Crassulacean acid metabolism (CAM) is a CO2-concentrating mechanism selected in response to aridity in terrestrial habitats, and, in aquatic environments, to ambient limitations of carbon. Evidence is reviewed for its presence in five genera of aquatic vascular plants, including Isoetes, Sagittaria, Vallisneria, Crassula, and Littorella. Initially, aquatic CAM was considered by some to be an oxymoron, but some aquatic species have been studied in sufficient detail to say definitively that they possess CAM photosynthesis. CO2-concentrating mechanisms in photosynthetic organs require a barrier to leakage; e.g., terrestrial C-4 plants have suberized bundle sheath cells and terrestrial CAM plants high stomatal resistance. In aquatic CAM plants the primary barrier to CO2 leakage is the extremely high diffusional resistance of water. This, coupled with the sink provided by extensive intercellular gas space, generates daytime CO2(p(i)) comparable to terrestrial CAM plants. CAM contributes to the carbon budget by both net carbon gain and carbon recycling, and the magnitude of each is environmentally influenced. Aquatic CAM plants inhabit sites where photosynthesis is potentially limited by carbon. Many occupy moderately fertile shallow temporary pools that experience extreme diel fluctuations in carbon availability. CAM plants are able to take advantage of elevated nighttime CO2 levels in these habitats. This gives them a competitive advantage over non-CAM species that are carbon starved during the day and an advantage over species that expend energy in membrane transport of bicarbonate. Some aquatic CAM plants are distributed in highly infertile lakes, where extreme carbon limitation and light are important selective factors. Compilation of reports on diel changes in titratable acidity and malate show 69 out of 180 species have significant overnight accumulation, although evidence is presented discounting CAM in some. It is concluded that similar proportions of the aquatic and terrestrial floras have evolved CAM photosynthesis. Aquatic Isoetes (Lycophyta) represent the oldest lineage of CAM plants and cladistic analysis supports an origin for CAM in seasonal wetlands, from which it has radiated into oligotrophic lakes and into terrestrial habitats. Temperate Zone terrestrial species share many characteristics with amphibious ancestors, which in their temporary terrestrial stage, produce functional stomata and switch from CAM to C-3 Many lacustrine Isoetes have retained the phenotypic plasticity of amphibious species and can adapt to an aerial environment by development of stomata and switching to C3 However, in some neotropical alpine species, adaptations to the lacustrine environment are genetically fixed and these constitutive species fail to produce stomata or loose CAM when artificially maintained in an aerial environment. It is hypothesized that neotropical lacustrine species may be more ancient in origin and have given rise to terrestrial species, which have retained most of the characteristics of their aquatic ancestry, including astomatous leaves, CAM and sediment-based carbon nutrition. In both terrestrial and aquatic CAM plants, dark CO2 fixation may result in net carbon uptake plus the conservation of carbon by refixation of respiratory CO2. In aquatic plants, CAM's contribution to the total carbon budget is variable. Exemplary studies of the contribution of CAM to the carbon budget, such as those by Boston and Adams, Madsen, and Robe and Griffiths for lacustrine species, are needed in a greater range of habitats. Quantitative estimates of the CAM contribution to the carbon budget are likely to provide more insights than attempts to typologically categorize variation with terms such as "idling," "cycling," AAM, SCAM, TAAM, and so forth. Although we have a reasonably good understanding of the selective factors favoring CAM in seasonal pools and oligotrophic lakes, there are other habitats (Section VII.C) where the role of CAM is not apparent. These species need to be examined in greater detail. Future research should focus on species with predictable diel acid fluctuations, but with characteristics that do not fit recognized criteria for CAM. Of particular interest is the seasonal pool species Downingia bella (Campanulaceae), which may reflect an innovative CAM mechanism. Other roles for dark CO2 fixation should be examined. Dark CO2 fixation may be important as a source of carbon skeletons for both carbon and nitrogen assimilation, particularly in nutrient-poor habitats. Of practical concern is the manner in which lake acidification and eutrophication alter carbon budgets (e.g., Robe & Griffiths, 1994). Also, in many parts of the globe aquatic CAM spe-` cies are threatened: I. andicola of Peru, for instance, is clearly threatened by habitat loss (Leon & Young, 1996), and two of the three primitive Isoetes, morphologically similar to the extinct Isoetites, are apparently extinct (Hickey, 1986). At the other extreme, the aquatic CAM Crassula helmsii is an aggressive alien (Dawson & Warman, 1987), in need of further studies such as those of Newman and Raven (1995) in a greater range of habitats. Isoetes, being the oldest lineage of CAM plants, potentially holds further interesting discoveries with respect to photosynthetic patterns. The most primitive species in the group are distinct in their lack of the typical terete "isoetid" leaf. These species are restricted to isolated sites in South America and have seldom been collected. They are apparently basal to the group, sharing the laminate leaf characteristic with the extinct and possibly ancestral Isoetites (Hickey, 1986). The hypothesized amphibious origin for CAM suggests the possibility that these primitive species may lack CAM. Further study of the photosynthetic metabolism and habitat characteristics of these would be a stimulating contribution to the story of aquatic CAM photosynthesis. Here, and in other aspects of aquatic CAM photosynthesis, a multitude of possibilities are presented with new molecular genetic techniques, now being applied to terrestrial CAM plants (Cushman & Bohnert, 1997).
