Sex Differences and Space Use of Breeding Common Ravens In

The Condor 106:529±539 ᭧ The Cooper Ornithological Society 2004

SEX DIFFERENCES IN SPACE USE OF BREEDING COMMON RAVENS IN WESTERN MARIN COUNTY, CALIFORNIA

JENNIFER E. ROTH1,2,4,JOHN P. K ELLY3,WILLIAM J. SYDEMAN2 AND MARK A. COLWELL1 1Department of Wildlife, Humboldt State University, Arcata, CA 95521 2Point Reyes Bird Observatory, 4990 Shoreline Highway, Stinson Beach, CA 94970 3Cypress Grove Research Center, Audubon Canyon Ranch, Marshall, CA 94940

Abstract. We estimated core areas and home-range sizes and evaluated sex differences in home-range size, seasonal variation in movements, and space use for breeding Common Ravens (Corvus corax) in western Marin County, California. There were no signi®cant interannual differences in average core area or average home-range size for either sex, although there were small-scale shifts in home-range use for 67% of females and 63% of males. There was no signi®cant difference in home-range size between the sexes. Home- range size was positively correlated with distance to food source. Ravens traveled shorter distances from the nest during the incubation ϩ nestling stage compared to other stages of the breeding season. Core areas were centered around nest sites and food sources, with signi®cantly aggregated locations for 83% and 100% of females in 2000 and 2001, respectively. Males were more variable in their use of space, with signi®cantly aggregated locations for 38% and 44% of males in 2000 and 2001, respectively. Individual variation in home- range size, movements, and space use was likely due to differences in the distribution of important food sources in the area. Key words: Common Raven, core area, Corvus corax, home range, seasonal variation, sex differences, space use.

Diferencias Sexuales en la UtilizacioÂn del Espacio en Corvus corax en el Oeste del Condado Marin, California Resumen. Estimamos los tamanÄos del aÂrea nuÂcleo y rango de hogar y evaluamos las diferencias entre sexos en el tamanÄo del rango de hogar, variaciones estacionales de los movimientos y utilizacioÂn del espacio en Corvus corax en el oeste del condado Marin, California. No hubo diferencias interanuales signi®cativas en el aÂrea nuÂcleo promedio ni en el rango de hogar promedio para ninguno de los sexos, aunque para el 67% de las hembras y el 63% de los machos hubo cambios de pequenÄa escala en el uso del rango de hogar. No hubo diferencias signi®cativas en el tamanÄo de los rangos de hogar entre sexos. El tamanÄo del rango de hogar estuvo positivamente correlacionado con la distancia a la fuente de alimento. Los cuervos viajaron distancias maÂs cortas desde el nido durante los perõÂodos de incubacioÂnydecrõÂa de polluelos en comparacioÂn con otros perõÂodos durante la eÂpoca reproductiva. Las aÂreas nuÂcleo se centraron en los sitios de nidi®cacioÂn y las fuentes de alimento, con posiciones signi®cativamente agregadas para el 83% y 100% de las hembras en 2000 y 2001, respectivamente. Los machos presentaron mayor variabilidad en la utili- zacioÂn del espacio, con posiciones signi®cativamente agregadas para un 38% y 44% de los machos en el 2000 y 2001, respectivamente. Las variaciones individuales en el tamanÄo de los aÂmbitos de hogar, movimientos y uso del espacio fueron probablemente dadas por las diferencias en la distribucioÂn de los recursos alimenticios importantes en el aÂrea.

INTRODUCTION and the impact of raven predation on popula- Common Ravens (Corvus corax) are an integral tions of several threatened or endangered wild- part of the environment throughout the western life species has led to questions regarding the United States. They are highly adaptable and effective management of raven populations have bene®ted from the presence of humans in (Little®eld 1995, Gaston and Elliot 1996, Boar- many areas (Marzluff et al. 1994). Concern over man 2003). These questions cannot be answered increasing raven populations (Sauer et al. 1997) without a thorough understanding of raven ecology and spatial distribution. Investigating the factors in¯uencing home- Manuscript received 10 September 2003; accepted 9 April 2004. range size and movements can provide impor- 4 E-mail: [email protected] tant information on how individuals are distrib-

