K. O. WinemillerTexas A&M Univtrsity

I. IntroductionII. Evolution of Fishes

III. Physiological Ecology of FishesIV. Life-History VariationV. Population RegulationVI. Ecological Interactions

VII. Community Patterns and Biodiversity

of the world's major marine fisheries has leveledoff over the past two decades.

Fish ecology is the study of fishes and their inter-actions with the physical and living componentsof their environments. Because fishes are the oldestand most species-rich lineage of living vertebrates,a particularly diverse variety of adaptations andecological relationships are observed among differ-ent species and even among different populationswithin some species. Fishes are found in a greatdiversity of habitats, including temporary ponds inarid environments, subterranean waters, mountainstreams, lakes, rivers, estuaries, coral reefs, pelagicocean waters, and oceanic abysses. On a globalbasis, the greatest number of living fishes occupymarine environments, but the highest speciesdensi-ties (number of species per unit area) tend to occurin fresh waters. These differences in species densi-ties may be related to greater habitat complexityand opportunities for geographical isolation of pop-ulations in freshwater ecosystems. Aquatic and es-tuarine habitats are very sensitive to pollution andhuman-induced landscape alterations, and as a re-sult the abundance and diversity of fishes have de-clined in many regions of the world. Despite in-creased fishing effort and advances in thetechnology of fishing, the collective annual catch

I. INTRODUCTION

Water covers over 70% of the earth's surface, soit is perhaps not surprising that fishes are the mostabundant and speciose group of living vertebrates(estimates of the number of living species rangebetween 20,000 and 30,000). Fishes occupy aquatichabitats ranging in elevation from mountain lakesand streams, approximately 5 km above sea level,to deep ocean trenches 11 km below sea level.Fishes are also the oldest vertebrate lineage; the firstagnathan fishes originated some 500 million yearsago. The long evolutionary history of fishes, cou-pled with the large quantity and diversity of aquaticand marine habitats, has resulted in tremendousvariation in the ecological characteristics of modemspecies of fishes. This article reviews the relation-ship between fishes and the living and nonlivingcomponents of their environments. Because it isimpossible to survey the full range of ecologicalcharacteristics of such a large and diverse group of

Copyrighl 0 \995 by Aademic Press, Inc.All righcs of reprodUCtion in any form reserved.Encyclopedia of Environmental Biology

Volume 249

50 fBHECOlOGY

organisms, select species and habitats will be usedto illustrate major similarities and differencesamong taxa and ecological settings.

Fi~h ecology is greatly influenced by the physicalproperties of water: density, heat capacity, viscos-ity, and miscibility as a solvent. Because water ismore than 750 times denser than air, fishes neednot invest great amounts of matter and energy intothe development and maintenance of a massiveskeleton to support the body. Most fishes possessa gas-filled swim bladder that allows them to main-tain neutral buoyancy over a range of water depthswith low energetic cost. A number of benthic fishesthat lack functional swim bladders [e.g., sculpins(Cottidae), many darters (Percidae) , and gobies(Gobiidae)] use fin movements and body undula-tions to rise above the substrate. These forays intothe water column require large energy expendi-tures, and as a result they tend to be of short dura-tion, _as when feeding or changing locations on thesubstrate. The high density of water also causessound to be transmitted much more efficiently thanthrough air. Sound attenuates slowly in water, sothat sonic signals can be transmitted over very longdistances. Fishes perceive sound with the innerear and the acousto-lateralis system. The acousto-lateralis system consists of pressure receptorshoused in tubes embedded in scales of the head andbody ~flanks. In minnows and suckers (Cyprini-formeS), tetras (Characiformes), and catfishes (Si-luriformes), a chain of small bones, called the We-berian ossicIes, tr~nsmits sound pressure from theswimJbladder to the inner ear. Sound pressure firstreceived by the body wall is transmitted to theswim bladder, which functions as an ear drum.Sounds are produced for intraspecific communica-tion by many minnows and catfishes, plus a varietyof fishes from families lacking Weberian ossicles.

Water achieves its maximum density at 4°C andcan stratify over relatively shallow depth gradients.Nearly all fishes are ectothermic and can exploittemperature gradients to locate their thermal op-tima or to increase their efficiency of energy utiliza-tion. By reducing vertical or horizontal mixing ofwater, thermal and density gradients also affectnutrient dynamics and ecosystem productivity. Incold climates, lakes may experience vertical mixing

(tUrnover) when they become isothermal duringthe spring and fall. This vertical mixing can releasenutrients locked in deeper waters into the lightpenetration layer (epilimnion) , where photosyn-thesis takes puce. Blooms of phytoplankton pro-duction often follow these seasonal lake turnovers.

Light attenuation occurs much more rapidly inwater than in air. Long wavelengths are the first tobe absorbed (within the first 25 m), so that objectsviewed at great depths appear blue due to re-flectance of shorter light wavelengths. Fishes areassumed to have color vision, especially those diur-nal species that have flashy species-specific or sex-specific color patterns. The retinal cone receptorshave been investigated in several fish species, suchas the walleye (Stizost~dion vitr~um) and goldfish(Carassius auratus). Many fishes have countershad-ing, darker pigmentation in the dorsal region fad-ing to a light-colored ventral surface. With illumi-nation from above, countershading results inuniform reflectance that masks the normal shadingcues that reveal object depth. Countershading is aparticularly important adaptation for concealmentin pelagic fishes.

Water has a high specific heat, meaning that itmust receive or lose more energy than air or rockto change the same number of degrees. Therefore,aquatic and marine habitats are more thermally buf-fered against sudden climatic changes than adjacentterrestrial environments. Because of its higher den-sity, water also conducts heat more rapidly thanair. Fishes are ectotherms, and the time lag betweena change in ambient temperature and subsequentchange in body temperature is very short.

Because water is more viscous than air, fishescan propel themselves by swimming. Drag influ-ences the energetics of alternative methods of loco-motion, with a fusiform body producing less dragthan stout or compressed body shapes. Water'shigh viscosity enhances the efficiency of suctionfeeding and, to a lesser extent, hinders grasping.Water's high density and viscosity facilitate passivedispersal of buoyant gametes, larvae, and evensome adult fishes. The high viscosity of water alsoaffects the architecture of aquatic and marine land.,scapes via sediment erosion, transport, and sub-strate scourinR.

51f84ECOlOGY

Water has been called the universal solvent, be-cause it dissolves a great variety of chemical sub-stances, including gases, salts, acids, bases, nutri-ents (nitrates, phosphates), and biological wastecompounds (carbon dioxide, ammonia). Becauseof the solvency of nutrients in water, the dynamicsof nutrient cycling and primary production tendto be faster in aquatic ecosystems compared withterrestrial systems. Partial pressures and diffusionrates of oxygen and carbon dioxide are lower forwater than for air. Even so, high CO2 concentra-tions can cause severe physiological stress or deathin many fishes. Acute or chronic hypoxia (lowconcentrations of dissolved oxygen) can pose a seri-ous problem for fishes in swamp, lake, and estua-rine habitats. Fishes from these habitats haveevolved a wide variety of respiratory adaptations,including aerial respiration with lungs (lungfishes),swim bladder (loricariid and clariid catfishes), gut(callichthyid catfishes), skin (some blennies and an-guillid eels), gill chamber (synbranchid eels),and supra branchial chambers (snakeheads andanabantids). During periods of acute hypoxia,many freshwater fishes rise to the surface and skimthe oxygen-rich surface film (aquatic surface respi-ration). The electrical conductivity of water per-mits several groups of fishes (African mormyri-forms, neotropical gymnotiforms) to utilizeelectrogeneration and reception for navigation,prey detection, and communication. At least threegroups of fishes (torpedinid rays, electrophorideels, and malapterurid catfish) have evolved theability to produce electrical shocks as mechanismsfor subduing prey and for defense.

species of hagfishes and lampreys of the ~lassesPteraspidomorphi and Cephalaspidomorp~i. Theclass Chondrichthyes contains about 800 liv~ng spe-cies of sharp, skates, rays, and chimaera$. Thevast majority of Chondrichthyes are restricted tomarine habitats, although some species, such as bullsharks (Carcharinus ltUcus) and sawfishes (Pristis),regularly enter freshwater habitats where they mayreside for extended periods. All the tetrapod verte-brates (amphibians, reptiles, birds, mammals) arederived from a primitive lineage within the Os-teichthyes. If all the known undescribed species andestimated undiscovered species were tallied, the to-tal number of bony fishes might be well ov~r2S,<XX>. Of the many dozens of new fish speciesdescribed each year, the vast majority come fromtropical latitudes, especially from freshwater hab-itats.

