The Condor92:229-236 Q The CooperOrnithological Society 1990

HABITAT, MOVEMENTS AND ROOST CHARACTERISTICS OF MONTEZUMA IN SOUTHEASTERN ’

MARK R. STROMBERG~ National AudubonSociety, Research Ranch Sanctuary,Box 44, E&n, AZ 85611 and Department of Environmental, Population and OrganismicBiology, Universityof Colorado,Boulder, CO 83009

Abstract. Movements,survivorship, covey size, roostingbehavior, and habitat use of MontezumaQuail ( montezumae)marked with radiotransmitters were studied in thefoothills of theHuachuca Mountains of southeasternArizona from October1986 through November1987. Coveys usedsmall areas(0.09-6 ha) and during midwinter, the same small area would be used for long periods. In late winter and early spring, coveys occupiedmuch larger areasby sequentially spending 3-10 days on adjacent, nonoverlapping areasas large as 50 ha. Daily movements were small (1 S-60 m) for most of the year, and movements between days were often 5 100 m. with radios persisted on the study site as long as 133 days, but most were followed for ~30 days. Covey size varied seasonally,with pairs observedfrom April-May through September. Individually followed coveys of up to six to eight birds declined in numbers from September through April. Slope, aspect, basal vege- tation cover, dominant plant species,and distance to the nearest tree was measuredat each location or flush site. When compared to randomly selectedpoints in the oak savanna habitat, quail preferred southeast-facinghillsides in tall grassesfor night roosts. Day-use areas were selectively on north-facing hillsides. Areas used by quail during the day were generallyon hillsides, about 16 m from the nearestoak tree and had grasscover intermediate between barren areas under oak trees and more dense grasscover farther away from the . Key words: Montezuma Quail; Cyrtonyx montezumae; habitat: movements;radio track- ing; Arizona.

INTRODUCTION terns of Montezuma Quail on land that has re- Montezuma Quail (Cyrtonyx montezumae) are covered from the effectsof grazing by domestic typically found in the understory vegetation of livestock. Specific comparisons were made to oak (Quercus spp.) and oak- (Pinus spp.) learn how Montezuma Quail use available hab- woodlandsof the southwesternUnited Statesand itat in relation to the range of habitats available Mexico (Leopold and McCabe 1957). Popula- (e.g., Brennan et al. 1987). tions of thesequail can be locally eliminated with removal of more than about 55% of the under- STUDY AREA AND METHODS story vegetation (combined annual productivity Observations were made on the ResearchRanch and standing crop) by livestock grazing (Ligon Sanctuary of the National Audubon Society 1927, Miller 1943, Wallmo 1954, Brown 1982). managed in cooperation with the Bureau of Land Previous researchon habitat use by Montezuma Management in Santa Cruz County in south- Quail has been restricted to sites grazed by live- eastern Arizona. This site has changed signifi- stock (Brown 1982) which has a dramatic influ- cantly after removal of livestock grazing and was ence on the landscape (Hastings and Turner very different from adjacent grazed lands (Bock 1965). The purpose of this paper is to describe et al. 1984). Brady et al. (1989) reported that the habitats, home ranges, and movement pat- cover of grasseson the study area increasedfrom 29% in 1969 to 85% in 1984 and that the number of plant speciesnearly tripled (from 22 to 63). * Received 18 August 1989. Final acceptance1 No- The largest increase in species number was in vember 1989. leafy forbs (from 10 to 35) and the most dramatic * Present address:Hastings Natural History Reser- increase in dominance was in the taller mid- vation, University of California-Berkeley, Museum of Vertebrate Zoology, 38601 E. Carmel Valley Road, grasses(Eragrostis intermedia, Bouteloua curti- Carmel Valley, CA 93924. pendufa). Changesin the vegetation on the study

