Patterns of Variation in Life History Among South American Fishes in Seasonalenvironments
Total Page:16
File Type:pdf, Size:1020Kb
:>ecologia (1989) 81 :225-241 (k(:i)logia D Springer-Verlag 1989 Patterns of variation in life history among South American fishes in seasonalenvironments Kirk O. WillelDiUer Departmentof Zoology, The University of Texas,Austin, TX 78712.USA s..mary. Ten traits related to life history theory weremea- demographic parameters (Steams 1983a; Reznick 1985; sured or estimatedfor 71 freshwaterfish speciesfrom two Peaseand Bull 1988). For example, models have incorpo- locations in the Venezuelanllanos. Multivariate statistics rated negativecorrelations betweencurrent investmentand and cluster analysis revealedthree basic endpoint patterns future investments in reproduction, juvenile survivorship bounding a two-dimensional continuum. A suite of attri- and clutch size. reproductive effort and adult survivorship, butesassociated with parental care and aseasonalreprodo<:- and mean generationtime and the intrinsic rate of increase. tion appeared to correspond to an equilibrium strategy. Alternative life history predictions can be testedby evaluat- A secondgroup of small fISheswas distinguishedby traits ing patterns of covariation among reproductive and demo- associatedwith rapid colonizing ability: early maturation. graphic parameters among heterospeciflCpopulations or continuous reproduction, and small clutches. The third conspecificdemes that have experienceddifferent environ- basic pattern was associatedwith synchronizedreprodo<:- mental conditions for long time periods. The presentstudy tion during the early wet season.high f~undity, absence constitutes one such test. characteriDng patterns of repro- of parental care. and breeding migrations. A subset of duction in tropical fishes with a seriesof dependentvari- mostly small fishes exhibiting little or no parental care. ables.The degreeof speciesclustering in multivariate space small clutches.and two to four month reproductiveseasons indicates the likelihood that certain combinations of traits was intermediatebetw~ the opportunistic (rapidly colon- will occur together in nature. Finally, I evaluate species izing) and seasonalstrategies. All ten life history variables clusters for associationwith phylogeny and ~Iogical fac- showedsignificant effectsof phylogeny.The cluster of spe- tors. ciescorresponding to the equilibrium group was dominated Tropical freshwaterfishes exhibit large diversity in mor- by siluriform fishes and perciforms of the Cichlidae. The phological, physiological. and ~logical attributes (Lowe- opportunistic cluster was dominated by cyprinodontiform McConnell 1987).Despite few comparative studiesof tropi- and characiform fishes. whereasthe seasonalcluster con- cal fish reproductive strategiesto date (Africa-Welcomme tained primarily characiform and siluriform fishes. Seven 1969; Central America-Kramer 1978; Asia-DeSilva et al. of nine traits weresignificantly correlatedwith body length. 1985),these diverse fish assemblagesprovide excellentsys- The three reproductive patterns are interpreted as being tems for evaluation of life history patterns. The great drain- adaptativewith respectto relative intensity and predictabi- age basinsof South America harbor the most diversefresh- lity of temporal and spatial variation in abiotic environmen- water fish assemblageson earth (ca. 2400described species). tal parameters.food availability. and predation pressure. Rather paradoxically, this ichthyofauna is derived from Key wonls:Llanos - Reproduction- Strategies- Tropical only a few basic stocks (speciesbelonging to the Characi- fishes-Venezuela formes. Siluriformes. and Cichlidae (Perciformes)comprise about 93% of all species,Lowe McConnell 1987).I investi- gated fish reproductive biology in the llanos of Venezuela during twelve co~tive months in 1984.The region has The diversity of methods by which organisms reprodtx:e an averageannual temperatureof 26.2° C and a highly sea- has long intrigued naturalists. One can argue that traits sonal pattern of rainfall, with 1189mm of precipitation oc- associatedwith reproduction should be subject to intense curring from June-Novemberand 178 mm from December- natural selection, as these directly affect the individual's May in 1984. Despite extreme changesin environmental geneticrepresentation within subsequentgenerations. Alter- conditions caused by seasonalrainfall. fishes exhibited a native modes of reproduction have obvious demographic wide rangeof reproductivecharacteristics in the llanos. This implications, sinceall populations experienceeither gradual report describesthese characteristics for 71 speciesand in- or periodic turn-overs.Alternative patternsof reproduction terprets the basic patterns within an ~logical context.