Local Effects of a Sedentary Grazer on Stream Algae

Freshwater Biotogy (1995) 33, 401-409

Local effects of a sedentary grazer on stream algae

ELIZABETH A. BERGEY Department of Entomology, University of Califomia, Berkeley, CA 94720, U.S.A.

Present address and address for correspondence: Department of Botany, University of Wisconsin, Madison, WI 53706, U.S.A.

SUMMARY 1. Larvae of the sedentary aquatic caterpillar Petrophila confusaiis (Walker) construct silken retreats around which they feed; outside these clearly demarcated grazed areas, stream algae are exposed to a variety of mobile grazers. Comparisons of the algal community inside and just outside grazed zones were made for third- and fifth-instar Petrophila in the South Fork of the Eel River, California. 2. Densities of both the filamentous macroalga Cladophora and diatom assemblages were significantly reduced within the grazed areas during both larval instars. Grazing of diatoms was taxonomically non-selective. 3. In spring, the grazed zones were relatively large (mean = 22.7 cm^) and visibly increased epilithic spatial patchiness. 4. Per cent composition of diatom assemblages inside and outside the grazed areas differed during the third instar but not during the fifth instar. During the third instar, the grazed zone contained more Synedra ulna (a conimon inunigrant) and less Gomphonenia olivaceum (a late-successional species).

Introduction Benthic algal grazers in freshwater habitats include a to this research emphasis. Whereas mobile grazers taxonomically wide array of animals, ranging from include vertebrates (e.g. Power, Matthews & Stewart, fish and amphibians to several groups of invertebrates 1985), crustaceans (e.g. Flint & Goldman, 1975), mol- (e.g. snails, insects and crustaceans). They can be luscs (e.g. Tuchman & Stevenson, 1991), insects (e.g. characterized as being mobile grazers or sedentary Feminella & Resh, 1991), oligochaetes (e.g. Hann, grazers. Mobile grazers change location in conjunction 1991) and protozoa (e.g. McCormick, 1991), sedentary with feeding and tend to focus their grazing on distinct grazers primarily include a small number of insects patches of periphyton (e.g. Hart, 1981; Kohler, 1984; [e.g. caterpillars of the genus Petrophila (Pyralidae: Vaughn, 1986; Steinman et ai, 1987; Scrimgeour et al, Lepidoptera; e.g. Tuskes, 1977), several chironomids 1991). In contrast, sedentary grazers feed in a localized (Chironomidae: Diptera; e.g. Pringle, 1985), and the area (e.g. Hart, 1985). Although not physically fixed caddisfly genus Leucotrichia (Hydroptilidae: Trichop- to the substrate, sedentary grazers construct attached tera; e.g. Hart, 1985)]. Although few in terms of retreats and graze the surrounding area and, often, numbers of taxa, sedentary grazers are often locally the retreat itself. Alternatively, fixed retreats may be abundant (Petrophila: Peterson & Grimm, 1992; chiron- shallow mines in aquatic macrophytes, with the animal omids: Hershey et al, 1988; leucotrichia: Hart, 1985) grazing epiphytes and filter feeding (e.g. Lamberti & and may be important components in local foodweb Moore, 1984). dynamics. Studies of benthic grazing have focused almost Most sedentary grazers produce a clearly demarc- exclusively on mobile grazers. Mobile grazers include ated grazed zone on the substrate around their retreats. a much larger number of taxa than sedentary grazers Outside this area, the algal commuruty is exposed to (McAuliffe, 1984), a factor that certainly contributes mobile grazers. Thus, algal populations inside and

