Factors Affecting Nest Site Location in Gila Woodpeckers

Home , Ambrosia deltoidea, Gila woodpecker

The Condor 86:73-78 0 The CooperOrnithological Society 1984

FACTORS AFFECTING NEST SITE LOCATION IN GILA WOODPECKERS

JEROME J. KOROL

AND RICHARD L. HUTTO

ABSTRACT. - We investigated the locations of Gila Woodpecker (Melanerpes uropygialis) nest sites in Organ Pipe Cactus National Monument, Arizona. A significantly greater proportion of saguarocacti in arroyos had woodpecker holes than did those on hillsides, ridgetops or desert flats. Saguarosused for nest-hole excavation were taller and more likely to have branches than randomly chosen saguaros.Both active and inactive nest holes were nonrandomly oriented, with the mean vectors of orientation being in northwesterly directions. Variation in nest-hole orientation probably is not due to seasonaldifferences in excavations becausethe orientation of springactive (late March) nest holeswas not significantly different from that of inactive holes. Some variation in nest-hole orientation is attributable to the structure of the saguaros.

Desert birds are often thought to be affected gions that receive summer and winter rains primarily by thermoregulatory constraints in (Shreve and Wiggins 1964). A Larrea triden- placing their nests (Ricklefs and Hainsworth tata-Ambrosia deltoidea associationdominat- 1969, Austin 1974). Recently, Inouye et al. ed the flats; columnar cacti such as Cereus (198 1) considered the nonrandom orientation thurberi, Ferocactuscovillei, and Opuntia big- of Gila Woodpecker (Melanerpes uropygialis) loveii predominated on the steeprocky slopes; nest holes in saguarocacti (Cereus gigunteus) and Cercidium microphyllum, Acacia con- to represent such an adaptation. In addition, stricta and Olneya tesota dominated the dry they suggestedthat variability about the mean creek beds (arroyos). Saguaroswere present in nest-hole orientation might reflect a thermal each of these physiographic areas. need for woodpeckers to use holes that are We assumedthat Gila Woodpeckers had ex- oriented differently at different times of the cavated most of the holes in the saguaroswith- year or season. in our study area. This assumptionfollows from Old woodpecker excavations constitute a three observations: 1) in this part of the mon- volume of valuable historical information that ument, the Northern Flicker (Colaptes aura- is unique and rarely available to animal ecol- tus) is also known to excavate holes in sagua- ogists. In this paper we use such information ros, but the holes of the two speciescan be to test the hypothesesthat Gila Woodpeckers distinguished by size (Gilman 19 15) and the use a random subset of saguarosfor nesting larger flicker holes were rare; 2) data from five (or roosting) purposes, and that the position 1-km line transectsconducted in this location and orientation of holes within those saguaros in March-April 1978 (unpubl. data) showed are random. Only the orientation of nest holes flickers to be six times rarer than Gila Wood- has received attention before this study (In- peckers; and 3) only one of 20 active wood- ouye et al. 198 1). We then compare the mean pecker nests that we found during our study orientation of inactive holes with that of holes belongedto a Northern Flicker. We also realize which were being actively used during early that holes may be excavated for roosting as spring, to test the hypothesis that the mean well as for nesting,but for simplicity and con- orientation of active holes shifts seasonally. formity with previously published work we re- Finally, we re-evaluate the possibility that fer to all holes as “nest holes.” thermal constraints alone can adequately ac- In order to characterize woodpecker nest count for the nest site locations of this wood- sites, we conducted transects in five different pecker species. physiographic areas: east-facing slopes, west- facing slopes, ridges, flats, and arroyos. All STUDY AREA AND METHODS transectswere 40 m wide and the length and We collected nest site information in Organ number of transectsin each area equalled the Pipe Cactus National Monument, Arizona distanceneeded to sample 100 saguaros.Each (3 1”57’N, 112”48’W) from 19-29 March 198 1. of the west- and east-facingtransects began at Vegetation was typical of Sonoran Desert re- the base of a slope and ended no closer than

