I 1 Smrrgm, Okrober 1992 Arch. Hydrobiol. 125 I 4 385-410 Hydrologic and trophic controls of seasonal algal blooms in northern California rivers By MARY E. Po-, University of California at Berkeley' With 15 figures and 1 table in the text Abstract Ckzdophoru glomeruta L., a dominant macroalga in lakes and rivers worldwide, undergoes a marked bloom-detachment-senescence cycle in unregulated rivers of north- ern California with natural winter flood, summer drought flow regimes. In two re- gulated channels which probably did not experience scouring floods, however, low standing crops of attached Cladopboru turf persisted throughout the year. The contrast between Ckzdophora phenology in regulated and unregulated rivers suggests that Cla- dophoru cycles may be extrinsically driven by factors related to the hydrograph. Pre- liminary data on seasonal patterns of animal abundance in regulated and unregulated channels suggest that winter flooding promotes Cladophoru blooms in rivers by reducing consumer densities. A working hypothesis relating hydrologic regime, food chain length, and algal phenology in rivers is advanced. This hypothesis predicts that pro- nounced algal bloomdetachment-senescence cycles will occur in unregulated rivers in Mediterranean climates following winter flooding, and that blooms will not occur in the absence of flooding in regulated channels, or in natural rivers during prolonged drought. Introduction In rivers which are sunlit, rock-bedded, and clear at low flow, attached algae are often dominant components of ecological communities. Cludophoru gfomeruta L., a filamentous green, may be the most common and cosmopolitan macroalga in temperate rivers throughout the world (BLIJM, 1956; WHITTON, 1970; WHAFSE et al., 1984). In sunlit rivers of the western United States, Cla- dophoru provides much of the physical structure in the habitat during the low flow season, and plays a driving role in food web dynamics (LAMBERTI & RF.SH, 1983; Gw, 1987; FEMINELU et al., 1989; POWER, 1990 a, b). In these rivers and elsewhere (notably in the Laurentian Great Lakes), Cludophoru blooms create severe management problems (BLUM, 1956, 1982; BELLIS, 1967; AUER et al., 1982; MILLNER & SWEENEY, 1982; GOLDMAN & HORNE, 1983). Despite their ecological and economic importance, factors regulating growth, detachment, and senescence cycles of Ckzdophoru are still not well understood (WHI-ITON, 1970; NEIL &JACKSON, 1982). ' Author's address: Dept. of Integrative Biology, University of California at Ber- keley, Berkeley, CA. 94707, USA. 25 jrchiv f. Hydrobiologie, Bd. 125 0003-9136/92/0125-0385 $6.50 0 1992 E. Schwcizerbm'scbe Vcrlagsbuchhmdlung. DJDXI Stuttgm 1 386 Mary E. Power This study addresses three questions about Ckdopbora in northern Cali- fornia rivers: 1. What is the seasonal cycle of Ckzdqhora in rivers with natural summer drought, winter flood hydrographs? 2. How does this cycle compare with Cladophoru phenology in regulated rivers with artificially stabilized flow? 3. How do abundance patterns of invertebrates and smallvertebrates asso- ciated with Cladophoru change seasonally in regulated and unregulated chan- nels? To develop hypotheses about factors governing timing, magnitude, and duration of algal blooms and mat detachment, I monitored Cludophora, physical factors, nutrients, and associated biota in four unregulated and two re- gulated rivers in northern California. Results from this survey complement ex- perimental studies on smaller spatial scales of controls on this dominant river dga (LAMBEXTI & ~SH, 1983; LIGON, 1986; FEMINELLA et d., 1989; POWER, 1990 a, b). Study Sites I monitored six rivers near sites gaged by the United States Geological Survey (USGS) (Fig. 1). These rivers differed in discharge, exposure to sun, and land use in their watersheds. Two rivers were regulated upstream of the monitored sites, either by a dam (Dry Creek) or by a water diversion (from the Eel River to the East Fork of the Russian River) and had artificially sustained summer baseflow (Fig. 2 e, f, Fig. 3). Dry Creek ex- perienced stable low flow throughout the year. Dividing discharge by channel drainage area indicates that releases from the Warm Springs Dam above Dry Creek were consid- erably less than natura winter flows in channels of similar drainage area in this region (Fig. 4). Consequently, the bed of Dry Creek was stable throughout the period of study. To compare the relative intensity of bed movement for the remaining five rivers, I used two empirical generalizations about gravel bedded riven. First, bankfull discharge typically has a recurrence interval of about 1.5 years (e.g. DUNNE & LEOPOLD, 1978). Sec- ond, significant gravel bed mobility often does not occur until the flow is close to bank- full discharge (Pmmn, 1978). In the East Fork Russian River, elevated winter flows oc- curred, but peak discharge in 1989 remained below bankfull discharge as estimated from flood frequency analyses (Fig. 5). Hence, it is likely that little bed movement occurred. This inference is consistent with visual observations that the bed of the East Fork Russian River showed no evidence of scour over the period of study. In contrast, the other four monitored rivers, which were unregulated (the South Fork Eel River, Elder Creek, Outlet Creek, and the Middle Fork Eel River), experienced the natural summer drought, winter flood hydrograph typical of streams in regions with Mediterranean climates (Fig. 2 a-d). Each unregulated river experienced flows equal or greater than bankfull discharge during the 1988- 1989 study period (Fig. 5). Visual observations over the winter confirmed that beds in these four rivers moved, and were subject to consid- erable scour. During the summer low flow season, from June through September, the four unre- gulated rivers experienced low or no discharge (Fig. 3). In contrast, flows in Dry Creek and the East Fork Russian River were maintained at levels ranging from 2-4 m's-' over Hydrologic and trophic controls of seasonal algal blooms 387 2 m Eureka \ Fig. 1. Location of the six study reaches (numbers 1,2,4-7) and of a diversion from the Eel River to the Russian River (3) that stabilizes flow in the East Fork Russian River (site 2). Drainage areas above the six monitored sites are, for the two regulated channels: (1) Dry Creek (USGS 11465000): 562km', and (2) East Fork Russian River (USGS 11461500): 239 km2. For the four unregulated channels, drainage areas are (4) Outlet Creek (USGS 11472200): 417kmz; (5) Middle Fork Eel (USGS 11473900): 1929 km2; (6) South Fork Eel (USGS 11475500): 114km' and (7) Elder Creek (USGS 11475560): 17 km'. the summer (Fig. 3). During October, the month of minimum discharge in natural chan- nels, most monitored sites had current velocities slower than 5cms-'. In the two re- gulated rivers, current velocities during October ranged from 0 to > 50cms-', and were fairly evenly represented among the monitored sites (Fig. 6). Despite the large variation in drainage areas and discharges of the six rivers (Fig. 1, legend), in summer months they were all easily waded. 