Summary of the thesis: Swimming performance and morphology of pregnant fish

A live-bearing mode of reproduction implies morphological and physiological constraints, which in fish might be influenced by the physical characteristics of water: a higher density which favours hydrodynamic body shapes and requires an extensive muscle system to propel the bodies through the dense liquid. This thesis aims to improve the understanding of the relationship between the morphological changes caused by pregnancy and the swimming performance of fish. I based my research on the live-bearing fish family Poeciliidae, which presents a wide variation of maternal provisioning that range from feeding their internally developing embryos exclusively from the nutrients that have been stored in the egg-yolk prior to fertilization (lecithotrophy), to almost exclusively nourish them through a placenta (placentotrophy).

A live-bearing mode of reproduction may increase the volume and mass of the female throughout the reproductive cycle increasing the costs of locomotion. Placentation may have evolved to reduce these costs as is proposed by the Locomotor cost hypothesis.
In Chapter 2, we explored the morphological aspects of this hypothesis, which predicts that females with a placenta have slender bodies than those without it, especially at the early stages of pregnancy, and that the difference diminishes over the course of gestation. This may indicate that females with a placenta would have a less-impaired swimming performance than those without it, which would increase their survival rate. We tested this hypothesis by comparing the three-dimensional body shape changes between two closely related species, one placentotrophic (Poeciliopsis turneri) and one lecithotrophic
(Poeciliopsis gracilis). For this, we developed and used a non-invasive photogrammetry method to document the three-dimensional shape and volume changes through pregnancy. Our findings supported the Locomotor costs hypothesis by showing that females with a placenta possessed more slender bodies in the early stages of pregnancy compared to the non-placental one, and that this advantage diminished over the course of gestation. Moreover, body thickness increased faster in the placental species than in the non-placental one. This provided the first empirical evidence for a possible adaptive morphological advantage of the placenta in live-bearing fish.

In both lecithotrophy and placentotrophy, the growing embryos increased the volume of the female and consequently its body surface and cross-sectional area. This could increase body drag, presumably causing a decrease of the maximum swimming speed and increasing the metabolic costs of locomotion. Ultimately this may decrease the fitness of the female by making her an easier predation target. However, the effects of pregnancy and the level of reproductive investment on the production of drag have not been evaluated so far in live-bearing fish. In Chapter 3, we examined the effect of the reproductive allocation increase (RAI, the increase in the proportion of body mass dedicated to reproduction) on the body drag of pregnant fish. We generated 3D printed models of females of P. gracilis in a straight body posture with different levels of RAI. And then we measured the drag and visualized the flow around them in a flow tunnel at different speeds. We confirmed that higher RAI resulted in increased drag, and more specifically we found that drag grew exponentially with higher RAI, while it grew in a power fashion with the increase of speed. The visualization of the flow around the fish showed that flow separation, one of the main sources of body drag, happened behind the abdominal region and that the separation i ncreased with bigger abdomens.

All of these measurements, be they morphological, biomechanical or physiological, must be analysed taking into account their effect on swimming performance and ultimately fitness, which is finally on which natural selection acts. In Chapter 4, we analysed the effect of pregnancy on the escape response, the main defence of many taxa against predator attacks. We elicited and measured the escape response of pregnant and virgin females of P. gracilis before and just after giving birth, and filmed it with three orthogonal high-speed cameras. We expected that pregnancy and the stage of gestation would have a significant
effect, as it is the case for other fish species, and we assumed that virgin females carry low to none reproductive allotments and therefore their morphology and physiology would not be affected by the burden of reproduction. Contrary to our predictions, we found that the stage of pregnancy and the reproductive stage (virgin vs pregnant) did not seem to have a significant effect on the escape response of P. gracilis, at least for the relatively small brood sizes (mean RAI: 13.5%) happening throughout the experiments.
The absence of significant differences could be due to the small body shape variation of P. gracilis throughout pregnancy which might be partially caused by superfetation (the ability of females to carry simultaneously multiple litters at different levels of embryonic development). Moreover, we found an unexpectedly high RA in virgin females, which might be one of the reasons for the absence of detectable differences between pregnant and virgin individuals. This suggests that the higher locomotor costs caused by carrying the unfertilized yolked eggs might affect also the locomotor performance of the virgin
females. These findings may have some bearing on the Locomotor cost hypothesis and the evolution of the placenta because they suggest that: (i) for a lecithotrophic, superfetatious species which has smaller changes in RA than a species with a placenta (Chapter 2), the increase in locomotor cost might be much smaller than for a species with a placenta, and (ii) that the placenta might lower the locomotor costs of females even before fertilization of their eggs, because they are exempt from the burden of carrying already all the nutrients needed for the embryo development, and can therefore keep a more slender body shape than virgin females who lack a placenta. The absence of significant differences in the escape response might further have been caused (at least partially) by the large variation in the escape responses presumably due to a suboptimal stimulus.

Escape responses and any other swimming behaviours are powered by the fish muscles. During pregnancy, the mobilization of energy from muscles to reproductive tissues and the extension of the abdomen, might negatively affect the contractile properties of the muscles and their power output as it has been shown for gravid females of other clades. Yet, no information is available for live-bearing fish on the possible morphological and physiological adaptations to their reproductive mode. In Chapter 5, we described for the first time the muscle morphology and the activity of genes that express essential components for muscle function for gravid fish. For pregnant and virgin females of P. gracilis, we stained with histological methods transverse sections of the abdominal region and performed real-time quantitative PCR on the muscles in the caudal peduncle to quantify the transcription levels of genes that are particularly expressed in muscles.
Our expectation was that pregnancy would induce changes in the muscle morphology and gene expression. Contrary to the expectations, we found no significant differences between pregnant and virgin females. However, the unexpectedly high reproductive allocation of virgin females decreases the morphological difference between them and the pregnant ones, which can result in similar production of drag (Chapter 3), and in general, similar locomotive needs (Chapter 4). And this might result in similar muscle architecture to cover those needs (Chapter 5). Another unexpected result was the finding of a newly reported extra zone of red and pink muscle fibers present in the belly region of pregnant and virgin individuals. However, males also possess this second region (though it is comparatively smaller), arguing against the idea that this region evolved to facilitate sustained swimming of females during pregnancy.

Our study shows that pregnancy causes changes in the body shape of females and that those changes might be related to the reproductive strategy of the species. We demonstrated how increasing reproductive allocation affects the drag production during coasting, and we expect that the scaling of drag, depending on RAI and speed that we found, may be indicative for the undulatory swimming as well. Moreover, our findings suggest that the lower morphological changes experienced by lecithotrophic, superfetatious species might be reflected in smaller or no significant effects on the muscle morphology and escape response. Our biomechanical approach is essential for the understanding of the evolution of complex traits such as placentation. However, to verify the morphological and performance advantages that different reproductive traits confer and that result in higher fitness, it is necessary to perform multispecies comparisons addressing different origins of the reproductive trait.