[731 74 JEROME J. KOROL AND RICHARD L. HUTTO

TABLE 1. Location of Gila Woodpecker nest holes rel- N= 70 ative to physiographicarea. SAGUAROS WITH HOLES SAGUAROS WITHOUT HOLES A0 Location & z East- west- f$;; facing 5 30 slope Ridgetop Flats Amyo B No. saguaros kF 20 with holes 3 4 6 6 30 No. saguaros without holes 97 96 94 94 70

4-6 6-6 a-10 IO-12 .I2 HEIGHT (M.) 30 m from a ridge. Ridge transects followed FIGURE 1. Frequencyhistograms representing the height the ridges, and arroyo transects followed ar- distribution of saguarosthat did or did not contain Gila royo beds and included saguaroswithin 20 m Woodpecker holes. of the nearest edge. On the flats, transectswere run basically north to south and were discon- tinued within 30 m of an arroyo. A rangefinder The saguarosthat were used for nestingpur- was used to determine transect widths. poses were significantly taller than would be To measure variables that might be impor- expected if the woodpeckers were using them tant in determining nest hole location and ori- in proportion to their availability; we found entation, we recorded the following informa- no saguaro shorter than about 4 m with a tion for each cactus encountered: height of woodpecker nest hole, and even restricting cactus; number, orientation, and height of analysis to saguarosthat were greater than or branches; number, orientation, and height of equal to 4 m in height, the taller saguaroswere holes; and direction to nearest arroyo. Nest used disproportionately more often than the hole orientation and direction to nearest point shorter saguaros(G = 82.3, P < .OOl; Fig. 1). in an arroyo were measuredby compassto the The mean height of saguaroswith holes was nearest 5”. Data were recorded in the same significantly (t = 9.4, P < .Ol) greater than manner for 20 saguarosoutside the transect those without holes (8.6 vs. 6.4 m, respective- area that contained active Gila Woodpecker ly). Inouye et al. (198 1) reported the mean nests (holes at which woodpeckers were seen height of saguaroswith nest holes to be 9.6 m. enteringand leavingperiodically). No data were Among saguarosthat were used,the number recorded for saguarosless than 4 m tall. of holes present in a given saguarowas signif- The statistical significanceof nest-hole and icantly positively correlated (Y = .33, P < .OOl) branch orientation was determined through with saguaroheight and increasedwith saguaro methods described by Batschelet (1965). This height at a rate that was slightly less than ex- technique allows one to test whether the ori- pected based on either saguarosurface area or entation of points is nonrandom, and provides volume (Fig. 2). Thus, we detected no pref- a relative index (r) that reflects the tightness erence for a particular height class since no of dispersion around a mean vector of orien- height class contained saguarosthat had dis- tation, with a value of 0.0 indicating a high proportionately more or fewer holes than pre- degreeof dispersionand 1.O indicating that all dicted from the regressionline. points are oriented in the same direction. Of saguarostaller than 4 m, the presenceof nest holes was significantly (G = 63.9, P < RESULTS .OOl) positively associatedwith the presence SAGUARO SELECTION of branches (Table 2). This association also The presence of woodpecker nest holes was held true when the five physiographic areas dependent upon physiographic area (Table 1). were considered separately. The conditional Significantly more saguaros in arroyos had woodpecker holes than did those in any of the other areas (G = 47.1, P < .OOl). The proportion of saguaroswith holes did not differ TABLE 2. The distribution of saguarouse relative to the presenceor absenceof branches. significantly among the four non-arroyo areas (G = 1.55, P > .05). The dependencebetween Nest holes presenceof holes and physiographic area per- Present Absent sists when the effects of saguaro height and presence of branches are kept constant (G = Branchespresent 46 Branchesabsent 24 33.1, P < .Ol). GILA WOODPECKER NEST SITES 75

