Macroinvertebrates are important members of aquatic ecosystems. They contribute significantly to the processing and cycling of nutrients (Merritt, Cummins, and Burton 1984) and comprise a major portion of the secondary productivity of streams and lakes (Benke 1984). In high elevation mountainous areas, macroinvertebrates have adapted to unique and often extreme environmental conditions (Mani 1968). in general, high elevation areas are typically cooler than lowland locations, and the length of time free of ice and snow is limited. Periods of macroinvertebrate development are often restricted and the availability of food resources can be reduced and irregular. In addition, dispersal and distribution of high mountain macroinvertebrates are complicated by topographic and physiographic barriers, and availability of appropriate habitats (Mani 1968).

Although nearshore macroinvertebrates play a potentially important role in lentic systems, knowledge of ecological factors influencing their distributions in high mountain lakes is limited. The distributions of nearshore macroinvertebrates in the high mountain lakes of North Cascades National Park Service Complex (NOCA), Washington, USA, were investigated to further understanding of baseline environmental conditions and factors influencing distribution. Prior to initiation of this research very little was known about taxonomic composition and distribution of nearshore macroinvertebrates in NOCA lakes. Research objectives included: 1) identification of potential abiotic factors affecting macroinvertebrate distribution; 2) elucidation of the relationship between macroinvertebrate distributions and the types of benthic substrates in nearshore microhabitats; 3) determination of the potential impact of vertebrate predators on macroinvertebrate distributions; and 4) assessment of the correspondence of benthic substrate and taxa distribution with a lake classification system (Lomnicky, unpublished manuscript; Liss et al. 1991) that grouped lakes based on regional physiographic characteristics of the terrestrial environment.

Macroinvertebrate species colonizing a lake originate from a species pool (Browne 1981, Wevers and Warren 1986). Successful colonization depends, in part, on the ability to overcome barriers to dispersal and arrive in a place conducive to continued survival (Barnes 1983; Sheldon 1984; Voshell and Simmons 1984; Bass 1992; MacKay 1992). Once organisms arrive at a site their ability to survive is determined, in part, by prevailing habitat conditions (Crowder and Cooper 1982; Gilinsky 1984; Frissell et al. 1986; and Wevers and Warren 1986). Southwood (1977) described habitat as a fundamental template for community organization. Browne (1981) and Friday (1987) identified lake age, isolation and surface area, and factors associated with overall habitat such as habitat diversity and biotic interactions as important in influencing successful colonization and species composition of lakes.

Temperature and water chemistry are important factors affecting the distribution, diversity, and abundance of aquatic organisms. Temperature has been implicated as a mechanism influencing spatial and temporal isolation (Ward 1992), and as one of several primary factors influencing life history patterns of aquatic insects (Sweeney 1984). Various aspects of water chemistry, e.g., acidity, dissolved oxygen, water hardness, etc. (Bell 1971; Pennak 1978; Fryer 1980; Faith and Norris 1989; Schell and Kerekes 1989; Foster 1991) also influence the distribution of freshwater taxa, although ascertaining the effects of selected chemical factors can be difficult (Ward 1992).

Vertebrate predators may also affect macroinvertebrate distribution. Most lakes of the northern Cascade Mountains were historically fishless. Salamanders (Ambystoma gracile and A. macrodactylum are typically the primary vertebrate predators in fishless high mountain lakes in NOCA (Liss et al. 1991). Taylor (1983) considers Ambystoma gracile (Baird) to be one of the top carnivores in fishless lakes of the Oregon Cascades. Trout have been introduced in Cascade Range lakes since the turn of the century (Jarvis 1987; Bahls 1992). Trout can reduce or eliminate A. gracile (Sprules 1972; Taylor 1983) and A. macrodactylum (Liss et al., in review) from lakes. However, in NOCA, A. gracile appears to be less vulnerable to fish predation than A. macrodactylum and thus capable of co-occurring with trout (see Amphibian section in this report).

Numerous potential effects of vertebrate predators on aquatic macroinvertebrates have been described. Salamander predation on macroinvertebrates has been documented (Efford and Tsumura 1973; Licht 1975; Taylor et al. 1988). Conflicting results concerning effects of salamanders on macroinvertebrate community composition and species abundance have been reported (Dodson 1970; Sprules 1972; Taylor et al. 1988; Petranka 1989). Fish have been shown to alter the size-structure (Post and Cucin 1984; Blois-Heulin et al. 1990; Chilton and Margraf 1990; Diehl 1992), species composition (Macan 1966, 1977; Johnsen 1978; Thorp and Bergey 1981; Gilinsky 1984; Healey 1984), and species abundance (Macan 1977; Healey 1984) of benthic macroinvertebrate communities. The attenuation and elimination of specific taxa (e.g., Chaoborus) has also been reported (Pope, Carter, and Power 1973; Stenson 1978). Macroinvertebrate abundance in ponds and small lakes may change relative to changes in fish abundance (Healey 1984).

NOCA, located in northern Washington, is 204,000 ha in area and encompasses parts of three major watersheds (Chilliwack-Nooksack, Skagit, Stehekin) containing hundreds of tributary drainages and >160 lake basins of importance to fisheries management. All lakes within the complex are considered oligotrophic. Glacial activity during the Pleistocene, and alpine glaciation approximately 2500 years ago formed a diversity of lake basin types (e.g., bench, cirque, ice scour, tarn, morainal deposition, landslide).

