 1.5 months after adults emerge. Therefore, if adults become active with the rains, warbles should appear during late July and August. Yearly, infestation rate was similar among females and males. However, when the data were separated into seasons a different pattern emerged. There were seasonal differences in infestation rates between sexes. Adult females were more heavily infested in the fall, whereas adult males were more infested in winter. This pattern was present during both years. Sealander (1961) found a male-biased infestation in Peromyscus leucopus. His analysis did not include the first part of the breeding season and did not include reproductive females. This pattern of male-biased infestation is in 118 agreement with my results since male biases should be more evident at the end of the breeding season. The hypothesis that botfly infestation is male-biased (Catts 1982, Xuhua and Millar 1990) was only partially supported in this study since male bias occurred only in winter. Female-biased infestation occurred in the fall. There is actually little evidence for sex biased infestations in the literature. Recently, Xuhua and Millar (1990) reviewed studies dealing with sex-related botfly infestations of Peromyscus leucopus; only 2 of 11 studies showed a significant male bias and some actually showed female biases. They suggested that the inconsistency of patterns of botfly infestation could be due to the common practice of pooling data from age classes, years and populations to increase sample size due to low infestation rates and low population densities of Peromyscus. Is botfly infestation related to motility? Some authors have proposed that mobile individuals should be more infested than sedentary ones (Timm and Cook 1979, Catts 1982, Xuhua and Millar 1990). Consequently, males should be more infested than females because they move more (Xuhua and Millar 1990), subadults should be more infested than adults for the same reason (Catts 1982) and juveniles should be less than adults (Xuhua and Millar 1990). Differential movements between the sexes have also been proposed as an explanation for higher infections of chiggers on male Microtus californicus and for higher number of ticks on male Apodemus sylvaticus (Healing and Nowell 1985). 119 Contrary to these expectations, I found either no relation between infestation rates and movements or a negative relation. Movements of healthy and infested individuals previous to the appearance of the botfly warble did not differ. However, females moved less during the breeding season and had higher infestation rates, while males moved less in the non-breeding season and had higher infestation during this time. Furthermore, this trend was supported by the patterns documented in the experimental grids, where higher infestation rates occurred among individuals, particularly females, who used traps near food stations. Experimental studies have documented reductions in female home ranges after food additions (Ostfeld 1987, Ims 1987a). The hypothesis that infestation results from higher motility was not supported, and I suggest that the opposite prediction is more likely. Individuals that move less and remain closer to the nest should be more susceptible to infestation, since the oviposition behavior of Cuterebrids is closely associated with specific host activities. Eggs are laid not randomly, but at entrances to or inside burrows, nests or runways (Bennett 1973, Catts 1967, Catts 1982, Capelle 1970, Timm and Cook 1979). Mice living in an area tend to use established burrows and may attract gravid female bot flies (Hunter et al., 1972). The maximum prevalence of botfly larvae (Rogenhofera bonaerensis) on the pampean mouse (Akodon azarae) occurred during the spring when rodent activity is close to the nest (Zuleta and Vignau 1990). Glicken and Schwab (1980) suggested flea reinfestation occurs more rapidly 120 in female Peromyscus maniculatus because they spend more time in the nest. Differential movements between the sexes have also been proposed as an explanation for higher infections of chiggers on male Microtus californicus and for higher number of ticks Ixodes trianguliceps on male Apodemus sylvaticus (Healing and Nowell 1985). In this study I documented