~291 230 MARK R. STROMBERG site were primarily attributed to the removal of to include a flexible nylon netting roof (below grazing (Brady et al. 1989). the wire mesh root) to minimize damage to the Coveys were locatedand use sitesmapped from birds’ heads while in the traps. Bishop (1964) 1983 to 1987. Birds were trapped and marked and Brown (1975) reported attempts to capture from October 1986 to November 1987. The Re- adult Montezuma Quail with funnel traps failed searchRanch is located on the northwestern side becausequail relied on underground plant parts of the Huachuca Mountains and includes rolling for food and would not respond to bait. Quail foothill grasslandand oak savanna at elevations on this site were consistently captured with crops from 1,400 to 1,560 m. Mean precipitation was full of bait grain. 43 cm/year, with most rain falling during mon- Montezuma Quail are cryptic and are virtually soon thunderstorms of July, August, and early impossible to see when they crouch and freeze September. Soils were gravelly loams with scat- in the oak woodland litter (Leopold and McCabe tered limestone outcrops. Dominant grassesin- 1957, Brown 1982). Thus, observations on free- cluded grama (Bouteloua spp.), curly mesquite ranging birds were limited, without radios, to (Hilaria belangeri), and plains lovegrass (Era- surprise flushes.Because preliminary studies in- grostisintermedia). Emory oak (Quercusemoryi) dicated that poncho-mounted transmitters (Heg- and Arizona white oak (Q. arizonica) were com- da1 and Colvin 1986, Cochran 1980) interfered mon in the study site. Both oaks occurred in with crouching and foraging behaviors and con- dramatically higher densities on north-facing tributed to mortality, I used backpacksto attach slopes, although a few were scattered on level transmitters. Radio transmitters were glued and alluvial flats at the base of rolling hills (Bonham tied to small tabs of auto upholstery material. 1972, Bock and Bock 1986). More detailed de- These tabs were tied on quail as backpackswith scriptions of the study site can be found else- ribbon loops beneath the wings. Radio packages where (Bahre 1977, Bock et al. 1984, Bock and weighed less than 4 g with a lo-cm antenna ar- Bock 1986, Brady et al. 1989). ranged to lie along the ’s back (Custom Te- Locating and capturing these quail proved ex- lemetry, Athens, Georgia). I used a hand-held tremely difficult, so several survey and trapping three-element Yagi antenna and radio receiver techniques were used, including using trained (Custom Electronics, Urbana, Illinois). When ra- dogs, groups of people walking through an area dio signals could not be found after extensive in close ranks, and continuous recording of en- ground searches,aerial surveys from fixed-wing counters. I used these data to locate every covey aircraft were conducted. in the central 2,33 l-ha area of the study site. Of 16 coveys consistently located on the study Traps were set in areas used by several coveys. site, radios were placed on birds in eight. Forty- In an effort to attract birds to traps, I briefly held six birds were captured and marked; 15 were chicks in cagesand recorded the repetitive cheep radio-marked. Only one bird per covey was ra- call they gave when separatedfrom their parents. dio-marked at a given time. I recorded similar calls from free-ranging chicks Birds were marked and releasedat the capture separatedafter flushing adults from the chicks. I site as quickly as possible because those held played back recordings of these chick calls and overnight or moved from one study site to another the descendingburr or trill of adults. Responses did not join local coveys, even if coveys were from known coveys to playback were erratic and within 100 m of repeatedradio relocations of the unpredictable, although elsewherein Santa Cruz bird. Small metal tags (National Band and Tag County, Arizona, Montezuma Quail would ap- Co., Newport, Kentucky) were placed through proach playbacks of adult vocalizations (R. Bow- the patagium to identify individuals. Birds were ers, pers. comm.). weighed and classedto age and sex at capture. Funnel traps (Reeves et al. 1968, Braun 1976) Young of the year were not classified by sex. were placed over commercial “mixed wild bird Radio-marked birds which joined flocks were seed” bait and tended every hour during trap- relocated. Birds were relocated usually only once ping. Drift fences made from woven hardware a day and then we left the covey use area to cloth, 0.3 m tall and 30 m long, extended from minimize disturbance to the birds. Night roosts each funnel entrance. Bait stations were main- were located in an effort to capture more birds tained throughout the year. Traps were opened from January to March 1987. Tall and dense between trapping sessions.Traps were modified grasson the ungrazed study site made it difficult HABITAT USE BY MONTEZUMA QUAIL 231