Lomnicky (unpublished manuscript; Liss et al. 1991) classified NOCA lakes according to vegetation zone, elevation, and location west or east of the crest of the Cascade Range (Table 1). West slope lakes are present in the Chilliwack-Nooksack and Skagit River watersheds and east slope lakes are located in the Stehekin River watershed. Vegetation zone can be viewed as an indicator of local climate and soil development (Lomnicky et al. 1989). Three vegetation zones were identified: forest (F), subalpine (S), and alpine (A). NOCA forest lakes occur in basins with open and closed canopy forests of mixed conifer with variable levels of understory development (Agee and Pickford 1985). Trees (e.g., western and mountain hemlock, Pacific silver fir, subalpine fir, Douglas fir, etc.) are predominant with an understory consisting typically of tall shrubs (e.g., Sitka alder, willows, and vine maple) and lowland grasses and herbs. Elevation was used to more finely differentiate the forest zone (Table 1). Vegetation in the subalpine zone is a mosaic of meadow and forest flora (Franklin and Dyrness 1973). The dominance of trees decreases relative to the forest zone. Trees occur more often in patches or clumps, and the presence of open areas inhabited by various communities of subalpine shrubs, herbs and other meadow plants increases. In alpine basins vegetation is limited and talus, snow, and glacial ice predominate. Low growing herbs and stunted, isolated clumps of conifers grow where conditions are amenable.

Environmental conditions also differ relative to the location of the Cascade crest. The west slope of NOCA tends to be wetter and cooler than the east slope (Lomnicky et al. 1989). The upper elevation boundary of all vegetation zones is higher on the east slope than on the west slope (Franklin and Dyrness 1973) and, consequently, east slope alpine lakes are rare.

A total of 41 lakes were studied from 1989 through 1991 (Table A.1, Appendix). Each lake had previously been assigned to one of six lake classification categories (Table 1). Lakes are located in remote and rugged terrain, and were accessed only by hiking or helicopter. Field seasons were restricted to the period after ice-out and before the return of inclement fall weather (approximately mid-June through mid-September). Thirty-eight of 41 lakes were sampled in 1989, 21 of 41 lakes in 1990, and 16 of 41 lakes in 1991. Of the 41 study lakes, 15 were sampled once, 18 were sampled two to four times, and eight lakes were sampled live to seven times throughout the duration of the study. Each lake sampled once or twice was visited during the middle to late portion of the field season.

Physical and Chemical Variables

Lake depth was determined with a handheld sonar gun along transects parallel and perpendicular to the long axis of a lake. Surface area was estimated by digitizing 7.5 min USGS topographical maps. Each time a lake was sampled water temperature was measured at 1 m intervals from the lake surface to the bottom over the deepest point in the lake using an Omega 871A digital thermo-couple. Maximum temperature at 1 m below the surface was defined as the highest temperature recorded at that depth during all visits (1989-1991). Water samples were collected at a depth of 1 m over the deepest point in a lake using a 1.5 L Van Dom bottle. Frozen filtered and unfiltered water samples were transported to the Forest Sciences Laboratory (CCAL), Oregon State University for chemical analysis.

The nearshore area was subdivided into three habitats: water surface, water column, and benthic stratum. The benthic habitat was composed of inorganic and organic substrates. Inorganic substrates included silt (ST), sand (SD), gravel (GR), cobble (CB), boulder (BL), and bedrock (BD). Categories were defined by particle size ranges presented in Ward (1992) based on modification of the Wentworth (1922) Scale by Cummins (1962). Organic substrates included emergent/submergent vegetation (ESV, e.g., grasses, sedges, vascular hydrophytes), coarse wood (CW, e.g., logs, snags, branches), fine wood (FW, e.g., twigs, wood pieces, needles), organic detritus (OD), and moss (MS).

Nearshore benthic substrates usually occurred in combinations (e.g., GR-SD CB, OD-CW-ESV, etc.) which were termed microhabitats. The nearshore benthic habitat of a lake was assessed by reconnaissance of the entire shoreline perimeter (Smith et al. 1981), at which time microhabitats and the substrates composing them were identified and mapped. All major lake perimeter structures (e.g., inlet and outlet streams, rock outcrops and cliffs, talus slopes, etc.) also were identified and mapped.

Macroinvertebrates

Macroinvertebrates were collected from the nearshore microhabitats in each lake. Qualitative samples were taken utilizing a sweepnet or by removing individuals from the water column and substrates with a hand net and forceps. Triplicate core samples were taken from benthic microhabitats using a 17 cm diameter metal sampling tube. The tube was placed in position over the chosen sampling site and depressed into the substrate. Material was extracted from the tube to an approximate depth of 5 cm and placed into a 250 pm sieve (U.S.A. Standard Tyler No. 60). Material in the sieve was rinsed with water removed from the tube with a plastic-baster. The material was then placed into a plastic container and handpicked for organisms. All organisms were preserved in 70% ethanol.

In the laboratory, organisms were identified to the lowest taxonomic level possible using a stereomicroscope. Fifty-two percent of all taxa were identified to genus. These were groups where larvae could be identified to the generic level with relative certainty (e.g., Ephemeroptera, Trichoptera, Coleopotera). An additional 43% were identified to family. Three groups of taxa (i.e., Hirudinea, Oligochaeta, and Turbellaria) were identified to Class, and nematodes were identified by Phylum.

