Methods for culling wild red deer in Scotland were compared by observing behaviour and collecting 164 post-mortem samples from wild deer shot: using a helicopter for deployment of stalkers and carcase removal (WH); by more than one stalker (WS+); by one stalker during the day (WSday); by one stalker at night (WSnight), and from farmed red deer shot in the field (FF) or killed at a slaughterhouse (FS). Deer were active before the first shot only in WH (27%) and WS+ (21%) (Fisher's exact test P<0.001). Forty-eight percent of deer culled by WH and 43% of WSnight deer moved away after the first shot without collapsing (Fisher's exact test P<0.01). Plasma cortisol concentrations in deer shot in WH (log10 mean 1.79 nmol/l SEM 0.059) were similar to those in FS (log10 mean 1.93 nmol/l SEM 0.032), but higher than WSnight (log10 mean 1.25 nmol/l SEM 0.046), WSday (log10 mean 1.42 nmol/l SEM 0.077) and FF deer (log10 mean 1.42 nmol/l SEM 0.070) (GLM P<0.001). WS+ culled deer (log10 mean 1.70 nmol/l SEM 0.101) had a greater plasma cortisol concentration than WSnight. WSnight deer (log10 mean 2.91 nmol/l SEM 0.049) had a greater muscle glycogen concentration than those culled via WH (log10 mean 2.56 nmol/l SEM 0.048), WS+ (log10 mean 2.64 nmol/l SEM 0.070), and WSday (log10 mean 2.65 nmol/l SEM 0.046); and FF (log10 mean 2.86 nmol/l SEM 0.076) had a greater concentration than WH (GLM P<0.001). Culling by WSday and FF were the most accurate and WSnight the least in achieving shot placement likely to have quickly killed the deer. There were no significant differences between methods in the numbers of wild deer that appeared to have died immediately after the first shot or were wounded. Wild deer were more disturbed by either WH or WS+ than with WSday.