2010 NORTHEASTERN NATURALIST 17(4):541–564 Freshwater Mussel (Bivalvia: Unionidae) Distributions and Habitat Relationships in the Navigational Pools of the Allegheny River, Pennsylvania Tamara A. Smith1,2 and Elizabeth S. Meyer3,* Abstract - The main-stem Allegheny River is nationally recognized for its freshwater mussel (Unionidae: Bivalvia) diversity; however, habitat disturbance and degradation may have triggered the decline and loss of mussel communities in the lower river, where lock and dam structures restrict the free flow of water and sand and gravel removal threaten limited habitat. We examined mussel diversity and abundance across 75 transects throughout navigational pools and recorded 21 live native mussel species, including federally endangered Pleurobema clava (Clubshell) and Epioblasma torulosa rangiana (Northern Riffleshell) and several species with state endangered or threatened status. Riverine species richness and counts were significantly higher in the most-upstream portions of the upper pools, indicating that areas with consistent flows and suitable substrate just downstream of the dams may provide refugia for riverine freshwater mussel species. Sand, gravel, cobble, boulder, and organic debris had significant positive effects on riverine and facultative counts, while clay, bedrock, and woody debris had significant negative effects. Silt and woody debris had significant negative effects on riverine species richness, and sand and gravel had significant positive effects. These data will help identify sensitive areas for future protection and provide baseline data for monitoring future trends. The protection of relatively shallow areas with suitable substrates not yet impacted by dredging operations will be important to sustain remaining freshwater mussel populations in these pools. Introduction Freshwater mussels are considered the most imperiled fauna in North America, with approximately 213 of the 297 recognized taxa considered endangered, threatened, or of special concern (Lydeard et al. 2004, NatureServe 2009, Ricciardi and Rasmussen 1999, Williams et al. 1993). Several species occur at their highest densities within their global range in the Allegheny River system, including Epioblasma torulosa rangiana (Northern Riffleshell) and Pleurobema clava (Clubshell), both federal and state endangered species. The Allegheny River contains two candidates for federal endangered listing status; Villosa fabalis (Rayed Bean) and Plethobasus cyphyus Rafinesque (Sheepnose). In addition, Epioblasma triquetra Rafinesque (Snuffbox), listed as endangered in Pennsylvania and 1Western Pennsylvania Conservancy, Northwest Field Station, 11881 Valley Road, Union City, PA 16438. 2Current address - US Fish and Wildlife Service, Twin Cities Ecological Services Field Office, 4101 American Boulevard East, Bloomington, MN 55425. 3Western Pennsylvania Conservancy, 800 Waterfront Drive, Pittsburgh, PA 15222. *Corresponding author - emeyer@paconserve.org. 542 Northeastern Naturalist Vol. 17, No. 4 two additional species of special concern in Pennsylvania, Amblema plicata (Three-ridge) and Pleurobema sintoxia (Round Pigtoe), are also found in the Allegheny River (NatureServe 2009). The diverse mussel fauna of the Allegheny River faces threats from habitat alteration, loss, and degradation. A series of 8 lock-and-dam structures were constructed on the lower Allegheny River during the late 1890s to the 1930s to create deep waters for navigation purposes. In addition, the Kinzua Dam, located 218 river kilometers upstream of Lock and Dam 9 near Warren, PA, is managed for flood control and to maintain flows during dry periods (US Army Corps of Engineers 2010a). Stream channel alterations and dams are documented threats to the viability of freshwater mussels (Watters 2000). Furthermore, active sand and gravel mining occurs within the navigational pools, and may permanently alter mussel habitat (Brown et al. 1998, Hubbs et al. 2003, Kondolf 1997, Meador and Layher 1998). However, areas within these channels not impacted by dredging and with consistent flows may serve as refugia for riverine freshwater mussel and host fish species. Early studies documented approximately 35 species in the Allegheny River (Ortmann 1919), and many of these species are now thought to be extirpated in this portion of the river. The portion of the Allegheny River downstream of the Kinzua Dam in Pennsylvania still maintains populations of approximately 30 unionid species (Villella and Nelson 2006). French Creek, a tributary to the Allegheny River, still holds over 26 species (Smith and Crabtree, in press). Recent project-specific surveys have been conducted in the navigational pools, usually as a direct response to dredging or construction permits (US Fish and Wildlife 2004). However, no comprehensive mussel study has been completed in the impounded navigational pools in the Allegheny River. The goals of this study were to initiate a comprehensive study of the freshwater mussel populations in the lower Allegheny River navigational pools 4 through 8 and to look for patterns in mussel distribution in relation to environmental variables and the position within the river. Identifying areas where rare mussels persist in these pools and environmental clues as to where others are likely to exist would provide baseline data that could inform decisions about protection and restoration efforts. For example, this data could be used to help decide locations of ecological reserves or make informed decisions on where to focus restoration efforts. Additionally, this information could lead to protection efforts for remaining freshwater mussel habitat, such as locations with designated limitations on commercial sand and gravel mining operations. Methods Study location The Allegheny River (610 km) flows from Potter County, PA north through Cattaraugus County in New York State, then flows south in 2010 T.A. Smith and E.S. Meyer 543 Pennsylvania until its confluence with the Monongahela River to form the Ohio River in Pittsburgh, PA. A series of locks and dams were constructed on the Allegheny River between 1897 and 1938, resulting in 116 km of slackwater navigation from Pittsburgh to just above East Brady, PA (US Army Corps of Engineers 2010b). This study took place in navigational pools 4, 5, 6, 7, and 8 (Fig.1). Figure 1. Transect locations in the Allegheny River navigational pools 4, 5, 6, 7, and 8. The insert magnifies a portion of pool 6, where transect locations were dense. River kilometers are abbreviated as RKM. 544 Northeastern Naturalist Vol. 17, No. 4 Transect selection Our surveys targeted relatively undisturbed areas believed to be suitable habitat for mussels. We sampled relatively shallow areas (<15 m deep). In 2005 and 2006, we determined potential habitat using an Eagle FishMark 320 depth finder. In 2007, we used newly collected bathymetric mapping data (Long and Chapman 2008) to determine potential habitat throughout pools 5 through 8. We selected transects where the channel had a gradual slope from shore out to mid-channel. Areas with abrupt changes in depths were suspected to have been recently dredged and were avoided if possible. Although we attempted to survey at least one transect per river km, some areas were avoided because of extremely deep water depths (e.g., portions of pools 5 and 7). Transects are represented by river km (RKM) as measured upstream from the confluence with the Monongahela River. Areas with recent surveys (i.e., RKMs 93.7 to 94.3) that used similar protocols (US Fish and Wildlife Service 2004) were also avoided to reduce handling disturbance, which can affect growth (Haag and Commens- Carson 2008). Survey methodology We followed protocols developed for the Ohio and Allegheny Rivers (ORVET 2004) to survey for mussels. Weighted transect lines (100 m by 1 m) were placed perpendicular to river flow. Each transect was divided into 10-m segments, and each segment was searched for at least 5 minutes and until all visible mussels were collected. Transect segments were generally not searched if the substrate was 100% bedrock or had at least 25 cm of silt deposition, although we did search some segments with thick silt substrate using less