Great Basin Naturalist Great Basin Naturalist

Volume 50 Number 2 Article 1

6-30-1990

Mayfly growth and population density in constant and variable Mayfly growth and population density in constant and variable

temperature regimes temperature regimes

Russell B. Rader Colorado State University, Fort Collins, Colorado

James V. Ward Colorado State University, Fort Collins, Colorado

Follow this and additional works at: https://scholarsarchive.byu.edu/gbn

Recommended Citation Recommended Citation Rader, Russell B. and Ward, James V. (1990) "Mayfly growth and population density in constant and variable temperature regimes," Great Basin Naturalist: Vol. 50 : No. 2 , Article 1. Available at: https://scholarsarchive.byu.edu/gbn/vol50/iss2/1

This Article is brought to you for free and open access by the Western North American Naturalist Publications at BYU ScholarsArchive. It has been accepted for inclusion in Great Basin Naturalist by an authorized editor of BYU ScholarsArchive. For more information, please contact scholarsarchive@byu.edu, ellen_amatangelo@byu.edu.

The Great Basin NaturalistPUBLISHED AT PROVO, UTAH, BY

BRIGHAM YOUNG UNIVERSITY

ISSN 0017-3614

VOLUME 50 30 JUNE 1990

MAYFLY GROWTH AND POPULATION DENSITY INCONSTANT AND VARIABLE TEMPERATURE REGIMES

Russen B. Raderl•2 and James V, 'Vard'

No.2

optimal with respect to growth, development,and body size. Fecundity, an essential compo­nent, but not the only factor defining repro­ductive success, should decrease in tempera­ture regimes warmer or cooler than optimal.O.ther factors, which may influence fecundityand mayor may not be influenced by temper­atw'e, can determine population size and dis M

tdbution (e.g., egg-hatching success, emer­gence success, mating success, feeding rates,assimilation efficiency, food quality and quan­tity, biotic interactions). The TE hypothesis,however, attempts to defme the influence oftemperature on population size and distribu­tion based only on the effect of temperatureon metabolism, growth, and therefore bodysize/fecundity.

••

ABSTRACT,-The thermal equilibrium hypothesis predicts that aquatic insect body sizelfecundityand, consequently.population density and hiomass will be maximized in geographic areas or along altitudinal gradients where the thermalregime is optimal with respect to growth and development. Seasonal growth analyses of three mayfly species,combined with detailed thermal descriptions, were used loexplore differences in hOOy size and fecundity at three siteswith similar elevations but difIerent temperature regimes. Site 1was located near the upper altitudinaldislribution foreach species, whereas sitf'-S 2 and 3 were located below a deepfirelense storage reservoir. The temperature pattern atsite 1 had rapid seasonal changes, with a short summer and a long., freezing winter. Site 2 demonstrated grddualseasonal changes combined with winter warm and summer OOQI tcmperahtre.s. Site 3 wa.~ intermediate with respect toseasonal change and winter harshness but had the highest maximum and mean annual temperatures, Mayfly develop­ment at site 1 was characlerized by slow growth during the sununer~autumnperiod, no growth dUring the winter, anda rapid increase during tht) spring-summer period. In contrast, growth at site 2 WliS continuous throughout the year,including: the winter. Growth l\t site 3 was either continuous across sites or rapid during the spring-summer t>eriod,deponding on the species, Based upon the interactions among temperature, body size, and metabolic (,,'Osts, thethermal eqUilibrium hY110thesis was suCC€ssfulat predicting body size and fecundity dilferenl..'Cs among sites. It was lesssuccessful at predicting variation in population density and biomass. Density-depend.ent and density~independent

sources of mortality, including temperature, may interrupt the translation ofhi~er fecundity into higher populationdensity and biomass.

