Env. Biol. Fish. Vol. 3, No. 1, pp. 7-31, 1978

Trophic and spatial interrelationships in the fish species of an Ontario temperate lake* Allen Keast Department o f Biology, Queen's University, Kingston, Ontario K7L 3N6, Canada Keywords: Annual cycle, Centrarchidae, Competition, Cyprinidae, Ecosystem, Faunal origins, Habitat, Ictaluridae, Percidae, Seasonal periodicity Synopsis Analysis of the fish faunas of Lake Opinicon and other small, cold temperate Ontario water bodies shows that the component species differ in body size, morphology, abundances, habitats, diurnal and seasonal habitat utilization patterns, diets, dietary changes with age, reproductive strategies, and population turnover rates. These differences are detailed. The number of species occurring in a lake is partly due to historic factors, the number of habitats available, and morphological, behavioral and ecological adaptations that, by channelling their owners towards alternative resources, permit species to co-occur. Diet overlap values between most species are low except for the congeneric bluegill and p u m p kinseed sunfishes, where values are moderate. These are the two commonest species in the lake and other parameters must increase the ecological differences between these two species. Lake Opinicon is a highly variable ecosystem. Part of this variability stems from the seasonal nature of the environment and the fact that different resources reach their peak abundance at different times of the year. Ecological overlap levels between fish species fluctuate greatly in the course of the season as species switch from, or move on to, different resources. Population levels in different habitats also vary seasonally. Species adaptations and interaction patterns were presumably evolved over a long period; most of these adaptations undoubtedly developed before the component species colonized the lake.

Introduction The present review, dealing with the fish faunas of small cold temperate Ontario waterbodies and specifiReceived 20.6.1 977 Accepted 31.5.1977 * This paper forms part of the proceedings of a mini-symposium convened a t Cornell University, Ithaca, N.Y., 18-19 May 1976, entitled 'Patterns of Community Structure in Fishes' (G.S. Helfman, ed.).

cally Lake Opinicon, is directed at determining: (i) how the component species use space and food resources, (ii) patterns of interaction between the species and their age classes and (iii) what controls the ecological roles and types of species that co-occur. Attention is given to the relationship between morphology and ecology, food specializations, resource division, habitat specializations, community groupings, and quantification of ecological overlap values between species. In association with the present review attention is drawn to studies of space utilization in waterways to the south of the Great Lakes where whole assemblages of congeneric species occur and where species richness is higher (Mendelson 1975, Werner et al. 1977). An experimental approach to the analysis of species relationships in lake fishes has been introduced by Werner & Ha11 (1976) and Werner (1977).

The study sites and previous research Lake Opinicon (Fig. 1 and 2) is broadly typical of a large number of small lakes in northeastern North America. It has an area of 890 ha, maximum depth of 11 m, is structurally diverse, and is ice-covered from December to April. It has been the focus of fish investigations since 1961. Published studies on its fauna and those of nearby water bodies include: resource division in a small bay community (Keast 1965) and in a stream (Keast 1966); body morphology relative to way of life (Keast & Webb 1966); feeding periodicity (Keast & Welsh 1968); feeding at low temperatures (Keast 1968a); and successional spawning (Amundrud et al. 1974). See also the work of Harker (1976) and Brown (1977) comparing the feeding and

diets of common species in habitats within a lake and between different lakes, reviews of the comparative feeding ecology of age groups within species (Keast 1977a-d), and studies of habitat utilization and cornmunity patterns (Keast & Harker 1977a,b, Keast et al. 1977).

Age and origin of the Ontario small lake faunas The small lakes and streams of eastern Ontario are part of the Great Lakes system. They are recent, postdating the northern withdrawal of the Wisconsin ice

CANADA

50 krn

- ) L ONTARIO

5-

" I 1

I 74-

Fig. I. Geographical location of Lake Opinicon and 0th water bodies discussed in the text.

Fig. 2. Habitat diversity in the eastern section of Lake Opinicon.

sheet some 11,000 years ago. The history of the freshwater systems of northeastern North America over the last 20,000 years has been complex. Ice sheets have advanced and retreated, large lakes have been created and contracted and, finally or most recently, eustatic rebound has progressively changed drainage patterns. Some 4,000 years ago, an arm of the sea temporarily extended up the St. Lawrence valley nearly to Lake Ontario (Hough 1958). Since European settlement, canals have been built and Lake Opinicon presently forms part of a canal system extending from the Ottawa River to Lake Ontario. Obviously opportunities for fish colonization and dispersal post-glacially have been great and hence it could be argued that many water bodies (including Lake Opinicon) today support their full potential complement of available morphological and ecological types. The Great Lakes fish fauna today consists of 172 species in 74 genera and 28 families (Hubbs & Lagler 1958). Of considerable zoogeographic interest is that this fauna is drawn from three foci, the Mississippi-Missouri basin to the south, the more northem cold glacial front lakes and Yukon-Alaska refugium, and the Atlantic coastal rivers. The greater bulk of the species are southern in origin. The Atlantic component, by contrast, is limited to a few species. The southern 'warm water' component, which is dominated by centrarchids, cyprinids, ictalurids, and the tribe Etheostomatini, and the northern 'cold water' component, composed of salmonids and coregonids, separate out broadly in terms of temperature preferenda and spawning seasons (Hubbs & Lagler 1958). It must be stressed that this subdivision, though valid, is a gross oversimplification, e.g. some species may have originally been northern but survived the last glaciation in the southern refuge. The yellow perch is an example (McPhail & Lindsey 1970). The species composition of individual lakes is related not only to latitude but lake depth and the habitats available. Contrast, for example, the faunas of the six water bodies shown in Figure 1 and Tables 1 and 2. Suggested areas of origin for the various species based on contemporary distribution patterns are given. Note that 25% of the species have both extensive southern and northern ranges and hence cannot readily be allocated to either category. To illustrate these different distribution patterns and provide a basis for discussions of adaptations, the distribution patterns of six of the commoner Ontario and Lake Opinicon fish species are shown in Figure 3. (For the scientific names of these species and those

discussed subsequently see Table 2). The four centrarchids (numbered 1-4) obviously are southern in origin and are near their northern range limits in Ontario. The pike, which is conspecific with the European pike, presumably entered North America via the Beringian land-bridge only recently. The muskellunge (Esox masquinongy) is apparently its original North American counterpart. The perch probably also came from the Palearctic region but rather earlier. Pike and perch both spawn at very low temperatures (Scott & Crossman 1973), which is characteristic of northern forms.

mouth. Bullheads are chemo-sensory, not visual, feeders (note the small eyes), prey being mainly detected by the barbels although the whole of the scale-less body is equipped with taste organs (Atema 1971). Much of the feeding is carried out whilst cruising over the bottom at a slight angle with barbels touching the substratum. When prey is detected the fish twists rapidly and seizes it with a snapping acPUMPKINSEED

Body morphology in relation to ecological role The general body morphology of 16 of the Lake Opinicon fish species is shown in Figure 4 and the mouth form of seven species is shown in Figure 5. Three major body shapes occur. Most species have a fusiform shape. This is physically the most efficient for rapid progression through a liquid medium, minimizing drag. It thus promotes speed and characterizes fish which like the small cyprinids, pursue their prey and/ or range widely. By contrast, the high, bilaterally compressed and short gibbous body form of the sunfish makes it 'a very stable swimmer, since its large lateral area limits rolling, while pitching rotations can be controlled by the pectoral and pelvic fin forces' (Harris 1938). This body form maximizes speed in turning while minimizing the power required, and permits delicate adjustments in a variety of planes through undulation of the fins (Alexander 1967). In this way the fish can position itself well with respect to prey. A third basic body type is that of the brown bullhead. This is sharklike in that the body is bilaterally compressed, permitting strong lateral contractions and powerful twisting movements, whilst the head is flattened and expanded to give a very broad

ROCK BASS

LARGEMOUTH

Fig. 3. Contemporary distribution patterns of the pumpkinseed (Lepomis gibbosus), bluegill (L. macrochirus), rock bass (Ambloplites rupestris), largemouth bass (Micropterus salmoides), yellow perch (Perca flavescens) and northern pike (Esox Eucius). Redrawn from Scott & Crossman (1973).

Table 1. Physical characteristics of six small Ontario water bodies. Data from Sheppard (1965), Brown (1977), and Terasme & Mirynech (1964). Name Kearney L. Gould L. Sydenham L. L. Opinicon Fish Lake Jones Creek (upper)

Surface area mZ

Maximum depth m

Mean depth m

Thermocline depth (July) m

3.2 x l o s 1.99 x lo6 4.5 x lo6 7.8 x lo6 2.5 x 1 0 6 small creek

30 62.1 36.4 11.0 3.3 1.8

9.1 21.8 6.8 4.9 2 1.0

8 7.5 9.1 isothermic isothermic isothermic

Mean pH (1 m) 6.8 7.5 8.0 8.2 -

Table 2. Fish species composition of six small Ontario water bodies. X in parentheses refer to species caught only on rare occasions. Kearney Lake Clupeidae Alosa pseudoharengus Salmonidae Salvelinus .fontinah S. namaycush Coregonidae Coregonus clupeaformis C. artedii Prosopium cylindraceum Umbridae Umbra limi Esocidae Esox lucius E. vermiculatus Cyprinidae Chrosomus eos Notemigonus crysoleucas Notropis cornutus N. heterodon N. heterolepis Pimephales notatus P. promelas Semotilus atromaculatus S. margarita Catostomidae Catostomus commersoni Ictaluridae Zctalurus natalis I. nebulosus Anguillidae Anguilla rostrata Cyprinodontidae Fundulus diaphanus Gadidae Lota lota Atherinidae Labidesthes sicculus Gasterosteidae Culaea inconstans Centrarchidae Ambloplites rupestris Lepomis gibbosus L. macrochirus Micropterus dolomieui M. salmoides Pomoxis nigromaculatus Percidae Perca flavescens Etheostoma exile E. nigrum Percina caprodes

Gould Lake

Brook trout1 Lake trout1 Lake whitefish1 Lake herring1 Round whitefish1 Central mudminnowz Northern pike1 Grass pickerel Redbellied daceZ Golden shiner4 Common shiner4 Blackchin shiner4 Blacknose shinerS Bluntnose minnow4 Fathead minnow4 Creek chub4 Pearl dace1 White sucker2 Yellow bullhead4 Brown bullhead4 American eel3 Banded killifish4

Brook silverside4 Brook stickleback' Rock bass4 Pumpkinseed4 Bluegi114 Smallmouth bass4 Largemouth bass4 Black crappie4 Yellow perch2 Iowa darter2 Johnny darter2 Log perch2

Species of presumed northern origin. Species with extensive range both to north and south of these water bodies. Atlantic origin. Southern species.

