FATTY ACIDS and STABLE ISOTOPES in ANTARCTIC SPONGES: DIET ANALYSIS of GUTLESS ANIMALS. a Thesis Presented to the Faculty Of

FATTY ACIDS AND STABLE ISOTOPES IN ANTARCTIC

SPONGES: DIET ANALYSIS OF GUTLESS ANIMALS.

A Thesis Presented to the Faculty of California State University, Stanislaus through Moss Landing Marine Laboratories

In Partial Fulfillment of the Requirements for the Degree Master of Science in Marine Science

By © Andrew Reichmann Thurber May 2005

CERTIFICATION OF APPROVAL

FATTY ACIDS AND STABLE ISOTOPES IN ANTARCTIC

SPONGES: DIET ANALYSIS OF GUTLESS ANIMALS.

By Andrew Thurber

______Dr. Nicholas Welschmeyer Date Professor of Biological Oceanography

______Dr. Stacy Kim Date Adjunct Professor of Benthic Ecology

______Dr. Pamela Roe Date Professor of Invertebrate Biology

iii

ACKNOWLEDGMENTS

This work would not have been possible without the support of my family, my committee, and the students, staff, and faculty of Moss Landing Marine Laboratories.

Rebecca Vega and the rest of my family were a constant source of support and joy throughout my time here. Invaluable comments on drafts of my proposal and thesis were made by Rob Leaf, Rhea Sanders, and Aroon Melwani. Dive buddies including Aaron

Carlisle, Dr. John Oliver, Dr. Kathy Conlan, Mike Donnellan, Jennifer Fisher, Dan

Malone, Dr. Stacy Kim, Rob Robbins, and Bob Zook helped regardless of how tired they

may have been. Aaron Carlisle was always willing to get dragged into the field at any

hour of day or night and for that I am grateful. Wade Smith, Dr. John Oliver, and Dr.

Rob Burton offered valuable advice throughout my time here at MLML and much of this

research was inspired by a class taught by and interactions with Dr. Burton. Dr. Stacy

Kim provided enthusiasm, opportunity, never ending support, and many a good time to

this research, not to mention advice along the way. Dr. Pamela Roe was amazing at

easing the life of a Moss Landing graduate student by always fixing any paperwork

problem I may have encountered or caused, not to mention being an invaluable member

of my committee. Dr. Nicholas Welschmeyer fostered the oceanographic side of my

research interests and helped me think in a more pelagic way. The MLML Benthic

Ecology lab provided a home and an endless amount of entertainment and support.

Kamille Hammerstrom provided the GIS maps and was there at the end to bounce ideas

iv off of. My initial interest in benthic ecology and the Antarctic were the result of interactions with Dr. Eric Vetter and Dr. Craig Smith. This research was supported by funds from the Earl and Ethyl Meyers Oceanographic Research Trust, the PADI

Foundation, the Dr. John H. Martin Memorial Scholarship, and the National Science

Foundation, Office of Polar Programs grant 0126319.

TABLE OF CONTENTS

Publication Rights...... iii

Acknowledgements...... iv

List of Tables ...... vii

List of Figures...... viii

Abstract...... x

Introduction...... 1

Materials and Methods...... 8

Results...... 14

Discussion...... 18

Conclusions...... 31

Literature cited ...... 32

Tables...... 41

Figures...... 46

LIST OF TABLES

TABLES PAGE

Table 1: The distribution of fatty acids, viable fecal coliform bacteria, and total bacteria as enumerated by epifluorescence microscopy at the given distances from the McMurdo Station outfall, Antarctica, listed from North to South. (---) indicates bellow detectable limit...... 42

Table 2: Statistical comparison between a selection of fatty acid biomarker concentrations in four species of sponge. Nested ANOVAs compare between species and linear regressions indicate differences in concentration with increasing distance from the McMurdo Station sewage outfall. Post hoc results are indicated on Figures 2,3, and 5...... 43

Table 3: Statistical comparison between groups of fatty acid biomarker concentrations in four species of sponge. Nested ANOVAs compare between species and linear regressions indicate differences in concentration with increasing distance from the McMurdo Station sewage outfall. Post hoc results are indicated on Figures 4 and 5...... 44

Table 4: Statistical comparison between the stable isotopic concentrations of four species of sponge. Nested ANOVAs compare between species and linear regressions indicate differences in concentration with increasing distance from the McMurdo Station sewage outfall...... 45

vii

LIST OF FIGURE

FIGURES PAGE

Figure 1: Locations of sponge and water collections at McMurdo Station, Antarctica. Current speed (cm s-1)and direction during the time of water collection is shown in the vector diagram as recorded by an S4 current meter...... 47

Figure 2: Percent composition of the FA 18:2(n-6) (top) and 18:1(n-9) (bottom) in four species of sponge at varying distances from the McMurdo Station outfall. Locations are listed from north to south and error bars are standard error. Location 166 for I. setifera had no replication and the lack of error bars reflects this. Differences in species are indicated by different letters over the group and a significant change with distance is indicated by an asterisk. Statistical results are presented in Table 2...... 48

Figure 3: Percent composition of the FA 22:6(n-3) (top) and 16:1(n-7) (bottom) in four species of sponge at varying distances from the McMurdo Station outfall. Locations are listed from north to south and error bars are standard error. Location 166 for I. setifera had no replication and the lack of error bars reflects this. Differences in species are indicated by different letters over the group and a significant change with distance is indicated by an asterisk. Statistical results are presented in Table 2...... 49

Figure 4: Percent composition of the polyunsaturated fatty acids (top) and bacterial fatty acids (15:0, 17:0, iso- and antiso-FA) (bottom) in four species of sponge at varying distances from the McMurdo Station outfall. Locations are listed from north to south and error bars are standard error. Location 166 for I. setifera had no replication and the lack of error bars reflects this. Differences in species are indicated by different letters over the group and a significant change with distance is indicated by an asterisk. Statistical results are presented in Table 3...... 50

Figure 5: The ratio of bacterial fatty acids to polyunsaturated fatty acids (top) and the abundance of 20:5(n-3) (bottom) in four species of sponge at varying distances from the McMurdo Station outfall. Locations are listed from north to south and error bars are standard error. Location 166 for I. setifera had no replication and the lack of error bars reflects this. Differences in species are indicated by different letters over the group and a significant change with distance is indicated by an asterisk. Statistical results are presented in Tables 2 and 3...... 51

viii

FIGURES PAGE

Figure 6: Principal Component Analysis (PCA) plot of the fatty acid constituents of four species of sponge at varying distances from the McMurdo Station sewage outfall...... 52

Figure 7a: Bi-plot of the eigenvectors that accounted for more than 10% of the variability in principal components one and two...... 53

Figure 7b: Bi-plot of the eigenvectors that accounted for more than 10% of the variability in principal components two and three...... 54

Figure 8: Stable isotopic concentration of four species of sponge at indicated distances from the McMurdo Station sewage outfall, Antarctica. Error bars indicate standard error. I. setifera data are expanded in the inset...... 55

Figure 9: Percent decrease of total bacterial abundance over time in a closed system containing the indicated species of sponge or a control of nothing. n=5 and error bars are standard error...... 56 .

ABSTRACT

The Antarctic has a diverse and conspicuous sponge community, but the diets of

Antarctic sponges have not been quantified or described. This study used three techniques, fatty acid analysis, stable isotope concentrations, and a laboratory based

feeding study, to address the diet of four species of Antarctic sponge: Homaxinella

balfourensis, Isodictya setifera, Kirkpatrickia variolosa, and Sphaerotylus antarcticus.

Sponges were sampled at distances between 115 and 840 m from the McMurdo station

sewage outfall to provide potential diet variability. The sewage effluent acted as a tracer

for particulates larger than bacteria. Sponge diet ranged from mostly bacteria, in the case

of I. setifera, to mostly flagellates, for H. balfourensis. The diet of K. variolosa was

intermediate between these two. The diet of S. antarcticus was not completely resolved

by this study; fatty acid analysis supported its similarity to K. variolosa yet the isotopic

analysis and feeding study separated it from the other sponges studied, suggesting that

symbionts were abundant enough in this species to confound the results. This study is the

first application of fatty acid analysis to determine diet composition of sponges, the first

stable isotopic analysis of Antarctic sponges identified to species, and the first indication

of differential utilization of microbial loop components by co-occurring sponges. The

role of sponges as a conduit for microbial resources to the benthic metazoan food web is

discussed.

INTRODUCTION

A unique benthic fauna surrounds the Antarctic continent, with large sponges locally abundant and ecologically important (Dayton et al., 1974; Starsman et al., 1999;

Dayton, 1990; Arntz et al., 1994). Sponges play various roles in Antarctic ecosystems: they provide a food source for asteroids, pycnogonids (Dayton et al., 1974) and nudibranchs (Wagele, 1989), host diatoms and cyanobacteria (Wilkinson, 1983; Sara et al., 1998), are a location where fish lay their eggs, and are the only known location of

Antarctic pentacrinoids (Barthel, 1997). Yet even with this importance to the rest of the ecosystem, very little is known about how these sponges subsist; growth rates, reproductive patterns, and even distribution of sponges in the high Antarctic have only been superficially assessed since the seminal work of Dayton et al. (1974). The basic life history question of diet in Antarctic sponges has only been examined by Kowalke (2000), who measured what sponges were physically capable of using as food, not what they actually consume. This study addresses the diet of four sponge species that are common in the high Antarctic, using stable isotope concentrations, fatty acid analysis, and bacterial consumption experiments.

