highly significantp < 0.001
All the fixed effects analyzed—month (year), flock size, flock size × THI, THI (year) and THI (month)—were highly significant ( p < 0.001) for all milk composition traits, with the exception of log 10 SCC ( Table 3 ).
All the fixed effects analyzed—month (year), flock size, flock size × THI, THI (year) and THI (month)—were highly significant ( p < 0.001) for all milk composition traits, with the exception of log 10 SCC ( Table 3 ).
A decreasing trend for bulk milk protein was observed for higher THI MAX classes −2d before sampling (from 62 to 74), but this decline was minor for bulk milk of small flocks, whilst it was greater for larger farms.
Somatic cell count did not have a clear trend, remaining nearly constant, whereas the total bacterial count showed higher inter-annual variability in June and July, conversely to what has been observed for dairy cattle, as dairy sheep farming is characterized by the high seasonality of lambing ( Figure 2 ).