--For several decades general opinion has suggested that southern flying squirrels (Glaucomys volans) have a negative effect on Red-cockaded Woodpeckers (Picoides borealis) through competition for cavities and egg/nestling predation. Complete removal of hardwood trees from Red-cockaded Woodpecker cavity tree clusters has occurred on some forests because southern flying squirrel abundance was presumed to be associated with the presence and abundance of hardwood vegetation. In some locations, southern flying squirrels have been captured and either moved or killed in the name of Red-cockaded Woodpecker management. We determined southern flying squirrel occupancy of Red-cockaded Woodpecker cavities in loblolly (Pinus taeda)-shortleaf (P. echinata) pine habitat (with and without hardwood midstory vegetation) and longleaf pine (P. palustris) habitat (nearly devoid of hardwood vegetation) during spring, late summer, and winter during 1990 and 1991. Flying squirrel use of Red-cockaded Woodpecker cavities was variable and was not related to presence or abundance of hardwood vegetation. Woodpecker nest productivity was not correlated with flying squirrel use of woodpecker cavities within clusters. In addition, we observed six instances where Red-cockaded Woodpeckers successfully nested while flying squirrels occupied other cavities in the same tree. Our results suggest that complete removal of hardwoods from woodpecker cluster areas in loblolly and shortleaf pine habitat may not provide benefits to the woodpeckers through reduction of flying squirrel numbers. Reduction of hardwood midstory around cavity trees, however, is still essential because of the woodpecker's apparent innate intolerance of hardwood midstory foliage. Received 3 Nov. 1995, accepted 21 Mar. 1996.

The Red-cockaded Woodpecker (Picoides borealis) is a cooperative breeder that lives in family groups composed of a breeding pair and fre- quently one to several helpers (Ligon 1970, Walters et al. 1988, Walters 1990). The activities of the group center around a cluster of cavity trees composed of living pines that contain one to several cavities and cavity starts. Cavities are excavated into the heartwood of pines that typically are infected with red heart fungus (Phellinus pini) (Conner and Locke 1982, Hooper 1988, Hooper et al. 1991, Rudolph et al. 1995). Cavity excavation in Texas requires an average of 1.8 y in loblolly pines (Pinus taeda), 2.4 y in shortleaf pines (P. echinata), and 6.3 y in longleaf pines (P. palustris) (Conner and Rudolph 1995). Pines selected for cavities in Texas usually exceeded 90 years of age (Conner and O'Halloran 1987, Wildlife Habitat and Silviculture Laboratory (Maintained in cooperation with the College of Forestry, Stephen F. Austin State University), Southern Research Station, U.S.D.A. Forest Service, Nacogdoches, Texas 75962. 697 698 THE WILSON BULLETIN ß Vol. 108, No. 4, December 1996 Rudolph and Conner 1991). The Red-cockaded Woodpecker is a keystone species of the fire-climax, southern pine ecosystems in that they are the primary species to excavate cavities in an otherwise cavity-barren envi- ronment relative to hardwood ecosystems (Conner and Rudolph 1995). Thus, the cavities that take Red-cockaded Woodpeckers a long time to create tend to be in relatively high demand by other cavity-using species (Dennis 1971, Rudolph et al. 1990b, Loeb 1993). As cavities near completion, Red-cockaded Woodpeckers peck shallow excavations, termed resin wells, around their cavity entrances. Continued pecking at resin wells causes a copious flow of resin down the bole of the pine (Ligon 1970). Woodpeckers also scale loose bark from the bole of the cavity tree and nearby pines. Although bark scaling and resin flow usually deters climbing by rat snakes (Elaphe obsoleta) (Jackson 1974, Rudolph et al. 1990a), the resin barrier does not deter southern flying squirrels (Glaucomys volans) which are frequent users of cavities with unenlarged entrances that are also preferred by Red-cockaded Woodpeck- ers (Rudolph et al. 1990b, Loeb 1993). Past studies have indicated a negative association between Red-cock- aded Woodpeckers and the density of hardwood midsttry and understory (Hopkins and Lynn 1971, Van Balen and Doerr 1978, Hovis and Labisky 1985, Conner and Rudolph 1989, Loeb et al. 1992). This has lead to widespread management programs that remove all hardwood vegetation from Red-cockaded Woodpecker cavity tree cluster areas (Conner and Rudolph 1991b). Although the negative effect of hardwood vegetation on Red-cockaded Woodpeckers is well documented, the mechanism that causes this negative relationship is poorly understood. One proposed mechanism for the hardwood effect is that southern flying squirrels, a