Water: Monitoring & Assessment
9.0 Fish Communities
Impacts on Quality of Inland Wetlands of the United States:
A Survey of Indicators, Techniques, and Applications of Community Level Biomonitoring Data
Excerpts from Report #EPA/600/3-90/073
(now out of print)
This discussion includes both adult and larval fish, both game and nongame species. Few freshwater fish spend their entire life in wetlands, and wetlands that seldom contain surface water (e.g., raised bogs) do not usually have fish. Although fish community structure has been widely described in lakes and rivers, and "indices of ecological integrity" which integrate community data have been developed and tested (Karr 1981), such efforts have not yet been transferred to wetlands. Advantages and disadvantages of use of fish as indicators are shown in Appendix A. The paper by Munkittrick and Dixon (1989) provides further discussion of the value of fish as indicators of ecosystem condition. They assert that fish populations, in general, respond to reduced food resources initially by a decline in fecundity, followed by reduced condition factor, an increase in mean age, and finally a drop in population level. They suggest that these characteristics might be used to indicate the "health" of a particular population, and in some cases, the types of stress that are impairing population health.
9.1 USE AS INDICATORS
Enrichment/Eutrophication. Nutrient enrichment can result in increased fish biomass (Colby et al. 1972, Gascon and Leggett 1977) and altered species diversity (Nakashima et al. 1977) in lakeshore wetlands. Increased biomass may result from increased biomass of invertebrate fish foods, these having increased as a result of increased attachment surfaces and detritus provided by nutrient-induced expansion of submersed wetland plant beds (Pardue 1973). If fish food is already abundant, eutrophication may result in population increases in addition to biomass increases (Nakashima and Leggett 1975). Omnivorous species may benefit the most from the increase in submersed plants (Camp, Dresser and McKee 1989). Walleye (Stizostedion vitreum) and Mosquitofish (Gambusia) are two of dozens of wetland species that tolerate eutrophic conditions (Dawson and Hellenthal 1986), but few species occur exclusively in eutrophic waters.
Organic Loading/Reduced DO. Among northern lacustrine wetlands, Rahel (1984) reported that the ratio of cyprinid to centrarchid fish was greater where winter anoxia occurred. Rivers downstream from sewage and industrial waste outfalls showed a decline in fish community richness in Illinois (Lewis et al. 1981) and in Louisiana (Gunning and Suttkus 1984). In the latter study, two species of darter, Ammocrypta vivax and Etheostoma histrio , were particularly intolerant of the effluents. A southern wetland exposed to treated wastewater experienced increased fish productivity and decreased fish species richness (Camp, Dresser, and McKee 1989). Fish habitat in another wetland, a cypress pond in Florida, was degraded by wastewater effluent (Jetter and Harris 1976).
The State of Florida's regulations for discharge of treated wastewater into wooded wetlands specify that the biomass of sport-commercial or forage fish shall not be allowed to decline by more than 10%; exceptions may be allowed if such declines can be attributed, through analysis of covariance, to other factors. The State also specifies that the biomass of rough fish shall not increase more than 25% unless the ratio of sport and commercial fish to rough fish is maintained; sampling protocols are specified. Florida regulations consider the following fish taxa to be most tolerant of treated wastewater: suckers (all Catostomidae), tilapia (all Chichlidae), gar (Lepisosteidae), bowfin (Amia calva), grass carp (Ctenopharyngodon idella), common carp (Cyprinus carpio), and gizzard shad (Dorosoma cepedianum). Table 13 includes some other species that tolerate relatively low levels of dissolved oxygen. .
| Amia calva | Bowfin |
| Cyprinus carpio | Common Carp |
| Eriomyzon sucetta | Lake Chubsucker |
| Etheostoma nigrum | Johnny Darter |
| Ictalurus melas | Black Bullhead |
| Ictalurus natalis | Yellow Bullhead |
| Ictalurus nebulosus | Brown Bullhead |
| Moxostoma carinatum | River Redhorse |
| Notemigonus crysoleucas | Golden Shiner |
| Notropis buchanani | Ghost Shiner |
| Notropis heterodon | Blackjaw Shiner |
| Notropis heterolepis | Blacknose Shiner |
| Noturus gyrnus | Tadpole Madtom |
| Umbra limi | Central Mudminnow |
Contaminant Toxicity. Declines in species richness and density of fish as a result of contaminants (oil, heavy metals, pesticides, etc.) have been widely documented in lakes and streams, but less often in wetlands. There exists a wealth of toxicological data from laboratory bioassays and tissue analyses. These include Johnson and Finley (1980), USEPA (1986), USEPA's "AQUIRE" database, and the US Fish and Wildlife Service's "Contaminant Hazard Reviews" series that summarizes data on arsenic, cadmium, chromium, lead, mercury, selenium, mirex, carbofuran, toxaphene, PCBs, and chlorpyrifos. However, relatively few field data are available for judging which wetland species are most sensitive.