a temporal reversal in infestation between sexes. Females were infested in the early part of the season whereas males were more infested in the last part of the season. It is not possible to assess temporal changes in infestation among sexes from the literature, since studies have either pooled sexes across seasons (Goertz 1966, Childs and Cosgrove 1966, Dunaway et al., 1967, Hunter et al., 1972) or have sampled only a very small fraction of the potential infestation period (Sealander 1961, Test and Test 1943, Xuhua and Millar 1990). The results of this study suggest that in addition to the problems underscored by Xuhua and Millar (1990), the practices of pooling seasons and presenting information from a small portion of the potential infestation period might obscure sex related patterns of infestation. Other ecological interactions may also be influenced by differences in the spatial distribution and movements of male and females. Daly et al., (1990) proposed a similar seasonal reversal in sex differential susceptibility to predation of kangaroo rats. During the breeding season male kangaroo rats are more mobile than females and suffer greater predation. Since life 121 tables of both males and females are very similar, females appear to suffer greater mortality outside the breeding season. Spatial patterns of infestation support the idea that botfly infestations are influenced by temperature and moisture (Catts 1982). Mice populations on areas with more cover had higher infestations. In turn, most infested individuals, both males and females, used exclusively microhabitats with higher cover. The effect of botfly parasitism on survival and reproduction of rodent populations seems highly variable. Survival of infested mice has been documented as worse than (Miller and Getz 1969, Boonstra et al., 1980), similar to (Iverson and Turner 1969, Hunter et al., 1972), or even better than (Wecker 1962, Goertz 1966, Getz 1970) that of non-infested mice. Likewise, reproduction in infested mice might be reduced (Clough 1965, Iverson and Turner 1969, Boonstra et al., 1980), or might be similar to non-infested mice (Getz 1970, Timm and Cook 1979). In this study, reduced survival was evident only in females. In contrast, during the food addition, infested females had higher survival than healthy females, since most of them were using food stations. This suggest that the pathology of botfly parasitism might be contingent on the physiological state of the hosts. Because of their physiological state females may be more susceptible while in breeding condition than when they are not breeding. 122 Differences in the impact of botfly infestations among small mammal species have been interpreted as related to the degree of adaptation of the host-parasite interaction. Parasites tend to lose virulence as the interaction evolves. Catts (1965) found greater site specificity in the location of larval warbles on native hosts and greater pathological effects on non-native host species. Some authors have suggested that Microtus species (M. pennsylvanicus, M. townsendii, M. montanus) and botflies may be assumed to have a shorter history, since they usually show lack of site specificity of larval warbles (Getz 1970, Hunter and Webster 1973, Boonstra et al, 1980). Therefore, the pathological effects should be greater. In turn, Timm and Cook (1979) proposed that the lack of negative effects of botflies on Peromyscus sp., is due to a coevolved relationship. Botfly warbles had little specificity in P. difficilis. However, Catts (1982) indicated that hosts may be killed by aberrant movements of larva during their internal migration to the warble site. All studies that have investigated the effect of infestation on host populations, including this one, compared survival of infested hosts once the warble was evident. The absence of differences in survival rates might lead to erroneous conclusions, since the critical stage might occur before the warble appears. Further studies might assess the impact of botfly parasitism by manipulative experiments. For example, prevention of infestations with anti-parasitic drugs (Lambin in process) might be more promising than removal of larvae from infested individuals (Munger and Karasov 1991). 