Southeastern Naturalist 629 C. Hartwig, et al. 22001188 SOUTHEASTERN NATURALIST 1V7o(4l.) :1672,9 N–6o4. 44 Temporal Changes in Diversity and Abundance of Mosquitoes (Insecta: Diptera: Culicidae) in a Small Ecological Preserve in North Carolina Carmony Hartwig1,*, Bruce Harrison1,2, Joshua York1, Elizabeth Brown1, Jay Bolin1, Parker Whitt3, Ryan Harrison4, Hugh Smith1, and Marlon Barber1 Abstract - We documented a changing diversity in mosquito species between 2 collection periods— 1994–1996 and 2013–2015—in a small (68-ha) ecological preserve in the piedmont of North Carolina. A short (22-y) ecological succession from abandoned farmland to developing forested wetland, and changes in precipitation clearly influenced differences in presence and abundance of species in the preserve. Thirty species were reported from the first period and 32 species in the second period. Of the 30 species found in 1994–1996, 3 species were not collected in the 2013–2015 period. Conversely, 6 species not reported previously were present in the 2013–2015 collections. From both periods, a total of 7172 mosquito specimens of 36 species were collected, representing 95% of species found in Rowan County, an area 2000 times larger than the Fred Stanback Jr. Ecological Preserve (FSJEP), and 54% of species recognized in North Carolina. These results demonstrate the advantages of studying mosquito diversity and abundance over time in small preserves, the impact of short-period environmental fluctuations and ecological succession on mosquito habitats, and the value of small wetland preserves for rare or uncommon species affected by habitat loss. Introduction Mosquitoes are beneficial to ecosystem function and community structure, serving not only as food sources for other organisms such as salamanders, frogs, lizards, birds, bats and insects, but as pollinators for thousands of plant species world-wide (Fang 2010). Several hundred mosquito species also serve as efficient vectors of infectious human pathogens. During the 1980–1990s, work was initiated to establish a permanent ecological preserve on the Catawba College campus (35°41'29''N, 80°29'4''W) in Salisbury, located in the central piedmont of North Carolina in Rowan County. This preserve was named the Fred Stanback Jr. Ecological Preserve (FSJEP) and was registered in 1999 as a wildlife refuge by the LandTrust for Central North Carolina. The city of Salisbury sits at an elevation of ~241 m. A North Carolina Department of Environment and Natural Resources (Raleigh, NC) 1994–1996 Culicidae mosquito survey of the preserve (unpubl.) was conducted prior to its designation as a registered NC Natural Heritage Area. 1Department of Biology, Catawba College, 2300 W Innes Street, Salisbury, NC 28144. 2Vector- borne Infectious Disease Laboratory, College of Health and Human Sciences, Western Carolina University, Cullowhee, NC 28723. 3NC Department of Agriculture and Consumer Services, 110 Dixieanna Drive, Winston-Salem, NC 27107. 4Vector Control, Forsyth County Department of Public Health, PO Box 686, Winston-Salem, NC 27102. *Corresponding author - chartwig@catawba.edu. Manuscript Editor: Robert Jetton Southeastern Naturalist C. Hartwig, et al. 2018 Vol. 17, No. 4 630 This seasonally inundated floodplain preserve currently harbors a diverse flora and fauna that has changed substantially from an early successional old-field plant community in the northern half of the FSJEP that existed when work was initiated in 1989. The FSJEP is part of the Catawba College campus, and is flanked by a steep hillside covered in mature forest on the west, Grants Creek and farm fields on the north, a greenway on the east, and an urban residential area on the south. Applying the plant community classification system of Schafale (2012), the following plant communities dominate the FSJEP: mature forested slopes are Mesic Mixed Hardwood Forest (Piedmont subtype), and the floodplain areas that comprise most of the preserve are Piedmont Levee Forest along Grants Creek and Piedmont Swamp Forest in the floodplain areas away from the creek levee. The FSJEP has changed dramatically since the original 1994–1996 Culicidae survey (Fig. 1) from numerous open fields with large, sunlit grassland spaces, ditches, temporary rain pools, thickets, and several trails (described in Baranski 1994), into a young wetland forest with a low canopy covering much of the area. In terms of canopy coverage, 26 ha (38.4% of total area) of open-canopy areas (fallow fields) existed at the time of the 1994–1996 survey period, while only 1.8 ha (2.6%) of open-canopy areas remained in 2013–2015. Besides a shift from open to shaded habitats in the northern half of the FSJEP over this period, there has also been trail construction, which has resulted in increased human use of the preserve. This natural area comprises only 68 ha (189 ac), but it has a rich flora and fauna. The most recent biological inventory of the preserve (J. Bolin and J. Cooley, Catawba College, Salisbury, NC, 2012, unpubl. data) identified: 303 vascular plants, 21 reptiles, 17 amphibians, 166 birds, 41 mammals, 67 butterflies, and 38 moths. Small preserves, such as the FSJEP, are invaluable in terms of understanding how habitat changes may affect populations of local flora and fauna, and serve as an overall indicator of the ecological health of an area. Mosquitoes are an abundant and integral part of a preserve’s insect fauna; thus, our goal was to ascertain how the dramatic environmental changes in the FSJEP over the 22–y period are related to mosquito diversity and abundance. We were specifically interested in examining the impact of ecological succession as it relates to canopy coverage and mosquitospecies habitat loss. Here we identify changes in the abundance and diversity of the mosquito populations in the FSJEP over a 20-y period (from 1994–1996 to 2013–2015) and discuss concurrent ecological and environmental changes that likely influenced population dynamics and habitat preferences of mosquito species. Materials and Methods During the original 1994–1996 Culicidae survey, both larval and human landing collections (HLC) were initiated throughout the preserve to ascertain the level of species diversity of abundant mosquitoes. HLC collections were taken at random throughout the