The North-eastern states of India, particularly Meghalaya, hold considerable freshwater aquaculture potential, yet fish production remains below expectation due to cool mid-hill climatic conditions (400-1200 m above MSL) that restrict the growing season to 7-8 months annually (Das et al., 2021a). Labeo gonius (Hamilton, 1822), popularly known as Kuria labeo or Ghonia, is a medium-sized cyprinid increasingly recognised as a promising species for aquaculture diversification in Northeast India (Jena and Das, 2011; Das et al., 2010). It tolerates relatively cooler water temperatures, is marketable at sizes of 100-300 g and fetches comparatively higher market prices than Indian major carps (Das et al., 2010; Bhuyan, 2003), though systematic nutritional studies under tank conditions remain scarce.

Feed cost, accounting for over 50% of total production expenses, is a critical constraint in semi-intensive aquaculture (Tacon and Metian, 2008). Rice bran, abundantly available across Northeast India, serves as a locally accessible energy and partial protein source, but its nutritional utility is constrained by high crude fibre, phytic acid content and lipase-induced rancidity (Putra et al., 2021). Fermentation using lactic acid bacteria (LAB) and yeast effectively reduces these anti-nutritional factors, improves protein content and enhances palatability (Siddik et al., 2024; Aroyehun et al., 2021; Dan et al., 2017) and recent work continues to confirm fermentation of low-value plant by-products, whether by LAB, yeast or their combination, as an effective, low-cost strategy for upgrading agro-industrial residues into functional aquafeed ingredients (Sun et al., 2026; Kalaiselvan et al., 2025; Raman et al., 2024). However, this literature addresses fermented feed ingredients mainly in lowland or cage-based systems in Labeo rohita (Islam et al., 2025; Das et al., 2024; Banerjee et al., 2023; Mandal and Ghosh, 2019; Shamna et al., 2015; Ghosh and Mandal, 2015) and tilapia (Alom et al., 2025; Putra et al., 2021) and no systematic investigation exists for L. gonius under controlled hill conditions. Curd, despite being a traditional and freely available source of LAB in North-eastern Indian households, has similarly not previously been evaluated as a fermentation inoculum for rice bran in a hill-aquaculture carp species. The present study was therefore undertaken to evaluate the effect of curd-fermented rice bran on growth performance, organ indices, apparent digestibility coefficients and muscle proximate composition of L. gonius at mid-hill altitudes of Meghalaya and is, to our knowledge, among the first to report such responses in this species and the first under mid-hill tank conditions of Meghalaya. It was hypothesised that graded replacement of unfermented rice bran with curd-fermented rice bran would progressively improve growth performance and nutrient digestibility in L. gonius fingerlings, relative to a diet containing unfermented rice bran.

Study location and duration
 
The experiment was conducted at the ICAR Research Complex for NEH Region, Umiam, Meghalaya for 60 days from February to April 2026.
 
Fermentation of rice bran
 
Raw rice bran was sun-dried to 8-9% moisture and ground to pass through a 0.5 mm sieve. A combined inoculum of fresh curd (dahi, freshly prepared and used within 24 hr of setting, 10% w/w) as a source of naturally occurring Lactobacillus spp. and baker’s yeast (Saccharomyces cerevisiae, 3% w/w) was mixed with the rice bran, sterile water added to 45-50% moisture, packed in sealed buckets and incubated at 25±2°C for 48 hr under near-anaerobic conditions. The fermented mass was sun-dried and oven-dried at 60°C for 8 hr to arrest microbial activity, then stored at room temperature until use. Fermentation was confirmed by a pH drop from 6.2 to 4.0, with crude fibre declining from 12.60% to 9.20% and crude protein increasing from 12.20% to 15.80% relative to unfermented rice bran. Direct microbial characterisation of the fermented rice bran was not performed in the present study; fermentation was instead confirmed indirectly through the pH and proximate compositional changes described above. Proximate composition of unfermented and fermented rice bran was estimated in triplicate following AOAC (2005) procedures and reported values represent the mean of three independent determinations.
 
Diet formulation, experimental design and sampling
 
Fingerlings (5.42±0.31 g; 30 fish/tank) were stocked after 15 days of acclimatisation, during which no mortality occurred, indicating that the fish were in good physiological condition and well adapted to the holding tank environment before the feeding trial began. Since fingerlings at this size are not externally sexable and gonadal differentiation in L. gonius is not complete until a considerably larger body size, sex could not be considered as an experimental factor; fish of mixed, undifferentiated sex were therefore distributed randomly across treatments.

