Aquatic Insect Adults as Indicators of Organochlorine Contamination

J. Great Lakes Res. 15(4):623-634 Internal. Assoc. Great Lakes Res., 1989

AQUATIC INSECT ADULTS AS INDICATORS OF ORGANOCHLORINE CONTAMINATION

Zsolt E. Kovats and Jan J. H. Ciborowski Department of Biological Sciences University of Windsor Windsor, Ontario N9B 3P4

ABSTRACT. The aquatic larval stages of many benthic invertebrates are useful indicators of organochlorine contamination. However, it is often difficult to obtain adequate material for gas chromatographic analyses using benthic sampling methods. Alternatively, one can collect the terrestrial adult stages of aquatic insects. We captured adult Trichoptera and Ephemeroptera at night at four contaminated sites on the Detroit and St. Clair rivers, and at several uncontaminated central Ontario locations. Modified Pennsylvania style light traps containing dry ice as a killing/preserving agent attracted large numbers of insects. Adult Trichoptera were active throughout the summer, but catches were greatest in late June. Hexagenia (Ephemeroptera: Ephemeridae) abundance was very low except during 2 weeks in late June. Air temperature, wind velocity, and wind direction greatly affected catches. Gas chromatographic analyses of central Ontario samples revealed low but detectable « 1 p,g/kg) levels of individual PCB congeners, and relatively high concentrations of some pesticides. Samples weighing 0.75 g dry weight yielded contaminant estimates as precise as larger sarnples. Neither storage time nor storage temperature influenced analyses. Adult insects collected near the Detroit and St. Clair rivers carried significantly higher body burdens of nearly all contaminants considered than adult insects from central Ontario sites. Spatial pattern of contaminants among insect samples corresponded to that in the sediments. Adult aquatic insects are effective alternatives to immature benthic insects or other invertebrates for preliminary surveys of organic contamination in aquatic habitats. ADDITIONAL INDEX WORDS: Toxic substances, chlorinated hydrocarbons, Trichoptera, benthos, benthic environment.

INTRODUCTION

erties of the contaminants also influence uptake (Larsson 1984, Reynoldson 1987). Consequently, benthic invertebrates are particularly useful as indicators of degree of sediment contamination (Kauss and Hamdy 1985, Tanabe et al. 1987). However, benthic invertebrate larvae in large lakes and rivers are difficult to sample, owing to their patchy distribution in sediment (Downing 1979), and the necessity for using specialized collecting equipment. Since animals are often collected with large amounts of sediment, extensive processing may be required prior to analysis. As a result, collection of adequate biomass (5 g fresh wt.) for analysis by gas chromatography (GC) is often impractical. Ciborowski and Corkum (1988) proposed collecting the winged, night-active adult forms of bottom-dwelling insects using ultraviolet lights as an alternative to benthic sampling. Because the

Benthic aquatic invertebrates living in contaminated habitats accumulate organochlorine (OC) compounds. Uptake of these contaminants has been documented in freshwater mussels (Kauss and Hamdy 1985), oligochaete worms (Oliver 1984), Chironomidae (Larsson 1984), crustaceans (Sanders and Chandler 1972), and caddisfly larvae (Bush et al. 1985). Because these animals comprise a significant proportion of the diet of predatory invertebrates and fishes, benthic invertebrates are an important transfer route between contaminated sediments and higher trophic levels. Bottom-dwelling larvae of aquatic invertebrates tend to accumulate OC compounds in proportion to the amounts present in the surrounding sediments, although specific attributes of the organisms, the type of sediment, and the chemical prop623

KOVATS and CIBOROWSKI

624

TABLE 1. Locations of sampling stations. Stations 1 and 2 are similar to Stations 2 and 4, respectively, of Ciborowski and Corkum (1988). Stn 1 2 3 4

River or Lake

Designation

Latitude (North)

Longitude (West)

Detroit R. Detroit R. St. Clair R. St. Clair R. Ausable R. Gull R. Balsam L. Scugog L.

River Canard East Windsor Sombra Sarnia Ausable R. Gull R. Balsam L. Scugog L.

42°11'48" 42°20'27" 42°42'02" 42°54'12" 43°05'10" 44°58'11" 44°34'46" 44°09'37"

83°06'13" 82°56'56" 82°29'03" 82°27'29" 81 °48'47" 78°40'58" 78°47'02" 78°49'02"

adults are shortlived and neither feed nor defecate, they retain the contaminant burden accumulated during their larval lifetime. Mauck and Olson (1977), Clements and Kawatski (1984), and Ciborowski and Corkum (1988) analyzed adults of the burrowing mayfly Hexagenia (Ephemeroptera) for OC contaminants and detected levels corresponding to those in the sediments of the lakes and rivers sampled. Analyses of caddisfly (Trichoptera) adults also yielded similar results (Ciborowski and Corkum 1988). However, all of these studies provided subjective estimates of local contamination. A monitoring tool that is to be of broad applicability requires standardized collection techniques and evaluation of the precision of the data thus generated. The objectives of this study were to design a light trap that permits efficient collection of adult aquatic insect samples, assess the seasonal availability of numerically dominant aquatic insect taxa, evaluate the analytic precision associated with samples of different sizes, and compare contaminant burdens in animals collected from contaminated vs. uncontaminated areas. MATERIALS AND METHODS Sample Collection Light trap collections of adult aquatic insects were made weekly from late spring until early fall at four Canadian sites along the Detroit and St. Clair rivers in southwestern Ontario (Table 1). These areas have been designated Areas of Concern and contain high concentrations of various industrial pollutants (International Joint Commission 1985). The sampling period extended from 19 May to 22