Bony fishes inhabit an incredibly wide range ofaquatic habitats, including marginal aquatic habi-tats like thermal springs (North American kill i-fishes of the genus Cyprinodotl, African cichlids ofthe genus Sarotherodon), underground waters [cave-fishes (Amblyopsidae), North American catfishgenera Satan and Trogloglanis (Ictaluridae»), torren-tial mountain streams [Asian loaches (Cobitidae),South American hillstream catfishes (Astroblepi-dae)], ephemeral savanna pools [African andSouth American killifishes (Rivuliidae), lungfishes(Ceratodontidae, Lepidosirenidae»), and tidal mud-fiats [mudskippers (Periophthalmidae»). Except fora few special cases, the generation of new fish spe-cies (speciation) is believed to have occurred by wayof geographical isolation and genetic divergenceof populations (allopatric speciation). Freshwaterhabitats provide numerous opportunities for geo-graphical isolation of populations, and on a globalbasis the density of fish species in fresh water ismuch greater than in marine habitats. About 40%of all fishes live in fresh water, which comprisesless than 0.01% of the earth's total volume of sur-face water. Populations can be split into isolatedsubunits when river drainages are divided by geo-logical and climatic changes, river captures (anasto-moses), confinement in separate lake basins, or for-mation of habitats hostile to dispersal betweenseparate river drainage basins. Species flocks occur

II. EVOLUTION OF FISHES

A. Phylogenetic Diversity

Fishes belong to the phylum Chordata and are de-rived from a common ancestor with the protochor-dates (acorn worms, sea squirts). The three groupsof living fishes are the Agnatha (jawless cartilagi-nous fishes), the Chondrichthyes (jawed cartilagi-nous fishes), and Osteichthyes (bony fishes). Agna-tha has by far the fewest living taxa: about 50

52~-f5fECOlOGY

in a number of lake basins (cyprinodontids in LakeTiticaca, Peru; atherinids in central Mexico; salmo-nids in the North American Great Lakes; cichlidsin the African Rift lakes). Evolutionary biologistsdebate the potential mechanisms of sympatric spe-ciation for generating these species flocks, espe-cially.the African cichlids. Most evidence seems tosupport a hypothesis of allopatric speciation withinisolated or partially isolated lake subbasins, orwithin preferred habitats separated by shoreline re-gions-<onuining hostile habitat. Geographical iso-lation of marine populations occurs by a variety ofgeographical and climatic changes, including for-mation of isthmuses and islands, tectonic move-ments of landmasses, zones offreshwater intrusion,and a host of oceanic currents that influence patternsof dispersal.

contain a number of specialized invertebrate- andfish-eating fishes, plus fishes that specialize on al-gae, detritus, wood, macrophytes, seeds, fruit, fishfins, fish scales, and the external mucus slime layerof other fishes.

Fishes have proven to be a particularly good groupfor the study of the relationship between functionalmorphology and ecology (ecomorphology). Bodyform and the sizes, shapes, and positions of fins de-termine swimming performance in fishes (Fig. 1).The bodies of benthic fishes tend to be cylindrical orstocky with a flattened belly region, or their bodiesare entirely depressed in the dorsoventral plane, asin skates, rays, and many catfishes. Soles, flounders,and other flatfishes are compressed in the lateral

SOtAh American CharaciOlml (CharaOdae. GastGo~)

B. Ecological Diversity Deep-boci8d

h ~ ~

/J~

Regions at higher latitudes generally contain fewerfish species than tropical areas. and this patternholds for both marine and freshwater ecosystems.Likewise. fish diversity tends to be lower at higheraltitudes compared with lowland areas in the samebiogeographic region. A variety of historical. cli-matic. and ecological factors interact to producethese gradients of fish species diversity. The basicmechanisms influencing fish biodiversity probablydo not differ significantly from models proposedfor terrestrial vertebrates. Compared with the trop-ics. regions at high latitudes and altitudes sufferedmore species extinctions during glacial epochs. ex-perience more frequent and more severe climaticchanges over evolutionary time scales. and gener-ally experience greater habitat changes with season.Consequently. coevolution. ecological specializa-tion. and adaptive radiations are much more appar-ent in tropical freshwater fish communities thanin arctic or temperate communities. Research hasdocumented more morphological variation andecological strategies in tropical fish assemblages rel-ative to fish assemblages in comparable physicalhabitats at higher latitudes. For example. arctic andtemperate freshwater fish assemblages are domi-nated by invertebrate-eating and fish-eating fishes.whereas tropical fishes in the same kinds ofhabitats

SIre8m-in8d . - - " ~

Surface-ori8nled

-~~

I~~~~r- "-~~~~

C-*8IAmet'aII~

~iJ:J

~

~

)0 >0,.

FIGURE I Ecological and morphological diversification o(body (Orin and swimming performance in South American characi-(orm fishes (above) and head (Orin and (eeding in Central Americancichlid fishes (below). (With permission (rom the National Ge0-graphic Society and Kluwer Academic Publishers.)

- IP8Ci88 (~)

~

53- --F94ECO1..OGY

ers), like sculpins (Cottidae) and scorpion -fishes(Scorpaenidae), use suction feeding rather thangrasping. Others, like the tarpon (Megalops), usean engulfing- mode of prey capture in which thepredator passes over its prey with jaws and gillcovers (opercula) flared. Some large-mouthedplankton feeders, such as paddlefish (Polyodonspathula) and whale sharks (Rhincodon typus) , usethis engulfing feeding mode, but in this case smallfood items are strained from the water passingthrough the gill openings using comblike gill rak-ers. In modem bony fishes like the North Americancrappie (Pomoxis), the Central American cichlid(P~tmia), the tarpon snook (Cmtropomus p~ctinatus),and the African thin-face cichlid (S~"anochromis an-gustic~ps), suction feeding is enhanced by highlyprotrusiblejaws. During jaw protrusion, the pre-maxillary and maxillary bones of the upper jawswing forward as the mouth opens and, with themandible, form a tubular gape. Fishes that primar-ily use biting or grasping often have less protrusiblejaws, and some, like piranhas (Pygoc~ntrus) and bar-racudas (Sphyratna), lack significant protrusibility

altogether.Depending on the ecological setting, piscivores

that feed on a wide variety of prey often can adopteither grasping, sucking, or biting prey capturemodes. North American bass (Micropttrus) and ma-rine groupers (Epin~ph~/us) feed on a wide varietyof invertebrate and fish prey using all three feedingmodes. Many small-mouthed invertebrate-feedingfiShes use a picking (biting) mode to glean insectsand small crustaceans from the substrate or vegeta-tion. The same species may also use suction feedingto capture small food items from the water columnor to glean surfaces. The jaw elements of sea horses,pipefishes, and trumpet fishes are fused to forma long tubular snout used for suction feeding. Inhabitats with soft bottom sediments, some fish for-age by grabbing a mouthful of sediment and siftingimmature aquatic insects from the sediments withtheir gill rakers or pharyngeal teeth (the latter lo-cated on the throat region). These digger/sifters,like .'earth-eating " cichlids (Gtophagus, Satanop-

uta) and marine mojarras (Diapttrus, Eucinostomus),usually have long snouts or highly protrusible jawsto reach deep into soft sediments. Cvoriniform

plane and rest on their sides on the substrate. Mid-water fishes that are compressed in the lateral planecan maneuver well in three dimensions. A laterallycompressed body and broad fins permit greater sta-bility in water, which is important for fishes that suc-tion feed on small food items in the water column orglean prey from surfaces. Nonh American sunfishes(i.qomis), many tropical cichlids, and coral reefdamselfishes and surgeonfishes illustrate this eco-morphological strategy. These deep-bodied fishesachieve slow but stable locomotion by sculling theirbroad pectoral fins. They can also perform rapidswimming by folding their broad fins close tothe body (reducing drag) and using undulatory pro-pulsion by laterally flexing the body and caudal(tail) fin.