388 Mary E. Power Hydrologic and trophic controls of seasonal algal blooms 389 S. Fk. Eel, 1988-1989 -mwl outlet Creek. 1988-1989 :a XpBi91Pp3~so Dry Creak, 1988-1989 I\ -m E. Fk. Russlan RIver, 1988-1989 Fig.2. Monthly discharge records (from the USGS) for the four natural (a-d) and two regulated (e, f) channels. Records for the South Fork Eel are from a currently monitored station near Leggett, CA, where the drainage area is 642 km', 5.6 times larger than at the study site near Branscomb, where USGS monitoring was discontinued. During 1968 - 1970 when Branscomb and Leggett stations were both monitored by the USGS, discharges were highly correlated (0.93 < r < 1.00 for 18 of 24 consecutive months, 0.27 < r < 0.78 during four transitional spring or fall months (May, June, August, and September). Discharge at Leggett was 5-6 times greater than at Branscomb during this period. The two regulated rivers also had similar proportional size distributions of bed sedi- ments (Fig. 7), but otherwise differed considerably in their physical and chemical con- ditions. Dry Creek is an open, sunlit stream bordered by scrubby willows (Sulzx spp.) and alders (Alnus rbombifoliu). The creek flows from the Warm Springs Dam impound- ing Lake Sonorna through vineyards with little topographic relief, and joins the Russian River (Fig. 1). Dry Creek receives, in addition to agricultural runoff, effluent from a Low Flow Discharge m P -0- SouthFkEel - ElderCreek - OutletCreek - Middle Fk. Eel East Fk. Russian May Jun Jul Aug Sep Fig.3. Mean monthly discharge during the summer low flow season in the four unre- gulated (solid lines) and two regulated (dashed lines) channels. Note that in August and September, when there was little or not flow even in the large unregulated rivers, flow continued in the two regulated channels. salmon hatchery just below Warm Springs Dam, upstream from monitored sites. Both sources contribute nutrients to the stream, and probably account for its relatively elevated levels of nitrate (Fig. 8). The East Fork Russian River near Ukiah receives water diverted from the Eel River through the Potter Valley Diversion (Fig. 1). The East Fork Russian does not receive agricultural runoff, but human habitations occur very close to the river just upstream from the monitored cross sections. This river had the second highest measured level of nitrate of the six studied here (Fig. 8). Insolation of the E. Fk. Russian, however, was much less than at Dry Creek due to shading during mornings and afternoons by tall alder trees and narrow valley walls. Outlet Creek is not regulated, but is otherwise heavily impacted by humans. Many human dwellings occur along the creek, which also receives sewage effluent from the town of Willits (Fig. 1). Of the unregulated streams, Outlet Creek has the highest meas- ured nitrate levels (Fig. 8). Outlet Creek is extremely open and sunlit, as the active, boulder-strewn channel kept open by winter floods is much wider than the wetted chan- nel during the summer low flow period. This is also the case along the monitored reaches of the Middle Fork Eel. The land around the Middle Fork Eel is sparsely settled by humans, and subject to light cattle grazing. Crystal-clear water and low nutrient levels (Fig. 8) suggest little human impact on this river at the monitored site. The South Fork Eel River flows for 5 km through a 3200 hectare forest preserve (the Northern California Coast Range Preserve) and is also relatively undisturbed by humans, al- though sparse settlements and a sawmill occur upstream from the monitored site. As in Outlet Creek and the Middle Fk., winter floods open a wider channel in the S. Fk. Eel than is wetted during low flow, so much of the river bed is sunlit. However, much of the channel is shaded during early morning and late afternoon by a tall bordering forests of mature Douglas fir (Psedotsugu menziesii) and redwood (Sequoia semperuirens), and by valley walls. Elder Creek is the least disturbed of all the monitored rivers, with the largest undisturbed Douglas fir forested watershed remaining in California (TRUSH, 1991). Its watershed was sparsely settled by homesteaders at the beginning of the cen- tury, but presently experiences minimal human impact. Elder Creek is shaded for most of its length by steep valley walls and oak (Lithocurpus densifonrs and Quercus spp.) and Douglas fir-dominated forest. 390 Mary E. Power South Fork Eel Outlet Creek 150 C -m Dry Creek e -mean 150 -rma -min I Elder Crook W a m g 1w East Fork Ruaalan f -mm W 2 150 < -mx Fig. 4. Monthly discharge from channels, divided by channel drainage area, to show the amount of precipitation discharged as runoff, illustrating the effect of storage by the Warm Springs dam in depleting runoff in Dry Creek. Methods I established three permanent cross-stream transects along monitored reaches of each river. At points at 0.5-m or 1.0-m intervals along each transect, I measured water depth, current velocity (with a low velocity threshold rotary current meter (Roy Olund, Dept. of Oceanography, Univ. Wash., Seattle Wa.)) and described substrate composition (classified by median particle size according to the modified Wentworth scale (HYNES, 1970, p. 24)). Using a face mask for underwater observation, I visually es- timated an area of 10 x 1Ocm' under each sampling point, and noted: the genera of dominant and subdominant algae; their height (I used a meter-stick or a 15cm ruler to r a ... 