Hydrobiologia 245: 53-64, 0 1992 Kluwer Academic

1992. Publishers.

Printed

53

in Belgium.

Influence of four grazers on periphyton communities associated with clay tiles and leaves Natalie K. Karouna

& Randall L. Fuller*

Biology Department,

Colgate University, Hamilton,

NY 13346, USA (*author for correspondence)

Received 14 August 1991; in revised form 22 January 1992; accepted 5 February 1992

Abstract This study assessed the individual effects of three mayflies (Paraleptophlebia sp., Ephemerella subvaria McDunnough and Epeorus sp.) and one caddisfly (Psilotreta sp.) on periphyton communities associated with clay tiles and leaves. Algal densities were estimated for leaf discs and tiles from experimental chambers (with individual grazers) and control chambers (i.e., no grazers). Scanning electron micrographs (SEM) of leaf discs and tiles also were taken for all mayily grazing experiments. Densities of algae on leaf discs were two to five times lower than on tiles. Mouthpart morphology influenced how different insects grazed the periphyton community. Paraleptophlebia had ‘typical’ collector-gatherer mouthparts and had no effect on diatom densities associated with leaves whereas diatom densities on grazed tiles were higher than densities on tiles from control chambers. Epeorus had ‘brusher’ mouthparts and had little impact on diatom densities regardless of substratum type. The other two grazers had the blade-like mandibles of a scraper. Psilotreta did not reduce the numerical abundance of diatoms on either substratum, but did alter community structure by significantly reducing densities of stalked Gomphonema olivaceum and large species of Navicula and Nitzschia; densities of smaller diatoms (Achnanthes spp) increased. However, E. subvaria reduced densities of most algal species regardless of size on both substrata and also significantly altered community structure. SEMs of substrata grazed by mayflies showed reductions in fungal hyphae on all grazed leaf discs, decreases in filamentous algal forms on grazed tiles, and greatly shortened stalks of G. olivaceum (Paraleptophlebia only). Thus, periphyton communities are different on leaves versus tiles and grazers with different mouthpart morphologies have varying effects on both algal and heterotrophic microbial community structure.

Introduction Until recently, the distribution and abundance of algae in streams was believed to depend primarily on physical and chemical factors (e.g., light, temperature, current velocity, substratum type, or nutrient availability) (Hynes, 1970; Tuchman & Blinn, 1979; Blinn et al., 1980; Horner & Welch, 1981; Stevenson, 1983). However, more recent studies suggest that periphyton communities in streams are influenced by herbivorous macroin-

vertebrates

(Lamberti

& Resh,

1983; Colletti

et al., 1987; Hill & Knight, 1987; Jacoby, 1987; Steinman et al., 1987; McCormick & Stevenson,

1989). Studies of grazing by snails have reported decreases in periphyton biomass, biovolume of algae, species diversity, and chlorophyll a concentrations as well as significant shifts in algal species composition (Steinman et al., 1987, McCormick & Stevenson, 1989). Stream insect grazers have produced similar responses including increases in primary productivity per unit biomass

54 (Lamberti & Resh, 1983; Hill & Knight, 1987). However, the extent of change in a grazed periphyton community varies with macroinvertebrate grazers. Differences in grazer mouthpart morphology have been used to explain differential effects on periphyton. For example, the radula of snails acts to scrape or rasp both loosely aggregated material as well as adnate diatoms that tightly adhere to substrata. Barnese et al. (1990) have shown that the radular structure of the prosobranch, Elimia livescens, was different from pulmonate snails and this resulted in very different grazing effects by E. Zivescens on the periphyton community compared to pulmonate snails. Mouthpart morphology of aquatic insect grazers is even more diverse, and ranges from mouthparts covered with dense brushes of hairs characteristic of collector-gatherers to blade-like mandibles of scrapers. Comparative studies of aquatic insect grazers have concentrated primarily on scraping caddisflies and mayflies, and results were similar to those observed for snails (Colletti et al., 1987; Lamberti et al., 1987; Steinman et al., 1987). However, less is known about the relative influence of insect collector-gatherers or ‘brushers’ (McShaffrey & McCafferty, 1986) on stream periphyton communities. This study was designed to assess the relative impact of four different insect grazers on periphyton communities in streams. The four grazers were different taxonomically and functionally (based on mouthpart morphology - see below). We also examined grazing effects on different periphyton communities (i.e., one community associated with clay tiles and another on leaf surfaces). We wanted to determine (1) whether or not these two substrata differed in algal species composition/abundance, and (2) the relative impact of different grazers when feeding on periphyton associated with different substrata.

