Forage particle size, rather than dietary iNDF concentration, affects chewing activity and lactation performance in lactating dairy goats Authors: Mostafa Soufizadeh, Rasoul Pirmohammadi, Younes Ali Alijoo, Hamed Khalilvandi-Behroozyar, Behzad Asadnezhad, Abdelfattah Zeidan Mohamed Salem Published: 2026 ABSTRACT Highlights•Forage PS, not dietary iNDF concentration, was the primary driver of DM intake, chewing activity, milk yield, and milk fat percentage in lactating dairy goats.•Long particles (20 mm) increased chewing time and milk fat but reduced feed intake, milk production, and feed efficiency.•Rumen pH and most fermentation variables remained unchanged across all treatments; only minor alterations in acetate and valerate were detected.•Goats maintained stable rumen function through marked feed sorting and behavioral adjustments, even on finely ground forage. Introduction Dietary fiber is essential for rumen health and productivity in high-yielding dairy animals. The concept of physically effective neutral detergent fiber (peNDF) represents the fiber fraction that stimulates chewing, mat formation, and rumen motility, traditionally estimated as total neutral detergent fiber (NDF) × the proportion of particles > 1.18 mm (Mertens, 1997). Although particle size (PS) drives chewing responses, fiber’s chemical features — digestibility, brittleness, and indigestible NDF (iNDF) — also affect ruminal fermentation (Li et al., 2017). Multi-omics studies demonstrate that peNDF levels alter rumen metabolites and microbial pathways linked to feed efficiency in goats (Zhou et al., 2022). Low-peNDF diets predispose animals to subacute ruminal acidosis (SARA) by reducing chewing and salivary buffering (Plaizier et al., 2008). Optimal peNDF thresholds vary by species; dairy goats differ from cattle in feeding behavior and rumen dynamics (Zhao et al., 2011). In dairy cows, higher peNDF levels reduce dry-matter (DM) intake and milk yield but may increase total-tract digestibility and milk efficiency (Cao et al., 2021). Increasing forage particle length elevates ruminal pH without affecting DM intake or milk yield. Most existing data concern cows, with limited evidence for small ruminants. Moreover, studies often characterize peNDF but not iNDF. We deliberately chose iNDF, rather than total NDF or ADF, as the second experimental factor because iNDF represents the truly indigestible fraction that is primarily responsible for rumen fill and influences particulate passage rate (Huhtanen et al., 2016; Krizsan et al., 2013). Total NDF includes both digestible and indigestible components, which would confound the physical filling effect with the supply of fermentable substrate. ADF, in turn, is a chemical fraction that does not directly quantify indigestibility and is less suited to predict rumen fill. By manipulating iNDF concentration independently from PS, our goal was to disentangle the physical (size) and chemical (indigestibility) dimensions of dietary fiber and to examine their unique and interactive effects on rumen function and performance. Dairy goats exhibit distinct feeding patterns — intake rate, diet selection, and eating behavior — preventing direct extrapolation from cattle (Lu et al., 2005). Factors such as fiber brittleness, moisture, and preservation method (silage vs. hay) also influence peNDF (Mertens, 1997). Dietary peNDF affects DM intake (DMI): low peNDF accelerates rumen passage and increases DMI (Teimouri Yansari et al., 2004; Zebeli et al., 2008), whereas excessive peNDF limits intake through rumen fill (Allen, 2000; Zebeli et al., 2010). In dairy goats, DMI declined linearly from 1.09 to 0.93 kg/day as peNDF rose from 2.2 to 13.1% DM (Zhao et al., 2011). Variation in ruminal pH often reflects concurrent changes in fermentable carbohydrate supply and volatile fatty acid (VFA) production. Whether ruminal pH decline under low-peNDF diets arises mainly from reduced chewing (and saliva buffering) or from increased fermentable carbohydrate intake remains unclear. This uncertainty is accentuated in goats, historically under-researched compared with cattle (Van Soest, 1994). iNDF, defining rumen fill and passage rate, interacts with PS to govern chewing and rumen function. While peNDF has been extensively studied (Li et al., 2014), iNDF as a separate NDF fraction is also well characterized (Krizsan et al., 2013). The present study investigates how forage PS (physical dimension) and iNDF concentration (chemical dimension) independently and interactively affect chewing activity, rumen fermentation, nutrient digestibility, and lactation performance in dairy goats. We hypothesized that increasing forage PS would increase chewing activity (both eating and ruminating time per unit of DM intake), elevate ruminal pH, and improve milk fat concentration without reducing DM intake. We further hypothesized that higher dietary iNDF concentration would reduce DM intake by increasing rumen fill, prolong rumen retention time, and potentially attenuate the decline in ruminal pH typically observed with finely ground forages. Regarding interactions, we anticipated that the positive effects of larger PS on chewing and rumen fermentation would be more pronounced at high iNDF concentrations, whereas a low iNDF concentration would diminish these PS-dependent responses. Materials and methods Animals, diets, and experimental design All experimental protocols were approved by the Animal Use Committee at Urmia University (Urmia, Iran). Sixty multiparous Mohabadi dairy goats (body weight: 55 ± 5 kg; 70 ± 12 days in milk) were randomly assigned to four treatment groups (n = 15 per group) in a 2 × 2 factorial design. Experimental diets varied in iNDF concentration (high vs. low) and forage PS (short: 2 mm vs. long: 20 mm), while maintaining constant chemical composition with adjusted ingredient proportions to achieve target iNDF levels. The resulting iNDF contrast between the high-iNDF and low-iNDF diets (approximately 18.8% vs. 15.3% of DM) represents a range of indigestible fiber that has previously been associated with measurable changes in rumen fill and particulate passage rate in dairy cattle (Huhtanen et al., 2007; Krizsan et al., 2010a, 2016). This difference was therefore expected to elicit detectable biological responses in the goats under study (Table 1). The two forage PSs were selected to represent contrasting degrees of physical effectiveness: a long hay (20 mm) that provides substantial physically effective fiber to stimulate chewing, rumination, and salivary buffering (Mertens, 1997), and a short hay (2 mm) that largely removes the physical structure of the forage, thereby minimizing chewing activity and accelerating ruminal passage (Teimouri Yansari et al., 2004; Zhao et al., 2011). Alfalfa hay was first chopped to a theoretical length of cut of 20 mm using a six-blade forage cutter (Niroo-Payeh, Iran). A portion of the chopped hay was then ground through a hammer mill (model 160, Arkan Gharb, Iran) fitted with a 2- mm screen to achieve the short PS. Actual PS distribution of the total mixed rations was verified weekly using the Penn State Particle Separator as described in Section 2.2.Table 1Ingredients and chemical composition (g/ 100 gDM) of the experimental diets.Table 1 dummy alt textTreatment diet1high iNDFlow iNDFItemshortlongshortlongIngredient, (g/ 100 gDM)Alfalfa hay, short33-33-Alfalfa hay, long-33-33Wheat Straw161622Orchard grass221616Corn Silage4444Barley Grain30302626Soybean Meal1.51.511Beet Pulp777.57.5Wheat Bran5599Sodium bicarbonate0.20.20.20.2Vitamin-mineral premix20.50.50.50.5Dicalcium phosphate0.40.40.40.4salt0.40.40.40.4Chemical composition3DM72.3172.2572.1072.05OM91.5091.3591.4991.48CP13.4013.3613.4113.40NDF35.2635.2935.2535.30ADF23.9023.9623.5124.05EE2.402.352.402.40ASH5.705.715.705.65NFC443.2443.2943.2443.25Lignin6.856.906.886.90iNDF240h18.8318.8515.2515.231SH= diet with short, LH= diet with long.2Vitamin-mineral mix (per kilogram): 450 mg of nicotinic acid, 2.25 g of Mn, 7.7 g of Zn, 1.25 g of Fe, 1.25 g of Cu, 10 mg of Se, 56 mg of I, 14 mg of Co, 1500 mg of vitamin E, 50,000 IU of vitamin D3, and 250,000 IU of vitamin A.3Values determined from analysis. All values except DM, %, are expressed on a DM basis.4NFC (%) = 100 – (NDF% + CP% + EE% + Ash%), where all values are expressed on a DM basis. Table 1Ingredients and chemical composition (g/ 100 gDM) of the experimental diets.Table 1 dummy alt textTreatment diet1high iNDFlow iNDFItemshortlongshortlongIngredient, (g/ 100 gDM)Alfalfa hay, short33-33-Alfalfa hay, long-33-33Wheat Straw161622Orchard grass221616Corn Silage4444Barley Grain30302626Soybean Meal1.51.511Beet Pulp777.57.5Wheat Bran5599Sodium bicarbonate0.20.20.20.2Vitamin-mineral premix20.50.50.50.5Dicalcium phosphate0.40.40.40.4salt0.40.40.40.4Chemical composition3DM72.3172.2572.1072.05OM91.5091.3591.4991.48CP13.4013.3613.4113.40NDF35.2635.2935.2535.30ADF23.9023.9623.5124.05EE2.402.352.402.40ASH5.705.715.705.65NFC443.2443.2943.2443.25Lignin6.856.906.886.90iNDF240h18.8318.8515.2515.231SH= diet with short, LH= diet with long.2Vitamin-mineral mix (per kilogram): 450 mg of nicotinic acid, 2.25 g of Mn, 7.7 g of Zn, 1.25 g of Fe, 1.25 g of Cu, 10 mg of Se, 56 mg of I, 14 mg of Co, 1500 mg of vitamin E, 50,000 IU of vitamin D3, and 250,000 IU of vitamin A.3Values determined from analysis. All values except DM, %, are expressed on a DM basis.4NFC (%) = 100 – (NDF% + CP% + EE% + Ash%), where all values are expressed on a DM basis. Ingredients and chemical composition (g/ 100 gDM) of the experimental diets. SH= diet with short, LH= diet with long. Vitamin-mineral mix (per kilogram): 450 mg of nicotinic acid, 2.25 g of Mn, 7.7 g of Zn, 1.25 g of Fe, 1.25 g of Cu, 10 mg of Se, 56 mg of I, 14 mg of Co, 1500 mg of vitamin E, 50,000 IU of vitamin D3, and 250,000 IU of vitamin A. Values determined from analysis. All values except DM, %, are expressed on a DM basis. NFC (%) = 100 – (NDF% + CP% + EE% + Ash%), where all values are expressed on a DM basis. The experiment consisted of a 10-day adaptation period followed by 60 days of data collection. Diets were formulated using the Small Ruminant Nutrition System (SRNS v1.9.5105) at a 45:55 concentrate-to-forage ratio on a DM basis. Goats were housed in individual tie-stalls equipped with separate feed bunks and waterers, and received total mixed rations twice daily (08:00 and 20:00) for ad libitum intake, with a target refusal rate of 10%. Feed offered and refused were weighed and recorded daily before the morning feeding. The DM content of the offered feed and refusals was determined by drying a representative subsample at 60 °C for 48 h, and DMI was calculated as the difference between the DM offered and the DM refused. Milking was performed twice daily using a milking machine. Feed sampling, PS distribution, and sorting behavior Samples of the total mixed ration (TMR) offered to each group were collected weekly during the experimental period, composited by treatment, and stored at −20 °C until analysis. Samples of feed refusals (orts) were collected daily from each individual goat, composited by animal over the 60-day collection period, and stored similarly. PS distribution of both offered TMR and refusal samples was determined in duplicate using the Penn State Particle Separator (Nasco, Fort Atkinson, WI) equipped with three sieves (19.0, 8.0, and 1.18 mm) and a bottom pan, as described by Kononoff et al. (2003). Approximately 200 g of as-fed sample was placed on the top sieve, and the separator was shaken horizontally at a frequency of 1.1 Hz and a stroke length of 17 cm for 40 complete cycles. The material retained on each sieve and the pan was then weighed, and the proportion of total dry matter (DM) retained on each fraction was calculated. The physical effectiveness factor (pef) was calculated as the proportion of particles retained on the 19.0 and 8.0 mm sieves (pef>8.0; Lammers et al., 1996) and as the proportion retained on the 19.0, 8.0, and 1.18 mm sieves (pef>1.18; Kononoff et al., 2003). Physically effective NDF (peNDF) of the diets was subsequently determined by multiplying the NDF content of the TMR (Table 2, Table 3) by the respective pef values. To quantify feed sorting behavior, a sorting index was calculated for each particle fraction as the ratio of actual intake of that fraction (g/d) to the predicted intake of that fraction assuming no sorting had occurred (g/d), according to the method of Leonardi and Armentano (2003). A sorting index greater than 1.00 indicates preferential consumption (sorting for), whereas a value <1.00 indicates rejection (sorting against) of that particle fraction.Table 2PS distribution1 of long and short particle alfalfa hay (mean ± SD; n = 8 weekly samples).Table 2 dummy alt textItemLongshort% of DM retained on sieve119.0 mm30.96 ± 1.50.49 ± 0.28.0 mm47.75 ± 2.026.78 ± 1.21.18 mm14.54 ± 0.840.56 ± 1.5Pan6.74 ± 0.532.15 ± 1.0Pef8.00.79 ± 0.020.27 ± 0.02Pef1.180.93 ± 0.010.68 ± 0.01peNDF8.036.53 ± 1.012.51 ± 0.5peNDF1.1842.80 ± 0.830.51 ± 0.61PS distribution of was measured using the Penn State Particle Separator (Kononoff et al., 2003); pef8.0 and pef1.18 = physical effectiveness factor determined as the proportion of particles retained on 2 sieves (Lammers et al., 1996) and on 3 sieves (Kononoff et al., 2003), respectively; peNDF8.0 and peNDF1.18 = physically effective NDF determined as NDF content of multiplied by pef8.0 and pef1.18, respectively.Table 3Effects of forage PS and dietary iNDF content on particle distribution, physical effectiveness factor (pef), and physically effective fiber (peNDF) contents of offered diets and feed refusals.Table 3 dummy alt textTreatment diet1high iNDFlow iNDFP- ValueItemshortlongshortlongSEMiNDFPSiNDF ×PSOffered FeedParticle composition, % DM retained on sieve219.0 mm4.7918.984.2619.820.3640.68<0.010.098.0 mm22.8640.2518.5038.460.637<0.01<0.010.071.18 mm30.5612.3028.6311.000.6600.04<0.010.64Pan41.7728.4648.5930.710.558<0.01<0.01<0.01Geometric mean3 (mm)2.885.832.445.620.055<0.01<0.010.08SD geometric mean3 (mm)3.283.893.224.020.0260.21<0.01<0.01Pef8.00.280.590.220.580.007<0.01<0.010.02Pef1.180.580.710.510.690.005<0.01<0.01<0.01peNDF8.010.0721.918.4921.550.3700.03<0.010.13peNDF1.1821.1526.0418.8625.370.282<0.01<0.010.02Feed RefusalsParticle composition, % DM retained on sieve219 mm0.149.550.029.000.3200.34˂0.010.568 mm10.7015.660.1713.620.3430.01˂0.010.171.18 mm10.727.528.816.820.3540.02˂0.010.16Pan88.0667.2691.01170.550.7790.01˂0.010.84Physical effectiveness factorpef₈.₀0.0100.250.0020.220.0040.04˂0.010.27pef₁.₁₈0.1200.320.090.290.0070.01˂0.010.11Physically effective NDFpeNDF₈.₀0.5311.110.079.90.2810.04˂0.010.25peNDF₁.₁₈5.2218.343.7812.680.3320.01˂0.010.751Short= diet with short hay (2 mm), Long= diet with long hay (20 mm), PS= PS.2PS distribution of was measured using the Penn State Particle Separator (Kononoff et al., 2003); pef8.0 and pef1.18 = physical effectiveness factor determined as the proportion of particles retained on 2 sieves (Lammers et al., 1996) and on 3 sieves (Kononoff et al., 2003), respectively; peNDF8.0 and peNDF1.18 = physically effective NDF determined as NDF content of multiplied by pef8.0 and pef1.18, respectively.3PS geometric mean and its SD for chopped hay and rations determined as recommended by ASAE. Table 2PS distribution1 of long and short particle alfalfa hay (mean ± SD; n = 8 weekly samples).Table 2 dummy alt textItemLongshort% of DM retained on sieve119.0 mm30.96 ± 1.50.49 ± 0.28.0 mm47.75 ± 2.026.78 ± 1.21.18 mm14.54 ± 0.840.56 ± 1.5Pan6.74 ± 0.532.15 ± 1.0Pef8.00.79 ± 0.020.27 ± 0.02Pef1.180.93 ± 0.010.68 ± 0.01peNDF8.036.53 ± 1.012.51 ± 0.5peNDF1.1842.80 ± 0.830.51 ± 0.61PS distribution of was measured using the Penn State Particle Separator (Kononoff et al., 2003); pef8.0 and pef1.18 = physical effectiveness factor determined as the proportion of particles retained on 2 sieves (Lammers et al., 1996) and on 3 sieves (Kononoff et al., 2003), respectively; peNDF8.0 and peNDF1.18 = physically effective NDF determined as NDF content of multiplied by pef8.0 and pef1.18, respectively. PS distribution1 of long and short particle alfalfa hay (mean ± SD; n = 8 weekly samples). PS distribution of was measured using the Penn State Particle Separator (Kononoff et al., 2003); pef8.0 and pef1.18 = physical effectiveness factor determined as the proportion of particles retained on 2 sieves (Lammers et al., 1996) and on 3 sieves (Kononoff et al., 2003), respectively; peNDF8.0 and peNDF1.18 = physically effective NDF determined as NDF content of multiplied by pef8.0 and pef1.18, respectively. Table 3Effects of forage PS and dietary iNDF content on particle distribution, physical effectiveness factor (pef), and physically effective fiber (peNDF) contents of offered diets and feed refusals.Table 3 dummy alt textTreatment diet1high iNDFlow iNDFP- ValueItemshortlongshortlongSEMiNDFPSiNDF ×PSOffered FeedParticle composition, % DM retained on sieve219.0 mm4.7918.984.2619.820.3640.68<0.010.098.0 mm22.8640.2518.5038.460.637<0.01<0.010.071.18 mm30.5612.3028.6311.000.6600.04<0.010.64Pan41.7728.4648.5930.710.558<0.01<0.01<0.01Geometric mean3 (mm)2.885.832.445.620.055<0.01<0.010.08SD geometric mean3 (mm)3.283.893.224.020.0260.21<0.01<0.01Pef8.00.280.590.220.580.007<0.01<0.010.02Pef1.180.580.710.510.690.005<0.01<0.01<0.01peNDF8.010.0721.918.4921.550.3700.03<0.010.13peNDF1.1821.1526.0418.8625.370.282<0.01<0.010.02Feed RefusalsParticle composition, % DM retained on sieve219 mm0.149.550.029.000.3200.34˂0.010.568 mm10.7015.660.1713.620.3430.01˂0.010.171.18 mm10.727.528.816.820.3540.02˂0.010.16Pan88.0667.2691.01170.550.7790.01˂0.010.84Physical effectiveness factorpef₈.₀0.0100.250.0020.220.0040.04˂0.010.27pef₁.₁₈0.1200.320.090.290.0070.01˂0.010.11Physically effective NDFpeNDF₈.₀0.5311.110.079.90.2810.04˂0.010.25peNDF₁.₁₈5.2218.343.7812.680.3320.01˂0.010.751Short= diet with short hay (2 mm), Long= diet with long hay (20 mm), PS= PS.2PS distribution of was measured using the Penn State Particle Separator (Kononoff et al., 2003); pef8.0 and pef1.18 = physical effectiveness factor determined as the proportion of particles retained on 2 sieves (Lammers et al., 1996) and on 3 sieves (Kononoff et al., 2003), respectively; peNDF8.0 and peNDF1.18 = physically effective NDF determined as NDF content of multiplied by pef8.0 and pef1.18, respectively.3PS geometric mean and its SD for chopped hay and rations determined as recommended by ASAE. Effects of forage PS and dietary iNDF content on particle distribution, physical effectiveness factor (pef), and physically effective fiber (peNDF) contents of offered diets and feed refusals. Short= diet with short hay (2 mm), Long= diet with long hay (20 mm), PS= PS. PS distribution of was measured using the Penn State Particle Separator (Kononoff et al., 2003); pef8.0 and pef1.18 = physical effectiveness factor determined as the proportion of particles retained on 2 sieves (Lammers et al., 1996) and on 3 sieves (Kononoff et al., 2003), respectively; peNDF8.0 and peNDF1.18 = physically effective NDF determined as NDF content of multiplied by pef8.0 and pef1.18, respectively. PS geometric mean and its SD for chopped hay and rations determined as recommended by ASAE. The geometric mean PS (GMPS) and its standard deviation were calculated according to ASABE (1992). Chemical analyses of feeds For chemical analysis, feed samples were dried at 60 °C for 48 h, ground through a 1.00-mm screen using a Wiley mill, and analyzed for DM, neutral detergent fiber (NDF), acid detergent fiber (ADF), and lignin according to Van Soest et al. (1991). Ash content was determined by combustion at 550 °C for 8 h, and crude protein (CP) was calculated as total nitrogen × 6.25, determined by the Kjeldahl method (AOAC, 2000). The iNDF content of feed samples was determined by in situ ruminal incubation using four ruminally cannulated lactating Mohabadi dairy goats for 240 h (Krizsan et al., 2013). Ankom F57 filter bags (4 × 5 cm; pore size: 25 µm) containing approximately 0.5 g of ground sample (1.00-mm screen) were incubated in the rumen to ensure about 20 mg of ground sample DM/cm². four bags were inserted in each of the goat’s rumen for each of the diets. Following incubation, bags were removed, rinsed thoroughly with cold water until the rinse water ran clear, and dried at 60 °C for 48 h. Residues were then analyzed for NDF content using heat-stable amylase and sodium sulfite (Mertens, 2002), and iNDF was expressed as the NDF remaining after 240 h of incubation. This extended incubation period is used to estimate the potentially indigestible fiber fraction under practical rumen conditions. Body weight and body condition score measurements Body weight (BW) was recorded weekly for each goat using a calibrated digital scale (accuracy ± 0.1 kg) immediately before the morning feeding. Body condition score (BCS) was assessed biweekly by two independent trained evaluators using a 0–5 scale (with 0.25-unit increments), where 1 = emaciated and 5 = obese, based on palpation of the lumbar vertebrae, short ribs, and sternal fat pads (as described by Hervieu et al., 1991). The final BCS for each goat was the average of the two evaluators' scores. Initial BW and BCS were measured at the start of the adaptation period (day 0), and final BW and BCS were measured on the last day of the data collection period (day 60). Changes in BW (ΔBW) and BCS (ΔBCS) were calculated for each goat as final value minus initial value, and mean changes per treatment were reported as positive or negative values. Milk yield and milk composition Milking was performed twice daily at 06:00 and 18:00 h using bucket milking machines (model HL-JN10, Chuangpu Animal Husbandry Technology Co., Ltd., Suzhou, China). Each treatment group was milked with a dedicated machine equipped with two milking clusters; both clusters were attached to goats of the same treatment group to prevent cross-contamination. The machines operated at a vacuum level of 50 kPa, a pulsation rate of 60–80 cycles/min, and a pulsation ratio of 60:40. Daily milk yield was recorded as the sum of