2011 NORTHEASTERN NATURALIST 18(1):61–72 Benthic Diatoms and Heavy Metals in East Foundry Cove and Constitution Marsh, NY, Post Superfund Restoration Betsey H. Hallihan1,2,* and Donald R. Roeder3 Abstract - East Foundry Cove and Constitution Marsh are located on the east side of the Hudson River, 85 km upriver from lower Manhattan. Between 1965 and 1971, Marathon Battery Company discharged an estimated 51,004 kg of particulate cadmium, and 1569 kg of soluble cadmium were discharged directly into East Foundry Cove, which became a Superfund site in 1994. Dredging and restoration of East Foundry Cove were completed in 1996. Cadmium concentrations were reduced from greater than 900 mg/kg to less than 3 mg/kg in East Foundry Cove. Constitution Marsh and South Cove were not included in Superfund restoration. For this study, 738 surface sediment samples were collected from 15 sampling sites on 8 occasions from 18 September 1997 to 30 October 1998. Diatoms were identified and enumerated for each collected sample. These data were analyzed using Bray-Curtis similarity, MDS ordination, analysis of similarity (ANOSIM), and similarity of percentages (SIMPER). Cadmium concentrations (mg/kg) were also determined for each sample. Results show the communities of diatoms were different in two stations (1 and 3) in East Foundry Cove from the remainder of the stations in the study. However, one East Foundry Cove station (2) was more similar to the stations in South Cove and the southern stations of Constitution Marsh. Three of the most abundant taxa, Navicula gregaria, Cyclotella meneghiniana, and Cocconeis placentula var. lineata, were present in similar proportions at all sampling sites. Navicula gregaria, however, occurred in greater relative abundance at stations 4 and 5 and in Constitution Marsh, adjacent to restored East Foundry Cove, than in any other stations. Introduction Marathon Battery Company was built in 1952 by the US Army Corps of Engineers to manufacture nickel-cadmium batteries under military contracts (Hazen 1981). Heavy metals were discharged into the Hudson River at Cold Spring Pier through the town’s waste treatment facility. In 1965, the company was ordered to discontinue using the town’s waste treatment facility, and thereafter until 1971 released heavy metal waste directly into Foundry Brook, which flows into East Foundry Cove. The factory closed in 1979. Resource Engineering (1983) estimated that 179 metric tons of cadmium were discharged from the factory into the Hudson River. They estimated that of this, 53 metric tons of cadmium were deposited into the sediments of East Foundry Cove, resulting in concentrations >900 mg/kg. As a result of these discharges, Foundry Cove earned the distinction of being the world’s largest cadmium dump site. 1Department of Biology, Lehman College, CUNY Bronx, NY. 2Current address - Department of Biology and Medical Lab Technology, Bronx Community College, City University of New York, Bronx, NY 10453. 3Department of Computing, Mathematics and Natural Science, Bard College at Simon’s Rock, Great Barrington, MA 01230. *Corresponding author - Betsey.Hallihan@bcc.cuny.edu. 62 Northeastern Naturalist Vol. 18, No. 1 An attempt to remove the most-contaminated sediments in 1973 was generally unsuccessful (Kneip and Hazen 1979). In 1975, about 30% of Foundry Cove still had surface cadmium levels in excess of 1000 mg/kg. Although by 1983 that area was reduced to 8%, the site was declared a priority one Superfund site, and dredging and restoration were completed in 1996 (EPA Record of Decision R02- 89/97). The relatively uncontaminated Constitution Marsh and South Cove were not included in the Superfund cleanup. Unfortunately, little data were collected on benthic diatoms prior to the cleanup. The New York State Department of Environmental Conservation (1985) used artificial substrates and reported Nitzschia parvula W. Sm., Coscinodiscus lacustris Grunow, and Melosira granulata (Ehrenb.) Ralfs as the dominant diatoms pre-restoration in East Foundry Cove. “Cyclotella meneghiniana and Stephanodiscus astrea were found in relatively high numbers in Tivoli Bays (their control site) but not at all in East (Foundry) Cove.” (The New York State Department of Environmental