For the feeding trial itself, three isoproteic (~25% crude protein) and isolipidic diets were prepared: T1 (0% FRB, control), T2 (50% FRB) and T3 (100% FRB), with the total rice bran fraction held constant at 25% across all diets (Table 1). These diets were assigned to fish held under a completely randomised design (CRD), with three replicate cement tanks per treatment (~800 L each; nine tanks in total). Fish were fed twice daily at 3% body weight, adjusted fortnightly, while water quality was maintained through a weekly exchange of 25-30% and continuous aeration that kept dissolved oxygen above 5 mg/L throughout. To estimate apparent digestibility, chromium oxide (Cr2O3, 1%) was incorporated into the diets as an inert marker during Days 50-60. All experimental procedures followed the guidelines of the Committee for the Purpose of Control and Supervision of Experiments on Animals (CPCSEA, 2021).



All fish were sampled on Day 0, Day 30 and Day 60 for individual weight and length measurements. Growth indices including SGR, ADG, FCR, condition factor (K), hepatosomatic index (HSI) and viscerosomatic index (VSI) were computed. Apparent digestibility coefficients (ADC) for dry matter, crude protein and crude lipid were estimated using the indirect marker method (Cho, 1982). At Day 60, dorsal white muscle samples pooled from five fish per replicate tank were subjected to proximate composition analysis following AOAC (2005) procedures. Water quality was monitored fortnightly using calibrated portable instruments and data were analysed by one-way ANOVA with Duncan’s multiple range test (DMRT) at p≤0.05 using SPSS v. 21.0 (IBM, USA). Prior to ANOVA, normality of residuals was verified using the Shapiro-Wilk test and homogeneity of variance using Levene’s test; all growth, digestibility and proximate composition datasets satisfied both assumptions (p≥0.05), justifying the use of parametric ANOVA throughout.

Water quality
 
Water quality parameters recorded fortnightly across all nine tanks are summarised in Table 2. Water temperature ranged between 18.2 and 24.5°C over the experimental period, with fortnightly treatment means of 21.3-21.5°C, consistent with mid-hill conditions at Umiam during February to April. Dissolved oxygen remained above 5 mg/L throughout and pH ranged from 7.08 to 7.13 across treatments. Free CO2 and total alkalinity varied between 4.09-4.22 mg/L and 82.3-84.1 mg/L, respectively. No significant differences (p>0.05) were observed in any water quality parameter among the three dietary treatments at any fortnightly sampling point, confirming that FRB-based diets did not adversely affect the tank environment.


 
Growth performance
 
Growth performance indices of L. gonius at Day 30 and Day 60 are presented in Table 3. Initial weights did not differ significantly among treatments (F2,6 = 0.03, p = 0.97). Significant treatment differences were already apparent by Day 30, with T3 (14.82±1.14 g) significantly heavier than T1 (10.96±0.94 g) and T2 (12.68±1.07 g) differing significantly from both. By Day 60, mean final weight followed the order T3 (26.14±2.11 g) > T2 (22.47±1.95 g) > T1 (18.63±1.82 g), with all pair-wise differences significant (T1 vs T3, p = 0.001; T1 vs T2, p = 0.006; T2 vs T3, p = 0.011). Relative to T1, final body weight was higher by 20.6% in T2 and 40.3% in T3, indicating a graded, dose-dependent growth response to increasing FRB inclusion. SGR ranged from 1.93±0.05%/day in T1 to 2.45±0.06%/day in T3 (F2,6 = 47.3, p<0.001; T1 vs T2, p = 0.008; T1 vs T3, p<0.001; T2 vs T3, p = 0.004) and FCR improved significantly with increasing FRB inclusion, declining from 2.82±0.09 in T1 to 1.98±0.06 in T3 (F2,6 = 52.1, p<0.001; T1 vs T2, p = 0.007; T1 vs T3, p < 0.001; T2 vs T3, p = 0.009). This represented an improvement in FCR of 18.1% in T2 and 29.8% in T3 relative to T1. ADG followed the same trend: T3 (0.35±0.01 g/day) > T2 (0.28±0.02 g/day) > T1 (0.22±0.01 g/day). Survival did not differ significantly among treatments (87.78-90.00%; F2,6 = 0.84, p = 0.47).