Sydenham Lake

Fish Lake

Jones Creek

Lake Opinicon

tion. Tube-dwelling chironomid larvae are a major item of diet. Adaptive modifications to the basic fusiform body type include changes in proportions and in the position, shape, and size of the fins (Fig. 4). The brook silverside, golden shiner, and mud minnow provide examples (Keast & Webb 1966). In the surface swimming silverside the dorsal body line is straight, and the dorsal fin low to minimize surface turbulence; the mouth is upturned. The small Cladocera-eating blackchin and golden shiners gain speed and maneuverability from their compressed fusiform bodies, long and slender caudal peduncles, deeply forked tails, and hgh mobile dorsal and paired fins. The thick, foreshortened body of the mud minnow is advantageous amongst tangled vegetation but, as the fleshy caudal peduncle and rounded caudal fin indicate, it is a slow swimmer. It feeds by stalking prey from a mid-water position. The mouths of fishes are highly adaptive and vary in size, shape and position (Keast & Webb 1966). In the bluntnose minnow, the mouth is very small in cross-section, tubular, sub-terminal, and the prey is taken bv suction. Small chironomids. cladocerans, and organic detritus are found in the stomachs. The banded killifish, by contrast, has a dorso-terminal mouth predisposing its owner to taking prey from ahead or above. In the sunfishes the mouth is terminal, protrusible, small and tubelike. The suctorial component dominates the feeding, the prey being

drawn into the mouth when the pharynx is expanded (Alexander 1967, Werner & Hall 1976). The yellow perch has a large, generalized terminal mouth. Zooplankton, amphipods and other small organisms, then larger-bodied Zygoptera and Anisoptera nymphs and, finally, fish are common food items. The piscivorous largemouth bass and pike have very large mouths. In the latter the head is attenuated into a broad snout, the marginal teeth are enlarged and recurved, and the mouth and throat are supplied with teeth that direct the slippery prey towards the stomach. The jaws are heavy, the premaxilla non-protrusible and prey is seized by a grabbing action. The homodont marginal teeth of most fish characteristically do little more than grasp prey. Mechanical breakdown of food is accomplished by the pharyngeal teeth at the back of the throat (see also Ogden & Lobel, these proceedings). These take the form of dorsal and ventral pads that move against each other in various ways. The dorsal and ventral pads of six fish species show considerable diversity (Fig. 6). In the pumpkinseed the pads are covered with molariform teeth and function as grinding mechanisms. This species is a mollusc and isopod feeder. The pads of the bluegill, by contrast, are covered with fine needlelike teeth, structures linked to a diet of small, softbodied prey such as Cladocera and chironomid larvae. In the black crappie, there is a complicated arrangement of smaller pads dorsally, whilst the lower pair is long, narrow, and diagonally set. The mode of action

*

BLUNTNOSE

MINYOW

sE4!& KILLIFISH

YELLOW

PERCH

z%

LARGEMOUTH

Fig. 4. Comparative body form of sixteen Lake Opinicon fishes. For scientific names see Table 2.

BULLHEAD

3ASS

Fig. 5. Mouth size and form in several Lake Opinicon fish species.

here has not been studied. Such a layout suggests that the pads are used with a fore and aft cutting action. The banded killifish, which feeds on a wide range of small-bodied organisms, has generalized pads that are equipped both with molariform and needle-like teeth. The size of the lower pad does not match that of the upper one. The pads of the piscivorous pike are insignificant but, like the mouth and throat, are equipped with backwardly directed teeth. Pike swallow their prey whole. The planktivorous alewife has simple pads. Most fish have generalized gill rakers. In typical sunfishes there are only 8-12 short rakers on the first pharyngeal arch. The alewife, black crappie, and brook silverside have large numbers of long rakers that function as a sieve to retain small food organisms in the pharynx when the water is ejected through the gdl slits. Brooks & Dodson (1965) found that, rather than passively sift organisms from the water column, alewives pursue the larger zooplankters individually. Emery (1973) has confirmed this. Plankton picking, and not passive filtering, is presumably also the feeding mode of the other two species.

PUMPKINSEED

BLUEGILL

B L A C K CRAPPIE

194mm

UPPER

The dominant family of fishes in Lake Opinicon is the Centrarchidae with five common species (plus a less common one). It is interesting, accordingly, to determine the pattern that morphological differentiation has taken here. Figure 7 provides a visual overview of seven basic body structures that have been modified in various ways to produce a series of distinctive adaptive types. The somewhat generalized rock bass has an enlarged eye (i.e. increased retinal area) for dim light vision. It is a crepuscular feeder taking Anisoptera nymphs and crayfish with its large strong mouth. The sunfishes, as noted, are adapted by their gibbous bodies and small tubular mouths for selectively obtaining small organisms; differences in the pharyngeal dentition channel the pumpkinseed and bluegill towards different resources. The largemouth bass has the classic attributes of the piscivore. The immediate function of the increased number of pyloric caeca (7-10 and dividing to form about 28) that are histologically identical with the proximal intestine is not apparent (Reifel 1973). The combination of gibbous body, high gill raker count, and somewhat enlarged mouth of the black crappie mark it as a mid-water feeder. Chaoboms larvae, the larger cladocerans, and small fish form its food (Keast 1968b). Body morphology thus channels each species towards an ecological role different from that of the other species. In energetic terms, as Werner (1977) has stressed, the ultimate significance of adaptive

. . .. .

.

BLACK

CRAPPIE

RADIATION PHARYNGEAL

PADS

Fig. 6. Pharyngeal pads of six fish species with different diets (total lengths of specimens are given in mm).

IN

THE CENTRARCHIDAE

Fig. 7. Seven characters of importance in feeding contrasted in the five common centrarchids of Lake Opinicon. Those characters representing specializations from the more generalized type are indicated by shading.

ments of sandy shallows, gravel and stony stretches, and extensive rocky shorelines. Weedbeds are well developed in inlets where water movement has deposited silt with a high organic content. The inshore weedbeds support as many as 15 aquatic plant species. The most prominent plants are Myriopkyllum, Ceratophyllum,Potamogeton robbinsi, P. richardsoni, P. illinoiensis, P. amplifolius, Najas and Chara. There are also areas of dense Potamogeton pusillus-P. pectinatus whose compact nature provides cover for large numbers of young fish. Less diversified weedbeds occur offshore in water up to about 2.8 m in depth. The

morphology is efficiency of resource utilization. The adaptive characters of a species are the 'tools' that it uses to obtain its resources.

Habitat specialization and the use of space Lake Opinicon has large areas less than 3 m in depth, many islands, and extensive areas of shoreline (Fig. 2). It is not sufficiently deep to support salmonids. The inshore areas that are important fish habitats are highly diversified, being composed of seg-

Table 3. The major habitats of Lake Opinicon fishes, June. The symbols indicate relative levels of abundance. For definition of habitats see Keast et al. (1971). Weedy shallows Alewife Mudminnow Northern pike Golden shiner Blackchin shiner XXX Bluntnose minnow X Yellow bullhead Brown bullhead Banded killifish XX Brook silverside X Rock bass XX Pumpkinseed XXX Bluegill XXX Smallmouth bass Largemouth bass X Black crappie X Yellow perch X Iowa darter Log perch

Weedy deep

X XXX X XXX

X XX XXX X XX XX

Sandy Mixed bottom shallows shallows

X X X X X X X XX XXX X

XXX X X XXX XXX X X

X X

X

Gravel beds

Rocky shelf

Cattail fringes

Submerged stumps

Open water inshore

Midlake water column

Open lake bottom

X X X XXX X X X

Table 4. Numbers of fish per 1000 mZ in different habitat types in Lake Opinicon, June and September, 1975. Values are the means of combined day and night strip transect counts. Bluegill Pumpkinseed Weedy shallows June Sept. Weedy

June Sept.

Rocky shelf

June Sept.

Gravel

June Sept.

Sandy shallow

June Sept.

Rock bass

Largemouth bass

Black crappie

Yellow perch

Blackchin Bluntnose shiner minnow

different habitat types have a patchy and discontinuous distribution in the lake (Fig. 2). As in other lakes the margins of Lake Opinicon are the most productive section. Eighty to 90% of the fish biomass and 68% of the invertebrate biomass occurs at depths of less than 2.5 m (Keast & Harker 1977b). By contrast, the central areas support only a limited population of alewives. The lake bottom here is composed of suspended sediments and is oxygen deficient, the only potential fish prey being large Chironomus larvae and Chaoboms. Fish are present in the lake margins from May to October, but spend the winter in deeper waters offshore. There is apparently no feeding during the period of ice-cover except in cold-adapted species like the yellow perch (Keast 1968a). The habitat preferences of the various Lake Opinicon fish species are summarized in Table 3. The various habitats are rated according to their level of importance to the species. These data have been obtained by selective netting (seine, gill, and trap) over several years. In addition, quantitative data on the fish communities of the various inshore habitat types have been obtained by the strip count method (Keast & Harker 1977a). The results are summarized in Table 4 and Figure 8. WEEDY SHALLOW

SANDY

ROCKY SHELF

GRAVEL

MID-LAKE

WEEDY DEEP

LARGEM MOUTH

lBLACKCHIN

GOLDEN

[?IIl ALEWIFE

BLUEGILL BASS PUMPKINSEED YELLOW PERCH ROCK BASS @KILLIFISH BLACK CRAPPIE SHINER

SHINER BLUNTNOSE MINNOW

SILVERSIDE

Fig. 8, Fish species compositions in terms of percent numerical abundance, six habitat types, Lake Opinicon. The results are based on day and night counts made during sixteen 1000 m transects between the 15th and 25th of June 1975 (Keast et al. 1977).