Conservation 1985). The objective of this research was to investigate whether diatom (Bacillariophyceae) communities in Constitution Marsh, South Cove, and East Foundry Cove differed, and what were the cadmium concentrations in those three areas post superfund remediation. Field Site East Foundry Cove, Constitution Marsh, and South Cove are located on the east side of the Hudson River, 85 km upriver from lower Manhattan, in Cold Spring, NY (Fig. 1). The coves and marsh are across the river from the northern section of West Point Military Academy and Storm King Mountain. Salinity at West Point averages five parts per thousand, but fluctuates seasonally, tidally, and with varying freshwater runoff. There are two high tides and two low tides occurring every 24 hours and 50 minutes, with an average range at West Point of 0.8 m (Limburg et al. 1986). Constitution Marsh, south of East Foundry Cove, is separated from the Hudson River by Constitution Island. The Marsh is a 161-ha contiguous area of emergent marsh plants, predominately Typha angustifolia L. (Narrow-leaved Cattail), Hibiscus moscheutos L. (Marsh Mallow), Pontederia cordata L. (Pickerelweed), and mud flats, but the water flows are quite different between the northern and southern areas of the Marsh. East Foundry Cove and South Cove are open water areas and have several freshwater streams flowing into them, thereby reducing the salinity below five parts per thousand. Only during periods of drought does the salinity increase above that level. Foundry Brook provides freshwater to East Foundry Cove, and Indian Brook flows into South Cove adjacent to collection site 12 (Fig.1). Methods All sampling was done after the Superfund cleanup and restoration of East Foundry Cove. Sampling began in September of 1997 and continued until October 1998 on a monthly or twice monthly regiment. In order to evaluate diatom 2011 B.H. Hallihan and D.R. Roeder 63 populations, it was necessary to follow this pattern for a minimum of one year (Ravera 1979). Sampling, however, could not be done during the months of November and December since Constitution Marsh and South Cove are an Audubon bird sanctuary and are off limits to humans during bird-hunting season (J. Rod, Former director of Constitution Marsh Center, 1997 pers. comm.). Stations 13, 14, and 15 were added during the 30 May 1998 collection as additional reference sites. They were then sampled until the end of the study period. Figure 1. Study site showing Constitution Marsh, Foundry Cove, and the location of the Marathon Battery Factory. Numbers 1–15 refer to sampling sites. 64 Northeastern Naturalist Vol. 18, No. 1 Surface sediments Surface sediments were collected in triplicate from 15 sampling sites (Fig. 1): three in East Foundry Cove, six in northern Constitution Marsh, two in southern Constitution Marsh, and four in South Cove with the LaMotte surface-sediment grab sampler. Collection depths varied from 1–4 cm, depending on the consistency of the sediment. Sediments were collected between low and mid-tide on the ebb tide or on the flood tide. At each sampling location, three replicates were collected within approximately 5 m of each other. All collected sediments were placed in ziplock bags, transported to Lehman College, and frozen for later analysis. Sediment composition and depths were not determined. A total of 18 collections were made from September 1997 to October 1998. Sediment digestion The sediments were digested of organic matter using the EPA method 3051 using a CEM (Matthews, NC) MDS 2000 microwave digester. This method uses a Teflon closed-vessel, temperature- and pressure-controlled system. The temperature inside each vessel was 175 °C. Two grams of wet-weight sediment were dried in a 100 °C oven in glass petri dishes. Once dried, 0.5 grams were weighed and placed in digestion tubes along with 10 ml of high-purity, metal-free, nitric acid. The sediment portion left after digestion, containing diatoms, was rinsed with distilled water, centrifuged, decanted, and washed three times with distilled water. After the final rinse, the digested sediment samples were added to 20 ml of distilled water. A few milliliters were pipetted onto cover slips in plastic evaporation trays and