Macroinvertebrates were categorized according to their habitat relationship (Figure 1): pleuston (water surface), plankton and nekton (water column), or benthos (bottom substrate) (Ward 1992). Benthic organisms were subdivided into substrate preference groups modified from Ward's (1992) general faunal categories. Organisms were placed in substrate preference groups based on life history information for each taxon. The determination of functional feeding group associations was based on information in Merritt and Cummins (1984).

Figure 1. Generalized model of the organization of nearshore macroinvertebrate communities in the lakes of the North Cascades National Park Service Complex, with Substrate Preference Groups (after Faunal Categories, Ward 1992) and Functional Feeding Groups (after Merritt and Cummins 1984).

Vertebrates

A preliminary analysis was conducted to determine if effects of vertebrate predators on nearshore macroinvertebrates were evident. A more comprehensive analysis of vertebrate predator impacts on macroinvertebrate distribution and abundance, utilizing data from 1989 through 1993, will be undertaken this year. Four vertebrate predation categories were identified for the preliminary analysis: a) no vertebrate predators (NVP, n=7); b) salamanders and no fish (S, n=6); c) non- reproducing trout with a moderate abundance of salamanders, usually A. gracile (FNR, n=5); and d) reproducing trout with salamanders absent or very low in abundance (FR, n=6). Only lakes for which the composition of the vertebrate predator complex was known were used in this analysis. Two species of salamander, Ambystoma gracile (Baird) and Ambystoma macrodactylum Baird, are the primary native vertebrate predators in NOCA lakes (Liss et al. 1991; see Amphibian section in this report). Salamander abundance was determined by snorkeling the lake perimeter and carefully searching through bottom material. The number of observed salamander larvae was recorded (Liss et al. 1991, in review). Cutthroat trout, Oncorhynchus clarki (Richardson), were present in seven of the fish lakes and rainbow trout, Oncorhynchus mykiss (Richardson), in four fish lakes. Presence of trout was assessed by gill netting, angling and snorkeling. Lakes with non-reproducing trout are periodically stocked and, in general, trout density is lower than in lakes with reproducing trout (Liss et al. 1991; see Fish section in this report).

The effects of vertebrate predators on the distribution of 15 nearshore macroinvertebrate taxa (Table 2) was examined. These taxa were either the most widely distributed among all forest and subalpine lakes, or taxa of particular interest (e.g., Chloroperlidae, Gerridae, Notonectidae, Potamonectes, Taenionema), and generally occurred in all vegetation zones both west or east of the Cascade crest. Since many taxa in NOCA are restricted in distribution to a few lakes, using more widely distributed taxa minimizes the possibility that absence of a taxon from a lake was due to factors not related to vertebrate predation.

F = Forest Zone
S = Subalpine Zone
W = West
E = East
FLW = Forest Low West
FHW = Forest High West
FHE = Forest High East

Statistical Analysis

The k-means clustering algorithm (NCSS version 5.03, Hintze 1992) was used to cluster lakes according to number of taxa, maximum temperature, and lake elevation.

Discriminant analysis (NOSS vers. 5.03, Hintze 1992) was used to examine how a set of predictor variables representing characteristics of the aquatic environment would influence the placement of lakes into the six classification categories (i.e., FLW, FHW, FHE, SW, SE, ALP) of the terrestrially-based lake classification (Lomnicky, unpublished manuscript; Liss et al. 1991). According to Tabachnick and Fidell (1989) discriminant analysis is an effective tool designed to elucidate the relationship between predictor variables and groups, and to examine the capacity of these combined variables to predict group membership. In this analysis the predictor variables (i.e., independent variables) consisted of: a) the proportion of microhabitats per lake in which each of 10 substrates were present and, b) the proportion of taxa per lake in each of 10 substrate preference groups (including pleuston and nekton). The dependent variable was made up of the terrestrially-based classification category designations for study lakes. The discriminant analysis program calculated multiple canonical variate scores for each lake, and ordination of these scores helped to show how individual lakes and lake categories were associated. The relationships between predictor variables and canonical variates were determined by calculating Pearson product-moment correlations (NOSS vers. 5.03, Hintze 1992). Finally, the usefulness of the aquatic- based predictor variables in predicting the membership of lakes in the terrestrially- based classification categories was elucidated, using discriminant analysis, by: 1) the construction of a classification matrix according to the assignment of lakes to the classification categories; 2) determination of classification accuracy by calculation of the percent reduction in classification error due to the predictor variables; 3) determination of the number of misclassified lakes; and 4) calculation of the number of lakes expected to be correctly placed into classification categories by chance alone (Tabachnick and Fidell 1989).

Multiple regression analysis (NCSS vers. 5.03, Hintze 1992) was used to determine the relationship of number of taxa per lake (dependent variable) and maximum temperature, elevation, surface area, and maximum depth (NCSS vers. 5.03, Hintze 1992). Confidence limits were set at 95%.

