(a) Study species and data collection

We conducted this study in Wytham Woods, Oxfordshire, UK (51°46′24″ N, 1°20′04″ W), a 424 ha semi-natural woodland surrounded by mixed arable pasture [38]. The resident high-density badger population (mean ± s.e. = 36 ± 3 badgers/km²; [39]) consists of large mixed-sex social groups (mean group size = 11, range = 2–29; [40]). Badgers have a polygynandrous mating system with high rates of extra-group paternity [41,42], where males exhibit seasonal peaks in testosterone levels [43,44]. Badgers are exposed to pathogens, such as coccidia, that negatively impact development and cause juvenile mortality [45–47].

Trapping was undertaken three times per year, for three consecutive days per social group in 2017 and 2018. Trapped badgers were anaesthetized using an intra-muscular injection of 0.2 ml ketamine hydrochloride per kg body weight [48]. Individuals were identified by a unique tattoo number on the left inguinal region, with capture date, social group affiliation and sex recorded. Age was determined as the difference between capture date and the 14th of February in the birth year, since implantation and parturition dates are highly synchronous in badgers [49–51]. Badgers first caught as adults were aged through tooth wear (scale 1–5), where a score of 2 typically indicates a 1-year-old adult [52]. Blood was collected through jugular venipuncture into vacutainers with EDTA anticoagulant. Badgers were released at their setts, after full recovery from anaesthesia. Additionally, bait-marking was conducted periodically to delimit social groups [53] and calculate group sizes using dispersal rules (see electronic supplementary material).

Immediately after blood collection, one drop of blood was smeared on a microscope slide. Slides were air-dried for 1 h then stained using Kwik-Diff (Thermo Scientific, Manchester, UK) according to the manufacturer's protocol. Leukocyte cell counts were conducted by the same observer (blind to group size and sex) by counting 100 cells per slide (4 repeats per slide, not consecutively to avoid bias; n = 82 slides, 23 individuals; 9 females, 14 males), at 40 x magnification using the battlement technique [54]. Cells were identified as neutrophils, eosinophils and basophils (i.e. granulocytes) or lymphocytes and monocytes (i.e. agranulocytes; [55]). Basophils (< 0.1%), eosinophils (1.4%) and monocytes (3.4%) were rarely observed, thus we only used neutrophils and lymphocytes to calculate the lymphocyte proportion from these data [56].