I I I I I I 4 6 8 IO 12 SAGUARO HEIGHT (M.) FIGURE 2. The number of Gila Woodpecker holes present in a saguaroin relation to saguaroheight. No height range contains saguaroswith disproportionately more holes than expected from the regression. probability that a cactuswould have branches ling 38” north of west (z = 5.26, Y = .49, P -c if it contained one or more nest holes was 66%, .O1). The inactive nestswere also significantly while the probability that a nest hole would be nonuniformly oriented (Fig. 4) with the mean present given the saguaroshad branches was vector of orientation equalling 17” north of much less (36%). west (z = 7.75, r = .21, P < .Ol). Moreover, We tested the hole, height, and branch vari- inactive nestswere not significantly differently ables for conditional independence using a oriented from the active nests (z = 5.26, P < 3-way analysisof association(Sokal and Rohlf .Ol). Although the distribution of active and 198 1: 750) and found that the presence of inactive nestswas significantly nonuniform, it holes depended upon height, given the presence or absence of branches (G = 37.8, P < .OOl). Furthermore, the presence of holes depended upon the presence of branches, given the height of the saguaro (G = 19.4, P <

.oo1). 0 NEST SITE CHARACTERISTICS WITHIN SAGUAROS We determined the nest site characteristicsof 20 active and 17 1 inactive woodpecker nest holes throughout the five physiographicareas. The relative height of nest holes within sa- guaroswas inversely related to saguaroheight such that there existed a maximum absolute 3J I height of about 7 m, beyond which wood- 4.0-5.9 6.0-7.9 8.0-9.9 10.0-11.9 t12.0 pecker excavations were rarely found (Fig. 3). CACTUS HEIGHT (M.) The active nest holes were significantly non- FIGURE 3. The mean absoluteheight (t 1 standardde- uniformly oriented around the saguaros(Fig. viation) of Gila Woodpecker nest holes in relation to sa- 4), with the mean vector of orientation equal- guaro cactusheight. 76 JEROME J. KOROL AND RICHARD L. HUTTO

0 = ACTIVE 0 L INACTIVE

FIGURE 4. The compassorientation of active and inactive Gila Woodpecker nest-hole entrances.Arrows indicate mean directions of orientation and arrow lengths are inversely proportional to the amount of dispersion about the mean (r).

