Strategies providing success in a variable habitat: I. Relationships of environmental factors and dominance of Cladophora glomerata

Authors


Correspondence: IngoEnsminger. E-mail: ingo.ensminger@uni-jena.de

ABSTRACT

Dominance of macrophytes and their response to environmental factors were studied in the river Ilm, Thuringia/ Germany with special reference to Cladophora glomerata (L.) Kütz. Macroalgae showed a growth peak in spring with C. glomerata, Ulothrix zonata Kütz., Lemanea fluviatilis (L.) C. Agardh and Audouinella sp. being the dominant species. Shortly after this peak, a rapid decline of macrophyte substrate coverage was observed. Only C. glomerata revealed a second growth peak in late summer/early autumn. Frequent disturbances of the macrophyte assemblage by floods resulted repeatedly in an almost complete wash out of benthic organisms. After summer floods C. glomerata was the species that recolonized the substrate. At high-light sites, faster recovery of C. glomerata was observed as compared to low-light sites. This is discussed in relation to the life cycle of C. glomerata. Among the physical and chemical parameters that were analysed, irradiance, current velocity, pH, soluble reactive phosphorus and ammonia-nitrogen accounted for most of the observed patterns of dominance of C. glomerata.

INTRODUCTION

Within the freshwaters of the temperate zones, Cladophora glomerata (L.) Kütz. is probably the most widespread macroalga and an important component of ecological communities ( Blum 1956, Whitton 1972, Dodds & Gudder 1992). Power (1990, 1992) emphasized the ecological importance of this filamentous green alga, as it plays an essential role in food webs and creates much of the physical structure, such as substrate for colonization or refuges within northern California rivers. The substrate suitable for colonization by benthic primary producers is usually limited and thus prone to competition between different species. As a result, C. glomerata can be found intermixed with other macroalgae ( Entwisle 1989; Power 1992) or even with mosses. Whitton & Buckmaster (1970) described the macroalga C. glomerata and the moss Fontinalis antipyretica Hedw. within the same habitats. They further mentioned that Rhynchostegium riparioides (Hedw.) Card. (quoted in Whitton & Buckmaster 1970 under its former name Eurhynchium riparioides) often grows within tufts of C. glomerata.

The presence of C. glomerata may serve as an indicator of good water quality ( Liebmann 1962), but the development of dominant stands is most often related to human impact and high nutrient loads ( Blum 1956; Auer et al. 1982 ; Dodds & Gudder 1992). Mass development of C. glomerata may affect the biocoenosis in various ways and can create severe management problems ( Blum 1956; Bellis & McLarty 1967; Auer & Canale 1982, Schönborn 1996). In the river Ilm/Thuringia (Germany, Fig. 1), river regulation, industrial pollution and municipal sewage effluents had a negative impact on the ecosystem until 1989. Cladophora glomerata was the most dominant primary producer and frequently developed large dominant stands. Since then, the shutdown of considerable parts of industrial facilities and improvement of municipal sewage treatments in the Ilm and its catchment area has dramatically improved ecological conditions especially in terms of nutrient loads ( Fig. 2). However, C. glomerata still remains the most dominant macroalga in the Ilm. Occasional mass developments of C. glomerata have been observed recently at different nutrient, light, and current conditions ( Schönborn 1996). Despite the ecological importance of C. glomerata, relationships between environmental factors that regulate growth cycles as well as physiological responses of C. glomerata to these factors are not well understood ( Whitton 1970; Dodds 1991a). It is clear that not only eutrophication may play an important role in the formation of C. glomerata-dominated benthic assemblages. Further factors that can effect dominance of macrophytes include, for example, disturbances ( Power 1992), grazers ( Dodds 1991b), life-cycle properties ( Kiirikki & Lehvo 1997), and irradiance ( Hanelt et al. 1997 ; Hanelt 1998).

Figure 1.

Thuringia (black area), the river Ilm and the two sampling locations upstream (A) and downstream of Stadtilm (B).

Figure 2.

