Vorobyovy gory, 119899 Moscow, Russia

ABSTRACT

The nature of phytoplankton community rearrangement caused by a change in N:P ratio is shown for both artificial laboratory microcosms and natural algocoenoses in vitro and in situ. The experiments have shown that high N:P weight ratios (20 to 50) favor the development of Chlorococcales, while an N:P lowering to the values of 5 to 10 leads to a community dominated by Cyanophyta. As shown by variation modelling, relative abundance of phytoplankton species should depend only on the relative amounts of N and P in the environment, so that the optimal N:P ratio for a given species is equal to the ratio of minimum cell requirements for these elements. An empirical proof of this law resides in the fact that for some species of Chlorococcales and Cyanophyta the ratios of requirements for N and P calculated in the experiment were close to their optimal environmental concentration ratios. For instance, an N:P increase from the value of 4 to 25-50 in the water of fish-breeding ponds has led to an increased abundance of Chlorococcales, due to mainly Scenedesmus quadricauda, having roughly the same requirements ratio. The N:P ratio may be recognized as an independent ecological factor, distinct from the nitrogen and phosphorus concentrations taken separately.

Keywords: Phytoplankton; Nutrients; Controlling the structure of community; Cell requirements.

A significant amount of papers present the results of laboratory chemostat experiments. Those experiments had the following common scheme: a varied sequence of biogenic element ratios was created by adding different substances to the medium and then the optimal ratios for the community members were found.

Holm and Armstrong (1981) grew a polyculture of two algae, the diatomic Asterionella formosa Hass. and the cyanobacterium Microcystis aeruginosa Kutz. emend Elenk. in the continuous cultivation regime. In the course of the experiment the atomic ratio Si:P ranged from 2 to 200. As the ratio increased, the biomass ratio of the two species changed from 1:99 to 96:4. It turned out that A. formosa is a species with a greater demand for silicon but gains an advantage over the rival in the condition of phosphorus deficiency. It is noteworthy that similar biomass ratios formed at different absolute concentrations of silicon and phosphorus but when their ratios in the medium were similar. For instance, the alga M. aeruginosa could dominate both at a ratio of 25:1 (98 % of the total biomass) and at a ratio of 100:10 (97 %), while A. formosa dominated both at 300:2 (93 %) and 100:0.5 (96 %).

Tilman (1977) studied a mixed continuous culture of diatoms A. formosa and Cyclotella meneghiniana Kutz. under a wide range of Si:P ratios. When the molar ratio of these biogenic elements was smaller than 6, A. formosa dominated, while when the ratio exceeded 90, C. meneghiniana evidently dominated.

The growth of the cyanobacterium Synechococcus Nag. in a polyculture with the green Scenedesmus quadricauda (Turp.) Breb. was studied in the condition of phosphorus deficiency, i.e., at high N:P ratios (Suttle and Harrison, 1988). In this situation Synechococcus suppressed its rival.

In experiments with a culture of two marine algae Skeletonema costatum (Grev.) Cl. and Phaeodactylum tricornutum Bohlin. the impact of temperature, light and biogenic conditions on the result of the competition was investigated.

S.costatum supplanted Ph.tricornutum at low N:Si ratios and high N:P ratios in the medium (de Pauw and Naessens-Foucquaert, 1991). An optimal N:Si:P ratio for S.costatum proved to be 25:25:1.

A variety of species responces under limiting substance variations has been shown at the monoculture level as well. When the productivity of two species of green algae, Scenedesmus quadricauda and Stigeoclonium tenue (Ag.) Kutz., were compared, the critical values of the N:P ratio turned out to be equal to 22 for the former and 17 for the latter (Vries and Klapwijk, 1987). The critical value is the lowest ratio value which, being exceeded, leads to phosphorus growth limitation.

Rhee (1978) concluded on the basis of his experiments with monocultures of Scenedesmus sp. that for this alga the N:P ratio equal to 30 is the growth stimulating one: at higher and lower ratio values the culture growth is limited by nitrogen and phosphorus, respectively. The optimal N:P ratio found for the cyanobacterium Anacystis nidulans P. Richt. is also 30 (Sirenko, 1972).

In our own experiments (Levich and Bulgakov, 1993) we studied optimal nitrogen-to-phosphorus ratios in nutrient environments by cultivating some microalgal species in laboratory polycultures. We attempted to control the species composition of a simplest artificial algocoenosis by varying the initial nitrogen-to-phosphorus concentration ratio.

