Sublethal Effects of Contamination on the Mediterranean Sponge Crambe

Marine Pollution Bulletin 46 (2003) 1273–1284 www.elsevier.com/locate/marpolbul

Sublethal eﬀects of contamination on the Mediterranean sponge Crambe crambe: metal accumulation and biological responses

E. Cebrian a,*, R. Martı a, J.M. Uriz a, X. Turon b

a Centre d’Estudis Avancßats de Blanes, CSIC, C/ Acces a la Cala St. Francesc, 14, 17300 Blanes, (Girona), Spain b Departament d’Invertebrats, Universitat de Barcelona, Avda. Diagonal 645, 08028 Barcelona, Spain

Abstract

The effect of low levels of pollution on the growth, reproduction output, morphology and survival of adult sponges and settlers of the sponge Crambe crambe were examined. We transplanted sponges from a control area to a contaminated site and measured the main environmental variables (chemical and physical) of both sites during the study period. Except some punctual differences in particulate organic matter, silicates, nitrates, and water motion, most environmental variables in the water were similar at both sites during the study months. Mainly copper, lead and OM concentrations in the sediment, and water motion were significantly higher at the polluted site and may be implicated in the biological effects observed: decrease in the percentage of specimens with embryos, increase in shape irregularity and decrease in growth rate. Individuals naturally occurring at the polluted site and those transplanted there for four months accumulated ten times more copper than either untouched or transplant controls. Although lead concentration in sediment did not differ between sites, native specimens from the contaminated site accumulated this metal more than untouched controls. Vanadium concentration also tended to increase in the sponges living at or transplanted to the contaminated site but this difference was not significant. C. crambe is a reliable indicator of metal contamination since it accumulates copper, lead and vanadium in high amounts. At the contaminated site, sponge growth, fecundity and survival were inhibited, whereas sponge irregularity ending in sponge fission was promoted. All these effects may compromise the structure and dynamics of the sponge populations in sheltered, metal-contaminated habitats. 2003 Elsevier Ltd. All rights reserved.

Keywords: Mediterranean; Sublethal contamination; Sponges; Crambe crambe; Heavy metals; Copper

1. Introduction than experimental, and contamination impacts are usually assessed from observed mortality or changes in Many studies on the effects of contaminants on the community structure, which are measured by structure aquatic environment have been performed in the lab- descriptors such as diversity indices, species richness, oratory using target organisms easy to maintain in lab- species abundance or presence/absence of indicator spe- oratory conditions (e.g. Brown and Ahsnullah, 1971; cies (Carman et al., 1995; Carballo and Naranjo, 2002). Kobayashi, 1980; Rainbow et al., 1980; Rainbow and When changes in the structure of benthic communi- Wang, 2001). Laboratory studies involve the exposure of ties are detected, we are restricted to measuring the le- organisms to well-defined concentrations of one or a few thal effects of pollution. However, most contaminants contaminants in controlled environmental conditions. enter marine waters as low-level chronic toxicants whose Unfortunately, it is difficult to predict the responses to harmful effects are not always obvious and take time to toxicants in the field from these laboratory experiments, appear (Young et al., 1979). Small doses of pollutants since they do not take into account a variety of complex affect the physiological functions and behaviour of or- environmental interactions (Cairns and Pratt, 1989). On ganisms without killing them (e.g. Newton and the other hand, field studies tend to be descriptive rather McKenzie, 1995; Agell et al., 2001). Yet, these cryptic effects compromise not only individual fitness but also population success. Most of these studies have been * Corresponding author. carried out on soft bottom invertebrates (e.g. Ozoh, E-mail address: [email protected] (E. Cebrian). 1990; Warwick et al., 1990) or plankton species