Mean annual NO3-N and soluble reactive phosphorus (SRP) concentrations in the Ilm between 1989 and 1998 upstream of Stadtilm. Data by Staatliches Umweltamt Erfurt. Open bars = NO3-N, closed bars = SRP.

The present work studied the response of C. glomerata to environmental factors in order to understand strategies that may provide for the alga’s success in a dynamic and variable habitat. Four sampling sites, which differed in terms of light conditions and the degree of municipal effluent inlets, were used to study the seasonal cycle of physical and chemical parameters and their relationships to important benthic macrophytes in the river Ilm. On the basis of the data obtained, factors that have strong effects on the dominance of C. glomerata in the Ilm were determined.

MATERIALS AND METHODS

Study sites

The Ilm is a small mountain river bordered by an incomplete tree strip that creates high-light and low-light habitats for benthic primary producers during the summer season. The stream-bed is covered mostly with rocks and cobbles that are suitable as substrate for attached macrophytes. Two locations in the metarhithral of the river were selected for observation and sampling. One location was situated upstream and the second downstream of the town of Stadtilm. Sampling was undertaken from May 1997 to May 1998 at 14 d intervals. Each location was close to sites regularly gauged by the Staatliches Umweltamt Erfurt and consisted of one open site (exposed to full sunlight and therefore referred to as high-light or HL site) and one shaded site (characterized by seasonally varying degrees of shade due to streamside vegetation and therefore referred to as low-light or LL site). Each sampling site represented a 10 m reach of the river ( Fig. 1).

Macrophytes: determination of frequency and coverage

The frequency of a certain species was estimated as the number of observations on the basis of the total number of samplings during the observation period; the relative frequency of macrophytes was calculated as the percentage of observations during the sampling period. Dominance of benthic macroalgae and mosses (also referred to as macrophytes) was estimated as coverage percentage of the river substrate with macrophytic thalli. A glass-bottomed box that covered a surface of 0·08 m2 was used for repeated estimates of coverage of the different species. According to Entwisle (1989) and Dethier (1984) percentage coverage was determined from visual estimates. At least 10 replicates of these visual estimates were then used to calculate mean substrate coverage percentage of the 10 m reach of each sampling site as a measure of dominance.

Length of thalli of C. glomerata was determined with a ruler with an accuracy of 0·5 cm, starting from the basal cells attached to the substrate up to the end of the filaments. Replicate measurements of five different plants at fixed points were used to calculate the mean.

Physical and chemical parameters

Field measurements of oxygen saturation, pH, conductivity, and water temperature were undertaken in the morning using a WTW-multimeter (WTW, Weilheim, Germany). The temporal and spatial variation of flow, according to Entwisle (1989), causes formation and destruction of riffle and pool structures. For that reason, current velocity was determined at each sampling site at five fixed points to give an estimate of changes of flow conditions. A magnetic-inductive flowmeter (Flowmate 2000; Marsh-McBirney, Frederick, MD, USA) was used to measure current velocity 5 cm above the substrate surface and to calculate mean current velocity. Discharge rate was measured by the Staatliches Umweltamt Erfurt at Gräfinau-Angstedt, which is located 10 km upstream of Stadtilm. Daily discharge rates were used to calculate mean and maximum values for the 14 d prior to the sampling date. Nutrient concentrations including soluble reactive phosphorus (SRP), NO3-N and NH4+-N were measured routinely by the Staatliches Umweltamt Erfurt at the two locations at Stadtilm. Total filterable phosphate was measured only infrequently at the upstream sites and was thus not included in this work. Hence, given values of P may be an underestimate at some times of the year ( Jarvie, Whitton & Neal 1998).

