Genetic selection toward mastitis resistance has not advanced substantially in the past. The detection of new traits that will improve selection is needed. The aim of this study was to examine the feasibility of using threshold concentration values of somatic cells as traits in selection against mastitis susceptibility. Quarter milk samples (n = 965) were collected at d 4 and d 11 post calving from Holstein and Jersey cows reared at the Dairy Education Unit at North Carolina State University. Cell counts for total leucocytes (TLC), neutrophils (N), macrophages (M), and lymphocytes (L) were obtained via differential somatic cell count. Intramammary infection with major and minor pathogens, as well as fungi, was determined via microbiological culturing on duplicate quarter milk samples. Log-transformed scores for the cell counts were calculated. MCMCglmm was used to fit various models for the following dependent variables: TLC, N, M, L, major pathogen presence, minor pathogen presence, fungi presence, and cell count thresholds for major pathogens. The animal and the week of sampling were treated as random, while breed, lactation, sampling day, sampling time, and quarter position were treated as fixed. The variance due to the animal was 0.56, 0.65, 0.47, 0.57, 12.57, 2.38, 10.22, 4.87, 3.84, and 4.41 for the variables TLC, N, M, L, major pathogen presence, minor pathogen presence, fungi presence, TLC threshold, N threshold, M threshold, and L threshold respectively. The variance due to the week of sampling obtained was negligible for all variables considered. The effect of sampling day was significant (P < 0.001) for the variables TLC, N, M, L, TLC threshold, N threshold, and M threshold. Lactation was significant for the presence of major pathogens. The effects of breed, sampling day, and quarter position were not significant for any variable. A second model included the additive genetic effect of the cow using pedigree information. Heritability results will be presented.