Help

Regular issues
Monographs
Special Issues



Southeastern Naturalist
    SENA Home
    Aim and Scope
    Board of Editors
    Staff
    Editorial Workflow
    Publication Charges
    Subscriptions

Co-published Journals
    Northeastern Naturalist
    Caribbean Naturalist

EH Natural History Home

A Comparison of Fixed-Width Transects and Fixed-Radius Point Counts for Breeding-Bird Surveys in a Mixed Hardwood Forest
James F. Taulman

Southeastern Naturalist, Volume 12, Issue 3 (2013): 457–477

Full-text pdf (Accessible only to subscribers.To subscribe click here.)

 

Site by Bennett Web & Design Co.
2013 SOUTHEASTERN NATURALIST 12(3):457–477 A Comparison of Fixed-Width Transects and Fixed-Radius Point Counts for Breeding-Bird Surveys in a Mixed Hardwood Forest James F. Taulman* Abstract - The fixed-width strip-transect and fixed-radius point-count survey methods for breeding birds were evaluated side-by-side in 2009, 2010, and 2011 in a 200-ha mixed hardwood forest surrounded by urban development in Parkville, MO. One 2-ha strip transect (80 m x 250 m) and four 0.5-ha fixed-radius plots (40 m radius, 150 m separation) were surveyed in adjacent riparian forest areas during May and June, 2009. In 2010, two additional sets of transects and corresponding circular-plot arrays were installed, bringing the total area surveyed by each method to 6 ha in 2010 and 2011. Abundance of individuals of all species was greater on circular-plot arrays compared with transects in both 2009 and 2010. Modeling the potential intersection of transect and circular-plot arrays on a background simulating a landscape distribution of bird territories at varying densities indicated that a dispersed array of circular survey plots may overlap more bird territories than contiguous strip transects, though both survey plots enclose the same total forest area. The fixed-radius point-count method appears to effectively sample a larger forest patch than the fixed-width transect method, possibly resulting in estimations of bird population parameters that are different between the two methods. Introduction A variety of survey methods have been developed and used by researchers attempting to describe characteristics of bird populations. The fixed-width striptransect method (Conner and Dickson 1980) consists of a rectangular area in a habitat through which the observer walks along a center line, recording birds seen or heard out to a specified distance on each side. A transect size of 80 m x 250 m is commonly used (Dickson et al. 1993, Thill and Koerth 2005, Watson 2004) because these dimensions provide a 2-ha area in each surveyed plot, and the 40-m distance from the observer to the plot boundary on each side is a good compromise between covering as much area as possible and still allowing detection of species that are quiet or otherwise cryptic (Alldredge et al. 2007, Hutto et al. 1986). The fixed-radius circular point-count method (Hutto et al. 1986) is also commonly applied to studies of bird populations (Buckland 2006, Carey 1988, Gregory et al. 2004, Pagen et al. 2000, Petit et al. 1994, Rodewald and Smith 1998, Tarvin et al. 1998). Circular-plot centers are typically either spread around a study area (Petit et al. 1994) or spaced along a linear transect through a habitat to be surveyed (Carey 1988, Gregory et al. 2004, Hutto et al. 1986). Fixed-radius circular plots are separated by rather large distances, 150 m (Manuwal and Carey 1991) to 200 m (Barber et al. 2001) or even 250 m and more (Ralph et al. 1993) *Department of Natural and Physical Sciences, Park University, 8700 NW River Park Drive, Parkville, MO 64152; james.taulman@park.edu. J.F. Taulman 2013 Southeastern Naturalist Vol. 12, No. 3 458 between plot centers, in order to reduce the possibility of recounting individual birds on adjacent plots. Both of these survey methods allow the creation of comparative indices describing 1) bird species richness (number of different species recorded per visit), 2) frequency of species detection (percent of replicated surveys on which a species was detected), 3) abundance (the mean numbers of individuals of a given species detected over all replicated surveys), 4) relative abundance (total numbers of individuals of a each species divided by the total numbers of all birds detected), and 5) densities of each species (mean number of birds of each species detected per survey divided by the plot area) (Conner and Dickson 1980, Hutto et al. 1986, Manuwal and Carey 1991, Thill and Koerth 2005). Species diversity and evenness indices can also be computed from these two survey methods (Conner and Dickson 1980). In comparing the benefits and limitations of each of these methods, Manuwal and Carey (1991) found both the fixed-width transect and fixed-radius point-count methods to be suitable for determining relative abundance, population trends, and densities. They favored the fixed-width strip-transect method over fixed-radius circular plots for estimation of species richness. Verner and Ritter (1985) found the fixed-width transect and fixed-radius point-count methods equally applicable to species richness estimates. They reported certain advantages in the fixedradius point-count method over the strip-transect surveys, such as 1) allowing better control of timing of the counting period, 2) allowing the observer to concentrate fully on bird detection and identification during the survey, without the distraction of having to walk through the plot, and 3) allowing a bird survey to be conducted in a small habitat patch. The strip-transect method was deemed more efficient at collecting total bird counts, however, because a given area can be surveyed in one timed bout, whereas a certain amount of additional time is necessary in traveling between an array of dispersed circular plots enclosing the same survey area as the transect. Verner and Ritter (1985) and Buckland (2006) concluded that transects provide superior results in estimating bird densities than circular-plot counts. Criteria used for selection of the survey method employed are often not stated (such as Noss 1991), but choosing between these two common methods has sometimes been based on shape of the habitat to be surveyed. A large, contiguous study area may be effectively surveyed with an array of fixed-radius circular plots spaced throughout it in order to provide full coverage (Petit et al. 1994). But small or narrow habitats might be more easily and adequately surveyed with a number of fixed-width strip transects (Noss 1991, Thill and Koerth 2005). Gregory et al. (2004) suggested using strip transects for large, heterogeneous, open habitats with conspicuous bird species, while advising that circular-plot point counts are better suited for forest habitats and more cryptic species. Ralph et al. (1993) stated that strip transects are very similar to point counts, but point counts are the more efficient method in forests. Despite describing many contrasting features favoring one method over another in different applications, Gregory et al. (2004) consider strip (line) transects and circular-plot arrays to be 459 J.F. Taulman 2013 Southeastern Naturalist Vol. 12, No. 3 equivalent variations on the transect method. They even suggest that strip transects and circular survey plots may be combined for a single survey. Fixed-width strip-transect and fixed-radius point-count survey methods have been used interchangeably in breeding-bird surveys without an evaluation of the comparability of