to capture birds on night roostswith trained dogs included in the analysis. Distance to the nearest as reportedfrom nearby grazedstudy sites(Brown oak tree (> 3 cm dbh) was measured with a range 1982). I flagged night roost sites as determined finder or tape. Height of vegetation was esti- by flushing the covey at night; habitat measure- mated by measuring the tallest understory plant ments were taken later during the day. I relocated above the four outermost comers ofthe quadrats. birds with transmitters by triangulation during Data from trap sites were not included in the daylight hours. Distances between radio trans- descriptions of habitat utilization. Ordinations mitters and the receiver were usually less than were based on slope, cover, distance to oak trees, 20 m. Birds with radios could be located within and presence/absenceof plant species(Beals 1984, 2-3 m* and could be mapped to an accuracy of McCune 1987). about 5 m*. Care was taken not to flush the covey A stratified random sample of the study site during daylight observations. I mapped locations was established with transects oriented north- of individual birds with transmitters on aerial south. Three transectswere placed roughly along photos (1’ = 1,000’). Sometimes, the bird in the shared boundaries of areas which would divide covey with the transmitter was inadvertently the study site into quarters. Fifty-one random flushed, as would happen when a member of the points were selectedalong the transectsand hab- covey without a transmitter was disturbed some itat variables were measured as above. These distance from the bird with the transmitter. After were compared to the sites used by the quail. such flushes, the number of birds in the covey Nonparametric, parametric, and radial statistics was recorded by walking the general area and were calculated (Siegel 1956, Zar 1984, respec- flushing all possiblequail, including the bird with tively) to compare sitesused by the quail to ran- the radio transmitter. Not all birds in the covey domly selected points. Sample sizes varied be- may have been flushed, so the number of birds tween analysesdue to missing data values; a few counted each time the covey was flushed was an birds escapedduring handling. estimate. Repeated counts of coveys usually dif- fered between days by only one or two birds. RESULTS AND DISCUSSION Original radio relocations of the covey member with the transmitter were marked and several PERSISTENCE ON THE STUDY SITE quantitative habitat measureswere taken at that Forty-six birds were captured with capture rates sample point (“day-use site”). When making varying from 0.008 to 0.12 bird/trap day. Re- hourly observations of covey movements, trian- capture rates were low, with only 19 birds pro- gulations were made from greater distances (ca. viding data on number of days the birds were 50 m) and I included only the hourly movement known to survive beyond their capture dates. recordsof coveys which were not flushed during Causes associatedwith the last observations of that day’s observations. Use areas were calcu- these 19 birds were: radio failure (7); raptor pre- lated by the minimum convex polygon method dation (5); recapture (4); and canid predation- (Jennrich and Turner 1969). most likely by coyote Canis latrans (3). Mean Six measureswere usedat eachflush site (roosts, number of days known alive was 28.4 (SE = 8.9). some day-use sites)or day-time radio relocations No differences were observed between number (day-use sites) to characterize the habitat used of days birds with or without radios persisted. by the quail. Slope was measuredwith a clinome- For example, bird #371 carried a radio trans- ter. Aspect was measured by locating a line per- mitter for 133 days. That radio was replacedupon pendicular to the slope at each sample point and recapture; 7 days later the bird was killed by a obtaining a compassreading for that line. Total predator (bobcat or coyote). Bird #370 was re- basal cover of vegetation was estimated visually captured after 124 days with only a wing tag. A from four 50- x 20-cm quadrats (Daubenmire transmitter was placed on this bird, and it was 1959) which were placed togetherto sharea com- killed by a Cooper’s Hawk (Accipiter coop-i) 2 mon comer centered on the flush site. Midpoints days later. The covey of this adult was subject of the ranges of each Daubenmire cover class to intense predation by an adult female Cooper’s were averaged. When the exact site of the bird Hawk. From 25 October to 12 November, pre- was not seen as it flushed, quadrats were tossed sumably the same hawk was located each day in into an area within a few meters from the flush a local use area of this covey. Four young of the site. Plant specieswith frequency < 1% were not year and two adults were found in this raptor’s 232 MARK R. STROMBERG