Temperature. because of its influence onmetabolism, growth, and reproductive suc­cess, is a dominant ecological determinant ofthe geographical and altitudinal distributionsofaquatic insects (e.g., Vannote and Sweeney1980, Ward and Stanford 1982, Sweeney1984, Ward 1986). The thermal equilibriumhypothesis (TE) is a conceptual model of theeffects of. temperatltl"e on aquatic insect

"metabolism, growth, b<;>dy size, and thereforefecundity (Sweeney ali~ Vannote 1978, Van­note and Sweeney 198Q). It predicts that pop­ulation density, distribution, and stahility(Connell and Sousa 1983) are determined byindividual reproductive success and will bemaximjzed in geographic areas or along altitu~

dinal gradients where the thermal regime is

IDepartmelit ofBiolol';Y, Colorado Stale Uni~sity. Fort Collins. Col.orada 80523,'Present addre1i: S~nab f\iycr Ero!ogy t..bontory, OnJ.....oer E. Aiken, Suuth C:lmllna 2.9&)3,

97

98 R B. RmEBAND J. V. WARD [Volume 50

The objective of this study was to analyzethe influence of temperature on the growthand body size/fecundity of three mayfly spe­cies and compare these results \Vith popula­tion size (density and biomass) data andthe predictiuns of the thermal equilibIiumhypothesis. Three sites were chosen on thesame ,iver, all \\ith similar elevations but dif­ferent temperature regimes. Site 1 was lo­cated near the upper altitudinal limit for eachspecies, whereas sites 2 and 3 were located ina wanner, more constant thermal regimedownstream from a deep-release reservoir.Specifically, we sought to test the followinghypotheses: (1) populalion size and body sizewill be smaller for each species at site 1 com­pared with sites 2 and 3; (2) population sizewill correspond wilh body size at the samesites (i. e., Site-specific ranks of body size andpopulation size should be the same); (3) sea­sonal patterns of growth will parallel seasonaltemperature patterns. Body size/fecunditywas compared lo examine its ability to explainamong-site diflerences in population size.

Several studies have demonstrated aposi­tive correlation between body size and fecun­dity in aquatic insects (see Clilfurd and Boer­ger 1974, Kondratieff and Voshell 1980,Sweeney and Vannote 1981). Therefore, weassume that larger mayflies produce moreeggs compared with smaller mayflies of thesame species. We did not attempt to deter­mine, as predicted by the TE hypothesis andnumerous other authors. whether an increasein population size is positively correlated withan increase in population stability. Consider­able evidence, however, indicates that largerpopulations are more stable than smaller pop­ulations (e.g., smaller populations are moresusceptible to extinction).

STuDY SITES AND INSEGrALTITUDINAL DlSTRlBUTIONS

The first hypothesis required us to a priorirank the study' sites according to how closelythey approximated the optimal temperatureconditions for the mayflies under investiga­tion. Our ranks were based on the altitudinaldistIibutions of the insects. Because the studysites were located near theu' upper limits, theunaltered temperature regime at site 1 wasconsidered cooler than the optimum neces­sary to maximize body Size/fecundity. There-

fore, the warmer, dam-impacted sites 2 and 3were assumed to be nearer the insects' opti­mal temperature regime. There was no a pri­ori reason to separate sites 2 and 3 with re­spect to their influence on growth, body size,and population size even though they hadvery different temperature regimes.

The study was conducted in the Upper Col­. orado River on the westem slope of the RockyMountains in the vicinity of Granby Reser­voir, a large (666 km"), deep-release storageimpoundment. Granby Reservoir is located38 km northwest of Denver, Colorado. Site 1was located in a third-order, free-flowing sec­tion of the river 4.0 kin above the reservoir;sites 2 and 3 were located 0.4 and 4.0 km,respectively. downstream from the dam. Al­though differentially influenced by streamregulatiol', all three sites had similar gradi­ents (0.006-0.009), canopy cover; geology,riparian vegetation, and elevation (2,593 m,2,454 m, and 2,426 m, respectively. (For com­plete site descriptions see Rader and Ward[1988).)

Three mayfly species were analyzed in thisstudy: Dronella grandis (Eaton), Ephemerellainfrequens (McDunnough), and Baetis tricau­datus (Dodds). Ward (1980, 1986) determinedthealti.tudinal distribution of macroinverte­brates, incl uding the mayflies of this study, inthe St. Vrain River, a free-flowing stream run­ning from the alpine tnndra to the plains ontbe eastern slope of the Rocky Mountains.Based on his results, site 1 ofthis study (uppermontane zone) was above the altitudinal dis­tIibulion fur D. grandis and very near theupper limits .for E. infi'equens and B. bicau­datus. All three species exhibited maximumdensities at lower elevations in the foothills 'orplains zones. Even though their altitudinalupper limits appear to be somewhat higher inthe Colorado River, probably because of itslarger size compared with the St. Vrain River,we concluded that all three study sites were•located near the upper altitudinal limit foreach of the three mayfly species.

