This study aimed to evaluate the chemical composition and somatic cells count (SCC) of milk from Holstein / Braunvieh (HB) cows of five different grade crossing (genetic groups): Seven ½ blood HB, 14 5/8 blood HB, 16 3/4 blood HB, 86 7/8 blood HB and 28 15/16 blood HB cows were used. Six samplings at seven-day intervals were performed. The quality of milk was assessed according to the fat, protein, lactose and defatted dry solids (DDE), urea and casein levels and somatic cells count. There was no significant difference (P>0.05) in fat content of milk among blood levels; however, the other nutrients in milk differed (P<0.05) among blood levels studied (lactose, DDE, protein, casein, urea). SCC was higher in milk produced by cows of 3/4 blood HB genetic composition. Variations in milk production among genetic groups were also verified, with 1/2 blood HB animals being more productive compared to other genetic compositions. Correlation of genetic composition, production volume, days in lactation, number of deliveries and SCC with the chemical components of milk was found. A positive correlation between genetic composition and DDE and negative with production volume and number of deliveries was found and the SCC value of 13.41% was above the recommended by Brazilian legislation for milk quality.

This study was carried out to determine which system of milkings and sampling methodology presenting chemical composition results safer and lower variation of the results. In Properties I and II, the fresh milk samples were collected in 40 ml flasks containing bronopol in order to compare the chemical composition of the milk, as adopting two sample collection methods, that is two and three milkings. In order to compare the chemical composition of milk and SCC, between two sampling methodologies, in two daily milkings (Property I) and three daily milkings (Property II), fresh milk samples were collected randomly totaling 300 samples on the property I and 312 samples on the property II. The milk chemical components were determined through analytical principle that is based on differential absorption of infrared waves by components of the milk. By analyzing the treatments, experimental design was used entirely randomized and averages were compared by Tukey test using the statistical program SISVAR (statistical analysis system). When evaluating the data obtained through variation coefficient (VC), the occurrence of experimental was observed, since most results showed low variation. However the distinct procedures for sampling of the fresh milk may present varied levels of the chemical composition and SCC, therefore overestimating or underestimating the values of the milk composition, when samples are collected at different times of the day and different number of milkings.