Depth is often used to stratify rivers in order to characterize the light conditions within a habitat. In this study incident photosynthetic active radiation (PAR) was determined instead, in order to give a direct estimate of the amount of available irradiance. Either a Li-190 quantum sensor or an underwater quantum sensor Li-192 (Li-Cor, Lincoln, NB, USA) was used to measure PAR. From determinations of PAR in unshaded air and simultaneously obtained underwater PAR values at the algal surface, the attenuation cA of PAR by streamside vegetation, light scattering and water turbidity was calculated. Data on daily global irradiance measured with a thermopile that did not select for specific spectral bands, were provided by Deutscher Wetterdienst Offenbach. The values were converted to mol photons m−2 d−1 ( Larcher 1994) and used to calculate the mean of the last 14 d (I14) before each sampling date. The effective irradiance (IE) as a measure of the proportion of irradiance reaching the algal surface, was estimated as IE = I14×cA.

Statistics

The software package SPSS release 9·0 (SPSS, Chicago, IL, USA) was used to perform statistical tests and multivariate procedures.

RESULTS

Changes of environmental factors

In order to detect effects of municipal effluents as well as effects of different light environments, one sampling location was chosen upstream and a second location downstream of Stadtilm. Each location was divided into one open and one shaded site. Thus, our data were tested for upstream/downstream effects in nutrients, for physical and chemical parameters as well as for the influence of irradiance ( Table 1). Daily sum of irradiance was highest during the summer period between April and September with values between 60 and 80 mol m−2 d−1 ( Fig. 3a). During the winter period between October and March, it was 10–20% of the summer values. This pattern was reflected directly by the water temperature ( Fig. 3a). Frequent flood events ( Fig. 3b, arrows) caused almost complete wash out of benthic algae with the exception of sparse basal cell fragments of C. glomerata attached to unmoved boulders and stones. The first flood event was observed in late July 1997. It was followed by pronounced discharge rates of more than 2 m3 s−1 until August and a further peak in October. Highest discharge rates occurred from January to mid-April.

Table 1.  Annual mean, maximum and minimum values of the physical and chemical parameters of the sampling sites during the observation period 1997–98
 Stadtilm upstreamStadtilm downstream
 Mean (± SE)MaximumMinimumMean (± SE)MaximumMinimum
  • a

    first row from open (HL) sites, second row from shaded (LL) sites.

  • b based on the atomic ratios of PO 43−- P, NO3 -N and NH4+-N.

Effective irradiance17·2 ± 2·341·11·430·7 ± 4·276·81·3
(mol m−2 d−1) a8·1 ± 1·331·51·16·5 ± 1·015·90·6
Water temperature (°C)9·0 ± 0·716·71·09·5 ± 0·0719·30·6
Current velocity (m s−1) 0·33 ± 0·041·020·090·50 ± 0·041·240·15
NH4±-N (mg l−1) 0·14 ± 0·010·300·000·48 ± 0·041·240·03
NO3 -N (mg l−1) 1·83 ± 0·133·391·132·04 ± 0·102·711·13
SRP (mg l−1) 0·24 ± 0·030·590·030·29 ± 0·030·650·05
N : P b16·2 ± 5·076·23·213·5 ± 2·651·03·2
Conductivity (μS cm−1) 321·2 ± 10·1439·0184·5734·6 ± 132·84660·0191·0
pH7·84 ± 0·048·707·417·89 ± 0·038·447·52
Oxygen saturation (%])106·9 ± 1·7148·093·4102·1 ± 1·1121·086·0
Figure 3.

Seasonal changes during the sampling period in (a) daily global irradiance and averaged water temperature; mean values of all sampling stations, and (b) discharge (arrows indicate major floods).