Sponge food is dependent on what is available in the water column, so local plankton dynamics must be known for an examination of diet. The sponges at McMurdo

Station, the location of this study, experience the archetypal Antarctic seasonal variation in phytoplankton abundance. Summer starts with a diatom bloom followed by a prymnesiophyte bloom (Phaeocystis sp.) that is an order of magnitude larger in biomass

1 than the diatom bloom (Stoecker et al., 1995). The abundance of heterotrophic

nanoplankton (including dinoflagellates, Mesodinium spp., Didinium spp., oligotrichs,

tintinnids, and other ciliates) increases throughout the summer, culminating in a

community composed mostly of athecate dinoflagellates (Stoecker et al., 1995).

Bacterioplankton (<2µm) and nanoplankton (<20µm) continue to be abundant, and in

other Antarctic locations, bacteria alone can account for up to 57% of the biomass of all

living non-zooplankton (i.e. phytoplankton, bacteria, and ciliates) at the end of the

summer (Becquevort et al., 2000). Although planktonic assemblage studies during winter

as far south as McMurdo are rare, at Signey Island, which is north of the Antarctic peninsula, plankton larger than 20µm were absent during the winter (Clarke and Leakey,

1996). The unavailability of plankton suggests the potential role of bacteria as a major source of food since bacterial production in the Antarctic periodically constitutes more than 40% of total production (Ducklow et al., 2001). It is conventionally presumed that filtering macrofauna in the Antarctic feed only during the summer phytoplankton bloom months, and for many species that do not have the capacity to filter smaller organisms this may be true. However, animals that are capable of feeding on microbe-sized plankton have the potential to feed year round. The studies cited above support the characterization of the southern ocean with a food web where protozoa, dissolved organic material (DOM), and bacteria all play key roles for larger fauna (Smetacek et al., 1990).

Pelagic plankton is not the only available carbon source for Antarctic benthic filter feeders; sea ice communities and sediments, through anchor ice formation, are potential sources of sustenance (Dayton, 1990). Sea ice ecosystems consist of an ice-

2 water interface dominated by diatoms (Archer et al., 1996) and internal brine channels

filled with microbial protists (specifically athecate dinoflagellates, chrysophyte

statocysts) and <5µm photosynthetic flagellates (Stoecker et al., 1992, 1993). These

organisms are released into the water column as sea ice breaks up, providing a potential

source of food (Stoecker et al., 1992). In Prydz Bay sea ice communities provided >20%

of annual carbon caught in sediment traps (Gibson et al., 1999). Sedimentary carbon,

either from phytodetritus from the Phaeocystis bloom (Gutt et al., 1998) or benthic

production/decomposition can be resuspended through anchor ice formation (Carey,

1987; Dayton, 1990). Anchor ice forms on the seafloor in shallow water (<15-30m)

using organisms or sediment as condensation nuclei. As the ice enlarges its buoyancy overcomes the attachment of the nuclei to the seafloor and it is lifted and incorporated into the overlying ice sheet. At some later date the ice may melt and deposit the original condensation nuclei. Although sediment on the seafloor is available all year as a food source, food from sediments deposited via anchor ice, and release of sea ice communities, is periodic and unavailable throughout the winter months while sea ice is forming instead of breaking up.

With bacterioplankton making up a significant portion of the planktonic biomass during much of the year, it would be advantageous for Antarctic fauna to be able to consume these <2µm carbon sources. In this light, it is not surprising that asteroid larvae from the Antarctic are capable of feeding on bacteria (Rivkin et al., 1991). Tunicates, members of another common filter feeding group in the high Antarctic (Dayton et al.,

1974), have been found to consume particles starting at 1.4µm which leaves much of the free-living bacteria out of their reach (Kowalke, 1999). Sponges are uniquely capable of

3 filtering down to the bacterial size range, between 0.2 and 2 µm (Reiswig, 1971, 1975;

Stewart and Klumpp, 1984), but only one study has looked specifically at species in the

Antarctic (Kowalke, 2000). The two Antarctic species that have been studied, Isodictya kerguelensis and Mycale accerata, were able to retain particles larger than 0.4 and 0.6

µm, respectively (Kowalke, 2000). For a discussion of sponge filtration apparatus and mechanisms see Kowalke (2000). Although the studies that have been done show the ability of sponges to filter this bacterial size range, none have examined the bacterial contribution to diet or what sponges utilize as their main food source.

Since sponges have no digestive tract, making gut content analysis impossible, analysis of stable isotopes and fatty acids (FA) may reveal what sponges utilize as food.

Stable isotope analysis takes advantage of predictable fractionation, or discrimination, of carbon and nitrogen isotopes by primary producers during fixation. This creates a unique signature depending on type of fixation in photosynthesis and isotopic availability

(reviewed in Lajtha and Michener, 1994). As the carbon moves up the food chain this signature is conserved, +1±1 ‰ for carbon and +3±1‰ for nitrogen each time it is consumed (Peterson, 1999; Zanden and Rasmussen, 2001; Macko et al., 1982). By knowing the relative abundance of 13C to 12C and 15N to 14N in the organism, the relative trophic level and diet overlap of the organism can be determined (Lajtha and Michener,

1994). Examination of stable isotope concentrations has been employed in the Antarctic to elucidate food chains and trophic interactions for several years (Wada et al., 1987;

Nyssen et al., 2002). The main sources of carbon available to sponges will range from

-31‰ to -21‰ from plankton and ca. –15‰ for sea-ice communities, the negative value indicating that the heavier isotope is less abundant in the tissue than the naturally

4 occurring ratio, represented by a standard (Gibson et al., 1999; Nyssen et al., 2002).

Nyssen et al. (2002) examined the isotopic signature of Antarctic sponges at the South

Shetland Islands; though they didn’t identify the sponges they analyzed, they did suggest that its isotopic signature was the result of a sea-ice derived food source.

To avoid some of the traditional pitfalls of stable isotope application (Gannes et al., 1997; Peterson, 1999), fatty acid analysis allows an independent assessment of food source (Graeve et al., 1994, 2001; Phleger et al., 1998). Most animals are not capable of synthesizing all of the FAs that they require so certain FAs are conserved from their food source (Watanabe et al., 1983). Different primary producers synthesize specific FAs, usually one of the essential poly unsaturated fatty acids (PUFAs), which are essential for their consumers. These FAs are metabolically expensive to produce so consumers incorporate FAs directly from their diet into their tissue rather than synthesize them de novo. By examining the FAs present in the tissues of a consumer the source of production can be identified. Naming conventions for FAs are based on chain length and the location of saturated bonds (so 18:1(n-7) is an 18 carbon chain with a single double bond on the 7th carbon from the tail.) The specific FAs that are of note in the Antarctic are 22:6(n-3), synthesized by flagellates, ciliates, and phytoplankton, and 20:5(n-3), synthesized by diatoms (Zhukova and Kharlamenko, 1999; Sargent and Whittle, 1981).

When material is recycled through bacteria the PUFAs are not conserved and the PUFA signature is replaced by a bacterial signature of 15:0, 17:0, iso and anteiso FAs, and

18:1(n-7) (Jeffries, 1972; Volkman et al., 1980; Zhukova and Kharlamenko, 1999). Only a subset of all bacteria are capable of containing PUFAs, specifically certain taxa that are present in the deep sea and in sea ice communities (Yano et al., 1997; Nichols and

5 McMeekin, 2002; Nichols et al., 1993). By analyzing the FA constituents of different

sponges the relative input of bacteria, diatoms, and flagellates/ciliates to the diet of

sponges can be resolved.

Although a combination of stable isotope and FAs will yield the taxa of the primary producer and its source, i.e. water column or sea-ice community, they lack the

ability to differentiate between bacteria that are symbiotic within the sponge and those

that are consumed from the water column. To discriminate between the two, sponge

tissue collection was carried out along the fecal pollution gradient at McMurdo Station.

The pollution gradient consists of macerated human waste discharged into the ocean with human fecal bacteria spreading down current after the initial deposition of larger particles

(Edwards et al., 1998; Conlan et al., 2000); the fecal bacteria are constantly diluted by seawater creating a gradient. Sewage has a unique isotopic composition (Conlan et al.,

2000; Burnett and Schaeffer, 1980; Van Dover et al., 1992) as well as FA content identified by 18:2(n-6) and 20:5(n-3) (Rieley et al., 1997). This creates a gradient of labeled particles that could be a proxy for naturally occurring water column bacteria. If consumption of bacteria occurs, it will be reflected in the isotopic and FA composition of the sponges, with sponges closer to the outfall reflecting more of a sewage signature than

sponges elsewhere. It is assumed that larger conglomerate pieces of fecal bacteria are not carried down current; this assumption can be verified by examining the relative sewage input to those species that consume larger particles along the gradient. To further address the role of symbiotic bacteria compared to consumption of water column bacteria, sponge filtration of water column bacteria was measured in the laboratory.

6 The objective of this research is to identify prey items of four species of Antarctic

sponge. I hypothesize that bacteria are a major food source for Antarctic sponges during the late winter and spring, the time of this study. This study is the first FA study and the first stable isotopic study on identified species of Antarctic sponges. Since these two techniques indicate the source of carbon utilized by these sponges this study will demonstrate the importance of the microbial loop to the benthic food web using sponges as the link between microbes and metazoans.

MATERIALS AND METHODS

Tissue collection for FA and stable isotope analysis sampling was carried out from 5 November to 23 November, 2002 at McMurdo Station, Antarctica (Figure 1).

