xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam 22 Ka 4500
controlfield tag 007 cr-bnu---uuuuu
008 s2010 flu s 000 0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0004635
Foraging decisions of nocturnal mice under direct and indirect cues of predation risk
h [electronic resource] /
by Robbin Capers.
[Tampa, Fla] :
b University of South Florida,
Title from PDF of title page.
Document formatted into pages; contains X pages.
Thesis (MS)--University of South Florida, 2010.
Includes bibliographical references.
Text (Electronic thesis) in PDF format.
Mode of access: World Wide Web.
System requirements: World Wide Web browser and PDF reader.
ABSTRACT: The perception of increased predation risk by nocturnal mice and other small mammals has been shown to reduce activity levels, particularly in foraging effort. Various cues of predation risk have been used in previous studies, but few have assessed the potential interactions between different types of cues. I conducted field, laboratory, and enclosure experiments using predator scents, artificial light, and microhabitat variables to determine the effects of direct and indirect cues of predation risk on foraging behavior in wild nocturnal mice. Experimental foraging trays served as artificial resource patches, and giving-up densities were measured in order to test for foraging persistence in patches exposed to cues of predation risk. Cotton mice (Peromyscus gossypinus) were used in laboratory and enclosure trials, and were the most common mice present at the sites used for field trials. Although previous foraging studies have used other Peromyscus species, this species has not been tested, but ranges over densely populated areas of the United States where artificial light could potentially affect its behavior. The perception of increased predation risk by nocturnal mice and other small mammals has been shown to reduce activity levels, particularly in foraging effort. Various cues of predation risk have been used in previous studies, but few have assessed the potential interactions between different types of cues. I conducted field, laboratory, and enclosure experiments using predator scents, artificial light, and microhabitat variables to determine the effects of direct and indirect cues of predation risk on foraging behavior in wild nocturnal mice. Experimental foraging trays served as artificial resource patches, and giving-up densities were measured in order to test for foraging persistence in patches exposed to cues of predation risk. Cotton mice (Peromyscus gossypinus) were used in laboratory and enclosure trials, and were the most common mice present at the sites used for field trials. Although previous foraging studies have used other Peromyscus species, this species has not been tested, but ranges over densely populated areas of the United States where artificial light could potentially affect its behavior. In outdoor and laboratory enclosures, cotton mice showed no aversive response to bobcat urine, cloths rubbed on cats, or snake sheds, but did exhibit avoidance of cat fur and artificial light. In the field experiment, mice showed a strong preference for covered microhabitats, but did not avoid bobcat urine or artificial light. Foraging in artificial resource patches also increased throughout the duration of the field experiment, possibly coinciding with a reduction in naturally-available forage. Mice in this population appear to use cover as their primary means of avoiding detection or capture by predators, though they do avoid artificial light and at least one fur-derived odor when their available options for escape are reduced.
Advisor: Earl D. McCoy, Ph.D.
t USF Electronic Theses and Dissertations.
Foraging Decisions o f Nocturnal Mice Under Direct a nd Indirect Cues o f Predation Risk by Robbin G. Capers A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science Department of Integrative Biology College of Arts and Sciences University of South Florida Major Professor: Earl D. Mccoy, Ph.D. Henry R. Mushinsky, Ph.D. Gordon Fox, Ph.D. Date of Approval: June 26, 2010 Keywords: habitat use P eromyscus gossypinus enclosures light, gi ving up density Copyright 2010 Robbin G. Capers
i TABLE OF CONTENTS LIST OF TABLES ii LIST OF FIGURES iii ABSTRACT iv INTRODUCTION 1 Optimal Foraging Theory 1 Artificial Light 2 Predator Scent 3 METHODS 6 Study Site 6 Study Specie s 8 Giving up Density 9 Field Experiment 10 Laboratory Experiment 1 3 Enclosure Experiment 1 4 RESULTS 19 Field Experiment 1 9 Laboratory Experiment 25 Enclosure Experiment 2 8 DISCUSSION 3 7 LITERATURE CITED 4 2 ABOUT THE AUTHOR END PAGE
ii LIST OF TABLES Table 1 Field experiment treatments. 1 2 Table 2 Enclosure experiment scent treatments. 1 5 Table 3 Within subjects ANOVA for field experiment. 19 Table 4 Between subjects ANOVA for field experiment. 20 Table 5 Within subjects ANOVA for laboratory experiment. 26 Table 6 Between subjects ANOVA for laboratory experiment. 2 7 Table 7 Paired t test for the bobcat urine treatment. 2 7 Table 8 Paired t test for the cat cloth treatment 27 Table 9 Paired t test for the cat fur treat ment. 27 Tab le 10 Within subjects ANOVA for enclosure experiment. 29 Table 11 Between subjects ANOVA for the enclosure experiment. 29 Table 12 Within subjects ANOVA with rabbit urine treatment removed. 30 Table 13 Between subjects ANOVA with rabbit uri ne treatment removed. 30 Table 14 Within subjects ANOVA for the cat fur treatment. 32 Table 15 Between subjects ANOVA for the cat fur treatment. 32 Table 16 Pairwise comparisons for enclosures within the cat fur treatment. 33 Table 17 Within subjects A NOVA for the bobcat urine treatment. 34 Table 18 Between subjects ANOVA for the bobcat urine treatment. 34 Table 19 Within subjects ANOVA for the rabbit urine treatment. 3 5 Table 20 Between subjects ANOVA for the rabbit urine treatment. 35 Table 21 Wit hin subjects ANOVA for the snake shed treatment 36 Table 22 Between subjects ANOVA for the snake shed treatment 36
iii LIST OF FIGURES Figure 1 Typical vegetation in the USF Ecological Research Area. 7 Figure 2 Cotton mouse ( Peromyscus gossypinus ). 9 F igure 3 Trapping and experiment stations for field experiments 11 Figure 4 Trapping stations used for enclosure experiments. 1 6 Figure 5 Enclosure design and completed enclosure. 17 Figure 6 Bar graph for field experiment treatments, grouped by micr ohabitat. 20 Figure 7 Bar graph for field experiment treatments, grouped by scent. 21 Figure 8 Bar graph for field experim ent treatments, grouped by light 22 Figure 9 Line graph of GUD against sampling date. 2 3 Figure 10 Linear regression of cover mic rohabitat GUD against study night. 2 4 Figure 11 Linear regression of open microhab itat GUD against study night. 25 Figure 12 Patterns of foraging in laboratory experiment. 26 Figure 1 3 Mean GUDs (g) for each scent, grouped by treatment. 28 Figure 1 4 Ba r graphs for the cat fur treatment, grouped by light. 31
iv Foraging Decisions o f Nocturnal Mice Under Direct a nd Indirect Cues o f Predation Risk Robbin G. Capers ABSTRACT The perception of increased predation risk by nocturnal mice and other small m ammals has been shown to reduce activity levels partic ularly in foraging effort. Various cues of predation risk have been used in previous studies, but few have assess ed the potent ial interactions between different types of cues I conduct ed field, laboratory and enclosure experiments using pred ator scents, artificial light and microhabitat variables to determine the effects of direct and indirect cues of predation risk on foraging behavior in wild nocturnal mice Experimental foraging tray s served as artificial resource patc hes, and giving up densities were measured in order to test for foraging persistence in patches exposed to cues of predation risk Cotton mice ( Peromyscus gossypinus ) were used in laboratory and enclosure trials, and w ere the most common mice present at the sites used for field trials. Although previous foraging studies have used other Peromyscus species, this species has not been tested, but ranges over densely populated area s of the United States where artificial ligh t could potentially affect its behavior. In outdoor a nd laboratory enclosures, cotton mice showed no aversive response to bobcat urine, cloths rubbed on cats, or snake sheds, but did e xhibit avoidance of cat fur and artificial light In the field experim ent, mice showed a strong preference for covered microhabitats but did not avoid bobcat urine or artificial light. Foraging in artificial resource patches also increased throughout the duration of the field experiment,
v possibly coinciding with a reduction in naturally available forage. Mice in this population appear to use cover as their primary means of avoiding detection or capture by predators, though they do avoid artificial light and at least one fur derived odor when their available options for escap e are reduced.