Current velocity was highest during periods of high discharge ( Fig. 4a) and revealed higher values at the downstream sites ( Table 1, P < 0·01, Mann–Whitney U-test). Levels of NH4+-N differed significantly between the upstream and downstream locations ( Table 1, Student’s t-test, P < 0·01). No significant differences were found for NO3-N and SRP. The observed variation in nutrients ( Fig. 4b–d) reflected the hydrographic situation, for example, high discharge rates resulted in lower concentrations due to dilution effects. No seasonal differences were found for NO3-N and NH4+-N, in contrast to SRP, which showed higher concentrations during summer (Mann–Whitney U-test, P < 0·01). Mean N : P ratios, based on the atomic ratios of NH4+-N, NO3-N and SRP, ranged between 16·2 (upstream sites) and 13·5 (downstream sites) but did not express a significant difference ( Fig. 4e, Table 1, Mann–Whitney U-test, NS). The test for seasonal differences showed a higher N : P ratio during summer than in winter (P < 0·05). Conductivity ( Fig. 4f, Table 1) was higher at downstream sites (P < 0·01). However, this difference reflected the temporary outlet of effluents of a saline-works that occasionally altered conductivity to extreme values rather than generally higher values. Thus, single peak values occurred frequently downstream of Stadtilm in June, August, September and October 1997. Neither seasonal nor upstream/downstream differences were observed for the values of pH ( Fig. 4g). Oxygen, as measured early in the morning, was usually oversaturated ( Table 1). Higher values were determined at the upstream compared with the downstream sites (P < 0·05, t-test), and during summer compared with winter (P < 0·05, Mann–Whitney U-test). From daily curves of oxygen saturation obtained in spring and summer, it was known that O2-levels never decreased below 95% (data not shown). The lowest measured value of oxygen saturation occurred in November (86%), when the Ilm was enriched in organic debris due to seasonal leaf fall.

Figure 4.

Seasonal variability of physical and chemical parameters at sampling sites upstream (○) and downstream (●) of Stadtilm. Current velocity (a), NH4+-N (b), NO3- N (c), PO43 −- P (d), N : P ratio (e), conductivity (f) and pH (g).

Patterns of macrophytic association

A total of nine different macroalgae and mosses was found at the four sampling sites ( Table 2). The most frequent species was C. glomerata that was present at 74·8% of the total of 107 samplings, followed by F. antipyretica with 47·7%. Medium frequencies were obtained for U. zonata, L. fluviatilis, Vaucheria sp., and Audouinella sp.; lowest frequencies were found for Batrachospermum sp., H. foetidus, and R. riparioides.

Table 2.  Frequencies of macrophytes at the four sampling sites (number of observations of a certain species on the basis of the total number of samplings during the observation period 1997–98. Number of samplings = 107)
SpeciesFrequencyrel. Frequency (%)
Cladophora glomerata (L.) Kütz. 8074·8
Fontinalis antipyretica Hedw.5147·7
Ulothrix zonata Kütz. 2220·6
Lemanea fluviatilis (L.) C. Agardh.2119·6
Vaucheria sp.1715·9
Audouinella sp.1312·1
Batrachospermum sp.2 1·9
Hydrurus foetidus (V.) Trev.1 0·9
Rhynchostegium riparioides (Hedw.) Card.1 0·9

Species diversity differed between the sites upstream and downstream of Stadtilm. The upstream sites showed a spring association composed of C. glomerata, U. zonata, L. fluviatilis and Audouinella sp. ( Fig. 5a & c) with the occasional appearance of small amounts of thalli of Vaucheria sp., Batrachospermum sp., H. foetidus, F. antipyretica, and R. riparioides. At the downstream sites only Vaucheria sp. was found in low numbers besides C. glomerata and U. zonata.

Figure 5.

Seasonal changes in cumulated coverage percentage of important (≥ 3%) macrophytes at open (a, b) and at shaded (c, d) sampling sites, and in length of Cladophora thalli (e, f). ○, effective irradiance; □, length of thalli at open sites; ▪, length of thalli at shaded sites.

The overall comparison of the substrate coverage data ( Fig. 5a–d) showed higher coverage percentages for C. glomerata as well as for the other macrophytes at the upstream, compared to the downstream location ( Table 3, Mann–Whitney U-test, P < 0·05). Highest values of substrate coverage were observed in May with up to 60% cumulated coverage due to dominant stands of C. glomerata, U. zonata, and L. fluviatilis followed by a dramatic decline of substrate coverage shortly after this spring bloom ( Fig. 5a–d). Lowest coverage was observed immediately after the flood events in early July 1997, January 1998, February 1998 and March 1998 at the downstream sites ( Figs 3b, 5b & d).