123 C h a p t e r 6 G E N E R A L C O N C L U S I O N S Both habitat structure and food influenced the population dynamics of rock mice. Demographic classes responded differently to both factors and had also different susceptibility to botfly parasit ism. In particular, breeding females were strongly influenced by food and habitat structure. Populat ion characteristics of rock mice are related to differences i n vegetation characteristics supporting the hypothesis that demographic parameters of habitat specialists are closely associated wi th habitat structure. Higher breeding densities, low juvenile and subadult recruitment in the breeding season and higher within-year stabil ity were found i n the most heterogeneous habitat (Chapter 2). Fur ther studies should investigate demographic correlates by manipulat ing habitat characteristics. Microhabitat use was strongly correlated wi th overstory cover, but var ied w i th demographic class and season. Degree of microhabitat part i t ioning was unrelated to either habitat heterogeneity or density. Juveniles and transients had larger microhabitat breadth i n the non-breeding season, and overlap between resident adults and transients was higher. Smal ler microhabitat breadth by adult females might be the result of 124 basic differences in selectivity by the sexes. Survival varied between microhabitats (Chapter 3). Several investigators have interpreted differences in microhabitat use by age, sex and resident classes as the result of social interactions. However, few studies have attempted to test this idea (Peters and Grubbs 1983, Desrochers 1989). Experimental removal of sex, age, or residence classes might enable us to distinguish between this and other hypotheses. Females responded most intensely to food additions. They had improved reproduction and improved survival during both wet and dry seasons. Food addition resulted in increased immigration, earlier breeding, and increased reproduction. The improvement in summer survival was evident only when analysis was undertaken at a finer scale: fed individuals showed longer residence times than unfed individuals (Chapter 4). Most experimental food additions have ignored the effect of spatial distribution of supplemented food. Further studies should investigate the effect of different food dispersion patterns on individual and population characteristics. Adult females were more heavily infested with botflies in the fall, whereas adult males were more infested in winter, during both years. Infestation rates were negatively related to motility. Higher infestation rates in both sexes coincided with periods of smaller movements. Infestation was also higher among individuals, particularly females, who used traps near food stations (Chapter 5). 125 M e t h o d o l o g i c a l p r o b l e m s Throughout this thesis I have underlined several methodological problems which hinder the detection of individual and population patterns in studies of small mammals. Many studies are restricted to summer and fall (Chapters 2, 3, 5). Demographic classes (transients and resident individuals, sexes, or ages) are often pooled in the analysis (Chapter 3, 5). Other problems are related to methodology of live-trapping. Researchers often underestimate animal movements and use trapping areas (< 1 ha) that are too small to document population phenomena (Chapter 2). Non-independent multiple captures are commonly used to increase sample sizes (Chapter 3). Finally, use of trapping grids as experimental units restrains replication (Chapter 4). I attempted to overcome these problems in several ways. I studied the population continuously through breeding and non-breeding seasons for 2.5 years. I used relatively large trapping areas, subdividing the population into demographic classes whenever possible. I eliminated all but the first capture of individuals during trapping sessions. Finally, I designed the experiments to utilize both subpopulations and areas as experimental units. 126 LITERATURE CITED Adler G.H. and M.L. Wilson. 1987. Demography of a habitat generalist, the white-footed mouse in a heterogeneous environment. Ecology 68:1785-1796. Alvarez T. and O.J. Polaco. 