The Fulton condition factor (K) was statistically similar across treatments at Day 0. By Day 30, K in T3 (1.23±0.06) was significantly higher than in T1 (1.12±0.05), with T2 (1.18±0.04) intermediate (Table 3). At Day 60, all three treatments differed significantly from each other (T1: 1.15±0.04, T2: 1.24±0.05, T3: 1.31±0.06; T1 vs T2, p = 0.021; T1 vs T3, p<0.001; T2 vs T3, p = 0.014) and within each treatment, the Day 60 value was significantly higher than the corresponding Day 0 value (T1, p = 0.032; T2, p = 0.002; T3, p<0.001).
 
Organ indices, digestibility and muscle composition
 
HSI and VSI at Day 60 are presented in Table 4. HSI differed significantly among treatments (F2,12 = 6.21, p = 0.014), being highest in T1 (1.67±0.12%) and lowest in T3 (1.44±0.11%), with T2 (1.53±0.09%) not differing significantly from either group on DMRT. VSI similarly declined with increasing FRB inclusion (F2,12 = 5.84, p = 0.017), from 6.82±0.43% in T1 to 6.21±0.41% in T3, with T2 (6.43±0.38%) intermediate and not significantly different from either extreme.



ADC for dry matter increased significantly with FRB inclusion (F2,6 = 38.4, p<0.001), from 64.32±1.84% in T1 to 76.84±1.92% in T3, with all pairwise differences significant (T1 vs T2, p = 0.004; T1 vs T3, p<0.001; T2 vs T3, p = 0.008). This corresponds to a relative gain in dry matter digestibility of 11.3% in T2 and 19.4% in T3 over T1. ADC for crude protein likewise improved significantly across all treatments (T1: 74.18±2.31%, T2: 81.52±1.71%, T3: 86.08±2.24%; T1 vs T2, p = 0.009; T1 vs T3, p<0.001; T2 vs T3, p = 0.021), an improvement of 9.9% in T2 and 16.0% in T3 relative to T1. ADC for crude lipid did not differ significantly among treatments (78.64-80.32%; F2,6 = 0.19, p = 0.83).

Muscle crude protein was significantly highest in T3 (61.4±1.1% dw) and lowest in T1 (56.2±1.2% dw), with all three treatments differing significantly from each other (T1 vs T2, p = 0.012; T1 vs T3, p<0.001; T2 vs T3, p = 0.006; Table 4). Muscle crude protein was thus 4.4% and 9.3% higher in T2 and T3, respectively, than in T1. Muscle crude lipid was significantly lowest in T3 (13.6±0.7% dw) compared to T1 (14.8±0.6% dw), while T2 (14.1±0.5% dw) did not differ significantly from either. Muscle moisture and ash did not differ significantly among treatments (p>0.05).

Water quality parameters remained within acceptable ranges for Labeo species culture throughout the experiment, with mean temperatures of 21.3-21.5°C reflecting characteristic mid-hill conditions at Umiam (Meghalaya) during February to April, consistent with observations of Debnath et al., (2025) and Das et al., (2021a). Although L. gonius tolerates cooler temperatures better than L. rohita (Bhuyan, 2003), the moderate SGR values relative to warmer plainland studies are partly attributable to reduced metabolic activity at lower temperatures. Dietary treatments did not interfere with any measured water quality parameter, confirming that FRB-based diets impose no adverse effect on the tank environment, consistent with findings for L. rohita and tilapia in cage systems (Islam et al., 2025; Alom et al., 2025).