Table 3.  Summary of macrophyte dominance expressed as coverage percentage of total substrate surface at the four sampling sites
  Coverage percentage (%)
  C. glomerataMacrophytesLength of C. glomerata
filaments (cm)
 SiteMean ± SEMaximumMean ± SEMaximumMean ± SEMaximum
Upstreamopen (HL)6·57 ± 1·7638·003·13 ± 1·5138·003·85 ± 1·1923·80
 shade (LL)6·28 ± 1·9938·002·74 ± 1·4438·002·59 ± 1·0323·00
Downstreamopen (HL)4·30 ± 1·6138·002·02 ± 1·4138·005·34 ± 2·1039·80
 shade (LL)3·92 ± 1·6638·000·89 ± 0·5313·002·93 ± 1·2522·60

The seasonal growth of C. glomerata matched the overall scheme of substrate coverage mentioned above. It showed maxima at both open and shaded sites in May. These were followed by a fast decline until the end of June, and, shortly thereafter, by complete disappearance of the alga. At the open sites, C. glomerata developed large stands with maxima in August and, upstream, in September ( Fig. 5a & b). After a further pronounced maximum in substrate coverage by C. glomerata during November at the upstream sites, the lowest values or even complete extinction of C. glomerata and other benthic macrophytes were found later in winter, followed by the spring development of mats consisting of C. glomerata, U. zonata, L. fluviatilis, and the stands of Audouinella sp.

The length of C. glomerata thalli did not show any significant differences between upstream and downstream sites, but the thalli from open sites (HL) were found to be longer than those from shaded sites (LL) ( Figs 5e & f; Table 3, Mann–Whitney U-test, P < 0·05).

Factors related to dominance of C. glomerata

Spearman’s rank correlation test revealed that dominance of C. glomerata was significantly correlated with that of U. zonata and F. antipyretica ( Table 4). Because temporal autocorrelation is often a problem in time-related data, the temporal autocorrelation was tested for each sampling site prior to multivariate analysis. The mean of the D-value of the sampling sites did not show any significant autocorrelation of the data (D = 2·4, Durbin–Watson-test).For further analysis a general linear model (GLM) was used with a forward selection method within a multivariate analysis of variance to test all kinds of cross-relations between the measured parameters and their effects on the dominance of C. glomerata, U. zonata, and F. antipyretica. The set of physical and chemical parameters included water temperature, oxygen saturation, conductivity, pH, global irradiance, effective irradiance, current velocity, NH4+-N, NO3-N, and SRP as well as mean, maximum and minimum discharge rate. Floods ( Fig. 3) were considered by the use of an index of days since the corresponding disturbance. Step-by-step all parameters without a statistically significant effect on dominance were eliminated from this analysis. The final set of parameters that had statistically significant influence on dominance of the three species consisted of pH, effective irradiance, current velocity, NH4+-N and SRP (Pillai–Spur, P < 0·05 for each of these five parameters).

Table 4.  Spearman’s rank order correlations between substrate coverage of C. glomerata and other macrophytes
  F.
antipyretica
Vaucheria.
sp.
L.
fluviatilis
R.
riparioides
U.
zonata
Batracho-
spermum
sp.
H.
foetidus
Audoui-
nella
sp.
  • **

    P < 0·01. Number of cases = 107.

C. glomeratacoefficient of0·256 **0·0600·143– 0·0290·401 **– 0·0410·0600·121
 correlation
 significance0·0080·5380·1420·7660·0000·6720·5410·214

After random selection of 50% of the observed cases a multiple regression analysis was applied that used the five parameters determined before to describe the dominance of C. glomerata. The resulting formula is given in Fig. 6. This phenomenological model was tested with the remaining 50% of cases, for example, the predicted coverage percentage of C. glomerata was calculated on the basis of the corresponding physical and chemical parameters and plotted against the observed values ( Fig. 6). The accuracy of the model is expressed by the R-squared value (which is in fact the squared Pearson’s correlation coefficient) R2 = 0·69 of the linear regression line. Using this formula, the reduced set of parameters gave reasonable estimates of trends in substrate coverage by C. glomerata in the Ilm.

