The main purpose of this study was to test the hypothesis that increased producer diversity can have similar (positive) effects on primary production and herbivore performance as has enrichment with light. To do so, we first reanalyzed data from two earlier experiments in which we manipulated light supply and producer diversity separately and in absence of grazers. We then described the grazer experiment, in which we manipulated light supply and algal species richness in a full factorial design in presence of a generalist grazer, the cladoceran Daphnia magna. As we were solely interested in assessing (and comparing) conjectured positive effects of enrichment and producer diversity on grazers, we tried to avoid confounding negative effects of these factors. In particular, we excluded high light intensities (which could lead to unfavorably high C:P ratios, and thus low food quality, of algal biomass) and we excluded algal taxa known to be toxic or inedible. Light supply was therefore constrained to ≤120 μmol photons PAR m−2 sec−1 and the algal species pool consisted exclusively of chlorophytes of similar size.
Experiments without grazers
Striebel et al. (2009b) measured short-term primary production and longer term biomass accrual of nine species of chlorophytes as a function of light supply. Methodological aspects of this experiment are very similar to the grazer experiment described below and are specified in detail in the original publication. Important features concerning design, replication, duration, and environmental conditions are also listed in Table 1. Seven of the nine chlorophyte species are shared with the grazer experiment and three of the light treatments cover a similar range of light supplies (Table 1). For the purpose of this article, we have therefore reanalyzed the effects of light supply on biomass accrual and the C:P ratio of these seven chlorophytes over the light supply range 10–110 μmol photons PAR m−2 sec−1.
Table 1. Comparison of treatment characteristics and environmental conditions in the reported experiments
| ||Striebel et al. (2009b)||Behl et al. (2011)||This study|
|Light treatments (μmol quanta PAR m−2 sec−1)||10, 20, 110||90||30, 60, 90, 120|
|Species richness treatments||1||1, 2, 3, 4||1, 2, 4, 8|
|Number of taxa in species pool||7||9||11|
|Phosphorus in culture medium (μg P L−1)||10||31||15|
|Culture volume (mL)||250||400||500|
|Average medium exchange rate (% day−1)||10||12.5||3|
|Total number of replicates||63||24||80|
|Initial algal biovolume (μm3 mL−1)||2.0||5.3||2.6|
Behl et al. (2011) measured biomass accrual of nine species of chlorophytes as a function of species richness. Methodological aspects of this experiment are, again, very similar to the grazer experiment described below and are specified in the original publication, the most important features being listed in Table 1. All nine chlorophyte species are shared with the grazer experiment. Behl et al. (2011) analyzed the effects of chlorophyte diversity on response parameters using Shannon diversity, whereas the grazer experiment was analyzed with species richness as the independent variable (see below). For the purpose of this article, we have therefore reanalyzed the data from Behl et al. (2011) based on species richness (range 1–4 chlorophyte taxa, Table 1).
We used 11 different strains of freshwater chlorophytes of similar edibility and size (Table 2). The strains originated from the SAG Culture Collection of Algae (Göttingen) and were precultured for several weeks under constant conditions in a freshwater medium (COMBO; 15.0 μg phosphorus L−1) appropriate for phytoplankton and zooplankton cultivation. We established a species diversity gradient with four diversity levels ranging from mono- to 8-spp. polycultures (1, 2, 4, and 8 different species). Each diversity level (except for the 11 monocultures) was replicated three times with different species compositions (no identical replicates), resulting in a total of 20 communities, randomly comprised of members from the species pool. We established a light intensity gradient with 30, 60, 90, and 120 μmol quanta m−2 sec−1 (measured with a LI-COR LI 191SA Quantum Sensor, Lincoln, Nebraska in front of the experimental units). This light intensity gradient is within the typical range experienced by phytoplankton in the mixed layer of a temperate lake. The two gradients (light and diversity) were fully cross-classified, yielding a total of 80 treatments.