September 1987 for the Detroit River sites and from 11 June to 31 August 1987 for the St. Clair River sites. Samples were also collected in central Ontario (18-19 June 1987 and 6-7 August 1987) from the Gull River near the town of Minden, at Balsam Lake, at Scugog Lake, and from the Ausable River (17 August 1987) near Arkona in southwestern Ontario (Table 1). These areas were presumed to be uncontaminated and material collected was used in detection limit studies. Collections obtained before 26 May were made using the method described by Ciborowski and Corkum (1988). Light Trap Design and Operation We modified the design of a standard Pennsylvania-type light trap (Frost 1957) so that we could catch and retain insects in good condition while minimizing the likelihood of incidental contamination. Typically, light traps have reservoirs containing organic solvents that immobilize and preserve captured insects. Such methods were unsuitable for samples collected for subsequent analysis using Gc. Our traps (Fig. 1) consisted of a galvanized iron bucket (top diameter 30 cm), with a 12-cm-wide cylindrical aluminum hardware cloth reservoir placed in the centre. Dry ice, (requiring approximately 1 kg h- I of operation) was packed around the reservoir. The release of CO 2 gas quickly anaesthetized trapped insects and the ice itself rapidly cooled (or froze) the sample. The reservoir prevented direct contact between animals and the dry ice. The mouth of the bucket was covered by a large funnel that emptied into the reservoir. The funnel was clipped into the bucket and secured by

AQUATIC INSECT ADULTS AS INDICATORS

A

B

c

625

approximately 2-5 m from the riverbank or lakeshore. Traps were operated for 2 h following sunset, since nocturnally active aquatic insects exhibit greatest flight activity during this period (Nimmo 1966). Most insects attracted by the light flew toward the top, collided with the vanes, and fell into the funnel, which guided them into the reservoir. The sheet served as a light reflector and a substrate for those insects that failed to enter the trap. Mayflies (Ephemeroptera) tended to alight and remain on the sheet rather than entering the trap. They were grasped by the wings and placed into separate sample jars or were manually added to the trap reservoir. At the end of the 2-h sampling period, the light was turned off and the contents of the reservoir were emptied into one or more pre-cleaned 500-mL amber glass specimen jars. The sheet was quickly folded to retain any insects that had landed on it. The specimen jars and the folded sheet were kept in a cooler containing dry ice during transport to the laboratory. Samples were stored at -20°C prior to sorting. Animals were removed from the sheet the morning following collection and were added to the sample jar. Sample Processing

FlG. 1. Diagram of light trap components. A, light assembly consisting of circular metal top plate, vanes and 12VDCJJ5W UV fluorescent lamp; B, funnel; C, pail containing metal hardware cloth reservoir and dry ice.

three flanges. A set of three 45-cm-tall, 15-cmwide, clear styrene vanes was mounted on the top of the funnel. The vanes were attached to a circular aluminum top plate. A 3-cm-diameter hole permitted placement and removal of a 45-cm 12V115W DC fluorescent long wave ultraviolet lamp at the axis of the vanes. The light was powered by two 6V lantern batteries connected in series, or by an automobile battery. The lantern batteries provided approximately 9 h of continuous service. Traps were precleaned with soap and water before use, and those parts with which the insects could come into contact were hexane-rinsed. Light traps were placed on a white sheet,

Samples were sorted by taxon at room temperature using hexane-rinsed forceps. Animals of different taxa were wrapped in hexane-rinsed aluminum foil packets and were stored at -20 or -70°C (see below) until further processing and analysis could take place. Invertebrates were identified to genus and included categories such as Hydropsyche (Trichoptera), Hexagenia (Ephemeroptera), Caenis (Ephemeroptera), "other" Ephemeroptera, or "Other Taxa" (mostly Diptera and Coleoptera). The total collection was weighed as well as the individual taxonomic groups. Sorting time increased with increasing diversity. One hundred randomly selected caddisflies in each sample were identified to genus and were preserved in 70070 ethanol for subsequent specific designation. One large sample of Hexagenia collected at Station 2 (East Windsor), and one sample of Hydropsyche collected in central Ontario (Gull R.), were each split and portions were stored at -20 and -70°C for use in evaluating the effect of storage temperature on the results of GC analyses.

