Southeastern Naturalist 757 D.B. McNair 22001155 SOUTHEASTERN NATURALIST 1V4o(4l.) :1745,7 N–7o7. 04 Breeding Distribution and Population Persistence of Loggerhead Shrikes in a Portion of the North Carolina Sandhills Douglas B. McNair* Abstract - This study provides the first comparison of historical (1979–1994) and recent (2012–2013) breeding distributions of Lanius ludovicianus (Loggerhead Shrike) at a single location in southeastern North America. I conducted roadside and area searches for Loggerhead Shrikes in the spring (April–early June 2012 and late March–early May 2013) in Richmond County, NC, and small areas of 4 adjacent counties. Most of the study area is within the Sandhills, a subregion of the Coastal Plain. I documented a total of 44–45 breeding territories in 2012–2013, of which 36–38 (82–84%) were in the Sandhills, where I estimated the recent breeding density to be 3.2–3.4 pairs/100 km2. In both years, a smaller percentage of confirmed historic breeding sites were occupied in the Piedmont compared to the Sandhills portion of the study area, and the Piedmont had a much lower breeding density (0.6–0.8 pairs/100 km2). Sixteen of 22 (73%) confirmed historical breeding sites were occupied (n = 10) or retained suitable breeding habitat (n = 6) for Loggerhead Shrikes in 2012–2013. The breeding population in the sampled portion of the Sandhills has apparently declined slowly yet is still fairly stable, an unexpected result based on sharp declines of Loggerhead Shrikes in the Atlantic Coastal Plain for over the past 40 years as documented by coarse-scale surveys (breeding bird surveys, Christmas bird counts, spring bird counts). This study reaffirms the importance of conducting fine-scale surveys to produce a precise estimate of a persistent population in a geographically restric ted area. Introduction The Sandhills is a subregion of rolling, sandy hills in the Inner Coastal Plain of the Carolinas. They are ancient beach dunes that generally divide the Piedmont from the Coastal Plain. Endangered and threatened avian species that occupy managed mature Pinus palustris Mill. (Longleaf Pine) forest in the Sandhills, such as Picoides borealis Vieillot (Red-cockaded Woodpecker) and Peucaea aestivalis Lichtenstein (Bachman’s Sparrow) have received the bulk of research attention (Brust et al. 2004; Carter et al. 1983; Kesler and Walters 2012; J. Marcus, The Nature Conservancy, Southern Pines, NC, pers. comm.). Few other avian species, other than gamebirds and recently Falco sparverius L. (American Kestrel; Anchor and Brust 2004, Miller et al. 2012), have received attention despite the importance of agricultural land that comprises slightly over 1/5 of the generally forested area of Richmond County, NC. This agricultural land is especially rich in landbirds, including migrant and wintering Pooecetes gramineus Gmelin (Vesper Sparrow) and Lanius ludovicianus L. (Loggerhead Shrike), both of which are listed as special-concern species in North Carolina (NCWRC 2008). Loggerhead *35 Rowell Road, Wellfleet, MA 02667; dbmcnair@gmail.com. Manuscript Editor: Jason Davis Southeastern Naturalist D.B. McNair 2015 Vol. 14, No. 4 758 Shrikes in North Carolina are considered permanent residents, perhaps supplemented by small numbers of migrant birds during winter (LeGrand et al. 2011). In South Carolina, abundance does not appear to vary between seasons, suggesting that the area does not receive a large influx of wintering migrants (DeMent et al. 2008, Gawlik and Bildstein 1993). Studies on the breeding distribution and ecology of Loggerhead Shrikes have been conducted in many areas of North America (Brooks and Temple 1990, Michaels and Cully 1998, Sauer et al. 1995, Yosef 1996) where Loggerhead Shrikes have sharply declined for over the past 40 years based on breeding bird survey (BBS), Christmas bird count (CBC), and spring bird count (SBC) data (Cade and Woods 1997; Lowe and Butcher 1990; Morrison 1981; Sauer et al. 1995, 2011; Westphal 2009; Yosef 1996). North Carolina now represents the northeastern limit of the Loggerhead Shrike’s breeding range in the Atlantic coastal states of North America, except for some increasingly isolated areas in Virginia and a few patches further north. Recent BBS data show that this index of current (2006–2010) Loggerhead Shrike abundance in North Carolina is greatest in the south-central Inner Coastal Plain (1–3 birds/route; Sauer et al. 2011), which includes the Sandhills subregion. These coarse-scale data generally agree with recent anecdotal assessments (LeGrand 2011) and SBC data collected by the Carolina Bird Club (e.g., Westphal 2009) which suggest that Loggerhead Shrikes have sharply declined in the North Carolina Piedmont, but that the population decline in the Coastal Plain may have leveled off during the past decade or so, with an area of local concentration in Southern Pines, Moore County, in the North Carolina Sandhills. Reproductive success has generally been high in the southeastern US and some other areas (Gawlik and Bildstein 1990; Luukkonen 1987; Yosef 1996, 2001; though see Collister and Wilson 2007 for information on reduced breeding success), so declines have generally been attributed to habitat loss through land-use changes, including agricultural