4 a s m 4 .c .s n X m n Hydrologic and trophic controls of seasonal algal blooms X m n 200 190 180 I 70 160 150 140 130 120 110 'O0#.O 1.1 1.2 1.3 1.1 1.5 1.6 1.7 1.8 1.0 2.0 Recurrence Interval (years) 1000 900 800 700 600 500 400 300 200 100 1.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.1 1.0 2.0 m R X I 391 20 18 16 14 12 10 8 6 4 2 4.0 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.0 2.0 Recurrence Interval (years) Recurrence Interval (Years) 500 450 400 350 300 250 200 150 100 50 '1.0 1.1 1.2 1.3 1.4 1.5 1.8 1.7 1.1 1.9 2.0 X a Recurrence Interval (years) Recurrence Interval (Years) Fig. 5. Recurrence intervals for momentary peak discharge showing the 1989 peak dis- charge in relation to that with a 1.5 year recurrence interval, which gives an estimation of the bankfull flow (DUNNE & LEOPOLD, 1978). Flood recurrence intervals were es- timated from 40 years of record in the E. Fk. Russian River; from a 25 year record in the S. Fk. Eel; from a 23 year record in Elder Creek; from a 34 year record in Outlet Creek; and from a 25 year record in the Middle Fk. Eel. measure modal height of filaments or maximum diameter of floating algal mats); and their density and condition, (categorized as indicated in Table 1; see POWER et al. (1985) 392 Mary E. Power South Fork Eel River 31 40 21 40 11 -20 5-10 <5 0 0.0 0.2 0.4 0.6 0 8 J 10 Elder 31-50 21.30 0.0 0.2 0.4 Creek -! 0.6 0.6 1.0 Hydrologic and trophic controls of seasonal algal blooms 393 00 L 0.1 0.2 0.3 0.4 0.6 0.6 0.7 Ylddle Fork Eel River e . '5 > <5 0 0 0.0 0.2 0.4 0.6 0 8 1 0 0.0 0.2 0.4 0 6 0.8 1.0 > 50 >E4 u 31.50 u 21-30 . . 31-y1 21.30 E - , 11.20 * 11-20 5-10 :: 5.10 I .- - e <5 > <5 0.0 0.2 0.4 0 6 0 8 1.0 00 0.2 0.4 0 6 0 8 1.0 Proportion of Sites Proportion of Sites Fig. 6. Proponional distributions of current velocities measured at sampling sites during October, at the time of lowest stream flow during the study period. Sample sizes (num- bers of sites sampled) were S. Fk. Eel: 49; Elder Creek: 32; Outlet Creek: 48; Middle Fk. Eel: 73; E. Fk. Russian: 50; and Dry Creek: 76. and POWER & STEWART (1987) for further methodological details). I also recorded the presence of macroscopically conspicuous animals within the observed area under each point, and collected representative samples of unfamiliar algae, macrophytes, and ani- mals for identification. At each transect site, I collected water samples in acid-cleaned 1-liter polyethylene botrles. To assess spatial variarion in nutrient concenrrations, six water samples were in- itially collected at each cross-section, near the surface and near the bed at three cross- channel positions. No cross-stream spatial variation was detected, even at low flow (I)OWI.IL Iwo~~), SO .Iitcr I'IXX WIIY (1111. t)r Iwt) s.III11dl.s pcr tr.i~lscl.t WCI'C COIIC~.IC~I OII .I 0.0 01 02 03 Ob 0.5 06 0.7 I Boyld.r-B1rosk I East Fork Rusrlan Rlver (Regulated) \\\\\\\\\\\\\\ ,,,,,,,,,,,,,, 11 \\\\\\\\\\\\\\ : P.M.. \'\'\'\'\'\'\'\'\'\'.'.'.'.' 0 0 0 I 0 2 0 3 0.4 0.3 0.6 0.7 Praporllon 01 SII.. \\\\\\\\\....\.\\ ,,,,,,*,,,,##,,, \\\\\\\.\\.\\\\\\ ,,,,,,,,,,,,,,,, \\\\\\.\\\\\\\\\\ u 0.0 0 I 0.2 0.3 0 4 0 6 0.8 0.7 . . . . . . . . , , I , , , , , , .\..\\\\ e- ,',',','#',',','# *\\\\\.\.. ,,,,,,,,,,, :: !B P*L*. ..................... \\\\\.\\.. Mu+S.rdD, , , , , , I 0.0 0.1 0.2 03 04 0.5 06 07 Dry Creek (Regulated) 1 . . . . . . . COMh .;\;.;.;.;.;.; \\\\\\\\\\,\ ,,,,,,,,,,, iB : : Phbl" .\.....\\\\\ ....................... t4Ld-S.m '4' D,, , , . , J 0.0 01 02 03 04 0.3 06 07 PrOpDrtlDn 01 SIl.. Fig. 7. Proportional distribution of bed sediments in mud-sand (median diameter < 2mm), pebble (2-64mm), cobble (65-256mm) and boulder to bedrock (> 256mm) size classes. Sample sizes (numbers of sites sampled) were S. Fk. Eel: 66; Elder Creek: 38; Outlet Creek: 48; Middle Fk. Eel: 75; E. Fk. Russian: 52; and Dry Creek: 76. (Sample sizes exceed those for current velocities because they include some sites too shallow for current velocity measurement). given sampling date. When low flow cut off marginal pools from the main channel, however, pool water was sampled separately. Water samples were filtered within 4-h of collection through Whatman GF/C, then through Gelman Metricel GN-6 0.45 micron filters, to remove particulates and algal cells. Filtration with Nalgene hand pumps was done at pressures of less than 10- 12psi to avoid rupture of cells. Filtered samples were stored frozen until analyzed chemically. Nitrate was analyzed using a hydrazine sulfate method (KAMPHAKE et al., 1967); am- monia was analyzed by a phenolhypochlorite method (SOLORZANO, 1969), and soluble reactive phosphorus was measured with a stannous chloride technique (American hhlk 1 Ic.iltli Asvociiition, 1985). These thrcc tcchniqucs Iiwc Iowcr wmitivity 394 Mary E. Power Hydrologic and trophic controls of seasonal algal blooms 395 80 "4-N T L N03-N I Fig. 8. Average concentrations with 1 SE of nutrients in the six monitored reaches. Samples were collected on survey dates; no seasonal variation was detected. Table 1. Categories for density and condition ranking of algae. Density: 0: none detected 1: 2: < 10% canopy cover 3: IO-25% canopy cover 4: 5: Condition (Cladophora): 0: detritus 1: 2: 3: 4: 5: a few filaments or small isolated patches (< 5 mm diameter) > 25% canopy cover, but can still see primary substrate substrate completely obscured by algal canopy detritus with a few recognizable filaments pale, weak, senescent in appearance discolored green: yellow or rusty with epiphyte or silt loads more green, filaments more robust deep, bright green with fresh-appearing growth thresholds of 1 pgA, 5pgA and 1 pgA, respectively. Nutrient absorbance readings were measured in 5 or lOcm cuvette cells on a Bausch and Lomb Spectronic 21 spectro- photometer. Low concentration samples were spiked with known concentrations of standard solutions to bring absorbance into the best measurement range for the instru- ment. Results i. Seasonal changes in Cladophora and other primary producers Cycles of bloom, detachment, senescence, and decomposition of Cla- dophoia were observed in all four unregulated rivers during the low flow pe- riod, but were more pronounced in the three sunlit channels (Fig. 9a-d). In Elder Creek, the most shaded of the natural channels, less algae accumulated. The accrual of algae during the low flow period was visually striking in all three sunlit channels, and locally in sunlit patches within Elder Creek. a0 60 40 20 0 A attached 0 loose Middle Fork Eel - .