J. N. Am. BenthoL Soc, 2003, 22(3)388-400O 2003 by The North American Benthological Society

Reduced detrital resources limit Pycnopsyche gentilis (Trichoptera:Limnephilidae) production and growth j1

!jSUSAN L. EGGERT1 ' . . _ t

' • • ' ' • ' • • " ' • • *rInstitute of Ecology, University of Georgia, Athens, Georgia 30602 USA •!

J. BRUCE WALLACE [•

Department of Entomology and Institute of Ecology, University of Georgia, Athens, Georgia 30602 USA

Abstract. Leaf inputs in temperate forest streams may limit caddisfly production because leafdetritus serves both as a food and case-material resource, V\fe estimated Pycriopsycftegenrilis productionin a stream experimentally decoupled from its riparian habitat and a reference stream for 8 y in thesouthern Appalachians. VVe also examined laboratory survivorship; growth, and case-building activ-ities of P. gentilis in substrate containing various quantities of leaf material Pycnopsyche gentilis pro-duction declined to 0 within 3 y of the start of litter exclusion. Abundance, biorhass, and productionof P. gentilis were positively related to leaf litter standing crops. Maximum individual length, of P.gentilis was reduced when annual leaf standing crops fell below 25 to 50 g AFDM/m2. Observationsof case construction for instars removed from their original leaf cases and kept in substrate with lew-leaf standing crop, showed mat P. gentilis was capable of rebuilding a case of available substrate andsurviving for 3 to 4 wk before dying of starvation. Survivorship and growth were significantly greaterfor larvae reared at high and intermediate leaf standing oops, than at low leaf standing crop. Olderinstars had higher survivorship rates but lower growth rates than younger instars in the low littersubstrates. Survivorship and growth rates were lower for some individuals farced to rebuild newcases, indicating an energetic cost associated with case-building activities. Our results demonstratethat the linkage between terrestrially derived organic matter and production of a caddisfly shredderwas a consequence of food availability.

Keywords: case-building behavior, caddisfly, organic matter, iriacroinvertebnttes, stream, resourcelimitation.

Detrital resources and benthic invertebrates in Many caddisfly genera are restricted to par-streams are closely linked (eg., Egglishaw 1964, ticular terrestrial biomes (Ross 1963). Larvae ofAnderson and Sedell 1979, Cummins etaL 1989, Pycnopsyche gentilis (MacLachlan) inhabit leafRichardson 1991, Dobson and Hildrew 1992, packs in temperate forest streams throughout-Wallace et aL 1999). Organic matter inputs from me eastern United States (Flint I960, Ross 1963,riparian habitats provide ~90% of the energy Mackay 1972, Mackay and Kalff 1973). Leaf lit-source of southern Appalachian headwater ter is a dominant part of their diet (Mackay andstreams (Webster et aL 1983). Rapid develop- Ka^ 1973). In headwater streams at Coweeta,ment in this region has hastened the decoupling && to 100% of their diet consists of leaf detritusof riparian habitats from streams and rivers. Pri- (Hutchens et aL 1997). The type of available or-vately owned land in the region has <60% for- i3™0 material influences habitat selection byest cover in the 30-m riparian zone (SAMAB Pycnopsyche spp. (Cummins 1964, Mackay and1996). The extent of riparian loss worldwide is **& 1973>- Second-through 4m-ihstar P. gentilisnot well known (Nilsson and Berggren 2000, construct 3-sided leaf disk cases (Mackay 1972).NRC 2002), although it ha^ been suggested that Vfb^f larval d?nsitlfs «? **&, 4th instars may>80% of riparian Ireas in North America and £?** ̂ J^ al jubstrates preinaturely

). The a b t y to successfully con-

to streams puts the sur- t^on from predators (Otto and s4sson 1980,vrval of sensitve stream mvertebrates such as Njslow md ̂ ^ m3f otto and Johanssonf

caddisflies at qak (Morse etaL 1993). 1995), provide ballast (Webster and Webster; 1943), and prevent desiccation (Zarnora-Munoz

1 E-mail address: eggerts@sparc.ecology.uga.edu and Svensson 1996).

388

2003] DETRITUS UMTTATION OF CADDBFLY PRODUCTION 389

The experimental exclusion of participate or-ganic matter from a forested headwater streamat Coweeta Hydrologic Laboratory, North Car-olina, reduced benthic leaf standing crops by •94%. Results from the first 4 y of litter exclusionsuggested that all shredder taxa did not re-spond similarly to detrital resource reductionduring the early years of the 8-y study (Wallaceet aL 1999). Here we report P. gentilis productionand resource supply for the entire 8-y study,which included an additional year of data fol-lowing small wood removal and 2 y of data af-ter large wood removal Another objective ofour study was to determine mechanisms re-sponsible for changes in P. gentilis productionthat occurred as a result of litter exclusion. Lowsurvivorship of this species could be the resultof a lack of available food resources, higher pre-dation rates, or respiratory stress caused by theabsence of leaf litter as a material for case con-struction for 2nd to 4th instars. In the laboratory,we manipulated levels of leaf standing crop(low, intermediate, and high) and larval casetype (field constructed case vs laboratory con-structed case) to examine the role of leaf detri-tus as a food resource and as case-constructionmaterial for P. gentilis. Specifically, we observedcase-building behavior to determine whetheryoung P. gentilis larvae could adapt to the lackof leaf material and survive in cases constructedfrom substrate containing little leaf material. Inaddition, we measured differences in growthrates and mortality of larvae with normal leafcases reared with various amounts of leaf litterto see whether differences in food availabilitycould account for the decline in P. gentilis pro-duction in the fiejd.