potential competitor for Red-cockaded Woodpecker cavities, are depen- dent on hardwood midsttry foliage. However, flying squirrels appear to prefer hardwood vegetation primarily as understory cover and as a food source (Bendel and Gates 1987). Contrary to popular belief, southern flying squirrels may avoid areas with dense midsttry foliage because it interferes with flight paths between boles of larger pines (Bendel and Gates 1987). The influence of plant species composition and midsttry and understory foliage densities in pine forests on southern flying squirrel abundance is not fully understood. To date, no published studies have demonstrated that southern flying squirrels have a negative effect on Red- cockaded Woodpecker populations, yet management programs that in- clude removal of southern flying squirrels from cavities and euthanasia (Gaines et al. 1995) are becoming more widespread. Several species of woodpeckers enlarge Red-cockaded Woodpecker cavity entrance tunnels by excavation and use the cavities (Conner et al. Conner et al. ß WOODPECKERS AND FLYING SQUIRRELS 699 1991, Neat et at. 1992). Some of these species, e.g., Pileated (Dryocopus pileatus) and Red-bellied (Melanerpes carolinus) woodpeckers, are con- sidered to be primarily associated with hardwood forests (Reller 1972, Conner et at. 1975). Metal plates that restrict the entrance diameter of Red-cockaded Woodpecker cavities (Carter et al. 1989) have been devel- oped for placement over enlarged cavities in hopes that some currently unsuitable cavities can be rehabilitated and on unentarged cavities to pre- vent enlargement. Although these plates may prevent further damage by larger species of woodpeckers, they will not deter the use of cavities by southern flying squirrels or other small species of woodpeckers which prefer smaller entrances. The competitive impact of southern flying squirrels on Red-cockaded Woodpeckers is largely hypothetical. If a detrimental impact is occurring, it may be exacerbated in small declining Red-cockaded Woodpecker pop- ulations such as those in eastern Texas that are also stressed by other factors such as isolation and forest fragmentation (Conner and Rudolph 1989, 1991a). Our objectives were to (1) determine the availability and use of Red- cockaded Woodpecker cavities during the nesting, late-summer, and win- ter seasons, (2) evaluate southern flying squirrel use of cavities in relation to species composition and structure of vegetation, and (3) explore pos- sible negative effects of southern flying squirrels on Red-cockaded Wood- pecker breeding success. STUDY AREAS AND METHODS The study was conducted on the Angelina (62,423 ha; 31ø15'N, 94ø15'W) and Davy Crockett (65,329 ha; 31ø21'N, 95ø07'W) National Forests from March 1990 to October 1991. We examined 11 Red-cockaded Woodpecker cavity clusters in open longleaf pine habitat, 10 clusters in loblolly-shortleaf pine habitat with all hardwood vegetation removed in the cluster area, and seven clusters in loblolly-shortleaf pine habitat with extensive hardwood vegetation present during 1990. We suspected that different seasons of the year may impose varying levels of competition for cavities. The breeding season (spring) is likely to be a season of potentially elevated competition, and competition at that time can decrease breed- ing success. The late sununer season may also be a critical period because new young have fledged and are searching for cavities for nocturnal roost sites. We sampled cavity occupants during winter to examine the possibility that thermal stress during the colder months may lead to increased demand for cavities. We climbed approximately 230 cavity trees using Swedish climbing ladders and examined them for occupancy during spring (April to May) of 1990 and 1991, late summer (August to October) of 1990 and 1991, and winter of 1990-1991 (December 1990 to February 1991). Only a few cavity trees were not climbed because of safety factors during each climbing season. Such trees were typically small-diameter, old, inactive cavity trees that were pri- marily hollow shells and whose cavities were of no use to Red-cockaded Woodpeckers for cavities. We lowered a small, high intensity light into each cavity chamber, examined con- tents with an oval mechanics mirror mounted on an extendable handle, and identified and 700 THE WILSON BULLETIN ß Vol. 108, No. 4, December 1996 counted cavity occupants. Often more than one southern flying squirrel was present in a cavity. When this occurred, a coat hanger wire (with the end bent around to prevent injury) was placed into the cavity and flying squirrels lifted out for counting. We used presence of chewed pine needles and fresh flying squirrel feces as an indicator of flying