Two nonparametric methods (NCSS vers. 5.03, Hintze 1992) were used to test for statistical significance (p<0.05). The Mann-Whitney Two Sample Test was used to compare sample means among vegetation zones for number of taxa per lake, maximum temperature per lake, proportion of microhabitats containing inorganic and organic substrates, and the proportion of taxa per lake in two general substrate preference categories (i.e., generalists and inorganic substrates; organic substrates). Sample means for pH, alkalinity, and [Ca²⁺] of lakes with and without gastropods and Potamonectes were also compared. Significant differences in the presence of macroinvertebrate taxa in lakes with different types and abundances of vertebrate predators were determined using Fisher's Exact Test.

Sixteen taxonomic groups of nearshore macroinvertebrates were present in NOCA lakes with a total of 88 taxa (Table 3). Trichoptera, Coleoptera, and Diptera had the greatest number of taxa per group. Several taxonomic groups were represented by only a single taxon.

Many taxa were restricted in distribution. Twenty-five percent of all taxa were collected from individual sites while 47% were restricted to <20% of all study lakes. Eighty-nine percent of all taxa, representing all 16 taxonomic groups, were collected from lakes in the forest zone (Table 3). Fifty-six percent of all taxa (12 groups) were collected from subalpine zone lakes and 20% (9 groups) from alpine lakes. Four groups (i.e., Gastropoda, Lepidoptera, Megaloptera, Odonata) were collected only from forest lakes, and the number of taxa in four groups (i.e., Coleoptera, Diptera, Ephemeroptera, Hemiptera) decreased from the forest to alpine zones. The number of amphipod, plecopteran, sphaeriid, and trichopteran taxa was nearly equivalent in forest

and subalpine lakes. Hirudinea, nematodes, oligochaetes and turbellarians were present in all zones.

Physical Variables and Macroinvertebrate Distribution

Number of taxa per lake was positively related to maximum temperature and negatively related to elevation (Table 4, Figure 2). Temperature was negatively related to elevation (r=-0.3927). Surface area and maximum depth did not directly affect number of taxa per lake.

Figure 2. Relationship of the number of taxa per lake and maximum water temperature.

The k-means clustering algorithm placed lakes in four groups (Table 5). Mean number of taxa per lake and maximum temperature increased from cluster I to cluster IV. Elevation varied little between clusters I, II, and Ill, but cluster IV, with the highest mean number of taxa and maximum temperature, had the lowest mean lake elevation.

Lake clusters were also related to vegetation zone. Lakes with the lowest number of taxa and lowest maximum temperature (cluster I) were predominantly alpine. Clusters II and III were composed principally of subalpine lakes and cluster IV was made up exclusively of forest lakes.

The relationship between number of taxa and maximum temperature was also demonstrated for lakes grouped according to classification categories (Table 6). At the vegetation zone-level, the number of taxa per lake decreased significantly from the forest to alpine zone (F>S, S>A, Mann-Whitney test, p<0.05). The mean maximum temperature per lake also decreased from the forest to alpine zone, with the difference in maximum temperature between forest and subalpine lakes significant at the <0.10 level (p=0.08, Mann-Whitney Test), while the difference between forest and alpine lakes was significant at <0.05 (p=0.002, Mann-Whitney Test). In general, lake categories with higher mean maximum temperatures tended to have higher mean numbers of taxa. For instance, FLW lakes had the highest number of taxa, highest maximum temperature, and lowest elevation. Conversely, lakes in the SW and ALP categories had the lowest number of taxa and lowest maximum temperatures.

Microinvertebrate-Habitat Substrate Relationships

Substrate composition of lake microhabitats varied between vegetation zones (Table 7). In general, the percent of microhabitats in each vegetation zone with at least one inorganic substrate was high. Yet, of the six inorganic substrates, ST, SD, GR, and CB were present in a greater percentage of microhabitats in subalpine and alpine zone lakes than in forest zone lake microhabitats. Conversely, the percent of microhabitats in forest zone lakes containing at least one organic substrate was significantly higher (Mann-Whitney Test, p<0.05) than in subalpine and alpine zone microhabitats. Organic substrates of all types occurred less frequently in the subalpine zone than in the forest zone. Organic substrates were virtually absent from alpine zone microhabitats.

E/S = Emergent/Submergent

Substrate preference group associations generally parallelled the vegetation zone trends shown for substrates (Table 7). The percent of taxa representing GENR and inorganic substrate preference groups (i.e., PSPE, LITH, and MINO) was significantly higher (Mann-Whitney, p<0.05) in alpine (71%) and subalpine (63%) lakes than in forest lakes (41%). Conversely, the percent of taxa representing organic substrate preference groups (i.e., PELO, XYLO, PHYT, and MORG) and preference groups associated with the water column and surface (i.e, PLEU and NEKT) was significantly higher (Mann-Whitney, p<0.05) in the forest zone (45%) than in the subalpine (24%) and alpine (23%) zones. The remainder of taxa in all three zones represented organisms associated with combined inorganic and organic substrate preferences (MMIX).

Microhabitat substrates and substrate preference group associations were also used, per discriminant analysis, to examine the relationship of lakes relative to the six classification categories. Five canonical variate scores for each study lake were calculated during the analysis and the first two variates retained 74% (i.e., variate one = 56%, variate two = 18%) of the among-group variability. Pearson correlations of the predictor variables with variates one and two (Table 8) revealed that canonical variate one was positively correlated with organic substrates, organic-based substrate preference groups, and water surface (pleuston) and water column (nekton) taxa, and negatively correlated with inorganic substrates (except BL), generalist taxa, and inorganic-based substrate preference groups. Patterns of correlation of substrates and substrate preference groups with canonical variate two were less clear. Predictor variables most strongly correlated with canonical variate two were MINO taxa and the substrate FW.

