2017_Krest_et_al_in_situ_tadpole_CBFO_rept.pdf

i

Use of an In Situ Tadpole Enclosure Study for Developing a Preliminary

Remediation Goal for Lead in Wetland Sediments at Prime Hook National

Wildlife Refuge, Milton, Delaware

CBFO-C16-01

Prepared by:

Sherry Kresta

Alfred E. Pinkneya

Michael A. Rutterb aU.S. Fish and Wildlife Service, Chesapeake Bay Field Office, 177 Admiral Cochrane Drive, Annapolis, Maryland 21401 bDepartment of Mathematics, Penn State Erie, The Behrend College, 5091 Station Road, Erie, PA 16563

U.S. Fish and Wildlife Service Chesapeake Bay Field Office 177 Admiral Cochrane Drive Annapolis, Maryland 21401

January 2017 ii

Executive Summary

The Broadkiln Sportsman’s Club operated a trap-shooting range adjacent to Prime Hook

National Wildlife Refuge, Milton, Delaware from 1962 until 1998. Approximately 8.9 hectares of refuge property fall within the shooting trajectory, portions of which received substantial lead shot deposition. In a 1998 screening, the U.S. Fish and Wildlife Service (USFWS) reported that drop zone surface soil samples had lead shot densities of 56 to 57,868 shot/ft2 (5 to 5376 shot/m2) and concentrations of 128 to 2745 mg Pb/kg dry weight (= parts per million (ppm). The

Department of the Interior (DOI) Central Hazardous Materials Fund (CHF) began funding the

Prime Hook Lead Shot Site investigations in 2000. As an interim remedial action, USFWS determined that soils from the upland area should be removed. In 2002-2003, the top 20 cm of soil and lead shot was excavated. The underlying 20 cm layer was stabilized by adding Portland cement, and the area was backfilled with clean soil and re-seeded with a mixture of grasses.

The adjacent forested wetland habitat consists of a braided stream (unnamed tributary)

/hummock environment supporting aquatic invertebrates, fish, amphibians, reptiles, and birds.

Although not as contaminated as the uplands, lead shot was visible on the forest floor and surface soil concentrations of 1000 to 5000 ppm Pb were detected. Because of concerns about the cost of additional cleanup and the damage to the wetland habitat, USFWS began a CHF-funded monitoring program to assess lead concentrations in groundwater and surface water at the site, and the potential for adverse effects on amphibians. A CHF-funded laboratory study demonstrated that exposure of southern leopard frog tadpoles (Lithobates sphenocephalus) to lead-spiked refuge sediments resulted in bioaccumulation, lethality, and malformations.

Here we report the results of an in situ study with the same species to test effects in the natural environment. Studies were conducted before the upland remediation (spring 2002), and afterwards in 2004, 2005, and 2011. The goals were to 1) determine whether toxic effects and bioaccumulation of lead would occur under natural conditions in the wetland and 2) if the results could be used along with literature and guidance to develop a preliminary remediation goal

(PRG) for possible sediment cleanup based on risk to amphibians. The specific objectives were to determine whether 1) lead was accumulating in the caged tadpoles; 2) lead was adversely affecting the frogs by inhibiting δ-aminolevulinic acid dehydratase (ALAD) activity; and 3) to iii compare residue and ALAD results before and after the upland cleanup. ALAD, a key enzyme involved in heme synthesis, is used as a sensitive biomarker of lead exposure and response in mammals, birds, fish, and amphibians.

The 3 to 8 week tests were conducted in the wetlands at a Hot Spot with an average sediment lead concentration of 4457 ppm, and at a Warm Spot (initially 207 ppm in 2002, then moved because of low water to a location with 13.4 to 75.1 ppm). Tests were also conducted at a reference location (Fred’s Pond) where sediment lead concentrations were 35.7 to 91.1 ppm.

In each year, there were statistically significant differences in mean lead concentrations in all three groups of tadpoles. At the Hot Spot, there was a >50% decrease in the mean (± one standard deviation) lead concentration in tadpoles exposed in 2002 (57.0±21.1 ppm) compared with the post clean-up years (2004: 17.4±6.1 ppm; 2005: 27.7±11.0 ppm; 2011: 28.0±15.3 ppm).

In year to year comparisons, both the Hot Spot and the Warm Spot tadpoles had significantly higher ALAD activity in 2004, 2005, and 2011 than in 2002. The Hot Spot tadpoles also had significantly higher ALAD activity in 2011 compared with those from 2005. Relative to Fred’s

Pond, the percent reduction in lead at the Hot Spot was 91% in 2002, 54% in 2004, 66% in 2005, and 16% in 2011. There was a significant negative linear correlation between ALAD activity and the natural log of lead concentrations in the tadpoles, but the relationship was weak (r = -0.53).

Bioaccumulation of lead and ALAD inhibition occurred at the Hot Spot where we calculated an average sediment concentration of 4457 ppm dry weight. The inconsistent sediment concentrations at the Warm Spot preclude the use of those results for developing a PRG. Based on the CHF-funded laboratory study and Environmental Protection Agency guidance, we propose 312 ppm lead as a PRG for sediments in Prime Hook wetland locations where amphibians breed, to reduce the risk of serious malformations.

iv

Acknowledgments

We appreciate the assistance of our colleagues at the Chesapeake Bay Field Office including

Robbie Callahan and Peter McGowan; the staff of Prime Hook National Wildlife Refuge including Art Coppola, Al Rizzo, George O’Shea, and Annie Larsen. The following researchers at Patuxent Wildlife Research Center made important contributions to the conception of the project, and the methods and performance of the ALAD assay: David Hoffman, Don Sparling, Natalie Karouna-Renier, Katie Schoen, Catherine Maddox, and Sandy Schultz. Funding was provided by the Department of Interior’s Central Hazardous Materials Fund.

v

Table of Contents

Executive Summary ...................................................................................................................... ii Acknowledgments ........................................................................................................................ iv

Table of Contents .......................................................................................................................... v

List of Tables ................................................................................................................................ vi List of Figures ............................................................................................................................... vi Introduction

Methods

Results

Discussion

References

TABLES

FIGURES

vi

List of Tables

Table 1. In situ exposure data; 60 tadpoles exposed at each location.

Table 2. Water quality data: median (minimum, maximum).

Table 3. Sediment lead concentrations (mg/kg dry weight = ppm) at the three enclosure locations.