left undisturbed until evaporation was complete. The cover slips were then mounted onto glass slides with Naphrax (refractive index 1.69), mounting medium. Cadmium concentrations from the same samples were determined using a Perkin Elmer (Norwalk, CT) ICP 3000 XL (inductively coupled plasma-optical emission spectrometry) with a one-milliliter automatic sampler. Cadmium was measured at 214.4 nm and 228.8 nm wavelengths. Diatom analysis A minimum of 300 diatom valves were counted and identified for each station replicate and collection date using a Series “R” Bausch and Lomb microscope equipped with a 98x fluorite objective and 12.5x oculars. The data from the three replicates (total of 900+ valves) were then combined for each collection, and these data from all collections were then pooled together for comparison between stations for the non-parametric analyses. Taxa were identified according to Barber and Haworth (1981), Dodd (1987), Hustedt (1930, 1939, 1953, 1959), Krammer and Lange-Bertalot (1986, 1988, 1991a, 1991b), Patrick and Reimer (1967, 1975), Stoermer et al. (1999), and Vinyard (1979). Metrics of biotic integrity The Bray-Curtis community similarity index (Bray and Curtis 1957, Clark and Warwick 2001) was calculated between stations and collection dates for 2011 B.H. Hallihan and D.R. Roeder 65 both diatom taxa and cadmium data. No data transformations were performed for Bray-Curtis similarity. These data were then used to calculate multivariate analyses including multidimensional scaling (MDS), cluster analysis (CA), similarity of percentage (SIMPER), and analysis of similarity (ANOSIM) using Primer 5 software (Clark and Warwick 2001). The MDS plots of diatom similarity, for pooled data from all 15 stations were overlaid with cadmium concentrations using “bubble plots” (Clarke and Warwick 2001). Two-way ANOVA were performed using Statview software to determine if significant differences in cadmium, using mean cadmium from the two measured wavelengths, dates, and station sites, existed. Stations one through twelve were used in calculating ANOVA. Results Cadmium concentrations Figure 2 shows the mean cadmium concentrations at each station during the study period. Note that the restored area (East Foundry Cove) stations 1–3 and the southern Constitution Marsh and South Cove stations (11–15) have reduced cadmium levels. Stations 4–10 in the unremediated upper Constitution Marsh contain slightly elevated levels of cadmium. The results from ANOVA (Table 1) show these differences between stations to be highly statistically significant. Figure 2. Mean Cadmium concentrations in sediments of sampling sites 1–15 between October 1997 and October 1998. 66 Northeastern Naturalist Vol. 18, No. 1 Diatom species composition One hundred eighty-eight taxa were identified. The more dominant taxa, herein defined as comprising ten percent of all taxa sometime during the study period, are Achnanthes chilidanos Hohn and Hellerman, A. haukiana Grun. in Cleve & Grun., A. haukiana var. rostrata Schulz, A. lanceolata (Breb. ex Kutz.) Grun. in Cleve & Grun., Actinocyclus normanii (Greg. ex Grev.) Hust. ex VanLand., Cocconeis neodimuata Krammer, C. placentula var. lineata (Ehrenb.) van Heurck, Cyclotella meneghiniana Kuetz, Navicula gregaria Donk., N. odiosa Wallace, N. salinarum var. intermedia (Grun. in van Heurck) Cleve, N. viridula (Kuetz) Ehrenb., Nitzschia frustulum (Kuetz) Grun. in Cleve & Grun., N. frustrulum var. perpusilla (Rabenh.) Grun. in van Heurck, and N. microcephala Grun. in Cleve & Grun. A complete list of diatoms and their relative abundances can be found in Hallihan (2006). Bray-Curtis community similarity index The diatom community similarity between stations is illustrated by a dendrogram (Fig. 3). The dendrogram shows three major clusters. Note the close similarity between stations four, five, six, eight, and nine (Constitution Marsh) in Figure 3. The first cluster includes stations 4–10 (sites with slightly elevated Table 1. ANOVA table of mean cadmium for stations 1–12 (East Foundry Cove, Constitution Marsh, and South Cove) d.f. Sum of squares Mean square F-value P-value R-square Day 17 241.820 14.225 7.194 less than 0.001 0.03489 Sites 11 4840.935 440.085 222.555 less than 0.001 0.69852 Day*sites 187 989.244 5.290 2.675 less than 0.001 0.14274 Unexplained 434 858.202 1.977 0.12383 Total 649 6930.201 Figure 3. Dendrogram of diatom community similarity of stations 1–15 using pooled data from all collections. 