~700 x 10-6 mol photons (g chla)-1 s-1, on the best performances measured by Bannister and Weidemann (1984). The maximum area projected by a single Asterionella cell is approximately 200 x 10-12 m2 or, with this chlorophyll content, 87 m2 (g chla)-1. The requisite active photon flux is calculable as 700 x 10-6 mol photons per 87 m2, or just 8 |mol photons m-2 s-1. This assumes all wavelengths of visible light are utilised but, if only half were usable, the growth-saturating light intensity would be similar to the level measured for Planktothrix agardhii.
It is not, at first sight, at all obvious that low-light adaptation should be related to algal morphology when it is functionally dependent upon appropriate enhancements in pigmentation. However, it is easily demonstrated that cell geometry and orientation raise the efficiency of light interception by the pigment complement (Kirk, 1975a, b, 1976). A spherical cell, d |m in diameter, has a volume, v = 4/3n (d/2)3, while the area that it projects is that of the equivalent disc, a = n(d/2)2. Because the carbon content is, primarily, a function of volume, the carbon-specific projection (ka) of spherical algae diminishes with increasing diameter. For example, we may calculate that, for a single cell of Chlorella (a = 12.6 x 10-12 m2; C content = 0.61 x 10-12 mol C), ka = 20.7 m2 (mol cell C)-1; for a spherical Microcystis colony (d = 200 | m), comprising 12 000 cells, each containing 14 pg C (Reynolds and Jaworski, 1978), in which a = 31.4 x 10-9 m2 for a content of 14 x 10-9 mol C, ka = 2.24 m2 (mol cell C)-1.
In the case of non-spherical algae that are flattened in one, like those of Pediastrum, or in two planes, like those of Closterium, Synedra or Asterionella, the area projected depends upon orientation. The maximum area projected is when the two longest axes are perpendicular to a unidirectional photon source. The typical cell of Asterionella in a colony lying flat on a microscope slide is ~65 |m in length and shows a tapering valve with an average width of ~3.3 |m. In relation to its approximate carbon content of 85 pg (7.08 x 10-12 mol C; Reynolds, 1984a), the maximum value of ka is ~ 28.2 m2 (mol cell C)-1. In other orientations relative to a single source of light, the area projected may greatly diminish. Kirk's (1976) calculations compensated for this but these are not followed here. In a turbid environment, much of the light available for interception is already scattered and at these low average intensities, changes of orientation prove to be of little consequence. Well-distributed light-harvesting complexes are everything.
These considerations emphasise the influential nature of the relationship between algal shape in the interception of light energy and the impact of algal size in governing its metabolism. Morel and Bricaud (1981) recognised this relationship some years ago, referring to the 'packaging' effect on pigment deployment (cf. Duy-sens, 1956), where the area projected by the pigments assumed the same relevance as the concentration of LHC receptors. It is thus inextricably linked to the contestable size allome-try of growth rate (with its -1/3 slope instead of the expected -1/4; see Section 5.3.1 above and Finkel, 2001). A general relationship between projection and morphometry is shown in Fig. 3.12. The independent variable is, again, the index msv-1, the product of the maximum cell dimension (m) and the surface-to-volume ratio. Note that it is a dimensionless property, length always cancelling out. For spheres, m = d, and msv-1 is a constant [d x 4n(d/2)2 ^ 4/3n(d/2)3 = 6]. For any shape representing distortion from the spherical form, msv-1 > 6. Figure 3.12 also shows that the smaller algae generally project large carbon-specific areas but larger ones have to be significantly subspherical to match the ka values of the smaller ones. It is especially interesting to observe that the algae that already project the greatest area in relation to their cell-carbon content are also mostly those with the maximum photoadaptive potential.
5.4.2 The effect of persistent low light intensities
Rather than experiencing alternations of dark interludes with windows of saturating light, phy-toplankton of so-called crepuscular habitats are exposed to variability that offers only windows of gloom. The algae forming metalimnetic swarms and deep chlorophyll maxima (DCM) in stable layers in seas and lakes experience the same circa-dian alternations of night and day perceived by terrestrial and littoral plants but, because they are located so relatively deep in the light gradient, the daytime irradiances they experience are low.