Acidification. Acidity clearly affects fish species richness in lacustrine wetlands (Jackson and Harvey 1989, Rahel and Magnuson 1983, Tonn and Magnuson 1982). Fish species richness declined among a series of lacustrine wetlands with progressively more acidic conditions (Rahel 1984, 1986). Various reviews (e.g., Ford 1989, Hastings 1984, Wiener et al. 1983) indicate that, in northern lakes and streams, species most susceptible to the effects of acidification include lake trout, brook trout, Atlantic salmon, smallmouth bass, walleye, burbot, and common shiner and various other species of minnows. Data on acidification effects in other regions and wetland types are limited.
Salinization. No quantitative, published information was found concerning the effects of salinization of wetlands on community structure of indigenous fishes.
Sedimentation/Burial. No quantitative information was found concerning the effects of sedimentation in wetlands on community structure of indigenous fishes. It is widely documented that one common wetland fish--carp, Cyprinus carpio--resuspends deposited sediments and in doing so, may alter community structure of wetland plants and invertebrates, as well as fish. Since the feeding and reproductive habits of most fish are well-documented, it might be possible to detect gross sedimentation by the density-weighted ratio of sediment-feeding/breeding species to intolerant species.
Vegetation removal. Removal of canopy of forested wetlands generally results in increased algal production and possible increases in herbaceous wetland plants. Removal of submersed macrophytes (e.g., "aquatic weed control") may similarly increase algae. As vegetation is thinned, herbivorous fish species can increase and those that depend on macrophytes for cover can decrease disproportionately (e.g., Homer and Williams 1986, Wiley et al. 1984). However, total abundance and biomass may change little (e.g., Boschung and O'Neil, Mikol 1985, Wile 1978), and with the removal of vegetation the juveniles of some species may become more vulnerable to predation (Peterson 1982).
In contrast, if submersed vegetation becomes too dense, species richness can decline. For example, Lyons (1989) presents data supporting the theory that extirpation of many shiner, darter, and minnow species was caused by invasion and excessive growth (as a consequence of of the exotic milfoil, Myriophyllum spicatum, in Lake Mendota, Wisconsin. Some experimental studies of macrophyte removal have shown declines in total forage fish standing crop, but increases in growth rates, at least initially, of predatory (piscivore) fish; density of six of eight sunfish species declined while density of two cyprinids increased (Bettoli 1987).
Physical alteration of channel structure within wetlands can reduce fish biomass and total production, both within riverine wetlands (e.g., Arner et al. 1976, Portt et al. 1986) and within lacustrine wetlands (e.g., Eadie and Keast 1984). Species assemblages also shift. Poe et al. (1986) suggested that percid-cyprinid-cyprinodontid assemblages had a stronger need for diverse habitats and a lower tolerance for habitat alteration than did assemblages of centrarchids. These investigators found that the percid-cyprinid-cyprinodontid assemblage dominated an area with an undisturbed littoral zone, high water quality and high species richness of aquatic macrophytes. A nearby altered site with bulkheaded shoreline, dredged area, degraded water quality, and low species richness of aquatic macrophytes was dominated by a centrarchid assemblage. Moring et al. (1985) found brook trout to be particularly sensitive to canopy removal in western floodplain wetlands. Brook trout were replaced by a greater dominance of white sucker, northern redbelly dace, blacknose dace, creek chub, and common shiner.
Abundance of fish larvae in a southeastern floodplain swamp stream was found to be 16 times higher in macrophyte beds than in open channels during the daylight hours (Paller 1987). Durocher et al. (1984) found a highly significant positive relationship (P<0.01) between percent submerged vegetation and largemouth bass (Micropterus salmoides). Any reduction below 20 percent of the total lake coverage of vegetation caused a decrease in recruitment and standing stock of bass.
