1. The tropical Cerrado savannah, the second-largest South American biome, is undergoing the most rapid land-use change in South America, including a wholesale shift from native vegetation to pasture and cropland. However, the consequences of this development for aquatic ecosystem functioning remain unclear. In this study, we investigated how agriculture affects chemical, physical and biological characteristics of headwater streams in the Brazilian Cerrado, and how these changes relate to an important ecosystem function, i.e. ecosystem metabolism. 2. Three paired samplings of pristine and agricultural streams were conducted considering the most abundant morphotypes of Cerrado streams. Nutrient concentrations, stream morphometry and hydrodynamics, sediment transport, as well as benthic microbial biomass (BMB), and its chlorophyll a (Chl-a) content were measured. In parallel, wholestream gross primary production (GPP) and community respiration (CR) were estimated using a diel open-channel dissolved O 2 change technique. 3. Agricultural streams had consistently higher nutrient concentrations, less variability in channel morphology, smaller channel cross-sectional areas, smaller transient storage zones, higher current velocities and higher boundary shear stresses than pristine streams. At base flow, agricultural streams exhibited a midstream band of shifting sediments, while pristine streams had stable sediments. Both agricultural and pristine streams were dominated by thick microbial mats. Due to differences in hydrodynamics and sediment stability, these microbial mats covered the entire stream bed in pristine streams, but were restricted to the stream margins in agricultural streams. 4. As a result, BMB of agricultural streams was diminished by a factor of 7.5 compared to pristine streams. Interestingly, CR of agricultural streams was diminished only by a factor of 2.0, because higher BMB-specific respiration efficiencies, potentially due to increased nutrient availability, compensated for lower BMB due to physical stress. Similarly, the 3.3-fold elevated GPP in agricultural compared to pristine streams was only to a minor degree due to increased benthic Chl-a, but mainly a result of 2.4-fold higher Chl-a-specific primary production efficiencies, potentially due to increased nutrient availability. 5. In conclusion, agriculture impacted the studied Cerrado streams through two antagonistic mechanisms: physical stress resulted in decreased BMB and whole-stream CR by preventing BMB accrual in the central streambed, but higher nutrient availability led to increased primary production and respiration efficiencies in marginal zones, as well as higher whole-stream GPP. Finally, measures of whole-stream metabolism were useful Correspondence: Bjö rn Gü cker,

1. Exotic invasive species modify natural food webs in a way frequently hard to predict. In several aquatic environments in Brazil the introduction of Oreochromis niloticus (tilapia) was followed by changes in water quality. Yet, because of its rapid and easy growth, this fish has been used in many aquaculture programmes around the country. 2. To measure the effects of tilapia on the phytoplankton community and on water conditions of a large tropical reservoir in south-eastern Brazil (Furnas Reservoir), we performed two in situ experiments using three controls (no fish) and three tilapia enclosures (high fish density). Abiotic and biotic parameters were measured at 4 day intervals for 28 days. 3. Fish presence increased nitrogen (N) and phosphorus (P) availability (ammonium 260 and 70% mean increase -first and second experiment; and total phosphorus 540 and 270% mean increase) via excretion. Nutrient recycling by fish can thus be significant in the nutrient dynamics of the reservoir. The higher chlorophyll a concentration in the experimental fish tanks (86 and 34 lg L )1 , first and second experiment, respectively) was the result of a positive bottom-up effect on the phytoplankton community (approximately 2 lg L )1 in the reservoir and control tank). 4. Because tilapia feed selectively on large algae (mainly cyanobacteria and diatoms), several small-sized or mucilaginous colonial chlorophyceans proliferated at the end of the experiments. Thus, the trophic cascade revealed strong influences on algal composition as well as on biomass. 5. Tilapia can contribute to the eutrophication of a waterbody by both top-down and bottom-up forces. In particular, by supplying considerable amount of nutrients it promotes the increase of fast growing algae. Tilapia must be used cautiously in aquaculture to avoid unexpected environmental degradation.

No abstract

Although the cyanobacterial toxin microcystin has been detected in Canadian fresh waters, little is known about its prevalence on a national scale. Here, we report for the first time on microcystin in 246 water bodies across Canada based on 3474 analyses. Over the last 10 years, microcystins were detected in every province, often exceeding maximum guidelines for potable and recreational water quality. Microcystins were virtually absent from unproductive systems and were increasingly common in nutrient-rich waters. The probable risk of microcystin concentrations exceeding water quality guidelines was greatest when the ratio of nitrogen (N) to phosphorus (P) was low and rapidly decreased at higher N:P ratios. Maximum concentrations of microcystins occurred in hypereutrophic lakes at mass ratios of N:P below 23. Our models may prove to be useful screening tools for identifying potentially toxic “hotspots” or “hot times” of unacceptable microcystin levels. A future scientific challenge will be to determine whether there is any causal link between N:P ratios and microcystin concentrations, as this may have important implications for the management of eutrophied lakes and reservoirs.