The circumstances of a cell placed at a constant depth and receiving a dielly fluctuating but low-intensity insolation differ from those of one receiving short bursts of high illumination, even though the daily photon flux might be similar. However, the ultimate objective - to maximise absorption of the photons available - does invoke certain similarities of response. All the organisms that successfully exploit stable layers have to be capable of maintaining vertical position with respect to the light gradient. They are either motile (e.g. the flagellated chrys-ophytes that form layers in oligotrophic, soft-water lakes and certain species of cryptophyte of slightly more enriched lakes), or they regulate buoyancy (as do some of the solitary filamentous Cyanobacteria, including, most familiarly, Planktothrix rubescens and other members of the prolifica group of species, and Planktolyngbya lim-netica). Buoyancy-regulating sulphur bacteria of the Chromatiaceae and Chlorobiaceae may stratify in the oxygen gradient, provided this lies simultaneously within a stable density gradient and is also reached by a few downwelling photons (usually <5 |imol m-2 s-1). Here, photoau-totrophs function on inputs of light energy that are invariably low. The extent of photoadaptation demanded of them depends essentially upon the depth in the light gradient at which the organism is poised. This determines the quantity of penetrating irradiance and the wavelengths of the residual light least absorbed at lesser depths (the quality of the irradiance). Beneath relatively clear layers, with absorption predominantly in the blue wavelengths, the pigmentation may be expected to intensify but without obvious chromatic adaptation. The less is the residual light in the red wavelengths, however, the more advantageous is the facultative production of accessory pigments, such as phycoerythrin and phy-cocyanin, to the organism's ability to function phototrophically at depth.
As discussed in Section 3.3.3, the adaptation is most plainly observed in depth-zonated populations of Cyanobacteria (Reynolds et al., 1983a) or in populations slowly sinking to greater depths
(Ganf et al., 1991). Taking as the most extreme case of photosynthetic efficiency from Post et al. (1985) for P. agardhii as a proven example of the low-light adaptation that is possible at 20 °C (viz. 0.54 mol C (mol cell C)-1 (mol photon)-1 m2; Section 5.4.1) and comparing it with the supposed basal rate of respiration of an alga of its dimensions, 0.079 (s/v)0325, i.e. ~0.064 mol C (mol cell C) d-1, or 0.74 x10-6 mol C (mol cell C)-1 s-1 at the same temperature, it is possible to deduce that compensation is achieved at ~1.4 |imol photons m-2 s-1, certainly in the order of 3-4 |imol photons m-2 s-1, if allowance for the dark period is accommodated.
Applying the results of laboratory experiments to the extrapolation of field conditions or to the interpetation of field data requires caution. In the context of the growth responses of phy-toplankton entrained in mixed layer, insolation may change rapidly, either increasing or decreasing at random (Fig. 3.14). Depending on the light gradient and on the depth of turbulent entrain-ment, phytoplankters might experience anything from a probable period of 30-40 minutes in effective darkness with a few minutes of exposure to high light (deep mixing, steep light gradient), to a similar time period of fluctuating light levels that are nevertheless adequate to support net photosynthesis throughout (mixing within the photic zone). The generic nature of the adaptive responses available, discussed here and in Section 3.3.3, is clearly aimed towards optimising growth against highly erratic drivers. However, the probabilistic, Eulerian aggregation of the responses of the whole population does not take account of the photoprotective and recovery behaviour on the growth rates of individual cells to what are sometimes sharp, sudden and possibly crucial changes in insolation.
There are experimental data that lend perspective to this issue. Litchman (2000) designed experiments that brought irradiance fluctuations to each of four cultured microalgae, in each of three ranges (15-35, 15-85 and 65-135 | mol m-2 s-1) and over three wavelengths of fluctuation (1, 8 and 24 h). Variations in the low-light range (15-35 |imol m-2 s-1, wherein growth rate is expected to be proportional) were fairly neutral. The growth rate of the diatom Nitzschia was slightly increased, that of the green alga Sphaero-cystis schroeteri was slightly depressed, compared to growth at a constant 20 |imol m-2 s-1. In the saturating range (65-135 |imol m-2 s-1), little effect was experienced, except that Anabaena flos-aquae was slightly increased over its performance at a steady 100 |imol m-2 s-1. Over the wide range of fluctuations (15-85 |imol m-2 s-1, spanning limitation to saturation), growth responses differed significantly from the behaviour under steady exposure. Growth of all species was maintained on the short (1-h) cycle; in each case, the relatively short exposure to high light remained within the capacity of the species-specific physiologies. On the longer cycles, however, growth rate was impaired in all species, though not all to the same extent. The reaction would once have been described as 'photoinhibition' but would now be better referred to photoprotection and the first steps toward photoadaptation (see Section 3.3.4). The experiments of Floder et al. (2002) investigated the influence of fluctuating light intensities (range 20-100 |mol m-2 s-1) on growth rates of natural phytoplankton assemblages collected from Biwa-Ko, imposed on cycles of 1, 3, 6 or 12 days. These ably illustrate the population responses consequential upon differential growth rates altering the composition of the assemblage.