Tunas, mackerels, salmon, and other fishes thatrely on rapid and sustained swimming generallypossess fusiform bodies with relatively narrow dor-sal and ventral fins. The more elongate, fusiformbody reduces drag in these strong swimmers. Dragis further reduced by a narrow caudal peduncle(region between the body and the caudal fin) anda caudal fin that is tall and forked to some degree(high aspect ratio). Pike, barracuda, and otherfishes that rely on stealth and a rapid swimmingburst to capture prey often have very elongate bod-ies and medial fins positioned posteriorly. Thesefishes can accelerate rapidly by bending the bodyin an s-shape, then shooting off like an uncoilingspring. These burst predators perform sustainedswimming by the normal undulating mechanismof lateral flexion of the body and tail. Eels haveextremely elongate bodies and reduced fins thatpermit swimming or burrowing by undulatory lo-comotion (lateral body flexion). The elongate eel-like morphology has evolved independendy in anumber of taxa, including anadromous eels, morayeels, and snipe eels (Anguilliforrnes) , neotropicalknife fishes (Gymnotiforrnes), paleotropical spinyeels (Mastacembelidae), and swamp eels (Synbran-

chidae).Much can be determined about a fish's feeding

behavior by examining its jaw morphology (Fig.1). Predatory fishes generally have large mouthsarmed with jaw teeth for grasping, piercing, orcutting the flesh of prey. Some piscivores (fish eat-

54~- -

fmof£Ca.OGY

the yon Bertalanffy curve to describe growth ratesin the following manner:

L, = L.(l-e-K('-'o»

fishes lack jaw teeth and manipulate prey primarilywith their pharyngeal teeth prior to ingestion.Fishes that feed on hard-bodied prey, like coralsand molluscs, usually have massive flattened pha-ryngeal teeth used for crushing.

Herbivorous and detritivorous fishes generallyhave relatively compact jaws for grasping and tear-ing plant material. Seed- and fruit-eating fishes usu-ally have multicuspid teeth for tearing and crush-ing. The relationship between powerful compactjaws;'\Jnulticuspid teeth, and a frugivorous diet iswell illustrated by a large number of neotropicalcharacids, including species of the genera Astyanax,Brycon, and Colossoma. Grazers have flat unicuspidor bicuspid teeth used for scraping algae from thesurfaces of rocks, wood, or vegetation. Great diver-sity in tooth morphology is often observed amongsympatric grazers, and different tooth morpholog-ies seem to select for different algae taxa in theirdiets. Herbivorous and detritivorous fishes usuallyhave long, coiled alimentary canals, and detriti-vores sometimes have crop and gizzard stomachchambers and pyloric ceca. The muscular gizzardcontains sand particles ingested along with detritusthat aid the mechanical breakdown of plant cellwalls. A long gut and pyloric ceca provide moresurface area for digestion of tough plant tissuesand absorption of nutrients. Some herbivorous anddetritivorous species, like the wood-eating SouthAmerican catfishes (Panaqut), have coevolved gutfaunas (microbial organisms) that aid the digestionof plant tissues.

where L, is length at age t, L,.. is the asymptoticlength, 10 is the hypothetical time when length iszero, and K is the rate of approach to L,... Somepiscivorous fishes show an accelerated growth rateduring the juvenile stage that coincides with anontogenetic shift from invertebrates to larger, moreenergy-rich fish prey. Because fishes are ecto-therms, each individual has an optimal temperaturefor food assimilation and growth. The biotic envi-ronment can also affect growth, with crowdingreducing growth via exploitation competition forlimited food or via interference competition forlimited space. The relationship between fish weight(W) and length (L) can be described by the equationW = aL., where a and b are constants. If growth

is isometric, with no change in shape with change insize, then b is approximately 3.0. The relationshipbetween fish length and fish size has been used asan index of condition of plumpness in fisheriesmanagement. Species-specific standards that de-scribe average or good condition have been derivedempirically by regression methods. Fish conditioncan be influenced by foraging rate, environmentalstress, reproductive state, and habitat character-istics.

Fishes can be aged by a variety of methods. allof which involve counting concentric bands that aredeposited on hard structures during their growthprocess. Spines. scales. and otoliths (mineral grainsin the inner ear) grow radially at different ratesdepending on seasonal ecological conditions, matU-ration, or reproduction. Minerals are deposited asdaily layers (circuli) that appear as rings in crosssection. During periods of rapid growth, the layersare thicker and the more opaque boundaries be-tween layers are spaced farther apart. During peri-ods of slow growth, as in winter or spawning sea-sons, the daily rings are tightly compacted andappear as bands. The bands on scales, spines, andotoliths correspond to annual age increments (an-nuli) in most temperate-zone fishes and many tropi-cal fishes from seasonal habitats. In young fishes,

III. PHYSIOLOGICAL ECOLOGYOF FISHES

Growth in fishes has been called indeterminate be-cause they exhibit some growth even after attain-ment of sexual maturation, even though the rateof growth usually declines with size and age. Asin any organism, a fish's growth potential is deter-mined genetically, and its realized growth is de-rived from the interaction between genes and envi-ronment (temperature, salinity, food quality andquantity). Fisheries ecologists frequently employ

55fISH ECOlOGY

dependent. Some reef-dwelling wrasses, such asthe cleaner fish Labroides dimidiatus, mature as fe-males and later change into males (protogyny). de-pending on tJteirstatus in the dominance hierarchyof a social group inhabiting a territory. When thedominant male is removed, the most dominantfemale rapidly begins to show hormonal and be-havioral changes followed by physiological andmorphological changes to male characteristics.

Several all-female species (gynogenetic species)have been identified within the live-bearing poecil-iid genera Poeciliopsis and Poecilia, and these appearto have orginated sympatrically via reproductiveisolation from genetic mechanisms following hy-bridization. Some of these all-female species mustmate with males of heterospecific species in orderfor the egg to develop into a genetic clone of itsmother. In other species, male genetic material isincorporated in offspring and then subsequentlylost during alternate generations, and no significantamount of the male's genetic material enters intothe gene pool of the all-female species (hybrido-

genesis).

IV. LIFE-HISTORY VARIATION

Fishes demonstrate more variation in life-historytraits than any other comparable taxonomic group(Table I). For example, clutch sizes range from1-2 in the longfm mako (Isurus paucus), thresher(Alopias), and sandtiger sharks (Eugomphodus taurus)to over 6x loB in the ocean sunfish (Mola mola).Even within species, individuals may vary consid-erably in their sizes and ages at maturity, growthrates, fecundities, and longevities. Fishes have beengrouped based on their spawning habitat andmethod of egg deposition, parental care, or brood-ing. At one end of the spectrum are nonguardingegg scatterers that usually have external fertilizationand group spawning. Pelagic spawners scatter theireggs into the water column. Eggs and early larvaeof pelagic egg scatters usually contain oil dropletsin the yolk that enhance buoyancy. In freshwaterhabitats, many nonguarding species scatter theireggs over vegetation, gravel, or stones. Some non-guarding fishes hide their broods. Trout and

otolith circuli often can be counted under a micro-scope to estimate their age in days.

