Selection of song perches by Cerulean Warblers.

ABSTRACT. Cerulean Warbler (Dendroica cerulea) populations have experienced serious declines in recent decades. Due to limited research on this species, a better understanding of habitat use and territory site selection at the regional level is needed. Within the Pleasant Run Unit of the Hoosier National Forest, Morgan-Monroe State Forest, and Yellowwood State Forest, a total of 43 Cerulean Warbler territories were mapped during 2004 and 2005. Within each territory, crown height, diameter at breast height (dbh), and species were recorded for trees used as song perches. Dbh, height measurements, and species were also recorded for trees within a 0.04 ha circle at the center of each territory. Cerulean Warbler song-perch trees were larger and taller than surrounding trees within territories. In addition, certain species were selected as song-perch trees, whereas others were avoided.

Keywords: Cerulean Warbler, Dendroica cerulea, song-perch trees, territory, Indiana

**********

The Cerulean Warbler is a species of conservation concern; Breeding Bird Survey (BBS) data have demonstrated an annual population decline of 3.7% between 1966-1996 (Hamel 2000a). Only five other North American breeding bird species showed greater declines. Until recently, very little study had focused on this species (Hamel 2000b). Because of its dependence on large tracts of mature deciduous forest for successful breeding, much of its decline is likely due to extensive loss and fragmentation of forest tracts for agricultural use (Oliarnyk & Robertson 1996; Hamel 2000a).

Robbins et al. (1992) demonstrated that the Cerulean Warbler is a canopy-dwelling species. Among closely related species, this small bird spends most of its time higher in the canopy than other wood warblers (Hamel 2000a). Some studies have suggested that Cerulean Warblers use larger trees as song perches (Lynch 1981; Robbins et al. 1992; Hamel 2000a). However, given the substantial variation in habitat across its breeding range (differences in tree species composition, size of available trees, and forest tract size) and in behavioral variables (territory size and site fidelity), inquiry into song-perch tree characteristics on the regional level is essential (Hamel 2000a; Jones & Robertson 2001; Roth 2004). The purpose of this study was to determine if song-perch trees are larger and taller than surrounding trees within Cerulean Warbler territories in southern Indiana. Also, selection of specific tree species as song-perches was investigated.

STUDY AREA

This study took place from 1 May to mid-August of 2004 and 2005 in the Pleasant Run unit of the Hoosier National Forest, Yellowwood State Forest, and Morgan-Monroe State Forest in Brown, Morgan, Lawrence, and Jackson counties, Indiana (Fig. 1). Historically, the Cerulean Warbler was one of the most abundant breeding warblers in the Ohio and Mississippi river valleys (Hamel 2000a). As a part of that area, forest blocks used in this study are among the largest and most unfragmented in southern Indiana.

[FIGURE 1 OMITTED]

METHODS

Bird surveys.--In each of 10 study sites, presence of male Cerulean Warblers was determined by walking seven transects within a 1.96 [km.sub.2] plot, with seven sampling points per transect, each point 200 m apart (Fig. 2). Transect point locations were recorded in Universal Transverse Mercators (UTMs) coordinates using Global Positioning System (GPS) receivers. To reduce edge influences, surveys were conducted > 50 m from roads. Bird surveys began 1 May, and were completed by 30 May. Surveys were conducted between 0530-1030 h, excluding rainy days (presence of precipitation), when cessation or reduction of vocalizing may occur. At each survey point, 3 minutes of listening for Cerulean Warbler vocalizations commenced, followed by a 15 second playback of a conspecific male song in each of the cardinal directions to elicit a vocal response. This was followed by an additional three minutes of listening before moving to the next survey point (Falls 1981). The compass bearing and distance of detected males was estimated from the nearest transect point.

[FIGURE 2 OMITTED]

Territory mapping.--Most male Cerulean Warblers were relocated after initial detection during surveys because they maintain territories during the breeding season (Hamel 2000a). Once surveys provided the initial location of a male, territories were mapped by flagging a minimum of 5 trees in which males vocalized and perched. Singing from territory boundaries is the primary means of defending a territory and attracting potential mates (Hamel 2000a). For the purpose of this study, trees in which males vocalize are called song-perch trees. A territory is defined as the area within the perimeter of song-perch trees. UTMs of song-perch trees were recorded for ease in returning to the territory and calculating the territory center.