A cornerstone role has long been accorded to nutrients in the regulation of productive capacity in the plankton and in shaping the species composition. A very large literature on the nutrient limitation of phytoplankton and the interspecific competition for resources reflects the importance of their availability in pelagic ecology, even if these key processes seem, at times, to have been misrepresented. The point has been made earlier that the least-available resource, relative to the minimum requirements of organisms, sets a finite 'carrying capacity' for the habitat. To establish the limiting role of nutrients on growth rate is more difficult, for two reasons. As has been shown in Chapter 4 and again in Section 5.3.3, most phytoplankters are, under ideal circumstances, able to take up nutrients far faster than they can deploy them. Moreover, they can continue to do so until very low resource concentrations are encountered. Even then, the luxury uptake in generations experiencing resource plenitude may support two or three generations born to resource deficiency.
At first sight, there is certainly a rapid transition from there being no competition for resources to there being little left over which to compete. There is a counter to this deduction, which refers back to the distinctions in the strategies of velocity, storage and affinity adaptation (Section 4.3.2). Naturally, fast-growing algae must be able to garner resources with equal velocity, from diminishing external concentrations. To be able to store resources in excess promises an advantage when external concentrations have been diminished, although it might be more beneficial to species that have the ability to migrate between the relative resource-richness of deep-water layers and resource-depleted but inso-lated surface waters. To survive, even to thrive, in waters chronically deficient in one (rarely more) major resource might well call for a superior competitive ability to win the scarce supplies and deny them to individuals of other species.
Evidence for the existence and implementation of these strategies is to be discussed later in this chapter. For the moment, the first task is to discern the resource levels that separate famine from bounty. It is convenient to consider these nutrient by nutrient.
Satisfaction of the alga-specific P requirements for growth has been suggested to rest upon the ability to maintain a stoichiometric balance of assimilates approximating to 1 P atom to every 106 of carbon. This determines the generalised requirement that 9.4 x 10-3 mol P (mol C)-1 is incorporated during each single replication time. As already indicated, the alga's uptake capacity is likely to be such that, under resource-rich conditions, it may achieve this in minutes rather than hours. It is not likely that any phytoplankter takes longer than its achievable generation time while external MRP concentrations exceed 0.13 x 10-6 M (4 |g P L-1). For many species, indeed, this could be true at MRP concentrations as low as 10-8 M (0.3 |g P L-1; see Sections 4.3.3, 5.3.3). Even then, the reserves accumulated during previous resource-replete generations may sustain one or two generations before the cell recognises impending shortages and perhaps three, or even four, before the cell quotas approach q0 and exhaustion. Fast-growing species, having high sv-1 (>1.3 |m-1, which, at about 30 °C, might permit specific growth rates of up to ~4 d-1, or ~50 x 10-6 s-1, to be attained), would be expected to reach exhaustion proportinately sooner. Yet averaging and normalising the total P requirements to the starting biomass carbon, the P demand of two generations (~0.5 x 10-6 mol P (mol initial cell C)-1 s-1) is still likely to be deliverable from a starting concentration of 0.03-0.13 x 10-6 M (1-4 |g P L-1).