1984. Estudio de los mamiferos capturados en La Michilia, sureste de Durango, Mexico. An. Esc. Nac. Cienc. Biol. Mex., 28:99-148. Anderson P.K. 1970. Ecological structure and gene flow in small mammals. Symp. Zool. Soc. Lond. 26:295-325. Anderson P.K. 1989. Dispersal in rodents: a resident fitness hypothesis. Am. Soc. Mammal. Spec. Publ. N. 9. 139pp. Barry W.J. 1976. Environmental effects on food hoarding in deermice (Peromyscus). J. Mamm. 57:731-746. Bell S.S., E.D. McCoy and H.R. Mushinsky. 1990. Habitat structure: the physical arrangement of objects in space. Chapman and Hall. London. 438 pp. Bennett G.F. 1973. Some effects of Cuterebra emasculator Fitch (Diptera: Cuterebridae) on the blood and activity of its host, the eastern chipmunk. J. Wildl. Dis. 9:85-93. Bierzychudek P. and V. Eckhart. 1988. Spatial segregation of the sexes of dioecious plants. Am. Nat. 132:34-43. Bondrup-Nielsen S. 1987. Demography of Clethrionomys gapperi in different habitats. Can. J. Zool. 65:277-283. Boonstra R. 1977. Effects of conspecifics on survival during population declines in Microtus townsendii. J. Anim. Ecol. 46:835-851. Boonstra R. and C.J. Krebs. 1978. Pitfall trapping of Microtus townsendii. J. Mamm. 59:136-147. Boonstra R., C.J. Krebs, and T. Beacham. 1980. Impact of botfly parasitism on Microtus townsendii populations. Can. J. Zool. 58:1683-1692. Boutin S. 1990. Food supplementation experiments with terrestrial vertebrates: patterns, problems, and the future. Can. J. Zool. 68:203-220. 127 Bowers M.A. and H.D. Smith. 1979. Differential habitat utilization by sexes of the deermouse, Peromyscus maniculatus. Ecology 60:869-875. Briggs J.M. 1986. Supplmental food and two island populations of Peromyscus leucopus. J. Mamm. 67:474-480. Brown J.H. and G.A. Lieberman. 1973. Resource utilization and coexistence of seed-eating desert rodents in sand dune habitats. Ecology 54:788-797. Capelle K.J. 1970. Studies in the life history and development of Cuterebra polita (Diptera: Cuterebridae) in four species of rodents. J. Med. Entomol. 7:320-327. Carleton M.D. 1989. Systematics and Development. In: Kirkland G.L. and J.N. Layne (eds.). Advances in the study of Peromyscus (Rodentia). Texas Tech Univ. Press, pp.7-141. Catts E.P. 1965. Host-parasite interrelationships in rodent botfly infections. Trans. 30 th N.A. Wildlife and Nat. Resources Conf. 30:184-196. Catts E.P. 1967. Biology of a California rodent botfly Cuterebra latifrons Coquillett (Diptera: Cuterebridae). J. Med. Entomol. 4:87-101. Catts E.P. 1982. Biology of New World bot flies: Cuterebridae. Annual Review of Entomology. 27:313-338. Childs H.E. Jr. and G.E. Cosgrove. 1966. A study of pathological conditions in wild rodents in radioactive areas. Am. Midi. Nat.: 76:309-324. Clough G. 1965. Physiological effect of botfly parasitism on meadow voles. Ecology 46:344-46. Clutton-Brock T.H. and P.H. Harvey. 1978. Mammals, resources and reproductive strategies. Nature 273:191-195. Clutton-Brock T.H. and S.D. Albon. 1985. Competition and population regulation in social mammals. In: Sibly R.M. and R.H. Smith. (Eds.). Behavioural Ecology. Ecological consequences of adaptive behaviour. Blackwell Scientific Publications, pp.557-575. Clutton-Brock T.H., E. Guiness and S.D. Albon. 1982. Red Deer. Behavior and Ecology of Two Sexes. University of Chicago Press. Chicago. U.S.A. 378pp. Cody M.L. 1985. An Introduction to habitat selection in birds. In: M.L. Cody (Ed.). Habitat selection in birds. Academic Press: 3-56.. 128 Cody M.L. 1985. Habitat selection in birds. Academic Press, Inc. U.S.A. pp.558. Comely J.E., D.J. Schmidly, H.H. Genoways and R.J. Baker. 1981. Mice of the genus Peromyscus in Guadalupe Mountains National Park, Texas. Occ. Papers Mus. Texas Tech Univ. 74:1-35. Daly M., M. Wilson, P.R. Behrends and L.F. Jacobs. 1990. Characteristics of kangaroo rats, Dipodomys merriami, associated with differential predation risk. Animal Behaviour 40:380-389. Davies N.B. and A. Lundberg. 1984. Food distribution and a variable mating system, in the dunnock Prunella modularis. J . Anim. Ecol. 53:895-912. De Laet J.V. 1985. Dominance and aggression in juvenile great tits, Parus major major L. in relation to dispersal. In: R.M. Sibly and R.H. Smith. Behavioural Ecology. Ecological consequences of adaptive behaviour. British Ecological Society. Blackwell Scientific Publications. Oxford. pp.375-380. Desrochers A. 1989. Sex, dominance, and microhabitat use in wintering Black-capped Chickadees: a field experiment. Ecology 70:636-645. Dunaway P.B., J.A. Payne, L.L. Lewis and J.D. Story. 