Alberto Mansilla M. ; Carlos Pedraza G.1; Paola Fajardo R. y Hernán Aguero E.
The objective of this study was to estimate the sub-clinical incidence of mastitis on dairy cows, defined as the count of somatic cells in milk (RCS) based on the California Mastitis Test (CMT) using correlation, linear regression and logit models. A database of five dairy herds from central Chile composed of 1200 mainly Holstein Friesian dairy cows was used. Information was systematized for each cow calculating the sum of CMT for individual quarters (SCMT), and transforming RCS in cellular ranges (RC), adding the average CMT value of the graduated milking recipient (CMTM). Linear correlations were calculated among these variables, linear regression CMTM and RC and the logit model that indicates the relationship in probabilistic terms: pj = 1 / (1 + EXP(b SCMT - ck)). Correlations were significant (P<0.01) with values between 0.37 and 0.78. Estimated linear regressions were RC = 0.74 SCMT for both kinds of animals. The logit function estimated a parameter b of 0.335 and R2 = 0.83 for heifers and b = 0.30 and R2 =0.63 for cows. The logit model proves to be a very accurate instrument for RCS estimation in extreme cases, with no mastitis presence, or mastitis close to clinical levels.
Key words: dairy cows, subclinical mastitis, diagnostic, somatic cell, logit model.
1 Instituto de Investigaciones Agropecuarias,Centro Regional de Investigación La Platina, Casilla 439, Correo 3, Santiago, Chile. E-mail: firstname.lastname@example.org.