Data compilations that compare growth rates, nutrient-uptake rates and their half-saturation coefficients are available (see Padisak, 2003), from which it ought to be a straightforward exercise to verify, on the basis of hard, experimental results, some of the above conjectures. The trouble is that many of the half-saturation concentrations refer to uptake by starved cells, rather than to that needed to half-saturate the requirement to maintain growth (Kr). These quantities require the analysis of a lot of batch cultures, across a range of concentrations, or the application of a semi-continuous technique in which the algal culture is diluted with test medium at a rate adjusted to keep the algal concentration constant (that is, to balance the rate of growth). As a consequence, there are few good data to confirm the half-saturation constants of P-limited growth. Rhee (1973), using a Scenedesmus, Ahlgren (1985), with Microcystis wesen-bergii, and Spijkerman and Coesel (1996a), using two desmids (Cosmarium abbreviatum and Staurastrum pingue), each found that the algae would grow at about one half the nutrient-saturated rate at 20 °C, in media supplying <6.0 |g P L-1 (0.19 | M). All but the Scenedesmus did so at <1.2 |g P L-1 (<0.04 |M). Cosmarium growth rate was half-saturated at 0.35 |g P L-1 (0.011 |M). Davies (1997), whose work with Asterionella has been highlighted earlier in the context of the interaction between P-uptake and cell quota
(Section 4.3.2), found that growth rate at 20 °C is half-saturated when the cell contains about 0.7 pg P (or about 0.003 mol cell P (mol cell C)-1). The external MRP concentration required to balance this quota is about 0.75 |g P L-1 (Kr = 0.024 | M). The extensive work of Tilman and Kilham (1976) using semi-continuous cultures of diatoms pointed to the half saturation of growth in Asterionella formosa falling between 0.02 and 0.04 | mol P L-1 (0.6-1.2 |g P L-1). However, the corresponding value for another diatom, Cyclotella menegh-iniana, was substantially higher than this (Kr = 0.25 |mol P L-1, or nearly 8 |g P L-1). Finally, in this context, Nalewajko and Lean (1978) used 32P-labelled phosphorus to track phosphorus uptake and turnover in cultures of Anabaena f los-aquae and Scenedesmus quadricauda, and remarked on the 'low' concentrations (~6 |g P L-1; ~0.2 |mol P L-1) at which full exchange activity is maintained.
Certainly, most of these data uphold the view that the onset of phosphorus 'famine', below which growth rate may be regulated by the supply of P, generally occurs at <0.1 |M P (~3 |g P L-1). This may not apply to all species: Scenedesmus and the Cyclotella of Tilman and Kil-ham are possible exceptions. For most species, allusions to growth-rate limitation of phytoplank-ton by phosphorus when external MRP concentrations exceed 0.1 |M P may be doubted. It is possible that the higher species-specific thresholds are a consequence of adaptation to the relatively P-rich habitats in which they occur. Conversely, differential species-specific affinities uptake (Kr values spanning 0.01-0.2 |M P) can be said to influence where species may live and how competitive they might be for truly limiting supplies of phosphorus.
Analogous calculations concerning algae and their nitrogen requirements are also available. Supposing the satisfaction of the alga-specific N requirements for growth to be similarly based upon the stoichiometric balancing against carbon in the atomic ratio 6.6 : 1, the generalised demand approximates to 151 x 10-3 mol N (mol C)-1. Again the DIN-uptake capacity is well up to this under nitrogen-replete conditions. From the information on DIN uptake in Chlorella (Section 5.3.3), it may be deduced that the nitrogen complement needed to sustain a doubling of mass could proceed 60 times more slowly and still fulfil the maximum growth rate. Moreover, the external concentration needed to supply DIN at this rate would be about 0.12 |imol N L-1, or 2 |g N L-1). The half-saturation constants for nitrate and ammonium uptake (KU) among small-celled oceanic phytoplankton are of similar order or only slightly higher (see Section 4.4.2) and are, thus, unlikely to experience symptoms of nitrogen-limited growth at external DIN concentrations >1-2 |g N L-1. On the other hand, the half-saturation constants for nitrate and ammonium uptake (KU) among larger diatoms of inshore waters may be up to an order greater again (0.5-5 |M N), so that problems of obtaining sufficient nitrogen to even half-saturate growth might be experienced at external DIN concentrations in the range 0.2-2 |M N (say 3-30 |g N L-1).
The lower limits of the range of DIN concentrations able to support phytoplankton growth in inland waters are less well researched. It was once deduced (Reynolds, 1972) that, based on the assumption that the appearance of nitrogen-fixing organelles (nostocalean heterocysts) at DIN levels of 20-25 |M provided a simultaneous advantage to the fixers over non-fixers, the nonfixers were experiencing supply difficulties. As a result of later observations but without varying the essential logic, this critical range was revised downwards to ~6-7 |M N (80-100 |g N L-1) (Reynolds, 1986b). As it now seems probable that heterocyst production responds to ammonia concentration and not DIN per se (discussion, section 4.4.3), it no longer follows that their appearance necessarily coincides with nitrogen shortage among the nitrate users. Current evidence indicates that they are competent to draw down DIN to concentrations to levels of 0.3-3 |M without growth-rate limitation.
