Environmental Biology of Fishes (2005) 72: 45-54
@ Springer 200~
Relationship between hatch date and first-summer
growth of five species oj
prairie-stream cyprinids Bart W. Durham & Gene R. Wilde Wildlife & FisheriesManagementInstitute, Mail Stop 2125, Texas Tech University Lubbock, TX 79409, U.S.A. (e-mail:
[email protected]) Received3 July 2003
Accepted 3 March 2004
Key words: growth increments, bet-hedging,overwinter mortality, young-of-the-year Synopsis
Stream fishes often exhibit a bet-hedging multiple spawning reproductive strategy. In many species,the reproductive seasonlasts severalmonths. This exposesyoung fishesto varying environmental conditions that may differentially affect growth. We studied the effect of hatch date on first-summer growth among members of a prairie-stream fish assemblage.The reproductive seasonin both years of the study was protracted, lasting from April through August. Due to intermittent stream-discharge,there were two distinct periods during which most speciessuccessfullyreproduced. In general, growth rate was greater among individuals with an early hatch date than among those with a later hatch date. Multiple regression models indicated that hatch date was related to growth in all study specieswith one exception(red shiner, Cyprinella lutrensis).The results of this study provide evidencethat young-of-the-yearof multiple spawning stream-fishspeciesthat are spawnedlate in the seasonmay grow at a slower rate than young spawned earlier in the season.
Introduction Reproduction by most fishesis seasonalin occurrence and timed to maximize the probability that offspring will have necessaryresourcesfor growth and survival (Cushing 1990, Conover 1992, Humphries et al. 2002). Timing of reproduction in some fishesis strongly linked to seasonalcycles in food availability (Cushing 1990, Leggett & DeBlois 1994),whereasfor others reproductive output appears to be associated with specific environmental conditions, such as temperature,photoperiod length, and water level or seasonalrainfall (Lowe-McConnell 1979, Munro 1990). Synchronization of reproduction with optimum conditions for growth and survival of offspring is more difficult for speciesthat inhabit variable environments. Streamsand rivers often are characterizedby ex-
tremely variable physical and chemical conditions and frequent disturbances such as flood and drought. Prairie and lowland streams, in particular, have been recognizedas highly variable environments that are characterized by rapid and unpredictable changes in discharge, temperature, pH, and dissolved oxygen concentration (Matthews 1987,1988). Fishes that inhabit prairie and lowland streams commonly exhibit a classicbet-hedgingstrategy of spreading reproductive output over an extended period, thereby increasingthe chancesthat a portion of the reproductive output will result in successful production of offspring (Lambert & Ware 1984, Rinchard & Kestemont 1996,Trippel et al. 1997).Becausethese fish spawn multiple clutches of eggs,severalsizesof offspring may be presentat any given time during the reproductive season.
46
Environmental conditions may differentially affect growth of young produced at different times. Individuals resulting from spawningepisodesearly in the season have the advantage of a longer growing season,compared to individuals spawned later (Keast & Eadie 1984,Conover 1992,Trippel et al. 1997). Becauseindividuals spawned later in the seasonhave a shorter first growing seasonthey may be subject to size-selectiveoverwinter mortality, which has beendocumentedfor a number of species(e.g., Hendersonet al. 1988,Shuter& Post 1990). Thus, hatch date and subsequentgrowing conditions may be critical determinants of reproductive successof prairie-stream fishes. In this study we examine first-summer growth of five members of a prairie stream-fishassemblage. Species studied are all members of the family Cyprinidae and include: Arkansas River shiner Notropis girardi, peppered chub Macrhybopsis tetranema,plains minnow Hybognathusplacitus, flathead chub Platygobio gracilis, and red shiner C}'prinella lutrensis. Except for red shiner, these fish are members of a reproductive guild that broadcast spawns non-adhesive, semi-buoyant eggs that are carried downstream by the current during incubation and larval development(Platania & Altenbach 1998). Red shiner spawns adhesive demersaleggsthat attach to substrate or other instream structures during development. Our study objective was to determine if growth of young-of-the-yearfish is related to hatch date. The null hypothesis underlying this investigation was that growth rates are similar for all young-of-theyear of a given speciesregardlessof hatch date.