is important that considerable dispersion ex- REFERENCE POINT isted (indicated by the relatively low r values) and that the nonrandom orientation was large- ly a statistical phenomenon; many holes were oriented in directions other than northwest. Over one-fourth of the saguaroswith nest holes also bore one or more “adjacent branches”-branches whose base was lower and tip higher than the height of the nest hole. The orientation of 17 active and inactive nest holes relative to the direction of pairs of adjacent branches was nonrandom (z = 16.2, Y = .98, P < .Ol); nest holes tended to bisect the acute angle formed by the directions of orientation of the two branches (Fig. 5). These 17 nest holes were themselvesrandomly oriented with respectto compassdirection (z = 0.12, r = .08, P > .05). Nest holes were randomly oriented with respectto the direction to the closestarroyo (z = 0.20, r = .ll, P > .05). DISCUSSION FIGURE 5. The orientation of Gila Woodpecker nest The disproportionately high use of saguarosin holes near a pair of branches. Reference point is the di- arroyos is not necessarilybecause the thermal rection that bisectsthe acute angle formed by orientation environment is most favorable there. An of the two branches. GILA WOODPECKER NEST SITES 77 equally plausible alternative is that the wood- ouye et al. 198 l), suggestingthat variability in peckers forage primarily in the arroyos. The nest hole orientation might be accounted for latter interpretation would be consistent with by seasonal differences in the orientation of Bent’s (1939) statement that the Gila Wood- newly excavated nest holes. Indeed, the ori- pecker feeds on food items that are either re- entation of nest entrances of Cactus Wrens strictedto, or much more common in, the wash (Campylorhynchus brunneicapillus; Ricklefs environment, such as insects in dead wood, and Hainsworth 1969, Austin 1974) and House fruits, berries, and bird eggs. Sparrows (Passer domesticus; McGillivray The reasons why taller saguarosare used 198 1) has been shown to differ significantly disproportionately more often than shorter with time of season.Inouye et al. (198 1) con- ones are not entirely clear, but we .suspectthat cluded that the variation in nest hole orien- there are problems associatedwith smaller sa- tation of Gila Woodpeckers may also result guarosthat may not exist with larger ones. For from seasonaldifferences in the optimal ori- example, excavating in a small structure may entation of nest and/or roost holes. If this were be difficult, predation rates may increasecloser true, we would expect the orientation of early to the ground, or thermal stressmay become (mid-March) neststo be clusteredfairly tightly a significant factor closer to the ground. Al- around a given direction, and we would also though the taller saguarosare used most often, expect that the direction would differ signifi- there also appearsto be an upper limit (Fig. 3) cantly from the mean direction of orientation to the height of nest excavations. Since struc- recorded for the entire group of inactive holes. tural change in saguarosis more likely to be This was not the case.The 20 active holes were associatedwith relative, rather than absolute, scatteredconsiderably in their orientation (r = height, the upper limit to hole height may be .49) and the mean direction of orientation did set by the energetic constraints of travel, just not differ significantlyfrom the mean direction as the upper limit in vertical flight of North- of all inactive nests.The variability in nest hole western Crows (Corvus caurinus) is likewise orientation can apparently not be explained on constrained (Zach 1979). the basis of seasonal shifts in the mean ori- The association between the presence of entation of nest holes that are used. branches in the saguarosis apparently due to One important variable contributing to the the woodpeckers’ preference for saguaroswith orientation of about 10% of the nest holes in branches. This follows from our finding that this study was the positioning of branchesrel- we were more likely to see branches, if a sa- ative to nest holes; holes were always exca- guaro had nest holes than we were to seea nest vated within the acute angle formed by adja- hole, if a saguarohad branches.The alternative cent branches when they were present. explanation for the associationwould be that In conclusion, the nest sites used by Gila the saguarostend to branch in responseto the Woodpeckers appear to be determined by sev- presenceof nest holes.The latter interpretation eral factors. Foraging constraintsprobably dic- would also be inconsistentwith the findings of tate which saguarosare used, while foraging, Yeaton et al. (1980), who concluded that predatory, thermoregulatory, and architectur- branching in saguarosis related to moisture al constraintsprobably all limit the height and availability. orientation of nest holes. We feel it unlikely The orientation of woodpecker nest holes that any single factor can adequately account has received considerable attention (see for the variability in location of Gila Wood- McClelland [ 19771 for a brief review). Nest pecker nest sites. hole orientations differ considerably among species and vary geographically within a ACKNOWLEDGMENTS species.The Gila Woodpecker appears to be We thank the personnel of Organ Pipe Cactus National unique, however, becauseit tends to excavate Monument for permission to conduct the study there and nest holes in a northwesterly direction. One for their many forms of assistance.We also thank the factor that has been important in accounting Zoology Department of the University of Montana for for nest hole orientation in a variety of wood- financial supportassociated with the 1981 Field Ecology pecker specieshas been the tendency for wood- course,and Joe McAuliffe for comments on a draft of the manuscript. peckers to excavate on the downward side of leaning trunks and branches (Conner 1975). We found so few nest holes in nonvertical por- LITERATURE CITED tions of saguarosthat this factor must be un- AUSTIN,G. T. 1974. Nesting successof the CactusWren important in determining the direction of nest in relation to nest orientation. Condor 7 1:216-2 17. BATSCHELET,E. 1965. Statisticalmethods for the analysis hole orientation in Gila Woodpeckers. of problems in animal orientation and certain bio- The nonrandom orientation of nest holes is logical rhythms. American Institute of Biological Sci- generally attributed to thermal constraints(In- ences,Washington, DC. 18 JEROME J. KOROL AND RICHARD L. HUTTO

