Factors influencing detection of the federally endangered diamond darter Crystallaria cincotta: implications for long-term monitoring strategies.
Population monitoring is an essential component of endangered species recovery programs. One species for which monitoring is needed is the federally endangered Diamond Darter Crystallaria cincotta. It has been argued there are two types of rare species: truly functional rare species and operationally rare species (McDonald, 2004). Truly functional rare species are those species that have a very small range and population size. Operationally rare species generally appear to be rare because they have sparse and/or clumped populations, are small, cryptic, or elusive, or current survey procedures are insufficient for detection (McDonald, 2004). The Diamond Darter is considered to be a rare species (USFWS, 2013). There is ample evidence the Diamond Darter's geographic distribution is in fact small, indicating it is a truly functional rare species. However, because of the Diamond Darter's small size, cryptic coloration, nocturnal behavior, and seemingly patchy distribution, along with limitations associated with current sampling methods, the species may appear to be more rare than it actually is. This may have also contributed to its late discovery as a species in 2008. Museum specimens indicate the Diamond Darter was once distributed throughout the Ohio River Basin, but it is now believed to be extirpated from the Muskingum River in Ohio; the Ohio River in Ohio, Kentucky, and Indiana; the Green River in Kentucky; and the Cumberland River drainage in Kentucky and Tennessee. Currently, it is known to exist only within the lower 50 km of the Elk River in West Virginia (Welsh and Wood, 2008; Welsh et al., 2009; Welsh et at., 2014).
Studies have found seasonal and diel plasticity in behavior and habitat use patterns of freshwater fish is common (Ali, 1992) and often strongly associated with environmental changes (Reeves et al, 2009). Seasonal plasticity in habitat use and behavior may help to increase survival and fecundity of fishes in dynamic seasonally-changing systems, and has been observed in numerous species, including darters (Jones et al, 1984; Hlohowskyj and Wissing, 1985; Harding et al., 1998). Diel plasticity has also been observed in numerous species and is usually attributed to individuals seeking thermal, velocity, or predator refuge (Mundahl, 1990; Smith and Fausch, 1997; Schaefer et al, 2003). Diel plasticity in behavior and habitat use has also been observed in fish species whose prey items become active during nocturnal periods (Reebs, 2002).
Failure to account for detection probability can be a major source of bias in presence-absence and count survey data, potentially resulting in false determination of species absence and underestimation of abundance, respectively (MacKenzie et al, 2002; MacKenzie et al., 2006; Kery and Royle, 2016). Ultimately, not accounting for detection probability may lead to misinformed management decisions, which can prove costly for an endangered species. Most fish monitoring programs either use counts of organisms as proxies for abundance or employ capture-recapture or removal sampling techniques. These techniques can often be costly in terms of effort and resources and may not be possible depending on the study organism. In particular, use of these survey techniques (which involve capturing and handling individuals) may not be a viable option for threatened and endangered species.
N-mixture models provide an alternative to traditional abundance estimation techniques by allowing abundance to be estimated without marking or removing individuals (Royle, 2004; Kery and Royle, 2016). Additionally, these models estimate detection probability independently from abundance and therefore allow for the inclusion of explanatory covariates for both parameters. This approach is appropriate when you have repeated count data or repeated detection/nondetection data from multiple sites and can assume sites are closed to changes in mortality, recruitment, and migration over the sampling interval (Royle, 2004; Dail and Madsen, 2011). A-mixture models have been used to estimate abundance of many wildlife species (e.g., Joseph el al., 2009; Couturier et al., 2013), as well as several fish species (e.g., Wenger and Freeman, 2008; Kanno et al,, 2014).
Our current knowledge regarding the historic and current distribution and abundance of the Diamond Darter is based on museum specimens and field surveys conducted since 2011 using a recently developed search method (Welsh et al., 2013). These surveys have confirmed that Diamond Darters use glide habitat (i.e., those areas of the river immediately upstream of riffles). However, qualitative observations indicate detection probabilities at known occupied sites show diel variation. In addition availability of Diamond Darters in glide habitat may vary seasonally.
The primary objective of this study was to determine if there is seasonal and diel variation in Diamond Darter detectability. Along with considering temporal factors, we also assessed five habitat variables that could influence detection of individuals. Our final objective was to estimate abundance at our study sites, resulting in the first site abundance estimates for the Diamond Darter.