2011 B.H. Hallihan and D.R. Roeder 67 cadmium). Stations 2 and 11–15 form a second cluster (reference sites except station 2), and stations 1 and 3 (restored East Foundry Cove) form a third cluster. Stations 1 and 3 in East Foundry Cove were not similar to other stations throughout the study, but there was a 74% similarity between these two stations. MDS ordination and cadmium concentrations The MDS plot (Fig. 4) shows the same data as Fig. 3 but with cadmium levels (“bubbles”) overlying the two dimensional representation of similarity. Note that the sites with elevated cadmium (Cluster A in Fig. 3) are closely associated in Figure 4. ANOSIM and SIMPER Results of ANOSIM indicate that the three clusters (A, B, and C) are different (Fig. 5). The global R of ANOSIM is 0.83, indicating a difference between the groups. More important are the R-values between clusters A, B, and C. The R-value between clusters A and B is 0.80 and is less than the global R. However, the comparison of clusters A and C has a R-value of 0.97, while the comparison between clusters B and C has a R-value of 0.99. A summary of SIMPER results is in Table 2. The similarity of taxa in cluster A is 84.9 percent. The average similarity of cluster B and C is 81.6% and 74.3%, respectively. The ten most common diatoms in each cluster are summarized in Table 2. Note that while there are similarities—e.g., Navicula gregaria is dominant in all three clusters—cluster A contains Bacillaria paradoxa Gmelin (= Bacillaria paxillifer (O.F. Müll.) Hendy) and the others do not, and Achnanthes haukiana var. rostrata was not a common diatom in cluster B but was in A and C. Discussion The cleanup and restoration of East Foundry Cove greatly reduced cadmium to levels found at the reference sites in South Cove. The exception to this is station 2, which is located on the southwestern shore of East Foundry Cove. The mean concentration of Cd there was 2.1 mg/kg. Although, this is not a significant amount, it Figure 4. MDS plot of pooled diatom data from all collections from stations 1–15 overlain with “bubbles” showing relative Cd concentrations. 68 Northeastern Naturalist Vol. 18, No. 1 is greater than stations 1 and 3 and also greater than the reference stations in South Cove. The ANOVA results confirm this. Station 2 is located in the area where tidal currents are greater during the ebb and flood tides. This area may be receiving metals from the Hudson River and/or from northern Constitution Marsh. The reference stations, 12–15, had very low levels of cadmium as was expected. The mean cadmium of these stations was ≤0.2mg/kg, with the exception of the 3 February 1998 collection where the mean concentration was 1.6 mg/kg. Cadmium concentrations were the greatest at station 4, which is located on the southeastern edge of Foundry Pond but was not subject to the restoration. The pond is separated from Foundry Cove by an old causeway. This is the closest site to East Foundry Cove in this study and is influenced by water fluctuations during the ebb and flood tides. The mean cadmium there was 8.6 mg/kg but varied throughout the collection period. The lowest cadmium level, 2.1 mg/kg, was on 6 January 1998, and the highest (13.5 mg/kg) during the 27 July 1998 collection. The temporal fluctuation of cadmium will be discussed in another publication. Two other studies have reported higher cadmium levels than are reported here. Advanced GeoServices Engineering P.C. (project number 95-219-04) was retained to do the long-term monitoring of a number of chemical parameters, including cadmium from 1995 through the present. In 1997–1998, they found between 0.39 and 104 mg/kg Cd (= 30.3) in East Foundry Cove, and 70.9–529 mg/kg Cd (= 175) in Constitution Marsh (Marano 2010). Mackie et al. (2007) also found higher levels of Cd in Constitution Marsh surface sediments (= 101.5 mg/kg ± 41.5) than in Foundry Cove’s western margin (= 14.3 ± 4.9 mg/kg), Foundry Cove’s northern margin (= 59.7 ± 16.8 mg/kg), or South Cove (= 5.4 ± 1.3 mg/kg) surface sediments. Figure 5. Analysis of similarity (ANOSIM) of pooled diatom data (global R) showing similarities between clusters identified in Figure 3. 