Skip to main content
Dryad logo

Quantification of faecal glucocorticoid metabolites as a measure of stress in the rock hyrax (Procavia capensis) living in an urban green space


Carlin, Elisabetta et al. (2021), Quantification of faecal glucocorticoid metabolites as a measure of stress in the rock hyrax (Procavia capensis) living in an urban green space, Dryad, Dataset,


Despite the abundance of rock hyrax (Procavia capensis) within South Africa's urban areas, there is not much information available about the effect of anthropogenic activities on rock hyrax wellbeing. To determine the potential impact of anthropogenic disturbance on adrenocortical activity, we conducted an ACTH challenge to identify a suitable enzyme-immunoassay (EIA) for measuring faecal glucocorticoid metabolite (fGCM) concentrations in the rock hyrax. This study identified an 11β-hydroxyaetiocholanolone EIA as the most suitable assay in this regard. The fGCM levels measured, indicate the physiological stress response in different rock hyrax populations, living in an area with varying degrees of anthropogenic activity (low, medium, high) within the National Botanical Garden of Pretoria, South Africa. The species’ habituation to human numbers (weekly mean number of people) was examined by determining individual flight initiation distance (FID). Seasonally, there were overall higher fGCM concentrations in late spring compared to winter. The fGCM concentrations, although not significantly different but possibly biologically relevant, in the section with the lowest anthropogenic disturbance were ~10% higher compared to those in the section with medium disturbance, and ~20% higher compared to those in the section with the highest disturbance. Animal FID did not differ significantly between seasons but they did differ significantly between sections, and decreased in accordance with fGCM concentrations. The non-invasive approach established in this study provides a foundation for assessing rock hyrax wellbeing, and can help better understand how anthropogenic presence is perceived as a stressor in this species.


For this study, we conducted an ACTH challenge test at the SANBI Mokopane Biodiversity Conservation Centre on 3 animals (two females and one male) to collect serum cortisol data and fGCM concentrations pre and post ACTH test. To confirm effective ACTH administration, we collected four blood samples (0.25-0.4 ml) from the brachiocephalic vein from each individual; one immediately before and three at 15, 30, 45 min post-injection. We also collected faeces for 46 h after the ACTH test, to measure fGCM concentrations. All the samples were collected using clean, sterile latex gloves, stored in single labelled tubes, and immediately frozen at –20°C, until transferred for analysis to the Endocrine Research Laboratory (ERL) at the University of Pretoria. To determine alteration in fGCM concentrations post-defecation, in order to get an indication of suitable collection intervals in the field without compromising adrenocortical activity analysis reliability, fresh faeces were collected on one day from a communal latrine at the SANBI Mokopane Biodiversity Conservation Centre. Faeces were collected using sterile gloves, homogenized into one single sample, subsequently divided into 21 sub-samples, and stored outside (20 - 27°C, 40 - 50% humidity, no precipitation). Following the methodology explained by Webber et al. (2018), sub-sample triplicates were frozen at –20°C at 0 h, 1 h, 2 h, 4 h, 8 h, 16 h, and 24 h. All frozen samples were sent to the ERL for enzyme immunoassay analysis. 

For the second part of the research, we collected faeces, using the same methodology, from a colony of free roaming Procavia capensis at the Pretoria National Botanical Garden (PBG). All samples were collected over 35 days in winter (July-August 2019) and over 25 days in late spring (October-November 2019), from 07:00 to 08:00. The intent was to monitor whether there was a correlation between human presence and procavia's fGCM concentrations, thus, as visitor numbers were not evenly distributed, due to various activities offered in distinct areas of the park, we divided PBG into three different sections (1,2,3) based on human activity present (low, medium, high). To evaluate the assumption of different levels of human numbers in different sections of the PBG, we used digital motion activated cameras (Cuddeback digital C.5.1- 1347 - Wisconsin, USA), with a trigger time of 0.25 seconds, and a detection zone of 20 m, to quantify human activity in the respective areas. Cameras were set up following Miller and colleagues (2017). Pictures were taken from 06:00 – 18:00 over a total of 50 days, split into two sampling periods (Period 1: 35 days in winter, July-August, and Period 2: 25 days in late spring, October-November 2019). Following Chapman and colleagues 2012, during the same study period of camera traping and faecal sampling (35 days in winter, July-August, and 25 days in late spring, October-November 2019), we also took FID measurements from animals living in each section, using a laser range finder (Nikon Forestry Pro laser range finder, accuracy: +18.2 mm, distance: 10-500 m). A total of 132 individual total measurements for colony one, 123 for colony two and 115 for colony three were collected, and daily median were calculated (n= 15) for each colony during both study periods.

Blood samples analysis: the ERL analysed 15 serum samples using a Cortisol EIA, originally described by Palme and Möstl (1997) and following procedures previously described (Ganswindt et al. 2012). Palme and Möstl (1997) provide detailed assay characteristics, including antibody cross-reactivities.