Vegetation measurement and analysis.--From early July to mid-August 2004 and 2005, vegetation was sampled within each of 43 territories, using the methods of James & Shugart (1970) outlined below. In territories, a 0.04 ha circular plot was marked at the approximate center of the territory. Diameter at breast height (dbh) and height were recorded for each tree in the plot with dbh [greater than or equal to] 10 cm, and species for each tree with dbh [greater than or equal to] 3 cm. A Nikon Laser 440[TM] compact rangefinder was used to determine tree heights. Dead trees were measured in the same manner as live trees. Dbh, height, and species were also recorded for all song-perch trees.

Dbh, height, and species were compared between trees from territory sample plots and song-perch trees. Results were calculated using all individual trees, as well as means calculated by territory, t-tests were used for comparison of individual trees, and paired t-tests were used for comparison of means. Chi-square analysis was used to compare tree species diversity between trees from sample plots and song-perch trees. Level of significance was set at P = 0.05.

RESULTS

Statistical analyses were computed using pooled data from 2004 and 2005. Mean values are reported as mean [+ or -] 1 SD. Mean number of Song-perch trees was 13.8 trees per territory (range 5-27). Cerulean Warblers used significantly larger ([bar.x] = 43.0 [+ or -] 14.1 cm, n = 594, P < 0.001, t = 19.26) and taller ([bar.x] = 26.9 [+ or -] 4.11 m, n = 591, P < 0.001, t = 18.10) trees for song perches than the average trees available within territories (2 = 27.4 [+ or -] 15.6 cm, n = 751; [bar.x] = 21.5 [+ or -] 6.04 m, n = 604, dbh and height, respectively) (Figs. 3, 4). Comparison of means by territory reflected the same pattern; perch trees were larger ([bar.x] = 44.3 [+ or -] 6.3 cm, n = 43) and taller ([bar.x] = 27.0 [+ or -] 2.2 m, n = 43) than surrounding trees ([bar.x] = 27.7 [+ or -] 3.71 m, n = 43; [bar.x] = 21.6 [+ or -] 2.0 m, n = 43, dbh and height, respectively) within territories (Figs. 5, 6).

[FIGURES 3-6 OMITTED]

Of the 39 tree species (including snags) present in territories, 12 were used more often than expected as song-perch trees (Table 1). Six species were used with less frequency than expected as song-perch trees (Table 1). Pawpaw (Asimina triloba), blue beech (Carpinus caroliniana), Ironwood (Ostrya virginiana), and grape (Vitis spp.), were never observed being used as song perches by Cerulean Warblers. Notably, 116 of 595 song perch trees in our study were white oak, and 105 were bitternut hickories. These two species made up 37% of all song perch trees in the study.

DISCUSSION

Within our study, only a handful (~2%) of song-perch trees within territories were < 15 m, and these were associated with canopy gaps or clearings. The vast majority of song-perch trees were relatively mature trees that were larger than those around them. Cerulean Warblers have long been known as canopy-dwelling birds (Robbins et al. 1992; Hamel 2000a). Robbins et al. (1992) noted that Cerulean Warblers were more often found in trees with larger dbh, and spent most of their time above the middle of the tree. In this study, song-perch trees had significantly larger dbh and height. Dbh and height are related, but there is much variation possible across individual trees, which gives value to comparison of both variables (Morey 1936).

Cerulean Warblers may also be selecting territories based on the presence of ideal song-perch trees. This will not always be apparent, as they will sing while foraging, preening, and in flight. However, when singing is the focus, they often choose and make return visits to exposed perches from which they sing (Hamel 2000a; KJ pers. obs.). Taller and larger trees may offer more high-quality song perches conducive to vocal projection. In a study of Golden-winged Warblers (Vermivora chrysoptera), song-perch trees were significantly larger than expected; and it was suggested that this song perch selection enhanced the birds' ability to display vocally and visually in mate attraction (Rossell 2001).