Figure 6.

Relationship between observed and predicted values of coverage percentage of Cladophora by the multiple regression model CP = –94·017 + (13·524 × pH) + (0·153 ×IE) – (8·842 ×CV) – (3·850 × NH4+-N) – (20·784 × SRP). (CP = coverage percentage of Cladophora;IE = effective irradiance; CV = current velocity). Regression ANOVAP < 0·05. The accuracy of the fit is given as R2 = 0·69 (a).

DISCUSSION

Physical and chemical parameters

The physical and chemical parameters of the Ilm provide data that illustrate the dynamics of the river. Seasonal variability of the measured parameters was of greater importance than variability between upstream and downstream sites receiving different municipal sewage effluents.

Irradiance and temperature made the period between April and September the most favourable for optimum phototrophic production and growth ( Fig. 3). Frequent flood disturbances caused scouring of benthic primary producers. This happened not only during winter, but also at the beginning and during the period of optimum production in summer. In particular, the early spring and summer disturbances are very important in terms of reproduction and dispersal as they do not promote the different algal life-cycle strategies in the same way (see below). A further consequence of the floods is the fact that low-light conditions are not exclusively restricted to the winter season due to low global irradiance levels and to sites covered by dense streamside tree vegetation in summer. Low-light conditions also appeared frequently as a consequence of high discharge and related increases in water turbidity, due to increased suspended sediments.

Differences in physical and chemical parameters between the two sampling locations were found for current velocity, NH4+-N and conductivity ( Fig. 4). The requirement of minimum current for the growth of some freshwater benthic algae has been demonstrated ( Dodds & Gudder 1992; Stevenson 1996). Current velocity was significant in the regression model ( Fig. 6), suggesting that it plays a key role among the factors affecting C. glomerata dominance in the Ilm. As the observed mean current velocity at the sampling stations ranged well within the optimum limits of 0·5 to 0·8 m s−1 for C. glomerata ( Schönborn 1996), it must be concluded, that the higher current velocity at the downstream site did not have negative effects on the dominance of C. glomerata.

With the exception of high discharge periods during winter absolute nutrient levels were within the ranges stated by Schönborn (1996) for optimal growth of C. glomerata. He quoted NO3-N levels of 2–3 mg l−1 and SRP levels of > 0·1 mg l−1 to be sufficient to induce mass production if growth is not limited by further physical and chemical parameters. Mean N : P ratios of 16 : 1 and 13 : 1 upstream and downstream, respectively, were comparable to the widely established Redfield ratio that assumes mean N : P requirements for most algae of 16 : 1 ( Redfield 1958). Taken together, this suggests that nutrients did not limit the macrophytes at the surveyed stretches for most of the observed year. High conductivity values that most probably resulted from the occasional loads of salts might be responsible as a stressor for the observed differences in species diversity as well as for the lower coverage values at the downstream sites ( Figs 4 & 5). Comparison of conductivity values from measurements in the morning and in the afternoon showed, that duration of the pulses was not longer than a few hours. Hence, effects are unlikely to be assessed by this study, because it must be assumed that conductivity pulses were only observed randomly and actually occurred more frequently than was reflected within our 2 week sampling scheme.

Macrophyte frequency and coverage

The success of any macrophyte species at the sampling sites depended mainly on the ability to cope with (i) frequent disturbances that lead to the extinction of vegetative cells, (ii) changing light environments, and (iii) variable temperatures.

