The experiment was conducted on Malhada das Cabaças Farm (14∘13′ S, 42∘46′51′′ W; 525 a.s.l.), where the average annual precipitation is 553.5 mm, and at the Guanambi slaughterhouse (FGI). Both locations are situated in Guanambi – BA, Brazil. All animal procedures were previously approved by the Ethics Committee at the State University of Southwestern Bahia, Brazil (approval no. 042/2013).

Thirty Santa Inês cull ewes with an average initial live weight of 44.2 ± 5.2 kg and an average age of 50 months were distributed in a completely randomized design. Sanitary measures against endo- and ectoparasites were taken and the ewes were housed in collective sawdust-bedded stalls (2.0 × 5.0 m) at the rate of six animals per stall. The stalls were equipped with collective polyethylene feeders (0.5 m/animal) and drinkers, which were cleaned daily.

Five high-concentrate experimental diets were prepared. The diets were formulated with 100 % concentrate, without roughage. The feed was supplied twice daily (half at 07:00 and the other half at 15:00) and adjusted daily to allow orts around 15 % of the total supplied, characterizing ad libitum intake. The diets were formulated as recommended by NRC (2007) and supplied so as to meet the daily requirements for an average daily gain of 300 g. Treatments were represented by the physical form of cottonseed (whole or crushed) and the inclusion of calcium lignosulfonate (Table 1) in the diet as follows: WCS – whole cottonseed; CCS – crushed cottonseed; WCSL – whole cottonseed with lignosulfonate (100 g/kg of cottonseed as fed); CCSL – crushed cottonseed with lignosulfonate (100 g/kg of cottonseed as fed); and CON – control diet, without cottonseed.

The experiment lasted 54 d, of which the first 12 were used for the animals to adapt to the collective stalls and dietary management. In the diets, the concentrate was added replacing hay of Digitaria decumbens Stent. cv. Transvala gradually, every 4 d, at the following ratios: 60:40 (day 1); 80:20 (day 5); and 100:0 (day 9). The first evaluation took place on the 13th day, when the total diet had 100 % concentrate, and the evaluation period was 42 d, of which the last 5 d were used for data collection.

Prior to slaughter, the ewes were weighed to determine slaughter weight (SW) and calculate the average daily weight gain (ADG).

Then, in vivo measurements of subcutaneous fat thickness (SFT) and loin-eye area (LEA) were performed using an ultrasound device (ALOKA^® SSD 500 v) with a 3.5 MHz linear transducer, a 12 cm acoustic probe, and a silicone coupler. After images were captured, the ewes were shorn and clipped in the region between the 12th and 13th thoracic vertebrae, on the left side. The best image of the loin-eye area was captured by measuring it from the medial side of the Longissimus muscle to its lateral side of the midline. The length (A, cm) and maximum depth (B, cm) of the muscle were measured to calculate LEA (in cm2) using the following ellipse formula: LEA = A/2 × B/2 × π (Silva Sobrinho et al., 2003). Using the same image, SFT was also determined (in mm) over the said muscle area.

Next, the following in vivo biometric measurements were taken with the animals standing on a flat surface: rump height, withers height, body length, chest width, chest girth, rump width, shank circumference, leg length, and shank length. Length and circumferences measurements were taken using a tape measure, whereas the width and depth variables were measured with a measuring stick. Body compactness was determined as the ratio between slaughter weight and body length.

At the time of slaughter, the animals were transferred to the slaughterhouse located in Guanambi – BA, Brazil, where they were slaughtered under veterinary inspection following the current norms established by the Regulation for Industrial and Sanitary Inspection of Products (Brazil, 2000).

After bleeding and skinning, the carcasses were eviscerated and the non-carcass components were separated into rumen–reticulum, omasum–abomasum, small and large intestines, heart, liver, kidneys, lungs–trachea–esophagus, tongue, blood, and external body components (head, feet, and skin). The gastrointestinal tract (GIT) was initially weighed full. Then it was emptied, washed and weighed again to determine the GIT content and calculate empty body weight (EBW = SW – GIT content). Non-carcass components were estimated relative to EBW.

Subsequently, the hot carcass weight (HCW) was determined and the hot carcass yield (HCY) was calculated by the following equation: HCY = HCW/SW × 100 (Cézar and Souza, 2007).

After 24 h of refrigeration in a cold room at 4 ∘C, the carcasses were weighed to determine cold carcass weight (CCW) and assessed subjectively for conformation [1 = Poor (concave); 2 = Regular (sub-concave); 3 = Good (straight); 4 = Very good (sub-convex); 5 = Excellent (convex)] and degree of fatness (1 = Emaciated; 2 =  Lean; 3 = Medium; 4 = Fat; 5 = Obese). The pelvic–renal fat, or visceral fat, was assessed by assigning scores from 1 to 3, where 1 = Low (both kidneys uncovered); 2 = Normal (one kidney covered); and 3 = Excessive (both kidneys covered).

