For sheep farmers it is very important to know the health status of the udder. Increasing SCC leads to a significant reduction in daily milk production, decrease in lactose and a moderate increase in fat and protein (Caria et al., 2016; Tančin et al., 2017; Baranovič et al., 2018) however, it significantly aggravates the coagulation properties of milk (Abdelgawad et al., 2016). Measuring the electrical conductivity (EC) of milk during milking has been studied in cattle as a low-cost mastitis detection method that can be easily automated (Romero et al., 2017). Milk normally has an EC of between 4.0 and 6.0 mS.cm^-1 (Ferrero, Valledor and Campo, 2014), but bacterial infection of the udder results in an increase in Na⁺ and Cl^- and decreases in the K+ levels (Kitchen, 1981), which causes an increase in EC. This is widely used as a method of monitoring mastitis infections. When measured conductivity is in extreme values (6.5 – 13.00 mS.cm^-1) at 18 °C, this indicates mastitis (Ferrero, Valledor and Campo, 2014). Caria et al. (2016) achieved a sensitivity of 73.08% and a specificity of 75.46% in their study, with an EC threshold of 4.84 mS.cm^-1 for sheep milk. There are only a few reports that have been published about the effect of mastitis on the conductivity of sheep's milk. This led us to a decision to evaluate the relationship between SCC and EC of milk in sheep reared in Slovakia as factors for monitoring subclinical mastitis on the basis of a bacteriological examination of udder health.

Samples from 590 udder halves of 295 machine milking ewes of different breeds from eight farms were collected during evening milking. Milk samples were collected aseptically after cleaning the teats, especially teat-ends with antibacterial wipes (GAMA Healthcare Ltd, UK). Sampling always started with the right udder half, the first two strips were placed separately, next 10 mL were used for EC measurement with a handheld conductometer Milk Checker N-4L (Oriental Instruments Co., Ltd., Japan) with compensation the measured EC on a standard temperature of 25 °C, 1 mL was aseptically gathered into sterile test tube for cytobacteriological analysis and an additional sample of 50 mL was taken for somatic cell count and a basic components analysis. Immediately after removal, the milk sample was stored in a portable refrigerator at 5 – 15 ° C. The samples were transported to the laboratory and refrigerated at 4 °C. Milk samples (inoculum 10 μl) streaked onto selective culture medium PM test (LabMediaServis s.r.o., CZ) were incubated at 37 °C for 24 h. Isolated strains of pathogens were then verified by typing with BBL Crystal® (Becton, Dickinson & Co., New Jersey, USA).

Somatic cell count was determined using a Somacount 150 (Bentley Instruments, Inc., Chaska, Minnesota, USA), milk composition was determined by MilkoScan FT120 (Foss, Hillerød, Denmark).

After the pathogen analysis, we found that 175 animals were free of the pathogen in the udder and in 120 animals the pathogen was present in at least one half of the udder. 76 animals (25.8%) from the “free of the pathogen” category had SCC < 1 × 10⁵ and EC ranging from 0.0 to 0.4. In total, 175 udder halves (29.7%) were infected, from that 55 animals were infected in both halves and 65 animals with only one half. In 14 samples (2.4%) major pathogens were detected (Staphylococcus aureus (5 samples), Streptococcus agalactiae).

The presence of the pathogen had an significant effect (F_(2;587) = 37.06; p <0.001) on the increase in electrical conductivity (Table 1), no significant differences were found between the minor and major pathogens. EC of the infected glands (n = 175) without considering the type of pathogen was (Mean ± SD) 5.3389 ± 1.2836 mS.cm^-1.

Similarly as above, the presence of the pathogens had an significant effect (F_(2;587) = 155.61; p <0.001) on the increase in LogSCC (Table 2), but the major pathogens increased the LogSCC level significantly higher than minor pathogens. This goes along with the results of other studies (Linage et al., 2017; Gonzalo, 2018).

The correlation between SCC and EC for all animals (Table 3) was higher (a moderate relationship) than that found by Caria et al. (2016) (r = 0.306) or Romero et al. (2017) (r = 0.33), but corresponds to the data reported by Peris et al. (1991). The strongest correlation was, similarly to Romero et al. (2017) (r = 0.25) in SCC ≥ 6 × 10⁵ class. This correlation may indicate that EC may be used in this class for mastitis detection. Also, differences in the EC were statistically significant (F(3;586) = 86.67; p <0.001) only between the SCC ≥ 6 × 10⁵ class and the other classes. When ordering EC according to Caria et al. (2016) significant differences between means (F_(1;588) = 261.08; p <0.001) were found. Lower value of EC in class SCC < 7 × 10⁵ as in classes 2 × 10⁵ ≤ SCC < 4 × 10⁵ or 4 × 10⁵ ≤ SCC < 6 × 10⁵ in classification above (Table 3) was caused by counting a greater number of cases from the SCC < 2 × 10⁵ class to this class.

The lactose and protein content was significantly affected by the presence of pathogens (Table 4) but without significant differences between minor and major pathogens groups. In the SCC class classification (Table 5), we found significant differences in lactose content only between the SCC ≥ 6 × 10⁵ class and the other classes. In the classifications according to Caria et al. (2016) and Barth, Burow and Knappstein (2008), the differences between the classes were statistically significant. However, there are no differences between classes in protein content.

The negative correlation between LogSCC and lactose content (Table 6) corresponds to findings from other authors (Scharch, Süß, and Fahr, 2000; Olechnowicz et al., 2009; Caria et al., 2016), but our values are lower than those reported by Olechnowicz et al. (2009) and Scharch, Süß, and Fahr (2000). The electrical conductivity showed a stronger correlation with the LogSCC than lactose and protein content reported, although it is still only a weak relationship.

We can argue that measuring the electrical conductivity of sheep milk may be a possible alternative for mastitis detection in sheep. EC can be useful in detecting animals with level of SSC greater than 6 × 10⁵ cells.mL^-1. But we can not estimate a threshold for healthy animals. Perhaps, if we obtain more data from animals in the 2 × 10⁵ ≤ SCC < 4 × 10⁵ and 4 × 10⁵ ≤ SCC <6 × 10⁵ (cells.mL^-1) categories, it will be possible to specify the threshold in the future. However, since electrical conductivity is influenced by several factors (Romero et al., 2017), it would be more appropriate to think about multiple individual assessments in milking parlours rather than using a portable device for mastitis detection.