Table 4. Lead concentrations (mg/kg wet weight=ppm) in whole body tadpoles after in situ exposures.

Table 5. ALAD results (nmol porphobilinogen per g. liver per hour at 37 ºC).

List of Figures

Figure 1a. General location of Prime Hook National Wildlife Refuge, showing the lead shot site.

Figure 1b. Map showing the distances between the Lead Shot Site, Fred’s Pond, and Rectangle Pond.

Figure 2. Map of the Lead Shot Site wetlands showing the Hot Spot and Warm Spots for the in situ exposures

Figure 3. Photo of the enclosure in Fred’s Pond.

Figure 4. Graph showing relationship between ALAD activity (nmol porphobilinogen per g. liver per hour at 37 ºC) vs. ln lead concentration in tadpole liver.

Introduction

Background

The Broadkiln Sportsman’s Club operated a trap-shooting range on a property adjacent to

Prime Hook National Wildlife Refuge in Milton, Delaware (Figure 1a), from 1962 until

1998 (Soeder and Miller 2003). Approximately 8.9 hectares of refuge property fall within the shooting trajectory, portions of which received substantial amounts of lead shot deposition. In a 1998 screening, the U.S. Fish and Wildlife Service (USFWS) reported that drop zone surface soil samples (top 2.5 cm) had lead shot densities of 56 to 57868 shot/ft2 (5 to 5376 shot/m2; n=12) and concentrations of 128 to 2745 mg Pb/kg dry weight (= parts per million (ppm) Pb, n=4; Murphy 1999).

Birds ingest lead shot into their gizzards as grit to aid in the digestion of food (Fisher et

al. 2006). Lead ammunition ingestion, exposure to lead contaminated soil, and secondary poisoning can cause death and sublethal toxic effects in birds and mammals (e.g., Eisler

1988, Vyas et al. 2000; Ma 1996; Pain 1996; Stansley and Roscoe 1996; Tranel and

Kimmel 2009). Surface water and sediment contamination can harm invertebrates, fish, amphibians, reptiles, and waterfowl (Hui 2002; Rabitto et al. 2005; Stansley and Roscoe

1996; Stansley et al. 1997; Tranel and Kimmel 2009).

The Department of the Interior (DOI) Central Hazardous Materials Fund (CHF) began funding the Prime Hook Lead Shot Site investigations in 2000. That year, the USFWS hired a contractor to eliminate wildlife access to the 0.24 hectare upland area with the greatest amounts of lead shot on the surface, by covering it with a synthetic geomembrane (Geotex 315 ST, U.S. Fabrics Inc., Cincinnati, OH). Concentrations under the cover were estimated to be greater than 10,000 ppm. The contractor, Harding ESE, sampled surface soil (top 5 cm) in the forested wetland adjacent to the cover along a 50 yard (45.5 m) square grid (Crowley and Richardson 2001). Concentrations in 64 samples ranged from 9 to 100,000 ppm Pb. Those within the first 100 meters of the edge of the covered area ranged from 77 to 100,000 ppm Pb. In 2001, USFWS conducted an ecological risk assessment and recommended cleanup goals of 421 ppm Pb and 7-9 pellets/ft2 (0.09 m2; Guy et al. 2004). The cleanup footprint consisted of 1.6-2.0 hectares of uplands and wetlands. As an interim remedial action, USFWS determined that soils from the membrane-covered area should be removed immediately. In 2002-2003, the geotextile was removed, and the top eight inches (20 cm) of soil and lead shot were excavated and landfilled. The underlying 20 cm layer was stabilized through addition of

Portland cement, and the area was backfilled with clean soil and re-seeded with a mixture of grasses (Shaw Environmental & Infrastructure 2003).

The adjacent forested wetland habitat consists of a braided stream (unnamed tributary)/hummock environment supporting aquatic invertebrates, fish, amphibians, reptiles, and birds. Although not as contaminated as the uplands, lead shot was visible on the forest floor (especially in the first ten meters adjacent to the upland). Surface soil concentrations of 1000 to 5000 ppm Pb were detected at many locations. In 2000 and

2001, Soeder and Miller (2003) monitored shallow ground water in wells installed in the upland, and in upgradient and downgradient locations within the forested wetland. They reported that lead concentrations in the ground water near a discharge zone into the unnamed tributary ranged from 10 to 1,000 times above natural background concentrations of 1 to 2 µg/L. They stated that acid rain and the acidic wetland environment may be the reasons that lead from the shot leached into the groundwater.

Because of concerns about the cost of additional cleanup and the damage the project would cause to the wetland habitat, USFWS began a CHF-funded monitoring program to assess lead concentrations in groundwater and surface water at the site, and the potential for bioaccumulation and adverse effects in amphibians. As part of the program, Pinkney and Callahan (2014) determined that from 2007 through 2014 there were infrequent detections of filtered lead (at ~5 to 10 µg/L) in upland wells and no detections in upgradient and downgradient wells in the wetland. However, surface water in the unnamed tributary had lead concentrations that far exceeded the chronic state ambient water quality standard, 0.93 µg/L calculated for site hardness of 40 mg/L CaCO3. For example, in the most recent sampling (December 2015), filtered lead concentrations ranged from 5.3 to 10,400 µg/L (USFWS, CBFO, unpublished data).

As part of the amphibian investigation, Sparling et al. (2006) surveyed the wetlands and observed adult green frogs (Lithobates clamitans), southern leopard frogs (L.

sphenocephalus), bullfrogs (L. catesbeianus), Northern cricket frogs (Acris crepitans), and Fowler’s toads (Anaxyrus fowleri) in the wetlands but did not observe tadpoles or egg masses. They conducted a laboratory sediment toxicity study with southern leopard frog tadpoles. Sediments collected from a reference location on the refuge with a lead concentration of 45 ppm dry weight, were spiked with reagent grade lead (II) acetate to obtain concentrations of 75, 180, 540, 2360, 3940, 5520, and 7580 ppm Pb. Tadpoles were exposed beginning at Gosner (1960) stage 26 (early limb bud) through metamorphosis. Tadpoles exposed to >3940 ppm Pb died within 2 to 5 days, with an

LC50 calculated as 3728 ppm Pb. All tadpoles exposed to >2360 ppm Pb displayed severe spinal problems, reduced femur and humerus lengths, and deformed digits.

Sparling et al. (2006) reported that tissue residues (body minus liver) increased with sediment concentration but only the tadpoles exposed to 2360 ppm Pb were statistically elevated above all other groups including the control.