the morning and evening milkings. Milk composition was assessed from representative samples collected over two consecutive days each week. Milk composition (total solids, non-fat solids, fat, protein, and lactose) was determined using a Foss MilkoScan FT6000 spectrometer (Hillerød, Denmark) at Rajzhan Ltd. milk laboratory (Karaj, Iran) following the standard procedure of ISO 9622:2017. Milk fatty acid analysis For milk fatty acid profiling, total lipids were extracted according to Folch et al. (1957) using a 2:1 (v/v) chloroform-methanol mixture. Fatty acid methyl esters (FAMEs) were prepared by base-catalyzed transmethylation with sodium methoxide (0.5 M in methanol), followed by acid-catalyzed esterification with boron trifluoride in methanol (14%). FAMEs were separated and quantified using an Agilent 6890 gas chromatograph (Agilent Technologies, Santa Clara, CA, USA) equipped with a flame ionization detector and a CP-Sil 88 fused silica capillary column (100 m × 0.25 mm i.d., 0.20-µm film thickness; Varian, Middelburg, Netherlands). Injector and detector temperatures were set at 250 °C and 255 °C, respectively. The oven temperature program was as follows: 70 °C for 1 min, ramped to 150 °C at 10 °C/min and held for 10 min, then to 190 °C at 3 °C/min and held for 10 min, and finally to 220 °C at 5 °C/min and held for 15 min. High-purity helium was used as the carrier gas at a constant flow rate of 1.0 mL/min. Peaks were identified by comparing retention times with those of authentic FAME standards (Supelco 37 Component FAME Mix, Bellefonte, PA, USA). For quantification, a known amount of nonadecanoic acid (C19:0; Nu-Chek Prep, Elysian, MN, USA) was added as an internal standard to each sample before lipid extraction. Individual fatty acid concentrations were calculated based on the peak area relative to the internal standard and expressed as g/100 g of total fatty acids. The internal standard also served to monitor extraction and methylation efficiency, with recoveries consistently exceeding 92%. Chewing activity Chewing activity (eating and ruminating) was recorded by direct visual observation at 5- min intervals over a 24- h period on days 19–20, 39–40, and 59–60 of the experiment (Krause & Combs, 2003). At each interval, two trained observers independently recorded the activity of every goat. The daily eating and ruminating times for each animal were then calculated as the average of the two observers' records. Inter-observer agreement was evaluated by comparing the two sets of records collected during one complete 24- h observation period, yielding a concordance of 96%. Total chewing time per kilogram of DM, NDF, peNDF>8, and peNDF>1.18 intake was calculated as described by Teimouri Yansari et al. (2004). Apparent total-tract digestibility Apparent total-tract nutrient digestibility was estimated using acid-insoluble ash (AIA) as an internal marker. Fecal grab samples (approximately 50 g) were collected rectally from each goat once daily before the morning feeding on days 55, 56, and 57 (the final week). The three samples per animal were composited by goat and analyzed for DM, NDF, ADF, and AIA using the same procedures applied to feed samples. The digestibility coefficient for each nutrient was calculated according to the standard formula (Van Keulen & Young, 1977): Digestibility (%) = 100 – 100 × [(% AIA in diet) / (% AIA in feces)] × [(% nutrient in feces) / (% nutrient in diet)] Rumen fluid sampling and analysis Rumen fluid samples (approx. 200 mL) were collected via a flexible PVC esophageal tube (1.2 m length, 10 mm internal diameter; Equivet, Denmark) connected to a stainless-steel suction pump. To minimize saliva contamination, the first 50 mL of aspirated fluid was discarded, and the subsequent fluid was inspected visually; only clear, non-foamy samples free of salivary sediment were retained. Any sample showing visible saliva was discarded and re-collection was attempted after a 30- min interval. The pH was measured immediately after collection using a calibrated pH meter (Titroline easy 5000, Schott AG, Mainz, Germany). The rumen fluid was filtered through four-layer cheese-cloth. The filtered rumen fluid was centrifuged at 15,000 g for 20 min at 4 °C, and 4-mL aliquots of the supernatant were mixed with one mL of 25% metaphosphoric acid and frozen at −20 °C until analysis for volatile fatty acid (VFA) and ammonia nitrogen content. Ammonia nitrogen (NH3—N) was measured, as reported by Broderick and Kang (1980). VFA concentrations were determined by gas chromatography (GC 6820, Agilent Technologies, Santa Clara, CA, USA) equipped with a flame ionization detector and a HP-FFAP capillary column (30 m × 0.25 mm i.d., 0.25 µm film thickness; Agilent). Helium was used as carrier gas at a constant flow rate of 1.0 mL/min. The injector and detector temperatures were set at 250 °C and 260 °C, respectively. The oven temperature program was as follows: 80 °C for 1 min, increased at 10 °C/min to 200 °C and held for 5 min. Individual VFA were identified by comparing retention times with a standard mixture (Volatile Free Acid Mix, Sigma-Aldrich, St. Louis, MO, USA). Blood sampling and plasma metabolites Blood samples (10 mL) were collected from the jugular vein every 20 days into vacuum tubes containing EDTA, centrifuged at 3000 × gfor 15 min at 4 °C, and plasma was harvested and stored at −20 °C until analysis. Plasma urea nitrogen (PUN) concentration was determined using a commercial kit according to the method of Khadem et al. (2009). Plasma non-esterified fatty acid (NEFA) concentration was measured using the Wako NEFA-C kit (Wako Pure Chemical Industries, Osaka, Japan) as described by Johnson and Peters (1993). Statistical analysis An a priori power analysis was conducted using the POWER procedure in SAS (version 9.1; SAS Institute Inc., Cary, NC) to determine the appropriate sample size. Based on expected effect sizes and variance estimates for primary production outcomes (DMI and milk yield) derived from previous literature, a total of 60 multiparous Mohabadi dairy goats (n=15 per treatment) was determined to provide greater than 80% statistical power to detect significant main effects and interactions at a Type I error rate of α=0.05. In this completely randomized design with a 2×2factorial arrangement, the residual error degrees of freedom for main between-subject effects were 56 [4treatments×(15−1)] This value exceeds the recommended threshold (10≤E≤20) of Mead’s resource equation, confirming adequate statistical power to detect biologically meaningful differences. Prior to model fitting, the assumptions underlying the statistical analyses were evaluated sequentially. Normality of residuals was assessed using the Shapiro–Wilk test (PROC UNIVARIATE; SAS Institute Inc., Cary, NC) in combination with visual inspection of quantile–quantile (Q–Q) plots. Homogeneity of variances across treatment groups was examined using Levene’s test. For variables violating the normality assumption, a base-10 logarithmic transformation was applied prior to analysis. Following transformation, normality and homogeneity of variance were reassessed to confirm that the transformed data satisfied model assumptions. Initial BW and parity were initially evaluated as covariates for all production and metabolic variables. As neither covariate reached significance (P > 0.20) and their inclusion did not alter the treatment effects, they were excluded from the final models to preserve degrees of freedom. Final BW and BCS changes were analyzed with initial values as covariates. All data were analyzed using SAS software (SAS Institute Inc., Cary, NC). For non-repeated parameters, such as the chemical composition, physical properties of the diets, data were analyzed using the GLM procedure according to the following model:Yijk=μ+Ni+Pj+(N×P)ij+ϵijkwhere Yijkis the dependent variable; μis the overall mean; Niis the fixed effect of the i-th iNDF concentration (high vs. low); Pjis the fixed effect of the j-th forage PS (short vs. long); (N×P)ij is the fixed interaction effect between iNDF concentration and PS; and ϵijkis the random residual error. A sorting index was calculated for each goat and each particle fraction (19.0, 8.0, 1.18 mm, and pan) according to Leonardi and Armentano (2003). Sorting indices were analyzed separately for each sieve fraction using a mixed model that included the fixed effects of iNDF concentration (high vs. low), forage PS (short vs. long), and their interaction, with animal included as a random effect. Least-squares means were compared using the Tukey adjustment. To determine whether feed sorting behavior was significantly different from random feeding (i.e., no selection), each treatment least-squares mean was tested against the null hypothesis of no sorting (H₀: μ = 1.00) using one-sample t-tests. Data involving repeated measurements over time — including BW and BCS changes, DMI, milk yield, milk composition, and rumen fluid and blood metabolites — were analyzed using the MIXED procedure according to the following model:Yijkl=μ+Ni+Pj+(N×P)ij+Tk+(N×T)ik+(P×T)jk+(N×P×T)ijk+Gl(ij)+ϵijklwhere Ni and Pj are the fixed effects of iNDF concentration and forage PS (as defined above), and (N×P)ij is their interaction; Tk is the fixed effect of the k-th sampling time; (N×T)ik, (P×T)jk, and (N×P×T)ijk are the two-way and three-way interactions among iNDF, PS, and time; Gl(ij) is the random effect of the l-th goat nested within the i-th iNDF and j-th PS combination, used to account for repeated measurements on the same animal; and ϵijkl is the random residual error.A first-order autoregressive covariance structure [AR (1)] was specified in the REPEATED statement to model the correlation between sequential measurements taken on the same animal over time. For all analyses, pairwise comparisons were performed using Tukey-adjusted least squares means. Statistical significance was declared at P≤0.05, and trends were noted at 0.058.0, pef>1.18) and the corresponding peNDF>8.0 and peNDF>1.18 values were higher in the long-particle diets (P < 0.01). The geometric mean PS rose from 2.44 to 2.88 mm for the short-particle diets to 5.62–5.83 mm for the long-particle diets (P < 0.01). The significant iNDF × PS interactions observed for several sieve fractions (e.g., pan and pef₁.