Hydrobiologia 369/370: 27–43, 1998. 27 M. Alvarez-Cobelas, C. S. Reynolds, P. Sanchez-Castillo & J. Kristiansen (eds), Phytoplankton and TrophicGradients. c 1998 Kluwer Academic Publishers. Printed in Belgium. Summer phytoplankton assemblages across trophic gradients in hard-water reservoirs M. J. Das´ ı, M. R. Miracle, A. Camacho, J. M. Soria & E. Vicente Departament de Microbiologia i Ecologia, Facultat de Biologia, Universitat de Val` encia, E-46100 Burjassot, Spain Key words: phytoplankton, trophic level, reservoirs, hypertrophic ecosystems, nutrients, diversity Abstract Summer phytoplankton assemblages are described and characterised according to their prevalence in a series of hard- water reservoirs of eastern Spain that had been classified in trophic categories on OECD criteria. Distribution patterns of phytoplankton species were ordinated statistically by principal components analysis (PCA). The first component was strongly related to trophic gradient and it particularly discriminated the eutrophic and hypertrophic reservoirs. The second component segregated life-forms, so that (1), on the oligo-mesotrophic side, large dinoflagellates were separated from small centric diatoms, unicellular chrysophytes and filamentous ullotrichales and, on the eu- hypertrophic side (2), colonial greens and large desmids were separated from unicellular volvocales and small centric diatoms. The large differences between eutrophic and hypertrophic reservoirs were also clearly identified in a second PCA, in which physical and chemical factors were used with the principal components solved from the phytoplankton data. From these results, a new trophic category was discerned, for which we propose the name ‘holotrophic’. This category applies to water bodies having the following main features: (1) concentrations of chorophyll, total P and total N in the range of the hypertrophic systems, but with much higher concentrations of dissolved phospho- rus and ammonia and (2) phytoplankton predominantly composed by unicellular green flagellates (Pteromonas, Chlamydomonas) and chlorococcales (Scenedesmus), without cyanobacterial blooms. Introduction Ortiz-Casas & Pe˜ na (1984) showed that OECD crite- ria (OECD, 1982), relating nutrient load to eutroph- ication responses, are valid for Spanish reservoirs. When broad-scale patterns of phytoplankton distrib- ution were explored in a set of more than 100 Spanish reservoirs, trophic level was also an important explana- tory factor (Margalef et al., 1982; Sabater & Nolla, 1991; Riera et al., 1992). However the resulting ordi- nations of the reservoirs shown in these works, based on phytoplankton species composition, consistently matched ordinations based on physical and chemical varables, in which two types of reservoirs, according to siliceous and calcareous rock dominance, were first established (Margalef et al., 1976). In the western part of the Iberian Peninsula, watersheds lie on igneous rocks or slates, which are responsible for the low alka- linities of reservoir waters in the region (less than 1 meq l 1 ). In contrast, watersheds in the eastern part, lie on limestone and sedimentary soils and reservoirs gener- ally have high alkalinities. Subgroups within these two sets of reservoirs were then distinguished by troph- ic level (indicated mainly by Chlorophyll-a), which became a clearer descriptor when only summer data was used (Estrada, 1975; Riera et al., 1992). Phytoplankton composition has been shown to be influenced by two composite factors: water hardness and trophic status. The outcome of their interaction was recognised long ago (Moss, 1973). The important role of these two main variables (essentially, pH-inorganic C availability, on the one hand; the nutrients, N and P, on the other) in regulating phytoplankton composi- tion has also been revealed in field experiments (e.g.