A number of fish species are confined to single habitats, or are more abundant in one habitat than in others. Thus blackchin shiner inhabit only the inshore shallow weedbeds where they are common, the log perch is a rare inhabitant of gravel segments, the mudminnow is confined to the muddy cat-tail margins, and the alewife is largely a mid-lake species. The golden shlner is mainly found in areas of open weedbed in deeper water. The banded killifish inhabits exposed sandy shallows and weedy sections, with many occurring in water less than 0.4 m in depth. In contrast with these 'habitat specialists', larger centrarchids like the bluegill, pumpkinseed, and largemouth bass occur in most inshore habitats. The bluegill is also by far the most numerous species (Table 4). The pumpkinseed has 1/10 to 115 the abundance of the bluegill and shows a greater preference for sandy and the shallower weedy sections. The rock bass, with densities characteristically less than 1/20 that of bluegills, is found mainly in rocky and weedy areas although night time dispersal occurs into sandy areas. Largemouth bass are present in small numbers. Most of those inshore are 0 and I age class fish. The marginal shallows support relatively fewer fish in autumn than in spring (Table 4), but the habitat preference patterns remain the same. Bluegill partly vacate the shallower waters for offshore areas in late summer and autumn. A few species show marked seasonal shifts in distribution. Thus the darter Etheostoma exile, an uncommon species, occurs in the inshore shallows only in May and September-October. The reason for this is not known. The brook silverside spawns in open water in June and July (Hubbs 1921); it is absent from inshore areas during this time. The bluntnose minnow, an inhabitant of the deeper weedy areas for most of the summer, partly disperses along the exposed rocky shoreline areas in September. Three species show marked die1 distributional shifts that take them from one habitat to another. The black crappie spends the day resting amongst the Myriophyllum and other weeds and after dark disperses into the deeper waters and, to a lesser extent, along the shoreline to feed. The stomachs of fish captured inshore after midnight are commonly filled with Chaoborus larvae - a prey that mainly occurs in waters of depths of 3.4 m and more. The nocturnal movements of rock bass involve even the smallest fish. Few rock bass are found in open sandy shallows by day but numbers move there to feed at night (Fig. 9). Brown bullheads rest by day on the bottom at depths of up to 6 m, and also in weedy inshore

areas with sunken logs and stumps. The fish disperse widely, including into the marginal shallows, to feed at night. With these exceptions the species compositions in the different habitats do not change by day and night. There are however marked changes in the numbers of bluegills and pumpkinseeds in a number of habitats after dark. These die1 variations, which are strongest early in the year, become progressively less so during the course of the summer. By August the numbers of pumpkinseed in the sandy shallow show no day to night changes. This contrasts with the rock bass, which continues to show day-night distribution differences throughout the ice-free months (Fig. 9). Emery (1973), working in a series of larger Ontario Lakes, found pronounced daylnight distributional shifts in a variety of fish species. Living and feeding space is divided in the vertical as well as the horizontal plane. The most obvious surface feeder is the brook silverside. The golden

SANDY AREA ROCK BASS

shiner, killifish, and bluegill also do some surface feeding. The bluntnose minnow, brown bullhead, and to some extent perch are bottom feeders. Most species feed either in midwater or utilize all depths. Underwater observations in lakes by Werner et al. (1977) have shown that 99% of the swimming of the pumpkinseed is near the bottom, whereas the bluegill usually swims well up in the water column. Such a vertical separation also characterizes young of the two species (Werner 1968). This difference is also broadly true in Lake Opinicon.

Community aggregations in Lake Opinicon Quantitative visual observations by means of repetitive strip counts show that the different habitats support characteristic species combinations in which the commoner species make up more or less consistent proportions at any time (Keast et al. 1977). The lake is thus made up of a mosaic of different habitats (Fig. 2) which differ in their species associations (Fig. 8). Bluegill dominate all the inshore associations but age classes of this species are differently represented in the various habitats (see later). It is a moot point as to whether these habitqt associations can be regarded as communities; the species may co-occur because of common needs in terms of cover and food. Nevertheless the fact that they are clumped in this way has important ecological implications.

Spatial segregation o f the age classes I

PUMPKINSEED

Total f i s h length ( m m )

Fig. 9. A comparison of diurnal and nocturnal distributional shifts of rock bass and pumpkinseed in a shallow sandy bay, Lake Opinicon. The former move into the area to feed after dark (August, 1975).

Larval and juvenile fish of the various species also partly separate out on the basis of habitat. The inshore open waters are occupied by the young of the yellow perch, log perch, black crappie, largemouth bass, pumpkinseed and bluegill, although this occurs successionally as a result of differences in spawning date (Amundrud et al. 1974, Keast 1977d). Weedbeds form the habitat of young blackchin shiner, golden shiner, banded killifish, and also sunfishes older than one month. The greatest density of juvenile alewife occurs in the deeper waters. All the longer-lived fish species show a degree of age class habitat separation. In the two sunfishes age classes 0-11 inhabit the weedbeds whilst the older fish are found mainly in exposed areas. This is a simplification, for the bluegill shows a complicated pattern of habitat separation by age class (Keast 1977a). In this species juveniles move from the open water

into the dense Potamogeton pusillus-pectinatus beds at the age of about one month, year classes I and I1 are mainly in the more open and diversified beds, a high proportion of year class 111-IV fish occur in loose aggregations in open waters offshore, and the largest fish tend to be in exposed rocky and other marginal areas along the shore. There are various differences in habitat utilization patterns with age between this species and the pumpkinseed (Keast 1977~). Age class habitat separation also characterizes black crappie, largemouth bass (year 0-1 fish in the shallower water), rock bass (young fish in the shallowest areas but hiding by day in crevices), and yellow perch. Separation does not occur, however, in

short lived species like the bluntnose minnow and banded killifish, where the schools are made up of individuals of varying ages and sizes. The Lake Ppinicon studies thus show that, despite the relative smallness of this lake, habitat specialization is marked. This applies both between and within species. Distributions may vary both diurnally and seasonally, but the preferences of the various species are fairly clear cut. The fact that some species change habitats as they grow suggests that space needs of individual species may be quite complex. These findings require testing in other lakes where the habitats, competitors, and, particularly, predator types and abundances are different.

Table 5. Food (% volume) of Lake Opinicon fish, based on grouped data for June, 1969-1971 (Harker 1976, Keast 1977, a-c and unpublished data). To keep the treatment as comparative as possible the 111 age group has generally been selected in the case of the longer-lived species, except for Micropterus, where the data are from year class I1 fish. L = larvae; P = pupae, N = nymphs.

Number of fish

63

73

28 2 3 4

2

4 8

3 24 1

29 3

5 14

4

Food organisms: Copepoda Cladocera Ostracoda Amphipoda Isopoda Decapoda Adult Diptera Chironomid (L) Chironornid (P) Chaoborus (L & P) Diptera (L & P) Trichoptera (L) Hemiptera Coleoptera (L) Ephemeroptera (N) Zygoptera (N) Anisoptera Hydracarina Mollusca Annelida Turbellaria Fish Filamentous algae Plant remains Detritus

1 4 4 1 3

1 10 6 33

3

Total body length ranges of the fish are: bleugill and rock bass, 90-1 10 mm, pumpkinseed and black crappie 95-115 mm, largemouth bass 95-130 mm, yellow perch 115-135 mm, log perch 80-115 mm, northern pike 200-250 mm, brown bullhead 70130 mm, banded killifish 60-80 mm, brook silverside 65-90 mm, bluntnose minnow 45-60 rnm, golden shiner 95-130 rnm, blackchin shiner 45-60 mm, alewife 95-125 mm, mudminnow 60-80 mm.

prey of weight 0.1-1.0 mg makes up the diet of the year class 0 fish, 1.O-70.0 mg the year class 111 fish, and 100-250 mg the year class VII-X fish. Figure 11 contrasts seasonal dietary shifts in three generalist and two specialist feeders. In the generalists the categories of prey may change seasonally but in the specialists only the proportions change. Most seasonal changes in diet can readily be explained by either (a) changing abundance levels in the major prey organisms and/or (b) organisms of appropriate body size (to the particular species) becoming available. For example because of their small mouths bluegill can only take small-bodied, newly hatched Anisoptera nymphs, whereas the largemouthed rock bass consume nymphs of larger size. Seasonal shifts in abundance of six major prey invertebrates are shown in Figure 12. Further details, including a discussion of sampling methods, are given in Keast (1977b, c). The high isopod consumption of yellow perch and pumpkinseed in May can be linked to the numerical peak of these prey items at that time, as can the

,,,

M

YELLOW

J

J

A

S

eating of amphipods by the two species from August onwards. The high autumn consumption of isopods by pumpkinseed is linked not to numbers but to the predominance of large-bodied prey individuals in the population at this time (Keast (1977~).The fish feed selectively on these larger prey. These examples contrast with cases where consumption of a resource is relatively uniform regardless of fluctuations in prey abundance. The consumption of chironomid larvae by several fish species throughout the summer is one example, as is the eating of gastropods by the larger pumpkinseeds.

Diet overlaps between year classes and between species The food relationships between year classes and spe. cies can be put into a quantitative framework by cal. culating diet overlaps. Where species differ markedly in their diet, overlap levels will be low. Where, how ever, common items make up a significant percentag of the diet of two species they will be high and (theo retically) the species will be in potential competitior should that resource become limiting. Diet overlaps are given for five common Lak~ Opinicon species that occur in the same habitat (Ta ble 6). Additional values are given for the bluegill anc

O

PERCH

I Fig. 11. Monthly changes (% volumes) of three generalist feeders, pumpkinseed (Lepomis gibbosus), bluegill (L. macrochirus), yellow perch (Perca flavescens), and two specialized feeders; alewife (Alosa pseudoharengus) and blackchin shier (Notropis heterodon). The fish were taken between the 15th and 25th of each month and belong to the same size categories as in Table 5. Individual monthly series are of 50-80 individuals and represent the grouping of individuals taken at equivalent times over a three year period to get a sufficiently large sample size. For original data on the f i s t three species, and rationale for grouping data from different years, see Keast (1977a,b). The data on Alosa and Notropis are unpublished.

M

J

J

A

S

O

N

60

TJ LO

E

M

J

,

J

A

S

O

N

r n o n t h

Fig. 12. Numerical fluctuations in standing crop of six cc mon benthic invertebrates important as fish food, Cow Isli Bay, Lake Opinicon (see Keast 1977a,b).

