Effects of the organochlorine insecticide lindane (γ-C6H6Cl6) on the population responses of the freshwater amphipod Hyalella azteca
Abstract
Adult populations of Hyalella azteca, with similar reproductive potential, were exposed in the laboratory to a range of lindane concentrations below the 240-h adult LC50 (26.9 μg/L) for a period of 35 d. The reproductive status of each population was assessed through quantification of the precopulatory guarding behavior of Hyalella, the number of gravid females, and their brood sizes. At the end of the experiment, recruitment within the H. azteca populations was assessed, and the body lengths of individuals were measured using image analysis. By employing a range of integrated and quantifiable response criteria, the influence of lindane at the individual and at the population level was determined. A lowest observed effect concentration of 13.5 μg lindane/L was identified. The experimental results are discussed in relation to the reproductive strategies adopted within the treatment groups and with respect to previous toxicity data.
INTRODUCTION
Many investigators have examined the intrinsic toxicity of chemicals using simple response criteria at the individual level of biological organization. Such studies typically involve the use of a small number of animals that are exposed to a range of concentrations, allowing the determination of median effect concentrations or times. The responses employed often include mortality and sublethal physiological or behavioral indicators of stress. The results of such investigations can be used in attempts to predict the effects of pollutant exposure at the individual or population level. However, to provide confidence in the predictive value of such techniques, especially where sublethal responses involving behavior are concerned, it is necessary to demonstrate relationships between the response and changes in the populations that have been exposed to pollutants.
The purpose of this study was to investigate such relationships with the amphipod Hyalella azteca Saussure. This detritivorous crustacean has frequently been used in aquatic toxicity investigations because of its abundance in aquatic ecosystems, ease of laboratory culture, and its sensitivity to a wide range of pollutants [1-6]. Its short generation time and high fecundity make it an especially suitable species for use in the assessment of chronic toxicant effects at the population level.
In a previous investigation, the acute toxicity of the insecticide lindane (γ-hexachlorocyclohexane) was determined using both neonate and adult H. azteca. Sublethal effects based on the disruption of the precopulatory mate guarding reproductive behavior of Hyalella were also examined [7]. The current study aimed to determine the effects of chronic lindane exposure on this amphipod species. The response criteria selected for this investigation included body size, number of gravid females, eggs per female, and the precopulatory behavior of H. azteca. Precopulatory behavior is of ecological importance as it is a prerequisite for reproductive success, is easily quantified, and its disruption has previously been demonstrated as a rapid and sensitive indicator of toxicant-induced stress [7-11]. It is proposed that examination of the precopulatory behavior of H. azteca during a chronic toxicity study would provide valuable information in conjunction with more conventional population response criteria, such as recruitment, on the mode of toxicant action at both the individual and the population level of this species. Results of this study are discussed with respect to other chronic investigations that have employed lindane as a toxicant.
MATERIALS AND METHODS
Culture method
Hyalella azteca was cultured in a controlled environment room maintained at 23 ± 2°C with a 16:8 h light:dark photoperiod and with “true light” broad-spectrum fluorescent tubes to mimic natural daylight. The animals were cultured in 40-L glass aquaria containing 30 L of dechlorinated water and with undergravel filtration. Horse chestnut leaves (Aesculus hippocastanum L.) that had been conditioned for at least 10 d in organically enriched dechlorinated water [12] and sterile cotton gauze were provided as sources of food and cover, and rabbit pellets were added every other day as a dietary supplement [4].
Adult Hyalella azteca were obtained from laboratory cultures maintained as described previously and randomly allocated to produce 12 experimental populations, each consisting of 10 gravid (ovigerous) females and 15 precopulatory pairs. The populations were randomly assigned to 12 6-L plastic tanks that contained 5 L of dechlorinated control water, or test solution, with nominal concentrations of 1, 7.5, and 15 μg lindane/L (i.e., four treatment groups consisting of three replicate populations). The tanks had been initially conditioned for 7 d in the appropriate lindane test solution to minimize loss of test chemical at the start of the study period. The experiment was performed over 35 d in a laboratory maintained at 23 ± 2°C with a 16:8 h light:dark photoperiod. Lindane concentrations in the test solutions were maintained by regularly measuring water concentrations and spiking the treatment tanks, when required, with the appropriate volume of 5 mg/L stock solution. Hyalella were provided with conditioned horse chestnut leaves and sterile cotton gauze as food and as substrate. The control water and lindane test solutions were aerated gently using air stones (25-mm cylinders) situated approx. 3 cm below the surface layer to minimize the surface entrapment of organisms. Lids were placed on the experimental tanks to reduce evaporation during the study.