preserve (walking on trails); this was not the primary sampling method, but we include data from both the original survey and the 2013–2015 period. Larval collections were made at random using mosquito dippers in the FSJEP from standing-water pools, ditches, summer rain pools, and tree holes throughout Southeastern Naturalist 631 C. Hartwig, et al. 2018 Vol. 17, No. 4 the preserve, and from abandoned trash along the perimeter of the northwestern slope in 2013–2015. Larval collections during both periods were made using mosquito dippers. Larvae were collected and identified, but not counted in the female- mosquito relative abundance calculations for the FSJEP during the 2 periods. To measure differences in mosquito species presence and activity within the FSJEP, 7 sites in 1994–1996 and 8 sites in 2013–2015 were used for monitoring females attracted to CDC light traps with CO2. Light traps were set in the afternoon and collected the following morning. Collection via standard CDC light traps equipped with a CO2 lure was the primary method for capturing females. During Figure 1. Catawba College FSJEP, with numbered locations for adult mosquito light-trap collection sites (sampling points). On left, aerial orthophoto of FSJEP in 1994 with 6 locations that were sampled; on right aerial orthophoto 20 y later, in 2014, with 8 total sampling locations. The orthoimagery indicates a dramatic shift from a large proportion (38.4%; 26 ha) of open-canopy (fallow field) habitat in 1994 to a largely closed canopy and early successional forest community in 2014, when only 1.8 ha (2.6%) of open-canopy habitat comprised of managed fields remained. Orthoimagery from the North Carolina Center for Geographic Information and Analysis, Raleigh, NC. Southeastern Naturalist C. Hartwig, et al. 2018 Vol. 17, No. 4 632 the 2013–2015 period, we supplemented adult sampling by the addition of a BGSentinel ® (BioGents USA, Dallas, TX) trap (near L2). We used the BG-Sentinel® trap for 4 collections, during which time the traps were active for ~20–24 h from one morning to the next. Traps using CO2 and other attractants are designed to collect females seeking blood hosts; thus, we anticipated species-specific differences in dispersal from eclosion sites, and selected sites to increase the diversity and abundance of species collected. During the 1994–1996 period, CDC light-trap collections of females (based on 16 trap-nights; 17 h each for a total of ~272 trap-h) were made from April through late September. In 2013–2015, we conducted 35 trap-nights of collections, active for ~16 h each trap night for a total of 560 light trap-h. Trapping in 1994–1996 occurred at 7 locations (the northwest [L3], open field [northern third of preserve, between L2 nd L3], northeast [L7], east end of the ditch [L6], middle of the western side [L2], middle of the eastern side [L5], and the southern end of the sewer-line trail near family residences [L1]), 6 of which we used during the 2013–2015 collection period (Fig. 1). For the 2013–2015 sampling, we added locations L4 (central sewer-line trail) and L8 (northwestern slope directly behind campus library) (Fig. 1). The 1994–1996 survey was carried out shortly after plans were initiated by the College to establish a formal ecological preserve. At that time, several trails had been constructed, wooden floodgates were prepared to control water movement, and a newly created large impoundment had just been completed. However, major development efforts were not initiated until 1998, when the open farm fields (L7 and L2) were left fallow to transition to more natural wetland areas and construction of raised trails, ponds, and impoundments began. At that time, most of the northern half of the preserve (L3 and L7) was in an early successional state with Andropogon virginicus L. (Broomsedge Bluestem), Rubus fruticosus L. (Shrubby Blackberry), Ambrosia spp. (ragweeds), Solidago spp. (goldenrods), small Pinus taeda L. (Loblolly Pine), Pinus virginiana Mill. (Virginia Pine), small Juniperus virginiana L. (Northern Red Cedar), Liquidambar styraciflua L. (Sweet Gum) saplings, Nyssa sylvatica Marsh. (Black Gum) trees, and numerous other grasses growing in these fields. The central part of the preserve (L2, L5, and L6) is slightly lower in elevation and has both semi-permanent and permanent water present. A major ditch (L6), with large Quercus spp. (oaks) and Carya spp. (hickory) trees on each side, crossed the preserve from west to east. This ditch was a source of slow-moving continuous seepage water from the hillside on the west side of the preserve, and provided water for the large impoundment. The ditch and associated line of trees created habitat on the side facing the northern open field for adults of certain mosquito species that favor forest edges, and also served as a major larval habitat for many mosquito species. In the mid 1990s, a higher water table was evident south (L1) and east (L5) of the central impoundment and many parts of the preserve remained wet throughout the year, primarily due to the runoff from Grants Creek (near L3 and L7) and overflow from Lake Baranski (centrally located in FSJEP; see Fig. 1). The plant Southeastern Naturalist 633 C. Hartwig, et al. 2018 Vol. 17, No. 4 community in this area consisted primarily of Salix nigra Marsh. (Black Willow), Fraxinus pennsylvanica Marsh. (Green Ash), and Acer rubrum L. (Red Maple). On the east side (L5), there were numerous open sunlit areas with species that require high light levels such as Alnus serrulata (Ait.) Willd. (Smooth Alder), Typha latifolia L. (Cattail), Juncus effusus L. (Common Rush), and Cephalanthus occidentalis L. (Buttonbush). In L1, Spyrogyra algae, Lemna spp. (duckweeds), and Potamogeton spp. (pondweeds) were abundant periodically. By the time of the 2013–2015 study period, the open-field character and early successional vegetation that required high light levels had transitioned to closed canopy, forested wetland with a understory dominated by shade-tolerant wetland plants such as Saururus cernuus L. (Lizard’s Tail) and Impatiens capensis Meerb. (Jewelweed). Currently, no Cattails or Common Rush, species characteristic of emergent wetlands, are present, and the closed-canopy forested wetland that dominates the FSJEP is classified as Piedmont Swamp