Following the subjective assessment of pelvic–renal fat, the kidneys and pelvic–renal fat were removed and their weights were recorded and subtracted from the hot and cold carcass weights. Next, the cold carcass yield (CCY) and chilling loss (CL) were calculated using the following formulas: CCY = CCW/SW × 100 and CL = (HCW – CCW) / HCW × 100.

Whole carcass length, internal carcass length, shank circumference, shank length, chest depth, chest width, chest girth, chest circumference, rump circumference, and rump width were measured. Carcass capacity was determined as the ratio between cold carcass weight and internal carcass length.

After the morphological evaluation, the carcasses were sectioned lengthwise, along the spine with an electric saw, which resulted in the separation of the right and left half-carcass. The left half-carcass was used for direct (on-carcass) measurements of LEA and SFT. Loin-eye area was determined by making a cross section between the 12th and 13th thoracic vertebrae and tracing the outline of the muscle on acetate sheet, in correspondence with the cranial portion of the loin. The maximum distance and depth of the muscle were measured with a ruler and calculated from the formula proposed by Silva Sobrinho et al. (2003), as previously mentioned in the ultrasound evaluation of LEA. The subcutaneous fat thickness in the carcass was measured (in mm) using a digital caliper at a 3/4 distance from the medial side of the Longissimus muscle, towards the side of the spinal process.

The effects of dietary treatments on the measured parameters were analyzed according to the following model: Y^ij=μ+Di+εij, where Y^ij is the measurement j of diet i, μ is the overall mean, Di is the fixed effect of diet i (i = C, WCS, CCS or CCSL), and εij is the random error term.

Additionally, the following contrasts were used to compare the effects of the different diets:

Contrast 1: cottonseed effect {CON vs. (WCS + CCS + WCSL + CCSL)}

All statistical procedures were performed using the PROC GLM procedure of the SAS software (SAS, 2014), adopting 0.05 as the critical probability of type-I error.

Neither the cottonseed processing method nor the inclusion of calcium lignosulfonate compromised the ADG or SW of the feedlot ewes. The control diet in comparison to the others (CON vs. other diets), cottonseed processing method (WCS + WCSL vs. CCS + CCSL), and calcium lignosulfonate inclusion in the high-concentrate diets (WCS + CCS vs. WCSL + CCSL) also did not influence (P>0.05) ADG (0.195 g) or SW (51.74 kg) (Table 2).

There were no differences (P>0.05) in body compactness (1.07 kg/cm), LEA (8.0 cm2), or SFT (1.70 mm) evaluated in vivo by ultrasound or in muscularity relative to carcass weight and final live weight (Table 2).

In comparison with the control diet without cottonseed, the inclusion of calcium lignosulfonate (WCS + CCS vs. WCSL + CCSL) and the cottonseed processing form (WCS + WCSL vs. CCS + CCSL) did not affect (P>0.05) the quantitative traits of HCW and HCY (24.8 kg and 47.8 %), CCW and CCY (24.2 kg and 46.8 %), or chilling loss (2.1 %) (Table 3). The treatments also did not affect (P>0.05) the subjective measurements of conformation (1–5), degree of fatness (1–5), or visceral fat (1–3) evaluated on the carcass, which averaged 2.6, 3.10, and 2.6 points, respectively (Table 3). There was also no effect of calcium lignosulfonate inclusion or the cottonseed processing method on LEA or SFT measured on the carcass, whose mean values were found to be 15.6 cm2 and 0.88 mm, respectively (Table 3).

The morphometric measurements evaluated on the carcass did not differ (P>0.05) between the treatment groups. The mean values of the evaluated variables were as follows: total carcass length: 56.8 cm; internal carcass length: 65.9 cm; shank length: 44.1 cm; chest depth: 26.1 cm; chest width: 29.6 cm; chest girth: 19.2 cm; rump circumference: 58.0 cm; and rump width: 19.1 cm (Table 4).

In the analysis of non-carcass components, neither calcium lignosulfonate inclusion nor the cottonseed processing method influenced (P>0.05) the full or empty GIT, rumen–reticulum, omasum–abomasum and empty intestines, or inguinal, pelvic, omental, renal, and total fats. However, calcium lignosulfonate inclusion in the diet increased the proportions of udder (P=0.0453) and liver (P=0.0458) relative to EBW (Table 5).