© 1995 BlackweU Science Ltd 401 402 £. A. Bergey outside the areas grazed by sedentary grazers will be extended into a pool. The area grazed by fifth instars influenced by differences in both grazing intensity was estimated by removing cobbles with Petrophila (high and low) and grazing strategy (sedentary and pupal cases and tracing the outline of the grazed zone mobile). Sedentary grazers are especially amenable to on to clear plastic. The 30 resulting tracings were cut studies of grazer-alga interactions because they have out and weighed. fixed locations and thus need not be confined in cages Cladophora biomass next to Pefrop/ii/o-grazed zones or artificial channels (as mobile grazers are often was measured by scraping Cladophora from rocks constrained). within 20 25-cm^ quadrats. Algal samples were frozen This study examines grazing by the sedentary cater- and later thawed, dried at 105 °C for 2 days, weighed, pillar Petrophila confusalis (Walker) through examina- ashed at 550 °C for 1 h, and reweighed. Samples of tion of epilithic algae within and adjacent to areas the ash were examined microscopically for diatom grazed hy Petrophila larvae. This comparison is made frustules. separately for the third- and fifth-instar larvae, in order In addition to making comparisons of algal popula- to evaluate possible ontogenetic changes in grazing. tions inside and outside areas grazed by Petrophila larvae, differences in the impacts of grazing by third- and fifth-{final) instar larvae were examined. Algal Methods samples associated with the fifth instar were collected The South Fork of the Eel River is a third-order stream on 10 June 1991 when some of the overwintering located in the Coast Range of northern California Petrophila were begiruiing to pupate. Algal samples (Mendocino Co.; latitude 39°44' longitude 123°38')- associated with the third instar were collected on 24 Local riparian vegetation is dominated by old-growth August 1991 when the next generation of Petrophila Douglas-fir {Pseudotsuga menziesii Franco) and, while was present. Sample collection and processing the banks are typically well shaded, the study reach methods were the same for both sets of samples. bordered an abandoned streambed, resulting in full Epilithic algae were collected with a brush sampler sunlight on the stream for about 7 h per day during {modified from Tuchman & Stevenson, 1980; area = summer. The 75-m-long study reach had a predomi- 1.53 cm^). For the third ins tar-associated samples, the nantly cobble-rubble substrate and the following sampler collected algae beneath the larval retreat (the characteristics during the study: width 9-12 m; depth retreat was removed prior to sampling and, once typically less than 1 m; spatial variation in current removed, its former location was not distinguishable velocity of 0.02-0.25 m s~^; and a water temperature in the field) and most of the grazed area around the of lS-17 °C in June, and 17-23 °C in August. retreat (diameter across grazed area and retreat was The life history of Petrophila confusalis in northern approximately 1.8 cm). For the fifth-ins tar-associated California has been described by Tuskes (1977, 1981). samples, the sampler was placed within the grazed Larvae build silken retreats over irregularities in rocks area surrounding the retreat (diameter across grazed or submerged wood, and graze both the underside of areas and the retreat was approximately 5.4 cm). the retreat (Kubik, 1981) and the surrounding area Samples outside the grazed area were taken adjacent (Tuskes, 1977). In preparation for pupation, fifth-instar to the grazed zone. Six {in June) or eight {in August) larvae form a flattened dome-like silk case, under replicate pairs of samples were collected and preserved which they pupate. Petrophita has two to three genera- in 6% formalin. In addition to algal samples, the larva tions per year in northern California (Tuskes, 1977). and the retreat or pupal tent associated with most The portion of the population studied consisted of samples were collected. overwintering fifth-instar larvae and third-instar lar- Processing epilithic samples included two steps: vae of tbe following generation. {i) a low-magnification quantification of filamentous Petrophila density was obtained by counting pupal algae and diatoms; and (ii) a high-magnification deter- cases {visible as light-coloured ovals in areas cleared mination of per cent composiHon of the diatoms. of Cladophora) in late June across 10 randomly