Environmental Toxicology and Chemistry, Vol. 22, No. 10, pp. 2476–2481, 2003 q 2003 SETAC Printed in the USA 0730-7268/03 $12.00 1 .00

A MODEL TO UNDERSTAND THE CONFOUNDING EFFECTS OF NATURAL SEDIMENTS IN TOXICITY TESTS WITH CHIRONOMUS RIPARIUS ALEXANDRE R.R. PE´RY,*† VANESSA SULMON,† RAPHAE¨L MONS,† PATRICK FLAMMARION,† LAURENT LAGADIC,‡ and JEANNE GARRIC† †Laboratoire d’e´cotoxicologie, Cemagref, 3bis quai Chauveau, CP 220, 69336 Lyon, Cedex 9, France ‡Laboratoire d’e´cotoxicologie aquatique, INRA, Campus de Beaulieu F-35042 Rennes Cedex, France ( Received 9 July 2002; Accepted 25 February 2003) Abstract—Recently, we built a model to link feeding input with growth, emergence, and reproduction of the midge Chironomus riparius exposed to an artificial sandy sediment. This model is based on assumptions about both feeding behavior and use of energy. Here, we show how it can be used for toxicity tests with natural sediments to understand and model the influence of sediment characteristics. We measured growth, emergence, and reproduction of chironomids exposed in beakers to four unpolluted natural sediments and three feeding conditions (no feeding, 0.2 mg Tetramint/larva/d, and 1.4 mg Tetramin/larva/d) and compared the results with data obtained on our artificial sandy sediment. Sediment characteristics had lower influence on growth than feeding level, but their influence could not be neglected. First, we could distinguish between sandy sediments and other sediments. This difference resulted in a significant delay of about 18 h in the growth curves. Second, in case of food limitation, chironomids could use the organic materials in the sediment, provided that the C:N ratio of the sediment was less than 14. Our model proved to be able to incorporate those two phenomena. As for reproduction, we observed a better reproduction (measured in number of eggs per mass) for natural sediments than for artificial sediments. We showed that this difference could be due to the lipid content of the natural sediments. Keywords—Chironomus riparius

Natural sediment

Growth

Reproduction

Model

ence survival, growth, and reproduction of the midges and interfere with feeding levels [8,10,11]. Ristola et al. [8] showed that organic content was a confounding factor but not the only one. In their experiment, larvae grew best in sandy sediments with low organic carbon content (0.5%). In this paper, we show how we can adapt our model to understand and quantify the influence of sediment physicochemical characteristics on chironomid growth, emergence, and reproduction. To reach this goal, we performed growth, emergence, and reproduction tests and built a model to describe the data. We also measured some characteristics of the sediments: particle size distribution, loss on ignition, and organic and nitrogen content. We finally show how our study can help carry out relevant toxicity tests with natural sediments or help analyze toxicity tests with natural sediments.

INTRODUCTION

Chironomids are commonly used to assess the toxicity of natural sediments [1,2] for two reasons. First, they represent a prominent part of benthic communities in all types of freshwaters. For instance, Berg and Hellenthal [3] reported an annual chironomid secondary production in an American stream (northern Indiana) that accounted for 80% of the total insect secondary production. Second, chironomids have a number of characteristics that make them valuable for toxicity tests [4]: The life stages are easy to identify, and the life history under laboratory conditions is short and well known. Here, we focus on Chironomus riparius (Meigen), which is a nonbiting midge widely distributed in the northern hemisphere at temperate latitudes. It can be found in both lentic and lotic environments, mostly in eutrophic or organic enriched waters [5]. Its life cycle comprises aquatic stages (egg, four larval instars within the sediment, pupa) and an aerial adult stage. The larvae feed on sediment-deposited detritus [6]. Feeding quantity can have a substantial influence on the outcome of assays. Sibley et al. [7] and Ristola et al. [8] showed the influence of food level on growth of C. riparius and on growth and reproduction of C. tentans, respectively. In a previous paper [9], we also showed this phenomenon and provided modeling tools to analyze the influence of food. Our models are based on assumptions about the use of energy by the chironomids, in favor of which we provided experimental arguments. Sediment physicochemical characteristics, like particle size distribution or organic matter content, can also greatly influ-

