ARTICLE IN PRESS

Ecotoxicology and Environmental Safety 67 (2007) 218–226 www.elsevier.com/locate/ecoenv Highlighted article The use of riparius larvae to assess effects of pesticides from rice fields in adjacent freshwater ecosystems

Mafalda S. FariaÃ, Anto´nio J.A. Nogueira, Amadeu M.V.M. Soares

CESAM & Departamento de Biologia da Universidade de Aveiro, Campus de Santiago, 3810-193 Aveiro, Portugal

Received 30 March 2006; received in revised form 10 November 2006; accepted 20 November 2006 Available online 16 January 2007

Abstract

A bioassay with larvae, using larval development and growth as endpoints, was carried out inside a rice field and in the adjacent wetland channel in Portugal, during pesticide treatments (molinate, endosulfan and propanil) to determine impact caused by pesticide contamination in freshwater ecosystems. The bioassay was also performed under laboratory conditions, to assess whether in situ and laboratory bioassays demonstrated comparable results. Growth was inhibited by concentrations of endosulfan (2.3 and 1.9 mgL1 averages) in water from rice field in both the field and laboratory, and by concentrations of endosulfan (0.55 and 0.76 mgL1 averages) in water from the wetland channel in the laboratory bioassay, while development was not affected. C. riparius larvae were not affected by molinate and propanil concentrations. The results indicate that endosulfan treatments in rice fields may cause an ecological impairment in adjacent freshwater ecosystems. The results also indicate that laboratory testing can be used to assess in situ toxicity caused by pesticide contamination. r 2006 Elsevier Inc. All rights reserved.

Keywords: Chironomus riparius; Bioassays; Endosulfan; Molinate; Propanil; Rice field; Ecological impact; Wetlands

1. Introduction and Mian, 1981), and may affect human health (Buxton and Stedfast, 1994; Sumner and McLaughlin, 1996). Several pesticides are globally used in agriculture and Rice, along with wheat and corn, is one of the three may enter aquatic environments through spray drift or crops on which many human subsists. Almost two billion runoff events, drainage or leaching, resulting in contamina- people depend on rice for their staple diet (Ronald, 1997). tion of surface and ground waters (Leonard et al., 1999; It is also an important crop in terms of pesticide use, Schulz et al., 2001a, b; Cerejeira et al., 2003). Pesticides used particularly herbicides. The use of pesticides during cereal in Portuguese agricultural areas have been found in both growth may affect the quality of the surrounding aquatic surface and ground waters. Insecticides (e.g., endosulfan, environment. In Portugal, rice is an irrigated crop and lindane, chlorfenvinphos) and herbicides (e.g., propanil, water is frequently discharged into surrounding water molinate, atrazine) have been detected in the surface waters bodies (Pereira et al., 2000). Studies designed to clarify the of three river basins (Tejo, Sado and Guadiana) from 1983 potential toxic effects of rice-associated pesticides on to 1999, and in ground water samples, collected from wells aquatic ecosystems are therefore an important tool for in seven different agricultural areas in the Tejo and Sado biological understanding and environmental management. basins between 1991 and 1998 (Cerejeira et al., 2003). However, the impact of pesticides on aquatic ecosystems is Waters contaminated with pesticides can cause toxic effects often difficult to assess because they degrade very quickly to aquatic flora and fauna (Forney and Davis, 1981; Mulla (Barry and Logan, 1998), can be absorbed onto the sediments (Peterson and Batley, 1993; Hamer et al., 1999) ÃCorresponding author. Fax: +351 234426408. and peak concentrations that coincide with storm runoff E-mail address: [email protected] (M.S. Faria). events may be difficult to measure (Cooper, 1996).

0147-6513/$ - see front matter r 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.ecoenv.2006.11.018 ARTICLE IN PRESS M.S. Faria et al. / Ecotoxicology and Environmental Safety 67 (2007) 218–226 219