population indices produced by each method. In order to directly examine the similarity of data produced by these two survey methods, I conducted a side-by-side evaluation of the strip-transect and fixed-radius circular-point survey techniques, employing both methods to survey adjacent riparian forest patches in a small mixed hardwood forest. My working hypothesis was that both methods would produce similar results with regard to bird population parameters commonly computed through fixed-area surveys, such as species richness and abundance. The intent of this study was not to evaluate the effectiveness of either of these survey methods in estimating true population parameters of bird species. Regardless of their predictive accuracy, both survey methods are still used by researchers to gain information about bird populations and to compare population indices among different sites (Hostetler and Main 2011, Link and Sauer 1998, Shriver et al. 2005), and the two methods are sometimes even considered equivalent and combined in single surveys (Gregory et al. 2004). Therefore, it is useful to compare both survey methods side by side in order to examine the similarity of data obtained under each protocol. Field-Site Description The study area surveyed is a hardwood forest of about 200 ha adjacent to the campus of Park University, in Parkville, MO (39.190°N, 94.667°W, WGS 84; Fig. 1). About 150 ha of this forest is owned by Park University, and another contiguous 50 ha is managed by the city of Parkville and the Missouri Department of Conservation. The terrain in this small hardwood forest is rolling, with riparian and upland portions and an elevation relief of about 75 m. Survey plots were placed in undisturbed riparian forest along creek channels. The study area is surrounded by a busy urban landscape with heavy freight-train traffic, a rock quarry, a small commercial area, a university campus, urban neighborhoods, and an airport nearby. The riparian forest overstory is dominated by Tilia americana L. (Basswood), but Celtis occidentalis L. (Hackberry), Ulmus americana L. (American Elm), Quercus muehlenbergii Engelm. (Chinquapin Oak), and Carya cordiformis Wangenh. (Bitternut Hickory) are also common. Ostrya virginiana Mill. (Eastern Hophornbeam) is a common midstory tree, and Asimina triloba L. (Pawpaw) is abundant in the understory. Materials and Methods Habitat description In order to compare results from breeding-bird surveys utilizing these two methods, it was necessary to examine the variation in features of the riparian forest habitat under consideration. Four 20-m x 20-m square macroplots J.F. Taulman 2013 Southeastern Naturalist Vol. 12, No. 3 460 (0.04 ha) were installed along the single strip transect in a systematic random method in 2009, and two more macroplots were placed in each of the additional transects in 2010. One macroplot was randomly placed in each of the four circular plots in 2009 and in four of the eight additional circular plots in 2010. At each macroplot, the following forest characteristics were described: ground cover, percent coverage of the plot by woody shrubs under 2 m high, understory horizontal vegetation density from ground level up to 3 m height, stem counts of trees 2.5–10 cm DBH, individual DBH measurements and identification of all trees >10 cm diameter (for a calculation of basal area), canopy cover, prism basal area (to compare with measured basal area), and height of the dominant tree on the plot. Ground cover was estimated visually for grass, leaves, down wood, rock, and bare ground/water, with a total coverage sum of 100%. Percent shrub cover was also estimated visually. Vegetative density was measured at four horizontal strata by estimating leafy coverage on a 0.5-m-square checkered board held at 10 m from plot center. An Figure 1. Bird survey areas, 2009–2011. The one shaded rectangle and 4 shaded circles represent the 2-ha transect and four 0.5-ha circular plots surveyed in 2009, 2010, and 2011. For the 2010 and 2011 seasons, two additional transects and eight new circular plots were surveyed in other similar riparian forest habitat, shown by the open figures. Six ha of forest area was surveyed in each set of three transect plots and 12 fixed-radius circular plots in 2010 and 2011. Positioning and dispersal of plots as shown was necessary to allow placement in suitable undisturbed riparian forest habitat within this small urban hardwood forest. 461 J.F. Taulman 2013 Southeastern Naturalist Vol. 12, No. 3 observer recorded the percentage of the board covered by vegetation while looking across horizontally at the board held with the center at heights of 0.25, 1, 2, and 3 m. The board was held at the four cardinal directions from plot center, and the four readings at each stratum were averaged. This is a modification of a method described by MacArthur and MacArthur (1961), and subsequently used by others (Thill and Koerth 2005), in which the board is moved away from the observer until it is 50% obscured by vegetation. The modified method used here, recording a variable percent coverage of the board at a standard distance, is much faster than the earlier method, and avoids the potential problem in a habitat with little vegetation of having to move the board to a great distance in order to achieve the needed 50% coverage (up to 50 m required in seedling stands studied by Thill and Koerth [2005]). This modified method also provides an easily interpreted description of vegetative density at the four understory strata that I have used effectively in previous research (Taulman 1999, Taulman and Smith 2004, Taulman et al. 1998). Canopy cover was measured using a spherical densiometer, a convex mirror with 24 grid squares held at waist level on which leafy canopy vegetation coverage was estimated. The mirror was read at the four cardinal directions from plot center, and the total squares covered were multiplied by 1.04 to provide a canopy-cover percentage estimate (24 x 4 x 1.04 ≈ 100). A 10-factor prism was used to estimate basal area from plot center. Height of the dominant tree was measured with a clinometer. Bird survey areas and procedures I was the only observer on all surveys in this study over all three years. In 2009, an 80-m x 250-m strip transect (2 ha) was defined along the creek in a ravine in the core of the Park University forest study area >100 m from any habitat edge (Fig. 1). Strip transects with these dimensions have been recommended and used by others for bird surveys in similar forest settings (Conner and Dickson 1980, Dickson et al. 1993, Thill and Koerth 2005). I walked down the center line of the strip transect, spending 32 min in passage, recording all birds seen and heard out to the transect boundary. Locations of birds were noted on the data sheet, and care was taken to ensure that birds seen or heard were not tallied twice. Some researchers have suggested an appropriate rate of about 1.0 km/hr for transit through a strip transect of similar size during a bird survey, or about 15 min for a transect of 250 m (Conner and Dickson 1980, Manuwal and Carey 1991). However, others have found it valuable to spend longer, from 20 min (Watson 2004) to about 30 min (Thill and Koerth 2005), to complete a strip transect survey of that size. The 32-min duration in this study also allowed equalization of observer effort for surveys of both the 2-ha strip transect and the 2 ha in the four fixed-radius point-count plots, which were surveyed for 8 min each. In 