The herbivores The herbivores in our experiments were abundant in local streams near Hamilton, New York where the grazers commonly coexist. The scraping cad-

disfly, Psilotreta sp. (Odontoceridae) has the typical blade-like mandible of a scraper; there are limited setae and no obvious hairs associated with the mouthparts (Figs lA, 2A). The other grazers were mayflies (Ephemeroptera) from three different families. Ephemerella subvaria McDunnough (Ephemerellidae) also has typical scraper mandibles, but in addition, it has short, pectinate setae associated with the labrum and maxillary palps that may also be used to remove material that is more tightly bound to the substratum (McShaffrey & McCafTerty, 1986) (Figs lB, 2B). Epeorus sp. (Heptageniidae) would be classified as a brusher (sensu McShafTrey & McCafferty, 1986) since it has long, pectinate setae associated with the labial palps, and setal fringes on the maxillary palps (Figs lC, 2C). The mouthparts of Paraleptophlebia sp. (Leptophlebiidae) are heavily fringed with long, fine, simple setae (Figs lD, 2D), and would be classified as a collector-gatherer. Thus, these four grazers represent distinctly different mouthpart morphologies. We hypothesized that collector-gatherers would have less of an impact on the periphyton community than scrapers and that scrapers would eliminate the larger diatom species and filamentous algae.

Methods Colonization of substrata Substrata used in these experiments were unglazed clay tiles (6.25 cm’) and red maple leaves (Acer rubrum L.). Clay tiles (N= 18) were arranged on overlapping pieces of duct tape (adhesive side up) to form a sheet (9 cm x 19 cm) equal in size to the face of a brick; the sheet was then secured with rubber bands placed between the tiles and around the brick. Whole red maple leaves collected from the forest floor (within a few days after abscission) were soaked in tap water until soft (< 1 h), and one layer of leaves was positioned on bricks and secured with four rubber bands. Bricks with different substrata (leaves or tiles) were placed in a deep (25 cm), slow-moving (10 cm s - ‘) stretch of Kingsley Creek, 0.4 km

55

Fig. 1. Scanning electron micrographs of the head capsules of four grazers showing mouthparts under low magnification. (a) Psilotretu, (b) Ephemerella subvaria, (c) Epeorus and (d) Paraleptophlebia. (L - labrum with pectinate setae; LP - labial palps; MP - maxillary palps).

above Lebanon Reservoir, Lebanon, N.Y. This section of Kingsley Creek is a second order stream with a partial canopy; substrata consisted of large cobbles of slate and the average water depth was 15-20 cm. For the first experimental trial, four bricks with tiles and four bricks with leaves were placed in the stream on 17 Sept. 1988 and retrieved one month later for use in the Psilotreta grazing experiments. In mid-October, a second set of bricks with tiles and leaves (6 bricks/ substratum type) were placed in the same stretch of Kingsley Creek, and allowed to colonize for one month prior to use in mayfIy grazing experiments.

Experimental

design

Grazing experiments were conducted in the laboratory in round, recirculating Plexiglas streams (15 cm diameter, 20-25 cm s - ’ current velocity) similar in design to those described by Mackay (1981). Spring water was used in the chambers. Once substrata were naturally colonized by algae and other heterotrophic microbes, leaf discs (6.25 cm’) were cut from the whole leaves using a cork borer (2.8 cm dia.). A thin (2 mm) circular strip of cork gasket was cut to fit into the bottom of the circular chambers and leaf discs were then secured with pins to the cork. Ten sub-

56

Fig. 2. subvaria,

Scanning electron micrographs of the mouthparts of all four grazers under high magnification. (a) Psilotretu, (c) Epeorus and (d) Paraleptophlebia.

strata (either leaf discs or tiles) were placed in a single chamber. In the Psilotreta trials, four chambers were used for each substratum type; three experimental chambers each containing 10 fifth instars of Psilotreta and one control chamber with the substrata but no Psilotreta larvae. For the mayfly trials, again three experimental chambers containing a single mayfly species and one control chamber (no mayfly larvae) were used per substratum type; however, densities for each mayfly species were different. For Epeorus, 20 early instars per chamber were added whereas for E. subvaria and Paraleptophlebia only 15 larvae were introduced per chamber. Mayfly densities were higher than for Psilotreta to reflect natural densities in Kingsley Creek (ca. 1300 Psilotreta