squirrel use. Unchewed pine needles in an enlarged cavity indicated use by fox squirrels (Sciurus niger). Cavity trees that were being simultaneously used by both southern flying squirrels and Red- cockaded Woodpeckers (in two different cavities) were examined closely during the woodq pecker nesting season to determine woodpecker fledging success and during other seasons to detect cavity usurpation by flying squirrels. We measured the entrance diameters of cav- ities and monitored cavities with restrictors in each cluster studied. Based on previous studies (Rudolph et al. 1990b), cavities were divided into those suitable for Red-cockaded Wood- pecker use (entrance diameters <7 cm in diameter) and those too enlarged to be acceptable to Red-cockaded Woodpeckers (entrances >7 cm in diameter). We measured reproductive success of Red-cockaded Woodpeckers in each cluster by determining the number of young fledged from nest cavities. Young were counted at 8, 20, and 23 days of age in each nest tree. Clusters were visited within a week of fledging to determine how many of the nestlings observed on day 23 successfully fledged. We also visited each cluster during August and September to determine number of surviving young. We made dawn and dusk visits to each woodpecker group to verify roost locations, band woodpeckers, and determine number of members in each family group during each climbing season. We also determined the number of Red-cockaded Woodpeckers roosting outside of cavities in the open. Vegetation measured in each cavity tree cluster (six points per cluster) included basal area of overstory and midstory pines and hardwoods using a one-factor metric prism, height of midstory and understory vegetation using a clinometer, canopy closure using a 4-cm di- ameter by 12 cm hollow tube, and foliage density of vegetation from the ground to 1 m, and 1 m to 2 m using a density board as described by MacArthur and MacArthur (1961). By spring 1991, the seven clusters which had a well-developed hardwood midstory during the first year of the study had received midstory treatment. All hardwood vegetation was removed from these seven clusters during the 1990-1991 winter giving them the same structural appearance as the 10 clusters in loblolly-shortleaf habitat that were initially with- out hardwoods. For each cavity tree cluster during each season we calculated the percentage of unenlarged cavities occupied by southern flying squirrels and those occupied by Red-cockaded Wood- peckers. Analysis of variance and Duncan's multiple range test were used to test for differ- ences in flying squirrel and Red-cockaded Woodpecker use of cavities among habitat treat- ments during each season (P = 0.05). We related fledging success with the proportion of unenlarged cavities (all unenlarged and available unenlarged cavities) occupied by flying squirrels with Spearman correlations (rs, P = 0.05). RESULTS Vegetation characteristics of cavity tree clusters.--Vegetation within the three treatments differed distinctly during the first year of the study (1990). Red-cockaded Woodpecker cavity tree clusters in longleaf pine habitat were nearly devoid of any hardwood vegetation or understory and midstory foliage except for grasses and forbs (Table 1), and the absence of hardwoods extended well beyond the boundaries of cluster areas. This was not the case in clusters located in loblolly-shortleaf pine habitat. Conner et al. ø WOODPECKERS AND FLYING SQUIRRELS 701 TABLE 1 VEGETATIVE CHARACTERISTICS (MEAN SD) OF RED-COCKADED WOODPECKER CLUSTER AREAS WHERE CAVITY OCCUPANTS WERE MONITORED IN LONGLEAF FINE, LOBLOLLY- SHORTLEAF FINE WITH NO MIDSTORY, LOBLOLLY-SHORTLEAF FINE WITH MIDSTORY PRESENT (PRE-TREATMENT 1990) AND MIDSTORY REMOVED (PoST-TREATMENT 1991) ON THE ANGELINA AND DAVY CROCKETt NATIONAL FORESTS IN EASTERN TEXAS Habitat variable Loblolly- Loblolly- Loblolly- shortleaf shortleaf shortleaf Longleaf no with midstory pine midstory midstory removed (N = 11) (N = 10) (N = 7) (N = 7) Overstory pine basal area a (m2/ ha) 14.0 (6.6) b 12.7 (3.5) b 10.9 (4.1) c 13.6 (3.7) b Overstory hardwood basal area (m2/ha) 0.1 (0.2) b 0.0 (0.0) b 1.0 (4.3) c 0.0 (0.3) b Midstory pine basal area (m2/ha) 0.2 (0.5) b 0.5 (1.2) b 1.8 (3.0) 1/2 0.3 (0.9) b Midstory hardwood basal area (m2/ha) 0.0 (0.0) b 0.1 (0.3) b 1.7 (1.7) c 0.2 (0.6) b Canopy closure (%) 46.3 (26.1) b 51.5 (16.9) b 55.8 (19.8) c 53.5 (15.8) be Overstory height (m) 23.1 (2.0) b 24.8 (2.8) 27.0 (3.7) a 25.3 (1.5) 1/2 Midstory height (m) 6.7 (5.4) b 3.3 (5.4) 12.4 (4.5) d 3.3 (7.3) c Understory height (m) 1.9 (0.9) b 1.6 (0.6) c 2.0 (0.8) b 1.5 (0.5) c Foliage density 0-1 rn (cm2/m 3) 0.2 (0.2) b 0.3 (0.1) b 0.3 (0.2) b 0.5 (0.7) Foliage density 1-2 rn (cm2/m 3) 