Forest, based on hydrology and characteristic vegetation. The dominant overstory trees include Green Ash, Red Maple, and to a lesser extent, Sweet Gum, Quercus phellos L. (Willow Oak), Quercus lyrata Walter (Overcup Oak), and Populus heterophylla L. (Swamp Cottonwood). After sorting and identification using Slaff and Apperson (1989) or unpublished drafts of Harrison et al. (2016), we stored specimens at -20 °C for later use. We used EZ4® dissection microscopes (Leica, Wetzlar, Germany) for initial species sorting and made final identification verifications for both females and larvae using a Leica S8APO® dissection microscope with 8x magnification. We identified larvae in ethanol in spot-well plates. We pointed, labeled, and preserved all female specimens in insect museum-drawers in a Cornell insect cabinet (BioQuip Products, Inc, Rancho Dominguez, CA.) at Western Carolina University, Cullowhee, NC, or in Wards Natural Science storage boxes (Ward’s Science, West Henrietta, NY) at Catawba College, Salisbury, NC, to be prepared as vouchers from both periods. The tribe Aedini generic and subgeneric names and their abbreviations follow Wilkerson and Linton (2015) and Wilkerson et al. (2015). Otherwise, we followed the anopheline and other non-Aedini culicine names in Knight and Stone (1977). Use of Culex pipiens complex throughout is based on knowledge that specimens in all but one county (Brunswick) of NC are hybrids of Culex pipiens L. x Culex quinquefasciatus Say (Harrison et al. 2016). Anopheles quadrimaculatus is listed as Anopheles quadrimaculatus s. l., because the 5 members of the An. quadrimaculatus species complex (Reinert et al. 1997) were not described until after the 1995–1996 sampling period. In order to evaluate relative differences in total mosquito diversity between the 2 collection periods (1994–1996 and 2013–2015), we calculated and compared the Shannon-Weiner diversity index for each period. Combination of our 2 studies was unplanned; thus, we did not apply statistical analysis of species abundance and diversity from the 2 collection periods (1994–1996 and 2013–2015) to resultant data sets. Instead, we used statistical analyses to compare only the top 10 species collected from the 1995 and 2014 years of record. We conducted a t-test with Southeastern Naturalist C. Hartwig, et al. 2018 Vol. 17, No. 4 634 Bonferroni correction to determine significant differences between the top 10 species from 1995 and 2014. Results A total of 7172 adult females were collected in the FSJEP during both sampling periods. These samples represent 2 subfamilies and 7 tribes that include 36 species in 9 genera (Table 1). The collections represent 95% of mosquito species known in Rowan County, an area of 1360 km2 (almost 2000 times the area of FSJEP), and 54% of the mosquito species recognized in North Carolina. The recent reduction of the previously recognized subfamily Toxorhynchitinae (Knight and Stone 1977) Table 1. Master list of mosquito species found in the FSJEP from 1994–1996 and 2013–2015 collections, including genus, subgenus, species, and author(s). 1 Aedes (Stegomyia) albopictus (Skuse) 2 Ae. (Ochlerotatus) atlanticus Dyar and Knab 3 Ae. (Ochlerotatus) canadensis canadensis (Theobald) 4 Ae. (Aedes) cinereus Meigen 5 Ae. (Ochlerotatus) dupreei (Coquillett) 6 Ae. (Ochlerotatus) fulvus pallens Ross 7 Ae. (Protomacleaya) hendersoni Cockerell 8 Ae. (Ochlerotatus) infirmatus Dyar and Knab 9 Ae. (Hulecoeteomyia) japonicus japonicus (Theobald) 10 Ae. (Ochlerotatus) sticticus (Meigen) 11 Ae. (Ochlerotatus) tormentor Dyar and Knab 12 Ae. (Protomacleaya) triseriatus (Say) 13 Ae. (Ochlerotatus) trivittatus (Coquillett) 14 Ae. (Aedimorphus) vexans (Meigen) 15 Anopheles (Anopheles) crucians A–genetic provisional species of Wilkerson et al. (2004) 16 An. (Anopheles) crucians E–genetic provisional species of Wilkerson et al. (2004) 17 An. (Anopheles) punctipennis (Say) 18 An. (Anopheles) quadrimaculatus sensu lato 19 Coquillettidia (Coquillettidia) perturbans (Walker) 20 Culex (Melanoconion) erraticus (Dyar and Knab) 21 Cx. (Culex) nigripalpus Theobald 22 Cx. (Culex) pipiens L. x Cx. (Culex) quinquefasciatus Sayhybrids 23 Cx. (Culex) restuans Theobald 24 Cx. (Culex) salinarius Coquillett 25 Cx. (Neoculex) territans Walker 26 Culiseta (Culiseta) inornata (Williston) 27 Cs. (Climacura) melanura (Coquillett) 28 Orthopodomyia signifera (Coquillett) 29 Psorophora (Psorophora) ciliata (Fabricius) 30 Ps. (Grabhamia) columbiae (Dyar and Knab) 31 Ps. (Janthinosoma) cyanescens (Coquillett) 32 Ps. (Janthinosoma) ferox (von Humboldt) 33 Ps. (Psorophora) howardii Coquillett 34 Ps. (Janthinosoma) mathesoni Belkin and Heinemann 35 Toxorhynchites (Lynchiella) rutilus septentrionalis (Dyar and Knab) 36 Uranotaenia (Uranotaenia) sapphirina (Osten Sacken) Southeastern Naturalist 635 C. Hartwig, et al. 2018 Vol. 17, No. 4 to a Tribe, Toxorhynchitini, in Subfamily Culicinae follows Harbach and Kitching (1998), Mitchell et al. (2002), and Rueda (2008). Mosquito diversity and abundance during 1994–1996 During this period, a total of 30 mosquito species were captured in the preserve in light trap, human landing, and larval collections (Table 2). HLC yielded 102 females, while 4943 females were collected in traps, providing 5045 total specimens. Twenty-seven species were collected in CDC light traps supplemented with CO2, 12 species in HLC collections, and 22 as larvae. All species collected by HLC and larval dipping were also represented in light-trap collections. Two species, Aedes tormentor and Culiseta inornata, were collected only as larvae, bringing the total number of species to 29. Although An. crucians Wiedemann is included in this list as a single species, it is now known to represent at least 5 provisional genetic species (Cockburn et al. 1993, Wilkerson et al. 2004). Two of those cryptic species (An. crucians A, E) were collected in the FSJEP in 1996, but not counted in those collections because they were not identified as rDNA ITS2 species until Wilkerson et al. (2004); however, they are listed in Table 2. Deleting the name, An. crucians Wiedemann, and adding the 2 provisional genetic species means there were 30 species known from the FSJEP before the 2013–2015 collections. Table 2. Culicidae taxa, female mosquito relative abundance (% of total collection for the respective periods) in FSJEP during 2 periods, and normal habitat preferences for the species. *See discussion and Table 3 regarding data for Aedes atlanticus and Aedes tormentor. Damaged specimens in the 2013–2015 period are reported as Ae. atlanticus/tormentor. [Table continued on following page.] Species abundance 1994– 2013– Adult species taxon 1996 2015 Larval habitat ANOPHELINAE Anopheles crucians sensu lato 32.93 7.01 Permanent freshwater ponds, lakes, and marshes An. punctipennis 1.43 3.24 Permanent and semi-permanent freshwater bodies of water An. quadrimaculatus sensu lato 0.22 0.94 Permanent freshwater ponds, lakes, and marshes CULICINAE AEDINI Aedes albopictus 0.10 0.52 Artificial and natural containers Ae. atlanticus n/a* 0.14* Temporary freshwater pools Ae. atlanticus/tormentor 19.52* 1.41* Temporary freshwater pools Ae. canadensis canadensis 3.55 11.24 Semi-permanent freshwater pools Ae. cinereus 0.00 0.05 Semi-permanent freshwater pools Ae. dupreei 0.18 0.66 Semi-permanent and temporary freshwater pools Ae. fulvus pallens 0.04 0.05 Temporary freshwater pools Ae. hendersoni 0.00 0.14 Tree holes rarely containers Ae. infirmatus 0.00 3.71 Temporary freshwater pools Ae. japonicus japonicus 0.00 0.80 Artificial and natural containers (rock and tree holes) Ae. sticticus 0.06 2.12 Still creek/riverine flood pools in woods Ae. tormentor n/a* 10.86 Temporary freshwater pools Ae. triseriatus 0.08 0.80 Artificial and natural containers Southeastern Naturalist C. Hartwig, et al. 2018 Vol. 17, No. 4 636 During the 1994–1996 study, mosquitoes were often very abundant in the FSJEP, and they were intolerable in the absence of proper clothing and repellents (B. Harrison, pers. observ.). The primary mosquito species collected from the northwest (L3), northern open field (between L3 and L2), and northeast (L7) sites were those characteristic of temporary rain pools, i.e., Ae. atlanticus, Ae. tormentor, Cx. nigripalpus, Cx. restuans, Cx. salinarius, Psorophora columbiae, Ps. cyanescens, Ps. ferox, and several long-distance flying species likely from the southern wetter Table 2 continued. Species abundance 1994– 2013– Adult species taxon 1996 2015 Larval habitat Ae. trivittatus 0.12 0.00 Creek/riverine flood pools in woods Ae. vexans 4.70 19.89 Temporary freshwater pools and marshes Psorophora ciliata 0.63 0.05 Temporary freshwater pools with other larvae in them (predaceous) Ps. columbiae 0.34 0.05 Small to medium temporary freshwater pools Ps. cyanescens 0.61 0.00 Temporary freshwater pools with grass Ps. ferox 12.24 28.82 Small temporary freshwater pools/still water ditches Ps. howardii 0.40 0.19 Temporary freshwater pools with other larvae in them (predaceous) Ps. mathesoni 0.00 0.09 Still creek/riverine flood pools in woods CULICINI Culex erraticus 3.23 2.40 Permanent freshwater with floating vegetation Cx. nigripalpus 0.06 0.00 Temporary freshwater pools and ditches Cx. pipiens complex 0.02 0.33 Temporary freshwater pools, ditches and artificial containers Cx. restuans 1.41 0.71 Freshwater pools, ditches, wheel ruts and artificial containers Cx. salinarius 3.94 1.08 Freshwater pools and marshes and brackish water marshes Cx. territans 0.06 0.85 Permanent and semi-permanent freshwater CULISETINI Culiseta inornata 0.00 0.28 Semi-permanent freshwater pools Cs. melanura 0.06 0.85 Root and stump holes from blown down trees in swamps MANSONIINI Coquillettidia perturbans 11.79 0.24 Permanent freshwater with emergent vegetation ORTHOPODOMYIINI Orthopodomyia signifera 0.02 0.00 Tree holes, used tires and larger artificial containers TOXORHYNCHITINI Toxorhynchites rutilus 0.00 0.09 Tree holes, large artificial containers, used tires septentrionalis URANOTAENIINI Uranotaenia sapphirina 2.24 0.42 Permanent freshwater with floating and other vegetation Southeastern Naturalist 637 C. Hartwig, et al. 2018 Vol. 17, No. 4 parts of the preserve, i.e., Ae. c. canadensis, Ae. vexans, An. crucians s.l., An. punctipennis, Cq. perturbans, and Cx. erraticus. Species most commonly collected in the middle section, including the ditch (L6), middle of the east side (L5), and middle of the west (L2) side sites were Ae. c. canadensis, Ae. atlanticus, Ae. tormentor, Ae. vexans, An. punctipennis, Cx. erraticus, Ps. ciliata, Ps. ferox, and Ps. howardii. However, the more sunlit areas in the middle of the east side (L5) contained significant populations of An. punctipennis, An. quadrimaculatus s.l., Cq. perturbans, and Uranotaenia sapphirina. The majority of specimens from the south end (L1) of the sewer trail in wet and shaded areas were An. crucians s.l., An. punctipennis, Cx. erraticus, Cx. territans, and Cs. melanura. Of particular interest, females of container-breeding, vector species, Ae. albopictus, Ae. triseriatus, and Cx. pipiens complex were uncommon in the preserve. Residences adjacent to the southern and northwestern ends of the preserve were likely sources of these species. In addition, various types of discarded containers that could serve as larval habitats for these species were occasionally found in the preserve. Most of these specimens were collected as larvae or in HLC collections. Larvae of the first 2 species were occasionally found in tree holes, while a few specimens of the last species were occasionally found in ditches. Mosquito diversity and abundance during 2013–2015 During the 2013–2015 period, we captured 32 mosquito species in the preserve using CDC light traps, the BG-Sentinel® trap, and larval collections. These combined collections provided 2127 total female specimens from light traps and the BG-Sentinel® trap. We collected 30 species in light traps, 1 species as larvae from a tree hole at L1 (Orthopodomyia signifera) and 1 species in the BG-Sentinel® trap near L2 (Toxorhynchites r. septentrionalis), for a total of 32 species. During the 2013–2015 period, the most abundant species at L1, L3, and L7 were Ae. c. canadensis, Ae. vexans, and Ps. ferox. We collected the rare (or uncommon) species Aedes dupreei, Ae. hendersoni, Cq. perturbans, Or. signifera, and Ps. howardii at L1, the most southern part of the preserve which borders a residential area. Ae. albopictus, Ae. fulvus pallens, Ae. hendersoni, Ae. triseriatus, and Cx. salinarius were less frequently collected and mixed among