(e.g. Brown and Ahsnullah, 1971; Riisgaard, 1979) and sisted of a vertical concrete wall from 0 to 5 m deep with scarce data are available on benthic filter feeders other similar facing and depth to control site. than mussels and clams (Bjerregaard and Depledge, From previous studies (Pinedo, 1998) organic matter, 1994; Abbe et al., 2000). Thus, the sublethal effects of most trace metals, TBT, and THC in the sediment at the contamination on many invertebrates, which dominate polluted site were no present at concentrations high most rocky assemblages, are unknown. enough to be considered pollutants. In contrast, copper Sponges are among the main constituents of sciaph- was at a mean concentration of 97 lggÀ1 at the polluted ilous, rocky communities in the Mediterranean benthos site (versus 6 lggÀ1 at the control site) and can be (Vacelet, 1979; Urizet al., 1992a,b), where they play a considered a contaminant, according to UNEP policies. paramount role in the energy transfer processes (Rei- Contamination in this harbour was expected to be due swig, 1971). According to the rare studies on sponge mainly to cleaning of small shipsÕ hulls, and urban responses to pollution (Alcolado and Herrera, 1987; sewage. Carballo et al., 1996; Perez, 2001), sponges are either resistant or susceptible to contaminants depending on 2.2. Environmental variables at the study sites the contaminant and the species considered. Here we examine the effect of low levels of pollution 2.2.1. Chemical parameters on the growth, reproduction output, morphology and During whole experiment period water was collected survival of adult sponges and settlers. These effects may weekly with a Niskins sampler from a depth of 3–4 m, at alter the species fitness and so its population dynamics, eight sampling points (four at the control site and four which will affect whole benthic assemblages. at the polluted site). We transplanted sponges from a control area to a For the analysis of particulate organic matter (or- contaminated site and measured the main environmen- ganic carbon and nitrogen), a 2 l sample of seawater was tal variables (chemical and physical) of both sites during screened through a 100 lm pore net to remove large the study period. Scattered individuals of Crambe plankton forms and detritus. Water was then passed crambe inhabited the contaminated site selected, which through a 0.22 lm diameter, GF/F glass fibre filter, allows us to state that the levels of contamination of this previously exposed to hydrochloric acid vapour for 48 h site were not lethal for the sponge. We subjected the in order to eliminate any inorganic material. Filters sponges transplanted to the same environmental condi- containing the organic matter were dried and analysed tions affecting the natural populations at both the con- with a C:H:N autoanalyser Eager 200. trol and contaminated areas for a four-month period. Filtered water was used to analyse nutrient salts, Since early life stages of benthic invertebrates are usually polycyclic aromatic hydrocarbons (PAHs) and heavy more susceptible to contamination than adult stages metals. Nutrient salts (nitrites, nitrates, phosphates and (Kobayashi, 1980), we also studied the survival rates of silicates) were analysed by colorimetric techniques settlers transplanted to both sites. (autoanalyser Technicon). Total PAHs were determined by gas chromatography coupled to mass spectrometry in the electron impact mode (GC-MS-EI) using a Fison 2. Material and methods GC8000 series chromatograph interfaced to a Fison MD 800 mass spectrometer (Sole et al., 2000). 2.1. Organism and study site 2.2.2. Sedimentation We studied the sponge C. crambe, a widespread sub- Gross sedimentation rates were assessed by placing littoral sponge in western Mediterranean (Vacelet, 1979; sediment traps, placed monthly during the experiment Urizet al., 1992a), which is well known from a biological (from February to June 1999) ðN ¼ 5Þ, at each study and ecological perspective (Becerro et al., 1994; Turon site. Traps remained in the water for three days. Dried et al., 1998; Urizet al., 1995). The encrusting growth sediment was weighed and then combusted at 300 C for habit of this sponge allowed us to determine sponge 48 h and the mineral residue was weighed. The organic growth and shape by measuring increases in area and matter was calculated by subtracting the mineral com- the perimeter/area ratio, respectively. Furthermore, C. ponent of the sediment from the total dry weight. To crambe also lives at some polluted sites, what makes it search for possible differences in granulometry, a con- suitable for studies of sublethal contamination. stant amount of sediment was suspended in a constant The study was carried out at the Blanes sublittoral volume of water, and an aliquot (20 ll) of this suspen- (NE Iberian Peninsula, western Mediterranean) (NE of sion was placed in a haemocytometer and examined Spain, 41 40.40 N, 2 48.20 E). The control site was a under a light microscope. Microscope fields ðN ¼ 3Þ, vertical rocky wall, from 2 to 10 m deep, facing west. selected at random, were captured in a digital video The polluted site, 500 m from the control, was on the camera connected to a computer and the images were inner side of the Blanes harbour breakwater and con- used for quantification of particle size-classes with the E. Cebrian et al. / Marine Pollution Bulletin 46 (2003) 1273–1284 1275

NIH Image program (public domain). The sediment was 2.4.2. Sponge growth, shape and survival classed in very fine (VF) (diameter <12 lm), and fine (F) Sponge growth, shape and survival of transplanted (between 12 and 30 lm in diameter), which can enter the individuals were assessed from monthly photographs sponge through its orifices, and medium (M) (between from which perimeter and area were calculated with 30 and 100 lm in diameter) and coarse (C) (diameter > NIH image program for Macintosh. 100 lm), which cannot enter the sponge via ostia (Gal- Growth was estimated from changes in area over era et al., 2000). time. Since the sponge was thinly incrusting, changes in area are good estimates of changes in biomass (Turon 2.2.3. Water motion et al., 1998). A monthly growth rate GRm was computed Water motion was assessed in various sea conditions by the formula:

(calm; slightly rough; rough; stormy). These measure- GRm ¼ðAm À AmÀ1Þ=AmÀ1; ments were not intended to represent the general trend where A and A are the areas in the month m and in in water motion of the study sites but were used exclu- m mÀ1 the previous month, respectively. sively for comparative purposes. We submerged CaSO 4 Sponge shape was approached from the ratio between balls ðN ¼ 5Þ at the control and polluted sites for three perimeter and area, which is an estimation of the sponge days, and measured their weight losses (Muus, 1968). irregularity, and may increase in stress conditions and Results are then expressed in g hÀ1 of CaSO . 4 precedes sponge fission (Turon et al., 1998). Sponge fission was recorded monthly for trans- 2.2.4. Irradiance planted individuals at both sites. Incident irradiance was measured by a sensor Licor- SPQA at the same time of the day, on several days, at 2.4.3. Reproduction both the working depth (about 4 m) and the sub- The percentage of individuals undergoing reproduc- superficial level. tion at both sites was recorded at the end of the experiment (June), just before larval release. We sampled the 2.3. Transplant experiment natural populations and the transplanted sponges at both sites. The specimens were dissected and examined A total of 40 sponge specimens were randomly taken under a stereomicroscope since embryos were mainly with their substrate from the control site. Twenty of located at the sponge base. We recorded the percentage these specimens were transplanted in situ to test possible of sponges that harboured embryos. transplant effects (transplant control, TC). The remain- Previously (in March), to verify that the sponges ing 20 specimens were placed in separated hermetic transplanted to the contaminated site were undergoing bowls underwater and transported to the polluted site, gametogenesis, we took small pieces (3 mm in diameter) where they were transplanted within 1 h of collection and processed them for histological observation. Sam- (harbour transplant, HT). Twenty more specimens from ples were fixed in 10% formalin/seawater, desilicified in the control site (control, C) and 20 more from the pol- 5% HF, dehydrated and embedded in paraffin. 5-lm luted site (harbour, H) were randomly selected, labelled, thick sponge sections (Autocut 2040 microtome), and left untouched until the end of experiment. At the stained following the MassonÕs Trichrome technique end of June, before the sponge larval release (Urizet al., (Martoja and Martoja, 1970), were examined under the 1998), we collected the specimens transplanted, and light microscope for the presence of gametes. those native to the control and polluted sites, for heavy metal analyses and examination for the presence of 2.4.4. Settler survival embryos. Larvae were obtained in the laboratory from ripe sponges by spontaneous release or by shaking the 2.4. Sponge descriptors sponges (Urizet al., 1998). Free larvae settled succes- sively on Petri dishes (10 larvae per dish), at sea tem- 2.4.1. Metal concentration perature and natural photoperiod. Upon larvae In general, accumulation of heavy metals (Cu, Pb, settlement, the Petri dishes were transferred to the Cd, Hg and V) were quantified using an inductively control and polluted sites ðN ¼ 15Þ, and the number of coupled plasma mass spectrometer (ICP-MS) Perkin living settlers was subsequently recorded weekly over a Elmer, Elan 6000. When the amount of heavy metals month. exceeded the optimum range of concentrations for ICP- MS, we used an inductively coupled plasma optical 2.5. Data analysis emission spectrometer (Thermo Jarrell Ash, ICAP 61E). Results are expressed in lggÀ1 of metal with respect to Differences in irradiance and medium grain size of the sponge tissue (dry weight). sediment between sites were assessed by t-tests. 1276 E. Cebrian et al. / Marine Pollution Bulletin 46 (2003) 1273–1284