TABLE 1. Frequency(%) of plant speciesat Mon- varied (Table 1). Siteswhere birds were relocated tezumaQuail day-use,random, and roostingsites. during the day differed (x2 = 32.3, 2 x 12, P < 0.001) from randomly selectedsites. Major con- Day-use Random Roost Species II= 116 n=51 n = 20 tributions to the chi-square value came from ap- parent avoidance of sites with a biennial forb, Boutelouacurtipendula 68.1 66.6 95.0 Heterothecasubaxillaris 27.6 47.1 60.0 camphor weed (Heterotheca subaxillaris), and Eragrostisintermedia 27.6 19.6 20.0 selection of sites with relatively tall grasses,in- Boutelouagracilis 21.5 11.7 5.0 cluding plains lovegrass(Eragrostis intermedia) Heteropogoncontortus 11.2 19.6 55.0 and wolftail (Lycurusphleoides). Roost and ran- 10.3 1.9 Lycurusphleoides domly selected sites were not significanlty dif- Sporoboluswrightii Aristidaspp. ferent (x2 = 8.0, 2 x 5, P > 0.05). These winter Boutelouahirsuta roost siteshad far fewer speciescompared to day- Hilaria belangeri use sites and were dominated by the tall grasses Panicumobtusum 2.6 0 including side oats grama (Bouteloua curtipen- Boutelouachondrosioides 2.6 0 dada) and tanglehead (Heteropogon contortus). The latter tall grasswas rare on the reserve and limited to south-facing slopes where it formed feedingroost during this period. Apparently, once dense, almost monotypic stands (Bock and Bock a raptor can locate a covey, it is able to localize 1986). its hunting efforts. Ordination of roost, day-use sites, and ran- domly selected sites revealed a complex inter- CAPTURE WEIGHTS, SEX RATIOS, action between cover, slope of the hillsides, and BREEDING SEASON distance from oak trees. Both Kendall’s and Weights of the birds captured indicated that the Pearson’s correlations were high (0.87-0.93) be- habitat was adequate to support the bird’s body tween the first axis (reflecting many variables) weight at levels reported by Leopold and Mc- and distance to oak trees. Cabe (1957). Mean female weight was 192.7 g Given this suggestive pattern, a series of (n = 14, SE = 3.79). Mean male weight was 208.8 regressions between several original variables g (n = 16, SE = 4.46). Mean weight reported by measured in the field were calculated. Linear Leopold and McCabe (1957) was 176 g and 195 regression between variables sampled at ran- g for females and males, respectively. Of 32 birds domly selected sites revealed that cover in- captured for which gender data were available, creasedwith distance from oak trees (R* = 0.093, 17 (53%) were male. This slight bias towards F = 5.05, P = 0.029, n = 51). At randomly se- more males is similar to that reported by Leopold lected points, slope decreasedwith distance from and McCabe (1957) and Brown and Gutierrez oak but analysis of variance indicated that the (1980). Mean male capture weight was greater relationship is not quite significant (R2 = 0.74, than mean female capture weight (t = 2.4 1, two- F = 3.32, P = 0.074). A similar analysis of data tailed, P < 0.05). Capture weight of young of the based on measuresat day-use sites showedagain year was 152.1 g (n = 14, SE = 6.51). Young of that cover had a significant positive correlation the year were captured only in October and No- with distance from oak trees (Rz = 0.126, F = vember. My data on capture weight of adult fe- 16.3, P -C 0.001, n = 116), and slope decreased males and young of the year are consistent with with distance from oak trees (R2 = 0.122, F = those reported by Leopold and McCabe (1957). 15.5, P < 0.001). Cover increased as relocation I observed breeding behavior (calling males) as sites were increasingly level (R2 = 0.042, F = early as March or April, probably in responseto 5.03, P < 0.05). Biologically, these patterns fit unusually wet winters. Nesting occurred after together. rains in July and August that resulted in green At this study site, oaks are at the lower limit vegetation; similar observations are reported by of their distribution on the Huachuca foothills. Brown (1982). Emory oaks at the lower elevations of the study site are far more common on north-facing hill- HABITAT USE sides(Sanchini 198 1, Bock and Bock 1986) which Frequency of dominant plant speciesfrom day- were rather steep.Areas at the base of oaks were use sites, randomly selectedsites and roost sites often relatively barren but surrounded by tall and HABITAT USE BY MONTEZUMA QUAIL 233