The aim of this study was to determine the effect of Arzila Marsh pesticides used in rice treatments on a selected aquatic macroinvertebrate and assess the potential risk of these chemicals to aquatic ecosystems. A bioassay using Chir- onomus riparius larvae and employing biological responses Wetland Channe (development, growth) as endpoints was deployed in a rice field and in the adjacent wetland channel during the period of pesticide spraying. Third instar larvae l s High were deployed in rice fields 48 h after each pesticide Rice Field Contaminated treatment (endosulfan, molinate, propanil). Bioassays Site using contaminated water from the rice fields and condi- tioned sediments were also conducted under laboratory Low conditions, at the same time as the in situ bioassay, to Contaminated assess whether in situ and laboratory tests demonstrated Site Reference 0 m 125 m comparable results. Site larvae have been recorded as pests of rice growing in many temperate countries (Surakarn and Yano, 1995). The larvae either attack the seed itself, or feed on the Fig. 1. Study sites where the in situ bioassay were performed. roots or shoots of young seedlings (Helliwell and Stevens, 2000; Stevens et al., 2000). Damage can arise through their tunnelling activity in the sediment which can destabilise the rice and potato fields. This herbicide is soluble in water but adsorbs root systems of young plants and increase turbidity, weakly to soil particles (Wauchope et al., 1992; Weed Science Society of America, 1994). It rapidly breaks down in water due to microbial activity reducing photosynthesis and slowing the growth of and has a low persistence in soil, with a field half-life of 1–3 days submerged seedlings (Helliwell and Stevens, 2000). But (Wauchope et al., 1992; Weed Science Society of America, 1994). Chironomids are also important macroinvertebrates in the ecology of the aquatic ecosystems. They are the most 2.2. Test organisms: culture conditions widespread and often abundant group of in freshwater environments (Pinder, 1995). They provide an C. riparius larvae used in the bioassays were obtained from laboratory important link between different trophic levels, and play a cultures established at the Biology Department, University of Aveiro. key role in recycling organic matter (Prat and Rieradevall, The culture unit was an enclosed transparent acrylic box (120 cm 1995; Garcı´a-Berthou, 1999). 60 cm 40 cm), containing all the conditions necessary to complete the whole life cycle of the chironomids and large enough to allow swarming and copulation of emerged adults (OECD, 2000). Cultures were 2. Material and methods maintained at 2072 1C, with a 14 h light: 10 h dark photoperiod. At the start of a new culture 120 newborn larvae were introduced into plastic beakers (27 cm 13 cm 11.5 cm) containing a 2 cm layer of organic 2.1. Study sites and pesticide treatment matter free natural sediment (ignited in a muffle furnace for 6 h at 450 1C) from the Bestanc-a river (north Portugal) and ASTM reconstituted hard In situ bioassays with C. riparius were carried out in rice fields within water (ASTM, 2000). This sediment was considered unpolluted because it the Natural Reserve of Arzila Marsh, a protected wetland located near the came from a river segment known to be free from pollution (Faria et al., Mondego river in Central Portugal (Fig. 1). The site designated as highly 2006). A suspension of ground TetraMins (Tetrawerke, Germany) of contaminated was located inside a rice field, representing one of the several 1 mg larvae1 day1 was added as the food source (Naylor and Rodrigues, rice fields that were treated with pesticides (Fig. 1). The control and 1995) and each beaker was gently aerated. Seven days later, larvae were low contamination sites were located in a channel bordering the rice fields transferred to new culture beakers with fresh media, sediment and food (Fig. 1). The control site was located upstream and the low contamination (60 larvae per beaker) until emergence. site was located downstream from the point where the channel received water from the (highly contaminated) rice fields.

Molinate (C9H17NOS) is a selective herbicide which belongs to the 2.3. In situ bioassays thiocarbamate class. It is used to control grassy weeds in rice fields (Hartley and Kidd, 1983; Meister, 1991) and is of low persistence in the Twenty four hours after each pesticide treatment in rice fields, five soil environment, with a field half-life of 5–21 days (Wauchope et al., chambers, with a layer of 5 cm natural sediment previously collected from 1992). It was the first pesticide to be introduced in the rice field in the study the Bestanc-a river, were placed in each site, to condition the sediment for site. Because it rapidly volatilizes if the soil is wet (Brooks, 1980), it was in situ and laboratory bioassays. Two additional chambers were deployed, ploughed into the dry soil before the water was allowed to inundate the to determine pesticide concentrations in the sediment on day 0 and on day site. After 3 weeks, endosulfan (C9H6Cl6O3S) was applied. Endosulfan is 6 of the in situ bioassay. Six additional chambers were placed in each site an organochlorine insecticide of the cyclodiene subgroup which acts as a to condition sediment during 24 h for laboratory tests. All chambers where contact poison for a wide variety of insects and is highly persistent in the placed in sites using a holding structure, constructed with 2 baskets environment (Helliwell and Stevens, 2000). This insecticide is frequently (10 21 25 cm) made of plastic coated metal wire. In situ chambers were used, associated with deltamethrin, by farmers in Central Portugal to modified from a design by Soares et al. (2005) consisting of PVC tube control chironomid larvae and crayfish Procambarus clarkii infestation in (20 cm high and 4.5 cm inner diameter) with four openings, two laterals rice fields. Approximately 1 month after endosulfan application, the (10 cm 7 cm), one at the top and one at the bottom. The openings were herbicide propanil (C9H9Cl2NO) was applied. Propanil is an acetanilide covered with 200-mm nylon mesh to allow contact with the sediment and post-emergence herbicide used to control numerous weeds that occur in water flow through the chambers. ARTICLE IN PRESS 220 M.S. Faria et al. / Ecotoxicology and Environmental Safety 67 (2007) 218–226

On the first day of the bioassays (day 0), 2 days after pesticide quently used associated with deltamethrin by farmers in Central Portugal, treatment, five third instar larvae of C. riparius were introduced into each water and sediment samples collected to determine endosulfan concentra- assay chamber using a plastic pipette via a plastic tube (tube size depended tions were also used to determined deltamethrin concentrations, but the on water depth) glued to the 200-mm nylon mesh that covered the assay results of deltamethrin analysis were below the detection limits chamber (glued with white thermal glue, Elis-Taiwan, TN122/WS, (0.002 mgL1). All pesticide concentrations were determined by the Taiwan, which has been shown to be nontoxic to cladocerans; Pereira accredited laboratory Innova-lab, Spain. et al., 1999), to avoid sediment disturbance. TetraMins (30 mg in suspension) was added to each of these chambers. The body length and head capsule width of 30 additional larvae were measured to 2.6. Statistical analysis determine initial body length and development stage of larvae. The conditioned sediment of the additional five chambers was collected for The statistical analysis comprised one-way analysis of variance use in the laboratory tests and the sediment from the other chambers, (ANOVA) followed, when necessary, by Tukey HSD multiple comparison previously added were collected for pesticide analysis. Water was tests to test for significant differences in responses of biological endpoints also collected at each site to use in laboratory tests and for pesticide (development and growth) between treatments and between laboratory analysis. and field exposures (Zar, 1996). One-way analysis of variance (ANOVA) At the end of the experiment (day 6) in each site, the surviving larvae followed by Dunnett tests was used to compare responses of biological were collected from the five chambers, killed and preserved in Von To¨rne endpoints between reference and contaminated (low- and high-) treat- conservant (Gama, 1964) for later measurement. Sediment from the sixth ments (Zar, 1996). Responses were tested for normality using the chamber was collected for pesticide analysis. Water samples from each site Kolmogorov–Smirnov test. Linear regression analyses were used to were also collected for pesticide analysis. Larval length and developmental investigate significant relationships between biological responses and stage were determined by measuring body length and head capsule width minimum and maximum water temperature. Statistical analyses were of larvae, respectively, using a stereomicroscope (Leica MS5, Leica performed using MinitabTM Release 14. Microsystems AG, Germany) fitted with a calibrated eye-piece micro- meter. Growth (body length increase) of larvae was calculated by subtracting the average initial length from each individual final length. 3. Results C. riparius larvae instar was determined according to Watts and Pascoe (2000). 3.1. Physical and chemical parameters

