490

J. Phycol.

39,

490–503 (2003)

EFFECTS OF METAL AND PETROLEUM HYDROCARBON CONTAMINATION ON BENTHIC DIATOM COMMUNITIES NEAR CASEY STATION, ANTARCTICA:

AN EXPERIMENTAL APPROACH

1

Laura Cunningham

Institute of Antarctic and Southern Ocean Studies, University of Tasmania, Box 252-77 Hobart, Australia, 7001

Jonathan S. Stark, Ian Snape

Human Impacts Program, Australian Antarctic Division, Channel Highway, Kingston, Tasmania 7050, Australia

Andrew McMinn

2

Institute of Antarctic and Southern Ocean Studies, University of Tasmania, Box 252-77 Hobart, Australia, 7001

and

Martin J. Riddle

Human Impacts Program, Australian Antarctic Division, Channel Highway, Kingston, Tasmania 7050, Australia

The effects of metals and petroleum hydrocar-bons on benthic marine diatom communities wereassessed using a manipulative field experiment atCasey Station, Antarctica. Uncontaminated, metalcontaminated, and petroleum hydrocarbon contami-nated sediments were deployed for 11 weeks duringthe 1999 austral summer. The treatments were de-ployed at three different locations: Brown Bay,which has elevated levels of anthropogenic contami-nants, and two uncontaminated reference locations,O’Brien Bay and Sparkes Bay, the latter of which hasnaturally occurring high concentrations of someheavy metals. At each location, significant differ-ences between the composition of diatom communi-ties recruiting to control and petroleum hydrocarboncontaminated treatments were observed.

Navicula di-recta

(Smith) Ralfs occurred at lower abundances inthe petroleum hydrocarbon contaminated treatmentsthan in the control treatments. In contrast,

Naviculacancellata

Donkin occurred at higher relative abun-dances in both contaminated treatments relative tothe control treatment. Interactions between treat-ment and location were also observed for severalspecies, including

Navicula glaciei

Van Heurck. Sig-nificant differences in the overall community com-position of diatom communities between controland metal contaminated treatments and metal con-taminated and petroleum hydrocarbon contami-nated treatments were only observed within BrownBay. The location of deployment also had a signifi-cant influence on the composition of diatom com-munities. Brown Bay had higher abundances of

Ach-nanthes brevipes

Agardh but lower abundances of

Navicula

aff.

cincta

(Ehrenberg) Van Heurck than ei-ther reference locations. This experiment demon-

strated that benthic diatom communities are sensi-tive to sediment contamination and would be suitablefor future monitoring work within this and other ar-eas of Antarctica.

Key index words:

benthic diatoms; community com-position; metal contamination; petroleum hydrocar-bons; Casey Station; Antarctica

Abbreviation:

SAB, special Antarctic blend diesel fuel

Antarctica is widely regarded as the last remainingwilderness; however, the integrity of this environmentis threatened by human activities. Increases in chemi-cal contaminants have been recorded in the marineenvironments around many of the permanent re-search stations (Lenihan et al. 1990). Detailed infor-mation on the extent of such contamination is avail-able for only a few sites, including McMurdo Station(Kennicutt et al. 1995, Crockett 1997) and Casey Sta-tion (Snape et al. 2001; Stark et al. in press). Informa-tion on the biological effects of such contamination isalso limited to a few studies, most of which have exam-ined the effects on faunal communities (Lenihan andOliver 1995, Stark 2000).

The benthic microflora of Antarctic regions con-sists almost entirely of diatoms (Gilbert 1991). Dia-toms have been shown to have narrow toleranceranges for many environmental variables and respondrapidly to environmental change, making them idealbioindicators (Reid et al. 1995). Contaminants can af-fect the growth rate and species composition of dia-tom communities, with subsequent implications fororganisms higher in the food chain (Stronkhurst et al.1994).

Recent studies (Snape et al. 2001, Scouller et al.2000, Stark et al. in press) have demonstrated higherconcentrations of some metals and petroleum hydro-carbons in sediments collected from bays adjacent to

1

Received 18 December 2001. Accepted 4 March 2003.

2

Author for correspondence: e-mail andrew.mcminn@utas.edu.au.

491

CONTAMINANTS AND DIATOM COMMUNITIES

Casey Station compared with locations further away.This contamination has been linked to both the oper-ation of a waste disposal site and hydrocarbon spills(Deprez et al. 1999). Brown Bay, adjacent to the wastedisposal site, is the worst affected, with some metals(including copper, lead, iron, and zinc) occurring inthe sediments at concentrations 10 to 100 timeshigher than background levels. Petroleum hydrocar-bons, derived from lubrication oil and special Antarc-tic blend diesel fuel (SAB), are present in the surfacesediments of Brown Bay at concentrations ranging be-tween 40 and 200 mg

�

kg

�

1

. In contrast, petroleum hy-drocarbons were not detected in sediments from ref-erence locations (Snape et al. 2001). Seawater collectedfrom Brown Bay also contained elevated concentra-tions of copper, cadmium, lead, and zinc (Deprez etal. 1999, Snape et al. 2001).

The species composition of diatom communitiesaround Casey Station differs significantly betweenBrown Bay and uncontaminated sites further awayfrom the station (Cunningham et al. 2000). Environ-mental factors, including grain size and total organiccontent of the sediments, also vary between these lo-cations (Stark et al. in press). No published informa-tion is available on the nature of predisturbance dia-tom communities that were present in Brown Baybefore contamination occurred. Subsequently, theobserved differences in diatom communities betweenBrown Bay and reference locations cannot be un-equivocally linked to contamination by metals and pe-troleum hydrocarbon.

Manipulative field experiments are the most appro-priate means of demonstrating causal relationshipsbetween environmental factors, such as contamina-tion, and observed biological patterns (Underwoodand Peterson 1988). Manipulative experiments usingmicrobenthic algae are very scarce (Ivorra et al. 1999)and generally consist of the translocation of diatomcommunities between sites. One problem associatedwith this approach is that the environmental factorsexamined are typically associated with the sites andmay not necessarily be independent of other pro-cesses acting at these sites. Artificially contaminatingsediments enable the influence of the contaminantsto be isolated from other sources of variability be-tween sites. Within this study, artificially contami-nated sediments were used to examine the effects ofmetal contamination and petroleum hydrocarboncontamination on benthic diatom communities.

The goal of this study was to assess whether benthicdiatom communities can be used as an indicator ofenvironmental impacts within the near-shore coastalregions of Antarctica. The objectives were as follows:

1. To determine whether a causal relationship be-tween differences in diatom community composi-tion and contamination by petroleum hydrocar-bons and/or metals could be established. It washypothesized that the diatom communities that de-veloped on contaminated sediments would be dif-

ferent from those in control uncontaminated sedi-ment.

2. To study the influence of location on the develop-ment of diatom communities and, in particular,the impact of contaminants in Brown Bay on thesecommunities. It was hypothesized that there wouldbe differences in diatom communities that devel-oped at the three locations and, furthermore, thatthese differences would be more pronounced be-tween communities that developed at Brown Baycompared with the reference locations.