intensification (Gawlik and Bildstein 1990, Yosef 1996). Gawlik and Bildstein (1990, 1993) studied Loggerhead Shrikes in the northcentral Piedmont of South Carolina, ~120 km west of the North Carolina Sandhills, but no studies in North Carolina have defined the precise breeding distribution and status of Loggerhead Shrikes in any region, county, or other defined geographic unit despite the legal status of this species and its importance as a bioindicator species of healthy diverse agricultural landscapes and sensitivity to processes of agricultural intensification (Gawlik and Bildstein 1990, Yosef 1996). Richmond County (adjacent to Moore County) has had intensively managed agriculture—primarily cotton farming—over a long period from the end of the American Civil War until the mid- 20th century. The human population of generally rural Richmond County, which lies mainly in the Sandhills, has increased only slightly (0.6%) from 1980 to 2014 (2014 estimate = 45,733; 37 individuals/km2; US Census Bureau 2015), although the total housing units (a better measure of urbanization; Brown et al. 2005) increased 22% from 1980 to 2011 (US Census Bureau 2015). From 1978 to 2012, farm acreage declined from 26,839 to 19,252 ha (reduction of 28%), with the mean size and number of farms each decreasing ~15% (US Department of Agriculture 2015), yet Southeastern Naturalist 759 D.B. McNair 2015 Vol. 14, No. 4 all types of pastureland declined more steeply, from 4505 to 2294 ha (loss of 2111 ha; 49%). Other grass- or forb-dominated agricultural habitats in the form of idle cropland, cropland used for cover crops, cropland used for soil improvement, and lands in conservation reserve or conservation reserve enhancement programs also declined from 2111 to 1567 ha (loss of 544 ha; 26%). All of these habitats are generally favorable for Loggerhead Shrikes (Smith and Kruse 1992), yet pastureland is the most important habitat component for shrikes in the Atlantic coastal states of southeastern North America, especially during the breeding season, and the loss of pasturelands is correlated with a sharp decline in indices of shrike abundance (Gawlik and Bildstein 1993). Loggerhead Shrikes in Richmond County have most often been reported in agricultural landscapes with some pasturelands centered around large residential yards where pairs have nested in coniferous or deciduous trees and shrubs. Yosef (1996) and Tryjanowski et al. (2003) writing about the Loggerhead Shrike and European shrikes, respectively, cautioned that it is difficult to compare historical and current distributions because the number of observers and level of effort expended to survey shrikes have increased dramatically. However, a careful assessment of historical data within the current context can provide some measure of change in Loggerhead Shrike breeding distribution and population persistence in the agricultural landscape of a portion of the North Carolina Sandhills. Based on conversion of agricultural land—including a decline in grass- or forb-dominated habitats in Richmond County—to industrial forest-land uses since 1950 (Brown et al. 2005; D.B. McNair, pers. observ.; US Department of Agriculture 2015), a limited negative influence of increasing urbanization, a sharp decline of Loggerhead Shrikes in the Piedmont compared to a less pronounced decline in the Coastal Plain of the Carolinas, and historical data in Richmond County from 1979–1994 as a population baseline (see results), I expected Loggerhead Shrikes in 2012–2013 to be uncommon to locally fairly common in the Sandhills and rare to locally uncommon in the Piedmont and near the Fall Line (contact zone of the Piedmont with the Sandhills and Coastal Plain). I also expected Loggerhead Shrikes to be most numerous in agricultural areas and less numerous in other land-use types including more-natural habitats such as managed Longleaf Pine forest, even though pine savannas constitute ancestral habitat for Loggerhead Shrikes in southeastern North America (Cade and Woods 1997). Methods Field-site description The study area was comprised of Richmond County, NC, and small areas of 4 adjacent counties (North Carolina: Anson, Montgomery, Scotland; South Carolina: Marlboro) for a total of 1772 km2. Moving clockwise around Richmond County (Puetz 1990a), the study area in eastern Anson County extended from the South Carolina line at County Road 1832 north to Morven, northeast along NC 145 to US 74, west to Lilesville, north on County Road 1704, west on County Road 1634 to NC 109, and north to the Great Pee Dee River at the Richmond County line. The survey area Southeastern Naturalist D.B. McNair 2015 Vol. 14, No. 4 760 in extreme southeastern Montgomery County extended from the Richmond County line along US 220 to County Road 1524, and east to Drowning Creek at the Moore County line until the creek reaches the Richmond County line. The study area in western Scotland County extended from Drowning Creek at the Moore County line south and west along US 15/501, encompassing Camp MacKall Military Reservation and crossing the Sandhills Gamelands, west along the southern boundary of the gamelands to County Road 1346, south along 1346 and County Road 1001 to Laurel Hill, west and south along County Roads 1152 and 1145 