2.4. Laboratory bioassays Physical and chemical conditions were more variable in the field than in the laboratory (Tables 1 and 2). Physical In all treatments, on day 0, field water (300 ml) was introduced into five and chemical parameters of water quality in the in situ chambers (glass flasks) with a conditioned sediment layer of 5 cm depth. bioassays were similar in reference and low contaminated five third instar larvae were introduced into each of the chambers and 30 mg of Tetramin, in suspension, was added to each chamber. The sites but differed between these sites and the highly chambers were provided with artificial aeration under a 14 h light: 10 h contaminated sites for the three pesticides (Table 1). dark photoperiod with a temperature of 20 1C. At the end of the Higher values of maximum temperature were observed in experiment (day 6), the surviving larvae were collected from the chambers, the highly contaminated site during the exposure period. killed and preserved in Von To¨rne conservant (Gama, 1964) for later Dissolved oxygen and pH values were also higher in the measurement. Measurement procedures were as described for in situ bioassays. At the same time, another chamber was also included in each highly contaminated site than in reference and low treatment without the addition of larvae and its water and sediment were contamination sites, during all pesticide treatments. In collected for pesticide analysis at the end of the experiment. The the laboratory bioassays, physical and chemical parameters treatments used were: (1) water and conditioned sediment brought from of water samples were similar in all pesticide treatments each site and (2) ASTM hard water (ASTM, 2000), control water, and (Table 2). natural sediment from the Bestanc-a river, used as a control treatment.

2.5. Physical and chemical analysis 3.2. Pesticide concentrations

In field and laboratory tests, hand-held field meters were used to Molinate and propanil concentrations and total endo- measure water pH (pH 330/SEF-2, Weilheim, Germany), conductivity sulfan concentrations, the sum of a and b isomers and (Cond 330i/SET, Weilheim, Germany) and dissolved oxygen (DO) endosulfan sulphate, in water and sediment during field (Oxi 330/SET, Weilheim, Germany) and a maximum–minimum thermo- meter was used to determine the minimum and maximum temperature and laboratory exposure are presented in Table 3. On day of water (880/847881, Electronic Temperature Instruments Ltd, UK) 0, the endosulfan concentration in water varied from 1 for day 0 and day 6. Water samples were collected on day 0 and on day 6 0.91–2.78 mgL . In the sediment of the rice field this was to determine nitrate, nitrite, ammonia and phosphate concentrations 0.62 mgkg1, while at the sites in the wetland channel of the water with a direct reading spectrophometer (DR/2000, Hach endosulfan was not detected in the sediment. On day 0, the company, USA). molinate concentration in the water column varied from The sediment and water samples collected for pesticide analysis were 1 placed in dark glass containers and preserved at 4 1C in the dark. The 4.63 to 0.31 mgL and in sediment it ranged from 1.43 to 1 analysis of endosulfan and molinate were carried out by organic solvent 1.09 mgL . On day 6, the endosulfan and molinate extraction with cartridges containing Oasiss Hydrophilic–Lipophilic concentrations in both water and sediment were higher in Balance Sorbent (HLB), a reversed-phase sorbent for all compounds, field than in laboratory. Propanil was not detected in followed by a concentration determination using liquid chromatography- sediment samples, however, it was detected in the water mass spectrometry (LC-MS). Propanil analysis was carried out by solid 1 phase extraction (SPE), and concentrations were determination by liquid from the rice field on day 0 (2.58 mgL ) and in the 1 chromatography-mass spectrometry (LC-MS). Since endosulfan is fre- laboratory mesocosms on day 6 (0.14 mgL ). ARTICLE IN PRESS M.S. Faria et al. / Ecotoxicology and Environmental Safety 67 (2007) 218–226 221

Table 1 Physical and chemical parameters of water in day 6 and day 0 (in parenthesis) of in situ bioassays

a Treatment Tempmin Tempmax pH DO Cond Nitrate Nitrite Amonia Phosphate Hardness 1 1 1 1 1 1 1 (1C) (1C) (mg L ) (ms cm ) (mg L ) (mg L ) (mg L ) (mg L ) (mg Ca CO3 L )