effort (<5 minutes segment). Substrate was characterized and maximum water depth was recorded for each 10-m segment. The percent of substrate types, defined by particle sizes in a modified Wentworth scale (Wentworth 1922), was visually estimated and recorded as percent area of each substrate type (silt, sand, gravel, cobble, boulder, bedrock, woody debris, and organic debris). In this study, woody debris was defined as relatively large pieces of wood (≥15 cm), while organic debris typically consisted of a mix of relatively small pieces of wood (<15 cm), leaves and other organic materials. Mussels within each segment were placed into mesh bags, identified, counted, and measured to the nearest mm in total length. All mussels were released via broadcast from the surface along the transect line immediately following processing, with the exception of federally listed species, which were placed back in the substrate by hand. All taxonomy followed Turgeon et al. (1998). We surveyed by SCUBA in pairs when conditions were safe for diving and the discharge measured on the Allegheny River at Kittanning, PA (US Geological Survey stream gage; www.waterdata.usgs.gov) was preferably 2010 T.A. Smith and E.S. Meyer 545 less than 8000 cubic feet per second (cfs). One dive was at 10,400 cfs and another was at 9130 cfs. The surveys did not follow rain events, and visibility was always >1 m. Analyses Each species found in this survey was assigned to a category based on accounts of the biology and habitat preferences (Parmalee and Bogan 1998) combined with observational experience specific to the Allegheny River. Riverine species were defined as those that tend to live in faster flowing systems and are less tolerant of silt. Facultative species were defined as species that tend to live in slower areas of otherwise flowing systems. Facultative species are silt tolerant to a degree, but tend not to occur in completely lentic habitats such as deep silt layers. Lentic species were defined as those that tend to live in non-flowing waters. We used a generalized linear mixed-effects model to analyze mussel count data and species richness as a function of environmental variables using R version 2.10.0 (R Development Core Team 2009). Fixed effects were maximum depth, distance from the nearest upstream dam, river km (RKM), and each of the substrate types. We also included maximum depth as a nonlinear factor (maximum depth squared) in the model. Transect was included in the model as a random effect, because the variance between transects was expected to be greater than within-transect (between segment) variance. We examined correlation between the covariates to ensure that colinearity would not influence the parameter estimates. To account for overdispersion (many zeros) in the data, we used a quasi-likelihood Poisson procedure, which allows the estimation of model parameters without fully knowing the error distribution of the response variable (McCullagh and Nelder 1989). Variables were considered significant at P = 0.05 level. We used logistic regression to analyze presence or absence of the 10 most common mussels (number of individuals >35) as a function of environmental variables using R version 2.10.0 (R Development Core Team 2009). Transect was included in the model as a random effect, and fixed effects were maximum depth, maximum depth squared, distance from the nearest upstream dam, river km (RKM), and each of the substrate types. Variables were considered significant at P = 0.05 level. We used size criteria to examine the number of juvenile mussels because of the ease and consistency of measuring lengths in the field. Juveniles were defined as individuals less than or equal to 30 mm in total length; other studies have used similar size limits (e.g., Chapman and Smith 2008; Mohler et al. 2006; Obermeyer 1998; Smith and Crabtree, in press). We used a cut-off of 20 mm for Rayed Bean, which is a naturally smaller species (Cummings and Mayer 1992). Sex ratios of sexually dimorphic species were also examined. The probability of detecting a mussel species relates to the sampling effort and search efficiency within the sample area, the distribution of 546 Northeastern Naturalist Vol. 17, No. 4 Table 1. Global and Pennsylvania State ranks for each species found during this study, and number of live found in each pool. Species only found as dead specimens in a particular pool are represented by asterisks. Key to global ranks: G5 = secure, G4 = apparently secure, G3 = vulnerable, G2 = imperiled, G1 = critically imperiled, T2 = subspecies. Key to state ranks: S5 = secure, S4 = apparently secure, S3 = vulnerable, S2 = imperiled, S1 = critically imperiled, SNR = not ranked. Ranks according to NatureServe (2009). Northern Riffleshell and Clubshell are listed as federally endangered and the Rayed Bean is a candidate for federal listing. Each species was categorized as primarily riverine (R), facultative (F), or lentic (L) according to Parmalee and Bogan (1998). Number live per pool Species Riverine, facultative, or lentic Global rank PA State rank 4 5 6 7 8 Actinonaias ligamentina (Lamarck) (Mucket) R G5 S5 1 148 611 490 1620 Alasmidonta marginata Say (Elktoe) R G4 S4 2 1 * Amblema plicata (Say) (Three-ridge) L G5 S2 S3 * * Elliptio dilatata (Rafinesque) (Spike) R G5 S4 * 114 1139 488 827 Epioblasma torulosa rangiana (Lea) (Northern Riffleshell) R G2 T2 S2 3 Dreissena polymorpha Pallas (Zebra Mussel) Exotic Exotic Exotic 73 2 Fusconaia flava (Rafinesque) (Wabash Pigtoe) F G5 S2 37 27 5 1 Fusconaia subrotunda (I. Lea) (Long-solid) R G3 S1 2 * Lampsilis cardium Rafinesque (Plain Pocketbook) F G5 S4 * 14 8 13 Lampsilis fasciola Rafinesque (Wavy-rayed Lampmussel) R G5 S4 1 Lampsilis ovata (Say) (Pocketbook) R G5 S3 S4 10 * 25 Lampsilis siliquoidea (Barnes) (Fatmucket) F G5 S4 2 35 121 31 49 Lasmigona costata (Rafinesque) (Fluted-shell) F G5 S4 25 124 91 48 Leptodea fragilis (Rafinesque) (Fragile Papershell) F G5 S2 1 2 26 18 40 Ligumia recta (Lamarck) (Black Sandshell) R G5 S3 S4 1 5 18 10 24 Pleurobema clava (Lamarck) (Clubshell) R G2 S1 S2 1 Pleurobema sintoxia (Rafinesque (Round Pigtoe) R G4 S2 3 1 Potamilus alatus (Say) (Pink Heelsplitter) F G5 S2 10 53 42 7 Potamilus ohiensis (Rafinesque) (Pink Papershell) F G5 SNR * Ptychobranchus fasciolaris (Rafinesque) (Kidneyshell) R G4 G5 S4 5 Pyganodon grandis (Say) (Giant Floater) L G5 S4 * 1 Simpsonaias ambigua (Say) (Salamander Mussel) F G3 S1 4 2 Strophitus undulatus (Say) (Creeper) F G5 S4 S5 1 3 Villosa fabalis (I. Lea) (Rayed Bean) R G1 G2 S1 S2 1 * 11 Villosa iris (I. Lea) (Rainbow) R G5 S1 * 2010 T.A. Smith and E.S. Meyer 547 sampling effort, and the abundance and spatial distribution of the species (Smith 2006). An unbiased estimation of abundance or density is only possible when the fraction of the survey area and the search efficiency is known or estimated, and in general, search efficiency increases with increasing search time (Smith 2006). We assumed equal detectability of all species for our analyses, calculated our overall search rate and survey area as a means of comparison to other studies, and calculated a mean overall search rate for our study from the search times we recorded for each 10-m segment. The fraction of the pools that was searched was calculated as the sum of the areas of each unit (10-m2 segments) in the entire survey (75 transects) divided by the total area of pools 4 through 8. Results A total of 75 transects were surveyed; six transects were sampled in pool 4, eight in pool 5, 31 in pool 6, 10 in pool 7, and 20 transects in pool 8 (Fig. 1, Table 1). Our mean search time per 10-m2 segment was 12.5 minutes (SE = 6.32), which gives a mean search rate of 1.25 min/m2. The fraction the total pool area surveyed was 0.08% (Table 2). Survey depths ranged from 0.5 to 13 m. No live mussels were found >10 m deep; over 95% of all live mussels were found in <6 m deep. Overall, we recorded 6403 live native Unionids from 21 species (Table 2). Sixty-one fresh dead and 2905 weathered dead shells were found, adding three species that were found only as shells. We also found the invasive exotic Dreissena polymorpha Pallas (Zebra Mussel) in 4 transects (Table 3). Table 2. Transect summary data for