[529] 530 JENNIFER E. ROTH ET AL. uted across the landscape. A home range is the derstanding of raven ecology and management area used by an animal during normal activities options in western Marin County, California. We such as food gathering, mating, and caring for used radio-telemetry to better understand the young (Burt 1943, Brown and Orians 1970). Sex factors in¯uencing the spatial distribution of ra- is one factor that may in¯uence home-range size vens in that area. Our objectives were to (1) es- and movements during the breeding season timate core areas and home-range sizes and (2) (Odum and Kuenzler 1955, Schoener 1968). In evaluate sex differences in home-range size, sea- ravens, females are largely responsible for in- sonal variation in movements, and space use cubation; males forage away from the nest and within home ranges for breeding ravens. return to feed females (Boarman and Heinrich 1999). As a result, it may be expected that fe- METHODS male home ranges will be smaller than male Western Marin County, California (38Њ04ЈN, home ranges during the breeding season. Stage 122Њ43ЈW; Fig. 1) has a Mediterranean climate, of the breeding season is another factor that may and elevations range from sea level to 784 m. in¯uence home-range size and movements dur- Lagoons, bays, estuaries, ocean beaches, and ing the breeding season (Odum and Kuenzler rocky cliffs characterize the western boundary. 1955, Schoener 1968). In ravens, attachment to Vegetation types include coastal saltmarsh, the nest begins during the preincubation stage, coastal beach dune, and northern coastal scrub increases during the incubation and nestling vegetation along the western boundary; grass- stages when both sexes have duties at the nest, lands, oak woodlands, and oak savannahs and decreases during the ¯edgling stage when throughout the low hills of the interior; and family groups begin to travel away from the nest mixed evergreens and bishop pines (Pinus mu- (Boarman and Heinrich 1999). Given these sea- ricata) at the higher elevations of the Coast sonal differences in the attachment of ravens to Range (Shuford 1993). Other vegetation types nests, it may be expected that the movements of found interspersed throughout the study area in- both sexes will be restricted during the incuba- clude freshwater marsh, riparian vegetation, and tion and nestling stages compared to the prein- stands of eucalyptus (Eucalyptus globulus), cubation and ¯edgling stages of the breeding Monterey cypress (Cupressus macrocarpa), and season. Monterey pine (P. radiata). The area includes Another important factor in understanding the land managed by Point Reyes National Sea- spatial distribution of individuals is knowledge shore, Tomales Bay State Park, Golden Gate Na- of space use within home-range boundaries tional Recreation Area, Audubon Canyon (Hayne 1949, Samuel et al. 1985). Resources are Ranch, and private agricultural land. Much of generally patchily distributed throughout the en- the area has been grazed by cattle since the mid- vironment, so animals are likely to use some 1800s (Shuford 1993). Ranches are still active portions of their home ranges more intensively on national, state, and private lands and are a than others (Hayne 1949, Samuel et al. 1985). dominant feature of the landscape. There are Within home ranges, the distributions of suitable several small towns (50 to 1500 people) and nest sites and reliable food sources are likely to many people visit the area from the nearby (20 determine areas of concentration (Odum and km) San Francisco metropolitan area. Human Kuenzler 1955, Samuel et al. 1985). Ravens ex- use of park lands is concentrated at national and ploit a wide variety of nesting habitats and food state park trailheads, picnic areas, boat ramps, items and concentrate in areas of abundant food and beaches. (Ewins et al. 1986, Engle and Young 1992). In areas with limited nesting habitat and concen- CAPTURE METHODS trated food sources, it may be expected that ra- We captured ravens prior to and during the vens will use some portions of their home ranges breeding season (December to July) from 1998 more intensively than others. to 2001. We captured ravens using a CODA En- Concern over the impact of raven predation terprises 86±6000 Netlauncher (Mesa, Arizona) on Snowy Plover (Charadrius alexandrinus), or M & M Fur Company Victor 1.5 Soft Catch Common Murre (Uria aalge), Great Blue Heron Coil Spring Traps (Bridgewater, South Dakota). (Ardea herodias), and Great Egret (Ardea alba) We baited trapping locations with carcasses, populations highlighted the need for a better un- meat scraps, or human foods for approximately COMMON RAVEN SPACE USE 531