located Samples were washed to remove formalin and transects in the 75-m-long study reach. Transects were adjusted to a volume of 4 ml. One millilitre of the 0.5-m wide and maximum depth was less than 1 m, well-shaken sample was added to a Sedwick-Rafter except in the two furthest upstream transects, which cell for algal quantification; the remainder of the © 1995 BlackweU Science Ltd, Freshwater Biology, 33, 401-109 Effects of a sedentary grazer on algae 403 sample was saved for diatom identification. Diatoms the two most upstream transects (3.64 and 4.50 (both living and non-living) were counted in 24 ran- individuals m~^), which were located in a pool with domly located fields at lOOx magnification with a deeper water (to 1,7 m) and slower velocity than the Whipple micrometer. Filamentous algae were identi- rest of the study reach. Excluding the pool transects, fied to genus and regarded as present or absent in mean Petrophila density was 35.64 individuals m"^ each of 100 randomly located Whipple disk fields. (SD = 10.12, n = 8). Some retreats were almost The remaining 3 ml of each sample were oxidized certainly missed in the visual counts; therefore values by; (i) warming in 30% hydrogen peroxide for 2 days; are probably underestimates. Third-instar larval densi- or (ii) heating in 30% hydrogen peroxide and adding ties were not estimated because retreats were better the catalyst potassium dichromate (an oxidative reac- concealed and difficult to count accurately. tion occurred within a few minutes). After rinsing, Algal abundances were lower inside than outside the diatoms were dried on coverslips and mounted in grazed area (Table 1). Relative numbers of Cladophora, Hyrax medium. Transects were counted at lOOOX green algal filaments other than Cladophora (primarily magnification until at least 350 individual diatoms Oedogonium, Mougeotia, and Zygnema), and blue—green had been identified. Frustule length was measured for filaments (primarily Anabaena, Calothrix, and Oscil- 100 diatoms per sample. Paired f-tests were used to latoria) were significantly lower (2-tailed paired f~test; compare mean sample sizes inside and outside the » = 6 or 8; P ^e 0.05) inside the grazed area for both Petrophila grazed areas. third- and fifth-instar larvae. Diatom densities were Examination of Petrophila guts by light microscopy also significantly lower inside the grazed areas and Petrophila retreats by scanning electron microscopy (P ^ 0.05). revealed that diatoms far outnumbered other algae; By the fifth instar, Petrophila activity resulted in an filamentous algae were rare in Petrophila guts (e.g. estimated reduction in standing crop of 0.7 mg dry four short segments of Cladophora were found in eight weight of periphyton cm~^ of streambed, or 7.5% of guts) and absent from the examined tents (« = 6). total stream periphyton (Table 2). This periphyton Therefore, entire Petrophila guts and whole retreats reduction occurred in the grazed areas, which aver- were oxidized with hydrogen peroxide and potassium aged 22.65 cm^ (SD = 5.52; n= 30), where losses were dichromate, a process that destroyed non-diatoma- much higher (estimated at 8.7 mg dry weight cm~^). ceous algae but effectively separated the diatoms. Observation suggested that Cladophora (because of its During this process the gut was destroyed and tents large size) accounted for most of the reduction in were oxidized into a loose clump of silk strands. standing crop. The high ash content of the ungrazed Diatom relative abundances were determined using periphyton samples suggests that loss of diatoms Hyrax mounts. To evaluate the effects of grazing, associated with Cladophora also contributed to the per cent compositions of diatoms in samples of gut reduced standing crop inside the grazed areas, because contents, retreats, and inside and outside the grazed the ash contained mostly frustules of the diatoms area on rocks were compared using cluster analysis Epithemia turgida (Ehrenb.) Kutz. and Cocconeis pedic- (SYSTAT; Euclidian distance and average linkage). In ulus Ehrenb. These species were common epiphytes addition, compositional differences inside and outside of Cladophora in the South Fork of the Eel River the grazed areas were evaluated using 2-way ANOVAS (E.A. Bergey, unpublished data) and also occurred (location inside or outside the grazed area and diatom abundantly on rocks. taxa), followed by Tukey's test on cell means, where appropriate (Zar, 1984). Diatom assemblage composition