- c QJ 401 20 P I1 , \" i, , , , , , , , , , East Fork Russian (Regulated) (Regulated) 40 20 0 S OND JFMAM J J A S S 0 N D J FMA M J J A S Month Fig. 9. Averaged heights or lengths of attached or loose (floating or deposited) algae in encountered at monitored sites in the six study reaches. Algal assemblages were do- minated by Cladopboru from June-September in the four unregulated reaches, and at all sample dates in the two regulated channels. All sample points are averages from 15 to 41 measurements of loose or attached algae on a given sampling date. From November to March, when the four unregulated streams were too high to wade, visual inspection re- vealed no conspicuous macroalgae. 396 Mary E. Power In contrast, no massive bloom or seasonal die-off of Chbpbwa or other algae occurred in either the shaded or the sunlit regulated river (Fig. 9 e, f). In both the E. Fk. Russian River and Dry Creek, short but healthy-appearing (green, robust) Cladophora turfs dominated primary producer assemblages throughout the year (Fig. 10e, f), although more sites were dominated by bare substrate or detritus in the more shaded E. Fk. Russian. Loose algae was not encountered along transects in the E. Fk. Russian River, with the exception of a single thin strand of drift algae, 20cm in length, in October. The relatively 1: ai - f Fig. 10. Proportion of monitored sites dominated on a given sampling date by various producer taxa, or by bare space. Three transects were monitored per channel per survey; the total numbers of sites under transects varied with river width and stage: Elder Creek: 21-37 sites; South Fork Eel: 38-55 sites; Outlet Creek: 45-50 sites; Middle Fork Eel: 63-78 sites; East Fork Russian: 52-55 sites; Dry Creek: 73-77 sites. Hydrologic and trophic controls of seasonal algal blooms 397 stable discharge of the two channels also permitted growth along their margins of vascular macrophytes including Potamogeton, Veronica, Nasturtium, Lemna, and Elodea. In the four unregulated rivers and creeks, November floods scoured all macroscopically conspicuous algae from the channels. Macroscopically visible algal growths first appeared on rock substrates in March. By April and early May, diatoms, cyanobacteria (Rivularia Nostoc, and Odlatoria), and the fil- amentous green alga Zygnemi dominated producer assemblages in three of the four unregulated rivers (Fig. 10). (Algal abundance and composition in the Middle Fork Eel were not monitored until June, when the river became low enough to wade). Zygnema grew as bright chartreuse turfs 10-40 cm long. By late May Zygnema was intermingled with Cladophora at many sites. By June, Zygnema was replaced by Cladophora at most sites. For example, in Outlet Creek, there was a large bloom of attached algae dominated by attached Zyg nema mixed with CIadophora that covered 60% of the bed (Fig. 10). By June, floating mats and attached turfs of algae dominated by Cladophora covered al- most 80% of the channel (Fig. 10). Zygnema in the South Fork Eel persists through the summer only in fast-flowing riffles (M. POWER, personal observa- tions from 1987- 1991). Cladophoru turfs first appeared in the South Fork only on bedrock and boulder (> 256 mm median diameter) substrates. After June, small attached Cladopbora filaments appeared on smaller cobbles and pebbles; these became abundant after late July (Fig. 11). Substrates with Attached Cladophora 40 n v) 0 Cobbles, Pebbles .- H Boulders, Bedrock v) c .c O 20 L 0) f 10 z 0 0 100 200 FMAMJ J AS0 Fig. 11. Number of monitored sites in the South Fork Eel, separated by substrate size, with attached Cladapboru. Pebbles and cobbles range from 16-256mm median diameter; boulders and bedrock substrates are > 256 mrn median diameter. 398 Mary E. Power Hydrologic and trophic controls of seasonal algal blooms 399 Standing crops of Cladophoru, indexed by height of attached turfs or length of detached mats peaked in midsummer and decreased afterwards (Fig. 9). Most filamentous green algae had disappeared or decomposed by early fall, before the onset of winter floods. In Outlet Creek, the warmest and most eutrophic of the unregulated rivers, fall senescence of Cladophoru was particularly pro- nounced. On Oct. 2, 1988, no living Cladophoru was found in the channel, al- though white Ckzdophoru "paper" was stranded along the shore. The bed of Outlet Creek at this time was covered with Cladophoru-derived detritus, with adnate Rivulurk spots on stones representing the only conspicuous living alga. 2. Seasonal changes in conspicuous fauna Abundances of visually conspicuous animals in the water column or on the surface of the river bed under monitored sites were estimated semi-quantita- tively with presence/absence observations. These observations suggest that the densities of particular guilds differed among the six rivers. Sessile primary con- sumers (eating primarily plants or detritus (LAMBERII & MOORE, 1984)) were abundant in the two regulated rivers on all sampling dates, while their frequency of occurrence rose, then declined in the two frequently monitored unregulated rivers (S. Fk. Eel and Elder Creek) over the low flow season (Fig. 12a, b). Mobile taxa initially dominated the primary consumer guilds in both of these rivers. Subsequently, in the South Fork Eel, predators increased, mobile primary consumers decreased, and sessile primary consumers increased. In the darker Elder Creek, similar trends occurred, but were seasonally delayed. Pre- dators and omnivores (animals eating both animal and plant-detrital foods) were less frequently observed in regulated than in unregulated channels (Fig. 12). More different taxa of conspicuous surface fauna in all trophic levels were ob served in the unregulated rivers (Fig. 13). All the trends suggested here require more observations with more rigorous sampling methods before they can be confirmed. Discussion The seasonal cycle of growth, detachment, and senescence of Ckzdophoru in unregulated rivers of northern California resembles its cycle in the Lau- rentian Great Lakes (BLUM, 1982) where in recent decades, massive Cludophoru blooms have occurred (AUER, 1982; MILLNER & SWZENEY, 1982; NEIL & JACK- SON, 1982). In both northern California rivers and the Great Lakes, Cladophoru initiates growth in the late spring, covering large portions of the bed where suitable stable substrates occur (AUER, 1982; DUFF et al., 1984; POWER, 1990 b). Cladophoru in both the lakes and in unregulated rivers attains its maximum biomass in mid-summer, then detaches to form floating mats. Senescence in late summer or fall is sometimes followed by a limited resurgence of growth A Mobile Drev 0.8 0.6 a South Fork Eel I- d Middle Fork Eel I a, 0.8 A Outlet Creek (I) .