1 INTRODUCTION Optimal Foraging Theory Much of the study of animal behavior is concerned with the pa rts of an environment and specific food resources that animals should use Animals require energy for all of their activities, but searching for and obtai ning food resources may reduce the time available for engaging in alternative activities such as finding mates or building nests, and increase t he risk of predation or injury Optimal foraging theory attempts to explain the relationship between an animal a nd its environment through the trade offs that animal might face in the pursuit of energy. MacArthur and Pianka (1966) and Emlen (1966) developed the first models of optimal foraging theory, which described the circumstances in which an animal should be ex pected to add a novel food source or resource patch to its diet. Charnov (1976) posited an additional method for determining the optimal use of patchy habitats by predators, which could include predators of both animals and plant s This patch choice model is similar to that developed by Mac Arthur and Pianka (1966), but explicitly utilizes the Marginal Value Theorem to determine that an individual should cease foraging in a resource patch when the patch is depleted to the point that higher returns per unit time can be found elsewhere. Resource depletion in a patch can be caused by the affected individual, or by interspecific or intraspecific competition, and may affect both individual patch use and the persistence of populations in an area, as proposed by Ma cArthur and Levins (1964). A reduction in foraging efficiency also may be caused by an increase in the risk s
2 of foraging in a particular patch. Higher perceived risk in one patch may increase the relative attractiveness of nearby patches with lower percei ved risks and cause patch shifting behavior. Although it is still largely unknown exactly how various species evaluate risk, an increased amount of time spent i n predator avoidance activities such as vigilance and refuging should reduce the energy gain per unit time spent in higher risk patch es (Lima 1998). If the consequent changes in prey behavior have fitness consequences it may be possible for top down population regulation to occur as a result not only of direct predation, but also of the perceived ri sk of predation. Much previous work in predator prey dynamics has focused on population level effects, with little regard to the behavioral responses of either predator or prey (Lima 1998 Brown et al. 1999). A better understanding of the behavioral respon ses of prey animals to predation risk could influence not only the conservation of populations, but could also inform models of predator prey dynamics and foraging behavior (Lima 1998). Artificial Light Artificial light in the nocturnal environment is an increasingly prevalent form of wildlife habitat disturbance (Rich and Longcore 2006). Discussions of habitat disturbance have long been biased toward spatial fragmentation, but the influence of light could potentially be as damaging to the habitat use of nocturnal animals as any physical fragmentation of that habitat. Furthermore, increased artificial lighting is an inherent part of some physical fragmentation, such as urbanization. This type of habitat disturbance may increase the costs of activity in aff ected areas, and potentially cause changes to population numbers in species sensitive to the disturbance. Previous studies have recognized artificial lighting as a cause of changes in circadian and circannual activities in birds (reviewed in Molenaar et al 2006), disruption in the foraging habits of bats (Reith 1982, Elangovan and Marimuthu 2001) and other small mammals (Kotler 1984c, Bird et al. 2004), and interference with the dispersal
3 patterns of sea turtles (McFarlane 1963) and carnivores (Beier 1995) A rtificial lighting can be disruptive in at least two ways: by disturbing the normal temporal activity patterns of animals and by making areas unsuitable for foraging, nesting, and movement as a result of a perceived increased risk of visual detection b y predators in an illuminated environment. M any species, including several members of Peromyscus ( P. polionotus leucocephalus, Bird et al. 2004; P. maniculatus, Travers et al. 1988) respond to increased levels of nocturnal ambient light by reduce d total fo raging eff ort (Bird et al. 2004; Halle 199 5; Kaufman and Kaufman 1982; Kolb 1992; Lockard and Owings 1974 a 1974 b d foraging in open areas (Kaufman et al. 1983; Kaufman and Kaufman 1982; Kotler 1984 a ; Price et al. 1984; Thompson 1982; Travers et al. 1988). Significant differences in foraging between dark and artificially or naturally (by moonlight) brightened habitat patches show that brightened areas are selected negatively by foragers, regardless of an abun dant food supply (Kotler 1984 c ; Justice 1961; Lockard and Owings 1974a). It has been presumed that these changes were a result of a perceived increase in predation risk (Brown et al. 1988; Epple et al. 1993; Kotler 198 4 a 198 4 b 1984 c ; Lockard and Owings 1 974 a ), but the effects of ambient light have rarely been tested in conjunction with other cues of predation risk in a comparative manner (but see Brown et al. 1988 and Orrock et al. 2003). Predator Scent Previous studies have shown that both activity lev els and feeding (Kaufman and Kaufman 1982; Kolb 1982; Lockard and Owings 1974 a 19 7 4 b reduced by cues of increased predation risk, particularly in open habitat patches (Kaufman and Kaufman 1982 Price et al. 1984 Travers et al. 1988). P redator scent, particularly predator urine, has commonly been used to approximate predator presence
4 in studies designed to test the effects of predator presence on prey foraging behavior. The effects of predator scent on prey activity have been tested in v arious ways and with disparate results (see Apfelback et al. 2005 for review), but a number of these studies have shown changes in small mammal foraging behavior (Herman and Valone 2000; Jacob and Brown 2000; Kats and Dill 1998; Kotler et al. 1993 ). A redu ction in foraging activity has been shown to be a common response to predator scent in many species (Abramsky et al. 1996; Brinkerhoff et al. 2005; Brown 1988; Kotler 1984 a 1984 b 1984 c ; Kotler et al. 1991; Lima 1998 ). For example, Epple et al. (1993) sho wed that mountain beavers ( Aplodontia rufa ) avoided feeding from bowls scented with predator odors, and reduced their total food consumption when predator odors were present. These results indicate that a perception of increased predation risk can reduce f oraging generally, and particularly in areas most associated with the cue of predation risk. In this study I chose bobcat ( Lynx rufus ) urine as a direct cue of predation risk, because bobcats are visually oriented predators of small mammals, and bobcat s ign is frequently encountered at the study site. A scent control was used to increase the likelihood that observed effects were caused by a response to predator scent and not just to any scent present. Eastern cottontail rabbit ( Sylvilagus floridanus ) urin e was chosen as the scent control because it is a small herbivorous mammal that should not be perceived as either a threat or a competitor to P. gossypinus Increased nocturnal light levels also would present an increased risk of detection for prey specie s, especially those commonly depredated by primarily visual predators. But does the risk of detection posed by artificial lighting affect prey foraging behavior on a similar level as a more direct representation of predation risk, such as the presence of p redator scent at a foraging site? Would a combination of these two influences, artificial lighting and predator scent, affect prey foraging to an even greater degree than either one alone, by giving the prey animal the perception that not only would they b e detected
5 more readily in a brighter environment, but that there is actually a predator present? In this study I compared foraging levels under artificial lighting, under simulated predator presence, and under a combination of both stimuli to further il lustrate the effects of both factors independently and in conjunction. The working hypotheses of this study were that independent exposure to artificial nocturnal lighting and predator scent would result in similar levels of foraging reduction from control levels, and that a combination of artificial light and predator scent would reduce total foraging activity and reduce foraging in open areas, to a greater degree than either cue alone.