Framework, goals, and objectives

Projects at the Prime Hook Lead Shot Site have been designed to determine the need for further remedial action to protect wildlife. Although the Sparling et al. (2006) laboratory study showed the potential for bioaccumulation and toxicity of the lead-contaminated sediments, an in situ study was needed to test effects in the natural environment. Here we describe exposures of southern leopard frog tadpoles conducted in the wetlands before the upland remediation (spring 2002), and afterwards in 2004, 2005, and 2011. The goals were to 1) determine whether toxic effects and bioaccumulation of lead would occur under natural conditions in the wetland and 2) if the results could be used along with laboratory data and literature to develop a preliminary remediation goal (PRG) for possible sediment cleanup based on risk to amphibians. The study objectives were to determine whether 1) lead was accumulating in the caged tadpoles; 2) lead was adversely affecting the frogs by inhibiting δ-aminolevulinic acid dehydratase (ALAD) activity; and

3) to compare residue and ALAD results before and after the upland cleanup. ALAD, a key enzyme involved in heme synthesis, is used as a sensitive biomarker of lead exposure and response in mammals, birds, fish, and amphibians (Stansley and Roscoe 1996;

Vogaitzis and Loumbourdis 1999; MacFarland 2005).

Methods

In situ exposures

We collected pre-limb bud (Gosner stage 24) southern leopard frog tadpoles from

Rectangle Pond (Figures 1a, b), on the Refuge about 6 km from the lead shot site with no known sources of lead. This pond was used as a source of background soil for the

Sparling et al. (2006) laboratory study in which total lead in surface sediment was 45 ppm Pb dry weight. After being examined for abnormalities and Gosner stage, batches of thirty tadpoles were placed in sealed plastic bags and transported to the study area.

Based on lead concentrations of 1,800 and 3,600 ppm Pb measured in two samples collected in 2000 by Crowley and Richardson (2001) at 200D, 50 L (200 yards deep and

50 yards to the left based on the middle of the five shooting houses), we set this location as the “Hot Spot”. A “Warm Spot” was established at 200D, 250L where Crowley and

Richardson (2001) detected lead at 150 ppm and 710 ppm Pb (Figure 2). These areas had water at a depth of about 0.3 m in the spring of 2002, which we considered adequate to maintain tadpoles through the exposure. The Warm Spot was rapidly drying through the

2002 exposure and it was moved in 2004 to 225D 0L. The new location had concentrations of 53 and 140 ppm Pb in 2000 (Crowley and Richardson 2001). The reference location was established at Fred’s Pond, an approximately ~0.4 hectare pond located about 3 km from the lead shot site, with no known sources of lead (Figure 1b).

At each location a 3.18 mm mesh nylon enclosure (Memphis Net and Twine Co., Inc., Memphis, TN), measuring 122 cm long by 122 cm wide by 61 cm deep with a divider in the center and cover on top, was secured into the sediment with wooden stakes (Figure

3). Large rocks were placed on the bottom to prevent the mesh from rising. Thirty tadpoles were placed in each half of the enclosure. The cover was secured with plastic cable ties to prevent predation by birds and mammals. At least 15 cm of the enclosure was above the water surface so that metamorphosed frogs had access to air.

The exposures were intended to last approximately 6 to 8 weeks, ending when the tadpoles reached Gosner stages 45 and 46 at Fred’s Pond. Once per week, the enclosures were opened and tadpoles were netted, observed and staged. We measured water quality parameters (temperature, dissolved oxygen, pH, and conductivity) with a YSI Model 56 multimeter (YSI, Yellow Springs, OH). On the last day of the exposure (see Table 1 for start and end dates), we opened the enclosures and counted tadpoles into plastic containers containing site water. The containers were placed in iced coolers and transported to the USFWS laboratory where they depurated overnight in spring water.

The following day, they were taken to the USGS Patuxent Wildlife Research Center laboratory (Laurel, MD) for processing. The tadpoles were euthanized with tricaine methanesulfonate (MS-222, Argent Laboratories, Redmond, WA), then placed into labeled 20 mL polyethylene scintillation vials and stored at -80°C.

Each year, we collected a single surface sediment sample (top 10 cm) from each location with an Ekman dredge to document lead concentrations. For the 2011 exposure, the sample was misplaced and we collected a new sample in 2013 (Table 2).

Liver ALAD analysis

Each tadpole was thawed, weighed, and staged. For tadpoles less than Gosner stage 42, the gut coil was removed to eliminate any lead-contaminated sediment that may have been ingested. Since tadpoles stop feeding during metamorphosis, the gut was left intact in tadpoles at Gosner stages >43. The liver was removed and weighed. The remainder of the body was placed in a chemically clean polyethylene jar and shipped on ice to a chemical laboratory for residue analysis (see below).

Liver ALAD activity was determined according to the colorimetric assay presented by

Fujita (1999), with details of buffer and reagent preparation in Patuxent Wildlife

Research Center (1991). Liver tissue was diluted with 0.1 M sodium phosphate buffer

(pH 6.7). The dilution factor was based on liver weight as follows: 0.010 – 0.015 g =

10:1; 0.016–0.025 g = 7:1; 0.026–0.035 g = 6:1; 0.036–0.045 g = 5:1; 0.046 –0.054 g =

4:1, > 0.055 g = 3:1. After the addition of buffer, the samples were homogenized using a handheld tissue homogenizer (for approximately 10 seconds; 2002, 2004, and 2005 samples) or (2011 samples) a Bullet Blender (Next Advance, Averill Park, NY, USA) utilizing stainless steel beads (0.9-2.0 mm blend). For each sample, a 200 µL aliquot of the liver/buffer mixture was transferred to a clean 1.5 mL centrifuge tube and placed on ice for 15 min. A blank, consisting only of 0.1 M sodium phosphate buffer, was also processed concurrently. Freshly prepared, cold ALAD substrate (133 µl) was added to each tube. ALAD substrate consisted of 42 mg of delta-aminolevulinic acid (Sigma-

Aldrich, St. Louis, MO) per 25 ml of 0.1M sodium phosphate buffer (pH 6.7). Two hundred µL of the liver/buffer mixture was placed in each tube, immersed in an ice bath.