Stations were chosen along the fecal pollution gradient at distances down current of

115m, 166m, 410m, 840m south of the outfall and one location on the north side at a

distance of 700m. All collections were from 20m water depth. Sponges of the species

Homaxinella balfourensis, Isodictya setifera, Kirkpatrickia variolosa, and Sphaerotylus antarcticus were collected at each location within a three meter radius. At the 410m site a transect perpendicular to the outfall was sampled due to low abundances of all sponge species. At the 840m site a ca. 20m wide transect was used because the sponges sought were patchily distributed.

Sponge tissues were collected using SCUBA. A small portion of the osculum was removed with a pair of scissors and placed in a mesh bag. Upon reaching the lab the sponges were sorted into like species and placed in 0.6 µm filtered seawater for 24 hours to remove the ambient seawater signature from the tissue. Following this they were placed into individual bags and frozen at –20oC until analysis. Exposure to air was

minimized.

Fatty acid analysis was carried out by Ellen Dickstien at the Bacterial

Identification and Fatty Acid Analysis Laboratory, University of Florida with the

Sherlock system (MIDI systems). Approximately 0.5 grams of tissue was placed in 2ml

of a sodium hydroxide, methanol, water solution (45g NaOH: 150ml EtOH: 150ml H2O

8 stock solution) at which point it was vortexed, heated to 100 degrees C for five minutes,

vortexed again, reheated to the same temperature for a twenty-five minutes, and cooled to

break the bond between FAs and their glycerol group, termed saponification. A methyl

group was added to the FAs by adding 4ml of a hydrochloric acid, methanol solution

(325ml 6.0N HCl : 275ml MeOH stock solution) at which point it was vortexed and heated for 10 minutes at 80 degrees C and then cooled rapidly. Finally the FAs were extracted from the rest of the solution by adding 1.25ml of a 1:1 mixture of hexane and methyl tert-butyl ether, tumbled for twenty minutes, and the bottom, or aqueous phase was discarded. Since these samples had a lot of debris present, the organic phase was transferred to a clean test tube and washed with four to five 3ml aliquots of a sodium hydroxide solution (10.8g NaOH: 900ml H2O stock solution) which were sequentially

added, tumbled for 5minutes, and decanted before adding the next aliquot. The solution

still appeared to have some of the aqueous layer present so anhydrous sodium sulfate was added before the sample was placed in a gas chromatograph (GC) vial. This sample was then injected on a 25mm x 0.2mm Ultra 2 column and flame ionization equipped Agilent

GC that heated the sample from 170 degrees C to 270 degrees C at a rate of five degrees

C per minute. The GC was calibrated by a standard (Microbial ID, Inc) and peaks were identified by comparison of retention time to mass spectrometer analyzed samples. This method was modified from Technical Note #1 available from MIDI systems.

Stable isotope analysis was carried out at the Stable Isotope Lab, University of

California, Davis after lipid removal and drying at Moss Landing Marine Labs. Lipids were removed by adding 5 ml methanol, 3 ml chloroform, and 2 ml distilled water, at which point they were allowed to sit overnight before the organic phase was discarded.

9 The samples were rinsed and then dried in an approximately 40 degree C oven. Samples

were then combusted on a Europa Hydra 20/20 ionic ratio mass spectrometer. Stable

isotopic results are presented in δ notation in reference to standards of Pee Dee

Belemnite, a marine limestone, and atmospheric nitrogen using the equation (Lajtha and

Michener, 1994)

13 15 δ C (‰) or δ N (‰) = [(Rsample/Rstandard) -1] x 1000

where R= 13C/12C or 15N/14N

Regression analyses were used on the arcsine square root transformed FA proportion data to analyze differences within species between sites and a nested ANOVA of species within site was used to compare and contrast the species, using a Tukey post hoc test. The ratio of bacteria to PUFA required an arcsine ((number)^(1/4))/100) transformation for the sake of normality. Stable isotopic results were analyzed using the same statistical tests except δ13C required no transformation and δ15N required square

root transformation. Statistical comparisons were carried out using Systat version 10

(SPSS Inc.). A principal component analysis (PCA) statistical comparison to look for the

overall trends of FA compositions was run on the entire data set of FAs which

incorporates all fatty acids present rather than just the predetermined biomarkers. The

PCA was run using Primer version 5.2.3 (Primer-E Ltd.)

Water Sampling

Water sampling was carried out to measure 1) fatty acid composition of

particulate matter 2) abundance of water column free living bacteria and 3) abundance of

total coliform bacteria. Water was collected at the same sampling locations as above as

well as directly at the sewage outfall. All sampling was done using a 10 L Niskin bottle

on five consecutive days from 17 November to 22 November, 2002. Locations were

sampled in increasing proximity to the sewage outfall to decrease contamination, and

bottles were washed in >300C fresh water prior to redeployment after outfall sampling.

All stations were sampled at approximately 2 meters above the bottom and sampling

coincided with the deployment of an S4 current meter (Ocean Systems) to indicate

prevailing currents during the time of sampling.

Fatty acid composition of the particulate matter was collected by filtering between

6 and 10 L of water through a precombusted Glass Fiber Filter (GF/F), nominal pore size of 0.6 µm. Water was filtered until the GF/F would not pass any more water. The filters

were frozen at –80oC as soon as filtration was finished. This sampling was not replicated

at each site but carried out at each location on consecutive days.

Microbial abundance was measured using the 4',6-Diamidino-2-phenylindole

(DAPI) techniques of Sherr et al. (2001). After preservation with 2.5% gluteraldehyde

samples were filtered within 48 hours. Ten milliliters of preserved water were vacuum

filtered onto 0.2 µm pore sized nucleopore membrane filter. After mounting on slides the

membrane filters were frozen and shipped to Moss Landing Marine Labs for later

11 epiflorescence analysis. Influence of the outfall on bacterial abundance was analyzed

using regression analysis with distance from the outfall being the independent factor and

bacterial abundance being the dependent factor. A significant result will support the

hypothesis that the outfall is contributing to the bacteria in the water column either

through sewage-derived dissolved organic matter (DOM) or by the fecal bacteria

themselves.

Fecal coliform abundance was measured using m-Coliblue 24 protocols

(Millipore Corp.).

Feeding experiments

Each of the sponge species was collected using scuba between November 28th and

December 7th, 2004 at either the 840m site (replicate 1, 2 and 3) or -700m site (replicates

4 and 5.) Sponges were individually placed in a 7.5 liter bucket after being removed from their substratum when necessary or placed in the bucket with the substratum if it would fit. After transport back to Crary Lab Aquarium at McMurdo Station they were

acclimated for >24 hours with a flow through seawater system in a water table to keep them at <2 degrees C. No epibionts were removed nor was the water aerated for fear of

1) disturbing of the sponges by additional handling or 2) bubbles impacting the sponges.

A control consisted of an empty bucket which was placed in the sea water table at the same time and had the same flow through regime. At the start of the experiment the flow through water was shut off and 15ml water samples were taken initially and at half hour intervals for two hours. These water samples were preserved with a 0.2µm filtered 10%

12 buffered formaldehyde solution. They were stained with DAPI and 15ml was filtered onto 0.22 µm poretics membrane filters (Osmonics Co.) and enumerated using epiflourescent microscopy. Although Sherr et al. (2001) recommend filtration within 24 hours this was not possible due to field supply limitations and the samples were stored for

2 months before filtration at Moss Landing Marine Laboratories. To enumerate the extremely low abundances of bacteria in certain samples, an 81µm wide transect was run until 200 bacteria were counted at which point the length of the transect was recorded.

This method produced reproducible results with low concentration of bacteria and very similar results to the method used by Sherr et al. (2001) when the abundances approximated normal seawater. Dry mass of sponges was recorded after finishing the experiment.

Species were compared by integrating under a standardized (count divided by initial abundance) curve of bacteria per ml and using a one way ANCOVA with mass of sponge as the covariant, using a Tukey post hoc multiple comparison analysis. The area under the curve technique was suggested by Crowder and Hand (1990).

RESULTS

Water Column

The outfall did not significantly impact the abundance of bacteria in the water

column (Table 1). The abundance of the fecal coliform bacteria varied from zero to 54

colonies per 100ml present at the outfall and coliforms were found at both the -700 and

840m sites, though in much lower abundances. The DAPI direct count of bacterial

abundance yielded abundances between 3.1x104 and 4.1x104 bacteria per ml and these

values were not increased in samples that had a positive result for fecal bacteria.

The fatty acid concentrations in the water column were impacted by the outfall in regard to the pollution indicator 18:2 (n-6) at the -700 m, 0m, and 166m locations.

These stations were sampled at the same time that fecal coliform bacteria were present, supporting the hypothesis that the sewage pollution indicator did represent the sewage plume in this ecosystem and was retained on a GF/F that has a nominal pore size of 0.6

µm. Although this supports that there was an impact on the microbial or larger community, the low abundance of fecal bacteria in relation to naturally occurring bacteria indicates that it is not necessarily the bacterial component of the ecosystem that is changed, rather just an aspect of the ecosystem larger than 0.6 µm. The occurrence of the fecal indicators at the -700m location is explained by an oscillating flow parallel to the

14 coast so the “up-current” direction was commonly down-current as recorded by the

current meter (Figure 1).

Fatty Acids

The fatty acids that decreased with distance from the outfall were 18:2(n-6) and

18:1(n-9) (Figure 2, Table 2). 18:2 (n-6), the sewage biomarker, was the most abundant

in H. balfourensis and least in I. setifera, where it was only present at the station nearest the outfall; it was never more than 5% of the fatty acid composition. 18:1(n-9) also

decreased in all species with distance from the outfall, and was the most abundant in H. balfourensis. 22:6 (n-3), 16:1(n-7), and PUFAs increased with distance from the outfall in K. variolosa and H. balfourensis, I. setifera, and K. variolosa respectively (Figures 3,

4, Tables 2, 3).