Endosulfan Ref 15.0 26.0 7.9 (7.9) 7.0 (6.8) 790 (730) 2.20 (2.13) 0.10 (0.05) 0.26 (0.59) 0.57 (0.37) 26.3 (34.5) Low 14.0 29.0 7.8 (7.9) 7.3 (6.9) 810 (733) 2.12 (2.23) 0.12 (0.05) 0.27 (0.50) 0.50 (0.42) 25.9 (40.9) High 12.5 40.0 8.8 (8.0) 9.7 (10.7) 667 (507) 0.95 (1.7) 0.02 (0.07) 0.20 (0.28) 0.01 (0.04) 15.8 (18.6) Molinate Ref 12.0 20.0 7.7 (6.8) 8.7 (8.5) 315 (389) 1.14 (1.32) 0.02 (0.02) 0.40 (1.14) 0.24 (0.15) 47.5 (44.0) Low 13.0 29.0 7.8 (6.5) 9.2 (8.6) 442 (352) 0.72 (1.54) 0.005 (0.01) 0.93 (1.10) 0.04 (0.26) 23.86 (32.8) High 7.0 32.0 7.2 (8.3) 10.2 (10.7) 397 (386) 0.43 (1.51) 0.002 (0.02) 0.38 (0.92) 0.02 (0.08) 16.2 (20.8) Propanil Ref 12.0 30.0 8.4 (7.9) 7.4 (7.8) 716 (751) 2. 03 (2.20) 0.02 (0.03) 0.28 (0.49) 0.11 (0.08) 22.6 (27.2) Low 17.0 31.0 8.2 (7.9) (7.3) 8.02 713 (781) 1.93 (2.11) 0.02 (0.04) 0.26 (0.45) 0.10 (0.07) 27.0 (27.7) High 18.0 47.0 8.9 (8.0) 10.5 (10.7) 636 (831) 0.12 (0.43) 0.02 (0.04) 0.51 (0.22) 0.03 (0.02) 15.1 (24.3)

Abbreviations: Temp: temperature; Cond: conductivity; DO: dissolved oxygen. aRef ¼ water from reference site and conditioned natural sediment, Low ¼ water from low contaminated site and conditioned natural sediment, High ¼ water from high contaminated site and conditioned natural sediment.

Table 2 Physical and chemical parameters of water determined in day 6 and day 0 (in parenthesis) of laboratory bioassays

Treatmenta Temp min (1C) Temp max (1C) pH DO (mg L1) Cond (ms cm1)

Endosulfan Control 17.8 18.0 7.6 (7.7) 7.0 (7.5) 666 (630) Ref 17.3 18.8 8.0 (7.7) 7.5 (7.2) 766 (737) Low 17.3 19.0 8.2 (7.6) 7.3 (7.7) 801 (830) High 17.4 18.7 8.3 (7.9) 7.2 (8.0) 714 (660) Molinate Control 19.1 20.4 8.4 (7.6) 8.1 (8.3) 624 (595) Ref 19.3 20.3 8.3 (7.8) 8.3 (8.9) 590 (423) Low 19.0 19.4 8.2 (7.9) 8.0 (8.9) 459 (390) High 19.0 19.4 8.2 (7.6) 8.5 (8.7) 490 (397) Propanil Control 19.8 20.5 8.0 (7.7) 7.5 (8.0) 667 (615) Ref 19.6 20.3 8.1 (7.9) 7.5 (7.3) 710 (733) Low 19.4 19.8 8.2 (8.0) 7.2 (8.0) 704 (730) High 19.7 20.0 8.0 (7.9) 7.9 (8.2) 675 (798)

Abbreviations: Temp: temperature; Cond: conductivity; DO: dissolved oxygen. aControl ¼ ASTM and not conditioned natural sediment, Ref ¼ water from reference site and conditioned natural sediment, Low ¼ water from low contaminated site and conditioned natural sediment, High ¼ water from high contaminated site and conditioned natural sediment.

3.3. Biological responses larvae changed from the third to the fourth instar and no significant (P40.05) differences were observed in head The herbicides molinate and propanil did not cause capsule width of larvae between treatments, although lower biological responses of C. riparius larvae (Fig. 2). No values were observed in the rice field channel (Fig. 3). significant (P40.05) differences in endpoints between treat- In situ, larval growth was significantly lower in the rice ments were observed in laboratory and field exposures to field than in the reference and low contaminated sites in the herbicides, although a trend of decreasing larval growth was wetland channel (Fig. 3), whilst no significant differences in observed from the highest to the lowest (reference) larval growth were observed between reference and low contaminated sites during molinate exposure in field (Fig. 2). contaminated sites (P40.05). The insecticide endosulfan had a negative effect on larval In the laboratory differences in larval growth were growth under field and laboratory exposure, but did not smaller but similarly significant between treatments, with affect larval development. During endosulfan exposure, all lower growth observed in the treatment with water and ARTICLE IN PRESS 222 M.S. Faria et al. / Ecotoxicology and Environmental Safety 67 (2007) 218–226

Table 3 Pesticide concentrations in water (mgL1) and sediment (mgKg1) treatments during in situ and laboratory bioassays

Treatment (d.l.)a Water (mgL1) Sediment (mgkg1)

day 0 day 6 (field) day 6 (lab) day 0 day 6 (field) day 6 (lab)

Endosulfan (0.002 mgL1) Reference 0.91 0.62 0.19 n.d. n.d. n.d. Low contaminated 1.18 0.81 0.33 n.d. n.d. n.d. High contaminated 2.78 1.82 0.94 0.62 0.11 0.10 Molinate (0.03 mgL1) Reference 0.31 0.81 0.21 1.10 1.09 0.10 Low contaminated 1.48 0.97 0.60 1.28 1.14 0.57 High contaminated 4.63 1.31 1.03 2.64 1.43 0.99

Propanil (0.03 mgL1) Reference n.d. n.d. n.d. n.d. n.d. n.d. Low contaminated n.d. n.d. n.d. n.d. n.d. n.d. High contaminated 2.58 n.d. 0.14 n.d. n.d. n.d.

Abbreviations: d.l.: detection limit, lab: laboratory, n.d.: not detected.