626

KOVATS and CIBOROWSKI Contaminant Analyses

Sizes of samples used for site comparisons ranged from 2-5 g fresh weight, depending on availability. Dry weights of samples used for GC analysis were estimated by weighing a 2-5 g portion of the sample, drying at 105°C for 24 h, and reweighing. Dry weights were then calculated by multiplying the fresh weight of the sample to be analysed by the dry weight:fresh weight ratio. Extractions and contaminant analyses were performed at the Great Lakes Institute analytical laboratory, University of Windsor. Samples were homogenized with mortar and pestle in 50 g Na ZS04 • Contaminants and lipids were extracted by solid-liquid column extraction using 20 g NazSO 4 and 300 mL 50070 dichloromethane (DCM)-50OJo hexane mixture as the solvent. The resulting extract was concentrated to 5 mL in a rotary evaporator and added to a Biobeads column (S-X3, 200-400 mesh, Bio-Rad Laboratories). Two fractions were eluted with 300 mL 450/0 DCM-55% hexane mixture. Solvent was evaporated from the first fraction and the remaining lipid residue was weighed. The second fraction, which contained all extracted OC compounds, was concentrated to 2 mL by rotary evaporator and cleaned by passage through a column containing 8 g Florisil (60/100 mesh, overlain with I g NaZS04). Fraction 1, eluted with 52 mL of hexane, was additionally concentrated to 2 mL. Fraction 2 was eluted with 65 mL 50% DCM-50OJo hexane and concentrated similarly. One p,L from each fraction was injected into the GC (Hewlett-Packard, model 5790A) equipped with a 25 m x 0.25 mm fused silica column and an electron capture detector. Specific conditions and methodology used during GC analyses were as outlined by Ciborowski and Corkum (1988). Concentrations of 29 contaminants (18 PCB congeners, pentachlorobenzene (QCB), hexachlorobenzene (HCB), octachlorostyrene (OCS), and 8 pesticides, Tables 2 and 3) were quantified based on peak patterns, and comparison to those in standard mixes of known concentrations. Recovery efficiencies of > 90 070 were obtained using uncontaminated samples spiked with specific concentrations of standard mixes. In addition to analyzing samples of similar taxa from sites with different degrees of contamination, the effects of storage time and temperature on analytical results were determined. Contaminant concentrations detected in fractions of the same sample analyzed at 60-day intervals were compared

as well as those in previously split samples stored at different temperatures. Samples collected at contaminated (Detroit River at Windsor, Ontario) and uncontaminated (Gull River, central Ontario) sites were analyzed to determine minimum sample size yielding acceptable variation in contaminant concentrations among triplicate subsamples. Increasing weights (0.09, 0.18, 0.38, 0.75, 1.5 g dry wt.) of animals were analyzed and separate coefficients of variation were calculated for each of the 29 compounds. These coefficients were plotted against sample dry weight. The weight at which the median coefficient of variation reached an asymptote was considered to be the minimum sample weight that yielded acceptable degree of variation. Separate estimates of minimum acceptable sample size were generated for samples from contaminated and uncontaminated areas.

RESULTS Trap Operation and Efficiency The light traps effectively collected large numbers of aquatic insects. Setup and operation of the traps was simple, and required no special skills. Single collections required an average of 2 kg of dry ice. However, up to 3 kg were needed on warm (> 26°C) or windy nights. Ideal nights for collecting insects were warm, humid, and calm. In June, there was an exponential relationship between caddisfly activity and air temperature (RZ = 0.59, p 15 km h- I ) catches were greatly reduced, often to fewer than 20 insects, and the light trap became susceptible to being blown over. In general, mayflies were more strongly affected by wind than were caddisflies. At wind velocities > 5 km h- 1 mayflies usually arrived with the wind. Below this wind velocity direction of arrival was independent of wind direction. Although detailed data are unavailable, relative humidity also seemed to influence catches; larger samples were collected on humid nights.

> e

,0

TABLE 2. Mean (± I S.E., n = 3) concentration of PCBs (p,g kg-I dry weight) in aquatic insect adults from various study sites. (ND = nondetectable). TAXON

LOCATION PCB 41 PCB 18 PCB 19 PCB 31 PCB 52 PCB 66 PCB 87 PCB 97 PCB 101 PCB 110 PCB 118 PCB 138 PCB 141 PCB 153 PCB 170 PCB 180 PCB 182 PCB 194

Trichoptera Ausab1e River Caenis Lake Scugog Hexagenia Balsam Lake Hydropsyche Gull River Hexagenia Sarnia (Site 4) Hexagenia Sombra (Site 3) Hexagenia Windsor (Site 2) Trichoptera R. Canard (Site 1)

NO NO NO NO NO

0.14 (0.140) NO 1.52 (0.106) NO

3.17 (0.376) NO 2.44 (0.408) NO 7.56 (0.261) 0.11 7.50 (0.105) (0.769)

NO

0.05 1.27 0.36 0.68 (0.053) (0.403) (0.094) (0.126) NO 1.13 4.20 2.26 2.04 (0.587) (0.208) (0.494) (0.300) NO NO 0.88 1.58 1.18 (0.191) (0.461) (0.191) NO 0.23 1.46 0.83 1.30 (0.230) (0.284) (0.126) (0.174) NO 1.68 6.34 3.25 3.35 (0.042) (0.599) (0.320) (0.516) NO 1.64 10.85 4.% 5.95 (0.330) (1.925) (0.107) (0.477) NO 2.75 7.48 4.61 6.54 (0.422) (0.143) (0.416) (0.571) 0.14 10.55 20.89 13.66 62.55 (0.143) (0.546) (2.228) (1.293)(20.174)

'PCB numbering follows Ballschmiter and Zell 1980.