abundant populations of An. quadrimaculatus s.l., and Cx. erraticus at L3 (northwest). At L7 (northeast), we collected more specimens of An. quadrimaculatus s.l., Ae. triseriatus, Cs. inornata, and Cs. melanura as well as small numbers of Ae. infirmatus, Cx. salinarius, and Cx. territans. Toward the southern end of the preserve, An. punctipennis, Ae. c. canadensis, Ae. vexans, and Ps. ferox were abundant at L4, while Ae. atlanticus, Cq. perturbans, Cx. territans, Ps. howardii, and Ur. sapphirina occurred at lower numbers. Similar to L4, we collected specimens of Ae. atlanticus and Ur. sapphirina at L5, toward the east where canopy cover is less prevalent. Collections from L5 shifted to favor high populations of Ae. tormentor, Ae. vexans, and Ps. ferox. Rare collections of Ps. mathesoni from the FSJEP occurred from L5 as well as L6, which is adjacent to a feeder ditch that has periodic flooding. Large numbers of Ae. c. canadensis, Ae. sticticus, and Ae. vexans characterized L6, as did low numbers of Cs. inornata and Ae. cinereus. Collections from L2 (near western Southeastern Naturalist C. Hartwig, et al. 2018 Vol. 17, No. 4 638 slope) consisted primarily of Ae. c. canadensis, An. crucians s.l., Ae. dupreei, Ae. j. japonicus, Ae. tormentor, Cx. pipiens complex, Ps. columbiae, and Ps. howardii. Finally, L8, the most western point of the preserve behind the library at the edge of the campus landscape, supported container-breeding species such as Ae. albopictus, Ae. j. japonicus, and Cx. restuans and a few specimens of Ae. infirmatus. Six of the 32 species collected during this period were not collected in the 1994– 1996 study (Ae. cinereus, Ae. hendersoni, Ae. infirmatus, Ae. j. japonicus, Ps. mathesoni, and Tx. r. septentrionalis). Considering the 6 species collected in 2013–2015, and the 3 species (Ae. trivittatus, Cx. nigripalpus, and Ps. cyanescens) collected only in 1994–1996, we observed an 8.3% increase in species richness between the 2 collection periods. Supporting these observations, the Shannon–Weiner diversity index calculated for both the 1994–1996 and 2013–2015 collection periods (H' = 2.07 and 2.29, respectively) demonstrates an increase in mosquito diversity in the latter period, even with the substantial drop in mosquito abundance (decrease of ~57.8% from 1994–1996 to 2013–2015). In order to make appropriate statistical comparisons of abundance between the 2 sampling periods, we selected the mid-term year of record for each study (1995 and 2014), and took the average abundance of the top 10 most-prevalent species per number of total light-trap nights (Table 3). Although we attempted more trap-nights in 2014 than in 1995, the number of females of each species analyzed represented only 30.9% of the females collected in 1995. Indeed, 3 species, An. crucians s. l., Cq. perturbans, and Cx. salinarius, were significantly less abundant (P ≤ 0.001) in 2014 than 1995. Table 3. Comparison of females collected per trap-night for most abundant species in years 1995 and 2014. 1995 (16 trap-nights) 2014 (20 trap-nights) Females Relative Females Relative per abundance per abundance # females trap-night ranking # females trap-night ranking Species (4330) (± SEM) (1994–1996) (1336) (± SEM) (2013–2015) Ae. atlanticus/tormentorA 976 61.0 ± 34.2 2 30 1.5 ± 1.0 10 Ae. atlanticusA n/a n/a n/a 3 0.2 ± 0.1 25 Ae. canadensis 171 10.7 ± 3.9 7 134 6.7 ± 3.2 3 Ae. infirmatus* 0 0.0 n/a 59 3.0 ± 1.1 6 Ae. sticticus 1 0.1 ± 0.1 21 39 2.0 ± 1.8 9 Ae. tormentorA n/a n/a n/a 226 11.3 ± 6.5 4 Ae. vexans 131 8.2 ± 3.7 5 285 14.3 ± 3.7 2 An. crucians s.l* 1479 92.4 ± 19.6 1 45 2.3 ± 1.2 5 Cq. perturbans* 405 25.3 ± 10.6 4 0 0.0 23 Cx. erraticus 125 7.8 ± 3.1 8 36 1.8 ± 0.9 8 Cx. salinarius* 192 12.0 ± 2.9 6 23 0.2 ± 0.1 11 Ps. ferox 591 36.9 ± 20.0 3 354 17.7 ± 13.1 1 AFemales of Ae. atlanticus and Ae. tormentor could not be differentiated morphologically before 2014 (Sither et al. 2013). Of the 2014 collections, 30 females were damaged or rubbed and were counted as Ae. atlanticus/tormentor. *Statistically significant difference between # of females collected between 1995 and 2014 collection years as determined via an independent t-test (P ≤0.001). Southeastern Naturalist 639 C. Hartwig, et al. 2018 Vol. 17, No. 4 The 6 species collected only during the 2013–2015 period, plus the 30 species previously collected in 1994–1996, total 36 species known from the FSJEP. Two other species not collected from the preserve, but recorded from Rowan County, are Aedes aegypti L. and Psorophora horrida (Dyar and Knab), for 38 total species known from this county—again indicating the great diversity of mosquitoes in this region. Discussion During the 1994–1996 study, 5045 females were collected. The 8 most-abundant species were: An. crucians s. l. (32.94%), Ae. atlanticus/tormentor (19.52%), Ps. ferox (12.25%), Cq. perturbans (11.79%), Ae. vexans (4.70%), Cx. salinarius (3.94%), Ae. c. canadensis (3.55%), and Cx. erraticus (3.23%), which represented 91.9% of the total specimens collected during that period (Table 2). We collected 2127 females during the 2013–2015 sampling period. The 8 most-abundant species were: Ps. ferox (28.82%), Ae. vexans (19.89%), Ae. c. canadensis (11.24%), Ae. tormentor (10.86%), An. crucians s. l. (7.01%), Ae. infirmatus (3.71%), An. punctipennis (3.24%) and Cx. erraticus (2.40%), which represented 87.2% of the total specimens during that period (Table 2). Of the 32 species collected between 2013 and 2015, Ae. cinereus, Ae. hendersoni, Ae. infirmatus, Ae. j. japonicus, Ps. mathesoni, and T. r. septentrionalis had not been collected previously. Aedes cinereus, Ae. hendersoni, and Ps. mathesoni are not commonly collected in North Carolina because of their unique habitats and narrow window of occurrence. Thus, the presence of these 3 species in the FSJEP likely indicates that the preserve is functioning as a refugium for them. Collectively, our data show distinct differences in species abundance during the 2 sampling periods. The fact that the open sunlit field present in 1994–1996 in the northern end of the preserve had developed into a lowland, wet, partially to heavily shaded forest by the 2013–2015 sampling period (Fig. 1) likely influenced the increase in species diversity. This change may be suggestive of a shift in