6 METHODS Study Site The University of South Flo rida's Ecological Research Area (Eco Area) is a 2 00 ha plot of preserved land in the midst of developed areas in suburban Tampa, Florida. Vegetative assemblages in the Eco Area include pine sandhills and flatwoods, cypress swamps, and bottomland hardwood f orests. Dominant vegetation (Figure 1) in my sample sites included sand live oak (Quercus genimata ) turkey oak ( Quercus laevis), longleaf pine ( Pinus palustris ), and slash pine ( Pinus elliottii ). The primary understory plants were saw palmetto ( Serenoa re pens ) and wiregrass ( Aristida stricta ) The site is burned periodically, and has a network of unpaved roads that serve as firebreaks and provide access to different parts of the site, though most access is by foot, which reduces the impact of researchers a nd other visitors to the site.
7 Figure 1 Typical vegetation in the USF Ecological Research Area.
8 Study Species The most abundant nocturnal mice at the study site are cotton mice ( Peromyscus gossypinus ). In the field exper iments, sample stations were chosen only if P. gossypinus was the only species trapped at that location. This restriction was made to reduce the potential of different species foraging at different stations. In enclosure experiments, only P. gossypinus wer e used. Peromyscus gossypinus (Figure 2 ) is a medium sized quadrupedal mouse that primarily inhabits bottomland hardwood forests and swamps (Pearson 1953). They are opportunistic omnivore s whose diet may be largely based on food availability but includ es invertebrates, seeds, fruits, and nuts (Calhoun 1941). P eromyscus gossypinus is considered semi arboreal (King 1968), and is primarily a nocturnal forager that spends its days in a nest In south central Florida P. gossypinus nests are typically either underground, often in abandoned gopher tortoise burrows or tree cavities, and are constructed of shredded saw palmetto ( Serenoa repens ) fibers, Spanish moss ( Tillandsia usneoides ), lichens, and cotton (Frank and Layne 1992). Potential predators of P. g ossypinus either confirmed or likely to exist in the study area include bobcats ( Lynx rufus ) southern black racers ( Coluber constrictor priapus ), yellow rat snakes ( Elaphe obsoleta quadrivittata ) feral cats ( Felis domesticus ), coyotes ( Canis latrans ), ba rred owls ( Strix varia ), and red shouldered hawks ( Buteo lineatus ).
9 Figure 2 Cotton mouse ( Peromyscus gossypinus ). Giving up Density Giving up density (GUD) is a common means of measuring perceived predation risk through the foraging activity of prey animals, and is described by Brown (1988) as the density of resources within a patch at which a n individual ceases foraging. I t represents the point where the benefits of continued foraging are outweighed both by any potential ris ks of that activity and the costs incurred by not engaging in alternative activities. Thus, the marginal value of a patch can be quantified using the giving up density. In a case where all conditions except th e perception of predation risk we re equal betwe en patches, a higher GUD in one patch would represent a higher perceive d predation risk in that patch. In this study I used GUD to evaluate the foraging decisions of nocturnal mice under conditions mimicking varying levels of predation risk in order to det ermine the effect o f perceived risk on patch use. To this end, I construct ed artificial
10 food patches, typically referred to as experimental feeding or foraging trays. Foraging trays are intended to serve as resource patches in the environment, and allow re searchers to accurately measure the availa ble resources in the patch before and after foraging activity. Field Experiment I conducted field based experiments to evaluate the foraging efficiency of mi ce in their natural environment To establish sampling s tations with consistent P. gossypinus presence, I set up trapping stations at 30 haphazardly selected sites, 25m apart. I placed two traps at each station and baited them for five nights (11 15 January 2009). Stations selected for inclusion in the experime nt were those at which I captured P. gossypinus on at least four of the five trap nights and did not catch any other species (Figure 3). Two e xperimental foraging trays at each of eight sampling stations served as resource patches which were exposed to d ifferent combination s of risk factors in a factorial design (Table 1) At each station o ne tray was placed under vegetative cover, usually saw palmetto ( Serenoa repens ) fronds, and the other was placed in the open a meter away, in order to examine the stre ngth of microhabitat preferences and to observe any potential microhabitat shifting caused by differences in perceived predation risk. Between the two trays, one meter away, I placed a battery powered lantern to serve as the source of artificial light.
11 Figure 3 Trapping and experiment stations for field experiments. C ircles were used in experiment, while squares were not used in experiment due to low trapping success.
12 Ta ble 1 Field experiment treatments. Treatment Scent Light 1 Rabbit Natural 2 Rabbit Artificial 3 Bobcat Natural 4 Bobcat Artificial Foraging trays consisted of clear plastic b oxes measuring 37 x 21.6 x 12.7 cm, and fitted with clear lids. Each tray h ad a 5.4 cm diameter hole cut in one side to allow mice to enter but exclude birds and other animals. Both trays were situated with the entrance holes facing toward cover and away from any roads or trails. I constructed scent dispensers from plastic film canisters attached to 8.9 cm nails with clear duct tape. Each canister had four 5 mm holes drilled around the top and was covered with a lid. This method followed a previously succes sful design used by Brinkerhoff et al. (2005). I applied scent s by pourin g 10 ml of either bobcat or rabbit urine on a single cotton ball in each scent dispenser, which I then placed immediately adjacent to the entrance hole of a foraging tray. Urine was procured from Sterling Fur & Tool Co., in Sterling, Ohio Stations were ran domly assigned to one of four treatments (Table 1). Each trial run consisted of five consecutive nights, during which the scent dispensers remained at the sampling stations and lanterns were turned on each day at dusk in accordance with a station's assigne d treatment. Each evening I baited trays with 5.0g of husked millet seeds mixed with one liter of dry, sifted sand, and each morning I sieved the remaining seeds from each tray, then dried a nd weighed them. Following the final night of each trial run, I re moved and thoroughly cleaned all scent dispensers, and baited all trays with millet seed for at least one night without applying any treatments. I then reassigned each station to a new treatment group until each station had received every treatment for two trial runs.