2011 B.H. Hallihan and D.R. Roeder 69 Our lower Cd concentrations might be a result of sampling and samplingsite differences. We sampled from the tidal channels within Constitution Marsh, East Foundry Cove, and South Cove. These channels were subject to intense tidal flow and subsequent sediment deposition. In addition, we sampled for epibenthic diatoms and as a result, did not usually penetrate the superficial sediments regularly. Advanced GeoServices Engineering sampled cores ≈15 cm deep (P.F. Marano, Advanced GeoServices Engineering P.C., West Chester, PA, pers. comm.) and Mackie et al. (2007) routinely sampled 0–5 cm of sediment. The results of ANOSIM indicate differences in the diatom communities between the clusters A, B, and C. Clusters A and B were less dissimilar (R = 0.8) than clusters A and C (R = 0.970 and clusters B and C (R = 0.99). This can also be seen on the dendrogram (Fig. 3) calculated from Bray-Curtis similarity. Table 2. SIMPER of clusters A (stations 4–10), B (2, 11–15), C (1, 3), and the dominant 10 diatoms contributing to the similarity of each cluster. Cluster / Taxa % Cluster A similarity (84.9%) Navicula gregaria 19.0 Cocconeis placentula var. lineata 7.7 Cyclotella meneghiniana 7.2 Navicula odiosa 4.9 Navicula lanceolata 4.4 Nitzschia brevissima 3.4 Achnanthes haukiana var. rostrata 3.1 Navicula viridula 3.1 Nitzschia frustrulum var. perpusilla 2.9 Bacillaria paradoxa 2.7 Cluster B similarity (81.6%) Navicula gregaria 12.4 Cyclotella meneghiniana 11.3 Cocconeis placentula var. lineata 9.9 Navicula salinarum var. intermedia 4.8 Actinocyclus normanii 4.4 Navicula lanceolata 4.3 Nitzschia frustrulum var. perpusilla 4.0 Navicula viridula 3.9 Navicula peregrina 2.9 Cocconeis neodiminuta 2.7 Cluster C similarity (74.3%) Navicula gregaria 14.6 Achnanthes haukiana var. rostrata 8.6 Cocconeis placentula var. lineata 6.4 Cyclotella meneghiniana 5.3 Nitzschia frustrulum var. perpusilla 5.2 Navicula viridula 5.1 Navicula lanceolata 4.4 Navicula odiosa 4.2 Achnanthes haukiana 3.1 Navicula salinarum var. intermedia 3.0 70 Northeastern Naturalist Vol. 18, No. 1 The results of SIMPER look more closely into these differences. SIMPER calculates similarity and also dissimilarity. Dissimilarity is also a means of investigating the differences between the dendrogram clusters designated as A,B, and C. The dissimilarity between clusters A and B is 24.3%, with Navicula gregaria contributing 11.6% towards this dissimilarity. The dissimilarity between clusters A and C is greater, with the average dissimilarity of 28.6%. Two diatoms, Achnanthes haukiana var. rostrata and Navicula gregaria, contribute 11.8% and 11.3%, respectively. The greatest dissimilarity is between clusters B and C, with 30.4% dissimilarity. Achnanthes haukiana var. rostrata contributes 13% towards this dissimilarity. Other taxa contribute less than seven percent to the dissimilarity between the clusters. SIMPER was used again to investigate the similarity and dissimilarity between the selected stations. Stations 4 and 5 had a similarity of 91.4%, while stations 13, 14, and 15 were 83.5% similar. The difference in the similarity of these is the relative abundance of Navicula gregaria, which accounts for 17% of this 91.4% similarity. Navicula gregaria contributes less, only 13%, to the similarity between stations 13, 14, and 15. Interestingly, the dissimilarity between stations four and five and 13, 14, and 15 is only 23.5%, and the contribution of N. gregaria is 10.2%. The other diatom contributing to the dissimilarity of these groups is Navicula salinarum var. intermedia. Its abundance is slightly greater in the South Cove stations of 13, 14, and 15. Two other clusters were then compared. Stations 4–9 were compared to stations 1, 2, and 3. The similarity of stations 4–9 was 86.2%, with N. gregaria, the most abundant taxon, contributing 18% to the similarity. Other taxa contributed less than 8% to the similarity between these stations. Stations 1, 2, and 3 were only 73.6% similar, with N. gregaria contributing 10% and Achnanthes haukiana var. rostrata contributing 