TABLE 1. Physical characteristics of reference andlitter exclusion streams at Coweeta Hydrologic Labo-ratory, North Carolina.

Methods

Study sites

Organic matter was excluded from Catchment(C) 55, at Coweeta Hydrologic Laboratory inMacon C6., North Carolina, USA, beginning inAugust 1§93 using 1.2-cm mesh gillnet canopyto exducre direct litter fall and lateral fencing toprevent blow-irr from the banks. The canopy be-gan at th£ top of the catchment and was openalong the| sides for aerial insect colonization.Small wcjody debris (<10 on diameter) Was re-moved by hand from the litter exclusion streamin August 1996. Remaining large wood was re-

Variable

CatchmentArea (ha)Elevation (m asl)Aspect

Channel length (m)Wetted width (m)Bankful channel area (m2)Gradient (cm/m)Canopy cover (%)Substrate (% composition)

MixedBedrock outcrop

Discharge (1985-2000, L/s)! Average

• Maximum . -Water temperature (1992-2000, °Q

Annual averageAnnual degree daysMaximumMinimum

Refer-ence

5.2829

South135

0.7-1.232727

88.9

7327

0,7332.13

12.043632070.7

Litterexclusion

75810

South170

1.2-1.637320

917

8713

13571.45

12.0436320.10.6

moved in August 1998. An unmanipulatedstream, C53, located within the same basin,served as a reference stream. Both catchmentsare similar with respect to aspect, elevation, dis-charge, water chemistry, and water temperature(Table 1).

Production and organic matter standing crop

Abundance and biomass of P. gentilis formixed-substrate habitats (substratum consistingof mixed silt, sand, pebble, and cobble) was es-timated in each stream using a 400-cm2 corer.Four random samples were collected monthlyfrom September 1992 to August 2000 from eachstream. Organisms and organic matter wereseparated into a >l-mm fraction and <l-mmfraction using nested sieves. Pycnopsyche gentilisfrom both fractions were sorted from organicmatter under 15X magnification and preservedin 7% formalin; Animals in the <l-mm fractionwere subsampled if necessary (Waters 1969). Allindividuals were identified, counted, and mea-sured to the nearest mm using a dissecting mi-croscope and graduated stage. Biomass was cal-culated using length-mass regressions (Benke et

390 S. L. EGGERT AND J. B. WALLACE [Volume 22

aL 1999). Secondary production estimates for P.gentilis were calculated using the size-frequencymethod (Hamilton 1969) with a cohort produc-tion interval correction of 275 (Benke 1979, Hu-ryn and Wallace 1988). Mean annual length andannual maximum length of P. gentHis were com-pared between streams over the 8-y study todetermine whether litter exclusion influenced P.gentilis growth and survivorship to emergenceCoarse particulate organic matter in each sam-ple was separated into categories (Le., leaf,wood, seeds, roots), dried at 60°C, weighed,ashed at 500°C, and reweighed to obtain ash-free dry mass (AFDM).

Laboratory experiment

Survivorship and growth of P. gentilis rearedwith high, intermediate, and low amounts (seebelow) of leaf litter were measured in the labo-ratory from December 1999 to February 2000 for1st and 3rd instars. Organisms were collectedfrom lower reaches of C53. First instars grew tothe 3rd-instat stage between December and Jan-uary. Newly collected 3rd-instar individualsgrew to the 4th-instar stage from January toFebruary. For each size class, 50 individuals (25in original field cases and 25 removed from fieldcases) were incubated individually in tea infus-ers (Toby TeaBoy Ltd., Aldridge, England) for 4to 6 wk. Toby Teaboys, 56 X 43-mm plastic bar-rels lined with 244-jun mesh netting, have beenused to measure insect growth rates in otherstudies (Benke and Jacobi 1986, Rosemond et aL2001). The bottoms of tea infusers for each treat-ment were filled with substrate (silt, sand, andpebbles) from the reference or litter exclusionstream. Naturally conditioned leaves of eachbreakdown-rate category (see following) fromthe reference stream were added to substrate inthe tea infusers according to mean proportionsof each breakdown category in previously mea-sured leaf litter inputs to C53 and C55 (fast [eg.,dogwood] = 3%, medium [eg., red maple] =30%, slow [e.g.̂ white oak] = 47%, very slow[e.g., rhododendron] = 20%; Cuffney et aL1990). Our high leaf treatment (268 g AFDM/m2) simulated mean leaf standing crop naturallyfound in the reference stream from December toFebruary. The low leaf treatment (0.4 g AFDM/m2) represented' the actual leaf standing crop inthe litter exclusion stream at the time our ex-periment was run. An intermediate leaf treat-

ment c of 13 g AFDM/m2 was also included,which simulated the amount of leaf litter pres-ent during the winter months of yeai 1 in thelitter exclusion stream, when the decline in P.gentilis production was first observed.!:

