In his classic critique of food-web ecology, Don Strong argued that trophic cascades (TCs) are ‘all wet,’ meaning, TCs occur more frequently and with greater strength in aquatic ecosystems characterized by low-diversity, linear food-chains.  In contrast, he argued that TCs are weak or non-existent in diverse systems where reticulate food-webs allow energy to flow along alternate pathways.  Two recent meta-analyses investigated Strong’s hypothesis and arrived at divergent conclusions; one found that increasing diversity reduces TCs whereas the second found the magnitude of TCs are not related to diversity.  These conflicting results highlight the need for controlled experiments designed to test Strong’s hypothesis.  We report results from manipulative experiments that measured the strength of TCs in 18 streams throughout California’s Sierra Nevada mountain range.  We then asked whether variation in the magnitude of TCs could be explained by variation in the diversity of algal species that support the base of the food-web.  To quantify the strength of TCs, we placed three standardized substrates (ceramic tiles) in each stream and used electrical exclusions to vary the number of trophic levels (TLs) with access to the substrates (1 TL = algae, 2 TLs = algae+invertebrate herbivores, or 3 TLs = algae+herbivores+fish). After a 60 day incubation period, we estimated the strength of a TC as the log ratio of algal biomass (as chlorophyll a) on the 2 vs. 3 TL tiles. 

Results/Conclusions
Predatory fish spent ~94% more time foraging on 3 TL tiles than on 1 and 2 TL tiles with electrical exclusions (ANOVA, F_2,309 = 306.2, p < 0.01), and herbivore densities were 79-87% lower on 1 and 3 TL tiles than on 2 TL tiles (ANOVA, F_{2, 57} = 8.1, p <0.01).  The reduction in herbivore densities translated to a mean 42-59% increase in the biomass of algal primary producers (ANOVA, F_{2, 51} = 3.4, p = 0.04).  The strength of TCs significantly decreased as the diversity of algae increased (R² = 0.23, p = 0.05), a trend consistent with Strong’s hypothesis that systems with greater biodiversity have dampened TCs.  No single algal species, functional, or taxonomic group appeared to be driving the magnitude of TCs among streams (MRPP, all p > 0.67), suggesting our results were general. Our data indicate that biodiversity conservation at the level of primary production has the potential to reduce or eliminate TCs, thereby preventing the runaway consumption that is the hallmark of a strong TC.