The role of bacteria as a food source differed between the species. Bacterial fatty

acids were highest in I. setifera and S. antarcticus, composing >9 % of their total FA

(Figure 4, Table 3). Examining the ratio of bacterial FA to nonbacterial PUFAs we see

that the input of bacteria to I. setifera was two to three times greater than that of PUFAs

(Figure 5). All of the species were significantly different from each other, with the

lowest ratios in H. balfourensis followed by K. variolosa, and S. antarcticus (Table 3).

The nonmicrobial input to diet distinguished the species of sponge as well.

Diatom markers, 20:5(n-3), the more specific marker, and 16:1(n-7), which occurs in

many taxa, were present but the third diatom marker 16:4(n-3) was completely absent

(Figures 3, 5). S. antarcticus appeared to have more 20:5(n-3) than the other species but

15 this was not significant (Table 2). The flagellate biomarker, 22:6(n-3), abundance was species specific but had the opposite trend of bacterial input (Figure 3) and a relatively large input.

Principal Component Analysis indicated that the greatest amount of variability in

FA concentration was between species rather than sites (Figure 6). Three axes accounted for 68% of the variance between samples with the PC1 separating H. balfourensis from the rest of the species. PC2 separated I. setifera from S. antarcticus and K. variolosa and these two species were resolved from each other by PC3. Only one of the I. setifera samples did not group with the rest of its species and instead grouped with S. antarcticus.

The FAs that changed with location were components of the PC3 but the majority of the variance was explained by PUFAs and bacterial signatures (Figure 7). The flagellate marker 22:6 (n-3) was one of the major components in separating the species.

Stable Isotopes

The stable isotopic values of four species of sponge ranged from between -21 and

-24 ‰ for carbon and 4 to 12 ‰ for nitrogen (Figure 8). H. balfourensis and S. antarcticus were both significantly different from each other as well as the other species

(Table 4). Proximity to the outfall shifted the carbon isotope values heavier except in I. setifera whose value did not vary with distance to the outfall (regression analysis, Table

4). There was no shift in the nitrogen isotopic concentration due to proximity to the outfall for any of the species.

16 Laboratory Study

The feeding study supported differences in bacterial consumption between the

species (Figure 9, F4, 20=5.180 P= <0.01). The covariate of weight was not significant

and as such it was excluded from the model to increase the power of the ANOVA. I.

setifera was the only species to significantly decrease the abundance of bacteria in

comparison to the control. K. variolosa, although not significantly different from the

control, was also not significantly different from I. setifera, supporting bacterial

consumption but at a lesser efficiency than I. setifera. The majority of the bacteria were consumed within the first half hour of the experiment in four of the five I. setifera experiments and after an hour the bacterial concentrations were <103 bacteria per ml from

2 x 104. There was a large amount of variability within S. antarcticus and H. balfourensis, but neither of them consistently decreased the abundance of bacteria in the water.

DISCUSSION

Fatty Acid (FA) concentrations indicate the size of plankton consumed by each

sponge species. Bacteria are picoplankton, between 0.2-2µm, so a bacterial signature

indicates the <2µm contribution to diet. The >2µm contribution to diet is represented by the presence of Poly Unsaturated Fatty Acids (PUFAs) including 22:6(n-3), which can be de novo synthesized by flagellates as well as ciliates and phytoplankton (Zhukova and

Kharlamenko, 1999). In this regard, I. setifera consumes primarily bacteria, S.

antarcticus and K. variolosa consumes both bacteria and flagellates, and H. balfourensis

consumes mostly flagellates (Figures 4 and 5). The Principal Component Analysis supports that S. antarcticus and K. variolosa have the greatest overlap in their FA composition and therefore diet.

The impact of the outfall on the stable isotope and FA composition of the sponge species supported a differential utilization of bacteria as well as species specific diets.

Since the water column’s sewage signature was in a size class greater than 0.6 µm, dissolved organic matter is dismissed as being the source of the sewage signature. There was no change in microbial abundance with distance to the outfall, regardless of coliform abundance, therefore it is concluded that the majority of the sewage signature was also in the size class larger than bacteria or >2µm. This is supported by I. setifera which had the greatest utilization of bacteria, but the lowest concentration of sewage FAs and no change in its stable isotope composition with distance to the outfall. All of the other species

18 were impacted by the outfall and at least part of their diet was composed of larger

particles which contain PUFAs. When combing this with the overlap in stable isotope

values we observe that S. antarcticus and H. balfourensis have unique diets and that K.

variolosa and I. setifera, although similar trophic levels, differ in their utilization of

larger particles.

Although these stable isotope results indicate differential particle size

consumption, their application to trophic structure was confounded by a lack of

knowledge about the isotopic composition of the small size classes of plankton. Stable

isotope analysis predicts a change of +1 ‰ for carbon and a +3 ‰ for nitrogen for each

step along a food web, but this has rarely been applied separately to nano- and pico-

plankton or animals that consume in this size fraction. Both H. balfourensis and S.

antarcticus exhibit more positive δ13C values compared to the other species, but a

corresponding positive increase in their δ15N values was only observed for S. antarcticus.

H. balfourensis’s δ15N decreased in comparison yet fed on nanoplanton which are one

step up the food web from bacteria, the source of food for the I. setifera. Rau et al.

(1990) addressed the stable isotopic concentration differences between a 3-8µm,

nanoplankton, and the <3µm, picoplankton, fraction in temperate oceanic water. During

three out of four dates the 3-8µm had a decrease compared to the <3µm size class in δ15N whereas the δ13C always increased with increasing particle size (Rau et al, 1990). If the

same trend holds true in Antarctica then the values of H. balfourensis may support that

this species was feeding on the larger (3-8µm) size fraction. To make any conclusions in

19 this regard, analysis of size fractioned water in McMurdo Sound would be required to

confirm the size spectrum isotope results.

The feeding study supported the conclusions from the isotopic and FA analysis as

well as dismissed the idea that the bacterial signature was caused by symbiotic bacteria.

H. balfourensis did not decrease the abundance of bacteria in the water column whereas I.

setifera did. K. variolosa qualitatively decreased the concentration of bacteria but this

change was not statistically different from the control or I. setifera. Since K. variolosa

had a large proportion of PUFAs compared to the bacterial signature, a generalist diet of

both bacteria as well as flagellates would cause this intermediate decrease in bacterial

abundance. This is in agreement with the stable isotope analysis. Although sponges are

known to have a suit of symbiotic bacteria associated with them (Reiswig, 1974; Sara et

al.1998), the decrease of water column bacteria by I. setifera discredits symbionts as the source of the bacterial signature in the analyzed tissues. S. antarcticus did not decrease the abundance of water column bacteria even though FA results indicated it would, suggesting that symbiotic bacteria may be the source of the bacterial signature in this species.

There are two possible errors in the interpretation of the feeding study: antimicrobial compounds within the sponges and damage to sponges during transport that restricted or halted feeding. Previous similar feeding studies have been viewed with skepticism because microbial decrease may be due to bactereriostatic, i.e. antimicrobial, compounds rather than sponge feeding (e.g. introduction of Reiswig, 1971 concerning

Claus et al., 1967; Madri et al., 1967). Three of the sponges studied here, excluding S. antarcticus, have been reported to have antibacterial compounds in their extracts

20 (McClintock and Gauthier, 1994). Although the bacteriostatic nature of the sponges

could cause the decrease observed in this study, the agreement of the laboratory

technique with the bacterial fatty acid abundances confirms bactivory. Damage during

transport could cause nonfiltration by the sponges present which would falsely indicate

nonbacterial consumption (Van de Vyver et al., 1990). There were no measurements of

filtration rate during the experiment and it is possible that a few of the specimens were

not filtering but this was not the case with K. variolosa or I. setifera since they decreased

the bacterial abundance in the water column. During and for approximately 12 hours

after transport, S. antarcticus contracted its osculae but they were fully extended

throughout the two-hour experiment, supporting that this sponge was sufficiently

acclimated and filtering water. H. balfourensis could have been nonfiltering, but its lack

of bactivory agrees with the FA results.

The three techniques in combination support that each species utilizes a different

diet. I. setifera is a bacteriovore with little utilization of the larger aspects of the microbial loop. K. variolosa makes use of the bacterioplankton but also utilizes the larger

size classes of plankton. H. balfourensis doesn’t utilize bacteria but instead consumes

nanoplankton. S. antarcticus had the least conclusive results. The stable isotopic analysis

indicated that S. antarcticus had a different source of nutrition than the others. No

change in the bacterial abundance during the feeding experiment but having a large

abundance of bacterial FAs suggest a symbiotic relationship with bacteria, and S.

antarcticus’s large variation in δ13C with the outfall suggests feeding on large particles.

The diet of this species was not sufficiently resolved.

21 Diet Specificity in Sponges

Diet specificity between different species of demosponge is not a new concept, but niche separation in diet has not previously been shown for sponges (Pile et al., 1997;

Reiswig, 1971,, 1974; Stewart and Klump, 1984). Specificity in diet has traditionally been linked to habitat either due to the availability of food resources (Reiswig, 1974) or sedimentation (Kowalke, 2000). Reiswig (1974), in contrast to this study, stated:

It appears that unlike most other animal groups, niche partitioning in sponges, is not by direct food specialization since the only significant dietary differences which has been detected (DOC requirement by Verongia) is attributed to the activity of symbionts. (Reiswig, 1974)

This study showed that sponges have niche separation in their diet even though they co- occur in the same habitat.

A mechanism for sponge differentiation in diet was suggested by Reiswig (1971).