The samples were then incubated for one hour at 37°C. To stop the enzyme reaction, the tubes were placed on ice and 133 µl of 10% trichloroacetic acid (TCA) (10% wt/vol in distilled water) was added to each tube and gently mixed for two minutes. Each tube was centrifuged at 4oC for 15 minutes at 3000 g and 100 µl of the supernatant was transferred to each of three wells of a 96-well plate. Then, 100 µl of Ehrlich’s reagent (Hurst and

Hughes 1985) was added to each well. The well plate was mixed for 30 seconds on the platform of a UV/visible spectrophotometer and read at 555 nm every 5 minutes for 20 minutes (15 minutes for 2011). Results were corrected for blank absorbance and the triplicate readings at each interval were averaged. ALAD activity was based on the last reading and calculated in nmol porphobilinogen formed per gram liver per hour at 37ºC according to Fujita (1999).

Chemical analysis

We removed the gut coil to avoid overestimating whole body lead concentrations by including gut material that is not incorporated into tissues (Burger and Snodgrass 2001).

After the gut coil and livers were removed, the tadpoles were stored at -20 ºC and shipped on ice to USFWS Contract Laboratories (Texas Environmental Research Laboratory, College Station TX for 2002 and 2004; Alpha Woods Hole, Inc., Mansfield, MA for 2005 and 2011) for lead analyses. Lead concentrations in sediments and tissues were determined by Inductively Coupled Plasma-Mass Spectrometry according to methods specified by the USFWS Analytical Control Facility (Shepherdstown, WV;

http://www.fws.gov/chemistry/acf_analytical_methods.htm). Quality control procedures included the analysis of duplicates, blanks, spikes, and standard reference materials

(http://www.fws.gov/chemistry/acf_qaqc.html). Sediment data were reported as mg

Pb/kg (ppm) dry weight and tadpole data were as mg Pb/kg wet weight. Detection limits for the tadpoles varied from 0.0008 to 0.266 mg/kg wet weight, with the higher detection limits in the tadpole samples with higher lead concentrations. Detection limits for sediments were <0.1 mg/kg dry weight except for highly contaminated samples which had to be diluted, in one case resulting in a detection limit of 196 mg/kg dry weight.

Data analysis and interpretation

Habitat differences between Fred’s Pond and the wetland locations (pond vs. pockets of standing water) precluded using survival and Gosner stage as endpoints. ALAD activity and lead residues were analyzed using ANOVA and analysis of covariance (ANCOVA)

(Sokal and Rohlf 2011). All statistical analyses were conducted using statistical computing software R (R Core Team 2015).

To satisfy the parametric assumptions, nmol porphobilinogen per gram per hour was transformed using the fourth root. ANCOVA was performed testing body weight, liver weight, and Gosner stage as possible covariates. Gosner stage was discretized by binning as stages 36-41 (developing), 42-45 (metamorph), and 46 (adult). Location comparisons for each year and year to year comparisons for each location were based on the Scheffe

(1999) multiple comparison procedure. In all analyses, we used α = 0.05 as the test for significance.

Whole body lead data were first analyzed by two factor ANCOVA testing body weight, liver weight, and Gosner stage as possible covariates and using natural log transformed values to meet parametric assumptions. Because no covariates were significant, we used two factor ANOVA for statistical comparisons. The Scheffe procedure was used to compare the three locations for each year and to compare results from among years at each location (Scheffe 1959). Pearson product-moment correlation (Sokal and Rohlf http://www.fws.gov/chemistry/acf_analytical_methods.htm http://www.fws.gov/chemistry/acf_qaqc.html

2011) was used to test for linear dependencies between nmol porphobilinogen per gram per hour and ln lead residue for each tadpole.

Results from the study were interpreted consistent with U.S. Environmental Protection

Agency (EPA) guidance on contaminated sediment management at hazardous waste sites

(EPA 2005). This framework requires development of remedial action objectives

(RAOs), which are intended to provide a general description of what the cleanup is expected to accomplish. For the Prime Hook Lead Shot Site, the RAO is to protect wildlife (including amphibians) residing in the wetlands from adverse effects due to exposure to lead. The next step is to use the results of this study, the literature, and EPA (2007) guidance to propose a PRG. Efromson et al. (1997) defined a PRG as an upper concentration limit for a chemical in a specific medium that is anticipated to be protective of human health or the environment.

Results

Exposure duration and water quality monitoring data

Although a 6 to 8 week exposure was intended, the depth of the water in the enclosures and the stage of development were the major considerations for deciding when to end the study. In 2002, 2004, and 2005 the exposures were for 44, 51, and 57 days, respectively

(Table 1). A cold spring delayed the start of the 2011 study until May 17, and then a lack of rain and dropping water levels forced us to end the study on June 8 (Day 22). Often the

Hot Spot and Warm Spot enclosures had water depth of only 10 to 15 cm.

Water quality data are summarized as median, minimum, and maximum for temperature, dissolved oxygen, pH, and conductivity for each location in each year (Table 2). The dissolved oxygen content of the waters in all locations was frequently less than 2 mg/L.

Lead concentrations in sediment

Sediment lead concentrations were >2,140 ppm Pb at the Hot Spot in each year’s sample

(Table 3). The Warm Spot had a lead concentration of 207 ppm Pb in 2002, after which it was moved (because of low water) and concentrations ranged from 13.4 to 75.1 ppm Pb.

The concentrations at Fred’s Pond ranged from 15.2 to 91.1 ppm Pb.

Lead concentrations in tadpoles

In each year, there were statistically significant differences in mean lead concentrations between all three groups of tadpoles (Table 4). Tracking the Hot Spot over time, we detected a >50% decrease in the mean (± one standard deviation) lead concentration between the tadpoles exposed in 2002 (57.0±21.1 ppm, n=12) and those sampled post clean-up (2005: 17.4±6.1 ppm, n=12; 2005: 27.7±11.0 ppm, n=9 2011: 28.0±15.3 ppm, n=14; Table 4). The mean lead concentrations in the 2004 and 2011 tadpoles were significantly lower than those from 2002; mean concentrations in 2005 tadpoles were not significantly lower than those from 2002 (p=0.08). For the Warm Spot, the mean concentrations ranged from 0.58 to 1.36 mg/kg and there were no significant differences when mean 2002 concentrations were compared with those from post-cleanup years (at the new Warm Spot). Mean lead concentrations in Fred’s Pond tadpoles were <0.18 ppm in each year.

ALAD activity

The ANCOVA model included liver weight and Gosner stage as covariates to account for the higher activity in the more developed tadpoles and in those with larger liver weight.