The Diamond Darter is known to exist only within the lower 50 km of the Elk River in West Virginia (Welsh and Wood, 2008; Welsh et al, 2009; Welsh et al, 2014). Our study focused on three glide areas (sites; Walgrove, 38.466494-81.4384; Reamer, 38.474105-81.3769; and Clendenin Compressor Station, 38.487817-81.3273; West Virginia) within their known geographical range (Fig. 1). These three sites were chosen because occupancy at these sites had previously been established, allowing us to focus on abundance and detection given known occupancy.
Sites were surveyed using a search method that employs the use of spotlighting at night with flashlights within wadeable sections of the river. This method has proven to be effective in glide habitats. These areas are shallow enough for a person to wade transects and have a smooth water surface, which allows the spotlighter to see through the water column to the substrate (Welsh et al, 2013). As a result of water level fluctuations in the river, the number of transects required to search a site varied by sampling night, with the range in the number of transects searched in a given night being 20-24 at Walgrove, 24-32 at Reamer, and 24-32 at Clendenin. Consequently, there was no standard number of transects sampled per sampling occasion. Counts of Diamond Darters were summed across all transects at a site to get a total count during a sampling occasion.
Potential diel variation in detectability was assessed by breaking up the time between dusk and dawn into three distinct time blocks based on time after sunset and time before sunrise. Sampling occurred 1 h after sunset, 2 h before sunrise, and in the middle of the night. The actual time of sampling varied throughout the summer as a result of day length changes. Because glide size varied among sites and depended on water levels, search times varied in order to thoroughly sample each glide area. Sampling took place in 3 d blocks. All three sites were sampled each night and each site was sampled at a different time block each night.
Time of night that sampling occurred was modeled using hours after sunset (h). To assess seasonal variation in detectability, we sampled in May, Aug., Sep., and Oct. (high river discharges prevented sampling during Jun. and Jul.) and modeled this covariate using day of year. The following covariates were recorded during each sampling event: water temperature, water gauge height at upstream gaging station, turbidity, and sky brightness. Temperature directly affects the metabolic rate of ectothermic organisms, and as a result, behavior and activity of these organisms often is contingent on the temperature of the surrounding environment (Ingersoll and Claussen, 1984). Water temperature was included because of its potential to influence detection of the darter.
Water gage height was included to capture the effects of water levels on detection. Water level can affect detection by: (1) decreasing the ability of the observer to see through deeper water (influencing detection when available) and (2) creating suboptimal habitat in glide areas as a result of increased velocity and other factors associated with higher water levels (influencing availability). Turbidity was included as a detection covariate because of its potential to affect both the behavior of the darter and the field of vision for the searchers. Most fish rely on their vision as their main source of sensory information, aiding in their ability to detect both predator and prey (Pitcher et al., 1993). Turbidity has been found to affect the visual abilities of fish; however, whether the effects are positive or negative largely depends on the species, the size class, and the interactive effects with available light (Utne-Palm, 2002; Stoner, 2004). Consequently, brightness of the sky (measured by a sky quality meter) was included as a covariate potentially influencing the behavior of the Diamond Darter, both on its own and as an interactive effect with turbidity (Utne-Palm, 2002). This interactive effect with turbidity may also influence observer detection of the darter and was included as a covariate for both reasons.
N-mixture models were used to estimate abundance and identify- important variables influencing individual detection. A pairwise correlation analysis of all predictor variables was conducted to identify any multicollinearity prior to Admixture modeling. Water temperature was included in candidate models as a quadratic term because preliminary analyses indicated the quadratic relationship with relative abundance was a better predictor than a simple linear relationship. Day of year was included in candidate models as a linear term and a quadratic term because of the potential for different seasonal effects influencing detection.
To estimate abundance at each site, we used a single-season Admixture model (Kery and Royle, 2016), with a binomial distribution for the observation process. A single-season model was used to analyze surveys conducted from May through Oct. under the assumption the population was closed during that period. Demographic closure was assumed based on analysis of population size structure throughout the season (Rizzo et al, 2017). Additionally, adjusted population estimates were similar throughout the sampling season, lending support to our assumption our populations were closed. Research by Kendall (1999) suggests if individuals occupy and vacate a site in a random manner, estimators used to explain detection and abundance/occupancy should remain unbiased. However, if there is random movement, resulting in the violation of geographic closure, then parameter values for detection and abundance/occupancy may be inaccurate. Inaccurate mean estimated probabilities would result in an increase in the variance of our estimates (MacKenzie et al, 2004).