1967. Incidence and effects of Cuterebra in Peromyscus. J. Mamm. 43:38-51. Dunbar R.I.M. 1985. Population consequences of social structure. In: Sibly R.M. and R.H. Smith. (Eds.). Behavioural Ecology. Ecological consequences of adaptive behaviour. Blackwell Scientific Publications. pp.507-519. Fairbairn D.J. 1977. The spring decline in deer mice: death or dispersal? Can. J. Zool. 55:84-92. Fordham R.H. 1971. Field populations of Peromyscus with supplemental food. Ecology 52:138-146. French N.R., D.M. Stoddart and B. Bobek. 1975. Patterns of demography in small mammal populations. In: Golley F.B., K. Petrusewicz and L. Ryszkowski (eds.). Small Mammals: their productivity and population dynamics, pp.73-102. Fretwell S.D. 1972. Populations in a seasonal environment. Princenton Univ. Press. Princeton, New Jersey. Galindo C. and C.J. Krebs. 1987. Population regulation in deer mice: the role of females. J. Anim. Ecol. 56:11-23. 129 Galindo-Leal C. 1991. Live-trapping vs. snap-trapping of deer mice: a comparison of methods. Acta Theriologica 35:361-367. Gauch H.G. 1982. Multivariate analysis in community ecology. Cambridge University Press. New York. 2198pp. Getz L.L. 1970. Botfly infestations in Microtus pennsylvanicus in southern Wisconsin. Am. Midi. Nat. 84:187. 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. Gilliam J.F. and D.F Fraser. 1988. Resource depletion and habitat segregation by competitors under predation hazard. In: Ebenman B. and L. Persson (Eds.). Size-structured populations: ecology and evolution. Springer Verlag. Berlin, pp: 173-184. Glazier D. 1985. Energetics of litter size in five species of Peromyscus with generalizations for other mammals. J. Ma mm. 66:629-642. Glendinning J.I. and L.P. Brower. 1990. Feeding and breeding responses of five mice species to overwintering aggregations of the monarch butterfly. J. Anim. Ecol. 59:1091-1112. Glicken A. and R.G. Schwab. 1980. Modes of ectoparasite re-infestations of deer mice (Peromyscus maniculatus). J. Wildl. Dis. 16:577-586. Goertz J.W. 1966. Incidence of warbles in some Oklahoma rodents. Am. Midi. Nat. 75:242-245. Hall R. 1981. Mammals of North America. 2nd ed. Ronald Press. New York. Halpin Z.T. 1981. Adult-young interactions in island and mainland populations of the deer mouse Peromyscus maniculatus. Oecologia 51:419-425. Hansen L.P. and G.O. Batzli. 1978. The influence of food availability on the white-footed mouse: populations on isolated woodlots. J. Mamm. 60:335-342. Harland R.M., P.J. Blancher and J.S. Millar. 1979. Demography of a population of Peromyscus leucopus. Can. J. Zool. 57:323-328. Healey M.C. 1967. Aggression and self-regulation of population size in deermice. Ecology 48:377-392. 130 Healing T.D. and F. Nowell. 1985. Diseases and parasites of woodland rodent populations. Symp. Zool. Soc. Lond. 55:193-218. Hilborn R., J.A. Redfield and C.J. Krebs. 1976. On the reliability of enumeration method for mark and recapture census of voles. Can J. Zool. 54:1019-1024. Hoffmeister D.F. 1986. Mammals of Arizona. University of Arizona Press : Arizona Game and Fish Dept. Arizona 602pp. Holbrook S.J. 1978. Habitat relationships and coexistence of four sympatric species of Peromyscus in northwestern New Mexico. J. Mamm. 59:18-26. Hunter D.M. and J.M. Webster. 1973. Aggregation behavior of adult Cuterebra grisea and C. tenebrosa (Fiptera: Cuterebridae) in deer mice. Can. Entomol. 105:1301-1307. Hunter D.M., R.M.F.S. Sadleir and J.M. Webster. 1972. Studies on the ecology of cuterebrid parasitism in deer mice. Can. J. Zool. 50:25-29. Hurlbert S.H. 1984. Pseudoreplication and the design of ecological field experiments. Ecol. Monog. 54:187-211. Ims R.A. 1987a. Responses in spatial organization and behaviour to manipulations of the food resource in the vole Clethrionomys rufocanus. J. Anim. Ecol. 56:585-596. Ims R.A. 1987b. Determinants of competitive success in Clethrionomys rufocanus. Ecology 68:1812-1818. Iverson S.L. and B.A. Turner. 