Turbidity/Shade. Increased turbidity, especially when it occurs over extended periods, generally decreases fish species richness and alters species composition (Menzel et al. 1984). Slight or moderate, seasonal increases in turbidity may or may not change fish density and biomass. Species commonly associated with elevated turbidity include carp, carpsuckers, black bullhead, green sunfish, and others (Menzel et al. 1984). Species apparently intolerant of elevated turbidity include fantail darter, smallmouth bass, northern hogsucker, rosyface shiner, hornyhead chub, southern redbelly dace, black redhorse, brook stickleback (Menzel et al. 1984) and many others listed in Plafkin et al. (1989). Also see above discussions of sedimentation/turbidity and vegetation removal.
Thermal Alteration. No quantitative information was found concerning the effects of thermal alteration on community structure of fishes specifically in wetlands. A shift toward warmer-water assemblages, e.g., carp, downstream from heated discharges seems inevitable.
Inundation/Dehydration. Virtually all fish depend on shallow-water habitats (i.e., generally wetlands) at some point in their life history. Some species depend more strongly than others on shallow areas and floodplain wetlands for feeding and reproduction. The proportion of highly-dependent species could theoretically be used as one indicator of hydrologic alteration of a wetland system.
Inundation alters the spatial and temporal distribution of suitable habitat, with unpredictable effects on floodplain-dependent species. Effects depend in part on habitat structure and soil chemistry of the areas being flooded, and whether inundation increases the exposure of isolated populations to predators or aggressive competitors. In southeastern floodplain wetlands many fish species benefit if water levels remain stable during the spawning period following seasonal inundation (e.g., Liston and Chubb 1984, Miranda et al. 1984). In the Florida Everglades, stable water levels resulted in increased fish community richness, diversity, biomass, average size of fish, and proportion of carnivorous species; however, fish density decreased (Kushlan 1976). In Mississippi River floodplain ponds, "days flooded" was the most significant factor in a multivariate regression for explaining total community biomass and biomass of catastomids, clupeids, crappies, cyprinids, and ictalurids; flooding in the sampled wetlands ranged from 24 to 115 days annually, with a mean of 81 (Cobb et al. 1984).
Dehydration reduces wetland fish diversity if it results in (for example) stranding of fish, anoxic conditions, cutting off of access, increased vulnerability to terrestrial predators, reduced area of productive periodically flooded areas, or altered food supply. However, periods of higher precipitation that follow droughts (or periods of inundation following partial drawdown) can result in increased fish production in wetlands; this could be due to increased nutrient availability or temporary elimination by drought of large competing or predatory invertebrates such as dragonfly larvae (Freeman 1989).
Where hydrologic alterations occur, the seasonality of their effects is critical in determining the effect they will have on fish community structure. Species considered by Mundy and Boschung (1981) to be most likely to decline with impoundment in Alabama floodplain wetlands were as follows: Bluehead Chub, Striped Shiner, Creek Chub, Creek Chubsucker, Frecklebelly Madtom, Crystal Darter, Scaly Sand Darter, and Redfin Darter.
Species that are most dependent on wetland portions of larger water bodies might be identified from existing regional literature (e.g., Crance 1988, Giese et al. 1987, Kwak 1988, Liston and Chubb 1984, Ross and Baker 1983, Tarplee 1975, Walker et al. 1985b) as well as from results of several ongoing studies of floodplain fish communities, e.g., studies being conducted by the Cooperative Fisheries Research Unit at Auburn University; the Corps of Engineers Waterways Experiment Station in Vicksburg, Mississippi; the U.S. Geological Survey in Tallahassee, Florida, and others.
Wetlands that normally contain surface waters but then are briefly dehydrated can, upon reflooding, support exceptionally high productivity and biomass of fish (Wegener et al. 1974, Welcomme 1979). However, this assumes fish have access into and out of the wetland as water levels change, and that sediments do not contain significant levels of oxidizable contaminants. Severely fluctuating water levels (i.e., causing repeated exposure of sediments every few hours or days) associated with hydropower generation or boat wakes can kill fish larvae (Holland 1987).
Fragmentation of Habitat. We found no explicit information on wetland fish community response to fragmentation of regional wetland resources. One can surmise that as the distance between wetlands containing fish becomes greater, and/or hydrologic connections become severed by dehydration or dams, species most dependent on floodplain habitats and/or which do not disperse easily might be most affected. The magnitude of the effect may depend on the size and intrinsic habitat heterogeneity of the wetlands that are being fragmented.