3. To examine whether the response of the diatomcommunities to experimental contamination wouldvary between the locations, specifically whether anincreased response would occur in diatom commu-nities from Brown Bay. It was hypothesized thatthere would be significant interactions between lo-cation and contamination.

A secondary aim was to assess the sensitivity of dia-tom communities to relatively low levels of chemicalcontamination and thus determine their suitability forfuture monitoring work within Antarctica.

materials and methods

Location description.

Casey Station is located at 66

�

17

�

S,110

�

32

�

E in the Windmill Islands (Fig. 1). Casey Station is thethird station to have been built in the region. The first station,Wilkes, was built in 1957 on the edge of the Clarke Peninsulabut was superseded in 1969 by what is now known as “Old Ca-sey.” Old Casey is located on Bailey Peninsula, 2 km south ofWilkes and 1 km east of the present Casey Station.

Refuse from the first two stations accumulated at local wastedisposal locations before change of policy in 1986 with wastesubsequently returned to Australia. The Thala Valley disposalsite received waste material from Old Casey Station between1969 and 1986. This material included domestic waste andwaste, such as batteries and fuel drums, from mechanical work-shops (Deprez et al. 1999). Despite an earlier clean-up effort,approximately 2500 m

3

of the waste generated by Old Casey sta-tion still remains in Thala Valley (Deprez et al. 1999, Snape etal. 2001).

In summer, a melt-stream flows through the Thala Valleywaste disposal site where water dissolves and entrains contami-nants before runoff enters into the adjacent Brown Bay. An esti-mated 8 m

3

of contaminated material associated with the tipwas removed by surface runoff and deposited into Brown Bayduring the 1998–1999 summer period (Cole et al. 2000). In ad-dition, hydrocarbon contaminants from soils surrounding theOld Casey mechanical workshops are carried into Brown Bay bythe movement of surface and subsurface waters (Snape et al.2001). Hydrocarbon contamination of Brown Bay may also re-sult from a number of small spills that have occurred during re-fueling operations.

Brown Bay is a small embayment at the southern end ofNewcomb Bay. Brown Bay has a maximum depth of approxi-mately 20 m with rocky sides grading to a muddy bottom (Stark2000). Close to the tip site, patches of sediment occur betweenboulders, discarded tip material, and areas of bare rock. Fur-ther away from the tip, areas of fine sediments are more exten-sive and relatively homogenous (Stark 2000) but are still occa-sionally interrupted by large boulders. Brown Bay is typically icefree for 1–2 months a year, generally between January and Feb-ruary. Sea ice in Brown Bay is rarely blown out during the win-ter period.

O’Brien Bay and Sparkes Bay are large bays situated severalkilometers south of Casey Station. These bays have a variety ofenvironments: ice cliffs dropping vertically into deep water,

492

LAURA CUNNINGHAM ET AL.

steep rocky shorelines, and shallower sediment-dominated ar-eas. These bays were chosen as reference locations because theyare unlikely to be contaminated by human activity and becausethey are relatively accessible. Naturally elevated levels of somemetals have been recorded within Sparkes Bay (Stark et al. inpress). Within these large bays, sampling locations were se-lected due to shared physical characteristics, such as sea-icecover, with Brown Bay. A small embayment on the northernside of Sparkes Bay was used as a reference location because ithas a very similar bathometry to that of Brown Bay, with therocky sides sloping down to a muddy bottom and a depth rangebetween 3 and 20 m (Stark 2001). The reference location inO’Brien Bay was also situated on the northern side of the bay.The sides consisted of steeply sloping bolder fields and moregently inclined rocky banks. Patches of poorly mixed mud andsand interspersed between gravel, cobbles, and bouldersformed the bottom sediments of this area (Stark 2000).

Little data are available on hydrodynamic conditions withinthe region, although Cathers et al. (1998) found currents to be

wind dominated, with little tidal influence. Resuspension ofbottom sediments is common within near-shore regions of Ant-arctica. Everitt and Thomas (1986) found that benthic diatomsconsistently formed a high proportion of species recorded inplanktonic communities at inshore localities near Davis Station.Within this study, it has been assumed that sufficient resuspen-sion of the sediment occurs to allow the colonization of sedi-ments 12 cm above the natural substrate.

Experimental method.

Sediment was collected from O’BrienBay by van Veen grab and divers using polyethylene buckets.The sediment consisted of very poorly sorted sands and finesands of which 20%–40% was less than 63

�

m (Stark et al. inpress). This sediment was uncontaminated (Snape et al. 2001,whereas the total organic carbon content varied between 1.5%and 3% (Stark et al. in press). The sediment was frozen at

�

20

�

Cfor 48 h, defrosted, and then sieved through a 500-

�

m mesh toremove fauna, with material coarser than 500

�

m discarded(on average,

�

11% of sediment). The remaining sediment wasthoroughly mixed to ensure homogeneity and then divided

Fig. 1. Map of the Windmill Is-lands showing Casey Station (rectan-gle), the Thala Valley waste disposalsite (triangle), and the experimentallocations used (asterisks).

493

CONTAMINANTS AND DIATOM COMMUNITIES

into three portions, each approximately 75 kg. One portion,the control treatment, was not further manipulated. One por-tion was contaminated by adding contaminated fine-grained tipmaterial, whereas the third portion was contaminated by add-ing petroleum hydrocarbons.

An analogue for metal contamination of marine sediments,such as that found in Brown Bay, was created by mixing materialfrom the Thala Valley waste disposal site with clean marine sedi-ment from O’Brien Bay. Approximately 50 kg of highly contami-nated waste disposal soil was sieved through a 500-

�

m mesh, thecoarse material discarded, and the fine material homogenizedand then added in a ratio of 1:100 to the O’Brien Bay sediment.

The hydrocarbon contaminated treatment was created byadding a mixture of 25 mL of lubricating oil, 50 mL of SAB,and 15 L of clean seawater to the sediment, forming a wetslurry. This was thoroughly mixed and left to equilibrate for 12h. The excess fluid was then siphoned off to remove any excesshydrocarbons that would have dispersed into the marine envi-ronment immediately on deployment. This treatment was in-tended to approximate petroleum hydrocarbon concentrationscurrently observed within Brown Bay.

Cylindrical plastic flowerpots (12 cm diameter

�

12 cmheight) were used as receptacles for the sediments. Three 8

�

8-cm holes were cut into the sides, and a hole of 9 cm diameterwas cut into the base of each flowerpot, all of which were thencovered with a 300-

�

m mesh (Fig. 2). These mesh inserts al-lowed exchange of water through the container to better simu-late natural conditions. Six pots were placed in a tray (60

�

35

�

cm) in a 2

�

3 arrangement; thus, the pots were separated fromeach other by approximately 10 cm (Fig. 2). The pots werethen secured to the tray using plastic cable ties. The pots werefilled with sediment on site immediately before deployment. Ateach of the three locations, three trays of each of the threetreatments were deployed randomly within a roughly circulararea (radius approximately 5 m). In all three locations the ex-perimental trays were placed onto muddy substrates. WithinBrown and Sparkes Bay the experimental trays were deployedat 12-m water depth, whereas in O’Brien Bay the experimenttrays were deployed at 12- to 15-m water depths.