The progressive improvement in growth indices and FCR with increasing FRB inclusion may be attributed to several interacting mechanisms associated with fermentation. Fermentation reduced the crude fibre content of rice bran from 12.60% to 9.20%, resulting in a lower overall dietary fibre content in T3 (7.68%) than in T1 (9.14%) and potentially improving energy availability, as cyprinids are inherently limited in their ability to digest high-fibre substrates (Putra et al., 2021). In addition, LAB-mediated phytase activity and enzyme production by yeast may have reduced phytic acid and other anti-nutritional factors, thereby improving mineral and nutrient bioavailability (Gatesoupe, 2008; Dan et al., 2017). The higher crude protein content of FRB (15.80% CP) compared with unfermented rice bran (UFRB) (12.20% CP) may partly reflect microbial biomass production during the combined LAB and S. cerevisiae fermentation, which could also have promoted more extensive degradation of anti-nutritional factors than fermentation with either microorganism alone (Siddik et al., 2024). This interpretation is supported by findings in rohu, where S. cerevisiae-fermented aquafeed showed increased crude protein, reduced crude fibre and anti-nutritional factors and consequently improved nutrient digestibility, growth performance and intestinal enzyme activity compared with unfermented feed (Das et al., 2021b). Similarly, fermented rice bran has recently been reported to promote microbial floc development and enhance growth and digestive enzyme activity in Indian spiny loach (Heniton et al., 2026), suggesting that the nutritional benefits of rice bran fermentation may extend beyond cyprinids. The use of freshly prepared curd as the inoculum further provides a plausible source of LAB, as freshly set home-prepared dahi has been reported to contain approximately 10⁷-10⁹  CFU/g of LAB, with Lactobacillus bulgaricus and Streptococcus thermophilus among the dominant bacteria, while homemade preparations may contain viable LAB populations of approximately 1.6 × 10⁸ CFU/mL (Sudheer et al., 2025). Although the microbial load and species composition of the fermented rice bran were not directly determined in the present study, the decline in pH from 6.2 to 4.0, together with the increase in crude protein and reduction in crude fibre after 48 h of incubation, is consistent with active microbial fermentation and provides indirect evidence that the fermentation process proceeded successfully. Nevertheless, direct enumeration and molecular identification of the dominant microbiota in future studies would be valuable for confirming this interpretation and establishing a clearer link between the microbial community and the observed nutritional improvements.

That the treatment separation in body weight and condition factor was already established by Day 30 suggests that the growth-promoting effect of FRB operates from early exposure rather than accumulating gradually, consistent with a proximate, digestion-linked mechanism rather than a slower adaptive response such as gut microbiota remodelling. The FCR improvement in T3 relative to T1 aligns with findings of Mandal and Ghosh (2019); Dan et al., (2017) and Shamna et al., (2015) in L. rohita and is consistent with the general finding that fermented plant ingredients improve feed conversion efficiency in carps (Alom et al., 2025; Das et al., 2024). Mechanistically, the improved FCR and digestibility with FRB most plausibly reflect a combination of three processes acting in sequence: partial pre-digestion of rice bran starch and protein by microbial extracellular enzymes during fermentation, which reduces the residual metabolic cost of digestion to the fish; hydrolysis of phytate-mineral and phytate-protein complexes by microbial phytase, which releases bound protein and minerals that would otherwise pass undigested through the gut; and a reduction in fibre-associated bulk that would otherwise limit gastric evacuation rate and gut transit time in a species with a relatively simple stomach and short intestine such as L. gonius. These effects are consistent with, rather than proven by, the present dataset and enzymatic or gut-microbiota analyses would be required to establish their relative contribution. Survival was satisfactory and consistent with values reported for L. gonius under similar tank conditions (Debnath et al., 2024).

The higher HSI in T1 (1.67%) relative to T3 (1.44%) suggests that anti-nutritional factors in unfermented rice bran imposed greater metabolic strain on the liver (Olsvik et al., 2010), with comparable HSI reductions reported in L. rohita (Banerjee et al., 2023), tilapia (Putra et al., 2021) and catfish (Shimul et al., 2024). Declining VSI with increasing FRB inclusion reflects reduced intra-coelomic fat deposition, consistent with lower muscle crude lipid in T3 and more efficient nutrient channelling toward somatic growth, a pattern similarly reported by Banerjee et al., (2023) and Putra et al., (2021).

The higher ADC for dry matter (76.84% in T3 vs. 64.32% in T1) and crude protein (86.08% vs. 74.18%) confirm that curd fermentation substantially improved the nutritional accessibility of rice bran for L. gonius. The protein ADC in T3 compares favourably with 87.20% reported for fermented rice bran in shrimp (Jannathulla et al., 2018) and with values for S. cerevisiae-fermented mustard oil cake in L. rohita (Islam et al., 2025) and is broadly consistent with protein digestibility values of 77.07-83.35% reported for fermented mesquite seed meal in Nile tilapia (Aroyehun et al., 2021). The absence of significant differences in ADC for crude lipid is expected, as LAB fermentation does not substantially alter lipid composition across isolipidic diets. Higher muscle crude protein in T3 (61.4% dw) relative to T1 (56.2% dw) reflects more efficient protein deposition consistent with improved protein digestibility, while lower muscle crude lipid in T3 likely results from reduced dietary lipid in FRB relative to UFRB and a protein-sparing effect from more efficient protein utilisation, patterns reported in L. rohita (Das et al., 2024; Dan et al., 2017), tilapia (Alom et al., 2025; Aroyehun et al., 2021) and catfish (Shimul et al., 2024). Muscle moisture and ash did not differ significantly among treatments, consistent with observations in L. rohita fed fermented linseed oil cake diets (Banerjee et al., 2023).