Availability of patches of relatively unaltered habitat with natural flow regimes, such as may occur in lower-order tributaries, can help sustain mainstem fish populations even when mainstem habitats are periodically subjected to pollution or extreme hydrologic alteration (e.g., Gammon and Reidy 1981). The distances between such patches may be important. In streams, individual non-anadromous fish over the course of a year seldom disperse more than a kilometer (Hill and Grossman 1987); however, substantially greater mobility (frequent movements of up to 12.7 km) was reported for fish inhabiting North Carolina floodplain wetlands (Whitehurst 1981).
In lakes, fish species diversity increases with increasing surface area and length of shoreline (Barbour and Brown 1974, Moyle and Cech 1982, Tonn and Magnuson 1982), probably as a result of increased habitat heterogeneity and thermal stratification (Eadie and Keast 1984).
Other Human Presence. Sport and commercial fishing comprise an obvious impact to certain wetland fish species, in some cases at the population level.
9.2 SAMPLING METHODS AND EQUIPMENT
Some factors that could be important to measure and (if possible) standardize among wetlands when monitoring anthropogenic effects on community structure of fishes include:
hydrologic access, water depth, winter ice cover, conductivity and baseline chemistry of waters and sediments (especially pH and dissolved oxygen), sediment type, current velocity, fishing pressure (harvest), stream order or ratio of discharge to watershed size (riverine wetlands), shade, amount and distribution of cover (logs, undercut banks, etc.), ratio of open water to vegetated wetland, and the duration, frequency, and seasonal timing of regular inundation, as well as time elapsed since the last severe inundation or drought.
Methods for sampling fish communities are described in Kushlan 1974b, Nielsen and Johnson 1985, Plafkin et al. 1989, Welcomme 1979, and many others.
Often, fish can be found in wetlands only during certain seasons of the year. If wetlands can be sampled only once, then the period just after seasonal rise in water levels, if it coincides with favorable temperatures, is usually recommended. In most regions, numbers of easily identifiable fish will be greatest late in the season due to annual recruitment of juveniles. However, caution is needed to time sampling to coincide with phenologies of particular taxa. Significant, regular events of fish life histories include migration, dispersal, territory establishment, spawning, and development (Brooks 1989).
Larval fish sampling is best accomplished at night to minimize sample bias due to fish avoiding the sampling gear (Chubb and Liston 1986). Schramm and Pennington (1981) also suggested nighttime sampling and showed a maximum of larvae at dusk, high diversity at night and dawn, and low diversity in the daytime. Nighttime samples were particularly important for collection of hiodontids, ictalurids, and percichthyids.
Equipment used in wetlands for fish sampling potentially includes seines, nets, trawls, electrofishing, ichthyocides, and various types of pot gear (Hocutt 1978, Nielsen and Johnson 1985, Plafkin et al. 1989). For sampling larval and egg stages, push-nets and modified plankton nets are often used (e.g., Meador and Bulak 1987), while in dense vegetation, suction pumps and light traps are often used. A study by Pardue and Huish (1981) evaluated techniques for collecting adult fish in forested wetland streams, and found that no single technique collected all species. Thus, they are best used in combination. Scientific collecting permits, available from state fish agencies, are generally required.
Electrofishing. temporarily stuns fish and thus allows them to be scooped into a bucket, identified and measured, and quickly released. Electrofishing equipment is commercially available, and permits for scientific collecting are typically required from state agencies. If the sampled wetland has clearly defined inlets and outlets, these may be blocked with nets to prevent fish from escaping ahead of the electrical field. Repeated passes are typically made. Electrical currents are not always used to stun fish; they may also be used to guide fish into nets or block their escape from a seining area (Nielsen and Johnson 1985).
Electrofishing can quickly obtain fish from many wetland habitats that are difficult to sample with nets, e.g., undercut banks, submersed plant beds. For quantification, data are best expressed as number of fish per unit area shocked. However, quantitative accuracy is good only for narrow, non-turbid channels. Morgan et al. (1988) reported that the effectiveness of electrofishing generally decreased as plant density or turbidity increased, due to the difficulty in locating and retrieving stunned fishes. Backpack shockers may be too bulky and unsafe for use in wetlands with extensive debris, very soft substrate, and/or ice. Boat-mounted shockers are limited by shallow water depths, debris, and ice.