~ and the degreeto which these are s~ful in different environmentalsettings have formed the basisfor both theo- Methods retical and empirical researchunder the headinglife history (Steams 1976; Southwood 1977; Horn 1978; Sibley and Collections Calow 1986a).In the most basic terms, life history theory During every month of 1984, fishes were collected from involvesconstraints, or trade-offs, among reproductiveand two locations in the state of Ponuguesa, Venezuela. The 226 more faunistically diversesite, Cano Maraca, is a swamp- servedin 10% fonnalin and depositedin the MHN (UNEL- creek located in the western llanos region (8°52'30" Lat. LEZ) and the TNHC (TMM). N; 69°53'30" Long. W). The area studied (termed an "es- tero") lies within low, flat terrain that experiencesextensive The data sheet flooding during the wettest months (June-August). During this time, the creek's broad floodplain is converted Several attributes related to reproduction and population from exposed,sun-baked soil and thorn-scrub habitat into structure weremeasured directly and other life history traits a productive marsh. The aquatic habitat is reduced to a werecalculated or estimatedfrom these.All preservedspec- network of pools (averagedepth ca. 1.0 m) during the driest imens were identified and measured for standard length months(December-May). Dissolved oxygen concentrations (SL). When available, 30 specimensof each speciesfrom are reducedduring this time, and many fishesrely on special eachmonthly collection at both Calio Maraca and C. V 01- respiratory adaptationsfor survival (Winemiller 1988). can were dissected.The condition of the gonad was coded The other site, Cano Volcan (8°59'15" Lat. N; basedin its relative sizeand color using the following cate- 69°53'30" Long. W), is a third order stream in the lowest gories: clear (1), translucent(2), opaque(3), small (1), medi- tier of the Andean Piedmont. This narrow stream flows um/small (2), medium (3), medium/large (4), and large (5). through deciduousforest and contains both pool and rime The overall gonad code for each individual was equal to habitats. Cano Volcan differs from C. Maraca in having (color code+size code) divided by two, and as a result, a slightly steepergradient, more stabledry seasondischarge, values ranged from 1.0 to 4.0. Although the same basic and a more forested watershed.Cano Volcan experiences criteria were usedin coding male and female gonads,inter- frequent but brief flash floods during wet seasondown- sexual and interspecific differences in the size, color, and pours. Physico-chemicaldata for each site during the year texture of fully mature gonads were taken into account. studied are given in Winemiller (1987). Undevelopedovaries were tiny transparent, globular or la- Fishes were collected by dipnet, seine (3.2 mm mesh/ mellar structures. Following a gradual transition, fully-de- 2.5 m length; 12.7mm/20 m), experimental gillnet, and veloped fish ovaries were large opaque-yellow or orange hook and line. At eachsite, an attempt was made to sample structures having a grainy appearancedue to the presence the entire fish community such that the sample for each of mature oocytes. Mature ovaries are cylindrical or lie speciesreflected its relative abundanceand population size as sheetsalong the lateral walls of the abdominal cavity. structureduring eachmonth. The collecting effort expended Immature testesappeared as transparent threadlike or min- each month was approximately equal for each site, con- ute lamellar structures in contrast to the smooth, opaque, sisting of three or four hauls of the 20 m seine in open milky-white appearance of mature testes. Fully-mature water habitats, and alternateuse of the 2.5 m seine,dipnets, testeswere either cylindrical. lamellar. or multilobed (the and gillnets for a minimum of 6 h within open water, shore- latter condition occursin catfishesof the Pimelodidae).For line, vegetation,and peripheralpool habitats. Samplingwas severalspecies, gonad codeswere regressedagainst gonado- terminated when two hours of collecting yielded no addi- somatic indices to test their reliability as overall indices tional rare species.Collections used for comparisonsof rela- of gonadal development(Fig. 1). tive population densitieswere made during either one or Diameters of ten of the largest oocytes in each mature two days betweenthe 11th and 28th of each month. Addi- ovary were determined using a dissectingmicroscope and tional collections were made on other dates in order to an ocular micrometer. Oocyte sizeswithin each examined supplementthe numbersof uncommon speciestaken in the ovary wereclassified as extremely uniform, moderately uni- standard samples.Except for very abundant species(for form, moderately variable, or extremely variable. The which a subsamplerepresentative of the abundance rank number of separateoocyte size classespresent