TABLE 2. Comparison of the number of sites used Data presentedhere show the quail prefer north- by Montezuma Quail on hillsides or level areas.Roost facing slopes and thus by association, are more and random sites differed (x2 = 8.75, P < 0.01). Day- likely to be near oaks. On rare occasions,I ob- use and random sites differed (x2 = 6.02, P < 0.05). served Montezuma Quail at least 3 km from any

Level Hillside trees, well out in open grassland. During the winter months, night roosts were Roost 0 20 small cups, about 15 cm in diameter, often at Random 17 34 Day-use 19 97 the base of a rock with tall grassoverhanging the shelf. Soil in the covey’s night roostshad several successivelyolder layers of fecal material, sug- gestingrepeated use. Roost sites were highly ag- dense grassesdescribed by Brady et al. (1989). gregated, limited to a small area (seven to 10 Germination and/or survivorship of many forbs roost sites in 1 ha) dominated by Heteropogon and grassesmay be inhibited by a layer of oak contortus. Roosts were at least 100 m distant leavesand oak litter. Experiments with oak leaves from areas used by coveys during the day. Ori- and litter to determine if they alone are capable entation and structure of roost sites suggested of reducing other plant growth would be useful. that they may serve effectively in collecting and Sites used by quail during the day faced north storing solar radiation, as the shelf and rock face (mean aspect of 16.3”, R = 0.32, n = 93) and were exposedto solar radiation during the day. these sites differed in aspect from randomly se- At night, the dense grass overhanging the rock lected sites (Rayleigh’s test, F = 10.6; df = 1, offered protection from radiation heat losses. I 122, P < 0.001). Most day relocations of quail did not see quail approach night roost sites, but were within 20 m of the nearest oak tree on steep I suspectedthat quail arrived just at dark because areas (see below) where total vegetative cover they were relocated well away from roost sites was relatively low. At higher elevations on the when it was no longer possible to see clearly. Huachuca Mountains, where Montezuma Quail Many roost sites known from winter observa- also occur, oaks are not as restricted to northerly tions in two widely separatedaggregations were exposures,and the relationships presented here checkedin the dry seasonin 1988 and 1989. No may not occur. evidence of summer use was seen. The night Night roostsand sites used by quail during the roosts described here were those used in winter day differed with regard to aspectfrom randomly only. selectedsites. Roost sites faced southeast(mean Previous research observed the associations angle: 137.6”, R = 0.86, n = 20) and differed between these quail and dense ground cover (Rayleigh’s test, F = 13.84; df = 1,48, P < 0.00 1) (Leopold and McCabe 1957, Brown 1982). In from randomly selectedsites which faced north- this study, quail selectedsteeper slopes for roost- east (mean angle: 74.4”, R = 0.45, n = 3 1). Many ing than random sites (t = 2.29, P < 0.05, Table sites used by the quail were level and could not 3). During the day, quail were relocated in sites be assignedan aspectvalue (Table 2). Roost sites with lower (t = 2.23, P < 0.05) understory cover were limited to hillsides (Table 2) and differed than randomly selected sites. Steep slopes had from random sites (x2 = 8.75, P < 0.01). Day- oaks and supported less understory vegetation. use sites were on hillsides far more than would Day-use sites tended to be nearer oak trees than be expected from the distribution of level sites randomly selectedsites, but one-way analysis of at randomly chosenpoints (x2 = 6.02, P < 0.05). variance indicated that the differenceswere not

TABLE 3. Sample sizes(n), means (X), and standarderrors (SE) of distancesfrom oak trees,cover values, slope, and vegetation height measurementsat various sites used by Montezuma Quail.