Diatom taxonomic richness inside and outside the Results Petrophila grazed zone was not affected by grazing during either instar (third instar, 27.0 and 29.7 taxa; Petrophila density and algal standing stock fifth instar, 29.8 and 30.0 taxa, respectively; paired t- The density of Hfth-instar larvae and pupae of Petro- tests, P > 0.05; n = 6), despite significant reductions phila was similar throughout the study reach (range = in diatom density with Petrophila grazing (Table 1). 25.14-48.39 individuals m~^), with the exception of Major shifts in diatom composition occurred

© 1995 Biackwell Science Ltd, Freshwater Biology, 33, 401-409 404 £. A. Bergey Table 1 Algal abundance inside and outside Petrophila grazed areas for third- (a) and fifth- (b) instar larvae. Values for Cladophora, green algal filaments other than Cladophora, and blue-green algal filaments are based on the number of presence-absence cx!currences; values for diatoms are numbers per cm' of rock. Two-tailed paired f-tests were run on log {x + 1) transformed data; the listed mean ± SE are untransformed. Significant P values {^ 0.05) are indicated with asterisks

Inside Outside Taxon d,f. t value P value Mean ± SE Mean ± SE

(a) Third instar Cladophora 5 12.^ 0.0001* 0,2 ± 0,2 29,0 ± 5,6 Other green algae 5 10.49 0.0001» 2.7 ± 1.3 48.8 ± 13.9 Blue-green algae 5 3.14 0.026* 49.0 ± 17.8 87.8 ± 30.9 Diatoms 5 4.07 0.010* 32.5 X 10^ ± 8.9 X lO-"^ 171.7 X 10^ ± 27.6 X 10-^ (b) Fifth instar Cladophora 7 239 aO48» 0.8 ± 0.3 11.6 ± 5.3 Other green algae 7 mi 0.027* 5.8 ± 1.3 41,9 ± 14.0 Btue-green algae 7 0.046* 21.1 ± 7.6 57.1 ± 14.4 Diatoms 5 0.001* 23.7 X 10^ ± 9.5 X 10^ 301,4 X 10-^ ± 80.4 X 10^

Table 2 Mean biomass (± SE) of periphyton on rocks outside sorex) showed a significantly different inside-outside the areas grazed by fifth-instar Petropbila [n = 20), and distribution, per cent compositions of the other taxa estimates of biomass inside the grazed areas and the reducHon in standing crop resulting from fifth-instar Petrophita grazing, during the third instar were recalculated without DW = dry weight and AFDW = ash-free dry weight Epithemia before further analysis. During the third instar, location inside or outside the grazed area, -2\ Algal biomass (mg cm difference among diatom taxa, and the interaction Outside Estimated loss from were significant (2-way ANOVA; d.f. = 1, f = 7.99, grazed Estimate inside grazing (per area of P = 0.005; d.f. - 18, f = 36.86, P = 0.0001; d.f. = 18, area grazed area* stTeam)+ F = 9.64, P = 0.0001, respectively; n = 6). Three taxa DW 9.34 ± 1,03 0,64 0.70 had significantly different densities inside and outside AFDW 4.31 ± 0,44 0,30 0.33 the grazed area (Tukey's test on cell means; P =s 0.05). Ash 5.03 ± 0,64 0.35 0.38 Cocconeis pediculus and Gomphonema olivaceum were * Estimated biomass inside grazed area - biomass outside proportionally more abundant outside the grazed grazed area x proportion of Cladophora abundance inside the areas and Epithemia sorex more abundant inside the grazed area to outside the grazed area. Using data from Table grazed areas. During the fifth instar, difference among 1, this proportion = 0.8/11,6 = 0,069. diatom taxa was significant (2-way ANOVA; d.f. = t Estimated loss from grazing = (difference in biomass outside 16, F = 17.41, P = 0.0001; n - 6), however, location and inside the grazed area) x proportion of the stream area grazed. Proportion of stream area grazed = mean grazed and the interaction were not (P > 0.05). area x retreat density = 22.65 an^ x 35.64 retreats m"^ = Diatom frustules inside areas grazed by third-instar 0.081 Petrophila were larger on average than those in adjacent ungrazed areas (paired f-tests; Table 4), the result of between the fifth larval instar (in mid-June) and the fewer small diatoms (e.g. Cocconeis pediculus) and third larval instar (in late August), reflecting temporal relatively more mid-sized (e.g. Epithemia sorex) and changes in stream assemblages (Table 3). For example, large diatoms (e.g. Cymbella tumida and Synedra ulna) Epithemia sorex increased in abundance with time, inside the grazed area. Mean diatom size during the Achnanthes minutissima decreased, and several species fifth instar was not significantly different inside and appeared and disappeared from the list of major taxa outside the grazed areas. During the third instar, (those comprising over 1% of the total number during growth form composition (i.e. per cent upright and an instar). stalked, prostrate, motile, or filamentous diatoms) Diatom assemblage composition differed inside and differed between inside and outside samples, primar- outside the grazed areas during the third instar, but ily because of differences in Epithemia sorex density not during the fifth instar (Table 3). Because the (Table 5). Inside samples had proportionally more dominant taxon during the third instar {Epithemia Epithemia and therefore more total motile forms than

Table 3 Mean percentage composition of abundant diatom taxa. Taxa with significantly different relative abundances inside and outside the Petrophila grazed areas are indicated with asterisks (2-way ANOVA on arcsine transformed data, P e 0.05; Tukey's test on cell means, P ^ 0.05; n = 6)