Freshwater fishes are hypertonic to their aquaticenvironment, and marine fishes are hypotonic tomarine waters that typically range between 32 and35 ppt (parts per thousand) solute concentration.Because the epithelia of the skin, mouth, and gillsoffishes are permeable to ions and water molecules,freshwater fishes' ions tend to diffuse out and watertends to diffuse into their bodies. Likewise, ionspassively diffuse into the body and water diffusesout of the bodies of marine fishes. Freshwater fishesuse their kidneys to eliminate excess water in diluteurine and gain lost salts in their diets and by activeion transport in branchial cells of the gill mem-branes. Marine fIShes produce concentrated urineand eliminate as little as 3 ml of urine per kilogramof body weight per day. Up to 90% of nitrogenouswastes may be eliminated through the gills of ma-rine fishes. Cartilaginous fishes in marine habitatshave elevated levels of urea in their body fluids,which reduces the diffusion gradient of water.Most fish species live out their entire lives in habi-tats that experience little variation in salinity. Thesespecies are generally stressed by salinity changes ofonly a few ppt and are intolerant of changes greaterthan 15-20 ppt. Salmon (Oncorhynchus, Salmo) ,American shad (A losa) , striped bass (Morone) ,American eels (Anguilla), and other diadromousfishes have the ability to make rapid osmotic adjust-ments by changing urine volumes and the uptakeor secretion of ions against diffusion gradients bythe gills.

Sex in most fishes is genetically or chromosom-ally determined, but sexual development in somespecies is influenced by environmental conditions.Sex ratio in the Siamese fighting fish (Betta splen-dens) is strongly influenced by temperature. A num-ber of marine fishes and a few freshwater speciesexhibit sex switching with age and size or in re-sponse to their social environment. Some groupers(Serranidae) and wrasses (Labridae) mature asmales and become females when they are olderand larger. This protandrous strategy maximizeslifetime fitness because mating is not highly selec-tive, and a small male can produce large quantitiesof sperm. but female fecundity is strongly size-

56 FISH ECOlOGY

TABLE IInterspecific Variation in North American Fish Ute-History Traits

Age of

maturity(years)

Length of

maturity"(mm)

Clutchsize'

(N_>

Eggdiameter

(mm)Longevity

(years)

9-106248

1-23

1-21

.l-44

3-42215

22Js1370486663

124085

129625974

90997611338450S80

790

3030147

414

2492

19147983

14

141.59.

543.4.

18.1.

226.200.

1.427.70.

81.213.

400.

3.352.650.754.705.501.602.801.251.305.003.752.151.200.891.421.75

208

3-4<I

4-S2?4?

"1 c

.2 ,223

4-S1

1-22

4-S

2743740352100

14903103559104W102870210710289

19864375SO906

124571

5701520445570

60

40

9

2

5

5

8

II

S

3

7

9

14

'9

33

7

II

tj

-17

3

20

16

10

12

3.755.7452.561.000

466.7012, 100

13.61969.600

68496.000

52150

1.900.0001.313

195.00083.000

3.500.00039.200

1.050.0009.300.0004.010.325

567640.000500.000

4.190.0003.329.000

2.551.000.900.806.500.9516.01.122.001.701.050.831.050.930.931.020.950.821.781.140.953.500.980.81

4

S2

4-S

Freshwater Species

Po/yodoll spalhllia (Polyodontidae)iLpisosttllS OSStllS (Lepisosteidae)Dorosollla ctpediolllllll (Clupeidae)Sa/lllo clarki (Salmonidae)

SallfelillllS 1141114Ycush (Salmonidae)Ulllbra lillli (Umbridae)Esox lllcillS (Esocidae)

Notmrigomls crysoltllcas (Cyprinidae)Pimtphala promelas (Cyprinidae)IctiobllS bllballlS (Catostomidae)ICf4/lIrIIS pllllCtohlS (Ictaluridae)

Alllblyopsis spe/aea (Amblyopsidae)upolllis I114crochi",s (Centrarchidae)Pollloxis _laris (Centrarachidae)Etheostolll4 spectobi/e (percidae)Stizostedioll vitrtlllll (Percidae)

Marine Species

Acipmsu oxyrhYllmlls (Acipenseridae)Allgui//4 rostrato (Anguillidae)A/osa pst&llloharrnglls (Clupeidae)Allmoa mitchilli (Engraulidae)

OIICOrhY"ChIlS tshawytscha (Salmonidae)OSlllmlS lIIoniax (Osmeridae)

Arillsftlis (Ariidae)Mu/llccills prodllctlls (Gadidae)

Lycodopsis pacifica (Zoarcidae)GastuosttllS a",/eat&lS (Gasterosteidae)KatsllwolIlIS pe/alllis (Scombridae)Amlllodytes alllericalllls (Ammodytidae)Polllatollllls saltotrix (Pomatomidae)Haelllll/oll allroliMatllm (pomadasyidae)Scianops ocellatllS (Scianidae)lAgodoll rholllboides (Sparidae)Crntropristis striata (Serranidae)

LutjallllS C4mpechalllls (Lutjanidae)Morone saxatilis (percichthyidae)

Microgobills gll/oslls (Gobiidae)Sebastes callrilllls (Scorpaenidae)

Ophiodoll e/ollgatlls (Hexagrammidae)Para/imthys drntatllS (Bothidae)Pstlldopltllrollectes alllericallUs (Pleuronectidae)

. Average for a population near the center of a species range.

. Maximum dutch size reported for a population near the center of a species range.

salmon use their caudal fins to bury their eggs ingravel (create spawning redds), and European bit-terling (Rhodeus sericeus) deposits their eggs in thegills of unionid clams. Brood-guarding occurs ina variety of forms. Ovoviviparity (live-bearing

without maternal nutritional contribution duringgestation) and viviparity (live-bearing with mater-nal nutritional contribution during gestation) arecommon in a variety of fish taxa, including manysharks, skates, rays, guppies and other poeciliids

531422912420051489(XK)(XK)136880(XK)

70104(XK)320(XK)

57.

FBHECOlOGY

body size with delayed maturation, long life span,large clutches, small eggs, and few spawning boutsduring a short reproductive season. fishes withwell-developed parental care tend to have largereggs and often have longer reproductive seasonsand serial spawning. Three primary life-historystrategies define the end points of a continuumderived from comparisons of diverse ecological andtaxonomic groupings of fishes worldwide (Fig. 2).At one end point, ptriodic-strattgists have delayedmaturation, mature at intermediate or large bodysizes, produce large clutches of small eggs, tend tospawn in annual episodes, and tend to exhibit rapidgrowth during the first year of life. Opportunistic-strattgists mature early at small sizes, produce small-to medium-sized clutches of small eggs, have mul-tiple spawning bouts each year, and grow rapidlyas larvae. Equilibrium-strattgists may be any size, buttend to produce small- or medium-sized clutchesof relatively large eggs, and often have broodguarding or maternal provisioning of nutrients to

developing embryos.Fishes near the periodic region of the life-history

continuum probably reap two major benefits fromdelayed maturation and large adult body size: ca-pacity to produce large clutches and enhanced adultsurvival during periods of suboptimal environmen-tal conditions, like winter and periods of reducedfood availability. Fishes with large clutches fre-quently spawn in synchronous bursts that coincide

'ir.:reas81g EnvWon~tal Dis~-:---Decreas81g Pre6cub8ty ~ Spatiot~1Variabaty of Resowces and M«tality Agents

FIGURE 2 ScMmatic represcnution of the triangular contin-uum of primary life-history stnregin in fishes and mvironmmulfaCtors as~ted with selection gradimrs. (Based on K. O.Win~miller and K. A. Rose (1992). Canal/. J. Fish. Aqua/. Sri.49(10), 2196-2218.1

(Poeciliidae), surfperches (Embiotocidae), and thecoelacanth Latimma chatumnat. The coelacanthgives birth to 1-5 very large, advanced offspringthat gain nutrition during gestation within the ovi-duct by consuming unfertiliz~d trophic ~ggs andperhaps even smaller less-advanced siblings.

Male guarding of the nest and brood is perhapsthe most common form of parental care in fishes.Most cichlids have biparental care, but a few spe-cies, like the dwarf cichlids (Apistogramma), havematernal care. External-bearing and mouth and gillchamber brooding are other forms of parental carein fishes. Males of several South American aspredi-nid and loricariid catfishes brood their ~mbryos onthe surface of their bellies, and mal~ pipefishes andsea horses (Syngnathidae) carry their eggs and lar-vae in a belly brood pouch. Males of th~ Australiannursery fishes (Kurtus) carry th~ir developing em-bryos on a hook that protrudes from their fore-heads. The developing embryos hang on either sideof the head in small clusters suspended by a twistedcord made from egg membranes looped throughthe eye of the hook. Oral brooding occurs in anumber of freshwater and marine families, includ-ing the ariid catfishes (Ariidae), bony tongues (Os-teoglossidae), cichlids (Cichlidae), and cardinalfishes (Apogonidae) [Stt LIfE HISTORIES.)