The experiments were carried out with batch microalgal cultures. The results of species growth in an artificial algocoenosis were estimated by their partial contributions to the total abundance or biomass of the community when the stationary phase of growth had been achieved by all species of the polyculture.

Four species of Chlorococcales participated in the first set of experiments: Scenedesmus quadricauda, Chlorella vulgaris Beyer, Ankistrodesmus falcatus (Corda) Ralfs and Ankistrodesmus sp.. The polyculture grew in two media with the following initial biogenic element concentrations: medium 1 - 11 mg L-1 nitrogen and 3 mg L-1 phosphorus (N:P = 3.5); medium 2 - 50 mg L-1 nitrogen and 2.5 mg L-1 phosphorus (N:P = 20).

In the second set of experiments the artificial community consisted of only two species of Chlorococcales: Scenedesmus quadricauda and Ankistrodesmus falcatus, while the nitrogen-to-phosphorus ratio took a larger set of values: medium 1: 4 mg L-1 nitrogen and 3.1 mg L-1 phosphorus (N:P = 1.3); medium 2: 14 mg L-1 nitrogen and 3.1 mg L-1 phosphorus (N:P = 4.5); medium 3: 34 mg L-1 nitrogen and 0.6 mg L-1 phosphorus (N:P = 57).

The storage-induced culture growth was small as compared with the entire growth, therefore the corresponding corrections were unable to essentially alter the species biomass distribution in all the three media. When the cultures were sown in all the media, the initial polyculture cell count consisted of 20 % S. quadricauda and 80 % A. falcatus. As the N:P ratio increased, S. quadricauda increased its relative abundance which at N:P = 1.3, 3.7 and 57 amounted to 40, 60 and 68 %, respectively.

The researcher's capabilities to control phytoplankton structure broaden when he works with multispecies natural algocoenoses. It becomes possible to analyze the influence of biogenic element ratios not only on separate phytoplankton species but also on whole groups of species, unified either by a systematic marker, or by size. Herewith the stationary laboratory condition enable a tough control of assigned environmental factors.

Sommer (1983), varying the environmental Si:P molar ratio from 4 to 30, cultivated natural multispecies (over 30 species) lake phytoplankton populations. In the deficient silicon condition the alga Mougeotia thylespora Ag. dominated. At approximately balanced ratios the most numerous were Koliella spiculiformis Hind, Synedra acus Kutz. and Asterionella formosa. At deficient phosphorus only S. acus was overwhelmingly dominant. The author observed these results independently of the natural inoculate composition. A comparison with the phytoplankton species abundance showed that when the values of biogenic element ratios were close to the experimental ones, the competition results in the natural reservoir were also similar.

Suttle and Harrison (1988) conducted a number of experiments with natural freshwater phytoplankton in laboratory conditions at N:P ratios equal to 5, 15 and 45. At N:P = 45 an absolute domination of the blue-green Synechococcus was observed. At lower ratios the most numerous were two diatoms, Nitzschia holsatica Hust. and Synedra radians (Kutz.) Hust., along with the green alga Scenedesmus sp.

Kilham (1986), analyzing the growth of phytoplankton from Lake Michigan in the laboratory at different environmental Si:P ratios, also concluded that not a certain specific resource is responsible for the species structure of an algocoenosis but just their partial quantities. He showed that under phosphorus limitation diatoms supplant the representatives of all other divisions from a community; however, the silicon concentration having been also decreased, the domination passes to the greens. In media with a high Si:P ratio (313 and 74) diatomic algae were dominant and the green ones were at low ratios (4.6 and 0.9). The cyanobacteria were eliminated from the experiment since very high nitrogen-to-phosphorus ratios (100) were under study.

Grover (1989), who cultivated a lake algae community using a semicontinuous cultivation, came to similar conclusions. When the medium was supplied with silicon and phosphorus in a ratio of 20, green algae dominated, while at Si:P = 80 diatoms did.

When an algae community from an acid lake (dominating species: the cyanobacterium Anabaena variabilis Kutz.) was inoculated in polyethylene bags, a pH increase and a decrease of the N:P ratio in the environment resulted in a phytoplankton biomass growth without domination change (Wilcox and De Costa, 1990). It was shown that the N:P ratio values higher than 25 were unfavorable for the growth of cyanobacteria.

Oligotrophic lake was fed by mineral fertilizers with the atomic N:P ratio equal to 50. After that samples were taken and the N and P consumption kinetics of two size fractions (smaller and bigger than 3 micrometers) were studied in laboratory. The consumed N:P ratio was calculated as the ratio of maximum consumption rates of these elements. It turned out that the small-size fraction (mainly Synechococcus spp.) is limited by nitrogen, whereas the large one (mainly Rhizosolenia spp. and Cyclotella spp.) by phosphorus. Hence for small-size species an optimal N:P ratio is higher than 50, while for large-size ones it is lower than 50 (Suttle et al., 1991).

