Logo: to the web site of Uppsala University

This study is the first to examine the brain gonadotropin-releasing hormone (GnRH) cell population phenotype in a protogynous and monandric sequentially hermaphroditic fish. Male ballan wrasse (Labrus berggylta) had on average higher numbers of GnRH-immunoreactive (GnRH-ir) cells within the brain preoptic area (POA) than females, a difference not found in GnRH-ir cells in other brain regions. Furthermore, in males, but not females, the number of these POA GnRH-ir cells correlated with body size. Maturational state (prespawning or postspawning) had marked effects on mean profile sizes (but not numbers) of both GnRH-ir cell bodies and cell nuclei, even when existing differences in body size and allometric relationships had been taken into account. Postspawning males tended to have larger GnRH-ir profiles in all brain regions relative to both prespawning males and females. Moreover, the GnRH-ir cell number in POA, and the cell body profile size in both POA and at the level of the anterior commissure, correlated with gonad size in spermiated prespawning males, indicating a relationship between both size and number of GnRH cells and male gonadal development. These results suggest that temporary changes in the size of brain GnRH-ir neurones are coupled to the male spawning cycle, and that permanent POA GnRH-ir cell number changes are involved in the process of sex change in sequential hermaphrodites. However, smaller males had no more preoptic GnRH-ir cells than equally sized females, which may argue against a proximate inducing role of GnRH cell number changes in naturally occurring sex reversal.

Phytoplankton and bacterioplankton dynamics in lakes are regulated by a multitude of factors such as light, temperature, water column mixing, nutrients, organic substrates, grazing, and food web structure. Tropical lakes are especially interesting to study because of the relative constancy of several of these factors. This thesis deals with interactions between phytoplankton, bacterioplankton and environmental factors in tropical, polymictic Lake Xolotlán, Nicaragua.

Phytoplankton biomass per unit area showed only small temporal and spatial variations, with an average of 65 g C m^-2. The photic zone:mixing depth ratio was relatively constant at 0.19, which is a critical lower limit for algal survival. Frequent mixing of the water column caused the algae to spend much time in darkness, thereby loosing by respiration what was gained by photosynthesis during short periods in the photic zone, and this caused high algal mortality.

Areal primary production was limited by the biomass of actively photosynthesizing algal cells in the photic zone and varied within the range 6.0-7.5 g C m^-2 d^-l. The depth of the photic zone was mainly regulated by light extinction by active algae, but also by dead algae. The rates of light saturated and light limited photosynthesis were close to theoretical maximum values.

Bacterial production (1 g C m^-2 d^-1) was comparable to known empirical models, but the bacterial biomass (5 g C m^-2) was much greater than predicted. This resulted in a very low production per biomass. Bacterial biomass and production were significantly correlated to algal degradation products suggesting that bacteria were regulated by algal mortality. Allochtonous organic carbon was an insignificant source of substrate for the bacteria.

This study shows that bacteria were sinks for organic carbon and regenerators of nutrients. By rapid decomposition of dead algae they made nutrients available for new algal growth, thus maintaining a high biomass specific primary production. Both algal and bacterial mortality by grazing was very low. The system was thus characterized by a tight coupling between algae and bacteria with little transfer of carbon and energy to higher trophic levels.

Cyanobacteria are present in most aquatic systems, and pelagic bloom-forming species often flourish in lakes. Since blooms of cyanobacteria are undesirable for several reasons, e.g. they can result in foul odours and tastes as well as toxicity, these organisms have been subject to an extensive amount of research around the globe. A number of hypotheses have been put forward to explain the strong competitive ability of cyanobacteria in phytoplankton communities. In this thesis results are presented from field and laboratory experiments on the influence of inorganic nitrogen and iron nutrition on cyanobacterial population dynamics. The results are discussed in the light of previously published theories.

It was hypothesised that cyanobacteria are favoured over eukaryotic phytoplankton when inorganic nitrogen is present in the form of ammonium rather than nitrate, and vice versa. This hypothesis was supported by three field-enclosure experiments performed in the moderately eutrophic Lake Erken in which it was found that cyanobacteria responded positively to additions of ammonium and phosphate, but did not respond to additions of nitrate and phosphate.