2 weeks before we attempted to capture ravens. tained locations for each bird across all time pe- Observers attended traps continuously through- riods. There were occasions when we were un- out each trapping session and removed birds able to locate birds. On these occasions, observ- from traps within a few minutes of capture. We ers conducted each search for 1 hr to ensure a held ravens for up to 60 min. minimum tracking effort for each bird. We measured the weight, culmen length, and Due to the distances between radio-tagged wing chord of each raven and sexed birds based birds and the limited road access in the study on size. Due to overlap in the sizes of female area, we were unable to locate birds in a ran- and male ravens, we con®rmed the sex of each domly assigned order. In 2000, we divided the bird by noting the presence of brood patches at birds into two groups, based on their proximity the time of capture and sex-speci®c behaviors to one another. The ®rst group consisted of birds (e.g., courtship displays, incubation, nest guard- residing along Sir Francis Drake Boulevard, and ing, etc.) throughout the breeding season (Boar- the second group consisted of birds residing man and Heinrich 1999). We used mouth color along Pierce Point Road (Fig. 1). Observers al- to age birds (Kertuu 1973). We individually col- ternated tracking effort between the two areas, or banded each bird and attached backpack- systematically searching for all birds known to mounted radio-transmitters (Model 5955, Ad- frequent an area. In each area, observers alter- vanced Telemetry Systems, Isanti, Minnesota) nated between the two ends of the road in de- weighing 21.6±22.2 g (Ͻ3% body weight) to termining the starting point for the tracking ses- adult birds. We made transmitter harnesses from sion. We set receivers (Model R4000, Advanced Te¯on ribbon. The harness design was similar to Telemetry Systems, Isanti, Minnesota) on a 4- that outlined in Buehler et al. (1995), except that sec scan that searched all frequencies, and we we used two 45-cm strands of ribbon to form tracked individual signals in the order heard. We the harness. We placed the strands along the added a third group of birds residing along scapulae, along either side of the sternum, and Highway 1 and the Marshall-Petaluma Road as behind the wings. We placed two 2.5-cm cross- the study progressed. pieces perpendicular to the sternum to connect To minimize bias associated with encounter- the two strands. The crosspieces were 2.5 cm ing frequencies along each route in the same se- apart, and we placed them approximately two- quence in 2001, we divided the birds into two thirds of the way posterior from the top of the groups (birds residing along Sir Francis Drake sternum. We color-coded each transmitter anten- Boulevard and Pierce Point Road and birds re- na using electrical heat-shrink tubing to identify siding along Highway 1 and the Marshall-Peta- individuals. We attempted to attach transmitters luma Road; Fig. 1), then further divided birds in to equal numbers of female and male ravens. each group into three subgroups of three or four. When we captured both members of a pair, we We established routes that rotated the three sub- radio-tagged either the female or male bird groups in each area through morning, midday, based on which sex was underrepresented in the and evening sampling periods and considered a sample. In 2001, we radio-tagged both members set of routes complete once we had searched for of one pair in order to obtain additional infor- each bird during each of the three periods. We mation on their impact on a nearby Great Blue randomly selected the order of the routes and the Heron and Great Egret colony. order that we looked for birds within each period. RADIO-TELEMETRY Each time we detected a bird, we recorded its To ensure coverage throughout the breeding sea- location on a 7.5Ј U.S. Geological Survey to- son, observers followed ravens from February pographic map. We con®rmed locations visually through July of 2000 and 2001. We did not con- whenever possible and triangulated locations duct more than one search for an individual in when we were able to con®dently determine the a given day in order to minimize temporal de- direction of the signal. We included estimated pendence of sequential observations. To account locations to reduce bias associated with limiting for diurnal patterns of movement, we divided analyses to easily accessible locations. We also each day into morning (®rst 4 hr of daylight), mapped nest sites and major food sources. Based midday (all hours between morning and even- on preliminary observations of ravens in west- ing), and evening (last 4 hr of daylight) and ob- ern Marin County (Roth et al., unpubl. data) and 532 JENNIFER E. ROTH ET AL.

FIGURE 1. Map of western Marin County, California, and roads used to track Common Ravens, 2000 and 2001 (California Gap Analysis Project 1999). results of previous studies, we de®ned major mined stage of the breeding season for each bird food sources as dairy ranches (Engle and Young through nest searching and nest observations in 1992), waterbird colonies (Little®eld 1995, Gas- 2001. For each marked raven, we divided the ton and Elliot 1996), and areas of concentrated breeding season into the following stages: pre- human activity where human foods were avail- incubation, incubation ϩ nestling, and ¯edgling able (Knight et al. 1993). Waterbird colonies in- (Boarman and Heinrich 1999). The preincuba- cluded Great Blue Heron, Great Egret, and tion stage lasted from 2±4 weeks prior to nest Common Murre colonies. Areas of concentrated building through the completion of egg laying. human activity included parking lots, picnic ar- The incubation ϩ nestling stage spanned the pe- eas, boat ramps, and beaches. We used the cen- riod between the onset of incubation and chick ters of colonies, parking lots, and picnic areas departure from the nest. The ¯edgling stage was and the point on the beach nearest the raven nest the period after chicks departed from the nest when mapping food sources. until the end of July. We used direct observations of adults or chicks on nests and behavioral BREEDING STAGE cues (e.g., chick vocalizations, chick feeding, In order to gather information on the effect of etc.) to determine nesting stage (Martin and breeding stage on raven movements, we deter- Geupel 1993). COMMON RAVEN SPACE USE 533