each pool. The mean, standard error of the mean (SE), minimum, and maximum number of individuals and number of species found per 100-m transect are given. Minimum, maximum, and mean maximum depths per 10-m segment in pool are also given. Numbers do not include Zebra Mussels. The fraction of the survey area was calculated as the sum of survey areas divided by the total area of pool and is given as a percentage. PL = pool length (km), PA = pool area (km2), # = number of transects, % = percent area surveyed. Segment maximum Count per Number of species depth (m) transect per transect Mean Mean Mean Pool PL PA # % Min Max (SE) Min Max (SE) Min Max (SE) 4 9.9 3154 6 0.03 1.8 9.8 4.4 1 5 2.8 1 3 1.8 (0.25) (0.60) (0.31) 5 9.5 2663 8 0.06 1.7 9.1 4.3 2 208 62.4 1 8 5.5 (0.18) (23.91) (0.78) 6 15.1 4910 31 0.12 0.5 7.9 3.5 1 262 69.3 1 10 6.2 (0.09) (11.73) (0.44) 7 11.1 3093 10 0.05 0.6 12.5 3.8 0 463 115.0 0 10 5.1 (0.34) (174.08) (1.18) 8 15.5 3954 20 0.09 1.2 13.1 4.0 0 492 133.3 0 10 4.2 (0.14) (39.65) (0.94) 548 Northeastern Naturalist Vol. 17, No. 4 Table 3. Summary data for each species found in our surveys including total numbers (count) and number of transects in which we found the species (transects occupied out of 75). Minimum, maximum, and mean total length with standard error (SE) mean total length is given. Sex of all individuals, number of juvenile mussels, and number of transects with juveniles is also given (sex and size data may be biased since we assumed equal detectability for all species). Length of all individuals (mm) Sex of all individuals Juveniles Species Count Transects occupied Min Max Mean (SE) F M Unknown Count Transects occupied Actinonaias ligamentina 2870 47 43.0 163.0 112.5 (0.65) 0 0 2870 0 0 Alasmidonta marginata 3 3 60.0 66.4 63.8 (1.94) 0 0 3 0 0 Elliptio dilatata 2568 53 13.0 144.0 104.0 (0.43) 0 0 2568 5 5 Epioblasma torulosa rangiana 3 2 42.0 48.0 44.7 (2.19) 2 1 0 0 0 Dreissena polymorpha 74 4 9.0 31.0 10.4 (0.36) 0 0 74 NA NA Fusconaia flava 70 24 30.0 30.0 71.1 (2.57) 0 0 70 1 1 F. subrotunda 2 1 53.2 60.3 NA 0 0 2 0 0 Lampsilis cardium 35 20 90.0 145.0 119.0 (3.30) 7 17 11 0 0 L. fasciola 1 1 86.0 86.0 NA 0 1 0 0 0 L. ovata 35 12 82.0 151.0 117.2 (3.49) 8 25 2 0 0 L. siliquoidea 238 49 8.0 158.0 113.2 (1.40) 54 97 87 1 1 Lasmigona costata 288 50 36.0 144.0 111.5 (0.90) 0 0 288 0 0 Leptodea fragilis 87 30 34.0 119.0 73.9 (2.43) 0 0 87 0 0 Ligumia recta 58 28 93.9 185.0 135.2 (3.18) 12 34 10 0 0 Pleurobema clava 1 1 31.9 31.9 NA 0 0 1 0 0 P. sintoxia 4 4 72.0 106.0 91.8 (7.35) 0 0 4 0 0 Potamilus alatus 112 36 49.0 181.0 116.0 (2.96) 0 0 112 0 0 Ptychobranchus fasciolaris 5 4 100.0 117.0 110.3 (3.61) 0 0 5 0 0 Pyganodon grandis 1 1 103.0 103.0 NA 0 0 1 0 0 Simpsonaias ambigua 6 2 31.0 44.0 36.0 (1.77) 0 0 6 0 0 Strophitus undulatus 4 4 43.0 76.0 56.5 (7.15) 0 0 4 0 0 Villosa fabalis 12 8 14.0 30.0 20.2 (1.50) 4 7 1 7 4 2010 T.A. Smith and E.S. Meyer 549 Thirteen out of 24 species were classified as vulnerable, imperiled, or critically imperiled at a global or state level, including the federally endangered Northern Riffleshell and Clubshell (Table 1). Thirteen species were classified as riverine, 9 as facultative, and 2 as primarily lentic species (Table 1). We found 14 total juveniles from four species: Elliptio dilatata (Spike), Fusconaia flava (Wabash Pigtoe), Lampsilis siliquoidea (Fatmucket), and Rayed Bean (Table 3). Both sexes were represented for each sexually dimorphic species, and in general, more males were found than females (Table 3). Species composition varied among pools. The most ubiquitous species were Actinonaias ligamentina (Mucket) and Spike, which were present in every pool (Table 1) and were the dominant species in pools 5 through 8. Mucket was found at 47 transects and accounted for about 44.3% of the total number of mussels found, while Spike was found at 53 transects and accounted for about 39.7% of the total number of mussels found (Table 3). Other species found at a relatively large number of transects but at low relative abundance were Lasmigona costata (Fluted Shell) (4.5%, 50 transects), and Fatmucket (3.7%, 49 transects), Leptodea fragilis (Fragile Papershell) (1.3%, 30 transects), Ligumia recta (Black Sandshell) (0.9%, 28 transects), and Wabash Pigtoe (1.1%, 24 transects). Potamilus alatus (Pink Heelsplitter) (1.7%, 36 transects) was the dominant species found in pool 4 and was the third-most abundant species in pool 5. Species with the most limited distributions were Northern Riffleshell, Fusconaia subrotunda (Long-solid), Lampsilis fasciola (Wavy-rayed Lampmussel), Pyganodon grandis (Giant Floater), Clubshell, and Ptychobranchus fasciolaris (Kidneyshell) (Table 3). Individual pool results Pool 4. Only six species were found, and total counts were generally low in pool 4 (Tables 1 and 2). Pink Heelsplitter accounted for 10 of the 17 live individuals found. Spike, Giant Floater, and Potamilus ohiensis (Pink Papershell) were found only as dead shells. Nearly half of the surveyed segments were comprised of over 50% silt substrate. Pool 5. Counts were low in pool 5 (Tables 1 and 2) except at the uppermost transects. Richness ranged from 1 to 8 per transect, with 10 species found throughout the pool. Substrate composition was generally a mix of cobble and gravel, with some silt. Mucket and Spike were the dominant species in the three transects in the upper pool (RKM 56.9, 57.3, and 57.4), where over 78% of the Unionids in this pool were found. Wabash Pigtoe and Pink Heelsplitter were dominant in transects surveyed in the lower portions of the pool between RKMs 50 to 54. Simpsonaias ambigua (Salamander Mussel) was found only as 4 live individuals in the lower pool, under large flat rocks in a transect that was dominated by gravel with a maximum depth of 2.3 m. Lampsilis cardium (Plain Pocketbook) was found only as weathered shells. We found 73 live Zebra Mussels; 72 were in one transect (RKM 52.9), and most were <10 mm in total length and attached to Unionids. 550 Northeastern Naturalist Vol. 17, No. 4 Pool 6. In pool 6, counts were high, and species richness was the highest (Tables 1 and 2). We found 17 live species throughout the pool, which is the most species we recorded in any of the pools. Substrate composition was generally a mix of sand, gravel, cobble, and boulder, except in a few transects that were dominated by silt (i.e., RKM 59.5, 59.9, and 64.2). Rayed Bean was found live in one transect in the upper pool, and Salamander Mussel was found live under a large flat rock in another transect in the upper half of the pool. In addition, one dead Villosa iris (Rainbow) was encountered. We found two juvenile Spikes between RKMs 71 and 73 and one juvenile Wabash Pigtoe near RKM 63. We found two live and two dead Zebra Mussels in pool 6. Pool 7. Counts were generally low in Pool 7 (Tables 1 and 2); however, it is the only pool in which we documented Three-ridge in our study, which was found as a weathered dead shell. No surveys occurred in the lower portions of pool 7 due to extreme water depths. Three transects surveyed below a railroad bridge, an area restricted from dredging operations, were unlike the other 7 transects surveyed in pool 7. In these transects, we documented 8 to 10 live species and 289 to 463 live individuals per transect. The substrate was comprised mainly of sand, gravel, cobble, and boulder substrates, and maximum depths ranged from 0.6 to 4.0 m. In this section of intact habitat below the railroad bridge, we found live Clubshell and the only juvenile mussel (Spike) in this pool. Outside of those areas, total abundances ranged from 0 to 29 live individuals per transect (mean total abundance = 10.4, SE = 4.15), with 0 to 7 live species per transect (mean species richness = 3.3, SE = 1.04). Substrate in these transects was dominated by silt and boulders, and maximum depths per segment ranged from 2.4 to 12.5 m. Mean maximum depth per 10-m segment in the three transects below the railroad bridge was 2.06 m (SE = 0.156), and mean maximum depth in all other transects was 5.33 m (SE = 0.482). Pool 8. Maximum depths per 10-m segment in pool 8 ranged from 1.2 to 13.1 m, with a mean maximum depth of 3.96 m (SE = 0.138, Table 2). Mean depth