It is often difficult in wetlands to confine the area being sampled, so fish may flee the advancing electrical field. Also, fish stunned by shockers are not necessarily representative of the general fish community. Collections tend to be biased toward larger individuals and species. Larvae are not captured. Some studies suggest that catchability of fish declines with successive passes through a wetland, with the effects lasting up to 24 hours.
Sampling efficiency can also be influenced by water quality. Pulsating, direct current (DC) units are effective in perhaps the widest range of conditions, but in the "soft" waters of many wetlands (particularly bog streams), AC units with outputs exceeding 500 volts might work just as well. Some investigators in small, confined soft-water wetlands have increased shocker effectiveness by placing salt blocks in the water, which increases conductivity. Extremely high conductivity can reduce effectiveness as well. Bosserman and Hill (1985) found that shockers were not effective in waters made highly conductive by acid mine drainage.
Seines. are robust nets, several meters long and with a width usually equal or greater than water depth, that are pulled by people or boats through shallow areas to confine and capture fish (often by herding them toward shore). When aquatic plants and debris interfere, seines can instead be placed in adjoining open areas and fish herded into the seines for capture (Nielsen and Johnson 1985). Seines are too ineffective for accurately estimating fish densities, but may allow a fair estimation of species richness and of relative dominance of species. Leidy and Fiedler (1985) used a 3 m long seine of 6 mm mesh to sample shallow streams. Ohio streams were sampled using a 4 ft x 8 ft "Common Sense" minnow sein with 1/8 inch mesh; about 30 seine hauls were required for thorough sampling (Tramer and Rogers 1973). For wetlands, Hocutt (1978) recommended the 5 ft x 10 ft "Common Sense" seine with 1/8 inch mesh. A mesh size of 1/8 inch mesh was recommended to capture smaller species and/or life stages. In situations where larger fish may outswim smaller seines, monofilament gill nets can be used for seining.
Sweep nets. (dip nets) can be used to capture fish as well as invertebrates. They can be effective for qualitative sampling in very confined, shallow, clearwater pools. Walker et al. (1985b) had limited success when dip-netting floodplain fish immobilized with spotlights at night. Studies using sweep nets include those by Leidy and Fiedler (1985) and Chubb and Liston (1986).
Other types of nets. are used to catch wetland fish, in a passive manner. All nets tend to be selective due to their design and thus usually provide the best catch results when used in combination. Fyke nets have been used to sample fish in wetlands (Nielsen and Johnson 1985, Swales 1982, Tonn 1985). Wetland vegetation is sometimes removed in a small area to make room for the net. Gill nets can be used to take a variety of fish and can be adapted to different depths (Hocutt 1978). These nets are very effective in wetlands, and are highly selective for particular size classes and species (Pennington et al. 1981). Gill nets of five mesh sizes between 2.54 and 12.7 cm were placed near shore in Ohio riverine areas by Hassel et al. (1988) and checked after 24 hours. Gill net selectivity produced catches dominated by relatively large species such as longnose gar and channel catfish. Trammel nets purportedly are less size selective than gill nets (Pennington et al. 1981), but select for fish species with rough surfaces and protrusions (Nielsen and Johnson 1985). Although traditional trawl nets are not effectively used in wetlands, Herke (1969) described a boat-mounted push-trawl useful for sampling marshes.
Lift nets. constructed with rectangular frames, hoops, or spreaders are set on the bottom below the water surface, then lifted to capture small schooling fish (Nielsen and Johnson 1985). Camp, Dresser and McKee (1989) reported on a lift net specifically designed for use in forested wetland systems. The 1 meter square net was made of two weighted PVC loops (a top and bottom) with netting of black fiberglass screen. When fully extended, the bottomless net measured 39.4 inches x 39.4 inches x 36 inches with a 6 inch flap along the base. The bottom frame was attached to the substrate and the top frame was connected to a rope and pulley system to allow the trap to be sprung (lifted) from a remote location without frightening any fish within the net. Small portable drop nets were used by Freeman et al. (1984) to sample fish in a heavily vegetated freshwater wetland. These collected significantly more fish per unit area than did seining. Large drop nets suffer problems of mobility, and when designed to be portable, create disturbance by movement and shadows (Freeman et al. 1984).