Before deployment sediment samples were taken from eachtreatment for all three locations. These samples were frozen at

�

20

�

C for later analysis of metal and hydrocarbon concentra-tions. In addition, the composition of the diatom communitiesalready present in the experimental sediments was determinedfor two samples from each location and for each treatment typewithin Sparkes Bay.

After a deployment period of 11 weeks the trays were re-trieved by scuba divers. Four pots from each tray were used forstudies on the recruitment of soft-sediment infauna (Stark et al.2003). Two scrapes (3 cm long by 0.5 cm deep) of sedimentwere taken from the surface of the pots; the position of thesescrapes was not predetermined and varied between pots. Thesesediment scrapes were preserved in glutaraldehyde and laterused for diatom analyses. The remaining sediments from thesetwo pots were frozen at

�

20

�

C for later analysis of metals andhydrocarbons in the sediments. A full description of the meth-ods used for the geochemical analyses is given in Stark et al.(2003).

Diatom preparation and identification.

Excess organic materialwas removed by digestion in a 10% hydrogen peroxide solutionfor 72 h. Excess liquid was decanted off, with the remainingslurry transferred to a centrifuge tube. Distilled water wasadded so that the volume of each tube was 10 mL. The sampleswere then centrifuged for 5 min at 3000 rpm. The supernatantwas decanted off, and the pellet was resuspended in 10 mL ofdistilled water. The centrifuging process was repeated twicemore. After the third treatment, the pellet was once again re-suspended in distilled water. This solution was diluted to ap-proximately 10% and pipetted onto coverslips. After air drying,the coverslips were mounted onto slides using Norland opticaladhesive 61 (Norland Products Inc., Cranberry, NJ, USA).

Examination of diatom taxa was undertaken using a ZeissKF2 light microscope (Jena, Zeiss, Germany) with 1000

�

mag-nification and phase contrast illumination. Identification wasprimarily based on Hasle and Syverston (1996), Roberts andMcMinn (1999), and Medlin and Priddle (1990). In an attemptto ensure that the diatoms included within the investigationwere exposed to the treatment types, species generally presentonly within planktonic or sea-ice communities were excludedfrom the data set. Subsequent to this, a minimum of 400 valveswas counted for each sample. The exact number of frustulescounted varied slightly between samples, with a maximum of520 benthic diatoms counted in any one sample. The relativeabundances of the benthic taxa were then calculated and usedin the statistical analyses.

Statistical analyses.

Multivariate analyses of community com-position were undertaken using nonmetric multidimensionalscaling and analysis of similarity (ANOSIM) procedures usingthe PRIMER software package (Plymouth Marine Laboratory,Plymouth, UK). A two-way crossed ANOSIM (with replication)was used to determine the overall influences of both the treat-ments and the locations. Individual one-way ANOSIMs were

Fig. 2. Diagrammatic details of the modified flower pots that were used as containers for the experimental sediments.

494

LAURA CUNNINGHAM ET AL.

used to examine the effects of treatment within each locationand the effects of location within each treatment type. This al-lows any interactions between the effects of treatment and loca-tion to be determined. No data transformation was used forthese analyses. Similarity matrixes were calculated using theBray-Curtis similarity measure.

Similarity percentages analyses (SIMPER) were used to de-termine the relative contribution of individual species to com-positional differences observed between diatom communities.Clarke and Warwick (1994) stated that species with a SIMPERratio greater than 1.3 are likely to be useful for discriminatingbetween groups. The 10 species with the highest average ratioswere selected as potential indicator species, which were thenused in the univariate analyses.

Natural logarithms were used to calculate Shannon-Wienerdiversity values, Simpson’s dominance index, Margalef’s speciesrichness values, and Pielou’s evenness values. Collectively thesevalues are defined as the structural parameters of the diatomcommunities.

Two-way analysis of variance (ANOVA) was used to determineif the potential indicator species or structural parameters variedsignificantly between groups. Cochran’s C test was used to checkthe assumption of homogeneity of variances before usingANOVA. Any data transformations that were necessary to fulfillthis assumption are noted in the results. Where ANOVAs indi-cated significant differences between samples, Student-NewmanKeuls (SNK) tests were used for multiple comparisons (

a

�

0.05). ANOVAs and SNK tests were performed using GMAV5(Institute of Marine Ecology, University of Sydney, Australia).

Comparisons between diatoms originally in the experimen-tal sediments and the developing communities were also un-dertaken using the methods described above.

results

Chemical contamination.

Contaminant concentrationswere higher in the metal and hydrocarbon treatmentsthan in the controls. The metal treatment had concen-trations 2 to 60 times that of the preexposure controlsediments, with the greatest increases occurring incopper, lead, zinc, and tin concentrations (Fig. 3).Comparisons of pre- and postexposure concentrationsindicate that the concentrations of several metals in-creased in contaminated experimental sedimentswithin Brown Bay during the 11-week deployment pe-riod, possibly reflecting further input from the ThalaValley waste disposal site during this time. Such an in-put of metals would be associated with sediment inputfrom the melt stream and the subsequent settling outof these particles onto the experimental sediments.Concentrations of manganese, tin, mercury, iron, andlead were also observed to increase in experimentalsediments, supporting the idea of further metal inputinto Brown Bay during the course of the experiment.

Despite the use of waste disposal fines as a contami-nant, concentrations measured in the metal treat-ment varied slightly from those previously measuredin Brown Bay sediments (Snape et al. 2001) (Table 1).Concentrations of most metals, including zinc, silver,arsenic, nickel, and chromium, were typically lower inthe experimental treatment than in Brown Bay sedi-ments. Lead was an exception, with higher mean con-centrations occurring in the experimental treatmentthan in Brown Bay sediments.

The petroleum hydrocarbon treatments had pre-deployment concentration ranges of 68.6 to 255.6

mg

�

kg

�

1

and 178.7 to 204.4 mg

�

kg

�

1

for SAB and lu-brication oil, respectively (Table 2). The lowest mea-sured value of SAB was in one of the Sparkes Baysamples, but this was believed to reflect sample heter-ogeneity and variability of the chemical extractionprocess rather than a consistent difference in contam-inant levels for the pots in this bay. Hydrocarbon con-centrations used in this experiment were similar tothose already present in Brown Bay, where fuel andlubrication oil are present in concentrations between41 and 200 mg

�

kg

�

1

(Snape et al. 2001). The concen-tration of petroleum hydrocarbons typically decreasedin the contaminated sediments during the course ofthe experiment. One exception to this was the con-centration of lubricating oil at O’Brien Bay; however,this was believed to reflect sample heterogeneity. Pe-troleum hydrocarbon concentrations were below de-tection limits (

�

0.001 mg

�

kg

�

1

) in the control treat-ments.

Diatom communities: a general description.

Forty-eightdifferent taxa were found in the diatom communities(Appendix). Only two of these species occurred at rel-ative abundances greater than 10%.

Achnanthes brevi-pes

occurred at these abundances in all samples, ex-cept in contaminated treatments within Brown Bay(5%–8%).