Life-history strattgitS result from trade-offs amongattributes, such as clutch size and egg size, that haveeither direct or indirect effects on reproduction andfitness. For example, one way to achieve a largerclutch is to partition the biomass available for re-production into smaller eggs. This functional con-straint contributes to the negative corr~lation be-tween clutch size and egg size frequently observedamong fish species. Larg~r clutches can also beattained by delaying reproduction and growing toa larger body size. Comparing both within andacross species, larger fishes tend to produce moreeggs per batch than do small~r fishes. Y et, ~xcep-tions to this rule are common. Many fishes withparental care, like the mouth-brooding gaftopsailcatfish (BagnlS marinus) , have much smaller clutchesthan those of smaller species that scatter their eggs,like anchovies (Anchoa mitch;",) and dace (Rh;n;ch-

thys atraINtus).Comparative life-history studies of fishes have

repeatedly identified associations of large adult~

58-

FGH£COlOGY

larvae. In contrast, the reproductive success of ma-rine broadcast spawners depends on rates of larvalencounters with suitable zones or patches. Massiveclutches of ~mall pelagic eggs undoubtedly enhancedispersal capabilities of wide-ranging marine fishesduring the early life stages. In a stable population,losses due to settlement in hostile habitats (calledadvection) ultimately are balanced by the survivalbenefits derived from the passage of some fractionof larval cohorts into suitable regions or habitats.

The opponunistic life-history strategy is associ-atcd with rapid population turnover rates and ahigh intrinsic rate of population increase (r, an indexof the potential for exponential growth). By havingamong the smallest rather than largest clutches,opportunistic-type fishes differ markedly from thetraditional model of r-strategists. Yet because oftheir small size, the relativt reproductive effon ofopportunistic strategists is actually high, despitethe fact that absolutt clutch size and egg size aresmall. In these small species. serial spawning some-times resultS in an annual reproductive biomass(i.e.. annual fecundity) that exceeds the female'sbody mass. Small fishes with early maturation andfrequent spawning are well equipped to repopulatehabitatS following disturbances and to sustain theirnumbers when faccd with continuously high mor-tality during the adult stage. Given their high in-trinsic rates of increase. opportunistic-strategistsare relatively efficient colonizers of disturbed habi-tats. A relative opportunistic-strategy is observedin the bay anchovy (Engraulidae), silversides (Ath-erinidae), annual killifishes (Rivuliidae), marsh kil-Ii fishes (Cyprinodontidae), and mosquito fishesand other freshwater live-bearers (Poeciliidae,Goodiidae). These small fishes often maintaindense populations in marginal or constantly chang-ing habitats and persist in the face of high predationmonality during the adult stage. Extreme examplesof the opportunistic-strategy appear to be morecommon in tropical fresh waters than in the tem-perate zone. and in shallow marginal habitats thanin deeper freshwater and marine habitats.

An equilibrium life-history strategy in fishes cor-responds largely with the suite of traits associatedwith the traditional K-strategy of adaptation to lifein resource-limited or density-dependent environ-

either with migration into favorable habitats orwith favorable periods within the temporal cycleof the environment, like the spring or rainy season.Cod" cobia. mackerels. tunas. and other marinefishes with tiny pelagic eggs and larvae also haveamong the highest fecundities. These species copewith large-scale spatial variation in the marine pe-lagic environment by producing huge numbers oftiny offspring. at least some of which are boundto thrive once they encounter favorable strata orpatches. On average. larval survivorship is ex-tremely low among highly fecund fishes in themarine environment, and the average larval fishprobably dies during the first week of life. Fastlarval growth rates reflect successful exogenousfeeding by the lucky survivors that encounter areasof relatively high prey density.

At higher latitudes. large-scale and cyclic tempo-ral variation in environmental conditions is a majorfactor selecting for the timing of reproduction.Highly fecund fishes can exploit predictable pat-terns in time or space by releasing massive numbersof progeny in phase with periods in which environ-mental conditions are most favorable for larvalgrowth and survival. Selection favors physiologicalmechanisms that enhance a fish's ability to detectcues that predict the periodic cycle (photoperiod.ambient temperature, solute concentrations). Intropical marine pelagic environments. large-scalevariation in space may represent a periodic signalas strong as the seasonal variation at temperate lati-tudes.c Research in physical oceanography hasshown patchy distributions for a variety of physicalparameters (salinity. temperature). primary pro-duction. and zooplankton due to upwellings,gyres, convergence zones, and other currents.

Many periodic-type species are migratory.Anadromous American shad (Alosa sapidissima)show more repeat spawning and devote a greaterportion of energy to migration at higher latitudeswhere environments are more variable and less pre-dictable. Anadromous sticklebacks (Gasttrosteusacultatus) have more periodic-type life-history traits(larger clutches, larger size at maturity) comparedwith conspecific freshwater populations. Byadopt-ing anadromy, adult fishes can find favorable envi-ronments for the development and survival of their

59FGHECOtOGY

that brood guarding is not a viable tactic for largefishes in oligotrophic systems. Data from arcticchar (Salvelinus alpinus) indicated that the g"rowingseason at higJt latitudes probably constrains age atmaturity and the frequency of spawning. Migra-tion to special spawning habitats and burial offertil-ized eggs (brood hiding) by salmon, char, and troutare forms of parental investment that carry largeenergetic and survival costs in relation to futurereproductive effort. Senescence associated withsemel parity in Pacific salmon might have evolvedas a consequence of the survival cost of returningto the sea after energetically costly upstream runsto fluvial habitats that enhance larval survivorship.By comparison, freshwater whitefishes (Coregonus.Prosopium) exhibit a perennial periodic-strategythat involves large clutches, small eggs, and annualspawning bouts.

A number of intermediate-sized fishes haveseasonal spawning, moderate clutch sizes, andnest guarding (North American ictalurid catfishesand sunfishes, Lepomis). Rockfishes of the easternPacific (Scorpaenidae) have large clutches andsmall eggs and bear living young. All of thesefishes lie between periodic and equilibrium endpoints of a triangular gradient. Small fishes withrapid maturation, small clutches, large eggs rela-tive to body size, and a degree of parental care(minnows of the genus Pimephales. madtoms,darters. sticklebacks, pipefishes, and sculpins)lie between opportunistic- and equilibrium-strategists. Similarly. small fishes with seasonalspawning, moderately large clutches, small eggs,and only one or a few bouts of reproduction perseason lie between opportunistic and periodicextremes of the gradient. It is important to notethat fishes with divergent life-history strategiesfrequently coexist in the same habitats. Eachspecies' morphology and feeding niche determinethe nature of the resource variation and predationthat it experiences; and morphological constraints.including features involved in feeding within par-ticular microhabitats. restrict the evolution of life-history features. In addition, phylogenetic con-straints will result in varying degrees of adaptivedivergence or evolutionary convergence towarda given adaptive suite of life-history traits.

ments (delayed maturation, large body size,clutches containing few large offspring). Largeeggs and parental care yield larger or more develop-mentally advanced juveniles at the onset of inde-pendent life. Live-bearing sharks and rays that bearrelatively large, advanced offspring would lie nearthe equilibrium end point of the life-history spec-trum. Among bony fishes, marine ariid catfishes(egg diameters 16-20 mm, oral brooding of eggsand larvae) and amblyopsid cave fishes (branchialbrooding of small clutches of relatively largeeggs) illustrate relatively extreme forms of thisequilibrium-strategy. Cave fishes inhabit amongthe most stable and resource-limited of aquatic hab-itats. Parental care tactics (including long gestationin live-bearers) tend to be more highly developedand widespread in tropical fresh waters and amongcertain reef-dwelling marine fishes, such as seahorses, surf perches, and sharks.