Differences in particulate organic matter (POM) and show significant differences between the control and soluble nutrients in the water, sedimentation rates, polluted sites during the study period but it varied dif- percentage of organic matter in the sediment and water ferently with time (significant interaction term) between motion between control and polluted sites were analysed sites (Table 1). PON ranged from 49 to 50 lgNlÀ1 (both by two-way ANOVAs. Metal accumulation within the sites) to 70 lgNlÀ1 (contaminated site). It also varied sponges was analysed by one-way ANOVA (Statistica differently with time at both localities, as revealed by a 4.1 package). The Tukey test was used for post hoc significant interaction between locality and time. The comparisons. Assumptions of normality and homo- t-test between sites for each time point indicated that geneity of variances were examined using the Kolmo- PON values were significantly higher at the contami- gorov–Smirnov and Barlett tests, respectively. Variables nated site only in May ðp < 0:01Þ. These differences did were rank-transformed (Conover and Iman, 1981; Po- not affect the C:N ratio, which followed the same trend twin et al., 1990) prior to the analysis when assumptions and varied significantly with time at both sites (Table 1). were not fulfilled. Differences in sponge growth rates, total area and the 3.1.3. Nutrient salts perimeter/area ratio were analysed by means of the two- As for the soluble nutrients, phosphates and nitrites level randomisation method based on Manly (1991) and varied with time (Table 1). However, while phosphates described in Turon et al. (1998) because data did not followed a parallel trend at both sites (no interaction meet the circularity assumption (MauchlyÕs sphericity term), nitrites varied differently with time at each site test) required by univariate and multivariate versions of (significant interaction term). Phosphate concentration repeated measures analysis of variance (Potvin et al., followed a similar trend at both sites, except in March 1990; Von Ende, 1993). The whole series of data was when values reached 0.25 lmol lÀ1 at the control site randomised 4999 times (plus the observed one) to ap- (Fig. 2A). Nitrites peaked in winter months (0.25–0.3 proximate the null hypothesis distribution of the sum of lmol lÀ1) and strongly decreased from April to May at squares for each factor and their interaction, and then both sites, reaching values as low as 0.16–0.18 lmol lÀ1 we examined how extreme were the observed values in in June (Fig. 2B). Silicate and nitrate concentrations this distribution. An effect was judged significant when followed parallel trend but significantly different at both the observed sum of squares was exceeded by less than localities and varied with time (Fig. 2C, D) (Table 1). 5% of the corresponding values in the randomisation Post hoc tests indicated that the significant differences series. between sites found in the ANOVA were exclusively due Differences in the frequency of individuals that in- to the higher values of nitrates in April ðp < 0:01Þ and cubated embryos between sites were analysed by means silicates in June ðp < 0:05Þ at the polluted site. of a 2 · 2 contingency table using the v2 statistic. The percentage of survivors for both settlers and 3.1.4. Hydrocarbons adult sponges at the polluted and control sites were Neither at the control nor at the polluted site were compared using GehanÕs Wilcoxon Test. polycyclic aromatic hydrocarbons (PAHs) detected in water at the depth (4 m) at which the experiment was conducted. 3. Results

3.1. Environmental parameters 3.1.5. Sedimentation Gross sedimentation rates (GSRs), defined as the 3.1.1. Light total amount of sediment per surface and time units The mean incident irradiance at the control site recovered in a sediment trap, are presented in Fig. 3A. (371 ± 26.04 lEmÀ2 sÀ1) (Mean ± SE) did not differ sig- They varied differently with time at both sites (Table 2). nificantly ðp ¼ 0:1123Þ from that at the polluted site T -tests between sites at each time point indicated that (439 ± 31.4 lEmÀ2 sÀ1). These mean values corre- sedimentation rates were higher at the control site in sponded to a percent surface light of 27% and 32%, re- March and April (p < 0:01 in both cases). spectively. The total amount of organic matter (Fig. 3B) varied significantly with time at both sites (Table 2). T -tests at 3.1.2. Organic matter each time point indicated that the values were higher at The time course of the particulate organic carbon the control site in March and April and at the polluted (POC), particulate organic nitrogen (PON) in the water, site in May and June ðp < 0:05Þ. and their ratio at both sites are shown in Fig. 1 and the The time course of the percentage of organic matter ANOVA results are listed in Table 1. POC concentra- in the sediment (Fig. 3C) was uniform and followed a tion ranged from 1300 to 1500 lgClÀ1 in February to parallel trend at both sites but it was significantly higher 2000–2100 lgClÀ1 in April (Fig. 1A). POC did not at the polluted site (Table 2). Post hoc analyses showed E. Cebrian et al. / Marine Pollution Bulletin 46 (2003) 1273–1284 1277