₁₈; P < 0.01) indicate that the effect of forage PS on the TMR particle distribution was not uniform across the two iNDF levels; instead, the magnitude of the PS effect differed between the high-iNDF and low-iNDF diets. Refusals after 24 h differed markedly from the offered TMR, showing major reductions in coarse particles and substantial increases in the pan fraction. In long particle diets, 19.0 mm retention declined from 18.98 to 19.82% in the offered feed to 9.0–9.6% in refusals, and 8.0 mm retention decreased from 38.46 to 40.25% to 13.62–15.66%. The pan fraction rose from 28.46 to 30.71% to 67.26–70.55%. Similar trends were observed for short particle diets, where the pan fraction increased from 41.77 to 48.59% to 88.06–91.01%. Compared with short PS, long PS significantly increased pef and peNDF in refusals (P < 0.01), and high iNDF also increased these values relative to low iNDF (Tabe 3, P < 0.01). Sorting indices for the four particle fractions are presented in Table 4. There was a significant main effect of forage PS on sorting of the material retained on the 19.0 mm (P = 0.008), 8.0 mm (P = 0.002), and 1.18 mm sieves (P < 0.001), and a tendency for the pan fraction (P = 0.09). Dietary iNDF concentration did not affect sorting (P ≥ 0.32), and no iNDF × PS interactions were detected (P ≥ 0.58). Goats receiving the short-particle diets did not show sorting behavior significantly different from random feeding (all indices close to 1.00; P > 0.05 vs. 1.00). In contrast, goats fed the long-particle diets actively sorted against the coarse particles retained on the 19.0 mm sieve (high iNDF: 0.85; low iNDF: 0.87; both P < 0.05 vs. 1.00), the 8.0 mm sieve (both 0.88; P < 0.05 vs. 1.00), and the 1.18 mm sieve (both 0.89; P < 0.05 vs. 1.00). Sorting of the pan fraction did not differ significantly from 1.00 in any treatment (P > 0.05), although numerical values (1.01–1.04) suggested a slight preference for the finest particles. Overall, feed sorting was driven exclusively by forage PS, with goats rejecting long particles irrespective of dietary iNDF concentration.Table 4Effects of forage PS and dietary iNDF content on feed sorting indices of Mohabadi dairy goats.Table 4 dummy alt textTreatment diet1high iNDFlow iNDFP- ValueItemshortlongshortlongSEMPSiNDFiNDF ×PS19.0 mm0.970.85*1.000.87 *0.030.0080.320.818.0 mm0.990.88 *1.000.88 *0.020.0020.550.741.18 mm1.000.89 *1.000.89 *0.02<0.0010.680.69Pan1.011.041.001.040.020.090.760.581Short= diet with short hay (2 mm), Long= diet with long hay (20 mm), PS= PS.⁎Sorting index differs significantly from 1.00 (no sorting) by one-sample t-test (P < 0.05). Table 4Effects of forage PS and dietary iNDF content on feed sorting indices of Mohabadi dairy goats.Table 4 dummy alt textTreatment diet1high iNDFlow iNDFP- ValueItemshortlongshortlongSEMPSiNDFiNDF ×PS19.0 mm0.970.85*1.000.87 *0.030.0080.320.818.0 mm0.990.88 *1.000.88 *0.020.0020.550.741.18 mm1.000.89 *1.000.89 *0.02<0.0010.680.69Pan1.011.041.001.040.020.090.760.581Short= diet with short hay (2 mm), Long= diet with long hay (20 mm), PS= PS.⁎Sorting index differs significantly from 1.00 (no sorting) by one-sample t-test (P < 0.05). Effects of forage PS and dietary iNDF content on feed sorting indices of Mohabadi dairy goats. Short= diet with short hay (2 mm), Long= diet with long hay (20 mm), PS= PS. Sorting index differs significantly from 1.00 (no sorting) by one-sample t-test (P < 0.05). Dry matter intake (DMI), milk production and composition Table 5 summarizes the effects of dietary iNDF concentration, forage PS, and their interaction on intake, milk production, and milk composition in Mohabadi dairy goats. Significant iNDF × PS interactions occurred only for NFC intake (P = 0.01) and peNDF>11.18 intake (P = 0.02); therefore, results focus primarily on main effects.Table 5Effects of the interactions between forage PS and different levels of diet iNDF concentration on intake and milk production in Mohabadi dairy goats.Table 5 dummy alt textTreatment diet1high iNDFlow iNDFP- ValueItemshortlongshortlongSEMiNDFPSiNDF ×PSIntake g/dDM1485.31347.81458.81319.615.870.110<0.010.95OM961.1872.1924.1836.08.23<0.01<0.010.96CP199.1180.1195.5176.82.26<0.01<0.010.45EE35.6431.6735.0131.671.290.31<0.010.31NFC2642.2583.5630.9570.65.23<0.01<0.010.01NDF524.2475.6534.6483.68.240.29<0.010.88ADF355.8322.9372.1336.75.370.02<0.010.81peNDF8.03149.6295.3123.8284.45.22<0.01<0.010.19peNDF1.183314.1351.0275.2334.84.01<0.01<0.010.02Milk yield, g/d135010501460105131.10.710.0360.70Fat, %3.734.063.633.910.0870.18<0.010.76Protein, %3.703.723.863.830.1480.370.990.87Lactose, %4.304.194.514.280.0630.040.020.36ASH, %0.820.830.850.850.0220.280.740.74FCM 3.5%4g/d140011451491112137.40.670.0280.61FCM yield efficiency,5g/g0.9430.8501.0220.8500.0310.07<0.010.4211Short= diet with short hay (2 mm), Long= diet with long hay (20 mm), PS= PS.2NFC = nanfiber carbohydrate calculated by the difference of 100 − (% NDF + % CP + % ether extract + % ash).3peNDF8.0 and peNDF1.18 = physically effective NDF determined as NDF content of mixed diet multiplied by pef8.0 and pef 1.18, respectively (Table 3). pef8.0 and pef1.18 = physical effectiveness factor determined as the proportion of particles retained on 2 sieves (Lammers et al., 1996) and on 3 sieves (Kononoff et al., 2003), respectively.4FCM₃.₅% (g/d) was calculated as milk yield (kg/d) × (0.432 + 0.1623 × fat%), then converted to g/d;.5FCM production efficiency (g FCM/g DMI) was calculated as FCM (g/d) divided by dry matter intake (g/d). Table 5Effects of the interactions between forage PS and different levels of diet iNDF concentration on intake and milk production in Mohabadi dairy goats.Table 5 dummy alt textTreatment diet1high iNDFlow iNDFP- ValueItemshortlongshortlongSEMiNDFPSiNDF ×PSIntake g/dDM1485.31347.81458.81319.615.870.110<0.010.95OM961.1872.1924.1836.08.23<0.01<0.010.96CP199.1180.1195.5176.82.26<0.01<0.010.45EE35.6431.6735.0131.671.290.31<0.010.31NFC2642.2583.5630.9570.65.23<0.01<0.010.01NDF524.2475.6534.6483.68.240.29<0.010.88ADF355.8322.9372.1336.75.370.02<0.010.81peNDF8.03149.6295.3123.8284.45.22<0.01<0.010.19peNDF1.183314.1351.0275.2334.84.01<0.01<0.010.02Milk yield, g/d135010501460105131.10.710.0360.70Fat, %3.734.063.633.910.0870.18<0.010.76Protein, %3.703.723.863.830.1480.370.990.87Lactose, %4.304.194.514.280.0630.040.020.36ASH, %0.820.830.850.850.0220.280.740.74FCM 3.5%4g/d140011451491112137.40.670.0280.61FCM yield efficiency,5g/g0.9430.8501.0220.8500.0310.07<0.010.4211Short= diet with short hay (2 mm), Long= diet with long hay (20 mm), PS= PS.2NFC = nanfiber carbohydrate calculated by the difference of 100 − (% NDF + % CP + % ether extract + % ash).3peNDF8.0 and peNDF1.18 = physically effective NDF determined as NDF content of mixed diet multiplied by pef8.0 and pef 1.18, respectively (Table 3). pef8.0 and pef1.18 = physical effectiveness factor determined as the proportion of particles retained on 2 sieves (Lammers et al., 1996) and on 3 sieves (Kononoff et al., 2003), respectively.4FCM₃.₅% (g/d) was calculated as milk yield (kg/d) × (0.432 + 0.1623 × fat%), then converted to g/d;.5FCM production efficiency (g FCM/g DMI) was calculated as FCM (g/d) divided by dry matter intake (g/d). Effects of the interactions between forage PS and different levels of diet iNDF concentration on intake and milk production in Mohabadi dairy goats. 1Short= diet with short hay (2 mm), Long= diet with long hay (20 mm), PS= PS. NFC = nanfiber carbohydrate calculated by the difference of 100 − (% NDF + % CP + % ether extract + % ash). peNDF8.0 and peNDF1.18 = physically effective NDF determined as NDF content of mixed diet multiplied by pef8.0 and pef 1.18, respectively (Table 3). pef8.0 and pef1.18 = physical effectiveness factor determined as the proportion of particles retained on 2 sieves (Lammers et al., 1996) and on 3 sieves (Kononoff et al., 2003), respectively. FCM₃.₅% (g/d) was calculated as milk yield (kg/d) × (0.432 + 0.1623 × fat%), then converted to g/d;. FCM production efficiency (g FCM/g DMI) was calculated as FCM (g/d) divided by dry matter intake (g/d). Increasing alfalfa hay PS from 2 mm to 20 mm significantly reduced DMI (1472 vs. 1334 g/d; P < 0.01), along with OM, CP, EE, NDF, and ADF intakes (all P < 0.01). Conversely, intake of physically effective NDF (peNDF>8.0) was substantially higher with long particle diets (289.9 vs. 136.7 g/d; P < 0.01), consistent with their greater coarse-particle content. Dietary iNDF concentration significantly reduced OM, CP, NFC, and ADF intakes, as well as peNDF>8.0 and peNDF>11.18 (all P < 0.01 except ADF, P = 0.02). Dry matter intake, EE intake, and NDF intake were not significantly affected by iNDF (P = 0.11, P = 0.31, and P = 0.29, respectively), and DMI was numerically higher in the high-iNDF group. Long PS decreased milk yield (1051 vs. 1405 g/d; P = 0.036) and 3.5% FCM yield (1133 vs. 1446 g/d; P = 0.028), while iNDF concentration did not influence milk output (P = 0.71). Milk fat percentage increased with long PS (3.99%vs. 3.68%; P < 0.01) but was not significantly affected by dietary iNDF concentration (3.90%vs. 3.77%; P = 0.18). Milk protein was unaffected (P ≥ 0.37). Lactose percentage decreased with long PS (4.24%vs. 4.41%; P = 0.02) and with high iNDF (4.25%vs. 4.40%; P = 0.04). Milk ash did not differ among treatments (P ≥ 0.28). Feed efficiency (FCM/DMI) was reduced by long PS (0.850 vs. 0.983 g/g; P < 0.01) and tended to decline with high iNDF diets (0.897 vs. 0.936 g/g; P = 0.07). Milk fatty acid profile Table 6 shows the effects of dietary iNDF concentration, forage PS (PS), and their interaction on the milk fatty acid (FA) profile of Mohabadi dairy goats. No significant iNDF × PS interactions were detected for any FA variables (P ≥ 0.14); therefore, interpretation focuses on main effects.Table 6Effects of the interactions between forage PS and different levels of diet iNDF concentration on milk fatty acid profiles in Mohabadi dairy goats.Table 6 dummy alt textTreatment diet1high iNDFlow iNDFP- ValueItemshortlongshortlongSEMiNDFPSiNDF ×PSC4:00.9151.0500.9201.0400.0200.73<0.010.91C6:02.5502.6652.5852.6350.0150.10<0.010.88C8:02.3302.1452.3052.1250.0130.86<0.010.18C10:09.7358.9709.8158.9050.0610.30<0.010.90C11:00.1950.2110.1980.2140.0020.93<0.010.38C12:04.9504.8854.9504.8500.0450.710.140.71C13:00.0810.0820.0800.0820.0010.990.361.00isoC14:00.1280.1260.1300.1200.0010.03<0.010.14C15:00.9970.9800.9870.9710.0050.930.030.15isoC15:00.3110.2280.3140.2130.0210.350.030.31C16:026.6828.6726.8427.920.4130.330.020.51C17:00.7190.6170.7280.6620.0190.400.010.23isoC17:00.3690.4280.3720.4630.0200.480.020.41anteisoC17:00.8720.7610.8820.7500.0040.07<0.010.91C18:08.8658.0508.7157.7450.1440.62<0.010.19cis-9 C18:120.4218.2121.1118.560.1200.05<0.010.87cis-11 C18:10.4080.4340.4200.4550.0110.690.050.22cis-12 C18:10.2380.2140.2290.2060.0040.84<0.010.14trans-9 C18:10.4290.5220.4280.5290.0120.75<0.010.78trans-10 C18:10.8651.4550.9451.3200.021<0.01<0.010.26trans-11 C18:13.9452.6214.0902.6200.0900.46<0.010.47trans-12 C18:10.2590.2470.2500.2520.0070.360.490.76C18:2n-62.9162.7652.9502.7150.0590.510.030.90C18:3n-30.5010.4800.5150.4310.0200.190.060.43C21:00.2730.2520.2740.2530.0070.940.040.8911Short= diet with short hay (2 mm), Long= diet with long hay (20 mm), PS= PS. Table 6Effects of the interactions between forage PS and different levels of diet iNDF concentration on milk fatty acid profiles in Mohabadi dairy goats.Table 6 dummy alt textTreatment diet1high iNDFlow iNDFP- ValueItemshortlongshortlongSEMiNDFPSiNDF ×PSC4:00.9151.0500.9201.0400.0200.73<0.010.91C6:02.5502.6652.5852.6350.0150.10<0.010.88C8:02.3302.1452.3052.1250.0130.86<0.010.18C10:09.7358.9709.8158.9050.0610.30<0.010.90C11:00.1950.2110.1980.2140.0020.93<0.010.38C12:04.9504.8854.9504.8500.0450.710.140.71C13:00.0810.0820.0800.0820.0010.990.361.00isoC14:00.1280.1260.1300.1200.0010.03<0.010.14C15:00.9970.9800.9870.9710.0050.930.030.15isoC15:00.3110.2280.3140.2130.0210.350.030.31C16:026.6828.6726.8427.920.4130.330.020.51C17:00.7190.6170.7280.6620.0190.400.010.23isoC17:00.3690.4280.3720.4630.0200.480.020.41anteisoC17:00.8720.7610.8820.7500.0040.07<0.010.91C18:08.8658.0508.7157.7450.1440.62<0.010.19cis-9 C18:120.4218.2121.1118.560.1200.05<0.010.87cis-11 C18:10.4080.4340.4200.4550.0110.690.050.22cis-12 C18:10.2380.2140.2290.2060.0040.84<0.010.14trans-9 C18:10.4290.5220.4280.5290.0120.75<0.010.78trans-10 C18:10.8651.4550.9451.3200.021<0.01<0.010.26trans-11 C18:13.9452.6214.0902.6200.0900.46<0.010.47trans-12 C18:10.2590.2470.2500.2520.0070.360.490.76C18:2n-62.9162.7652.9502.7150.0590.510.030.90C18:3n-30.5010.4800.5150.4310.0200.190.060.43C21:00.2730.2520.2740.2530.0070.940.040.8911Short= diet with short hay (2 mm), Long= diet with long hay (20 mm), PS= PS. Effects of the interactions between forage PS and different levels of diet iNDF concentration on milk fatty acid profiles in Mohabadi dairy goats. 