(b) Ephemerellu

m - 2 versus 1900-2200 m - 2 for individual mayfly species, R. Fuller, unpublished data); densities in the experimental chambers were different among mayllies to maintain equal biomass (ca. 5 mg total dry weight/chamber) for all mayfly species among treatment and still remain within the range of natural maylly densities observed in Kingsley Creek. All chambers for both trials were placed in an environmental room at 4 “C under a 12:12 light-dark photoperiod and each trial lasted 10 d. Algal identification and quantification was made for leaf discs and tiles initially (prior to use in chambers), and for substrata from control and experimental chambers at the end of an experimental period. At least three randomly selected

57 substrata from each chamber were used in our analysis. Individual tiles and leaf discs were scraped with a scalpel, scrubbed with a toothbrush and this material placed into separate bottles for each chamber (substrata from the control chambers were kept in separate bottles 3 substrata/bottle). Each substratum was then placed in a Whirlpak bag with 30 mls of distilled water and sonicated in a Ultramet III sonic cleaner water bath for 17 min.; the sonicated material was added to the bottle containing the scrapings, and the total sample was preserved with Lugol’s solution. Algae were quantified by filtering known volumes on to Millipore filters (0.45 pm pore size); the filters were dried, cleared with immersion oil and permanent slides made. Identifications were made with a Zeiss Standard Lab 06 phase contrast microscope at 1000 x , and algal cells were enumerated at 100 x . At least 200 cells per slide were counted which required a minimum of 5 Whipple grids. Scanning electron microscopy (SEM) was used to examine the surface of the initial, control and grazed substrata from the maylly trials and for characterizing grazer mouthpart morphology. Leaf discs and tiles were preserved in FAA (formalin, ethyl alcohol and acetic acid). Substrata were taken through a graded series of ethanol before being dried in a critical point dryer using CO, as a transitional fluid and finally sputter coated with gold-palladium. Grazers were preserved in FAA but were dried only in the graded ethanol series and then air-dried prior to being sputter coated. Examination of all grazer mouthparts and substrata was made with an IS1 Model 40 scanning electron microscope. Grazing effects were assessed separately for the Psilotretu trial and the mayfly trials because of differences in the time frames for algal colonization. Algal densities for both trials were logarithmically transformed (In n + 1; where, yt = algal density). For Psilotreta, comparisons of algal density for the more abundant taxa were made for each substratum type between control (n = 1, substrata were replicates) and experimental conditions (n = 3, chambers were replicates) using a t-test. Because the control chambers were the

same for all of the mayfly trials, comparisons among the mayfly species and controls for each algal taxon/substratum type were made using a oneway ANOVA (SAS GLM procedure). When significant differences were found, comparisons between the means were determined using Scheffe’s multiple comparison test. Comparisons of algal density between leaf and tile substrata from control chambers were made using a t-test. Statistical results should be viewed with some caution because control chambers were not true replicates (i.e., pseudoreplicates, Hurlbert, 1984).

Results Algal communities on leaves versus tiles In our experiments, there was a distinct difference between the algal communities on leaves and tiles. Although species composition and proportions of species did not differ significantly between these two substrata, leaf discs supported significantly lower (p < 0.00 1, t-test, df = 4) algal densities than tiles from control chambers and from substrata analyzed prior to use in experiments (Table 1). Although most filamentous algal forms were destroyed during the scraping/sonification process, examination of gently scrubbed tiles and leaves suggested an abundance of filamentous algae (primarily Oedogonium) on control tiles, but no similar representation of filamentous algae was apTable 1. Mean ( + 1 SE; n = 3) Diatom1 densities on leaf disks and tiles from control chambers at the end of an experiment and from substrata prior to their use in experiments (Initial).

Total number of diatoms x lo3 cmd2

trials Initial Final control chambers

Leaves

Tiles

2641&312 3254+255

5275& 111 4519k236

Psilotreta

Mayfly trials Initial Final control chambers

163 & 59 613 + 62

3732k 138 3296~ 23

58

parent on leaf discs from control chambers. Some differences in the relative abundances of specific taxa were observed between algal communities on leaves versus tiles, but these differences were not consistent when comparisons were made between the Psilotreta and mayfly trials.