We conducted experiments aimed at studying the impact of different ratios of mineral forms of nitrogen and phosphorus on the species and size structure of an algocoenosis when the latter, taken from a fish-breeding pond, was subject to batch cultivation (Levich et al., 1992). Water from a fish-breeding pond was placed in retorts where later mineral forms of nitrogen and phosphorus were added in different quantitative combinations. Five values of the N:P ratio were obtained altogether, with the natural background taken into account: 2, 5, 20, 50, 100. The original pond phytoplankton biomass were the same in all retorts. To eliminate the effect of zooplankton grazing, the water to be studied was prior to the experiment let through a meshed net with mesh size corresponding to minimal zooplankton size and then placed in the darkness for two days. One of the retorts served as a control one, i.e., the original biogenic element concentration in it was equal to the background one. In the course of the experiment all the retorts were kept out of doors. The ultimate biomass was analyzed both for high phytoplankton taxons (Chlorococcales, Chlorophyta, Bacillariophyta and Cyanophyta) and at the level of genera and species. The taxons at the level of genera and species were divided into dominating ones (whose biomass was not less than 20 % of the entire biomass on the sixth day of the experiment at least in a single retort), unrepresentative ones (with a biomass less than 1 %) and subdominant ones (all the rest). Biomass of size classes of algae were also compared.

As early as in the thirties of the present century Pearsall (1930, 1932) was one of the first ones to observe that different chemical element abundance ratios, including N:P and Ca+Mg:Na+K, promote the formation of different phytoplankton communities in the nature.

Smith (1983), having analyzed the situation in 12 lakes of the world, found the domination of cyanobacteria during the periods when the nitrogen-to-phosphorus ratio had values less than 25. At atomic N:P ratios greater than 25 green and diatomic microalgae dominated. In some of the lakes under study this ratio was increased due to the removal of phosphorusabundant sewage; herewith the fraction of green and diatomic, non-nitrogen-fixing algae also increased. Tilman (1982) calls the results revealed by Smith "a dramatic impact of the nitrogen-tophosphorus ratio on the taxonomic composition of lake algocoenoses".

Later Smith (1986) summarized the materials from 22 lakes and took into account the light conditions and morphometry of the reservoirs. The inclusion of these factors in addition to the nitrogen-to-phosphorus ratio did not change the conclusion concerning the significant impact of the biogenic element ratio on the species structure of the plankton.

Pick and Lean (1987), reviewing a number of studies on the influence of different nutrients on cyanobacteria, concluded that for this group of algae, apart from high temperature, relative stability of the water layer and carbon and iron concentrations, certain values of the N:P ratio and the light conditions are important as well.

Schindler (1977) for many years carried out an experimental study of small fertilized lakes. In one of the lakes, fed with fertilizers with the atomic N:P ratio equal to 30 for six years running, algae from the genus Scenedesmus dominated the plankton throughout the experiment. After that the nitrogen- to-phosphorus proportion in the fertilizer was lowered to the value of 11 and the plankton became dominated by cyanobacteria, mainly Aphanizomenon gracile (Lemm.) Elenk. In another lake fertilizer was fed with the N:P ratio equal to 11 throughout the experiment. That resulted in the domination of the nitrogen-fixing cyanobacteria of the genus Anabaena. Schindler emphasizes that the cyanobacteria not only dominate at low N:P ratios in the unfertilized water of the lakes under study, but their domination may be as well provoked by additional fertilizing with proper element ratios added in the course of the experiment. Thus, when the N:P ratio in the fertilizer was as low as 5, it turned out that the phytoplankton biomass increase in the experimental lake (by a factor of 4 to 8 as compared with a non-fertilized lake) took place almost entirely at the expense of cyanobacteria, mainly of the genus Anabaena (Findley and Kasian, 1987).