Feeding specializations and the subdivision of food resources

Diets and food separations In an initial study on the feeding of the fish inhabiting Birch Bay, Lake Opinicon (Keast 1965), it was concluded that (1) the various species separated out to a significant extent both in terms of taxonomic categories and sizes of food organisms consumed; (2) whilst there were overlaps, most species ate one or two food types to a disproportionately greater extent than others; this was true even of generalists like bluegill; (3) the fauna was a blend of (a) specialist feeders that subsist on 3-4 basic prey types irrespective of whether the prey are rare or common, and (b) generalists that at any time may contain up to 8-10 different prey types; (4) older and younger age classes within species may sometimes differ strikingly in their diet (e.g. the rock bass); and (5) diets usually vary seasonally. Analysis of seasonal fluctuations in the resource base showed that such changes could commonly be explained in terms of a population peak of a particular prey organism, the presence of prey of a particular size or alternatively, loss of a prey type due to its emergence or its becoming too large to be consumed. An overview of the diets of 16 major species in Lake Opinicon on the basis of later data can be found in Table 5. The table demonstrates that there is both

PUMPKINSEED

BLUEGILL

ROCK BASS

LARGEMOUTH BASS YELLOW PERCH

Fig. 10. Diet changes with age in five common species for the month of June. Size of segments represents percentage volume of each prey type in the stomachs. The data are based on the analysis of the stomach contents of 40-80 individuals of each year class of each species.

a significant degree of food specialization and of diet overlap between fish species. Clidocera are take; by many species and make up at least 19% of the stomach content volume in nine of the 16 species. Chironomid larvae are prominent in the diet of six. However within both categories there is some separation at the generic level. Bluegill and blackchin shiner eat the cladoceran Bosmina in the shallow inshore waters, yellow perch take the largerbodied genera such as Simocephalus and Ceriodaphnia, whilst alewife obtain mostly Daphnia from the deeper waters. Because of difficulties in identification, most chironomid larvae were not identified beyond family level. Black crappie feed largely on the free-swimming Procladius whereas brown bullhead consume largely tube-dwelling types, apparently by sifting them from the mud (Keast 1968b). Relative specialist feeders include the largemouth bass, pike, brook silverside, bluntnose minnow, alewife, black crappie, and blackchin shiner. At the other extreme the two sunfishes, brown bullhead, banded killifish, and mud minnow are generalists. Longer lived fish species change their diets progressively as they grow. Details of some of these changes are given in Keast (1977 a-c). They can readily be catalogued because fish populations at these latitudes have series of distinct age classes, growth (linear increase) being limited to about 10 weeks per vear. Dietary changes with age are shown for five common species in Figure 10. In the pumpkinseed these mainly involve the progressive replacement of chironomid larvae by molluscs. In bluegill, by contrast, Cladocera dominate the diet of the younger fish whereas the older fish take more Trichoptera larvae, vegetable material, and small-bodied Anisoptera nymphs; the diet of these older fish is also more diversified. In these two species, and the yellow perch, the age changes in diet are gradual and involve shifts in the proportions of the major foods eaten. In the rock bass, by contrast, there are three distinct age-specific diets that involve shifts from chironomid larvae and other small prey to Anisoptera nymphs,, and medium-sized insects and, finally, decapods and small fish. The largemouth bass is already predominantly piscivorous at the end of the first summer and subsequent diet changes involve the taking of successively larger fish. It is characteristic of all fish to take prey of increasing size as they grow. In the bluegill prey of 0.1-0.75 mg weight is taken by year class 0 fish, and 0.5-15.0 mg by year class VI fish. In the rock bass

Table 6. Diet overlap values (Levins Index) for the month of June for fish of various species of equivalent ages (see Table 5). Values are for species occurring in (a) the sandy shallow areas of Lake Opinicon compared with (b) two pairs that replace each other in different habitats. The top figure in each cell represents the overlap A ij, and the bottom figure the reciprocal A ji.

Pumpkinseed

Bluegill 0.29 10.67

j Largemouth bass 0.02 10.0 1 0.01 10.03 0.24 10.3 1

Rock bass 0 ' 2 0 / ~23 . 0.24 10.19

Bluegill Rock bass i Largemouth bass

Banded killifish 0.15 10.16

Black crappie 0. 05 0.04

0.3910.20 0.01 10.03 0.01 /0.01

0.2010. 08 0.26 10.20 0.25 10.23 0.08 10.07

Bluegill 0.26 10.56

Mudminnow

Banded killifish (b) Blackchin shiner i Alewife

j Golden shiner 0.73 10.87

Alewife 0.651 1.06

0'80/0.58

pumpkinseed, the two species whose diets have most in common (and that are the two commonest species in the lake) in Table 7. The latter comparisons are on the basis of monthly values and for three different year class groups I, 111 and VI-VIII). The overlap index of Levins (1968) is used. On this scale, which ranges from 0 to slightly over 1, an overlap of 0.3 or less is insignificant, and one of 0.7 or more is considered high. In overlap calculations involving pairs of species the overlap value of the first on the second may differ from the reverse. This is because the overlap (i.e. 'impact') of a generalist feeder on a specialist will be greater than the reverse (Cody 1974). For this reason pairs of reciprocals, not just means, are given. The overlap formula of Levins (1968) is as follows:

i

Brown bullhead

the bluegill and pumpkinseed are the overlaps potentially significant (Table 6). When the overlap values of the latter two species on each other are compared (Table 7) four observations emerge: (1) almost all are in the medium range, (2) they vary from month to month, (3) in a few cases the values are quite high, occasionally low and, (4) there is generally little difTable 7. Monthly variation in diet overlap between three equivalent year class groups of pumpkinseed on bluegill. Data are based on % volumes of different taxonomic categories of foods in the stomachs. The top figure in each cell represents the overlap of pumpkinseeds on bluegills, and the bottom figure the overlap of bluegills on pumpkinseeds (index of Levins 1968). Year classes of fish

I

A ij (niche overlap) =

t' pih pjh/ h=21 pih2 h=1

where cyij is the overlap of species j on species i relative to the niche breadth of species i; Pih (h =1, ..., S) is the proportion of a particular item h in the diet of species i; and Pjh is the proportion of a particular item h in the diet of species j. May (1975) shows that, of the niche calculations available, the Levins formula seems to be the most appropriate, exhibiting good mathematical reasoning and biological relevance. As would be expected from their rather different diets the diet overlap values for rock bass, largemouth bass, killifish and black crappie on each other and relative to the two sunfish are minor; only between

0.3710.49

Mudminnow

May

0.42

111 0.66

0.37

VI-VIII 0.22

0.67

0.36

June July August

0.26

0.53 0.13

September October

0.5 1 0.23

0.20

0.66

0.54 0.43

0.32 0.29

0.32

0.35 0.24

0.55 0.50

0.33

0.74

ference between the reciprocal pairs. Reference to the original food data on the species (tabulations in Keast, 1977a, b) reveals the reasons for the moderate overlap values between the pumpkinseed and bluegill. Twenty-two to 50% of the diet items by volume of these year class 111 fish is shared. The actual figures for May, when the overlap figures are high, is 66%; in September, when overlaps are low, 23% of the diet items are shared. The reasons for the occasional marked disparity in the overlap values between the reciprocal pairs is discussed elsewhere (Keast 1977~).The disparity occurs apparently because in these months one member of the pair took a disproportionate amount of a single resource and thus emerged as a specialist feeder at these times. This is illustrated in the pie diagram for year class I pumpkinseed and bluegill (month of June) in Figure 10. Diet overlap phenomena must, of course, be interpreted relative t o the fluctuating resource base. In previous reviews of the monthly changes in the diets of the year classes of bluegill and pumpkinseed (Keast 1977a, c), it was found that overlap values increased at two times: (a) when a favoured resource peaked in abundance and several age classes were simultaneously attracted to it, and (b) in July when, as a result of emergences, etc., the diversity of available prey organisms fell. It has been noted elsewhere (e.g. Keast 1965) that the peaking of favoured resources like amphipods may also simultaneously attract several species of fish. The data in Figure 12, which shows the seasonal peaks of six major prey organisms, might be expected to explain some of the higher diet overlap levels in Table 7. For example, overlap figures for year class I11 bluegill and pumpkinseed are high in May because chironomid larvae and several lesser items are being shared; bluegill did not however take the isopods that peaked at this time. This observation reinforces previous evidence that the food preferences of these two species are basically different. Co-occurring species interact, of course, in many dimensions, not just in diet. The bluegill and pumpkinseed also show partial habitat separation. Diet and habitat may be combined into a single overlap value by taking the mean of the sum of the two values ('summation alpha' of Cody 1974). When this is done for Lake Opinicon data, combined overlap values of approximately 0.5 result (Keast 1977~).This figure corresponds well with those obtained by other workers for co-occurring species pairs, vide 0.56 for grassland birds (Cody 1968), 0.47 for Australian lizards (Pianka 1969) and 0.51 for blackbird species in Washington (Orians & Horn 1969).

Most researchers who utilize overlap calculations warn that the method is subject to severe limitations. It is, for example, mathematically impractical to derive calculations for more than two niche dimensions at once. Invariably not just two or three but a multiplicity of factors are involved. The figure of 0.5 for combined overlap between bluegill and pumpkinseed may be too high because (a) identification of the prey organisms was only taken down to the ordinal level, a low level of resolution; (b) diet overlap levels fluctuate seasonally, sometimes dropping to 0.2-0.3; (c) the fish feed, at least in part, at different heights in the water column; and (d) it seems probable that the larger members of both species can adjust to declining food levels by changing distribution patterns. Seasonal shifts in distribution (Keast & Harker 1977a) from the shallows into deeper waters may be resource based, notwithstanding that many fish tend to maintain home ranges (Gerking 1958). The most telling argument that pumpkinseed and bluegill are basically different ecologically comes from the great numerical abundances of the two species; it is obvious that real overlap values between the two must be small. Similar problems of explaining successful co-occurrence despite seeming high ecological overlap values have confronted other workers (e.g. Fraser 1976). Diet overlap calculations may be used to test if species occupying a common habitat are really more distinct ecologically than those occupying different habitats. Table 6 gives figures for pairs of species in the inshore sandy shallows of Lake Opinicon and for two sets of species that largely, or partly, replace each other in different habitats. The expected difference is borne out. The planktivorous but spatially separated blackchin shiner, golden shiner, and alewife have high overlap values, whilst bluegill, brown bullhead, and mudminnow have moderately high values. These are obviously not examples of one species competitively displacing another from a habitat, although the presence of the one may reinforce the habitat specialization of the other. The alewife only colonized Lake Opinicon in 1964.