Stauroneis

cf.

wislouchii

occurred at theseabundances only within Brown Bay and did not ex-ceed 5% at either reference location. Relative abun-dances between 5% and 10% were recorded for

Navic-ula glaciei

and the

Achnanthes delicatula

complex at alllocations.

The composition of the diatom communities thatformed on the experimental sediments was quite simi-lar between locations and treatments. In addition tothe species described above, each community had ap-proximately 10 species that occurred at relative abun-dances between 1% and 5%, typically including sev-eral species of

Navicula

and

Cocconeis.

Pseudostaurosira

cf.

wislouchii

,

Synedropsis

cf.

hyperboreoides

, a

Nitzschia

species, and two varieties of

Fragilaria construens

alsogenerally occurred at these abundances. The diatomcommunities that developed on the experimental sed-iments generally contained a further 25 species whoserelative abundances was below 1%.

Influence of sediment contamination on diatom commu-nities.

A two-way crossed ANOSIM showed that thetreatment type significantly influenced the composi-tion of the diatom communities, with a global R valueof 0.375 (

P

�

0.0001). Although significant differ-ences occurred between all three treatments, the con-trol and petroleum hydrocarbon contaminated treat-ments showed the highest level of differentiation(Table 3).

Responses to sediment contamination by the dia-tom communities varied between locations. WithinBrown Bay each of the three treatments resulted insignificantly different community compositions (Ta-ble 3). In O’Brien Bay and Sparkes Bay, however, sig-nificant differences in community composition wereonly detected between the petroleum hydrocarbon

495

CONTAMINANTS AND DIATOM COMMUNITIES

contaminated treatments and control treatments (Ta-ble 3, Fig. 4).

The differences observed as a result of contamina-tion were predominantly due to changes in speciesabundances rather than to the presence or absence ofindividual species. Only a few species responded tothe sediment treatments in the same manner at all lo-cations. A two-way ANOVA and subsequent SNK tests(Table 4) indicated that

Navicula directa

decreased inrelative abundance in the diatom communities fromhydrocarbon treatments, relative to communities fromthe control treatments regardless of the location ex-amined.

Navicula

cf.

cancellata

also responded to con-tamination in a consistent fashion at all locations, in-

creasing in abundance in response to both metal andpetroleum hydrocarbon contamination relative tocontrol treatments. The relative abundances of

Stauro-neis

cf.

wislouchii

,

Navicula glaciei

,

Navicula tripunctata

var.

schizonemoides

,

Fragilaria construens

var.

venter

, andthe

Achnanthes delicatula

complex also varied signifi-cantly between treatments types; however, significantinteractions between location and treatment typewere found for these species (Table 4).

Despite differences in the relative abundances ofthe species present, no statistically significant differ-ences in the species diversity, richness, dominance, orevenness values were detected between the treatmentsoverall (Table 5).

Fig. 3. Metal concentrations (inmg�kg�1) in experimental sediments.Open bars indicate the control treat-ment; solid bars represent the metalcontaminated treatment. Preexposureconcentrations are shown in gray; post-exposure concentrations are shown inblack.

496

LAURA CUNNINGHAM ET AL.

Influence of location on the diatom communities.

The lo-cation of deployment had the greatest influence onthe composition of the diatom communities, with aglobal R value of 0.734 (

P

�

0.0001. All three loca-tions had significantly different community composi-tions (Fig. 5), although structural parameters werenot statistically different.

O’Brien Bay and Sparkes Bay had the most similarcompositions regardless of the treatments compared(Table 6). The greatest dissimilarity in diatom com-munities from metal contaminated sediments oc-curred between O’Brien Bay and Brown Bay. In con-trast, the greatest dissimilarity between diatomcommunities from both the control and petroleumhydrocarbon contaminated treatments occurred be-tween Sparkes Bay and Brown Bay.

The differences in composition observed betweenlocations were produced by differences in the relativeabundances of a number of species. Consistent differ-ences in the relative abundances of several species oc-curred between locations, regardless of treatmenttype. For example,

Navicula

aff.

cincta

occurred athigher relative abundances within Brown Bay than ineither of the reference locations (Table 5). In con-trast,

Achnanthes brevipes

and

Synedropsis

cf.

recta

oc-curred at significantly higher relative abundanceswithin both the reference locations than in BrownBay.

Cocconeis fasciolata

and

Navicula perminuta

bothhad significantly higher abundances in O’Brien Baythan in Brown Bay (Table 5). Significantly higher rel-ative abundances of

Navicula directa

occurred inO’Brien Bay than at either Sparkes Bay or Brown Bay.

The relative abundance of several other species var-ied between locations, but these differences were not

consistent across treatment types. Significant interac-tions between location and treatment were observedfor

Navicula tripunctata

var.

schizonemoides

,

Navicula gla-ciei

,

Stauroneis wislouchii

, and the

Achnanthes delicatula

complex (Table 5). For example,

N. tripunctata

var.schizonemoides was observed to have higher relativeabundances in samples from Sparkes Bay for the hy-drocarbon treatment (Table 5). Similarly, O’BrienBay had significantly higher relative abundances of N.glaciei than Sparkes Bay or Brown Bay, within themetal and hydrocarbon treatments (Table 5).

Comparison between developing and remnant diatomcommunities. Remnant diatom communities refer tothose diatoms originally present in sediment collectedfrom O’Brien Bay, which represent the naturally oc-curring communities from this location. These com-munities were significantly different from the commu-nities that developed on the deployed sedimentsduring the experiment (Fig. 6). Structural parameterswere significantly different between the remnant com-munities and those that developed on the experimen-tal sediments. The remnant communities had higherdominance values, with lower diversity and evennessvalues than the developing communities (Table 7).

A number of species were present in the experi-mental sediments that were not recorded in the rem-nant communities: Anomoeoneis cf. follis var. hannae,Amphora sp. a, Ctenophora pulchella, Hantschia sp. a,Navicula muticopsis, Navicula perminuta, and two uni-dentified pennates. In contrast, Pleurosigma sp. a waspresent in the remnant communities but was not ob-served in the communities that developed on the ex-perimental sediments.

SIMPER analyses also revealed that the developingcommunities had lower relative abundances of Cocco-neis costata, C. fasciolata, Fragilaria construens var. venter,and Pseudostaurosira cf. brevistriata. The developingcommunities had higher relative abundances of Nav-icula tripunctata var. schizonemoides, Stauroneis cf. wis-

Table 1. Metal concentrations in sediment samples fromBrown Bay (from Snape et al. 2001 and those of the metalcontaminated treatment mean (mg�kg�1 SE, 3 replicates).

Brown Bay Metal treatment

Mean SE Mean SE

Copper 11.67 8.68 12.29 1.11Zinc 52.77 15.82 39.75 3.21Lead 28.99 29.00 60.15 18.16Silver 0.32 0.32 Below detectionCadmium 0.88 0.53 0.47 0.01Nickel 3.60 1.54 0.65 0.041Chromium 5.42 1.98 0.45 0.043Antimony 0.60 0.33 0.42 0.13Arsenic 17.25 3.13 7.49 0.44

Table 2. Concentrations (mg�kg�1) of petroleum hydrocarbonsin experimental sediments, pre- and postdeployment.