13 The weight of millet seeds remaining each morning was used untransformed as giving up densities (GUDs). Repeated measures analysis of v ariance (ANOVA) was performed on the GUD data, with scent (bobcat or rabbit) and light ( natural or artifici al ) as fixed, whole plot factors, and microhabitat (open or cover) as a within subjects factor. Giving up densities for open and covered trays were compared with study night through linear regression analysis to determine how foraging changed over time dur ing the study. Daily temperatures, humidity, and rainfall were obtained from the University of South Florida Weather Station, locate d within 5 km of the study site The fraction of the moon that was illuminated each night was obtained from the US Naval Obse rvatory. Weather and moonlight data were compared to GUD through linear regression analysis to determine the possible influence of these factors on foraging behavior All an alyses were performed in SPSS 18 Laboratory Experiment I conducted indoor enclosu re based experiments in the laboratory from 11 20 March 2009 to investigate the effect of predator scent in a controlled environment. Scents used in this experiment were 10ml of bobcat urine, a 5 x 5 cm cloth that was and 10g of cat fur in a cheesecloth envelope. These scents were paired with controls of water, a clean cloth, and an empty chee secloth envelope, respectively. Laboratory enclosures consisted of 61 x 32 x 43 cm glass terrari a covered with black paper in or der to prevent mice from seeing out. Lids were constructed of 1/4 in hardware cloth to prevent escape, and 60w lamps set to a 12 hour daylight schedule were placed above the enclosures. Floors of terrari a were covered with aspen shavings to provide a natur al substrate that does not confer a strong scent. Mice were provided with one 20 x 15 x 9 cm plastic nest box in the center of one side of the terrarium, and
14 two 8 x 2 cm ceramic food bowls located in the corners of the terrarium on the wall opposite the n est box. Water was provided ad libitum through a wall mounted bottle placed in the center of the wall opposite the nest box For each scent treatment, five mice were randomly selected from those captured overnight at trapping stations marked in Figure 4 I then weighed and determined the sex of each mouse to ensure that they were adults and transferred them to terraria in the laboratory immediately after retrieval from the traps. In each terrarium, one bowl was randomly chosen as experimental and one as co ntrol, and scent treatments were applied. In the cloth and fur treatments, the scent cue or appropriate control was taped to the wall of t he terrarium immediately above each food bowl, and in the bobcat urine treatment, 10ml of either bobcat urine or water was poured into a P etri dish placed below the food bowl Fiv e grams of millet seed were placed in each bowl. The following morning the remaining seeds were weighed and the mice were released to their site of capture. The weight of seeds remaining was comp ared between scent treatments using repeated measures ANOVA, with bowl (control or experimental feeding bowl) as the within subjects factor. Mean differences within each scent treatment were compared using paired t tests Enclosure Experiment I conducted outdoor enclosure based experiments from 22 May 13 June 2009 in order to examine foraging preferences of mice in a natural setting that controlled for forag er density and food availability As in the laboratory experiment, this experiment was set up as a c hoice test, where mice were allowed to forage from containers scented with either one of several predator scents or a control substance (Table 2 ) while exposed to eith er natural or artificial light.
15 Table 2 Enclosure experiment s cent treatments Experimental Substance Control Substance Bo bcat urine (10 ml) Distilled water (10 ml) Rabbit urine (10 ml) Distilled water (10 ml) Southern black racer shed (10g) Empty scent dispenser Domestic cat fur (10 g) Empty scent dispenser Trappin g stations were established at sites previously used in field experiments, as well as three others that proved to be reliable sources of P. gossypinus (Figure 4). Each morning I checked all traps and released any mice that were not P. gossypinus I randoml y chose five mice of those captured, weighed and determined the sex of each mouse, and transferred them singly into enclosures prepared with fresh water, bedding, and scent dispensers. During their captivity, each mouse was provided with 5g of millet seed mixed with one liter of dry, sifted sand in each of two foraging trays. Food was offered only from dusk until dawn of the next day to determine the effect of the treatments on nocturnal foraging activity and not allow mice to temporally shift their foragin g activity. All mice were held for 24 hours, then released to their capture sites after I sieved the seeds from their feeding trays in the same manner as in the field experiments.
16 I constructed five 0.6 m 3 enclosures fro m " galvanized wire mesh. Each enclosure was fitted with a lid made of the same wire mesh riveted to aluminum angle pieces. In addition, I attached an eight inch strip of aluminum flashing around the top of each enclosure so that mice could not easily esc ape by climbing out the top of the enclosure while the lid was removed (Figure 5 ). Figure 4 Trapping stations used for enclosure experiments, marked with white circles.
17 A B Figure 5 Enclosure design (A) and completed enclosure (B) Each enclosure was located at least 10 m from any other enclosur e, and under s hade to prevent overheating I wired the corners of each enclosure to iro n posts, which I drove 0.3 m into the ground to prevent any movement of the enclosures. The bottoms were then covered with approximately 75 mm of sand, dry leaves, and tw igs in order to provide a natural substrate Enclosures were equipped with one nest box, placed in the center of a haphazardly selected wall, constructed from a 17 x 11 x 7 cm plastic box with an opaque lid and a 5 cm diameter entrance hole on one side. E ach day, nest boxes were filled with fresh nesting material consisting of dry leaves and Spanish moss ( Tillandsia usneoides ). In each of the opposing corners I placed a 15 x 15 x 13 cm foraging tray filled with one liter of sand. Each tray had an opaque li d and a 5 cm diameter hole cut in the side facing the nest box. In the center of each enclosure I placed a 15 cm diameter plastic tray, which I filled with clean water each day. Several saw palmetto ( Serenoa repens ) fronds were also placed in each enclosur e to provide cover for mice while out of their nest boxes.
18 Scent dispensers were constructed as in the field experiments and again placed immediately adjacent to the entrance holes of each foraging tray. Each day, one dispenser contained an experimental substance and the other contained its corresponding control substance. I alternated the position of the experimental and control substances each day. The southern black racer shed was gathered in the field immediately following shedding, and domestic cat fur was acquired from Temple Terrace Animal Hospital, in Temple Terrace, Florida. To provide artificial light, I placed a single battery powered lantern outside each enclosure, centered between the two foraging trays. Lanterns were turned on at dusk on ea ch night designated as an artificial light night, and left in place on natural light nights. Seeds remaining each morning were used untransformed as giving up densities. I ran a repeated measures ANOVA in SPSS 18, using GUDs from the control and experimen tal trays and a within subjects factor, and scent, light, and enclosure and between subjects factors. I then ran separate ANOVAs on each scent, again using subjects factors, and light and enclosure as between subjects factors.
