Someone once said plants are the ugly stepchildren of the toxicological world. This was not out of lack of respect for plants, but rather reflected the common assumption that aquatic plants were less sensitive than aquatic fauna to chemicals. We now know this is not a valid generalization. Still, too little attention is given to what information would be needed to protect plants. Three primary issues arise when dealing with protection goals for aquatic plants: what species to test, the relevance of the effects we detect, and the level of protection required. Our experience is primarily with aquatic plants—although much of what follows applies to terrestrial plants (and to animals as well).
The first issue deals with what species or groups we should test. Algae belong to 4 separate Kingdoms: the Bacteria (cyanobacteria, prochlorophytes), the Protozoa (e.g., euglenoids, dinoflagellates), the Chromista (e.g., diatoms, brown algae), and Plantae (e.g., green algae, desmids, red algae). If the few aquatic lichens are included, there is a fifth Kingdom—the Fungi. The Kingdom Plantae also includes nonvascular bryophytes (hornworts, liverworts, mosses), seedless vascular plants (true ferns, club mosses, horsetails), and flowering vascular plants (monocots, dicots). When evaluating the effects of a toxicant on the aquatic plant community, we are lucky if there are data available for a duckweed and a couple of microalgal species. If there are more data available, it is usually because the compound of interest is a known or suspected phytotoxicant. For compounds not considered phytotoxic, plant toxicity results often are considered nonessential. The assessment goal is not to have an even representation of all major aquatic plant taxonomic or morphological groups. The goal, rather, is to have as accurate a portrayal as practical of the distribution of plant sensitivities to a given toxicant.
A recent toxicological evaluation using species sensitivity distributions (SSDs) of the representativeness of the 5 standard aquatic plant species used for the registration of pesticides in the United States—4 microalgal species and 1 of duckweed—shows conclusions using these data are reasonably similar to those using data sets with significantly more species (http://www.regulations.gov/#!documentDetail;D=EPA-HQ-OPP-2011-0898-0012). This analysis included a limited number of compounds, so there is the need for more comparisons; however, the task of representing the sensitivity of aquatic plants may not be as daunting as one might imagine. Nevertheless, there is a catch. Although there is no evidence to suggest that the underrepresented plant groups are uniquely sensitive, neither is there any evidence that they are not. As stated above, the true goal of the compilation of aquatic plant toxicity results is to determine a distribution of sensitivities relevant to the majority of aquatic plants. The validity of this distribution, however, cannot be fully assessed without consideration of plants that are rarely, if ever, tested. This is especially true for marine and freshwater macrophytes. In addition, plants have a wide variety of morphologies, physiologies, and life history patterns. Plants often have more than one free-living, “adult” phase. In some species, the gametophyte generation dominates; in others, the sporophyte generation dominates. Few, if any, studies are available comparing the relative sensitivities of different life history patterns or different life stages.
The second issue relative to protection goals is the toxic effect that is biologically meaningful, and which biological variables are sensitive and relevant. Should we focus on EC50 or EC20 values, or something else? It can make a big difference (Figure 1). What might appear as a minimal community effect based on an SSD using EC50 values can have a much greater effect with other percentiles. Some argue that because several aquatic plants (e.g., phytoplankton) rapidly reproduce and grow, we can use a greater effect percentile in SSDs due to the potential of recovery. Although this is worth considering, there perhaps is a good reason why their growth is so fast—herbivores graze them quickly. An issue related to determining a preferred percent effect is the selection of appropriate biological variables. Survival is rarely assessed. Sublethal effects have included biomass-related (e.g., germination, root elongation, total biomass, leaf injury, pigment content), activity-related (e.g., CO2 uptake, O2 evolution, variable fluorescence), and biochemical parameters (e.g., adenosine 5′-triphosphate levels, enzyme activities). In practice, the variables measured usually relate to growth, but there is no reason not to consider multiple biological variables and point estimates in plant SSDs.
Some scientists advocate consistency in endpoint selection for SSDs, but the approach should be a distribution of “sensitivity” and not a distribution of identical endpoints. There may be a need to use different endpoints for different groups. Vegetative growth may be sufficient for phytoplankton, but other species may need something different. For example, wild rice is an important freshwater annual species—effects on seed germination or flowering are important parameters. Unlike animal chronic toxicity tests (that may include full or partial life cycle tests, and early life stage tests), many aquatic plant tests are generally analogous to testing the growth of “adults.” This is a potential issue because sexual reproduction in seaweeds can be the most sensitive life history stage, and the relative sensitivity of sexual reproduction in freshwater algae or aquatic vascular plants in general is unknown. Using a variety of endpoints may be necessary to represent the true range of sensitivity within an aquatic plant community.
Issue number 3 is the often cited assumption of functional redundancy. There are 2 ways to think about the relationship between species diversity and ecosystem function, the “functionalist” and “compositionalist” perspectives (Rosenfeld 2002). If the goal is to protect ecosystem function, then there has to be agreement on what functions to protect. This perspective often manifests itself as some aspect of functional redundancy. This means “that different species perform the same functional role in ecosystems so that changes in species diversity does not affect ecosystem functioning” (Loreau 2004). Clearly, the focus cannot be just some specific physiological rate or amount of biomass, because no one would argue that equivalent rates or biomasses between an alga responsible for a harmful bloom would be comparable to a nontoxic species. In fact, there is evidence that different strains (one toxic and the other not) of the same species may be functionally distinct. Strains of the cyanobacterium Microcystis aeruginosa have different sensitivities to Cd—with the toxic strain more tolerant (Zeng et al. 2009). Additionally, different communities of microalgae living on surfaces in streams may look similar, but have very different fatty acid compositions, which in turn could significantly affect macroinvertebrate consumers (Torres-Ruiz et al. 2007). Even if there could be agreement on what function to consider, one cannot automatically assume the most sensitive species in a community are the redundant ones.
The other perspective is one of protecting the biodiversity of an ecosystem. The assumption is that if we protect the composition of an ecosystem, then we protect that system's functions. This is probably true, except it is not practical to protect everything all the time (and regulations generally do not attempt to do so) (USEPA 1985). The issue is where we draw the line with respect to plants. With animal species SSDs, the lower 5th percentile is often used. With plants, some have made the case that the 5th percentile may be too conservative because of either functional redundancy or the fact that many aquatic plants, especially phytoplankton, grow so fast, which enables their recovery from temporary growth reductions. However, as stated earlier, functional redundancy is not a well-tested hypothesis for aquatic plants and a fast growth rate may be a result of rapid grazing.
Finally, there is increasing recognition of the importance of plants to the risk assessment process; however, research efforts are few and the pace is slow. Groups are working on recommendations for enhancing assessments of risk to aquatic macrophytes (Maltby et al. 2010). The efforts of these groups, although significant, are limited to freshwater macrophytes and focus on herbicides. Attention needs to expand beyond known and suspected phytotoxins to include commonly detected toxicants in freshwater and saltwater environments. Aquatic animal toxicity data for phytotoxins guard against unintended consequences to animals. We should make the same effort relative to plants and nonphytotoxins—it is not an insurmountable task.