STATISTICAL ANALYSES ship between home-range size and distance be- Prior to data analysis, observers transferred all tween the nest and nearest food source prior to locations onto electronic topographic maps, cre- conducting the ANCOVA. Individual ravens ating maps for each bird in each year using were the unit of analysis. We restricted the anal- ArcView GIS (ESRI 1998). We then used the ysis to the 2001 data because home-range esti- ®xed-kernel method to calculate core areas and mates were based on greater than 60 locations home ranges for each bird. We de®ned core area for all birds, providing the most accurate and as the area, or group of subareas, encompassing comparable home-range estimates obtained dur- 50% of the probability distribution. We de®ned ing the 2 years of study. Additionally, we had home range as the area, or group of subareas, information on nest locations and could calcu- encompassing 95% of the probability distribu- late covariate values for more individuals in tion. Kernel methods are considered superior to 2001. We excluded three birds prior to analysis: the minimum convex polygon method because one member of the radio-tagged pair to ensure they measure how intensively an animal uses independence of data, one bird whose nest we different parts of its range (Kernohan et al. were unable to ®nd, and one bird that renested 2001) and do not include large areas not used after a failed attempt, which made it impossible by the animal (White and Garrott 1990). Kernel to objectively determine which nest site to use methods are also considered superior to the har- as a reference point. Additionally, we excluded monic-mean method because they are nonpara- one bird on the basis of residual plots and re- metric and are not sensitive to grid size or spac- gression diagnostics that indicated that the home-range size associated with that bird was ing (White and Garrott 1990). The ®xed-kernel both an outlier and highly in¯uential (A®® and method provides home-range estimates that are Clark 1999). The outlier was the result of one more accurate and precise than adaptive-kernel male that frequently visited a distant food source estimates when the appropriate smoothing pa- and, therefore, had a large home range relative rameter is used (Kernohan et al. 2001). We used to the distance between the nest and nearest food least-squares cross-validation to choose smooth- source. We log transformed home-range esti- ing parameter values that minimized the dis- mates in order to normalize the data. crepancy between estimated and true densities We assessed site ®delity between years using without making assumptions regarding the un- multiresponse permutation procedures (MRPP), derlying distribution of the data (Kernohan et al. a nonparametric technique that determines 2001). In all but one case, we estimated home whether two sets of locations come from a com- ranges for breeding birds with greater than 30 mon distribution (Kernohan et al. 2001), in locations per year in order to increase the reli- BLOSSOM (Cade and Richards 2001). Speci®- ability of estimates (Kernohan et al. 2001); we cally, MRPP compares average intragroup dis- used 20 locations for one bird in 2000 due to a tances to average distances obtained from all late capture. Kernel methods are relatively in- other possible groupings (Biondini et al. 1988). sensitive to the effects of outliers (Kernohan et MRPP can assess home-range shifts between al. 2001), and we included all locations in esti- years by grouping the locations from each year mation of core areas and home ranges. We used and using distances between locations within the Animal Movement extension to ArcView each year as the basis for comparison (Kernohan GIS to make the calculations (Hooge and Ei- et al. 2001). MRPP is superior to methods for chenlaub 1997). We paired core-area and home- comparing home-range centroids because of its range estimates between years for the subset of ability to assess expansion or contraction of a birds that were followed in both years. We then home range when the centroid remains the same used paired t-tests to compare average core area (Kernohan et al. 2001). Additionally, MRPP is and average home-range size between years for superior to methods assessing home-range over- each sex. lap because it does not include areas that are not We used ANCOVA to compare female and used in a given year in its assessment (Kernohan male home-range sizes during the breeding sea- et al. 2001). son, with distance between the nest and nearest We used repeated-measures ANOVA to ex- food source as the covariate (Kuehl 2000). We amine variation in raven movements in relation used linear regression to establish the relation- to stage of the breeding season (Kuehl 2000). 534 JENNIFER E. ROTH ET AL.

TABLE 1. Average core area (50% kernel) and home-range (95% kernel) size of breeding Common Ravens in western Marin County, California, in 2000 and 2001. Annual variation was not statistically signi®cant for either sex.

2000 2001 Mean Ϯ SE Range Mean Ϯ SE Range Females (n ϭ 6) Core area (ha) 26.2 Ϯ 11.2 1.4±62.9 24.4 Ϯ 14.8 2.1±96.5 Home range (ha) 135.6 Ϯ 54.2 19.7±304.2 140.3 Ϯ 74.9 15.7±484.5 Males (n ϭ 8) Core area (ha) 11.7 Ϯ 4.1 2.6±31.4 12.5 Ϯ 2.6 2.6±93.3 Home range (ha) 117.9 Ϯ 51.6 26.6±437.0 89.0 Ϯ 15.2 28.0±485.8

Speci®cally, we compared average distances be- home-range sizes were compared for the subset tween locations and nests during the preincu- of birds that were followed in both years (Table bation, incubation ϩ nestling, and ¯edgling stag- 1; paired t-tests, all P Ͼ 0.5). Individual birds es for females and males. Individual birds were remained in the same part of the study area be- the unit of analysis. We restricted analysis to the tween years, although small-scale home-range 2001 data because we could not determine the shifts were evident for several birds (Table 2). timing of each stage for most birds in 2000. We MRPP indicated signi®cant shifts in the distri- included only successful breeders in the analysis bution of locations for 67% (4 of 6) of females in order to meet the requirement for a balanced and 63% (5 of 8) of males between years. design. For this analysis, we again excluded one Sex differences. Core areas and home-range member of the radio-tagged pair to ensure in- sizes were highly variable for both sexes (Roth dependence between subjects, as well as the one 2003). Female home-range size averaged 179.9 bird whose nest we were unable to ®nd. We log Ϯ 74.7 ha (n ϭ 7) and male home-range size transformed distances prior to analysis in order averaged 80.3 Ϯ 20.5 ha (n ϭ 5) when home- to normalize the data. We used Tukey multiple range sizes were estimated for all birds in 2001. comparison procedures for pairwise compari- There was a positive relationship between home- sons of differences among means. range size and distance between the nest and We conducted nearest-neighbor analyses to nearest food source (r2 ϭ 0.64, P ϭ 0.002; Fig. evaluate space use within each bird's home 2). There was no signi®cant difference in home- range. Speci®cally, we used these analyses to range size between the sexes when distance be- determine whether locations within the bound- tween the nest and nearest food source was used aries of each home range exhibited random, uni- as a covariate (ANCOVA: F1,9 ϭ 0.3, P ϭ 0.59). form, or clumped distributions (Clark and Evans 1954). We were not able to combine data from SEASONAL VARIATION IN MOVEMENTS the 2 years of the study, and we conducted near- Stage of the breeding season had a signi®cant est-neighbor analyses separately for each year effect on the distance ravens traveled from the using the Animal Movement extension to nest, but sex did not (ANOVA: season: F2,14 ϭ ArcView GIS (Hooge and Eichenlaub 1997). We 27.9, P Ͻ 0.001; sex: F1,7 ϭ 0.03, P ϭ 0.86). included all locations within the boundaries of There was a sex ϫ season interaction (F2,14 ϭ the home range in each analysis. Means Ϯ SE 4.7, P ϭ 0.03). Females traveled shorter dis- are presented throughout the text. We assumed tances from the nest during the incubation ϩ statistical signi®cance if P Ͻ 0.05. nestling stage than in either the preincubation or ¯edgling stages (Tukey MCP, both P Ͻ 0.001; RESULTS Fig. 3). Males traveled shorter distances from CORE AREA AND HOME-RANGE SIZE the nest during the incubation ϩ nestling stage Annual variation. There were no signi®cant in- compared to the preincubation stage (P Ͻ 0.03) terannual differences in average core area or av- and similar distances during the incubation ϩ erage home-range size for either sex when nestling and ¯edgling stages (P Ͼ 0.2). There COMMON RAVEN SPACE USE 535