After 16 d, the survival and reproductive behavior of the populations were monitored by transferring the contents of each tank into a white plastic tray (33 × 24 × 5 cm). The number of adults, gravid females (discernible with the naked eye), and precopulatory pairs were recorded, but the small size and large number of neonates present precluded an assessment of recruitment. Test populations were promptly returned to their respective treatment tanks.
The experiment was terminated after 35 d, at which time the population in each tank was again poured carefully into a tray and the number of gravid females and precopulatory pairs counted. All the animals were then preserved in 70% alcohol and later counted after serial sieving with 2-, 1-, and 0.5-mm mesh sieves. The main purpose of this procedure was to sort the populations into size classes before image analysis to establish the body length distributions of individuals in the experimental populations. The images of individuals from each population were obtained using a Panasonic CCTV color camera (WV-CL320/B) (Matsushita Communication Industrial Company, Tokyo, Japan) in combination with a stereomicroscope connected to an IBM-compatible computer (486SX-33 MHz, 8 MB RAM, SVGA, and a 16-bit Iris video digitising card). The video images were scanned and subsequently captured using Colour Vision 1, True Colour Image Workbench Version 1.17 (Cocoon Software, Roosendaal, The Netherlands) and were stored as computer files. The body lengths of the amphipods were later measured using SigmaScan Image Analysis software (Jandel Scientific Software, San Rafael, CA, USA). Each individual was measured from behind the eye to the tip of the third uropod along the curve of the dorsal surface. The system, once calibrated, stored the body lengths (mm) in memory before being exported in ASCII format to disk. The exported files were then read into Origin Version 3.5 (Microcal Software, Northampton, MA, USA) for statistical and graphical purposes. Following the length measurement of gravid females, the number of eggs was assessed by dissection of the ventral brood pouch using fine forceps.
Toxicant and water quality analyses
The concentrations of lindane in test solutions were determined by extracting aqueous samples (0.45-μm filtered; Whatman International, Maidstone, Kent, UK) into n-hexane and analyzing the solutions by gas liquid chromatography using a Pye Unicam 4500 (Cambridge, Cambridgeshire, UK) with a 5% SE-30 packed column and electron capture detection.
Water quality was maintained throughout the study by weekly replacement of 50% of the control and test solution volume using water siphon tubes (3-mm-diameter PVC) with 5-cm-diameter plastic funnel attachments. The funnel mouths were covered by fine mesh filter (0.1 mm) to prevent the loss of animals from the test systems. Water quality parameters (temperature, pH, conductivity, and dissolved oxygen) were measured with portable meters and water hardness samples (filtered through 0.45 μm) fixed with ARISTAR® nitric acid (BDH, Poole, UK) at the 1% level and measured by flame atomic absorbance spectrophotometry (Model 457, Instrumentation Laboratory, Wilmington, MA, USA) once a week.
| Nominal concentration (μg/L) | Mean measured concentration (μg/L) | SE | Range |
|---|---|---|---|
| 1.0 | 1.07 (n = 24) | 0.04 | 0.87–1.66 |
| 1.0 | 1.05 (n = 24) | 0.04 | 0.83–1.52 |
| 1.0 | 1.08 (n = 24) | 0.05 | 0.80–1.66 |
| 7.5 | 6.94 (n = 23) | 0.12 | 5.97–8.27 |
| 7.5 | 7.03 (n = 23) | 0.068 | 5.84–7.90 |
| 7.5 | 6.83 (n = 23) | 0.113 | 5.84–7.90 |
| 15.0 | 12.96 (n = 23) | 0.274 | 10.76–16.20 |
| 15.0 | 13.87 (n = 23) | 0.244 | 11.67–16.10 |
| 15.0 | 13.68 (n = 23) | 0.269 | 11.93–16.20 |
Data analysis
Specific data requirements were necessary to permit the use of certain statistical tests. One-way analysis and two-way analysis of variance (ANOVA) required homogeneity of variances that was tested using a Bartlett's Box computer macro [13] and normality that was checked using a method in the statistical package Minitab Version 9.2 (Minitab, State College, MA, USA) equivalent to the Shapiro-Wilk test. Data obtained for the number of precopulatory pairs and gravid females following 35 d exposure to control water and lindane treatments were log10 transformed to meet these requirements.
