Many marine organisms form photosymbioses with zooxanthellae, but some, such as the medusozoans, are less well known. Here, we summarize the current knowledge on the diversity of zooxanthellate jellyfishes, to identify key traits of the holobionts, and to examine the impact of these traits on their ecology. Photosymbiosis with zooxanthellae originated at least seven times independently in Medusozoa; of these, five involve taxa with medusae. While most zooxanthellate jellyfishes are found in clades containing mainly non-zooxanthellate members, the sub-order Kolpophorae (Scyphozoa: Rhizostomeae) is comprised-bar a few intriguing exceptions-of only zooxanthellate jellyfishes. We estimate that 20-25% of Scyphozoa species are zooxanthellate (facultative symbiotic species included). Zooxanthellae play a key role in scyphozoan life-cycle and nutrition although substantial variation is observed during ontogeny, or at the intra-and inter-specific levels. Nonetheless, three key traits of zooxanthellate jellyfishes can be identified: (1) zooxanthellate medusae, as holobionts, are generally mixotrophic, deriving their nutrition both from predation and photosynthesis; (2) zooxanthellate polyps, although capable of hosting zooxanthellae rarely depend on them; and (3) zooxanthellae play a key role in the life-cycle of the jellyfish by allowing or facilitating strobilation. We discuss how these traits might help to explain some aspects of the ecology of zooxanthellate jellyfishes-notably their generally low ability to outbreak, and their reaction to temperature stress or to eutrophication-and how they could in turn impact marine ecosystem functioning. 6 Box 3: Glossary We give here some of the technical terms used in this review. For more information on jellyfish anatomy, development and taxonomy, see Arai 1997 and Bouillon et al. 2006.

There has been an upsurge of interest in trait-based approaches to zooplankton, modelling the seasonal changes in the feeding modes of zooplankton in relation to phytoplankton traits such as size or motility. We examined this link at two English Channel plankton monitoring sites south of Plymouth (L4 and E1). At L4 there was a general transition from diatoms in spring to motile microplankton in summer and autumn, but this was not mirrored in the succession of copepod feeding traits; for example the ambushing Oithona similis dominated during the spring diatom bloom. At nearby E1 we measured seasonality of food and grazers, finding strong variation between 2014 and 2015 but overall low mesozooplankton biomass (median 4.5 mg C m-3). We also made a seasonal grazing study of five copepods with contrasting feeding modes (Calanus helgolandicus, Centropages typicus, Acartia clausi, Pseudocalanus elongatus and Oithona similis), counting the larger prey items from the natural seston. All species of copepod fed on all food types and differences between their diets were only subtle; the overriding driver of diet was the composition of the prey field. Even the smaller copepods fed on copepod nauplii at significant rates, supporting previous suggestions of the importance of intra-guild predation. All copepods, including O. similis, were capable of tackling extremely long (>500 µm) diatom chains at clearance rates comparable to those on ciliates. Maximum observed prey:predator length ratios ranged from 0.12 (C. helgolandicus) up to 0.52 (O. similis). Unselective feeding behaviour and the ability to remove highly elongated cells have implications for how copepod feeding is represented in ecological and biogeochemical models.

The Plymouth L4 time plankton series in the Western English Channel is a textbook example of a shallow, stratifying shelf sea system. Over its 30 yr of weekly sampling, this site has provided a diverse and contrasting suite of numerical and conceptual models of plankton bloom formation, phenology, and seasonal succession. The most recent of these papers, Kenitz et al. (2017) has initiated this comment, partly because we feel that it has presented a slightly misleading picture of the plankton composition at this site, and of a robust, recurring seasonal succession. We address this by illustrating the extent of inter‐annual variability in phenology that occurs at the site, and which needs to be captured better within models. However our main aim is to foster a much better integration of the variety of top‐down and bottom‐up processes that have all been suggested to be key in driving seasonal succession. Some of these, particularly the multiple grazing and growth controls contributing to the so‐called “loophole hypothesis” may be complementary, but others, such as the role of copepod feeding traits in driving species succession (Kenitz et al. 2017) offer testable competing hypotheses. The basic assumptions and outputs of all these models need to be validated more critically, both against time series data and process studies that include the finding of unselective feeding. We suggest that the variability in plankton phenology (and not just mean timing and amplitude) could be used to diagnose the performance of alternative models of plankton succession.