A Poisson distribution was chosen for the state process after comparing Akaike Information Criterion ([AIC.sub.c]) values and residual diagnostic plots for three distributions (i.e., Poisson, zero-inflated Poisson, and negative binomial; Kery and Royle, 2016). Model goodness-of-fit was evaluated using a parametric bootstrap of the Pearson chi-square statistic (Mazerolle, 2016). The goodness-of-fit test indicated the data were overdispersed, and we accounted for this by inflating the estimated standard errors and 95% confidence intervals (CI) based on the [??] value (i.e., 4.7; Kery and Royle, 2016).
A total of 10 a priori models were constructed based on our knowledge of darter biology. The number of models included was partially restricted by the small sample size. Model complexity was minimized in an effort to make interpretation simple, allowing researchers to identify those variables most responsible for variation in individual Diamond Darter detectability. Because we were interested in estimating abundance by site, site was included as a covariate of abundance in all 10 models. All models, except one, included only one detection variable. The exception was a model that included the interacdve effect of skybrightness and Turbidity. Site was also included in one model as a detection variable, allowing us to determine if there was evidence that detectability varied among sites due to factors that were not measured (Table 1).
Candidate models were ranked according to Akaike's information criteria corrected for small sample size ([AIC.sub.c]). The model with the lowest [AIC.sub.c] was considered to be the most parsimonious among the collection of candidate models, and all models within seven [AIC.sub.c] values of the minimum were considered to have some support (Burnham and Anderson, 2002). A likelihood ratio test was used to verify that the best-fitting detection model was significantly better than the intercept-only detection model (Bolker et al., 2009). Abundances at all three sites were estimated using the best-fit model.
We completed 45 surveys over the three study sites and detected Diamond Darters during 41 of those surveys, resulting in a naive species detection probability of 0.91. Counts at the three study sites ranged from 0 to 42 Diamond Darters detected during a single survey event. Predicted abundances varied widely among the three sites. The Clendenin site had the highest abundance ([??] = 96.1 [58.9-156.8]), followed by the Reamer site ([??] =68.8 [39.3-120.3]), and then the Walgrove site ([??] = 6.5 [2.2-19.6]).
The water temperature model was the only detection model with substantial support ([AIC.sub.c] = 353.5, [w.sub.i] = 0.996; Table 1). The model had a positive co-efficient for water temperature ([beta] = 0.5615, SE = 0.05) and a negative co-efficient for water [temperature.sup.2] ([beta] = -0.0127, se = 0.00). Water temperatures near 22 C resulted in the highest detection probability. Detection probability when surveying at the optimal temperature was approximately 6% and 7.5% greater than when surveying at 16 C and 29 C, respectively (Fig. 2). Both water temperature and water [temperature.sup.2] were statistically significant in the analysis (P < 0.05), and the likelihood-ratio test was statistically significant ([chi square] = 27.34, P < 0.001), indicating water temperature significantly impacted detection probability.
Our study indicated water temperature was an important environmental factor for detecting the endangered Diamond Darter. Presumably, water temperature influenced the availability of fish to be detected rather than the probability of an observer detecting an individual that was available. There are a number of reasons why temperature may alter the behavior or micro-distribution of Diamond Darters, affecting their availability for sampling. One potential explanation involves the effect of temperature on metabolic rate. Fish activity, feeding, and other fundamental behaviors may vary with temperature changes that occur within a normal range of physiological tolerance (Stoner, 2004). Multiple studies have found with most species of fish, food consumption normally increases steadily with temperature and then decreases rapidly--possibly because of limitations in the ability of the respirator)' and circulatory systems to meet the high oxygen demand of respiring tissues (Brett, 1979; Jobling, 1997). While this potential explanation is plausible, we currently have no evidence Diamond Darters are feeding during these night-time surveys, or whether this activity would be influenced by the temperature range we surveyed. Therefore, additional research is necessary to determine causation.
Another explanation that may explain why temperature is affecting detection probability-may involve darters shifting their microhabitat use in an attempt to thermoregulate. In a study conducted by Ingersoll and Claussen (1984), this behavior was noted in some species of darters living in stream/riverine systems; however, the shift in habitat use occurred between summer and winter months and not throughout the summer. In a laboratory study conducted by Ruble et al. (2014), Diamond Darters spent more time buried under the sand substrate at temperatures below 15 C and were noted as being consistently above the substrate at water temperatures above 21 C. This lab study did not allow water temperature to exceed 25 C. Because of this, it is difficult to directly relate the results of their study with ours, although it does lend some support to temperature influencing Diamond Darter detection probability.