RESULTS
The lindane concentrations measured in each of the test treatment tanks are shown in Table 1. The mean lindane concentrations (with 95% confidence intervals) used when describing the effects of the 1.0-, 7.5-, and 15-μg lindane/L nominal treatments were 1.07 (1.02–1.12), 6.9 (6.8–7.07), and 13.5 (13.19–13.82) μg lindane/L, respectively. The mean values (with standard error) for water quality parameters measured during the 35-d study included the following: temperature 22.3°C (0.04), pH 7.77 (0.7), conductivity 359.6 (1.4) μS/cm, dissolved oxygen 83% (0.7) of the air saturation value, and total water hardness 102.1 (0.9) mg/L as CaCO3. No significant differences (ANOVA, p > 0.05) between the treatment groups were found for these parameters.
The mean population size determined with standard error for each treatment group at the end of the 35-d study is presented in Figure 1. Lindane exposure influenced the recruitment of certain experimental populations. An approx. 19-fold size increase was observed in control populations during the 35-d study with a 16-, 15-, and 5-fold increase occurring in populations exposed to 1.07, 6.9, and 13.5 μg lindane/L, respectively. Parametric statistical analysis determined that populations exposed to 13.5 μg lindane/L were significantly smaller than control populations (ANOVA, p < 0.001; Tukey-Kramer, p < 0.05) after 35 d.
Histograms (Fig. 2) of individual body length distributions in the final populations clearly illustrate the high fecundity of Hyalella with large numbers of individuals in the lower-size class ranges. In the control water, 1.07-, and 6.9-μg lindane/L treatment groups, the populations were composed mainly of recruited juveniles (body length < 3 mm). However, in the H. azteca populations exposed to the highest lindane treatment of 13.5 μg/L, recruitment was dramatically reduced in comparison with other treatment groups.
The mean H. azteca population size with standard error following exposure to control water and lindane treatments for a 35-d period. Horizontal lines indicate the standard error around the control mean. * indicates a significant difference from control values; ANOVA, p < 0.001; Tukey-Kramer, p < 0.05.
The length frequencies were statistically analyzed by dividing the body length data set into four body length categories: 0 to 2.5, 2.5 to 5.0, 5.0 to 7.5, and 7.5 to 10.5 mm. Two-way ANOVA was employed and significant differences were identified with size class, experimental treatment group, and for the interaction of size class and treatment group (p < 0.001). Overall differences therefore exist in the mean number of individuals in the different size classes and there are differences present in size distribution that are related to experimental treatment. One-way ANOVA with multiple comparison procedures determined that there were significantly less amphipods of a body length 0 to 2.5, 2.5 to 5.0, or 7.5 to 10.5 mm from populations exposed to the 13.5-μg/L lindane treatment when compared to control values (ANOVA, p < 0.05; Tukey-Kramer, p < 0.05). However, no significant difference was determined between the number of individuals occupying the 5.0- to 7.5-mm body length category (Table 2).
The success of the Hyalella populations might be attributable to the effects of control and test conditions on the survivorship and/or behavioral strategies adopted by individuals. Survival of the initial adult populations was not significantly affected following 16 d of exposure to control water and lindane test solutions (ANOVA, p = 0.202), with the mortality observed ranging from approx. 5 to 15%. However, lindane exposure did influence the reproductive performance of individuals in the experimental populations. Observation of the precopulatory guarding behavior of sexually mature Hyalella exposed to 13.5 μg lindane/L revealed a highly significant level of impairment (ANOVA, p < 0.001; Tukey-Kramer, p < 0.05) in comparison with control populations at the 16-d sampling period (Fig. 3A). Furthermore, the number of gravid females (those not in a precopulatory pairing) present in the populations exposed to 13.5 μg lindane/L was also significantly reduced (ANOVA, p = 0.069; Tukey-Kramer, p < 0.1) (Fig. 3B). Although the number of neonates in each population was not quantified, during the examination at 16 d qualitative observations were made. These observations indicated that the production of offspring in populations subjected to the highest lindane treatment was reduced.