2009, four fixed-radius circular plots (40 m radius) were established along a creek in a ravine adjacent to the one containing the strip transect, with a ridge of about 30 m height separating the two ravines. All birds seen and heard within J.F. Taulman 2013 Southeastern Naturalist Vol. 12, No. 3 462 40 m of the center line of the strip transect, and within the 40-m-radius circular plot, were recorded (Conner and Dickson 1980, Gregory et al. 20 04). Distance to the 40-m plot or transect boundary was measured with a laser rangefinder (Bushnell Yardage Pro, Bushnell Corp., Overland Park, KS). Species known to be migrants or local residents but non-breeders in the forest, such as Vermivora peregrine W. (Tennessee warbler) and Branta canadensis L. (Canada Goose), respectively, were not recorded (Petit et al. 1994). Adjacent circular-plot centers were separated by at least 150 m and sampled for 8 min each from the plot center , as recommended by Manuwal and Carey (1991) and Tarvin et al. (1998). The 40-m maximum detection distance has also been used in similar forests by Dickson et al. (1993), Dickson et al. (1995), Petit et al. (1994), Petit et al. (1995), and Thill and Koerth (2005), and that detection distance is within the recommended range of 35–50 m of Conner and Dickson (1980) but less than the 50-m radius used by Rodewald and Smith (1998) and recommended by Ralph et al. (1993). However, Ralph et al. (1993) did suggest that the survey-plot radius can be reduced in densely vegetated or noisy forests to as little as 25 m. In support of a shorter detection distance, Alldredge et al. (2007) found that errors in estimating the distance of birds using auditory detection were highly variable, depending on the species and orientation of the bird with respect to the observer. They concluded that even trained observers were unable to accurately estimate distance of birds in the range of 65 to 86 m. Vegetation density in the forest often precluded visually observing birds at the boundary of the survey area, forcing reliance on hearing songs and calls. Alldredge et al. (2007) investigated the factors producing errors in judging distance to singing birds, such as orientation of the bird in relation to the observer and volume of the song in different species. In order to reduce the bias associated with misjudging distance in birds near the boundary of the transect or circular plot, I omitted every other song or sighting very near the boundary, assuming that as many of the distant birds were just outside the survey area as were just within it. I personally conducted all surveys in order to avoid problems with consistency of data among different observers (Buckland 2006; Conner and Dickson 1980; Johnson 1995, 2008; Manuwal and Carey 1991). More importantly, any observer bias, such as differences in detectability of species, was similar between the two survey methods under consideration. Because comparison of data produced by each method was the aim of the study, any bias inadvertently entering into data collection was applied consistently across methods and should not reduce the validity of comparisons of indices between methods. Conner and Dickson (1980) and Manuwal and Carey (1991) advised surveying at least 8 ha of forest in order to encompass enough habitat for a valid bird survey. For the 2010 and 2011 seasons, the survey area was enlarged as much as possible to cover the riparian habitat in this forest while maintaining required separation between circular plots in point-count arrays. This enlargement consisted of adding two new 2-ha transects and eight new 0.5-ha fixed-radius circular 463 J.F. Taulman 2013 Southeastern Naturalist Vol. 12, No. 3 survey areas, separated again by at least 150 m (Fig. 1). The 2009 areas were also included in the 2010 survey, producing a total of 6 ha surveyed in 3 strip transects and 6 ha in twelve 0.5-ha circular plots. The isolated urban forest in this study did not have sufficient riparian forest habitat to permit a larger survey area for additional sets of both transects and circular -plot arrays. I conducted bird surveys in 2009 from 17 May through 24 June between the times of 0630 and 1000, in 2010 from 18 May through 15 June between 0630 and 1100, and in 2011 from 1 June through 30 June between 0630 and 1030 (Conner and Dickson 1980, Dickson et al. 1993, Hutto et al. 1986, Rodewald and Smith 1998). Field activity was begun later in 2011 in order to better coincide with arrival of resident breeding birds and to avoid non-breeding migrants that had been detected during the early days of the 2009 and 2010 surveys. In 2009, I surveyed both the strip transect and the four fixed-radius plots once per day on 14 different visits, in order to compensate somewhat for the small forest area surveyed (Carey 1988) and to reduce within-treatment variation (Conner and Dickson 1980). I completed 10 replications in 2010 and 8 in 2011 for both the three transects as well as the twelve 0.5-ha circular study areas. Though time constraints did not allow the same number of repeated visits during all years, the same number of surveys was conducted and equal effort expended for each survey method within a given season. I visited both the transect and circular- plot array on each survey in 2009, and I visited two sets of s tudy areas each day during 2010 and 2011: two transects and eight associated circular plots. The sequence of surveys at the transects and circular-plot arrays was alternated, starting on the strip transect one morning and then beginning o n the fixed-radius plots the next time out. No surveys were conducted during rain or high-wind conditions (Manuwal and Carey 1991). Hutto et al. (1986) recommend multiple counts at a given study area (25 used in their landscape study), but with no replications at any particular survey point, and Gregory et al. (2004) suggested visiting plots no more than four times. However, many researchers have recommended and used 8–12 replications of surveys on each plot (Carey 1988; Conner and Dickson 1980; Dickson et al. 1993, 1995; Noss 1991; Tarvin et al. 1998; Thill and Koerth 2005). Ralph et al. (1993) suggested a single visit to each plot during a season for point counts, but advised 8–12 visits to survey areas in spot-mapping surveys where information on densities and distribution of territories in small patchy habitats is sought. The goal in the present study was to evaluate as fully as possible the different population indices that might be obtained by the strip-transect and circular- plot survey methods, primarily species richness and abundance. Multiple visits to each site allowed the possibility of encountering individuals in territories only partially overlapped by the survey plots to produce more accurate index of overall species abundance, after averaging counts per visit. Statistical analyses Density-board data and overstory-tree basal areas were compared separately among the macroplots within the transects and within the fixed-radius J.F. Taulman 2013 Southeastern Naturalist Vol. 12, No. 3 464 point-count areas in order to examine continuity of understory vegetative density and basal areas within each sampling area. Vegetative density and tree basal areas were compared between strip transect and fixed-radius macroplots. The Kruskal- Wallis ANOVA was used for comparisons within and between survey areas. The Mann-Whitney U test was used to compare all other habitat variables between the groups of strip-transect and fixed-radius