Teaboys were randomly placed in 32-L tubsof aerated stream water, which was 'replacedweekly. Water temperature was maintained at12.5°C Timers controlled the lightdark cycle at12 h Iighfcl2 h dark. Additional conditionedleaves collected from the reference stream wereadded to each teabby as needed to maintainconstant leaf standing crops in each treatmentAt the beginning of the experiment, individualsremoved from their field cases were observedon a daily basis, to record case-building activi-ties. Case-building behavior of additional indi-viduals was videotaped using a video cameraequipped with a macro lens. Survivorship wasrecorded weekly. Initial mass was estimated bymeasuring head capsule widths (HCW) at thestart of the experiment using an ocular microm-eter. Initial wet masses were not measured be-cause of the reluctance of larvae to voluntarilyreturn to their cases (Le., head capsule widthscould be measured while larvae were in theircases). HCWs were converted to initial AFDMusing a HCW-AFDM regression equation de-veloped from individuals collected in the ref-erence stream: AFDM (g) = 0.0008128x HCW(mm)- 0.000203 (r2 = 0.67, p < 0.001, n = 35).AFDM was measured for each individual at theend of the experiment Individual instantaneousgrowth rates (KR) were calculated using theequation:

IGR = (In W, - In WJ/t

where W, is the individual AFDM (mg) at timet (d), and W0 is the initial estimate of individualAFDM (mg).

Differences in survivorship between organicmatter levels and case treatments were deter-mined using 2-way repeated measures ANOVAof arcsine-transformed % survival values (SigmaStat, version 2.0, SPSS, Chicago). Differences inindividual growth rates between organic matterlevel and case treatment were examined for eachsize class using 2-way ANCOVA (JMP, version4.0.4, SAS, Gary, North Carolina) with initialHCW as the covariate because smaller larvaetend to grow faster than larger larvae (Perry etaL 1987). Differences in mean survivorship and

2003] DETRITUS LIMITATION OF CADDISFLY PRODUCTION 391

350n

0.30-1

m

ReferenceLitter exclusion

o.oo

PreTmt

Yeari

il Litter Litter exclusion Litter exclusion- ii; exclusion + small wood + large wood

•• •? - ' ^ - ' 3 ; i -"• '"• ; '.;'•'• • • . . • • ' • — - ' ' ' • • . - • • ' • ' • removal ' • .removal- ^Fia l.;'iMean annual abundance (A), biomass (B), and production (Q of PycmpsychegentUis in mixed sub-

strates ofjeference and litter exclusion streams from 1992 to 2000. Arrows indicate start of treatment periods.Pretmt ^iipretreatment year Years 1,2, and 3 = first 3 y of litter exclusion in Catchment (Q 55. Years 4 and 5

= small wood removal with ongoing litter exclusion in C55. Years 6 and 7 = large wood removal wilh ongoing'':' inC55. ' • - • • ' - . - • ! ' - . • - . • . ' . , . - . - . • '

392 S. L. EGGERT AND J. B. WALLACE [Volume 22

ReferenceLitter exclusion

Litterexclusion

Litter exclusion+ small wood

removal

Litter exclusion+ large wood

removalFIG. 2. Mean monthly leaf standing crop in mixed substrate of reference and litter exdusion streams from

1992 to 2000. Dashed line indicates 25 g AFDM/m2. Years as explained in Fig. 1.

growth rates between treatments were com-pared using Tukey's multiple comparison tests.

Results

Production and leaf standing crop .

At the start of die study, P. gentilis abundancein mixed substrate was similar between streams(Fig. 1A). A greater proportion of large instarsin the reference stream compared to the litterexclusion stream accounted for higher biomassand production in the reference stream duringthe pretreatment year (Fig. IB, Q. Production ofP. gentilis in the litter exdusion stream declinedto <1% of the pretreatment level by the end ofthe 2nd year of litter exdusion, and was reducedto 0 within 3 y (Fig. 1C). In the reference stream,P. gentilis production ranged from 032 to 0.98 gnr2 jr1 over thfe 8-y period. Leaf standing cropsin the mixed substrate of the reference streamincreased each autumn and declined during;spring and summer (Fig. 2). Five months after

the start of litter exdusion, leaf standing crop inthe treatment stream declined to <25 g AFDM/m2 and remained at least that low for the next7y.

Regression analyses suggested that mean an-nual leaf standing crop was an important de-terminant of P. gentilis production. C55 datafrom 1985 and 1986 (Lugthart and Wallace 1992)and 1989 and 1990 (Whiles and Wallace 1995)were induded with C55 data collected in ourstudy and strong positive relationships werefound between leaf standing crop and mean an-nual abundance, biomass, and production of P.gentilis. The data induded leaf standing cropsthat ranged over 2 orders of magnitude (Fig. 3).Maximum individual length of P. gentilis wasdramatically reduced at mean annual leaf stand-ing crops below 25 to 50 g AFDM/m2 (Fig. 4).Maximum individual length of individuals col-lected in the reference stream and in the litterexdusion stream before treatment ranged from16 to 18 mm (Fig. 5A). After litter exdusion,

2003] DETRITUS LIMITATION OF CADDISFLY PRODUCTION 393

250 i

100 150 200 250

'' Leaf standing crop (g AFDM/m2)

Fie. 3. Long-term relationships between mean an-nual abundance (A), ash-free dry mass (AFDM) (B),and production (Q, and mean annual leaf standingcrop in Catchment 55 for 1985,1986,1989,1990, and1992 to 1995. n = 12 for each regression. Data pointslabeled 1-? designate litter exclusion years. Abun-dance - 6.864 + Q.831x (leaf standing crop), r2 = 0.77,p < O.OQl;jbiomass = 3.200 + 0.478x (leaf standingcrppfc rf =$0.86, p < Q.OOl; production = -0.0747 +3.345X (leafstanding crop), r1 = 0.85, p < 0.001.. . . , - , . , . . . . . . .

maximum individual length in the treatmentstream declined to 1-3 mm. The difference be-tween streams for mean annual individuallength was similar in both streams over the 8-ystudy (Fig. 5B).