0.1 (0.1) t 0.1 (0.1) b 0.2 (0.1) a 0.1 (0.2) cd Common letters indicate nonsignificant differences (ANOVA, Duncan's multiple range test [P = 0.05]). Clusters in habitat where hardwoods had recently been removed were quite similar to longleaf habitat in the actual cluster area (Table 1), how- ever, a virtual wall of hardwood midstory foliage was encountered at the edges of each cluster where midstory removal and thinning of overstory pines had ceased. Clusters that had not yet received hardwood removal treatment still had substantial hardwoods in the overstory, midstory, and understory (Table 1). During the winter of 1990-1991, hardwoods and midstory trees in the seven untreated clusters were removed, changing those clusters into a vegetative condition similar to the loblolly-shortleaf clusters that had received midstory treatment prior to the study (Table 1). Faunal use of Red-cockaded Woodpecker cavities.--A variety of ver- tebrates and invertebrates were observed using Red-cockaded Woodpeck- er cavities during the study. Although observed in cavities infrequently, American Kestrels (Falco sparverius), Eastern Screech-Owls ( Otus asio), Pileated Woodpeckers, Wood Ducks (Aix sponsa), and fox squirrels typ- ically used cavities which had both the entrance and cavity chamber en- larged. Eastern Screech-Owls were observed in three cavities with en- trances <7 cm in diameter, but the entrances of these three cavities had 702 THE WILSON BULLETIN ß Vol. 108, No. 4, December 1996 been slightly enlarged and were between 6.5 and 7 cm in diameter. Red- bellied Woodpeckers are known to conflict with Red-cockaded Wood- peckers over cavities (Neal et al. 1992, Kappes and Harris 1995) but were observed using unenlarged cavities only once during spring 1991 and on four occasions during winter. Mud-daubers (Sphecidae) were typically found in inactive cavities. Their mud chambers were tolerated or pecked off when Red-cockaded Woodpeckers began to use a cavity containing mud-dauber nests. The presence of mud-daubers or their nests did not appear to interfere with Red-cockaded Woodpecker use of cavities. However, the presence of pa- per wasps (Vespidae), particularly large nests, and honey bees (Apis mel- lifera) did prevent Red-cockaded Woodpecker use of cavities. Broad- headed skinks (Eumeces laticeps), five-lined skinks (E. fasciatus), and gray tree frogs (Hyla versicolor\chrysoscelis) were observed occasionally within inactive enlarged and unenlarged cavities. Southern flying squirrel use of woodpecker cavities.--Red-cockaded Woodpeckers preferred unenlarged cavities (Table 2); they used cavities with greatly enlarged entrances (>7 cm) in only two instances, both dur- ing late summer 1990. As previously noted by Rudolph et al. (1990b) and Loeb (1993), southern flying squirrels also prefer entrance diameters 7 cm. Thus, the southern flying squirrel exhibited extensive overlap in cavity use with Red-cockaded Woodpeckers; it was observed in relatively high numbers and also used primarily unenlarged cavities (Table 2). In most clusters, however, empty unenlarged and enlarged cavities were available throughout the year for either Red-cockaded Woodpeckers or flying squirrels to use (Tables 2, 3). Southern flying squirrel use of Red-cockaded Woodpecker cavities dur- ing the woodpecker breeding season (spring) was high, but dwindled greatly by late summer during both 1990 and 1991 (Table 2). The number of cavities used by Red-cockaded Woodpeckers was somewhat higher during late summer than during the breeding season. Red-cockaded Woodpeckers were present in greater numbers in the late summer because young woodpeckers had recently fledged from nest cavities and many were now roosting in cavities. We detected very few significant differences in the percentage of unen- larged cavities used by Red-cockaded Woodpeckers and southern flying squirrels among habitat treatments (Table 3). During spring 1990 southern flying squirrels used unenlarged Red-cockaded Woodpecker cavities at a higher frequency in longleaf pine habitat than in loblolly-shortleaf habitat where hardwood midstory vegetation was absent (Table 3). During spring 1991 the percentage of empty unenlarged cavities in 1oblolly-shortleaf pine habitat without midstory was significantly lower than in longleaf Conner et al. ß WOODPECKERS AND FLYING SQUIRRELS 703 704 THE WILSON BULLETIN ß Vol. 108, No. 4, December 1996 TABLE 3 PERCENTAGE OF UNENLARGED CAVITIES WITHIN EACH CLUSTER OCCUPIED BY RED-COCKADED WOODPECKERS BY SOUTHERN FLYING SQUIRRELS, OR EMPTY (MEAN + SD) IN LOBLOLLY- SHORTLEAF PINE HABITAT WITHOUT HARDWOOD MIDSTORY VEGETATION (N = 10), LOBLOLLY-SHORTLEAF PINE HABITAT WITH HARDWOOD MIDSTORY VEGETATION (PRE- AND POST-REMOVAL, N = 7), AND