Hair cortisol was extracted and analyzed adopting the protocol defined in Yamanashi et al. (2013); Yamanashi et al. (2016) using 50 mg of pulverized and minced samples. The experiments pertaining to recovery rate and hC variations across body were not performed due to limited access to samples and sensitivity of elephants towards body hair, respectively. The collected tail-hair samples were extracted by adding one ml of 80% methanol to each weighed sample and incubating these samples for 24 h at 50 °C (with shaking at 700 rpm; Incubator SIC-320LW, AS ONE Co., Osaka, Japan). After incubation, extracted samples were centrifuged at 3,000 rpm for 2 mins at 25 °C and 600 µl supernatant was transferred into a 2 ml tube. The tubes with the supernatant were opened and kept for drying (to evaporate methanol) inside the box with silica gel which was placed over the hot plate at 50 °C. Once the sample vials were completely dried (∼20 to 24 h), these dried extracts were reconstituted with 200 µl of the EIA phosphate buffer, were vortexed for a minute and stored at −20 °C for further analyses.

The levels of tail-hair cortisol (hC) were analyzed using the cortisol enzyme-linked immunosorbent assay (EIA) with FKA404E antibody and FKA403 antigen (horseradish peroxidase conjugated cortisol; Cosmo Bio Co., Ltd.) based on Kinoshita et al. (2011). Sample extracts were diluted in the ratio of 1:2 and assessed in duplicate. Inter-assay and intra-assay coefficients of variation (CV) were measured to assess the plate-to-plate consistency and the variance between data points within an assay, respectively; they were found to be 10.23% and 3.77% for the cortisol assay using pulverized hair samples, and 10.78% and 3.42% while using the minced samples, respectively (see Fig. S3 for the standard curve and parallelism). The software Microplate Manager 6 (MPM) was used for calculating the cortisol concentrations from the obtained optical densities.