1Short= diet with short hay (2 mm), Long= diet with long hay (20 mm), PS= PS. Forage PS significantly influenced most milk FA. Compared with short particles (2 mm), long particles (20 mm) increased several de novo synthesized FA, including C4:0, C6:0, C8:0, C10:0, and C11:0 (all P < 0.01), as well as C16:0 (P = 0.02). Conversely, long PS reduced C18:0 and cis 9 C18:1 (P < 0.01). Biohydrogenation intermediates also shifted: trans 9 C18:1 and trans 10 C18:1 increased (P < 0.01), whereas trans 11 C18:1 (vaccenic acid) decreased markedly (P < 0.01). The concentration of cis 12 C18:1 declined (P < 0.01), while cis 11 C18:1 tended to increase (P = 0.05). Odd- and branched-chain FA (OBCFA) were also affected by PS. Long particles reduced isoC14:0 (P < 0.01), C15:0 (P = 0.03), isoC15:0 (P = 0.03), C17:0 (P = 0.01), anteisoC17:0 (P < 0.01), and C21:0 (P = 0.04), but increased isoC17:0 (P = 0.02). Polyunsaturated FA showed smaller responses, with lower C18:2n-6 (P = 0.03) and a tendency toward lower C18:3n-3 (P = 0.06) in milk from goats fed long PS diets. Dietary iNDF concentration had fewer effects on the FA profile. High iNDF reduced isoC14:0 (P = 0.03) and cis 9 C18:1 (P = 0.05), and increased trans 10 C18:1 (P < 0.01). A tendency toward lower anteisoC17:0 was also observed (P = 0.07). Other FA, including de novo synthesized short-chain FA and C18:0, were unaffected by iNDF concentration (P > 0.10). Body weight and body condition score Initial BW and BCS averaged 54.8 ± 5.0 kg and 2.71 ± 0.15 across treatments. All goats increased BW and BCS during the 60-day trial (Table 7). Final BW did not differ by iNDF concentration (P = 0.34), forage PS (PS; P = 0.21), or their interaction (P = 0.81). Final BCS was likewise unaffected by iNDF (P = 0.40), PS (P = 0.31), or their interaction (P = 0.65).Table 7Effects of the interactions between forage PS and different levels of diet iNDF concentration on BW and BCS changes of Mohabadi dairy goats.Table 7 dummy alt textTreatment diet1high iNDFlow iNDFP- ValueItemshortlongshortlongSEMiNDFPSiNDF ×PSInitial BW, kg54.554.555.055.05.0²———Final BW, kg56.355.857.557.03.2²0.340.210.81BW change, kg+1.8+1.3+2.5+2.00.180.030.040.58Initial BCS³2.702.702.722.720.15———Final BCS³2.852.822.922.890.120.400.310.6511Short= diet with short hay (2 mm), Long= diet with long hay (20 mm), PS= PS. Table 7Effects of the interactions between forage PS and different levels of diet iNDF concentration on BW and BCS changes of Mohabadi dairy goats.Table 7 dummy alt textTreatment diet1high iNDFlow iNDFP- ValueItemshortlongshortlongSEMiNDFPSiNDF ×PSInitial BW, kg54.554.555.055.05.0²———Final BW, kg56.355.857.557.03.2²0.340.210.81BW change, kg+1.8+1.3+2.5+2.00.180.030.040.58Initial BCS³2.702.702.722.720.15———Final BCS³2.852.822.922.890.120.400.310.6511Short= diet with short hay (2 mm), Long= diet with long hay (20 mm), PS= PS. Effects of the interactions between forage PS and different levels of diet iNDF concentration on BW and BCS changes of Mohabadi dairy goats. 1Short= diet with short hay (2 mm), Long= diet with long hay (20 mm), PS= PS. Body weight change was influenced by both dietary factors, with no interaction (P = 0.58). Low iNDF diets led to greater weight gain (+2.25 kg) than high iNDF diets (+1.55 kg; P = 0.03). Short PS (2 mm) produced numerically higher gains than long PS (20 mm), and PS affected weight change significantly (P = 0.04). Chewing activity Eating time (min/d) was significantly increased by both high dietary iNDF concentration and long forage PS (P < 0.01; Table 8). Goats fed high-iNDF diets spent more time eating than those on low-iNDF diets (148 vs. 137 min/d, averaged across PSs), and long PS markedly increased eating time compared with short PS (199 vs. 86 min/d). Ruminating time (min/d) was significantly increased by long PS (P < 0.01), whereas the effect of iNDF did not reach significance (P = 0.06). Total chewing time (eating + ruminating) was therefore greater with long PS (P < 0.01) but did not differ between iNDF levels (P = 0.66). Chewing time per kilogram of NDF intake increased with both long PS and high iNDF (P < 0.05), while chewing effort per unit of peNDF>8 decreased with long PS (P < 0.01). This pattern suggests that longer forage particles independently stimulated chewing activity, and that high iNDF diets required additional oral processing per unit of fermentable fiber consumed. This apparently contradictory result suggests that PS independently stimulates chewing activity beyond its effect on peNDF content, likely due to the greater physical resistance of longer fibers. Furthermore, ruminating time per kilogram of peNDF>1.18 increased with PS (P<0.01), indicating that goats adjusted their chewing patterns to maintain optimal rumen function when offered coarser diets. These findings are consistent with Allen’s (1997) hypothesis that chewing activity reflects a balance between PS effects and rumen fill.Table 8Effects of the interactions between PS and different levels of diet iNDF concentration on chewing activity in Mohabadi dairy goats.Table 8 dummy alt textTreatment diet1high iNDFlow iNDFP- ValueItemshortlongshortlongSEMiNDFPSiNDF ×PSEatingMin/ d91205811923.30<0.01<0.010.62Min/g of DM0.060.150.050.140.0020.02<0.010.57Min/g of NDF0.170.430.150.390.0100.01<0.010.55Min/g of ADF0.250.630.210.570.0160.01<0.010.56Min/g of peNDF8.020.610.690.650.670.0240.790.060.25Min/ g of peNDF1.1820.290.580.290.570.0100.44<0.010.63RuminatingMin/ d3524613454435.730.06<0.010.37Min/g of DM0.230.340.230.330.0040.25<0.010.69Min/g of NDF0.670.970.630.910.0170.03<0.010.65Min/g of ADF0.991.420.911.310.0290.01<0.010.54Min/g of peNDF8.022.361.562.751.560.0610.01<0.010.01Min/ g of peNDF1.1821.121.311.241.320.0200.01<0.010.03ChewingMin/ d41367740170516.940.66<0.010.27Min/g of DM0.280.500.270.530.0120.36<0.010.17Min/g of NDF0.791.420.741.460.0350.85<0.010.29Min/g of ADF1.162.101.061.090.0460.29<0.010.36Min/g of peNDF8.022.782.293.212.480.1340.04<0.010.40Min/ g of peNDF1.1821.321.931.442.110.0640.04<0.010.6911 Short= diet with short hay (2 mm), Long= diet with long hay (20 mm), PS= PS.2peNDF8.0 and peNDF1.18 = physically effective NDF determined as NDF content of mixed diet multiplied by pef8.0 and pef1.18, respectively (Table 3). pef8.0 and pef1.18 = physical effectiveness factor determined as the proportion of particles retained on 2 sieves (Lammers et al., 1996) and on 3 sieves (Kononoff et al., 2003), respectively. Table 8Effects of the interactions between PS and different levels of diet iNDF concentration on chewing activity in Mohabadi dairy goats.Table 8 dummy alt textTreatment diet1high iNDFlow iNDFP- ValueItemshortlongshortlongSEMiNDFPSiNDF ×PSEatingMin/ d91205811923.30<0.01<0.010.62Min/g of DM0.060.150.050.140.0020.02<0.010.57Min/g of NDF0.170.430.150.390.0100.01<0.010.55Min/g of ADF0.250.630.210.570.0160.01<0.010.56Min/g of peNDF8.020.610.690.650.670.0240.790.060.25Min/ g of peNDF1.1820.290.580.290.570.0100.44<0.010.63RuminatingMin/ d3524613454435.730.06<0.010.37Min/g of DM0.230.340.230.330.0040.25<0.010.69Min/g of NDF0.670.970.630.910.0170.03<0.010.65Min/g of ADF0.991.420.911.310.0290.01<0.010.54Min/g of peNDF8.022.361.562.751.560.0610.01<0.010.01Min/ g of peNDF1.1821.121.311.241.320.0200.01<0.010.03ChewingMin/ d41367740170516.940.66<0.010.27Min/g of DM0.280.500.270.530.0120.36<0.010.17Min/g of NDF0.791.420.741.460.0350.85<0.010.29Min/g of ADF1.162.101.061.090.0460.29<0.010.36Min/g of peNDF8.022.782.293.212.480.1340.04<0.010.40Min/ g of peNDF1.1821.321.931.442.110.0640.04<0.010.6911 Short= diet with short hay (2 mm), Long= diet with long hay (20 mm), PS= PS.2peNDF8.0 and peNDF1.18 = physically effective NDF determined as NDF content of mixed diet multiplied by pef8.0 and pef1.18, respectively (Table 3). pef8.0 and pef1.18 = physical effectiveness factor determined as the proportion of particles retained on 2 sieves (Lammers et al., 1996) and on 3 sieves (Kononoff et al., 2003), respectively. Effects of the interactions between PS and different levels of diet iNDF concentration on chewing activity in Mohabadi dairy goats. 1 Short= diet with short hay (2 mm), Long= diet with long hay (20 mm), PS= PS. peNDF8.0 and peNDF1.18 = physically effective NDF determined as NDF content of mixed diet multiplied by pef8.0 and pef1.18, respectively (Table 3). pef8.0 and pef1.18 = physical effectiveness factor determined as the proportion of particles retained on 2 sieves (Lammers et al., 1996) and on 3 sieves (Kononoff et al., 2003), respectively. Rumen fermentation parameters, total tract digestibility and blood metabolites Table 9 summarizes rumen fermentation, nutrient digestibility, and plasma metabolite responses to dietary iNDF concentration and forage PS (PS). Ruminal pH, measured 4 h after the morning feeding, ranged from 6.47 to 6.72 and was not significantly influenced by iNDF concentration, forage PS, or their interaction (P ≥ 0.06; Table 9). Total VFA concentration did not differ among treatments (P ≥ 0.67). For the molar proportions of individual VFA, long PS significantly increased the proportion of acetate (55.95 vs. 53.95 mol/100 mol, averaged across iNDF levels; P = 0.02). No significant iNDF × PS interaction was detected for acetate (P = 0.79).The proportions of propionate + isobutyrate, butyrate, and isovalerate were unaffected by treatment (P ≥ 0.25). Valerate was significantly reduced by long PS (0.75 vs. 0.91 mol/100 mol averaged across iNDF levels; P = 0.02), with no effect of iNDF or interaction (P ≥ 0.60). Ruminal ammonia-N concentration did not differ among diets (P = 0.28).Table 9Effects of the interactions between PS and different levels of diet iNDF concentration on Rumen fermentation, Apparent digestibility and blood metabolites in Mohabadi dairy goats.Table 9 dummy alt textTreatment diet1high iNDFlow iNDFP- ValueItemshortlongshortlongSEMiNDFPSiNDF ×PSTotal VFA, mM79.9080.2781.2280.426.400.790.670.68VFA, mol/100 