LONGLEAF PINE HABITAT (N = 11) IN EASTERN TEXAS Habitat treatment Loblolly-shortleaf Loblolly-shortleaf Loblolly-shortleaf with midstory post midstory without midstory (pre-removal) removal Longleaf Variable SD g SD g SD ß SD Spring 1990 Red-cockaded (%) 56.4 b 26.3 43.5 b 24.6 50.3 u 33.7 Flying squirrel (%) 16.9 18.8 38.8 c 28.0 44.8 31.0 Empty (%) 14.7 16.6 14.6 u 18.3 3.8 b 8.6 Spring 1991 Red-cockaded (%) 48.2 17.2 29.5 14.7 35.5 26.3 Flying squirrel (%) 41.6 14.8 38.4 15.0 28.5 20.9 Empty (%) 7.2 13.0 27.2 1/2 15.1 30.9 1/2 14.7 Summer 1990 Red-cockaded (%) 64.0 b 34.8 60.8 25.8 46.9 b 30.7 Flying squirrel (%) 7.0 9.5 14.4 14.1 6.1 11.2 Empty (%) 25.0 23.8 21.4 24.5 46.3 b 28.4 Summer 1991 Red-cockaded (%) 60.4 b 17.5 49.7 24.9 58.8 u 20.1 Flying squirrel (%) 7.8 10.4 23.0 23.2 19.6 b 13.2 Empty (%) 26.2 u 15.1 25.0 10.3 21.6 21.3 Winter 1990-1991 Red-cockaded (%) 62.8 u 25.2 38.2 c 12.6 47.1 c 26.3 Flying squirrel (%) 16.3 19.3 25.7 b 12.5 19.4 b 15.2 Empty (%) 18.9 21.2 27.8 11.4 28.1 22.9 a Common superscript letters following means indicate nonsignificant differences among habitat treatments (ANOVA, Duncan's multiple range test, P = 0.05). pine or loblolly-shortleaf pine from which midstory had been recently removed (Table 3). Southern flying squirrel use of Red-cockaded Wood- pecker cavities was not related to the presence or absence of hardwood midstory (Table 3). Thus, treatment specific and annual use of cavities by flying squirrels appears to be minimally affected by hardwood mid- story abundance. Southern flying squirrel use of cavities in the loblolly-shortleaf habitat with midstory was greater than their use of loblolly-shortleaf habitat with- out hardwood vegetation during spring 1990, although not significantly Conner et al. ß WOODPECKERS AND FLYING SQUIRRELS 705 TABLE 4 NUMBER OF RED-COCKADED WOODPECKERS ROOSTING IN THE OPEN DURING SPRING AND LATE SUMMER 1990 AND 1991 IN LOBLOLLY-SHORTLEAF FINE HABITAT WITHOUT HARDWOOD MIDSTORY VEGETATION (N = 10 CLUSTERS), WITH HARDWOOD MIDSTORY VEGETATION (PRE- AND POST-HARDWOOD REMOVAL, N = 7 CLUSTERS), AND LONGLEAF PINE HABITAT (N = 11 CLUSTERS) IN EASTERN TEXAS Season Habitat treatment Loblolly-shortleaf Loblolly-shortleaf without midstory with midstory Longleaf No. woodpeckers No. woodpeckers No. woodpeckers Spring 1990 1 1 a 6 Summer 1990 3 5 a 4 Spring 1991 1 3 b 0 Summer 1991 2 2 b 2 a Pre-midstory removal treatment within these clusters. b Post-midstory removal treatment within cluster areas completed during winter 1990-1991. so (Table 3). However, the percentage of unenlarged cavities used by flying squirrels remained the same during spring 1991 even though hard- wood midstory vegetation had been removed (Table 3). Flying squirrel use of unenlarged cavities increased in the loblolly-shortleaf area without midstory even though no habitat alteration occurred (Table 3). Both the percentage of cavities used by flying squirrels and the abundance of flying squirrels counted in Red-cockaded Woodpecker cavities decreased be- tween spring and late summer in 1990 and 1991 (Table 2, 3). We did not make a detailed survey of the crowns of nearby pines and hardwoods in the woodpecker cluster areas, but strongly suspect that flying squirrels were spending the hot, late summers in leaf nests rather than woodpecker cavities, as also observed by Muul (1974). During winter, the percentage of unenlarged cavities and available unenlarged cavities used by southern flying squirrels was relatively sim- ilar in all habitat treatments (Table 3). Empty unenlarged cavities were readily available in clusters in all habitat treatments during winter, sug- gesting that cavity availability did not create a competitive problem for Red-cockaded Woodpeckers during winter. Extra-cavity roosting by woodpeckers.--Extra-cavity roosting as de- scribed by Hooper and Lennartz (1983) is a possible indicator of insuf- ficient cavity availability for Red-cockaded Woodpeckers. In general, very few Red-cockaded Woodpeckers were observed roosting in the open (Ta- ble 4). Typically, when Red-cockaded Woodpeckers roosted in the open, there were empty cavities available within their cluster areas. Spring 1990 in the longleaf pine habitat appeared to be exceptional in this regard. 706 THE WILSON BULLETIN ß Vol. 108, No. 4, December 1996 With the exception of longleaf pine habitat in spring 1990, Red-cockaded Woodpeckers appeared to roost in the open more during late summer than during the breeding season (Table 4). Flying squirrels were very abundant during spring 1990 in the longleaf habitat (Table 2) and empty cavities were few, suggesting that a few Red-cockaded Woodpeckers may have been forced temporarily to roost in the open. Many recently fledged young Red-cockaded Woodpeckers did not roost in cavities during the late sum- mer. Because many unenlarged empty cavities were available for these woodpeckers to use during late summer, roosting in the open appears to be voluntary and may have been in response to the typical high air tem- peratures during August and September. Red-cockaded Woodpecker fiedging success.