TABLE 2. Home-range shifts of breeding Common Ravens in western Marin County, California, from 2000 to 2001.

2000 2001 Mean Mean Sex distance (m)a nb distance (m)a nb ␦c P Females (n ϭ 6) 902 69 413 75 ±1.7 0.07 271 56 620 77 ±40.9 Ͻ0.001 1009 41 923 74 ±13.2 Ͻ0.001 456 62 317 76 ±2.2 0.04 409 60 330 78 0.4 0.54 1306 60 1922 63 ±5.4 0.003 Males (n ϭ 8) 530 20 696 77 ±11.5 Ͻ0.001 258 50 336 79 0.4 0.56 376 34 720 77 ±4.4 0.006 1239 57 709 72 ±5.2 0.003 374 63 684 67 ±9.3 Ͻ0.001 439 38 592 73 0.1 0.43 1144 51 923 68 ±0.6 0.18 379 34 419 75 ±4.8 0.004 a Mean distance between radio-telemetry locations. Mean distances for the two years were compared to mean distances obtained from all other possible groupings using multiresponse permutation procedures. Signi®cant results indicate a shift in the distribution of locations between years. b Number of radio-telemetry locations. c Standardized test statistic based on a Pearson Type III distribution. were no signi®cant differences between the pre- gated locations for 3 of 8 and 4 of 9 individuals incubation and ¯edgling stages for females (P Ͼ in 2000 and 2001, respectively (Fig. 4). All birds 0.5) or males (P Ͼ 0.1). had a core subarea centered around their nest site in both years of the study (Fig. 5). In some SPACE USE WITHIN HOME RANGES cases, concentrated food sources occurred within Locations were signi®cantly aggregated within those core subareas. For birds with multiple cen- the home ranges of most females in both years ters of activity, additional core subareas were (5 of 6 and 7 of 7 females in 2000 and 2001, centered around food sources, including dairy or respectively; Fig. 4). Males were more variable horse ranches, waterbird colonies, or areas of in their use of space, with signi®cantly aggre-

FIGURE 3. Sex differences and seasonal variation in FIGURE 2. Relationship between home-range size distance (mean Ϯ SE) between radio-telemetry loca- (ha) and distance (m) between the nest and nearest tions and nests of Common Ravens in western Marin food source for breeding Common Ravens in western County, California, 2001. Data were log transformed Marin County, California, 2001. prior to analysis. 536 JENNIFER E. ROTH ET AL.

