The objective of the study was to determine the effect of varying degrees of overmilking on teat-end hyperkeratosis, milk production variables and indicators of udder health during late lactation. This was examined by assessing the effect of four end-of-milking criteria on 181 spring-calving, mixed-age Holstein-Friesian cows, at an average 217±24 d in milk, over a six-week period. The four treatments were: remove cluster once milk flow rate fell to 0·2 kg/min plus 5 s (Ovr0), plus 120 s (Ovr2), plus 300 s (Ovr5), and plus 540 s (Ovr9). Daily measurements included individual cow milk yield, milking duration, overmilking duration, maximum milk flow rate, milk flow rate at cluster removal and the number of cluster re-attachments. Individual cow bulk milk samples were collected weekly at AM and PM milkings to determine composition (fat, protein and lactose) and somatic cell count (SCC; AM only). Teat-end hyperkeratosis score was assessed at weeks 0, 3, 5 and 6. At week 6 mean teat-end hyperkeratosis score of the Ovr2 treatment was not greater than Ovr0, whilst Ovr5 was greater than Ovr2 and Ovr9 was greater than Ovr5 and Ovr2. Milk production, milking characteristics and SCC were not different between treatments, except milking duration and milk flow rate at cluster removal. However, higher teat-end hyperkeratosis scores may have a longer-term impact on indicators of udder health if teat-end condition reaches severe levels. Results indicate that to minimise changes in teat-end condition overmilking should be limited to 2 min, which has implications for milking management in large parlours not fitted with automatic cluster removers.