In 2002, ALAD activity was significantly depressed in tadpoles from the Warm and Hot

Spots compared with those from Fred’s Pond (Table 5, showing untransformed means and standard deviations). In 2004 and 2005, only tadpoles from the Hot Spot had lower activity. In 2011, there was no significant difference in ALAD activity in either the Warm or Hot Spot tadpoles relative to those from Fred’s Pond. In year to year comparisons, both the Hot Spot and the Warm Spot tadpoles had significantly higher ALAD activity in

2004, 2005, and 2011 than in 2002. The Hot Spot tadpoles also had significantly higher

ALAD activity in 2011 compared with those from 2005. Relative to Fred’s Pond, the activity was lower by 91% in 2002, 54% in 2004, 66% in 2005, and 16% in 2011.

Relationship between lead residues and ALAD activity

There was a significant negative linear correlation between ALAD activity and the natural log of lead concentrations in the tadpoles. The p value was (<0.001) and the correlation coefficient was -0.53 (Figure 4).

Discussion

Variability in exposures to lead

The Hot Spot concentration in 2004 of 76,700 ppm Pb was likely due to the presence of spent shot fragments. Crowley and Richardson (2001) reported a surface lead concentration of 29,000 ppm Pb at 200D, 0L in 2000 in a sample with 77 surface pellets, whereas a resampling of the same location (lead shot not counted) had a concentration of

130 ppm Pb. Thus, it is likely that the concentrations reported in 2002 (5820 ppm Pb), 2005 (2140 ppm Pb), and 2011 (5410 ppm Pb) are more representative of the hot spot than the 2004 concentration. We use the average of these three years, 4457 ppm Pb, for estimating effect concentrations for the purposes of developing a PRG.

Moving the Warm Spot after 2002 (when the lead concentration was 207 ppm Pb) decreased exposure since sediment concentrations thereafter ranged from 13.2 to 75.1 ppm Pb. In fact, in all post cleanup years (2004, 2005, and 2011), the sediment lead concentration at the Warm Spot was slightly less than that at Fred’s Pond (Table 3). The change in exposure makes it difficult to compare across years at the Warm Spot and the overlap in sediment lead concentrations with Fred’s Pond makes the degree of lead exposure uncertain. Therefore, we focus the discussion and interpretation of results on the

Hot Spot in comparison with Fred’s Pond.

A second confounding factor in the study was the shorter exposure in 2011 (22 days vs.

44–57 days), which may have decreased the magnitude of lead accumulation and ALAD inhibition at the Hot Spot, adding uncertainty to comparisons with the other years. Only one published study was found that addressed the effects of exposure duration on lead accumulation in tadpoles. Berzins and Bundy (2002) reported that whole body Pb concentrations in African clawed frog, Xenopus laevis, tadpoles exposed to 20 ppb Pb in water and 90 ppm Pb in sediments were more than twice as high after 6 weeks —

5.11±3.50 ppm (mean ± one standard deviation) — vs. 3 weeks (2.16±1.5 ppm), and 5 weeks (2.27±1.30 ppm) but these differences were not statistically significant.

Water quality

Dissolved oxygen concentrations were low; median concentrations ranged between 0.91 and 2.02 mg/L in 2002, 2004, and 2005 and 0.96 to 6.20 mg/L in 2011. In general, the dissolved oxygen content was similar among the three locations (except in 2011 when it was higher at Fred’s Pond; Table 2). It is unlikely that the hypoxic conditions affected tadpole development because ranid tadpoles can compensate by breathing air (Ultsch et

al. 1999).

Comparisons of Pb residues and ALAD activity across time and space

Focusing on the Hot Spot, it is likely that the upland cleanup conducted in 2002, after the first monitoring study, reduced the input of lead. This is demonstrated by the significant decrease in tadpole lead concentrations from 57.0 ppm (pre-cleanup) in 2002 to less than half that concentration in the post-cleanup years (Table 4). Although the sediment concentration in the Hot Spot sample does not show much change over time (but is only based on a single sample per year), it is likely that the removal of the upland soil the 2002 lowered the inputs of lead through runoff. Lead entering the Hot Spot from runoff of upland soil and detritus would be taken up by tadpoles, which are known to accumulate lead from the water column and through ingestion of sediment, detritus, and food organisms (Sparling and Lowe 1996; Lefcort et al. 1998; Berzins and Bundy 2002). As stated above, moving the Warm Spot after 2002 to a location with sediment concentrations less than those at Fred’s Pond precludes a comparison of pre-and post-cleanup responses. Thus, the analysis is focused on the Hot Spot and Fred’s Pond.

The decrease in the degree of ALAD inhibition over time at the Hot Spot from 91% inhibition relative to Fred’s Pond in 2002 to only 16% inhibition in 2011, suggests decreasing exposure, possibly resulting from the 2002 cleanup of the upland. However, the 2011 results may be affected by the shorter exposure duration of only 22 days. No studies that analyzed ALAD activity in controlled exposures with tadpoles were found in the literature. Fink and Salibian (2005) reviewed toxicological studies of the effects of lead on adult amphibians. They cite studies with wild-caught adult frogs and several laboratory studies. Stansley and Roscoe (1996) measured liver ALAD in green frogs

(Lithobates clamitans) collected from the shot fall zone at a trap and skeet range (75000 ppm Pb in soil) and a control site (74 ppm Pb in soil). Liver ALAD was depressed 64%.

In a study with adult marsh frog, Pelophylax (formerly Rana) ridibunda, Vogiatzis and

Loumbourdis (1999) reported that laboratory exposure for one month at one-tenth of the

LC50 resulted in a 90% inhibition of hepatic ALAD. Other studies have measured ALAD inhibition in the blood of toads injected with lead compounds (Fink and Salibian 2005).

Thus, based on the literature and the current study, it is clear that exposure to lead results in ALAD inhibition in both tadpoles and adult frogs and toads. The degree of inhibition is likely related to both the lead concentration and the duration of exposure. Although we reported a statistically significant negative correlation between ALAD activity and Pb concentration, the correlation was weak (r= -0.53). There was a much clearer separation of the Hot Spot tadpoles from the other locations based on lead concentration than on

ALAD (Figure 4).