METHODS

Temperature, Growth, arid Body Size"

Water temperature was measured continu­ously at each site fur 18 months using Ryan90-day thermographs. Each thermograph waschecked against a Weksler hand-held ther­mometer on a monthly basis and calibrated

1990] MAYFLY GROWfH AND'fEMPERATURE CONSTANCY 99

prior to placement and following retrievaLDaily mean temperatures were used to calcu­late annual mean temperatures, annual coeffi­cient 'of variation, annual degree days, num­berofdays less than 3 C, number ofdays equaltoO C, length of spring-summer and summer­autumn peliods, and rate of spring-summerincrease and summer-autumn decline.

Annual growth rate analyses and generaltemperature descriptors (e.g., accumulateddegree days, mean annual temperature,etc.) cannot explain site-specific variation inaquatic insect body size and fecundity be­cause they average over important seasonalinformation. Seasonal growth rate analysescombined with seasonal temperature profilescan, however, provide insights into the rela­tionships between temperature, growth, andbody size.

Temperature pro61es for each site wereseparated into three periods: (1) spring-sum­mer, (2) summer-autumn, and (3) winter(Fig. 1). The winter period was defmed bymean daily temperatures <3 C in order toinclude the winter warm temperatures atsite 2. The end of the spring-summer period!beginning of the summer-autumn period wasset at 15 August, based on temperature peaksapparent at sites 1 and 3 (Fig. 1). Therefore,the spring-summer period began when themean daily temperature exceeded 3 C andended 15 August. The summer-autumn pe­riod began 15 August and ended when themean daily temperature dropped below 3 C.

Growth, defined as the monthly increase inmean biomass of individuals collected persampling date, was determined for D. grandisand E. infrequens. Growth for B. tri.caudatuswas not analyzed because of difficulty in as­signing intermediate-sized instars to the cor­rect generation.

Site-specific differences in seasonal growthrates were determined by regressing monthlymean individual biomass estimates against thenumber of Julian days accumulated over thetlu'ee separate growth periods (spring-sum­mer, summer-auhl1nn, and winter). A slopescomparison test (analysis of covariance) wasused to determine among-site differences inseasonal growth rates. No boansformation wasnecessary because growth was linear over theshort seasonal periods.

Population Size and Body Size

Estimates of population density and bio-

mass were based on four Surber samples(0.09'm each, 240 flom mesh) collected monthlyacross the width of the stream at each site andfour artificial substrates. Artificial substratesconsisted ofclay bricks (23 X 19 X 9.5 cm) thathad been in the streams for one month beforebeing sampled. Most of the Surber samplesenclosed natura) substrate particles similar insize to the artiflcial brick substrates. There­fore, sampling units from both techniqneswere combined because the samphng areaswere approximately equal. A simple t testindicated that the mayfly population meansbased on Surber samples of natural substratewere not significantly (P = .63) different frompopulation means based on the artificial sub­strate samples. Following identification andenumeration, all nymphs were divided intoO.l-mm size classes, based on maximum headcapsule width, and dried at 60 C for 48 hours.Mean annual population biomass was deter­mined by summing biomass estimates for allsize classes across all sampling dates. Meanannual population density was determinedafter summing the number of individuals ineach size class at each site. Head-capsulemeasurements were also used to constructsize-frequency plots for life-history determi­nations, including the number of annualgenerations produced. Complete life-historyinformation for these species at each site canbe found in Rader and Ward (1989).

Female body size (dry weight biomass) oflate instars was estimated by taking the meanof the three largest size classes collected fromeach site. Late mayfly instal'S have a full com­plement of mature eggs.