mosquito species populations that prefer partial to heavily shaded oviposition and larval habitats. A comparison of the species abundance data by location between the 2 periods supports this suggestion. In contrast to 1994–1996 records of species collections at L1, L3, and L7 being dominated by species such as Ae. atlanticus, Ae. tormentor, Cx. nigripalpus, Ps. columbiae, and Ps. cyanescens, the most abundant species in 2013–2015 were Ae. c. canadensis, Ae. vexans, and Ps. ferox. During 2013–2015, we collected the rare species Ae. dupreei, Ae. hendersoni, Cq. perturbans, Or. signifera, and Ps. howardii at L1, the most southern part of the preserve, which borders a residential area. With the exception of L3, these locations (L1 and L7) run south to north along the multi-use trail and have transitioned since 1994–1996 to become more seasonally inundated wetland areas that receive runoff from Grants Creek and overflow from Lake Baranski. We collected a small number of Ae. atlanticus and Ur. sapphirina females at L5 in 2013–2015, which has transitioned to a seasonally flooded landscape from the open, sunlit, permanent water source with floating vegetation described in 1994–1996. Indeed, collections from L5 shifted from significant Southeastern Naturalist C. Hartwig, et al. 2018 Vol. 17, No. 4 640 populations of An. punctipennis, An. quadrimaculatus s.l., Cq. perturbans, and Ur. sapphirina in 1994–1996 to favor high populations of Ae. tormentor, Ae. vexans, and Ps. ferox, and the rarer collection of Ps. mathesoni. Two locations, L6, which is a feeder ditch that has periodic flooding, and L2, which is on the trail located between the western slope and Lake Baranski, did not show great shifts in predominant mosquito populations collected; Ae. c. canadensis, Ae. sticticus, Ae. vexans, Ae. tormentor, and Ps. howardii were abundant in both 1994–1996 and 2013–2015 collections. This finding is not surprising, as these areas have experienced the least successional changes compared with other locations. One thing that became obvious during our study was the much higher numbers of the 8 most-abundant species collected per 16 trap-nights in 1995, compared to the abundance of those species collected in 20 trap-nights in the 2014 (Table 3). Although more trap-nights were attempted in 2014, the number of females of each species analyzed represented only 30.9% of the females collected in 1995. Three species—An. crucians s. l., Cq. perturbans, and Cx. salinarius—declined significantly between 1995 and 2014. This finding initiated concern that species abundance levels were declining overall in the FSJEP until we examined precipitation levels for those years. Salisbury receives an average of 106.81 cm (42.05 in) of rain per year. In 1995, 81.69 cm (32.16 in; 77%) of the mean annual precipitation fell in the 6-month collection period (April–September), while in 2014, only 63.12.cm (24.85 in; 59%) of the mean annual precipitation fell in the same 6 months. Typically, June and July are hot and dry in Salisbury, but in 1995 the FSJEP received over 36.19 cm (14.25 in) of rainfall in those 2 months, while in 2014 only 15.62 cm (6.15 in) fell in those months. Abundance of many species in 1995 was dramatically higher than in 2014, but there was little difference in species diversity, except for the loss of 3 species (Ae. trivittatus, Cx. nigripalpus, and Ps. cyanescens) that utilized the open sunlit habitats which were present in 1995 but not 2014. The loss of those 3 species was overshadowed by a gain of 6 species in 2014 that were not collected in 1995. Given the stark differences in rainfall between the 2 sampling periods we cannot rule out the possibility that the abundance differences we observed were highly influenced by a shortage of breeding habitats for mosquitoes. Indeed, this factor combined with the ecological succession and loss of open-canopy breeding pools are most likely key elements that have led to the overall decline in mosquito populations in the FSJEP. It should also be noted that diversity and abundance measurements of mosquito species attracted to traps are highly variable based on the type of trap and on environmental factors such as precipitation, temperature, wind, and humidity present each time the trap is used. Mosquito-trap catches normally vary widely from day to day and year to year. One year may be a banner year for a species and the next year it may or may not show up (B. Harrison, pers. observ.). Thus, comparing changes in small sample sizes like 16 specimens in 1995 to 3 specimens in 2014 should be considered variations typical of mosquito sampling. Despite changes in overall preserve ecology and total mosquito numbers over 22 years, An. crucians s. l., Ae. atlanticus/tormentor, Ae. c. canadensis, Ae. vexans, Southeastern Naturalist 641 C. Hartwig, et al. 2018 Vol. 17, No. 4 Cx. erraticus, and Ps. ferox were among the most abundant species in both collection periods (Table 2). Based on abundance ranking, stably abundant species in 2013–2015 included Ae. infirmatus, Ae. sticticus, Ae. tormentor, and Ae. vexans, while declining species included An. crucians s.l., Cq. perturbans, and Cx. salinarius. There was a striking increase of Ae. vexans from 5th in abundance in the 1994–1996 period to 2nd in abundance in 2013–2015. Also of notable interest, An. crucians s. l., was ranked 1st in 1994–1996, but dropped to 5th in abundance in 2013–2015. The abundance of An. crucians s. l., collected during 1994–1996 may have been due to exceptionally high precipitation levels in 1995 that created optimal eutrophic water conditions for this species (McLaughin et al. 1987). An even more dramatic decline occurred in Cq. perturbans, which was 4th in abundance in 1994–1996 and decreased to 23rd in abundance in 2013–2015, and Cx. salinarius, which was 5th in abundance in the first period, but declined to 11th in abundance in 2013–2015. During the 1994–1996 sampling period, several large larval collections of Ae. atlanticus were made, while only 1 larva of Ae. tormentor was collected. At that time, females of Ae. atlanticus and Ae. tormentor could not be separated by morphological characters; therefore, specimens looking like these 2 species were called Ae. atlanticus/tormentor. However, in 2014, morphological separation of females