Case building and survivorship

Pycnopsyche gentUis larvae reared in high leafsubstrate followed the typical progression ofcase-building activity reported by Mackay(1972) and Mackay and Kalff (1973). First in-stars constructed cases of fine sand grains be-fore converting to 3-sided leaf disk cases 1 to 2wk later. Videotapes showed that young P. gen-tUis used mostly flakes of pyrite, which theycarefully selected, fit in place, and attached withsilk The selection, fitting, and attachment pro-cess required ~5 min per particle. Larvae com-pletely rebuilt cases within 4 to 8 h. All 2nd to4th instars maintained their leaf cases until the5th instar, when they gradually switched tolarge sand grain cases. In contrast, all of the lar-vae reared in substrate lacking leaf litter initiallybuilt fine sand grain cases and did not switchto the leaf disk cases. These larvae continuedadding larger sand grains to the tops of cylin-drical cases, and they used bits of wood andseeds if these materials were present in the sub-strate

Survivorship of larvae kept in high leaf stand-ing crops ranged from 84 to 95% during theexperiment (Fig. 6). Only 28 to 32% of 1st in-stars survived through wk 6 in the intermediatelevel of leaf substrate (Fig. 6A). Ninety to 95%of 3rd instars were still alive through the end ofwk 4 in the intermediate leaf substrate (Fig. 6B).Results from repeated measures ANOVA indi-cated that survival of 1st instars was greatest forlarvae growing at high followed by intermediateand low leaf standing crops regardless of casetype (ANOVA, p < (X001, Tukey test, p < 0.05)(Table 2, Fig. 6A). For 3rd instars, the effect ofleaf standing crop depended on case type (Table2, Fig. 6B). More larvae with rebuilt cases sur-vived at high leaf standing crops relative to lar-vae with field cases (Tukey test, p = 0.02) (Fig.65). At low leaf standing crops, survivorshipover the 4-wk experiment was significantlygreater for larvae with field cases than thosewith rebuilt cases (Tukey test, p = 0.04) (Fig.6B).

394 S. L. EGGERT AND J. B. WALLACE [Volume 22

20-

* Referenceo Litter exclusion

100 150 200 250

Leaf standing crop (g AFDM/m2)

300

FIG. 4. Relationship between leaf standing crop and maximum individual body length (y = 17.642 (1 -0.97"); r2 = 0.96, p < 0.0001) for Pycnopsyche gentUis collected in mixed substrates of reference and litter exclusionstreams from 1992 to 2000. n = 16.

Larval growth

First instars grew to 3rd instars, and 3rd in-stars grew to 4th instars during the growth ex-periments. IGRs of 1st instars with field cases inthe high leaf treatment were 1J5X greater thanin the intermediate leaf treatment (Fig. 7, AN-COVA followed by Tukey test, p < 0.05). For 1stinstars forced to rebuild cases, there was a 2.4Xdifference in growth rates between high and in-termediate treatments (Fig. 7, Tukey test, p <0.05). Differences in IGR between high and in-termediate leaf standing crops for 3rd instar P.gentUis were 1.5X for organisms with field cases(Tukey test, p > 0.05) and 1.6X for those withrebuilt cases (Fig. 7, Tukey test, p > 0.05).Growth rates were significantly lower for 1stand 3rd instars in low leaf standing crops thanin intermediate or high leaf standing crops (Fig.7, Tukey test, p < 0.05). The laboratory data cor-roborated results observed from field data. Aswith die maximum individual length data fromthe field (Fig. 4), a decline in laboratory growthrates of P. gentttis occurred when leaf standingcrop declined,'to <25 to 50 g AFDM/m2 (Fig.7).

Discussion

Response to reduced leaf inputs

Production of other shredders (eg., Peltoper-lidae, Leuctra spp., Lepidostoma spp., Tipula spp.)in the litter exclusion stream declined followingthe start of the experiment, but not as quicklyas observed for P. gentUis (Wallace et aL 1999).Pycnopsyche gentUis production (0.255 g AFDMm~2 y-1) composed 14% of total shredder pro-duction in the litter exclusion stream before thestart of the manipulation. Production of P. gen-tUis averaged 0.011 g AFDM nr2 y1 during the7-y exclusion period, representing 2% of totalshredder production. Pycnopsyche gentUis pro-duction in the reference stream made up 21% oftotal shredder production (average productionover 8 y = 0.731 g AFDM nr2 jr1) and wassimilar to mat reported by Stout et aL (1993)and Stone and Wallace (1998) for other Ist-orderreference streams at Coweeta Hydrologic Lab-oratory.

