**4. Discussion**

As stated in the introduction section, one of the main limitations in studying wildlife reproductive physiology lies within the nature of the sampling possibilities. It is indeed almost impossible to collect repeated samples on the same animal in order to obtain a dynamic profile of specific analytes throughout the different phases of gestation. This is why collecting samples upon killing during the hunting season may represent a good, if not the only, way to perform this kind of research [19]. This study confirms the chance to collaborate with hunters and local biometric centers.

It is widely recognized that analyzing hormones in hair provides information regarding a longer timespan when compared to blood analyses [19,20], but being able to specifically identify such timespan can be challenging. Indeed, due to the nature of the sampling, it is not always possible to completely trim the chosen area to calculate the growth rate of the hair in wildlife animals prior to a subsequent hair sampling. The red deer undergoes two pelages, also known as moultings, one in spring and one in autumn, approximately in May and October, respectively [28,29]. Therefore, the hair collected in the present study belongs to the winter coat and should provide info regarding few months before trimming.

The hormones analyzed in the present study, in both matrices, showed wide ranges suggesting strong individual variations. The differences in progesterone levels may be related to the fact that the sampling took place throughout more than one month and that hinds may be in different phases of pregnancy (approximately from the third up to the fifth month of gestation). Moreover, it was already proved that, in this species, the levels of plasma progesterone fluctuate and correlate with the number of corpora lutea present [14]. On the other hand, cortisol variations, already described in literature for the red deer [20], are related to a multitude of environmental, seasonal, and physiological factors.

The mean level of blood progesterone recorded in the present study (3.86 ng/mL) is in accordance with previous studies performed in pregnan<sup>t</sup> red deer hinds farmed in paddocks [14,30]. Such agreemen<sup>t</sup> of data seems to support the possibility to use data collected from farmed animals also for wildlife. Such statement is extremely preliminary and particular attention will have to be paid if the analytes were to concern stress patterns, which will necessarily be influenced by husbandry conditions and environmental factors. Overall, relatively higher levels of plasma progesterone have to be expected in pregnan<sup>t</sup> hinds when compared to non-pregnan<sup>t</sup> animals, as already reported [31].

The mean hair progesterone level was 73.05 pg/mg with high individual variations most likely explained by the variations detected in plasma levels. As of today, to the best of the authors' knowledge, no studies were published assessing hair levels of this hormone in red deer and cervids in general. It is therefore quite challenging to discuss such findings. What can be said, on the basis of the present study, is that the analytical approach seems to be reliable and relatively easy to perform. Hair progesterone

was assessed in cows in a recent study were no differences between pregnan<sup>t</sup> and non-pregnan<sup>t</sup> animals were found [32]. The authors hypothesized that the cause behind the lack of differences may be the short period between calving and the next successful insemination [32]. Again, it is impossible to compare the results with the ones obtained in the present study in light of the physiological differences between the species and the zootechnical pressure dictated by industry demands.

When compared to progesterone, cortisol is overall more influenced by the conditions in which the animals live [15]. Blood cortisol levels change quickly throughout the day and are therefore relatively unreliable, but quantification in matrices such as feces and hair seems to be a good indicator of chronic stress, potentially also related to the reproductive status of the animal [16,23]. Moreover, cortisol, in all matrices, shows high seasonality variations in red deer [33] as in other wild ungulates [34].

Our results show, as expected, highly variable plasma CORT levels ranging from 0.38 to 2.97 ng/mL. This is likely to be imputed to several "acute" factors and the variability of this hormone pattern itself [22,34]. On the other hand, hair CORT levels are the reflection of the continuous incorporation of the hormone into the hair shaft [15], and thus less influenced by acute stressors. This is why this matrix has been addressed in the last years as highly efficient in wildlife [19]. A study performed on Italian red deer reported mean values of hair CORT of 5.75 pg/mg in females [20], almost 20 times higher than the mean value obtained in the present study (0.29 pg/mg). Factors influencing cortisol production during pregnancy are extremely variable and partially unknown [35]. What is known is that cortisol increases in the late phase of a pregnancy [23], and that, in cows, parturition in the month preceding sampling increases hair cortisol levels [36]. Animals sampled in the present study were approximately at the third month of gestation, and therefore in the first phase of pregnancy. It is possible to suppose that hair levels of CORT would have been higher in case of later phases of gestation. The correlation between hair CORT and chronic rather than acute stress was further confirmed by a study carried out on a red deer population from the central Italian Alps, Sondrio Province [20].

Our correlation analysis did not highlight any relationship between the analyzed parameters aside from a significant correlation (ρ = 0.659; *p* = 0.042) between hair P4 and the date of death. This correlation may be explained by the ongoing pregnancy, characterized by high levels of P4. Therefore, it is likely that hinds killed later on have had higher accumulation of this hormone in the hair. A weak correlation, not statistically significant, was also found between the plasma levels of the two hormones, P4 and CORT (ρ = 0.612; *p* = 0.060). Nonetheless, it is important to acknowledge that the low sample size may be responsible for a weak power of the statistical analysis.

The present study strengthens the use of hair for endocrinological evaluations in wild animals, in light of the easy and non-invasive sampling procedure and the representation of longer time periods in one single sample [15]. Our purpose was to describe the concentration of progesterone and cortisol, from hair and blood, of the wild population of red deer females (*Cervus elaphus* L., 1758) living in the Apennine area, during a particular physiological situation such as gestation. In conclusion, performing such studies on hunted animals seems to be useful in better understanding the physiology of wildlife animals and may provide new useful data that can benefit the species and the scientific community.

**Author Contributions:** A.E. and M.L.B. conceptualized the study; A.E. coordinated the sampling procedures; M.B. and N.G. performed the analyses; A.E. and D.V. analyzed the data; D.V. and C.A. drafted the manuscript; N.G., M.L.B., and A.P. revised and approved the final draft. All authors have read and agreed to the published version of the manuscript.

**Funding:** The research was funded by RFO-UNIBO.

**Acknowledgments:** Authors would like to thank the volunteers of "URCA Bologna" Association and the "ATC BO 3" for their help in the sampling process and Stefano Mattioli for the age estimations.

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