Particle uptake occurs in three separate locations within sponges: archaeocytes, choanocytes, and microvilli (Johnston and Hildemann, 1982). Archaeocytes consume particles 1µm and larger while roving within the sponge cellular matrix. Choanocytes filter to approximately the 1µm size range with their collar being the location of microvilli and uptake of particles in the 0.1µm size class. Reiswig (1971) hypothesized the different abundance and motility of archaeocytes were responsible for the ability of sponges to consume larger particles and this, combined with differential volumes of water filtered, separate the species’ diets. A different composition of these tissue types could explain the results here, with I. setifera having few archaeocytes and H.

22 balfourensis being dominated by archaeocytes and either reduced or morphologically

different microvilli, although this is pure speculation. Selectivity for bacterial type,

although consistently thought to be impossible (e.g. Kowalke, 2000), has been

demonstrated (Van de Vyver et al., 1985), and this too could cause differentiation in diet.

Niche separation due to resource partitioning has been key point of study in

marine ecology for many years (Paine, 1964). The demosponge fauna around the

Antarctic is diverse and co-occurs with other filter feeding taxa, specifically octocorals,

hydroids, tunicates, and bryozoans. The different taxa appear to separate their diets between species as well as larger taxonomic grouping (Orejas et al., 2001; Orejas, 2001;

Kowalke et al., 2001). Niche partitioning between groups has been observed in other locations with more consistent planktonic communities (Stewart and Klump, 1984). In the Antarctic the co-occurrence of many different species with similar feeding strategies espouses the importance of niche partitioning through particle size utilization and specificity, as observed in this study.

Dissolved Organic Matter as a Food Source for Sponges

DOM has been shown to be a food source for sponges as well as Antarctic larval invertebrates. Reiswig (1974) found that one species of tropical sponge must use DOM to account for approximately 70 % of its metabolic requirements. Nonarthropodian marine invertebrate larvae have been shown in a variety of studies to uptake DOM from seawater (Manahan, 1990). The role of amino acids from DOM has been the focus of the majority of research and it has been shown in three phyla that DOM is utilized to

23 supplement growth (Manahan, 1990; Shilling and Bosch, 1994). FAs are components of

DOM and may act as a food resource, impacting the analysis undertaken here. No

studies, to my knowledge, have looked at the DOM fatty acid composition of Antarctic seawater. One of the few studies to address this, from the equatorial Pacific, found that between 7 and 91% of the FAs in seawater were B-hydroxy acids that are characteristic of marine bacteria (Wakeham et al., 1997). Furthermore, the FA constituents of the dissolved fatty acids are independent of phytoplankton growth, and PUFAs are not a dominant component of DOM FAs (Kattner and Brockmann, 1990). Since microbial

lipids can be the most abundant portion of the DOM pool of lipids, the microbial loop is

key in DOM nutrition regardless of whether direct consumption of bacteria occurs or not.

In this study the utilization of DOM by the sponges could be the source of the

bacterial FA signature, but this would not account for the decrease in microbial

abundance observed during the feeding experiment. I. setifera and to a lesser degree K.

variolosa consumed bacteria as shown by the feeding experiment. The exception for this

may be S. antarcticus that showed a bacterial FA signature but no impact on the bacterial

content of seawater during the feeding experiment. The relative abundance of diatom

biomarkers, FAs that could be present in the DOM pool, as well as bacterial markers,

could be the result of utilization of DOM directly, since if there were a transfer through

bacterial symbionts it would destroy the diatomaceous PUFA signature. Two alternative

hypotheses that could account for the presence of both diatom and bacterial FAs are 1)

DOM uptake through symbiotic bacteria and direct feeding on diatoms, which are at the

uppermost size limit of sponge feeding abilities, and 2) symbiotic diatoms causing the

diatomaceous FA signature and the unique stable isotope signature.

Feeding Study Limitation on Use

The feeding study had problems in its design due to long preservation and time point selection that limited the analyses that could be applied. Preserved bacteria with a visible nucleoid are steady over a period of seven months but the overall bacterial count decreases by ca. 25% during 2 months of preservation (Vosjan and van Noort, 1998). In this regard the feeding study only addresses the abundance of nucleoid visible bacteria since filtration was delayed. This explains why the abundances were less (1 to 2 x 104 average initial values compared to 3 x 104 counts from the 2002 samples that were

filtered within 48 hours of sampling.) Since controls and all treatments were treated the

same, the implications of this is that bacteria with a visible nucleoid are consumed by the

sponges and consumption of nonvisible nucleoid bacteria is not addressed.

The time scale missed the exponential decrease in the microbial abundance,

restricting accurate measurement of filtration or clearance rate. I. setifera filtered >50%

of the bacteria out of the water within the first half hour, after which point the water had

an approximate steady state abundance of bacteria. This lack of temporal resolution was

likely the reason why mass was not a significant covariate, since rate of clearance is mass

dependent, assuming the doubling rate of the bacteria was insignificant to the volume

filtered. The appropriate future technique would be a flow-through system measuring the

microbial count in exhalent water, similar to what was done by Riisgard et al. (1993).

This technique is highly invasive to the sponges and although accepted in the literature

does increase the likelihood of confounded results.

Alternative Food Bases to Water Column Processes

Both the stable isotopes as well as the FAs indicated that sea ice communities were not a source of food for sponges during this time of year. Sea ice communities have high concentrations of PUFAs, (40-50%) and between 17-21 % of the FA concentration is composed of 20:5(n-3); the 22:6 (n-3) biomarker is only between 0.4-1% of the total sea ice community fatty acids (Nichols et al., 1993). Although high PUFAs were observed in H. balfourensis, 20:5(n-3) was not an abundant constituent. Sea-ice bacteria can also synthesize PUFAs, specifically 22:6(n-3) (Bowman et al., 1998). Since bacterial biomarkers had the opposite trend of 22:6(n-3), although not conclusive, this supports that sea ice was not a source of food at this time of year. Nyessen et al. (2002) used their heavy isotope as a tracer for sea ice community input since sea ice values are commonly about -15‰. The heaviest, i.e. least negative, value of the carbon stable isotope observed in sponge tissue was about -21‰ and away from the outfall was <-22‰. These values also refute sea ice as a source of food for sponges.

Tracers in Sponges: Known and New

The stable isotopic analyses presented here represent the first numbers for high

Antarctic (>64 degrees south) sponges. Both carbon and nitrogen values had the same range as those from King George Island (Nyssen et al., 2002). Palmer Station sponges

26 had δ13C values of -27, which is much lighter than any of the results here even though the

δ15N values were similar (Dunton, 2001). Kaehler et al. (2000) also used two non-

specified species of sponge and found a diet that didn’t vary in carbon but did vary in

nitrogen, by approximately two per mil. In contrast to these studies the sponges analyzed

in this study had both the lipids removed and species identified. Trial runs supported that

lipid removal greatly decreased the variability within the species of sponge studied

(author, unpublished data).

The FA composition of Antarctic sponges has not been previously reported nor

has the technique been applied for dietary purposes to populations elsewhere in the

world. FA analysis has been applied to sponges for purposes of analyzing biosynthetic

pathways (Carlalleira et al., 1998; Carlalleira and Pagan, 2001; Hahn et al., 1988),

chemotaxonomy (Lawson et al., 1986), and as a base of a food web though the sponges

diet was not discussed (Dembitsky et al., 1994). In all of these studies bacterial FA

signatures were present but the overall composition was usually simply referred to

“unusual” (Dembitsky et al., 1994). In this regard, the current studies' application of FA

results to Antarctic sponges is novel.

The application of tracers to sponges is an inexpensive way to analyze time

integrated diet, increasing our knowledge as a whole of the phylum. Many habitats where

sponges thrive, such as the deep sea, do not lend themselves to such thorough treatments

in diet analysis as are possible elsewhere (e.g. Reiswig, 1971). The time required by such

treatments also restricts the number of species analyzed and our knowledge about energy

flow through different components of the ecosystem. Because of this, the behavior of a

27 few species of sponge has been extrapolated to an entire phylum of organisms. Tracers

can also be applied to analyze how applicable these extrapolations are. The Antarctic is a

location where this sort of approach is critical, considering the >300 species of

demosponges (Sara et al., 1998), where food resources are limited and vary temporally,

and the expense and logistical problems of extensive laboratory work pose additional

challenges.

Ecological Implications of Diet Specificity

Antarctic sponges are traditionally thought of as slow growing, and as such,

relatively constant in their abundance; H. balfourensis is the exception to this paradigm.

The unique diet of H. balfourensis may help explain its different growth style compared

to the other monitored sponges. The population of H. balfourensis booms and busts in

the McMurdo Sound area in sync climatic oscillations (Dayton, 1989). Although the oscillations were from anchor ice caused by up-welled super cooled water, the oscillations could impact food availability by exposing the sponges to deeper, older, more food depleted water, similar to what is experienced on the west side of the sound (Dayton and Oliver, 1977). Deeper water in the Antarctic has reduced bacterial abundance and thus may decrease the abundance of bactivorous flagellates (Ducklow et al., 2001;

Church, 2003). During the periods of upwelling food limitation of larger particles may hasten the population decline that has been observed of H. balfourensis.

The other species of sponges grow slowly as predicted of Antarctic fauna. K. variolosa is more consistent of the Antarctic slow growth paradigm with two of the five

28 specimens followed for ten years increasing in volume by 5-10%, whereas none of the 23 followed for three years, nor any of the other three species followed for ten years had measurable growth (Dayton, 1979). Only one of 72 specimens of S. antarcticus followed for ten years showed increase in size. I. setifera was not an abundant member of the community so was not followed for the ten year period of study although two settled and grew five and eight cm during the eight years that they were recorded (Dayton, 1979).