Of course most fishes are associated not with aparticular end point strategy, but rather with inter-mediate strategies within the triangular gradient ofprimary life histories. Some of the largest periodic-type fishes, the sturgeons (Acipenseridae) and pad-die fish (Polyodontidae), have relatively large eggs,which reduces their theoretical maximum clutchsize. Salmon and trout possess even larger eggs(4-6.5 mm diameter) and smaller clutches thanfishes exhibiting the extreme periodic-strategy. Yetamong populations of coho salmon (Oncorhynchuskisutch), egg size declines and clutch size increaseswith increasing latitude. Selection seems to favorlocal optima in egg size with clutch size adjustmentsresulting from physiological constraints andecological performance. Relative to periodic-strategists with larger clutches and smaller eggs,salmon and trout apparently have evolved a moreequilibrium-strategy of fewer but larger offspringat the onset of independent life. In eutrophic ecosys-tems, a pulse of primary and secondary productionduring short summers probably favors a periodic-type strategy at high latitudes. Studies of severalfish species with large ranges along the easternNorth American coast show that juvenile growthrates are actually faster for fishes at higher latitudeswhere the growing season is shorter. Yet, thegrowing season at high latitudes may be so short

60 FISH ECOlOGY

V. POPULATION REGULATION resulting in r ~ In ~ 'x m.)/ T. The relative rate ofpopulation increase is directly dependent on fecun-dity, timing of reproduction, and survivorship dur-ing both imm~ture and adult stages. Averaged overmany generations, the three parameters (Ix, mx. T)must balance, or populations would decline to ex-tinction or would grow to precariously high densi-ties and eventually crash.

The three end point life-history strategies offishes referred to previously are associated withtrade-offs among age at maturation (a positivelycorrelated with T), fecundity, and survivorship.The periodic-strategy corresponds to high valueson fecundity and age at maturity axes (the latter isa correlate of population turnover rate) and a lowvalue on the juvenile survivorship axis. The oppor-tunistic-strategy of high population turnover ratevia rapid maturation corresponds to low values onall three axes. The equilibrium-strategy corre-sponds to low values on the fecundity axis andhigh values on the age of maturity and juvenilesurvivorship axes.

Clearly, the periodic-strategy maximizes age-specific fecundity (clutch size) at the expense ofoptimizing turnover time (turnover times arelengthened by delayed maturation) and juvenilesurvivorship (maximum fecundities are attained byproducing smaller eggs and larvae). Large bodysize enhances adult survivorship during suboptimalconditions and permits storage of energy and bio-mass for future reproduction. Iteroparity permitsa fish to sample its environment several times until,sooner or later, reproduction coincides with favor-able conditions and strong recruitment occurs. Vir-tually all ecosystems exhibit either spatial or tempo-ral variation that is to some degree predictable. Thismay be especially true in freshwater and marineenvironments, because batch spawning of largeclutches is predominant among bony fishes world-wide. Spawning by these periodic-type fishes isusually annual and synchronous, so that genera-tions are often recognized as discrete annual co-horts. Yet, correlations between parental stockdensities and densities of young-of-the-year re-cruits have been shown to be negligible in thesefishes. Recruitment often depends on climatic con-ditions that influence water currents. larval reten-

The study of population regulation in fishes hasbeen .. dominated by the search for density-dependent recruitment in commercial stocks. Sev-eral simple models describe density-dependent re-lationships between stock abundance (expressed aseither spawning adults or egg cohorts) and theabundance of recruits (expressed as age 1 fishes. orcohorts when they enter the fishery or becomevulnerable to sampling gear). These include thewell-known Ricker and Beverton-Holt stock! re-cruitment models. By and large. data from largemarine and lake fisheries conform very poorly todensity-dependent. stock-recruit models. This hasspurred greater examination of potential density-independent factors that influence recruitment.such as climatic fluctuations and seasonal changesin marine Currents. Evidence for density-dependentpopulation dynamics has been obuined from a va-riety of sources. including studies of local popula-tion dynamics. niche relationships. and predator-or resource-regulated community dynamics. [St't'POPULATION RECULAllON.)

In terms of life-history strategies. fitness can beestimated by either V.. the reproductive value ofan individual or age-class. or by '. the intrinsic rateof natural increase of a population or genotype.Each of these fitness measures can be expressed asa function of three essential components: survivor-ship. fecundity. and the onset and duration of re-productive life. In the case of reproductive value.

wV x = mx + L (/",.,)/1.

Isx+1

where for a stable population mz is age-specificfecundity, Iz is age-specific survivorship, and Cd isthe last age-class of active reproduction. When xis equal to ct, the age of first reproduction, repro-ductive value is equivalent to the lifetime expecta-tion of offspring, and contains survivorship, fecun-dity, and timing components. The intrinsic rate ofpopulation increase can be approximated as r ~

In Rot T, where Ro is the net replacement rate, Tis the mean generation time, and Ro = I Ix mx,

61f54ECOLOGY

tion zones, productivity of prey patches, and a hostof other environmental factors that determine earlygrowth and survival. Accurate predictions of re-cruitment by periodic-type species in large marineecosystems require understanding of physicaloceanography plus the ability to forecast weatherconditions. Because weather cannot be predictedover long time intervals, fisheries projections oftenrely on short-term estimates of juvenile cohortstrength several weeks or months followingspawning, rather than long-term estimates basedon parental stocks.

Many periodic-type fishes spread their reproduc-tive effort over many years (or over large areas),so that high larval/juvenile survivorship during oneyear (or in one area) compensates for the many badyears (or areas). For example, anadromous femalestriped bass (Morone saxatilis) live up to 17 yearson average and produce an average clutch of4 x 10' eggs every year or two. This requires sur-vivorship of roughly 3 X 10-8 during the egg tomaturation interval to maintain a stable population.Most years probably result in a larval survivorshipapproaching zero for most females. For an individ-ual female, the fitness payoff comes only duringone or two spawning acts over the course of anormal life span. In species like striped bass, thevariance in larval survivorship that serves as inputfor population projections lies well beyond ourability to measure differences in the field. Manage-ment of exploited populations of long-lived, highlyfecund fishes requires the maintenance of criticaldensities of adult stocks and the protection ofspawning habitats during the short spawning sea-son. Because recruitment is largely determined byunpredictable interannual environmental variation,this critical density is impossible to determine withany degree of precision. And because most larvaenever recruit into the adult population even underpristine conditions, it follows that some spawningmust proceed unimpeded each year if strong re-cruitment is to occur during the exceptional year.

The opportunistic life history maximizes the in-trinsic rate of population growth (r) through a re-duction in the mean generation time (T). Fishesexhibiting an opportunistic-type strategy are oftenassociated with shallow marRinal habitats. These

edge habitats are the kinds of environments thatexperience the largest and most unpredictablechanges on small temporal and spatial scales.Changes in precipitation and temperature inducemajor alter~tions in water depth, substrate charac-teristics, and productivity in shallow aquatic habi-tats. Population density estimates for small fishesin shallow marginal habitats, like headwaterstreams and salt marshes, has shown large monthlyvariation. In the absence of chronic intense preda-tion and resource limitation, opportunistic-typepopulations can quickly rebound from localizeddisturbances.

Because they tend to be small and occur in shal-low marginal habitats, opportunistic-type fishesare not usually exploited commercially. Some im-portant commercial species, like gulf menhaden(Brelloortia patron us) , are intermediate betweenopportunistic- and periodic-strategies. Yet, smallfishes are often the most important food resourcesfor larger piscivorous species. Bay anchovies (An-choa mitchiU,) and silversides (Menidia menidia) in-habit relatively stable habitats, yet they suffer highadult mortality from predation. Given the capacityfor opportunistic-strategists to sustain losses duringall stages oflife spans that are typically rather short,one of the keys to their management is protectionfrom large-scale or chronic perturbations that elim-inate important refugia.