MATERIALS AND METHODS

Sampling sites and sediment characterization All the selected sediments came from unpolluted sites regularly analyzed by our institute (Cemagref, Lyon, France) for contaminants to confirm their potential use as reference sediment. We collected three of them with a 1-L grab sampler, the first one in Port-Galland (France) in a tributary of the River Ain and the second and third ones in Beaujeu (France) near the source of the River Ardie`res. The fourth sediment we used came from Lake Aiguebelette (France) and was provided by the Laboratoire des Sciences de l’Environnement in Vaux-enVelin (France). The sediments are designated in this paper as Port-Galland, Beaujeu I, Beaujeu II, and Aiguebelette. Sediments were 2-mm sieved and homogenized before use. Particle size distribution was determined in the Laboratoire d’e´codynamique des Sediments of our institute by wet sieving.

* To whom correspondence may be addressed ([email protected]). 2476

Influence of sediment on tests with Chironomidae

The percentage of dry matter, loss on ignition, nitrogen, and organic content were determined by the Laboratoire de Chimie des Eaux with an elemental analyzer (Carlo Erba Instruments EA 1108 CHNS-O, Milan, Italy).

Growth experiments Chironomids had been cultured prior to the experiments according to standard methods. For each experiment, 10 2-dold organisms were added at random to 0.55-L beakers (0.11 L of silica sand and 0.44 L water). Midges were fed with Tetramint fish food (Tetrawerke, Melle, Germany). Each day of measurement, three beakers per feeding and sediment conditions were taken to measure length, using a binocular microscope Olympus SZH10 (Olympus, Melville, NY, USA) fitted with a calibrated eyepiece micrometer. The organisms had to be killed with formalin before measuring their length. The beakers were set in a water bath at 218C with a 16:8h light:dark photoperiod. Conductivity, temperature, pH, and amount of dissolved oxygen, nitrates, and nitrites were measured daily. Because the amount of food could dramatically affect water quality, we used an aeration system. Three growth experiments were carried out. Experiment 1: Natural sediment without feeding. Chironomids were exposed to the natural sediments we had selected without any feeding to assess the intrinsic potential of the organic content of the sediments. Measurements were done at days 0, 4, 6, 8, and 10 after the beginning of the experiment. Experiment 2: Natural sediment with feeding ad libitum. Chironomids were exposed to the natural sediments we had selected with a feeding level of 1.4 mg Tetramin/larva/d, which leads to no food limitation during the whole life cycle of C. riparius [9]. In this way, organic content should have little influence, which made possible the assessment of the effects of other confounding factors, like particle size distribution. Measurements were done at days 0, 2, 3, 4, 5, 6, and 8 after the beginning of the experiment. Experiment 3: Natural sediment with a little amount of food. Chironomids were exposed to the natural sediments we had selected with a feeding level of 0.2 mg Tetramin/larva/d. Measurements were done at days 0, 2, 3, 4, 5, 6, and 8 after the beginning of the experiment.

Emergence and reproduction experiment Four beakers per feeding and sediment conditions were used for the emergence and reproduction experiment in the same conditions as for growth tests with feeding levels 0.2 and 1.4 mg/larva/d. Four more beakers were used with an artificial sediment (silica sand), a feeding diet of 1.4 mg/larva/d, and an additional input of 0.2 mg/larva/d of cholesterol (SigmaAldrich Chemie, Steinheim, Germany) to test if any difference in reproduction between natural and artificial sediments could be due to lipids. We chose lipids because Tetramin has a high glucide and protein content but a poor lipid content. The beakers were covered to prevent adults from escaping. Emergence was measured every day. The females were put into 1-L bottles, with 0.1 L water, with males from laboratory culture in a ratio of three males per female. No more than six females were placed in the same bottle. After mating and oviposition, each egg mass was removed and put into a 5-ml tube with 2 ml H2SO4, 2 N, overnight. The following day, the tubes were agitated to dissociate the eggs and then were counted using a binocular microscope. For each mass, measurements

Environ. Toxicol. Chem. 22, 2003

2477

were made three times to reduce experimental errors, and we took into account the average of the measurements.

