Biodiversity of trait covariation in teleost fishes

Phenotypic variance is considered the root of every evolutionary process. This variance is not limited to characteristics that can be directly observed, for example behavioural or physical ones, but can be detected in all aspects of animal’s biology, from physiology to cognition. In the last few decades, studies that simultaneously focused on the variance of more than one character found that they are often non independent, but the phenotypic score of multiple traits tend to covary among individuals. Some of these covariations are well studied (i.e., pace-of-life-syndrome and behaviour-cognition covariation), but others are less known. In this thesis, I present an original work aimed to study these covariations across the main biology’s domains (life history, physiology, behaviour and cognition) in order to gain a deeper understanding of their possible evolutionary causes, consequences and factors that might influence them. In order to do this, I centred my studies on teleost fish, a broad infraclass of vertebrates that have proved to be excellent models for addressing complex biological questions and which allow me to have, among its advantages, the large phenotypic screenings necessary for my work. However, given the varied objectives of each study, I selected different teleost species, each offering unique advantages. In fact, despite being part of the same larger group, the chosen species represented distinct models due to substantial differences in various aspects, as for instance ecological adaptations or life history. With exemption and modifications where needed, the subjects of different studies were tested in batteries of tests to measure phenotypic diversity, with a focus for cognitive tests, which emerged as key in the covariation among traits. In the first study, I explored covariations across domains using wild-caught tench (Tinca tinca). I then explored whether these covariations are affected by the sex of the individuals using a species with large sexual dimorphism (study 2; guppy, Poecilia reticulata) and whether they affect fitness using a species in which this parameter could be inferred based on migration success (study 3; European eel, Anguilla anguilla). Last, I studied whether diversity of traits covariation is affected by human activities, focusing on the problems of hybridisation with non-native species (study 4; marble trout, Salmo marmoratus). The results of all five studies provided sufficient evidence to support our initial hypothesis of the presence of a complex network of covariation among traits, both within and across domains. The first study in tench strengthened our idea of the crucial role that cognition can play in trait’s covariance, also opening up the discussion as to whether the invasive success of this species in places where they are introduced compared to their disappearance in their original locations may be related to differences in traits in the populations. In guppy, the results suggested how the notable network difference in the two sexes could be traced back to difference in ecological and evolutionary pressures. The study made in European eel, an endangered species, highlighted the importance of traits covariation for migration and settlement in new environment, at the same time giving us important information that could be implemented in future conservatory actions. Finally, the comparative study of the endangered marble trout and its hybrid showed us the crucial role of adaptation to certain environments for traits’ covariance, as we detected a significant disruption of them in the hybrid line. In conclusion, this thesis brings new knowledge on the covariation between traits and the role of cognition in these associations, with a focus on the role of different factors in them. In the light of it, future studies not only on fish, but on other models as well, should take into account during their investigation the role of traits’ variance and covariance.