Utilization of pico-plankton may be an evolutionary advantage for long lived sponge species, allowing them to compete with fast growing species, such as H. balfourensis.

Water Column Bacteria in Benthic Ecosystems

In the last thirty years, the one of the greatest revelations in our understanding of

the ocean is the importance of heterotrophic microbes in transferring organic matter in

the sea (Azam et al.1983, 1994; Legendre and Le Fevre, 1995; Azam, 1998; Landry,

2002). The majority of attention has been on the pelagic ecosystem where the pathway of

microbial carbon transfer is through protazoan grazers (Sherr and Sherr, 1994; Landry

and Kirchman, 2002; Smetacek et al., 1990 but see Vargas and Gonzalez, 2004), with few studies addressing what role pelagic microbes play for benthic organism (exceptions being Bak et al, 1998; Orejas et al., 2001; Gili et al., 2001). The ability of benthic organisms to consume directly on the bacteria could be a crucial step in energy transfer for ecosystems where bacteria are an abundant source of food including the deep sea, tropical latitudes, and polar ecosystems. The techniques used here could readily be applied to these ecosystems to address the role of bacterioplankton where the more

29 traditional filtration experiments are not feasible but in many instances filter feeders are

present. The sponges studied here provide food to other benthic fauna including the

asteroids Odontaster meridionalis, Odontaster validus, Perknaster fuscus antarcticus,

Acodontaster conspicuous, and the nudibranch Doris kerguelenensis (Dayton et al.,

1974). As such, these sponges act as a buffer to the intense seasonality experienced by

Antarctic fauna, and sponges in the deep-sea and tropical latitudes may be a key link in energy transfer for those food webs as well.

CONCLUSIONS

The diets of four species of sponges were described using three separate methods, each of which indicated species-specific diets. This is the first study to find niche partitioning in diet between species of sponges that occur at the same location. The application of fatty acid analysis to address sponge diet was utilized for the first time. A similar application on a larger selection of fauna would increase our knowledge of energy flow in the Antarctic as well as the role of the microbial loop in the benthic filter feeding food web.

LITURATURE CITED

Archer, S.D., R.J.G. Leacky, P.H. Burkill, M.A. Sleigh, and C.J. Appleby. 1996. Microbial ecology of sea ice at a coastal Antarctic site: Community composition, biomass and temporal change. Mar. Ecol. Prog. Ser. 135:179-195.

Arntz, W.E., T. Brey, and V.A. Gallardo. 1994. Antarctic Zoobenthos. Oceano. Mar. Biol. Ann. Rev. 32:241-304.

Azam, F. 1998. Microbial control of oceanic carbon flux: the plot thickens. Sci.280:694- 696.

Azam, F., T. Fenchel, J.G. Field, J.S. Gray, L.A. Meyer-Reil, and F. Thingstad. 1983. The ecological role of water-column microbes in the sea. Mar. Ecol. Prog. Ser. 10:257-263.

Azam, F., D.C. Smith, G.F. Steward, and A Hagstrom. 1994. Bacteria-organic matter coupling and its significance for oceanic carbon cycling. Microb. Ecol. 28:167- 179.

Bak, R.P.M., M. Joenje, I. de Jong, D.Y.M. Lambrechts, and G. Nieuwland. 1998. Bacterial suspension feeding by coral reef benthic organisms. Mar. Ecol. Prog. Ser. 175:285-288.

Barthel, D. 1997. Fish eggs and pentacridnoids in the Weddell Sea hexactinellids: further examples for the structuring role of sponges in Antarctic benthic ecosystems. Pol. Biol. 17:91-94.

Becquevort, S., P. Menon, and C. Lancelot. 2000. Differences of the protozoan biomass and grazing during spring and summer in the Indian sector of the southern ocean. Pol. Biol. 23:309-320.

Bowman, J.P., J.J. Gosink, S.A. McCammon, T.E. Lewis, D.S. Nichols, P.D. Nichols, J.H. Skerratt, J.T. Staley, and T.A. McMeekin. 1998. Colwellia demingiae sp. nov., Colwellia hornerae sp. nov., Colwellia rossensis sp. nov. and Colwellia psychrotropica sp. nov.: psychrophilic Antarctic species with the ability to synthesize docosahexaenoic acid (22:6 omega 3). I. J. Systematic Bacteriol. 48:1171-1180.

Burnett, W.C. and O.A. Schaeffer. 1980. Effect of ocean dumping on the 13C/12C ratios in marine sediments from the New York Bight. Est. Coastal. Mar. Sci. II 605-611.

Carey, A.G. 1987. Particle flux beneath fast ice in the shallow southwestern Beaufort Sea, Arctic Ocean. Mar. Ecol. Prog. Ser. 40:247-257.

Carlalleira, N.M. and M. Pagan. 2001. New methoxylated fatty acids from the Caribbean sponge Callyspongia fallax. J. Nad. Prod. 64:620-623.

Carlalleira, N.M., R. Colon, and A. Emiliano. 1998. identification of 2- methoxyhexadecanoic acid in Amphimedon compressa. J. Nat. Prod. 61:675-676.

Church, M.J., E.F. DeLong, H.W. Ducklow, M.B. Karner, C.M. Preston, and D.M. Karl. 2003. Abundance and distribution of planktonic Archaea and Bacteria in the waters west of the Antarctic Peninsula. Limnol. Oceanogr. 48:1893-1902.

Clarke, A. and R.J.G. Leakey. 1996. The seasonal cycle of phytoplankton, macronutirents, and the microbial community in a near shore Antarctic marine ecosystem. Limn. Oceanogr. 41:1281-1294.

Claus, G., P. Madri, and S. Kunen. 1967. Removal of microbial pollutants from waste effluents by the redbeard sponge. Nat. 216:712-714.

Conlan, K.E., G.H. Rau, G.A. McFeters, and R.G. Kvitek. 2000. Influence of McMurdo Station Sewage on Antarctic marine benthos: Evidence from stable isotopes, bacteria, and biotic indices. In: Davison W., C. Howard-Williams and P. Broady (eds.), Antarctic Ecosystems: Models for Wider Ecological Understanding. New Zealand Natural Sciences, Caxton Press, Christchurch, New Zealand. pp 315-318.

Crowder, M.J. and D.J. Hand. 1990. Analysis of Repeated Measures, Chapman and Hall, London. p 257.

Dayton, P.K. 1979. Observation of growth, dispersal and population dynamics of some sponges in McMurdo Sound, Antarctica. In: Levi C. and N. Boury-Esnault (eds.) Sponge Biology. Centre National de la Recherche Scientifique, Paris. pp. 271- 288.

Dayton, P.K. 1989. Interdecadal variation in an Antarctic sponge and its predators from oceanographic climate shifts. Sci. 243:1484-1486.

Dayton, P.K. 1990. Polar benthos. In: Polar Oceanography, Part B: Chemistry, Biology, and Geology, Academic Press. pp. 631-685.

Dayton, P.K. and J.S. Oliver. 1977. Antarctic soft-bottom benthos in oligotrophic and eutrophic environments. Sci. 197:55-58.

34 Dayton, P.K., W.A. Newman, R.T. Paine, and L.B. Dayton. 1974. Ecological accommodation in the benthic community at McMurdo Sound, Antarctica. Ecol. Monogr. 44:105-128.

Dembitsky, V.M., T. Rezanka, and A.G. Kashin. 1994. Comparative study of the endemic freshwater fauna of Lake Baikal. 6. Unusual fatty acid and lipid composition of the endemic sponge Lubomirskia baicalensis and its amphipod crustacean parasite Brandtia (Spinacanthus) parisitica. Comp. Biochem. Physiol. 109B:415-426.

Ducklow, H., C. Carlson, M. Church, D. Kirckman, D. Smith, and G. Steward. 2001. The seasonal development of the bacterioplankton bloom in the Ross Sea , Antarctica, 1994-1997. Deep-Sea Res. II 48:4199-4221.

Dunton, K.H. 2001. δ15N and δ13C measurements of Antarctic peninsula fauna: trophic relationships and assimilation of benthic seaweeds. Amer. Zool. 41:99-112.

Edwards, D.D., G.A. McFeters, and M.I. Venkatesan. 1998. Distribution of Clostridium perfringens and fecal sterols in a benthic coastal marine environment influenced by the sewage outfall from McMurdo Station, Antarctica. Appl. Environ. Microbiol. 64:2596-2600.

Gannes, L.Z., D.M. O’Brien, and C. Martinez del Rio. 1997. Stable isotopes in animal ecology: assumptions, caveats, and a call for more laboratory experiments. Ecol. 78:1271-1276.

Gibson, J.A.E., T. Trull, P.D. Nichols, R.E. Summons, and A. McMinn. 1999. Sedimentation of 13C-rich organic matter from Antarctic sea-ice algae: A potential indicator of past sea-ice extent. Geol. 27:331-334.

Gili, J-M., R. Coma, C. Orejas, P.J. Lopez-Gonzalez, and M. Zabala. 2001. Are Antarctic suspension-feeding communities different from those elsewhere in the world. Pol. Biol. 24:473-485.

Graeve, M., G. Katner, and W. Hagen. 1994. Diet induced changes in the fatty acid composition of Arctic herbivorous copepods: experimental evidence of trophic markers. J. Exp. Mar. Biol. Ecol. 182:97-110.

Graeve, M., P. Dauby, and Y. Scailteur. 2001. Combined lipid, fatty acid, and digestive tract content analyses: a penetrating approach to estimate feeding modes of Antarctic amphipods. Pol. Biol. 24: 853-862.