Cladophora glomerata was found to be the most frequent and dominant macrophyte at any of the four sampling sites. Only in early spring did it seem to compete with U. zonata, L. fluviatilis, and eventually, Audouinella sp. for the substrate surface. In U. zonata zoosporogenesis of vegetative cells is stimulated at temperatures above 10 °C and usually results in the rapid decline of the population ( van den Hoek, Jahns & Mann 1993). This temperature dependent zoosporogenesis of U. zonata may have led to the very early decline of its population, because by the end of May the water temperature rose to more than 10 °C ( Figs 3 & 5). At the same time, C. glomerata showed an increased dominance ( Fig. 5).

Biomass losses as a consequence of floods have been observed by Sand-Jensen et al. (1989) in a Danish stream and Entwisle (1989) in two Australian creeks. The high frequency of these disturbances during the observation period in the Ilm is in fact most favourable for C. glomerata (see Figs 4 & 5). In comparison with other algae, its life cycle provides the capacity for rapid colonization and further growth. It is characterized by two forms of reproduction, attached akinetes and zoospores. The former are developed mainly in autumn as resting cells that provide overwintering of the population. The factors that induce their formation are not clear yet. Rosemarin (1985) mentioned temperature and, to a lesser extent, reduced light and nutrient depletion. He observed in autumn at temperatures between 7 and 10 °C in a depth of 0·5 m the initial formation of akinetes, although in the fringe zone C. glomerata still showed vegetative cells until December. The germination of akinetes happened in May, well before the peak development of sporangia in June. Rosemarin (1985) therefore concluded that the attached akinetes result in a perenniating effect that provides an efficient spring inoculum which will give C. glomerata an important advantage in comparison with species that might grow faster or use P more efficiently.

Pronounced sporulation was found to start in C. glomerata in June ( Rosemarin 1985), nevertheless zoosporangia seem to be produced all through the year ( Hoffmann & Graham 1984). Kiirikki & Lehvo (1997) stated that propagules of C. glomerata are released during or just after the period of active growth. This fits well to our own observations of ‘bleached’ cells due to zoospore release and the rapid decline of the first late-spring peak of C. glomerata growth ( Fig. 5). As zoospores contribute to the dispersal of the population they play an important role in colonization during summer, especially after disturbances. After flood disturbances, growth and recolonization of C. glomerata were reduced at shaded sites ( Fig. 5). This may have resulted from light-limited growth rates of C. glomerata fragments that survived on rocks at protected crevices as well as from the limitation of zoospore establishment from upstream sites. Lorenz, Monaco & Herdendorf (1991) stated that there is a minimum light requirement of 25 μmol quanta m2 s−1 for the establishment of C. glomerata zoospores.

A hint at the importance of critical light values for zoosporogenesis is given by the fact of longer C. glomerata filaments during summer at open sites. Factors that induce zoosporogenesis are still under debate. High temperatures (15–20 °C) and shortened photoperiod (8 h) are mentioned by Hoffmann & Graham (1984). Mean water temperature of the Ilm during summer was usually around 13·2 °C ( Fig. 4a, Table 1) and was thus within the 12 °C optimum range for growth given by Schönborn (1996). The rapid decline after the spring bloom in June 1997 was related to the highest water temperature observed (14·3 °C, Figs 3a & 5). Temporary reductions of effective irradiance as a result of increasing streamside vegetation cover or because of periods of decreased global irradiance as observed during the early summer in 1997 may also have triggered zoosporogenesis. Small-scale variations in C. glomerata dominance, that occurred between summer and late autumn showed a relation to the level of effective irradiance. This was most obvious at the upstream open and the upstream shaded site but also appeared at the downstream shaded site ( Figs 5a, c & d, respectively).

Determination of factors related to dominance of C. glomerata

With regard to the linear regression formula obtained, it is clear that the effective irradiance was found to be a factor that has significant positive effects on the dominance of C. glomerata. The positive statistical relationship to pH may be interpreted in two ways. On the one hand, higher coverage values usually are related to higher photosynthetic production and carbon uptake by primary producers, which will in turn affect the carbon-balance and induce higher pH ( Cheney & Hough 1983). This positive correlation between dissolved oxygen concentration, pH value and photosynthetic activity was confirmed by measurements of the daily course of these parameters in C. glomerata (unpublished results).On the other hand, C. glomerata is known to rely on high pH values ( Whitton 1970; Schönborn 1996).