Ordination of lake scores for canonical variates one and two (Figure 3) revealed three major groups of lakes (i.e., A=FLW, D=SW, and B-C-E-F=FHW-FHE-SE-ALP). Relative to the other lake groups, FLW lakes had the highest mean proportional presence of organic substrates in microhabitats and the highest mean number of taxa in organic-based substrate preference groups (including pleuston and nekton). SW lakes had the highest mean proportion of taxa in inorganic-based substrate preference groups (i.e., PSPE, MINO, and LITH). ALP lakes (i.e., F) were more similar to SW lakes than to the other lake groups and had the highest mean proportions of inorganic substrates and generalist taxa relative to all other groups. The mean proportions of inorganic and organic substrates and substrate preference groups calculated for FHW (i.e., B), FHE (i.e., C), and SE (i.e., E) lakes, in general, fell between those calculated for FLW, and SW and ALP lakes, and tended to be more similar to each other than to the other three groups. These results indicate that differences between groups can be associated with differences in the proportional presence of inorganic and organic substrates and substrate preference groups.

Figure 3. Ordination of benthic substrates (A) and lakes by vegetation zone (B) based on axes derived from DECORANA analysis of substrates matrix.

Discriminant analysis also helped to demonstrate the usefulness of aquatic-based predictor variables in the prediction of the membership of lakes in classification categories based on characteristics of the terrestrial environment. A classification equation based on the predictor variables was calculated for each classification category (i.e., FLW through ALP), and a classification matrix was constructed during the assignment of lakes to the categories. During this process, each lake was predicted to be a member of the classification category to which it had been previously assigned according to the terrestrially-based classification. Therefore, no lakes were misclassified and the accuracy in classification, based on the percent reduction in classification error, was 100%. Further, calculation of the number of lakes (i.e., 8.5) expected to be correctly placed into classification categories by chance alone appeared to be sufficiently low enough to support the conclusion that the classification equations developed by this analysis were useful in assigning lakes to categories (Tabachnick and Fidell 1989). These results demonstrate the utility of using variables associated with lake substrates and their associated taxa in the prediction of the membership of lakes in the terrestrially-based classification categories.

All functional feeding groups were present in each vegetation zone (Table 9). Predators were predominant in forest categories. Subalpine (especially SW) and alpine lakes tended to have a more equitable distribution of taxa among functional groups than forest lakes. Scrapers, with mouth parts adapted for removing periphyton from rock surfaces (Lamberti and Moore 1984), increased in SW lakes possibly due to the relatively greater importance of larger inorganic substrates (i.e., CB and BL) in the microhabitats of these lakes. Although the mean percent of detritivore taxa (i.e., oligochaetes and sphaeriids) varied among categories, they were widely distributed among all lakes. Oligochaetes occurred in 98% and sphaeriids in 61% of all lakes.

*PRED: Predator
SCAV: Scavenger
COGA: Collector-Gatherer
SHRD: Shredder
SCRP: Scraper
DETR: Detritivore

Habitat Conditions and Distribution of Individual Taxa

Snails representing four Gastropoda families were collected only from three FLW lakes. This limited distribution was directly associated with pH, alkalinity, and [Ca²⁺]. Mean pH (based on calculation of mean free [H⁺]), alkalinity, and [Ca²⁺] were significantly higher in the three FLW lakes with snails than in 31 lakes without snails (Mann-Whitney Test, p<0.05). The beetle, Potamonectes griseostriatus (DeGeer) was present in eight subalpine lakes and only one forest lake. The difference in mean pH per lake between the lakes with Potamonectes (pH=6.40) and the forest lakes without this taxon (pH=7.45) was significant (Mann-Whitney Test, p=0.002), indicating that this species might be restricted from forest lakes, in part, because of higher pH levels. The difference in mean pH between subalpine lakes with and without Potamonectes was not significant, suggesting that factors other than pH (e.g., microhabitat substrates) might be responsible for restricting the distribution of this beetle in the subalpine zone. Odonates were collected only from forest lakes. This may have been related, in part, to the increased availability of microhabitats with organic substrates (e.g., ESV, OD, and FW) which are preferred by odonates (Pennak 1978; Corbet 1980; Merritt and Cummins 1984). Callibaetis (Baetidae) was the dominant mayfly in the forest zone, occurring in 93% of lakes compared to 40% of subalpine zone lakes. Callibaetis was not present in alpine lakes. This pattern appears associated with the need for this genus to be near aquatic vegetation and its tolerance for higher water temperatures and fluctuations in pH (Edmunds, Jensen, and Berner 1976; Hafele and Hughes 1981). Ameletus (Siphlonuridae), associated with clean inorganic substrates (Edmunds et al. 1976) such as GR and CB, was most prevalent in the subalpine (75% of lakes) and alpine (83% of lakes) zones, and present in only 33% of forest lakes. The trichopterans Limnephilus and Halesochila (Limnephilidae), and Polycentropus (Polycentropodidae) were present in a greater percent of forest lakes (87%, 53%, 53%, respectively for each taxon) than subalpine and alpine lakes combined (27%, 31 %, 8%, respectively for each taxon). Polycentropus tends to be more tolerant of warmer water temperatures and Halesochila requires habitats with detritus and bottom sediments soft enough for burrowing (Wiggins 1977). Desmona mono (Denning), Psychoglypha, and Ecclisomyia (Limnephilidae) were present in a greater percent of subalpine and alpine lakes (73%, 58%, 35%, respectively for each taxon) than forest lakes (27%, 40%, 20%, respectively for each taxon). D. mono is a LITH organism, and Psychoglypha and Ecclisomyia are reported to be cold-adapted species (Anderson 1976; Wiggins 1977).