Model for growth in artificial sandy sediments Our growth model is based on the two following assumptions, which have been tested successfully in Pe´ry et al. [9]: First, midges have low maintenance costs. In other words, all the food absorbed is converted into biomass. Second, the midge is isomorphic during its development. During the growth period, length remains proportional to width. As a consequence, the weight of an organism is proportional to the cube of its length. We use a discrete model with a time unit of 1 d for three reasons. First, the photoperiod is a very important factor for chironomids [12], and the day would be a relevant time unit with respect to the behavior of these organisms. Second, we fed the animals once per day. Third, when organisms are food limited, we know what we put only at day j, and we know that no food remains at day j 1 1. We do not have information about the way food density varies between consecutive feeding events.

Model for growth in nonlimiting food conditions Under conditions of no food limitation, we showed [9] that growth (as length) can be described by the following equation:

d 3 d l } l2 ⇔ l 5 a dt dt

(1)

where l is the mean length of the organisms, a is a constant depending on the instar and on the sex of the chironomids, and dt represents a 1-d period. Equation 1 represents the uptake of energy, which is proportional to the surface of the organism and its use to increase the volume, which is proportional to the cubic length of the organism because of assumption 2 (isomorphism). No difference in length was observed between male and female larvae until the fourth instar, at which point a significant difference appears [12]. During this period, a is thus expected to be different between males and females, but no difference between sexes is expected for second and third instars. Length growth is assumed to stop as soon as a maximum length, called lmax, is reached. In Pe´ry et al. [9], we found that for second and third instars, growth was linear with respective growth rates of 0.81 and 1.42 mm/d. For the fourth instar, the growth of males and females was linear with respective growth rates 1.72 and 2.21 mm/d. The limit mean length for males was 11.36 mm and for females 13.72 mm. The second and third stages lasted 2 d each.

Model for food limited growth As maintenance costs are low (assumption 1), the daily increase in weight is, in case of food limitation, just equal to the daily amount of food introduced into the beaker. The equation describing the situation is thus

Wn11 2 Wn 5 g · (ln311 2 ln3 ) 5

u 3Q N

(2)

where Q (in mg) represents the daily quantity of food introduced into one beaker, N the number of larvae, ln (in mm) the larval length at day n, Wn (in mg) the individual weight at day n, u the percentage of Tetramin that can effectively be incorporated by the chironomids, and g the proportionality factor

2478

Environ. Toxicol. Chem. 22, 2003

A.R.R. Pe´ry et al.

between the weight and the cube of the length. The estimate of u and g were, respectively, 0.691 and 0.523 in Pe´ry et al. [9]. Food limitation had another consequence. Under a feeding level of 0.3 mg/larva/d, the second instar lasted 3 d instead of 2 d in ad libitum feeding conditions. To make growth pattern predictions, we calculate the length day after day. To do this, we use Equation 1 to deduce length at day j 1 1 from length at day j assuming that the organisms are not food limited. We then calculate the cubic length increase associated with this length increase and check with Equation 2 that this is coherent with the actual feeding input. If not, we conclude that the organisms are food limited and use Equation 2 to deduce length at day j 1 1 from length at day j. In Pe´ry et al. [9], the model for food limitation proved to be able to predict length growth pattern for a large range of diets and organism densities.