Third instar Eifth instar Taxon Inside Outside Inside Outside

Achnanthes lanccolata Breb. ex Kutz. 2.17 1.71 Achnanthes minutissima Kutz. 2.60 6.07 23.42 29.85 Cocconeis pediculus Ehrenb. 0.89* 14.15' 4.43 7.10 Cocconeis placentula Ehrenb. 1.56 2.57 1.59 0.79 Gifchtelta meneghiniana Kiitz. 1.00 1.19 Cymbella tsimida (Breb. ex Kiitz.) V. H. 3.55 1.09 Diatoma vulgare Bory 3.79 4.47 Epithemia admta (Kutz.) Breb. 3.59 1.60 8.00 3.92 Epithemia sorex Kutz. 56.92- 22.79' 12.72 11.45 Epithemia turgida (Ehrenb.) Kutz. 7.92 8.97 Fragilaria capucina Desm. 1.31 1.88 Fragilaria construens (Ehrenb.) Grun. 2.09 3.41 Gomplwnettia olivaceum (Lyngb.) Kiitz. 0.35* 4.69* 2.13 2.85 Gomphonenm nr intricatum KUtz. 4.96 2.06 Navicula menisculus (Grun.) Grun. 2.05 3.86 2.37 2.71 Navicula salimrum Grun. 1.84 3.90 1.34 0.80 Nitzschia dissipata (Kutz.) Grun. 1.32 0.82 1.77 1.93 Nitzschia fonticota Grun. 3.47 5.74 1.17 0.95 Nitzschia frusltiluni (Kutz.) Grun. 3.24 5.15 6.28 8.20 Nitzschia fruslulum var. pcrminuta Grun. 3.44 2.33 Nitzschia pateaceae Grun. 2.12 2.89 1.35 1.02 Rhoicosphenia cunxita (Kutz.) Grun. ex Rabh. 1.16 4.76 3.62 3.18 Synedra ulna (Nitz.) Ehrenb. 1.71 0.97

Table 4 Mean diatom size inside and outside the grazed area during the third and fifth instars of Petrophila. Results from 2-tailed paired f-tests are given. Diatom sizes (means and standard errors) are lengths along the longitudinal axis in micrometres

Inside Outside Larval instar d.f. t value P value Mean ± SE Mean ± SE

Third instar 5 5.25 0.003 31.89 ± 1.23 25.78 ± 0.41 Fifth instar 5 0.53 0.62 26.85 ± 2.90 25.79 ± 3.46

outside samples; no other differences were significant. Table 5 Percentage composition by growth form (mean ± SE) Growth form data associated with the fifth instar were of diatom assemblages inside and outside Petrophila grazed not analysed because the diatom compositions inside areas during the third instar. Because of the abundance of and outside the grazed area were not significantly Epithemia spp. , this taxon was separated from the remainder of the motile taxa and, for statistical tests, proportions were different. Eig. 1 illustrates the relative size of Petrophila recalculated with Epithemia spp. excluded. Asterisks indicate mouthparts and representative diatoms. Whereas significant differences between inside and outside samples (2- smaller diatoms (e.g. Achnanthes minutissima) are quite way ANOVA on arcsine transformed data, P « 0.05; Tukey's small relative to the insect's mandibles, the larger test on cell means, P « 0.05; n = 6) diatoms (e.g. Epithemia turgida and Synedra ulna) are Growth form Inside Outside large enough that selective feeding (or avoidance) appears possible. Epithemia spp. 62.0 ± 4.0' 25.9 ± 5.6' Other motile forms 15.8 ± 3.1 24.7 ± 2.7 The cluster diagram of per cent composition of Prostrate forms 6.4 ± 0.9 24.8 ± 4.1 diatoms in guts, on retreats and inside and outside Upright or stalked forms 11.7 ± 1.1 19.7 ± 2.3 grazed areas on rocks for the third instar (Fig. 2) Filamentous forms 3.8 ± 0.7 4.6 ± 1.0 includes two large clusters, designated A and B.