macroplots. Because the area in an array of four fixed-radius point-count plots was the same as the area contained in a single fixed-width strip transect, bird data from the four fixed-radius circular plots in 2009 were combined to compare with the transect data. In 2010 and 2011, data from all 12 circular plots were combined each year for comparisons with the 3 strip transects. In bird surveys, a comparative index of species richness was described as the number of species detected in each survey area on a given day. Average bird abundance was analyzed using the Mann-Whitney U test. Bird abundance was considered to be the number of individuals of each species detected in each of the areas on a given visit, with the average calculated over all surveys providing an estimator of the population of a given species in that defined survey area. Where a survey plot overlapped a small portion of a bird’s territory, multiple visits increased the likelihood of encountering one or both members of that breeding pair in the part of their territory intersected by the survey plot. Frequency of detection of each species was also computed for each of the survey areas over the course of the 14 visits in 2009, 10 surveys in 2010, and 8 visits in 2011. Species richness and abundance estimates were compared with the Mann-Whitney U test. Frequency of detection, and numbers of birds in, each species detected on both the transect and circular-plot arrays (23 species in 2009, 29 in 2010, and 23 in 2011), were compared in a pairwise manner with the Wilcoxon Matched Pairs test. Shannon’s diversity indices (H') for both survey areas were compared using the Mann-Whitney U test. The Margalef’s index of community diversity and species evenness indices were also computed for the species detected on both the transect and circular-plot array survey areas (Carey et al. 1991, Magurran 2003, Stainfield 2009). An α = 0.05 significance level was used for all tests. Simulation exercise Because fixed-radius circular-plot centers were at least 150 m apart, the distance from the center point of the first to the fourth circular plot was at least 450 m. Though the actual area surveyed was the same in both the transects and circularplot arrays, the additional forest area within which an array of 4 circular plots was dispersed may have overlapped the territories of more birds of a given species than were sampled within the contiguous 2 ha of a strip-transect survey area. In order to examine the possible differences in the number of breeding-bird territories that could be intersected by 2-ha transects and arrays consisting of 4 circular plots of 0.5 ha separated by 150 m, I created a model of a landscape with a background simulating territories of pairs of breeding birds of a given species, and tested the ways in which transects and circular plot arrays could 465 J.F. Taulman 2013 Southeastern Naturalist Vol. 12, No. 3 overlap those territories. The area of a typical territory of a breeding bird pair was defined in the model as 5.86 ha, and represented by a circle 273 m in diameter (Fig. 2). The territory size used in this model is the mean for 16 breeding bird species found in this forest whose territory sizes were also reported by Hamel (1992) for similar forests in the southeastern US. Because pairs of breeding birds attempt to exclude others from their territories, territories in the model do not overlap one another. The first iteration of this model portrayed a landscape saturated with territories of a bird species, with each territory abutting another on 4 sides, assuming maximum utilization of the habitat. Onto this background, I overlaid a 20-m x 20-m grid, numbered each grid location, and then randomly placed scaled 80-m x 250-m (2 ha) strip transects into the gridded territory array. I also randomly assigned an azimuth to each transect (1°–360°) in each trial placement. The number of territories overlapped by each transect was recorded in 100 trial placements. I next placed into the bird-territory array four different configurations of four circular plots scaled to 0.5 ha each (40 m radius, 2 ha total) and separated by 150 m (Fig. 2). The four configurations of circular-plot arrays were tested in order to examine whether any one of a number of possible Figure 2. Model simulating the possible overlap of 2-ha strip transects and arrays of four 0.5-ha circular plots, in four typical configurations, on a background of breeding bird territories. Territory size for a pair of breeding birds in this model is scaled at 5.86 ha (273 m diameter), an average of the territory sizes for 16 species of breeding birds in southeastern forests of the United States, reported by Hamel (1992), which were also detected in surveys in this study. Different hatch patterns in circular plots correspond to the four different array configurations tested. J.F. Taulman 2013 Southeastern Naturalist Vol. 12, No. 3 466 circular-plot array distributions overlap more bird territories than any other (see Loehle et al. [2005] and Mitchell et al. [2006] for examples of circular-plot distribution along grid locations). I repeated insertions of each of the four circular- plot configurations 30 times (120 total trials), again assigning a random grid location and random azimuth each time. Assumptions in this test were that the researcher is not aware when installing a strip transect or circular-plot array how that survey area will intersect existing or future breeding bird territories, and that a random placement of scaled survey areas onto a simulated background of bird territories in a model can serve to investigate the phenomenon of actual territory overlap by survey areas in a natural setting. I tested the number of territories overlapped by the 4 circular-plot configurations with ANOVA to determine if one plot configuration overlapped more bird territories than another. I then compared the number of bird territories overlapped in the array by transects and circular plots using t he Student’s t-test. I next randomly removed 50% of the bird territories from the original background and repeated the random placement of 100 transects and 120 circular-plot arrays using the four different circular-plot arrays, simulating the possible overlap of territories by the two survey methods in a patchy landscape not fully saturated with territories. Finally, I repeated the simulation exercise by removing 75% of the territories from the original background array to compare the possible overlap of territories by the transects and circular-plot arrays in a habitat where bird territories were at a 25% density . Statistical analyses were performed using Statistica (Statsoft 2011). Results Habitat comparisons Horizontal vegetative density was not different at any of the four measured strata in comparisons between the 8 transect and 8 circular area macroplots combined from 2009 and 2010, or in comparisons between the 4 transect and 4 circular area macroplots in 2009. In the 2010 macroplots, the 0.25-m stratum density (% coverage of a board viewed at a distance of 10 m) was higher on circular plots than transect plots (circular plots: x = 34.7%, SE = 6.2; transects: x = 20.9, SE = 4.6; U = 74.5, P = 0.04). In comparisons of the 4 circular-area macroplots in 2009 with the 4 added circular area plots in 2010, the vegetation density at the 2-m stratum was greater on the 2010 macroplots (2009: x = 4.7%, SE = 1.9; 2010: x = 20.8%, SE = 6.4; U = 63, P = 0.01). Vegetative density at the 0.25-stratum was