19 RESULTS Field Experiments Treatments were applied at 16 foraging trays for a total of 20 nights. Trays with no signs of foraging were excluded from all analyses, but trays with footprints or droppings were included even if s igns of extensive foraging were not present. There was a highly significant effect of microhabitat (p < 0.001), with mice preferring covered trays in all treatments (Table 3 ) although t here were no significant effects of scent or light (Table 4 ). Similar patterns were seen for scent and light treatments within each microhabitat, but these differences were not significant (Figure 6). GUDs for scent were nearly identical as averaged over all other factors (Figure 7 ), while average GUD for light treatments w as higher than dark treatments (Figure 8 ). Table 3 Within subjects ANOVA for field experiment. Source Type III Sum of Squares df Mean Square F Sig. Microhabitat 69.160 1 69.160 78.177 .000 Microhabitat Scent .136 1 .136 .154 .698 Microhabitat Light .269 1 .269 .304 .586 Microhabitat Scent Light .781 1 .781 .883 .355 Error(Microhabitat) 24.770 28 .885
20 Table 4 Between subjects ANOVA for field experiment Source Type III Sum of Squares df Mean Square F Sig. Intercept 514.779 1 514.779 154.620 .000 Scent .072 1 .072 .022 .884 Light 3.768 1 3.768 1.132 .296 Scent Light .008 1 .008 .003 .960 Error 93.221 28 3.329 Figure 6 Bar graph for field experiment t reatments, grouped by microhabitat.
21 Figure 7 Bar graph for field experiment treatments, grouped by scent.
22 Figure 8 Bar graph for field experiment treatments, grouped by light. When averaged over all treatments, giving up density decreased over time during the course of the experiment (Figure 9) This pattern held for both covered (p < 0.001, Figure 10 ) a nd open trays (p < 0.001, Figure 11 ) In the covered trays, study night accounted for almost 9 0 percent of the variation among GUDs over time although this figure was lower for the open trays. Linear regressions were also run against moon illumination; minimum, maximum, and mean temperature; and maximum and mean humidity, but none of these w as sig nificant.
23 Figure 9 Line graph of GUD against sampling date.
24 Figure 10 Linear regression of cover microhabitat GUD against study night. y = 0.140x + 3.720 R = 0.947 R 2 = 0.896 P < 0.001
25 L aboratory Experiment W hen all scent treatments were analyzed together t here was a marginal difference between the c ontrol and ex perimental bowls ( P = 0.068) and a significant interaction between the bowl and scent treatment s (P = 0.014; Table 5). N o significant difference existed in the total amount of millet seed eaten each night among the scent treatments (Table 6). Patterns of f oraging are illustrated in Figure 12. Mice significantly preferred bobcat urine to its control (P = 0.015 ; Table 7) and ate more from bowls scented with cat cloths than with clean cloths, although this difference was not significant (P = 0. 170 ; Table 8 ) Average giving up density for the cat fur treatment was lower than that for the control, but this also was not significant ( P = R = 0.763 R 2 = 0.582 P = < 0.001 Figure 11 Linear regression of open microhabitat GUD against study night. y = 0.80x + 4.995
26 0. 230 ; Table 9 ). One mouse was excluded from this analysis because it removed the cat fur from the terrarium wall and took it in to the nest box but an additional mouse was captured and tested in order to maintain an N of 5 for all scent treatments Figure 12 Patterns of foraging in laboratory experiment. Table 5 Within subje cts ANOVA for laboratory experiment. Source Type III Sum of Squares df Mean Square F Sig. Bowl 3.809 1 3.809 4.019 .068 Bowl Treatment 11.722 2 5.861 6.183 .014 Error(factor1) 11.375 12 .948
27 Table 6 Between subjects ANOVA f or laboratory experiment. Source Type III Sum of Squares df Mean Square F Sig. Intercept 217.244 1 217.244 411.249 .000 Treatment .835 2 .418 .791 .476 Error 6.339 12 .528 Table 7 Paired t test for the bobcat urine treatment Paired Differences t df Sig. (2 tailed) Mean Std. Deviation Std. Error Mean 95% Confidence Interval of the Difference Lower Upper ExperimentGUD ControlGUD 2.02400 1.10776 .49540 3.39946 .64854 4.086 4 .015 Table 8 Paired t test for the cat cloth treatment. Paired Differences t df Sig. (2 tailed) Mean Std. Deviation Std. Error Mean 95% Confidence Interval of the Difference Lower Upper ExperimentGUD ControlGUD 1.08400 1.45098 .64890 2.88562 .71762 1.671 4 .170 Table 9 Paired t test for the cat fur treatment. Paired Differences t df Sig. (2 tailed) Mean Std. Deviation Std. Error Mean 95% Confidence Interval of the Difference Lower Upper ExperimentGUD ControlGUD .97000 1.53458 .68629 .93544 2.87544 1.413 4 .230
28 Enclosure Experiment In the enclosure experiments, mice generally foraged less under artificial than natural light, and less in experimental than control trays (Figure 1 2 ). This trend was si gnificant for light (P = 0.01 0 ; Table 1 1 ). Light was also significant when the herbivore scent treatment (rabbit urine) was removed from the analysis (P=0.030; Tables 12 13). No effect was seen for scent when all treatments were analyzed together (Table 11 ) or when predator treatments were analyzed without the rabbit urine treatment (Table 13) meaning that overall foraging was not different among the scent treatments Figure 13 Mean GUDs (g) for each scent, grouped by treatment. Error bars represent 95% conf idence intervals.
29 Table 10 Within subjects ANOVA for enclosure experime nt. Source Type III Sum of Squares df Mean Square F Sig. Tray .235 1 .235 2.400 .129 Tray Light .064 1 .064 .658 .422 Tray Scent .377 3 .126 1.283 .293 Tray Enclosure .382 4 .095 .976 .432 Tray Light Scent .073 3 .024 .247 .863 Tray Lig ht Enclosure .294 4 .074 .752 .563 Tray Scent Enclosure 1.658 12 .138 1.412 .201 Tray Light Scent Enclosure .758 12 .063 .646 .790 Error(Tray) 3.915 40 .098 Table 11 Between subjects ANOVA for the enclos ure experiment. Source Type III Sum of Squares df Mean Square F Sig. Intercept 491.086 1 491.086 1837.68 5 .000 Light 1.947 1 1.947 7.286 .010 Scent .261 3 .087 .325 .807 Enclosure 1.617 4 .404 1.513 .217 Light Scent .445 3 .148 .556 .647 Light En closure 1.198 4 .300 1.121 .360 Scent Enclosure 3.918 12 .326 1.222 .303 Light Scent Enclosure 7.385 12 .615 2.303 .024 Error 10.689 40 .267
30 Table 12 Within subjects contrasts with rabbit urine treatment removed. Sou rce Type III Sum of Squares df Mean Square F Sig. Tray .274 1 .274 2.809 .104 Tray Light .067 1 .067 .690 .413 Tray Scent .296 2 .148 1.519 .236 Tray Enclosure .161 4 .040 .414 .797 Tray Light Scent .069 2 .034 .353 .706 Tray Light Enclosure .482 4 .120 1.234 .318 Tray Scent Enclosure .546 8 .068 .699 .690 Tray Light Scent Enclosure .670 8 .084 .858 .561 Error(Tray) 2.829 29 .098 Table 13 Between subjects contrasts with rabbit urine tr eatment removed. Source Type III Sum of Squares df Mean Square F Sig. Intercept 354.867 1 354.867 1224.73 .000 Light 1.511 1 1.511 5.216 .030 Scent .233 2 .116 .402 .673 Enclosure .950 4 .237 .819 .523 Light Scent .445 2 .223 .768 .473 Light Encl osure .828 4 .207 .715 .589 Scent Enclosure 3.350 8 .419 1.445 .220 Light Scent Enclosure 7.377 8 .922 3.183 .010 Error 8.403 29 .290
31 In the cat fur treatment, mice foraged less under artificial light and in experimental trays (Figure 14 ). W hen analyzed separately, a significant difference existed in GUD between control and experimental feedi ng trays for cat fur (P=0.028; Table 1 4 ). T here was also a significant effect of light (P=0.008) and enclosure (P=0.04 9), and a marginal interaction betw een light and e nclosure (P=0.063; Table 1 5 ). Pairwise comparisons showed enclosures A and B to have significantly different mean GUDs (Table 16). Figure 14 Bar graphs for the cat fur treatment, grouped by light.
