The teats of 18 cows were externally exposed to infection with mastitis organisms by dipping them daily for 1 or 2 weeks in a mixed suspension of Streptococcus dysgalactiae and Str. uberis at the start, the middle or the end of the dry period. The teat sinuses of quarters which remained uninfected after external exposure at the beginning or the middle of the dry period were then infused with the same strains of streptococci. Of the 9 quarters infected following external exposure 8 occurred in animals exposed at the start and one in the middle of the dry period. Thirty-six new infections occurred after infusion of bacteria into the teat sinuses of 38 quarters. In a further trial with 10 cows, Staphylococcus aureus and Str. zooepidemicus were inoculated into the distal 3 mm of the streak canals of 5 cows immediately after drying-off and into those of 5 cows which had been dry for 28 weeks. Animals were slaughtered 48 h later and infection determined by teat puncture. Five infections occurred in cows which were at the start of the dry period and only one in the cows dry for 28 weeks. It is suggested that these differences in the rates of new infection between the early dry period and a very extended dry period are due to differences in the ease with which bacteria can penetrate the teat canal. At the later stage bacterial growth through the teat canal appeared to be inhibited.The unmilked udder is particularly susceptible to infection in the 3 weeks following drying-off (Neave, Dodd & Henriques, 1950). Indeed, rates of new infection during the dry period exceed those of lactation, which may be due to the absence of the flushing out effect of regular milking (Neave et al. 1968; Thomas etal. 1972). However, this could not account for the large difference in the rate of new infection between the early part and the later part of the dry period. This difference may be due to the higher level of exposure to bacteria at the beginning of the dry period, i.e. bacteria remaining on the teat skin from the preceding lactation period. Alternatively, progressive changes in the composition of the udder secretion during the dry period (Wheelock et al. 1967;Smith, Wheelock & Dodd, 19676) may influence the establishment of infection after bacterial penetration of the streak canal. It is also possible that changes within the streak canal, such as the formation of a natural seal, may make bacterial penetration more difficult in later stages of the dry period. The aim of the experiments described in this paper was to investigate further the changes in susceptibility to new infection during the dry period. Two experiments were carried out; the first was designed to expose the udders of dairy cows to similar numbers of pathogenic bacteria either at the beginning, the middle or towards the end of the dry period and to measure the resulting numbers of new infections. The second