1.0 Molinate - Development Control Propanil - Development Control Reference Reference 0.8 Low contaminated Low contaminated High contaminated High contaminated

0.6

0.4

0.2 Head capsule width (mm) + SE 0.0 8 Molinate - Growth Propanil - Growth

6

4

2

Body length increase (mm) + SE 0 Laboratory In situ Laboratory In situ

Fig. 2. Mean values and standard errors of biological responses, development (head capsule width) and growth (body length increase) of C. riparius larvae exposed to molinate and propanil treatments (control, reference low and high contamination; see Table 2), in the laboratory and in the in situ bioassays.

conditioned sediment from the rice field in comparison to tamination sites (P40.05). Comparison with the control treatments with water and conditioned sediment from treatment revealed an inhibition of larval growth in high, reference and low contamination sites (Fig. 3). As was low and reference treatments of the insecticide (Fig. 3). observed in situ, laboratory bioassay found no significant Regression analyses indicated that larval development was differences in growth between the treatments with water significantly negatively related (r2 ¼ 0.72, Po0.01) to and conditioned sediment from reference and low con- maximum temperature during in situ exposure to pesticides ARTICLE IN PRESS M.S. Faria et al. / Ecotoxicology and Environmental Safety 67 (2007) 218–226 223

1.0 4. Discussion Development Control Reference 4.1. Pesticide concentrations 0.8 Low contaminated High contaminated On day 0 of the in situ bioassay, 2 days after endosulfan 0.6 and molinate spraying of rice fields, the wetland channel was contaminated due the spray drift and/or runoff. The highest pesticide concentrations were found in the rice field, 0.4 as expected, since this site was subject to direct application (Table 3). The lowest pesticide concentrations were found 0.2 in the reference site, located in the wetland channel. Although pesticide concentrations in the reference site were Head capsule width (mm) + SE lower than in the low contamination site, the differences in 0.0 concentration between the two sites were lower than we 7 expected, since the reference site and the low contamina- Growth tion site were located upstream and downstream, respec- 6 * tively, from the contaminated rice field effluent. This might * be explained by the occurrence of a reflux of water between 5 # # ** * the two sites, which were in close proximity to each other in # # # the same channel. This effect may be greater when water 4 from the rice fields is discharged into the channel. 3 Differences in the molinate and endosulfan concentrations in water and conditioned sediment between field and 2 laboratory treatments on day 6 may have been caused by the input of additional endosulfan during in situ exposure 1 period or they may have resulted from the storage and Body length increase (mm) + SE transport of water and sediment from field to the 0 laboratory. Laboratory In situ

Fig. 3. Mean values and standard errors of biological responses, 4.2. Biological responses development (head capsule width) and growth (body length increase) of C. riparius larvae exposed to endosulfan treatments (control, reference, C. riparius larvae were not affected by herbicides, low and high contaminated; see Table 2), in the laboratory and in the in suggesting that this species is not a suitable to assess situ bioassays. Asterisks indicate significant differences between treat- herbicide contamination.The decreasing trend of the ments (* ¼ Po0.05; ** ¼ Po0.01, Tukey tests); hash marks indicate significant differences between treatments and the control (] ¼ Po0.05; growth response to molinate from the highest to the lowest ]]] ¼ Po0.0001, Dunnett tests). (reference) levels of contamination observed in the in situ bioassay (Fig. 2) might be explained by differences in minimum and maximum temperature between treatments (Table 1). Although no significant relationships between (head capsule width ¼ 0.623–0.00325 maximum tempera- larval growth and temperature were observed in this study, ture), but it was not related to maximum temperature Pe´ry and Garric (2006) found high significant relationships during laboratory exposure (P40.05). Regression analyses between the growth of C. riparius larvae not exposed to did not reveal significant relationships (P40.05) between contamination and temperature. head capsule width and minimum temperature or between In contrast to the herbicides, the insecticide endosulfan larval growth and temperature during in situ and negatively affected growth of C. riparius larvae, conform- laboratory exposure to pesticides. ing to expectations given that chironomids represent a Overall growth was generally greater in laboratory target group of insecticide application. Growth was compared to in situ bioassays for the same pesticides negatively affected by endosulfan contamination in both treatments (Fig. 3). Significant differences (Po0.05) in the laboratory and in situ bioassays (Fig. 3). However, while biological responses of C. riparius larvae, exposed the inhibition of growth in laboratory can be related to the to most all pesticide treatments between in situ and presence of endosulfan concentrations, since physical and laboratory bioassays were observed. The exceptions chemical conditions were similar between treatments, (P40.05) were found for growth of larvae exposed to inhibition of growth in the rice field during in situ exposure endosulfan in the reference treatment and in larvae exposed may also have been influenced by differences in physical to propanil in the low contaminated treatment, and, also, and chemical conditions, in particular the high values for development of larvae exposed to molinate in all recorded for maximum temperature. This may have caused treatments. an additional stress to C. riparius larvae. Landrum et al. ARTICLE IN PRESS 224 M.S. Faria et al. / Ecotoxicology and Environmental Safety 67 (2007) 218–226