The dependence of cyanobacteria abundance on low N:P ratios and sufficient phosphorus supply was obtained as well during a study of phytoplankton blooming in Lake Kennedy (Stockner and Shortreed, 1988). Studies in the Soyang lake in Korea, where the role of cyanobacteria increased after a gradual phosphorus concentration increase in 1984-1989 (with a high content of nitrogen) and the corresponding decrease of the N:P ratio from 100 to 50, led to a similar conclusion (Cho et al., 1990). A reverse effect was obtained in the South African hypertrophic impoundment (Haarhoff et al., 1992): measures aimed at water clearing from industrial sewage resulted in the N:P ratio increase in the water from 4 to 25. After that the entire phytoplankton biomass lowered, while the formerly dominating cyanobacteria (Microcystis aeruginosa) were replaced by green algae. An inverse dependence between the number of cyanobacteria and the value of the N:P ratio in lake ecosystems was proved by the method of statistical correlation analysis (Varis, 1991).

So far we dealt with the impact of mineral forms of nitrogen and phosphorus on the phytoplankton structure; however, according to McQueen and Lean (1987), the increase of the annual percentage of cyanobacteria can be induced by the decrease of the ratio between the nitrate nitrogen to the whole amount of phosphorus. Conversely, when this ratio was greater than 5, a mass development of cyanobacteria was never observed.

There are certain data (Blomqvist et al., 1989; Klapwijk, 1990) indicating a deep restructuring of algocoenoses as a result of a changed N:P ratio in the water. Thus, when the Swedish acid lake Njupfatet (the general background abundance of nitrogen and phosphorus were, respectively, 200 and 4 mg L-1, N:P = 50) was fed with 100 mg L-1 of nitrogen and 10 mg L-1 of phosphorus (the N:P ratio equals 10), it resulted in a significant biomass growth of Merismopedia tenuissima Lemm., Peridinium inconspicuum Lemm. and Dictyosphaerium botritella Kom. et Perm., although only the first of these species had been among the dominant ones before the additions were made (Blomqvist et al., 1989). The ratio of inorganic forms of N and P in the basin of the Rhine drastically decreased in the recent 45 years (Klapwijk, 1990). The result is that for these years the phytoplankton species composition became poorer: several taxons disappeared, while the size of the others became many times smaller.

The joint interaction of N, P and Si may exert influence on the composition of a natural algocoenosis as well. Long-term changes in the hydrochemistry of the Rhine delta, expressed in the decrease in the Si:N and Si:P ratios, promoted a mass development of dinoflagellates and cyanobacteria (Admiraal and Vlugt, 1990). One can get rid of this phenomenon at the expense of lowered nitrogen and phosphorus loading.

Shamess et al. (1990) came to the conclusion that the phytoplankton blooming may be controlled not only by the relationship between nitrogen, phosphorus and silicon but also by some other substances and even microelements. In their experiments the molar concentration ratios of sulphate and the molybdenum cation played a dominant role in the community of nitrogen-fixing cyanobacteria. Herewith, as the authors stress, the concentrations of these ions taken separately did not affect the nitrogen fixing process.

The species and size composition of phytoplankton communities is not indifferent to variations of biogenic element ratios in the environment. However, the question remains of how great is the significance of this factor as compared with the others, in particular, with the absolute concentrations of N, P and Si. Such a comparative analysis can be carried out using the data of the on-location pond experiment described in the previous section (Levich and Bulgakov, 1992).

In the 1987 the fertilization system did not incorporate the spring "warming-up" (beginning of fertilization in early April with N:P=4 during initial 2 weeks) and fertilizer addition with an increased frequency; however, that had no effect on the algal community structure regulation phenomenon. The inserted absolute fertilizer doses were different in the experiment and the control over the whole season, while the phytoplankton restructuring effects showed up only in the middle of the year, i.e., just as soon as the N:P ratio increased. At the same time, the absence of a low N:P ratio in the spring period of the third season of the experiments resulted in an earlier (in early June) detection of a biomass redistribution of the Chlorococcales and the Cyanophyta. In the first two seasons, when the fertilizer N:P ratio equal to 4 determined the corresponding substance concentration ratio in the water for a sufficiently long time, the biomass redistribution happened later, in late June or July. Hence it apparently follows that the relationship of biomass of taxons and size groups depends above all not on the increased nitrogen concentration in the water but on the increased nitrogen concentration with respect to phosphorus concentration. The same is indicated by studies of other authors (Holm and Armstrong, 1981; Shamess et al., 1990).