Distancefrom oak (m) Cover (%) Slope Vegetationheight (cm) n II SE n R SE n + SE n + SE Roost 20 19.0 2.74 20 49.2 5.63 20 20.7 1.06 20 49.5 2.34 Random 51 20.9 2.24 51 54.3 3.49 51 13.5 1.89 51 42.3 2.37 Day-use 116 16.4 1.75 116 45.4 2.31 116 15.1 1.21 116 41.9 3.62 234 MARK R. STROMBERG

significant (Table 3). Slope of day-use sites was Observations of hourly movements of a covey not different than slopeofrandomly selectedsites over several days revealed that baited trap sta- (Table 3). Height of vegetation in roost, random, tions were rarely approachedby birds with trans- and day-use sites did not differ. This is probably mitters, even when the traps were removed. I because all sampling sites (roost, random, and would suggestthat tests of various baits with day-use) had been protected from grazing for radio-tagged birds may be productive in deve- many years and all sites supported taller grasses. loping a more efficient means of trapping birds Overall density and height of the vegetation on and thus facilitate future studies. the reserve was markedly higher than that on On a given day, coveys tended to stay in small surroundinggrazed land (Brady et al. 1989). Mean areas. A covey at Post Canyon was relocated at grasscover on the reserve, for example at roost 30-min intervals for 3 days. On 21 November, sites(49.2%) and day-use sites(45.4%) was much the mean distance this covey of eight moved higher in 1987 than the averageunderstory cover between observation intervals was 18.6 m (SE = of thesevegetation associationswhen grazing was 8.2, n = 10). On 28 November, the mean distance halted in 1968 (11.7% to 29%; Bonham 1972). moved between observations was 15.1 m (SE = No data were taken on comparableadjacent lands 3.5, n = 12). On 12 December, when six covey with quail coveys, but there the grasseswere members were known, the mean distance moved clearly clipped low as a result of intensive grazing was 63.8 m (SE = 46.4, IZ= 15). Aflockin Turkey by . Creek was similarly tracked in January 1987 with a mean of hourly moves of 20.1 m (SE = 3.6, n MOVEMENTS = 9). Daily movement patterns were revealed by fre- quent observations of birds with radio trans- DAILY AND SEASONAL SPACING PATTERNS mitters. These birds were extremely tolerant of Members of a covey were flushed between 0.5 close approach, often flushing only after being and 15 m from each other during the day. Re- approached to within 1 m. Many birds with peated observations documented small “use transmitters were approachedto within a square areas” which lasted from a few days to months meter, but could not be seen until they moved. (Table 4). A “use area” is a minimum convex On two occasions, birds remained motionless polygon estimate of spatial use within this time while allowing us to capture them as they period. Basically, a covey used the same small crouchedat the base of a clump of grass.Leopold area day after day. For most ofthe year, the covey and McCabe (1957) also observed the tendency would restrict their activity to a small area and of these quail to restrict their flight. From Jan- then (often overnight) a covey would shift to uary to March, three coveys were relocated on another, nonoverlapping area which became the 27 pairs of sequential days. Mean distance be- new use area where they restricted their daily tween relocations was 97.8 m (SE = 15.1). From movements. Longer observations revealed that 28 March to 31 May, bird #368 was relocated coveyshad several of theseuse areas.I combined with its covey as it increasedits wanderings. Dur- successiveuse areasinto a covey’s “home range.” ing that period, distance moved between days Separate coveys (that is, a covey without a bird almost doubled (X = 194.9 m, SE = 56.8, IZ = carrying a radio) were sometimes flushed near a 11). Only pairs were flushed during June and July covey under study. Although I have no evidence (Table 4). This period corresponds to the time of exclusive use of the smaller use areas by a of pair formation and nest construction (Leopold given covey, coveys seemed to generally stay and McCabe 