The differential use of resources in time The seasonal component in fish ecology and in the ecological separation of species has not received the attention it merits (see Emery, these proceedings). Yet a seasonal factor dominates the cold-temperate latitudes, limiting most life processes in poikilotherms

to the short summer months. In most fish species, feeding, reproduction, and growth are limited to at most 5-6 months of the year. As might be anticipated, physiological differences between southern and northern derivative species would be expected to have far reaching effects on their seasonal activity cycles and temporal patterns of resource utilization. The 'annual cycles' of a represen-

tative of each group are contrasted in Figure 13. Bluegill taken from under the ice in mid-winter contain no food and hence are not winter feeders. They move inshore in mid-May as the water warms, with feeding commencing at 8-10' C. At this time they are noticeably thin and the caloric content of their tissues is 3900-4200 Kcal per gram, compared to 4500 Kcal at the end of summer (Keast, unpubl.).

m'.1 deposit pre-winter fat

weight Increase

growth

1

(

Linear)

move inshore

[ Y E L L O W PERCH

1

move

m] deposit pre-winter fat

g r o w t h (linear) f e e d i n g

///////////////////////////////////////

1

\

ice

M O N T H Fig. 13. Annual cycles of bluegill sunfish and yellow perch. The bluegill belongs to the Centrarchidae, a family of southern origin while the yellow perch is a member of the Percidae, a more cold adapted family.

The summer weight-to-length ratio (condition factor) is restored by about the end of May. Growth (length increase) commences in early June and continues for about 10 weeks, until mid-August. The prewinter subcutaneous fat levels are restored by the end of July. Our data suggest that the food uptake rate in May and June is about 2.5-3.2% of body weight per day and falls to 1.8-2.2% in July, August, and September. Bluegill taken at the beginning of spring have very small ovaries; enlargement begins shortly after the commencement of feeding in mid-May. By the approach of spawning in midJune, at a water temperature of 22"C, the ovaries of mature females average 10% of body weight. This figure fluctuates but remains high for the next three weeks as successive batches of eggs mature and are spawned. With the completion of spawning in midJuly the gonosomatic index drops sharply. Bluegill remain inshore until the early autumn when a prewinter withdrawal to deeper water begins. Yellow perch, by contrast, feed throughout the winter. No data are available as to the amount of food eaten. Ovarian development, which begins at the end of summer, is accelerated during the winter; by the time the ice goes out yellow perch are ready to

spawn. Prior to spawning in late April or early May the gonosomatic index reaches 25% of the body weight; all eggs mature simultaneously in this species. There is a mass spawning of some five days duration at a water temperature of about 10" C. Perch do not show any loss of body condition at the end of winter, because the body is then distended by the large gonads. As far as is known the duration of the period of length increase in yellow perch is about the same as that of bluegill. It is apparent from the above that the different temperature adaptations of these two species lead to a degree of temporal ecological separation. The energy invested in gonad development in yellow perch is obtained partly in mid-winter, when bluegill are not feeding. As discussed elsewhere (Keast 1977d), the spawning seasons of the various Lake Opinicon fishes differ to an extent that may permit their young to harvest a common food resource in sequential fashion. Young yellow perch appear first (late April or early May), followed by those of log perch and, more or less in succession if the water warms gradually, by young black crappie, rock bass, pumpkinseed, and bluegill. Yellow perch, as noted, have a single mass spawning

Table 8. Diets of juvenile fishes in the inshore open waters of Lake Opinicon. Data are average numbers of organisms per 18 fish, based on samples from 50 stomachs. Food organism length (mm)

Cyclops

Nauplius 0.1-0.2

0.1-0.2

0.2-0.3

Bosmine 0.3-0.7

0.1-0.2

Fish total length (mm) Yellow perch - May 14-15 10 7 11 9 Log perch 4 6 40

-

May 14-15 & 21-23 3 10

1 1

Black crappie - June 12- 17 18 4 25 16 35 8 Pumpkinseed plus bluegill

-

July 17-

0.2-0.4

Sida 0.1-0.4

Diaphano- Other soma 0.3-0.4

whilst the others, including the log perch spawn two or three times. Since the largest numbers of eggs in small species seem to be released at the first spawning, the degree of overlap of the different species is only partial. Pike spawn before yellow perch but in a different habitat. When they first hatch the young of all species are dependent on larval and small cyclopoid copepods, the only abundant prey of sufficiently small size (Table 8). At an age of 7-10 days Bosmina and larger cladocerans are incorporated into the diet. Equivalent size juveniles of the various species have a more or less common diet for the first few weeks of life. Hence temporal separation is of basic importance. An interesting facet of resource use ecology, and one that merits proper investigation, is the extent to which different optimum temperatures for feeding and growth may serve to partly separate species ecologically. Such differences, for example, may well be expected in species of northern and southern origin. Some years ago the writer ran a series of experiments in which the food uptake and growth rates for six species were tested at temperatures from 5 to 30' C (discussed briefly in Keast 1970). These experiments showed some species differences. Mudminnows achieved optimum conversion at 20' C and performed very poorly at 30° C, whereas bluegill and banded killifish performed optimally at 25' C and did very well at 30' C. Mudminnows fed well at l o 0 C, whilst bluegills did not. There is evidence that certain fish species feed under the ice and that some commence to feed earlier in the spring than others (Keast 1968a). In general larger-bodied poikilotherms can operate at lower temperatures than smaller ones and this may materially effect feeding rates in spring and fall (Heckrotte 1967, Asplund 1974).

The use of space and other resources: species interactions and a summary The present review demonstrates the following: (1) Each species has certain basic morphological attributes that adapt it for a particular role and channel it away from potential competitors. To judge from the long representation of some taxonomic groups in the fossil record, many of the adaptive structures are probably old: the Cyprinidae, for example, are known from the Paleocene, the pikes from the Middle Eocene and centrarchids from the Miocene (Scott & Crossman 1973). (2) The lake is not a single habitat, rather it is a

mosaic of different habitats or subhabitats. The fish respond to this mosaic by showing habitat specializations or preferences. Species compositions and ratios vary from one habitat to the next. These assemblages are hardly 'communities' in the classic sense; presumably they are groupings of species with similar needs (e.g. cover). (3) Habitat separations not only characterize different species but also characterize age groups of species. Age group separation presumably serves to minimize intraspecific competition, but may mean that the species encounters more complex interspecific competitive situations. (4) Some species are generalist feeders, others specialists. In the former the diet varies seasonally as the fishes respond to population peaks of prey organisms. Specialists utilize the same narrow range of prey types but the proportions vary on a seasonal basis. Species may be generalists at one time of life, specialists at another. Most fish species have one or two prey types that they take to a proportionately greater degree than other species. As with habitat preferences, some species show marked dietary changes with age. (6) Ecological overlap values are low for the different species with the exception of those between the congeneric bluegill and pumpkinseed. Feeding overlap values may be high when species occur in different habitats, as is the case for planktivores. Alewives inhabit the mid-lake regions and the golden shiners frequent weedy areas in deeper water. Diet overlap levels between these species are high. Diet overlap between co-occurring species may fluctuate seasonally, a phenomenon that reflects the varying patterns of resource utilization. When a resource peaks, it commonly attracts two or more species and year classes; a common diet hence does not necessarily indicate increased competition. (7) Ecological overlap values based on only two parameters ( e g food and habitat) may be artificially high. Co-occuring specks differ ecologically in a wide spectrum of ways, e.g. sizes of foods eaten, height in water column at which they feed, and so on. This concept of multidimensional differences has to be invoked to explain the co-occurrence of bluegill and pumpkinseed, which have much in common and are also the most abundant species in the lake. Whether or not one defines 'niche' in terms of Hutchinsonian hyperspace (Hutchinson 1958) or in more simple terms of food and hrabitat utilization, it is apparent that the concept of niche in fishes is more complicated than in homeotherms. Protracted attain-

Table 9. The total number of food "niches" and distinct habitats occupied by the various species of Lake Opinicon fish, based on diet and habitat differences between year classes. The divisions are necessarily somwhat arbitrary. Species

Number of distinct diets

Number of different habitats occupied at different intervals of life

Alewife Mudminnow Northern pike Golden shiner Blackchin shiner Bluntnose minnow Yellow bullhead Brown bullhead Banded killifish Brook silverside Rock bass Pumpkinseed Bluegill Largemouth bass Black crappie Yellow perch Log perch Number of Linnean species - 17 Number of "ecological" species - 24-27

ment of maturity and indeterminate growth mean that the species may be composed of not one but several size groups with differing ecologies. If one uses the term niche t o mean characteristics of diet and habitat, as has commonly been done for birds, it can be argued that some fish species simultaneously occupy several niches. Table 9 lists for each Lake Opinicon species the number of distinct diets and habitats occupied at different times of life. The criteria are differences of the kind that separate taxonomically distinct species, i.e. the diets of years classes 0, 11, and V-VI rock bass are as different from each other as are those of rock bass and yellow perch, or yellow perch and bluegill. Argued in this way it could be said that the 17 Linnean species of fish listed are actually the equivalent of 24-27 'ecological' species. This approach is obviously arbitrary, for ecological differences between age classes may be marked or slight. The concept is introduced, however, t o highlight a feature of fish faunas that must be considered in any consideration of interspecific competition.