Certain limitations of the present study warrant mention. The trial was conducted over a single 60-day growth phase, which, while sufficient to demonstrate clear treatment separation in growth and digestibility, does not capture performance across a full grow-out cycle to marketable size, nor across the seasonal temperature variation that mid-hill pond systems in Meghalaya experience through the year. The trial was further conducted in cement tanks under controlled feeding and water exchange, which, although necessary for precise measurement of digestibility and organ indices, does not fully replicate the more variable temperature, natural food availability and management regime of earthen pond culture under which most small-scale hill farmers in the region operate. The relatively modest number of replicate tanks per treatment (three), while adequate for the growth and digestibility endpoints reported here, limits the statistical power available for detecting smaller effect sizes in secondary indices such as VSI. Direct microbial characterisation of the fermented rice bran was also not undertaken; while indirect evidence supports active fermentation, a quantitative and species-level link between the fermenting microbiota and the observed nutritional outcomes remains to be established. Extrapolation of the present findings to commercial-scale, full-cycle culture should therefore be made with appropriate caution and the results are best regarded as establishing proof of concept rather than a final production recommendation.

Complete replacement of unfermented rice bran with curd-fermented rice bran significantly improved SGR (2.45±0.06 %/day), FCR (1.98±0.06), condition factor (1.31±0.06), ADC for dry matter (76.8±1.9%) and crude protein (86.1±2.2%) and muscle crude protein (61.4±1.1% dw) in Labeo gonius fingerlings under hill aquaculture conditions of Meghalaya, while reducing HSI (1.44±0.11%) and VSI (6.21±0.41%) relative to the unfermented control. Survival and water quality remained unaffected across treatments. An indicative cost comparison, based on approximate local bulk prices of raw rice bran (~ INR 15/kg), fresh curd (INR 100/kg) and baker’s yeast (INR 200/kg), estimated complete diet costs at INR 62, 64 and 66/kg for T1, T2 and T3, respectively. Applying recorded FCR values of 2.82, 2.31 and 1.98, feed cost per kg weight gain was estimated at INR 175, 148 and 131 for T1, T2 and T3, respectively, representing a reduction in feed cost per unit weight gain of 15.4% for T2 and 25.1% for T3 relative to the unfermented control. This cost advantage should, however, be interpreted with some caution before large-scale adoption is recommended. The present estimate accounts only for feed ingredient prices and does not include the additional labour, fermentation infrastructure and processing time that curd fermentation demands at farm level, nor the batch-to-batch variability in curd microbial load that is likely under uncontrolled, non-standardised farm conditions. For smallholder and semi-intensive farmers in Meghalaya, who typically process rice bran in small, irregular batches, the labour cost of fermentation may partly offset the feed cost saving demonstrated here and the economics would benefit from validation under actual farm-level conditions rather than the standardised laboratory protocol used in the present trial. With this caveat, the estimated 25.1% reduction in feed cost per kilogram weight gain in T3, combined with the significantly improved growth and feed conversion efficiency, indicates that curd-fermented rice bran has practical potential as a low-cost, locally sourced feed ingredient for commercial carp farming in hill regions, provided that a simple, farmer-adoptable fermentation protocol can be standardised and validated at pond scale. Future research should quantify anti-nutritional factor reduction through fermentation, extend trials over a full growing season in earthen pond conditions and explore combined fermentation of rice bran with mustard oil cake for L. gonius culture in NE India. A direct microbiological characterisation of curd-fermented rice bran would strengthen future work in this line by establishing a quantitative and reproducible link between the fermentation process and the nutritional and growth responses reported here and by clarifying the extent to which batch-to-batch variability in curd microbiota affects fermentation outcomes.

The authors gratefully acknowledge the Director, ICAR Research Complex for NEH Region, Umiam, Meghalaya, for approving and supporting the project entitled “Nutritional enhancement of fish feeds through microbial fermentation technique” (2025-2028), under which this study was conducted.

The authors declare that they have no known competing financial interests or personal relationships that could have influenced the work reported in this paper.