We found large variability in mean estimated abundance among the three study sites. Although the study was not designed to quantify predictors of abundance, one potential explanation is variation in microhabitat conditions at glide locations. Welsh et al. (2013) found Diamond Darters appeared to be associated with substrate that was primarily composed of sand. Based on our results, it is likely this association is due to higher-abundances at sandier glide sites (i.e., Clendenin and Reamer). Future research should explicitly investigate factors (e.g.., substrate type) that influence Diamond Darter occupancy and abundance.
The results of this study can be used to improve sampling efficiency and abundance estimation for Diamond Darters. We suggest future surveys seek to sample during optimal water temperatures, as well as include measurements of water temperature for use as a detection covariate in abundance estimates. In addition we found time of night and season were not strong predictors of detection probability. Therefore, timing of surveys is flexible within the ranges we studied, assuming water temperature is appropriate. Future research is needed to determine the optimal number of survey replications for Diamond Darter abundance estimation.
Acknowledgments.--The authors wish to thank NiSource and U.S. Fish and Wildlife Service for funding. We thank Joni Aldinger, Brian Crabill, Kevin Lambert, and Rich Raesly for field assistance. Any use of trade, firm, or product names is for descriptive purposes only and does not imply endorsement by the U.S. Government. This study was performed under the auspices of West Virginia University IACUC protocol 12-0205.
Ali, M. A (ed.). 1992. Rhythms in fishes. Plenum Press, New York, New York, U.S.A. 348 p.
Burnham, K. P. and D. R. Anderson. 2002. Model selection and multimodel inference: a practical information-theoretic approach. Springer-Verlag New York, Inc., New York, New York, U.S.A. 488 p.
Bolker, B. M., M. E. Brooks, C.J. Clark, S. W. Glance, J. R. Poulsen, M. H. H. Stevens, and J. S. S. White. 2009. Generalized linear mixed models: a practical guide for ecology and evolution. Trends Ecol. Evol., 24:127-135.
Brett, J. R. 1979. Environmental Factors and Growth, p. 599-675. In: W. S. Hoar, D.J. Randall, and J. R.
Brett (eds.). Fish Physiology. 8th ed. Academic Press, Inc, New York, New York, U.S.A.
Couturier, T., M. Cheyean, A. Bertolero, G. Astuc, and A. Besnard. 2013. Estimating abundance and population trends when detection is low and highly variable: A comparison of three methods for the Hermann's Tortoise. J. Wildlife Manage., 77:464-462.
Due, D. and L. Madsen. 2011. Models for estimating abundance from repeated counts of an open metapopulation. Biometrics 67:577-587.
Harding, J. M., A. J. Burky, and C. M. Way. 1998. Habitat preferences of the Rainbow Darter, Etheostoma caeruleum, with regard to microhabitat velocity shelters. Copeia, 4:988-997.
Hlohowskyj, I. and T. E. Wissing. 1985. Seasonal changes in the critical thermal maxima of fantail (Etheostoma flahellare), greenside (Etheostoma blennoides), and rainbow (Etheostoma caendeum) darters. Can. J. Zoolog., 63:1629-1633.
Ingersoll., C. G. and D. F. Claussen. 1984. Temperature selection and critical thermal maxima of the fantail darter Etheostoma flahellare, and johnny darter, E. nigrum, related to habitat and season. Environ. Biol. Fish., 11:131-138.
Jobling, M. 1997. Temperature and growth: modulation of growth rate via temperature change. Soc. Exp. Biol. Sem. S., 61:225-254.
Jones, R. N., D. J. Orth, and O. E. Maugham. 1984. Abundance and preferred habitat of the Leopard Darter, Percina pantlierina, in Glover Creek, Oklahoma. Copeia. 1984:374-384.
Joseph, L. N., C. Elkin, T. G. Martin, and H. P. Possingham. 2009. Modeling abundance using Admixture models: the importance of considering ecological mechanisms. Ecol. Appl., 19:631-642.