FIGURE 4. Sex differences in the spatial distribution of locations of breeding Common Ravens in western Marin County, California, 2000±2001. Each data point FIGURE 5. Frequency with which ®xed-kernel core represents the summary of a radio-tagged raven's lo- subareas included nest sites and four types of concen- cations. Index values less than or greater than one rep- trated food sources for Common Ravens in western resent aggregated and uniform patterns, respectively. Marin County, California, 2000±2001. One or more Values falling outside of the dashed lines deviate sig- core subareas are represented for 14 birds in 2000 and ni®cantly (P Ͻ 0.05) from a random distribution. 16 birds in 2001. concentrated human activity, in all but one case. Ravens traveled shorter distances from the In 2000, one bird had a core subarea that was nest during the incubation and nestling stage not clearly associated with a nest site or any of compared to other stages of the breeding season. the food sources de®ned in this study. Seasonal variation in movements is likely due to the greater attachment of both sexes to the nest DISCUSSION during the incubation and nestling stages. A va- SEX DIFFERENCES AND SEASONAL riety of other birds have been shown to reduce VARIATION IN SPACE USE their home-range size and travel shorter dis- Home-range sizes did not differ between female tances from the nest during this time (Grahn and male ravens in western Marin County. 1990, Marzluff et al. 1997, Plissner et al. 2000). These results are consistent with a comparison Within home ranges, the distribution of female of female and male ravens in southern California locations tended to be more aggregated than the (Linz et al. 1992). Though sex has been shown distribution of male locations. The greater at- to in¯uence home-range size in some species tachment of females to the nest during the in- (e.g., Jenkins and Benn 1998), sex roles may not cubation period likely accounts for this differ- differ enough during the breeding season to af- ence. fect overall home-range size in ravens. Alter- Home-range structure was highly variable for natively, similar home-range sizes may re¯ect both sexes; individual home ranges were either courtship, pair-bond reinforcement, increased continuous or discontinuous and had one or vigilance by the male during egg laying and in- more subareas of concentration. All birds had a cubation, or other social behaviors associated core subarea centered around their nest site. In with nesting (Boarman and Heinrich 1999). Ad- some cases, the nearest food source was within ditionally, home-range size may not be an ap- that core subarea. When there was more than propriate measure for examining sex differences one area of concentration, additional core sub- in space use because it fails to account for de- areas were centered around concentrated food tailed movements that occur within an indivi- sources in all but one case. Thus, it appears that dual's home range. For example, other research- the distribution of nest sites and food sources ers have found sex differences in linear dis- in¯uences space use within individual home tances traveled (Plissner et al. 2000) and time ranges for ravens in western Marin County. Ro- spent near the nest (Gerrard et al. 1992) that lando (1998) concluded that individual differ- were not evident in home-range comparisons. ences in food preference and the availability and COMMON RAVEN SPACE USE 537 distribution of food resources within individual bias associated with these factors is likely very home ranges were responsible for similar vari- small, though, given the open terrain and great ation in home-range structure for European Jays visibility of our study area. Finally, even though (Garrulus glandarius). Other studies have doc- we tried to sample ravens from a variety of hab- umented wide variation in raven diets that re- itats, captures tended to occur in accessible areas ¯ects foods available within individual home along the coast and these results may be in¯u- ranges (Ewins et al. 1986, Stiehl and Trautwein enced by larger habitat or landscape features. 1991). There were no signi®cant interannual differences in average core area or average home- HOME-RANGE SIZE AND ANNUAL VARIATION range size for either sex. In addition, there were Elsewhere, Common Raven home-range esti- no large-scale shifts in home-range placement mates have ranged from 120 to 125 200 ha between years. This result is consistent with (Boarman and Heinrich 1999). Home ranges in studies of other corvids. Grahn (1990) reported western Marin County were at the lower end of that European Jays did not move more than a this spectrum. Several biological factors may few hundred meters between years. Similarly, have contributed to this result, including the Black-billed Magpies (Pica hudsonia) often presence of abundant and reliable food sources changed nest sites, but remained in the same lo- in the area. Ravens in western Marin County cale between years (Dhindsa and Boag 1992). feed extensively on grain, calf carcasses, and af- Site ®delity may be bene®cial because it en- terbirths found at many local dairies (Roth et al., hances familiarity with an area, which aids in unpubl. data). Additionally, ravens obtain food food acquisition, access to cover, and territory from human refuse at beaches, parking lots, and defense (Greenwood and Harvey 1982). dumpsters in the area and prey on eggs and Despite the high degree of home-range over- young waterbirds at local colonies (Roth et al., lap between years, signi®cant small-scale shifts unpubl. data). All of the ravens in this study in home-range placement were evident for some nested within 3.7 km of one of these food sourc- birds. In some cases, these shifts were associated es and did not have to range widely in search of with changes in nest locations between years food. The positive correlation between home- that were substantial enough to in¯uence the lo- range size and distance to food source in this cations of core subareas. While we did not spe- study is consistent with other research indicating ci®cally investigate the effect of nesting success that prey abundance and distance to food and on site ®delity, success did not appear to in¯u- water in¯uence home-range size and movements ence changes in the locations of nests or core in birds (Schoener 1968, Engle and Young 1992, subareas between years. Other researchers have Marzluff et al. 1997). Finally, we studied only failed to ®nd a relationship between reproduc- breeding adults during the breeding season. In tive success and nest-site ®delity for corvids general, juveniles and nonbreeding adults range (Dhindsa and Boag 1992). Other factors that farther than breeding birds, and home ranges may account for some