The use of spectral photometry for detection of mastitis milk

By using automatic milking system reliably sensors are necessary for detecting altered milk. Previous own investigations resulted a clear detection of blood and colostrums in bovine milk by using spectral photometry (Wiedemann, 2003). The objective of this trial was to prove if this technique can also detect mastitis milk.
This study was carried out on the research farm “Hirschau” (TU Munich- Weihenstephan) using a VMS® that milked 45 cows (Red Holstein x Simmental). The research was split up in two experiments: In part A the quarter specific milk fractions (portions of 500 ml) of 40 animals for two days (with and without secretion disorders) were collected. Because of the influence of teat cleaning to milk ejection samples were taken both with and without previous udder cleaning. In part B the quarter individual foremilk samples from 16 cows (with and without secretion disorders) were collected and analysed over a period of nine days (280 milkings). The udder health was classified by using the SCC (every milking) and the cyto-bacterial analysis of foremilk sample (first and last day). The spectral photometric measurement was done by using the instrument spectro-color (DrLange, Germany). Before warming all milk samples to an uniform temperature of 37°C to do the photometric measurement aliquots of 2 ml were frozen for further analysis. Also they were examined for fat, protein, lactose and SCC in a standard test.
Because of the high correlation between milk fat content and milk colour (Bergann & Schick ,1999) it was important, to find out which bands in the visible spectrum are mainly affected by the fat content. The analysis showed that the fat content was reflected in particular in the red wavelengths (620-700 nm). In contrast the band between 400 and 520 nm is only slightly influenced by fat. Therefore the analysis of milk samples regarding to detect altered milk were based on the band between 400 and 620 nm.
The investigations in part A resulted that the foremilk (<500 ml) of diseased quarters reduced the spectral reflectance in the band 400 to 520 nm in comparison to the normal producing quarters. In the ongoing milk fractions (>500 ml) no differences in spectral reflectance between healthy and diseased quarter were found.
Based on the previous results, in part B only the first 300 ml of each foremilk (without pre-stimulation) were taken for analysis. The mean of average reflectance in the band 400-520 nm from milk of healthy quarters was 44.4 % (SEM = 0.042). In addition a significant difference between the average reflectance of milk from diseased quarters (i.e. more than 400,000 cells/ml or positive bacterial findings) and healthy quarters (P<0.05) was observed. At the moment it is possible to detect more than 70 % of all diseased quarters by using the method of inter quarter comparison (specificity = 85 %). Currently the aliquots are analysed regarding to electrical conductivity. A comparison of sensitivity and specificity of both parameters will be presented in the final paper.
The momentary results have shown that the spectral photometry is another useful tool for detecting mastitis milk if the importance of the foremilk sample and the blue band (400 to 520 nm) are taken into account.

Corresponding author:

Martin Wiedemann Landtechnischer Verein in Bayern e.V.
Vöttinger Straße 36
85356 Freising Germany
Phone +49 8161 71-5254
Fax +49 8161 71-4048

© Animal Sciences Group - Wageningen UR. Last update: 20-02-2008 10:03. Mail any questions and remarks to: webmaster.asg@wur.nl