Kanno, Y, B. H. Letcher, J. C. Vokoun, and E. F. Zipkin. 2014. Spatial variability in adult brook trout (Salvilinus fontinalis) survival within two intensively surveyed headwater stream networks. Can. J. Fish. Aquat. Sci., 71:1010-1019.
Kendall, W. L. 1999. Robustness of closed capture-recapture methods to violations of the closure assumption. Ecology, 80:2517-2525.
Kery, M. and J. A. Royle. 2016. Applied Hierarchical Modeling in Ecology: Analysis of distribution, abundance and species richness in R and BUGS, Volume 1. Academic Press/Elsevier Inc, San Diego, USA. 808 p.
MacKenzie, D. L. J. D. Nichols, G. B. Lachman, S. Droegl, J. A. Royle, and C. A. Langtimm. 2002. Estimating site occupancy rates when detection probability are less than one. Ecology. 83:2248-2255.
Mackenzie, D. I., J. D. Nichols, J. A. Royle, K. H. Pollock, J. E. Hines, and L. L. Bailey. 2006. Occupancy Estimation and Modeling: Inferring Patterns and Dynamics of Species Occurrence. Elsevier, San Diego, California, U.S.A. 324 p.
Mackenzie, D. I., J. A. Royle, J. A. Brown, and J.D. Nichols. 2004. Occupancy estimation and modeling for rare and elusive populations, p. 149-171. In: Thompson, W. L., (ed.). Sampling rare or elusive species; concepts, designs, and techniques for estimating population parameters. Island Press, Washington, DC., U.S.A.
Mazerolle, M. J. 2016. Package 'AICcmodavg': model selection and multimodel inference based on (Q)AIC(c). R package version 2.0-4. <https://cran.r-prqject,org/web/packages/AICcmoda vg/>. Accessed 4 Apr 2016.
McDonald, L. L. 2004. Sampling rare populations, p. 11-42. In: Thompson, W. L., (ed.). Sampling rare or elusive species; concepts, designs, and techniques for estimating population parameters. Island Press, Washington, DC., U.S.A.
Mundahl, N. D. 1990. Heat death offish in shrinking stream pools. Am. Midi. Nat. 123:40-46.
Pitcher, T. P., D. M. Gutherie, and W. R. A. Muntz. 1993. Role of vision in fish behavior, p. 89-121. In: T. P. Pitcher (ed.). Behavior of Teleost Fishes. 2nd ed. Chapman and Hall, London, United Kingdom.
Reebs, S. G. 2002. Plasticity of diel and circadian activity rhythms in fishes. Rev. Fish. Biol. Fisher., 12:349-371.
Reeves, G. H., J. B. Grunbaum, and D. W. Lang. 2009. Seasonal variation in diel behavior and habitat use by age 1+ Steelhead (Oncorhynchus mykiss) in Coast and Cascade Range streams in Oregon, U.S.A. Environ. Biol. Fish., 87:101-111.
Rizzo, A. A., S. A. Welsh, and P. A. Thompson. 2017. A paired-laser photogrammetric method for in situ length measurements of benthic fishes. N. Am. J. Fish. Manage, 37:16-22.
Royle, J. A. 2004. N-mixture models for estimating population size from spatially replicated counts. Biometrics, 60:108-115.
Ruble, C. L., P. L. Rakes, J. R. Shute, and S. A. Welsh. 2014. Captive propagation, reproductive biology, and early life history of the Diamond Darter (Crystallaria cincotta). Am. Midi. Nat., 172:107-118.
Schaefer, J. F., E. Marsh-Matthews, D. E. Spooner, K. B. Gido, and W. J. Matthews. 2003. Effects of barriers and thermal refugia on local movements of threatened leopard darter, Percina pantherina. Environ. Biol. Fish., 66:391-400.
Smith, R. K. and K. D. Fausch. 1997. Thermal tolerance and vegetation preference of Arkansas Darter and Johnny Darter from Colorado plains streams. T. Am. Fish. Soc., 126:676-686.
Stoner, A. W. 2004. Review Paper: Effects of environmental variables on fish feeding: implications for the performance of baited fishing gear and stock assessment. J. Fish. Biol., 65:1445-1471.
USFWS (U.S. Fish and Wildlife Service). 2013. Endangered and threatened wildlife and plants; endangered species status for the Diamond Darter. Federal Register, 78:45074-45095.
Utne-Palm, A. C. 2002. Visual feeding of fish in a turbid environment: Physical and behavioral aspects. Mar. Freshw. Behav. Phy., 35:111-128.