Laboratory competition experiments with mixed cultures of a cyanobacterium and a green alga showed that the former was favoured by highly frequent, small pulses of inorganic nitrogen either as ammonium or nitrate, whereas the green alga was favoured by larger pulses applied less frequently. Based on these results and the fact that, regardless of the inorganic nitrogen source, it was largely eukaryotic phytoplankton that responded to the nutrient additions in the field-enclosure experiments in Lake Erken, the hypothesis regarding inorganic nitrogen and cyanobacterial success was reformulated. It is suggested that in water bodies in which eukaryotic phytoplankton are disfavoured owing to nitrogen limitation, cyanobacteria are likely to be strong competitors unless limited by other elements.

The importance of iron availability as a factor limiting cyanobacterial growth was studied in three field-enclosure experiment in Lake Erken. Despite additions of phosphate and inorganic nitrogen every second day, cyanobacterial biomass decreased. However, in cases where dissolved iron was also supplied, extensive outgrowths of cyanobacteria occurred. This result has implications for the management of wastewater treatment plants in which iron is used to precipitate phosphate.

Neural responses to anoxia were studied in two fish species, the rainbow trout(Oncorhynchus mykiss) and the crucian carp (Carassius carassius). The rainbow troutbrain reacted to anoxia by a massive release of excitatory amino acids and K⁺ into theextracellular compartment, in a fashion very much like the brain of mammals duringanoxia. By contrast, the crucian carp is extremely tolerant to anoxia, which is probablyrelated to an ability to down-regulate the ATP use in order to avoid a fall in ATP levelsand save on the glycogen stores. This study shows that the crucian carp lowers itsspontaneous locomotor activity by ~50% in response to anoxia, which corresponds to anestimated fall in body energy consumption by 35-40%. Furthermore, in anoxia the heatproduction of isolated telencephalic slices was found to be reduced by 37%,corresponding to a 31% reduction in ATP turnover rate, although energy charge wasmaintained. The anoxic crucian carp did not down regulate the rate of tram-membraneK⁺ flux in brain ("channel arrest"), a phenomenon that thought to be an energy savingstrategy in the anoxia-tolerant fresh water turtle Trachemys scripta. In normoxia thecrucian carp maintained ion-homeostasis even though the glycolysis was inhibited withiodoacetate, whilst in anoxia, a functioning glycolysis was found to be totally essential.The opening of K_ATP-channels in mammalian neurons has been suggested to prolong thesurvival time of energetically depressed neurons during short term energy deficiency bycounteracting depolarisation and glutamate release. However, I was unable to find anyevidence for such channels in the brains of crucian carp and rainbow trout. Finally, it wasfound that in anoxia, the crucian carp retina suppresses the electric response to visualstimulation by ~80%, and that there is a corresponding decrease in evoked potential in theoptic tectum of the same magnitude. These changes were reversible and are likely toreflect a strategy for reducing ATP consumption during anoxia.

Bacterioplankton are among the smallest and most abundant organisms in lake ecosystems. Due to the lack of relevant methods, the taxonomic composition of bacterioplankton has been studied less thoroughly than that of larger organisms. The work described in this thesis has focused on analyzing temporal and spatial variations in bacterioplankton community composition, especially in relation to environmentalparameters. The community composition was determined by denaturing gradient gel electrophoresis (DGGE) of PCR-amplified (polymerase chain reaction) 16S rDNA (16S rRNA encoding DNA).

Temporal changes within a bacterioplankton community in a mesotrophic brownwater lake were observed over a two-year period. These changes appeared to occur slowly and gradually, without any apparent coupling to season. The inflow of water from the drainage area was one possible factor causing the observed changes.

The bacterioplankton community composition at different depths in two stratified lakes was investigated in a combined field survey and field experiment. The results suggest that differences between strata in the amount of dissolved oxygen, the availability of phosphorus, and temperature were important factors contributing to differences in community composition.