Grazing eflects on algal community structure

Grazer mouthpart morphology significantly influenced the impact that each grazer had on algal community structure. Algal communities on leaf discs grazed by Psilotreta showed significant reductions (p< 0.01, t = 4.98, df = 4) in two, large Navicula and Nitzschia species (Naviculaperegrina (Ehr.) Kutz., Navicula exigua greg., Nitzschiapalea

Psilotreta

grazing

leaf

discs

[7 m

440 330 220 cl

1 110

6

0

l-7 0

G. oliv.

N. exlg. N.

N.

pere.

Control Grazed

(Kutz.) and Nitzschia f@ixmis (W.S.))(Fig. 3). However, there was no effect (p> 0.20, t = 1.48, df = 4) on densities of two small Achnanthes species (Achnanthes lanceolata (Breb.) Grun. and Achnanthes minutissima Kutz.) and the adnate diatom, Cocconeis when compared with algal communities on control leaf discs (Fig. 3); there also was a significant increase (p< 0.05, t = 2.81, df = 4) in Achnanthes linearis (W.Sm) Gnu-r. (another small diatom species). Tiles grazed by Psilotreta had reduced densities of the same large species of Navicula and Nitzschia (p < 0.05, t = 3.1, df = 4) (Fig. 4); in addition, the density of the small diatom, A. lanceolata increased (p< 0.05, t = 2.94, df = 4) whereas densities of other algal species were not influenced (Fig. 4). Grazing by E. subvaria had the greatest impact on algal communities regardless of substratum type. On grazed leafdiscs, all algal species showed significant decreases (p < 0.05) in density when compared with controls except A. linearis, which

** ILL

palea

N.

Psilotreta

grazing

tiles

220

220

165

165

0 m

Control Grazed

I

fili.

X 37501

03 4401

T G. oliv.

N. exig.

X

N.

palea

6000 4000 1 3000

cocc.

* 2000

A. line.

A. min. Totals lane. Fig. 3. Mean densities of select diatom species on ungrazed leaf disks and disks grazed by Psilotreta. (Bars are 1 S. E. of the mean; * = significantly different from adjacent bar at p= 0.05; G. oliv. = Gomphonema olivaceum, N. pere = Navicula peregnka, N. exig. = Navicula exigua, N. palea = Nitzschia palea, N. fili = Nitzschiafiliformis, Cocc. = Cocconeis, A. lane. = Achnanthes lanceolata, A. line = Achnanthes linear& A. min. = Achnanthes minutissima, Totals = total diatom numbers). A.

N. fib.

N. pere.

1000 AIlLrlLM

cocc.

A. lone.

0 A. line.

A. min.

Ill Totals

4. Mean densities of select diatom species on ungrazed clay tiles and tiles grazed by Psilotreta. (Bars are 1 S.E. of the mean, * = significantly different from adjacent bar at p = 0.05; see Fig. 3 heading for an explanation of species abbreviations). Fig.

59 Ephemerella 75

grazing

leaf discs 75

1

1

0 D

Control Grazed

$1‘ii,;ii 1:’ *iI,*, N. orypto.’

G. oh.

X

N. palea

N. fili.

N. pew.

v,

b 75-

n

E 2

Ephemerella

on tiles

0 m

500

Control Grazed

400 300 c\l 1 200

E 0 100 m i 00 V

N. Ct-#XO.

5000

;200,

N. palea

N. fili.

1

4000

3 150

Z

-

ti. OIIV.

