## Abstract

Individual quarters are in one of a number of distinct health states, such as 'uninfected', 'infected' and 'recovering from infection'. Somatic cell count (SCC) of an individual quarter is strongly dependent on this health state. The objective of this study was to evaluate whether these states can be recognised in the distribution of SCC and utilised. A dataset with 316,426 SCC testday records of 41,567 cows on 407 farms was analysed. The statistical mode of the distribution (the peak) was the same for a wide range of subsets of data; i.e. grouped by parity, stage of lactation or previous history, but the means varied. Such a case where a change in the mean is largely caused by a shift of records from the peak to the tail, is strong evidence for multiple underlying distributions. We interpreted the peak to be related to cows with four uninfected quarters. The frequency distribution of SCC observations was best described with a mixture of four normal (N1, N2, N3 and N4) distributions and one exponential (E) distribution. Analysis of a second independent dataset with 1,546,570 records on 58,070 cows yielded similar parameters of these five distributions. We interpreted N1 as uninfected cows, N2 as cows recovering from an infection, N3 as cows with a non-persistent infection and N4 and E as cows with persistent infections. From the probability that an observation belongs to an underlying distribution, we defined several new SCC traits. For example, 'Mastitis suspected' is the probability that at least one SCC in a lactation originated from N3, N4 or E. Estimating the percentages of records in each underlying distribution for a herd may also be a useful tool for veterinary advice.

Original language | English |
---|---|

Title of host publication | Mastitis Control: From Science to Practice |

Publisher | Wageningen Academic Publishers |

Pages | 227-234 |

Number of pages | 8 |

ISBN (Print) | 9789086860852 |

DOIs | |

Publication status | Published - 1 Dec 2008 |

## Keywords

- Frequency distribution
- Monitoring
- Somatic cell count