0.42 (0.109) 0.82 (0.213) 0.64 (0.165) 0.57 (0.065) 2.14 (0.393) 3.47 (0.244) 3.03 (0.307) 8.64 (0.751)

1.62 (0.453) 10.85 (2.175) 2.21 (0.218) 2.67 (0.509) 8.69 (0.670) 16.63 (1.212) 12.92 (0.977) 49.08 (3.379)

0.67 (0.210) 2.77 (0.629) 1.41 (0.660) 1.80 (0.554) 3.44 (0.280) 5.63 (0.151) 5.95 (0.635) 16.40 (0.670)

1.75 (0.397) 5.19 (0.861) 2.01 (0.263) 3.06 (0.114) 7.91 (0.857) 13.57 (0.611) 10.10 (0.640) 15.54 (7.930)

2.04 (0.688) 22.89 (4.007) 2.17 (0.330) 5.40 (0.589) 7.70 (0.803) 14.34 (1.163) 18.14 (0.864) 64.22 (9.046)

0.45 2.81 (0.162) (0.816) 7.77 32.47 (1.431) (6.597) 0.58 3.58 (0.139) (1.118) 0.76 5.06 (0.035) (0.484) 2.03 13.21 (0.366) (1.342) 4.49 23.19 (0.252) (1.908) 6.10 21.80 (0.330) (0.143) 23.27 97.55 (2.480) (11.027)

0.55 (0.279) 7.90 (1.513) 0.55 (0.038) 1.08 (0.286) 1.46 (0.737) 3.50 (0.130) 4.80 (0.667) 29.89 (3.726)

1.20 (0.261) 22.39 (3.593) 1.68 (0.200) 2.35 (0.028) 4.84 (0.605) 9.59 (1.878) 10.42 (0.526) 63.80 (6.635)

2.09 (0.887) 18.11 (3.996) 1.16 (0.333) 1.82 (0.249) 1.86 (0.354) 3.33 (0.563) 7.33 (2.634) 48.52 (6.161)

0.53 (0.266) 3.75 (0.794) 1.26 (0.198) 0.89 (0.186) 2.91 (0.125) 4.68 (0.542) 4.49 (0.640) 7.42 (7.417)

...~ ...Z ~

00 t'I"j ~

""3

> o ~

~

00

>

...Z

00

...o ~

a ~

00

QI

!'oJ

......

0'1

l>J

QC

TABLE 3. Mean (± 1 S.E., n = 3) concentration of pesticides and other organochlorine compounds (p.g kg- J dry weight) in representative adult aquatic insect samples at various sites. (Hept. = Heptachlor, H. Epox. = Heptachlor Epoxide, ND = non-detectable). TAXON Trichoptera Caenis Hexagenia Hydropsyche Hexagenia Hexagenia Hexagenia Trichoptera

LOCATION Ausable River Lake Scugog Balsam Lake Gull River Sarnia (Stn. 4) Sombra (Stn. 3) Windsor (Stn. 2) R. Canard (Stn. 1)

070 LIPID

QCB

14.10

0.05 (0.046) 0.38 (0.197) 0.85 (0.183) 0.16 (0.158) 3.81 (1.604) 4.75 (0.503) 10.23 (0.777) 2.39 (0.088)

6.68 13.58 15.02 13.29 14.17 20.86 21.63

HCB

OCS

0.73 1.10 (0.093) (0.598) 0.12 0.81 (0.061) (0.064) 1.52 0.33 (0.744) (0.330) 0.23 1.88 (0.038) (0.106) 27.48 4.78 (3.846) (2.323) 69.97 23.13 (3.271) (3.318) 16.64 120.18 (13.882) (1.619) 21.02 30.81 (0.934) (3.498)

HEPT.

p,p'-DDE

p,p'-DDT

a-BHC )'-BHC

ALDRIN

H. EPOX.

DIELDRIN

0.10 (0.100) 0.98 (0.984) 0.26 (0.263) 0.21 (0.109) ND

45.70 (31.255)

3.16 (2.029) ND 1.03 (1.035) 11.57 (0.808) ND

1.49 2.27 (0.412) (1.339) 2.38 0.50 (0.882) (0.351) 4.72 3.71 (1.083) (2.062) 8.28 3.53 (0.424) (0.919) 14.47 5.70 (1.039) (0.196) 12.43 6.30 (1.881 ) (0.649) 16.71 10.26 (1.376) (2.447) 6.49 12.70 (0.200) (2.209)