(1999) verified that toxicity of several organophosphate endosulfan entering organism through food ingestion insecticides (e.g., azinphos methyl, disulfon, fensulfothion, may increase its toxic effects. Nagel and Loskill (1991) terbufos) to C. riparius larvae varied with pH and found that where endosulfan had been taken up by the prey temperature. Similarly, Lydy et al. (1999) found a positive items of small fish, the resultant body load in fish was five correlation between temperature and toxicity of the orders of magnitude higher than the concentrations in the organophosphate insecticides, chlorpyrifos and m-para- ambient water. Bahner et al. (1977) similarly observed that thion to Chironomus tentans. significant quantities of the organochlorine insecticide Contrary to growth, larval development was not affected chlordecone (kepone) were transferred from preys to by the insecticide. No significant differences were observed predatory fish. In aquatic birds, concentrations of organo- in larval head capsule width between treatments in chlorine contaminants can be up to 100 times greater in laboratory or between different sites in the in situ body tissue than in the surrounding water (Anderson and exposures. The lowest head capsule widths were observed Hickey 1976; Norstrom et al., 1976). in the rice field, the highly contaminated site (Fig. 3), and The sublethal effects (growth inhibition) on C. riparius also for chironomids subject to in situ exposure to propanil larvae caused by endosulfan contamination in the wetland (Fig. 2) and appear to be related to higher maximum channel of the Natural Reserve of Arzila Marsh, resulting temperature at that site (40 and 47 1C) compared to the from pesticide treatment, may be indicative of ecological reference (26 and 30 1C) and the low contamination sites impairment in this protected wetland channels and (29 and 31 1C) (Table 1), with regression analysis indicating similarly indicate the ecological risk that the endosulfan a negative relationship between head capsule width and spray drift and runoff from rice fields constitute to maximum temperature. Frouz et al.’s (2002) detailed study freshwater ecosystems. of the effects of temperature on the developmental rate of Chironomus crassicaudatus demonstrated that larval devel- 4.4. Laboratory versus in situ bioassays opment increased rapidly with increasing temperature up to 20 1C, slowed between 20 and 27.5 1C, and decreased at Higher growth of C. riparius larvae found during temperatures higher than 27.5 1C. laboratory bioassays compared to in situ bioassays, has also been documented by other authors (e.g., Castro et al., 4.3. Relevance of C. riparius larvae bioassay to assess 2003) and maybe be explained by the occurrence of high ecological impact of pesticides temperatures and the increased variability in physical and chemical conditions under condition of in situ exposure Head capsule with of C. riparius larvae appeared to be Nevertheless, a similar biological response to pesticides, more influenced by physical and chemical conditions of principally inhibition on larval growth, was found under exposure, especially by temperature, than by the insecticide both laboratory and field exposures to endosulfan, endosulfan, suggesting that development is not a suitable providing evidence for comparability of the respective measurement to assess endosulfan contamination. In experimental designs. Tucker and Burton (1999) also found contrast, larval growth was highly affected by endosulfan significantly higher lethal effects on C. tentans larvae contamination for both in situ and laboratory exposures, exposed to contaminated rivers surrounded by an agricul- indicating that bioassays with C. riparius larvae using tural area compared to larvae exposed to water and growth as an endpoint is sensitive and a suitable tool to sediment from the same rivers but under laboratory assess the impact of endosulfan contamination in fresh- conditions. These authors similarly attributed this differ- water ecosystems. ence to the variation in the parameters of physical and Growth impairment in chironomid larvae may lead to chemical water-quality, highlighting the potential for bias reduced fitness with a subsequent decrease in population in extrapolating laboratory results to field-based scenarios. abundance (Liber et al., 1996; Sibley et al., 1997). Because In the laboratory bioassay it was also possible to detect chironomids are typically one of the predominant macro- toxicity effects in larval growth resulting from endosulfan invertebrates in freshwater ecosystems (Vos, 2001) and are contamination of the wetland channel, by comparing the important prey items for birds and fish (Van de Bund, growth of larvae in treatments with water and conditioned 1994; Prat and Rieradevall, 1995; Garcı´a-Berthou, 1999), sediment from the wetland channel with the control the detrimental effects on the growth of chironomids larvae treatment, while this was not possible in the in situ caused by a contaminant may have ramifying effects exposure due to absence of a control. Although, in situ throughout the ecosystems. Organochlorine insecticides, bioassays have several advantages: (1) they eliminate biases such as endosulfan, are lipid-soluble and tend to accumu- associated with laboratory-to-field extrapolation; (2) they late in the fatty tissues of organisms. This problem can be reduce sampling-related artefacts; (3) they allow stressor further compounded by biomagnification, which results in concentrations to fluctuate naturally (Tucker and Burton, increased concentrations of contaminants in higher trophic 1999), laboratory bioassays can be important for the levels. Although endosulfan in the water column can be assessment of low levels of contamination by enabling the toxic to fish (Maier-Bode, 1968; Jonsson et al., 1993) and comparison under controlled conditions. This was the case aquatic birds (National Library of Medicine, 1987), of the study area because all wetland channels within the ARTICLE IN PRESS M.S. Faria et al. / Ecotoxicology and Environmental Safety 67 (2007) 218–226 225 limits of Natural Reserve of Arzila Marsh receive inputs accumulation, loss, and transfer through estuarine food chains. agriculture fields. Furthermore, laboratory bioassays with EPA-600/J-77-074. water and sediment collected in situ provide an effective Barry, M.J., Logan, D.C., 1998. The use of temporary pond microcosms for aquatic toxicity testing: direct and indirect effects of endosulfan on way to discriminate effects caused by stress contamination community structure. Aquat. Toxicol. 