greater on the 2009 transect plots compared with the 2010 transect macroplots (2009: x = 34.7%, SE = 6.2; 2010: x = 20.9, SE = 4.6; U = 55, P = 0.006). Within-area plot comparisons to test the homogeneity of macroplots in transect and circular-plot forest patches showed no differences in horizontal vegetation density at any stratum in 2010. At the 3-m stratum among the four strip-transect macroplots in 2009, one macroplot had substantial leafy vegetation and the other three had very little (K-W H [3, n = 16], P = 0.019). On the 467 J.F. Taulman 2013 Southeastern Naturalist Vol. 12, No. 3 four fixed-radius circular-plot macroplots, horizontal vegetative density at both the 1-m and the 3-m strata was different among plots in 2009 (H [3, n = 16], P = 0.025 and P = 0.035, respectively). Basal areas of overstory trees were not different among the circular macroplots but were different among the 4 strip-transect macroplots in 2009 (H [3, n = 58], P = 0.024). However, basal area of overstory trees was not different between the groups of transect and circular survey macroplots in 2009, 2010, or in comparisons between the 8 transect plots and 8 circular plots combined for 2009 and 2010. Shannon’s diversity indices were not different for overstory tree species between the strip transect and circular area macroplots in 2009 or 2010 (2009: H'transect = 2.32, H'circular = 2.13, U = 86, P = 0.60, ntransect = 15, ncircular = 13; 2010: H'transect = 1.93, H'circular = 2.08, U = 59, P = 0.49, ntransect = 11, ncircular = 13 ). Evenness values for overstory tree species in the strip transect and circular area macroplots were similar (2009: 0.857 and 0.829, respectively; 2010: 0.806 and 0.812, respectively). All other vegetative variables showed no differences in comparisons between fixed-width strip-transect and fixed-radius point-count macroplots. Bird surveys In the 2009 survey, a total of 31 species of resident breeding birds were detected, 28 in the strip transect and 26 in the four fixed-radius circular plots. Twenty-three species were found in common in both survey areas. There was no difference in species richness, frequency of detection, or in diversity indices between the 2-ha transect and the circular-plot array. However, the mean number of individual birds detected during the 14 visits was greater in the fixed-radius circular plots than in the strip transect (transect: birds/ha = 10.79; circular array: birds/ha = 15.36; U = 10.5, P < 0.001) (Table 1). And the mean numbers of individuals detected of the 23 common species, considered pairwise, were also greater in the fixed-radius point-count plots (transect = 10.32, circular array = 14.93; z = 3.00, P = 0.003). In the 2010 survey of 6 ha in each of three 2-ha strip transects and twelve 0.5-ha circular-plot arrays, 36 bird species were encountered during 10 surveys. Thirty-one species were found on the three strip transects, 34 species on the twelve fixed-radius circular plots, and 29 species were common to both survey areas. Frequency of detection of the 29 common species was not different between strip transect and fixed-radius circular plots (z = 1.634, P = 0.10). Shannon’s diversity indices were similar between the transects and circular-plot arrays in 2010 (transect = 2.475, circular plots = 2.751; U = 520, P = 0.93). Both the Margalef’s index of community diversity (transect = 8.052, circular plots = 8.271) and evenness indices (transect = 0.721, circular plots = 0.780) were numerically similar between the two survey types. However, the mean numbers of individual birds detected in 2010 of the 29 common species were again greater on the circular-plot arrays compared with the transects (transect: birds/ha = 7.45, circular plot arrays: birds/ha = 9.68; z = 3.03, P = 0.002). Both species richness and abundance of birds of all species detected J.F. Taulman 2013 Southeastern Naturalist Vol. 12, No. 3 468 Table 1. Species richness and abundance by survey for strip transects and point-count circular-plot arrays during 2009, 2010, and 2011. Abundances are total individual birds detected per survey divided by the total surveyed area. Transects Circular Plots Survey Species richness Abundance Species richness Abundance 2009 1 17 12.50 13 14.00 2 16 10.50 12 12.50 3 16 15.00 16 17.00 4 9 7.50 13 13.00 5 15 12.00 15 17.00 6 13 11.50 14 15.00 7 10 8.50 12 14.50 8 14 12.0 15 15.50 9 12 8.00 13 12.00 10 12 9.50 12 14.50 11 11 11.0 16 21.00 12 10 9.50 12 18.50 13 15 12.00 12 16.50 14 14 11.50 11 14.00 Mean 13.1 10.791 13.3 15.361 SE 0.68 0.54 0.44 0.66 2010 1 16 7.67 24 10.50 2 20 9.67 16 7.33 3 17 7.33 23 12.00 4 17 7.33 18 10.50 5 17 8.50 19 11.00 6 15 6.50 18 10.67 7 16 7.00 21 9.00 8 15 7.83 18 9.00 9 18 7.33 20 9.83 10 19 8.67 23 12.00 Mean 17.02 7.783 20.02 10.183 SE 0.52 0.29 0.84 0.46 2011 1 16 6.67 12 6.33 2 14 6.33 17 6.17 3 18 7.50 14 5.83 4 16 7.00 15 7.00 5 12 4.83 15 6.50 6 16 8.00 18 8.00 7 16 8.17 14 6.33 8 15 5.67 17 8.67 Mean 15.4 6.77 15.3 6.85 SE 0.63 0.41 0.70 0.35 1In 2009, abundance was significantly greater on circular -plot arrays. 2In 2010, species richness was significantly greater on circular -plot arrays. 3In 2010, abundance was significantly greater on circular -plot arrays. 469 J.F. Taulman 2013 Southeastern Naturalist Vol. 12, No. 3 over the 10 surveys were greater on the fixed-radius circular plots (richness: x = 17.0/6 ha for transects, x = 20.0/6 ha for circular plots, U = 16.5, P = 0.012; abundance: x = 7.78 birds/ha for transects, x = 10.18 birds/ha for circular plots, U = 8.5, P = 0.002). Considering the 2010 survey data from only the original 2-ha transect and four circular plots surveyed in both 2009 and 2010 in order to eliminate an effect possibly produced by the addition of new survey areas in 2010, the 2-ha circularplot array in 2010 still produced indices of species richness and abundance that were significantly higher than those from the 2-ha transect (richness: U = 17.0, P = 0.013; abundance: U = 21.5, P = 0.031). Comparing 2009 and 2010 data to investigate an effect by year, only the 2-ha transect and four circular-plots array surveyed during both years were considered. Only the first 10 replications of the total of 14 visits from 2009 were included in the analysis to ensure an equal survey effort to compare with 2010 results. Abundance was significantly greater on the circular-plot array during 2009 compared with the circular-plot array in 2010 (U = 6.00, P = 0.001). Species richness was greater on the transect plot in 2009 compared with the same transect results in 2010 (U = 21.5, P = 0.034). In 2011, thirty-two species of breeding birds were recorded, 26 on the strip transects and 29 in the circular plot arrays. Twenty-three species were found on both survey areas. There were no differences in abundances of individuals detected between transect and circular plots, either considering all birds detected per survey (transect = 6.77, circular plots = 6.86, U = 31.0, P = 0.96) or in pairwise tests of species found on both transect and circular plot surveys (z = 0.21, P = 0.83, Nt and Nc = 23). Likewise, neither frequencies of species detection nor species richness over the 8 surveys were different in comparisons of transect and circular-plot data. Comparing transect surveys in 2010 with those in 2011 showed no differences in abundances or frequency of detection. Circular-plot data were also similar from 2010 to 2011. Simulation trials In simulation trials of the number of bird territories intersected by the 4 different