Collection of faecal samples

The dung piles (n = 80, three to nine boli each) were collected from the 10 study individuals for five consecutive days (13–17 January 2013) between 04.00 and 09.00 h. Following defecation by any of the elephants, the dung piles were collected and sampled immediately (hereafter known as ‘time 0’ samples) before being randomly assigned to one of the 2 × 2 treatment combinations [wet–shade (n = 21 faecal piles), wet–sun (n = 20), dry–shade (n = 17) and dry–sun (n = 22)]. For ‘wet’ samples, 1 litre of purified water was poured over the dung pile just after the time 0 sample was taken; for ‘dry’ samples, no water was added; for ‘shade’ samples, the dung piles were placed in an area under a tree canopy (70–100% canopy cover); and for ‘sun’ samples, the dung piles were placed in an open field (0% canopy cover). At night (19.00–04.00 h) and during two brief episodes of rain, we covered all the samples with plastic sheets (placed over each dung pile) to avoid rainfall affecting our experimental treatments. During the experiment, the nearest weather station (Felda PPP Tun Razak, 03°50′N, 102°34′E) recorded daily averages of 6.4 ± 2.5 h of sunlight, 25.5°C temperature, 4 ± 8 mm of rainfall and 80 ± 2% relative humidity.

To collect faecal subsamples (n = 685 from the 80 dung piles) for hormone analysis, we used scissors to cut three small openings in different parts of the dung pile (usually on different boli) and removed faecal matter from the centre of the bolus using forceps; we then placed the sample in a ziplock bag, mixed thoroughly, and kept it frozen at −20°C until laboratory analysis. Each dung pile was coded with a unique identity (ID) number. Hormone sample bags were labelled with dung pile ID, date and time of sampling. Faecal subsamples were taken at a range of times, with sampling occurring more intensively within the first half of the day of defecation to capture initial changes in fGCM and less frequently thereafter, as follows: time 0 (fresh defecation), 0–2, 2–4, 4–6, 6–8, 8–11 and 11–16 h and 1, 1.5 and 2 days after defecation. Dung piles collected at the start of the day (e.g. at 04.00 h) were sub-sampled more frequently than dung piles collected later in the day (e.g. at 09.00 h), hence the number of subsamples differed between dung piles.