41, 101–124. from effects caused by stress of other physical and chemical Brooks, G.T., 1980. Perspectives of the chemical fate and toxicity of conditions. Thus, when in situ toxicity testing is used in pesticides. J. Environ. Sci. Health B 15, 755–793. combination with laboratory toxicity testing and physico- Buxton, D.E., Stedfast, D.A., 1994. Estimating the vulnerability to chemical characterization, a more realistic assessment of pesticide contamination of drainage basins used for public supply in New Jersey. New directions in pesticide research, development, pesticide effects to freshwater ecosystems can be made. management and policy. In: Weigman, D.L. (Ed.), Proceedings of the Fourth National Conference on Pesticides. Virginia Water 5. Conclusions Resources Center, Blacksburg, Va, pp. 184–199. Castro, B.B., Guilhermino, L., Ribeiro, R., 2003. In situ bioassay chamber and procedures for assessment of sediment toxicity with Chironomus Impairment of larvae growth by endosulfan contamina- riparius. Environ. Pollut. 125, 325–335. tion in the wetland channel adjacent to rice fields, suggests Cerejeira, M.J., Viana, P., Batista, S., Pereira, T., Silva, E., Vale´rio, M.J., that insecticide spray in rice fields may cause ecological 2003. Pesticides in Portuguese surface and ground waters. Water Res. impairment in adjacent freshwater ecosystems. These 37, 1055–1063. results also lends support to the use of bioassay based on Cooper, B., 1996. Central & North West Regions Water Quality Program, 1995/96 Report on Pesticides Monitoring. Department of Land & C. riparius larvae using growth as the endpoint to assess Water Conservation, Technical Services Directorate, Water Quality ecological impairment caused by insecticide applications to Services Unit. rice fields. The study also indicates that bioassays with C. Faria, M.S., Re´, A., Malcato, J., Silva, P.C.L.D., Pestana, J., Agra, A.R., riparius larvae conducted in the laboratory provide a Nogueira, A.J.A., Soares, A.M.V.M., 2006. Biological and functional valuable auxiliary tool to the interpretion of in situ responses of in situ bioassays with Chironomus riparius larvae to assess river water quality and contamination. Sci. Total Environ. 371, bioassays based on this species. 125–137. Forney, D.R., Davis, D.E., 1981. Effects of low concentrations of Acknowledgments herbicides on submerged aquatic plants. Weed Sci 29, 677–685. Frouz, J., Ali, A., Lobinske, R.J., 2002. Influence of temperature on developmental rate, wing length, and larval head capsule size of The authors would like to thank Anabela Simo˜es and the pestiferous Chironomus crassicaudatus (Diptera: Chironomi- ‘‘ICN—Instituto da Conservac-a˜o da Natureza’’ for allow- dae). J. Econ. Entomol. 95, 699–705. ing the execution of this study in the Natural Reserve of Gama, M.M., 1964. Colembolos de Portugal Continental. PhD Thesis. Arzila Marsh, Reinaldo Maia for consenting to conduct Universidade de Coimbra, Coimbra. Garcı´a-Berthou, E., 1999. Food of introduced mosquito fish: ontogenetic this study in his rice field during the pesticide spray period, diet shift and prey selection. J. Fish. Biol. 55, 135–147. Joa˜o Malcato for help with nutrient analysis, Pedro Reis Hamer, M.J., Goggin, U.M., Muller, K., Maund, S.J., 1999. Bioavail- for help in the field, Ricardo Lopes, Philippe Ross and ibility of lambda-cyhalotrin to Chironomus riparius in sediment–water Kieran Monaghan for revising and comments on the and water only systems. Aquat. Ecosyst. Health Manage. 2, 403–412. manuscript. The authors also thank ‘‘Fundac-a˜oparaa Hartley, D., Kidd, H., 1983. The Agrochemicals Handbook. Royal Society of Chemistry, Nottingham, England. Cieˆncia e a Tecnologia’’, for providing a Ph.D. grant to Helliwell, S., Stevens, M.M., 2000. Efficacy and environmental fate of Mafalda S. Faria (SFRH/BD/2935/2000). alphacypermethrin applied to rice fields for the control of chironomi- dae midge larvae (Diptera: Chironomidae). Field. Crops Res. 67, 263–272. Funding sources: This research was supported by the Jonsson, C.M., Toledo, M.C.F., 1993. Acute toxicity of endosulfan to the Fundac-a˜o para a Cieˆncia e a Tecnologia (FCT) via a fish Hyphessobrycon bifasciatus and Brachydanio rerio. Arch. Environ. Ph.D. studentship grant to M.S. Faria (SFRH/BD/2935/ Contam. Toxicol. 24, 151–155. 2000). Landrum, P.F., Fisher, S.W., Hwang, H., Hickey, J., 1999. Hazard This study involved experimental that were evaluation of ten organophosphorus insecticides against the midge, Chironomus riparius, via QSAR. SAR QSAR Environ. Res. 10, conducted in accordance with national and institutional 423–450. guidelines for the protection of human subjects and Leonard, A.W., Hyne, R.V., Lim, R.P., Chapman, J.C., 1999. Effect of welfare. endosulfan runoff from cotton fields on macroinvertebrates in the Namoi river. Ecotox. Environ. Saf. 42, 125–134. Liber, K., Call, D.J., Whiteman, F.W., 1996. Effects of Chironomus References tentans larval growth retardation on adult emergence and ovipositing success: implications for interpreting freshwater sediment bioassays. Anderson, D.W., Hickey, J.J., 1976. Dynamics of storage of organo- Hydrobiologia 323, 155–167. chlorine pollutants in herring gulls. Environ. Pollut. 10, 183–200. Lydy, M.J., Belden, J.B., Ternes, M.A., 1999. Effects of temperature on ASTM, 2000. Test methods for measuring the toxicity of sediment- the toxicity of M-parathion, chlorpyrifos, and pentachlorobenzene to associated contaminants with freshwater invertebrates. E 1706-00. In: Chironomus tentans. Arch. Environ. Contam. Toxicol. 37, 542–547. Annual Book of American Society for Testing and Materials Maier-Bode, H., 1968. Properties, effect, residues, and analytics of the Standards. ASTM, Philadelphia, USA. insecticide endosulfan. Res. Rev. 22, 10–44. Bahner, L.H., Wilson Jr., A.J., Sheppard, J.M., Patrick Jr., J.M., Meister, R.T., 1991. Farm Chemicals Handbook ‘91. Meister Publishing Goodman, L.R., Walsh, G.E., 1977. Kepone bioconcentration, Company, Willoughby, OH. ARTICLE IN PRESS 226 M.S. Faria et al. / Ecotoxicology and Environmental Safety 67 (2007) 218–226