1995 Blackwell Science Ltd, Freshwater Biology, 33, 401-^)9 406 E. A. Bergey Cluster A includes samples from inside the Petrophila grazed areas and the retreats, indicating that although the substrate differs (rock and silk, respectively), Petro- phila affects these two areas similarly. Samples outside the Petrophila grazed area are grouped in cluster B and are thus distinct from those inside the grazed area and retreats. This distinction between diatom composition inside and outside Petrophila grazed areas is consistent with the analysis of diatom per cent composition. Gut samples were sometimes dominated by taxa that were not dominant on rocks, such as Nitzschia frustulum or Fragilaria construens, and occur most frequently as outliers from the two main clusters (A and B). Gut 100/jm content counts could have been biased by differential gut retention of diatom species or by the timing of Fig. 1 Comparative sizes of mandibles of (a) fifth-instar and feeding (several of the larvae had few or no diatoms (b) third-instar Petrophila, and (c) representative diatom taxa in the gut). (from small to large: Achnaniheii minutissima, Epithemia sorex; Samples taken when fifth-instar larvae were present Epithemia lur^ida; and Synedra ulna). Mandible drawings are based on SEM photographs. show a different pattern of clustering (Fig. 3). The different sample types (inside the grazed area, outside, gut contents and retreat samples) do not each cluster

Distance 0.0

In In In Tent In Cluster In A Tent Tent Tent Tent Out In Tent Gut Gut Fig. 2 Cluster diagram of diatom Out samples associated with third-instar Petrophila. In, samples inside the Cluster Out Petrophila grazed zone; Out, outside the B Out Petrophila grazed zone; Tent, samples from Petrophila pupal tents, and Gut, Out samples from Petrophila guts. Two main Out dusters {A and B) are indicated. Gui Euclidian distance and average linkage were used for clustering. Gut 1995 BlackweU Science Ltd, Freshwater Biology, 33, 401-409 Effects of a sedentary grazer on algae 407