Negative regression coefficients were found for NH4+-N and SRP, which is surprising as C. glomerata development is usually associated with nutrient enrichment by phosphorus and nitrate ( Sand-Jensen et al. 1989 ; Dodds 1991a; Schönborn 1996). According to Schönborn (1996) the observed phosphate level in the Ilm was well within the range for optimum production of C. glomerata. Furthermore, toxic effects of ammonia, which can result from the formation of NH3 at high concentrations, pH > 9, and high temperatures, can be excluded. The observed concentrations were well below critical values for C. glomerata mentioned by Robinson & Hawkes (1986). SRP as well as NH4+-N do not only effect macrophytes, but in turn these nutrients themselves are affected by uptake of benthic photo-autotrophs. Because the amounts of nutrients were saturating to uptake of C. glomerata, they were not likely to be limiting. Thus, the negative coefficients may represent a situation in which the actual demand of C. glomerata drives down inorganic nutrient concentrations during times of high biomass.

Within weed-beds of macrophytes the current increases growth and photosynthetic rates by increased transport of nutrients ( Madsen & Søndergaard 1983a; Dodds 1991b). Dodds (1991b) found at velocities higher than 8 cm s−1 that photosynthetic rates of C. glomerata decreased because of compaction and inhibition of transport of materials into and away from the algal tufts. Mean current velocity of our data was about one magnitude higher but it is very well within the optimum given by Schönborn (1996). This discrepancy might be due to the difference in measuring current velocity. Madsen & Søndergaard (1983a) and Dodds (1991b) measured this parameter within the tufts, whereas in the present study as well as that by Schönborn (1996), the parameter was measured as open water velocity above the tufts. From Madsen & Warncke (1983b), it can be concluded that current within the tufts was about the same magnitude, as they showed that current velocities in weed-beds can be reduced by more than 90% of open water velocities. In C. glomerata, Dodds (1991c) found 0·4 m s−1 open water current velocities to be reduced to 0·3 m s−1 within the first 2 cm of an algal tuft.

Contrary to data published by Dodds (1991a), who described the effects of floods on dominance of C. glomerata, floods did not show any statistical relation in our analysis. Because we recognized the effects of scouring by floods visually during the sampling period, future work should try to integrate the effects of flood disturbances using additional data, e.g. the size of rocks that C. glomerata is attached to ( Dodds 1991a, Power 1992).

The quality of the model for dominance of C. glomerata is expressed by the R-squared value in Fig. 6. An accuracy of R2 = 0·69 of the multiple regression analysis ( Fig. 6) suggests that the equation fits the data sufficiently, although there are several circumstances that limited the model’s accuracy. Thus, for practical reasons a linear regression model was used instead of a non-linear model. Furthermore there are still numerous unexplained effects, for example the influence of the pulsed peaks in conductivity at the downstream sites or the influence of temperature on zoosporogenesis. Further important factors that influence C. glomerata and other macrophytes have not been included in this study, such as the effects of grazers and epiphyte grazers ( Dodds 1991b). The role of spatial and seasonal changes of the irradiance environment is another area that is of great importance considering strategies providing success of benthic macrophytes. The ability of C. glomerata to respond to this factor in terms of photosynthesis is within the focus of a second paper presented within this volume (Ensminger, Hagen & Braune 2000).

ACKNOWLEDGMENTS

This study was supported by the Deutsche Forschungsgemeinschaft (DFG) Graduate Study Group Analysis of Function and Regeneration of Degraded Ecosystems. Data on discharge and nutrients were kindly provided by Staatliches Umweltamt Erfurt. Deutscher Wetterdienst Offenbach provided data on global irradiance. Two anonymous reviewers provided helpful suggestions on the manuscript. Jens Schumacher (Jena) gave valuable critics on the statistics, and Erica Froneberg (Vienna) kindly reviewed the first English draft.

Ancillary