7% towards the similarity. When comparing these two clusters, N. gregaria contributed 14.6% towards the dissimilarity and A. haukiana var. rostrata contributed 9.1%. Other taxa contributed less than five percent towards the dissimilarity. Three species identified in the SIMPER analysis (Table 2) do contribute to this dissimilarity. Actinocyclus normanii, Navicula peregrina (Ehrenb.) Kuetz., and Cocconeis neodiminuta were found only in cluster B at greater than 2.5% relative abundance. Only in cluster A did Nitzschia brevissima Grun. in van Heurck occur at greater than 2% relative abundance, while Navicula minima Grun. in van Heurck was found only in cluster C at a relative abundance greater than 2%. These results indicates that the relatively rare species are contributing significantly to the similarity and dissimilarity between the three groups. In this study, ANOSIM demonstrates that there are differences between the stations in East Foundry Cove, Constitution Marsh, and South Cove. However, SIMPER discloses that there are no striking differences in community structure. The results indicate that stations 4 and 5 had larger relative abundances of Navicula gregaria, which contributed greatly to the 91.4% similarity between these stations. These are not in close proximity spatially to each other, as was earlier noted. They do both contain the highest concentrations of cadmium. Several authors have reported heavy-metal tolerant species of diatoms (e.g., Ivorra 2000). In laboratory experiments, she reported that “Navicula gregaria performed significantly better in the P (phosphorus) Cd + P and Cd treatments ….” 2011 B.H. Hallihan and D.R. Roeder 71 Most of these studies use mesocosms and microcosms, comparative analysis of diatom communities upstream and downstream from a point source of pollution, and translocation experiments. So, the outcome of this study is more difficult to interpret or compare to those, but raises the following questions for further research. Does this mean that this diatom is more tolerant or that the other species are more sensitive to higher levels of cadmium and nickel? Is this an example of pollution-induced community tolerance as indicated by Blanck et al. (1988)? Finally, we hypothesize that sediment or substrate differences might explain the differences observed in the diatom communities. For instance, why were the diatoms at station 2 more similar to the reference sites than in the adjacent East Foundry Cove stations? Sediment characterization was not part of this study. However, we observed casually throughout the study that station 2 sediments were consistently composed of fine silty-clay particles, as were those in stations 4–15. The sediments at stations 1 and 3 varied over the course of study. Kelaher et al. (2003) discussed in detail the sediments in East Foundry Cove, North Cove (North of Cold Spring), and South Cove, and concluded that sediment compaction greatly influences macrobenthic assemblages. We believe these differences should be examined in future research. Acknowledgments We would like to thank Thomas Jensen and Michael Baxter from Lehman College, Thomas Ruehli and Balori Paulori at Bronx Community College, Paul Marano of Advanced GeoServices P.C., West Chester, PA , and all the field assistants from Bard College at Simon’s Rock. Literature Cited Barber, H.G., and E.Y. Haworth. 1981. A guide to the morphology of the diatom frustule with a key to the British freshwater genera. Freshwater Biological Association Scientific Publication #44, Ambleside, Cumbria, UK. Blanck, H., Wangberg, S.-Å., and S. Molander. 1988. Pollution-induced community tolerance: A new ecotoxicological tool. In J. Cairns, Jr. and J.R. Pratt (Eds.). Functional testing of aquatic biota for estimating hazards of chemicals. ASTM STP 988. American Society for Testing of Materials, Philadelphia, PA. Pp. 219–230. Bray, J.R., and K.R. Curtis. 1957. An ordination of the upland forest communities of Southern Wisconsin. Ecological Monographs 27:325–349. Clarke, K.R., and R.M. Warwick. 2001. Change in marine communities: An Approach to Statistical Analysis and Interpretation. 