molAcetate54.3856.2255.5257.820.8550.210.020.79Propionat+isobutyrate216.8416.6217.2316.941.0500.340.550.25Butyrate27.5025.5026.4426.430.9180.860.250.34Valerate0.880.740.930.750.0600.600.020.74Isovalerate1.461.681.581.640.2010.850.490.68pH6.606.726.506.470.0780.060.580.35NH3—N, mM12.0112.2910.779.811.830.280.590.58Digestibility %DM75.3574.5474.6971.621.200.160.130.36OM75.2471.6676.5973.230.8060.09<0.010.88CP72.5771.0473.0272.551.030.360.350.61EE47.7447.3549.0447.550.5890.220.130.36NDF48.5147.1151.8051.381.160.050.360.22ADF44.1842.5546.4347.081.140.040.340.83Blood metabolitesGlucose mg/dl62.5058.1866.5159.914.050.490.780.20Triglyceride mg/dl25.6124.9522.9623.151.560.180.880.79Cholesterol mg/dl56.1659.1158.9356.362.160.990.930.22Urea mg/dl11.9811.839.0310.070.9190.020.630.53Albumin g/dl3.774.073.884.060.1410.720.110.69Total protein g/dl5.81b6.02a6.14a5.84b0.1160.540.710.04NEFA mmol/l0.290.270.290.260.0090.890.030.79BHB mmol/l0.500.490.480.490.0080.500.930.331Short= diet with short hay (2 mm), Long= diet with long hay (20 mm), PS= PS.2Under the chromatographic conditions employed, isobutyrate partially co-eluted with propionate; the two are therefore reported as a combined value (propionate + isobutyrate). Table 9Effects of the interactions between PS and different levels of diet iNDF concentration on Rumen fermentation, Apparent digestibility and blood metabolites in Mohabadi dairy goats.Table 9 dummy alt textTreatment diet1high iNDFlow iNDFP- ValueItemshortlongshortlongSEMiNDFPSiNDF ×PSTotal VFA, mM79.9080.2781.2280.426.400.790.670.68VFA, mol/100 molAcetate54.3856.2255.5257.820.8550.210.020.79Propionat+isobutyrate216.8416.6217.2316.941.0500.340.550.25Butyrate27.5025.5026.4426.430.9180.860.250.34Valerate0.880.740.930.750.0600.600.020.74Isovalerate1.461.681.581.640.2010.850.490.68pH6.606.726.506.470.0780.060.580.35NH3—N, mM12.0112.2910.779.811.830.280.590.58Digestibility %DM75.3574.5474.6971.621.200.160.130.36OM75.2471.6676.5973.230.8060.09<0.010.88CP72.5771.0473.0272.551.030.360.350.61EE47.7447.3549.0447.550.5890.220.130.36NDF48.5147.1151.8051.381.160.050.360.22ADF44.1842.5546.4347.081.140.040.340.83Blood metabolitesGlucose mg/dl62.5058.1866.5159.914.050.490.780.20Triglyceride mg/dl25.6124.9522.9623.151.560.180.880.79Cholesterol mg/dl56.1659.1158.9356.362.160.990.930.22Urea mg/dl11.9811.839.0310.070.9190.020.630.53Albumin g/dl3.774.073.884.060.1410.720.110.69Total protein g/dl5.81b6.02a6.14a5.84b0.1160.540.710.04NEFA mmol/l0.290.270.290.260.0090.890.030.79BHB mmol/l0.500.490.480.490.0080.500.930.331Short= diet with short hay (2 mm), Long= diet with long hay (20 mm), PS= PS.2Under the chromatographic conditions employed, isobutyrate partially co-eluted with propionate; the two are therefore reported as a combined value (propionate + isobutyrate). Effects of the interactions between PS and different levels of diet iNDF concentration on Rumen fermentation, Apparent digestibility and blood metabolites in Mohabadi dairy goats. Short= diet with short hay (2 mm), Long= diet with long hay (20 mm), PS= PS. Under the chromatographic conditions employed, isobutyrate partially co-eluted with propionate; the two are therefore reported as a combined value (propionate + isobutyrate). No iNDF × PS interactions were detected for digestibility variables (P ≥ 0.22). Among blood metabolites, only total protein showed a significant iNDF × PS interaction (P = 0.04; Table 9); all other plasma variables were unaffected by the interaction (P ≥ 0.18). DM digestibility was unaffected by iNDF (P = 0.16) or PS (P = 0.13). OM digestibility was significantly lower with long PS than with short PS (72.4%vs. 75.9%; P < 0.01) and tended to be higher in low-iNDF than in high-iNDF diets (P = 0.09), with no iNDF × PS interaction (P = 0.88). CP and EE digestibilities were unchanged (P ≥ 0.22). Dietary iNDF concentration significantly influenced fiber digestibility. Low iNDF diets increased NDF digestibility (51.6%vs. 47.8%; P = 0.05) and ADF digestibility (46.8%vs. 43.4%; P = 0.04). Particle size did not affect NDF (P = 0.36) or ADF digestibility (P = 0.34). Blood metabolites were minimally affected by diet. Glucose, triglycerides, cholesterol, albumin, total protein, NEFA, and BHB were unchanged (P ≥ 0.18). Blood urea nitrogen was higher in goats fed high iNDF diets (11.91 vs. 9.55 mg/dL; P = 0.02), with no PS effect (P = 0.63) or interaction (P = 0.53). Discussion PS distribution of diets and feed sorting behavior PS distributions of offered and refused TMR demonstrated strong selective feeding in Mohabadi goats. Across treatments, goats consistently avoided coarse forage and favored fine, energy-dense particles, reflecting their browsing-type anatomy (Fedele et al., 2002; Lu, 1987). Selectivity was most pronounced in the long-PS diets, where the 19- mm fraction declined from 18.98 to 19.82% in the offered TMR to 9.0–9.6% in refusals (∼50% rejection), while the pan fraction (<1.18 mm) increased. Thus, goats reshaped their diet during consumption and reduced actual intake of long forage despite its intentional inclusion. Although sorting lowered realized peNDF intake, treatment contrasts were maintained. Goats on long-PS diets still consumed nearly double the peNDF>8.0 (289.9 vs. 136.7 g/d), confirming that intended fiber-related effects persisted. The observed iNDF × PS interactions in particle-size distributions likely reflected structural differences in alfalfa (e.g., greater rigidity in high-iNDF forage), but feeding behavior itself was driven mainly by PS. Sorting behavior provides context for intake and production outcomes. Preferential consumption of fine particles left increasingly stem-rich refusals, reducing palatability and slowing processing, contributing to the lower DMI observed with long PS diets—consistent with rumen fill and palatability principles (Allen, 2000). Still, the longer fiber consumed increased chewing and salivary buffering, supported a more stable rumen environment, and promoted acetate production, explaining the higher milk fat percentage and enhanced de novo fatty acid synthesis in long-PS diets. A methodological takeaway is that refusal composition must be measured directly; relying solely on the offered TMR would overestimate peNDF intake and distort diet–response interpretations, consistent with cattle research (Leonardi & Armentano, 2003; Maulfair et al., 2011). Sorting also occurred in short-PS diets, indicating an inherent preference for fine particles and concentrate (Lu et al., 2005). Overall, Mohabadi goats engage in strong, consistent sorting that reduces—but does not eliminate—the effects of forage PS, shaping intake patterns, rumen conditions, and milk production responses. Dry matter intake (DMI), milk production and composition Forage PS exerted a more pronounced influence on intake, milk production, and milk composition in Mohabadi goats than the moderate variation in dietary iNDF. The absence of PS × iNDF interactions confirms that PS acted independently of the indigestible fiber fraction. Long PS (20 mm) reduced DMI (1472 to 1334 g/d; −9.4%), consistent with rumen fill constraints (Allen, 2000; Zebeli et al., 2010) and with studies showing higher DMI with reduced PS (Kahyani et al., 2013; Li et al., 2014; Zhao et al., 2011). All nutrient intakes declined proportionally, whereas peNDF>8.0 intake nearly doubled (136.7 to 289.9 g/d), indicating markedly higher physically effective fiber load. High iNDF further reduced OM, CP, EE, and NFC intakes, but its effects were subordinate to PS. Milk yield decreased by approximately 25% with long PS (1051 vs. 1405 g/d), primarily due to reduced OM and NFC intake, with no compensatory improvement in FCM, aligning with Zhao et al. (2011). Unlike some responses in cattle (Krause et al., 2002; Kononoff et al., 2003), goats appear more sensitive to rumen fill because of smaller rumen capacity. Milk fat percentage increased significantly with long PS (3.99%vs. 3.68%; P < 0.01). Although milk fat was numerically higher with high iNDF diets (3.90%vs. 3.77%), this difference did not reach statistical significance (P = 0.18). The higher milk fat on long-PS diets is consistent with the greater chewing activity and the significantly elevated ruminal acetate proportion observed in those treatments (P = 0.02; Table 9), which together would favour acetate supply for de novo mammary fatty acid synthesis, although a direct causal link was not established. Conversely, lactose percentage declined with long PS (4.24%vs. 4.41%) and high iNDF (4.25%vs. 4.40%), likely reflecting reduced propionate supply and glucose availability for lactose synthesis (Reynolds et al., 2003). Milk protein and ash were unaffected. Feed efficiency (FCM/DMI) decreased with long PS (0.850 vs. 0.983 g/g) and tended to decrease with high iNDF (P = 0.07). This reduction is attributable to increased energetic costs of chewing and rumen motility with high peNDF, consistent with observations in cows (Zebeli et al., 2008) and goats (Zhao et al., 2011). Overall, increasing forage particle size from 2 mm to 20 mm reduced DMI, milk yield, and feed efficiency while increasing milk fat. Dietary iNDF concentration significantly influenced peNDF supply and the intake of several nutrients, but its effects on production traits were modest and did not reach statistical significance. These findings underscore that particle size was the primary driver of production responses, whereas iNDF played a more important role in modulating nutrient intake and fiber digestibility. The results also highlight the heightened sensitivity of goats to rumen fill compared with dairy cattle. Milk fatty acid profile This study demonstrates that forage PS had a stronger physiological impact on milk fat synthesis and ruminal biohydrogenation in Mohabadi goats than the moderate range of iNDF tested, with no interactions indicating that PS acted independently of dietary iNDF. Long forage particles significantly increased the proportions of several de novo synthesized short- and medium-chain FA: C4:0 (P < 0.01), C6:0 (P < 0.01), C8:0 (P < 0.01), C10:0 (P < 0.01), C11:0 (P < 0.01), and C16:0 (P = 0.02). In contrast, C12:0 and C13:0 were not significantly affected by PS (P = 0.14 and P = 0.36, respectively). The increases in these short- and medium-chain FA reflect higher peNDF and the associated increases in chewing, salivary buffering, and rumen pH stability, which promote cellulolytic activity and a greater supply of acetate for de novo mammary FA synthesis, consistent with Li et al. (2014), who linked low peNDF to reduced chewing and altered VFA profiles. Dietary iNDF concentration did not significantly affect any of these FA (P ≥ 0.10), except for a tendency toward higher C6:0 with high iNDF (P = 0.10). A central finding was the shift in biohydrogenation pathways: long PS significantly reduced trans-11 C18:1 (P < 0.01) and increased trans-10 C18:1 (P < 0.01). This pattern is characteristic of diet-induced milk fat depression (MFD) in cows but is less frequently observed in goats. Despite goats’ lower inherent MFD sensitivity, the results indicate partial redirection of biohydrogenation