of the small-scale shifts may be smaller during the breeding season include similar small-scale shifts in the distri- (Grahn 1990, Linz et al. 1992). Additional bio- bution of desirable food and water resources. logical factors that may in¯uence home-range Given the apparent association between ravens size, but were not addressed in this study, in- and agricultural activities and areas of concen- clude body size, population density, intraspeci®c trated human use in western Marin County, var- and interspeci®c interactions, and vegetation iation in grazing or harvesting rotations or in the type (Schoener 1968, Grahn 1990, Rolando distribution of and access to human foods most 1998). likely affected space use. Other researchers have Our study design also may have contributed reported similar associations (AndreÂn 1992, En- to the small home ranges of ravens in western gle and Young 1992, Knight et al. 1993). Vari- Marin County. Home ranges may be underesti- ation in the phenology or reproductive success mated to some extent due to dif®culty of track- of nearby waterbird colonies may have had a ing birds in less accessible areas. Loss of anten- similar effect. nae from six transmitters made tracking dif®cult Our results suggest that sex and season in¯u- when birds were not visible and also may have ence space use for breeding ravens and that caused home ranges to be underestimated. The much of the individual variation in home-range 538 JENNIFER E. ROTH ET AL. size, movements, and space use is due to the on desert tortoises. Environmental Management distribution of food sources in the area. These 32:205±217. BOARMAN, W. I., AND B. HEINRICH. 1999. Common Ra- results may have important implications for re- ven (Corvus corax). In A. Poole and F. Gill [EDS.], source managers concerned with the effects of The birds of North America, No. 476. The Birds increasing raven populations on sensitive spe- of North America, Inc., Philadelphia, PA. cies. For example, many of the food sources BROWN,J.L.,AND G. H. ORIANS. 1970. Spacing pat- used by ravens in western Marin County are di- terns in mobile animals. Annual Review of Ecol- ogy and Systematics 1:239±262. rectly related to human activities, and resource BUEHLER, D. A., J. D. FRASER,M.R.FULLER,L.S. managers may be able to manage the distribu- MCALISTER, AND J. K. D. SEAGER. 1995. Captive tion or local abundance of ravens by reducing and ®eld tested radio attachment techniques on their access to those food sources. Effectively Bald Eagles. Journal of Field Ornithology 66: managing raven distribution or abundance may, 173±180. BURT, W. H. 1943. Territoriality and home range con- in turn, reduce predation pressure on sensitive cepts as applied to mammals. Journal of Mam- species (Boarman 2003). In addition, our results malogy 24:346±352. indicate that breeding ravens occupy the same CADE, B. S., AND J. D. RICHARDS. 2001. User manual home ranges for multiple years, with little an- for BLOSSOM statistical software. U.S. Geolog- nual variation in overall home-range size and ical Survey Midcontinent Ecological Science Cen- placement. Therefore, management actions that ter, Fort Collins, CO. CALIFORNIA GAP ANALYSIS PROJECT. 1999. Gap analy- discourage ravens from utilizing food sources or sis of mainland California: an interactive atlas of occupying selected areas may have long-lasting terrestrial biodiversity and land management. In- effects. teractive CD-ROM Version. University of Cali- fornia, Biogeography Lab, Santa Barbara, CA. ACKNOWLEDGMENTS CLARK, P. J., AND F. C. EVANS. 1954. Distance to nearest neighbor as a measure of spatial relationships We would like to extend our sincere thanks to every- in populations. Ecology 35:445±453. one involved in this project. The insights of S. Allen, DHINDSA,M.S.,AND D. A. BOAG. 1992. Patterns of L. George, R. Golightly, G. Hendrickson, and J. Marz- nest site, territory, and mate switching in Black- luff improved this work. E. Neatherlin provided in- billed Magpies (Pica pica). Canadian Journal of struction in trapping techniques. J. Awkerman, R. Dymzarov, M. McCaustland, E. Punkay, and A. Rossi Zoology 70:633±640. provided invaluable ®eld assistance. D. Adams, S. Al- ENGLE,K.A.,AND L. S. YOUNG. 1992. Movements and len, L. Campbell, K. Fehring, M. Flannery, G. Frasier, habitat use by Common Ravens from roost sites S. Guers, A. Meehan, and M. Ruhlen provided addi- in southwestern Idaho. Journal of Wildlife Man- tional assistance with raven trapping. D. Dobkin, H. agement 56:596±602. Powell, and two anonymous reviewers provided help- ESRI. 1998. ArcView GIS Version 3.1. Environmental ful comments on earlier drafts. Several private land- Systems Research Institute, Inc., Redlands, CA. owners graciously allowed access to their property EWINS, P. J., J. N. DYMOND, AND M. MARQUISS. 1986. during the course of this study. We conducted animal The distribution, breeding, and diet of Ravens capture and handling with the approval of the Hum- Corvus corax in Shetland. Bird Study 33:110± boldt State University Institutional Animal Care and 116. Use Committee (# 00\01.W.37.A). Point Reyes Bird GASTON,A.J.,AND R. D. ELLIOT. 1996. Predation by Observatory, Audubon Canyon Ranch, Point Reyes Ravens Corvus corax on Brunnich's Guillemot National Seashore, and the California Department of Uria lomvia eggs and chicks and its possible im- Fish and Game provided support for this project. This pact on breeding site selection. Ibis 138:742±748. is PRBO contribution no. 1123. GERRARD, J. M., A. R. HARMATA, AND P. N. GERRARD. 1992. Home range and activity of a pair of Bald LITERATURE CITED Eagles breeding in northern Saskatchewan. Jour- nal of Raptor Research 26:229±234. AFIFI,A.A.,AND V. C LARK. 1999. Computer-aided multivariate analysis. 3rd ed. CRC Press, Boca GRAHN, M. 1990. Seasonal changes in ranging behav- Raton, FL. ior and territoriality in the European Jay Garrulus ANDREÂ N, H. 1992. Corvid density and nest predation g. glandarius. Ornis Scandinavica 21:195±201. in relation to forest fragmentation: a landscape GREENWOOD, P. J., AND P. H. HARVEY. 1982. The natal perspective. Ecology 73:794±804. and breeding dispersal of birds. Annual Review of BIONDINI, M. E., P. W. MIELKE, AND E. R. REDENTE. Ecology and Systematics 13:1±21. 1988. Permutation techniques based on Euclidean HAYNE, D. W. 1949. Calculation of size of home range. analysis spaces: a new and powerful statistical Journal of Mammalogy 30:1±17. method for ecological research. Coenoses 3:155± HOOGE, P. N., AND B. EICHENLAUB. 1997. Animal 174. Movement extension to ArcView, Version 1.1. BOARMAN, W. I. 2003. Managing a subsidized predator U.S. Geological Survey, Alaska Biological Sci- population: reducing Common Raven predation ence Center, Anchorage, AK. COMMON RAVEN SPACE USE 539

