*4.4. Integrated Assessment*

Interpreting cortisol fluctuations in free-living vertebrates is certainly a complex practice, particularly when applying alternative matrices for hormone assessment [6,9]. This is why linking cortisol levels to other endpoints of the stress responses can significantly enhance the current understanding on the ecology of stress [76].

In the present study, fluctuations in skin mucus cortisol levels between habitats paralleled those detected in blood, the traditional matrix used for hormone assessments in fish [77]. Relative to habitat quality, changes of the hormone in skin mucus also coincided with variations in the hematological parameters, except for MN levels. Furthermore, the amount of cortisol in skin mucus was directly proportional to frequencies of abnormal erythrocytes (IE and ENA) and to the well-established stress index N:L ratio. Red blood cells are highly sensitive to landscape disturbances [78] and, more specifically, to environmental pollution [79,80]. Accordingly, the measurement of abnormal erythrocytes has been successfully used to assess the health status of the Catalan chub [11,42] and many other fish species [12,13,17]. In the same context, WBC counts, particularly the N:L ratio, increases in individuals exposed to heavy metals and other contaminants proportional, indeed, to the circulating cortisol levels [16]. Given the very clear e ffect of pollution on leukocyte and erythrocyte profiles, the strong linkages detected in this study provide new evidence that the measurement of cortisol in skin mucus could be potentially used as a biomarker of habitat quality in freshwater fish residing polluted environments.

The use of a robust sample size is recommended in natural settings where individuals are exposed to di fferent environmental conditions [9,81]. However, capturing a relevant number of individuals in wild conditions may not always be possible, especially in highly degraded habitats such as the one included in the present field experiment. An important limitation of our study design may, therefore, be the small sample size in one of the polluted habitats, which requires results of this site be interpreted with caution.

While physical and chemical techniques are commonly applied for wetland monitoring [39], methodologies that provide information about how animals are influenced by their environment are only occasionally used. The incorporation of biomarkers into the constructed wetlands' managemen<sup>t</sup> would provide complementary data to the conventional analyses. Hence the demonstrated sensitivity of the methods evaluated in the present study to di fferent pollution gradients could be exploited for biomonitoring the wetlands systems' performances. Indeed, the non-invasive measurement of cortisol in skin mucus would largely improve our understanding about the link between the detected chemical concentrations and the biological e ffects observed.

**Author Contributions:** Conceptualization, C.E., M.I., D.V., and M.L.-B.; methodology, A.C., P.S., and O.T.-P.; software, A.C.; validation, A.C. and O.T.-P.; formal analysis, A.C.; investigation, A.C. and P.S.; resources, M.L.-B. and D.V.; data curation, A.C.; writing—original draft preparation, A.C.; writing—review and editing, A.C., P.S., O.T.-P., M.L.-B., and D.V.; visualization, O.T.-P. and M.L.-B.; supervision, M.L.-B., C.E., M.I., and D.V.; project administration, D.V.; funding acquisition, C.E., M.I., and D.V.

**Funding:** This research was financially supported by Aigües de Barcelona and sponsored by Diputació de Barcelona (DiBa) in the context of the Biobesòs Project (Evaluation of the wetlands' e fficacy to remove pollutants and improve the biodiversity levels in the Besòs river and its surroundings).

**Acknowledgments:** The authors thank Pau Fortuño, Narcís Prat, Núria Flor, and Núria Cid from Universitat de Barcelona (Barcelona, Spain) for their excellent help with sampling and technical assistance. We are grateful to Juan Carlos Ruiz and María Monzó from Aigües de Barcelona, and Anna Riera and Antoni Maza from DiBa for the support to make this project possible.

**Conflicts of Interest:** The authors declare no conflicts of interest.