0.08 (0.083) ND

4.65 (1.119) ND 2.75 (0.160) 7.77 (0.062) 8.65 (0.277) 8.73 (0.749) 14.94 (0.574) 20.95 (1.587)

16.75 (4.447) 3.54 (0.751) 3.92 (1.114) 22.40 (1.252) 19.87 (1.610) 12.24 (4.235) 31.42 (3.294) 70.54 (4.224)

-

ND ND -

0.63 (0.274)

28.10

(3.473) 21.46 (5.371) 26.31 (0.528) 30.40 (5.053) 36.33 (2.838) 32.76 (4.399) 62.53 (3.339)

-

ND -

ND 27.44 (4.588)

-

0.31 (0.313) ND ND -

ND ND ND -

~

0

-< ~ fJJ ~

=

Q.

n 0==

~

0 ~

fJJ ~

629

AQUATIC INSECT ADULTS AS INDICATORS

300 200

300

• ~

•

-~ .c. en

Detroit River - Station 1. (River Canard) --- Station 2. (East Windsor)

.<:

l/l

~

100

200

!!J Vl Vl

50

<{ ~

0

100

iIi

-ell '-

en

Vl Vl <{ ~

o

m

MAY

JUNE

JULY

AUG

SEPT

10 5

~ 100 l/l

Ii

!!J

1\ 1\ 1\ I \ I I I I I \ I I

~

•

SI Clair River

,

.<:

50

Vl Vl <{ ~

--- Station 3. (Sombra) Station 4. (Sarnia)

I

0

ffi

16

18

20

22

24

26

Air Temperature (Ge) FIG. 2. Relationship between air temperature and catch of Trichoptera in light traps during June, 1987. Regression equation takes the form Ln(l + Biomass) = 0.582(Temperature) - 9.96 (R 2 = 0.59, P < 0.01).

Sample Composition Caddisflies and mayflies comprised the bulk (> 800/0 by fresh weight) of the samples collected at all sites. Most of the Trichoptera captured belonged to the families Hydropsychidae and Leptoceridae. Hexagenia was the numerically dominant mayfly at Stations 2, 3, and 4. Few or no mayflies were caught at Station 1 during the sampling period. There were noticeable differences in diversity among the samples obtained at the different sites. The River Canard sampling station was located near marshland. Catches at this site contained considerably more representatives of other insect orders than those at the other sites. All other sampling stations were situated near rocky shores or breakwalls. Samples collected at those locations were dominated by Trichoptera and required less sorting time than those from the River Canard site.

MAY

JUNE

JULY

AUG

SEPT

FIG. 3. Seasonal variation in abundance of Trichoptera in light traps, May to September, 1987. Top: Detroit River. Bottom: St. Clair River. Note that scales differ on Y-axes.

Numbers of Caenis latipennis (Ephemeroptera: Caenidae) caught at the four stations were variable and depended on wind velocity and direction. This tiny, shortlived mayfly was found to be occasionally very abundant, especially at sampling stations located near marshes. However, the large numbers of individuals required for analysis (1,000-2,000/ replicate), the amount of sorting time required to separate Caenis from other animals, and its weather-dependent availability may reduce the usefulness of Caenis in contaminant monitoring.

Seasonal Variation Caddisfly biomass (fresh wt./trap/2 h) collected on single nights was plotted against calendar date for all sites (Fig. 3). Trap collections during most of May 1987 yielded very small « 2 g) samples. The first large sample (16 g) was collected on 26 May at Station 1. Thereafter, catch sizes increased steadily through June to a maximum on 23 June at both Detroit River sites (Fig. 3). Samples collected at the St. Clair River sites were much smaller, and only began to increase toward the end of June. In

630

KOVATS and CIBOROWSKI

general, caddisfly numbers peaked in late June, and declined through July and August with several minor peaks occurring later in the season. Numbers of Hexagenia were low during most of the summer in both rivers except for a 2-week period in late June. At Station 1, however, the maximum number of Hexagenia caught during the 2 h collecting period was only 11 animals. Contaminant Analyses Samples of all taxa from all sites contained detectable concentrations of most of the contaminants we monitored. At least 85070 of the contaminants occurred at detectable levels in samples from uncontaminated sites. Detailed results of contaminant analy'ses (excluding spiked samples, and storage time/temperature experiments) are listed in Tables 2 and 3. Ciborowski and Corkum (1988) analysed adult aquatic insects collected along the Detroit and St. Clair rivers for PCB congeners and three other OC contaminants. Their principal component (PC) analysis of the compounds distinguished five independent groups of contaminants. Concentrations of contaminants within groups were correlated with one another (r~0.47). To simplify presentation of our data set, we will describe comparative results of one compound representative of each of the three major groups described by Ciborowski and Corkum (1988) and an additional group, the pesticides: 1. Highly chlorinated PCBs (hexa-, hepta-, and octachlorobiphenyls). Representative compound: PCB 180 (2,2', 3,4,4',5,5'-heptachlorobiphenyl) 2. Less highly chlorinated PCBs (tetra- and pentachlorobiphenyls). Representative compound: PCB 66 (2,3',4,4'-tetrachlorobiphenyl) 3. Other OC contaminants (pentachlorobenzene (QCB), hexachlorobenzene (HCB), octachlorostyrene (OCS». Representative compound: HCB. 4. Pesticides (aldrin, dieldrin, heptachlor epoxide, p,p'-DDE, p,p'-DDT, a-BHC, )'-BHC) Representative compound: dieldrin. Differences in storage temperature and length of storage time of tissue samples had no significant effect on results of GC analyses (pairwise t-tests, p > 0.05). Samples may be stored at -20°C for