32 Table 14 Within subjects ANOVA for the cat fur treatment. Source Type III Sum of Squares df Mean Square F Sig. Tray .543 1 .543 6.636 .028 Tray Light .059 1 .059 .725 .415 Tray Enclosure .189 4 .047 .579 .685 Tray Light Enclosure .213 4 .053 .651 .639 Error(Tray) .818 10 .082 Table 15 Between subjects ANOVA for the cat fur treatment. Source Type III Sum of Squares df Mean Square F Sig. Intercept 130.755 1 130.755 894.691 .000 Light 1.600 1 1.600 10. 948 .008 Enclosure 2.052 4 .513 3.510 .049 Light Enclosure 1.861 4 .465 3.183 .063 Error 1.461 10 .146
33 Table 16 Pairwise comparisons for enclosures within the cat fur treatment (I) Enclosure (J) Enclosure Mean Dif ference (I J) Std. Error Sig. 95% Confidence Interval Lower Bound Upper Bound A B .3706 .12861 .047 .0033 .7380 C .3159 .12861 .121 .0514 .6833 D .1941 .12861 .563 .1733 .5614 E .3253 .12861 .104 .0420 .6926 B A .3706 .12861 047 .7380 .0033 C .0547 .12861 .993 .4220 .3126 D .1766 .12861 .648 .5439 .1908 E .0453 .12861 .997 .4126 .3220 C A .3159 .12861 .121 .6833 .0514 B .0547 .12861 .993 .3126 .4220 D .1219 .12861 .876 .4892 .2455 E .009 4 .12861 1.000 .3580 .3767 D A .1941 .12861 .563 .5614 .1733 B .1766 .12861 .648 .1908 .5439 C .1219 .12861 .876 .2455 .4892 E .1313 .12861 .844 .2361 .4986 E A .3253 .12861 .104 .6926 .0420 B .0453 .12861 .997 .3220 .4126 C .0094 .12861 1.000 .3767 .3580 D .1313 .12861 .844 .4986 .2361
34 When the remaining scent treatments were analyz ed separately, no significant differences were seen for experimental versus control trays, light, or enclosure (Tables 1 7 22 ), though there was a significant interaction between light and enclosure in the bobcat urine treatment ( P = 0.014; Table 1 7 ). Tabl e 17 Within subjects ANOVA for the bobcat urine treatment. Source Type III Sum of Squares df Mean Square F Sig. Tray 9.000E 5 1 9.000E 5 .001 .977 Tray Light .004 1 .004 .040 .845 Tray Enclosure .393 4 .098 .995 .454 Tray Light Enclosure .373 4 .093 .943 .478 Error(Tray) .988 10 .099 Table 18 Between subjects ANOVA for the bobcat urine treatment. Source Type III Sum of Squares df Mean Square F Sig. Intercept 116.008 1 116.008 554.427 .00 0 Light .286 1 .286 1.365 .270 Enclosure .590 4 .147 .705 .606 Light Enclosure 4.495 4 1.124 5.371 .014 Error 2.092 10 .209
35 Table 19 Within subjects ANOVA for the rabbit urine treatmen t Source Type III Sum of Squares df Mean Square F Sig. Tray .333 1 .333 2.069 .181 Tray Light .337 1 .337 2.092 .179 Tray Enclosure 1.344 4 .336 2.088 .157 Tray Light Enclosure .099 4 .025 .153 .957 Error(Tray) 1.609 10 .161 Table 20 Between su bjects ANOVA for the rabbit urine treatment. Source Type III Sum of Squares df Mean Square F Sig. Intercept 132.532 1 132.532 732.942 .000 Light .071 1 .071 .395 .544 Enclosure 1.059 4 .265 1.464 .284 Light Enclosure .686 4 .172 .949 .476 Error 1.80 8 10 .181
36 Table 21 Within subjects ANOVA for the snake shed treatment Source Type III Sum of Squares df Mean Square F Sig. Tray .000 1 .000 .003 .956 Tray Light .004 1 .004 .035 .855 Tray Enclosure .673 4 .168 1. 481 .279 Tray Light Enclosure .249 4 .062 .547 .705 Error(Tray) 1.137 10 .114 Table 22 Between subjects ANOVA for the snake shed treatment. Source Type III Sum of Squares df Mean Square F Sig. Intercept 119.232 1 119. 232 241.838 .000 Light .166 1 .166 .338 .574 Enclosure 3.077 4 .769 1.560 .259 Light Enclosure .613 4 .153 .311 .864 Error 4.930 10 .493
37 DISCUSSION The four treatments in the field experiment were designed to represent varying levels of perceived predation risk. Foraging effort was expected to coincide with this perception, with less foraging occurrin g in treatments representing hi gher levels of predation risk. In addition, because foraging under cover should have reduced the percepti on of predation risk in all treatments, I expected that foraging would occur at lower levels in open feeding trays than in covered trays. In keeping with these assumptions, the difference in GUDs between cover and open microhabitats should have been greate r at treatment arrays representing higher perceived predation risk, representing an interaction between predation risk and microhabitat. In practical terms this would have meant that mice shifted microhabitats in respons e to increased predation risk. In m y field experiment however, the only significant difference in GUD was between open and covered microhabitats. While many small mammal species have been shown to reduce foraging under increase d nocturnal lighting conditions or in the presence of predator cues, it cannot be assumed that the same behavior will characterize all species in all locations. In response to differences in predator types, availability of natural cover, and proximity to urban areas (with concomitant light pollution), the primary mean s of avoiding predation may vary in different prey populations. One possible explanation for a lack of response to predation risk cues is that P. anti predator behavior is not targeted toward specific predator types. This
38 response may differ f rom other populations and species that have been studied because of the abun dance of cover available to mice and the wide range of predators that inhabit Central Florida pine sandhills, flatwoods, and bottomland hardwood forests that dominate the study si te. In the population studied here, it appears that predation may be avoided primarily through the use of cover while foraging. In many previous studies showing significant effects of either predator scent or light on foraging behavior, the vegetative comm unity provided less abundant cover (e.g deserts, grasslands, beach dunes) In contrast, t he high availability of cover in my study area may allow the prey species to utilize cover more effectively, thus allowing them to remain active even in situations th at represent high levels of predation risk. A n environment with a high abundance of predators of varying types could leave prey species with few predator specific options for evading predation. The use of cover while active, in combination with observed be haviors such as burrowing, running, and standing still, may mitigate the risk of predation for these mice better than any avoidance strategy targeted to one particular predator species or another (Lima and Bednekoff 1999, Powell and Banks 200 4 Verdolin 20 06) In this case, availability of cover may be of great importance in the persistence of When given fewer choices for avoiding predation or using alternate food sources such as in the enclosure experiment mice showed a gr eater response to predator cues. Artificial light significantly reduced foraging in enclosures, a pattern that held for all scents tested, including rabbit urine. The increased response to artificial light in the enclosure s as contrasted with the field ex periment, follows a pattern of more pronounced behavioral effects in enclosure experiments than field experiments generally (Reviewed in Verdolin 2006). This could be due to the availability of natural forage in field experiments, as well as the ability of prey animals in field conditions to use a wider