BENT, A. C. 1939. Life histories of North American SHREVE, F., AND I. L. WIGGINS. 1964. Vegetation and woodpeckers.U.S. Natl. Mus. Bull. 174. flora of the Sonoran Desert. Stanford Univ. Press, CONNER, R. N. 1975. Orientation of entrancesto wood- Stanford, CA. pecker nest cavities. Auk 92:371-374. SOKAL,R. R., AND F. J. ROHLF. 1981. Biometry. W. H. GILMAN, M. F. 1915. Woodpeckersof the Arizona low- Freeman and Co., San Francisco. lands. Condor 17:151-163. YEATON, R. I., R. KARBAN, AND H. B. WAGNER. 1980. INOUYE, R. S., N. J. HUNTLY,AND D. W. INOUYE. 1981. Morphological growth patternsof Saguaro(Carnegiea Nonrandom orientation of Gila Woodpecker nest en- giguntea:Cactaceae) on flats and slopesin Organ Pipe trancesin saguarocacti. Condor 83:88-89. CactusNational Monument, Arizona. Southwest.Nat. MCCLELLAND.B. R. 1977. Relationshinsbetween hole- 25:339-349. nestingbirds, forest snags,and decay-inwestern larch- ZACH, R. 1979. Shell dropping: decision making and Douglas-fir forestsof the northern Rocky Mountains. optimal foraging in Northwestern Crows. Behaviour Ph.D. diss., Univ. of Montana, Missoula. 63:106-l 17. MCGILLIVRAY,W. B. 1981. Climatic influenceson pro- ductivity in the House Sparrow. Wilson Bull. 93: 196- Departmentof Zoology, Universityof Montana, Missoula. 206. Montana 59812. Received 22 February 1983. Final ac- RICKLEFS,R. E., AND F. R. HAINSWORTH.1969. Tem- ceptance30 June 1983. peratureregulation in nestlingCactus Wrens: the nest environment. Condor 7 1:32-37.

The Condor 86:78 0 The CooperOrnithological Society 1984

RECENT PUBLICATIONS

Breeding Biology of the Adblie Penguin.-David G. Ain- The Arctic Skua/A study of the ecologyand evolution of ley, Robert E. LeResche,and William J. L. Sladen. 1983. a seahird.-Peter O’Donald. 1983. Cambridge University University of California Press, Berkeley. 244 p. $27.50. Press,Cambridge. 324 p. $49.50. Arctic Skuas(known to Seabirdsare longer-lived than most landbirds, and must us as Parasitic Jaegers,Stercorurius parasiticus) occur in therefore be studied for longer periods in order to under- three main phenotypesthat vary in frequencyfrom mostly stand their life history. Here is the outcome of fifteen dark in the southern colonies to almost entirely pale in seasonsof fieldwork in the penguinrookeries at Cape Cro- the north. The ecologicaland genetic factors maintaining zier, Ross Island, Antarctica. Individually marked, known- this stable, clinal polymorphism are the chief subject of agebirds were studiedthroughout, and many of them were this book. It is largelybased on original, long-term research observed during several seasonsafter being banded. In- conductedat the Fair Isle Bird Observatory, between the tegratingthese findings with those from other studies,the Shetlandand Orkney islands.The stageis set with chapters report describesbreeding activities in detail and analyzes on the numbers and distribution of birds, their feeding how age, experience,and other factors influence breeding behavior, and breeding biology. The following chapters success.It then examines the size and age structureof the treat genetics, demography, and sexual selection, incor- Crozier population, and finally, the variablesthat regulate porating general theory where needed to understand the it. The book is crammed with data (drawn from impres- specificanalysis. Despite this material, the argument is at sively large samples), yet it is well written, with chapter an advanced level and, to be graspedfully, calls for some summariesand other passagesto guidethe reader. It makes prior knowledge of population genetics. Several appen- a major contribution to our knowledgeof the Ad&lie Pen- dicesgo further into the mathematicaland statisticalmeth- guin and sheds light on the life history patterns of other ods for analyzing bird populations and mating patterns. large seabirds. Between the lines, it shows the kinds of Informative as it is about jaegers, the book will probably biological questions that can explored only by long-sus- be of wider interest to those who are investigating poly- tained investigations.Illustrations, appendices,references, morphism, selection, or mating in other birds. Illustra- index. tions, references,index.