Conclusions The major issue in the present study is how all of these differing ecological attributes and strategies combine to form a lake fish fauna such as that of

Lake Opinicon. Firstly it is apparent that the total number of species present is in part a reflection of the number of habitats available and diversity of the lake. In aquatic environments species diversity may be linked to substrate diversity (Kohn 1967, Abele 1974, Roth 1976, Emery, these proceedings). Secondly can Lake Opinicon be regarded as a system where resources are limited and competition is important? Because of the highly seasonal nature of the system I suggest not. Different resources are not produced at a uniform rate, but peak numerically at different times. A succession of 'superabundant' resources is thus produced and species feeding on such are not in competition. The 'feeding season' of the fish then could be thought of as times of 'superabundant' food alternating with times of reduced food abundance. A generalist like the bluegill is able to harvest a number of resource types in turn as they peak in abundance. A poikilotherm can, of course, survive without feeding in times of scarcity, thus increasing its options in the way of resource and space utilization. No effort has been made to determine if bluegills suffer from food shortages at certain times. If in fact food is not limiting some other factor must be keeping populations in check. The Lake Opinicon system is obviously characterized by a great deal of adaptability and 'give and take'. Low fish predation levels in winter ensure high food availability in early spring, when the fish need this to restore bodily condition and initiate growth. The entry of new generations of prey invertebrates through the summer ensures a constant replenishment of food resources which species can respond to according to their morphology and preferences. Such capacities as ability to-vary diet with prey availability, having more than one 'food niche' and multiple spawning should be noted here. In conclusion, it is appropriate to ask how the morphological adaptations and space and food utilization patterns described for Lake Opinicon compare with those characterizing the fish of water bodies in other geographic areas. Compared to a tropical situation a number of observations mav be made. In Lake Opinicon, the ratios of Linnean families to genera and genera to species (9 families, 15 genera, 17 species) are high and infer that at these northern latitudes a minimum degree of morphological differentiation is characteristic. Morphological diversification does not approach that of tropical systems (Fryer & Iles 1972, Lowe-McConnell 1975). Note that even marginal teeth may show striking,morphological diversification in the tropics (see photographs in Liem & Osse 1975).

Werner: You have singled out arbitrary year classes of species (e.g. year class 111) to make comparisons of the diet of different species and in overlap calculations between bluegill and pumpkinseed have limited yourself to equivalent age classes. Since the morphological component equipment of species differs so widely in the community many species will utilize common resources at different ages and this will determine when the relevant competitive interactions will occur. Have you attempted to identify such age classes where diets or habitat use are 'most similar' and evaluate competition at such points? Do these constitute 'bottlenecks' to the recruitment of particular species? Keast: In this paper I have attempted to develop a broad review only, limiting myself mostly to the kinds of differences and degrees of separation that occur. The morphological component does indeed suggest that, in longer lived species, it may not be equivalent age groups but ones of disparate age that have the most similar diets. For example in mouth width the year class I rock bass is the equivalent of the year class VI bluegill. These two groups do indeed have more in common in diet (both take chironomid larvae, amphipods) than do the old rock bass and old bluegills. Rock bass are not a common species, however. I wonder if one should not look to the interactions of cyprinids and young bluegill for examples of the kind of thing you mention. I confess I have a lot more thinking to do about this kind of thing. Incidentally Fraser (1976) found that in two species of Plethodon salamanders 'competition' was between the adults of one and young of another species. Larkin: Large volumes of observational data can usually be interpreted in a variety of ways, and however ingenious the interpretations may be they are not convincing. Dr. Keast has evidently achieved a rare degree of familiarity with Lake Opinicon, and is led by a mass of evidence to the conclusion that the 17 species of fish in the lake have sorted out their partitioning of resources. He states that 'it can be argued that' many water bodies, including Lake Opinicon today support their full compliment of species, given the available fauna. This may well be true, but it would be useful to have something other than circumstantial evidence. From the data presented it would be as convincing to argue that there was obviously room to shoehorn in at least a dozen more species, especially if they were to be relatively rare in the lake.

Keast: The point is well made. I was talking in general terms because in virtually all lakes in eastern Ontario of a minimum size one finds bluegill, p u m p kinseed, largemouth bass, black crappie, and rock bass, for example. Secondly various species must repeatedly have had the opportunity to establish in Lake Opinicon but have failed to do so. White suckers (Catostomus commersoni) occur in streams near the lake but only rarely has one been found in the lake. The troutperch (Percopsis omisc;omaycus)is taken by gdl net in the centre of the lake every few years (including a gravid female this year) but does not become established. Walleye and lake trout occur in adjacent lakes of the Rideau system but not in Lake Opinicon (which is apparently too shallow and warm). Actually one new species has established itself in Lake Opinicon in later years, the alewife in 1964. I would agree with Larkin that the food and space utilization data reviewed in this paper could be taken as prima face evidence that additional species could be 'shoe-horned' into Lake Opinicon. But how many would establish permanently, how many only survive a few years? To think of a lake having an upper number of species, or species constapcy, as I have tried to do, may be naive. Possibly we should think of a lake in terms of island biogeographic theory, i.e. that there is constant species turnover, at least where lesser species are concerned. One investigator (Keddy 1976) has attempted to explain the occurrence of aquatic plants in terms of the theory. Note also that water bodies change in time due to silting, weed growth, etc., i.e. what a lake has to offer fish in the way of habitat is changing. This could lead to the establishment of new species and extinction of others. Zaret: I would argue that in these lakes the preponderance in both numbers and biomass of one generalist species, the bluegill, filling several existing niches whch might someday be occupied by other, perhaps several, species is proof in itself that the lake faunas are not saturated. Supposing again you took a series of 20 different Ontario lakes, what range would you find for the ratio of number of species per lake divided by number of species par 20 lakes? Keast: That is a good point about the bluegill. However, alternatively one might argue that the shear breadth of the bluegill's niche represents an optimum adaptation for life at these latitudes and there might be a high selective premium for retaining it in present form. The calculations you suggest would be interesting to develop. I have not attempted anything like that.

Larkin: Keast seems to consider that an adult bluegill is an adult bluegill. He does not seriously entertain the notion that the bluegills of Lake Opinicon may not do the same things as the bluegills in other lakes. Bearing in mind the flexibility that most fishes show in diet and habitat, it would seem reasonable to expect that the various species would make different ecological accommodations to each other in different localities. Keast: My discussions here have been limited to the bluegill in Lake Opinicon. It is a highly adaptable species and would be expected to vary ecologically from lake to lake. For example in the Michigan lakes studied by Werner & Hall (1976) the smaller, sympatric green sunfish (Lepomis cyanellus) displaces the bluegill from the inshore areas. At the same time there is a limit to how much variation one should expect. A survey of the literature on its food from localities as far apart as Kentucky, Tennessee, Virginia and Minnesota (Keast 1977a) showed that chironomid larvae, other small aquatic insects, and zooplankton dominate its diet throughout. However it must also be stressed that the proportions of these different basic food items may vary greatly. This is the case even in different parts of Lake Opinicon: Harker (1976) for example, found different proportions between habitats. Again, bluegill that wander into deeper water and are taken in gdlnets have been found filled with Cladocera - or flying ants that have fallen into the water - at times when the bulk of the population inshore was feeding on a mixed diet of chironomid larvae, amphipods, and Cladocera. I, too, have always been impressed by the ecological adaptability of fish. In Larkin's (1956) review of interspecific competition in fish - a paper incidentally that all ecologists should read - the point was made that fish are more adaptable than birds and mammals. I agree with you. It is a subject that we should take up for detailed discussion some time. For one thing indeterminate growth conveys great plasticity to fish. A stunted or runt population, held down to the size of say the normal year class I1 cohort will be limited to the diet of these small fish. I suspect in Alabama where bluegill reach sizes of half a pound they move into a larger prey size category that one never sees in the north. There are a number of papers providing documented examples of fish species changing their ecology in the presence of a competitor. Nilsson (1963), Andrusak & Northcote (1971), and Werner & Hall

(1976) provide examples. Brown (1977) has recently demonstrated that the yellow perch was a food generalist in a cool deep lake where centrarchids were uncommon but a specialist in Lake Opinicon where there are large populations of bluegill and pumpkinseed. Note that theoretical ecologists have long acknowledged that the niche of a species may differ in the presence or absence of a competitor (Hutchinson 1958, 1959, Colwell & Futuyma 1971). One should equally, however think of the potential adaptability of a species to different physical and environmental conditions. It is becoming very important today to try to get information on how a species will 'perform' under different sets of circumstances. Without this we cannot predict what effect the introduction of a new competitor or a change in lake quality like euthrophication will be. Note here the tremendous changes that have taken place in the fish fauna of Lake Ontario in recent years (Christie 1973). Brothers: I would like Keast's feelings on the general validity or applicability of the conclusions he has drawn concerning relative abundance and presence or absence of species. Keast: I feel that they have general application in the broader sense. But I would like to see them tested (a) over a range of Ontario lakes where the conditions and faunas are similar and, (b) in water bodies to the south of the Great Lakes where faunas are richer and species compositions differ. Most of the species studied here have extensive ranges and one needs to investigate their ecology against different backgrounds of competitor species. Brothers: Infraspecific groupings, species, guilds and even communities are sometimes described as being space, food, predation or 'disaster' limited or controlled. Is it possible to make such generalizations with respect to cold temperate lakes? Keast: I think one must always separate out ultimate and proximate factors. I would not expect to find immediate year-to-year evidence of this sort of thing because what we have today is the result of a long period of evolution. Such factors must, however, have been important in the moulding process. What I am trying to say is this. I have never seen any obvious cases of fish being emaciated due to starvation and overpopulation, of groups of fish being 'forced' into an unsuitable habitat (deep water for an inshore species) by obvious space or food limitations, or of

True herbivores are virtually absent in these cold temperate regions (the stream-dwelling dace, Chrosomus eos, is almost a true herbivore but takes insects in spring). Tropical waters have a great diversity of small-bodied fish species. In cold temperate waters the ecological equivalent of these are the young age forms of the larger species. Witlun the North American continent the greater diversity of fish species in waterways to the south of the Great Lakes has been noted. This increase is presumably accommodated by a finer degree of habitat division. (Mendelson 1975, Page & Schemske, unpubl.) In contrast to this, insular Britain has only a small number of freshwater fish species. A survey of the comparative feeding data published in Hartley (1947) for the River Cam assemblage shows little diet separation. This lack of separation may be a consequence of low species numbers. In these regions and others, it would be interesting to look at the relationships of morphological adaptations between co-occurring species, perhaps by using measures of 'morphological distance' (e.g. Schoener 1965, Karr & James 1975, Smith, these proceedings). In this manner, it might be possible to gain a greater understanding of species coexistence patterns in different fish assemblages.