and NC 381 to Gibson, and west along NC 79 to the Marlboro County line. In extreme northern and northwestern Marlboro County (Puetz 1990b), the study area extended from the Scotland County line at County Road 714; generally west along County Roads 166, 257, and 258 to the Richmond County line; south along SC 177 and west along County Roads 280, 37, and 630; south along County Roads 113 and 97 to US 1; and south and west to the Great Pee Dee River at the Chesterfield County line. The only location within the study area in 2012–2013 (not during the historical period) where access was completely restricted centered on barracks where basic training was conducted at Camp MacKall Military Reservation. Access to most of the grounds outside Morrison Correctional Institution in 2012–2013 was partially restricted; it housed adult medium-security inmates (absent during the historical period). The study area is dominated by the Sandhills subregion (1130.4 km2) of the Coastal Plain, but also includes a small area of the Inner Coastal Plain (12.4 km2) in northwestern Scotland County (centered on and near Gibson), and a larger area of the Piedmont (628.9 km2) in northwestern Richmond and eastern Anson counties. The Fall Line generally separates the Piedmont from the Coastal Plain, but transitional zones between the Piedmont and the Sandhills and Coastal Plain are commonly irregular and broad—as much as 5 km wide (US Department of Agriculture 1999). Within the study area, this contact zone especially occurs near the Great Pee Dee River in southwestern Richmond, southeastern Anson, and northwestern Marlboro counties, respectively. For this study, I considered this contact zone to be areas along and near the Fall Line where reddish or reddish-tinted soils occur with the Piedmont. 1979–1994 surveys I collected anecdotal data on the breeding distribution and breeding status (e.g., confirmed breeding evidence) of Loggerhead Shrikes within the study area defined above for the period 1979 to 1994 (parts of 13 years represented). I considered Loggerhead Shrikes to have been at potential breeding sites if adult birds were present from March to August; I also included records from February if I had direct breeding evidence. I classified Loggerhead Shrike territories as confirmed breeding sites if I detected nests or adults with recently fledged juveniles. I considered single males and/or females or pairs that occupied a site in at least 2 y or for a minimum of 20 d in 1 y to be on breeding territories. I treated all other sites as uncertain breeding locations. I recorded the physiographic region, county, nearest town, and precise site (e.g., section of county road) in which the breeding-season location occurred. Using Google Earth 2015 (version 7.1.2), I examined aerial photographs Southeastern Naturalist 761 D.B. McNair 2015 Vol. 14, No. 4 from 1993 to 2013 and made repeated site visits to assess the potential suitability of available habitat at confirmed historical breeding sites without documented presence of Loggerhead Shrikes in 2012–2013. These data (1979–1994) are henceforth called historical data. 2012–2013 surveys Tryjanowski et al. (2003) emphasized that use of 3 different survey methods could severely underestimate detection of shrike species in Poland compared to their most sophisticated method, which included nest searches and color banding that estimated precise shrike densities. Pasinelli et al. (2011), however, demonstrated that a variety of shrike survey methods, even for less-well-studied populations (e.g., Lübcke 2007 cited in Pasinelli et al. 2011), still had negligible sampling variance and provided precise density estimates. Kéry and Schmidt (2008) estimated detection probability of Lanius collurio L. (Red-backed Shrike) in 1-km quadrats to be high, with 2.7 surveys per season yielding a 95% probability of detecting this species, if present. Following Kéry and Schmidt (2008) and Pasinelli et al. (2011), I used strict criteria in 2012–2013 to ensure that Loggerhead Shrike counts would have a high level of detection and be precise. All Loggerhead Shrikes were unmarked so I could not measure site fidelity, yet presence in the same breeding locations between 2 divergent periods can provide information to help understand the cues used by Loggerhead Shrikes for habitat and land use. In 2012–2013, I conducted a systematic survey of Loggerhead Shrike breeding distribution in the same study area, focusing on the Sandhills subregion where I expected most shrikes to be present. I included all historical breeding locations in my 2012–2013 surveys. Consistent coverage and effort was comprehensive in both years, even though I deemphasized collection of direct breeding evidence because of time constraints imposed by covering such a large area. I used roadside surveys supplemented by area searches to reduce roadside-survey biases of under-detection of shrikes. Loggerhead Shrikes are slightly larger yet similar to Lanius collurio L. (Red-backed Shrikes), with males especially using conspicuous perches on exposed parts of hedges, trees, and shrubs in their territories during the breeding season. In the North Carolina Sandhills, overhead utility wires along roadsides (Gawlik and Bildstein 1993), fences, and other