Differences in community composition were also detected between five lakes differing in trophic status and humic content. The results indicated that the amount of potential grazers (mixotrophic flagellates and microzooplankton) present in the lakes had an impact on bacterioplankton community structure. The abundance of these organisms appeared to be coupled to the trophic status of the systems, indicating that trophic status at least indirectly influenced the composition of the bacterioplankton communities.

Differences in bacterioplankton community composition were also found in five lakes of similar trophic status but different humic content. In these lakes grazers again appeared to be important, as did the size of both the lakes and their drainage areas. The coupling to humic content was less evident.

In summary, many physical, chemical, and biological factors within the pelagic environments as well as in surrounding ecosystems appear to influence the taxonomic structure of bacterioplankton communities, which was found to vary in time and space.

This thesis deals with how reproductive traits and reproductive behaviour in femalebrown trout (Salmo trutta) are linked to offspring performance.

In the small stream Tullviksbäcken, the choice of spawning sites was stronglyinfluenced by short-term fluctuations in stream flow and by substrate quality. Coarsegravel gives high egg and alevin survival and was preferred, but substrate selectivityvaried among years. Stream accessibility in the autumn, as determined by fluctuations instream flow and sea level, affected both the number of spawners and the timing ofspawning.

In Tullviksbäcken, the distribution of stocked trout two months after emergence wasbest explained by a mathematical model that considered fry to consist of one componentwith resident behaviour and one component with dispersive behaviour. This is opposedto the general view that fry dispersal behaviour is uniform within populations. Dispersionfrom the redds was spatially limited and this may be an adaptation to poor survival farfrom the redds. Initially small sibling trout appeared to be more dispersive than wereinitially large siblings. Egg size had no general effect on fry performance in localaggregations, and trout survival was not affected by differences in egg size. Thus, theprevailing 'bigger-is-better' paradigm concerning larval fish performance wascontradicted. However, initially large trout appeared to be favoured in stream sectionswith high biomass of underyearling trout and good growth opportunities, and it issuggested that selection for initial size may occur locally.

Resident trout in the stream Jörlandaån had relatively larger eggs than both sympatric migratory trout and trout from five other Swedish streams. The fitness in residentJörlandaån females is probably enhanced by the production of extremely large offspring, possibly through increased fry survival relative the more numerous fry descending from larger migratory parents.

In conclusion, redd-site selection is important for egg and alevin survival but also forjuvenile performance, because fry will not disperse far from their original redd andselection pressures may vary locally. Female egg production strategies may also influenceoffspring survival. Thus, the presented results demonstrate a strong coupling betweenfemale reproductive pattern and offspring performance in brown trout.

Phytoplankton have adapted to survive in the variable and heterogeneous environment that is characteristic of lakes in temperate regions. One strategy for survival during periods unfavorable for growth is to form resting propagules. Dinoflagellates are bloom-forming phytoplankton which have the ability to form resistant and long-lived resting cysts. The role of cysts in the ecology and life history of dinoflagellates was investigated in this thesis. First, it was tested whether cysts have the ability to assimilate phosphorus (P) during dormancy, which could improve the survival of newly germinated cells when returning to the water column. Secondly, the role of resting cysts in regulating the seasonal succession of dinoflagellates in Lake Erken was examined. Also. biotic and abiotic factorscontrolling the timing of germination were tested.

Cysts of the marine dinoflagellate Scrippsiella trochoidea which were incubated in P enriched medium attained a P content twice that of cysts incubated in P deplete medium. Cysts of the freshwater dinoflagellate Ceratium hirundinella had only a slightly higher P content when incubated in P rich medium. Some uptake may be taking place (although adsorption to the cyst wall cannot be excluded), but not enough to provide newly excysted cells with a competitive advantage over other phytoplankton. Phosphorus may, however, play a role in cyst survival and subsequent germination processes.

Nine species of dinoflagellates were found in Lake Erken, of which one new thecate phototrophic species, Peridinium euryceps Rengefors et B. Meyer was encountered. This is a cryophilic-and a cyst-forming species, with a distinct and extremely flattened shape.

