Milkoscan 133b

At the beginning of the trial, pasture samples were collected along transects according to methods from Freschi et al. (2015 (link)Freschi et al. ( , 2016 (link)Freschi et al. ( , 2017) ) (link) and Rizzardini et al. (2019) . Grass samples were dried in a stove, homogenized, and analyzed. Samples of the pelleted feeds and dried pasture were ground in a hammer mill with a 1-mm screen and analyzed using the following AOAC International (2004) procedures: DM (method 934.01), fat (method 920.39), ash (method 942.05), CP (method 954.01), crude fiber (method 945.18), ADF and ADL (method 973.18), and NDF (method 2002.04).
Individual milk samples were collected twice per day (at 0700 and 1800 h), stored at 4°C, and immediately transported to the laboratory for the analysis of fat, protein, lactose, and total solids with an infrared milk analyzer (Milkoscan 133-B, Foss Electric, Hillerod, Denmark) previously standardized for goat milk. Cheese samples were homogenized and lyophilized. Fat content was determined by the Soxhlet method (AOCS, 1996) and expressed as fat in DM. Total nitrogen was determined by the Kjeldahl method (AOAC, 1995) and expressed as protein content (nitrogen content × 6.38). In milk and cheese samples, ash content was detected after burning a lyophilized sample in a muffle furnace at 550°C for 5 h.

To determine the DM content of feed, refusals and feces, samples were dried in a forced-air oven (60°C, 48 h), and subsequently ground in a Wiley mill 3 mm diameter (Arthur H. Thomas, Philadelphia, PA). Organic matter was determined by incineration (550°C for 3 h), fat, total nitrogen (AOAC International, 2005) , and CP content. The NDF and ADF were determined using the ANKOM technique (Van Soest et al., 1991) with α-amylase and uncorrected for ash. Milk samples were analyzed using a MilkoScan 133B (Foss Electric, Hillerød, Denmark) to obtain the values of protein, fat, TS, and SNF. Fatty acids of the dietary components were separated by the Soxhlet method (AOAC International, 2005) .

Pasteurized bovine, caprine, and ovine bulk milks were lyophilized (lyophilization cycle -50°C for 24 h; Lio5P, 5 Pascal s.r.l., Trezzano, Milan, Italy) and the resulting lyophilized powders stored at 4°C. Bovine, caprine, and ovine milk samples were analyzed for chemical composition using an infrared spectrophotometer (Milko Scan 133B; Foss Electric, Hillerød, Denmark) . Casein and whey protein fractions were obtained according to IDF (1993) .

Residual DM concentration of dried and ground samples was determined in a forced-air oven at 105°C for 16 h. Silage DM concentration was corrected for the loss of volatile compounds as described by Huida et al. (1986) . Ash concentration was determined in a muffle furnace according to AOAC method 942.05 (AOAC International, 2012) . Nitrogen concentration in feed samples was determined using a Dumas-type elemental N analyzer (Leco FP-428, Leco Corporation, St. Joseph, MI). The NDF concentration was analyzed in the presence of Na 2 SO 3 , as described by Van Soest et al. (1991) , using an Ankom 220 Fiber Analyzer (Ankom Technology, Macedon, NY). Heat-stable α-amylase was used for starch containing samples. Indigestible NDF (iNDF) in feeds and feces was determined based on 12-d in situ incubations in the rumen of dairy cows as described by Ahvenjärvi et al. (2006) . Grass silage OM digestibility was determined based on pepsin-cellulase OM solubility as described by Nousiainen et al. (2003) and modified by Huhtanen et al. (2006b) . Volatile fatty acid concentration in grass silage was determined by GC as described by Huhtanen et al. (1998) . Milk samples were analyzed for DM, protein, fat, and lactose concentration using an infrared milk analyzer (MilkoScan 133B, Foss Electric, Hillerød, Denmark) .