Pot gear. (fish traps) of wood, wire mesh, and/or acrylic plastic have been routinely used by several experimenters. Traps can be used in a variety of areas of moderate depth and/or heavy cover, and when baited, are strongly selective for particular species and size classes (Pennington et al. 1981). Studies that used fish traps in wetlands include Finger and Stewart (1988), Tonn and Magnuson (1982), Walker et al. 1985b.
Ichthyocides are poisons (preferably biodegradable) that can be used to destructively sample the entire fish population of a wetland. They are undoubtably the most efficient tools for obtaining both quantitative and qualitative fish samples. However, when used by inexperienced collectors, problems may outweigh benefits (Hocutt 1978). Examples of use in wetlands include studies by Durocher et al. (1984) and Walker et al. (1985b).
Also, radiotelemetric methods can be used to track individuals (e.g., Savitz et al. 1983) and estimate potential wetland dependency.
9.3 SPATIAL AND TEMPORAL VARIABILITY, DATA GAPS
In general, quantitative data on wetland fish community structure has not been uniformly collected from a series of statistically representative wetlands in any region of the country. Thus, it is currently impossible to state what are "normal" levels for parameters such as fish density, species richness, biomass, Index of Biotic Integrity (IBI, Karr 1981) and their temporal and spatial variability, in any type of wetland.
A data set that is perhaps the closest to meeting this objective was collected from a series of relatively pristine Arkansas rivers that are mostly bordered by wetlands (Giese et al. 1987). These fish data were collected in part from streams flowing through relatively pristine floodplain wetlands, and thus help serve as a regional baseline for bottomland hardwood wetlands. Although data on fish density were not developed, up to 36 species per stream were found and community structure of relatively pristine streams was defined. In nearby Kentucky, a riverine slough wetland supported at least 12 species (Bosserman and Hill 1985).
In submersed wetland plant beds, up to 255 fish per 10m2 may be present (Morgan et al. 1988). On the floodplain of the Kankakee River in Illinois, 481 fish were captured during 4800 hours of trapping (Kwak 1988). In one of the few studies of larval fish communities, Chubb and Liston (1986) reported densities of up to 32.2 larvae per m3 from Great Lakes emergent wetlands.
For stream fish studies, coefficients of spatial variation have ranged from about 50 to 150 percent (Eberhardt 1978). In submersed vegetation, this coefficient may range from 9 to 80 percent (Morgan et al. 1988). Studies that have compared fish communities among wetlands (spatial variation) have largely been conducted along the lower Mississippi River, and include:
Baker et al. 1988, Cobb and Clark 1981, Cobb et al. 1984, Conner et al. 1983, Felley and Hill 1983, Hall 1979, Lowery et al. 1987, Mathis et al. 1981, Pennington et al. 1980, and others.
Only a few studies (Clady 1976, Freeman 1989, Kushlan 1976, Lyons 1989) have quantified year--to-year or long-term variation in fish community structure in wetlands, but conceivably unpublished data may be available from sites of the U.S. Department of Energy's National Environmental Research Park system, as well as the following sites of the National Science Foundation's Long Term Ecological Research (LTER) program (that contain studied wetlands): Illinois Pool 19 site, Illinois-Mississippi Rivers sites, New Hampshire Hubbard Brook riparian forest, Oregon Andrews Experimental Forest riparian forest, and Michigan Kellogg Biological Station site. Temporal (year-to-year) variation in western riparian fish communities was quantified by Platts and Nelson (1988). Although state fishery agencies undoubtedly have long-term data on average biomass or length of captured game fish, these data may not have been systematically collected from wetland sites, and do not include all wetland fish species.
Quantitative data on community composition of wetland fish appears to be most available for lacustrine aquatic bed (herbaceous) wetlands, western riparian wetlands, and southeastern bottomland hardwood systems. Apparently such data are least available for riverine herbaceous wetlands and for riparian wetlands in other regions.
Even qualitative lists of "expected" fish in wetlands of various types do not appear to have been compiled, although regional distribution of fish is relatively well-documented (e.g., Hocutt and Wiley 1988; Lee et al. 1980). Some publications in the "community profile" series of the USFWS (Appendix C) mention particular taxa known to occur in wetlands, and wetland fish are listed in the ERAPT database (Dawson and Hellenthal 1986), in Niering (1985), and in the "Vertebrate Characterization Abstracts" database managed by The Nature Conservancy and various state Natural Heritage Programs. Quantitative data are generally most available for harvested species, and less available for non-game species.