This paper reviews the rapidly expanding literature on the ecological effects of cyanobacterial toxins. The study employs a qualitative meta-analysis from the literature examining results from a large number of independent studies and extracts general patterns from the literature or signals contradictions. The meta-analysis is set up by putting together two large tables--embodying a large and representative part of the literature (see Appendix A). The first table (Table A.1) reviews the presence (concentrations) of different cyanobacterial toxins in the tissues of various groups of aquatic biota after exposure via different routes, experimentally in the lab or via natural routes in the environment. The second table (Table A.2) reviews the dose dependent effect of toxins on biota. The great majority of studies deal with the presence and effects of microcystin, especially of the MC-LR congener. Although this may partly be justified--MC-LR is an abundant and highly toxic protein--our review also emphasizes what is known about (i) other MC congeners (a number of studies showed a preferred accumulation of the less toxic variant MC-RR in animal tissues), (ii) nodularin (data on a range of biota from studies on the Baltic Sea), (iii) neurotoxins like anatoxin-a(s), which are conspicuously often present at times when mass mortalities of birds occur, (iv) a few studies on the presence and effects of cylindrospermposin, as well as (v) the first examples of ecological effects of newly identified bioactive compounds, like microviridin-J. Data were reorganized to assess to what extent bioconcentration (uptake and concentration of toxins from the water) or biomagnification (uptake and concentration via the food) of cyanobacterial toxins occurs in ecosystems. There is little support for the occurrence of biomagnification, and this reduces the risk for biota at higher trophic levels. Rather than biomagnification biodilution seems to occur in the foodweb with toxins being subject to degradation and excretion at every level. Nevertheless toxins were present at all tropic levels, indicating that some vectorial transport must take place, and in sufficient quantities for effects to possibly occur. Feeding seemed to be the most important route for exposure of aquatic biota to cyanobacterial toxins. A fair number of studies focus on dissolved toxins, but in those studies purified toxin typically is used, and biota do not appear very sensitive to this form of exposure. More effects are found when crude cyanobacterial cell lysates are used, indicating that there may be synergistic effects between different bioactive compounds. Aquatic biota are by no means defenseless against toxic cyanobacteria. Several studies indicate that those species that are most frequently exposed to toxins in their natural environment are also the most tolerant. Protection includes behavioral mechanisms, detoxication of MC and NODLN by conjugation with glutathione, and fairly rapid depuration and excretion. A common theme in much of the ecological studies is that of modulating factors. Effects are seldom straightforward, but are dependent on factors like the (feeding) condition of the animals, environmental conditions and the history of exposure (acclimation and adaptation to toxic cyanobacteria). This makes it harder to generalize on what is known about ecological effects of cyanobacterial toxins. The paper concludes by summarizing the risks for birds, fish, macroinvertebrates and zooplankton. Although acute (lethal) effects are mentioned in the literature, mass mortalities of--especially--fish are more likely to be the result of multiple stress factors that co-occur during cyanobacterial blooms. Bivalves appear remarkably resistant, whilst the harmful effects of cyanobacteria on zooplankton vary widely and the specific contribution of toxins is hard to evaluate.