All phototrophic dinoflagellates in Lake Erken rested as cysts during the major part of the year. Timing of germination was mainly regulated by an internal maturation period, a biological clock, and temperature. Furthermore, C. hirundinella cysts did not germinate when incubated in anoxic conditions. Cysts of Peridinium aciculiferum, a winter species, did not germinate in the presence of exudate from zooplankton (predators). These results suggest that cyst dormancy may also be a predator avoidance strategy for some species.

To summarize, the results of the studies presented in this thesis show that resting cysts play a significant role in the seasonal succession of dinoflagellates. Also, the cyst stage appears to be a survival strategy for escaping herbivory. Furthermore the results suggest that nutrient uptake may occur during the cyst stage, which in turn can affect the survival of the cyst and the newly germinated cell.

Phosphorus (P), a key element in lake primary production and eutrophication, is accumulating in agricultural soil, urban landfills and lake sediments. This accumulation increases the risk for non-point losses of P to ground- and surface waters.

Forms and amounts of potentially mobile P were determined in digested sewage sludge, soil, and lake sediment in order to predict maximum future non-point P-losses from these P-accumulating systems. To determine concentrations of potentially mobile P a new method was developed in which water differing in oxygen status is passed through samples enclosed in flow-through chambers. Released P was labeled potentially mobile P, and under oxic and anoxic conditions accounted for 20-30% and 95%, respectively, of the total P cement in digested sewage sludge after passage of water corresponding to 10 years of rain. P released from soil mixed with sludge accounted for 30-40% of the total P content.

In the Lake Erken sediments, there was 4.8 g/m² of potentially mobile P, which was made up of 3.7 g/m² NaOH-nrP (organic-P) and 1.1 g/m² BD-P (Fe-P). Organic-P mobilized under anoxic conditions in deeper sediment layers, probably representing bacterial-associated P, was adsorbed by iron (Fe) in the oxic surface sediments. Fe-P then acted as the main P-source to the water column during spring mixing.

Analysis of sediments from six Washington lakes previously treated with aluminium (A1) showed that added A1 is buried in the sediment together with adsorbed P at a ratio of 11:1 (by weight).

Sludge landfills and fertilized soils contains large amounts of potentially mobile P compared with lake sediments. Non-point P-losses from terrestrial systems thus have the potential to cause the eutrophication of lakes and coastal areas. Based on the results obtained in this work, its possible to estimate maximum future P losses from certain types of P-accumulating systems. This knowledge may be crucial in the development of cost effective measures for reducing P losses to surface waters. The recycling of P within terrestrial systems can help substantially to both reduce P-losses to aquatic systems and preserve P for generations to come.

Mercury and radioactive caesium have received immense attention in Sweden during the past decades, as common contaminants in fish. This thesis examines how the Hg and ¹³⁷Cs levels in fish from 79 circumneutral lakes are affected by various environmental factors. In addition, the observed contaminant levels in the fish are compared to observations from more acidiclakes.

Partial least square regression models (PLS) were used to assess the influence of 46-53 environmental predictors on Hg and ¹³⁷Cs levels in perch (Perca fluviatilis L.) and roach (Rutilus rutilus L.). Linear regression intercepts of contaminant concentration vs. fish size were used as estimates of lake specific contaminant level in fish, in the cases when the contaminant content had a clear size-dependency.

Differences in land use in the catchment areas were found to be a good predictor of both Hg and ¹³⁷Cs levels in the fish. Lowest levels were found in lakes draining arable land. High Hg levels in both species were predominantly found in forest lakes, whereas high ¹³⁷Cs levels were observed in wetland lakes.

The Hg and ¹³⁷Cs levels in the fish were negatively correlated to the cation concentration in lake water, but had a positive relationship to the amount of humic matter, i.e. low contaminant levels in hard-water lakes and high levels in brown-water lakes. No effect of lake pH was found on Hg or ¹³⁷Cs levels in either species. The Hg contents in perch and roach were of the same magnitude as in fish from more acidic lakes. The radiocaesium deposition in the close vicinity of lakes had a more profound effect on the ¹³⁷Cs levels in fish than more distantdeposition.