per 10-m segment was 3.58 m (SE = 0.098) in the upper portion of the pool (RKM 94.6 to 98.6) and 4.29 m (SE = 0.250) downstream of RKM 91.9. Species richness and abundances above RKM 94.6 in pool 8 were unlike segments surveyed downstream of RKM 91.9. No transects were surveyed between RKM 91.9 and 94.6 in this study. Surveyed transects between RKM 94.6 and 98.6 had total counts ranging from 222 to 492 live individuals (mean total abundance = 329.6, SE = 38.52), with 6 to10 live species (mean species richness = 8.9, SE = 0.52). These transects in the upper portion of the pool generally had a mixture of sand, gravel, cobble, and boulder substrate with maximum depths less than 4.9 m. Live Northern Riffleshell was found in two transects in the upper portion of pool 8 in segments with a mixture of sand, gravel, cobble, and boulder substrate and at maximum depths of approximately 3.5 m. Transects that contained live Northern Riffleshell had 2010 T.A. Smith and E.S. Meyer 551 relatively high species richness (10 species) and total counts (358 and 483). Rayed Bean was found live at several locations in the upper portion of pool 8, and we also documented Wavy-rayed Lampmussel in the upper part of the pool. In addition, one dead Alasmidonta marginata (Elktoe) was found. We found 10 juvenile mussels in the upper portion of pool 8: seven Rayed Bean, one Fatmucket, and two Spikes. Transects surveyed below RKM 91.9 in pool 8 had total abundances ranging from 0 to12 live individuals (mean total abundance = 2.3, SE = 1.06) and 0 to 4 species (mean species richness = 1.0, SE = 0.39). Transects in the lower portion of the pool had generally deeper segments and high percentages of silt. No juvenile mussels were found at these downstream transects. In-stream habitat relationships The relationships between counts and richness to maximum depth are not linear. The results of the mixed effects analyses show that maximum depth had significant positive effects on total counts and richness (Table 4), riverine species counts and richness (Table 5), and facultative species counts and richness (Table 6). Conversely, maximum depth squared shows signifi- cant negative effects on total, riverine, and facultative counts and species richness (Tables 4, 5, and 6). These analyses show the significance of the patterns observed: low counts and richness at low depths, their peak at intermediate depths (approximately 4 m), and subsequent decline to zero at the highest maximum depths (Fig. 2). Table 4. Results of the generalized linear mixed-effects model (656 observations, 74 groups). Response variables were total mussel counts and total species richness. The estimate, standard error (SE), and t-value is given for each fixed effect. Variance and standard deviation (std. dev.) are given for each random effect. Environmental variables were considered significant at P < 0.05 level (shown in bold). Total count Species richness Fixed effects Estimate SE t-value Estimate SE t-value Intercept -0.642 0.214 -3.000 -6.03E-01 2.41E-01 -2.505 Clay -0.003 0.001 -2.370 7.05E-03 3.04E-03 2.317 Silt 0.000 0.001 -0.740 -2.77E-05 1.33E-03 -0.021 Sand 0.001 0.000 1.150 2.67E-03 1.37E-03 1.952 Gravel 0.002 0.000 3.890 4.18E-03 1.35E-03 3.106 Cobble 0.001 0.000 2.870 2.00E-03 1.22E-03 1.640 Boulder 0.002 0.000 4.190 2.50E-03 1.42E-03 1.756 Bedrock -0.022 0.001 -21.370 -6.10E-03 2.43E-03 -2.508 Organics 0.069 0.004 19.090 3.50E-02 6.71E-03 5.215 Woody debris -0.022 0.001 -28.360 -1.07E-02 2.62E-03 -4.068 Max. depth 0.831 0.016 52.690 5.30E-01 4.17E-02 12.691 Max. depth2 -0.099 0.002 -45.170 -6.51E-02 4.98E-03 -13.057 RKM 0.028 0.003 10.540 1.22E-02 2.39E-03 5.086 Distance dam -0.371 0.011 -33.940 -2.22E-01 1.05E-02 -21.145 Random effects Variance Std. dev. Variance Std. dev. Transect 0.107 0.327 0.072 0.267 Residual 0.051 0.226 0.133 0.364 552 Northeastern Naturalist Vol. 17, No. 4 Table 5. Results of the generalized linear mixed-effects model (656 observations, 74 groups). Response variables were total riverine mussel counts and total riverine species richness. The estimate, standard error (SE), and t-value is given for each fixed effect. Variance and standard deviation (Std. Dev.) are given for each random effect. Environmental variables were considered significant at P < 0.05 level (shown in bold). Riverine total count Riverine species richness Fixed effects Estimate SE t-value Estimate SE t-value Intercept -2.695 0.141 -19.170 -1.783 0.173 -10.326 Clay -0.070 0.004 -19.820 -0.011 0.005 -2.199 Silt 0.002 0.000 5.190 -0.005 0.001 -4.426 Sand 0.002 0.000 7.310 0.004 0.001 4.045 Gravel 0.004 0.000 12.440 0.007 0.001 6.410 Cobble 0.002 0.000 9.010 0.003 0.001 2.657 Boulder 0.004 0.000 12.830 0.005 0.001 4.232 Bedrock -0.022 0.001 -35.640 -0.003 0.002 -1.637 Organics 0.057 0.003 20.310 -0.011 0.009 -1.214 Woody debris -0.021 0.000 -46.820 -0.009 0.002 -4.030 Maximum depth 0.827 0.010 82.600 0.419 0.035 12.103 Maximum depth2 -0.101 0.001 -69.790 -0.056 0.004 -13.370 RKM 0.052 0.002 30.040 0.022 0.002 14.486 Distance dam -0.449 0.007 -59.950 -0.198 0.007 -28.250 Random effects Variance Std. dev. Variance Std. dev. Transect 0.042 0.206 0.022 0.148 Residual 0.015 0.122 0.056 0.237 Table 6. Results of the generalized linear mixed-effects model (656 observations, 74 groups). Response variables were total facultative mussel counts and total facultative species richness. The estimate, standard error (SE), and t-value is given for each fixed effect. Variance and standard deviation (Std. Dev.) are given for each random effect. Environmental variables were considered significant at P < 0.05 level level (shown in bold). Facultative total count Facultative species richness Fixed effects Estimate SE t-value Estimate SE t-value Intercept -2.695 0.141 -19.170 -0.808 0.254 -3.179 Clay -0.070 0.004 -19.820 0.011 0.004 3.165 Silt 0.002 0.000 5.190 0.001 0.002 0.347 Sand 0.002 0.000 7.310 0.004 0.002 2.137 Gravel 0.004 0.000 12.440 0.003 0.002 1.961 Cobble 0.002 0.000 9.010 0.003 0.001 2.262 Boulder 0.004 0.000 12.830 0.001 0.002 0.395 Bedrock -0.022 0.001 -35.640 -0.007 0.004 -1.680 Organics 0.057 0.003 20.310 0.034 0.007 5.175 Woody debris -0.021 0.000 -46.820 -0.007 0.004 -1.991 Maximum depth 0.827 0.010 82.600 0.424 0.053 7.945 Maximum depth2 -0.101 0.001 -69.790 -0.056 0.006 -8.85 RKM 0.052 0.002 30.040 0.004 0.002 1.808 Distance dam -0.449 0.007 -59.950 -0.172 0.010 -17.163 Random effects Variance Std. dev. Variance Std. dev. Transect 0.042 0.206 0.046 0.214 Residual 0.015 0.122 0.125 0.354 2010 T.A. Smith and E.S. Meyer 553 Distance from the nearest upstream dam had significant negative effects on total counts and total species richness, as well as riverine and facultative counts and richness (Tables 4–6, Fig. 2). River kilometer (RKM) had significant positive effects on total, riverine, and facultative species richness and counts (Tables 4–6, Fig. 2), meaning that segments in the upper pools (higher RKM) had relatively high counts and species richness. Gravel, boulder, and organic debris had significant positive effects on total species richness and total counts (Table 4). Cobble had significant positive effects on total counts, while clay, bedrock, and woody debris had significant negative effects (Table 4). Sand and clay had significant positive effects on total species richness, while bedrock and woody debris had significant negative effects (Table 4). Sand, gravel, cobble, boulder, organic debris, and silt had significant positive effects on riverine and facultative counts, while clay, Figure 2. Scatterplot matrix of total riverine and facultative species counts (River. Count and Fac.Count, respectively) and riverine and facultative species richness (River.SPP and Fac.SPP, respectively) and maximum depth (m; Max.Depth), distance (km) from the nearest upstream dam (Dist.Dam), and river km (RKM) per 10-m segment. 