artificial perch sites preferred by Loggerhead Shrikes were numerous, and provided additional, conspicuous perch sites. I conducted surveys along all primary (federal and state highways), secondary (paved and traversable unpaved county and town roads), and accessible tertiary roads (other traversable unpaved public roads and private roads) within the study area that crossed potential Loggerhead Shrike breeding habitats. Except on the busiest highways, I generally drove 30 km/h or less to locate birds, stopping frequently; these surveys included frequent trips on private tertiary roads (e.g., on farms which were generally accessible). I remained a minimum of 15 minutes at each potential site on each visit. To increase detection rates, I walked all areas that could not be reached by road to within 200 m of potential Loggerhead Shrike habitats. I mapped all surveyed sites using GPS. Southeastern Naturalist D.B. McNair 2015 Vol. 14, No. 4 762 In 2012, I surveyed Loggerhead Shrikes on 33 days from 7 April to 4 June, a 2-month period that extended from pair formation to fledging of first young and the beginning of renesting attempts (D.B. McNair, unpubl. data; Yosef 1996). In 2013, I surveyed for Loggerhead Shrikes on 29 days from 31 March to 2 May, a compressed 5-week period that avoided most renesting attempts. I classified territories as occupied on the basis of breeding behavior as well as nest detection. However, except for finding nests and adults with fledged young, the criteria for designating an occupied breeding territory in 2012–2013 were very strict compared to the historical period. I visited all historical breeding locations of Loggerhead Shrikes a minimum of 4 times each over a minimum of 20 days during the peak-breeding period to document whether shrikes still occurred at these historic sites. Using the same methods, I also surveyed all other areas of potential habitat to determine if Loggerhead Shrikes occupied “new” (i.e., previously unrecorded) breeding locations in 2012–2013. Unless I found a nest or adults with fledged young, a minimum of 3 registrations over a minimum of 20 days at each site were required to qualify as an occupied shrike breeding territory. I classified Loggerhead Shrikes at locations that did not meet these criteria as visitors, or uncertain breeders not on occupied territories. Search effort during surveys in both years was almost identical (2012: 3510 miles, 199.5 h; 2013: 3490 miles, 200.5 h). I recorded the dominant land-use type within a 500-m radius of a nest or center of shrike activity for each Loggerhead Shrike location. Gawlik and Bildstein (1993) correlated indices of Loggerhead Shrike abundance with the percentage of available pastureland as the most important habitat component in the Piedmont of South Carolina; they noted that this relationship has also been found to apply in many other areas and in many states. Pastures were frequent components of the agricultural- habitat matrix of many Loggerhead Shrike breeding territories in the North Carolina Sandhills, and could also be a component of other land-use types. In this exploratory effort, I simply categorized the dominant land-use type, irrespective of whether or not the sites contained pastures or pasture-like components as part of the habitat matrix. Data analysis I used chi-square tests to analyze and compare nominal data on the breeding distribution of Loggerhead Shrikes among geographic regions between the 2 divergent time periods. I computed a two-tailed t-test, assuming unequal variances, to analyze differences in detection rates of Loggerhead Shrikes on breeding territories between the 2 recent years based on the first detection of presence at each site in 2012–2013. I compared these results with those of Kéry and Schmidt (2008) for Red-backed Shrikes. Finally, I used Mann-Whitney U-tests to examine data from uncertain breeding-season territories or visitation sites in 2012–2013 to determine the minimum length of stay (days) and the date of first occurrence at sites visited in one or both years compared to sites visited in one year but used as breeding territories or confirmed breeding sites in the alternative year. Statistical significance was set at P = 0.05. Otherwise, I present descriptive statistics. Southeastern Naturalist 763 D.B. McNair 2015 Vol. 14, No. 4 Results 1979–1994 Data I documented a total of 58 historical Loggerhead Shrike locations in the study area—43 (74%) in the Sandhills, 14 (24%) in the Piedmont and along or near the Fall Line, and 1 in the Inner Coastal Plain just below the Sandhills. Forty-one (71%) of these 58 locations were in Richmond County, and 36 (88%) of those were in the Sandhills. Breeding territories. I documented breeding territories at 28 (48%) of 58 historical locations in 2012–2013; 22 (38%) of the locations were confirmed breeding sites (nests or recently fledged young; Fig. 1). Seven (32%) of the 22 confirmed breeding locations occurred along or near the Fall Line, the other 15 (68%) were in the Sandhills. Twelve of 22 (55%) confirmed historical breeding sites were not used during 2012–2013, even though 6 of these 12 sites (50%) retained suitable breeding habitat. In contrast, the remaining 10 confirmed historical breeding sites (45%) were used recently (9 of 10 in both years; 1 in 2012 only). Two of the 7 confirmed historical breeding sites along or near the Fall Line (29%) were active in 2012–2013, whereas 8 of the 15 confirmed historical sites in the Sandhills (53%) were active (χ2 = 0.39, df = 1, P > 0.05). Thus, a total of 16 of 22 (73%) confirmed historical breeding sites were occupied or retained suitable breeding habitat for Loggerhead Shrikes in 2012–2013. The loss of suitable breeding habitat at 6 confirmed historical sites (27%) occurred in agricultural (4 sites) and non-agricultural landscapes (2 sites). Uncertain breeding-season territories. The remaining 30 (52%) historical locations constituted uncertain breeding season territories or visi tation sites. 