N. pere.

X

t 50.

grazing

i

3000 1

l-l

cooo. A. line. A. min. Totals Fig. 5. Mean densities of select diatom species on ungrazed leaf disks and disks grazed by Ephemerella subvaria. (Bars are 1 S.E. of the mean; * = significantly different from adjacent bar at p = 0.05; see Fig. 3 heading for an explanation of species abbreviations; N. crypto. = Navicula cryptocephala).

cooo. A. line. A. min. Totals Fig. 6. Mean densities of select diatom species on ungrazed clay tiles and tiles grazed by Ephemerella subvaria. (bars are 1 S.E. of the mean; * = significantly different from adjacent bar at p = 0.05; see headings of Figs. 3 and 5 for an explanation o species abbreviations).

was not affected (Fig. 5). Comparisons of algal communities between grazed and control tiles also showed significant reductions (~~0.05) in all algal species excluding A. linearis and A. minutissima (Fig. 6). No algal species increased in abundance as a result of E. subvaria grazing. The brusher, Epeorus, had less impact on periphyton communities. When Epeorus grazed leaf discs, both Gomphonema olivaceum Kutz. and N. filiformis densities were significantly reduced (p < 0.05), whereas N. palea and A. linearis densities increased significantly (p< 0.05) (Fig. 7); densities of other algal taxa were not influenced by Epeorus grazing activity. The same algal species growing on clay tiles showed similar responses to Epeorus grazing (Fig. 8). Paraleptophlebia had no negative impacts on algal densities. However, we observed a significant increase (p < 0.05) in the density of N. palea on grazed leaf discs while densities of all other algal species were not influenced (p> 0.20) by

comassociated with clay tiles grazed by Paraleptophlebia showed increases in densities of A. linearis and A. minutissima which resulted in an increase in overall algal density; no algal species showed any decreases in density (Fig. 8).

Paraleptophlebia grazing (Fig. 7). Periphyton

munities

SEM studies of grazing effects by maypies

Leaf discs from control chambers had thick accumulations of flocculent material, numerous fungal hyphae, and diatoms (Fig. 9a). Leaf discs grazed by mayflies showed varying degrees of disruption of this community. Epeorus and Paraleptophlebia removed much of the flocculent and filamentous matter but they had less impact on the algal community (Fig. 9c,d). In contrast, E. subvariu eroded most of the material attached to the leaf surface exposing the striated texture of the leaf (Fig. 9b).

60 Mayflies 75

grazing

leaf discs

1

75

0 6Dl W

1

Mayflies

Control Epeorus Pa~alepto.

grazing

tiles

0 f?iD i?Z3

500

Control Epeorus Parole@0

400 50

cl k M 0 7 X

cl k o100 M 300

200

25 0

G. oliv.

0c-

N. crypto.

N. palea

N. fili.

N. pere.

i! :

E z’

75,

8801

0I

X (II L

G. oh.

N. crypto.

N. palea

N. fili.

N. pere.

*

~200

E 2150 50 100 25 50

Cocc.

A. line.

A. min.

cocc.

Totals

Fig. 7. Mean densities of select diatom species on ungrazed leaf disks and disks grazed by Epeorus and Paraleoptophlebia. (Bars are 1 SE. of the mean; * = significantly different from the other two adjacent bars at p = 0.05; see headings of Figs. 3 and 5 for an explanation of species abbreviation; Paralepto. = Paraleptophlebia).

Similar effects were seen with SEMs of tiles except there were fewer fungal hyphae on control tiles and more filamentous algae and stalked G. olivuceum (Fig. lOa). Tiles grazed by Puraleptophlebia had few filamentous algae and the G. olivaceum cells present had very short stalks relative to stalks from control titles; also, larger diatom species (e.g., Cocconeis, N.Jiliformis, etc.) were abundant (Fig. 10d). Similar effects were seen for Epeorus (Fig. 10~). Tiles grazed by E. subvariu had few diatoms present and no filamentous algae (Fig. lob).

Discussion There were distinct differences in the effects of each grazer on algal species composition and/or abundance. Also, we observed significantly lower diatom densities in periphyton communities as-

A. line.

A. lane.

A. min.

Totals

Fig. 8. Mean algal densities of select diatom species on ungrazed clay tiles and tiles grazed by Epeorus and Paraleptophlebia. (Bars are 1 SE. of the mean; * = significantly different from the other two bars at p = 0.05; see headings of Figs. 3 and 5 for an explanation of species abbreviations; Paralepto.

