Oxygen availability in aquatic habitats is a major ecological factor influencing the distribution of fishes. The capacity of fish to survive, and even thrive, under conditions of low oxygen varies markedly among and within species, with more tolerant individuals being able to exploit low oxygen habitats. Fishes of East Africa provide outstanding examples of inter- and intraspecific variation in hypoxia tolerance. Previous research on B. neumayeri has shown that populations inhabiting low oxygen swamps differ in respiratory behaviour , gill morphology [27–29], and tissue metabolic capacity , all of which presumably enhance the hypoxia tolerance of individuals from these swamp populations. Herein, we evaluate the influence of source population and acclimatization to low oxygen in determining growth and metabolic potential of B. neumayeri from habitats differing in oxygen availability.
Among the various morphological and enzymatic measurements reported, only liver LDH activity was significantly related to site of origin, with individuals collected from hypoxic habitats displaying greater activities. This observation is consistent with our earlier observation that B. neumayeri from Rwembaita Swamp had higher liver LDH activity than fish from the normoxic Njuguta River . In the current study, the effect of site of origin was observed after four weeks of field acclimatization to different oxygen habitats, whereas in the previous study, the effect of collection site was measured immediately after collection as well as after long term laboratory acclimation to normoxia. Although LDH plays a critical role in anaerobic metabolism, it is also involved in the process of gluconeogenesis by catalyzing the conversion of lactate into pyruvate, which can then be used for glucose production. Consequently, the role of higher LDH activity in livers from fish collected in the hypoxic Rwembaita Swamp may be in the clearance of blood lactate and the provision of glucose for metabolism by extra-hepatic tissues such as heart and brain, which are important in the maintenance of an organism's homeostasis [30–32].
The observation of differences between the Rwembaita Swamp and the inlet stream population populations of B. neumayeri in liver LDH activities after field acclimatization (this study) and laboratory acclimation , suggest a potential genetic component. This is supported by recent results using nuclear markers (CK and TPI-A) that show significant genetic divergence between Rwembaita Swamp and Inlet Stream West populations of B. neumayeri (Harniman unpublished data). Although we found strong evidence for plastic effects in some enzymatic traits, local adaptation to divergent oxygen regimes may also contribute to the observed trait variation as has been observed in other studies of fishes experiencing divergent selective environments [33–35].
One theoretical expectation arising from local adaptation is that organisms adapted to their native environment should perform "better" than organisms transplanted from their native environment and acclimatized to a novel habitat. This expectation would be experimentally supported in our study by a significant origin by acclimatization interaction with respect to potential fitness correlates such as growth. In this study, growth rates were negative in all groups, regardless of site of origin or acclimatization treatment, ranging from - 0.1% to - 0.15% body mass per day. It can be argued that the negative growth observed in this study was due to the artificial conditions of the experimental enclosures (e.g., increasing stress, reducing food availability); however, for B. neumayeri population within this system, it is quite natural for fish to become trapped in small pools and experience mass loss during the dry season [1, 27]. In a mark and recapture study of the foraging ecology of B. neumayeri during the summer season, Schaack and Chapman  reported negative growth rates similar to those reported here.
Interestingly, mass loss was greater, on average, for fish acclimatized to the hypoxic Rwembaita Swamp compared to those in the normoxic Inlet Stream West, irrespective of site of origin. The restrictive effect of low oxygen on growth is consistent with numerous laboratory and field studies demonstrating reduced growth in fish under hypoxia . Low growth under hypoxia in other species has been attributed to reduced rates of ingestion or assimilation efficiency or both [16, 39, 40]. Previous work on B. neumayeri demonstrated a trade-off between respiratory and trophic morphology, such that relatively larger gilled swamp fish spent more time handling food items than smaller gilled stream fish . In the current experiment, this trade-off did not result in lower growth in fish originating from hypoxic swamps when compared to fish from the normoxic stream in either acclimatization treatment. This result indicates that under these experimental conditions, this putative trade-off was without adverse effects on growth. It is possible that other, unmeasured fitness tradeoffs contribute to the maintenance of differences between these populations in LDH, haematocrit, and morphological traits such as gill size .
Acclimatization to differing oxygen levels also affected tissue levels of metabolic enzymes. In particular, PFK was greater in heart and skeletal muscle from individuals acclimatized to the low oxygen swamp site, while CCO was lower in heart after hypoxic acclimatization. Phosphofructokinase is an important regulatory enzyme of glycolysis, and changes in PFK activity are likely to impact the flux through the glycolytic pathway. Higher PKF activities after hypoxic acclimatization, therefore, could represent an increased capacity for glycolytic ATP production. Cytochrome c oxidase, on the other hand, catalyzes the final step in mitochondrial electron transport to oxygen, and its rate is important in determining rates of aerobic ATP production. The significant effect of acclimatization on heart CCO was mirrored by a non-significant trend toward lower muscle citrate synthase (P = 0.08), a mitochondrial enzyme catalyzing the first step of the Kreb's cycle. The overall effect of acclimatization to low oxygen, therefore, was an increase in glycolytic capacity (heart and muscle PFK) and a trend towards decreased aerobic capacity (heart CCO and muscle CS). Results from laboratory acclimation of fish to hypoxia have yielded similar, if sometimes mixed results [reviewed by 2]. Nevertheless, the current data are an important field validation of predictions based upon energetic considerations and laboratory studies.
The present study employed a reciprocal transplant experimental design that is not commonly used in physiological studies. There are certainly challenges to this field experiment, for example the escape of fish from one replicate cage and the sporadic loss of individuals from other cages, likely due to predators that could enter the cages through the open tops (e.g., birds, snakes, and spiders). Another potential limitation of the study is that it was carried out at specific locations of known DO level, and to extrapolate our results more broadly it would be necessary to replicate this study in other systems having divergent oxygen concentrations. In addition, it is possible that the cage restriction itself altered the response of the fish to the acclimatization treatment. However, free swimming B. neumayeri collected at this and other Kibale field sites had negative growth rates , condition factor , and tissue LDH activities  similar to those measured in fish after the 4-week acclimatization treatment. Despite these concerns, our results suggest that reciprocal transplant experiments provide a powerful tool in uncovering physiological responses to naturally occurring ecological stresses.