Eventually, the ability of heterocystous cyanobacteria to fix nitrogen when DIN is simultaneously depleted influences the recruitment of to natural communities (Riddolls, 1985) but the concentration threshold favouring Cyanobacte-rial dominance, between 2 and 6 |M N, remains imprecise. In many of the nitrogen-deficient but simultaneously phosphorus-poor lakes of the Andean-Patagonian lakes investigated by Diaz and Pedrozo (1996), the phytoplankton biomass is demonstrably (by assays and by mathematical regession) constrained by nitrogen availability. The significant incidence of dinitrogen-fixing Cyanobacteria is relatively restricted to lakes, such as Bayley-Willis and Fonck, that have higher TP contents (and where, incidentally, the SRP is drawn down in summer to growth-limiting levels as identied above, viz. to <0.1 |M P). Elsewhere, the unsupplemented combined levels of nitrate-, nitrite- and ammonium- nitrogen levels stand at <0.8-1.0 |M (<11-14 |g N L-1), perhaps falling in summer to <0.2 |M (see Diaz et al., 2000).
The revised verdict on the DIN levels representing the onset of nitrogen limitation among non-fixers is 0.1 |M (1.5 |g N L-1) for oceanic picoplankton and 1-2 |M (15-30 |g N L-1) for many microplankters in lakes and seas. If other conditions are satisfied (see Section 4.4.3), nitrogen fixers may avoid altogether the constraint of low DIN concentrations.
Silicon limitation of growth in diatoms is absolute and well understood. Lund's (1949,1950) classical studies on Asterionella in Windermere and its relationship with the availability of silicon helped to establish the importance of nutrients in planktic ecology. The clarity of the effects and the precision of the chemical thresholds were, and are still, impressive. That they were never emulated in the studies of other elemental requirements is attributable to the nature of the nutrients and the margin around the empiricism of the requirements that the organisms introduce through storage and luxury uptake. As stated (Section 4.6), the diatoms are the biggest planktic consumers of silicon; they take up no more than is required to build the silica valves of the frustules of the current generation, and the silicon polymer is laid down under close genetic supervision. All this leads to the silicon requirement of each new cell of a given species and size being readily predictable. The concentrations of silicic acid capable of delivering the silica requirement have been found through observation and experiment. The consequence of silicon 'limitation' is easy to detect as its consequence is that cells cannot complete the growth cycle. Moreover, failure of the putative cell to build its new frustular valve is fatal.
In the case of Lund's Asterionella, the average complement of 140 pg (SiO2)n cell-1 could be satisfied from a concentration equivalent to >0.5 mg (SiO2) L-1. Translation of the units notwithstanding, both quantities have been abundantly verified. Solving the regressions of Jaworski et al. (1988), an Asterionella cell 65 |m in length has a probable Si content of 65 pg (2.3 pmol cell-1). Using the experimental data of Tilman and Kil-ham (1976), the silicic acid concentration that will half-saturate the Si requirement is 3.9 | M (equivalent to 109 |g Si L-1, or 0.23 mg (SiO2)n
Equivalent data from many other experiments, reviewed in Tilman et al. (1982) and in Sommer (1988a), show that growth rates of freshwater diatoms at 20 °C tend to be half-saturated at between 0.9 (in Stephanodiscus min-utus) and 20 |M Si (in Synedra filiformis). That of Cyclotella meneghiniana is half-saturated at 1.4 |M Si (Tilman and Kilham, 1976). Among the investigated clones of marine Thalassiosira pseudo-nana and T. nordenskioeldii, half-saturation of the growth rate at 20 °C tends to occur in the range 0.2 to 1.5 | M (data of Paasche, 1973a, b).