--We examined Red-cock- aded Woodpecker fledging success to explore the possibility that inter- actions with southern flying squirrels reduced woodpecker nest produc- tivity. Because southern flying squirrel use of woodpecker cavities was uniformly high over all habitat treatments and years, our ability to eval- uate the influence of squirrel use of cavities on fledging success through comparisons across habitats was limited. Fledging success was slightly higher in loblolly-shortleaf habitat with hardwood vegetation (pre-hardwood removal) than in the loblolly-short- leaf habitat without hardwood vegetation during 1990 (Fig. 1). Fledging success remained somewhat higher in these cluster areas in 1991 (post- treatment) even though the hardwood vegetation had been removed prior to the 1991 breeding season. Longleaf pine habitat, relatively devoid of hardwood vegetation, and often considered the premiere habitat for the woodpecker, had a slightly lower fledging success than either loblolly- shortleaf habitats during both 1990 and 1991 (Fig. 1). Excluding nests where eggs failed to hatch, we failed to detect any significant differences in fledging success among habitat treatments (Kruskal-Wallis X 2 approx- imation, X 2 = 1.42, P = 0.49). We compared the proportion of all unenlarged cavities used by flying squirrels and the proportion of available unenlarged cavities (open cavities not used by Red-cockaded Woodpeckers) that contained flying squirrels with woodpecker fledging success (Fig. 1). During both 1990 and 1991 we observed no relationship between southern flying squirrel occupancy and habitat condition (abundance of hardwood midstory) or Red-cock- aded Woodpecker fledging success. Red-cockaded Woodpecker fledging success per habitat treatment during the two breeding seasons (N = 6) was not correlated with the percentage of all unenlarged cavities occupied by southern flying squirrels (rs = -0.08, P = 0.87) or the percentage of available unenlarged cavities (those not used by Red-cockaded Wood- peckers) occupied by southern flying squirrels (r s = -0.46, P = 0.35). If Conner et al. ß WOODPECKERS AND FLYING SQUIRRELS 707 u- 2.4 2.2 2.0 1.8 1.6 1.4 1.2 1.0 0.8 0.6 0.4 0.2 0.0 1990 LOBLOLLY-SHORTLEAF WITH MIDSTORY (PRE REMOVAL) [] RCW FLEDGED PER CLUSTER [] PROP. ALL CAVITIES WITH G.V. [] PROP. OPEN CAVITIES WITH G.V. LOBLOLLY-SHORTLEAF LONGLEAF WITHOUT MIDSTORY 2.6 2.4 u') 1.4 O-r- 0.8 re I- 0.6 0.4 0.2 0.0 1991 LOBLOLLY-SHORTLEAF WITH MIDSTORY NOW REMOVED [] RCW FLEDGED PER CLUSTER [] PROP. ALL CAVITIES WITH G.V. [] PROP. OPEN CAVITIES WITH G.V. LOBLOLLY-SHORTLEAF LONGLEAF WITHOUT MIDSTORY HABITATTREATMENT FIG. 1. Comparisons of Red-cockaded Woodpecker fiedging success with the proportion of all unenlarged Red-cockaded Woodpecker cavities occupied by southern flying squirrels (Glaucomys volans, G. V.) and available unenlarged cavities (those not used by Red-cock- aded Woodpeckers) occupied by southern flying squirrels in Ioblolly-shortleaf pine habitat with hardwood midstory present (pre- and post-hardwood removal), loblolly-shortleaf pine habitat without hardwood midstory throughout the study, and longleaf pine habitat during the 1990 and 1991 breeding seasons on the Angelina and Davy Crockett National Forests in eastern Texas. 708 THE WILSON BULLETIN ß Vol. 108, No. 4, December 1996 clusters are treated as the sample unit (N -- 44), fledging success is still not correlated with either the percentage of all unenlarged cavities oc- cupied by flying squirrels (rs -- -0.18, P -- 0.23), or the percentage of available unenlarged cavities occupied by flying squirrels (r = -0.11, P = 0.48). Potential for flying squirrel predation on woodpeckers.--During our 2 year study we observed 6 instances where Red-cockaded Woodpeckers nested and produced young in cavities while southern flying squirrels were occupying other cavities in the same pine tree. In only one instance were eggs lost (to unknown causes), but the woodpeckers renested and successfully fledged young from the same cavity. Young fledged suc- cessfully from all five of the other nest cavities. DISCUSSION Competition between flying squirrels and woodpeckers.