2.1. Animal Diet and Management
The experiment was conducted at the Laboratory of Goat Rearing, Research Center for Human, Social, and Agricultural Sciences, Federal University of Paraíba, at Campus III, located in Bananeiras—PB, Brazil. The study was analyzed and approved by the Committee on Ethics in the Study and Research of Deontology of the Federal University of Vale do São Francisco (case no. 0007/131014).
Twenty-four multiparous Alpine goats were used, with an average live weight of 38 ± 4 kg and an average lactation of 30 days. These animals were distributed entirely at random. The experimental duration was a 14-day introduction followed by three periods of 16 days giving a total duration of 64 days. The goats were kept in an intensive system and allocated individual cages with 1.26 m2 of space equipped with troughs and water fountains and with floors made of railed wood crates.
The ingredients used for the formulation of the diet had the following composition: buffel grass hay (
Cenchrus ciliaris L.) and a base made of soybean meal, corn bran and mineral 50:50 (concentration:volume) (
Table 1). The experimental rations were formulated to meet the milk production requirements of 2 kg/day [
12]. The ration was available in two daily servings after milking at 7:00 and 15:00 h. The dry matter intake was controlled according to the diet provided and the leftovers, which were adjusted daily to allow 20% of leftovers.
The treatments corresponded to increasing levels of total dissolved solids in the water provided to the animals, which was created using sodium chloride (NaCl), distributed into four electrical conductivity levels: 1.0, 5.0, 9.0, and 13.0 dS m
‒1. The conductivity and temperature of each treatment were read daily using a conductivity meter (Digimed, São Paulo, Brazil) with an allowance of a difference of 5% in the limit of each treatment. To convert the electrical conductivity of the water to parts per million (ppm), or milligrams per liter of total dissolved solids (mg L
−1 TDS), we multiplied 1 dS m
−1 by 640 mg L
−1 [
15]. The treatments were converted to milligrams per liter in the following proportions: T1—640, T2—3188, T3—5740, and T4—8326 mg L
‒1 TDS (
Table 2). Water troughs were washed once weekly to prevent NaCl accumulating on the borders, which might affect the salt concentration of each treatment.
During the entire experiment, samples of water from each treatment were collected weekly, conditioned in labeled plastic bags and subsequently frozen until they were analyzed. The samples were sent to the Geo-Environmental Laboratory of Embrapa Semiarid, where chemical analyses for bicarbonate, chlorides, calcium, magnesium, potassium, and sodium were conducted and where the electrical conductivity of the collected water samples was also measured.
Sodium and potassium analyses were performed using the flame photometry method, where the water was diluted 0, 10, 100, and 1000 times, and each sample had four replicates; then, each sample was read after removal of all impurities. The equipment was washed with distilled water. For titration of the chlorides, the Mohr method, which relies on the titration of the water sample with silver nitrate using potassium chromate as the endpoint indicator, was used. To titrate the carbonates and bicarbonates, the volumetric or titrimetric alkaline method was used, using sulfuric acid to determine the levels of carbonates and bicarbonates and the indicators phenolphthalein and methyl orange to titrate the carbonates and the bicarbonates, respectively [
16]. The calcium and magnesium analyses were performed through complexometric titration, i.e., by using EDTA (Ethylenediaminetetraacetic acid) to complex calcium and magnesium at alkaline pH levels. For the analysis of calcium, murexide was used as an indicator after the sum of the Ca + Mg was determined using Eriochrome Black T, and the Mg concentration was determined by the difference [
17].
2.2. Milk Sampling and Analysis
For the sampling, the udders of the animals were sanitized before the milking. The milk samples were collected in three distinct 16-day periods, with collections taken on the 15th and 16th day of each experimental period by manual milking carried out at 6:00 and 14:00 h in the milking room. Aliquots of each animal’s milk were obtained as compound samples proportional to the milk production for each milking shift (70% of the milk collected in the morning and 30% in the afternoon). Polyethylene bottles of 300 mL were used for the physicochemical analysis, and 100 mL bottles were used for the analysis of minerals; the milk samples were kept under freezing temperatures (−18 °C) for further analysis.
The physicochemical analyses were performed at the Laboratory of Bromatology of the Health Sciences Center of the UFPB at the end of the experimental period. The levels of density, acidity, protein, lipids, lactose and ash were determined according to the methodology described by the AOAC [
18]. For moisture, the percentage of reduction of the total dry extract (TDE) was calculated, and the defatted dry extract (DDE) was determined by calculating the TDE minus the percentage of fat [
18].
The mineral analyses were carried out at the Laboratory of Chemical Analysis of Food of the UFPB. For the analysis of phosphorus and chlorides, the ash solutions were extracted and then stored in 100 mL glass pots. The phosphorus analysis was obtained by the colorimetric method, which is based on the reaction between the phosphorus of the mineral solution and ammonium molybdate, producing ammonium phosphomolybdate. The amount of phosphorus was determined by measuring the intensity of the blue color that is produced by the formation of the phosphomolybdate. For this analysis, a spectrophotometer (Coleman, model 33D, Santo André, São Paulo, Brazil) with a wavelength of 650 nm was used [
19]. For the titration of chlorides, the Mohr method was used [
18]. This test was based on the precipitation of chlorides in the form of silver chloride at pH 8.3 and in the presence of potassium chromate as an indicator. The end of the reaction was indicated by the formation of a brick-red precipitate of potassium chromate, and the concentration of chloride and sodium present in the sample was obtained through the following equation:
The determination of calcium was conducted by the titrimetric method based on the complexation reaction of EDTA and calcium. The potassium content was obtained by flame photometry using the MERCK
® brand potassium standard and a flame photometer (Tecnow 7000, São Paulo, Brazil) [
18].
2.3. Sensory Analysis
For sensory analysis, samples of milk were collected from each treatment, and five sub-samples were prepared of one liter each. The milk was pasteurized, and sensorial analysis was made two days after cold storage. The sensory analysis was performed in individual booths in controlled environmental conditions at a temperature of around 23 °C [
20].
The sensory evaluation was carried out with an internal panel consisting of 20 evaluators (aged 20 to 40 years), and 11 evaluators were selected and trained according to the methodology of Noronha [
21]. Said subjects were selected for their sensory ability and trained for descriptive analysis according to the standard flavor profile guidelines set by ISO 6564:1985. Panel training sessions were performed to familiarize the assessors with the language and goat milk. The samples were described using the Quantitative Descriptive Analysis (QDA) technique [
20]. The QDA test was administered using a five-point scale ranging from 1 (little) to 5 (very much) regarding the following attributes: external odor aspects (overall intensity, goat’s milk odor (butter/rancid and aromatic)), aftertaste (overall intensity and persistence), overall acceptability (disliked very much or liked very much), and flavor (overall intensity, butter/rancid, goat, aromatic) (
Table 3).
The three testing sessions (trained panel and consumer testing) were conducted in individual booths under conditions in accordance with ISO 8589 (facilities) and ISO11037 (lighting). Each assessor was served four samples (each pot coming from a treatment) coded with three-digit random numbers, served immediately after being taken out of refrigerated storage with lids containing 50 mL aliquots of the duly pasteurized milk. Assessors were asked to use low-salt crackers and water to clean their palates between the assessed samples.