Microcystis is a genus of freshwater cyanobacteria, which causes harmful blooms in ecosystems worldwide. Some Microcystis strains produce harmful toxins such as microcystin, impacting drinking water quality. Microcystis colony morphology, rather than genetic similarity, is often used to classify Microcystis into morphospecies. Yet colony morphology is a plastic trait, which can change depending on environmental and laboratory culture conditions, and is thus an inadequate criterion for species delineation. Furthermore, Microcystis populations are thought to disperse globally and constitute a homogeneous gene pool. However, this assertion is based on relatively incomplete characterization of Microcystis genomic diversity. To better understand these issues, we performed a population genomic analysis of 33 newly sequenced genomes mainly from Canada and Brazil. We identified 17 Microcystis clusters of genomic similarity, five of which correspond to monophyletic clades containing at least three newly sequenced genomes. Four out of these five clades match to named morphospecies. Notably, M. aeruginosa is paraphyletic, distributed across 12 genomic clusters, suggesting it is not a coherent species. A few clades of closely related isolates are specific to a unique geographic location, suggesting biogeographic structure over relatively short evolutionary time scales. Higher homologous recombination rates within than between clades further suggest that monophyletic groups might adhere to a Biological Species-like concept, in which barriers to gene flow maintain species distinctness. However, certain genes-including some involved in microcystin and micropeptin biosynthesis-are recombined between monophyletic groups in the same geographic location, suggesting local adaptation.

Cylindrospermopsis raciborskii (Wołosz.) Seenayya et Subba Raju is a planktonic filamentous cyanobacterium whose sudden worldwide proliferation and ability to produce toxins are a reason for concern. In this paper, we suggest that its ecological dominance might be explained by antagonistic interaction with other phytoplankton species due to production of allelopathic metabolites. To test this hypothesis, experiments were run with exudates of natural phytoplankton and C. raciborskii strains isolated from Lagoa Santa, a small natural lake in southeastern Brazil, where this species has become dominant in recent years. The exudates were added to different algal species obtained from the same environment and maintained in culture. After 24 h incubation, PAM fluorometry was used to compare control and treatment photosynthetic responses (relative electron transport rate) to the dissolved extracellular products. Results indicate that most of the target species were sensitive to C. raciborskii exudates, which showed strong inhibitory effects on their photosynthetic activities. These results provide evidence that allelopathy may offer a competitive benefit to C. raciborskii and contribute to its stable dominance in Lagoa Santa. A potential allelopathic advantage could also help to explain the geographic expansion of this species at midlatitudes.

a b s t r a c tThis study presents data on seasonal changes of the phytoplanktonic community of a small tropical lake, Lagoa Santa, in the Brazilian cerrado. Temporal variation in the community structure was small and we observed a permanent dominance of the cyanobacteria Cylindrospermopsis raciborskii. This dominance could be associated to the stability of several environmental conditions. Especially water temperature, always within the optimal growth range for C. raciborskii, and the constant mixed water column may have been important factors driving to the long term dominance of this species in the lake. Remarkable seasonal variation in rainfall, as well as the occasional and non-seasonal variation in nutrient concentrations, were important to explain changes in environmental and biological variables, but were not related to the dominance of C. raciborskii. Pearson's correlation and PCA could just in part explain the stable dominance of this cyanobacterium. Meteorological and chemical factors seem to have no clear control on the variability of the phytoplankton dynamics in Lagoa Santa. The weak relationship between C. raciborskii abundance and environmental variables points to additional intrinsic factors associated to this species that may be important in structuring the phytoplankton assemblage.

A series of 22 lakes in southern Quebec spanning a wide trophic range were sampled to develop models of changes in cyanobacterial abundance and toxicity. All lakes contained toxic cyanobacteria, and epilimnetic toxin content, expressed as microcystin equivalents, was best predicted by total nitrogen concentration and total phosphorus concentration (TP). Although phytoplankton biomass increased linearly with increases in TP among lakes, toxigenic biomass increased as greater than the squared power of TP. The only potentially toxigenic genera whose biomass was correlated with microcystin concentration were Microcystis and Anabaena. Surprisingly, the best model for toxicspecies biomass was based on epilimnetic nitrogen. The level of the hepatotoxin microcystin per unit biomass in these organisms did not vary markedly among lakes, supporting the idea that environmental factors control the occurrence, but have only a limited effect on the toxicity, of potentially toxic species.Résumé : Nous avons échantillonné une série de 22 lacs du sud du Québec couvrant une large gamme trophique afin de mettre au point des modèles du changement d'abondance et de toxicité des cyanobactéries. Tous les lacs contiennent des cyanobactéries toxiques et le contenu de toxines dans l'épilimnion, exprimé en équivalents de microcystine, est prédit le mieux par les concentrations d'azote total (TN) et de phosphore total (TP). Bien que la masse du phytoplancton s'accroisse de façon linéaire en fonction de l'accroissement de TP parmi les lacs, la biomasse des toxigènes augmente plus rapidement que la puissance au carré de TP. Les seuls genres potentiellement toxigènes dont la biomasse est en corrélation avec la concentration de microcystine sont Microcystis et Anabaena. De façon inattendue, le meilleur modèle de la biomasse des espèces toxiques est basé sur l'azote de l'épilimnion. La concentration chez ces organismes de l'hépatotoxine microcystine par unité de biomasse ne varie pas considérablement d'un lac à l'autre, ce qui appuie la proposition que les facteurs du milieu contrôlent la présence des espèces potentiellement toxiques, mais qu'ils ont un effet limité sur leur toxicité.[Traduit par la Rédaction] Giani et al. 2109