Much of the research on Cerulean Warbler habitat has focused largely on canopy structure, with less attention to tree species composition (Hamel 2000a, Jones and Robertson 2001, Weakland and Wood 2005). Oliarnyk (1996) and Hamel (2000b) reported no preference in Cerulean Warblers for tree species in nesting or foraging in Ontario and Tennessee populations, but song-perch tree preferences were not investigated. Among 13 species of foliage-gleaning birds in floodplain forests in southern Illinois, the Cerulean Warbier was the second most selective bird, closely following the Yellow-throated Warbler (Dendroica dominica), in tree species usage (Gabbe et al. 2002).

Gabbe et al. (2002) found that Cerulean Warblers in Illinois showed the strongest preference for shellbark hickory (Carya laciniosa) and bitternut hickory (Carya cordiformis). Tree species that were most strongly avoided were red maple (Acer rubrum) and blue beech. Shellbark hickory had such a low density overall in our study area, its usage by Cerulean Warblers cannot be compared to the study in southern Illinois. We also found that males selected bitternut hickories as song perches together with white oaks. Red maples were avoided, and blue beeches were never used for song perches. It appears that Cerulean Warblers are adaptable in their use of tree species across their breeding range, but where breeding areas have tree species in common, Cerulean Warblers appear to be fairly consistent in species selection.

Thirty percent of trees in territories were sugar maples, yet only 2.4% of song perch trees were of this species. On the other hand, white oaks and bitternut hickories represented only 5.1% and 3.7% of trees within territories, respectively, but 19.5% and 17.6% of song perch trees were of those species. The appearance of tree species selectivity in our study area may be due to a relationship between species and size parameters. Sugar maples may not have been used in proportion to their abundance because many of them were not canopy trees, and being a canopy-dwelling species, Cerulean Warblers would not encounter them. However, even in territories where sugar maples made up the majority of canopy trees, some males avoided them completely during our observations. The only sugar maples recorded as song perch trees were concentrated in the territories of what appeared to be a few exceptional birds, most of which were unpaired. Further study addressing tree species importance, or at least importance of certain crown types associated with groups of species, would be worth pursuing.

This study demonstrated that in southern Indiana, Cerulean Warblers are utilizing the largest and tallest trees in their territories as song perches. These trees may offer individuals some advantage in territory defense and mate attraction, acoustically and/or visually. Cerulean Warblers in southern Indiana also used tree species disproportionately to their availability, just as they have been demonstrated to do in Illinois (Gabbe et al 2002). Exploration of more specific aspects of song perch selection (e.g. documentation of perch heights relative to tree heights, approximate girth of perch branches or twigs, foliage cover on perch) throughout the breeding range would be helpful in deepening our understanding of the specific habitat needs of Cernlean Warblers (Robbins et al. 1992).

ACKNOWLEDGMENTS

Funding for this study was provided by the Indiana Academy of Science, Ball State University Office of Academic Research and Sponsored Programs, U.S. Fish and Wildlife Service, U.S. Forest Service, and Sigma Xi. This study would not have been possible without the field assistance of Kirk Roth, April Howard, Sarah Register, Matt Kellam, and Corey Shaffer. We thank John Castrale, one anonymous reviewer, and the Editor of the Proceedings for their comments and suggestions which greatly improved the quality of the manuscript.

LITERATURE CITED

Falls, J.B. 1981. Mapping territories with playback: An accurate census method for songbirds. Studies in Avian Biology 6:86-91.

Gabbe, A. P., S. K. Robinson, & J. D. Brawn. 2002. Tree-species Preferences of Foraging Insectivorous Birds: Implications for Floodplain Forest Restoration. Conservation Biology 16:462-470.

Hamel, RB. 2000a. Cerulean Warbler (Dendroica cerulea). In The Birds of North America, No. 511 (A. Poole and E Gill, eds.). The Birds of North America, Inc., Philadelphia, Pennsylvania.

Hamel, RB. 2000b. Cerulean Warbler status assessment. U.S. Fish and Wildlife Service, Minneapolis, Minnesota.

James, EC. & H.H. Shugart. 1970. A quantitative method of habitat description. Audubon Field Notes 24:727- 736.

Jones, J. & R.J. Robertson. 2001. Territory and nest-site selection of Cerulean Warblers in eastern Ontario. The Auk 118:727-735.