Production of P. gentUis in the referencestream varied annually. Based on the relation-ship between 1? gentUis production and meanannual leaf standing crop, it appears that muchof the variation in production was a result of

2003] DETRITUS LIMITATION OF CADDEFLY PRODUCTION 395

? 20-, AE, 18-

1) 14-§ 12'- 10-E 8-

i r| J:2 o j

x-s 1°"C gE

o> 6 "

0 -

-i

_

ni i i

B

n* * * n1 n1

•̂i Reference1 1 Litter exclusion

n1 1 In.-* i iPreTmt 1 2

i l l1 1 13 4 5

A Year A

In 1i6

t

Ini7

Litter Litter exclusion Litter exclusionexclusion + small wood t large wood

removal removalFIG. 5. Annual maximum body length (A) and mean annual body length (B) for Pycnopsyche gentilis in

reference and litter exclusion streams from 1992 to 2000. Years as explained in Fig. 1.

annual differences in leaf standing crop. Annualleaf standing crop varies with the timing andfrequency of storms; thus, if is not constantfrom one, year to the next (Wallace et aL 1997).Mean, monthly leaf standing crop in the refer-ence sbsam was i>25gAFDM/m2 for 94 of 96ino?; in ̂ thfestudy/ Pycnopsyche: gentilis, larvaegrow coritinuously, from October to June, the pe-riod; when leaf standing oops are >25; gAEDM/m\ Leaf standing crop in the referencestream fell below 25 g AEPM/m2 most oftenduring the months of July> August; and Septem-ber, after;jleaves with fast and medium break-down raies had disappeared from the stream.Pycnapsy&ie gentilis larvae pupate in July andAugust, Ijnd, therefore, are. not affected by thelate-sumrner decline in leaf standing crop., Average length of P. gentilis in both streamsreflected! ̂ n abundance of small instars relativeto large iitetars, except for the pretreatment yearin each stream. The 94% reduction in maximumindividual lengths of exclusion stream larvae

that larvae in the detritus-limitedstream were not completing larval developmentSmall instars appeared in the litter exclusionstream each year because of recolonization byaerial adults, but did not survive to emergence

When leaf standing crops in the litter exclu-sion stream declined to <25 g AFDM/rfl2, P.gentilis did not switch to alternate food resourc-es such as wood, as did other shredder taxa inthe stream (Hall et aL 2000, Eggert 2003). Hutch-ens et aL (1997) reported that leaf material madeup 86 to 100% of the diet of P. gentUis in anotherCoweeta stream. Hutchens et aL (1997) found noevidence of P. gentilis switching to alternate foodresources. Mackay (1972) found that P. gentilis inWest Creek, Quebec, used only leaf material forcase materials, food, and habitat When offeredtwigs as food in the laboratory, P. gentilis pre-ferred leaf material (Mackay and Kalff 1973). Incontrast to the specialized behavior of P. gentilis,P. hiculenta used twigs readily as a food andcase-building resource (Mackay and Kalff 1973).

396 S. L. EGGERT AND J. B. WALLACE

High leaf-field case

[Volume 22

O High leaf-rebuilt case

A

A £ Intermediate leaf-field caseA' Intermediate leaf-rebuilt case• Low leaf-field caseQ Low leaf-rebuilt case

FIG. 6. Percent survival of Pycttopsyche gentttis 1st instars (A) and 3rd instars (B) reared with high (268 gAFDM/m2), intermediate (13 g AFDM/m2), and low (0.4 g AFDM/m2) levels of leaves. 'Tield case" larvae hadfield-constructed cases. "Rebuilt case" larvae were removed from field cases before start of experiment andforced to rebuild cases. ,

Response to low leaf standing crop in the laboratory

Mortality as a result of predation was not afactor in our laboratory experiments. Larvaesuccessfully rebuilt cases of inorganic materialsin the absence of leaves and survived for up to3 wk, which indicates that respiratory functionsin the atypical;; cases were not impaired to thepoint of causing death by respiratory stress.Otto and Svensson (1980) showed that inorganiccases of the caddisfiy Potamoptylax cingulatuswere energetically more expensive to build thanleaf-disk cases because leaf-disk cases have few-?er pieces to be assembled. A similar energetic

cost may have contributed to the higher mor-tality rates and slower growth rates observed inour study for larvae that were forced to rebuildcases of mineral material instead of the usualleaf material.

First instars suffered higher mortality rates atlow resource levels than did 3rd instars. Otto(1974) reported higher fat content (12-14%) in3rd through 5th instars of P. cingulatus than in1st instars (9%). hi addition, fat and energy con-tent of P. cingulatus larvae was related to thepresence of preferred food resources in thestream (Otto 1974). In our study, higher fat con-

2003] DETRITUS LIMITATION OF CADDBFLY PRODUCTION 397

TABLE 2. Results of 2-way repeated measures AN-OVAs for survivorship of 1st and 3rd instar Pycnop-syche gentilis reared at 3 levels of leaf standing crop(high, intermediate, and low) and 2 case types (field-constructed or laboratory-constructed) over the 4 to 6wk laboratory study in winter 1999 to 2000. * = p <0.05, *** = p < 0.001.

