A representative raw milk sampling should be carried out with a view to a reliable ascertainment several quality criteria. The aim was to determine if the agitation, place of milk gathering and the time of cooling in the expansion tank interferes in milk quality. Pilot assay was developed at the Sistema de Pesquisa e Desenvolvimento em Pecuária Leiteira in Embrapa Clima Temperado. All 30 Jersey milk was evaluate (milk tank), on which composition analyzes (fat, protein, lactose and total solids), Somatic Cell Count (SCC) and Total Bacterial Count (TBC), as well as analyzes of Unstable Non Acid Milk (UNAM), the alcohol test and the titratable acidity test (Dornic). The main treatments were defined as Automatic Shaking (AA), with samples collected only with tank automatic shaking; and Control Agitation (AC), with samples collected after horizontal shaking (paddle) for five minutes and vertical shaking using a hand shaker. The secondary treatments were defined by the place where the collections were made, below (B) or above (C). The content of fat, total solids and SCC were the parameters that showed the greatest variability in all treatments. The results of alcohol and titratable acidity tests were lower in AA treatment than in AC treatment. The treatment that showed the lower standard deviations in all the evaluated variables was the AC, C, which indicates greater precision of the results.