554 Northeastern Naturalist Vol. 17, No. 4 bedrock, and woody debris had significant negative effects (Tables 5 and 6). Silt, clay, and woody debris had significant negative effects on riverine species richness and sand, gravel, cobble, and boulder had significant positive effects (Table 5). Clay, sand, gravel, cobble, and organic debris had signifi- cant positive effects on facultative species richness, while woody debris had significant negative effects (Table 6). Results from the logistic regression analyses on presence-absence data show that distance from nearest upstream dam had significant negative effects and RKM had significant positive effects on Mucket, Spike, Fatmucket, Fragile Papershell, Black Sandshell, Lampsilis ovata (Pocketbook), and Plain Pocketbook presence (Table 7). River km (RKM) had significant negative effects on Wabash Pigtoe and Pink Heelsplitter presence (Table 7). Maximum depth had significant positive effects on Mucket (Est. = 1.211, SE = 0.528, Z value = 2.294, Pr(>|z|) = 0.022) and Spike (Est. = 2.149, SE = 0.539, Z value = 3.990, Pr(>|z|) = 0.000) and Fragile Papershell (Est. = 1.121, SE = 0.570, Z value = 1.967, Pr(>|z|) = 0.049). Maximum depth squared had significant negative effects on Mucket (Est. = -0.142, SE = 0.061, Z value = -2.335, Pr(>|z|) = 0.020), Spike (Est. = -0.225, SE = 0.059, Z value = -3.838, Pr(>|z|) = 0.000), and Fatmucket (Est. = -0.102, SE = 0.049, Z value = -2.091, Pr(>|z|) = 0.037). These results indicate a nonlinear relationship between maximum depth and Mucket, Spike, and Fatmucket. In other words, after an intermediate depth (approximately 4 m), the presence of these three species declines. Silt had significant negative effects on Pocketbook (Est. = -0.074, SE = 0.037, Z value = -1.988, Pr(>|z|) = 0.047) and Plain Pocketbook (Est. = -0.040, SE = 0.020, Z value = 2.011, Pr(>|z|) = 0.044) presence. Sand (Est. = 0.041, SE = 0.016, Z value = 2.513, Pr(>|z|) = 0.012) and gravel (Est. = 0.032, SE = 0.015, Z value = 2.097, Pr(>|z|) = 0.036) had significant positive effects on Mucket presence. All other environmental variables were not statistically significant. Table 7. Results of the logistic regression (656 observations, 74 groups) on the presence or absence of the 10 most common species (number of individuals > 35). Environmental variables were considered significant at P < 0.05 level; fixed effects not presented in this table or in the text were not significant. Distance from nearest upstream dam was not significant for F. flava (Wabash Pigtoe) and river kilometer (RKM) was not significant for L. costata (Flutedshell). Distance from nearest upstream dam RKM Est. SE Z value Pr(>|z|) Est. SE Z value Pr(>|z|) A. ligamentina -0.612 0.116 -5.281 0.000 0.079 0.025 3.146 0.002 E. dilatata -0.675 0.118 -5.723 0.000 0.096 0.028 3.457 0.001 F. flava -0.061 0.082 -0.745 0.456 -0.058 0.021 -2.745 0.006 L. cardium -0.221 0.097 -2.276 0.023 0.031 0.015 2.061 0.039 L. ovata -0.362 0.175 -2.061 0.039 0.101 0.025 3.993 0.000 L. siliquoidea -0.211 0.051 -4.125 0.000 0.021 0.011 1.940 0.052 L. costata -0.215 0.059 -3.640 0.000 0.026 0.014 1.923 0.054 L. fragilis -0.166 0.064 -2.615 0.009 0.055 0.015 3.665 0.000 L. recta -0.262 0.080 -3.260 0.001 0.039 0.013 3.022 0.003 P. alatus -0.227 0.065 -3.474 0.001 -0.059 0.013 -4.447 0.000 2010 T.A. Smith and E.S. Meyer 555 Discussion The Pittsburgh District of the US Army Corps of Engineers operates eight locks and dams on the Allegheny River for navigation of commercial vessels. Dams fragment mussel habitat by inhibiting longitudinal movement of host fishes and glochidia (Watters 1996). Furthermore, the impoundments provide habitat for invasive species such as Zebra Mussels, which are a documented threat to freshwater mussels (Biggins et al. 1995, Ricciardi et al. 1998, Strayer and Malcom 2007) and were present in this study. The detrimental effects of dams and impoundments on freshwater ecosystems has been widely documented (e.g., Bates 1962; Baxter 1977; Blalock and Sickel 1996; Bogan 1993; Chessman et al. 1987; Kondolf 1997; Parmalee and Hughes 1993; Porto et al 1999; Richter et al. 1997; Sickel et al. 2007; Vaughn and Taylor 1999; Watters 1996; 2000; Williams and Fuller 1992). The lock-and-dam structures on the Allegheny River have altered the river from free-flowing, well-oxygenated riffles and runs into a series of deep, slower-flowing pools or lakes (Ortmann 1909). Of the entire 22.6 million-m2 mapped area in pools 4 through 8, 37.1% is deeper than 6 m and 19.5% is deeper than 9 m (E. Long, Western Pennsylvania Conservancy, Blairsville, PA., pers. comm.; Long and Chapman 2008). Of the surveyed segments, 10.2% were in areas deeper than 6 m and 2.1% were in areas deeper than 9 m. The results of our study show that 95% of the live individuals were found in the relatively shallow areas (less than 6 m) of these pools, and zero live individuals were found in the fourteen segments surveyed in areas deeper than 9 m. Generally, shallow depths are productive biologically due to solar penetration, dynamic currents, and higher dissolved-oxygen levels (Allan 1995, Vannote et al. 1980). However, there have been mussels documented from other systems in water deeper than 9 m (e.g., James 1985, Reigle 1967) where bottom water flow was adequate and the substrate composition was favorable. Therefore, although depth may provide us with some indication where mussels persist in the Allegheny River, depth should not be used as the only indicator of mussel presence in a system. Habitat parameters combine to create suitable habitat for mussel populations (Parmalee and Bogan 1998, Strayer 2008). Longitudinal shifts in community composition and abundance occur naturally in river systems (Vannote et al. 1980), and we observed a change from an abundance of facultative species more typical of slow-moving rivers in the lower pools to a dominance of riverine species in the upper pools. Pink Heelsplitter, for example, the dominant species in pool 4, gradually decreased in relative abundance going upstream and was not present in our surveys in pool 8. In contrast, the upper pools were dominated by Mucket and Spike and contained species more typical to smaller waterways, such as Elktoe and Wavy-rayed Lampmussel. In addition to the shift from riverine to facultative species throughout the navigational pools of the Allegheny River as it flows towards its confluence, we documented similar changes 556 Northeastern Naturalist Vol. 17, No. 4 within each pool. Riverine species richness and counts were higher in the upper portions of the pools, indicating that areas with consistent flows and suitable substrate just downstream of the dams may provide refugia for riverine freshwater mussel species. In the Allegheny River navigational pools, impoundments are confounded with river bottom disturbance. Sand and gravel extraction can significantly alter the chemical, physical, and biological components of mined streams and rivers (e.g., Brown et al. 1998, Meador and Layher 1998, Nelson 1993). Altered substrate and flow resulting from gravel extraction can reduce or eliminate mussel populations (Hubbs et al. 2003). Dredging removes sand and gravel substrate, and the deep depressions that remain often fill with silt and debris (Brown et al. 1998) unsuitable for colonization by riverine mussels. These deep portions of the river may not be subjected to any water currents and therefore have depleted dissolved-oxygen levels. Recent studies show significantly higher levels of total dissolved solids, turbidity, arsenic, selenium, and zinc in river water following mining operations in the Allegheny River (Murray et al. 2008). Additionally, river islands and shoals, which provide important habitat for mussels, are also affected by dredging due to increased erosion from altered flow regimes in close proximity to islands (Kondolf 1997). In addition to the direct effects to mussels, suspended sediments from excavation activities have led to the loss or reduction of fish and macroinvertebrate spawning, rearing, and foraging habitat (Brown et al. 1998), which means the potential loss of the fish hosts needed to complete freshwater mussel lifecycles. It can take decades for a river to recover from sand and gravel mining without remediation (e.g., Kanehl and Lyons 1992). It is not known how much time is needed to replenish substrate in a system where the natural migration of gravel and sand from upstream sources is impeded by impoundments, such as in the lower Allegheny River. This study provides some evidence that the alteration and loss of habitat due to sand and