Faeces analysis: Frozen faeces were lyophilized, and the resulting dry faeces pulverized and sifted with a wire filter to remove any undigested material (Ganswindt et al. 2002). Steroid extraction was performed by dissolving 0.050 - 0.055 g of faecal powder in 3.0 ml of 80% ethanol, with each sample being subsequently vortexed for 15 min and then centrifuged for 10 min at 1500 g (Ganswindt et al. 2002). Lab technicians decanted supernatants into 1.5 ml safe-lock microcentrifuge tubes, labelled, and frozen at -20°C until further analysis. To determine the suitability of the EIAs for fGCM quantification, a subset of the ACTH challenge faecal extracts (n = 13 per animal), were measured for immunoreactive fGCM concentrations using five different enzyme-immunoassays (EIA): (i) a 11-Oxoaetiocholanolone I EIA (detecting 11,17 dioxoandrostanes) (lab code 72a), (ii) a 11-Oxoaetiocholanolone II EIA (detecting fGCMs with a 5β-3α-ol-11-one structure) (lab code 72T), (iii) a 5α-pregnane-3β,11β,21-triol-20-one EIA (detecting fGCMs with a 5α-3β,11β-diol structure) (lab code 37e), (iv) a Cortisol EIA (lab code CSL), and (v) a 11β-hydroxyaetiocholanolone EIA (detecting fGCMs with a 5β,3α,11β-diol structure) (lab code 69a). Detailed assay characteristics, including cross-reactivities, can be found in Palme and Möstl (1997) for the 11-oxoetiocholanolone I and cortisol EIAs, Möstl et al. (2002) for the 11-oxoetiocholanolone II EIA, Touma et al. 2003 for 5α-pregnane-3β,11β,21-triol-20-one EIA and Frigerio (2004) for the 11β-hydroxyaetiocholanolone EIA. The sensitivity of the EIAs used were 1.2 ng/g dry weight (DW) (11-oxoaetiocholanolone I, 11-oxoaetiocholanolone II and cortisol), 2.4 ng/g DW (11β-hydroxyaetiocholanolone), and 4.8 ng/g DW (5α-pregnane-3β,11β,21-triol-20-one). The cortisol and 11β-hydroxyaetiocholanolone EIA) for analysing the steroid extracts from the fGCM stability post-defecation test. Serial dilutions of faecal extracts gave displacement curves that were parallel to the respective standard curves (relative variation (%) of the slopes of respective trend lines <5%) for both assays. Subsequently, faecal samples collected at the PBG were analysed by using only the 11β-hydroxyaetiocholanolone EIA. All hormone analyses were conducted at the ERL.

Statistcal analysis: We used descriptic statistics to analyse the pattern in serum GC and fGCM concentrations following ACTH administrations and to describe alterations in fGCM concentations post defecation. Changes in fGCM concentrations post-defecation, determined with the cortisol and 11β-hydroxyaetiocholanolone EIA, were further analysed using One-way Analysis of Variance (ANOVA), followed by Holm-Sidak’s t-test as post hoc. Individual serum GC concentrations, collected at the point of ACTH injection (time 0h), and individual median fGCM concentrations determined from samples collected up to 26 h pre-ACTH administration, were set as baseline (100%). Subsequently, we calculated relative changes (%) in serum GC concentrations for individual samples collected at 15-, 30-, and 45-min post-injection. Differences in peak fGCM concentration 15-22 h post-ACTH administration were identified by comparing individual peak values with individual median pre-injection (baseline) values and normalised those by expressing these differences in percent. Peak values with an increase of >100% above baseline were deemed considerable (Young et al. 2017). Variations in human numbers and FID for each study period were determined by using a generalised linear model (GLM), using the weekly mean number of people and, calculating daily median FID for each colony, as a response variable and section and season as predictor variables. An interaction between season and section was included to the model. The GLM analyses were followed by a Tukey’s post hoc analysis, conducted at 95% family wise confidence levels. A GLM was also used to calculate whether the covariates section (1, 2, 3) and season (late spring and winter) significantly influenced variations in fGCM concentrations. Subsequently, the use of a Tukey’s post hoc test, conducted at 95% family wise confidence levels, followed by Bonferroni adjustment, helped to determine whether significant differences in seasonal fGCM concentrations for each section occurred. Data are presented as mean ± standard deviation (SD), and significance was taken at 0.05. Data were statistically analysed using R Studio software (R Core Team 2016, 4.0.2 version).

Usage Notes

Please note that animals ID are different in the dataset compared to those used in the manuscript:

Animal A5-1 in the dataset corresponds to Female F1 in the manuscript

Animal A5-2 in the dataset corresponds to Male M1 in the manuscript

Animal A5-3 in the dataset corresponds to Female F2 in the manuscript


University of Pretoria