CALIBRATION AND ACCURACY CONTROL IN DETER MINING SOMATIC-CELL COUNT IN COWS MILK BY MEANS OF THE METHOD FOSSOMATIC

The presented work deals with the correspondence of determining the so matic cell count (PSB) in raw and thermally treated milk. Milk samples (n = 10) have been used with PSB from 112 to 721 thousand/ml, the ari thmetic average was 329 +/- 212 and the geometric average 274 thousand /ml. Samples of raw (S) and pasteurized (P) milk were measured non-con served (N) and conserved by means of K2Cr2O7 (D) and bronopol (BSMT) a nd by means of devices (Fossomatic 90, Foss Electric, Denmark) with di fferent levels of discrimination (DL), as well as by means of the dire ct microscopic method (MM). The question of the principle of calibrati ng the method F is being discussed in relation to the existence of sou rces of fluorescence impulses, whether of a parasitic origin (''non-ce ll'', i.e. chemical, biochemical, resp. microbial origin), or the cell (somatic) origin.Further the question of potential oscillations of th e frequency display of parasitic and cell impulses is being discussed from the point of view of fluorescence peak and under the effect of ce rtain factors - cow's individuality, mastitis, thermal treatment of mi lk (Fig. 1 and 6). At all DL levels used, there are - with exceptions - significant differences (P < 0.05) between the samples S and P. The samples P show always higher values: with increasing DL the registered PSB decreased markedly and significantly in samples S and non-signifi cantly in samples P; PSB of samples P decreased significantly with gro wing DL with subthreshold values of the device F used which applies ge nerally for differently preserved samples (Tab. I and III). It follows from the results that thermal denaturing effects during pasteurizatio n increase the colouring ability of cell nuclei and displace the distr ibution of cell impulses, probably increase the frequency and change a lso the distribution of parasitic impulses (Fig. 2 and 6). Differences between the samples D and N and BSMT and N within the samples S were greater and significant in some cases and they were small and predomin antly non-significant between BSMT and D within S, as well as with all preservation methods within the samples P (Tab. II and IV). As expect ed, the differences in the direct microscopical method between the var iants S and P was slight(P > 0.05, Tab. V). In case an optimal thresho ld according to MM and the curve of threshold S (Fig. 2, D, I) = 0.98 was determined for the results mentioned, a distinct intercept change applies for the equation P (Tab. VI, Fig. 3). In opposite, if the thre shold has been defined according to the curve P = 3,90, a distinct cha nge of slope applies for the equation S (Tab. VI, Fig. 4). That is why the average values of PSB divert distinctly and significantly (P < 0. 01 and P < 0.001) from MM (Tab. VI) in both cases. A good closeness of the relationship of determining PSB between MM and F with different D L has been assessed (Tab. VI, r = P < 0.001). A potential adjustment o f DL according to samples P during routine measurement of samples S is , therefore, not appropriate. Fig. 6 shows an attempt to describe chan ges in the distribution of impulses due to pasteurization. As correct can be considered only a delimination, resp. a potential adjustment of DL according to the re suits of MM in samples measured in a routine w ay, i.e. with a similar distribution of the size of fluorescence impul ses. For the control of calibration stability in time (without the aim of adjusting DL), also repeatability of measurement in material with changed distribution of impulses (thermally treated milk) can be used. Lintner et al. (1984) and Szijarto, Barnum (1984) used pasteurized mi lk just only for the control of calibration stability in time and not for a