gravel dredging activities has had an adverse effect on the freshwater mussel fauna of the Allegheny River navigational pools. The upper portion of pool 8, which has not been mined, had consistently high counts and species richness, evidence of recruitment, and an undisturbed habitat comprised of a mix of sand, gravel, cobble, and boulders. Conversely, total counts and richness in pool 8 were low in areas with past or current commercial mining permits (in the lower half of the pool). In these areas, maximum depths were generally deeper than in non-dredged areas, and silt dominated the substrate composition. Similarly, only in the protected area of pool 7 did we find an abundant and diverse mussel community. Much of pool 7 is highly impacted by commercial sand and gravel dredging, resulting in relatively deep water up to 14 m (Long and Chapman 2008). Divers observed a high percentage of silt in these areas; thick silt substrate is unsuitable for most riverine mussels. Relatively shallow depths indicate undisturbed substrate in this pool; for example, since no dredging is allowed within 500 feet of any 2010 T.A. Smith and E.S. Meyer 557 bridge, pier, or abutment (US Army Corps of Engineers 2007), the substrate was intact and unaltered under the railroad bridge and, subsequently, the species richness and counts were higher there than those observed in the deeper, dredged areas. Pool 6 has a moratorium on dredging operations that has been in place since December of 1985, which may account for the relatively high abundances and species diversity recorded there. However, this type of restriction may not be feasible for resource managers who are faced with the challenge of protecting imperiled species while not impeding commercial operations. Depth-limited dredging restrictions or limitations to areas that have been previously altered may be one option to allow commercial mining in areas where it has already occurred while still protecting remaining mussel habitat. The protection of any relatively shallow areas with intact substrate will be important to sustain any remaining freshwater mussel populations in these pools. Of the 65 species of freshwater mussels that have been reported from Pennsylvania, at least 11 (17%) have not be collected in the past 25 years and are considered historic or possibly extirpated (PNHP 2009). The range of all 11 of those species included the currently pooled sections of both the Allegheny and Ohio River mainstems. Nationally, over the past 100 years, over 30 species are considered to have gone extinct, with at least 70 others in danger of extinction (Lydeard et al. 2004, Ricciardi and Rasmussen 1999, Williams et al. 1993). Species such as Clubshell and Northern Riffleshell are critically imperiled throughout most of their range (NatureServe 2009), but exhibit the highest densities in their range in the middle portion of the Allegheny River (Villella and Nelson 2006). Clubshell is a species that formerly occupied the Monongahela River tributaries, Ohio River tributaries, the Allegheny River, and several tributaries of the Allegheny River in Pennsylvania (Ortmann 1919, USFWS 1994). Clubshell are currently only known from the Shenango River (Bursey 1987), the French Creek Watershed (Smith and Crabtree, in press), the middle Allegheny River upstream of our study sites (Villella and Nelson 2006), and from one live individual found in the upper portion of pool 8 (USFWS 2004). In Pennsylvania, Northern Riffleshell was historically known from the Allegheny River, Leboeuf Creek, Conewango Creek, French Creek, and the Shenango River, (Ortmann 1919, USFWS 1994); it is currently only known from French Creek (Crabtree and Smith 2009; Mohler et al. 2006; Smith and Crabtree, in press), Conewango Creek (Evans and Smith 2005), and the Allegheny River (Villella and Nelson 2006). In this study, these two species were found only in areas with intact substrate and relatively high species richness and counts. This result indicates that there are positive species interactions occurring in those areas. Rare species in particular have been documented to profit energetically from living in species-rich communities (Vaughn et al. 2008). Although the relatively large number of transects surveyed within pool 6 increased our probability 558 Northeastern Naturalist Vol. 17, No. 4 of detecting rare species there, Northern Riffleshell was not found in pool 6 during our surveys. There is some evidence that Northern Riffleshell may exist in pool 6; it was recently found as a dead specimen along the shore below lock 7 (PNHP files 1985, 1995a, 1995b, 2002). Much of the habitat we observed in pool 6 seemed suitable for this species, and in combination with habitat protection and augmentation, it may be possible for this species to successfully re-establish in this pool. The documentation of live Salamander Mussel is particularly important in Pennsylvania, as it was historically known from the Middle Allegheny- Redbank drainage (Clarke 1985), but may be extant in the French Creek and Lower Monongahela River drainages (PNHP 1995b; Smith and Crabtree, in press). This species is easy to overlook because it is a small-sized mussel and typically found under large flat rocks, presumably where it was deposited by its primary host, Necturus maculosus Rafinesque (Mudpuppy) (Parmalee and Bogan 1998). Early surveys in streams and rivers in western Pennsylvania were devoid of this species (Ortmann 1909), and the first documented occurrence in the Allegheny River was in 1969 (Bogan and Locy 2009). It has since been collected in just a few locations in the Allegheny River navigational pools (Bogan and Locy 2009). Several species we expected to see in our surveys were absent, although they have been recently found both upstream and downstream of the pools surveyed in this study. Although Sheepnose, a candidate for federal listing, was absent from our surveys, the middle Allegheny River is considered a stronghold for this species (Villella et al. 2008) and it is present in the Ohio River (Zeto et al. 1987). Other species absent from our study that have been documented in the Allegheny River upstream of our study are Utterbackia imbecillis Say (Paper Pondshell) and Quadrula cylindrica Say (Rabbitsfoot) (Villella and Nelson 2006). The low densities typical of these species in nearby streams (i.e., Mohler et al. 2006; Smith and Crabtree, in press) indicate that more effort may have been needed to detect those species to determine their presence in these pools. Three-ridge and Kidneyshell, which we found but in low numbers, are present in great numbers in upstream tributaries (i.e., Chapman and Smith 2008; Smith and Crabtree, in press; Smith and Horn 2006). These two species are known from small streams to big rivers in a variety of habitats (Parmalee and Bogan 1998), so it is unclear why they are were not more common in our study. Until this study, Pink Papershell was previously documented only once in the Allegheny River, and to our knowledge this is the second time it has been recorded in Pennsylvania (PNHP 2002). Pink Papershell has not been documented upstream (Villella and Nelson 2006), but is present in the Ohio River drainage in Ohio (Watters 1995) and West Virginia (Zeto et al. 1987) and is typically found in slow-moving water (Parmalee and Bogan 1998). Although the ORVET protocol we used is a useful way to get an initial look at species presence and rough estimates of relative abundances, it is not 2010 T.A. Smith and E.S. Meyer 559 adequate for finding all rare species that may be present (ORVET 2004). Our mean search rate of only 1.25 min/m2 was less than the surface search rate that Smith et al. (2001) determined was necessary to detect Clubshell 31% of the time and Mucket 70% of the time in surveys upstream of the navigational pools. Increasing search time and search area would have increased detectability of species in this study (Smith 2006). We recommend that any further studies, particularly in areas of proposed projects, use enough effort to detect rare species (e.g., Smith 2006) and incorporate quadrat-based studies to get unbiased estimates of species densities, sex ratios, and the number of juveniles present (Strayer and Smith 