Gosner stage, ALAD, and lead residues

Whereas Gosner stage was a significant covariate for ALAD activity, i.e. there was greater activity at later stages of development, this was not the case for lead accumulation. Although the tadpoles were more developed at Fred’s Pond than at the

Warm and Hot Spots, we cannot assert that lead exposure was associated with slower development due to the habitat differences. In the Sparling et al. (2006) laboratory study, the tadpoles exposed to 2380 ppm Pb took a significantly longer time to reach metamorphosis than those exposed to 45 or 75 ppm Pb. An in situ study with similar habitats or a mesocosm study would be needed to further assess that endpoint.

Preliminary Remediation Goal (PRG) for lead in sediment at the wetland for reducing risks to amphibians

Bioaccumulation of lead and ALAD inhibition occurred at the Hot Spot where we calculated an average sediment concentration of 4457 ppm Pb dry weight. The inconsistent sediment concentrations at the Warm Spot preclude the use of those results for developing a PRG. Sparling et al. (2006) was the only study we identified that exposed tadpoles to lead in contaminated sediments and measured both effects and tissue residues. The added value of this study for proposing a PRG for Prime Hook wetlands is that it utilized refuge sediments in the laboratory spiking experiment. The following endpoints were determined: LC50—3728 ppm Pb (all dry weight), threshold for significant bioaccumulation—2360 ppm Pb; increased time to metamorphosis—2360 ppm Pb; EC50 for spinal malformation—579 ppm Pb; threshold for spinal malformations—540 ppm Pb. Sparling et al. (2006) stated that survivorship remained relatively high through 180 ppm Pb and decreased at 540 ppm Pb and higher. At and above 540 ppm Pb, nearly 50% were malformed, and the malformations affected all parts of the skeletal system including spinal curvature, and curved and truncated femurs and other long bones. At 180 ppm Pb, only about 10% were malformed and these were primarily digit abnormalities.

We propose to use the Sparling et al. (2006) results to calculate a preliminary remediation goal for wetland locations where amphibians breed to reduce the risk of serious malformations, recognizing the considerable uncertainty in basing the PRG on this one study. From the study, a sediment concentration of 180 ppm Pb was the No Observed

Adverse Effect Concentration (NOAEC) and 540 ppm Pb was the Lowest Observed

Adverse Effect Concentration (LOAEC) for serious skeletal malformations. EPA (2007) recommends selecting a PRG within the range of the NOAEC and LOAEC. As a midpoint, they recommend calculating a geometric mean rather than an arithmetic mean because environmental data are frequently log-normally distributed. In this case the geometric mean value would be 312 ppm Pb, which we propose as a PRG.

Mean background lead concentrations in Delaware ranged from 4.8 to 23.2 ppm at 8 reference sites at which 20 discrete samples were collected per site (DNREC 2012).

However, none of the sites were located close to Prime Hook NWR. Thus, it is reasonable to select the highest lead concentration at Fred’s Pond 91.1 ppm Pb, which is within the DNREC (1999) state background range of 30 to 100 ppm Pb, as an estimate of background Pb concentration at the refuge. EPA (2002) states that at Superfund hazardous waste sites, it generally does not clean up to concentrations below background.

The 312 ppm PRG is more than three times the estimated background concentration for the refuge. A final remediation level, based on the PRG, may incorporate additional site-specific factors affecting Pb bioavailability such as grain size distribution, pH, and oxygen content.

Recommendations

We documented bioaccumulation and enzyme inhibition in southern leopard frog tadpoles exposed at the Hot Spot in the wetlands where the average concentration was

4457 ppm. USFWS monitoring data (Pinkney, unpublished data and Pinkney and

Callahan 2014) reported water column lead concentrations that exceed the Delaware

Water Quality Standard by several orders of magnitude. Based on the current study, evaluation of the literature, and EPA guidance, we propose 312 ppm lead in sediments as a PRG for avoiding risk to amphibians residing in the wetlands at Prime Hook NWR.

Later steps in the remediation process will require refining the PRG into a Remedial Goal

(RG), which considers site-specific data, and ultimately a cleanup level, which takes into account additional criteria including cost (EPA 2005). Through the CHF Program, USFWS has contracted for an Engineering Evaluation/Cost Analysis (EECA) performed according to EPA (1993) guidance. The EE/CA, scheduled to begin in early 2017, will determine the extent and magnitude of lead contamination in the wetland and will evaluate remedial alternatives.

References

Berzins, D.W. and J.K. Bundy. 2002. Bioaccumulation of lead in Xenopus laevis tadpoles from water and sediment. Environment International 28: 69-77.

Burger, J. and J. Snodgrass. 2001. Metal levels in southern leopard frogs from the

Savannah River Site: location and body compartment effects. Environmental Research

Section A 86:157-166.

Crowley, D.M., and Richardson, N. 2001, Preliminary assessment-site investigation, shotgun range, Prime Hook National Wildlife Refuge, Milton, Delaware: Report prepared under contract no. 1448–98695–98–C007 for U.S. Fish and Wildlife Service, Annapolis, MD by Harding ESE, Engineering and Environmental Services, Pittsburgh, PA, 25 p.

Delaware Department of Natural Resources and Environmental Control (DNREC). 2012.

Statewide soil background study: report of findings. DE-1348. Division of Waste and

Hazardous Substances, Site Investigation & Restoration Section, New Castle, DE 61 p.

Efroymson, R.A., G.W. Suter II, B.E. Sample, and D.S. Jones. 1997. Preliminary remediation goals for ecological endpoints. Prepared for U.S. Department of Energy

Office of Environmental Management, Lockheed Martin Energy Systems, Oak Ridge, TN.

Eisler, R. 1988. Lead hazards to fish, wildlife, and invertebrates: a synoptic review. U.S.

Fish and Wildlife Service. Biological Report 85(1.14). 134 p.

Fink, N.E. and A. Salibian. 2005. Toxicological studies in adult amphibians: effects of lead. Applied Herpetology 2:311-333.

Fisher, I.J., D.J. Pain, and V.G. Thomas. 2006. A review of lead poisoning from ammunition sources in terrestrial birds. Biological Conservation 131:421-432.

Fujita, H. 1999. Measurement of δ-aminolevulinate dehydratase activity. In: Current

Protocols in Toxicology, M. Maines, L.G. Costa, E. Hodgson, D.J. Reed, and I.G. Sipes, editors. John Wiley & Sons, New York. pp. 8.6.1-8.6.11.

Gosner, K.L. 1960. A simplified table for staging anuran embryos and larvae with notes on identification. Herpetologica 16:183-190.