Acknowledgements The writer would like to express his sincere thanks to the National Research Council of Canada for funding this study through the years. He would also like to thank Queen's University, Kingston, Ontario, for the use of the facilities at its research station at Chaffey's Locks, Ontario, and to Dr. R. J. Robertson, Director of that station. Ms. Jennifer Harker drew the diagrams and calculated diet overlap values and the writer would also like to express thanks for this. The following kindly read and suggested improvements for this manuscript: E. Brothers, G. Helfman, P. Larkin, S. Levin, E. Werner and T. Zaret.

Questions and answers Kaufman: Do you feel that habitat preference is being selected for to avoid competition for food or is it reinforced by some other extrinsic factor like predators? Keast: Certainly there is abundant evidence that the

use of cover to minimize predation is all important to the younger and smaller bodied fish. The bigger fish may occur in weedbeds but the younger ones avoid exposed places. But this also helps keep the year classes of bluegill and pumpkinseed apart. Werner: There is considerable intraspecific size (age) class segregation by habitat (e.g. in the bluegill) which you indicate minimizes intraspecific competition. Much of this separation is because the smaller size classes are associated with progressively denser weedbeds, wluch can be interpreted as a response to relative predation pressures. How do you reconcile and weigh the action of these two factors in determining habitat use among these species. Keast: That is a very good point. To determine the influence of predation in these habitat utilization patterns we need to study them in different lakes where the numbers and diversity of prqdators are different. We also need to study space use in the different age classes of bluegills in lakes where the available habitats are basically different. As far as I can determine there has not yet been an analysis of the influence of a fish predator on the behaviour of a fish prey, certainly nothng at the level of the study of Stein and Magnuson (1976) for the effect of predatory smallmouth bass on crayfish habitat selection and activity. Zaret: Your paper is based on the implicit assumption that the community is organized around competitive interactions, or at least the avoidance of same. Even the Levins equation is one which assumes that there are no higher-order interactions (i.e. from predators). My own studies of fish communities in the tropics have indicated that predation has an overwhelming effect on the community organization, from determining die1 patterns of activity, to habitat locations, and even to certain morphologies that will be successful in the presence of predators. What aspects of the fish's niche do you feel will be most determined by piscivorous preddtors, and how would you go about examining the im ortance of predation 9 in these cold temperate communities? Keast: There is now much evidence that predators, by reducing the abundance of a dominant species, will increase or help maintain species diversity (e.g. Connell & Orias 1964, Paine 1966). This has also been argued and discussed for African fishes (Worthington 1954, Fryer & Iles 1972). There is considerable evidence of morphological features being predator induced (see discussions of 'knti-predator' devices in Lowe-McConnell 1975, McPhail 1977, Zaret %

1977). In the case of Lake Opinicon it could be argued that the protective spines of the centrarchids and ictalurids, the schooling of cyprinids and killifish, the midwater aggregations of the 100-140 mm long bluegill, the habit of resting immobile in the shallows at night, the short intense period of reproduction of the bluegill producing so many young as to 'swamp' predators, are all any anti-predator devices. To what extent then might piscivores be responsible for habitat utilization patterns and 'specializations' in Lake Opinicon? The lake has two top piscivores, the pike and largemouth bass. The larger individuals of these take centrarchids up to lengths of about 160 and 150 mm, respectively. Four other species, (yellow bullhead, yellow perch, rock bass and black crappie) may be regarded as 'secondary' piscivores in that only later in life do they take significant amounts of fish prey, and then only prey up to the length of 60-80 mm. One would not really rate the yellow perch as a piscivore until it reaches a length of about 150 mm (Keast 1977b) but I have caught 70 mm long perch in the denser weedbeds in August literally gorged with 20-30 mm long sunfish. Several fish species, including alewife and the brown bullhead may take larval and juvenile fish. Lake Opinicon thus has a balanced 'suite' of piscivores. Whilst the larger ones keep mostly to the open water, young perch and pike do in fact pursue prey in the inshore weedbeds so that occupation of these does not afford complete protection. The Lake Opinicon piscivores are an integral part of the fauna and they obviously impose certain constraints on 'community' development and habitat utilization patterns. I see interspecific competitive interactions or, rather, the need to minimize this, as the major selective force in morphological and ecological specialization, predation as a less important one. The point made by Dr. Werner and yourself about it being predation that keeps young fish in sheltered habitats is certainly valid. Certainly it is most desirable that the predator effect be tested. Connell (1975), in his review of the relative influence of predation and competition on community structure, suggests three general methods for testing predator effects: the natural field experiment (which in this case would be to compare lakes with and without the predator), develop a predictive model and test it, and the controlled field experiment. This last approach would be difficult as there is no ready way of eliminating a fish predator from a lake. I would concentrate on the first method.

Zaret: Recent arguments (Werner & Hall 1976, Werner 1977) imply that fishes' habitat separations within guilds are far more effective or important than food separation in preventing species overlaps. Do you agree that in fishes habitat separation is more common and more important? Keast: In cold temperate lakes I would say that both are equally important. Where one gets 7 or 8 species occurring together in a weedbed they show a fair degree of dietary separation but then there are also replacements or 'counterparts' in different habitats. Werner: It has several times been noted that overlap in the diet of closely related fish species declines as resource levels decline seasonally. Do you find evidence for this in your studies? How do you interpret the overlap values in lieu of quantitative relations to resource levels? Keast: This does not occur in Lake Opinicon as far as I can see. Zaret & Rand (1971) found that food overlaps in fish in lowland streams of Panama were at a minimum during the dry season when food abundance was at its lowest. In these cold temperate lakes the major feature is not that all prey species are abundant at one season and rare at another. Rather, the different resources peak successively. In May many resources are abundant - there has been little fish predation during winter and numbers have built up. But even at the very end of the season, in October, numbers of certain organisms (amphipods, for example) are still high. If there is any time when resource diversity is reduced it is July after most aquatic insects have emerged and when the new insect generations are still too small to be of much use to fish. There is more, not less, diet overlap then, at least between age classes within species. We recently completed a set of diet overlap calculations for the various Lake Opinicon fish species, allowing three size categories for each of the longer-lived ones. I have not really had time to analyze this or, especially, go into the underlying reasons for fluctuating diet overlap levels between species. One point may be made here. With the obvious exception of the bluegill-pumpkinseed situation I really doubt if shared diets between species are ever as marked at these latitudes as they are in the tropics, allowing the habitat separations discussed. I have now looked at the stomach contents of a couple of dozen Amazon fish species taken at Leticia, Colombia in February, half-way through the period of river rise and found that many of the characins were feeding on the same things.

predator populations in these lakes growing so large that they imperil the numerical status of prey. Now what one might find in a recent man-made lake where new species are entering and displacing the earlier arrivals - and all is turmoil - is quite different. Again I can conceive of the fish populations of a shallow lake that never provided more than marginal conditions, being wiped out in one year by oxygen deficiency due to thick winter snow cover. But in an ecologically diversified lake and with a fauna that has been millions of years evolving would one expect, in the short term, to observe examples of space or disaster control of numbers? These are my thoughts, anyway. I confess I am also saying that I don't know what controls the numbers of bluegills and perch in Lake Opinicon from year to year.

Lobel: Do you think the fact that in many bluegill populations the individuals tend to become stunted in size relates at all to the fact that they might be adapted to a superabundant food resource? Keast: No, I thlnk it is because the age classes (which in the bluegill have much the same diet) cannot separate out spatially. I have never worked with a stunted population of fish. However, I recently pulled together food data on yellow perch (Keast 1977b), a species that is also very prone to produce stunted populations. Data on perch from Lake Opinicon and other water bodies in North America showed that (a) there was a general trend for perch to feed on zooplankton up to a length of about 50-60 mm (end of year class 0), consume larger prey such as Ephemeroptera and Odonata nymphs up to a length of about 150- 160 mm, then switch to a diet composed largely of fish; and (b) virtually all populations had normal growth up to the end of year class 0; after that growth rates varied greatly from lake to lake. I wonder if this was not related to food since perch are pretty adaptable in terms of temperature. I then looked through the literature on stunting. A most interesting study is that of Eschmeyer (1937) on the stunted perch populations of the Pigeon River pothole lakes of Michigan, which only reached an average length of 90-92 mm in year class 11, and then died. There was conclusive evidence of starvation in many individuals, such as a disproportionately large head to body ratio, emaciated appearance of the body, and poorly developed gonads. Very small organisms predominated in the stomachs. If 'stage 2' prey, the large-bodied Odonata and Ephemeroptera nymphs, are in short supply all age classes of perch would be forced to compete for small

bodied prey. Apart from everything else this would be highly inefficient to the larger fish in terms of energy outlay. Odonate nymphs have a long generation time and must be very vulnerable to elimination by predators, especially in a homogeneous waterbody with few places to hide. If this line of reasoning is correct then stunting in bluegill would occur as a result of overpopulation in a water body that is so homogeneous that the age classes cannot separate spatially. In this case as noted, there is not much age class diet separation. There is a lot of evidence, incidentally, that stunted fish are usually genetically normal. Remove them to another water body, or reduce the population size by the introduction of a predator, and they grow normally.