39 variety of anti predator responses that may allow th em to forage more readily under higher predation risk. Only one of the predator scents tested in any of the experiments elicited an aversive response. In laboratory and enclosure experiments, mice avoi ded bowls scented with cat fur. This finding agrees with previous studies showing fur derived odors to be more effective in reducing foraging and other measures of prey animal activity than urine or feces deri ved odors (See Apfelbach et al. 2005 for a review). These odors may be more useful to prey animals because of their volatility, since odors that evaporate sooner indicate a predator that is more likely to still be in the area. In addition to avoiding cat f ur scent, mice in the labora t ory experiment showed a preference for bobcat urine and cloths that had been rubbed on cats. Not only was this the case when average GUDs for control and scented bowls were compared but every mouse tested in the bobcat urine t reatment and all but one mouse tested in the cat cloth treatment preferred foraging in the bowl scented with predator odor. These results are not readily explainable, as even if they were the result of cross contamination of scents between bowls, I would e xpect to have seen little preference between control and experimental bowls. Instead this unanticipated behavior may have resulted from altered motivat ions in an unnatural environment, a situation behavioral researchers should be especially mindful of when conducting experiments in laboratory settings. In addition to anti predator behavior, mice in the field experiment showed temporal variation in their foraging habits. L ow natural food resources as the stu dy continued may have lead to the increase in forag ing over time seen in this experiment (Lima 1998). GUDs for each separate treatment show general patterns of decline over time but also exhibit a high degree of variability Averaging GUDs over all treatments shows th e same pattern much more clearly and may be a reflection either of a general increased acceptance of the foraging trays themselves, or of a reduction in natural ly
40 available forage and consequent increase in marginal value of the artificial foraging trays. T o gain additional insight into the e ffect of reduced food availability on foraging, future researchers should manipulate seed densities in artificial foraging trays or conduct longer term studies which could provide more information about changes in foraging behavior through seasonal chang es and associated changes in resource availability. Additional metrics of activity could be m easured in order to acquire a broader perspective of the effects of perceived predation risk on prey animal behavior. Past studies have utilized sand tracking, li ve trapping, and either direct or camera mediated observation (Kaufman and Kaufman 1982). These techniques could be useful in future studies in combination with the use of experimental feeding stations. Studies designed to investigate anti predator behavi or naturally need to incorporate cues from native predators, but in an area with many predators of various types, experimental designs can become increasingly complex as cues from additional predator species are included. Also, mimicking different types of predators may require the use of very different cues, from scents, to calls, to replicas of the predators themselves. This may make comparisons between prey responses more difficult, as a respond to different cues. Conclusions It is clear from this study that Peromyscus gossypinus prefers foraging under cover, and may use cover as its primary means of avoiding detection or capture by predators. For this population, the use of cover may be more reliable and useful than other anti predator behaviors due to the large number and variety of predators, and the abundance of available cover, in its environment. This differs from the environments of many other populations tested in similar studies. A greater response to predator cues or
41 nocturnal lighting may be seen when prey species have less opportunity to remain concealed while active. In environments where animals must cross open areas to reach foraging grounds, increased predation risk may requ ire lower activity levels than in those environments where it is possible for animals to move among nests and foraging areas under cover. Cover may be especially useful in areas with a high density and variety of predators, such as for small mammals in pen insular Florida, because it allows them to avoid detection from most predators at once, without needing to identify or assess individual threats. Though cover appears to be the primary means of avoiding predation, this species does recognize and respond t o some predator odors. A fur derived odor reduced foraging, which may indicate that fur derived odors provide more valuable information than odors derived from other sources. This could be a result of the more volatile and ephemeral nature of these odors a s compared especially t o urine or feces derived odors, which could make them a more useful indicator of a predator that is still in Laboratory and enclosure experiments produced stronger effects of predator cues, i ncluding unexpected results in the laboratory experiment. This indicates that controlled environments may intensify behavioral responses to stimuli, or even change natural be haviors altogether. While it is often beneficial to use controlled environments to amplify behavioral responses, care must be taken when experimental settings may limit natural behavioral repertoires or alter motivations, leading to responses that do not In a similar vein, giving up densi ties are a useful tool for estimating the value of a resource patch to a forager. However, it is imperative that researchers remain aware that providing artificial food sources, as well behaviors in a way that makes interpretation and extrapolation of these behaviors difficult.
42 LITERATURE CITED Abramsky, Z., E. Strauss, A. Subach, A. Riechman, and B. P. Kotler. 1996 The effect of barn owls ( Tyto alba ) on the activity and microhabita t selection of Gerbillus allenbyi and G. pyramidum .Oecologia. 105:313 319. Apfelback, R., C. D. Blanchard, R. J. Blanchard, R. A. Hayes, and I. S. McGregor. 2005. The effects of predator odors in mammalian prey species: A review of field and laboratory st udies. Neuroscience and Biobehavioral Reviews 29:1123 1144. Beier, P. 1995. Dispersal of juvenile cougars in fragmented habitat. Journal of Wildlife Management 59:228 237. Bird, B. L., L. C. Branch, and D. L. Miller. 2004. Effects of coastal lighting on foraging behavior of beach mice. Conservation Biology 18:1435 1439. Brinkerhoff, R. J., N. M. Haddad, and J. L. Orrock. 2005. Corridors and olfactory predator cues affect small mammal behavior. Journal of Mammalogy. 86:662 669. Brown, J.S. 1988. Patch u se as an indicator of habitat preference, predation risk, and competition. Behavioral Ecology and Sociobiology 22:37 48. Brown, J. S., B. P. Kotler, R. J. Smith, and W. O. Wirtz II. 1988. The effects of owl predation on the foraging behavior of heteromyid rodents. Oecologia 76:408 415. Brown, J. S., J. W. Laundr, and M. Gurung. 1999. The ecology of fear: optimal foraging, game theory, and trophic interactions. Journal of Mammalogy 80:385 399. Calhoun, J. B. 1941. Distribution and food habits of mammals in the vicinity of Reelfoot Lake Biological Station, III. Discussion of the mammals recorded from the area. Journal of Tennessee Academy of Science 16:207 225. Charnov, E. L. 1976. Optimal foraging, the marginal value theorem. Theoretical Population Biolo gy 9:129 136. Elangovan, V. and G. Marimuthu. 2001. Effect of moonlight on the foraging behaviour of a megachiropteran bat Cynopterus sphinx Journal of Zoology, London 253:347 350. Emlen,J. T. 1966. The role of time and energy in food preference. Amer ican Naturalist 100:611 617.