It is well-known that at low nitrogen-to-phosphorus ratios in pond water, resulting in nitrogen limitation, the cyanobacteria, able to assimilate nitrogen from the atmosphere, gain an advantage against the representatives of other taxons. Thus, in hypertrophic lakes a minimal value of the N:P concentration ratio preceded the blooming outbursts of nitrogen fixing species (Barica, 1990). And conversely, the nitrogen enrichment of a reservoir resulted in a slower phytoplankton growth where the nitrogen fixing cyanobacterium Aphanizomenon flos-aquae (L.) Ralfs was dominant (Yelizarova and Korolyova, 1990). It is known, however, that even not nearly all cyanobacteria are able to fix nitrogen. This is true primarily for the representatives of the genera Anabaena and Aphanizomenon having heterocysts. Some algae, having no heterocysts, are also able to fix nitrogen, for instance, the genus Oscillatoria (Carpenter and Price, 1976; Bryceson and Fay, 1981). In the ponds studied by the present authors (Levich and Bulgakov, 1992) these genera were never dominant in the biomass. The difference between the experiment and the control was detected by the decreased biomass of the genera Merismopedia, Phormidium and, in some seasons, Aphanothece and Microcystis, not endowed by a nitrogen fixing mechanism. By other data (Varis, 1992), a limited nitrogen supply in a lake is advantageous not only to nitrogen fixing cyanobacteria, but also to the genus Microcystis. Consequently, cyanobacteria, like other algal taxons, respond to relative rather than absolute amounts of biogenic elements.

No doubt that the difference in responses of different representatives of a microalgal community to a changed substrate factor ratio in the environment is underlain by a certain adaptive physiological mechanism acting at the cell level. Such a mechanism is described by the conception of phytoplankton requirements for mineral nutrition components (Levich et al., 1986; Levich, 1989). By this conception, a species' requirement for nitrogen, phosphorus, etc., is equal to the diminution of the corresponding element in the environment related to a single cell or to a unit amount of biomass. In other words, it is the content of a substance assimilated in a cell (a quota) if respiration, excretion, etc. are neglected. A quota is a specific quantity for a species and for a single species it may vary between a maximum and minimum values in the process of population development. The minimum quota is such a minimum amount of a substrate inside a cell that it is still able to divide. In the logistic growth curve a reached minimum quota means that a steady state phase has been achieved. It is tempting to conclude that if a certain species possesses a certain ratio of minimum quotas of, for instance, nitrogen and phosphorus, then this ratio indicates the degree of its competitiveness at one or another value of the N:P ratio in the water. The biogenic element concentration ratio, closest to the ratio of minimum quotas, is optimal for a given species. Similar ideas have been put forward by Rhee and Gotham (1980) who found optimal N:P ratios for nine microalgal species and stressed that these ratios correspond to the ratios of requirements for these elements.

In a variational mathematical model of a phytoplankton community (Levich, 1980) the optimization theorem was proved (Levich et al., 1983) which asserts that the relative abundance of species should depend only on the ratios of environmental resources available to the community, while the optimal resource ratio for a given species is equal to the ratio of minimum cell quotas of these resources for the same species. As a matter of fact, it was the hypothesis contained in the formulation of the optimization theorem that motivated both some of the experimental work described above and a bibliographic search for information on the role of biogenic element ratios.

Empirical proof of such a correspondence is contained in the fact that the ratios of calculated (Levich and Artiukhova, 1991) demands for N and P for a number of species of Chlorococcales and Cyanophyta were close to their optimal concentration ratios measured in laboratory and on-location experiments (Levich et al., 1992; Levich and Bulgakov, 1992,

1993). In particular, an increase the N:P ratio in the fertilized pond water to the values of 25-50 resulted in an increased abundance of Chlorococcales, mainly due to the species S. quadricauda which possesses roughly the same requirements ratio. It should be stressed that the dominance of Chlorophyta, Cyanophyta, etc. in the phytoplankton does not at all mean an entire identity of quota ratios of all member species of these divisions. It is sufficient that just some species (or even a single species, as is the case with S. quadricauda), the most mass ones, of the corresponding taxon be close in the above characteristic.

Thus, the environmental biogenic element concentration ratio should be recognized as an independent abiotic factor restricting the growth of populations in a community and consequently affecting its structure. It should be emphasized that we have measured the minimum cell quota ratios with a very low accuracy, so that actually only certain ranges of values have been determined. On this basis, as well as according to the general ecological principles (the Shelford rule), one should conclude that a certain range of values of the N:P ratio is optimal for a given species, with the corresponding upper and lower allowed bounds.

The aforesaid does not mean that nitrogen, phosphorus and silicon taken separately cease to be factors limiting the phytoplankton growth. Just their interaction turns out to be so strong (being cased by the interaction of cell requirements) that the ratios of these elements becomes a factor of greater significance than are the absolute concentrations.

Table 1. Fertilization pattern in experimental and control ponds

Remark. "-" Fertilization was begun after the fish stocking.