We have investigated the responses of con tests to machine milking in a stu
dy of relatively newly installed commercial milking parlours fitted with on
e of two types of milking cluster. The first was a common type with a large
clam volume (> 200 mi), 15-16 mm i.d. long milk tube, 10 mm short pulse tu
be, cluster weight < 3.2 kg and used alternate pulsation. The second was a
more traditional type with a 150 mi claw bowl volume, 13.5 mm i.d. long mil
k tube; 8 mm short pulse tube, cluster weight similar to 3.5 kg and used si
multaneous pulsation. We scored similar to 50 cows in each of 20 herds, all
within 60 s of cluster removal, for changes from the premilking teat condi
tion: teat colour (creation of reddening or blueness), firmness, thickening
at the base of the test associated with the position of the liner mouthpie
ce, and whether the tent duct orifice was open. There were statistically si
gnificant differences in the proportion of cows displaying these four alter
ations in teat condition between herds using the two types of cluster. The
more common type of cluster was always associated with better tent conditio
n. The cause and effect of poorer teat condition hare not been fully establ
ished and are likely to be multifactorial. The principal risk factors may b
e cluster weight, overmilking, vacuum applied during the overmilking phase
and the design of the liner mouthpiece.