1969. The effect of Cuterebra spp. on weight, survival and reproduction in Microtus pennsylvanicus. Manit. Entomol. 2:70-75. Jameson E.W. 1950. Determining fecundity in male small mammals. J. Mamm. 31:433-436. Jarman P.J. 1974. The social organization of antilope in relation to their ecology. Behaviour 48:215-267. Jones S.W. 1985. Habitat selection in raptorial birds. In: M.L. Cody (Ed.). Habitat selection in birds. Academic Press: 159-190. Jongman R.H., C.J.F. ter Braak and O.F.R. van Tongeren. 1987. Data analysis in community and landscape ecology. Pudoc Wageningen. The Netherlands. 131 Kaufman D.W. and G.A. Kaufman. 1989 . Population Biology. In: Kirkland G.L. and J.N. Layne (eds.). Advances in the study of Peromyscus (Rodentia). Texas Tech University Press, pp.233-270. Kaufman D.W., M.E. Peak and G.A. Kaufman. 1985. Peromyscus leucopus in riparian woodlands: use of trees and shrubs. J . Mamm., 66:139-143. Kincaid W.B. and G.N. Cameron. 1985. Interactions of cotton rats with a patchy environment: dietary responses and habitat selection. Ecology 66:1769-1783. Kirkland G.L. and J.N. Layne (eds.). 1989. Advances in the study of Peromyscus (Rodentia). Texas Tech University Press. U.S.A. Klenner W. 1991. Red squirrel population dynamics. II. Settlement patterns and the response to removals. J. Anim. Ecol. 60:000-000 Krebs C.J. 1989. Ecological methodology. Harper and Row, Publishers. New York. 654pp. Krebs C.J. and R. Boonstra. 1984. Trappability estimates for mark-recapture data. Can. J. Zool. 62:2440-2444. Krebs J.R. and N.B. Davies. 1981. An introduction to behavioral ecology. Sinauer Associates Inc. U.S.A. Krohne D.T. 1989. Demographic characteristics of Peromyscus leucopus inhabiting a natural dispersal sink. Can. J. Zool. 67:2321-2325. Krohne D.T. and R. Baccus. 1985. Genetic and ecological structure of a population of Peromyscus leucopus. J. Mamm. 66:529-537. Llewellyn J.B. and S.H. Jenkins. 1987. Patterns of niche shift inmice: seasonal changes in microhabitat breadth and overlap. Am. Nat. 129:365-381. Lusk S. J.G. and J.S. Millar. 1989. Reproductive inhibition in a short-season population of Peromyscus maniculatus. J. Anim. Ecol. 58:329-341. M'Closkey R.T. 1975. Habitat dimensions of white-footed mice, Peromyscus leucopus. Am. Midi. Nat. 93:158-167. M'Closkey R.T. 1985. Patterns of microhabitat use and seed collection within populations of heteromyid rodents. Can. J. Zool. 63:325-330. Mares M.A., T.E. Lacher, M.R. Willig, N.A. Bitar, R. Adams, A. Klinger, and D. Tazik. 1982. An experimental analysis of social spacing in Tamias striatus. Ecology 63:267-273. 132 Merkt J.R. 1981. An experimental study of habitat selection by the deer mouse Peromyscus maniculatus, on Mandarte Island B.C. Can. J. Zool. 59:589-597. Merritt J.F. 1974. Factors influencing the local distribution of Peromyscus californicus in northern California. J. Mamm. 55:102-114. Metzgar L.H. 1971. Behavioural population regulation in the woodmouse Peromyscus leucopus. Am. Midi. Nat. 86:435-447. Mihok S. 1979. Behavioral structure and demography of subarctic Clethrionomys gapperi and Peromyscus maniculatus. Can. J. Zool. 57:1520-1535. Millar J.S. 1989. Reproduction and development. In: Kirkland G.L. and J.N. Layne (eds.). Advances in the study of Peromyscus (Rodentia). Texas Tech University Press, pp. 169-232. Miller D.H. and L.L. Getz. 1969. Botfly infections in a population of Peromyscus leucopus. J . Mammal 50:277-283. Montgomery W.I. 1989. Peromyscus and Apodemus: patterns of similarity in ecological equivalents. In: Kirkland G.L. and J.N. Layne (eds.). Advances in the study of Peromyscus (Rodentia). Texas Tech University Press, pp.293-366. Morris D. 1984. Sexual differences in habitat use by small mammals: evolutionary strategy or reproductive constraints? Oecologia 65:51-57. Morse D.H. 1974. Niche breadth as a function of social dominance. Am. Nat. 108:818-830. Morse D.H. 1985. Habitat selection in North American Parulid warblers. In: M.L. Cody (Ed.). Habitat selection in birds. Academic Press: 131-158 Munger J.C. and W.H. Karasov. 1991. Sublethal parasites in white-footed mice: impact on survival and reproduction. 1991. Can. J. Zool. 69:398-404. Nadeau J.H., R.T. Lombardi, and R.H. Tamarin. 1981. Population structure and dispersal of Peromyscus leucopus on Muskeget Island. Can. J. Zool. 59:793=799. Noyes R.F., G.W. Barret and D.H. Taylor. 