configurations of arrays of four 0.5-ha circular plots, there were no differences in the number of bird territories overlapped by any of the four array shapes (F = 1.85, df = 3, P = 0.14, n = 30; Fig. 2). Therefore, results from tests on the four groups of 30 circular-plot trials in the model were combined to compare the average numbers of territories intersected by all 120 circular-plot array placements with results from the placement of 100 transects in the model. The mean bird territories intersected by the 100 transect placements (x̅ = 2.22, SE = 0.06) were significantly fewer than the mean number of territories overlapped by the 120 circular-plot arrays (x̅ = 3.23, SE = 0.07) in the simulation exercise (t = -10.45, df = 218, P < 0.001, nt = 100, nc = 120). The ratio of territories overlapped by transects compared with circular-plot arrays was 0.687. The abundance of individual birds in this forest apparently declined over the three years of this study, from an estimated 10–15 birds per ha in 2009 to less J.F. Taulman 2013 Southeastern Naturalist Vol. 12, No. 3 470 than 7 birds per ha in 2011 (Fig. 3). In order to investigate whether a lower saturation of territories on the landscape could alter the differences in territory overlap by plots representing the two survey methods, I conducted two more simulation exercises. Overlap of bird territories by transect and circular-plot arrays was examined at reduced background territory densities of 50% and 25%. I assumed that as bird abundance decreases in a forest, the density of territories of a given species will also decrease. Circular-plot arrays still showed a significantly greater overlap of territories than transects at both the 50% and 25% background territory density levels. At a 50% territory density, mean territories overlapped by transects was 1.03 (SE = 0.08, n = 100), and by circular-plot arrays was 1.58 (SE = 0.09, n = 120) (t = -4.6, P < 0.001; Fig. 4). At a background territory density of 25%, mean territories overlapped by transects was 0.51 (SE = 0.06, n = 100) and by circular plot arrays was 0.75 (SE = 0.07, N = 120) (t = -2.55, P = 0.01; Fig. 5). The ratios of territories overlapped by transects compared with circular-plot arrays remained consistent with that from the saturated landscape: 0.652 at 50% density and 0.680 at 25% density. Figure 3. Change in abundance of individuals of all bird species recorded during surveys in 2009, 2010, and 2011. Values are total individual birds detected per survey divided by the total surveyed area. 471 J.F. Taulman 2013 Southeastern Naturalist Vol. 12, No. 3 Discussion The surprising difference in results yielded by these two survey methods was the greater abundance of individual birds detected in the fixed-radius point-count arrays compared to the strip transect results in 2009 and 2010 (mean numbers of birds detected per ha in 2009: strip transects = 10.79, circular plots = 15.36; in 2010: strip transects = 7.78, circular plots = 10.18). In the 2010 survey, species richness estimates were also greater in the circular-plot arrays compared with the transect results. The working hypothesis that these two survey methods would produce similar estimates of bird abundance was disproved by these results in 2009 and 2010. Interestingly, in 2011, comparisons of abundance, species richness, and frequency of detection showed no significant differences between transect and circular-plot arrays, thus lending support to the hypothesis of similarity between survey methods during that year. The simulation model with a possible distribution of territories of pairs of a given breeding-bird species, and trials showing ways in which 2-ha strip transects and sets of four 0.5-ha circular-plot arrays might overlap those territories, illustrated the potential for a circular-plot array to intersect more bird territories Figure 4. Model simulating overlap of 2-ha strip transects and arrays of four 0.5-ha circular plots on a background of breeding bird territories at 50% saturation of territories on the landscape. Fifty percent of the original background territories were randomly selected and removed from the model. J.F. Taulman 2013 Southeastern Naturalist Vol. 12, No. 3 472 and effectively survey a larger forest patch than a transect of the same total area (Figs. 2, 4, 5). An assumption in these three simulation exercises was that a pair of breeding birds occupies each territory and attempts to exclude conspecifics from that area. Because each territory intersected by a survey plot provides the potential for an observer to record both the male and female of that breeding pair, if one survey method results in more territories overlapped than another method, that could translate into estimates of higher abundance and possibly higher species richness for the survey method that encompasses more terri tories. However, the 2011 results in this study illustrate that in any given instance the placement of strip transects and circular-plot arrays may overlap existing bird territories in a habitat in about equal proportions and produce similar field survey results. The potential effect of higher abundance estimates from the circular-plot array, as shown by the simulation exercise, does not necessarily occur in each field implementation of these survey methods. The simulated placement of transect and circular-plot arrays into a model landscape with a lower density distribution of breeding-bird territories examined the possibility that territory overlap of the two survey methods might be different in a landscape with lower bird density. The result indicated that the ratios of Figure 5. Model simulating overlap of 2-ha strip transects and arrays of four 0.5-ha circular plots on a background of breeding bird territories at 25% saturation of territories on the landscape. Seventy-five percent of the original background territories were randomly selected and removed from the model. 473 J.F. Taulman 2013 Southeastern Naturalist Vol. 12, No. 3 territories overlapped by transect and circular-plot arrays remain consistent even in landscapes with sparse territory density. In trial placements of both survey area types, circular-plot arrays still overlapped significantly more territories than the transects of equal area, even at 50% and 25% territory density levels. Interestingly, the ratio of the perimeter of an 80-m x 250-m strip transect (660 m) to the total perimeter of four 40-m-radius circular plots (1005.3 m) is 0.656, a ratio very close to the consistent ratio of territories overlapped by transect and circular-plot arrays in the simulation exercises. This similarity may just be coincidental, however. Were 4 circular plots not separated in space, but lined up so that the boundary of one met the next in line, three could be contained entirely within a single 80-m x 250-m transect and the fourth circular plot would extend 70 m into adjacent landscape outside the transect. While no simulation was undertaken to determine the difference in territories overlapped by this configuration of circular plots and transects, the number of territories intersected by the circular-plot array would certainly not be 35% greater than those intersected by the transect in that case. The dispersal of circular plots in an array, with 150 m or more between plot centers, is undoubtedly the primary reason for the greater number of territories intersected by circular plot arrays compared with transects over a large number of trials. While the size, shape, and