Distance 0.0 20.0

Tent Gut Out Tent Out In tn In Gut Out Fig. 3 Cluster diagram of diatom Out samples associated with fifth instar In Petrophila. In, samples inside the Petrophila grazed zone; Out, outside the In Petrophila grazed zone; Tent, samples from Petrophila pupal tents, and Gut, Out samples from Petrophila guts. Euclidian In distance and average linkage were used for clustering, Out together, but instead appear in an unpattemed array. Moore, 1984; Hill & Knight, 1988) in algal removal This is consistent with the analysis of diatom per cent and handling. Hence, grazers typically remove large composition for the fifth instar, which showed no or stalked diatoms and leave small prostrate forms significant differences inside and outside grazed areas. attached to the substrate (e.g. Jacoby, 1987; Steinman The jumbled order of sample types within clusters, et ai, 1987). Third-instar Petrophila, with their smaller and the small differences between clusters, may result mouthparts, might be expected to graze more effici- from spatial variation exceeding differences between ently (and hence more 'selectively') than fifth-instar 'treatments' (location inside or outsides the grazed larvae. Indeed, third-instar Petrophiia removed the zone). small prostrate diatom Cocconeis pedicutus, which has been described as grazer resistant (Dudley, 1992). However, third-instar larvae also left several large Discussion diatoms (e.g. Cymbella tumida and Synedra ulna), and Grazing differences between instars mean diatom size was larger within the grazed zone, rather than smaller as expected (e.g. Jacoby, 1987; Diatom assemblage composition inside and outside Steinman et ai, 1987), suggesHng that differences in the Pctrophila-^tazed zones differed during the third selective grazing do not explain the ins tar-related but not the fifth instar. Possible reasons for this instar difference in diatom assemblages inside and outside difference include: (i) differences in selective grazing the grazed areas. between instars (ii) temporal changes in the mobile grazer assemblage, and (iii) differences in grazing Alternative possibilities of larval instar-specific graz- frequency between instars. ing pattems involve temporal changes in grazing, Because diatoms are small relative to the insect either outside the grazed area (e.g. compositional grazer, 'selectivity' in grazing is closely tied to the changes in mobile grazer assemblages), or inside the mechanical efficiency of the mouthparts (Lamberti & grazed area (i.e. changes in grazing frequency). Popu- © 1995 Blackwell Science Ltd, Freshwater Biology, 33, 401-409 408 £. A. Bergey lation changes in grazers other than Petrophila between algal density and possibly changes in algal composi- June and August could be involved in altering relative tion (especially of filamentous algae) around the abundance of diatoms inside and outside the grazed retreats. This study of Petrophila grazing also suggests areas. Unfortunately, data were not collected on that the reduction in density of diatoms may be non- changes in other grazer populations or their individual selective for taxa. grazing effects. Sedentary grazers increase the spatial heterogeneity Because aquatic holometabolous insects tend to of benthic algal communities by producing heavily increase feeding and growth rates during the final grazed areas that are distinct from adjacent areas. larval instar (e.g. Oemke, 1983; Hart, 1985), fifth-instar These grazed patches may have the following charac- larvae may graze more frequently than third-instar teristics relative to adjacent areas: (i) lower algal larvae. Immigrating diatoms therefore might have density (this study); (ii) higher light levels, resulting more success in establishing during the third than the from the removal of filamentous algae (personal obser- fifth instar, and indeed, areas grazed by third-instar vation); (iii) possible intermittent exposure to nutrients larvae had relatively more Synedra ulna, a common released by the grazer (e.g. Power 1991); and (iv) immigrator (Tuchman & Stevenson, 1991; Peterson & unknown changes in local water flow patterns. While Grimm, 1992), and the area not grazed has more many of these characteristics would be beneficial to a Gomphonema oUvaceum, a late successional species variety of epilithic algae, the intense grazing is cer- (McCormick & Stevenson, 1991). tainly a dominant influence and detrimental to most exposed algae. I Sedentary grazing Acknowledgments In this study, it had been expected that large differences between the alga! assemblages would be found inside I thank Trish Steel for assistance with field-work and and outside the Petrophila-grazed zone (i.e. differences Eugene Stoermer with diatom identification. Discus- caused by sedentary and mobile grazers). However, sions with Vince Resh were helpful, and suggestions although Petrophila-^razed areas had a consistently from Vince Resh and Michael Winterbourn improved much lower algal density (especially of filamentous the manuscript. Funding was provided by the Penny algae), changes in per cent diatom composition were Devlin Foundation and ARCS Foundation. mostly minor, especially during the fifth instar. This result differs somewhat from that obtained by Hart (1985) in a study of grazing by the sedentary microcad- References disfly Leucotrichia pictipes (Banks), probably because Dudley T.L. (1992) Beneficial effects of herbivores on the algal communities were different. In Augusta stream macroalgae via epiphyte removal. Oikos, 65, Creek, Michigan, Hart (1985) demonstrated the exclu- 121-127. sion of the invasive blue-green alga Microcoleus Feminella J.W. & Resh V.H. (1991) Herbivorous vaginatus (Vauch.) Gomont from Leucotrichia grazed caddisflies, macroalgae, and epilithic microalgae: areas, leaving the otherwise dominant Schizothrix- dynamic interactions in a stream grazing system. diatom community intact. In the Eel River, no similar Oecologia, 87, 247-256. invasive algae were observed, and the dominant com- Flint R.W. & Goldman C.R. (1975) The effects of a benthic munity of Cladophora and diatoms was numerically grazer on the primary productivity of the littoral zone reduced and diatom composition was altered (third of Lake Tahoe. Limnology and Oceanography, 20,935-944. Harm B.J. (1991) Invertebrate grazer-periphyton inter- instar) or not (fifth instar). Other studies of sedentary actions in a eutrophic marsh pond. Freshwater Biology, aquatic insects as algal grazers have dealt primarily 26, 87-96. with chironomids. Wiley & Warren (1992) demon- Hart D.D. (1981) Foraging and resource patchiness: field strated algal reduction in the grazing territories of experiments with a grazing stream insect. Oikos, 37, Cricotopus spp. and Winterbourn (1990) provides an 46-52. SEM photograph showing what appears to be a grazed Hart D.D. (1985) Grazing insects mediate algal inter- zone around the retreat of an orthoclad chironomid. actions in a stream benthic community. Oikos, 44,40-46. The overall pattern for sedentary grazers is reduced Hershey A.E., Hiltner A.L., HuUar M.A.J., Miller M.C, © 1995 BlackweU Sdence Ltd, Freshwater Biology, 33, 401-409 Effects of a sedentary grazer on algae 409