= Paraleptophlebia).

sociated with leaf surfaces versus clay tiles, however the species composition of each substratum type was similar. Potential explanations for differences in algal densities between these two substrata may include physical differences such as surface texture or chemical/biological factors that might be different between substrata. Leaves release significant amounts of dissolved organic matter (DOM) upon entering streams (cu. 25 y0 of total leaf weight lost in the first 24 h, Lock & Hynes, 1976); this DOM includes tanins and other organic chemicals that may inhibit or delay algal growth. Furthermore, nutrients may become limiting to algae because of more efficient uptake by microheterotrophs (Paul & Duthie, 1989). SEMs of leaves from control chambers showed accumulations of flocculent material and many fungal hyphae; the latter may negatively affect algae through shading, chemical inhibition or nutrient competition. Regardless of the mecha-

61

Fig. 9. Scanning electron micrographs of leaf surfaces from (a) control (b) Ephemerella subvaria, (c) Epeorus and (d) Paraleptophlebia.

nism(s), leaf substrata did support lower algal population densities than clay tiles. The variable effects of different grazers on periphyton communities may have been related in part to differences in mouthpart morphology. Psilotretu and especially E. subvaria markedly reduced algal densities and these species both fit the general ‘scraper’ category. Furthermore, when E. subvaria grazed the periphyton on leaf discs there was significant erosion of the leaf surface (sometimes as much as 80% of the total surface area had been eroded). Thus, E. subvaria may act as a shredder when feeding on leaf substrata, however grazed patches on leaves demonstrates their ability to scrape material without totally eroding the leaf surface. Psilotreta did not erode

chambers (ungrazed)

and leaf surfaces grazed by

leaves to the same extent as E. subvaria, and it did not have as great an impact on the periphyton community regardless of substratum type. It is difficult to ascertain whether the mandibles, the pectinate setae, or some other mouthpart are eroding leaf surfaces, but it is clear that E. subvaria is more effective at removing materials (algae or upper layers of leaf surfaces) from substrata than the other grazers tested. Algal species composition also was influenced by grazing of Psilotreta and E. subvaria. Psilotreta successfully removed filamentous algae (Oedogonium) and the larger Navicula and Nitszchia species, however, smaller species (e.g., Achnanthes spp) and the adnate diatom, Cocconeis did not show reductions in density. This periphyton com-

62

Fig. 20. Scanning electron micrographs of clay tile surfaces from (a) control chambers (ungrazed) and tile surfaces grazed by (b) Ephemerella subvaria, (c) Epeorus and (d) Paraleptophlebia.

munity was similar in physiognomy to periphyton communities that were grazed by a heptageniid mayfly (Colleti et al., 1987) and the snail Juga silicula (Lamberti et al., 1987; Steinman et al., 1987). In contrast, grazing by E. subvariu totally disrupted the entire periphyton community. All of the more abundant algal species on leaves were reduced by E. subvaria probably because of erosion of the leaf surface. Only on clay tiles did we observe the smaller Achnanthes species escape significant consumption by E. subvariu; all of the larger algae were reduced. The surviving periphyton community consisted of a monolayer of small, sparsely populated diatoms similar to a periphyton community described by Lamberti et al. (1987) that had been grazed by a scraping

caddisfly. Our data, in conjunction with those of other studies, suggest that individual scrapers as members of the same guild do not always produce the same levels of disturbance in periphyton communities. Instead, sensitivity of algal species to grazing is specific for individual scraping macroinvertebrates. Epeorus had less of an effect on the diatom community than E. subvmiu or Psilotreta in part because of its brusher mouthpart morphology. Epeorus reduced densities of only a few, large algal species; however, most algae were not influenced by this grazer or densities increased (especially for some of the smaller diatoms). Although all maylly grazers in our experiments were in their early to middle stages of development,

63 E. subvaria was slightly larger than Epeorus and this may have influenced the impact each grazer had on the periphyton community. We maintained similar grazer biomass among the mayflies but similar biomass did not result in similar impacts on periphyton communities. In the future, investigations concerning the impact of different developmental stages should be considered. The increase in algal density resulting from Paraleptophlebia grazing was from an increase in A. minutissima, a small diatom species. This species is known to increase in relative abundance in grazed periphyton communities but this increase usually is in conjunction with decreases in the larger, overstory diatoms (Sumner & McIntire, 1982; Colletti et al., 1987; Hill & Knight, 1987; Steinman et al., 1987). In our study, no diatom species was reduced by Paraleptophlebia grazing, although the filamentous alga, Oedogonium was not found on grazed tiles. Also, SEMs demonstrated that G. olivaceum had very short stalks relative to stalk length on control tiles. Thus, the periphyton community was disrupted to some extent, and this may have been sufficient to allow A. minutissima to increase. In conclusion, our results suggest that all primary consumers do not necessarily reduce their food source. Whether other collector-gatherers besides Paraleptophlebia can stimulate growth of periphyton is not known but it warrants further examination. Also, our laboratory experiments and those of others (Lamberti et al., 1987; Colletti et al., 1987; Steinman et al., 1987) suggest the potential for marked reductions in periphyton abundance by grazers, but more complex field studies should be conducted to examine the influence of different development stages of grazing macroinvertebrates on periphyton communities.