The 100-fold range in uptake thresholds has ecological consequences, to be considered in the next section. For the moment, the deduction is that silicon concentrations begin to interfere with the growth of diatoms at concentrations below ~0.5 mg Si L-1 (say 1 mg L-1 as equivalent silica, or ~20 |M). In most lacustrine instances, growth-limiting concentrations are encountered mainly below about 0.1 mg Si L-1 (say <4 |M) and, in the sea, below 0.03 mg Si L-1 (<2 |M).
5.4.5 The effect of resource interactions: nutrients and light
The able demonstrations by Tilman and his co-workers of the interspecific differences in the capability of diatoms to take up the silicon and phosphorus required to sustain their growth from relatively low external concentrations also established an important conceptual theory of resource-based competition. Simply, if two species, A and B, having similar resource-saturated growth rates, are cultured together in a gradient of growth-limiting concentrations of a resource S1, the one with the higher uptake rate (VU) is clearly able to sustain a faster rate of growth at low concentrations of [S1] than the other. The theoretical growth performances of A and B are shown in Fig. 5.6a. Against a second resource, S2, however, it is species B that performs better at low concentrations (Fig. 5.6b). Placed together in a medium deficient in both S1 and S2, it is possible, depending on the relative concentrations of either nutrient, for the species to be simultaneously limited by different nutrients. Plotting against the concentrations of both resources (Fig. 5.6c), the respective limitations can be used to predict competitive outcomes of variable resource combinations (Fig. 5.6d). Thus, at low concentrations of (S1), an outcome will always be favoured in which A dominates; at low concentrations of ions of (S2), it will be B that is favoured. At slightly higher concentrations of both resources, A and B may coexist successfully, while the one (B) remains S1-limited and the other is still limited by S2. This prediction was precisely the outcome of their investigations (Tilman and Kilham, 1976) of phosphorus- and silicon-limited growth between Asterionella formosa (Kr(P), ~0.03 |mol P L-1, Kr (Si) 3.9 |mol Si L-1) and Cyclotella meneghiniana (Kr(P), ~0.25 |mol P L-1, Kr (Si) 1.4 |mol Si L-1); Cyclotella dominated over Asterionella in mixed cultures at low Si : P ratios. The opposite was true at high Si : P ratios. On the basis of later investigations of other species, Tilman et al. (1982) emphasised the differential competitive abilities of diatoms to Si : P by plotting the experimentally solved Kr(P) against Kr(Si) for each (see Fig. 5.6e). The plot ably arranges species on the basis of Si: P preferences.
For most other plankton, silicon is a minor nutrient and, not surprisingly, they tolerate Si:P ratios very much lower than Cyclotella's limit of 5.6 (Holm and Armstrong, 1981; Sommer, 1989). However, it is the relationships between the nitrogen and phosphorus requirements that have aroused enormous interest, especially in the context of a widely held belief that low
Resource competition and species interactions. Parts (a) and (b) compare the nutrient-limited growth rates two species of phytoplankter, Sp. A and Sp. B, against low, steady-state concentrations of resources S| and S2. Growth of either may be limited (c) by the availability of either resource. Tilman's theory of resource-based interspecific competition acknowledges that the uptake constraints acting on Sp. A and Sp. B differ sufficiently for (d) Sp. A to dominate over Sp. B when [S| ] is low and Sp. B to do so when [S| ] is low, but Spp. A and B do not compete when limited by different resources. The relative competitive abilities of named diatoms for silicate and phosphate, as determined by Tilman et al. (1982), are shown in (e): A.f., Asterionella formosa; C.m., Cyclotella meneghiniana; D.e., Diatoma elongatum; F.c., Fragilaria crotonensis; S.f., Synedra filiformis; S.m., Stephanodiscus minutulus; T.f., Tabellaria flocculosa. In (f), the effects on phytoplankton assemblages in a selection of natural lakes of differing N and P availabilities are represented: C, Crose Mere, and W; Windermere, are in UK; E is Esrum, Denmark; K, Kasumiga-Ura, and S, Sagami-Ko, Japan; Me, Mendota, T, Tahoe, and Wa, Washington, in USA; Mg is Maggiore, Italy/Swizerland and Ml, Malaren, in Sweden. Area 1 applies to low-P lakes, dominated by diatoms and chrysophytes; area 2 covers lakes in which nitrogen-fixing Cyanobacteria are abundant through substantial parts of the year; area 3 lakes are are dominated by Microcystis for long periods. The composite combines figures redrawn from Reynolds (1984a) and Reynolds (1987b).
