The objective of this study was to estimate the effects of recurrent episodes of gram-positive and gram-negative cases of clinical mastitis (CM) on milk production in Holstein dairy cows. We were interested in the severity of repeated cases in general, but also in the severity of the host response as judged by milk production loss when a previous case was caused by a similar or different microorganism. The results were based on data from 7,721 primiparous lactations and 13,566 multiparous lactations in 7 large dairy herds in New York State. The distribution of organisms in the CM cases showed 28.5% gram-positive cases, 31.8% gram-negative cases, 15.0% others, and 24.8% with no organism identified. Mixed models, with a random herd effect and an autoregressive covariance structure to account for repeated measurements, were used to quantify the effect of repeated CM and several other control variables (parity, week of lactation, other diseases) on milk yield. Our data indicated that repeated CM cases showed a very similar milk loss compared with the first case. No reduction of severity was present with increasing count of the CM case. Gram-negative cases had more severe milk loss compared with gram-positive and other cases irrespective of the count of the case in lactation. Milk loss in multipara (primipara) due to gram-negative CM was approximately 304 kg (228 kg) in the 50 d following CM. This loss was approximately 128 kg (133 kg) for gram-positive cases and 92 kg (112 kg) for other cases. The severity of a second case of gram-negative CM was not reduced by previous cases of gram-negative CM in multipara and only slightly less severe in a similar scenario in primipara cows. Similarly, a previous gram-positive case did not reduce severity of a second or third gram-positive case. Hence, our data do not support that immunological memory of previous exposure to an organism in the same generic class provides protection for a next case of CM with an organism in the same class.