Dynamics of Variations of Calibration Line Setups at Determination of Somatic-Cell Counts on Fossomatic Apparatus with Various Discriminator Levels

Veterinarni Medicina 38, 1993, 497-509

Analysis of the relationship between direct and indirect method of determination of somatic cell counts (SCC) in samples of raw milk (R) and pasteurized milk (P) has turned out necessary. A Fossomatic method is probably the most widespread and the most precise method of SCC determination all the world round used for the purposes of payments for milk. We tried to find an explanation to the fact that Fossomatic (F) apparatuses ''identical'' with each other in readings of control P samples may show clearly different readings of SCC determination in R milk. The results can help to elucidate some factually observed phenomena in SCC readings at apparatus checks. Lintner et al. (1984) and Szijarto and Barnum (1984) used the system of apparatus SCC determination checks by means of control P samples. Such a control was applied in time, but most probably not to adjustment of the apparatus discriminator level. Identical samples of R and P milk as for SCC were analyzed on a Fossomatic 90 apparatus (Foss Electric, Denmark) at various discriminator levels (DL), and the results were confronted with a reference procedure of direct microscopical method (MM). The samples analyzed on the apparatus were stored for 24 hours at 4-degrees-C and they were either unpreserved (U) or preserved with K2Cr2O7 (D) or with bronopol (BSMT). The results in Tabs. II and III and in Fig. 1 show that the setups DL = 1.0 - 1.5 can be taken as threshold (optimum) for R samples preserved with BSMT and D, DL < 1.0 as underthreshold and DL > 1.5 as suprathreshold. DL setup determined in a certificate for the given apparatus is 1.4. The lower threshold value would apply to U sample (Tab. I), while quite different thresholds would apply to P samples (Tabs. I, II, III). The values of standard deviations of the mean differences (sd) between F and MM readings of both R and P samples, and between F, R readings and F, P readings (Tabs. I, II, II), logically increase in general with departures of the slope of respective regression and calibration lines from the ideal value 1. The large Fossomatic range of DL (Fig. 1

FULL PAPER on request
« back