Adaptation of the model for organic content of natural sediments For growth in ad libitum feeding conditions, we do not expect any influence of sediment organic content. If no food is given, growth can be possible in natural sediment. We assume that organic matter content does not vary during the test, which means that feeding level is constant during all the test, then Equation 3 describing length growth is similar to Equation 1 with an energy uptake proportional to the surface of the organism and a proportionality factor, b, which depends on the instar and on organic content [13]:

d 3 d l } l2 ⇔ l 5 b dt dt

(3)

For feeding in limiting conditions, we propose a model that links the basic model and the observations in no-feeding conditions. When the chironomids are food limited, we consider that the organisms first use the food input. The length increase can be calculated with Equation 2. We estimate the time needed to incorporate this food with Equation 1: If dl is the calculated length increase, then the time to incorporate the food t (in days) is

t5

dl a

Table 1. Particle size distribution (in percentage) and organic content for the natural sediments and the artificial sediment (silicate) we used. For the latter, we take into account all the food introduced in the beaker since the beginning of the experiment. All sample sites are located in France Sample Port-Galland Beaujeu I Aiguebelette Beaujeu II Silicate

Beaujeu II Port-Galland Beaujeu I Aiguebelette Silicate 1 low feeding level Silicate 1 high feeding level

,2,000 to .500 mm

,500 to .200 mm

0.4 0.6 7.7 67.7 0

0.4 1.2 17.6 27.5 0

Loss on ignition (%)

Organic carbon (%)

Organic nitrogen (%)

C:N

,0.2 3 8.6 22.5

,0.1 0.35 0.79 1.13

8.6 10.9 19.9

,1 10 20.5 40.5

Statistical methods To compare observations for a given time, a given feeding level, and a given sediment and model predictions, we performed Student’s tests with the value predicted and the mean of the length measures. With three beakers, about 30 length measures should be taken per data point. We previously checked, using the analysis of variance method, that no significant difference (p . 0.05) existed between the results from the three beakers.

Model for emergence and reproduction For emergence data, we showed [9] that the mean day of emergence is equal to 4.5 d plus the day for which the maximum length is reached. Using our growth model, we can then deduce the theoretical mean emergence time.

7.1 27.6 42.1 1.6 90

,50 mm

92.1 70.6 32.7 3.2 10

0.0007

0.0012

5.9

0.049

0.0084

5.9

For reproduction data, we showed [9] that the amount of eggs produced per female was proportional to the amount of food given per remaining female larvae during the period between the end of the growth of the females and emergence, which may be a specific period of energy investment into reproduction. The number of eggs per mass reached a plateau for high feeding levels. Thus, the equation describing the number of eggs per mass was

N 5 max(Q 3 b, Nmax)

(5)

where Q is the amount of food available between the end of the growth and the pupal stage, N the mean number of eggs per mass as a function of Q, Nmax the maximum mean number of eggs, and b the food–eggs conversion factor. Our model provided a very good description of the data with Nmax eggs per mass and b 5 629.2 eggs/(mg/d).

(4)

We believe that the rest of the time (1 2 t), the chironomids feed on natural sediment. We consequently multiply (1 2 t) by the length growth rate estimated in experiment 1, b, and we add this length increase to the one due to food.

,200 to .50 mm

RESULTS

Sediment characteristics Particle size distribution and organic content of the sediments are presented in Table 1. We can classify the sediments from the finer to the coarser: Port-Galland–Beaujeu I–Silicate (artificial sediment)–Aiguebelette–Beaujeu II. We can classify the sediments from the poorer to the richer in organic materials: Beaujeu II–Port-Galland–Beaujeu I–Aiguebelette. We also indicate the organic content of silicate with low and high feeding levels. To present it, we take into account all the food introduced during the experiment, which overestimates the actual organic content along the experiment. However, even when food was applied, the organic content of the silica sand was less than that of any of the natural sediments.