Vertebrate Predators

The distribution of three of 15 nearshore macroinvertebrates appeared to be affected by vertebrate predators (Table 10). Three taxa (i.e., Taenionema, Ameletus, and Desmona) were found in significantly fewer lakes with predators (Fisher's Exact Test, p<0.05). The stonefly, Taenionema, was not collected from lakes with salamanders or trout but was found in 5 of 7 lakes with no vertebrate predators. The mayfly, Ameletus, was present in a significantly greater number of lakes without vertebrate predators (Fisher's Exact Test, p=0.002). Its restricted distribution in these lakes may be due more to salamanders than trout, since no significant difference between lakes with and without fish was found and some FNR lakes can have well established populations of the salamander A. gracile. The caddisfly, Desmona, may be limited in distribution by trout. Desmona was found in a significantly higher number of lakes without fish (Fisher's Exact Test, p=0.006). Differences among vertebrate predation categories for the other 12 taxa were not significant.

1. NVP : No vertebrate predation
2. VP : Vertebrate predation (lakes with salamanders and fish)
3. S : Salamanders and no fish
4. FNR : Non-reproducting fish with salamanders
5. FR : Reproducing fish with minimal or no salamanders
6. NF : No fish (NVP + S)
7. F : Fish (non-reproducing and reproducing)

The distribution of nearshore macroinvertebrates in high mountain lakes is affected by interrelated factors (Figure 4) associated with life history, habitat, and trophic relationships (Wevers and Warren 1986). The life history of freshwater macroinvertebrates is determined by those innate and acquired capacities which promote reproductive success and survival including: 1) the ability to disperse and colonize other locations; 2) the ability to acquire necessary resources for growth and maintenance; 3) adult fecundity; and 4) intra- and interspecific behavior (Oliver 1979; Butler 1984; Sweeney 1984; Wevers and Warren 1986). Inorganic and organic substrates, as well as water chemistry and temperature regime impart organization and structure to habitats. Trophic organization is related to the energy and material relations that interconnect the parts of a community (Liss and Warren 1980). Competition and predation are two types of interactions that can affect acquisition and utilization of resources and, therefore, successful colonization and persistence (Friday 1987; Bass 1992; Mackay 1992).

Figure 4. Factors affecting the presence of macroinvertebrates in high mountain lakes.

Lakes in the northern Cascade Mountains are isolated in rugged terrain in a region which is tectonically active (Press and Siever 1982) and geologically young (McKee 1972). Lakes often lack physical connection via inlet and outlet tributaries, and basin characteristics (e.g., geology, climate, elevation, aspect, vegetation, etc.) typically vary between sites (Lomnicky et al. 1989). These conditions may limit dispersal and contribute to the limited distribution of many macroinvertebrates in NOCA study lakes.

The distribution of taxa within an area reflects the concordance of habitat conditions and habitat requirements of organisms. Habitat diversity contributes to the persistence of organisms in lentic systems and enhances species diversity (Liss and Warren 1980; Crowder and Cooper 1982; Gilinsky 1984). In the present study, three fundamental aspects of habitat affected the distribution of nearshore macroinvertebrates: 1) maximum water temperature; 2) water chemistry variables, especially pH, alkalinity, and [Ca²⁺]; and 3) the types of substrates present in the nearshore areas of lakes.

The mean number of nearshore macroinvertebrate taxa inhabiting a lake was directly related to maximum temperature. Forest zone lakes tended to have warmer average maximum temperatures and higher mean numbers of taxa than subalpine and alpine zone lakes. Departures from this general trend appeared related, in part, to differences in elevation and availability of benthic microhabitat substrates.

Temperature influences aquatic insect life history, especially larval growth, adult size and fecundity (Sweeney 1984), and potentially controls many aspects of insect development (Vannote and Sweeney 1980). According to Ward and Stanford (1982), cool headwaters of streams may have been the ancestral habitat for many aquatic insect groups, and evolutionary adaptation to different thermal conditions made colonization of lower river reaches and lentic systems possible. An increasing number of taxa became associated with wider daily and annual temperature variations, creating a greater range of temperature requirements for cool- and warm-adapted species.

