Abstract

Background:
The early postpartum period is characterised by competing physiological demands that substantially increase the risk of metabolic disease and intramammary infection in dairy cows.

Aim:
This study investigated the temporal behaviour of milk electrical conductivity (MEC) and milk fat content (MFC) in 17 Algerian smallholder dairy cows over 30 consecutive days commencing at day 30 postpartum, and quantified the predictive relationship between MEC and MFC by linear regression.

Methods:
Nine Montbéliarde and eight Prim'Holstein cows were monitored at a semi-intensive farm in Chlef wilaya, Algeria. MFC was determined by the Gerber butyrometric method and MEC by portable conductimetry. Daily measurements were grouped into four weekly periods. Descriptive statistics, Spearman rank correlation, and ordinary least-squares (OLS) linear regression with heteroscedasticity-consistent (HC3) standard errors were applied to 526 individual observations.

Results:
MFC did not differ significantly across weekly periods (W1: 16.82 ± 14.39 g/L; W2: 15.89 ± 7.80 g/L; W3: 16.61 ± 8.87 g/L; W4: 18.45 ± 10.08 g/L; all p > 0.05). MEC was persistently elevated above the healthy reference range (9.38–10.70 mS/cm). The Breusch–Pagan test detected significant heteroscedasticity (LM = 115.73, p < 0.001). The HC3-corrected model confirmed a statistically significant negative association between MEC and MFC (slope = −3.722 g/L per mS/cm; SE = 0.692; t = −5.38; p < 0.001; R² = 0.249).

Conclusion:
MEC and MFC reflect distinct but partially overlapping physiological processes. Their combined use is supported as a complementary low-cost screening strategy for udder health surveillance in resource-limited dairy production systems, pending validation against somatic cell count data.

Key words: Heteroscedasticity; Linear regression; Milk electrical conductivity; Milk fat content; Subclinical mastitis.