extended periods (at least 2 mo) without appreciable effects on analytical results. Results of experiments performed to determine minimum sample size are presented in Figure 4. Median of coefficients of variation calculated from contaminant concentrations of all 29 compounds in replicate samples of Hexagenia from Station 2 (Windsor, Detroit River) declined to an asymptote (20%) at 0.38 g dry weight (25 animals, Fig. 4A). The proportion of non-detectable compounds also declined to approximately 20% at the same weight. Based on these results, a minimum sample dry weight of 0.38 g yields acceptable variation in areas of known high OC contamination. Similar analyses of samples of Hydropsyche from the Gull River suggest that a minimum sample dry weight of 0.75 g (170 animals) is appropriate (Fig. 4B). The larger biomass estimate required reflects the lower concentrations of contaminants in these samples. Use of the larger sample size is advised when analyzing samples from previously unstudied areas. Concentrations of the four representative compounds in Hexagenia at three sites along the Lake Erie-Lake Huron connecting channel were compared to those at Balsam Lake (Fig. 5). Concentrations of most compounds were significantly higher at the St. Clair River and Detroit River sites than at Balsam Lake (p
631

AQUATIC INSECT ADULTS AS INDICATORS 160 ~

0

120

0

>

-'

ell

0

u

80

ell

0.38

009

wt.{g)

n=30

80

ell

40

• " ,

I

z

.

••

0.38

0.75

·• I

~ 15

Dry wt. (g)

n=30

ell

-g ....

60

U

C 0

018

• I

80

ell

ell "0

!

Q

40

Dry

....0 ....

((

u

018

009

¢

0

I

40

.D

80

-'

. 0

0

B

120

~

>

-

160

A

.... ,

. - . --- -----_.

' .........

'

.....

'-

....

~

009

....

ell

'

.....

20

60

U

0.18

038

075

1.5

Dry wt (g)

~

40

I

C

o

z

20

~

009

0.18

0.38

Dry wt.

075

15

(g)

FIG. 4. Coefficients of variation (top) and percent non-detectable compounds (bottom) plotted against sample dry weight for (A) Hexagenia at the Detroit River (Windsor), (regression equation: % Non-detectable = 25.33(Dry weightto. 430), and (B) Trichoptera at the Gull River (central Ontario), (regression equation: % Non-detectable = 14.23 (dry weight)-o.488). Median values are represented by large open circles.

Concentrations of the four representative compounds were significantly greater in Hydropsychidae from Station 1 (Detroit River) than in hydropsychid samples from the Ausable and Gull rivers (p < 0.001, pairwise t-tests, Fig. 6). Concentrations of the pesticides dieldrin and p,p'-DDE (Table 3) in samples from uncontaminated sites were considerably higher than those of the other contaminants. DISCUSSION Relatively few studies have investigated the use of adult aquatic insects as indicators of contamination (Mauck and Olson 1977, Clements and Kawatski 1984, Ciborowski and Corkum 1988),

and none have described protocols adequate for use in a general monitoring program. Our results indicate that given appropriate weather conditions, enough adult Trichoptera and Ephemeroptera can be collected by light traps to permit analysis by GC for OC contaminants. The collecting equipment is relatively inexpensive and requires no special skills to operate. The traps were not successful in providing unassisted catches of Ephemeroptera, which were attracted to the lights but required hand collection. Light traps yielded larger numbers of Trichoptera with considerably less effort than the manual method used by Ciborowski and Corkum (1988). However, light traps suffer from some of the same limitations as manual collection methods

632

KOVATS and CIBOROWSKI DIELDRIN

HCB

150

.

0'

..::.c

100

20

50

10

;:)

2 3 4

C 0

...... Cl

.... ...... c

<11

u

c 0 u

PCB 66

6 5 4

PCB 180

Site

BL

15

40 20

5

DR AR GR

c o

"0.... ......