Physicochemical conditions Temperature was constant (21 6 0.38C), as was pH (between 8.1 and 8.4). Conductivity was between 350 and 420 ms/cm, and the percentage of dissolved oxygen was always above 65%. Nitrate levels rose twice slightly above 2 mg/L during the growth experiments with food and natural sedi-

Influence of sediment on tests with Chironomidae

Fig. 1. Mean length of chironomids in natural sediments without feeding as a function of time (diamonds: Aiguebelette; triangles: PortGalland; white circles: Beaujeu I; black circles: Beaujeu II). The line represents the growth in silicate with a low feeding level (0.2 mg/ larva/d). All sample sites are located in France.

ments. Each time, we consequently renewed 4/5 of the water in the beakers.

Growth experiments Experiment 1. Survival was low, with less than 20% for Aiguebelette sediment, about 60% for Port-Galland sediment, and about 90% for Beaujeu I and Beaujeu II sediments. These results could be partly due to the fact that small organisms are very difficult to recover from natural sediments. Some undetermined physicochemical parameter or other stressor could be involved. Indeed, growth was very slow (Fig. 1) compared to growth in silicate with a low amount of feeding (0.2 mg Tetramin/larva/d). We can distinguish between Aiguebelette sediment, for which no growth was detected after 10 d of experiment, and Beaujeu I, Beaujeu II, and Port-Galland sediment, which allowed larval development even without feeding. The estimate, obtained with linear description, of the growth rates for these sediments was 0.33, 0.34, and 0.37 mm/ d, respectively, for Port-Galland, Beaujeu I, and Beaujeu II. These growth rates were not significantly different from each other (p . 0.05). Experiment 2. When comparing observed data and the model for silicate for feeding input of 1.4 mg/larva/d, we realized that only the observations for Beaujeu I and Beaujeu II were always not significantly different (p . 0.05) from the model. For Port-Galland and Aiguebelette, we observed a quasi-constant decay in length between model and observed curve. Growth seemed to begin in those two sediments hours after what was observed with silicate, the growth curves remaining parallel during the whole experiment. We estimated the delay using a least-squares method at 0.6 mm (which is equivalent to 18 h for second-instar larvae or 7 h for fourth-instar larvae). When we introduced this delay in the model, the fitting became quite correct (Fig. 2): Only 2 out of 18 observed mean lengths were different (Student’s tests performed with length measures, p , 0.05) from the model predictions (Aiguebelette at day 8 and Beaujeu I at day 4). Experiment 3. For the test with a low feeding level, no significant difference was observed between Aiguebelette data and the predictions using the model with a decay but not taking into account organic content, except for one data point (Fig. 3). This is not a surprise because experiment 1 showed that it was not possible for the chironomids to compensate for a lack of food in sediments of high organic carbon content. On the contrary, in Beaujeu I, Beaujeu II, and Port-Galland sediments,

Environ. Toxicol. Chem. 22, 2003

2479

Fig. 2. Mean length of chironomids in natural sediments with feeding ad libitum as a function of time (diamonds: Aiguebelette; triangles: Port-Galland; white circles: Beaujeu I; black circles: Beaujeu II). The plain line represents the model for growth in silicate, and the dotted line represents this model with a decay of 0.6 mm. The letters a and b1 indicate significant differences between model and, respectively, Aiguebette and Beaujeu I data. All sample sites are located in France.

eight out of nine of the mean lengths were significantly higher (p , 0.05) than what the model predicts (Fig. 3) as soon as the organisms were food limited (at day 7, as shown in Pe´ry et al. [9]). This observation was not really surprising, for experiment 1 indicated that the organic material of these sediments could be used for growth of the chironomids. Figure 4 presents the data together with the predictions obtained with the model taking into account the influence of organic content. Now, only one among the nine data points for the period of food limitation is significantly different from the model predictions.