distribution of breeding-bird territories in nature will not be exactly as depicted in the model used here, these simulation exercises provide one way to evaluate possible overlap of breeding-bird territories by transects and circular-plot survey arrays. There is a smaller chance of detecting one or both birds in a breeding territory when the transect or circular plot only partially overlaps a territory. However, this partial overlap of territories at the edge of the survey area applies to both the transect and point-count methods. The consistently larger number of territories intersected by a dispersed array of circular survey plots, as shown in all simulations, combined with a number of replicated visits, can be expected to produce greater abundance estimates compared with the transect method over time. In comparison to fixed-width strip transects of equal area, the larger forest area effectively surveyed by an array of fixed-radius circular plots, the possible overlap of more bird territories by a dispersed group of circular plots as indicated by simulation trial results, and the higher bird abundances seen during the years 2009 and 2010 in this study on the circular plot arrays, all suggest that these two survey methods may not be as similar as previously believed (Gregory et al. 2004, Ralph et al. 1993). Though a few of the measured habitat variables were different between macroplots within the groups in the strip-transect and fixed-radius areas, the lack of differences in comparisons of the majority of habitat variables between the transect and circular-plot survey areas and the similarity of overstory tree species richness and diversity lead me to conclude that these riparian forest patches were comparable for the purposes of the bird-community analysis undertaken here. J.F. Taulman 2013 Southeastern Naturalist Vol. 12, No. 3 474 Watson (2004) found that whole-patch searches were more efficient at estimating species richness than timed transect surveys, but he used stopping rules that allowed for many repeated area searches in order to arrive at a species richness estimate. He acknowledged that the standard 20–30 min effort normally spent on transect surveys was insufficient to detect a majority of species occurring in a forest. In this study, I conducted multiple replicated surveys in order to increase the probability of detection of less-commonly occurring species, those that had lower detectability, and those that might have arrived later in the breeding season to establish territories. Because bird territories may be viewed as “random events in space” (Johnson 1995:2) and a given transect or circular plot may incorporate a single territory or overlap portions of one or more territories of a given species, multiple surveys also allowed detection of individuals that might only occasionally be present in the part of their territory covered by the survey plot. Averaging numbers of individuals and species over the total number of visits reduced the additive effect of multiple surveys in the present study. Surveys of both fixed-width transects and fixed-radius circular plots suffer from biases due to differential detectability of bird species resulting from such factors as variability in observer skill levels; changing environmental parameters, such as temperature and wind velocity; and behavioral and physical characteristics of bird species that render some more or less conspicuous than others (Buckland 2006, Johnson 2008, Rosenstock et al. 2002). Johnson (2008) concluded that none of the current means of adjusting bird survey methods to mitigate these biases are uniformly effective in removing possible sources of error. He advised researchers to acknowledge possible shortcomings in survey methods to be employed and to attempt to control for those sources of error through appropriate study design. I have attempted to address the common sources of bias associated with the survey methods investigated in this study through control of sampling protocols. The maximum detection distance was 40 m for both survey methods, and I was the only observer, so any bias in species’ detectability was similar for each survey method under consideration. The forest habitat surveyed was similar for both methods. Surveys using each method were undertaken in similar environmental conditions and time of day. Moreover, results were compared between survey methods by species in a pairwise manner for species detected by both survey methods, further reducing the bias in detectability among different species. Buckland (2006) recommended omitting female birds from survey data because they are less detectable than males. However, in this study, a single observer conducting all surveys provided consistency in identification, multiple replications were conducted at each survey area, and pairwise comparisons were made of species richness, abundance, and frequency of detection between commonly observed species in each survey type. Therefore, any difference in detection between males and females was consistent in surveys undertaken by both methods, and with respect to all species, throughout the study. Because the goal of this study was to compare population indices produced by the two survey methods, not to validate the efficiency of either method at estimating true species 475 J.F. Taulman 2013 Southeastern Naturalist Vol. 12, No. 3 population parameters, a difference in male and female detectability did not bias the results of the comparison of methods in this study . The species richness found in the hardwood forest in this study (26 for transects, 29 for circular plots, and 32 total in 2011; 31 for transects, 34 for circular plots, and 36 total in 2010; 29 for transects, 26 for circular plots, and 32 total in 2009) compares favorably with the results of Dickson et al. (1995) for hardwooddominated wide streamside zones in Texas, where 32 resident bird species were detected. Thill and Koerth (2005) found between 17.4 and 24.7 species in uneven aged pine-hardwood forests in Texas, though over 50% of their reported species were migrants. Barber et al. (2001) detected 26 resident species on surveys of forests in Arkansas under a range of silvicultural conditions. The two survey methods under investigation in this study appeared equally effective at allowing detection of resident bird species (mean species per survey detected over 6 ha in 2011: strip transect = 15.38, circular-plot arrays = 15.38; over 6 ha in 2010: strip transect = 17.00, circular-plot arrays = 20.00; over 2 ha in 2009: strip transect = 13.21, circular-plot array = 13.29). Further side-by-side comparisons of these two methods in different forest types and larger landscapes would allow researchers to better evaluate their differences and applicability to specific research goals. Acknowledgments I am grateful to D.L. Williams, M. Tounzen, and D. Epperson for providing assistance in macroplot measurements. The assistance of S.L. Johnson, who detected a nesting hummingbird during a class field trip that was later recorded in a circular-plot survey, is greatly appreciated. I am very thankful for the generosity of Google Earth Pro in providing licenses and permissions to use imagery. The comments of five anonymous reviewers are also appreciated. Literature Cited Alldredge, M.W., T.R. Simons, and K.H. Pollock. 2007. A field evaluation of distancemeasurement error in auditory avian point-count surveys. Journal of Wildlife Management 71:2759–2766. Barber, D.R., T.E. Martin, M.A. Melchiors, R.E. Thill, and T.B. Wigley. 