Bacterioplankton growth is affected by physical, chemical, and biological factors. Theavailability of energy and carbon sources, as well as inorganic nutrients, are of specialinterest when studying regulation of heterotrophic bacterioplankton in lakes.Bacterioplankton was studied in two non-humic lakes, oligotrophic Lake Njupfatet andeutrophic Lake Erken. In both lakes, experiments were conducted to test the effects ofinorganic nutrients, organic carbon, and metazoan zooplankton on heterotrophicbacterioplankton growth.

Generally, bacterioplankton production was primarily limited by inorganic nutrients inboth lakes. However, there was often a close balance between limitation by inorganicnutrients and organic carbon. Organic carbon alone or in combination with phosphorusbecame limiting during early summer and during the autumn circulation period in LakeErken. In Lake Njupfatet, organic carbon was in short supply during late summer. In bothlakes, metazoan zooplankton stimulated bacterial production, suggesting that processeswithin the water column were important in providing bacteria with inorganic nutrients andorganic compounds. In Lake Njupfatet, bacterial production seemed tightly coupled to theactivity of other planktonic biota, such as grazing and phytoplankton growth. In LakeErken, no evident coupling was found between bacterioplankton and phytoplankton.Despite high production, bacterial biomass and abundance usually remained low in bathlakes, indicating that grazing pressure on bacteria was high. The effect of different forms ofinorganic nitrogen on bacterial growth was studied in Lake Erken. Heterotrophicbacterioplankton production responded more strongly to additions of ammonium thannitrate, suggesting that the inorganic form of nitrogen is important to take intoconsideration when studying regulation of bacterial growth.

Altogether the studies show that even in lakes where the input of allochthonous organicmatter is low, as in Lake Njupfatet and Lake Erken, inorganic nutrients and not organic carbon is primarily limiting bacterioplankton growth. This indicates that bacterialproduction in freshwaters may generally be limited by inorganic nutrients, since manylakes have higher concentrations of organic carbon than those investigated in this thesis.

Herbivorous crustacean zooplankton link phytoplankton and other microorganisms withhigher trophic levels in pelagic food webs. Crustaceans show little intraspecificvariation in somatic C:N:P elemental ratios, which may have implications for theregulation of zooplankton growth and for nutrient cycling.

Nucleic acids were found to constitute the largest pool of P in Daphnia andEudiaptomus, This proportion of the total P pool decreased with increasing size anddecreasing reproductive rate. In daphnids, 14% of the P was allocated to the carapace,which may result in a significant loss of P from the epilimnion during periods of highDaphnia abundance.

Algal P content was an important factor determining the body growth rate ofDaphnif, but at low algal C:P ratios, the addition of polyunsaturated fatty acids (PUFA)to the diet resulted in earlier maturation. It is concluded that the two major hypotheses for explaining nutrient-limited growth, i.e. that growth is limited mainly by theavailability of P and PUFA, respectively, are not mutually exclusive.

Effects of grazers on the nutrient status and growth of phytoplankton and bacterio-plankton were assessed in laboratory and field experiments. Grazing by Daphniaresulted in changes in the stoichiometry of phytoplankton and bacterioplankton.Grazing by Eudiaptomus stimulated bacterial growth. Similar responses havepreviously been explained as trophic cascades via bacterivorous protozoa, but this studyshows that the response may also be explained by zooplankton nutrient regeneration. Ina field study, bacterioplankton production was frequently limited by P at the same timeas phytoplankton were N-limited. During early summer, when bacteria were co-limitedby P and C, zooplankton stimulated bacterioplankton production.

This thesis shows that zooplankton are affected by the stoichiometry of their food,and that zooplankton affect the nutrient status and growth of their food organisms. Thefeedback mechanisms involved are not yet fully understood, and their importance innatural systems has not been extensively explored. In future food-web research, weneed to recognise differences between species in terms of their roles in nutrient cycling.Approaches aimed at improving the understanding of pelagic food webs based onecological stoichiometry appear to hold promise for such work.