Abstract
<jats:p>Steroid hormone concentrations change with reproductive state and time. Yet short-term adrenal patterns remain unclear in managed felids. We reanalysed 41 dated faecal glucocorticoid immunoreactivity (FGI) measurements from one adult female snow leopard over 53 calendar days. A separately prepared behaviour record identified a reconstructed five-day peri-mating episode among 16 overlapping dates. We compared phases and fitted continuous-time autoregressive candidate models. We also fitted local temporal models without behaviour labels. Further checks used exact temporal nulls, complete calendar windows, zero- to two-day behavioural lags and a hypothetical multiplicative-imprecision stress test. Episode FGI was 4.419 times recovery FGI (date-conditional 95% bootstrap interval 2.014-8.631; Cliff's delta = 0.80; exact Mann-Whitney P = 0.063). The top full-series model included smooth time, episode and recovery terms (AICc weight = 0.364). An episode-and-recovery model retained weight = 0.256. Their adjusted episode-to-recovery ratios were 5.155 and 3.923. In the complete 10-24 February series, a quadratic model peaked on 17 February (coefficient-simulation interval 16-18 February). AICc-selected segmentation placed changes on 13 and 21 February. The episode ranked third among 20 complete five-day windows. It ranked first among seven complete five-day/next-four-day structures. Same-day and +1-day behaviour discrimination each had AUC = 0.80. The +2-day AUC was 0.575. Simulated independent multiplicative imprecision preserved the phase ratio but changed the five-day rank. The record supports a transient temporal association in this female. It does not show a reproductive-specific effect, validated assay response, pregnancy marker, welfare diagnosis or species-level pattern.</jats:p>