Lynch, J.M. 1981. Status of the Cerulean Warbler in the Roanoke River basin of North Carolina. Chat 45:29-35.

Morey, H.F. 1936. Age-size relationships of Hearts Content, a virgin forest in northwestern Pennsylvania. Ecology 17:251-257.

Oliarnyk, C.J. 1996. Habitat selection and reproductive success in a population of Cerulean Warblers in southeastern Ontario. M.S. thesis, Queen's University, Kingston, Ontario, Canada.

Oliarnyk, C.J. & R.J. Robertson. 1996. Breeding behavior and reproductive success of Cerulean Warblers in southeastern Ontario. Wilson Bulletin 108:673-684.

Robbins, C.S., J.W. Fitzpatrick & P.B. Hamel. 1992. A warbler in trouble: Dendroica cerulea. Pp. 549-562 In Ecology And Conservation Of Neotropical Migrant Landbirds (J. M. Hagan, III. and D. W. Johnston, eds.). Smithsonian Institution Press, Washington, D.C.

Rossell, C.R. Jr. 2001. Song perch characteristics of Golden-winged Warblers in a mountain wetland. Wilson Bulletin. 113:246-248.

Roth, K.L. 2004. Cerulean Warbler breeding biology. M.S. thesis, 46 Pp. Ball State University, Muncie, Indiana.

Weakland, C.A. & EB. Wood. 2005. Cerulean Warbler (Dendroica Cerulea) Microhabitat and Landscape-Level Habitat Characteristics in Southern West Virginia. The Auk 122:497-508.

Manuscript received 27 December 2005, revised 7 March 2006.

Kelly Jones and Kamal Islam: Department of Biology, Ball State University, Muncie, Indiana 47306-0440 USA

Corresponding author: Kamal Islam: Department of Biology, Ball State University, Muncie, Indiana 47306-0440 USA; Phone: [phone]; fax #: [phone]; e-mail: [e-mail]

Table 1.--Thirty-eight tree species and snags identified within
Cerulean Warbler (Dendroica cerulea) territories in southern Indiana
during 2004 and 2005 in the Hoosier National Forest, Yellowwood State
Forest, and Morgan-Monroe State Forest. An "X" indicates significantly
higher than expected frequencies of tree species in Cerulean Warbler
territory sample plots vs. song-perch trees. An asterisk indicates
exclusive occurrence within the category.

                                                 Territory
                                                  sample     Song-perch
   Tree species           Scientific name          plots       trees

Red maple             Acer rubrum                   X
Sugar maple           Acer saccharum                X
Downy Juneberry       Amelanchier arborea
Hercules-club         Aralia spinosa
Pawpaw                Asimina triloba               X *
Blue beech            Carpinus caroliniana          X *
Bitternut hickory     Carya cordiformis                          X
Pignut hickory        Carya glabra                               X
Shagbark hickory      Carya ovata                                X
Eastern redbud        Cercis canadensis
Flowering dogwood     Cornus florida                X
American beech        Fagus grandifolia             X
White ash             Fraxinus americana
Green ash             Fraxinus pennsylvananica                   X
Honey locust          Gleditsia triacanthos
Black walnut          Juglans nigra                              X
Eastern red cedar     Juniperus virginiana
Spicebush             Lindera benzoin
Tulip poplar          Liriodendron tulipifera                    X
Black gum             Nyssa sylvatica               X
Ironwood              Ostrya virginiana             X *
Short-leaf pine       Pinus echinata
Eastern white pine    Pinus strobus
Sycamore              Platanus occidentalis                      X
Bid-toothed aspen     Populus grandidentata
Black cherry          Prunus serotina
White oak             Quercus alba                               X
Chinquapin oak        Quercus mueldenbergii                      X
Chestnut oak          Quercus prinus                             X
Red oak               Quercus rubra                              X
Black oak             Quercus velutina                           X
Winged sumac          Rhus copallina
Black locust          Robinia pseudoacacia
Sassafrass            Sassafrass albidum
Basswood              Tilia americana
American elm          Ulmus americana
Red elm               Ulmus rubra
Grape species         Vitis spp.                    X *
  snag                --                            X