SAB Lubricating oil

Pre Post Pre PostO’Brien Bay 242 92 184 194Brown Bay 255 26 204 122Sparkes Bay 68 56 178 154

Numbers given are the average of three replicate samples.

Table 3. ANOSIM results for comparison of control, metalcontaminated, and petroleum hydrocarbon contaminatedtreatments, both overall and within each of the three locations.

Treatments compared R value Significance level

OverallControl Metal 0.254 0.3%Control Hydrocarbon 0.635 0.0%Hydrocarbon Metal 0.242 0.2%

Sparkes BayControl Metal �0.074 71.6%Control Hydrocarbon 0.569 0.2%Metal Hydrocarbon 0.113 18.6%

Brown BayControl Metal 0.633 0.2%Control Hydrocarbon 0.900 0.2%Metal Hydrocarbon 0.395 0.4%

O’Brien BayControl Metal 0.183 11.9%Control Hydrocarbon 0.353 1.0%Metal Hydrocarbon 0.217 5.2%

Significant results shown in bold.

497CONTAMINANTS AND DIATOM COMMUNITIES

louchii, Navicula glaciei, and Achnanthes brevipes thanthe remnant communities.

discussionContamination of marine sediments, at levels simi-

lar to those produced by station activities, can resultin changes to the composition of diatom communitiesliving in and on these sediments. Both the metal andpetroleum hydrocarbon contamination significantlyaffected the diatom communities; however, the re-sponse was complex and the location of deploymentinfluenced the observed effects. The diatom samplescollected from the experimental sediments wouldhave initially contained some of the remnant diatomcommunity as well as the developing communities.That the developing communities were so differentfrom the remnant community indicates that the dia-toms that developed on the sediments were in suffi-cient number to overcome much of the influence thatthe diatoms remaining from the original sedimentswould have had on the relative abundances. Thus, itcan be concluded that the observed differences incommunity composition are a response to the experi-ment itself and not simply variations present withinthe sediments at the time of deployment. The absenceof a live diatom community at the time of deployment

would explain the increased diversity and richness val-ues observed in the developing communities. It is pos-sible that a longer deployment period may have en-abled a more natural diatom community to develop.

The design of the experimental units used in thisexperiment may account for some of the variation ob-served between the remnant diatom communities andthose that developed on the experimental sediments.The experimental sediments were raised 12 cm abovethe natural substrate. This may have caused a bias to-ward species that are more easily resuspended. Addi-tionally, some of the diatom species included in theanalysis occur in both benthic and under-ice commu-nities. These species could colonize the experimentalsediments either via resuspended sediment or as“rain” from the sea ice, which may have resulted inhigher abundances of these species, relative to speciesthat only occur in sediments. The method of deploy-ing manipulated sediments used in this experimentwas successful and would be applicable for future in-vestigative and monitoring work within Antarctica.

This study is one of the first to examine the effectsof either petroleum hydrocarbon or metal contamina-tion of marine sediments on the microphytobenthosthat inhabit the sediments. The majority of work ex-amining the sensitivity of marine diatoms to pollution

Fig. 4. Two-dimensional nonmet-ric multidimensional scaling ordina-tion of diatom communities showingseparation into location groups withineach treatment type. Reference loca-tions are depicted by open symbols,whereas the contaminated location hasa solid symbol.

498 LAURA CUNNINGHAM ET AL.

has been based on phytoplankton (Hsiao 1978). Al-though the use of periphyton for pollution monitor-ing has increased in recent years, this has typicallybeen related to water quality, not sediment contami-nation (Ivorra et al. 1999).

Contamination by petroleum hydrocarbons had alarger effect on diatom community composition thanmetal contamination. This could be related to the solu-bility and diffusion of the two contaminant types thatare likely to be different in interstitial pore water. Metalbioavailability is controlled by a large number of com-plex interacting factors, most importantly for the resultspresented here, the free ion concentrations (Eriksen etal. 2001). Many metals are known to have appreciablyhigher solubilities under the reducing conditions thatwere found at a depth of approximately 1 cm in the ex-perimental sediments. It is therefore likely that a steepgradient in the free ion concentrations would havebeen present below the sediment–water interface. Be-cause diatoms are essentially restricted to the surfacelayer of the sediments, they would potentially have beenexposed to only slightly elevated metal concentrationsrelative to the control treatments. In contrast, the petro-

leum hydrocarbons would have been less affected by re-ducing conditions and thus would have been moreevenly distributed throughout the sediment profile.

The response of diatom communities to petroleumhydrocarbon contamination within this experimentdoes not appear to be directly related to concentra-tion. Sparkes Bay had the lowest initial hydrocarbonconcentrations; final concentrations of hydrocarbonswere lowest in Brown Bay. Despite this, differences be-tween control and hydrocarbon treatments were mostpronounced in Brown Bay, with Sparkes Bay showingthe second strongest response. It therefore seems un-likely that the differences in the hydrocarbon concen-tration are responsible for the varying responses ob-served between locations.

The concentrations of hydrocarbons used in thisexperiment were selected to represent petroleum hy-drocarbon concentrations present within Brown Bayyet are still well below concentrations that have beenreported elsewhere in Antarctica (Lenihan et al. 1990).Concentrations of hydrocarbons as low as 0.04 �g mL�1

have previously been demonstrated to inhibit diatomgrowth (Siron et al. 1991); however, this is subjectiveand related to the sensitivity of an individual species(Ostgaard et al. 1984). Toxic effects of petroleum hy-drocarbon contamination on diatom communities in-crease with exposure time (Hsiao 1978). Compositionaldifferences, with marked changes in the presence orabsence of species, have previously been observed be-tween control communities and those exposed to ei-ther light crude oil or diesel-based oil cuttings (Plante-Cuny et al. 1993). In contrast, the presence or absenceof particular species contributed little to differencesin community composition in our experiment, proba-

Table 4. ANOVA and SNK results for comparisons of species between locations, between treatments, and interactions between variables.

Location Treatment

Achnanthes brevipes Overall O � S � BAchnanthes delicatula complex Overall S � B � O Sparkes C � M � H

Control S � B � OCocconeis fasciolata Overall O � BFragilaria construens var. venter Overall M � C � H

Metal B � O Brown M � C � HNavicula aff. cincta Overall B � S � ONavicula cancellata Overall B � S � O Overall H � M � CNavicula directa Overall O � B � S Overall C � HNavicula glaciei Overall O � B � S Overall C � M � H

Metal O � B � S Brown C � M � HHydrocarbon O � B � S O’Brien M � C � H

Navicula perminuta Overall O � BNavicula tripunctata var. schizonemoides Overall S � B � O Sparkes C � H

Control B � OBHydrocarbon S � B � O

Pseudostaurosira cf. brevistriata Control S � B Sparkes C � HMetal B � S � O Brown M � H � CHydrocarbon O � S

Stauroneis cf. wislouchii Overall B � S � O Overall H � C � MControl B � S � O Brown H � M � CMetal B � S � OHydrocarbon B � S � O

Synedropsis cf. recta Overall S � O � B

O � O’Brien Bay, S � Sparkes Bay, B � Brown Bay.