- 0.4 -11 1 114 I 0.0 o.2uJ SONDJFMAMJ JAS UJ S 0 N DJ FM A MJ J A S Month Fig. 12. Seasonal changes in abundance estimates of members of three consumer guilds. "Mobile Prim" - mobile primary consumers (e.g., mayflies, mobile caddisflies, snails, tadpoles, sphenid beetle larvae); "Sessile Prim" - attached or retreat-dwelling primary consumers (e.g., sessile caddisflies and tube- or tuft-dwelling chironomids); "Preds and Omnivs" are taxa that eat live animal tissue (e.g., roach (minnows), steelhead parr, mites, stoneflies, naucorids (hemipterans), crayfish, flatworms). Lines are drawn be- tween October and August observations in the East Fork Russian because during a January census of algae when fauna not recorded, the presence of large numbers of hydropsychids was noted qualitatively. In the four unregulated channels, winter flows precluded wading transects, but from the bank, the absence of fauna and algae on scoured bed material could be detected. (MILLNFX & SWEENEY, 1982; POWER, unpublished data). In winter, Cludophoru is eliminated from shoreline habitats in the Great Lakes by ice scour and from river beds by scouring floods. In Lake Michigan, perennating basal rhizoidal holdfasts on deeper substrates overwinter and vegetatively produce the first re- growth in the spring (BLWM, 1982). A similar pattern has been described for Cladophoru in the Baltic Sea by WAERN (1952, cited in BLUM (1982)), who dis- tinguished populations of "perennial Cladopbora" in deep habitats from 400 ' 0.0- 0.0- 0 1, Mary E. Power Mobile Primary Consumers 0- 0- 8nU * 8 Hydrologic and trophic controls of seasonal algal blooms 401 I ." 0.8- 0.6 - 0.4- Elder Creek Sessile Primly Consumers Middle Fork Eel River .." 0.0- 0.6- 0.4 - 0.2- Outlet Creek I Predators and Omnivores GI wauln. H mW 0 a- O N- 8 mb 8 M.d I3 -dl South Fork Eel River 0.8 I ." 0.8' 0.8- 0.4- "" { Mobile Primary Consumers 08 08 0.4 02 00 M O 8 2 .n ""PSggSile primary Consumers Sessile Primary Consumers 0.6 o'81 n 0.4 02 0.0 O < c 0.2j 1 .O Predators and Omnivores Predators and Omnivores Predatocs and Omnivores 0 mwm?a 0.8 o'81 0.4 1 0.4 0.2 0.0 s D 0.2 1 0.0 1 O M 2 i? $ Fig. 1312 Fig. 13/1 summer "hydrolittoral Cfudopboru" at the waterline. Shoreline Cludopboru ap pears to recruit from spores released by the deeper summer populations, and does not overwinter. Spring regrowth of Cludopboru in California rivers also appears to be from vegetative regrowth by basal holdfasts that overwinter on boulder or bedrock substrates. Attached Cladopboru first occurs in late spring only on large boulder and bedrock substrates (Fig. 11), which experience less scour and overturn during bed movement (SOLISA, 1979; POWER & STEWART, 1987). From March through August of 1989, Cfudophoru in the South Fork Eel did not develop on unglazed tiles placed above the bed where they were inac- cessible to most grazers, despite luxuriant growth of the alga on bedrock im- mediately beneath tiles. This observation also suggests limited recruitment from zoospores during that year. The annual bloom-detachment-decomposition cycle of Cludophoru in the Great Lakes and in unregulated California rivers contrasts with the constant 26 dchiv f. Hydrobiologie, Bd. 125 402 Mary E. Power 0.6' 0.4- 0.2- Hydrologic and trophic controls of seasonal algal blooms 403 East Fork Russian River (Regulated) 1 .o Mobile Primary Consumers Sessile Prhry Consumers O imuiium hyhop.* 0.6 1 .o Predators and Omnivores 0.6 0- 0.4 m .- 0 IMCh "'1 0.01 - 1 x '" 3 s d Fig. 13/3 Dry Creek (Regulated) 0.6 n Predators and Omnivores o.8/ Fig. 13. Frequencies of occurrence of taxa (numbers of sightings per number of sites monitored) in the three consumer guilds in the six monitored rivers. low standing crops of the alga maintained in the two rivers with artificially re- gulated flow. The contrast suggests that Cladophoru cycles in rivers may be ex- trinsically driven by factors related to the hydrograph. Most physicc-chemical factors were probably more favorable for Cfa- dophoru growth in the two regulated rivers. Their more constant hydrographs would reduce mortality from scour in winter and from stranding and desicca- tion in summer. Temperature was more constant in these channels (between 9- 18 "C (data from the USGS and POWER, unpublished data)). This tempera- ture interval is close to that yielding maximum production of Lake Michigan Cludophoru: 10-23 OC; above temperatures of 23 "C, standing crops declined (GRAHAM et al., 1982). During summer low flow, shallow stagnant areas of the unregulated rivers in this study could warm to 28-30 "C on a daily basis. The sustained flow in regulated channels during summer (Figs. 3, 6) would also maintain higher fluxes of nutrients known to favor growth of filamentous green algae (WHITFORD & SCHUMAKER, 1964); in addition, measured nitrate con- centrations were higher in the regulated than in the unregulated rivers (Fig. 8). In North America, phosphorus is more often limiting in eastern and mid- western regions (SCHMLER, 1978), while nitrogen limitation often occurs in western regions (GOLDMAN, 1981). In northern California, N: P ratios are low (typically around 2 (Fig. 8)), and nitrogen can limit algal growth (HILL & KNIGHT, 1988; POWER, 1991). Despite apparently more favorable physico-chemical conditions, Clu- dophoru blooms were not observed in regulated channels as they were in the rivers with natural hydrographs. Instead, low but viable standing crops of Clu- dophoru were maintained throughout the year. These observations suggest that interaction of the hydrograph with biotic factors: consumers, self-limitation, and/or epiphytes, may determine whether or not Cludophoru cycles in rivers. A working hypothesis Although grazing has been considered unimportant as a potential