of these species became possible using Sither et al. (2013) and drafts of Harrison et al. (2016), and we identified 226 specimens of Ae. tormentor compared to 3 specimens of Ae. atlanticus. We labeled 30 specimens as Ae. atlanticus/tormentor that were too damaged for positive identification. These data strongly suggest that Ae. tormentor is currently more common than Ae. atlanticus in the FSJEP. Both species utilize shaded forest pools for oviposition; thus, increased shade due to forest succession should not have favored one species over the other. However, Roberts and Scanlon (1975) determined that Ae. atlanticus utilizes open fields to search for blood meals from small rodent hosts, while Ae. tormentor remains in the forest to locate blood-meal hosts. The loss of the old-field habitat in the northern half of the FSJEP over the last 22 y might have altered the abundance of Ae. atlanticus in the preserve. During this combined study, determining the presence and abundance of vector species of human arbovirus pathogens was not a major focus, but we would be remiss in not mentioning them. Based on mosquito-borne pathogens documented in North Carolina (NC Department of Health and Human Services, Raleigh, NC, unpubl. data), La Crosse, West Nile, eastern equine encephalomyelitis, Cache Valley, and St. Louis encephalitis viruses have been found as autochthonous human cases in the state. Eleven mosquito species known to vector 1 or more of these viruses have been found in FSJEP. The 4 most-important vector species in NC are uncommon in the preserve, and their abundance levels in FSJEP has remained low over the last 22 y: Ae. albopictus (0.10% in 1994–1996; 0.52% in 2013–2015), Ae. j. japonicus (0% in 1994–1996; and 0.80% in 2013–2015), Ae. triseriatus (0.08% in 1994–1996; 0.80% in 2013–2015), and Cx. pipiens complex (0.02% in 1994–1996; 0.33% in 2013–2015) (Table 2). These 4 important vector-species occur primarily Southeastern Naturalist C. Hartwig, et al. 2018 Vol. 17, No. 4 642 as container-breeding species in urban/suburban areas. The exceptionally low numbers of these vectors is likely a direct reflection of ongoing FSJEP efforts to eliminate trash and containers of all types in the preserve. Nearly all of the specimens of those 4 species were females collected close to adjacent residences outside the preserve. Three of those 4 species also utilize tree holes as larval habitats (Harrison et al. 2016). Low tree holes (below 2 m) were infrequently observed, and higher canopy tree holes (above 2 m) were not sampled in the FSJEP. In addition, the introduced Ae. j. japonicus (Asian Bush Mosquito), did not occur in the US during the 1994–1996 collection period (Peyton et al. 1999). The first collections of this species in North Carolina occurred in 2003 (Gray et al. 2005). By 2004, its distribution in NC had expanded rapidly into a large number of counties including Rowan, and has been reported 16 km north of the FSJEP. Due to a lack of mosquito surveillance, this species was not collected in the FSJEP until 2013, and then only small numbers were collected at site L8 behind the college library. This finding suggests that only small populations have become established in the preserve and/ or they are coming from outside the preserve. We suspect loss of open sunlit habitats due to forest succession is the direct cause for the decline of 4 species: Cq. perturbans, Cx. nigripalpus, Ps. columbiae, and Ps. cyanescens. Although sample sizes for the latter 3 taxa are too small for analysis, the situation of the first species is notable. A total of 595 Cq. perturbans were collected in 1994–1996, but we detected only 5 in 2013–2015. A logical reason for this decline is the loss of sun-loving Cattail plants, which have nearly disappeared in the FSJEP, but may still be present in wetlands adjacent to the preserve. These plants are a preferred host for the unique larvae of this species, which insert their siphon into the roots of the plant to obtain oxygen through the plant aerenchyma. Another very uncommon species, Cs. melanura, increased from 3 in 1994–1996 to 18 in 2013–2015, possibly due to an increase in shaded permanent water sites in the FSJEP. During data analysis, we noted a great similarity in species diversity between the FSJEP and the diversity and abundance of mosquitoes collected in 2 Carolina Bay habitats in South Carolina (Ortiz et al. 2005). Carolina Bays are circular or elliptical bodies of shallow water in palustrine habitats in the coastal plains of North and South Carolina. There are few similarities in plant ecology and topography between the Carolina Bay habitats and the FSJEP. However, there are distinct similarities between the mosquito communities. Ortiz et al. (2005) sampled 2 Carolina Bays and found that the 4 most-abundant species in the first bay, Savage Bay Heritage Preserve (28 ha [70 ac]) were An. crucians s. l. (36.4%), Ae. c. canadensis (20.4%), Ae. atlanticus/tormentor (7.8%), and Cs. melanura (7.3%), which is similar to the 4 most abundant species collected in the FSJEP in 1995: An. crucians s. l. (32.93%), Ae. atlanticus/tormentor (19.52%), Ps. ferox (12.24%), and Cq. perturbans (11.79%). The second bay sampled by Ortiz et al. (2005) was Woods Bay State Park (623 ha [1540 acres]). The 4 most abundant species in that bay were An. crucians s. l. (67.4%), Cq. perturbans (9.9%), Ae. c. canadensis (6.1%), and Ur. sapphirina (4.8%), which is also similar to the 4 most common species collected in the FSJEP in 1995. Although the 4 most abundant species collected in the FSJEP in 2014—Ps. Southeastern Naturalist 643 C. Hartwig, et al. 2018 Vol. 17, No. 4 ferox (28.82%), Ae. vexans (19.89%), Ae. c. canadensis (11.24%), and Ae. tormentor (10.86%)—were less abundant than in 1995, probably due to a drier mosquito season, the species composition is also similar to the collections in the Carolina Bays. Data in this study demonstrate obvious changes in diversity and abundance of mosquito species over a 22-y period in the 0.68-km2 FSJEP. Primary changes occurring in the preserve involved secondary plant succession, that largely eliminated open field habitats (from 26.0 ha in 1994 to 1.8 ha in 2014; Fig. 1) and shifted the mosquito fauna toward those that favor shadier habitats. The very healthy and diverse mosquito