Vestal J.R., Lock M.A., Rundle S. & Peterson B.J. (1988) Power M.E., Matthews W.). & Stewart A.J. (1985) Grazing Nutrient influence on a stream grazer: Orthocladius minnows, piscivorous bass, and stream algae: dynamics microcommunities respond to nutrient input. Ecology, of a strong interaction, Ecology, 66, 1448-1456. 69, 1383-1392. Pringle CM. (1985) Effects of chironomid (Insecta: Hill W.R. & Knight A.W. (1988) Concurrent grazing effects Diptera) tube-building activities on stream diatom com- of two stream insects on periphyton. Limnology and munities. Journal of Phycology, 21, 185-194. Oceanography, 33, 15-26. Scrimgeour G.J., Culp J.M., Bothwell M.L., Wrona F.J. & Jacoby J.M. (1987) Alterations in periphyton character- McKee M.H. (1991) Mechanisms of algal patch deple- istics due to grazing in a Cascade foothill stream. tion: importance of consumptive and non-consumptive Freshwater Biology, 18, 495-508. losses in mayfly diatom systems. Oecologia, 85,343-348. Kohler S.L. (1984) Search mechanism of a stream grazer Steinman A.D., Mclntire CD., Gregory S.V, Lamberti in patchy environments: the role of food abundance. G.A. & Ashkenas L.R. (1987) Effects of herbivore type Oecologia, 62, 209-218. and density on taxonomic structure and physiognomy of algal assemblages in laboratory streams, lournal of Kubik CL. (1981) The ecology o/Parargyractis bifascialis the North American Benthological Society, 6, 175-188. (Lepidoptera: Pyralidae) in the Chippeu^a River, Isabella County, Michigan. MSc thesis. Central Michigan Univer- Tuchman M.L. & Stevenson R.J. (1980) A comparison of clay tile, sterilized rock, and natural substrate diatom sity, Mount Pleasant, Michigan. communities in a small stream in southeastern Michi- Lamberti G.A. & Moore J.W. (1984) Aquatic insects as gan, USA. Hydrobiologia, 75, 73-79. primary consumers. The Ecology of Acjuatic Insects (eds Tuchman N.C & Stevenson R.J. (1991) Effects of selective V. H. Resh and D. M. Rosenberg). Praeger Press, grazing by snails on benthic algal succession. Journal New York. of the North American Benthological Society, 10, 430-443. McAuliffe J.R. (1984) Resource depression by a stream Tuskes P.M. (1977) Observations on the biology of Parargy- herbivore: effects on distributions and abundances of ractis confusaiis, an aquatic pyralid (Lepidoptera: Pyrali- other grazers. Oikos, 42, 327-333. dae). Canadian Entomologist, 109, 695-699. McCormick P,V, (1991) Lotic protistan herbivore selectiv- Tuskes P.M. (1981) Factors influencing the abundance ity and its potential impact on benthic algal assem- and distribution of two aquatic moths of the genus blages, journal of the North American Benthological Society, Parargyractis (Pyralidae). Journal of the Lepidopterists' 10, 238-250. Society 35, 161-168. McCormick PV. & Stevenson R.J. (1991) Mechanisms of Vaughn C.C (1986) The role of periphyton abundance benthic algal succession in lotic environments. Ecology, and quality in the microdistribution of a stream grazer, 72, 1835-1848. Helicopsyche boreaiis (Trichoptera: Helicopsychidae). Oemke M.P. (1983) Interactions between a stream grazer Freshwater Biology, 16, 485-493. and the diatom flora. Proceedings of the 4th International Wiley M.J. & Warren G.L. (1992) Territory abandonment, Symposium on Trichoptera (ed. ]. C. Morse). Dr W. Junk theft, and recycling by a lotic grazer: a foraging strategy Publishers, Tlie Hague. for hard times. Oikos, 63, 495-505. Peterson CG. & Grimm N.B. (1992) Temporal variation Winterbourn M.J. (1990) Interactions among nutrients, in enrichment effects during periphyton succession in algae and invertebrates in a New Zealand mountain a nitrogen-limited desert stream ecosystem, journal of stream. Freshwater Biology, 23, 463-474. the North American Benthological Society, 11, 20-36. Zar J.H. (1984) Biostatistical Analysis, 2nd edn. Prentice- Power M.E. (1991) Shifts in the effects of tuft-weaving Hall Inc., Engiewood Cliffs, New Jersey, midges on filamentous algae. American Midland Natural- ist, 125, 275-285. (Manuscript accepted 17 November 1994)