control on Cladophoru in the Great Lakes (AUER & CANALE, 1982), grazing by ver- tebrates and invertebrates can limit Cludophoru accrual in California rivers and streams (LAMBERTI & RESH, 1983; FFMINELLA et al., 1989; POWER, 1990 b). Spring blooms of Ckzdophora in unregulated rivers may occur because the alga recovers from winter scour before the densities of grazers build up. Severe winter floods reduce stream invertebrates, many of which overwinter as larvae (FEMINELJA & RESH, 1990; RESH et al., 1989). Following winter flooding, Cla- dophoru enjoys a window of time in late spring and early summer with favorable growth conditions in a community with only one functionally sig- nificant (sensu FRETWELL, 1977) trophic level: producers unchecked by her- bivory (Fig. 14). Subsequently, as river levels drop and consumer densities build up, Cludophoru biomass will be grazed down where rivers have two or four functional trophic levels (unrestrained grazers, or where predators of pre- dators release grazers (POWER, 1990 b)). The alga will persist longer where it is 404 Mary E. Power JFMAMJJASOND scouring spring grazing, scouring floods bloom senescence floods Fig. 14. Predicted patterns of seasonal fluctuation in algal biomass produced by interac- tions of hydrologic and trophic regimes in regulated and unregulated river channels. ungrazed (one trophic level), or where third-level predators hold grazers in check (Fig. 14, POWER, 1990 b). If large growths of Cladophoru accrue, they will eventually slough and se- nesce, whether grazed or not. Cldophoru cells near basal portions of filaments under massive growths are in dark, stagnant environments, and are likely to weaken, detaching the colony (STEVENSON & STOERMER, 1982 a). In addition, Cludophoru, especially when near or at the water surface, accrues heavy loads of epiphytic diatoms (the rough surfaces of its cell walls make good settling substrates (LOWE et al., 1982)). Heavy epiphyte overgrowth in late summer has been attributed to the leakage of nutrients by senescing Cludophoru (FITZGE- RALD, 1969). Epiphytes may also induce senescence, however, as they can shade host filaments, block nutrient uptake, or damage cell walls (STEVENSON & STOERMER, 1982 a, b). In unregulated rivers, these interactions among macro- algae and epiphytes occur under increasingly unfavorable flow, nutrient, and thermal regimes as flows decline over the summer. A combination of physico- chemical and biotic stresses may account for the senescence and decomposition of Cladophoru before the onset of winter floods. Higher trophic levels may ac- celerate or retard the detachment and senescence of algae following blooms (Fig. 14), but they probably cannot prevent it. If chronically grazed, however, Cfudophoru may never build up sufficient biomass to become self-limited. Low standing crops, well beneath the water surface, will also be slower to accrue epiphytes. In addition, algivores such as midges (POWER, 1991), mayflies (DUDLEY, in press), and tadpoles (KUPFERBERG Hydrologic and trophic controls of seasonal algal blooms 405 et al., manuscript) can groom Cfudophoru, removing epiphytes from host fil- aments. In the regulated rivers, where constant, low, viable standing crops of Cfudophoru persisted throughout the year, densities of primary consumers, particularly sessile taxa, also remained high year round (Fig. 12e, f and M. POWER, personal observations). Potential predators were less frequently observed in regulated than in un- regulated rivers during the low flow season (Fig. 12). One explanation for this observation is that some predatory taxa such as stoneflies or salmonids may re- quire natural hydrologic fluctuation and/or natural seasonal thermal regimes to complete their life cycles (WARD & STANFORD, 1979). However, salmonids and stoneflies were observed in the regulated channels surveyed here, (Dry Creek was the release stream for a salmon hatchery), suggesting that physical factors did not exclude these predators. Another hypothesis to explain low predator abundances in the regulated channels is that under prolonged stable flow regimes, primary consumers with effective defenses against predators, such as sessile taxa protected by cases or retreats, come to dominate (Fig. 15), and provide an inadequate food base for predators. Thus, sessile primary con- sumers with attached cases or retreats (shown for one of the common Eel River taxa to enhance resistance to predators (POWER et al., 1992, see also Jo- HANSSON, 1991 and references therein) were usually more abundant than mobile primary consumers in both regulated channels (Figs. 12, 13)). In the most frequently surveyed unregulated channel, the South Fork Eel, mobile primary consumers initially dominated following winter floods, but were eventually outnumbered by sessile primary consumers as predators built up (Fig. 12 a). A similar but weaker, and seasonally delayed, trend occurred in the more shaded, less productive Elder Creek (Fig. 12 b). Less algae accrued in Elder Creek than in other unregulated channels (Fig. 9), probably because it was the most shaded of all the monitored channels. In addition, of the four un- regulated channels, Elder Creek had the smallest peak flood relative to its bankfull discharge (Fig. 5), so Overwintering grazers might have suffered less from scour. Mobile, but not sessile grazers were relatively high during the first sample date in Elder Creek (Fig. 12 b). Further study is needed to determine the relative importance and possible interplay of productivity and disturbance regimes in limiting algal accrual in these channels. An implicit assumption underlying the hypothesis presented above is that stream invertebrates that are more resistant (or faster to recover from) scour- ing floods are less resistant to predators. In channels with frequent, large dis- charge fluctuations, disturbance-resistant taxa with traits like high mobility, short life histories, and high intrinsic rates of increase should dominate (GRAY & FISHER, 1981). In channels with prolonged stable flow, taxa protected from predators by traits like heavy cases, retreats, or sessile or sedentary habits which make potential prey inconspicuous and keep them in or near refuges, 406 E c m D 2 a g W 0 0 Mary E. Power Retreat-dwelling caddis, midges Heavy-cased