toward the trans-10 pathway, consistent with Toral et al. (2015) and with the species-specific considerations highlighted by Chilliard et al. (2014). High dietary iNDF further increased trans-10 C18:1 (P < 0.01), but did not significantly affect trans-11 C18:1 (P = 0.46). The observed shifts are compatible with a hypothetical transition in the biohydrogenating community, but no rumen microbial profiling was performed in this study (see Section 4.7). The proportion of cis-9 C18:1 was significantly reduced by long PS (P < 0.01) and tended to be lower with high iNDF (P = 0.05). C18:0 was also significantly lower with long PS (P < 0.01) but was not affected by iNDF (P = 0.62). These declines likely reflect reduced C18:0 availability for Δ9-desaturation and/or inhibition of Δ9-desaturase by trans-10 C18:1, consistent with mechanisms described in cows, ewes, and goats (Della Badia et al., 2021; Kadegowda et al., 2008; Shingfield et al., 2010). Responses of odd- and branched-chain fatty acids (OBCFA) were generally consistent with possible microbial shifts, but the interpretation remains hypothetical in the absence of direct rumen microbial data. Long PS significantly reduced C15:0 (P = 0.03), iso C15:0 (P = 0.03), C17:0 (P = 0.01), anteiso C17:0 (P < 0.01), and C21:0 (P = 0.04), while iso C17:0 was significantly increased (P = 0.02). Iso C14:0 was reduced by both long PS (P < 0.01) and high iNDF (P = 0.03). Because iso C14:0, iso C13:0, and C15:0 have been proposed as markers of subclinical acidosis (Kupczyński et al., 2024), their numerical decrease with long PS is consistent with a more stable rumen environment, although, as discussed in Section 4.7, a single time-point pH measurement cannot definitively rule out SARA. Lower anteiso C17:0 has been associated in the literature with Prevotella and Butyrivibrio spp. (Kaneda, 1991; Vlaeminck et al., 2006). If these associations hold under the present conditions, the observed decrease would be compatible with a shift in the abundance of these bacterial groups. However, the absence of direct rumen microbial profiling means that this interpretation remains speculative. Species differences in OBCFA proportions (e.g., lower values in goat vs. buffalo milk) reinforce the need for species-specific interpretation. The generally small effects of iNDF suggest that under a 45:55 forage-to-concentrate ratio, goats compensated for higher iNDF through adjustments in intake and feeding behavior. However, the significant reduction in iso C14:0 (P = 0.03) and the trend for lower cis-9 C18:1 (P = 0.05) with high iNDF imply that truly indigestible fiber may modestly reduce energy availability or alter the profile of fatty acids reaching the mammary gland. The polyunsaturated fatty acids C18:2n-6 and C18:3n-3 showed small but statistically significant or tendency responses: C18:2n-6 was significantly lower with long PS (P = 0.03), while C18:3n-3 tended to be lower (P = 0.06). Neither FA was affected by iNDF (P ≥ 0.19). Species differences provide important context. Toral et al. (2016) reported that shifts from trans-11 to trans-10 pathways are more pronounced in cows, explaining why MFD occurs more readily in cattle. Bernard et al. (2017) showed that mammary lipogenic gene expression in goats is less responsive to starch-rich or PUFA-rich diets, consistent with the comparatively modest shifts observed here. In conclusion, increasing forage PS from 2 mm to 20 mm significantly altered milk FA composition by enhancing de novo short-chain FA synthesis and promoting a shift toward the trans-10 biohydrogenation pathway, with little modification by iNDF. These findings highlight the central role of the physical effectiveness of fiber in shaping mammary lipogenic substrate supply and the profile of fatty acids reaching the mammary gland in goats. Body weight and body condition score All goats gained BW and BCS during the 60-day experiment despite being in mid-lactation (70 ± 12 DIM), indicating that all diets supplied sufficient nutrients to support both milk production and tissue accretion. Body weight change was affected by both iNDF concentration and forage PS. Goats fed low iNDF diets gained more weight (+2.25 vs. +1.55 kg), consistent with their greater DMI (Table 5). Higher intakes of organic matter, crude protein, and non-fiber carbohydrates increased net energy supply beyond maintenance and milk synthesis requirements. This additional energy supported both higher milk yield (Table 5) and greater body reserve replenishment, indicating improved nutrient availability rather than a shift in energy partitioning away from lactation. PS also influenced BW change (P = 0.04). Diets with shorter particles promoted higher DMI and organic matter digestibility, likely due to faster ruminal particle breakdown and passage rate, allowing greater energy consumption and numerically higher weight gains within each iNDF level. However, final BW and BCS did not differ among treatments, suggesting that the observed differences in weight gain were small and did not translate into measurable differences in final body composition after 60 days. These responses align with established patterns of nutrient partitioning in mid- to late-lactation goats, where dietary energy progressively shifts from milk synthesis toward body reserve restoration as lactation advances (Morand-Fehr et al., 2007). In this study, goats fed low iNDF diets produced more milk yet still gained more BW than goats fed high iNDF diets, indicating that energy intake exceeded combined requirements for maintenance and lactation, a response commonly observed in well-fed dairy goats during mid-lactation (Cannas et al., 2004). Goats fed high iNDF diets had lower intake and milk yield but still gained weight, suggesting that reduced milk synthesis allowed greater energy allocation to body reserves. This pattern is consistent with the buffering capacity of ruminant nutrient partitioning, whereby energy is directed toward tissue gain when milk production is limited by diet (Friggens et al., 1998). Even goats fed the long-particle, high-iNDF diet gained 1.3 kg, supporting evidence that goats in later lactation typically maintain or increase BCS when fed ad libitum unless severely underfed (Chilliard et al., 1991). Importantly, no treatment caused weight loss, indicating that even high iNDF diets met maintenance energy requirements and supported modest gain. The 45:55 concentrate-to-forage ratio and ad libitum feeding likely contributed to this outcome, as reflected by rumen pH values above 6.4 across treatments, indicating stable rumen fermentation. In summary, BW gains reflected the normal physiological response of mid-lactation goats under ad libitum feeding. Low iNDF and shorter PS slightly increased weight gain through higher intake and nutrient availability, highlighting that both chemical (iNDF) and physical (PS) forage characteristics influence energy balance and body reserve dynamics in dairy goats, although their effects were more pronounced on production traits than on final body composition. Chewing activity Both dietary iNDF concentration and forage PS significantly influenced eating time (min/d), with high iNDF and long PS each increasing the time goats spent eating. Ruminating time was primarily driven by PS (long > short), while iNDF showed only a numerical trend (P = 0.06). Total chewing time was therefore higher with long forage particles but did not differ between iNDF levels. Chewing time per kilogram of NDF intake rose with PS, whereas chewing effort per unit of peNDF>8 decreased (P < 0.01). Ruminating time per kilogram of peNDF>1.18 also increased (P < 0.01), indicating that PS stimulates chewing independently of peNDF content, likely due to the greater physical resistance of longer fibers. This allows goats to adapt chewing behavior to maintain rumen function. Chewing is the primary driver of PS reduction, affecting digestion and passage rate, and is influenced by forage PS, NDF, iNDF, feed quality, and intake (Zali et al., 2015). Consistent with earlier work, longer particles markedly increased ruminating and total chewing activity (Li et al., 2014; Teimouri Yansari et al., 2004; Zhao et al., 2009). Grinding forage below 3 mm, conversely, greatly reduces chewing (Kononoff, 2002; Teimouri Yansari et al., 2004). Goats show notable tolerance to PS variation. They generally spend more time eating and ruminating than cattle, likely due to smaller size and weaker chewing musculature, which may necessitate finer particle reduction before ruminal clearance (Van Soest, 1994). As a result, DMI is constrained by interactions among PS, NDF/iNDF fractions, and physical fill. Yet goats can maintain stable ruminal pH without high peNDF, possibly by increasing eating frequency (Van Soest, 1994). Elevated chewing increases salivary buffering (Allen, 1997), and similar buffering responses have been observed in goats and cattle fed alfalfa of different PSs (Zhao et al., 2011; Zebeli et al., 2007). Greater chewing stimulation with longer particles was accompanied by a significantly higher ruminal acetate proportion (P = 0.02; Table 9), a pattern consistent with improved conditions for fiber‑degrading microbes, although a direct causal link was not established in the present study. These outcomes support Allen’s (1997) view that chewing reflects a balance between PS and rumen fill. Thus, alfalfa PS and iNDF concentration better predict chewing behavior than NDF alone, confirming that high-peNDF diets more strongly stimulate mastication than low-peNDF diets. Rumen fermentation parameters, total tract apparent nutrient digestibility and plasma metabolites Ruminal pH, measured 4 h after the morning feeding, ranged from 6.47 to 6.72 and was not significantly affected by iNDF concentration, forage PS, or their interaction (Table 9). A tendency for a lower pH in the low-iNDF diets (P = 0.06) likely reflected their higher NFC intake, but all values remained well within the normal physiological range for dairy goats (Cannas et al., 2004). Data indicating that severe pH depression was not evident at this sampling point. However, because SARA is defined by sustained periods of low pH (typically pH < 5.8 for >3 h per day; Gozho et al., 2005), a single time-point measurement cannot definitively exclude its occurrence. Continuous pH monitoring would be required to fully characterize the diurnal rumen pH profile and to rigorously evaluate SARA risk under these dietary conditions. Total VFA concentration did not differ among treatments. Long PS significantly increased the molar proportion of acetate (P = 0.02; Table 9). This result is consistent with the greater chewing activity observed on long-particle diets, which would promote salivary buffering and cellulolytic fermentation, although a direct causal pathway was not established. The absence of a significant iNDF × PS interaction (P = 0.79) indicates that the acetate response to particle size was independent of the dietary iNDF level. Previous studies have reported similarly modest or