References cited Abele, L. G. 1974. Species diversity of decapod crustaceans in marine habitats. Ecology 55: 156-161. Alexander, R. McN. 1967. Functional design in fishes. Hutchinson, London. 160 pp. Amundrud, J. R., D. J. Faber & A. Keast. 1974. Seasonal succession of free-swimming perciform larvae in Lake Opinicon, Ontario. J. Fish. Res. Board Can. 31: 1661-1665. Andrusak, H. & T. G. Northcote. 1971. Segregation between adult cutthroat trout (Salmo clarki) and dolly varden (Salvelinus malma) in small coastal British Columbia lakes. J. Fish. Res. Board Can. 28: 1259- 1268. Asplund, K. K. 1974. Body size and habitat utilization in whitetail lizards (Cnemidophorus). Copeia 1974: 695-703. Atema, J. 1971. Structures and functions of the sense of taste in the catfish Ictalurus natalis. Brain Behav. Evol. 4: 273-294. Brooks. J. L. & S. I. Dodson. 1965. Predation.,bodv. size.,and composition of plankton. Science 150: 28-35. Brown, K. E. 1977. Feeding and growth performance of yellow perch (Perca flavescens) in relation to different backgrounds of resources and potential competitor species. M.S. Thesis, Queen's University, Kingston, Ontario. 155 pp. Christie, W. J. 1973. A review of the changes in the fish species composition of Lake Ontario. Great Lakes Fish. Comm. Tech. Rep. No. 13, Ann Arbor, Michigan. 65 pp. Cody, M. L. 1968. On the methods of resource division in grassland bird communities. Am. Nat. 102: 107-147. Cody, M. L 1974. Competition and the structure of bird communities. Princeton University Press, Princeton, New Jersey. 318 pp. Colwell, R. K. & D. J. Futuyma. 1971. On the measurement of niche breadth and overlap. Ecology 52: 567-576. Connell, J. H. 1975. Some mechanisms producing structure in natural communities. pp. 460-591. In: M. L. Cody and J. M. Diamond (ed.), Ecology and Evolution of Communities. Harvard Univ. Press, Cambridge, Mass. Connell, J. H. & E. Orias. 1964. The ecological regulation of species diversity. Am. Nat. 98: 399-414. Emery, A. R. 1973. Preliminary comparisons of day and

night habits of freshwater fish in Ontario. J. Fish. Res. Board Can 30: 761-774. Eschmeyer, R. W. 1937. Some characteristics of a population of stunted perch. Papers Mich. Acad. Sci. Arts & Lett. 22: 613-628. Fraser, D. F. 1976. Coexistence of salamanders in the genus Plethodon: a variation of the Santa Rosalia theme. Ecology 57: 238-251. Frver. G. & T. D. Iles. 1972. The cichlid fishes of the Great iakks of Africa; their biology and evolution. Oliver & Boyd, Edinburgh, 641 pp. Gerking, S. D. 1958. The restricted movement of fish populations. Biol. Rev. 34: 221-242. Harker, J. M. 1976. Prey selection relative to availability and size: a comparison of the feeding of four inshore fish assemblages in Lake Opinicon, Ontario. M.S. thesis, Queen's University, Kingston, Ontario. 152 pp. Harris, J. E. 1938. The role of the fins in the equilibrium of the swimming fish. 11. The role of the pelvic fins. J. Exptl. Biol. 15: 32-47. Hartley, P. H. T. 1947. Food and feeding relationships in a community of fresh-water fishes. J. Anim. Ecol. 17: 1-14. Heckrotte, C. 1967. Relationships of body temperature, sue, and crawling speed of the common garter snake, Thamnophis s. sirtales. Copeia 1967: 759-763. Hough, J. L. 1958. Geology of the Great Lakes. Univ. Illinois Press, Urbana, Ill. 313 pp. Hubbs, C. L. 1921. An ecological study of the life history of the freshwater atherine fish Labidesthes sicculus. Ecology 2: 262275. Hubbs, C. L. & K. F. Lagler. 1958. Fishes of the Great Lakes region. Univ. Mich. Press, Ann Arbor. 21 3 pp. Hutchinson, G. E. 1958. Concluding remarks. Cold Spring Harbor Symp. Quant. Biol. 22: 415-417. Hutchinson, G. E. 1959. Homage to Santa Rosalia or why are there so many kinds of animals. Am. Nat. 43: 145-158. Karr, J. R. & F. C. James. 1975. Ecomorphological configurations and convergent evolution. pp.258-291. In: M. L. Cody & J. M. Diamond (ed.), Ecology and Evolution of Communities. Harvard Univ. Press., Cambridge, Mass. Keast, A. 1965. Resource subdivision amongst cohabiting fish species in a bay, Lake Opinicon, Ontario. Proc. 8th Conf. Gt. Lakes Res., Univ. Mich.: 106- 132. Keast, A. 1966. Trophic interrelationships in the fish fauna of a small stream. Proc. 9th Conf. Gt. Lakes Res., Univ. Mich.: 51-79. Keast, A. 1968a. Feeding of some Great Lakes fishes at low temperature. J. Fish Res. Board Can. 25: 1199-1218. Keast, A. 1968b. Feeding biology of the black crappie, Pomoxis nigromaculatus. J. Fish. Res. Board Can. 25: 285-297. Keast, A. 1970. Food specializations and bioenergetic interrelations in the fish fauna of some small Ontario waterways. pp. 377-411. In: J. H. Steele (ed.), Marine Food Chains. Oliver and Boyd, Edinburgh. Keast, A. 1977a Mechanisms expanding niche width and minimizing intraspecific competition in the rock bass and bluegill sunfish (Centrarchidae), Evolutionary Biology 10: 333-395. Keast, A. 1977b. Feeding and food overlaps between the year classes relative to the resource base, in the yellow perch, Perca flevescens. Env. Biol. Fish. 2: 55-70. Keast, A. 1977c. Feeding interrelations between age groups

of pumpkinseed sunfish (Lepomis gibbous) and comparisons with the bluegill sunfish (L. macrochirus). J. Fish. Res. Board Can. (in press). Keast, A. 1977d. The beginning of exogenous feeding, foods and feeding relationships of young fish in the first weeks of life, open water community, Lake Opinicon, Ontario. Env. Biol. Fish. (in press) Keast, A. & J. Harker. 1977a. Strip counts as a means of determining densities and habitat utilization patterns in lake fishes. Env. Biol. Fish. 1: 181-188. Keast, A. & J. Harker. 1977b. Fish and benthic invertebrate abundance relative to water depth in an Ontario lake. Env. Biol. Fish. (in press) Keast, A., J. Harker & D. Turnbull. 1977. Fish habitats and species associations, Lake Opinicon, Ontario. Env. Biol. Fish. (in press) Keast, A. & D. Webb. 1966. Mouth and body form relative to feeding ecology in the fish fauna of a small lake, Lake Opinicon, Ontario. J. Fish. Res. Board Can. 23: 1845-1874. Keast, A. & L. Welsh. 1968. Daily feeding periodicities, food uptake rates, and dietary changes with hour of day in some lake fishes. J. Fish. Res. Board. Can. 25: 1133-1 144. Keddy, P. A. 1976. Lakes as islands: the distributional ecology of two aquatic plants, Lemna minor L. and L. triscula L. Ecology 57: 353-359. Kohn, A. J. 1967. Environmental complexity and species diversity in the gastropod genus Conus on Indo-West Pacific reef platforms. Am. Nat. 101: 25 1-259. Larkin, P. A. 1956. Interspecific competition and population control in fresh-water fish. J. Fish. Res. Board. Can. 13: 327-342. Levins, R. 1968. Evolution in changing environments. Princeton Univ. Press, Princeton, New Jersey. 120 pp. Liem, K. F. & J. W. M. Osse. 1975. Biological versatility, evolution and food resource exploitation in African cichlid fishes. Amer. Zool. 15: 427-454. LoweMcConnell, R. H. 1975. Fish communities in tropical freshwaters; their distribution, ecology and evolution. Longman, London. 337 pp. McPhail, J. D. 1977. A possible function of the caudal spot in characid fishes. Can. J. Zool. 55: 1063-1066. McPhail, J. D. & C. C. Lindsey. 1970. Freshwater fishes of northwestern Canada and Alaska. Fish. Res. Board Can. Bull. 173: 381 pp. May, R. M. 1975. Some notes on estimating the competition matrix. Ecology 56: 737-741. Mendelson, J. 1975. Feeding relationships among species of Notropsis (Pisces: Cyprinidae) in a Wisconsin stream. Ecol. Monogr. 45: 199-230. Nilsson, N. A. 1963. Interactions between trout and char in Scandinavia Trans. Am. Fish. Soc. 92: 276-291. Orians, G. H. & H. Horn. 1969. Overlap in foods and foraging of four species of blackbirds in the potholes of central Washington. Ecology 50: 930-938. Page, L. M. & D. W. Schemske. 1977. The effect of interspecific competition on the distribution and sue of darters of the subgenus Catonotus. Unpubl. ms. Paine, R. T. 1966. Food web complexity and species diversity. Amer. Natur. 100: 65-75. Pianka, E. R. 1969. Sympatry of desert lizards (Ctenotus) in Western Australia Ecology 50: 1012-1030. Pianka, E. R. 1975. Niche relations of desert lizards.

pp. 292-314. In: M. L. Cody & J. M. Diamond (ed.), Ecology and Evolution of Communities. Harvard Univ. Press, Cambrdige, Mass. Reifel, C. W. 1973. A macroscopical, histological and carbohydrate histochemical study of ten teleostean alimentary canals. M.S. thesis, Queen's University, Kingston, Ontario. 189 pp. Roth, R. R. 1976. Spatial heterogeneity and bird species diversity. Ecology 57: 773-782. Scott, W. B. & E. J. Crossman. 1973. Freshwater fishes of Canada Fish. Res. Board Can. Bull. 184. 920 pp. Schoener, T. W. 1965. The evolution of bird size differences among congeneric species of birds. Evolution 19: 189-213. Sheppard, J. D. 1965. The food relationships of five cohabiting Cyprinidae in Kearney Lake, Ontario. B.S. Honours Thesis, Queen's University, Kingston, Ont. 88 pp. Stein, R. A. & J. H. Magnuson. 1976. Behavioral response of crayfish to a fish predator. Ecology 57: 75 1-761. Terasme, J. & E. Mirynech. 1964. Postglacial chronology and the origin of deep lake basins in Prince Edward County, Ontario. Proc. 11th Conf. Gt. Lakes Res., Univ. Mich.: 161-169. Werner, E. E. 1977. Species packing and niche complementarity in three sunfishes. Am. Nat. 111: 554-578. Werner, E. E. & D. J. Hall. 1976. Niche shifts in sunfishes: experimental evidence and significance. Science 191: 404-406. Werner, E. E., D. T. Hall, D. R. Laughlin, D. J. Wagner, L. A. Wilsmann & F. C. Funk. 1977. Habitat partitioning in a freshwater fish community. J. Fish. Res. Board Can. 34: 360-370. Werner, R. G. 1968. Ecology of liinetic bluegill (Lepomis macrochirus) fry in Crane Lake, Indiana. Am. Midl. Nat. 81: 164-181. Worthington, E. B. 1954. Speciation of fishes in African lakes. Nature, Lond. 173: 1064-1067. Zaret, T. M. 1977. Inhibition of cannibalism in Cichla ocellaris and hypothesis of predator mimicry among South American fishes. Evolution 3 1: 421 -437. Zaret, T. M. & A. S. Rand. 1971. Competition in tropical stream fishes: support for the competitive exclusion principle. Ecology 52: 336-342.