mobile caddis Hydrologic and trophic controls of seasonal algal blooms 407 spring and early summer before densities of animals build up. Initially, primary consumer guilds should be dominated by mobile, fast-growing species like mayflies. Such taxa will be replaced by predator-resistant consumers as predator densities build up following disturbance. The hypothesis predicts that heavy winter rains associated with El Nino events will trigger particularly pro- nounced algal bloom-sloughing-senescence cycles in unregulated rivers. In the absence of winter flooding, predator-resistant primary consumers will come to dominate, producing food webs with two functional trophic levels: herbivores unrestrained by predators. Under this trophic regime, Cladophora should not attain sufficient biomass to undergo the bloom-detachment-senescence cycle. Tests of this working hypothesis will require long-term monitoring, closer study of the performance of invertebrates with various traits exposed to dis- charge fluctuations and predation, and more rigorous faunal surveys of differ- ent types of channels. The visual scanning and coarse taxonomic resolution I used in field surveys yield only a superficial impression of faunal distributions. These methods were applied consistently among the six rivers studied, how- ever, and provide a preliminary basis for comparison. These observations motivate more careful study, because comparisons of communities and com- munity assembly in rivers that differ in size, productivity, and discharge-me- diated disturbance regime may provide insights about key processes that struc- ture river communities. The cyclical accrual of large standing crops of macroalgae is not only one possible outcome of these processes, it is also a re- gulator of interactions among higher trophic levels, and of exchange between rivers and their watersheds (FISHER et al., 1982; POWER, 1990 a). Sessile midges, pyralids, caddis L m VI VI c .- a: Thin-cased snails and mobile caddis Mobile non-armored taxa (e.g. mayflies) Resistance to Scour or Dessication 3 Fig. 15. Diagram of a working hypothesis concerning adaptations that may represent tradeoffs for aquatic invertebrates between resistance to discharge fluctuation and re- sistance to predation. Sessile and sedentary habits are hypothesized to decrease the abil- ity of individuals to avoid areas of scour or desiccation. In addition, these traits, as well as allocation to armor, would slow food acquisition by individuals, reducing the growth rates of their populations that would permit recovery from floods or desiccation. Mobile, non-armored taxa should permit individual mobility and high population growth and recovery rates, at the expense of vulnerability to fish and other predators. In the surveyed northern California rivers surveyed in this study, Petrophila (Paragyrac- tidae: Lepidoptera) and hydropsychid caddis larvae (Trichoptera) would be representa- tive sessile retreat dwellers; Dicosmoecw gzfvipes (Limnephilidae: Trichoptera) would re- present heavycased species; Physel(a (Gastropoda) would represent thin-cased species; and a variety of mayfly species (Ephemeroptera) would represent mobile, non-armored taxa. should dominate. (The tendency for beds of rivers experiencing low stable dis- charge to become choked with fine sediments should enhance the advantages of sessile or cased invertebrates over mobile invertebrates, whose spatial ref- uges from predators dwindle as pore spaces in the bed fill up.) Because many traits conferring predator-resistance are likely to reduce both mobility and al- location to growth and reproduction which allow taxa to escape or recover from disturbance, resistance to one of these major sources of mortality for stream invertebrates should often increase vulnerability to the other (Fig. 15). To the extent that these tradeoffs involve energy allocations, transitions from dominance by disturbance-resistant to predator-resistant primary consumers should occur earlier in channels with higher primary productivity. In summary, hydrologic and trophic controls of Cludophoru in northern California rivers are hypothesized to interact. Winter floods that reduce pop- ulations of invertebrates permit river algae to accrue large standing crops in Acknowledgement I would like to thank BILL RNNEY and HABTE KIFLE for help with field work, HEATH CARNEY and RHEA WILLIAMSON for help with chemical analyses, KENNETH MARKHAM and JOHN PALXIER of the USGS for providing hydrologic records, BILL DIETRICH for help with flood frequency analyses, MICHAEL PARKER for valuable discussion and comments on the manuscript, and the California State Water Resources Center (W-726) and the National Science Foundation (Ru-8600411, BSR-9106881) for financial support. References American Public Health Association, Amerian Waterworks Association and Water Pollution Control Federation (1980): Standard Methods for the Examination of Water and Wastewater. - APHA, Washington, D.C. AUER, M. T. (1982): Preface, Ecology of Filamentous Algae. - J. Great Lakes Res. 8: 1-2. AUER, M. T. & CANALE, R. P. (1982): Ecological studies and mathematical modeling of Cladophora in Lake Huron: III. The dependence of growth rates on internal phos- phorus pool size. - J. Great Lakes Res. 8: 93-99. 408 Mary E. Power BELLIS, V. J. & MCLARTY, D. A. (1967): Ecology of Cfudophoru gfomeruta (L.) KUTZ. in BLUM, J. L. (1956): The ecology of river algae. - Bot. Rev. 22: 291-341. southern Ontario. - J. Phycol. 3: 57-63. - (1982) Colonization and growth of attached algae at the Lake Michigan water line. - J. Great Lakes Res. 8: 10- 15. DUDLEY, T. L. (1992): Beneficial effects of herbivores on stream macroalgae via epiphyte removal. - Oikos (in press). DUFF, J. H., TRISKA, F. J. & Omurn, R. S. (1984): Denitrification associated with stream periphyton: chamber estimates from undisrupted communities. - J. En- viron. Qual. 13: 514-518. DUNNE, T. & LEOPOLD, L. B. (1978): Water in Environmental Planning. - Freeman and Company, San Francisco, CA. FEMMELLA, J. W. & RESH, V. H. (1990): Hydrologic influences, disturbance, and intra- specific competition in a stream caddisfly population. - Ecology 71: 2083 -2094. FEMINELLA, J. W., Poum, M. E. & %H, V. H. (1989): Periphyton responses to inver- tebrate grazing and riparian canopy in three northern California coastal streams. - Freshw. Biol. 22: 445-457. FISHER, S. G., GRAY, L. J., GRIMM, N. B. & BUSCH, D. E. (1982): Temporal succession in a desen stream ecosystem following flash flooding. - Ecol. Monogr. 