Length distributions (mean with standard error) of H. azteca populations following a 35-d exposure period to (A) control water, (B) to 1.07 μg lindane/L, (C) 6.9 μg lindane/L, and (D) to 13.5 μg lindane/L.
| Size class (mm) | Lindane treatment (μg/L) | Mean no. individuals (SE) | One-way ANOVAa |
|---|---|---|---|
| <2.5 | Control | 400 (18.7) | |
| 1.07 | 363 (43.0) | p = 0.002, <0.05 | |
| 6.9 | 331 (58.5) | ||
| 13.5 | 81 (29.2)* | ||
| 2.5–5.0 | Control | 230 (25.7) | |
| 1.07 | 181 (28.6) | p = 0.002, <0.05 | |
| 6.9 | 170 (5.7) | ||
| 13.5 | 60 (3.8)* | ||
| 5.0–7.5 | Control | 51 (9.2) | |
| 1.07 | 53 (3.8) | p = 0.400, >0.05 | |
| 6.9 | 57 (6.6) | ||
| 13.5 | 41 (3.2) | ||
| >7.5 | Control | 30 (1.5) | p = 0.022, <0.05 |
| 1.07 | 18 (4.5) | ||
| 6.9 | 25 (1.0) | ||
| 13.5 | 11 (9.0)* |
- aTukey-Kramer multiple comparison. *indicates a significant difference from control values.
At the end of the 35-d study, the precopulatory behavior of Hyalella (Fig. 4A) was again examined and found to be significantly affected by lindane (ANOVA, p = 0.028). Multiple comparison procedures revealed that the number of precopulatory pairs present in the 13.5-μg lindane/L treatment was significantly lower when compared to control population values (Tukey-Kramer, p < 0.1). The number of gravid females (Fig. 4B) was also significantly reduced in populations exposed to the 13.5-μg lindane/L treatment (ANOVA, p = 0.015; Tukey-Kramer, p < 0.05). However, examination of the number of eggs per female (Fig. 4C) did not reveal a significant difference in brood size following lindane exposure (ANOVA, p = 0.630).
DISCUSSION
A range of integrated and quantifiable response criteria was employed during this investigation, as it was thought that through examination of the growth, reproductive success, and behavior of H. azteca links would be revealed between the toxicity process at individual and population levels of biological organization.
Lindane significantly reduced the growth of certain populations of Hyalella during the 35-d experiment and influenced the growth and development of individuals within these affected populations. Following chronic exposure at 13.5 μg lindane/L, the initial populations of 15 precopulatory pairs and 10 gravid female H. azteca increased by approx. five times, whereas a 19-fold increase took place over the same period in populations that had been maintained under control conditions. It can be hypothesized that pollutant-stressed populations comprise individuals experiencing increased metabolic demands to maintain basic homeostatic and reproductive processes [14]. Growth has commonly been used to provide an indication of the fitness of individuals and populations in toxic environments as it represents a composite of all physiological and biochemical processes [15, 16]. For example, lindane has previously been reported to reduce juvenile growth of the European amphipod Gammarus pulex (L.) at 6.1 μg/L in a 14-d study [17]. Therefore, it is likely that H. azteca exposed to lindane in this study might have been unable to sustain the same level of fitness or energetic balance as control individuals. It is probable that toxicant-exposed H. azteca curtailed or modified their reproductive strategies to maximize individual survival until more favorable environmental conditions prevailed or tolerance developed [8]. In the present study, the frequency of precopulatory pair formation in the populations exposed at 13.5 μg lindane/L was significantly reduced, probably because of the high metabolic cost of carrying females for extended periods in combination with toxicant-induced effects on the fitness of male Hyalella [7] (Fig. 4A). It has previously been reported that amphipods are more likely to separate and abandon precopulatory pairings during a toxicant exposure [7, 8, 18-21]. The implications of delayed copulation are that a female exoskeleton hardens, including her oviducts, which must be flexible to allow passage of eggs through to the brood pouch where fertilization occurs [11]. Abandonment of precopulatory behavior at a critical period, such as immediately prior to or during a female's molt, might prevent egg release and fertilization.
The mean number (with standard error) of (A) precopulatory pairs and (B) gravid females present in H. azteca populations following a 16-d exposure period to control water and lindane treatments. * indicates a significant difference from control values; (A) ANOVA, p < 0.001; Tukey-Kramer, p < 0.05; (B) ANOVA, p < 0.069; Tukey-Kramer, p < 0.1.
The mean number (with standard error) of (A) precopulatory pairs; (B) gravid females and (C) eggs per gravid female present in H. azteca populations following a 35-d exposure period to control water and lindane treatments. * indicates a significant difference from control values; (A) ANOVA, p < 0.028; Tukey-Kramer, p < 0.1; (B) ANOVA, p < 0.015; Tukey-Kramer, p < 0.05.