C <11

u

C

2~~:1.-A 2 3 4

60

DR AR GR

12 10

4

2 1

20

DIELDRIN

BL

8 6

3

HCB

10

~'-"-"-- ~

0'

80

25

30

2 3 4

o u

BL

FIG. 5. Concentrations of selected contaminants in Hexagenia at three Detroit and St. Clair river stations and at Balsam Lake (BL). Vertical bars represent 1 S.E.

in that sizes of catches are greatly affected by prevailing environmental conditions, especially temperature and wind velocity. The traps were very effective for Trichoptera, yet we do not know if all local genera of Trichoptera are equally attracted, or whether sensitivity varies with behavioral state of the animals. Trichoptera were available in sufficient numbers to permit unequivocal contaminant detection from June to mid-August. Hexagenia numbers were relatively low throughout the summer, exhibiting only one pronounced peak in late June. Caenis was available sporadically, and only at sites located near marshes. We recommend careful timing of mayfly collections, at or near the peak emergence period, to ensure sufficiently large samples for contaminant analysis. Results of contaminant analyses suggest that considerable year-to-year variation in contaminant body burdens may occur. Comparison of our data with those of Ciborowski and Corkum (1988, Table 3), who also sampled along the St. Clair and Detroit rivers, reveals a significant (2-12 fold) increase in concentrations of most PCBs in Trichoptera at Station 1 (River Canard), although HCB, OCS, and QCB body burdens remained

PCB 180 60 40

4 2

Site

(n = 3).

16 14 12 10 8 6

DR AR GR Site

DR AR GR Site

FIG. 6. Concentrations of selected contaminants in Trichoptera from the Detroit River and two uncontaminated sites (DR: Detroit River at Windsor, AR: Ausable River, GR: Gull River). Vertical bars represent 1 S.E. (n = 3).

unchanged. At Station 2 (Windsor), concentrations of QCB, HCB, and OCS remained constant in Hexagenia, while those of PCBs increased slightly during the same time period. St. Clair River Hexagenia exhibited an opposite trend with respect to PCBs. Animals at Station 4 (Sarnia) harbored at least 50070 lower amounts of most PCBs than those collected the year before, whereas PCB body burdens at Station 3 (Sombra) were slightly higher. Although these results might suggest downstream displacement of PCBs, other factors might also account for the observed trend. For example, our findings may reflect decreasing contaminant levels in the upper St. Clair River sediments due to degradation with time. Active dispersal by the insects is another possibility. The Hexagenia adults collected at Station 4 could have originated from the south end of Lake Huron, which is less contaminated than the St. Clair River. Larvae of this animal have been collected primarily from lentic habitats in the St. Clair drainage basin (Edsall et af. 1988). Because both our own data

AQUATIC INSECT ADULTS AS INDICATORS and those of Ciborowski and Corkum (1988) represent single time-point catches, it is unclear whether these differences reflect random temporal variation or actual trends. Long-term monitoring would be necessary to draw specific conclusions. Nevertheless, the spatial pattern that we detected in contaminant distribution corresponds to that reported in the sediments for at least one compound. This implies that biases due to dispersal effects may be relatively minor. Preliminary studies evaluating dispersal abilities of Lake St. Clair Hexagenia and Detroit River hydropsychid caddisflies also support this conclusion (Kovats and Ciborowski, In prep.); mean dispersal distance (± S.E.) of Hexagenia was estimated as 2,570 (± 228) m, and that of caddisflies as 970 (± 178) m. Passive transport by wind is minimized by the insects' behavior - adult aquatic insects tend to remain on vegetation during windy conditions (Johnson 1969). Mauck and Olson (1977) and Clements and Kawatski (1984) were also able to detect a spatial trend in contaminant body burdens in Hexagenia trapped at sampling sites 10-20 km apart. Their results are consistent with ours, indicating that the spatial resolution of this technique is acceptable despite the possibility that both active and passive insect dispersal may occur. Comparisons of contaminant concentrations between animals from industrialized vs. rural areas confirmed that aquatic insect adults are sensitive indicators of OC contamination. Although samples from uncontaminated sites contained detectable amounts of most (> 85010) of the compounds studied, with the exception of some pesticides, OC concentrations were much lower at uncontaminated (control) sites than at contaminated sites. We attribute the compounds detected in samples from non-industrial sites to agricultural inputs (Watt 1988) and aerial deposition (Metcalfe and Macdonald 1988). Overall, our results indicate that adult aquatic insects are useful and reliable as indicators of OC contamination. Monitoring adults offers several advantages over use of sediment-dwelling immature forms. Animals are available in large numbers during at least one time of the year. Sampling is simple and inexpensive, requiring considerably less effort than benthic sampling techniques such as use of Petersen or Ponar grabs. Adequate biomass of adult insects from light traps can be obtained and processed much more quickly than an equivalent volume of benthic animals. In addition, all samples consist of only one life stage of the ani-