52: 93- 110. FITZGERALD, G. P. (1969): Some factors in the competition or antagonism among bac- teria, algae, and aquatic weeds. - J. Phycol. 5: 351-359. FRETWELL, S. D. (1977): The regulation of plant communities by food chains exploiting them. - Persp. Biol. Med. 20: 169-185. GOLDMAN, C. R. (1981): Lake Tahoe: two decades of change in a nitrogen deficient oligo- trophic lake. - Verh. Internat. Verein. Limnol. 21: 45-70. GOLDMAN, C. R. & HORNE, A. J. (1983): Limnology. - McGraw-Hill, New York, N.Y. GRAHAM, J. M., Am, M. T., CANALE, R. P. & HOFFMAN, J. P. (1982): Ecological studies and mathematical modeling of Cladophoru in Lake Huron. IV. Photosynthesis and respiration as functions of light and temperature. - J. Great Lakes Res. 8 100- 11 1. GRAY, J, J. & FISHER, S. G. (1981): Postflood recolonization pathways of macroinver- tebrates in a lowland Sonoran Desert stream. - Am. Midl. Nat. 94: 421 -424. GRIMM, N. B. (1987): Nitrogen dynamics during succession in a desen stream. - Ecol- HILL, W. R. & KNIGHT, A. W. (1988): Nutrient and light limitation of algae in two northern California streams. - J. Phycol. 24: 125- 132. HYNES, H. B. N. (1970): The Ecology of Running Waters. - Univ. of Toronto Press, Toronto. JOHANSSON, A. (1991): Caddis larvae cases (Trichoptera, Limnephilidae) as antipredatory devices against brown trout and sculpin. - Hydrobiologia 211: 185- 194. KAMPHAKE, L. S., HANNAH, S. A. & COHEN, J. M. (1967): An automated analysis for ni- trate by hydrazine reduction. - Wat. Res. l: 205-216. KUPFERBERG, S. J., Msas, J. C. & Po=, M. E.: Variation in natural algal and detrital diets: Effects on larval anuran life history. Manuscript. LAMBERTI, G. A. & RESH, V. H. (1983): Stream periphyton and insect herbivores: An ex- perimental study of grazing by a caddisfly population. - Ecology 64: 1124- 1135. LmBEnn, G. A. & MOORE, J. W. (1984): Aquatic insects as primary consumers. - In: RESH, V. H. & ROSENBERG, D. M. The Ecology of Aquatic Insects, pp. 164- 194. - Plenum, New York, N.Y. ogy 68: 1157-1170. Hydrologic and trophic controls of seasonal algal blooms 409 LIGON, F. K. (1986): The response of algal communities in streams to timber harvest ac- tivities. - M.S. thesis, Univ. Cali., Berkeley. Lomz, R. C. & HERDENDOW, C. E. (1982): Growth dynamics of Cfudopboru glomerura in Western Lake Erie in relation to some environmental factors. - J. Great Lakes Res. 8: 42-53. LOWE, R. L., ROSEN, B. H. & KINGSTON, J. C. (1982): A comparison of epiphytes on Bungia urropurpurea (Rhodophyta) and Ckzdophoru gfomeruru (Chlorophyta) from Northern Lake Michigan. - J. Great Lakes Res. 8: 164- 168. MILLNER, G. C. & SWEENEY, R. A. (1982): Lake Erie Ckzdophoru in perspective. - J. Great Lakes Res. 8: 27-29. MILLNER, G. C., SWEENEY, R. A. & FREDERICK, V. R. (1982): Biomass and distribution of Cladophoru glomerura in relation to some physical-chemical variables at two sites in Lake Erie. - J. Great Lakes Res. 8: 35-41. NEIL, J. H. &JACKSON, M. B. (1982): Monitoring Cladophoru growth conditions and the effect of phosphorus additions at a shoreline site in Northeastern Lake Erie. - J. Great Lakes Res. 8: 30-34. PARKER, G. (1978): Self-formed straight rivers with equilibrium banks and mobile bed. Part 2. The gravel river. - J. Fluid Mechanics 89: 127- 146. POWER, M. E. (1990a): Benthic turfs vs. floating mats of algae in river food webs. - Oikos 58: 67-79. - (1990b): Effects of fish in river food webs. - Science 250: 411-415. - (1990 c): Seasonal and hydrologic controls of algal blooms in northern California rivers. - Tech. Compl. Rep., Water Res. Ctr., Univ. Calif., Riverside, Riverside, CA. - (1991): Shifts in the effects of tuft-weaving midges on filamentous algae. - Am. Midl. Nat. 125: 275-285. Po-, M. E., hms, J. C. & PARICER, M. S. (1992): Variation in the vulnerability of prey to predators: community-level consequences. - Ecology (in press). POWER, M. E., hmws, W. J. & STEWART, A. J. (1985): Grazing minnows, piscivorous bass and stream algae: dynamics of a strong interaction. - Ecology 66: Pomn, M. E. & STEWART, A. J. (1987): Disturbance and recovery of an algal assemblage following flooding in an Oklahoma stream. - Am. Midl. Nat. 117: 333-345. RESH, V. H., JACKSON, J. K. & MCELRAVY, E. P. (1989): Disturbance, annual variability, and lotic benthos: examples from a California stream influenced by a Mediter- ranean climate. - Mem. 1st. Idrobiol. 47: 297-317. SCHINDLER, D. W. (1978): Factors regulating phytoplankton production and standing crop in the world's freshwaters. - Limnol. Oceanogr. 23: 478-486. SOLORZANO, L. (1969): Determination of ammonia in natural waters by the phenol- hypochlorite method. - Limnol. Oceanogr. 14: 799-801. SOIJSA, W. P. (1979): Disturbance in marine intertidal boulder fields: the nonequilibrium maintenance of species diversity. - Ecology 60 1225- 1239. STEVENSON, R. J. & STOERMER, E. F. (1982a): Abundance patterns of diatoms on Cla- dophoru in Lake Huron with respect to a point source of wastewater treatment plant effluent. - J. Great Lakes Res. 8: 184- 195. - (1982b): Seasonal abundance patterns of diatoms on Cfudopboru in Lake Huron. - J. Great Lakes Res. 8: 169- 183. TRUSH, W. J. (1991): The influence of channel morphology and hydrology on spawning populations of steelhead trout in South Fork Eel tributaries. - Ph. D. disserta- tion, Univ. Calif., Berkeley. 1448 - 1456. - 410 Mary E. Power, Seasonal algal blooms WARD, J. V. &STANFORD, J. W. (1979): Ecological factors controlling stream zoobenthos with emphasis on thermal modification of regulated streams. - In: WARD, 1. V. & STANFORD, J. W.: The Ecology of Regulated &earns, pp. 35-55. - Plenum, New York, N.Y. WHARFE, J. R., TAYLOR, K. S. & MONTGOMERY, H. A. C. (1984): The growth of Clu- WHITFORD, L. A. & SCHUMACHER, G. J. (1961): Effect of current on mineral uptake and WHITTON, B. A. (1970): Biology of Cladopboru in freshwaters. - Water Res. 4: 457 -476. dophom glomeruzu in a river receiving sewage effluent. - War. Res. 18: 971 -979. respiration by a freshwater alga. - Limnol. Oceanogr. 6: 423 -425. Submitted: 29 August 1991; accepted: 21 April 1992.