RESULTS

Temperature, Growth, and Body Size

A progressive increase in accumulated de­gree days and mean annual temperatures wasfound from site 1 to site 3 (Table 1). Site 1 wascharacterized by rapid seasonal changes intemperature and a long, freezing (0 C) winter(Table 1, Fig. 1). In contrast, site 2 demon­strated gradual seasonal changes combinedwith winter warm and summer cool tempera­tures. Site 3 was intermediate with respect torates of seasonal change but bad the greatestamount of tllermal energy (largest number ofaccumulated degree days. largest maximumtemperature, and largest annnal mean tem­perature).

100 R. B. RADER AND J. V. WARD [Volume 50

Winter

Winter

IIIII

• ., I

, I I ....

,lit" I" •, ,I' I ,

"'I.,

Summer - Autumn

IIIIII

ISummer -" Autumn... "•

" , ""'1'\"I'" II. ,

t \, I I-,.,' . I ',', '. . ,I,,·• •,

" .I'· •I' I'.,ff,

Site 1

Site 2

Spring - Summer

, '•

8

2

6

6

2

8

14

12

1

IIIIII

: .1•" ', ..'.I ''. ",.,

I ~ ~,r

°1-rr~"""""""'I-rnTTf,.,.;.,,-,+,...........-h::~~,...,....I~""';'T'M'"....-j-. ...........-n-......-to-."::;"'h-MAY JUN JUL AUG SEPT - OCT NOV DEC JAN FEB MAR APR MAY

14

10

-u!- 12W0:::>~0:WCL:,;WI-0:W

~~

Winter

..

,,

Spring - Summer

6

': "summer - I• "of' Autumn'. ," I,I'""" .

• , • II' • 01\"\ ,t. \. .IV -J

.... 'I' I~.. 'I \' '" •

.'. • I."I '" ... \ II ' \.', I

f "...I ""il.I ' "",'1"',. ,,'. ,I I ':\' ~"OI to ,:" ,. .. ,I ' ...,....... .,'

~+~:"h'=+=':rt-","c.-I'forr"r."h+:'+=:o::t-'"!:':'-!'"1-':!:",k'=+I~'~~"\rr°FEB MAR ' APR' MAY' JUN JUL AUG' SEPT OCT' NOV' I DEC' JAN 'FEB

2

8

10

12

o""""....r,........+.,.....Tt-r.,..j.,-rtr=.f,:,~+....""""l*.,.,.,I-M-o.....-t~mTTT'~,.,.,...-rj-o.,.,.._14 MAY JUN JUL AUG SEPT' OCT 'NOV" DEC.' JAN FEB' MAR' APR MAY

Site 3

Fig. 1. Temperature profiles for each site during 1981-1982. jndividual points represent daily means. (See text forexplanation ofcriteria used to determine seasonal separations. Note that the x-axes have been adjusted to facilitate acomparison of the length ofeach period at each site.)

1990] MAYFLY GROWTH AND TEMPERATURE Co 'STANCY 101

TABLE 1. Temperature chard.cteristics ofeach site.

Seasonal growth rates were analyzed to ex­plain among-site differences in body size foreach species. The seasonal pattern of growthfor D. grandis and E. infrequens at site 1 wascharacterized by slow growth during summer­autumn, no growth dUling winter, and a rapidincrease during the spring-summer period(Figs. 2, 3). In contrast, growth at site 2 wascomparatively fast and continuous througheach seasonal period including winter. Atsite 3, D. grandis demonstrated a continuousgrowth pattern similar to that observed atsite 2, whereas E. infrequens demonstrated aspring-summer pulsed pattern of growthmore similar to that at site 1.

The seasonal growth rate (0.0108 mg/day)for the early instars of E. infrequens dw'ingthe summer-autumn period was significantly(P = .0001) faster at site 2 than at sites 1 and 3(0.004 and 0.003 mg/day, respectively), whichwere not significantly different (Fig. 2). Thistrend continued into the winter period whenthe growth rate at site 2 (0.022 mg/day) wasagain significantly (P = .0001) faster than atsites 1 and 3, where growth rate was not differ­ent from zero. By the end of Winter, the site 2population had completed over 76% of itsgrowth, and larvae were over five times largerthan those at sites 1 and 3 (Fig. 2). Therefore,the larger body sizes at site 2 can be attributedto rapid growth starting at egg hatch and con­tinuing through the winter. During thespring-summer growth period, individuals atsites 1 and 3 grew significantly (P = .0001)faster (0.034 and 0.033 mg/day, respectively)than individuals at site 2 (0.014 mg/day).However, the body sizes oflate instar larvae atsites 1 and 3 were still considerably smallerthan those at site 2.