Differences in the reproductive behavioral strategies of H. azteca exposed to 13.5 μg lindane/L may have led to significantly reduced numbers of gravid females in these experimental populations in comparison with control populations (Fig. 4B). Further investigations are required to unequivocally demonstrate the relationship between disruption of precopulatory pairs of H. azteca and a decrease in gravid Hyalella in toxicant stressed populations.
Interestingly, the brood size of gravid females was not significantly influenced following the 35-d lindane exposure (Fig. 4C), although the fecundity of these individuals might be modified by lindane-induced effects on the viability of the brooded eggs and the possiblity of increased egg abortion rates. Because the 240-h LC50 for neonate H. azteca has previously been determined as 9.8 μg lindane/L [7], it is also possible that mortality of young produced during the study might have influenced population recruitment. Support for the interpretation of the results of this experiment can be found in the observations of a 17-week study conducted with Gammarus fasciatus in which no significant precopulatory pairing occurred following exposure to 17.7 μg lindane/L and in which, of only two gravid females observed, no young were produced [22]. In addition, exposure of G. fasciatus to 8.6 μg lindane/L in the same experiment caused reduced precopulatory pairings and the mortality of all young following their release from brood pouches.
Although the reproductive effort of certain populations exposed to lindane was significantly reduced, an increase in population size was seen in all the treatment groups during the 35-d study. Examination of the body length frequencies revealed that significantly fewer individuals were present in the size categories 0 to 2.5, 2.5 to 5.0, and 7.5 to 10.5 mm in the populations exposed to 13.5 μg lindane/L compared to control values, whereas no difference was identified in the 5.0- to 7.5-mm size category. This size class was probably composed of adults from the initial adult population that suffered a reduction in their normal growth rate because of toxicant-induced stress, possibly to permit a degree of reproduction to occur.
In the populations exposed to 13.5 μg lindane/L, it is hypothesised that the young produced during the 35-d study were mainly the result of the reproductive efforts of the original adults. Two discernible peaks in the number of young recruited to these populations are present (Fig. 2D), and it is proposed that the structured precopulatory behavior exhibited by H. azteca, in combination with an overall reduction in amphipod growth rates, has been responsible for this pattern of young production. In the control water, 1.0-, and 7.5-μg/L treatment groups, the general recruitment pattern is different because of the reproductive efforts of both the individuals recruited during the study and the initial population of mature amphipods; H. azteca possess a short life cycle, and consequently in lowstress environments recruited individuals have a rapid growth rate and attain sexual maturity relatively quickly. However, exposure at 13.5 μg lindane/L caused a reduction in the Hyalella growth rate and thereby increased the time required for the young that were produced to reach sexual maturity.
The present investigation examined a range of H. azteca responses and has generated a lowest observed effect concentration (LOEC) of 13.5 μg lindane/L (measured concentration). This value is similar to the LOECs produced from other chronic lindane toxicity studies conducted with freshwater crustaceans: 19 μg/L for Daphnia magna in a 64-d study and 8.6 μg/L in a 17-week study conducted with Gammarus fasciatus based on survivorship and reproductive success [22]. Furthermore, an LOEC of 9.9 μg lindane/L was generated in a life cycle study conducted using Chironomous riparius (Insecta) [23], demonstrating the sensitivity of H. azteca as a toxicity test organism. In sublethal investigations that have examined effects of lindane toxicity on the direct and indirect disruption of the precopulatory behavior of H. azteca, LOECs of 24.4 and 17.3 μg lindane/L were determined, respectively [7]. These studies were conducted over 24 and 4 h, respectively, and demonstrate that sensitive and relevant information can be obtained through the use of behavioral bioassays at the individual level when suitable response criteria are employed.
Investigations examining the sublethal effects of toxicants on populations are often labor intensive. However, they can reveal an ecologically useful insight into toxicity processes and information suitable for the interpretation of behavioral and sublethal effects at both the individual and the population level of a species. It can be concluded that the reduction in population growth observed during this study resulted from a combination of toxicant effects: disruption of the reproductive behavior patterns of adult H. azteca and a reduction in the growth of recruited individuals and consequently their delayed sexual development. For species such as H. azteca with relatively short generation times, large discrepancies in recruitment patterns soon arise.
Acknowledgements
This investigation received financial support from Zeneca Agrochemicals, Ecological Risk Assessment Section, Jealott's Hill Research Station, Bracknell, Berkshire, United Kingdom. The authors wish to thank Timothy Kedwards and Una Goggin for assistance with the image analyses and the University of Ghent and Zeneca Agrochemicals for supplying Hyalella azteca for the establishment of laboratory cultures at Cardiff.