Guy, C.P., M. Huston, S. Krest, and D. R. Murphy. 2004. Ecological risk assessment for the Prime Hook National Wildlife Refuge lead shot site, Milton, Delaware. CBFO-C03-

04. U.S. Fish and Wildlife Service, Annapolis, MD.

Hui, C. 2002. Lead distribution throughout soil, flora, and an invertebrate at a wetland skeet range. Journal of Toxicology Environmental Health Part A 65:1093–1107.

Hurst, P.L., and P.M. Hughes 1985. Poor correlation between erythrocytic 5-aminolevulinic acid dehydratase activity and blood lead concentration in a group of 11-year-old children. Clinical Chemistry 31:1244-1245.

Lefcort, H., R.A. Meguire, L.H. Wilson, and W.F. Ettinger. 1998. Heavy metals alter the survival, growth, metamorphosis, and antipredatory behavior of Columbia spotted frog

(Rana luteiventris) tadpoles. Archives of Environmental Contamination and Toxicology

35:447-456.

Ma, W. 1996. Lead in mammals. In: Environmental Contaminants in Wildlife, Interpreting Tissue Concentrations. Beyer W.N., G.H. Heinz, and A.W. Redmon-

Norwood, editors. Lewis Publishers: Boca Raton, FL. pp. 281–296.

McFarland, C.A. 2005. Delta-aminolevulinic acid dehydratase (ALAD) characterization and use in metal exposure assessments. Dissertation, Texas Tech University, Lubbock, TX

Murphy, D.R. 1999. Prime Hook National Wildlife Refuge lead shot survey. U.S. Fish and Wildlife Service, Chesapeake Bay Field Office, Annapolis, MD. CBFO-C99-02.

Pain, D.J. 1996. Lead in waterfowl. In: Environmental Contaminants in Wildlife, Interpreting Tissue Concentrations. Beyer W.N., G.H. Heinz, and A.W. Redmon-

Norwood, editors. Lewis Publishers, Boca Raton, FL. pp. 251-264.

Patuxent Wildlife Research Center. 1991. Technical operating procedure. ALAD (delta-aminolevulinic acid dehydratase) determination in red blood cells. Laurel, MD. 7 p.

Pinkney, A.E. and C.R. Callahan. 2014. Ground water and surface water monitoring of the Prime Hook Lead Shot Site, Prime Hook National Wildlife Refuge, Milton, Delaware

(October 2007-December 2014). U.S. Fish and Wildlife Service, Chesapeake Bay Field

Office, Annapolis, MD. CBFO-C14-01.

R Core Team. 2015. R: A language and environment for statistical computing. R

Foundation for Statistical Computing, Vienna, Austria.

Rabitto, I.S., J.R.M. Alves Costa, H.C. Silva de Assis, E. Pelletier, F.M. Akaishi, A.

Anjos, M.A.F. Randi, and C.A. Oliveira Ribeiro. 2005. Effects of dietary Pb(II) and tributyltin on neotropical fish, Hoplias malabaricus: histopathological and biochemical findings. Ecotoxicology and Environmental Safety 60:147-156.

Scheffé, H. 1959. The Analysis of Variance, John Wiley & Sons, New York.

Shaw Environmental & Infrastructure. 2003. Prime Hook National Wildlife Refuge, Milton, DE. Lead shot removal Phase II. Final report. Shaw Environmental &

Infrastructure, Monroeville, PA.

Soeder, D., and C. Miller. 2003. Ground-water Contamination from lead shot at Prime

Hook National Wildlife Refuge, Sussex County, Delaware. USGS Water Resources

Investigation Report 02-4282.

Sokal R, and F. Rohlf. 2011. Biometry, 4th ed., W.H. Freeman, New York.

Sparling, D.W. and T.P Lowe. 1996. Metal concentrations of tadpoles in experimental ponds. Environmental Pollution 91:149-159.

Sparling, D.W., S. Krest, and M. Ortiz-Santaliestra. 2016. Effects of lead-contaminated sediment on Rana sphenocephala tadpoles. Archives of Environmental Contamination and Toxicology 51:458-466.

Stansley, W., and D.E. Roscoe. 1996. The uptake and effects of lead in small mammals and frogs at a trap and skeet range. Archives of Environmental Contamination and

Toxicology 30:220-226.

Stansley, W., M.A. Kosenak, J.E. Huffman, and D.E. Roscoe. 1997. Effects of lead contaminated surface water from a trap and skeet range on frog hatching and development. Environmental Pollution 96:69-74.

Tranel, M.A., and R.O. Kimmel. 2009. Impacts of lead ammunition on wildlife, the environment, and human health—A literature review and implications for Minnesota. In:

Ingestion of Lead from Spent Ammunition: Implications for Wildlife and Humans. R. T.

Watson, M. Fuller, M. Pokras, and W. G. Hunt, editors. The Peregrine Fund, Boise, ID.

pp. 318-337.

Ultsch, G.R., D.F. Bradford, and J. Freda. 1999. Physiology: Coping with the environment. In: Tadpoles. The Biology of Anuran Larvae, R.W. McDiarmid and R.

Altig, editors. The University of Chicago Press, Chicago, IL. pp. 189-214.

U.S. Environmental Protection Agency (EPA). 1993. Guidance for conducting non-time-critical removal actions under CERCLA . Office of Solid Waste and Emergency

Response, Washington, DC. EPA 540-R-93-057.

U.S. Environmental Protection Agency (EPA). 2002. Role of background in the

CERCLA cleanup process. Office of Solid Waste and Emergency Response, Washington, DC. OSWER 9285.6-07P, April 26, 2002.

U.S. Environmental Protection Agency (EPA).2005. Contaminated sediment remediation guidance for hazardous waste sites. Office of Solid Waste and Emergency Response, Washington, DC. EPA 540-R-05-012.

U. S. Environmental Protection Agency (EPA). 2007. The rule of five: A novel approach to derive PRGs; presented by M.S. Greenberg at the National Defense Industrial

Association (NDIA) Joint Services Environmental Management Conference & Exhibition

(JSEM) in Columbus, Ohio, May 22.

http://proceedings.ndia.org/jsem2007/4039_Greenberg.pdf

Vogiatzis, A.K., and N.S. Lombourdis. 1999. Exposure of Rana ridibunda to Pb. I: Study of Pb accumulation in various tissues and hepatic delta-aminolevulinic acid dehydratase activity. Journal of Applied Toxicology 19:25–29.