nitrogen-to-phosphorus ratios favour the (usually unwelcome) dominance of Cyanobacteria (Smith, 1983). Rhee (1978) suggested that the evident mutual competition along N : P gradients is influenced by differing N : P optima in the cells of various species. In a major programme of laboratory experimentation, Rhee and Gotham (1980) showed systematic differences in the ratios of species-specific optimal N and P quotas dur ing normal growth (when both storage effects and deficiencies should have been minimal). Outcomes ranged between 7, for the diatom Stephanodiscus binderanus, and 20-30, for three species of Chlorococcales. For the Cyanobacterium Microcys-tis aeruginosa, it was 9. These optima differ from the ideal ('Redfield') N : P stoichiometry centred at 16 molecular and from the suggested (Section 4.4) range of normality of 13-19. Given the range in physiological variability in N : C and P : C content of individual cells, extremes of N : P of 4 and 108 are theoretically tenable but values within the range 10 to 30 are scarcely indicative of nutrient stress.
Nevertheless, the findings of Rhee and Gotham (1980) do not violate any supposition about the importance of N : P ratios in favouring Cyanobacteria or otherwise. However, there are difficulties (see below) in applying resource ratios to either the interpretation or the prediction of the composition of natural communities. Besides, ratios of available resources change quite rapidly through time, without necessarily precipitating immediate changes in species composition. If the total nitrogen and phosphorus resources delivered to and present within a water body are considered, some broad compositional trends are discernible. The distribution of lakes shown in Fig. 5.6f against axes of total N and total P separates those that will support large populations of nitrogen-fixing Cyanobacteria as being low-N-high-P habitats, from those that support non-fixing Microcystis (high N, high P) and those that seem never to be dominated by bloom-forming genera (low-P lakes).
To be any more specific about predictions of differentiated growth or the composition of the population structures that it might yield needs much more information. One component obviously lacking from the previous paragraph is the carbon input and the solar-energy income that is essential to its compounding with nutrient resources into biomass. In developing a hypothesis about the deterministic importance of the light : nutrient ratio in lakes, Sterner et al. (1997) showed that the relationship between the mean mixed-layer light level (equivalent to the calculation of I* in Section 3.3.3) in each of a number of lakes and its corresponding total P concentration is a good probabilistic predictor of the C : P content of the seston and of the efficiency of resource use by the system as a whole. They further hypothesised that the seston C : P ratio influenced the pathways of secondary production (essentially the food-web consumers of primary product) and thence, biassed the intensity of nutrient recycling, the strength of micro-bial processing and, indeed, the structuring of the entire ecosystem.
These plausible deductions invoked parallel developments in the appreciation of organismic stoichiometry, prompted, in part, by the work of Hessen and Lyche (1991) on the differential elemental make up (chiefly C:N:P) of zooplankton. Stoichiometric differences between the trophic components imply consequences for the system as a whole. For instance, animals with a relatively low N : P content feeding on algal foods having a high N:P make-up will retain P preferentially and, so, recycle wastes with a yet higher N:P content. This approach has been developed further by J. Elser and co-workers (usefully reviewed in Elser et al., 2000). They were able to demonstrate striking connectivities among the molecular sto-ichiometries of growing cells of a wide range of algal species, their rates of growth and the evolutionary pressures underpinning their life-history traits. That is to say, the evolution of fast or slow growth rates and the allocation of the catalysing (P-rich) RNA molecules are inextricably interlinked. The matching of evolutionary traits, from their molecular bases to the environments in which whole organisms function and interact, now has a wide following. The recent book by Sterner and Elser (2002) on Ecological Stoichiometry conveys and nourishes much of this excitement. Besides providing an alternative perspective on ecosystem function and a persuasive argument for unifying concepts in its understanding, the book invokes pertinent explorations of just what goes on inside the living cell to yield recognisable structural stoichiometries in the first place and what activities lead to (relatively modest) departures therefrom.