fauna discovered in the preserve comprises 95% of known mosquito species in Rowan County, and 54% of the entire mosquito fauna in North Carolina. This large diversity of 1 small arthropod faunal component in a small preserve, when combined with the known diversity of the overall flora and fauna cataloged in the preserve (J. Bolin and J. Cooley, Catawba College, Salisbury, NC, 2012, unpubl. data) demonstrates the complexity, health, and ecological importance of small preserves. This preserve exists beside the Catawba College campus; thus, it provides numerous opportunities for research projects and for training biology and environmental health students in surveillance, identification, data-management skills, and ultimately the complexities of the vast numbers of floral and faunal habitats that can exist in a small area. Moreover, the results of this study suggest future opportunities to document the biological importance of ecologically rich areas such as small urban/ suburban parks, walkways, and preserves. Acknowledgments Volunteer student assistants deserving special thanks are Pamela Casdorph, Gong Cheng, Brian Ethridge, Shannon Garrick, Matthew Jordan-Steele, Lisa Moore, Hannah Przelomski, Cameron Reader, Jennifer Rivera, Tyler Sabolovic, Ashley Wagoner, Kenneth West, and Garrett White. We also appreciate the support of Catawba College, specifically, the Department of Biology and Environmental Programs, and the following individuals for their advice and assistance: Dr. George Drum, Dr. Constance Rogers-Lowery, Dr. John Wear (Catawba College), and Dr. Marc Milne (University of Indiana). We also thank Dr. Nolan Newton, previous Head, Public Health Pest Management, NC Department of Environment and Natural Resources, for his support during the 1990s. Literature Cited Baranski, M.J. 1994. Natural areas inventory for Rowan County, North Carolina. Conservation Trust for North Carolina, Raleigh, NC. Cockburn, A.F., J. Zhang, O.P. Perera, P. Kaiser, J.A. Seawright, and S.E. Mitchell. 1993. A new species of the Anopheles crucians complex: Detection by mitochondrial DNA polymorphism. Pp. 32–35, In D. Borovsky and A. Spielman (Eds.). Host-regulated development mechanisms in vector arthropods. Proceedings of the 3rd Symposium. Vero Beach, FL. Fang, J. 2010. Ecology: A world without mosquitoes. Nature 466:432–434. Gray, E.W., B.A. Harrison, M.L. Womack, J. Kerce, C.J. Neely, and R. Noblet. 2005. The discovery and establishment of Ochlerotatus japonicus japonicus (Theobald) in Georgia and North Carolina. Journal of the American Mosquito Control Association 21:144–146. Southeastern Naturalist C. Hartwig, et al. 2018 Vol. 17, No. 4 644 Harbach, R.E., and I.J. Kitching. 1998. Phylogeny and classification of the Culicidae (Diptera). Systematic Entomology 23:327–370. Harrison, B.A., B.D. Byrd, C.B. Sither, and P.B. Whitt. 2016. The Mosquitoes of the Mid- Atlantic Region: An Identification Guide. Mosquito and Vector-borne Infectious Diseases Laboratory Publication 2016-1, Western Carolina University, Cullowhee, NC. 201 pp. Knight, K.L., and A. Stone. 1977. A Catalog of the Mosquitoes of the World (Diptera: Culicidae), 2nd Edition. Thomas Say Foundation, Entomological Society of America Volume 6. 611 pp. McLaughin, R.E., M.F. Vidrine, and O.R. Willis. 1987. Distribution of Anopheles quadrimaculatus and An. crucians larvae within rice field habitats in southwestern Louisiana. Journal of the American Mosquito Association 3:574–578. Mitchell, A., F.A.H. Sperling, and D.A. Hickey. 2002. Higher-level phylogeny of mosquitoes (Diptera: Culicidae): mtDNA data support a derived placement for Toxorhynchites. Insect Systematics and Evolution 33:163–174. Ortiz, D.I., A. Wozniak, M.W. Tolson, and P.E. Turner. 2005. Arbovirus circulation, temporal distribution, and abundance of mosquito species in two Carolina bay habitats. Vector-Borne and Zoonotic Diseases 5:20–32. Peyton, E.L., S.R. Campbell, T.M. Candeletti, M. Romanowski, and W.J. Crans. 1999. Aedes (Finlaya) japonicus japonicus (Theobald), a new introduction into the United States. Journal of the American Mosquito Control Association 15:238–241. Reinert, J.F., P.E. Kaiser, and J.A. Seawright. 1997. Analysis of the Anopheles (Anopheles) quadrimaculatus complex of sibling species (Diptera: Culicidae) using morphological, cytological, molecular, genetic, biochemical, and ecological techniques in an integrated approach. Journal of the American Mosquito Control Association 13 (Supplement): 1–102. Roberts, D.R., and J.E. Scanlon. 1975. The ecology and behavior of Aedes atlanticus D and K and other species with reference to Keystone virus in the Houston area, Texas. Journal of Medical Entomology 12:537–546. Rueda, L.M. 2008. Global diversity of mosquitoes (Insecta: Diptera: Culicidae) in freshwater. Hydrobiologia 595:477–487. Schafale, M.P. 2012. Guide to the Natural Communities of North Carolina, 4th Approximation. North Carolina Natural Heritage Program, Raleigh, NC. 208 pp. Sither, C.B., V.E. Hopkins, B.A. Harrison, B.J. Bintz, E.Y. Hickman, J.S. Brown, M.R. Wilson, and B.D. Byrd. 2013. Differentiation of Aedes atlanticus and Aedes tormentor by restriction fragment length polymorphisms of the second internal transcribed spacer. Journal of the American Mosquito Control Association 29:376–379. Slaff, M., and C.S. Apperson. 1989. Key to the mosquitoes of North Carolina and the mid- Atlantic states (AG-412). Agricultural Extension Service, North Carolina State University, Raleigh, NC. 38 pp. Wilkerson, R.C., and Y-M. Linton. 2015. Elevation of Pseudoskusea, Rusticoidus, and Protomacleaya to valid subgenera in the mosquito genus Aedes based on taxon naming criteria recently applied to other members of the Tribe Aedini (Diptera: Culicidae). Parasites and Vectors 8:668. DOI:10.1186/s13971-015-1247-x. Wilkerson, R.C., J.F. Reinert, and C. Li. 2004. Ribosomal DNA ITS2 sequences differentiate 6 species in the Anopheles crucians Complex (Diptera: Culicidae). Journal of Medical Entomology 41:392–401. Wilkerson, R.C., Y-M. Linton, D.M. Fonseca, T.E. Schultz, D.C. Price, and D.A. Strickman. 2015. Making mosquito taxonomy useful: A stable classification of Tribe Aedini that balances utility with current knowledge of evolutionary relationships. PLoS ONE 10(7):e0133602. DOI:10.1371/journal.pone.0133602.