Source of variation df MS

1st instarTimeLeaf standing cropLeaf standing crop X

TimeCase typeCase type X TimeLeaf standing crop X

Case typeResidual

3rd instarTimeLeaf standing cropLeaf standing crop X

TimeCase typeCase type X TimeLeaf standing crop x

Case typeResidual

5 2945.62 96322 19-29**

10 499.41 33 0.115 29.7

2 49.9 2.0110 24.8

3 1740.62 3397.4 1559*

6 607.91 IS 0503 3.0

2 306.1 632.*6 492

tent accumulated by older instars may haveserved as an energy reserve for larvae rearedwith limited resources for food and case-build-ing activities.

IGRs of 3rd instars were lower than those of1st instars, regardless df resource level Mackay(1972) reported a similar pattern of growth forP. g^flis,in'West Creek. Our 3rd-instar AFDMgrowth rates (0.0100/d) were lower than drymass growth rates measured by Hutchens et al(1997) for 3rd-ihstar P. gentilis reared on birchleaves (O.pl74^03i3/d), but were similar tothose reared on a diet of white dak (0.0084-0.0233/dj Growth rates of 1st instars forced torebuild a:case were also lower than those left intheir original case Dudgeon (1987) also foundthat when Potycentropus flaoomaculatus larvaewere forced to rebuild nets, they lost more massthan undisturbed larvae. The difference in IGRfor field cjase vs rebuilt case organisms was .larg-er for 1st instars than for 3rd instars, again in-dicating that older larvae had more internal en-

ergy reserves available to them when resourceswere reduced.

Our experiment did not separate the use ofleaf litter as a food resource and case material,but our results showed that growth rates of 1st-and 3rd-instars of P. gentilis that were not re-quired to rebuild atypical cases were signifi-cantly lower at low leaf standing crops than atthe high and intermediate levels. This reductionin growth was most apparent for young instars.In addition, the 95 to 100% mortality for indi-viduals of both age classes that had field-builtcases suggests that food was insufficient at thelowest leaf standing crop treatment (litter exclu-sion stream) for P. gentilis to survive.

Linkages between P. gentilis and leaf inputs

Results of our study showed that production,survivorship, and growth of P. gentilis werelinked to leaf litter standing crop. Pycnopsychegentilis tracks its primary food resource closelyand does not switch to alternate energy sourcesin these streams. As a result, this species couldnot complete its life cycle in the stream lackingdetrital inputs, thus lowering the overall diver-sity of the ecosystem. Researchers have hypoth-esized that limnephilid caddisfly diversity is re-lated to resource availability (Hint I960, Ross1963, Wiggins and Mackay 1978, Mackay andWiggins 1979). Pycnopsyche is limited to temper-ate deciduous forests of eastern North America(Flint 1960, Ross 1963). In southern Appala-chians forested streams, P. gentilis is tightlylinked to terrestrial leaf inputs and may be morespecialized hi its resource use than other shred-ders (Cummins 1964, Mackay 1972, Mackay andKalff 1973).

Our field data suggest that at least 25 to 50 gAFDM/m2 df leaf standing crop are requiredduring larval development (October to June) forP. gentilis to survive and maintain its functionalrole in a stream. The energetic needs df othershredder taxa in these streams remain un-known, but additional shredder production willrequire even higher levels of organic matterstanding crops. Because of differences in litterbreakdown rates, the presence of fast, medium,and slow decaying leaf species all may be nec-essary to sustain production of obligate leafshredders like P. gentilis throughout the year (Pe-tersen and Cummins 1974, Webster and Ben-field 1986, Cummins et al. 1989, Stout et aL

398 S. L. EGGERT AND J. B. WALLACE [Volume 22

0.05 i

0.04-

§, 0.03o:O

I 0.02

0.01 -

0.00

• 1 st instar - field caseo 1st instar-rebuilt caseT 3rd instar-field casev 3rd instar-rebuilt case

if

100 150 200 250

Leaf standing crop (g AFDM/m )

FIG. 7. . Relationship between leaf standing crop and mean (±1SE) instantaneous growth rates (IGR) of Isl-and 3rd-instar Pyaiopsyche gentilis reared in substrate with high (268 g AFDM/m2), intermediate (13 g AFDM/m2), and low (0.4 g AFDM/m2) levels of leaves. 'Tield case" larvae had field-constructed cases (1st instar r2 =0.99, p = 0.03; 3rd instar r3 = 0.99, p = 0.03). "Rebuilt case" larvae were removed from field cases before startof experiment and forced to rebuild cases (1st instar r2 = 0.99, p = 0.01; 3rd instar r* = 0.99, p = 0.02).

1993, Grubbs and Cummins 1996). We suggestthat small stream restoration efforts consider thereplacement of riparian vegetation in a mannerthat maintains both leaf litter diversity and totalinputs to accommodate stream detritivores thatdepend on leaf litter as a food resource.

Acknowledgements

We thank the technicians and students fromthe Wallace, Meyer, and Webster labs for theirassistance with the 8-y litter exclusion study.Derrick Carter jhelped with the growth experi-ment Suggestions by Darold Batzer, StephenGdlladay, John Hutchens, Judy Meyer, Amy Ro-semond, John Richardson, and 2 anonymous re-viewers improved the paper. Wayne Swank, Bri-an Kloeppel, ahd the staff at Coweeta Hydro-logic Laboratory provided site support The Na-tional Science Foundation (Grants DEB-9207498and DEB-9629268) funded this research.