2012–2013 Data I documented a total of 67 Loggerhead Shrike locations in the study area in 2012–2013. Fifty-six (84%) locations were in the Sandhills, 8 (12%) were in the Piedmont and along or near the Fall Line, and 3 (4%) were in the Inner Coastal Plain just below the Sandhills (Fig. 1); 44 locations were used both years, whereas 9 of 53 (17%) and 14 of 58 (24%) locations were used only in 2012 and 2013, respectively. Most birds occurred in 3 core areas centered on northeastern Richmond County (Ellerbe–Norman), east-central Richmond County (Hoffman), and southeastern Richmond County (south of Rockingham and Hamlet) into adjacent Scotland County (to Gibson). Fifty-one (76%) of these 67 locations were in Richmond County, and 48 (72%) were in the Sandhills. The proportion of locations in the Sandhills compared to locations in the Piedmont (including along and near the Fall Line) and the Inner Coastal Plain was not significantly different between the 2 survey periods (2012–2013 vs 1979–1994; χ2 = 1.1, df = 1, P > 0.05). In 2012– 2013, 22 (33%) Loggerhead Shrike locations were used during both time periods (17 of 51 in Richmond County; 33%). Breeding territories. Forty-five (85%) of 53 and 44 of 58 (76%) Loggerhead Shrike locations in 2012 and 2013, respectively, represented birds on breeding territories; 32 of 57 (56%) breeding territories were used both years, whereas 13 Southeastern Naturalist D.B. McNair 2015 Vol. 14, No. 4 764 Figure 1. Location of 22 historical (1979–1994) confirmed breeding sites and 67 recent (2012–2013) breeding-season locations of the Loggerhead Shrike in Richmond County, NC (and small areas of 4 adjacent counties). Large, open circles refer to historical confirmed breeding sites; small, filled squares represent 19 recent confirmed breeding sites in 1 or both years; filled diamonds show 38 other recent breeding territories in 1 or both years, excluding confirmed breeding sites, and filled Xs indicate 10 recent uncertain breeding-season territories or visitation sites in 1 or both years, excluding confirmed breeding sites and other recent breeding territories. Southeastern Naturalist 765 D.B. McNair 2015 Vol. 14, No. 4 and 12 breeding territories were used only in 2012 and 2013, respectively. I confirmed breeding at a total of 19 (33%) of these 57 territories (Fig. 1), of which 15 (26%) were in Richmond County. During the 2012–2013 survey period for the 44–45 breeding territories, the estimated Loggerhead Shrike population density in the study area was 2.5 pairs/100 km2. Loggerhead Shrike density was 0.6–0.8 pairs/100 km2 and 3.2–3.4 pairs/100 km2 for the Piedmont/Fall Line area and the Sandhills, respectively. Loggerhead Shrikes on breeding territories were first detected during my initial surveys at 40 of the 45 shrike locations (89%) in 2012 and at 41 of 44 locations (93%) in 2013; shrikes were first detected during the 4th visit of surveys at the 5 remaining breeding territories in 2012 and twice during the 2nd visit and once during the 3rd visit at the 3 remaining breeding territories in 2013. Three of the 5 latter sites in 2012 were also historical sites (1 was also used in 2013); one of these had partially restricted access in 2012 (Morrison Correctional Institution). The mean detection rate for 1st detection of presence was slightly higher (more efficient) in 2013 compared to 2012, but the difference was not significant (2013: 1.09 visits ± 0.05 SE, 95% CI = 0.98–1.20, n = 44; 2012: 1.33 visits ± 0.14 SE, 95% CI = 1.05–1.61, n = 45; t = 1.59, P = 0.12). I detected 5 (9%), 42 (74%), and 10 (18%) Loggerhead Shrike breeding territories along primary, secondary, and tertiary roads (including supplemental areasearches that departed from these roads), respectively. Thirty-nine (68%) of the 57 Loggerhead Shrike pairs on breeding territories in 2012–2013 were within the agricultural land-use type. The number of breeding territories in other land-use types occupied were as follows: 6 (11%) in rural housing developments, 3 in industrial parks, 2 at raceways (in mowed grassy parking areas and open Longleaf Pine forest), 1 at an airport, 1 at a prison, 1 in a cemetery, 1 on a power-line right-of-way, 1 in a forest clearcut, 1 in a young pine plantation, and 1 in a managed mature Longleaf Pine forest (Gamelands Field-trial Grounds). Sand pits and military drop-zones were not used as established breeding territories during 2012–2013 but had been used once each during the historical period. Uncertain breeding-season territories. I classified a total