--Our obser- vations in eastern Texas suggest a minimal competitive impact of southern flying squirrels on Red-cockaded Woodpeckers. Because we did not mea- sure woodpecker fledging success over a wide range of flying squirrel abundance, however, our results may not be definitive. Competition from southern flying squirrels in Texas is likely transient and occurs as isolated events during ecological "bottle-necks." If such competition occurs at all in eastern Texas, the effects are subtle rather than overwhelming. The effect of southern flying squirrels on any healthy woodpecker population is probably minimal to non-existent. Specifically, we have not seen (1) a relationship between woodpecker fledging success and flying squirrel use of cavities, (2) Red-cockaded Woodpeckers forced to roost in the open because of a squirrel-caused shortage of unenlarged cavities, or (3) regular squirrel predation on Red- cockaded Woodpecker eggs and young even when both woodpeckers and flying squirrels occupied the same cavity tree. Relationships among woodpeckers, squirrels, and hardwood vegeta- tion. We did not observe a strong relationship between southern flying squirrel abundance and presence of hardwood vegetation. Flying squirrels were common in cavities in longleaf pine habitat with almost no hard- wood vegetation. This finding, however, does not negate the necessity to reduce hardwood vegetation within woodpecker cluster areas. Past studies have clearly demonstrated the negative effects of excessive hardwood midstory on woodpecker populations (Van Balen and Doerr 1978, Hovis and Labisky 1985, Conner and Rudolph 1989, Loeb et al. 1992). Thus, we strongly urge that reduction (not elimination) of hardwood vegetation within Red-cockaded Woodpecker cluster areas be continued. Our results indicate that complete or partial removal of all hardwoods Conner et al. ß WOODPECKERS AND FLYING SQUIRRELS 709 will likely not affect the use of Red-cockaded Woodpecker cavities by southern flying squirrels. What our study suggests is that southern flying squirrels are not the cause of harmful effects on Red-cockaded Wood- peckers associated with the presence of hardwood vegetation within their cluster areas. As we have suggested before (Conner and Rudolph 1991b), Red-cockaded Woodpeckers may have an innate avoidance of areas with extensive hardwood vegetation as a result of their adaptation to the south- ern fire-climax pine ecosystem. A selective advantage may accrue for Red-cockaded Woodpecker pairs that avoid habitat with abundant hard- wood vegetation because such areas may support greater numbers of other species of woodpeckers that can easily out-compete Red-cockaded Wood- peckers for cavities or destroy the cavities they excavate. Another possible reason why Red-cockaded Woodpeckers have an aversion to hardwoods is that they may provide predators access to cavities (Walters 1990). We saw no negative effect of southern flying squirrels on Red-cockaded Woodpeckers, nor have any other studies demonstrated such an effect. We strongly discourage removal and euthanasia of southern flying squir- rels in woodpecker clusters because of the complete lack of evidence that it would benefit Red-cockaded Woodpeckers. If removal of southern fly- ing squirrels is deemed necessary, it should be based on site-specific data that statistically demonstrates a severe competitive problem. In such in- stances, control of flying squirrels should last only as long as the wood- pecker population is small and vulnerable to sudden extirpation. ACKNOWLEDGMENTS We thank J. R. Walters, R. G. Hooper, E C. James, J. D. Ligon, S. C. Loeb, B. Parresol, and J. E Taulman for constructive comments on an early draft of the manuscript. Partial funding was provided by a Challenge Cost Share Agreement (#19-90-008) with the Resource Protection Division, Texas Parks and Wildlife Dept. LITERATURE CITED BENDEL, P. R. AND J. E. GATES. 1987. Home range and microhabitat partitioning of the southern flying squirrel (Glaucomys volans). J. Mammal. 68:243-255. CARTER, J. H. III, J. R. WALTERS, S. H. EVERHART, AND P. D. DOERR. 1989. Restrictions for Red-cockaded Woodpecker cavities. Wildl. Soc. Bull. 17:68-72. CONNER, R. N., R. G. HOOPER, H. S. CRAWFORD, AND H. S. MOSBY. 1975. Woodpecker nesting habitat in cut and uncut woodlands in Virginia. J. Wildl. Manage. 39:144-150. --AND B. A. LOCKE. 1982. Fungi and Red-cockaded Woodpecker cavity trees. Wilson Bull. 94:64-70. -- AND K. A. O'HALLORAN. 1987. Cavity-tree selection by Red-cockaded Woodpeckers as related to growth dynamics of southern pines. Wilson Bull. 99:398-412. -- AND D.C. RUDOLPH. 1989. Red-cockaded Woodpecker colony status and trends on the Angelina, Davy Crockett, and Sabine National Forests. U.S. Dept. Agric., For. Serv. Res. Pap. 50-250. 710 THE WILSON BULLETIN ß Vol. 108, No. 4, December 1996 AND ---. 1991a. Forest habitat loss, fragmentation, and Red-cockaded Wood- pecker populations. Wilson Bull. 103:446-457. AND . 