The growth rate (0.064 mg/day) of early

Site 1 Site 2 Site 3Characteristic

Annual degree days (C)Mean annual temperature (C)C.V. (%)Minimum (C)Maximum (C)Days <3CD3YS=OCRate ofspring-summer i ("Clda}')Length ofspring-summer (days)Rate ofsummer-autumn! eC/day)Length ofsummer-autumn (days)

11303.61190.0

18.01911530.1499

0.2070

1729 20824.7 5.753 641.8 0.09.8 18.2155 101o 17

0.05 0.0694 177

0.05 0.10115 87

instars of D. grandis during the summer­autumn period was significantly differentamong sites (P = .0001), being greatest at site2 and slowest at site 1 (0.035 mg/day). Wintergrowth was also faster at site 2, with an aver­age rate of 0.064 mg/day, followed by site 3(0.044 mg/day) and then site 1 (0.015 mg/day).Winter growth at site 1 was not significantlydifferent from zero. Spring-summer growth atsite 1, however, was significantly (P = .0001)faster than at sites 2 or 3, with the fastestseasonal rate of increase for this study (0.119mg/day). Spring-summer growth rates atsites 2 and 3 (0.076 and 0.055 mg/day, respec­tively) were not significantly different.

Baetis tricaudatu8 was univoltine at site 1,but bivoltine with slow and fast seasonal gen­erations at sites 2 and 3. In contrast, D. gran­dis and E. infrequens had univoltine, slowseasonal life cycles at each site. Complete life­history data for each species can be found inRader and Ward (1989).

As predicted by the first hypothesis, meanannual density and biomass of each specieswere much greater at the warmer sites belowthe dam than at site 1 (Table 2). Populationdensity and biomass of Baetis t,'icaudatus andE. infrequens were largest at site 2, followedby sites 3 and 1. Maximum density andbiomass of D. grandis were greatest at site 3,followed by sites 2 and 1.

Contrary to the predictions of the secondhypothesis, dry weights of the largest instars,and population sizes of each species, did notcorrespond when' ranked by sites (Table 2).Although the largest late instars of E. infre­quens occurred at site 2, where the densitywas greatest, the largest instars ofB. l7icauda­tus and D. grandis did not Occur in the largestpopulation. Body size and population size cor­responded in only two other instances; thesmallest late instars of B. tricaudattls and theintermediate-sized late instars of D. grandisoccurred in the small- and intermediate-sizedpopulations, respectively, at site 1.

The largest B. tricaudattls larvae occurredat site 3, whereas its maximum populationdensity and biomass occurred at site 2. LateB. tricaudatus instars at sites 2 and 3 wereover two times larger than late instars at site 1(Table 2). The largest-sized larvae of D. gm1V­dis occurred at site 2. Site 3, which had thelargest population of D. grandis, had thesmallest late instars (Table 2).

102 R. B. RADERANDJ. V. WARD [Volume 50

1.5

SUMMER-AUTUMN

1.0

•••

280

•

250220

.8..• lI. _-

._-1----...:-- ---

190

0.00 30 60 90 120 150

4.0

WINTER 0

3.0 •

•'g

2.0 '0•

8 • 80

1.0 •

0.0160

0.5

(f)(f)«2-l«::Jo>-oz-zL52

6.0

5.0

4.0

SPRING-SUMMER

8·o

.....• • • . . • Ie

• • •• • • 8 ....

o• • •• •

o·0

o

3.0

2.0

1.0

----f __-I__- .J--­- --. _-1 •

0i---•

370310 340

JULIAN DAYS

O.O+-------+-------+--------j280

Fig. 2. Seas(mru growth rate patterns for E1)lIemerella infrequens. Site 1 is rcprc:scnted hy a solid !incand diamonds,sites 2 and 3 by (l dotted line and circles, and a dashed line and triangles, respectively. Each symbol (dialllond, circle,and triangle) represents the mean biomass for a single sample.