Acknowledgements We would like to thank Dr R. A. Hoham for his assistance with the identification of the algae and Dr W. J. Oostenink for assistance with the SEM photographs. Dr R. G. Wetzel and an anonymous reviewer made many valuable comments on

an earlier draft of this manuscript. The Colgate Research Council provided funds for the SEM work and construction of the experimental chambers.

References Barnese, L. E. & R. L. Lowe, 1990. Comparative grazing efficiency of pulmonate and prosobranch snails. J. N. Am. Benthol. Sot. 9: 35-44. Blinn, D. W., A. Fredericksen & V. Korte, 1980. Diatom community structure on three different rock substrata in a lotic system. Brit. Phycol. J. 15: 303-310. Colletti, P. J., D. W. Blinn, A. Pickart & V. T. Wagner, 1987. Influence of different densities of the mayfly grazer Heptageniu criddlei on lotic diatom communities. J. N. Am. Benthol. Sot. 6: 270-280. Hill, W. R. & A. W. Knight, 1987. Experimental analysis of the grazing interaction between a maylly and stream algae. Ecology 68: 1955-1965. Horner, R. R. & E. B. Welch, 1981. Stream periphyton development in relation to current velocity and nutrients. Can. J. Fish. aquat. Sci. 38: 449-457. Hurlbert, S. H., 1984. Pseudoreplication and design of ecological field studies. Ecol. Monogr. 54: 187-211. Hynes, H. B. N., 1970. The ecology of running waters. Univ. Toronto Press, Toronto. Jacoby, J. M., 1987. Alterations in periphyton characteristics due to grazing in a Cascade foothill stream. Freshwat. Biol. 18: 495-508. Lamberti, G. A. & V. H. Resh, 1983. Stream periphyton and insect herbivores: an experimental study of grazing by a caddisfly population. Ecology 64: 1134-l 135. Lamberti, G. A., L. R. Ashkenas, S. V. Gregory & A. D. Steinman, 1987. Effects of three herbivores on periphyton communities in laboratory streams. J. N. Am. Benthol. Sot. 6: 92-104. Lock, M. A. & H. B. N. Hynes, 1976. The fate of ‘dissolved’ organic carbon derived from autumn-shed maple leaves (Acer saccharum) in a temperature hard-water stream. Limnol. Oceanogr. 21: 436-443. Mackay, R. J., 1981. A miniture laboratory stream powered by air bubbles. Hydrobiologia 83: 383-385. McCormick, P. V. & R. J. Stevenson, 1989. Effects of snail grazing on benthic algal community structure in different nutrient environments. J. N. Am. Benthol. Sot. 8: 162-172. McShatTrey, D. & W. P. McCafferty, 1986. Feeding behavior of Stenacron interpunctatum (Ephemeroptera:Heptageneidae). J. N. Am. Benthol. Sot. 5: 200-210. Paul, B. J. & H. C. Duthie, 1989. Nutrient cycling in the epilithon of running waters. Can. J. Bot. 67: 2302-2309. Steinman, A. D., C. D. McIntire, S. V. Gregory, G. A. Lamberti & L. R. Ashkenas, 1987. Effects of herbivore type and

64 density on taxonomic structure and physiognomy of algal assemblages in laboratory streams. J. N. Am. Benthol. Sot. 6: 175-188. Stevenson, R. J., 1983. Effects of current and conditions simulating autogenically changing microhabitats on benthic diatom immigration. Ecology 64: 15 14- 1524.

Sumner, W. T. & C. D. McIntire, 1982. Grazer-Periphyton interactions in laboratory streams. Archiv fur Hydrobiologie 93: 135-157. Tuchman, M. L. & D. W. Blinn, 1979. Comparison of attached algal communities on natural and artificial substrata along a thermal gradient. Brit. Phycol. J. 14: 243-254.