of 19 sites as uncertain Loggerhead Shrike breeding-season territories or visitation sites in 1 or both years. Ten of 19 sites (53%) were only used as uncertain breeding-season territories, with 3 sites visited both years, 2 sites visited only in 2012, and 5 sites visited only in 2013 (= visitor only group). In contrast, 9 of 19 (47%) sites were used as uncertain breeding-season territories in 1 year (3 sites in 2012, 6 sites in 2013), but were also used as confirmed breeding sites (n = 1) or breeding-season territories (n = 8) in the alternative year (= alternative group). Differences in the median length of stay and date of first occurrence between these 2 groups were not significant (length of stay: visitor only group = 1 day, alternative group = 14 days, Mann-Whitney U = 30; date of first occurrence: visitor only group = 26 April, alternative group = 7 April, Mann-Whitney U = 82.5; critical value of Mann-Whitney U-test = 28). However, visiting Loggerhead Shrikes had a minimum stay of 13 days longer and arrived 17 days earlier at sites that had confirmed breeding or were used as breeding-season territories in the alternative year—either 2012 or 2013. Southeastern Naturalist D.B. McNair 2015 Vol. 14, No. 4 766 Three of the 4 longest stays at the 10 sites that Loggerhead Shrikes visited only in 2012–2013 occurred at all 3 sites that were visited both years. This total included 1 historical site where breeding had been confirmed in managed Longleaf Pine forest with pastures at the North Carolina Gamelands Field-trial Grounds headquarters in 1988–1989 and 1993. Shrikes remained at this site for a minimum of 6 and 1 days, respectively, in 2012 and 2013 even though macrohabitat and land-use had apparently remained unchanged (examination of aerial photos from 1993 to 2013 on Google Earth [version 7.1.2], and D.B. McNair, pers. observ.). Similarly, there was a Loggerhead Shrike breeding territory in 1981 and 1983–1984 at another historical site nearby in Derby, Richmond County. In 2012, a pair was present at this site in the same hedgerow among a similar matrix of agricultural habitats for a minimum of 17 days but departed shortly after mechanized agricultural-management activities occurred within their core territory; a pair remained on their breeding territory here in 2013. Two confirmed historical breeding sites in a matrix of agricultural habitats and pine forest in 1989–1990 at Scholl, along or near the Richmond–Scotland county line, were used as visitation sites in 1 recent year, but used as breeding territories in the alternative year in 2012–2013. Finally, a Loggerhead Shrike pair abandoned a young pine plantation after 1 day in 2012 when 1 territorial adult Accipiter cooperii Bonaparte (Cooper’s Hawk) unsuccessfully attempted to capture a shrike; shrikes were not detected at this site in 2013. Discussion My study is the first in the Carolinas or the rest of southeastern North America that has re-assessed the breeding distribution of Loggerhead Shrikes between 2 different time periods. In contrast, many studies of shrike species in Europe have documented their fine-scale breeding distribution and status over many years (Kuczyński et al. 2010, Pasinelli et al. 2011), including records of increasing and declining populations. My survey protocol in 2012–2013 did not allow me to estimate any detection probability function, but my mean survey time for first detection of presence with a 95% confidence interval was 1.09–1.33 visits, more efficient than 2.7 visits on surveys for the Red-backed Shrike (Kéry and Schmidt 2008), although the latter species is probably less visible compared to the Loggerhead Shrike. My protocol allocated 3 Loggerhead Shrike locations to the uncertain breeding-season territory category in 2012, and 2 in 2013, even though these sites had been used as confirmed historical breeding sites by shrikes and still retained suitable breeding habitat. My protocol may have been too strict, but I wanted to ensure that Loggerhead Shrike locations that were legitimate breeding territories in 2012–2013 were rigorously documented and that my population estimate was precise. I detected shrikes efficiently along primary and secondary roads that traversed the Gamelands, but my detection rate may have been lower along tertiary roads in managed Longleaf Pine forests where anthropogenic perch sites are generally scarce. Most Loggerhead Shrike detections along tertiary roads from 2004–2012 were concentrated at the Field-trial Grounds where managed Southeastern Naturalist 767 D.B. McNair 2015 Vol. 14, No. 4 fields and open Longleaf Pine stands are maintained (J. Marcus et al., unpubl. data), which I covered in 2012–2013. I under-recorded the number of Loggerhead Shrike breeding territories, but not locations per se, during the historical period because of insufficient effort. Nonetheless, Loggerhead Shrikes declined along or near the Fall Line where 5 confirmed historical breeding sites were not used in 2012 (71% reduction), which is consistent with a sharp decline reported for the Piedmont of North Carolina where the species is now rare or locally very uncommon (LeGrand et al. 2011, Sauer et al. 2011). The Loggerhead Shrike breeding population in my study area within the North Carolina Sandhills