Literature Cited

ANDERSON, N. H., AND J. R. SEDELL. 1979. Detritusprocessing by macroinvertebrates in stream eco-systems. Annual Review of Entomology 24351-377.

BENKE, A. C1979. A modification of the Hynes meth-od for estimating secondary production with par-ticular significance for multivoltine populations.Limnology and Oceanography 24:168-171.

BENKE, A. C, A. D. HURYN, L. A. SMOCK, AND J. B.WALLACE 1999. Length-mass relationships forfreshwater macroinvertebrates in North Americawith particular reference to the southeasternLtaited States. Journal of the North American Ben-tholpgical Society 18308-343.

BENKE, A. C., AND D. L JACOBL 1986. Growth rates ofmayflies in a subtropical river and their implica-tions for secondary production. Journal of theNorth American Benthological Society 5:107-114.

CUFFNEY, T. E, J. B. WALLACE, AND G J. LUGTHAKT.1990. Experimental evidence quantifying the roleof benthic invertebrates in organic matter dynam-

2003] DETRITUS LIMITATION OF CADDISFLY PRODUCTION 399

ics of headwater streams. Freshwater Biology 23:281-299.

CUMMINS; K. W. 1964. Factors limiting the microdis-tribution of larvae of the caddisflies Pycnopsychelepida (Hagen) and Pycnopsyche guttijer (Walker) ina Michigan stream (Trichoptera: Limnephilidae).Ecological Monographs 34271-295.

CUMMINS, K. W., M. A. WILZBACH, D. M. GATES, J. B.FERRY, AND W. B. TAUAFERRO. 1989. Shreddersand riparian vegetation. BioScience 3924-30.

DOBSON, M., AND A. G. HILDREW. 1992. A test of re-source limitation among shredding detritivores inlow order streams in southern England. Journalof Animal Ecology 61:69-77.

DUDGEON, D. 1987. A laboratory study of optimal be-haviour and the costs of net construction by Po-b/centropusflaoomacuSatus (bisecta: Trichoptera: Po-tycehtropodidae). Journal of Zoology 211:121-141.

EGGERT,S. L.2003. Resource use by detritivorousinac-roinvtti'UjLrdltis in southern Appalachian head-water streams. PhD Thesis, University of Georgia,Athens, Georgia.

EGGUSHAW, H. J. 1964. The distributional relationshipbetween bottom fauna and plant detritus instreams. Journal of Animal Ecology 33:463-476.

FUNT, Q S. 1960. Taxonomy and biology of Nearcticlimnephilid larvae (Trichoptera), with special ref-erence to species in eastern United States. Ento-mologica Americana 40:1-120.

GRUBBS, S. A., AND K. W. CUMMINS. 19%. Linkagesbetween riparian forest composition and shred-der vdtinism. Archiv fur Hydrobiologie 137-39-58.

HALL, R. Q, J. B. WALLACE, AND S. L. EGGERT. 2000.Organic matter flow in stream food webs withreduced detrital resource base. Ecology 813445-3463.

HAMILTON, A. L, 1969. On estimating annual produc-tion. Limnology and Oceanography 14:771-782.

HUKYN, A. D, AND J. B. WALLACE. 1988. Communitystructure of Trichoptera in a mountain stream:spatial patterns of production and functional or-ganizatioiL Freshwater Biology 20:141-155.

HUTCHENS, J. J,rE. E ftSNttELO, AND JJH. WEBSTER.1997. Diet and growth of a leaf-shredding cad-disfly in southern Appalachian streams of con-trasting disturbance history. Hydrobiologia 346:

. 193-201. ;«• ; . : • • - : - - - . . . ; . . . . -0 •LUGTHARtj G. J., AND J. B. WALLACE. 1992. Effects of

disturbance on benthic functional structure andproduction in mountain streams. Journal of weNoruf; American Benfeological Society 11:138-

• : -164. j E j ' - ' •' ' " • ' ; - • '••••'• ' • - • . : ;\-;.-'MACKAY, ik. J. 1972. Temporal patterns in life history

and Sight behaviour of Pycnopsyche gentiUs, P. bi-culentH, and P. scabripamis (Tridu^tera: Limne-philidae). Canadian Entomologist 104:18194-1835.

MACKAY, R. J. 1977. Behavior of Pycnopsyche (Trichop-

tera: Limnephilidae) on mineral substrates in lab-oratory streams. Ecology 58:191-195.

MACKAY, R. J., AND J. KALFF. 1973. Ecology of tworelated species of caddis fly larvae in the organicsubstrates of a woodland stream. Ecology 54:499-511.

MACKAY, R. J., AND G. B. WIGGINS. 1979. Ecologicaldiversity in Trichoptera. Annual Review of En-tomology 24:185-208.

MORSE, J. C, B. P. STARK, AND W. P. MCCAFFERTY.1993. Southern Appalachian streams at risk: im-plications for mayflies, stoneflies, caddisflies, andother aquatic biota. Aquatic Conservation: Marineand Freshwater Ecosystems 3293-303.

NAIMAN, R. J., H. DECAMPS, AND M. POLLOCK. 1993.The role of riparian corridors in maintaining re-gional biodiversity. Ecological Applications 3:209-212.