Although many water chemistry variables have wide natural ranges (Ward 1992) to which numerous macroinvertebrates are tolerant, certain variables (e.g., pH, alkalinity, and water hardness) may have a particularly important influence on distribution of taxa in aquatic systems. The level of successful emergence in aquatic insects (Bell 1971), species diversity in Crustacea (Fryer 1980), and species richness of macroinvertebrates in lakes (Schell and Kerekes 1989) have all been shown to be positively correlated with pH. Foster (1991) identified three species of Coleoptera in Britain, including P. griseostriatus, that were restricted to acidic conditions in standing water. Some macroinvertebrates also are limited by water hardness. For example, mollusks and some crustaceans have high physiological demands for calcium and therefore specific hardness levels (Ward 1992). In NOCA, the distribution of gastropods and P. griseostriatus appeared to have been limited by habitat requirements related to aspects of water chemistry, (e.g., pH, alkalinity, and [Ca²⁺]).

The influence of benthic habitat substrates on the distribution and diversity of macroinvertebrates in aquatic systems has been associated with substrate particle size in stream habitats (Wevers and Warren 1986), spatial heterogeneity of habitat in the littoral zone of a pond (Gilinsky 1984), and the substrate preference of organisms in aquatic systems in general (Ward 1992). Substratum and substrate particle size influence the abundance and distribution of aquatic insects (Minshall 1984). Crowder and Cooper (1982), Gilinsky (1984), and Diehl (1992) found that habitats with submerged vegetation support higher abundances of benthic macroinvertebrates and greater species richness.

In the forest zone in NOCA, organic substrates were present in a high percentage of microhabitats. Of the three forest categories, FLW lakes had the highest proportion of microhabitats with organic substrates and the highest proportion of taxa comprising organic-based substrate preference groups. As elevation increased in the forest zone (i.e., from FLW to FHW and FHE), and as vegetation zone changed to subalpine and alpine, the proportion of microhabitats containing organic substrates decreased.

As the presence of inorganic substrates increased in nearshore microhabitats so to did the proportion of taxa representing inorganic-based substrate preference groups. This change, which occurred in higher elevation forest lakes (i.e., FHW and FHE), as well as in subalpine and alpine zone lakes was due, in part, to increases in the number of LITH and MINO taxa.

Aquatic systems should be viewed in a broad, integrative framework which expresses the physical environment (Frissell et al. 1986; Lomnicky et al. 1989). Lomnicky (unpublished manuscript; Liss et al. 1991) classified lakes in NOCA according to physiographic characteristics of the terrestrial environment. These characteristics are vegetation zone, crest position (regional topography), and elevation within the forest zone. They indirectly reflect climatic regime and geology. Climate is expressed, in part, by cycles of precipitation and temperature which influence vegetation complexes. Regional topography influences the development of microclimatic conditions within individual lake basins. Variations in the physical and chemical weathering of different lithologies contribute to the formation of different complexes of weathering products (e.g., types of rock fragments, solutions, and secondarily produced new materials; OIlier 1984), as well as to the differential genesis and development of soils and vegetation.

In our research, predictor variables based upon characteristics of the aquatic environment were found to be useful in assigning lakes to terrestrially-based classification categories. This outcome is representative of patterns in the data that are consistent with the hypothesis that abiotic and biotic conditions within lakes are dynamically interconnected with characteristics and processes in the terrestrial environment. Physiographic factors in the terrestrial environment determine characteristics of lake habitat including quantity and type of inorganic and organic substrates, temperature regime, and water chemistry which, in turn, influence distribution and abundance of lake biota.

Macroinvertebrate communities in forest, subalpine, and alpine lakes reflected the relationship between changing habitat conditions and the habitat requirements of organisms. These changes occurred along a continuum associated with vegetation zone, crest position, and elevation (Table 11). Forest zone lakes, in general, had high numbers of taxa and elevated water temperatures. A typical forest lake community was dominated by predators, and included organisms (e.g., Callibaetis, Gastropoda, Halesochila, leeches, Limnephilus, Odonata, Megaloptera, Polycentropus, Sphaeriidae, etc.) tolerant of warmer water temperatures and requiring complex habitats composed, in part, of fine particulate substrates and aquatic vegetation. Taxa associated with these habitat parameters were dominant in FLW lakes and present to a lesser extent in FHW and FHE communities. Relative to FLW lakes, FHW and FHE lakes had a higher percentage of microhabitats containing inorganic substrates, decreased water temperatures, and taxa tolerant of these conditions. In subalpine zone lakes, the number of taxa and water temperature were reduced relative to forest lakes. The percentage of microhabitats with inorganic substrates was higher than in forest zone lakes and microhabitats containing organic substrates were greatly reduced. Taxa (e.g., Ameletus, Desmona, Psychoglypha, Ecclisomyia, small-bodied dytiscid beetles, etc.) associated with inorganic substrates and colder water temperatures increased in subalpine communities, and taxa tolerant of warmer water and varying habitat conditions diminished. Predators became less dominant and the proportion of taxa per functional group became increasingly more equitable possibly due, in part, to a general decrease in the number of taxa per lake. SW communities, more so than communities in SE lakes, tended to exhibit a strong relationship between low taxa diversity and reduced water temperature. Alpine lake communities further illustrated the attenuation of macroinvertebrate diversity and distribution related to reduced water temperature and diminished habitat complexity, especially the lack of organic substrates. ALP lakes were inhabited almost exclusively by organisms adapted to colder temperatures and microhabitats composed of inorganic substrates, and the percent of taxa per functional group in this zone was more "even" than in the forest and subalpine zones.