Material and methods Studyarea The Canadian River is located in south central United States and is characterized by sand substrate and shallow braided channels typical of Great Plains prairie streams (Matthews & Hill 1980). The Canadian River is the southernmost tributary of the Arkansas River and rises on the easternslope of the Sangrede Cristo Mountains of northeast New Mexico and southern Colorado. The river flows generallyeastward through eastern New Mexico. Texas. and Oklahoma where it dis-
charges into the Arkansas River (Dolliver 1984). The study portion of the Canadian River is approximately 218km in length and is impounded upstream by Ute Reservoir in New Mexico and downstream by Lake Meredith in Texas (Figure 1). River discharge is variable and is affected by seasonalrainfall, periodic releasesfrom Ute Reservoir, irrigation return flows, and discharge from a major tributary, Revuelto Creek (Bonner & Wilde 2000). Arkansas River shiner, peppered chub, plains minnow, flathead chub, and red shiner (the five speciesstudied herein) collectively comprise 90% of the fish assemblagein this portion of the Canadian River (Bonner &Wilde 2000). Sampling
We collected young-of-the-yearfish from two sites on the Canadian River (Figure 1). Site 1 is located at the junction of U.S. Highway 287 (Potter County, Texas;35028' 13" N, 101052'45" W). Site 2 is located near the junction of U.S. Highway 385 (Oldham County, Texas; 350 31' 25" N, 1020 15' 42" W). We collected fish from May through Septemberin 2000 and 2001 with a small mesh seine (1.8 x 3.4m, 1-mm mesh). We made collections bimonthly in May and September and weekly during June through August unless river conditions prevented sampling. We preserved all samplesin 95% ethanol in the field. Laboratory
procedures
We identified fish to speciesand measured total length (TL) to the nearestmillimeter. We removed the largest pair of otoliths, the sagittae, from each fish. Procedures for otolith preparation and counting of daily growth increments generally follow the methods outlined by Secoret al. (1992) and Miller & Storck (1982). We attached one otolith to a clear glass microscope slide with thermoplastic cementand stored the other otolith for use in the event that the first one was broken or lost. We ground otoliths of larger specimenswith fine grit sandpaperand polished them (Taubert & Coble 1977, Secor et al. 1992) to make growth incrementsmore visible. Two experiencedreaders independently counted daily growth increments for all otoliths using a compound light microscope at 40x magnification. We acceptedage estimates
47
Figure J. Study area and location of two collection sites on the Canadian River, Texas. Collection sites are indicated by arrows.
from the two readersthat were within 10% of each other and assignedfinal daily agesas the mean of these two estimates. When initial readings varied by more than 10%,a third readercounted growth increments.If the third reading fell within 10% of one or both of the other two readings,we assigned a mean of the two closestreadingsas the final age estimate (Miller & Storck 1982). However, if the third reading was within 10% of the two initial readings and fell exactly between the two initial readings, all three readers re-read the otolith and we either assigned a final age according to the guidelines above or excluded the otolith from analyses. We excluded all otoliths for which readingscould not be reconciled to within 10%by any two readers. Cyprinid fishesproduce daily and annular otolith increments (Victor &Brothers 1982, Escot & Granado-Lorencio 2001). There is a strong relationship betweennumber of daily incrementsand otolith diameter, which allows validation of daily ages by regressing age on otolith diameter (Folkvord et al. 2000). Among the species we studied otolith diameter, measuredto the nearest 0.0 I mm along the longestaxis (Secoret al. 1992),
showeda strong positive correlation with assigned daily ages(r = 0.81-0.95, p < 0.001). Data analysis Hatch date was determined for each fish as the difference,in days, betweendate of collection and age of the fish (determined by the number of daily rings). Mean daily increasein length, a commonly used measure of growth, was calculated for each fish by dividing TL by age. Because growth in young fishes is generally non-linear, for example fish at five days old add more body length per day than fish at 50 days old, and becauseour samples included fish of varying ages, we adjusted length for age using linear regressionafter log-transformation of length and age.We usedmultiple linear regressionto assessthe effects of length and hatch date on growth.