The ADG and SW of the feedlot ewes evaluated in this study were considerably higher than the 0.034 and 45.21 kg, respectively, reported by Souza et al. (2010) in Barriga Preta feedlot ewes fed high-concentrate diets, demonstrating the higher potential of Santa Inês ewes for performance and meat production.

Rump and withers heights were similar (P>0.05) between the treatments, averaging 72.3 and 73.2 cm, respectively (Table 4). This indicates that the animals had a uniform body height, which is a very important trait to accurately determine other biometric measurements in vivo and on the carcass (Table 5) (Pinheiro and Jorge, 2010; Pinheiro et al., 2015). Supporting these assumptions, the results of our study were practically identical to those observed by Pinheiro and Jorge (2009), who found 69.5 and 68.0 cm for rump and withers heights, respectively, in Santa Inês sheep with the same SW (43 kg).

The morphometric measurements of the cull ewes did not differ between the treatments (P>0.05). As the tested diets did not influence these measurements, it is inferred that there was a common body condition pattern, considering that a selection process was performed to acquire animals at approximate ages, which was achieved by evaluating their dental arch. According to Garcia et al. (2003), another explanation is that cull animals are mostly at a stage when their physiological growth has already concluded and their body condition will be a determining factor for their biometrics. Yet another factor that may explain the similar results is that ADG and SW were not affected. This supports the affirmation of Marques et al. (2008), who asserted that biometric measurements are little influenced by nutritional management provided that the animals are slaughtered with the same weight.

The average body compactness indices found in the ewes of the present study were higher than the 0.61 described by Pinheiro and Jorge (2010). This may have been due to the better nutritional composition of the diet supplied in this study (17.38 vs. 11.81 % CP), which resulted in higher weight gain (SW = 51.74 vs. 43.34 kg; Pinheiro et al., 2015) and, consequently, greater muscle deposition in the carcass (LEA = 15.58 vs. 11.92 cm2; Pinheiro et al., 2009). Body compactness is an important parameter, as it can influence the product's selling price and its visual perception. In addition, it favors the consumption of this meat, since higher body compactness indices translate into higher proportions of muscles and fat in the animal (Costa Júnior et al., 2006).

The LEA and SFT results obtained by ultrasound may be attributed to the fact that the animals were slaughtered at similar body weights. This finding corroborates the following inference drawn by Osório (2002): when carcasses have similar weights and amounts of fat, almost all body regions have similar proportions, irrespective of the breed. Loin-eye area and SFT are traits used as indicators of muscularity and fatness, respectively (Souza Júnior et al., 2013; Al-Jammas et al., 2016).

In the current study, HCY and CCY were higher than the 45.0 % and 44.7 % reported by Pinheiro et al. (2009) and 45.2 % and 43.7 % described by Pelegrini et al. (2008) in Santa Inês and ideal feedlot cull ewes, respectively. These results confirm the presuppositions that when fed diets of better nutritional quality and specialized in and having potential for meat production, feedlot ewes exhibit higher performance and, consequently, higher carcass yields.

The morphometric traits of a carcass vary according to breed, sex, and diet (Murta et al., 2009). Therefore, the similarity found between the treatment groups in this study can be explained by the similar composition of the diets, since the animals shared the same sex, breed, age, and live weight. In addition, growth in adult animals is limited, which results in similar body measurements (Garcia et al., 2003).

According to Pinheiro et al. (2009), in sheep, changes in skin and udder weight are directly related to the animal's age and physiological stage. The liver and renal fat, on the other hand, are more affected by compensatory gain. Thus, part of the weight changes is the result of the recovery of the liver's metabolic activity and, consequently, the increased weight of these organs. This may explain the heavier weight of the udder and liver in the animals that received diets with inclusion of calcium lignosulfonate, which exhibited a higher absolute weight at slaughter (Table 4).

Calcium lignosulfonate improves nitrogen metabolism by increasing recycling (Cirne et al., 2020). This effect probably contributes to increasing liver activity and, as a consequence, animal size. Additionally, this enhancement can increase the number of cells in the udder or its activity by increasing its size.

The cottonseed-containing diets, whose neutral detergent fiber (NDF) content was higher due to the ingredient, were expected to result in a larger GIT content and, consequently, a higher volume of its components, given the longer residence time of this feed in the rumen. Alves et al. (2013) and Medeiros et al. (2008) observed this phenomenon in their experiment, where they describe higher GIT contents in Santa Inês and Morada Nova sheep, respectively, which were fed a high-fiber diet. Nonetheless, the increase in GIT content is thought to be greater when diets rich in roughage-derived fiber are used. Moreover, because the animals received diets with a similar physicochemical composition and were subjected to the same pre-slaughter fasting time, this response may be considered biologically normal.