Table 5. ANOVA results for structural parameters of diatomcommunities for comparisons between treatments, locations,and the interactions between variables.

Location Treatment L � T

Richness 0.70 0.09 0.06Diversity 0.79 0.43 0.37Evenness 0.85 0.72 0.66Dominance 0.83 0.75 0.41

499CONTAMINANTS AND DIATOM COMMUNITIES

bly reflecting the comparatively low levels of hydrocar-bon contamination used in our study (400 mg�kg�1 asopposed to 45,000 mg�kg�1).

The diversity of diatom communities has previouslybeen reported to vary as a result of exposure to petro-leum hydrocarbons (Vargo et al. 1982, El-Dib et al.2001). Metal contamination has also been shown to in-fluence the structure of diatom communities, result-ing in lower diversity and richness but increased domi-nance values (Crossey and La Point 1988). Within thisstudy, significant differences as a result of petroleumhydrocarbon contamination and metal contaminationwere only observed in the composition, not the struc-

ture, of diatom communities. The low concentrationsof hydrocarbons and the nature of the metal contami-nation used in this experiment may explain the lack ofobserved structural response to contamination withinour experiment. Crossey and La Point (1988) exam-ined the effects of water column contamination onbenthic diatom communities, whereas this experimentexamines the influence of sediment. Contaminationvia the sediments has previously been demonstrated toproduce a more muted response than contaminationvia the water column (Peres et al. 1997). Under the ex-perimental conditions used in this study, diversity andassociated measures such as dominance are not useful

Fig. 5. Three-dimensional non-metric multidimensional scaling ordi-nation of diatom communities show-ing the separation of treatment typeswithin each location. Control treat-ments are indicated by an open circle;metal contaminated sediments are de-picted by an asterisk; hydrocarbon con-taminated sediments are indicated by asolid triangle. Stress levels indicatedthat a three-dimensional ordinationwas required; ordination axes as indi-cated.

500 LAURA CUNNINGHAM ET AL.

for monitoring effects of either metal or petroleum hy-drocarbon contamination of marine sediments at thelevels currently occurring around Casey Station.

Compositional changes in benthic diatom commu-nities resulting from metal contamination have previ-ously been documented in lakes (Ruggiu et al 1998)and rivers (Ivorra et al. 1999). Both of these studiesfound Achnanthes spp. to be metal tolerant, increasingin abundance with increased copper (Ruggiu et al.1998), zinc, and cadmium concentrations (Ivorra etal. 1999). Our results did not indicate any significantchanges in the relative abundances of Achnanthes spp.as a result of either metal or petroleum hydrocarboncontamination. The relative abundances of Achnan-thes brevipes were higher at control locations than inBrown Bay; however, this could have resulted fromfactors other than contamination.

Several species did respond significantly to contam-ination. Navicula cancellata appears to be pollution tol-erant, because significantly higher abundances of thisspecies were observed within both the metal contami-nated and hydrocarbon contaminated treatments, rel-ative to the control treatment. In contrast, the lowerrelative abundance of Navicula directa within the hy-drocarbon treatments, relative to the controls, suggestthat this species may be sensitive to hydrocarbon con-tamination. Navicula glaciei may also be sensitive to hy-drocarbon contamination, although because this re-sponse was only observed within Brown Bay, it ispossible there are other contributing factors involved.Further investigation of the responses of these speciesto varying contaminant concentrations may enablethem to be used as indicator species within the re-gion. As a genus, Navicula is considered to be highlytolerant of pollution (Palmer 1969); however, our re-sults clearly indicate that different species of this ge-nus can have varying responses to the same contami-nants.

It has previously been suggested that small specieswill become dominant in communities exposed tochemical stress (Kinross et al. 1993). Increased abun-dances of small forms of Navicula spp. have beenrelated to organic enrichment and eutrophication

Table 6. ANOSIM results for comparison of samples deployedat each location, both overall and for each treatment type.

Treatments compared R value Significance level

OverallSparkes Bay Brown Bay 0.631 0.0%Sparkes Bay O’Brien Bay 0.414 0.0%Brown Bay O’Brien Bay 0.553 0.0%

ControlSparkes Bay Brown Bay 0.850 0.2%Sparkes Bay O’Brien Bay 0.607 0.5%Brown Bay O’Brien Bay 0.766 0.5%

Metal ContaminatedSparkes Bay Brown Bay 0.563 0.2%Sparkes Bay O’Brien Bay 0.357 1.9%Brown Bay O’Brien Bay 0.757 0.2%

Hydrocarbon ContaminatedSparkes Bay Brown Bay 0.974 0.2%Sparkes Bay O’Brien Bay 0.767 0.2%Brown Bay O’Brien Bay 0.945 0.2%

All results were significant.

Fig. 6. Two-dimensional nonmetricmultidimensional scaling ordinationindicating the differences in composi-tion observed between the preexistingnatural communities from O’Brien Bayand the diatom communities that de-veloped on the experimental sedi-ments. Stress � 0.13.

Table 7. ANOVA results for structual parameters betweenrecruiting and developing communities.

Remnant communities

Developingcommunities P value

Diversity 2.903 0.007 3.091 0.021 0.00005Dominance 0.073 0.0592 0.0003 0.012Evenness 0.789 0.004 0.8457 0.009 0.000001

Results shown are for the calculated value SE (if greaterthan 0.0001).

501CONTAMINANTS AND DIATOM COMMUNITIES

(Kelly and Whitton 1995) and to zinc and cadmiumpollution (Ivorra et al. 1999). This trend was not ob-served within our results. Although the relative abun-dances of Navicula directa (a large species, 70–120 �min length in our samples) decreased in treatmentsexposed to hydrocarbon treatments, so too did therelative abundances of Navicula glaciei (8–45 �m inlength).

Further comparisons of the present work to otherstudies is hampered by a lack of published informa-tion. Most previous studies of benthic diatom re-sponses to chemical contaminants have examinedfreshwater, not marine, species (Ivorra et al. 1999,Ruggiu et al. 1998). The pollution tolerances of estua-rine and coastal diatom species to pollution are essen-tially unknown; data on this topic are so scarce it pre-vents the development of marine diatom indices forpollution monitoring purposes (Sullivan, 1999).

The two reference locations, Sparkes Bay andO’Brien Bay, developed the most similar diatom com-munities, although that of Sparkes Bay was intermedi-ate between O’Brien Bay and Brown Bay. It is possiblethat this reflects metal concentrations, with O’BrienBay having only background levels, Sparkes Bay hav-ing naturally high levels of several metals, and BrownBay having very high levels of a range of metals.

The Brown Bay samples had a community composi-tion significantly different from that of the combinedcommunities of the reference locations. This may be areflection of the hydrocarbon contamination presentwithin Brown Bay. This suggestion is supported by theexperimental results indicating significant changes inthe composition of diatom communities as a result ofhydrocarbon contamination. In addition to sedimentcontamination, differences in water chemistry couldalso be influencing the diatom communities of BrownBay. Leaching of contaminants from the waste dis-posal site has increased metal concentrations in boththe sediments and water column of Brown Bay (Snapeet al. 2001). Although previous studies (Ruggiu et al.1998, Ivorra et al. 1999) have demonstrated thatmetal contaminants within the water column can af-fect the composition of benthic species, no such studyhas been undertaken for the diatom flora of theWindmill Islands.