35 Gutt, J., A. Starmans, and G. Dieckmann. 1998. Phytodetritus deposited on the Antarctic shelf and upper slope: its relevance for the benthic system. J. Mar Res. 17:435- 444.

Hahn, S., I.L. Stoilov, T.B. Tam Ha, D. Raederstorff, G.A. Doss, H. Li, and C. Djerassi. 1988. Biosynthetic studies of marine lipids. 17.1 The course of chain elongation and desaturation in long-chain fatty acids of marine sponges. J. Am Chem. Soc. 110:8117-8124.

Jefferies, P.F. 1972. Fatty acid ecology of tidal marsh. Limnol. Oceanogr. 17:433-440.

Johnston, I.S. and W.H. Hildemann. 1982. Cellular organization in the marine demosponge Callyspongia diffusa. Mar. Biol. 67:1-7.

Kaehler, S., E.A. Pakhomov, and C.D. McQuaid. 2000. Trophic structure of the marine food web at the Prince Edward Islands (Southern Ocean) determined by δ13C and δ15N analysis. Mar. Ecol. Prog. Ser. 208:13-20.

Kattner, G. and U.H. Brockmann. 1990. Particulate and dissolved fatty acids in an enclosure containing a unialgal Skeletonema costatum (Greve.) Cleveculture. J. Exp. Mar. Biol. Ecol. 141:1-13.

Kowalke, J. 1999. Filtration in Antarctic ascidians-striking a balance. J. Exp. Mar. Biol. Ecol. 242:233-244.

Kowalke, J. 2000. Ecology and energetics of two Antarctic sponges. J. Exp. Mar. Biol. Ecol. 247:85-97.

Kowalke, J., M. Tatian, R. Sahade, and W. Arntz. 2001. Production and respiration of Antarctic ascidians. Pol. Biol. 24:663-669.

Lajtha, K. and R. Michener. 1994. Stable isotopes in Ecology and Environmental Science. Blackwell Scientific Publishing, New York. 336p.

Landry, M.R. 2002. Integrating classical and microbial good web concepts: evolving views from the open-ocean tropical Pacific. Hydrobiologia 480:29-39.

Landry, M.R. and D.L. Kirchman. 2002. Microbial community structure and variability in the tropical Pacific. Deep-Sea Res. II 49:2669-2693.

Lawson, M.P., P.R. Berquist, and R.C. Cambie. 1986. The cellular localization of long chain fatty acids in sponges. Tiss. Cell. 18:19-26.

36 Legendre, L. and J. Le Fevre. 1995. Microbial food webs and the export of biogenic carbon in the oceans. Aquat Microb Ecol. 142:3-17.

Macko, S.A., W.Y. Lee, and P.L. Parker. 1982. Nitrogen and carbon isotope fractionation by two species of marine amphipods: laboratory and field studies. J. Exp. Mar. Biol. Ecol. 63:145-149.

Madri, P.P., C. Claus, S.M. Kunen, and E.E. Moss. 1967. Preliminary studies on the Escherichia coli uptake of the red-beard sponge (Microciona prolifera Verrill). Life Sci. 6:889-894.

McClintock, J.B., and J.J. Gauthier. 1994. Antimicrobial activities of Antarctic sponges. Antarct. Sci.. 4:179-183.

Manahan, D.T. 1990. Adaptations by invertebrate larvae for nutrient acquisition from seawater. Amer. Zool. 30:147-160.

Meziane, T., L. Bodineau, C. Retiere, and G. Thoumelin. 1997. The use of lipid markers to define sources of organic matter in sediment and food web of the intertidal salt- marsh-flat ecosystem of Mont-Saint-Michel Bay, France. J. Sea. Res. 38:47-58.

Nichols, D.S., P.D. Nichols, and C.W. Sullivan. 1993. Fatty acid, sterol and hydrocarbon composition of Antarctic sea ice diatom communities during the spring bloom in McMurdo Sound. Antarc. Sci. 5:271-278.

Nichols, D.S. and T.A. McMeekin. 2002. Biomarker techniques to screen for bacteria that produce polyunsaturated fatty acids. J. Microbiol. Met. 48:161-170.

Nyssen, F., T. Brey, G. Lepoint, J. Bouquegneau, C. De Broyer, and P. Dauby. 2002. A stable isotope approach to the eastern Weddell Sea trophic web: focus on benthic amphipods. Pol. Biol. 25:28-287.

Orejas, C. 2001. Role of benthic cnidarians in energy transfer processes in the southern ocean marine ecosystem (Antarctica). Ber. Polarforsch. Meeresforsch. 395:1-186.

Orejas, C., J. M. Gili, P.J. Lopez-Gonzalez, and W.E. Arntz. 2001. Feeding strategies and diet composition of four Antarctic cnidarian species. Polar. Biol. 24:620-627.

Paine, R.T. 1964. Food web complexity and species diversity. Am. Naturalist 100:65-75.

Peterson, B.J. 1999. Stable isotope as tracers of organic matter input and transfer in benthic food webs: A review. Acta Oecologica 20:479-487.

Phleger, C.F., P.D. Nichols, and P. Virtue. 1998. Lipids and trophodynamics of Antarctic zooplankton. Comp. Biochem. Physiol. B 120:311-323.

Pile, A.J., M.R. Patterson, M. Savarese, V.I. Chernykh, and V.A. Fialkov. 1997. Trophic effects of sponge feeding within Lake Baikal’s littoral zone. 2. Sponge abundance, diet, feeding efficiency, and carbon flux. Limnol. Oceanogr. 42:178- 184.

Rau, G.H. J-L. Teyssie, F Rassoulzadegan, and S.W. Fowler, 1990. 13C/12C and 15N/15N variations among size-fractionated marine particles: Implications for their origin and trophic relationships. Mar. Ecol. Prog. Ser. 59:33-38.

Reiswig, H.M. 1971. Particle feeding in natural populations of three marine desmosponges. Biol. Bull. 141:568-591.

Reiswig, H.M. 1974. Water transport, respiration and energetics of three tropical marine sponges. J. Exp. Mar. Biol. Ecol. 14:231-249.

Reiswig, H.M. 1975. Bacteria as food for temperate-water marine sponges. Can. J. Zool. 53:582-589.

Reiswig, H.M. 1990. In Situ feeding in two shallow-water hexactinellid sponges. In: Rutzler, K. (ed): New Perspectives in Sponge Biology. Smithsonian Institute. 504-510.

Rieley, G., C.L. Van Dover, and G. Eglinton. 1997. Fatty Acids as sensitive tracers of sewage sludge carbon in a deep-sea ecosystem. Environ. Sci. Technol. 31:1018- 1023.

Riisgard, H.U., S. Thomassen, H. Jakobsen, J.M. Weeks, and P.S. Larsen. 1993. Suspension feeding in marine sponges Halichondria panacea and Haliclona urceolus: effects of temperature on filtration rate and energy cost of pumping. Mar. Ecol. Prog. Ser. 96:177-188.

Rivkin, R.B., M.R. Anderson, and D.E. Gustafson. 1991. Ingestion of phytoplankton and bacterioplankton by polar and temperate echinoderm larvae. Ant. J. U.S. 156-158.

Sara, M., G. Bavestrello, R. Cattaneo-Vietti, and C. Cerrano. 1998. Endosymbiosis in sponges: relevance for epigenisis and evolution. Symbiosis 25:57-70.

Sargent, J.R. and K.J. Whittle. 1981. Lipids and hydrocarbons in the marine food web. In: Longburst (ed) Analysis of Marine Ecosystems. Academic Press, London. pp. 491-533.

Sherr, B. and E. Sherr. 1994. Bactivory and herbivory: key roles of phagotrphic portists in pelagic food webs. Microb. Ecol. 28:223-236.

38 Sherr, B., E. Sherr, and P. del Giorgio. 2001. Enumeration of total and highly active bacteria. In. Paul J. (ed.), Methods in Microbiology, Volume 30: Marine Microbiology, pp.129-135.

Shilling, F.M. and I. Bosch. 1994. ‘Pre-feeding’ embryos of Antarctic and temperate echinoderms use dissolved organic material for growth and metabolic needs. Mar. Ecol. Prog. Ser. 109:173-181.

Smetacek, V., R. Scharek, and E.M. Nothig. 1990. Seasonal and regional variation in the pelagial and its relationship to the life history cycle of krill. In: Kerny R. and G. Hempel (eds.), Antarctic Ecosystems: Ecological Change and Conservations. Springer, Berlin pp.103-114.

Starmans, A., J. Gutt, and W.E. Arntz. 1999. Mega-epibenthic communities in Arctic and Antarctic shelf areas. Mar. Biol. 135:269-280.

Stewart, V. and D.W. Klump. 1984. Evidence for food-resource partitioning by kelp-bed filter feeders. Mar. Ecol. Prog. Ser. 16:27-37.

Stoecher, D.K., K.R. Buck, and M. Putt. 1992. Changes in the sea-ice brine community during the spring-summer transition, McMurdo Sound, Antarctica. I. Photosynthetic protists. Mar. Ecol. Prog. Ser. 84:265-278.

Stoecher, D.K., K.R. Buck, and M. Putt. 1993. Changes in the sea-ice brine community during the spring-summer transition, McMurdo Sound, Antarctica. Phagotrophic protists. Mar. Ecol. Prog. Ser. 95:103-113.

Stoecker, D.K., M. Putt, and T. Moisan. 1995. Nano- and Microplankton dynamics during the spring Phaeocystis sp. bloom in McMurdo Sound, Antarctica, Mar. Biol. Ass. U.K. 75:812-832.