1. Substrates in > 35% of microhabitats and listed in decreasing order of proportional presence
   
2. Listed by decreasing proportion of taxa/group
   
3. Groups representing 67-76% of taxa/category; COLE=Coleoptera; DIPT=Diptera; EPHE=Ephemeroptera; GAST=Gastropods; HEMI=Hemiptera; ODON=Odonata; PLEC=Plecoptera; TRIC=Trichoptera

Salamanders and trout have been implicated as affecting macroinvertebrate populations. Ambystoma gracile and Ambystoma macrodactylum larvae prey upon nearshore macroinvertebrates (Anderson 1968; Efford and Tsumura 1973) which are also important prey of trout (Efford and Tsumura 1973; Johnsen 1978; Stenson 1981). Ambystoma have been shown to affect the size-structure and species composition of zooplankton communities in Colorado alpine ponds (Dodson 1970; Sprules 1972). The elimination of prey species or groups of species by fish in high mountain lakes has also been documented (Nilsson 1972; Walters and Vincent 1973; Dawidowicz and Gliwicz 1983; Bahls 1991). Reimers (1958) reported that brook trout (Salvelinus fontinalis (Mitchell)) were able to deplete mayfly and caddisfly populations in a small, high-altitude lake in the eastern Sierra Nevada, California, USA. Johnsen (1978) has demonstrated that trout predation upon various groups of macroinvertebrates (e.g., surface organisms, macrobenthos, plankton) changed with weather conditions, season, and prey availability. Trout also induce changes in behavioral responses of nearshore macroinvertebrates (Pierce 1988; Feltmate and Williams 1989; Feltmate, Williams, and Montgomerie 1992). Feltmate and Williams (1989) and Feltmate et al. (1992) found that stoneflies selected darker substrates and changed diurnal activity patterns in the presence of trout. Odonates and stoneflies increased predator avoidance behavior (Pierce 1988; Feltmate and Williams 1989), and odonates decreased use of exposed microhabitats in the presence of fish (Pierce 1988).

Results from the present study indicate that the distributions of three taxa may be affected by vertebrate predators. Several factors may mitigate the impact of vertebrate predators on nearshore macroinvertebrates in NOCA lakes. Prey species and predators that have coexisted over long periods of time have developed relationships which are tightly integrated and relatively stable (Thorp 1986). Members of the salamander family Ambystomatidae are indigenous to North America (Nussbaum et al. 1983), and have probably been present in Cascade Range lakes for some time. Taylor et al. (1988) suggested that although larval Ambystoma talpoideum (Holbrook), Eurycea quadridigitata (Holbrook), and Notophthalmus viridescens (Rafinesque) of different size selectively preyed upon planktonic and benthic macroinvertebrates in a temporary pond in South Carolina, USA, they did not adversely affect macroinvertebrate community composition or abundance. Petranka (1989) determined that Ambystoma opacum (Gravenhorst) larvae did not reduce zooplankton biomass in 11 natural ephemeral ponds in North Carolina, USA. Trout, having more recently been introduced into NOCA lakes, may have a greater potential of adversely impacting macroinvertebrate distributions. Luecke (1990) found that cutthroat trout introduced into Lake Lenore, Washington, USA, decreased densities of dominant macroinvertebrates in profundal and pelagic regions, but not in the littoral zone. This outcome was attributed to increased refugia in the littoral zone, as well as to the relationship between salmonid mouth structure, prey-search pattern and the ability to access partially concealed prey. Bahls (1991) found that large water-column organisms (e.g., Chaoborus and Coleopteran and Hemipteran taxa) in lakes of the Selway Bitterroot Wilderness, Idaho, USA, were present in fewer lakes with trout. Large water-column taxa were restricted in distribution in NOCA, Nine of 10 of these taxa were found in <20% of study lakes with five taxa collected only from individual locations. The potential for detecting impacts of trout predation on these taxa through the comparison of their distributions in fish and fishless lakes in NOCA was limited by their attenuated distributions.

A high percentage of nearshore macroinvertebrates were restricted in distribution in northern Cascade Range lakes. Distributions were influenced by ecological factors such as water temperature, benthic habitat substrates and taxa substrate preferences, water chemistry parameters, and, to a limited extent, the type and presence of vertebrate predators in lakes. Discriminant analysis was used to examine how variables representing characteristics of the aquatic environment (i.e., nearshore microhabitat substrates and their associated taxa) influenced the predicted placement of lakes into six lake classification categories based on physiographic characteristics of the terrestrial environment. This analysis helped to demonstrate the interconnection between processes in the terrestrial environment and the physical, chemical, and biological conditions in lakes. Future research interests should include a more specific examination of those aspects of temperature regime which are important in influencing macroinvertebrate distributions. Investigation of the impact of vertebrate predators on macroinvertebrate distribution, abundance and behavior should also be continued. Perhaps the best way to assess effects of fish on lentic macroinvertebrates is through some form of experimental manipulation of lakes. The distributions of offshore organisms needs to be elucidated. This information would complement what is already known concerning nearshore macroinvertebrate distributions. Finally, research designed to examine the contribution of macroinvertebrates to the secondary productivity of high mountain lakes, as well as to elucidate the specific roles of macro invertebrates in the processing and cycling of nutrients would contribute significantly to a more complete understanding of high mountain lake macroinvertebrate ecology.