Twenty-four Holstein cows, producing at least 21 kg of milk/d, were us
ed in two replicate experiments to determine the effect of presence or
absence of pulsation on loss of teat canal keratin during machine mil
king. Left quarters were milked without pulsation and right quarters w
ere milked with pulsation. On d 0 and 10, keratin was collected from o
ne left and from one right teat canal of each cow prior to milking and
from the remaining two teat canals after milking. Milk was collected
for assessment of SCC and bacteriological status on d 0 and approximat
ely every 3 d until d 18. Quantity of keratin recovered before milking
on d 10 did not differ between teats milked with or without pulsation
, but loss of keratin because of milking was greater from teats milked
with pulsation. By d 7, 30% (12 of 43) of quarters milked without pul
sation had become infected, but no (0 of 47) quarters milked with puls
ation were infected. By d 14 to 16, new infections had increased to 68
% (28 of 41) of quarters milked without pulsation and 2% (1 of 43) in
quarters milked with pulsation; mean SCC in pulsationless quarters inc
reased sevenfold relative to pulsation quarters. Protein and water con
tent of keratin did not differ because of treatment, and changes in li
pid composition were minor. Histological analysis of the teats of 4 co
ws indicated that the mean diameter of the teat canal, within 2 h afte
r milking, was greater without pulsation than with pulsation (680 vs.
483 mum).