POC

2200 20 ( Temperature A 2000 16 1800 -1 12 1600 g·L

µ 8 1400 4 1200 ) 1000 0 February March April May June

PON 80 20 eprtr ( Temperature B 70 16

-1 12 60 g·L

µ 8 50 4 eprtr ( Temperature ) 40 0 February March April May June

C:N 50 20 45 C 16 40 12 -1 35

g·L 8

µ 30 25 4 ) 20 0 February March April May June

Control Site Polluted Site Temperature (Cº)

Fig. 1. Time course of mean POC (A) and PON (B) concentrations and the C:N ratio (C) during the experiment at the control (solid symbol) and polluted (empty symbol) sites. The time course of the seawater temperature is indicated by a shaded line. Vertical bars are standard errors. that the differences between sites were due to the values 3.2. Sponge descriptors of the first two months (Tukey test, p < 0:05 and p < 0:01, respectively). 3.2.1. Metal accumulation The medium grain size of the sediment was signifi- Copper, lead and vanadium were the only heavy cantly larger (t-test, p < 0:05) at the control site (Fig. 4). metals detected in the sponge tissues (Fig. 7). Copper Coarse sediment over 100 lm in diameter was only accumulation was significantly higher (Tukey test, present in the controls traps and never exceeded 5% of p < 0:05) in the specimens naturally occurring at the the total (Fig. 5). The medium-sized fraction (M) (30– polluted site (H) and in those transplanted there for four 100 lm in size) was about 20–30% in all the traps except months (HT) than in both the specimens from the in June at the contaminated site. The fine (F) fraction control site (untouched control, C) and those trans- (grain size 12–30 lm) was the main component of the planted in situ (transplant control, TC) (Table 4, Fig. sediment at both sites in all the sampled months except 7A). Lead concentration (Fig. 7B) was significantly in June at the polluted site, when the fraction smaller higher (Tukey-test, p < 0:05) in the specimens naturally than 12 lm (VF) accounted for most of the sediment occurring at the polluted site than in those at the control (Fig. 5). site. Although not statistically significant (Table 4), vanadium concentration also tended to be higher in the sponges native or transplanted to the contaminated site 3.1.6. Water motion (Fig. 7C). Water motion level was significantly higher at the control site (Fig. 6) and varied similarly at both sites (no 3.2.2. Growth interaction between site and time) depending on the sea At the beginning of the experiment, the sponge condition (Table 3). Nevertheless, no differences were mean area was nearly the same at both sites. There- found (Tukey-test, p > 0:05) between sites in very calm after, the area of the specimens transplanted to the and stormy sea conditions (Fig. 6). control site (TC) increased slightly, while that of the 1278 E. Cebrian et al. / Marine Pollution Bulletin 46 (2003) 1273–1284

Table 1 Phosphates Two-way ANOVAs for site (control and polluted sites) and time (four 0.35 20 A months) eﬀects on organic carbon (POC), particulate organic nitrogen 0.3 (PON), C:N ratio and dissolved nutrients (phosphates, nitrites, sili- 16 ( Temperature 0.25 cates and nitrates) 0.2 12 Variable Factor DF F p 0.15 POC (lglÀ1) Site 1 2.319 0.130 8 º Time 4 29.110 0.000 0.1 C) 4 Site and time 4 4.218 0.003 0.05 Error 134 0 0 PON (lglÀ1) Site 1 4.068 0.045 February March April May June Time 4 4.283 0.002 Nitrites 0.35 20 Site and time 4 30.668 0.018 B Error 134 0.3

16 Temperature C:N Site 1 2.391 0.124 0.25 Time 4 10.492 0.000 0.2 12 Site and time 4 0.745 0.562 0.15 8 (

Error 134 º 0.1 C) Phosphates Site 1 0.013 0.969 4 0.05 (lmol lÀ1) Time 4 8.970 0.000 Site and time 4 0.577 0.856 0 0 Error 153 February March April May June Silicates Nitrites Site 1 0.664 0.337 2.5 20 (lmol lÀ1) Time 4 88.532 0.000 C

Site and time 4 5.714 0.008 2 16 Temperature Error 153 1.5 12 Silicates Site 1 7.524 0.002 (lmol lÀ1) Time 4 29.522 0.000 1 8 ( º Site and time 4 0.728 0.380 C) Error 153 0.5 4

Nitrates Site 1 4.012 0.013 0 0 (lmol lÀ1) Time 4 4.658 0.000 February March April May June Site and time 4 1.027 0.846 Nitrates Error 153 3 20 D 2.5 16 Temperature 2 sponges transplanted to the contaminated site (HT) 12 decreased meaningfully (Fig. 8A, Table 5). The site 1.5 8 (

(polluted versus non-polluted sites) had a signiﬁcant 1 º C) eﬀect on the sponge monthly growth rate (GRm): while 0.5 4 growth rates of the individuals transplanted to the 0 0 control site (TC) were close to zero, growth rates of the February March April May June specimens transplanted to the polluted site were negative throughout experiment (Table 5, Fig. 8B). Month Control Site Polluted Site Temperature