43 Epple, G., J. R. Mason, D. L. Nolte, and D. L. Campbell. 1993. Effects of predators odors on feeding in the mountain beaver ( Aplodontia rufa ). Journal of Mammalogy 74:715 722. Frank, P. A. and J. N. Layne. 1992. Nests and dayt ime refugia of cotton mice (Peromyscus gossypinus) and golden mice ( Ochrotomys nuttali ) in south central Florida. American Midland Naturalist 127:21 30. Halle, S. 1995. Effect of extrinsic factors on the activity of root voles, Microtus oeconomus Journa l of Mammalogy 76:88 99. Herman, C. S. and T. J. Valone. 2000. The effect of mammalian predator scent on the foraging behavior of Dipodomys merriami Oikos 91:139 145. Jacob, J. and J. S. Brown. 2000. Microhabitat use, giving up densities and temporal activity as short and long term anti predator behaviors in common voles. Oikos 91:131 138. Kats. L. and L. M. Dill. 1998. The scent of death: chemosensory assessment of predation risk by prey animals. Ecoscience. 5:361 394. Kaufman, D. W., and G. A. Kau fman. 1982. Effect of moonlight on activity and Dipodomys ordii ). Journal of Mammalogy 63:309 312. Kaufman, D. W., S. K. Peterson, R. Fristik, and G. A. Kaufman. 1983. Effect of microhabitat features on habitat use by Peromyscus leucopus. American Midland Naturalist 110:177 185. King, J. A. 1968. Psychology. In: Biology of Peromyscus (Rodentia), J. A. King (Ed.) pp. 496 542. Special Publication 2. The American Society of Mammalogists. Kolb, H. H. 1992. The effect of moonlight in the wild rabbit (Oryctolagus cuniculus). Journal of Zoology (London) 228:661 665. Kotler B. P., J. S. Brown, and W. A. Mitchell. 1993. Environmental factors affecting patch use in two species of gerbilline rodents. Journal of Mammalogy. 74:614 620. Kotler, B. P. 1984 a Risk of predation and the structure of desert rodent communities. Ecology 65:689 701. Kotler, B. P. 1984 b Harvesting rates and predatory risk in desert rodents: a comparison of two communities on different continents. Jo urnal of Mammalogy 65:91 96. Kotler, B. P. 1984 c Effects of illumination on the rate of resource harvesting in a community of desert rodents. American Midland Naturalist 111(2):383 389. Kotler, B. P. J. S. Brown, and O. Hasson. 1991. Factors affecting g erbil foraging behavior and rates of owl predation. Ecology. 72:2249 2260.
44 Lima, S. L. 1998. Nonlethal effects in the ecology of predator prey interactions: what are the ecological effects of anti predator decision ma k ing? BioScience 48:25 34. Lima, S. L and P. A. Bednekoff. 1999. Temporal variation in danger drives antipredator behavior: the predation risk allocation hypothesis. American Naturalist 153:649 659. Lockard, R. B., and D. H. Owings. 1974a. Seasonal variation in moonlight avoidance by banner tail kangaroo rats. J of Mammalogy 55:189 193. Lockard, R. B., and D. H. Owin gs. 1974 b. Mo on related surf ace activity of bannertail (Dipodomys spectabilis) and Fresno (D. nitratoides) kangaroo rats. Animal Behaviour 22:262 273. MacArthur, R. and R. Lev ins. 1964. Competition, habitat selection, and character displacement in a patchy environment. Zoology 51:1207 1210. MacArthur, R. and E. Pianka. 1966. On optimal use of a patchy environment. American Naturalist 100:603 609. McFarlane, R. W. 1963. Disori entation of loggerhead hatchlings by artificial road lighting. Copeia 1963(1):153. Molenaar, J. G., M. E. Sanders, and D. A. Jonkers. 2006. Road lighting and grassland birds: local influence of road lighting on a black tailed godwit population. Ecologica l Consequences of Artificial Night Lighting. Island Press, Wash. D. C. O'Farrell, M. J. 1974. Seasonal activity patterns of rodents in a sagebrush community. J. of Mammalogy 55:809 823. Orrock, J. L., B. J. Danielson, and R. J. Brinkerhoff. 2003. Rodent foraging is affected by indirect, but not by direct, cues of predation risk. Behavioral Ecology. 15(3):433 437. Pearson, P. G. 1953. A field study of Peromyscus populations in Gulf Hammock, Florida. Ecology 34:199 207. Powell, F. and P. B. Banks. 2004. Do house mice modigy their foraging behavior in response to predator odours and habitat? Animal Behaviour 67:753 759. Price, M. V., N. M. Waser, and T. A. Bass. 1984. Effects of moonlight on microhabitat use by desert rodents. J. of Mammalogy 65:353 356. Reith, C. 1982. Insectivorous bats fly in shadows to avoid moonlight. J. of Mammalogy 63(4):685 688. Rich, C. and T. Longcore. 2006. Introduction. Ecological Consequences of Artificial Night Lighting. Island Press, Wash. D. C.
45 Thompson, S. D. 1982. Structure and species composition of desert heteromyid rodent species assemblages: effects of a simple habitat manipulation. Ecology 63:1313 1321. Travers, S. E., D. W. Kaufman, and G. A. Kaufman. 1988. Differential use of experimental habitat patches by foraging Peromyscus maniculatus on dark and bright nights. J. of Mammalogy 69:869 872. Verdolin, J. L. 2006. Meta analysis of foraging and predation risk trade offs in terrestrial systems. Behavioral Ecology and Sociobiology 60:457 464.
ABOUT THE AUTHOR Robbin G. Capers was born in Boulder, Colorado and raised in Azle, Texas. She attended Texas A&M Univer sity, in College Station, Texas, where she earned a B achelor of Science in Wildlife and Fisheries Sciences with a concentration in Wild ife Ecology a nd Management. She has served as a member and officer in the Texas A&M University chapter of the Society for Conservation Biology and the Biology Graduate Student Organization at the University of South Florida. She currently resides in Alaska.