was concentrated in 3 areas, but has slowly declined over approximately the last 30 years despite a substantial loss of all types of pasturelands as well as other types of grassy or forb-dominated habitats. This slowly declining yet fairly stable population in the Sandhills is not predicted by CBC or BBS data from North Carolina (Sauer et al. 1995, 2011), demonstrating the value of using a strict survey protocol at a fine scale to produce a precise population estimate of a resident species still occupying favorable land-use types and macrohabitats in a geographically restricted area. In southern Indiana, where breeding populations of migratory Loggerhead Shrikes restricted to 2 physiographic regions have sharply declined (Burton and Whitehead 2002), a core population declined less (reduction of 13%) than other populations (reduction of 60%) from 1988 to 2000. The breeding population of Loggerhead Shrikes in my study area in the Sandhills probably constitutes a core population, declining less precipitously compared to most other areas in North Carolina. Nonetheless, the breeding density of Loggerhead Shrikes in the North Carolina Sandhills appeared to be low in 2012–2013 although few comparisons are available in North America (cf., Yosef 1996, where roadside survey densities are actually indices of abundance). All but 2 territorial pairs in the North Carolina Sandhills were separated by a minimum distance of 1000 m, so any potential confusion of pair identity that could occur in dense populations of unmarked Loggerhead Shrikes rarely occurred. Behavioral observations pinpointed core-activity areas for pairs that occurred close together, including probable nests in trees in rural farmyards that I did not access because of privacy concerns. Despite a low population density, population persistence (not site fidelity per se) at breeding territories in 2012–2013 compared to historical data was approximately 33%. Considering that the 2 time periods were separated by 18–19 years, population persistence appears to be high in my study area. The recent annual measure of population persistence between 2012 and 2013 was 56%, and longer length-of-stay and earlier arrival of Loggerhead Shrikes at territories used in alternative years and reuse of some historical sites suggests population persistence between both recent years may even have been higher. Annual re-occupancy of breeding territories of continental Loggerhead Shrike populations has ranged from 41% to 73%, including 60% in Virginia (Luukkonen 1987, Yosef 1996). Takagi (2003) and Tryjanowski et al. (2007) suggested that high site-fidelity of migratory Lanius cristatus L. (Brown Shrike) in Japan and the Red-backed Shrike in Western Europe, respectively, was Southeastern Naturalist D.B. McNair 2015 Vol. 14, No. 4 768 due to the patchy, isolated character of breeding habitats in intensive agricultural landscapes that inhibited shrike dispersal. High site-fidelity in my study area within the North Carolina Sandhills, coupled with local dispersal of unattached shrikes, suggests that breeding territories are saturated. Acknowledgments I thank C.J. Randel for preparing Figure 1; J.A. Collazo, J. Ma rcus, and P. Tryjanowski for their reviews of a penultimate version of the manuscript; and 2 anonymous individuals for their reviews of the submitted version of the manuscript. Literature Cited Anchor, S., and K. Brust. 2004. Investigation of the American Kestrel (Falco sparverius) on the Fort Bragg Military Reservation, Hoke County, NC. Annual report. Sandhills Ecological Institute, Southern Pines, NC. 28 pp. Brooks, B.L., and S.A. Temple. 1990. Habitat availability and suitability for Loggerhead Shrikes in the Upper Midwest. American Midland Naturalist 123:75–83. Brown, D.G., K.M. Johnson, T.R. Loveland, and D.M. Theobald. 2005. Rural land-use trends in the conterminous United States, 1950–2000. Ecological Applications 15:1851–1863. Brust, K., J.H. Carter III, K.B. Beck, J.R. Walters, and P.D. Doerr. 2004. Status of the Red-cockaded Woodpecker on private and public lands in the North Carolina Sandhills longterm-study sites, 1992–2001. Pp. 116–126, In R. Costa and S.J. Daniels (Eds.). Red-cockaded Woodpecker: Road to Recovery. Hancock House Publishers, Blaine, WA. 744 pp. Burton, K.M., and D.R. Whitehead. 2002. Productivity and survival of the Loggerhead Shrike in Indiana. Bird Populations 6:13–20. Cade, T.J., and C.P. Woods. 1997. Changes in distribution and abundance of the Loggerhead Shrike. Conservation Biology 11:21–31. Carter, J.H., III, R.T. Stamps, and P.D. Doerr. 1983. Status of the Red-cockaded Woodpecker in the North Carolina Sandhills. Pp. 24–29, In D.A. Wood (Ed.). Red-cockaded Woodpecker Symposia II Proceedings. Florida Game and Freshwater Fish Commission and US Fish and Wildlife Service, Atlanta, GA. 112 pp. Collister, D.M., and S. Wilson. 2007. Contributions of weather and predation to reduced breeding success in a threatened northern Loggerhead Shrike population. Avian Conservation and Ecology 2(2):11. Available online at http://www.ace-eco.org/vol2/iss2/art11/. Accessed June 2012. DeMent, S.H., K.S. Latimer, and K.A. Hobson. 2008. Survey of wintering Loggerhead Shrikes in South Carolina including stable hydrogen-isotope feather analysis. Chat 72:1–6. Gawlik, D.E., and K.L. Bildstein. 