The equilibrium-strategy in fishes is roughlyequivalent to the traditional K-selection model ofevolution in density-dependent and resource-limited environments. For example. some stream-dwelling darters (Etheostoma) and madtoms (No-fUnIS) probably have fewer refuges in shallow rifflesduring periods of reduced stream flow. If refugesare limited, individuals may be forced to competefor depleted food supplies in areas immediately sur-rounding refuges. In the East African rift lakes,many brood-guarding and mouth-brooding cich-lids have home ranges that cover only a few squaremeters. In the marine environment. parental careis most frequently seen in small fishes associatedwith the benthos or structure (damselfishes, pipe-fishes, sea horses, eelpouts. some gobies). Com-pared with opportunistic- and periodic-strategists.equilibrium-strategists ought to experience lower

62 fISH ECOlOGY

temporal variation in population density and con-form better to stock-recruit models. ?

Because equilibrium-strategists produce smallnumbers of offspring. early survivorship must berelatively high for these populations to avoid lo-cal extinction. When parental care is involved.survivorship during early life stages depends onboth the condition of adults and their nestinghabitat. Relatively few equilibrium-type fishesare commercially exploited on a massive scale.Sev-eral species exploited by sport fisheries exhibitbrood" guarding and are intermediate betweenequilibrium- and periodic-strategies. includinglingcod (Ophiodon elongatus) and the North Ameri-can sunfishes. Management of exploited stocks ofequilibrium-type fishes requires the maintenanceof undegraded habitats and adult stock densitiesthat promote surplus yields that can be fished andreplaced via natural compensatory mechanisms.

cialized defenses against predators (other than hid-ing and schooling), yet their populations thriveowing to their rapid population turnover times andhigh growth potentials.

Piscivores c~ have major effects on the density,population structure, and behavior of their prey.The addition of piscivores (largemouth bass, Mi-cropterus) to North American lakes has been shownto decrease the abundance of small planktivorousfishes (minnows), which in turn leads to wholesalechanges in zooplankton, phytoplankton, and nutri-ent dynamics. Because mouth gape and the diame-ter of the throat limit the size of prey ingested,piscivores that swallow their prey whole are oftenvery size selective. Size selectivity can alter the sizeand age structure of prey populations. For example,small gizzard shad are more vulnerable than largerconspecifics to predation by largemouth bassand other piscivorous fishes. Shad populations inpredator-dense habitats tend to be dominated byolder age-classes comprising larger individuals. InEurope, the size/age structure of roach (Rutilus)populations is often influenced by predatory perch(Perca). Piscivores can also influence the competi-tive interactions between prey species. Walleye(Stizostedion) are efficient predators of both young-of-the-year perch (Perca) and gizzard shad (Doro-soma). In temperate lake ecosystems, young perchcompete with shad for zooplankton. If shad densi-ties are high, competition can cause a competitivebottleneck for perch recruitment into older age-classes. At high densities of gizzard shad, youngperch grow more slowly and remain vulnerable towalleye for a longer period of time.

Effects of piscivores on aquatic ecosystems areclearly demonstrated by introductions of exoticpredators. The predatory sea lamprey has contrib-uted to the decline of a number of fishes of theLaurentian Great Lakes, including lake trout andendemic whitefishes (Coregonus). The introductionof peacock bass (Cichla) into Lake GatUn, Panama,resulted in major changes in the aquatic food weband local extirpation of several native fish species.In Africa, the introduction of predatory Nile perch(LAtes) into Lake Victoria ranks among the mostdevastating exotic species introductions in history.Prior to the introduction of Nile perch, the lake's

VI. ECOLOGICAL INTERACTIONS

A. Predation

Predation has a major influence on fish evolution,population dynamics, and community structure.The diversity of feeding mechanisms of fishes isderived from the mechanical trade-offs involved inharvesting different kinds of food resources. In theabsence of stable food resources and interspecificcompetition, generalized feeding is often favored.Many fishes show seasonal diet shifts, in which agreater variety of prey is consumed during periodsof prey abundance. In diverse fish assemblages,predatory fishes usually exhibit divergence in feed-ing tactics and diet. Piscivores may feed by a sit-and-wait tactic (flounders, scorpion fishes), stealthand rapid pursuit (barracudas, pikes), solitary ac-tive searching (snappers, black bass), or activelysearching in schools (tunas, piranhas). Divergencein piscivore foraging tactics has coevolved with anequivalent amount of evolutionary and ecologicaldivergence in prey escape tactics. Prey species avoid

predation by crypsis, mimicry, hiding, schooling,spines, bony plates, and aggression. Small fiShes,like anchovies and killifishes, often have few spe-

63f&fEca.OGY

fish community was dominated by more than 400species of endemic haplochromine cichlids. LakeVictoria is now dominated by Nile perch, a nativeminnow, and an introduced species of tilapia (adetritivorous/herbivorous cichlid), and a greatnumber ofhaplochromine cichlids are now extinct.The complex food web supported by the ecologi-cally diverse haplochromines has been replaced bya simple food web linking detritus-shrimp, tilapia,minnows-Nile perch. Recent evidence indicatesthat the entire lake ecosystem's nutrient dynamicsmay have been destablized by the replacement ofhaplochromine biomass and diversity with shrimp,minnow, and tilapia biomass.

By influencing the behavior of their prey, pisciv-orous fishes can have subtle effects on ecosystemstructure and function. Recent studies have demon-strated that fishes are able to perceive predationrisks in different habitats, and that they alter theiruse of habitats in accordance with the relative costsand benefits of foraging gains versus predationrisks. Experiments have shown that algae-feedingstone rollers (Campostoma) will avoid stream poolscontaining largemouth bass. As a consequence,pools that contain bass develop and maintain morebenthic algae than pools without bass. In NorthAmerican streams, dace (Rhinichthys) will avoidpools that containing large piscivorous creek chubs(Stmotilus). Stream areas containing higher concen-trations of food induce dace to take greater preda-tion risks. Similarly, in the presence of largemouthbass (Mic,opttrus), bluegill sunfish (Ltpomis) spendmore time foraging in vegetation than in open wa-ter areas of ponds where foraging is moreprofitable.

fish assemblages in the tropics show more interspe-cific variation in ecomorphology than assemblagesfrom comparable habitats at higher latitudes. Re-source segregation in response to interspecific com-petition is a lifcely causal factor for this pattern. Ina study of four tropical fish assemblages, fisheswere clustered into feeding guilds and specieswithin guilds were segregated in their use of foodresources more than would be expected by chance.Many studies have shown high levels of resourcesegregation, particularly in fishes' use of habitat.Considerable diet and habitat overlap have beenobserved during early stages of ontogeny and dur-ing seasonal increases in resources in many differentaquatic and marine ecosystems. Declines in fishdiet or habitat overlap during periods of naturalresource depression provide comparative evidencefor interspecific competition.

Interspecific competition has been inferred fromobserved niche shifts in response to species intro-ductions in natural and experimental systems. Theintroduction of planktivorous alewife (A/osa PStU-doharengus) into Lake Michigan resulted in nativebloaters (Coregonus hay') shifting to forage in deeperwaters on larger and more benthic invertebrateprey. The bloater's shift in ecology was accompa-nied by a reduction in the length and number ofgill rakers formerly used for straining planktonfrom the water column. In experimental ponds,bluegill sunfish (Lepomis macrochirus), green sunfish(L. cyanellus), and pumpkinseed sunfish (L. gibbo-Jus) use a similar range of habitats and food re-sources when stocked as single species. When allthree are stocked together, bluegills feed more inopen water on zooplankton, green sunfish feedmore in the vegetation on aquatic insects, andpumpkinseeds feed more on benthic invertebrates.The growth rates of each species are slower inmultispecies ponds compared with similar single-species ponds, offering further evidence ofinterspe-cific competition for resources.

B. Competition

Competition occurs when the supply of resourcesis insufficient to meet the demand of two or moreconspecific consumers (intraspecific competition)or consumer populations (interspecific competi-tion). Food and habitat have been identified as theresources that most often limit individual fitnessand population densities. Historical competition isfrequently inferred from comparisons of speciescharacteristics within natural assemblages. Diverse

VII. COMMUNITY PATTERNSAND BIODIVERSITY

In general. regional fish diversity is positively cor-related with drainage basin area and negatively cor-

64 FISH ECOlOGY

tat gradients. Stream systems in many differentregions of the world show clear patterns of faunalchange and species turnover in relation to longitu-dinal fluvial gradients. Many small species are re-

,-stricted to headwater tributary streams, wherestream flow is highly variable and habitats are verydynamic. As one moves downstream into larger,more stable, and more productive habitats, a fewof the headwater species are retained and new spe-cies are added.