Significant changes in multivariate communitycomposition resulting from metal contaminationwere only observed within the polluted site, BrownBay. There are several possible explanations for this.Concentrations of many metals within experimentalsediments in Brown Bay increased during the experi-ment, presumably indicating further input fromThala Valley. It is possible that the diatoms were af-fected by the resultant higher levels but that the con-centrations used in the experimental treatmentsthemselves were too low to produce a detectable ef-fect. In other words, the diatom communities may beable to tolerate low levels of metals, but once a partic-ular threshold has been reached, diatom communitycomposition is significantly affected.

Alternatively, the pollution history of Brown Baycould be influencing the development of diatom com-munities. The level of preexisting stress in an area canstrongly influence the outcome of an experiment(Berge 1990). Communities in areas already stressedmay be more susceptible to the influence of furtherstressors (contamination or disturbance) than com-munities from reference locations (Underwood 1989).Thus, the contamination already present in BrownBay could increase susceptibility of the diatom com-munity to the sediment contamination imposed bythis experiment. This hypothesis is supported by thelarge compositional differences between diatom com-munities from both metal and hydrocarbon contami-nated treatments in Brown Bay and all other samples.Synergism between hydrocarbon and metal contami-nation within Brown Bay may also be contributing tothe greater effects observed within this bay.

This study has demonstrated that the compositionof diatom communities can be influenced by bothmetal and hydrocarbon contamination at concentra-tions comparable with pollution levels caused by stationactivity. Differences between control and contaminatedlocations and between control and hydrocarbon con-taminated treatments were apparent after only 11weeks, indicating diatom communities may be a use-ful tool for rapidly detecting impacts caused by petro-leum hydrocarbon and metal contamination. The re-sponse of diatom communities to a range of differentcontaminant concentrations would need to be deter-mined before the method could be used for routinemonitoring. If diatoms are used as a monitoring toolwithin the region, caution must be exercised to en-sure that the influence of location can be distin-guished from the effects of contamination.

This work was carried out at the Institute of Antarctic andSouthern Ocean Studies, in conjunction with the AustralianAntarctic Division, with the financial support of a TasmanianUniversity Strategic Scheme Scholarship awarded to LauraCunningham, an Australian Antarctic Division Ph.D. Scholar-ship awarded to Jonathan S. Stark. Logistic support was pro-vided by the Antarctic Science Advisory Committee (ASACProject No. 2201) awarded to Martin J. Riddle. Field supportfrom Andrew Tabor, Paul Goldsworthy, and J. Davidson, pro-vided through the Human Impacts Program, Australian Antarc-tic Division, was essential to this project and is gratefully ac-knowledged. Thanks also to M. Callinan and N. Babicka, whoassisted with sediment preparation, and to A. Revill (CSIRO,Hobart) for the hydrocarbon analyses. The assistance of JohnCox in the preparation of Figure 1 is much appreciated.

Berge, J. A. 1990. Macrofauna recolonization of subtidal sediments.Experimental studies on defaunated sediments contaminatedwith crude oil in two Norwegian fjords of unequal eutrophica-tion status. I. Community Responses. Mar. Ecol. Prog. Ser. 66:103–15.

Cathers, B., Tate, P. M. & Morris, C. E. 1998. Wind driven circula-tion of coastal waters at Casey Station, Antarctica. Antarctic Re-search Report No. 199 Australian Antarctic Division, AustraliaGovernment Press, Canberra, 164 pp.

Clarke, K. R. & Warwick, R. M. 1994. Change in marine communi-ties: an approach to statistical analysis and interpretation. Nat-

502 LAURA CUNNINGHAM ET AL.

ural Environmental Research Council, Plymouth Marine Labo-ratory, Plymouth, 144 pp.

Cole, C. M., Snape, I., Gore, D. B., Revill, A. T. & Riddle, M. J. 2000.Contaminants in the Antarctic III: chemical and physical pro-cesses that influence contaminants in cold regions. In Hugh-son, T. & Ruckstuhl, C. [Eds.] ISCORD 2000: Proceedings of theSixth International Symposium on Cold Region Development, Hobart.International Association of Cold Region DevelopmentStudies, Tasmanian Government Press, Hobart, pp. 122–131.

Crockett, A. B. 1997. Water and wastewater quality monitoring, Mc-Murdo Station, Antarctica. Env. Monit. Assess. 47:39–57.

Crossey, M. J. & La Point, T. W. 1988. A comparison of periphytoncommunity structure and functional responses to metal. Hydro-biologia 162:109–21.

Cunningham, L., McMinn, A., Riddle, M., Stark, J. & Snape, I. 2000.The ecology of, and human impacts on, near-shore benthicmarine algal mats in the Windmill Islands, Antarctica. The Sec-ond International Conference on Applications of Micro- andMeioorganisms to Environmental Problems: Program and Ab-stracts. Avalon Institute of Applied Science, Canada, p. 36.

Deprez, P. P., Arens, M. & Locher, H. 1999. Identification and pre-liminary assessment of contaminated sites at Casey Station,Wilkes Land. Polar Record 35:299–316.

El-Dib, M. A., Abou-Waly, H. F. & El-Naby, A. H. 2001. Fuel effecton the population growth, species diversity and chlorophyll (a)content of freshwater microalgae. Int. J. Environ. Health Res. 11:189–97.

Eriksen, R. S., Mackey, D. J., van Dam, R. & Nowak, B. 2001. Copperspeciation and toxicity in Macquarie Harbour, Tasmania: aninvestigation using a copper ion selective electrode. MarineChem. 74:99–113.

Everitt, D. A. & Thomas, D. P. 1986. Observations of seasonalchanges in diatoms at inshore localities near Davis Station, EastAntarctica. Hydrobiologia. 139:3–12.

Gilbert, N. S. 1991. Primary production by benthic microalgae innearshore marine sediments of Signy Island, Antarctica. PolarBiol. 11:339–46.

Hasle, G. R. & Syverston, E. E. 1996 Identifying marine diatoms. InTomas, C. R. [Ed.] Identifying Marine Diatoms and Dinoflagellates.Academic Press, Harcourt Brace & Company, New York, pp. 5–385.

Hsiao, S. I. C. 1978. Effects of crude oils on the growth of arctic ma-rine phytoplankton. Environ. Pollut. 17:93–107.

Ivorra, N., Hettelaar, J., Tubbing, G. M. J., Kraak, M. H. S., Sabater,S. & Admiraal, W. 1999. Translocation of microbenthic algalcommunities used for in situ analysis of metal pollution in riv-ers. Arch. Environ. Contam. Toxicol. 37:19–28.

Kelly, M. G. & Whitton, B. A. 1995. The trophic diatom index: anew index for monitoring eutrophication in rivers. J. Appl. Phy-col. 7:433–44.

Kennicutt, M. C., McDonald, S. J., Sericano, J. L., Boothe, P., Ol-iver, J., Safe, S., Presley, B. J., Liu, H., Wolfe, D., Wade, T. L.,Crockett, A. & Bockus, D. 1995. Human contamination of themarine environment—Arthur Harbour and McMurdo Sound,Antarctica. Environ. Sci. Technol. 29:1279–87.