1957). From July through October, apart. From November to January, the use areas the mean distance moved between successive were about l-5 ha (Table 4). During late winter days dropped to 83.8 m (SE = 9.84, n = 7). From and early spring, the use areaswould shift rapidly October through December, the mean distance and relatively far from each other, and thus home moved between days was 79.2 m (SE = 47.4, n rangeswere far larger (up to 50 ha). This spring = 5). Many additional sequential relocations home range expansion coincided with the time which spanned more than 24 hr agreedwith this when covey size reduced to pairs and reflectspair pattern of reduced movements. Multiple, non- spacing(Table 4). Data here on covey size agree overlapping use areas were not observed at these with the general range observed by Leopold and times; use areas defined home range of coveys McCabe (1957) from four to 10. from June to November (Table 4). I have no data to suggestthat coveysever form HABITAT USE BY MONTEZUMA QUAIL 235

TABLE 4. Summary of use and home range areasof oak woodland, and could not be expectedto sup- eight coveys of Montezuma Quail. Covey sizes listed port coveys. The most densely occupied habitat are the maximum counted in repeated flushes during had up to four coveys/259 ha. a given observation period. Covey one had three dif- ferent use areas,covey two had five different use areas, During census surveys of the reserve, coveys etc. The single use areas for coveys four to seven de- of Montezuma Quail were consistently relocated fined their home ranges. in the same small areas, not only in 1987, but from year to year. From 1983 to 1987, coveys HOlIE Relo- Use area Covey range were found in 16 specificsites (usually within the Period Days cations (ha) size (ha) same 50-m* area) with great regularity. Site fi- 7-14 Jan 8 4 1.04 3 6.56 delity between yearsand during a given year seem 27 Jan-l Feb 6 1.11 4 - especially characteristic of Montezuma Quail. 2-6 Feb : 7 2.46 4 - BecauseMontezuma Quail have such small use 20-27 Jan 8 5 0.47 2 12.8 areas, and have such high site fidelity, frequent 28 Jan-2 Feb 6 5 5.56 6 - and intense hunting pressure, particularly with - 2-9 Feb 8 6 0.79 3 trained bird dogs,can lead to virtual elimination 10-16 Feb 7 8 1.43 3 - 18-24 Feb 7 5 0.76 2 - of quail where hunter density is high, and thus should be considered as a conservation issue by 28 Mar-4 Apr 4 0.34 9 50.65 8 Apr-15 Apr : 3 0.17 5 - land managers. 17 Apr-6 May 20 3 1.69 3 - 9 May-14 May 6 4 1.36 2 - ACKNOWLEDGMENTS 15 May-31 May 16 7 6.67 2 - Fundingfor this study was provided by the National 7 Jun-12 Jun 6 5 1.65 2 1.65 Audubon Society, Joe and Helen Taylor, and the Re- 7 Jun-15 Ott 124 5 1.87 2 1.87 searchRanch Foundation. Bob Hazen shared the ser- 16 Sep-26 Sep 11 7 1.64 2 1.64 vices of his trained hunting dogs. Rukin Jelks III pro- 17 act-23 Ott 7 5 1.32 12 1.32 vided aircraft time and accompanied the author on 16 Nov-19 Nov 4 4 0.09 8 6.26 many successfulquail searches.Interns who assisted 20 Nov-26 Nov 7 14 0.70 8 - in the field researchinclude Brent Bibles, Chris Mono- 28 Nov-29 Nov 2 14 0.09 8 - vich, Duncan Halley, and David Manry. Suggestions 9 Dee-12 Dee 4 14 0.12 6 - and assistancefrom Carl Bock, Jane Bock, Barbara Stromberg, and Jane Church were most helpful. Paul Zimmerman provided radio transmitters, photo- graphs, and field assistance.Pat O’Brien and Terry Johnson (Arizona Game and Fish Department) pro- large aggregations. From June to October, pairs vided usefuladvice, field assistance,and state permits. or coveysremained sedentaryin small areas,often Reviews by Carl Bock, Clait Braun, Rocky Gutierrez, and Jerry Hupp clarified the manuscript. During this smaller than 2 ha (Table 4). Such small use areas study, the Arizona state office of the Bureau of Land typically had many (5-20/m2) small, scattered Management, particularlyDean Bibles, was most help- pits dug at the base of plants which produce tu- ful in establishingpermanent protection of the study bers (Zpomea sp.) and clumps of grass. These site from subdivision