Table 2. Chlorophyte species used in monoculture and polycultures experiments and their mean biovolumes and cell sizes
|Chlorophyte species||Maximum cell diameter (μm)||Mean cell biovolume (μm³)||In polyculture|
| Chlamydomonas reinhardtii ||10.4||385.6||4b; 8a,c|
| Monoraphidium minutum ||6.7||104.5||4a; 8a,b,c|
| Scenedesmus obliquus ||17.7||294.8||4a,c; 8a,b,c|
| Selenastrum capricornutum ||9.5||113.8||4a,c; 8a,b|
| Desmodesmus subspicatus ||8.6||162.2||2c; 4b; 8a,b,c|
| Golenkinia brevispicula ||11.9||907.9||2a; 4b; 8a|
| Haematococcus pluvialis ||16.5||1203.0||2c; 4c; 8b,c|
| Staurastrum tetracerum ||35.0||1641.0||8a,b,c|
| Tetraedron minimum ||8.7||315.3||2b; 4c; 8b,c|
| Crucigenia tetrapedia ||7.1||150.5||2b; 4a; 8a,b|
|Pediastrum simplex (single cells)||17.1||1125.4||2a; 4b; 8c|
All treatments were inoculated with an identical total algal biovolume (2.62 × 106 μm³ mL−1 equaling 0.5 mg particulate organic carbon, POC, L−1), and different species contributed equal biovolumes to communities with two or more species. All inocula were grown under dim light conditions for 1 day before each treatment received a founder population of eight age-synchronized neonate Daphnia magna (maximum 12 hours after birth) from our laboratory stock. The communities (500 mL) were exposed to the experimental treatments in 650-mL cell culture flasks over an 11-day period, with a 10% medium exchange on days 3, 5, and 9. Temperature was constant at 20 ± 0.5°C with a 16-h-light/8-h-dark photoperiod regime. All communities were gently shaken twice a day to prevent algae from sinking and accumulating at the bottom of the culture flask. Daphnia populations were monitored qualitatively on a daily basis to follow reproduction and mortality events.
We sampled each algal community on day 1 (before adding the neonates), day 6, and day 11 (after removing all daphnids). Samples were poured through an 80-μm mesh net to retain daphnids, exuviae, and large detrital particles. As a measure of algal biomass, POC was determined after filtration onto precombusted and acid-washed glass-fiber filters (Whatman GF/C, Whatman International Ltd, Kent, U.K.) by elemental analysis (Elemental Analyzer, EA 1110 CHNS, CE Instruments, Wigan, U.K.). Particulate phosphorus (PP) was measured after sulfuric acid digestion followed by molybdate reaction. Algal biomass C:P ratios (more precisely “seston C:P ratios”) were calculated as the molar ratio of POC:PP. (Strictly speaking, POC and PP values include, beside living algal biomass, all kinds of particulate matter, which are retained on the GF/C filter, e.g., large bacterial colonies. To be precise, we will use the term “seston” instead of “algae” throughout the manuscript, where it is appropriate.) Additionally, we fixed an aliquot of each sample with Lugol's iodine to determine initial (day 1) and final (day 11) phytoplankton composition by inverted microscopy using Utermöhl chambers. A minimum of 100 cells of every species was counted by scanning at least five perpendicular transects or 20 randomly distributed, distinct fields. AnalySIS software (Pro 2.11.006, Soft Imaging Software GmbH) was used to determine biovolumes of cells by measuring two-dimensional live pictures; biovolumes were calculated from geometric shapes according to Hillebrand et al. (1999) or our own adjustments.
Daphnia body lengths were measured at the onset of the experiment (50 neonates not used in the experiment) and at the end (day 11, all surviving founder individuals and juveniles that hatched during the experiment). Length measurements were obtained electronically employing a microscope combined with a video system (ALTRA20 Soft Imaging System) and cellP software (Olympus Soft Imaging Solutions GmbH, Germany). Body length was defined as the distance from the upper edge of the compound eye to the base of the apical spine. Individual dry mass was calculated using the empirical length–mass relationship W = 11.824 × L2.236, where W is dry mass (μg), and L is body length (mm) (Stibor 1995). On day 4, the day we first detected females with eggs, all founder individuals were scanned for eggs in their brood chambers, and the number of gravid females was determined.
Data processing and statistics
Effects of light supply and algal species richness on response variables were analyzed with simple (experiments without grazers) or multiple (grazer experiment) linear regression on log transformed data (Table 3). As single data values cannot easily be determined in multiple linear regression plots, the same data are shown also as 2D linear regression plots together with their respective statistics (Figs. S1–S3, online appendix). When response variables included zero values, data were log(x+n) transformed, where n is the smallest detectable unit. Thus, n = 1 in case of the numbers of gravid and surviving founder Daphnia individuals, and n = average biomass of an individual Daphnia in case of final Daphnia biomass. Algal variables (biomass and C:P ratio) were averaged over days 6 and 11 to better reflect average food conditions for Daphnia (separate analyses of days 6 and 11 did, however, reveal qualitatively similar patterns). All statistical analyses were performed with SigmaPlot 11.0 (2008), Systat Software, Inc. Daphnids suffered complete mortality in 14 of the 80 communities. We included these communities in the statistical analyses of Daphnia responses, but excluded them from the analyses of algal responses. Results of algal statistics were, however, very similar whether those communities were included or excluded. To test for interactive effects of light enrichment and species richness on response variables, we additionally performed two-way ANOVAs on all data from the grazer experiment. None of the interactions were statistically significant (n = 66 or 80, all P > 0.23). We, therefore, do not report the ANOVA statistics in the results section.