Fish biodiversity is threatened by a number offactors. The introduction of exotic species can elim-inate native species directly via competition andpredation or indirectly via alteration of the foodweb. On a global basis, the tiny North Americanmosquito fish (Gambusia affinis) and herbivorousAfrican tilapias (Tilapia, Sarotherodon, Oreochromis)may be the most widespread and damaging exoticspecies. Carp (Cyprinus carpio), largemouth bass(Micropterus salmoides) , rainbow trout (Oncorhyn-chus mykiss), and brown trout (Salmo solar) havebeen introduced worldwide for sportfishing. Theeffects of these introductions on native species havevaried greatly, but in most areas the full impact onthe native fauna is unknown. Exotic introductionscan result in genetic contamination of locallyadapted fish populations, as when hatchery salmonare introduced to rivers containing native stocks.Overfishing and environmental degradation froma variety of agents have caused the decline of fishpopulations and species diversity in virtually allmajor aquatic and marine ecosystems. Fluvial andestuarine ecosystems are among these most affectedby pollution and other anthropogenic perturba-tions in regional landscapes. Deforestation andplowing cause increased soil erosion, runoff, andsedimentation, all of which degrade fish habitat.Trace contaminants like lead and PCBs can attainhigh concentrations in fish tissues through the pro-cess of bioaccumulation. Elevated nutrient levelsfrom pollution and agricultural runoff increase pri-mary productivity of aquatic ecosystems, whichfrequently leads to higher biomass of less desirablespecies like carp and gar and declines in game fisheslike largemouth bass. In arid regions, fragile aquatichabitats have been completely destroyed by re-moval of subsurface water for irrigation, industry,

related with latitude. South American fresh waterscontain the highest regional freshwater fish diver:.sity, and the southwestern Pacific Ocean containsthe greatest diversity of marine fishes. Within fresh-water families having little or no salinity tolerance,species and populations show conservative biogeo-graphic patterns in which recently divergent taxaare restricted to disjunct drainage basins. On alonger time scale, distributions of fish families re-flect dispersal across continental landmasses viastream captures and freshwater intrusions intocoastal marine habitats. For example, catastomidsuckers range from eastern Asia across the BeringStrait into North America and south to centralMexico; and centrarchid sunfishes are restricted toeastern, central, and southern portions of NorthAmerica. Minnows, carp, and other cyprinid fishesare found on all continents except South America,Australia, and Antarctica. Marine taxa have distri-butional patterns that reflect biogeographical pro-cesses on different scales. The Central Americanisthmus separates recently divergent sister speciesof numerous coastal marine taxa. Several marinefamilies have circumglobal distributions that arerestricted by latitude, such as the anadromous sal-monids in the Northern Hemisphere and their sisterclade, the anadromous argentiniids in the SouthernHemisphere [See BIODIVERSITY.]

At a local level, fish species composition resultsfrom both historical regional processes and con-temporary ecological processes. Analyses of lakeassemblages in Canada showed that geographicaldistance explained the level offish assemblage simi-larity more than local habitat conditions. In con-trast, streams located only a few kilometers apart,but with different habitats, may share only a verysmall fraction of their species. For example, sometropical freshwater fishes are restricted to "black-water" streams that flow through regions con-taining nutrient-poor soils. Other species inhabitonly nutrient-rich "whitewater" rivers that sup-port high aquatic primary production and biomassof invertebrates and fishes. Species composition ofthese assemblages correlates more with habitat con-ditions than geographical distance. Several studieshave used multivariate statistical methods to exam-ine fish assemblage composition in relation to habi-~

65~ ECOlOGY

and municipalities. When they block movement tocritical spawning habitats, dams can eliminate localpopulations of migratory fishes in little more thana single generation. [See MARINE BIOLOGY, HUMAN

IMPACTS. ]Fishes can serve as sensitive indicators of water

quality and ecosystem health. Small species thatadjust well to laboratory conditions, such as thefathead minnow (Pimephales promelas) and the me-daka (Orizias latipes), are used extensively in toxi-cological research. Species that have broad geo-graphical ranges and are highly sensitive to habitatdegradation are useful as bioindicators. For exam-ple, many North American stream-dwelling dart-ers (Percidae) require oxygen-rich water flowingover hard substrates. These species are among thefirst eliminated from habitats degraded by pollu-tion, siltation, and other landscape perturbations.In addition, the structure of local fish communitieshas been used as an indicator of ecosystem stress.Under normal conditions, diverse fish faunas willcontain a mixture of detritivorous. insectivorous.omnivorous, and piscivorous fish species. Often-times, piscivores will be supported by greater num-bers and a greater variety of fishes at lower trophiclevels. An index of biotic integrity was developedto evaluate ecosystem change as a result of anthro-pogenic perturbations. This index compares thestructure of fish communities in degraded habitatswith the structure of communities in similar butunperturbed habitats from the same biogeographic

region.

Catadromous Of or pertaining to species that spend mostof their lives in fresh water and migrate to the sea tobreed.

Demenal Of or pertaining to the deep-water regions nearthe bottom or substrate of aquatic and marine environ-.-ments.

Diadromous Of or pertaining to species that migrate be-tween fresh water and the sea.

Fluvial Of or pertaining to flowing freshwater ecosystemssuch as streams and rivers.

Iteroparity Condition of performing repeated episodes ofreproduction over the lifetime.

Marine Of or pertaining to oceanic or coastal environ-ments. typically containing solute concentrations greaterthan 30 parts per thousand.

Pelagic Of or pertaining to the open-water region of fresh-water and marine environments.

Pharyngeal teeth Throat teeth that occur on pads on vari-ous gill arch elements in many fishes.

Semelparity Condition of performing a single bout of re-production during the lifetime.

Stock Local population of fishes. usually referring to onethat is exploited as a commercial or recreational resource.

Glossary

Anadromous Of or pertaining to species that spend mostof their lives in the sea and migrate to fresh water tobreed.

Assemblage Group of species populations coexisting in alocal area. usually pertaining to species belonging to thesame higher taxonomic grouping (e.g.. fish assemblage.minnow assemblage. trout assemblage).

Benthic Of or pertaining to the bottom or substrate regionof freshwater and marine environments. Category offishes that inhabit the bottom region (e.g.. many cat-fishes. sculpins. darters).

Bibliography

Breder, C. M., Jr., and Rosen, D. E. (1966). "Modes ofReproduction in Fishes." Garden City, N.Y.: NatUralHistory Press.

Hilborn, R., and Walters, C. J. (1992). "Quantitative Fish-eries Stock Assessment." New York.: Chapman & Hall.

Lowe-McConnell, R. H. (1987). "Ecological Studies inTropical Fish Communities." Cambridge, England:Cambridge University Press.

Marshall, N. B. (1971). "Explorations in the Life of Fishes,"Harvard Books in Biology, No.7. Cambridge, Mass.:Harvard University Press.

Matthews, W.J., and Heins, D. C., eds. (1987). "Commu-nity and Evolutionary Ecology of North AmericanStream Fishes." Norman: University of OklahomaPress.

Moyle, P. B., and Cech,J.J.,Jr. (1988). "Fishes: An Intro-duction to Ichthyology." Englewood Cliffs, N.J.: Pren-tice-Hall.

Nelson,J. S. (1984). "Fishes of the World," 2nd ed. NewYork: John Wiley & Sons.

Sale, P. F., ed. (1991). "The Ecology of Fishes on CoralReefs." New York.: Academic Press.

Winemiller, K. 0., and Rose, K. A. (1992). Patterns oflife-history diversification in North American fishes: Im-plications for population regulation. Canad. J. Fish.Aqua,. Sli. 49(10), 2196-2218. .

Wooton, R. J. (1990). "Ecology of Teleost Fishes." NewYork: Chapman & Hall.