1982. Social structure of feral house mouse (Mus musculus L.) populations: effects of resource partitioriing. Behav. Ecol. Sociobiol. 10: 157-163. 133 Ostfeld R.S. 1985. Limiting resources and territoriality in microtine rodents. Am. Nat. 126:1-15. Ostfeld R.S. 1986. Territoriality and the mating system of California voles. J. Anim. Ecol. 55:281-301. Ostfeld R.S. and L.L. Klosterman. 1986. Demographic substructure in California vole population inhabiting a patchy environment. J. Mamm. 67:693-704. Ostfeld R.S., W.Z. Lidicker and E.J. Heske. 1985. The relationship between habitat heterogeneity, space use and demography in a population of California voles. Oikos 45:433-442. Parmenter R.R. and J.A. MacMahon. 1983. Factors determining the abundance and distribution of rodents in a shrub steppe ecosystem: the role of shrubs. Oecologia 59:145-156. Payne J.A., P.B. Dunaway, G.D. Martin, and J.D. Story. 1965. Effects of Cuterebra angustifrons on plasma proteins of Peromyscus maniculatus. J. of Parasitology 51:1004-1008. Peters W.D. and T.C. Grubbs. 1983. An experimental analysis of sex-specific foraging in the Downy Woodpecker Picoides pubescens. Ecology 64:1437-1443. Petticrew B.G. and R.M.F.S. Sadleir. 1974. The ecology of deer mouse Peromyscus maniculatus in a coastal coniferous forest. I. Population dynamics. Can. J. Zool. 52:107-118. Pulliam R.H. 1988. Sources, sinks and population regulation. Am. Nat. 132:652-661. Rosenzweig M.L. 1973. Habitat selection experiments with a pair of coexisting heteromyid rodent species. Ecology 54:111-117. Rubenstein D.I. 1981. Individual variation and competition in the everglades pygmy sunfish. J. Anim. Ecol. 50:337-350. Sadleir R.F.M.S. 1984. Ecological consequences of lactation. Acta Zool. Fennica 171:179-182. Sadleir R.M.F.S. 1965. The relationship between agonistic behaviour and population changes in the deermouse Peromyscus maniculatus (Wagner). J. Anim. Ecol. 34:331-352. Sadleir R.M.F.S. 1974. The ecology of deer mouse Peromyscus maniculatus in a coastal coniferous forest. II. Reproduction. Can. J. Zool. 52:119-131. 134 Savidge I.R. 1974. Social factors in dispersal of deer mice (Peromyscus maniculatus) from their natal site. Am. Midi. Nat. 91:395-405. Schoener T.W. 1977. Competition and the niche. In: C. Gans and D.W. Tinkle (Eds.). Biology of the Reptilia. Academic Press, London, pp.35-136. Seagle S.W. 1985. Patterns of small mammal microhabitat utilization in Cedar glade and deciduous forest habitats. J. Mamm. 66:22-35. Sealander J.A. 1961. Haematological values in deer mice in relation to botfly infection. J. Mammal. 42:57-60. Smith C.C. and O.J. Reichman. 1984. The evolution of food caching by birds and mammals. Ann. Rev. Ecol. Syst. 15:329-354. Smith M.H., M.N. Manlove and J.Joule. 1978. Spatial and temporal dynamics of the genetic organization of small mammal populations. In: D.P. Snyder. (Ed.) Populations of small mammals under natural conditions. Spec. Publ. Ser. Pymatuning Lab. Ecol. Univ. Pittsburgh Press, Pittsburgh, U.S.A. 5:99-113. Sokal R.R. and F.J. Rohlf. 1981. Biostatistics. 2nd ed.Freeman, San Francisco. Soule M. 1973. The epistasis cycle: a theory of marginal populations. Ann. Rev. Ecol. Syst. 4:165-187. Stueck K.L. and G.W. Barrett. 1978. Effects of resource partitioning on the population dynamics and energy utilization strategies of feral house mice (Mus musculus) populations under experimental field conditions. Ecology 59:539-551. Sullivan T.P. 1979. Demography of populations of deer mice in coastal forest and clear-cut Cogged) habitats. Can. J. Zool. 57:1636-1648. Sullivan T.P. and C.J. Krebs. 1981. An irruption of deer mice after logging of coastal coniferous forest. Can. J. For. Res. 11:586-592. Sullivan T.P., D. Sullivan and C.J. Krebs. 1983. Demographic responses of a chipmunk (Eutamias townsendii) population with supplemental food. J. Anim. Ecol. 52:743-755. Sutherland W.J. and G.A. Parker. 1985. Distribution of unequal competitors. In: R.M. Sibly and R.H. Smith. Behavioural Ecology. Ecological consequences of adaptive behaviour. British Ecological Society. Blackwell Scientific Publications. Oxford, pp.255-273. 135 Swank S.E. and W.E. Oechel. 