Van De Vyver, G.V., B. Vray, S. Balaouane, and D. Toussanint. 1990. Efficiency and selectivity of microorganism retention by Ephydatia fluviatilis. In: Rutzler, K. (ed): New Perspectives in Sponge Biology. Smithsonian Institute. 511-515.

Van Dover, C.L., J.F. Grassle, B. Fry, R.H. Garritt, and V.R. Starczak. 1992. Stable isotopic evidence for entry of sewage derived organic material into a deep-sea food web. Nat. 360:153-156.

Vargas, C.A. and H.E. Gonzalez. 2004. Plankton community stricture and carbon cycling in a coastal upwelling system. I. Bacteria, microprotozoans and phytoplankton in the diet of copepods and appendicularians. Aqaut. Microb. Ecol. 34:151-164.

39 Volkman, J.K., R.B. Johns, F.T. Gillan, G.J. Perry, and H.J. Bavour. 1980. Microbial lipids of an intertidal sediment 1. Fatty acids and hydrocarbons. Geochem. Cosmochim. Acta 44:1133-1143.

Vosjan, J.H. and G.J. van Noort. 1998. Enumerating nucleoid-visible marine bacterioplankton: bacterial abundance determined after storage of formalin fixed samples agrees with isopropanol rinsing method. Aquat. Microb. Ecol. 14:149- 154.

Wada, E., M. Terazaki, Y. Kabaya and T. Nemoto. 1987. 15N and 13C abundance in the Antarctic ocean with emphasis on the biogeochemical structure of the food web. Deep-Sea Res. 34:829-841.

Wagele, H. 1989. Diet of some Antarctic nudibranchs ,Gastropoda, Opisthobranchia, Nudibranchia. Mar. Biol. 100:439-441.

Wakeham, S.G., J.I. Hedges, C. Lee, M.L. Peterson, and P.J. Hernes. 1997. Composition and transport of lipid biomarkers through the water column and surficial sediments of the equatorial Pacific Ocean. Deep-Sea Res. II 44:2131-2162.

Watanabe, T., C.S. Katajima, and S. Fujita. 1983. Nutritional values of live organisms used in Japan for mass propagation of fish: a review. Aquaculture 34:115-134.

Wilkinson, C.R. 1983. Net primary productivity in coral reef sponges. Science 219:410- 412.

Yano, Y., A. Nakayama, and K. Yashida. 1997. Distribution of polyunsaturated fatty acids in bacteria present in the intestines of deep-sea fish and shallow-sea poikilothermic animals. Appl. Environ. Microbiol. 63:2572-2577.

Zanden, M.J. and J.B. Rasmussen. 2001. Variation in δ15N and δ13C trophic fractionation: Implications for aquatic food web studies. Limnol. Oceanogr. 46:2061-2066.

Zhukova, N.V. and V.L. Kharlamenko. 1999. Sources of essential fatty acids in the marine microbial loop. Aquat. Microb. Ecol. 17:153-157.

40 TABLES Table 1: The distribution of fatty acids, viable fecal coliform bacteria, and total bacteria as enumerated by epifluorescence microscopy at the given distances from the McMurdo Station outfall, Antarctica, listed from North to South. (---) indicates bellow detectable limit.

Distance from Outfall Fatty Acid -700 0 115 166 410 840 9.0 3.36 1.51 15.36 1.5 ------10.0 1.98 ------1.37 ------12.0 3.4 7.15 7.89 3.07 ------14.0 6.93 10.51 10.97 5.6 14.03 --- 15.0 ------2.04 ------16.0 24.63 28.62 36.81 29.37 44.26 43.76 18.0 15.13 17.68 28.97 18.13 30.33 38.9 16:1 (n-7)c 10.97 5.3 --- 3.3 ------16:1 (n-8)c ------2.4 ------18:1 (n-9)c 8.84 18.28 --- 10.87 ------18:1 (n-9)t 3.71 5.48 --- 2.95 ------18:2 (n-6)c 3.41 5.03 --- 2.48 ------18:3 (n-6)c ------1.77 ------22:1 (n-7)c/22:3 (n-3)c 3.09 ------24:1 (n-3)c --- 0.44 ------24:1 (n-9)c Nervonic 6 ------C14 N Alcohol ------6.5 11.38 --- C16 N Alcohol ------2.71 ------C25 N Alcohol 4.81 ------2.24 ------ISO 17:1 (n-5)c ------1.98 ------Unknown 20.588 3.76 ------17.34 Unknown 25.052 ------1.72 ------

Total Bacteria Average Microbial Abundance (cells ml-1) 3.1x104 3.3x104 3.6x104 3.2x104 3.4x104 4.1x104 St Error 2.4x103 6.8x102 8.9x102 1.6x103 2.3x103 1.1x104

Fecal Bacteria Coliform abundance (cells per 100ml) 1.3 13.5 0.8 1.5 0.0 0.3

42 Table 2: Statistical comparison between a selection of fatty acid biomarker concentrations in four species of sponge. Nested ANOVAs compare between species and linear regressions indicate differences in concentration with increasing distance from the McMurdo Station sewage outfall. Post hoc results are indicated on Figures 2,3, and 5.

Nested ANOVA Regression analysis FA Species Station (species) Species Coefficient x 104 P R2 16:1 (n-7) F3,69= 51.5 p<0.01 F15,69= 1.9 p=0.04 H. balfourensis 0.163 0.30 0.05 I. setifera 1.065 0.01 0.34 K. variolosa 0.473 0.10 0.13 S. antarcticus 0.109 0.66 0.01 18:1 (n-9) F3,69= 25.7 p<0.01 F15,69= 12.9 p=0.01 H. balfourensis -2.395 <0.01 0.69 I. setifera -2.062 <0.01 0.48 K. variolosa -2.415 <0.01 0.50 S. antarcticus 0.601 <0.01 0.28 18:2 (n-6) F3,69= 13.9 p<0.01 F15,69= 15.5 p<0.01 H. balfourensis -2.542 <0.01 0.66 I. setifera -1.544 <0.01 0.68 K. variolosa -1.158 <0.01 0.31 S. antarcticus -0.775 <0.01 0.35 20:5 (n-3) F3,69= 5.8 p<0.01 F15,69= 3.5 p<0.01 H. balfourensis 0.224 0.74 0.10 I. setifera 1.158 0.07 0.19 K. variolosa 0.455 0.34 0.05 S. antarcticus -0.517 0.50 0.02 22:6 (n-3) F3,69= 109.9 p<0.01 F15,69= 2.4 p=0.01 H. balfourensis 1.224 <0.01 0.34 I. setifera 0.761 0.24 0.08 K. variolosa 1.310 0.03 0.23 S. antarcticus 0.304 0.23 0.07

43 Table 3: Statistical comparison between groups of fatty acid biomarker concentrations in four species of sponge. Nested ANOVAs compare between species and linear regressions indicate differences in concentration with increasing distance from the McMurdo Station sewage outfall. Post hoc results are indicated on Figures 4 and 5.

Nested ANOVA Regression analysis FA Species Station (species) Species Coefficient x 104 P R2 Bacteria to F3,69= 85.6 p=0.01 F15,69= 5.5 p<0.01 H. balfourensis -27.415 0.19 0.07 PUFA I. setifera -11.841 0.44 0.04 Ratio K. variolosa 4.307 0.02 0.26 S. antarcticus -3.114 0.03 0.22 Bacteria F3,69= 99.2 p<0.01 F15,69= 5.37 p<0.01 H. balfourensis -0.947 0.18 0.08 Fatty I. setifera 0.195 0.55 0.02 Acids K. variolosa 0.313 0.44 0.03 S. antarcticus -0.277 0.40 0.04 PUFA F3,69= 61.5 p<0.01 F15,69= 3.0 p<0.01 H. balfourensis 0.388 0.44 0.03 I. setifera 0.734 0.22 0.09 K. variolosa 1.272 0.02 0.24 S. antarcticus -0.041 0.91 0.00

44 Table 4: Statistical comparison between the stable isotopic concentrations of four species of sponge. Nested ANOVAs compare between species and linear regressions indicate differences in concentration with increasing distance from the McMurdo Station sewage outfall.

Isotope Nested ANOVA Regression analysis Species Station (Species) Species Coefficient x 104 P R2 15 δ N F3,67= 89.0 p<0.01 F16,67= 0.38 p=.98 H. balfourensis 0.217 0.84 0.00 I. setifera -0.128 0.75 0.00 K. variolosa -1.408 0.33 0.05 S. antarcticus -0.283 0.89 0.00 13 δ C F3,67= 24.9 p<0.01 F16,67= 8.25 p<0.01 H. balfourensis -8.219 <0.01 0.40 I. setifera -2.837 0.16 0.10 K. variolosa -7.748 0.02 0.21 S. antarcticus -11.657 0.04 0.20

45 FIGURES

2 12 10 8 6 4 2 0 270o 0 0 90o 024681012 2

180o

Figure 1: Locations of sponge and water collections at McMurdo Station, Antarctica. Current speed (cm s-1) and direction during the time of water collection is shown in the vector diagram as recorded by an S4 current meter.

Figure 6: Principal Component Analysis (PCA) plot of the fatty acid constituents of four species of sponge at varying distances from the McMurdo Station sewage outfall.

52 Figure 7a: Bi-plot of the eigenvectors that accounted for more than 10% of the variability in principal components one and two. 53 Figure 7b: Bi-plot of the eigenvectors that accounted for more than 10% of the variability in principal components two and three.

55 Figure 9: Percent decrease of total bacterial abundance over time in a closed system containing the indicated species of sponge or a control of nothing. n=5 and error bars are standard error.