Fig. 2. Time course of mean concentration for phosphates (A), nitrites 3.2.3. Shape (B), silicates (C) and nitrates (D) during the experiment period at the The site also had a significant effect on sponge ir- control (solid symbol) and polluted (empty symbol) sites. The time regularity, as measured by the perimeter/area ratio. This course of the seawater temperature is indicated by a shaded line. Vertical bars are standard errors. ratio was similar for individuals from both sites at the beginning of the experiment but increased significantly with time among the individuals transplanted to the site (1.94 ± 0.38) compared to the control site (0.15 ± polluted site (Table 5). Conversely, individuals trans- 0.08; Mean ± SE). planted to the control site maintained that ratio more or less constant. Transplantation per se did not enhance 3.2.4. Adults survival sponge irregularity (Fig. 8C). Survival in transplanted individuals was 90% (pol- Sponge fission was significantly enhanced (t-test, luted site) and 85% (control) during the first month of p < 0:001) in individuals transplanted to the polluted the experiment (Fig. 9). No decrease in survival was E. Cebrian et al. / Marine Pollution Bulletin 46 (2003) 1273–1284 1279

Gross sedimentation rate Table 2 80 20 Two-way ANOVAs for site and time eﬀects on gross sedimentation A rates (GSRs), organic matter, and the percentage of organic matter in

16 Temperature 60 the sediment

1 Variable Factor DF F p

d- 12 · 2 1

-2 À À 40 GSRs (g m d ) Site 1 0.257 0.000 g.m ( Time 3 24.161 0.855

8 º C) Site and time 3 5.961 0.003 20 4 Error 26 Organic matter Site 1 4.918 0.035 0 0 (g mÀ2 dÀ1) Time 3 2.453 0.085 March April May June Site and time 3 4.529 0.011 Total organic matter in the sediment Error 26 20 20 18 B Organic matter Site 1 6.397 0.017 16 16 (percentage) Time 3 2.264 0.104 Temperature 14 Site and time 3 2.133 0.120 12 12 Error 26 1 d-

· 10 -2

8 8 ( º C) g.m 6 4 4 detected from March to May at the control site, while 2 some of the sponges transplanted to the polluted site 0 0 died. Mortality increased in June among individuals March April May June transplanted to the contaminated site, in parallel with an Percent organic mater in the sediment increase in both water temperature and the amount of 100 20 C fine sediment (Fig. 5). Nevertheless, the differences in survival between sites were not significant (GehanÕs 80 16 Temperature Wilcoxon Test, p ¼ 0:086). 60 12 3.2.5. Reproduction

40 8 ( º Percentage %

C) The percentage of sponges harbouring embryos in June was lower among individuals transplanted (25%) 20 4 and naturally occurring (11.7%) to the polluted site than 0 0 in those untouched (82%) and transplanted (50%) to the March April May June control site (Fig. 10; v2, p < 0:01). Light microscope Control Site Polluted Site Temperature examination of paraﬃn sections of the transplanted sponges indicated that all of them underwent gameto- Fig. 3. Time course of gross sedimentation rates (A), total organic genesis in March. Thus, the low number of sponges matter (B) and percentage of organic matter (C) in the sediment during the experiment at the control (solid symbol) and polluted (empty harbouring embryos among those transplanted to the symbol) sites. The time course of the seawater temperature is indicated polluted site is due to causes other than the absence of by a shaded line. Vertical bars are standard errors. initial reproduction.

Fig. 4. Mean medium grain-size (lm) at the control and polluted sites. Boxes represent standard errors; vertical bars are standard deviations. 1280 E. Cebrian et al. / Marine Pollution Bulletin 46 (2003) 1273–1284

Table 3 Two-way ANOVA for site (control and polluted sites) and sea condition (calm, slightly rough, rough and stormy) eﬀects on water motion Variable Factor DF F p Water motion Site 1 41.556 0.000 (g hÀ1) Sea conditions 3 17.412 0.000 Site and sea con. 3 0.929 0.436 Error 37

contaminated site, the presence of organometallics (e.g. TBT) was discarded because the use of these products is prohibited in antifouling paints for small ships. The mean values of POM (carbon and nitrogen) and dissolved nutrient salts at both sites were within the range for coastal Mediterranean waters (Mann, 1982; Ballesteros, 1992) and the differences found were due to Fig. 5. Granulometric sorting of the sediment at the control and pol- very punctual values, and thus, they do not appear to be luted sites for different experiment months. Very fine sediment (VF < 12 lm), fine sediment (12 lm < F > 30 lm), medium sediment (30 relevant enough to exert significant effects on the bio- lm < F > 100 lm) and coarse sediment (C <100 lm). logy of the sponge. Only organic matter in the sediment and water movement can affect the sponge biology along with, of course, the higher concentration of copper at 3.2.6. Settlers survival the polluted site. A similar pattern of settler mortality was observed at The lower water motion detected at the polluted site the control and polluted sites (GehanÕs Wilcoxon Test, is inherent to harbours and is usually associated with p ¼ 0:47). Most of the settlers disappeared during the high sediment rates, which can affect negatively sponges first week of the experiment at both sites (Fig. 11). (Verdenal, 1986). However, in our case, the gross sedi- Thereafter, the number of settlers remained more or less ment rates (GSRs) were higher at the control site in constant until the end of the experiment (four weeks). winter months, probably owing to the re-suspension of coarse particles under storms (Greemare et al., 1998). The 4. Discussion sedimentation rates recorded at both sites were within the range reported for other Mediterranean areas We aimed to experimentally address the sublethal (Buscail, 1991; Greemare et al., 1998) and therefore do effects of metal contamination on the widespread Med- not suggest particular adverse effects for the sponge. iterranean sponge C. crambe by means of a transplant Only during the last month of the experiment (June), the experiment. However, since environmental parameters fine fraction was conspicuously higher at the polluted other than metals could also vary between the control site. Thus, the fine sediment, although it may be dele- and the polluted sites selected, we also analysed the most terious for the sponge in summer by clogging its inhal- relevant water and sediment variables at both sites. ant orifices, did not affect the sponge variables measured Given the small size of the harbour selected as the during the previous months.