Welsh, S. A. and R. M. Wood. 2008. Crystallaria cincotta, a new species of darter (Teleostei: Percidae) from the Elk River of the Ohio River drainage, West Virginia. Zootaxa, 1680:62-68.
Welsh, S. A., R. M. Wood, and K. R. Sheehan. 2009. Threatened fishes of the world: Crystallaria cincotta Welsh and Wood 2008 (Percidae). Environ. Biol. Fish., 84:191-192.
Welsh, S. A., D. M. Smith, and N. D. Taylor. 2013. Microhabitat use of the Diamond Darter. Pent. Freshw. Fish., 22:587-595.
Welsh, S. A., A. A. Rizzo, and P. A. Thompson. 2014. Distribution and abundance of the Diamond Darter in Elk River, West Virginia. Project report submitted to West Virginia Division of Natural Resources, Elkins, West Virginia.
Wenger, S. J. and M. C. Freeman. 2008. Estimating species occurrence, abundance, and detection probability using zero-inflated distributions. Ecology, 89:2953-2959.
Submitted 8 July 2016
Accepted 20 February 2017
AUSTIN A. RIZZO (1)
Division of Forestry and Natural Resources. West Virginia University, 322 Percival Hall, Morgantown 26506
DONALD J. BROWN
Division of Forestry and Natural Resources, West Virginia University, 322 Percival Hall, Morgantown 26506; US. Forest Service, Northern Research Station, PO Box 404, Parsons 262S7
STUART A. WELSH
US. Geological Sunny, West Virginia Cooperative Fish and Wildlife Research Unit, PO Box 6125, Morgantown 26506
PATRICIA A. THOMPSON
Division of Forestry and Natural Resources, West Virginia University, 322 Percival Hall, Morgantown 26506
(1) Corresponding Author: telephone: (315) 276-6289: f AX: (304)-293-4826: e-mail: firstname.lastname@example.org
Caption: Fig. 1.--Map showing the three study sites (Walgrove, Reamer, and Clendenin Compressor Station) on the Elk River, West Virginia used to estimate abundance and factors affecting detection probability of the endangered Diamond Darter Crystallaria cincotta
Caption: Fig. 2.--The relationship between water temperature and per-individual detection probability of the endangered Diamond Darter Crystallaria cincotta based on 45 surveys conducted at three sites in die Elk River, West Virginia. Bands show the 95% confidence interval at each water temperature
Table 1.--Ten candidate models for explaining variation in detection probability during surveys of the endangered Diamond Darter Crystallaria cincotta. For each model the Akaike Information Criterion adjusted for sample size ([AIC.sub.c]), the delta [AIC.sub.c], [AIC.sub.c] weight ([w.sub.i]), the number of parameters (k), and the maximum log likelihood [-2 log (L) ] are given. The variables evaluated include water temperature (WTEMP), turbidity, amount of light (SQM), gage height, hours after sunset, site, and day of year Covariates Abundance Detection K Site (wtemp + [wtemp.sup.2]) 6 Site (turbidity x sqm) 7 Site gage height 5 Site turbidity 5 Site sqm 5 Site ~ 4 Site hours after sunset 5 Site site 5 Site dav of year 5 Site (day of year + dav of [year.sup.2]) 6 Abundance [AIC.sub.c] [DELTA] [w.sub.i] -2 log (L) [AIC.sub.c] Site 353.5 0.00 0.996 -181.234 Site 365.0 11.54 0.003 -186.703 Site 368.2 14.70 0.001 -189.083 Site 372.0 18.53 0.000 -190.998 Site 375.6 22.11 0.000 -192.789 Site 377.8 24.34 0.000 -194.904 Site 378.1 24.60 0.000 -194.034 Site 378.7 25.28 0.000 -193.874 Site 379.8 26.34 0.000 -194.904 Site 380.8 27.34 0.000 -194.904
|Printer friendly Cite/link Email Feedback|
|Author:||Rizzo, Austin A.; Brown, Donald J.; Welsh, Stuart A.; Thompson, Patricia A.|
|Publication:||The American Midland Naturalist|
|Date:||Jul 1, 2017|
|Previous Article:||Diet composition of age-0 fishes in created habitats of the Lower Missouri River.|
|Next Article:||Effects of woody biomass harvests on a population of plethodontid salamanders in Southeast Indiana.|