1990J MAYFLY GROWTH AND TEMPERATURE CONSTANCY 103

Fig. 3. Seasonal growth rate patterns for Dronella grandis. (For further description, see Fig. 2 legend.)

104 R B. RADER A OJ. V. WARD [Volume 50

TABLE 2. Mean annual population density (#'s m-2) and biomass (mg m-2). plus mean indjvidual size (mg drywt.) for!ale instal'S of each species. Values in parentheses for the population parameters are the percentage of meanrepresented by the standard error. Values in parentheses for size estimates indicate the number of individuals used todetermine each mean.

Species

B. tricaudatfJS

E. infrequeTls

Site 1 Site 2 Site 3

Population size Body size Population size Body size Population size Body sizeDensity Biomass Density Biomass Density Biomass

754 142.6 0.730 7720 1501.7 0.550 4018 1162.8 1.960(21%) (17%) (31) (18%) (21%) (35) (18%) (27%) (35)

367 502.6 3.500 1034 4964.7 4.1>30 644 1188.5 2.800(25%) (23%) (23) (23%) (25%) (35) (24%) (26%) (25)

102 1585.9 14.930 80 2368.4 18.110 278 3355.4 14.220(31%) (18%) fiG' (10%) (22%) (9) (14%) (20%) (17), ,

DISCUSSION

Temperature, Growth, and Body Size

The winter warm and summer cool condi­tions ofsite 2 allowed rapid continuous growthof E. infrequens and D. grandis, which pro­duced larger instars and greater fecundity.Site-specific explanations of growth pallernsand body-size differences for D. grandis andE. infrequens are consistent with the TE hy­pothesis. Vannote and Sweeney (1980) pro­posed that the seasonal pallern of growth foraquatic insects, as for other small ectotherms(e.g., Phillipson 1981), may be detennined bythe interaction between temperature andbody size. Smaller instars, which have a largesurface-to-volume ratio, will have a highermetabolic rate than larger instars at the sametemperature. At site 1 the smallest D. grandisand E. infrequens instars appeared during thewarmest months of the year (July, August,and September). High summer-autumn tem­peratures coincident with small instars atsite 1 likely resulted in large metabolic costsand, therefore, slow growth rates and possiblyhigh mortality rates. Both species ceased togrow during the freezing winter tempera­tures. In the spring, winter survivors experi­enced a rapid increase in temperature andthus a relatively short period (99 days or less)to complete growth and maturation. All elsebeing equal, the magnitude and length ofsummer-autumn temperatures when coinci­dent with early instars, plus the rate of vernalrise, limit growth and body Size/fecundity andprobahly have an important influence on thegeographic distribution and upper altitudinallimits of aquatic insects. This may be espe-

cially applicable to cool-adapted boreal spe­cies (see Edmunds 1982).

Early instars at site 2 began growth in muchcooler summer-autumn temperatures; meta­bolic costs were low and growth rates fastcompared to those at sites 1 and 3. Wintertemperatures, which varied slightly around2 C, were not sufficiently cold to inhibitgrowth, which continued at a rapid pace.Growth appeared near completion before thevernal rise in temperature, thus leavingplenty of time for maturation and emergence.The rapid completion of growth probably re­sulted in the extended emergence of E. infre­quellS and the addition ofa second generationof B. trica'uWtus at site 2.

The early instars of D. grandis Uuly) and E.infrequens (August and September) begangrowth at site 3 during tbe warmest months ofthe year. August and September were, onaverage, 7-8 C warmer at site 3 than site 2.Although E. infrequens did not grow, D.grandis early instars grew rapidly during thesummer-autumn growth period. Because D.grandis early instars were approximately twotimes larger than E. infrequens early instars,they probably had lower metabolic costs. Thisallowed them to grow at the warm summer­autumn temperatures. Although both speciesstopped growing during winter at site 1 butcontinued to grow during winter at site 2, onlyD. grandis continued winter growth at site 3.The fact tbat D. grandis was over three timeslarger than E. infrequens at the beginning ofwinter may explain its ability to grow in winterat site 3 in contrast to E. infrequens.