Fig. 6. Water motion as measured by CaSO4 losses at the control site and polluted sites for all the sea conditions studied. E. Cebrian et al. / Marine Pollution Bulletin 46 (2003) 1273–1284 1281

Fig. 7. Mean concentration of copper (A), lead (B), and vanadium (C) present within the sponges tissues. C, untouched control; CT, transplant control; H, harbour native; HT, harbour transplant. Boxes represent standard errors; vertical bars are standard deviations. Mean concentration, which proved not signiﬁcantly diﬀerent in a Tukey test were joined by horizontal lines.

Table 4 ronmental variables that differed between the control One-way ANOVAs for sponge treatment (untouched control, trans- and the polluted sites. plant control, native at the polluted site and transplanted to the pol- Sponges filter a large volume of water and accumu- luted site) effect on metal accumulation late heavy metals (e.g. Patel et al., 1985; Verdenal et al., Variable (lggÀ1) Factor DF F p 1990; Perez, 2001; Hansen et al., 1995). However, ac- Copper Treatment 3 34.231 <0.000 cumulation seems to depend on the metal and the spe- Error 12 cies considered (Perez, 2001). Our results indicate that Lead Treatment 3 3.901 0.037 the sponge C. crambe efficiently concentrates heavy Error 12 metals from seawater. Specimens from the control and Vanadium Treatment 3 2.460 0.112 contaminated sites accumulate copper, lead and vana- Error 12 dium, although the two last metals were at low concentrations in both sites. Differences in copper were particularly relevant between specimens transplanted to the polluted and to the control sites. Moreover, the Sheltered conditions, copper concentration in the higher organic contents of the sediment at the contam- water, together with organic matter in the sediment inated site during the two last months of the experiment when the temperature increases above 16 C (i.e. late may have promoted copper and lead accumulation by May and June), appear to be the most significant envi- metal binding, which would favour further ingestion by 1282 E. Cebrian et al. / Marine Pollution Bulletin 46 (2003) 1273–1284

Area Adults survival 18 100 A 16 14 80 12 60

2 10

cm 8 40 6 % survivors 4 20 2 0 0 March April May June March April May June

GRm Fig. 9. Percentage of survivors among sponges transplanted to the 0.4 B control (solid symbol) and polluted (empty symbol) sites. 0.2

-0.2 Reproduction

-0.4 Native Transplant -0.6 (Control site) (Control Site) 82% 18% 50% 50% -0.8 March April May June

Perimeter/Area

8 C Transplant Native (Polluted Site) (Polluted site) 6 25% 75% 11,8% 88,2%

0 Containing embryos Without March April May June Control Site Polluted Site Fig. 10. Percentage of sponges harbouring embryos in the various treatments. Fig. 8. Time course of the mean sponge area (A), monthly growth rate (B) and perimeter/area ratio (C) of the sponges transplanted to the control site (solid symbol) and to the polluted site (empty symbol). Vertical bars are standard errors. Settlers survival 100

Table 5 80 Signiﬁcance levels obtained by randomisation for the repeated measures analyses of the area, monthly growth rate and perimeter/area 60 ratio

%s survivors 40 Source of Percentage of randomisation SS that variation exceeds the observed sum of squares 20

Area GRm Perimeter/area 0 1st week 2nd week 3rd week 4th week Site 0.134 0.000 0.000 Time 0.279 0.180 0.256 Control Site Polluted Site Site and time 0.382 0.027 0.985 Fig. 11. Settler survival (%) at the control (solid symbol) and polluted SS: Sums of squares. (empty symbol) sites. Vertical bars are standard errors. the sponge cells.There were no signiﬁcant diﬀerences in taminated site for four months (transplants) and those copper contents between sponges inhabiting the con- living there for more than three years (native sponges). E. Cebrian et al. / Marine Pollution Bulletin 46 (2003) 1273–1284 1283