Results The five study species produced offspring throughout a protracted spawning season, from
48 2000
2001
1.1.
1.1
0.9.
0.9.
0.7.
0.7 .
0.5.
0.5.
0.3.
0.7.
0.3. 0.1. 1.1. 0.9. 0.7.
0.5.
0.5 .
'!o 0.3. E
0.3 .
0.1
1.1 0.9.
-
.s
0.1
0.1'
1.1.
:5 OJ c: 0.9.
0.9 .
0.7.
0.7 .
1.1.
.E!. .5 Q)
0.5.
m 0.5. Q) 'u 0.3. cn
.5 :?::o
0.3 .
0.1.
0.1
"ffi 1.1.
1.1
"C
c: 0.9. m Q) ~ 0.7.
0.9 . 0.7 .
0.5.
0.5.
0.3.
0.3 .
0.1.
0.1 .
1.1.
1.1 .
0.9.
0.9. 0.7.
0.7 ;
0.5.
0.5
0.3.
0.3
0.1
0.1
Apr
May
Jun
Jul
Aug
Sap
Apr
May Jun
Jul
Aug
Sap
Hatch date Figure 2. Mean daily increasein length for young-of-the-year prairie-streamcyprinids from the Canadian River, Texas in 2000 and 2001.
April through August during both years of the study. In eachyear, there were two distinct periods of successfulreproduction separatedby a period in which no reproductive successwas evident (Figure 2). This pattern appeared to be associated with periods of drought, which occurred during the
reproductive season of both years (Figure 3). Mean daily increasein length (growth) was highly variable in all speciesin both years of the study and ranged from 0.22 to 0.83 mm day-I for Arkansas River shiner, 0.26-0.75mmday-1 for peppered chub, 0.22-1.0 mm day-I for plains
49
Figure 3. Canadian River dischargeduring 2000 and 2001 reproductiveseasons.
minnow, 0.25--0.76mm day-I for flathead chub, and O.18---Q.88 mm day-I for red shiner (Figure 2). Mean daily increasein length for plains minnow in 2000 and Arkansas River shiner, plains minnow, flathead chub, and red shiner in 2001 appearedto belessfor individuals spawnedand hatchedduring the second period of successful reproduction (Figure 2). There was a significant positive (p < 0.001) relationship betweenlog-length and log-age for all species in both years of the study, with log-age explaining 46--86%of the variation in log-length (Table 1). The slopes of all regressionswere significantly different (p < 0.001) from].O indicating that growth is not isometric. To fully assessthe effect of hatch date on growth we usedmultiple regressionto adjust for the affects of age on length. Multiple regressionmodels were significant (p < 0.00]) for Arkansas River shiner and plains minnow in 2000 and for Arkansas River shiner, peppered chub, plains minnow, and
flathead chub in 2001 (Table 2), showing that growth rates of thesespecieswere related to hatch dates. Including hatch date in these models explained an additional I % (Arkansas River shiner, 2000)to II % (flathead chub 2001)of the variation in length. Hatch date was negatively related to growth rate in Arkansas River shiner, plains minnow, and flathead chub, indicating that individuals spawned later in the reproductive season grew at a slower rate than individuals spawned earlier in the season.There was a positive relationship between hatch date and length in peppered chub, indicating that individuals spawned later in the seasongrew at a greaterrate than those spawnedearlier.
Discussion Hatch date significantly affected growth rate in most species studied. Except among peppered
50 Table J. Linear regressionmodels for total length and age (independentvariable) of young-of-the-year prairie-stream fish collected from the Canadian River, Texas during the 2000 and 2001 reproductiveseasons. Species
Year
Intercept
Slope
Rangeof ages (days)
?
Arkansas River shiner
2000 2001
0.520
0.628 0.764
14-115 13-65
0.74 0.63