2nd Edition. PRIMER-E, Plymouth, MA. Dodd, J.J. 1987. Diatoms (The Illustrated Flora of Illinois). Southern Illinois University Press, Carbondale and Edwardsville, IL. Hallihan, B.H. 2006. Impacts of heavy metals on diatom (Bacillariophyceae) communities in Constitution Marsh and Foundry Cove Cold Spring, New York, USA. Ph.D. Dissertation. The City University of New York, New York, NY. 330 pp. Hazen, R.E. 1981. Cadmium in an aquatic ecosystem. Ph.D. Dissertation. New York University, New York, NY. 135 pp. Hustedt, F. 1930. Bacillariophyta (Diatomeae). In A. Pascher (Ed.). Die Susswasser- Flora Mitteleuropas. Heft 10. Gustaz Fischer Verlag, Jena, Germany. 466 pp. 72 Northeastern Naturalist Vol. 18, No. 1 Hustedt, F. 1939. Systematische und okolgische Untersuchungen uber die Diatomen- Flora von Java, Bali and Sumatra nach dem Naterial der Deutschen Limnologischen Sundra-Expedition III. Die okolgischen Factorin und ihr Einfluss auf die Diatomeenflora, Archiv für Hydrobiologie, Supplement 16:274–394. Hustedt, F. 1953. Diatomeen aus dem Naturschutzpark Seeon. Archiv für Hydrobiologie 47:625–635. Hustedt, F. 1959. Die Diatomeenflora der Unterwesser von der Lesummündung bis Bremerhaven mit Berücksichtigung des Unterlaufs der Hunte and Geeste. Veröff. Inst. Meeresforsch. Bremerhaven 6:13–176. Ivorra, N. 2000. Metal induces succession in benthic diatom consortia. Ph.D. Dissertation. Universiteit van Amsterdam, Amsterdam, The Netherlands. 163 pp. Kelaher, B.P., J.S. Levinton, J. Oomen, B.J. Allen, and W.H. Wong. 2003. Changes in Benthos following the clean-up of a severely metal-polluted cove in the Hudson River estuary: Environmental restoration or ecological disturbance? Estuaries 26(6):1505–1516. Kneip, T.J., and R.E. Hazen. 1979. Deposit and mobility of cadmium in a marsh-cove ecosystem and the relation to cadmium concentration in biota. Environmental Health Perspectives 28:67–73. Krammer, K., and H. Lange-Bertalot. 1986. Süßwasserflora von Mitteleuropa. Band 2. Bacillariophyceae. Teil 1. Naviculaceae. Gustav Fischer Verlag, Stuttgart, Germany. Krammer, K., and H. Lange-Bertalot. 1988. Süßwasserflora von Mitteleuropa. Band 2. Bacillariophyceae. Teil 2. Bacillariaceae, Epithemiaceae, Surirellaceae. Gustav Fischer Verlag, Stuttgart, Germany. Krammer, K., and H. Lange-Bertalot. 1991a. Süßwasserflora von Mitteleuropa. Band 2. Bacillariophyceae. Teil 3. Centrales, Fragilariaceae, Eunotiaceae. Gustav Fischer Verlag, Stuttgart, Germany. Krammer, K., and H. Lange-Bertalot. 1991b. Süßwasserflora von Mitteleuropa. Band 2. Bacillariophyceae. Teil 4. Achnanthaceae, Kritische Ergänzungen zu Navicula (Lineolatae) and Gomphonema. Gustav Fischer Verlag, Stuttgart, Germany. Limburg, K.E., M.A. Moran, and W.H. McDowell. 1986. The Hudson River Ecosystem. Springer-Verlag, New York, NY. 331 pp. Mackie, J.E., S.M. Natali, J.S. Levinton, and S.A. Sanudo-Wilhelmy. 2007. Declining metal levels in Foundry Cove (Hudson River, New York): Response to localized dredging of contaminated sediments. Environmental Pollution 149:141–148. Marano, P.F. 2010. Project number 95-219-04 sampling event report, year fourteen— October 2009, Long-term Monitoring Program, Marathon Remediation Site January 12, 2010. Advanced Geoservices Engineering P.C., West Chester, PA. New York State Department of Environmental Conservation. 1985. Marathon Battery federal superfund site, Cold Spring, NY remedial investigation\feasibility study. Draft Remedial Investigation Report. Albany, NY Patrick, R., and C.W. Reimer. 1967. The Diatoms of the United States, Exclusive of Alaska and Hawaii. Volume 1. Monograph of the Academy of Natural Sciences of Philadelphia, Number 13. 668 pp. Patrick, R., and C.W. Reimer. 1975. The Diatoms of the United States, Exclusive of Alaska and Hawaii. Volume 2, Part 1. Monograph of the Academy of Natural Sciences of Philadelphia, Number 13. 213 pp. Ravera, O. 1979. Consideration on the effects of pollution at community and population level. Experientia 35:710–713. Resource Engineering. 1983. Preliminary site background data analysis of Foundry Cove. Report prepared for Vinson and Elkins, Houston, TX. Stoermer, E.F., R.G. Kreis, Jr., and N.A. Andresen. 1999. Checklist of diatoms from the Laurentian Great Lakes. II. Journal of Great Lakes Research 25:515–566. Vinyard, W. 1979. Diatoms of North America. Mad River Press, Eureka, CA.