To bring us back specifically to phytoplank-ton, if these attractive theories of resource ratios and biological stoichiometry are to be helpful to understanding how pelagic communities function, then it is important first to separate just what is cause and what is effect. It is necessary to emphasise the distinctions among the ratios of algal cell quotas, the ratios of resource availability and supply, and the competitive abilities of algae to take up elements at low concentrations. Taking cell quotas first, we have to accept that the ratio N : P = 16 is certainly not absolute, that it is subject to a margin of physiological variability, including in the complement of RNA, and that there may indeed be systematic, inter-phyletic differences in the optimal elemental balances. Nevertheless, the range of normality in the elemental composition of cells, from bacteria to elephants, is relatively quite narrow, reflecting general similarities in the cell complements of protein and nucleic acid (Geider and La Roche, 2000). From Chapter 4, it is plain that a factor of 50% variation in either the N or the P content is hardly exceptional, yet it yields a full range in N:P from 5 to 36. For the purpose of estimating stoi-chiometrically the phytoplankton-carrying capacity of the nutrients in a given habitat, less error attaches to the adoption of a mean complement of 16 to 1 than to the correct estimation of the base of bioavailable nutrients (Reynolds and Maberly, 2002). However, it is more straightforward (and more illuminating) to examine the support resource by resource. To be able to form a standing phytoplankton biomass equivalent to, say, 106 |mol C L-1 (1.27 mg C L-1, ~25 |g chl a L-1) ostensibly requires the supply of 16 |mol N and 1 |mol P L-1 (i.e. 224 |g N, 31 |g P L-1). If the water body can fulfil only 10 | mol N and 0.1 |mol P L-1 (note, N : P = 100), it is obvious that the growth demand will first exhaust the phosphorus, at a rather smaller chlorophyll yield than 25 |g chl a L-1. Long before this maximum is reached, the biomass : P yield is stretched, according to Eq. (4.15), to 12.2 |g chla L-1; the biomass has a probable content of carbon equivalent to ~51 |mol C and 7.7 |mol N L-1, but no more than the originally available 0.1 |mol P L-1. We deduce a biomass N : P ratio of 77, correctly inferring the obviously severe phosphorus limitation on the biomass. The magnitude of the eventual P-limited quota constraint on the biomass is predictable on the basis of the initial phosphorus availability. That it was phosphorus, rather than nitrogen, that would impose the eventual limit is implicit in the starting resource ratio.
Now, if the water body could fulfil 1 | mol P but only 1.6 |mol N L-1 (i.e., 31 |g P, 22.4 |g N L-1; N:P 1.6), growth would be expected to exhaust the nitrogen, for the production of 32 | mol biomass C L-1 (assuming the minimum C : N ratio quoted in Section 4.4) and which, stoi-chiometrically, would have a phosphorus content equivalent to not less than 0.3 | mol P L-1
and a probable biomass N : P (5.3) indicative of N-limitation. The residual P concentration (arguably ~0.7 |mol P L-1) confirms that P is certainly not a constraint and it is also sufficient to support the activity of nitrogen-fixing Cyanobacteria. These could grow to the limit of the phosphorus capacity (47 |g chla L-1), only to now be dominated by nitrogen-fixing Cyanobacteria having an intracellular N : P ratio of ~30. Note that, provided it is not itself limited by some other factor, nitrogen fixation forces the nitrogen-deficient system to the capacity of its phosphorus supply (cf. Schindler, 1977).
Now let us consider the role of interspecific competition in the way different species might simultaneously satisfy their resource requirements to sustain their growth rates. Here we bring into sharp focus the uptake capabilities of the algae themselves and the sorts of threshold concentrations at which they fail to be able to take up specific nutrients as fast as maximum consumption would demand (1-2 | mol N, 0.1 | mol P L-1 (Section 5.4.4). The corollary of this is that if the concentrations of bioavailable N and P are significantly greater than these thresholds, neither imposes a rate-limiting constraint upon the growth of any alga. Several species could grow simultaneously and, while each satisfies its requirement, they are not in mutual direct competition (sensu Keddy, 2001; see Box 4.1). Each performs to its capacity (or to some independent regulation). At this stage, the ratio of the available resources is quite irrelevant to the regulation of species-specific growth rates.
This deduction arouses persistent controversy. Yet it is readily verifiable in the laboratory through the measurement of the early exponential increase of a test alga in prepared media offering nutrients in differing mutual ratios but at initially saturating concentrations. I am unaware of any publication that draws attention to this behaviour. However, I am most grateful to Dr Catherine Legrand, of the university of Kalmar, Sweden, for her permission to reproduce a graph that she presented at the 1999 Meeting of the American Society of Limnology and
Was this article helpful?