Vyas, N.B., J.W. Spann, G.H. Heinz, W.N. Beyer, J.A. Jaquette, and J.M. Mengelkoch.

2000. Lead poisoning of passerines at a trap and skeet range. Environmental Pollution

107:159-166.

TABLES

Table 1. In situ exposure data; 60 tadpoles exposed at each location.

Year Location Grid position

Dates Days Surv-ivors

Gosner stage range at end of exposure

2002 4/30-6/13 44 Fred’s Pond 16 37-46

Warm spot 200D, 250L

10 37

Hot spot 200D, 50 L

12 <36-37

2004 5/4-6/24 51 Fred’s Pond 12 36-46

Warm spot 225D, 0L 20 <36-45 Hot spot 200D, 50

L 22 <36-43

2005 5/5-7/7 57 Fred’s Pond 1 +9 wilda 44-46

Warm spot 225D, 0L 42 <36-46 Hot spot 200D, 50

L 9 <36-42

2011 5/17-6/8 22b Fred’s Pond 7 39-45

Warm spot 225D, 0L 43 <36-39 Hot spot 200D, 50

L 43 <36-39 aOnly one individual found inside enclosure, 9 wild frogs were found outside and assumed to have escaped.

bEnded exposure because of low water levels in the two wetland locations.

Table 2. Water quality data: median (minimum, maximum). DO = dissolved oxygen.

Temp. (ºC) DO (mg/L) pH (units) Conductivity (msiemens/cm)

2002 (n=7) Fred’s Pond 20.4 (13.2, 26.4) 0.91 (0.54, 5.23) 5.83 (5.56, 6.00) 0.148 (0.108, 0.241) Warm Spot 17.2 (13.8, 23.4) 1.51 (0.52, 2.45) 5.36 (5.23, 5.73) 0.088 (0.042, 0.740) Hot Spot 19.7 (14.4, 23.9) 1.40 (0.52, 4.49) 5.40 (5.31, 5.70) 0.068 (0.058, 0.110)

2004 (n=8) Fred’s Pond 22.3 (18.6, 26.2) 1.44 (0.40, 3.47) 6.00 (5.27, 6.66) 0.180 (0.100, 0.222) Warm Spot 19.4 (14.7, 22.8) 1.19 (0.81, 3.03) 6.16 (5.94, 6.43) 0.130 (0.096, 0.155) Hot Spot 20.7 (17.9, 25.7) 1.34 (0.92, 2.09) 6.04 (5.80, 6.40) 0.110 (0.058, 0.700)

2005 (n=8) Fred’s Pond 21.0 (13.5, 25.5) 2.02 (0.69, 6.17) 6.01 (5.64, 6.17) 0.135 (0.112, 0.142) Warm Spot 18.2 (11.7, 23.4) 1.11 (0.96, 3.15) 6.01 (5.57, 6.33) 0.117 (0.109, 0.136) Hot Spot 19.0 (12.6, 24.7) 1.81 (0.81, 3.21) 5.81 (5.59, 6.28) 0.097 (0.072, 0.780)

2011 (n=3) Fred’s Pond 24.3 (21.8, 29.2) 6.20 (1.25, 7.40) 5.94 (5.64, 5.97) 0.684 (0.353, 1.03) Warm Spot 19.8 (16.6, 21.1) 0.96 (0.84, 1.20) 5.07 (4.89, 6.20) 0.100 (0.075, 0.225) Hot Spot 19.4 (17.0, 20.8) 0.98 (0.60, 1.30) 4.95 (4.77, 5.15) 0.100 (0.080, 0.140)

Table 3. Sediment lead concentrations (mg/kg dry weight = ppm) at the three enclosure locations. Concentrations in bold exceed the proposed preliminary remediation goal (PRG) of 312 ppm.

Sampling Date

Location Conc.

5/21/2002 Fred’s Pond 35.7 Warm Spot 207

Hot Spot 5,820

6/16/2004 Fred’s Pond 44.3 Warm Spot 13.4

Hot Spot 76,700

7/7/2005 Fred’s Pond 15.2 Warm Spot 14.8

Hot Spot 2,140

6/25/2013 Fred’s Pond 91.1 Warm Spot 75.1

Hot Spot 5,410

Table 4. Lead concentrations (mg/kg wet weight=ppm) in whole body tadpoles after in situ exposures. Based on two factor ANOVA (site and year) on natural log-transformed data, groups with different capital letters are significantly different according to Scheffe’s procedure (p<0.05).

Location Freq. det. Mean + Standard Deviation

Statistics

Fred’s Pond 14/14 0.18+0.04 A All groups had significantly different (p<0.05) mean lead concentrations

Warm Spot 5/5 1.34+0.50 B Hot Spot 12/12 57.0+21.1 C

Fred’s Pond 12/12 0.06+0.01 A All groups had significantly different (p<0.05) mean lead concentrations

Warm Spot 12/12 0.65+0.33 B Hot Spot 12/12 17.4+6.1 C

Fred’s Pond 9/9 0.08+0.02 A All groups had significantly different (p<0.05) mean lead concentrations

Warm Spot 20/20 1.36+0.59 B Hot Spot 9/9 27.7+11.0 C

Fred’s Pond 5/7* 0.049±0.043 A All groups had significantly different (p<0.05) mean lead concentrations

Warm Spot 14/14 0.58±0.21 B Hot Spot 14/14 28.0±15.3 C

By year (only showing comparisons with significant (p<0.05) differences) Fred’s Pond 2002 vs. 2004 (p=0.03), 2002 vs. 2011 (p=0.01) Warm Spot 2004 vs. 2005 (p=0.01), 2005 vs. 2011 (p=0.003) Hot Spot 2002 vs. 2004 (p<0.001), 2002 vs. 2011 (p=0.02)

*Detection limits (mg/kg wet weight): 2002—0.0001 to 0.150 (based on need to dilute Hot Spot samples); 2004—0.003 to 0.08; 2005—0.018 to 0.18; 2011—0.022 to 0.266. In 2011, two samples were reported as non-detected (at <0.069 and <0.045); both were treated as 0 for the purpose of calculating the mean and standard deviation.

Table 5. ALAD results (nmol porphobilinogen per g. liver per hour at 37ºC). ANCOVA was performed with site and year as the main effects, site: year interactions significant, and included liver weight and Gosner stage as significant covariates. Groups with different capital letters are significantly different (p<0.05) according to Scheffe’s procedure. Statistics performed on fourth root transformed data.

Location N Mean +…