Kinross, J. H., Christofi, N., Read, P. A. & Harriman, R. A. 1993. Fil-ametous algal communities related to pH in streams in Tro-ssachs, Scotland. Freshw. Biol. 30:301–17.

Lenihan, H. S. & Oliver, J. S. 1995. Anthropogenic and natural dis-turbances to marine benthic communities in Antarctica. Ecol.Appl. 5:311–26.

Lenihan, H. S., Oliver, J. S., Oakden, J. M. & Stephenson, M. D.1990. Intense and localised benthic marine pollution aroundMcMurdo Station, Antarctica. Mar. Poll. Bull. 21:422–30.

Medlin, L. K. & Priddle, J. [Eds.] 1990. Polar Marine Diatoms. British

Antarctic Survey. Natural Environment Research Council,British Antarctic Survey, Cambridge, UK, 214 pp.

Ostgaard, K., Hegseth, E. N. & Jensen, A. 1984. Species dependentsensitivity of marine planktonic algae to Ekofisk crude oil un-der different light conditions. Bot. Mar. XXVII:309–18.

Palmer, C. M. 1969. A composite rating of algae tolerating organicpollution. J. Phycol. 5:78–82.

Peres, F., Coste, M., Ribeyre, F. Ricard, M. & Boudou, A. 1997. Ef-fects of methylmercury and inorganic mercury on periphyticdiatoms communities in freshwater indoor microcosms. J. App.Phycol. 9:215–27.

Plante-Cuny, M. R., Salen-Picard, C., Grenz, G., Plante, R., Alliot, E.& Barranguet, C. 1993. Experimental field study on the effectsof crude oil, drill cuttings and natural biodeposits on micro-phyto- and macrozoobenthic communities in a Mediterraneanarea. Mar. Biol. 117:355–66.

Reid, M. A, Tibby, J. C., Penny, D. & Gell, P. A. 1995. The use of di-atoms to assess past and present water quality. Aust J. Ecol. 20:57–64.

Roberts, D. & McMinn, A. 1999. Diatoms of the saline lakes of theVestfold Hills, Antarctica. Biblio. Diatomol. 44:1–83.

Ruggiu, D., Luglie, A., Cattaneo A. & Panzani, P. 1998. Palaeoeco-logical evidence for diatom response to metal pollution inLake Orta (N. Italy). J. Paleolimn. 20:333–45.

Scouller, R. C., Stark, J. S., Snape, I., Riddle, M. J. & Gore, D. B.2000. Contaminants in the Antarctic Environment. V. Accumu-lation in marine sediments. In Hughson, T. & Ruckstuhl, C.[Eds.] Proceedings of the 6th International Symposium on Cold Re-gion Development. Tasmanian Government Press, Hobart, pp.136–9.

Siron, R., Giusti, G., Berland, B., Morales-Loo R. & Pelletier, E.1991. Water-soluble petroleum compounds: chemical aspectsand effects on the growth of microalgae. Sci. Tot. Environ. 104:211–27.

Snape, I., Riddle, M. J., Stark, J. S., Cole, C. M., King, C. K., Du-qense, S. & Gore, D. B. 2001. Management and remediation ofcontaminated sites at Casey Station, Antarctica. Polar Rec. 37:199–214.

Stark, J. S. 2000. The distribution and abundance of soft-sedimentmacrobenthos around Casey Station, East Antarctica. PolarBiol. 23:840–50.

Stark, J. S., Riddle, M. J., Scouller, R. C. & Snape, I. Human impactsin Antarctic marine soft-sediment assemblages: correlations be-tween multivariate biological patterns and environmental vari-ables. Est. Coast. Shelf Sci. (in press).

Stark, J. S., Snape, I. & Riddle, M. J. 2003. The effects of petroleumhydrocarbon and metal contamination of marine sedimentson recruitment of Antarctic soft-sediment communities: a fieldexperimental investigation. J. Exp Mar Biol Ecol 283:21–50.

Stronkhurst, J., Vos, P. C. & Misdrop, R. 1994. Trace metals, PCBsand PAHs in benthic (epipelic) diatoms from intertidal sedi-ments: a pilot study. Bull. Environ. Contam. Toxicol. 52:818–24.

Sullivan, M. J. 1999. Applied diatom studies in estuaries and shallowcoastal environments. In Stoermer, E. F. & Smol, J. P. [Eds.]The Diatoms: Applications for the Environmental and Earth Sciences.Cambridge University Press, Cambridge, 469 pp.

Underwood, A. J. 1989. The analysis of stress in natural popula-tions. Biol. J. Linn. Soc. 37:51–78.

Underwood, A. J. & Peterson, C. H. 1988. Towards an ecologicalframework for investigating pollution. Mar. Ecol. Progr. Ser. 46:227–34.

Vargo, G. A., Hutchins, M. & Almquist, G. 1982. The effect of lowchronic levels of No. 2 fuel oil on natural phytoplankton com-munities in microcosms. I. Species composition and seasonalsuccession. Mar. Environ. Res. 6:245–64.

503CONTAMINANTS AND DIATOM COMMUNITIES

Appendix. List of species names and authors.

Species name Author

Achnanthes brevipes AgardhAchnanthes delicatula complex (Kutzing) GrunowActinocyclus actinochilus (Ehrenberg) SimonsenAmphora ovalis var. affinis KutzingAmphora ovalis var. ovalis (Kutzing) KutzingAnomoeoneis cf. follis var. hannae ReimerAuricula compacta (Hustedt) MedlinCocconeis costata GregoryCocconeis fasciolata (Ehrenberg) BrownCocconeis pinnata GregoryCocconeis schuettii Van HeurckCtenophora pulchella (Ralfs ex Kutzing) Williams and RoundDactyliosolen antarcticus CastracaneDiploneis splendida (Gregory) CleveFallacia marnieri (Manguin) Witkowski, Lange-Bertalot and MetzetlinFragilaria cf. construens var. pumila GrunowFragilaria construens var. venter (Ehrenberg) GrunowNavicula aff. cincta (Ehrenberg) Van HeurckNavicula cf. cancellata DonkinNavicula directa (Smith) RalfsNavicula glaciei Van HeurckNavicula muticopsis Van HeurckNavicula perminuta Grunow in Van HeurckNavicula tripunctata var. schizonemoides (Van Heurck) PatrickNitzschia aff. hybrida Grunow in Cleve and GrunowPinnularia quadratarea (A. Schmidt) ClevePleurosigma aff. obscurum W. Smith emend SterrenburgPorosira pseudodenticulata (Hustedt) JousePseudogomphonema kamtschatica (Grunow) MedlinPseudostaurosira cf. brevistriata (Grunow in Van Heurck) Williams and RoundStauroneis cf. wislouchii Poretzky and AnisimowaSynedropsis cf. hyperboreoides Hasle, Syverston et MedlinSynedropsis cf. recta Hasle, Medlin et SyverstonTrachyneis aspera EhrenbergTrigonium arcticum (Brightwell) Cleve