NATIONAL RESEARCH COUNCIL (NRC). 2002. Riparianareas: functions and strategies for managementA report of the National Research Council. Na-tional Academy Press, Washington, DC.

NttSSON, C., AND K. BERGGREN. 2000. Alterations ofriparian ecosystems caused by river regulation.BioScience 50:783-792.

NBLOW, K. H., AND M. C. MOLLES. 1993. The influenceof larval case design on vulnerability of Umw-phSusfrijole (Trichoptera) to predation. Freshwa-ter Biology 29:411-417.

OTTO, C. 1974. Growth and energetics in a larval pop-ulation of Potamophylax ringulatus (Steph.) (Tri-choptera) in a south Swedish stream. Journal ofAnimal Ecology 43:339-361.

Ono, C, AND A. JOHANSSON. 1995. Why do somecaddis larvae in running waters construct heavy,bulky cases? Animal Behaviour 49:473-478.

OTTO, C, AND B. S. SVENSSON. 1980. The significanceof case material selection for the survival of cad-dis larvae. Journal of Animal Ecology 49:855-865.

PERRY, W. Bv E. F. BENFIELD, S. A. PERRY, AND J. R.WEBSTER. 1987. Energetics, growth, and produc-tion of a leaf-shredding stonefly in an Appala-chian mountain stream. Journal of the NormAmerican Benthological Society 6:12-25.

PETERSEN, R. C, AND K. W. CUMMINS. 1974. Leaf pro-cessing in a woodland stream. Freshwater Biolo-gy 4345-368. • • • • ' • ,

RICHARDSON, J. S. 1991. Seasonal food limitation ofdetritivores in a montane stream: an experimentaltest Ecology 72:873-887.

ROSEMOND, A. P., CM. PRINGLE, A. RAMIREZ, ANDM. J. PAUL. 2001. A test of the top-down and bot-tom-up control in a detritus-based food web.Ecology 822279-2293.

Ross, H. H. 1963. Stream communities and terrestrialbiomes. Archiv fur Hydrobiologie 59235-242.

SOUTHERN APPALACHIAN MAN AND THE BIOSPHERE(SAMAB). 1996. The Southern Appalachian As-

400 S. L. EGGERT AND J. B. WALLACE [Volume 22

sessment Aquatics Technical Report Report 2 of5. US Department of Agriculture, Forest Service,Southern Region, Atlanta, Georgia.

STONE, M. K., AND J. B. WALLACE. 1998. Long-termrecovery of a mountain stream from dear-cut log-ging: the effects of forest succession on benthicinvertebrate community structure. Freshwater Bi-ology 39:151-169.

STOUT, B. M, E. F. BENFELD, AND J. R. WEBSTER. 1993.Effects of forest disturbance on shredder produc-tion in southern Appalachian headwater streams.Freshwater Biology 2959-69.

WALLACE, J. B., T. F. CUFFNEY, S. L. EGGERT, AND M.R. WHILES. 1997. Stream organic matter inputs,storage, and export for Satellite Branch at Cow-eeta Hydrologic Laboratory, North Carolina,USA. Pages 67-74 in J. R. Webster and J. L. Meyer(editors). Stream organic matter budgets. Journalof the Norm American Benthological Society 16.

WALLACE, J. B., S. L. EGGERT, J. L. MEYER, AND J. R.WEBSTER. 1999. Effects of resource limitation on adetrital-based ecosystem. Ecological Monographs69:409-442.

WATERS, T. F. 1969. Subsampler for dividing largesamples of stream invertebrate drift Limnologyand Oceanography 14:813-815.

WEBSTER, D. A., AND F. C WEBSTER. 1943. Influence ofwater current on case weight in larvae of the cad-

disfly Goera cokorota (Banks). Canadian Entomol-ogist 75:105-108. h

WEBSTER, J. R., AND E. F. BENHELD. 198$. Oscularplant breakdown in freshwater ecosystems. An-nual Review of Ecology and Systemancs 17567-

'594 . . - . • " • : . ! > : • -WEBSTER, J. R., M. E. GURTZ; J. J. HAINS, JJIL, MEYER,

W. T. SWANK, J. B. WAIDE, AND J. B.j WALLACE.1983. Stability of stream ecosystems, Pages 355-395 in J. R. Barnes and Q W. Minshall (editors).Stream ecology. Plenum Press, New York.

WHILES, M. R., AND J. B. WALLACE. 1995. Macromver-tebrate production in a headwater stream duringrecovery from anthropogenic disturbance and hy-drologk extremes. Canadian Journal of Fisheriesand Aquatic Sciences 5212402-2422.

WIGGINS, G. B., AND R. J. MACKAY. 1978. Some rela-tionships between systematics and trophk ecolo-gy in Nearctic aquatic insects, with special refer-ence to Trichoptera. Ecology 59:1211-1220.

WILLIAMS, D. D, A. F. TAVARES, AND E. BRYANT. 1987.Respiratory device or camouflage?—A case forthe caddisfiy. Oikos 50:42-52.

ZAMORA-MUNOZ, C, AND B. W. SVENSSON. 1996. Sur-vival of caddis larvae in relation to their case ma-terial in a group of temporary and permanentpools. Freshwater Biology 36-23-31.

Received: 26 November 2002Accepted: 16 June 2003