Mulla, M., Mian, L., 1981. Biological and environmental impacts of Sibley, P.K., Monson, P.D., Ankley, G.T., 1997. The effect of gut contents insecticides malathion and parathion on non-target biota in aquatic on dry weight estimates of Chironomus tentans larvae: implications for ecosystem. Res. Rev. 78, 101–135. interpreting toxicity in freshwater sediment toxicity tests. Environ. Nagel, R., Loskill, R., 1991. Bioaccumulation in aquatic systems— Toxicol. Chem. 16, 1721–1726. contributions to the assessment. VCH, Verlagsgesellschaft, Weinheim, Soares, S., Cativa, I., Moreira-Santos, M., Soares, A.M.V.M., Ribeiro, R., Germany. 2005. A short-term sublethal in situ sediment assay with Chironomus National Library of Medicine, 1987. Hazardous Substances Databank. riparius based on postexposure feeding. Arch. Environ. Contam. TOX-NET, Medlars Management Section, Bethesda, MD. Toxicol. 49, 163–172. Naylor, C., Rodrigues, C., 1995. Development of a test method for Chironomus Stevens, M.M., Fox, K.M., Warren, G.N., Cullis, B.R., Coombes, N.E., riparius using a formulated sediment. Chemosphere 31, 3291–3303. Lewin, L.G., 2000. An image analysis technique for assessing Norstrom, R.J., Risebrough, R.W., Cartwright, DJ., 1976. Elimination of resistance in rice cultivars to root-feeding chironomidae midge larvae chlorinated dibenzofurans associated with polychlorinated biphenyls fed (Diptera: Chironomidae). Field Crops Res. 66, 25–36. to mallards (Anas platyrhynchos). Toxicol. Appl. Pharmacol. 37, 217–228. Sumner, M., McLaughlin, M., 1996. Adverse impacts of agriculture OECD, 2000. Guidelines for testing of chemicals-proposal for a new on soil, water and food quality. In: Naidu, R., et al. (Eds.), guideline 219. Sediment–water chironomid toxicity test using spiked Contaminants and the Soil Environment in the Australasia-Pacific water. Draft document, O.E.C.D., Paris. Region. Kluwer Academic Publishers, Dortrecht, The Netherlands, Pereira, A.M.M., Soares, A.M.V.M., Gonc-alves, F., Ribeiro, R., 1999. pp. 125–181. Test chambers and test procedures for in situ toxicity testing with Surakarn, R., Yano, K., 1995. Chironomidae (Diptera) recorded from zooplankton. Environ. Contam. Toxicol. 18, 1956–1964. paddy fields of the world a review. Acta Dipterol 18, 1–20. Pereira, T., Cerejeira, M.J., Espı´rito-Santo, J., 2000. Use of microbiotests Tucker, K.A., Burton, G.A., 1999. Assessment of nonpoint-source runoff to compare the toxicity of water samples fortified with active in a stream using in situ and laboratory approaches. Environ. Toxicol. ingredients and formulated pesticides Environ. Toxicol 15, 401–405. Chem. 18, 2797–2803. Pe´ry, A.R.R., Garric, J., 2006. Modelling effects of temperature and Van de Bund, W., 1994. Food Web Relations of Littoral Macro- and feeding level on the life cycle of the midge Chironomus riparius:an Meiobenthos. PhD thesis. University of Amsterdam, Netherlands. energy-based modelling approach. Hydrobiologia 553, 59–66. Vos, J.H., 2001. Feeding of detritivores in freshwater sediments. PhD Peterson, S.M., Batley, G.E., 1993. The fate of endosulfan in aquatic Thesis, Department of Aquatic Ecology and Ecotoxicology, Institute systems. Environ. Pollut. 82, 143–152. for and Ecosystems Dynamics, University of Amsterdam, Pinder, L.C.V., 1995. Biology of freshwater chironomidae. Ann. Rev. Netherlands. Entomol. 31, 1–23. Watts, M.M., Pascoe, D., 2000. A Comparative Study of Chironomus Prat, N., Rieradevall, M., 1995. Life cycle and production of Chironomidae riparius Meigen and Chironomus tentans Fabricius (Diptera:Chirono- (Diptera) from Lake Banyoles (NE Spain). Freshw. Biol. 33, 511–524. midae) in aquatic toxicity tests. Arch. Environ. Contam. Toxicol. 39, Ronald, P.C., 1997. Making rice disease-resistant. Sci. Am. 277, 100–105. 299–306. Schulz, R., Peall, S.K.C., Dabrowski, J.M., Reinecke, A.J., 2001a. Wauchope, R.D., Buttler, T.M., Hornsby, A.G., Augustijn-Beckers, Current-use insecticides, phosphates and suspended solids in the P.W.M., Burt, JP., 1992. SCS/ARS/CES Pesticide properties database Lourens River, Western Cape, during the first rainfall event of the wet for environmental decisionmaking. Rev. Environ. Contam. Toxicol. season. Water SA 27, 65–70. 123, 1–157. Schulz, R., Peall, S.K.C., Dabrowski, J.M., Reinecke, A.J., 2001b. Spray Weed Science Society of America, 1994. Herbicide Handbook, seventh ed. deposition of two insecticides into surface waters in a South African Champaign, IL. orchard area. J. Environ. Qual. 30, 814–822. Zar, J.H., 1996. Biostatistical analysis, third ed. Prentice-Hall, NJ.