absent changes in VFA profiles of dairy goats fed varying peNDF levels (Zhao et al., 2011), reinforcing the view that goats are less responsive than cattle to dietary fiber manipulation (Zebeli et al., 2010). The molar proportions of propionate + isobutyrate, butyrate, and isovalerate were not affected by treatment (P ≥ 0.25). In contrast to many studies with dairy cows, where increasing forage PS consistently elevates the acetate-to-propionate ratio (Krause & Combs, 2003; Zebeli et al., 2008), we observed no significant shift in this ratio. This aligns with the results of Zhao et al. (2011), who reported stable VFA profiles across a wide peNDF range in lactating goats, and may reflect species-specific adaptations such as more frequent meals or greater chewing efficiency per unit of feed, which help maintain a stable ruminal environment (Lu et al., 2005; Van Soest, 1994). Valerate was the only VFA significantly affected by PS as a main effect, with lower proportions in goats fed long-particle diets (0.75 vs. 0.91 mol/100 mol; P = 0.02). Valerate is a minor fermentation product arising from carbohydrate and protein degradation; its biological significance is typically small, and a reduction of this magnitude is unlikely to have meaningfully altered energy or metabolite supply. A similar decrease in valerate with high-fiber diets was noted by Bhandari et al. (2008) in goats, although the underlying mechanisms remain unclear. In the absence of direct microbial profiling, we cannot attribute this change to specific bacterial groups, as discussed in Section 4.7, Limitations. Ruminal ammonia-N concentration did not differ among treatments, which is consistent with earlier work where moderate changes in physically effective fiber did not alter ammonia levels in goats (Teimouri Yansari et al., 2004; Zebeli et al., 2008). Collectively, the fermentation data indicate that the rumen environment was remarkably stable across the dietary contrasts employed. This stability likely arose from the goats’ ability to adjust feeding behaviour (e.g., increased eating time, sorting) in response to changes in PS, thereby buffering the rumen against large fluctuations in fermentation end-products. Importantly, the single time-point ruminal pH measurement cannot capture diurnal patterns, and therefore the current data cannot definitively exclude the occurrence of subacute ruminal acidosis. Although the pH values observed at 4 h post-feeding give no indication of a depressed rumen environment, continuous pH monitoring would be required to rigorously evaluate SARA risk. This limitation is explicitly acknowledged in Section 4.7. Moreover, the apparent discrepancy with cattle studies, where finely ground forages frequently induce SARA (Khafipour et al., 2009a), reinforces the view that peNDF thresholds established for dairy cows are not directly transferable to dairy goats (Lu et al., 2005; Zhao et al., 2011). Digestibility responses revealed that iNDF—not PS—was the primary factor determining NDF and ADF digestibility. Low-iNDF diets markedly increased total-tract NDF and ADF digestibility, consistent with the lower indigestible fiber content of these diets (Nousiainen et al., 2009). These improvements paralleled higher OM intake and greater BW gain (Table 5, Table 7), indicating increased substrate availability for microbial fermentation. OM digestibility also tended to be higher with low iNDF (P = 0.09), though the effect was smaller than for the fiber fractions. In contrast, OM digestibility was significantly reduced by long PS (72.4%vs. 75.9%; P < 0.01). This reduction likely reflects a combination of lower total DMI on long-PS diets (Table 5) and the marked feed sorting behaviour observed (Table 4): goats fed long-PS diets consumed a greater proportion of coarse, stem-rich material, which may have depressed overall OM digestibility. The significant effect on OM but not on NDF or ADF digestibility suggests that the impact of particle size was mediated through changes in the intake of non-fibre components rather than through altered fibre fermentation per se. For the fibre fractions, the expected increase in surface area with short particles was likely offset by a faster ruminal passage rate, resulting in no net change in NDF or ADF digestibility (Al-Saiady et al., 2010; Shaver et al., 1986). Similar compensating effects have been reported in sheep (Hogan & Weston, 1967; Thomson et al., 1972). Feed sorting (Table 4) probably further reduced the actual differences in consumed particle size, contributing to the stability of fibre digestibility across PS treatments. Blood metabolites further indicated stable energy balance. Glucose, NEFA, and BHB were unchanged, consistent with uniform BCS and the positive weight gain observed across all treatments (Table 9). These results match studies showing that moderate forage processing changes typically do not alter metabolic profiles in dairy goats (Bhandari et al., 2008; Kononoff & Heinrichs, 2003). A significant iNDF × PS interaction was observed for total protein (P = 0.04): long PS slightly increased total protein under high iNDF but decreased it under low iNDF (Table 9). The biological relevance of this isolated interaction is unclear, but it may reflect minor shifts in protein metabolism or hydration status that were not captured by the other plasma variables measured. The higher BUN concentration in high-iNDF goats (11.91 vs. 9.55 mg/dL; P = 0.02) likely reflects greater hepatic urea production due to enhanced amino acid catabolism for energy, since ruminal ammonia-N did not differ (P = 0.28). Lower DMI, milk yield, and BW gain in high-iNDF goats support this mild energy deficit. This interpretation aligns with reports of elevated BUN when goats consume high-fiber diets (Rapetti et al., 2014). Nevertheless, BUN values remained within normal physiological limits. Overall, iNDF concentration was the dominant driver of fiber digestion and nitrogen metabolism, with low iNDF improving NDF and ADF digestibility and reducing BUN. In contrast, forage PS had minimal effects, likely due to compensatory shifts in passage rate and feed sorting. These findings highlight the importance of considering both chemical and physical forage characteristics in dairy goat ration formulation, while emphasizing that the indigestible fiber fraction exerts the strongest influence under the conditions tested. Limitations Several limitations should be acknowledged. First, no rumen microbial profiling was performed; therefore, inferences about specific bacterial groups or biohydrogenation pathways remain hypothetical. Second, rumen pH was measured at a single time-point post-feeding, which precludes definitive conclusions about subacute ruminal acidosis; continuous monitoring would be required for a rigorous assessment. Third, the dietary iNDF contrast (∼3.5 percentage points) was modest, and a wider range might have revealed stronger effects. Fourth, feed sorting behaviour altered the consumed diet composition, potentially introducing individual variation in nutrient supply. Fifth, individual pre-trial milk yield and composition were not recorded; therefore, baseline homogeneity of these production traits could not be formally tested, although randomization, balanced days in milk, and the adaptation period likely minimized any pre-existing differences. Finally, these findings apply to multiparous Mohabadi goats in mid-lactation and may not extrapolate directly to other breeds, parities, or production systems. Conclusions In multiparous Mohabadi dairy goats, forage PS dominated intake, chewing activity, milk yield, milk fat percentage, and feed efficiency, whereas dietary iNDF concentration mainly influenced fiber digestibility and blood urea nitrogen, with no significant effect on production traits. Long particles (20 mm) reduced DMI and milk yield, increased milk fat and chewing time, and lowered feed efficiency; short particles (2 mm) improved feed conversion and decreased chewing without compromising ruminal pH. Most rumen fermentation variables remained unchanged, with only a significant main effect of PS on acetate (P = 0.02) and a minor valerate shift detected; however, continuous pH monitoring would be required to definitively exclude subacute acidosis. Goats maintained rumen stability through feed sorting and behavioral adjustments, enabling efficient use of finely ground forage. These findings underscore the need for species-specific fiber recommendations, as thresholds derived from cattle cannot be directly transferred to goats. Future work integrating continuous pH recording, rumen microbial profiling, and a broader iNDF gradient will be essential to resolve the underlying mechanisms. Ethical statement All animal procedures were conducted with the approval of the Animal Care and Use Committee of Urmia University (Approval No. IR3012606), in accordance with the guidelines of the Iranian Council on Animal Care. The study followed the ARRIVE guidelines for reporting in vivo experiments, and all experimental procedures complied with applicable institutional regulations and standards. Plant material collection and use were carried out in accordance with relevant institutional, national, and international guidelines. Declaration of generative AI and AI-assisted technologies in the manuscript preparation process During the preparation of this work, the authors used AI-assisted language editing to improve clarity, structure, and readability. After using this tool, the authors reviewed and edited the content as needed and take full responsibility for the content of the manuscript. Ethics approval and consent to participate All animal procedures were conducted with the approval of the Animal Care and Use Committee of Urmia University (Approval No. IR3012606), in accordance with the guidelines of the Iranian Council on Animal Care. The study followed the ARRIVE guidelines for reporting in vivo experiments, and all experimental procedures complied with applicable institutional regulations and standards. Plant material collection and use were carried out in accordance with relevant institutional, national, and international guidelines. Financial support This study was partially supported by the 10.13039/501100007298Urmia University PhD program by Grant no. 2606. CRediT authorship contribution statement Mostafa Soufizadeh: Writing – original draft, Investigation, Formal analysis, Conceptualization. Rasoul Pirmohammadi: Writing – review & editing, Supervision, Project administration, Funding acquisition. Younes Ali Alijoo: Writing – review & editing, Supervision, Project administration, Data curation. Hamed Khalilvandi-Behroozyar: Writing – review & editing, Writing – original draft, Validation, Supervision, Methodology, Investigation, Funding acquisition, Conceptualization. Behzad Asadnezhad: Writing – review & editing, Writing – original draft, Software, Formal analysis. Abdelfattah Zeidan Mohamed Salem: Writing – review & editing, Supervision. Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.