These site-specific explanations of growthare consistent with the TE hypothesis

1990] MAYFLY GRO\'TH AND 'TEMPERATURE CONSTANCY 105

suggesting that growth rate and consequentlybody size and fecundity are determined bythe length of time individuals are exposed to aspecific optimal range oftemperatures. Otherfactors, however, that may also influencemayfly growth rates (e.g., food abundane",;Sweeney et al. 1986) were altered by the ef­fects of stream regulation. For example, con­stant flow conditions and the addition ofplanktonic diatoms from the reservoir en­hanced food quality and quantity at sites 2 and3 (Rader and Ward 1989).

If summer temperatures increase meta­bolic costs, causing growth to slow or stop,then the rate of vernal rise and autumn de­cline determines the amount oftime individu­als are exposed to optimal temperatures and,therefore, the amount of time available forgrowth. Growth of both D. grandis and E.infrequens continued as long as temperaturesremained between 2 C and 10 C. However,when temperatures exceeded this range,growth slowed or stopped. Where growth wascontinuous (site 2), temperatures were alwayswithin this range. The optimal temperaturerange for these two species appears to lie be­tween 2 C and 10 C.

Population Size and Body Size

''''inter warm and summer cool conditionsat site 2 and the long spring-summer period atsite 3 probably accounted for the multivoltinelife cycle ofB. tricaudatus at these sites. Stan­ley and Short (1988) suggested that populationsize may remain unaltered or even increase inwarmer than optimal conditions if fastergrowth rates and shorter generation timescompensate for smaller body sizes and lowerfecundity. For some aquatic insects, warmerthan optimal temperatures may offer a trade­off between body size/fecundity and genera­tion time. Will they maximize reproductiveeffort by producing fewer, larger individuals(slow growtb and a univoltine life cycle) ormany, smaller individuals (fast growth andmultivoltine life cycle)? These data demon­strated that 8aetis may have the genetic plas­ticity necessary to respond to such tradeoffs.As temperatures approached optimality(site 2) from cooler conditions (site 1), bothvoltinism and body size increased within thesame population. When comparing site 1 withthe much warmer and very different tempera­ture regimes downstream from the reservoir,

we found that our data support the predictionsof the TE hypothesis. The largest body sizeand population size of each species occurreddownstream from the reservoir. However,site-specific comparisons of body size andpopulation size did not correspond as pre­dicted by the TE hypothesis. The thermalregimes at sites 2 and 3 were apparently suffi­ciently similar that temperature did not havean overriding influence on population sizes.The TE hypothesis assumes that higher fecun­dity is equivalent to larger population size.Sources of mortality at every stage ofdevelop­ment, eggs, nymphs, and adults (Sweeneyand Vannote 1982, Butler 1984, Peckarsky1984, Gilliam et al. 1989), which may varyacross sites, may interrupt the translation ofhigher fecundity into higher population den­sity and biomass. Numerous !actors, in addi­tion to the influence of temperature on bodysize/fecundity, will undoubtedly influencethe geographic or altitudinal variation inaquatic insect population size and stability.Toward the center, and probably over most ofa species range, other sources ofmortality anddeterminants of reproductive success shouldhave a greater influence on aquatic insect pop­ulation size. Temperature at the edge of aspecies range may, however, be more limit­ing, compared to other factors, in determin­ing the extent of a species distribution.

ACKNOWLEDGMENTS

J V. McArthur made helpful comments onan earlier draft ofthis paper. Data analysis andmanuscript preparation were supported bycontract DE-AC09-76SROO-819 between theU.S. Department of Energy and the Univer­sity ofGeorgia. Data collection was supportedby a research grant to J. V. Ward, ColoradoState University, from the Colorado Experi­ment Station.

REFERENCES

BUTLEf;, M. G, 1984. Life histories of aquatic insects.Pages 24-55 in V. H. Resh and D. M. Rosenburg,eds" The ecology ofaquatic insects. Praeger Pub~lishers, New York.