Table 3. Simple and multiple linear regression statistics (log y = a + b × log SR + c × log Light) describing the influence of algal species richness (SR) and light intensity (Light) treatments on several independent algal and Daphnia response variables (y)
| y || n ||Overall regression||Coefficients (b, c)||Ratio of (SPRCSR/SPRCLight)|
| r² || p || a ||Log SR (SEM)|| P ||Log Light (SEM)|| P |
|Striebel et al. (2009b)|
|a||Log algal biomass||61||0.30|| ||2.64|| || ||0.37 (0.07)|| c || |
|b||Log seston molar C:P ratio||61||0.22|| ||1.97|| || ||0.42 (0.10)|| c || |
|Behl et al. (2011)|
|c||Log algal biomass||24||0.53|| ||3.94||0.11 (0.02)|| c || || || |
|d||Log biovol.-specific absorbance||24||0.29|| ||−6.09||0.51 (0.17)|| b || || || |
|e||Log POC-specific absorbance||24||0.11|| ||0.89||0.25 (0.15)||n.s.|| || || |
|f||Log algal biomass d 6&11||66||0.28|| c ||1.37||0.26 (0.11)|| a ||0.81 (0.19)|| c ||0.53|
|g||Log seston molar C:P ratio d 6&11||66||0.26|| c ||0.69||0.29 (0.12)|| a ||0.83 (0.20)|| c ||0.58|
|h||Log No. of surviving founders||80||0.26|| c ||−0.51||0.37 (0.10)|| c ||0.58 (0.16)|| c ||0.98|
|i||Log No. of gravid founders (day 4)||80||0.23|| c ||−0.41||0.18 (0.05)|| c ||0.24 (0.08)|| b ||1.15|
|j||Log Daphnia biomass (founders)||80||0.32|| c ||1.34||0.45 (0.10)|| c ||0.68 (0.16)|| c ||0.99|
|k||Log Daphnia biomass (juveniles)||80||0.28|| c ||−0.39||0.86 (0.21)|| c ||1.24 (0.32)|| c ||1.06|
|l||Log founders relative yield||31||0.42|| ||−0.25||0.62 (0.13)|| c || || || |
|m||Log juveniles relative yield||28||0.05|| ||−0.06||0.30 (0.26)||n.s.|| || || |
In the grazer experiment, standardized partial regression coefficients (SPRCs) were used as a measure of the relative contributions of light supply and species richness to the response variables. SPRC was calculated as
where b is the regression coefficient of the independent variable x (light or species richness), and s is the standard error of the independent (x) and dependent (y) variables, as determined in the multiple regression (Table 3). The relative contributions of light supply and species richness to a response variable was calculated as the ratio SPRCSR/SPRCLight, where ratios >1 indicate a larger relative contribution of species richness (SR) and ratios <1 indicate a larger relative contribution of light supply.
To further explore whether effects of algal species richness propagated to the herbivore level, we calculated the relative biomass yield of Daphnia as the ratio of observed Daphnia biomass in a the given polyculture Pi to the Daphnia biomass expected from monocultures of the algal species contributing to polyculture Pi as
where ZPi is Daphnia biomass in polyculture Pi with i algal species, ZMj is Daphnia biomass in monoculture of algal species j, and kj(Pi) = 1/i is the proportional contribution of algal species j to total algal biomass in polyculture Pi at the start of the experiment. For statistical analyses, relative yield was log transformed. Thus, overyielding occurred when the log transformed ratio was positive, and underyielding when it was negative. The occurrence of zero values (no surviving Daphnia) was addressed in two ways: either by excluding zero values from the analysis or by addition of the average biomass of one individual Daphnia to the values of ZPi and ZMj prior to log transformation. Results were similar and we only report the ones where zero values were excluded from the analysis.