Emergence and reproduction experiment Emergence. Table 2 presents for males and females the comparison between the predictions obtained by our model (the one that takes into account organic content) and the observed mean emergence times, emergence time being the time between the beginning of the experiment and emergence. As our model for food limitation is discrete, our confidence interval for predictions cannot be less than 1 d. Consequently, we could not detect any significant difference between observed data and model predictions. The results confirmed the ability of

Fig. 3. Mean length of chironomids in natural sediments with feeding 0.2 mg/larva/d as a function of time (diamonds: Aiguebelette; triangles: Port-Galland; white circles: Beaujeu I; black circles: Beaujeu II). The plain line represents the basic model for growth in silicate, and the dotted line represents this model with a decay of 0.6 mm. The letters a, b1, b2, and pg indicate significant differences between model and, respectively, Aiguebette, Beaujeu I, Beaujeu II, and PortGalland data. All sample sites are located in France.

2480

Environ. Toxicol. Chem. 22, 2003

A.R.R. Pe´ry et al.

Fig. 4. Mean length of chironomids in natural sediments with feeding 0.2 mg/larva/d as a function of time (triangles: Port-Galland; white circles: Beaujeu I; black circles: Beaujeu II). The plain line represents the model for growth in silicate that incorporates the effect of sediment organic content, and the dotted line represents this model with a decay of 0.6 mm. The letter b1 indicates significant difference between model and Beaujeu I data. All sample sites are located in France.

Beaujeu I, Beaujeu II, and Port-Galland sediment to provide better growth than our artificial sediment under conditions of food limitation. Reproduction. No significant difference was observed between the number of eggs per mass among the three natural sediments. In high feeding conditions, it was 505 6 110, 520 6 110, 495 6 100, and 540 6 120 for, respectively, PortGalland, Beaujeu I, Beaujeu II, and Aiguebelette (more than 12 data per sediment). In low feeding conditions, it was 305 6 40, 350 6 50, and 320 6 50 for, respectively, Port-Galland, Beaujeu I, and Beaujeu II (more than 10 data per sediment). Unfortunately, we could not obtain enough reproduction data for Aiguebelette (only two). These values were significantly higher (p , 0.01) than what we obtained with silicate in the same feeding conditions (410 in ad libitum conditions and 245 in the low feeding conditions). The observed increase for both feeding levels is about 25%. When feeding was supplemented with lipids, the difference was no longer significant (p . 0.05), with a mean number of eggs per mass of 496 6 68 for ad libitum conditions and 340 6 40 for low feeding conditions. DISCUSSION

Influence of the sediment particles on chironomid growth When feeding conditions were not limiting, we observed a delay between Beaujeu I, Beaujeu II, and silicate sediments compared to the other ones. As food was provided ad libitum, organic content should not be responsible for this phenomenon. Particle size distribution is also a poor descriptor of the phenomenon because Beaujeu I and Beaujeu II have very different particle size distributions (see Table 1). We suggest that chi-

Table 2. Emergence time of the chironomids as a function of sex and feeding level. Model predictions are indicated between brackets. All sample sites are located in France

Sample Port-Galland Beaujeu I Beaujeu II Aiguebelette

High feeding level

Low feeding level

Males

Females

Males

Females

11.1 (11.5) 11.1 (11.5) 11.2 (11.5) 10.9 (11.5)

11.5 (11.5) 11.9 (11.5) 12.0 (11.5) 11.7 (11.5)

13.7 (13.5) 13.2 (13.5) 13.1 (13.5) 14.5 (14.5)

16.8 (16.5) 16.3 (16.5) 16.3 (16.5) 18.0 (18.5)

ronomids show a delay in the growth pattern when exposed to sediment that was not sandy. Indeed, Port-Galland sediment is silt, and despite its particle size distribution, Aiguebelette sediment is a very dark sediment that can hardly be considered sandy. The other sediments are visually very close to silicate artificial sediment. Other authors indicate the importance of this factor on chironomid growth [8,10]. Ristola et al. [8] found that chironomids grew better, whatever the feeding level was, in a sandy sediment compared to three finer muddy sediments. Winnel and Jude [14] have also shown in an in situ sampling that chironomids had a preference for sandy sediments. Naylor and Rodrigues [15] proposed an explanation. The particle type could influence the difficulty for the chironomids to build tubes. As this building of tubes is of vital importance (the authors observed that larvae that have been disturbed resumed feeding only once they had constructed a new tube), the difference in times to construct tubes could explain the decay observed since the beginning of the experiment. Here, when introducing such a decay in our model, we could describe properly the growth data.