This suggests the existence of an intracellular control of can attribute the sublethal effects on the biology of the metal concentration in the sponge (Philp, 1999). sponge mainly to water movement, copper content and As for the biological variables, negative effects were organic matter in the sediment. They inhibited sponge observed in the sponges transplanted to the polluted growth, increased sponge irregularity and fission and site. The sponge growth rate was negative in these decreased fecundity. All these effects may compromise sponges, while it was close to 0 in the transplantcontrol the structure and dynamics of the sponge populations in during the study, as expected, since the sponge does not moderately metal-contaminated habitats. grow in winter (Turon et al., 1998). The sponge shape was more irregular in the individuals transplanted to the contaminated site than in Acknowledgements transplant controls, which may be a result of sponge stress (Agell et al., 2001). As reported for other inver- We thank the S.C.T. of the University of Barcelona tebrates (Turon and Becerro, 1992), changes in the pe- for MOP and heavy metal analyses, and S. Pla for nu- rimeter of C. crambe also appear to respond to trient salts analyses. M.P. Piscitelli helped us in the environmental pressures (Becerro et al., 1994). fieldwork and image treatment of the sediment granul- The percentage of individuals containing embryos ometry, G. Agell performed paraffin sections and I. was notably lower in the sponges at the polluted site Abreu was our captain in winter cruises from the Blanes than in transplant and untouched controls. Although harbour to the Sta. Anna Point. We also thank R. Coma this suggests that contamination hinders embryo pro- for technical support on image treatment. This re- duction, and copper has been proved to be responsible search was funded by grants MAR98-1004-CO2 and for inhibition of embryo development in other inverte- REN2001-2312-CO3/MAR from the Spanish Govern- brates (Bellas, 2001), a lower density of individuals at ment to MJU and XT, and BIOMARK (Mast-III the contaminated site may also shrink fertilisation suc- CT97-0118) from the European Union to MJU. It also cess and contribute to the lower fecundity observed. benefited from fellowships to RM (CIRIT) and EC Manipulation during transplantation also seems to af- (CSIC). fect sponge reproduction since the number of transplanted individuals that incubated embryos was 20% lower than that of the untouched individuals at the References control site. Stress-induced gamete re-absorption may be responsible. Abbe, G.R., Riedel, G.F., Sanders, J.G., 2000. Factors that influence Sponge survival was relatively high until June. Mor- the accumulation of copper and cadmium by transplanted eastern tality was higher at the control site than at the harbour oysters (Crassostrea virginica) in the Patuxent River, Maryland. during the first month of the experiment, probably Marine Environmental Research 49, 377–396. Agell, G., Uriz, M.J., Cebrian, E., Marti, R., 2001. Does stress owing to the exposed conditions of the control site. proteins induction by copper modify natural toxicity in sponges? Conversely, mortality increased drastically at both sites Environmental Toxicology Chemistry 20 (11), 2588–2593. in June, in concordance with an increase in temperature Alcolado, P.M., Herrera, A., 1987. Efectos de la contaminacioon sobre over 20 C. Mortality rates similar to those recorded at las comunidades de esponjas en el litoral de la Habana, Cuba the control site have been reported for natural popula- Reporte de Investigacioon del Instituto de Oceanologııa. Academia de Ciencias de Cuba 68 (1). tions of C. crambe in the same area (Turon et al., 1998). Ballesteros, E., 1992. Els vegetals i la zonacio litoral: espeecies, Survival of the three-weeks old settlers was similar at comunitats i factors que influeixen en la seva distribucioo. Institut both sites, which indicates that they endured the con- dÕEstudis Catalans, Barcelona. ditions of the polluted site and explains the presence of a Becerro, M.A., Uriz, M.J., Turon, X., 1994. Trends in space natural population of the sponge in this site. However, occupation by the encrusting sponge Crambe crambe: variation in shape as a function of size and environment. Marine Biology 121, there was a noticeable mortality during the first week of 301–307. the experiment (70%). This high mortality is similar to Bellas, J.D., 2001. Evaluacioon de la calidad del agua en ecosistemas that observed for newly settlers in the field (Urizet al., costeros mediante criterios bioloogicos: bioensayos con gametos, 1998) and to that reported for early post-metamorphic embriones y larvas de la ascidia solitaria Ciona intestinalis. PhD stages of several invertebrates (Gosselin and Qian, University of Vigo, Vigo, Spain, 182 pp. Bjerregaard, P., Depledge, M.H., 1994. Cadmium accumulation in 1997). Littorina littorea, Mytilus edulis and Carcinus maenas: the influence To summarize, the sponge C. crambe appears to be a of salinity and calcium ion concentrations. Marine Biology 119, suitable biomonitor of metal contamination thanks to 385–395. its notable capacity to accumulate copper, lead and Brown, B., Ahsnullah, M., 1971. Effect of heavy metals on mortality vanadium in its tissues, and because it experiences de- and growth. Marine Pollution Bulletin 2, 182–185. Buscail, R., 1991. Le cycle du carbone sur une marge continental: grees of behavioral and physiological responses such as aspects biogeeochimiques du transfert de la matieere organique a changes in shape, growth rates, reproduction, which can lÕinterface eau-seediment. Theese Doctorat dÕEtat, Universiteede be easily monitored (Phillips and Rainbow, 1994). We Perpignan. 1284 E. Cebrian et al. / Marine Pollution Bulletin 46 (2003) 1273–1284