JENKINS, A. R., AND G. A. BENN. 1998. Home range 1997. The in¯uence of habitat, prey abundance, size and habitat requirements of Peregrine Falcons sex, and breeding success on the ranging behavior on the Cape Peninsula, South Africa. Journal of of Prairie Falcons. Condor 99:567±584. Raptor Research 32:90±97. ODUM,E.P.,AND E. J. KUENZLER. 1955. Measurement KERNOHAN, B. J., R. A. GITZEN, AND J. J. MILLSPAUGH. of territory and home range size in birds. Auk 72: 2001. Analysis of animal space use and move- 128±137. ments, p. 125±166. In J. J. Millspaugh and J. M. PLISSNER, J. H., L. W. ORING, AND S. M. HAIG. 2000. Marzluff [EDS.], Radio tracking and animal pop- Space use of Killdeer at a Great Basin breeding ulations. Academic Press, San Diego, CA. area. Journal of Wildlife Management 64:421± KERTUU, M. E. 1973. Aging techniques for the Com- 429. mon Raven (Corvus corax principalis) Ridgeway. ROLANDO, A. 1998. Factors affecting movements and M.Sc. thesis, Michigan Technical University, home ranges in the Jay (Garrulus glandarius). Houghton, MI. Journal of Zoology 246:249±257. KNIGHT, R. L., H. A. L. KNIGHT, AND R. J. CAMP. 1993. ROTH, J. E. 2003. Sex differences in space use of Raven populations and land-use patterns in the breeding Common Ravens (Corvus corax)in Mojave Desert, California. Wildlife Society Bul- western Marin County, California. M.Sc. thesis, letin 21:469±471. Humboldt State University, Arcata, CA. SAMUEL, M. D., D. J. PIERCE, AND E. O. GARTON. 1985. KUEHL, R. O. 2000. Design of experiments: statistical principles of research design and analysis, second Identifying areas of concentrated use within the edition. Brooks and Cole, Paci®c Grove, CA. home range. Journal of Animal Ecology 54:711± 719. LINZ, G. M., C. E. KNITTLE, AND R. E. JOHNSON. 1992. Home range of breeding Common Ravens in SAUER, J. R., J. E. HINES,G.GOUGH,I.THOMAS, AND B. G. PETERJOHN. 1997. The North American coastal southern California. Southwestern Natu- breeding bird survey results and analysis. Version ralist 37:199±202. 96.4. Patuxent Wildlife Research Center, Laurel, LITTLEFIELD, C. D. 1995. Demographics of a declining MD. ¯ock of Greater Sandhill Cranes in Oregon. Wil- SCHOENER, T. W. 1968. Sizes of feeding territories son Bulletin 107:667±674. among birds. Ecology 49:123±141. MARTIN,T.E.,AND G. R. GEUPEL. 1993. Nest-moni- SHUFORD, W. D. 1993. The Marin County breeding toring plots: methods for locating nests and mon- bird atlas: a distributional and natural history of itoring success. Journal of Field Ornithology 64: coastal California birds. California Avifauna Se- 507±519. ries 1. Bushtit Books, Bolinas, CA. MARZLUFF, J. M., R. B. BOONE, AND G. W. COX. 1994. STIEHL, R. B., AND S. N. TRAUTWEIN. 1991. Variations Historical changes in populations and perceptions in diets of nesting Common Ravens. Wilson Bul- of native pest bird species in the west. Studies in letin 103:83±92. Avian Biology 15:202±220. WHITE,G.C.,AND R. A. GARROTT. 1990. Analysis of MARZLUFF, J. M., B. A. KIMSEY,L.S.SCHUECK,M.E. wildlife radio-tracking data. Academic Press, San MCFADZEN,M.S.VEKASY, AND J. C. BEDNARZ. Diego, CA.