1990. Reproductive success and nesting habitat of Loggerhead Shrikes in north-central South Carolina. Wilson Bulletin 102:37–48. Gawlik, D.E., and K.L. Bildstein. 1993. Seasonal habitat use and abundance of Loggerhead Shrikes in South Carolina. Journal of Wildlife Management 57:352–357. Kéry, M., and B.R. Schmidt. 2008. Imperfect detection and its consequences for monitoring for conservation. Community Ecology 9:207–216. Kesler, D.C., and J.R. Walters. 2012. Social composition of destination territories and matrix habitat affect Red-cockaded Woodpecker dispersal. Journal of Wildlife Management 76:1028–1035. Southeastern Naturalist 769 D.B. McNair 2015 Vol. 14, No. 4 Kuczyński, L., M. Antczak, P. Czechowski, J. Grzybek, L. Jerzak, P. Zablocki, and P. Tryjanowski. 2010. A large-scale survey of the Great Grey Shrike, Lanius excubitor, in Poland: Breeding densities, habitat use, and population trends. Annales Zoologica Fennici 47:67–78. LeGrand, H.E., Jr., J. Haire, A. Iyoob, and T. Howard. 2011. Birds of North Carolina: Their distribution and abundance. Loggerhead Shrike. Available online at www.carolinabirdclub. org/ncbirds/accounts.php. Accessed June 2012. Lowe, J.D., and G.S. Butcher. 1990. Population dynamics of the Loggerhead Shrike in winter: 1963–1987. Birdscope 4:4. Lübcke, W. 2007. Langjährige Bestandserfassung (1984–2005) des Neuntöters (Lanius collurio) im Bereich des MTB-Viertels 4820/II Bad Wildungen. Vogelkundliche Hefte Edertal 33:72–79. Luukkonen, D.R. 1987. Status and breeding ecology of the Loggerhead Shrike in Virginia. M.Sc. Thesis. Virginia Polytechnic Institution and State University, Blacksburg, VA. Michaels, H.L., and J.F. Cully Jr. 1998. Landscape and fine-scale habitat associations of the Loggerhead Shrike. Wilson Bulletin 110:474–482. Miller, M.P., T.D. Mullins, J.W. Parrish Jr., J.R. Walters, and S.M. Haig. 2012. Variation in migratory behavior influences regional genetic diversity and structure among American Kestrel (Falco sparverius) populations in North America. Journal of Heredity 103(4):503–514. Morrison, M.L. 1981. Population trends of the Loggerhead Shrike in the United States: A statistical study of Christmas Bird Count reports provides qualified confirmation of the species’ blue-listing. American Birds 35:754–757. North Carolina Wildlife Resources Commission (NCWRC). 2008. Protected wildlife species of North Carolina. Available online at http://www.ncwildlife.org. Accessed June 2012. Pasinelli, G., M. Schaub, G. Häfliger, M. Frey, H. Jakober, M. Müller, W. Stauber, P. Tryjanowski, J.–L. Zollinger, and L. Jenni. 2011. Impact of density and environmental factors on population fluctuations in a migratory passerine. Journal of Animal Ecology 80:225–234. Puetz, C.J. 1990a. North Carolina. County Map Books. L yndon Station, WI. 156 pp. Puetz, C.J. 1990b. South Carolina. County Map Books. L yndon Station, WI. 128 pp. Sauer, J.R., S. Orsillo, and B.G. Peterjohn. 1995. Geographic patterns and population trends of breeding and wintering Loggerhead Shrikes in North America. Proceedings of the Western Foundation of Vertebrate Zoology 6:128–141. Sauer, J.R., J.E. Hines, J.E. Fallon, K.L. Pardieck, D.J. Ziolkowski Jr., and W.A. Link. 2011. The North American breeding bird survey results and analysis 1966–2010. Version 12.07.2011. USGS Patuxent Wildlife Research Center, Laurel, MD. Available online at http://www.mbr-pwrc.usgs.gov/bbs/bbs.html. Accessed June 2012. Smith, E.L., and K.C. Kruse. 1992. The relationship between land-use and the distribution and abundance of Loggerhead Shrikes in south-central Illinois. Journal of Field Ornithology 63:420–427. Takagi, M. 2003. Philopatry and habitat selection in Bull-headed and Brown shrikes. Journal of Field Ornithology 74:45–52. Tryjanowski, P., M. Hromada, M. Antczak, J. Grzybek, S. Kuźniak, and G. Lorek. 2003. Which method is most suitable for censusing breeding populations of Red-backed (Lanius collurio) and Great Grey (L. excubitor) shrikes? Ornis Hungarica 12–13:223–228. Tryjanowski, P., A. Golawski, S. Kuźniak, T. Mokwa, and M. Antczak. 2007. Disperse or stay? Exceptionally high breeding-site infidelity in the Red-backed Shrike, Lanius collurio. Ardea 95:316–320. Southeastern Naturalist D.B. McNair 2015 Vol. 14, No. 4 770 United States Census Bureau. 2015. 2014 estimates. Available online at http://www.census. gov/. Accessed April 2015. United States Department of Agriculture. 1999. Soil Survey of Richmond County, North Carolina. Natural Resources Conservation Service. Available online at http://www.usda. gov. Accessed March 2012. United States Department of Agriculture. 2015. National Agricultural Statistics Service. Census of Agriculture. Available online at http://agcensususda.gov/Publications/. Accessed April 2015. Westphal, M. 2009. 2009 spring migration counts in North Carolin a. Chat 73:77–104. Yosef, R. 1996. Loggerhead Shrike (Lanius ludovicianus). No. 231. In A. Poole (Ed.). Birds of North America. Cornell Laboratory of Ornithology, Ithaca, NY. Available online at http://bna.birds.cornell.edu/bna/species/231/articles/introduction. Accessed June 2012. Yosef, R. 2001. Nesting ecology of resident Loggerhead Shrikes in south-central Florida. Wilson Bulletin 113:279–284.