for home sites. characteristic diggingsare described in detail by Leopold and McCabe (1957). No obvious food LITERATURE CITED resourcewas particularly more abundant in the BAHRE, C. J. 1977. Land-use history of the Research observed use areas than in adjacent similar hab- Ranch, Elgin, Arizona. J. Ariz. Acad. Sci. itats. Winter diet was not analyzed in detail, but 12(suppl.):l-32. my observations of crop contents from 45 birds BEALS,E. W. 1984. Bray-Curtis ordination: an effec- tive strategyfor analysisof multivariate ecological harvested by hunters included larval insects, data. Adv. Ecol. Res. 14:1-55. vegetation, and acorns. These foods are similar BISHOP, R. A. 1964. The Meam’s Quail (Cvrtonyx to those reported by Bishop and Hungerford montezumae mearnsi) in southern Arizona. (1965) except that amplijhlia, an impor- M.Sc.thesis. Univ. Arizona, Tucson. BISHOP, R. A., AND C. R. HUNGERFORD.1965. Sea- tant plant to quail in grazed sites, was virtually sonal food selection of Arizona Meams Ouail. J. absent on this study site. Wildl. Manage. 29:813-819. Overall density of quail can be estimated by BOCK,J. H., AND C. E. BOCK. 1986. Habitat rela- using the number of quail/covey (Table 4) in the tionships of some native perennial grasses in southeasternArizona. Desert Plants 8:3-14. most intensively surveyed 2,331 ha on the re- BOCK,C. E., J. H. BOCK,W. R. KENNEY,AND V. M. serve where 16 coveys were consistently relo- HAWTHORNE.1984. Responsesof birds, rodents cated from 1983-1987. Not all of this area was and vegetation to livestock exclosure in a semi- 236 MARK R. STROMBERG

desert grasslandsite. J. Range Manage. 37:239- lemetry, p. 679-698. In A. Y. Cooperrider led.], 242. Inventory and monitoring of wildlife habitat..U.S. BONHAM,C. D. 1972. Ecologicalinventory and data Den. Inter.. Bur. Land Manaae.. Washinnton. DC. storage-retrievalsystem for the Research Ranch. JE-;H, R. I.; AND F. B. TURNER.’ 1969. Measure- Range Sci. Dep., Colorado State Univ., Sci. Series, ment of non-circular home range. J. Theor. Biol. 14. Fort Collins. 22~227-237. BRADY,W. W., M. R. STROMBERG,E. F. ALDON,C. D. LEOPOLD,A. S., AND R. A. MCCABE. 1957. Natural BONHAM, AND S. H. HENRY. 1989. Responseof history of the Montezuma Quail in Mexico. Con- a semidesert grasslandto 16 years of rest from dor 59:3-26. grazing. J. Range Manage. 42:284-288. LIGON,J. S. 1927. Wildlife of , its con- BRAUN,C. E. 1976. Methods for locating, trapping servation and management. New Mexico State and banding Band-tailed Pigeons in Colorado. Game Commission. Santa Fe. Colorado Div. Wildl., Spec. Rep. 39. Denver. MCCUNE,B. 1987. Multivariate analysison the PC- BRENNAN,L. A., W. M. BLOCK,AND R. J. GUT&RRE~. ORD system. Holcomb Res. Inst., HRI Rep. 75. 1987. Habitat useby Mountain Quail in northern Indianapolis, IN. California. Condor 89:66-74. MILLER,L. 1943. Notes on the Meams Quail. Condor BROWN,R. L. 1982. Effects of livestock grazing on 45:104-109. Meams Quail in southeasternArizona. J. Range REEVES,H. M., A. D. GEIS, AND F. C. KNIFL~N. 1968. Manage. 35727-732. Mourning dove capture and banding. U.S. Dep. BROWN,D. E., AMYR. J. GUTIBRREZ.1980. Sexratios, Int., Fish and Wild. Ser., Spec. Sci. Rep. Wildl. sexualselection, and sexualdimorphism in . 117. J. Wildl. Manage. 44: 198-202. SANCHINI,P. J. 1981. Population structure and fe- COCHRAN,W. W. 1980. Wildlife telemetry, p. 507- cundity patterns in Quercus emoryii and Q. ari- 520. In S. D. Schemnitz [ed.], Wildlife manage- zonica in southeasternArizona. Ph.D. diss. Univ. ment techniques manual. The Wildlife Society, Colorado, Boulder. Washington, DC. SIEGEL,S. 1956. Nonparametric statisticsfor the be- DAUBENMIRE,R. 1959. A canopy-coveragemethod havioral sciences.McGraw-Hill, New York. of vegetationalanalysis. Northwest Sci. 33:43-64. WALLMO,0. C. 1954. Nesting of Meams Quail in HASTINGS,J. R., AND R. M. TURNER. 1965. The southeasternArizona. Condor 56: 125-128. changingmile. Univ. Arizona Press, Tucson. ZAR, J. H. 1984. Biostatisticalanalysis. Prentice-Hall, HEGDAL,P. L., AND B. A. COLVIN. 1986. Radiote- Englewood Cliffs, NJ.