Influence of organic matter content on chironomid growth, emergence, and reproduction Our study confirmed that feeding has a greater influence than sediment organic matter content, even if, in the experiment with a feeding of 0.2 mg/larva/d, food always represented less than 10% of the total amount of organic materials. This is no surprise, for it has been shown that food is considerably better used by chironomids if it is available on the sediment surface than if it is mixed with the sediment [15]. Nevertheless, we also showed that the influence of sediment organic matter content cannot be neglected under conditions of food limitation. Adults emerged 1 or 2 d earlier when the larvae could use the organic content of the sediment. We also showed that growth was possible without feeding in certain conditions. We can give an explanation for the results we found. Growth without feeding was possible for Beaujeu I, Beaujeu II, and Port-Galland but not for Aiguebelette. However, the latter sediment contained more organic nitrogen and organic carbon than the two other sediments. Consequently, these factors are poor descriptors of the potential use of organic materials by the chironomids. On the contrary, the C:N ratio appears to be a better descriptor. In our study, Port-Galland and Beaujeu I had a comparable C:N ratio (10.9 and 8.6, respectively), but the C:N ratio of Aiguebelette was 19.9. Ristola et al. [8] also studied four natural sediments and observed growth without feeding only for a sediment with a C:N ratio of 13.6. The C:N ratios of the other sediments were 62.5, 25.6, and 19.7. The C:N ratio has also been used by other authors to explain the influence of food quality. Vos et al. [16] showed that food with high C:N ratios (between 10 and 12.6) were correlated with a lower length for the chironomids compared to food with lower C:N ratios (between 4.7 and 5.9). Similar results were found by Dorgelo and Leonards [17] for the snail Potamopyrgus jenkinsi (E.A. Smith). They found that growth rate plotted against the C:N ratios of the different diets they used followed an optimum curve with a maximum at a C:N value of 15.9. In this study, by adapting our model to incorporate the influence of organic content, we were able to describe the growth pattern and to predict the emergence data. More experimental work is needed, however, to more precisely define the contribution of C:N ratio to growth.

Influence of sediment on tests with Chironomidae

Using our results to analyze toxicity tests with natural sediments Our study not only determined the confounding factors in sediment toxicity tests but also proposed models to take them into account and quantify their influence. As it is difficult to find reference sediments with the same characteristics that tested sediments, our models could be used to assess if a difference in the endpoint measurements is due to toxicity or sediment characteristics. In addition, we observed a significant difference between natural sediments and silicate concerning reproduction (mean number of eggs per mass) but no difference between the natural sediments we used. A reference sediment could thus underestimate the reference mean number of eggs per mass (by about 20%) for a tested natural sediment. We showed that lipids might be responsible for this phenomenon or at least could help compensate for the difference between artificial and natural sediments. Two possibilities exist to take this phenomenon into account in reproduction toxicity tests. We can either add lipids during the test or try to assess a relationship between lipid content of the sediment and reproduction data. We found that relatively poor sediments (Beaujeu II and Port-Galland) allowed the increase of reproduction efficiency, suggesting that only a little amount of lipid is needed. More work is necessary here. An interesting perspective for us could thus be to study the influence of food types on chironomid growth to study if C:N and lipid contents are sufficient descriptors of feeding influence or if other descriptors, such as protein content, are needed. Acknowledgement—The authors would like to thank three anonymous reviewers for their comments; Bernard Migeon, who built the experimental system; and Marc Bray and Herve´ Que´au, who helped perform the experiments. REFERENCES 1. Giesy JP, Rosiu CJ, Graney RL, Henry MG. 1990. Benthic invertebrates bioassays with toxic sediment and pore water. Environ Toxicol Chem 9:233–248. 2. Pellinen J, Soimasuo R. 1993. Toxicity of sediments polluted by the pulp and paper industry to a midge (Chironomus riparius Meigen). Sci Total Environ 151(Suppl.):1247–1256. 3. Berg MB, Hellenthal RA. 1992. The role of chironomidae in energy flow of a lotic system. Neth J Aquat Ecol 26:471–476.