1991b. Effects of midstory reduction and thinning in Red-cockaded Woodpecker cavity tree clusters. Wildl. Soc. Bull. 19:63-66. -- AND . 1995. Excavation dynamics and use patterns of Red-cockaded Wood- pecker cavities: relationships with cooperative breeding. Pp. 343-352 in Red-cockaded Woodpecker: recovery, ecology and management (D. L. Kulhavy, R. G. Hooper, and R. Costa, eds.). College of Forestry, Stephen E Austin State Univ., Nacogdoches, Texas. --, D. L. KULHAVY, AND A. E. SNOW. 1991. Causes of mortality of Red- cockaded Woodpecker cavity trees. J. Wildl. Manage. 55:531-537. DENNIS, J. V. 1971. Species using Red-cockaded Woodpecker holes in northeastern South Carolina. Bird-Banding 42:79-87. GAINES, G. D., K. E. FRANZREB, D. H. ALLEN, K. S. LAVES, AND W. L. JARVIS. 1995. Red- cockaded Woodpecker management on the Savannah River site: a management/research success story. Pp. 81-88 in Red-cockaded Woodpecker: recovery, ecology and man- agement (D. L. Kulhavy, R. G. Hooper, and R. Costa, eds.). College of Forestry, Ste- phen E Austin State Univ., Nacogdoches, Texas. HOOPER, R. G. 1988. Longleaf pines used for cavities by Red-cockaded Woodpeckers. J. Wildl. Manage. 52:392-398. -- nrD M. R. LENNARTZ. 1983. Roosting behavior of Red-cockaded Woodpecker clans with insufficient cavities. J. Field Ornithol. 54:72-76. --, AND H. D. MUSE. 1991. Heart rot and cavity tree selection by Red- cockaded Woodpeckers. J. Wildl. Manage. 55:323-327. HoPVaNS, M. L. AND T. E. LYNN, JR. 1971. Some characteristics of Red-cockaded Wood- pecker cavity trees and management implications in South Carolina. Pp. 140-169 in The ecology and management of the Red-cockaded Woodpecker (R. L. Thompson, ed.). Bur. Sport Fish and Wildl. and Tall Timbers Res. Stn., Tallahassee, Florida. HOVlS, J. A. AND R. E LABISKY. 1985. Vegetative associations of Red-cockaded Woodpecker colonies in Florida. Wildl. Soc. Bull. 13:307-314. JAClCSON, J. A. 1974. Gray rat snakes versus Red-cockaded Woodpeckers: predator-prey adaptations. Auk 91:342-347. KAPPES, J., JR. AND L. D. HARRIS. 1995. Interspecific competition for Red-cockaded Wood- pecker cavities in the Apalachicola National Forest. Pp. 389-393 in Red-cockaded Woodpecker: recovery, ecology and management (D. L. Kulhavy, R. G. Hooper, and R. Costa, eds.). College of Forestry, Stephen E Austin State Univ., Nacogdoches, Texas. LIGON, J. D. 1970. Behavior and breeding biology of the Red-cockaded Woodpecker. Auk 87:255-278. LOEB, S.C. 1993. Use and selection of Red-cockaded Woodpecker cavities by southern flying squirrels. J. Wildl. Manage. 57:329-335. --, W. D. PEPPER, AND A. T. DOYLE. 1992. Habitat characteristics of active and aban- doned Red-cockaded Woodpecker colonies. So. J. Appl. For. 16:120-125. MACARTUR, R. H. ANt) J. W. MACARTUR. 1961. On bird species diversity. Ecology 42: 594-598. MuuL, I. 1974. Geographic variation in the nesting habits of Glaucomys volans. J. Mammal. 55:840-844. NEAL, J. C., W. G. MONTAGUE, AND D. A. JAMES. 1992. Sequential occupation of cavities by Red-cockaded Woodpeckers and Red-bellied Woodpeckers in the Ouachita National Forest. Arkansas Acad. Sci. 46:106-108. REI, LER, A. W. 1972. Aspects of behavioral ecology of Red-headed and Red-bellied wood- peckers. Am. Midl. Nat. 88:270-290. Conner et al. ß WOODPECKERS AND FLYING SQUIRRELS 711 RUDOLPH, D. C., H. KYLE, AND R. N. CONNER. 1990a. Red-cockaded Woodpeckers vs rat snakes: the effectiveness of the resin barrier. Wilson Bull. 102:14-22. , R. N. CONNER, AND J. TURNER. 1990b. Competition for Red-cockaded Woodpecker (Picoides borealis) roost and nest cavities: the effects of resin age and cavity entrance diameter. Wilson Bull. 102:23-36. -- AND . 1991. Cavity tree selection by Red-cockaded Woodpeckers in relation to tree age. Wilson Bull. 103:458-467. , --, AND R. R. SCHAEFER. 1995. Red-cockaded Woodpecker detection of red heart infection. Pp. 338-342 in Red-cockaded Woodpecker: recovery, ecology and man- agement (D. L. Kulhavy, R. G. Hooper, and R. Costa, eds.). College of Forestry, Ste- phen F. Austin State Univ., Nacogdoches, Texas. U.S. FISH AND WILDLIFE SERVICE. 1985. Red-cockaded Woodpecker recovery plan. U.S. Fish and Wildl. $erv., Atlanta, Georgia. VAN BALEN, J. B. AND P. D. DOERR. 1978. The relationship of understory vegetation to Red-cockaded Woodpecker activity. Proc. Annu. Conf. Southeast. Assoc. Fish and Wildl. Agencies 32:82-92. WALTERS, J. R. 1990. Red-cockaded Woodpeckers: a 'primitive' cooperative breeder. Pp. 69-101 in Cooperative breeding in birds (P. B. Stacey and W. D. Koenig, eds.). Cam- bridge Univ. Press, London, United Kingdom. , P. D. DOERR, AND J. H. CARTER, III. 1988. The cooperative breeding system of the Red-cockaded Woodpecker. Ethology 78:275-305.