1991. Interactions among the effects of herbivory, competition, and resource limitation on chaparral herbs. Ecology 72:104-115. Taitt M.J. 1981. The effect of extra food on small rodent populations: I. Deermice {Peromyscus maniculatus). J. Anim. Ecol. 50:111-124. Taitt M.J. and C.J. Krebs. 1981. The effect of extra food on small rodent populations: II. Voles (Microtus townsendii). J. Anim. Ecol. 50:1251-137. Taitt M.J. and C.J. Krebs. 1985. Population dynamics and cycles. In: Tamrin R.H. (ed.). Biology of new world Microtus. Amer. Soc. Mammal. Spec. Public. 8:567-620. Taitt M.J., J.H.W. Gipps, C.J. Krebs and Z. Dundjerski. 1981. The effect of extra food and cover on declining populations of Microtus townsendii. Can. J. Zool. 59:1593-1599. Temeles E. 1986. Reversed sexual size dimorphism: effect on resource defense and foraging behaviors of non-breeding northern harriers. Auk 103:70-78. Terman C.R. 1968. Population dynamics. In: King J.A. (Ed.). Biology of Peromyscus (Rodentia). Spec. Publ., Amer. Soc. Mamm., 2:1-593. Test F.H., and A.R. Test. 1943. The effect of botfly larvae on reproduction in white-footed mice Peromyscus leucopus. The American Midland Naturalist 101:211-217. Thompson S.D. 1982. Structure and species composition of desert heteromyid rodent species assemblages: effects of a simple habitat manipulation. Ecology 63:1313-1321 Timm R.M. and E.F. Cook. 1979. The effect of botfly larvae on reproduction of white-footed mice, Peromyscus leucopus. Am. Midi. Nat. 101:211-217. 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., M.R. Perrin and S.L. Iverson. 1975. Winter coexistence of voles in spruce forest: relevance of seasonal changes in aggression. Can. J. ZooL 53:1004-1011. Van Home B. 1981. Demography of Peromyscus maniculatus populations in serai stages of coastal coniferous forest in southeast Alaska. Can. J. Zool. 59:1045-1061. 136 Van Home B. 1982. Niches of adult and juvenile deer mice (Peromyscus maniculatus) in serai stages of coniferous forest. Ecology 63:992-1003. Van Home B. 1986. Summary: when habitats fail as predictors -the researcher's viewpoint. In: Verner J., M.L. Morrison and C.J. Ralph (Eds.). Wildlife 2000: modeling habitat relationships of terrestrial vertebrates. University of Madison Press. Madison, Wisconsin, :257-258. Vemer J. , M.L. Morrison and C.J. Ralph (Eds.) 1986 Wildlife 2000: modeling habitat relationships of terrestrial vertebrates. University of Madison Press. Madison, Wisconsin, U.S.A. Wecker S.C. 1962. The effects of botfly parasitism on a local population of the white-footed mouse. Ecology 43:561-565. Werner E.E. and J.F. Gilliam. 1984. The ontogentic niche and species interactions in size-structured populations. Ann. Rev. Ecol. Syst. 15:393-425. White G.C., D.R. Anderson, K.P. Bumham and D.L. Otis. 1982. Capture-recapture and removal methods for sampling closed populations. Los Alamos National Laboratory, Los Alamos, New Mexico. 235pp. Wilkinson L. 1988. SYSTAT: The System for Statistics. Evanston. II. SYSTAT, Inc. U.S.A. Wilson D.E. 1968. Ecological distribution of the genus Peromyscus. Southwestern Naturalist 13:267-274. * Wolff J.O. 1985. The effects of density, food, and interspecific interference on home range size in Peromyscus leucopus and Peromyscus maniculatus. Can. J. Zool. 63:2657-2662. Wolff J.O. 1989 . Social Behavior. In: Kirkland G.L. and J.N. Layne (eds.). Advances in tile study of Peromyscus (Rodentia). Texas Tech University Press, pp.271-291. Wolff J.O. 1989. Social Behavior. In: Kirkland G.L. and J.N. Layne (eds.). Advances in the study of Peromyscus (Rodentia). Texas Tech University Press, pp.271-291. Wolff J.O. and D.M. Cicirello. 1989. Field evidence for sexual selection and resource competition infanticide in white-footed mice. Anim. Behav. 38:637-642. 137 Xuhua X. and J.S. Millar. 1990. Infestations of wild Peromyscus leucopus by botfly larvae. J. Mamm. 71:255-258. Zar J.H. 1984. Biostatistical Analysis. 2nd ed. Prentice Hall, Inc. U.S.A. pp.718. Zuleta G.A. and M.L. Vignau. 1990. Botfly parasitism (Rogenhofera bonaerensis) (Diptera, Cuterebridae) in the Pampean Grassland Mouse (Akodon azarae), in Argentina. J. Wildlife Diseases 26:11-17. 138 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.831.1-0100480/manifest

Comment

Related Items