Environ. Toxicol. Chem. 22, 2003

2481

4. Anderson RL. 1980. Chironomidae toxicity tests—Biological background and procedures. In Buikema AL, Cairns J Jr, eds, Aquatic Invertebrate Bioassays. STP 715. American Society for Testing and Materials, Philadelphia, PA, pp 70–80. 5. Beck WM. 1977. Environmental requirements and pollution tolerance of common freshwater chironomidae. EPA/600/4-77-024. U.S. Environmental Protection Agency, Cincinnati, OH. 6. Rasmussen JB. 1984. Comparison of gut content and assimilation efficiency of fourth instar larvae of two coexisting chironomids, Chironomus riparius Meigen and Glyptotendipes paripes (Edwards). Can J Zool 62:1022–1026. 7. Sibley PK, Benoit DA, Ankley GT. 1997. The significance of growth in Chironomus tentans sediment toxicity tests: Relationship to reproduction and demographic endpoints. Environ Toxicol Chem 16:336–345. 8. Ristola T, Pellinen J, Roukolainen M, Kostamo A, Kukkonen JVK. 1999. Effect of sediment type, feeding level, and larval density on growth and development of a midge (Chironomus riparius). Environ Toxicol Chem 18:756–764. 9. Pe´ry ARR, Mons R, Flammarion P, Lagadic L, Garric J. 2002. A modeling approach to link food availability, growth, emergence, and reproduction for the Midge Chironomus riparius. Environ Toxicol Chem 21:2507–2513. 10. Ankley GT, Benoit DA, Balogh JC, Reynoldson TB, Day KE, Hoke RA. 1994. Evaluation of potential confounding factors in sediment toxicity tests with tree freshwater benthic invertebrates. Environ Toxicol Chem 13:627–635. 11. Suedel BC, Rodgers JH Jr. 1994. Response of Hyalella azteca and Chironomus tentans to particle-size distribution and organic matter content of formulated and natural freshwater sediments. Environ Toxicol Chem 13:1639–1648. 12. Ineichen H, Riesen-Willi U, Fisher J. 1979. Experimental contributions to the ecology of Chironomus (Diptera): II. The influence of the photoperiod on the development of Chironomus plumosus in the 4th larval instar. Oecologia 39:161–183. 13. Kooijman SALM. 2000. Dynamic Energy and Mass Budgets in Biological Systems. Cambridge University Press, Cambridge, UK. 14. Winnell MH, Jude DJ. 1984. Associations among Chironomidae and sandy substrates in nearshore Lake Michigan. Can J Fish Aquat Sci 41:174–179. 15. Naylor C, Rodrigues C. 1995. Development of a test method for Chironomus riparius using a formulated sediment. Chemosphere 31:3291–3303. 16. Vos JH, Ooijevaar MAG, Postma JF, Admiraal W. 2000. Interaction between food availability and food quality during growth of early instar chironomid larvae. J North Am Benthol Soc 19: 158–168. 17. Dorgelo J, Leonards PEG. 2001. Relationship between C/N ratio of food types and growth rate in the snail Potamopyrgus jenkinsi (E.A. Smith). J North Am Benthol Soc 20:60–67.