2001. Nesting success of birds in different silvicultural treatments in southeastern US pine forests. Conservation Biology 15:196–207. Buckland, S.T. 2006. Point-transect surveys for songbirds: Robust methodologies. Auk 123:345–357. Carey, A.B. 1988. The influence of small streams on the composition of upland bird communities. Pp. 153–162, In K.J. Raedeke (Ed.). Streamside Management: Riparian Wildlife and Forestry Interactions. University of Washington, Institute for Resources, Seattle, WA. Contribution No. 59. 277 pp. Carey, A.B., M.M. Hardt, S.P. Horton, and B.L. Biswell. 1991. Spring bird communities in the Oregon coast range. Pp. 123–144, In L.F. Ruggiero, K.B. Aubry, A.B. Carey, and M.H. Huff (Tech. Coords). Wildlife and vegetation of unmanaged Douglas-Fir forests. General Technical Report PNW-285. USDA Forest Service Pacific Northwest Research Station, Portland, OR. 533 pp. J.F. Taulman 2013 Southeastern Naturalist Vol. 12, No. 3 476 Conner, R.N., and J.G. Dickson. 1980. Strip-transect sampling and analysis for avian habitat studies. Wildlife Society Bulletin 8:4–10. Dickson, J.G., R.N. Conner, and J.H. Williamson. 1993. Neotropical migratory bird communities in a developing pine plantation. Proceedings of the Annual Conference of the Southeastern Association of Fish and Wildlife Agencies 47:439–446. Dickson, J.G., J.H. Williamson, R.N. Conner, and B. Ortego. 1995. Streamside zones and breeding birds in eastern Texas. Wildlife Society Bulletin 23:750–755. Gregory, R.D., D.W. Gibbons, and P.F. Donald. 2004. Bird census and survey techniques. Pp. 17–55, In W.J. Sutherland, I. Newton, and R.E. Green (Eds.). Bird Ecology and Conservation: A Handbook of Techniques. Oxford University Press, New York, NY. 408 pp. Hamel, P.B. 1992. Land Manager’s Guide to Birds of the South. The Nature Conservancy, Southeastern Region, NC. 437 pp. Hostetler, M.E., and M.B. Main. 2011. Florida monitoring program: Transect method for surveying birds. Available online at http://edis.ifas.ufl.edu/uw164. Accessed 2 August 2012. Hutto, R.L., S.M. Pletschet, and P. Hendricks. 1986. A fixed-radius point-count method for nonbreeding and breeding-season use. Auk 103:593–602. Johnson, D.H. 1995. Point counts of birds: What are we estimating? Pp. 117–123, In C.J. Ralph, J.R. Sauer, and S. Droege (Tech. Eds.). Monitoring bird populations by point counts. General Technical Report PNW-149. USDA Forest Service, Pacific Northwest Experiment Station, Portland, OR. 181 pp. Johnson, D.H. 2008. In defense of indices: The case of bird surveys. Journal of Wildlife Management 72:857–868. Link, W.A., and J.R. Sauer. 1998. Estimating population change from count data: Application to the North American breeding bird survey. Ecological Applications 8:258–268. Loehle, C., T.B. Wigley, S. Rutzmoser, J.A. Gerwin, P.D. Keyser, R.A. Lancia, C.J. Reynolds, R.E. Thill, R. Weih, D. White, Jr., and P.B. Wood. 2005. Managed forest landscape structure and avian species richness in the southeastern US. Forest Ecology and Management 214:279–293. MacArthur, R.H., and J.W. MacArthur. 1961. On bird species diversity. Ecology 42:594–598. Magurran, A.E. 2003. Measuring biological diversity. Blackwell Publishing Co., Malden, MA. 264 pp. Manuwal, D.A., and A.B. Carey. 1991. Methods for measuring populations of small diurnal forest birds. Pp. 1–23, In A.B. Carey and L.F. Ruggiero (Eds.). Wildlife-habitat Relationships: Sampling Procedures for Pacific Northwest Vertebrates. General Technical Report PNW-278. USDA Forest Service, Pacific Northwest Experiment Station, Portland, OR. Mitchell, M.S., S.H. Rutzmoser, T.B. Wigley, C. Loehle, J.A. Gerwin, P.D. Keyser, R.A. Lancia, R.W. Perry, C.J. Reynolds, R.E. Thill, R. Weih, D. White, and P.B. Wood. 2006. Relationships between avian richness and landscape structure at multiple scales using multiple landscapes. Forest Ecology and Management 221:15 5–169. Noss, R.F. 1991. Effects of edge and internal patchiness on avian habitat use in an oldgrowth Florida hammock. Natural Areas Journal 11:34–47. Pagen, R.W., F.R. Thompson III, and D.E. Burhans. 2000. Breeding and post-breeding habitat use by forest migrant songbirds in the Missouri Ozarks. Condor 102:738–747. 477 J.F. Taulman 2013 Southeastern Naturalist Vol. 12, No. 3 Petit, D.R., L.J. Petit, T.E. Martin, R.E. Thill, and J.F. Taulman. 1994. Breeding birds of late-rotation pine-hardwood stands: Community characteristics and similarity to other regional forests. Pp. 103–116, In J.B. Baker (Compiler). Ecosystem Management Research in the Ouachita Mountains: Pretreatment Conditions and Preliminary Findings. General Technical Report SO-112. USDA Forest Service, Southern Research Station, New Orleans, LA. 259 pp . Petit, D.R, L.J. Petit, V.A. Saab, and T.E. Martin. 1995. Fixed-radius point counts in forests: Factors influencing effectiveness and efficiency. Pp. 49–56, In C.J. Ralph, J.R. Sauer, and S. Droege (Tech. Eds.). Monitoring Bird Populations by Point Counts. General Technical Report PSW-149. USDA Forest Service, Southwest Research Station, Albany, CA. 181 pp. Ralph, C.J., G.R. Geupel, P. Pyle, T.E. Martin, and D.F. DeSante. 1993. Handbook of field methods for monitoring landbirds. General Technical Report PSW-144. USDA Forest Service, Southwest Research Station, Albany, CA. 41 pp. Rodewald, P.G., and K.G. Smith. 1998. Short-term effects of understory and overstory management on breeding birds in Arkansas oak-hickory forests. Journal of Wildlife Management 62:1411–1417. Rosenstock, S.S., D.R. Anderson, K.M. Giesen, T. Leukering, and M.F. Carter. 2002. Landbird counting techniques: Current practices and an alternat ive. Auk 119:46–53. Shriver, W.G., A.L. Jones, P.D. Vickery, A. Weik, and J. Wells. 2005. The distribution and abundance of obligate grassland birds breeding in New England and New York. Pp. 511–518, In C.J. Ralph and D. Rich (Eds.). Bird Conservation Implementation and Integration in the Americas: Proceedings of the Third International Partners in Flight Conference. 2002 March 20–24; Asilomar, CA. General Technical Report PSW–191. USDA Forest Service, Southwest Research Station, Albany, CA. 651 pp. Stainfield, J. 2009. Margalef’s index of community diversity. University of Plymouth, Department of Geographical Sciences, Plymouth, UK. Available online at http://www. gees.ac.uk/Projects/outputs/fieldsim/con_margalef.htm. Accessed 1 August 2012. Statsoft, Inc. 2011. Statistica ™. Statsoft, Inc., Tulsa, OK. Tarvin, K.A., M.C. Garvin, J.M. Jawor, and K.A. Dayer. 1998. A field evaluation of techniques used to estimate density of Blue Jays. Journal of Field Ornithology 69:209–222. Taulman, J.F. 1999. Selection of nest trees by Southern Flying Squirrels (Glaucomys volans) in Arkansas. Journal of Zoology 248:369–377. Taulman, J.F., K.G. Smith, and R.E. Thill. 1998. Demographic and behavioral responses of Southern Flying Squirrels to experimental harvests in Arkansas. Ecological Applications 8:1144–1155. Taulman, J.F., and K.G. Smith. 2004. Home range and habitat selection of Southern Flying Squirrels in fragmented forests. Mammalian Biology 68:1–17. Thill, R.E., and N.E. Koerth. 2005. Breeding birds of even- and uneven-aged pine forests of eastern Texas. Southeastern Naturalist 4:153–176. Verner, J., and L.V. Ritter. 1985. A comparison of transects and point counts in oak-pine woodlands of California. Condor 87:47–68. Watson, D.M. 2004. Comparative evaluation of new approaches to survey birds. Wildlife Research 31:1–11