2003). More precise and quantitative measurements of environmental variables could be made at the quadrat level; however, there has not been much evidence that fine-scale habitat preferences exist. Other environmental variables, such as water flow and oxygen levels, may be more important to measure and monitor, for example, in areas with known reproducing populations. Understanding spatial and temporal dynamics of mussel populations will help us understand the benefits of any conservation efforts, comprehend the consequences of disturbance, and determine if existing regulations are sufficient to protect the mussel communities. We hope the data presented in this study will provide baselines for future monitoring and conservation efforts. Relatively intact areas with reproducing populations, for example, could be monitored and compared to other areas of the river that are being considered for relocation, introduction, or habitat-restoration efforts. Acknowledgments This study was funded by a US Fish and Wildlife Service State Wildlife Grants Program Grant T-2 administered through the Pennsylvania Fish and Boat Commission (PAFBC). Supplemental funding from the United States Fish and Wildlife Service (USFWS) Pennsylvania Field office was used to conduct field work in 2005. Thanks to Patricia Morrison (USFWS), Janet Butler (USFWS), and Robert Anderson (USFWS) for oversight and dive training. We thank the PA Fish and Boat Commission (PAFBC) for use of their dive boat and Robert Morgan (PAFBC) and Doug Fischer (PAFBC) for help with boat operations in 2005. A portion of the 2007 surveys were subcontracted to Environmental Science, Inc. Thanks to the Colcom Foundation for funding the purchase of a research vessel and thanks to Eric Chapman (Western Pennsylvania Conservancy [WPC]) for maintaining the boat. We thank the US Army Corps of Engineers for allowing us lockage and the Rosston Eddy Marina and Nautical Mile Marina for accommodations and boat repair. Thanks to Scott’s SCUBA in Freeport, PA and to Divers World in Erie, PA for being accommodating with special equipment rental and maintenance needs. Thanks to Darran Crabtree (The Nature Conservancy), Mary Walsh (Pennsylvania Natural Heritage Program [PNHP]), Jeremy Deeds (PNHP) and Nevin Welte (PAFBC) for reviewing drafts of this manuscript and to Charles Bier and Erin Stacy (PNHP) for additional help. This manuscript was substantially improved thanks to the comments of David Smith (USGS), Beth Gardner (USGS) and two anonymous reviewers. Special thanks to additional WPC/PNHP SCUBA dive crew members Ryan Evans, Zachary Horn, Nicole Rhodes, Mary Walsh, Erik Weber, and Jacob Winkler. 560 Northeastern Naturalist Vol. 17, No. 4 Literature Cited Allan, D.J. 1995. Stream Ecology: Structure and Function of Running Waters. Kleuwer Academic Publishers, Springer, Norwell, MA. 388 pp. Bates, J.M. 1962. The impact of impoundment on the mussel fauna of Kentucky Reservoir, Tennessee River. American Midland Naturalist 68:232–236. Baxter, R.M. 1977. Environmental effects of dams and impoundments. Annual Review in Ecological Systems 8:255–283. Biggins, R.G., R.J. Neves, and C.K. Dohner. 1995. Draft national strategy for the conservation of native freshwater mussels. US Fish and Wildlife Service, Washington, DC. 26 pp. Blalock, H.N., and J.B. Sickel. 1996. Changes in mussel (Bivalvia: Unionidae) fauna within the Kentucky portion of Lake Barkley since impoundment of the lower Cumberland River. American Malacological Bulletin 13:11–116. Bogan, A.E. 1993. Freshwater bivalve extinctions (Mollusca: Unionoida): A search for causes. American Zoology 33:599–609. Bogan, A.E., and D.D. Locy. 2009. Current distribution of the Salamander Mussel, Simpsonaias ambigua (Say, 1825), in Pennsylvania. Ellipsaira 11:11–12. Brown, A.V., M.M. Lyttle, and K.B. Brown. 1998. Impacts of gravel mining on gravel bed streams. Transactions of the American Fisheries Society 127:979–94. Bursey, C.R. 1987. The unionid (Mollusca: Bivalvia) fauna of the Shenango River in Mercer County, Pennsylvania. Proceedings of the Pennsylvania Academy of Science 61:41–43. Chapman, E.J., and T.A. Smith. 2008. Structural community changes in freshwater mussel populations of Little Mahoning Creek, Pennsylvania. American Malacological Bulletin 26:161–169. Chessman, B.C., D.P. Robinson, and K.G. Hortle. 1987. Changes in the riffle macroinvertebrate fauna of the Tanjil River, southeastern Australia, during construction of Blue Rock Dam. Regulated Rivers Resource Management 1:317–329. Clarke, A.H. 1985. The tribe Alasmidontini (Unionidae: Anodontinae). Part II: Lasmigona and Simpsonaias. Smithsonian Contributions to Zoology 399:1–75. Crabtree, D.L., and T.A. Smith. 2009. Population attributes of an endangered Mussel, Epioblasma torulosa rangiana (Northern Riffleshell), in French Creek and implications for its recovery. Northeastern Naturalist 3:339–354. Cummings, K.S., and C.A. Mayer. 1992. Field Guide to Freshwater Mussels of the Midwest. Illinois Natural History Survey Manual 5, Champaign, IL. 194 pp. Evans, R., and T.A. Smith. 2005. Northern Riffleshells located in Conewango Creek, Allegheny River Basin, Pennsylvania, with a note on new records for the Zebra Mussel. Ellipsaria 7: 6–7. Haag, W.R., and A.M. Commens-Carson. 2008. Testing the assumption of annual shell ring deposition in freshwater mussels. Canadian Journal of Fisheries and Aquatic Sciences 65:493–508. Hubbs, D., S. Lanier, D. McKinney, D. Sims, and P. Black. 2003. Evaluation of abandoned commercial sand and gravel dredge sites as freshwater mussel habitat on the lower Tennessee River. Tennessee Wildlife Resources Agency, Aquatic Habitat Protection Program, Nashville, TN. ii + 30 pp., + 2 appendices. 2010 T.A. Smith and E.S. Meyer 561 James, H.R. 1985. Distribution, biomass, and production of the freshwater mussel Hyridella menzeisi (Gray), in Lake Tapua, New Zealand. Freshwater Biology 15:307–314. Kanehl, P., and J. Lyons. 1992. Impacts of in-stream sand and gravel mining on stream habitat and fish communities, including a survey on the Big Rib River, Marathon County, Wisconsin. Wisconsin Department of Natural Resources Research Report 155, Madison, WI. Kondolf, G.M. 1997. Hungry water: Effects of dams and gravel mining on river channels. Environmental Management 21:533–551. Long, E., and E.J. Chapman. 2008. A unique approach to bathymetry mapping in a large river system using GIS tools to evaluate hidden habitat. ArcUser. Available online at http://www.esri.com/news/arcuser/0708/eli-river.html. Accessed Spring 2008. Lydeard, C., R.H. Cowie, W.F. Ponder, A.E. Bogan, P. Bouchet, S.A. Clark, K.S. Cummings, T.J. Frest, O. Gargominy, D.G. Herbert, R. Hershler, K.E. Perez, B. Roth, M. Seddon, E.E. Strong, and F.G. Thompson. 2004. The global decline of nonmarine mollusks. BioScience 54:321–330. McCullagh, P., and J.A. Nelder. 1989. Generalized Linear Models. Chapman and Hall, CRC, London, UK. Meador, M.R., and A.O. Layher. 1998. Instream sand and gravel mining: Environmental issues and regulatory process in the United States. Fisheries 32:6–13. Mohler, J.W., P. Morrison, and J. Haas. 2006. The mussels of Muddy Creek on Erie National Wildlife Refuge. Northeastern Naturalist 13:569–582. Murray, M., A. Michanowicz, K. Ferrer, S. Good, R. Suphagaphan, C. Christen, E. Talbot, and C. Volz. 2008. A comparison of water quality variables and concentrations of selected toxic and nuisance elements at various distances and depths downstream from Allegheny River gravel mining operations during periods of activity and inactivity. Poster presentation: Ohio River Basin Consortium for Research and Education Conference, ORBCRE Scientific Symposium 2008 October 29–31, 2008, Pittsburgh, PA. NatureServe 2009. NatureServe Explorer: An online encyclopedia of life [web application]. Version 6.1. NatureServe, Arlington, Virginia. Available online at http:// www.natureserve.org/explorer. Accessed 30 December 2009. Nelson, K.L. 1993. Instream sand and gravel mining. Pp 189–196, In C.F. Bryan and D.A. Rutherford (Eds.). Impacts on Warmwater Streams: Guidelines for Evaluation. Southern Division, American Fisheries Society, Little Rock AR. 285 pp. Obermeyer, B.K. 1998. A Comparison of Quadrats Versus Timed Snorkel Searches for Assessing Freshwater Mussels. American Midland Naturalist 139:331–339. Ohio River Valley Ecosystem Team Mollusk Subgroup (ORVET). 