CUFFORD, H. F .• AND H. BOERCER. 1974. Fecundity ofmayflies (Ephemeroptera), with species referenceto mayflies of a brown-water stream of Alberta,Canada. Canadian Entomologist 106: 1111-1119.

106 R. B. RADERANDJ. V. WARD [Volume 50

CONNELL,]. H.. ANO W. P. SOGSA. 1983. On the evidem:eneeded to judge eeo}ogl(:al stability or persistence.American Natu......tJist 121: 789-824.

EDMUNDS. G. F. 1982, Historical and life history factors inthe biogeography of mayflies. American Zoologist22: 371--374.

CI.L1JAM, 1- F., D. F. FRASER, .\ND A. M. SABAT. 1989.Strong effects of foraging minnows on a streambenthic community. Ecology 70: 445-4,52.

KONl1Jl.ATIEFP, B, C .. ,\NO J. R. VOSHELl.., JR. 1980. Lifehistory and ecoJo~y of Stenonerna modestum(Banks) (Ephemeroptera: HeptagenHdae) in Vir­ginia, USA. Aqulltic Insects 2: 177-189,

PECICARSKY, B. 1... 1984. Predator-prey interactions amongaquatic insects. Pages 196-254 ill V. H. Resh QndD. M. Rosenburg, eds., The ecology of aquaticinsects. Praeger Publishers. New York.

PfllllIPSON.J. 1981. BiocnergeLicoptions and phylogeny.Pages 20-45 in C. R. Townsend and P. Calow,eds., Physiological ecology. Blaclrn-ell ScienLificPublishers. Sunderland, Massachusetts.

RADER, R B., AND J. V. WAnD. 1988. Influence of regula·tion on envi,roomental conditions and themacroinvertebrate t'OUlmunity in the upper Colo·rado River. Regulated Rjvers 2: 597-618.

__. 1989. Influence ofimpoundments on may(}y diets,life hi'itorics, and production. Journal of the NorthAmerican Benthologica! Society 8: 64-73.

STANUY, E. R, AND R A. SHORT. 1988. Temperature ef·feds on wurmwatcr stream insects: a test of thethermal equilibrlmn hypothesis, Oikos 52:313-320.

SWgENEY, B. W. 1984. Factors influencing life·historypatterns ofaquatic insects. Pages 56-100 in V. H.Resh and D. M. Rosenburg. cds., The ecology of

aquatic insects. Pmeger Publishers. New York.SWEENEY, B. W.. ANI> H. L. VANNOTE. 1978. Sire variation

and the distrihution of hemimetabolous aquaticinsects: two th(:rmal equilibrium hypotheses. Sci·ence 200: 444-446.

__ .1981. Ephemerella mayflies of White Clay Creek:bioeu(.·rgetic and ecological relationship!' amongsix (..'ocxisting .~pccies. Ecology 62: 13S3-I369.

__ . 1982. Population synchrony in mayflies: a preda­tor satiation hypothesis, Evolution 36: 810-821.

SWEENl':V, B. W., H. L. VANNOTE, AND P. 1. DODDS. 1986.Effects oftcmperatnre and food quality on growthand development ofa mayfly. Lept(",lllebia inter­media. Canadian Journal of Fisheries and AquaticSciences 43: 12-18.

VANNOTI:, R. L., AI'D B. W. SWEENEY. 1980. Geog.....phicanalysiS of tllermal e<luilibria: a conceptual modelfor evaluating the effect of natural and modifiedthemlal regimes on aquatic insect communities.American Naturalist 115: 667-695.

WARD, J. V. 1980. Abundance and altitudinal distributionof Ephemeroptera in a Rocky Mountain stream.Pages 169-] 77 in J. F. Flannagan and K. E. Mar·shaH, <XIs .. Advances in Ephemcrol!tera biology.Plenum Publishers, New York.

__ . 1986. Altitudinal 7..Onation in a Rocky Mountainstream. Archiv fur Hydrohiologie, SupplementBand 74: 133-J99.

WARD, J. V" AND]. A. STANFORD. 1982. Thermal responsein the evolutionary ecology of aquatic insects. An­nual Review of Entomology 27; 97-117.

Received 1 May 1990Accepted 15June 1990