633

mals collected, alleviating problems associated with differential contaminant uptake due to variation in length of contaminant exposure time. Drawbacks of our approach include weatherdependent sampling, restricted seasonal availability of adults of some taxa (Hexagenia), and the possibility of bias caused by dispersal. Most sampling programs require appropriate weather conditions. Dispersal and wind factors may complicate localization of sample sources, although their effects are likely to be of relatively minor importance. Also, dispersal by the adult insects renders our method unsuitable for contaminant pointsource determination. Although monitoring of adults will not replace the use of benthic insects in the assessment of precise locations of elevated contaminant levels and their effects, we suggest that collection of adults may represent an efficient alternative to benthic sampling for preliminary or synoptic surveys, or for long-term monitoring of contaminants in aquatic habitats. ACKNOWLEDGMENTS We wish to thank Stephen Pernal for assistance with collections and sorting. Dr. G. D. Haffner made gas chromatographic facilities available at the Great Lakes Institute. Dr. L. D. Corkum provided helpful comments on the manuscript. Dr. R. Lazar kindly advised us on analytical procedures. This research was funded by the Ontario Ministry of the Environment, Research Advisory Committee. The views and ideas expressed in this paper are those of the authors and do not necessarily reflect the views and policies of the Ministry of the Environment, nor does mention of trade names or commercial products constitute endorsement or recommendation for use. REFERENCES Ballschmiter, A., and Zell, M. 1980. Analysis of polychlorinated biphenyls (PCB) by glass capillary gas chromatography. Fres. Z. Anal. Chem. 302:20-31. Bush, B., Simpson, K. W., Shane, L., and Koblintz, R. R. 1985. PCB congener analysis of water and caddisfly larvae (Insecta: Trichoptera) in the upper Hudson River by gas capillary chromatography. Bull.

Environ. Contam. Toxicol. 34:96-105. Ciborowski, J. J. H., and Corkum, L. D. 1988. Organic contaminants in adult aquatic insects of the St. Clair and Detroit rivers, Ontario, Canada. J.

Great Lakes Res. 14:148-156.

634

KOVATS and CIBOROWSKI

Clements, J. R., and Kawatski, J. A 1984. Occurrence of polychlorinated biphenyls (PCB's) in adult mayflies (Hexagenia bilineata) of the upper Mississippi River. J. Freshwat. Ecology 2:611-614. Downing, J. A. 1979. Aggregation, transformation, and the design of benthos sampling programs. J. Fish. Res. Board Can. 36:1454-1463. Edsall, T. A., Manny B. A, and Raphael, C. N. 1988.

The St. Clair River and Lake St. Clair, Michigan: an ecological profile. Biological Report 85(7.3) U.S.E.P.A Frost, S. W. 1957. The Pennsylvania insect light trap. J. Econ. Entomol. 50:287-292. International Joint Commission. 1985. Report on Great Lakes Water Quality 1985. Great Lakes Water Quality Board. Windsor, Ontario, Canada. Johnson, C. G. 1969. Migration and Dispersal of Insects by Flight. London: Methuen. Kauss, P. B., and Hamdy, Y. S. 1985. Biological monitoring of organochlorine contaminants in the St. Clair and Detroit rivers using introduced clams, Elliptio complanatus. J. Great Lakes Res. 11 : 247-263. Larsson, P. 1984. Transport of PCS's from aquatic to terrestrial environments by emerging chironomids. Environ. Pol/ut. 34:283-289. Mauck, W. L., and Olson, L. E. 1977. Polychlorinated biphenyls in adult mayflies (Hexagenia bilineata) from the upper Mississippi River. Bul/. Environ. Contam. Toxicol. 17:387-390. Metcalfe, C. D., and Macdonald, C. R. 1988. An ecosystem approach to the monitoring of PCBs in pristine Ontario lakes. In Proc. Ontario Ministry of the

Environment 1988 Technol. Trans. Conj., Toronto, Ontario. Nimmo, A. P. 1966. The arrival pattern of Trichoptera at artificial light near Montreal, Quebec. Quaest. Entomol. 2:217-242. Oliver, B. G. 1984. Uptake of chlorinated organics from anthropogenically contaminated sediments by oligochaete worms. Can. J. Fish. Aquat. Sci. 41:878-883. Pugsley, C. W., Hebert, P. D. N., Wood, G. W., Brotea G., and Obal, T. W. 1985. Distribution of contaminants in clams and sediments from the HuronErie Corridor. I - PCBs and octachlorostyrene. J. Great Lakes Res. 1l:275-289. Reynoldson, T. B. 1987. Interactions between sediment contaminants and benthic organisms. Hydrobiologia 149:53-66. Sanders, H. 0., and Chandler, J. H. 1972. Biological magnification of a polychlorinated biphenyl (Arochlor 1254) from water by aquatic invertebrates. Bul/. Environ. Contam. Toxicol. 5:257-263. Tanabe, S., Tatsukawa, R., and Phillips, D. J. H. 1987. Mussels as bioindicators of PCB pollution: a case study of uptake and release of PCB isomers and congeners in green-lipped mussels (Perna viridis) in Hong Kong waters. Environ. Pol/ut. 47:41-62. Watt, W. E. 1988. The effects of argicultural drainage on sediment and water quality loadings. In Proc.

1988 Ontario Ministry of the Environment Technol. Trans. Con!, pp. 319-321. Toronto, Ontario. Younos, T. M., and Weigmann, D. L. 1988. Pesticides: a continuing dilemma. J. Water Pol/ut. Control 60: 1199-1205.