2004. Draft Protocol for Mussel Surveys in the Ohio River Where Dredging/Disposal/Development Activity Is Proposed. Frankfort, KY. 4 pp. Ortmann, A.E. 1909. The destruction of freshwater fauna in western Pennsylvania. Proceedings of the American Philosophical Society 48:90–110. Ortmann, A.E. 1919. A monograph of the naides of Pennsylvania. Memoirs of the Carnegie Museum, Vol. VIII, Pittsburgh, PA. Parmalee, P.W., and A.E. Bogan. 1998. The Freshwater Mussels of Tennessee. University of Tennessee Press, Knoxville, TN. 328 pp. 562 Northeastern Naturalist Vol. 17, No. 4 Parmalee, P.W., and M.H. Hughes. 1993. Freshwater mussels (Mollusca: Pelecypoda: Unionoida) of Tellico Lake: Twelve years after impoundment of the Little Tennessee River. Annals of Carnegie Museum 62:81–93. Pennsylvania Natural Heritage Program (PNHP). 1985. ACC files. Presence record of Simpsonaias ambigua in French Creek. Crawford Co., PA. Documented by Charles Bier (WPC) and T.L. Schumann. September 1, 1985. vailable online at http://www.naturalheritage.state.pa.us/invertebrates.aspx. Accessed 8 August 2008. PNHP. 1995a. ACC files. Presence record of Epioblasma torulosa rangiana. Armstrong Co., PA. Documented by Charles Bier (PNHP) and G.T. Bier. 26 Aug 1995. vailable online at http://www.naturalheritage.state.pa.us/invertebrates. aspx. Accessed 8 August 2008 PNHP. 1995b. ACC files. Presence record of Simpsonaias ambigua in French Creek. Crawford Co., PA. Documented by Charles Bier (WPC). September 11, 1995. vailable online at http://www.naturalheritage.state.pa.us/invertebrates.aspx. Accessed August 8, 2008. PNHP. 2002. ACC files. Presence record of Potamilus ohiensis in the Allegheny River pool 5 from Civil and Environmental Consultants, Inc. for Hanson Aggregates PMA, Inc. July 15, 2002. vailable online at http://www.naturalheritage. state.pa.us/invertebrates.aspx. Accessed 26 March 2008. PNHP. 2009. PNHP website. Available online at http://www.naturalheritage.state. pa.us/invertebrates.aspx. Accessed December 30, 2009. Porto, L.M., R.L. McLaughlin, and D.L.G. Noakes. 1999. Low-head barrier dams restrict the movements of fishes in two Lake Ontario streams. North American Journal of Fisheries Management 19:1028–1036. R Development Core Team. 2009. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. Available online at http://www.R-project.org. Accessed 26 October 2009. Reigle, N. 1967. An occurrence of Anodonta (Mollusca, Pelecypoda) in deep water. American Midland Naturalist 87:530–531. Ricciardi A., R.J Neves, and J.B Rasmussen. 1998. Impending extinctions of North American freshwater mussels (Unionoida) following the Zebra Mussel (Dreissena polymorpha) invasion. Journal of Animal Ecology 67:613–619. Ricciardi, A., and J.B. Rasmussen. 1999. Extinction rates of North American freshwater fauna. Conservation Biology 13:1220–1222. Richter, B.D., D.P. Braun, M.A. Mendelson, and L.L. Master 1997. Threats to imperiled freshwater fauna. Conservation Biology 11:1081–1093. Sickel, J.B., M.D. Burnett, C.C. Chandler, C.E. Lewis, H.N. Blalock-Herod, and J.J.Herod. 2007. Changes in the freshwater mussel community in the Kentucky portion of Kentucky Lake, Tennessee River, since impoundment by Kentucky Dam. Journal of the Kentucky Academy of Science 68:68–80. Smith, D.R., R.F. Villella, and D.P. Lemarie. 2001. Survey protocol for assessment of endangered mussels in the Allegheny River, Pennsylvania. Journal of the North American Benthological Society 20:118–132. Smith, D.R. 2006. Survey design for detecting rare freshwater mussels. Journal of the North American Benthological Society 25:701–711. Smith, T.A., and D. Crabtree. In press. Freshwater mussel (Unionidae: Bivalvia) distributions and densities in French Creek, Pennsylvania. Northeastern Naturalist. 2010 T.A. Smith and E.S. Meyer 563 Smith, T.A., and Z.J. Horn. 2006. Freshwater mussel (Unionidae) spatial distributions in the New York portion of the Allegheny River Watershed. February 16. Submitted to The Nature Conservancy, Union City, PA. 67 pp. Strayer, D.L. 2008. Freshwater Mussel Ecology: A Multifactor Approach to Distribution and Abundance. University of California Press, Berkeley, CA. Strayer, D.L., and H.M. Malcom. 2007. Effects of Zebra Mussels (Dreissena polymorpha) on native bivalves: The beginning of the end or the end of the beginning? Journal of the North American Benthological Society 26:111–122. Strayer, D.L., and D.R. Smith. 2003. A guide to sampling freshwater mussel populations. American Fisheries Society Monograph 8, Bethesda, MD. Turgeon, D.D., J.F. Quinn, Jr., A.E. Bogan, E.V. Coan, F.G. Hochberg, W.G. Lyons, P.M. Mikkelsen, R.J. Neves, C.F.E. Roper, G. Rosenberg, B. Roth, A. Scheltema, F.G. Thompson, M. Vecchione, and J.D. Williams. 1998. Common and Scientific Names of Aquatic Invertebrates from the United States and Canada: Mollusks, 2nd Edition. American Fisheries Society, Special Publication 26, Bethesda, MD. US Army Corps of Engineers (USACE). 2007. Record of decision. Final environmental impact statement commercial sand and gravel dredging operations Allegheny River (mile 0.0–69.5) and Ohio River (mile 0.0–40.0) Pennsylvania. Pittsburgh District, Pittsburgh, PA. 25 pp. USACE. 2010a. Kinzua Dam and Allegheny Reservoir. Available online at http:// www.lrp.usace.army.mil/rec/lakes/kinzuala.htm. Accessed January 25, 2010. USACE. 2010b. Navigation. Available online at http://www.lrp.usace.army.mil/nav/ nav.htm#notice. Accessed 25 January 2010. US Fish and Wildlife Service (USFWS). 1994. Clubshell (Pleurobema clava) and Northern Riffleshell (Epioblasma torulosa rangiana) recovery plan. Hadley, MA. 68 pp. USFWS. 2004. Allegheny River Mussel Survey Pool 8—River miles 58.8 to 58.2. US Fish and Wildlife Service Pennsylvania Field Office, State College, PA. August 16. 25 pp + appendicies. US Geological Survey. Web application. Available at http://www.waterdata.usgs. gov. Accessed 15 October 2007. Vannote, R.L., G.W. Minshall, K.W. Cummins, J.R. Sedell, and C.E. Cushing. 1980. The river continuum concept. Canadian Journal of Fisheries and Aquatic Sciences. 37:130–137. Vaughn, C.C., and C.M. Taylor. 1999. Impoundments and the decline of freshwater mussels: A case study of an extinction gradient. Conservation Biology 13:912–920. Vaughn, C.C., S.J. Nichols, and D.E. Spooner. 2008. Community and foodweb ecology of freshwater mussels. Journal of the North American Benthological Society 27:409–423. Villella, R.F., and R. G. Nelson. 2006. Mussel survey of the Upper Allegheny River: Phase one and phase two sampling. US Geological Survey, Leetown Science Center, WV. 47 pp. Villella, R.F., E. Pendelton, and W. Lellis. 2008. Bridging the information gap for mussels. Endangered Species Bulletin 33:40–42. Watters, G.T. 1995. A Field Guide to the Freshwater Mussels of Ohio. Revised 3rd Edition. Ohio Department of Natural Resources, Division of Wildlife, Columbus, OH. 122 pp. 564 Northeastern Naturalist Vol. 17, No. 4 Watters, G.T. 1996. Small dams as barriers to freshwater mussels (Bivalvia, Unionoida) and their hosts. Biological Conservation 75:9–85. Watters, G.T. 2000. Freshwater mussels and water quality: A review of the effects of hydrologic and instream habitat alterations. Pp. 1–14, In R.A. Tankersley, D.A.Warmolts, G.T. Watters, B.J. Armitage, P.D Johson, and R.S. Butler (Eds.). Freshwater Mollusk Symposia Proceedings. Ohio Biological Survey, Columbus, OH. Wentworth, C.K. 1922. A scale of grade and class terms for clastic sediments. Journal of Geology 30:377–392. Williams, J.D., and S.L. Fuller. 1992. Effects of impoundments on freshwater mussels (Mollusca: Pelecypoda: Unionidae) in the main channel of the Black Warrior and Tombigbee rivers in western Alabama. Alabama Museum of Natural History Bulletin 13:1–10. Williams, J.D., M.L. Warren, K.S. Cummings, J.L. Harris, and R.J. Neves. 1993. Conservation status of freshwater mussels of the United States and Canada. Fisheries 18:6–22. Zeto, M.A., W.A. Tolin, and J.E. Schmidt. 1987. The freshwater mussels (Unionidae) of the upper Ohio River, Greenup and Belleville Pools, West Virginia. Nautilus 101:182–185.