Cairns Jr., J., Pratt, J.R., 1989. The scientific basis of bioassays. and man-induced variability in the Bay of Blanes. PhD. University Hydrobiologia 188, 5–20. of Barcelona, Barcelona. Carballo, J.L., Naranjo, S., 2002. Environmental assessment of a large Potwin, C., Lechowics, M.J., Tardiff, S., 1990. The statistical analysis industrial marine complex based on a community of benthic filter of ecophysiological response curves obtained from experiments feeders. Marine Pollution Bulletin 44 (7), 605–610. involving repeated measures. Ecology 71, 1389–1400. Carballo, J.L., Naranjo, S., Garcııa-Goomez, J.C., 1996. Use of marine Rainbow, P.S., Scott, A.G., Wiggins, E.A., Jackson, R.W., 1980. sponges as stress indicators in marine ecosystem at Algeciras Bay Effect of chelating agents on the accumulation of Cadmium by the (southern Iberian Peninsula). Marine Ecology Progress Series 135, barnacle Semibalanus balanoides and complexation of soluble Cd, 109–122. Zn and Cu. Marine Ecology Progress Series 2, 143–152. Carman, R.K., Fleeger, J.W., Means, J.C., Pomarico, S.M., McMillin, Rainbow, P.S., Wang, W.-X., 2001. Comparative assimilation of Cr, D.J., 1995. Experimental effects of polynuclear aromatic hydro- Se and Zn by the barnacle Elminius modestus from phytoplankton carbons on an estuarine sediment food web. Marine Environmental and zooplankton diets. Marine Ecology Progress Series 218, 239– Research 40, 289–318. 248. Conover, W.O., Iman, R.L., 1981. Rank transformation as a bridge Reiswig, H.M., 1971. Particle feeding in natural populations of three between parametric and nonparametric statistics. American Sta- marine demosponges. Biological Bulletin 141, 568–591. tistics 35, 124. Riisgaard, H.U., 1979. Effect of copper on volume regulation in the Galera, J., Turon, X., Uriz, M.J., Becerro, M.A., 2000. Microstructure marine flagellate Dunaliella marina. Marine Biology 50, 189–193. variation in sponges sharing growth form: the encrusting demo- Solee, M., Morcillo, Y., Porte, C., 2000. Stress proteins response in sponges Dysidea avara and Crambe crambe. Acta Zoologica 81, 93– tributyltin-exposed clams: laboratory and field experiments. Bulle- 107. tin Environmental Contamination Toxicology 64, 852–858. Gosselin, L.A., Qian, P.I., 1997. Juvenile mortality in benthic Turon, X., Becerro, M., 1992. Growth and survival of several ascidian invertebrates. Marine Ecology Progress Series 146, 265–283. species from the northwestern Mediterranean. Marine Ecology Greemare, A., Amouroux, J.M., Charles, F., Medernach, L., Jordana, Progress Series 82, 235–247. E., Nozais, C., Vetion, G., Colomines, J.C., 1998. Temporal Turon, X., Tarjelo, I., Uriz, M.J., 1998. Growth dynamics and changes in the biochemical composition of particulate organic mortality of the encrusting sponge Crambe crambe (Poecilosclerida) matter sedimentation in the Bay of Banyuls-sur-Mer. Oceanologica in contrasting habitats: correlation with population structure and Acta 21, 783–792. investment in defence. Functional Ecology 12, 631–639. Hansen, I.V., Weeks, J.M., Depledge, M.H., 1995. Accumulation of Uriz, M.J., Maldonado, M., Turon, X., Martıı, R., 1998. How copper, zinc, cadmium and chromium by the marine sponge reproductive output, larval behaviour, and recruitment contribute Halichondria panicea Pallas and the implications for biomonitor- to adult spatial patterns in Mediterranean encrusting sponges? ing. Marine Pollution Bulletin 31 (1–3), 133–138. Marine Ecology Progress Series 167, 137–148. Kobayashi, N., 1980. Comparative Sensitivity of various developmen- Uriz, M.J., Martin, D., Rosell, D., 1992a. Relationship of biological tal stages of sea urchins to some chemicals. Marine Biology 58, and taxonomic characteristics to chemically mediated bioactivity in 163–171. Mediterranea littoral sponges. Marine Biology 113, 287–297. Manly, G., 1991. Randomisation and Monte Carlo Methods in Uriz, M.J., Rosell, D., Martin, D., 1992b. The sponge population of Biology. Chapman & Hall, London. the Cabrera archipelago (Balearic Islands): characteristics, distri- Mann, K.H., 1982. The role of microorganisms in coastal waters. In: bution and abundance of the most representative species. PSZNI Anderson, D.J., Greig-Smith, P., Pilteka, F.A. (Eds.), Ecology of Marine Ecology 113, 101–117. Coastal Waters. A System Approach. University of California Uriz, M.J., Turon, X., Becerro, M.A., Galera, J., Lozano, J., 1995. Press. Patterns of resource allocation to somatic, defensive, and repro- Martoja, M.R., Martoja, M., 1970. Teecnicas de Histologia Animal. ductive functions in the Mediterranean encrusting sponge Crambe Toray-Masson, Barcelona. crambe (Demospongiae, Poecilosclerida). Marine Ecology Progress Muus, B.J., 1968. A field method for measuring exposure by means of Series 124, 159–170. plaster balls. Sarsia 34, 61–68. Vacelet, J., 1979. La place des spongiaires dans les systemes trophiques Newton, L.C., McKenzie, J.D., 1995. Echinoderms and oil pollution: marins. In: C. Leevi, Boury-Esnault N, (Eds.), Biologie des A potential stress assay using bacterial simbionts. Marine Pollution Spongiaires. Editions du CRNS, Paris. Bulletin 31, 453–456. Verdenal, B., 1986. Spongiculture en Meediterraneee Nord-Occidentale: Ozoh, P.T.E., 1990. Volume regulation and accumulation of copper aspects culturaux, molysmologiques et eeconomiques. PhD thesis, in whole and decapitated Hediste (Nereis) diversicolor under UniversiteedÕAix-Marseille II. Centre dÕOceeanologique de Mar- varying salinity copper regimes. Water Air Soil Pollution 49, seille, Station Marine dÕEndoume, Marseille. 215–230. Verdenal, B., Diana, C., Arnoux, A., Vacelet, J., 1990. Pollutant levels Patel, B., Balani, M.C., Patel, S., 1985. Sponge ‘‘sentinel’’ of heavy in Mediterranean commercial sponges. In: Ruutzler,€ K. (Ed.), New metals. The Science of the Total Environment 41, 143–152. Perspectives in Sponge Biology. Smithsonian Institution Press, Perez, T., 2001. QualiteedelÕeenvironnement marin littoral: eetude des Washington, pp. 516–524. spongiaires pour la bioevaluation des peuplements de substrats Von Ende, C.N., 1993. Repeated measures analysis: growth and other durs. PhD University de la Meediterraneee, Marseilles, France, 229 time dependent measures. In: Sheiner, S.M., Gurevitch, J. (Eds.), pp. Desing and Analysis of Ecological Experiments. Chapman & Hall, Phillips, D.J.H., Rainbow, P.S., 1994. Biomonitoring of Trace Aquatic London, pp. 113–137. Contaminants, second ed Chapman & Hall, London. Warwick, R.M., Platt, H.M., Clarke, K.R., Agarrad, J., Gobin, J., Philp, R.B., 1999. Cadmium content of the marine sponge Microciona 1990. Analysis of macrobenthic and meiobenthic community prolifera, other sponges, water and sediment from the eastern structure in relation to pollution and disturbance in Hamilton Florida panhandle: possible effects on Microciona cell aggregation Harbour, Bermuda. Journal of Experimental Marine Biology and and potential roles of low pH and low salinity. Comparative Ecology 138, 119–142. Biochemistry and Physiology C-Pharmacology Toxicology and Young, S.J., Buschbom, R.L., Gurtisen, J.M., Joyce, S.P., 1979. Effects Endocrinology 124, 41–49. of copper on the sabellid polychaete, Eudistylia vancouveri: I. Pinedo, S., 1998. Structure end dynamics of Western Mediterranean Concentration limits for copper accumulation. Archives Environm soft-bottom communities along a disturbance gradient. Natural Contamination Toxicology 8, 97–106.