The experiment was a 2×3 factorial cross-over design, involving two sheep breeds (Red Maasai and Dorper) by three combinations of GIN infection and diet (hereafter named ‘experimental treatments’): Infected + ad libitum feeding (I-adlib), uninfected + ad libitum feeding (Un-adlib), and uninfected + restricted feeding (Un-restd). Experiment was conducted over two sequential time periods (P1 and P2), all combinations of breed by experimental treatment measured in each time period. Each experimental treatment had 4 lambs of each breed (balanced for sex) making a total of 24 lambs (twelve lambs of each breed). Rhodes grass (Chloris gayana) hay of poor quality was used as feed to mimic the prevailing quality of diets for sheep in small-holder farming systems in SSA.

Each period lasted 36 days (Day 1 to 36). Lambs within each breed were randomly allocated to the three experimental treatments in P1, balanced by entry LW and sex within each breed. The first experimental period (P1) was followed by a 21-day wash out period where all the lambs were dewormed, and faecal examination was done weekly to ensure the lambs remained worm free. Afterwards, lambs were re-allocated to their new experimental treatments for P2.

In order to conduct the CH₄ measurements for each animal on the same days post infection (Days 34–36) and given the availability of only three animal respiration chambers, the lambs were divided into batches of three (one lamb from each treatment), the batches were also separated by breed. Therefore, each batch had either 3 Dorper or 3 Red Maasai. The batches entered the experiments in a staggered manner, every three days.

Acclimatization to the feed, feeding and housing was conducted over the quarantine period (defined below).

The lambs were selected from a flock of lambs born in 2020 at Kapiti Research Station 60 km South-East of Nairobi. The lambs were six to seven months old and their LW were 25.3 ± 2.9 and 27.6 ± 3.6 kg, for Red Maasai and Dorper, respectively. On arrival at the Mazingira Centre the lambs were treated with anthelmintics (10% Albendazole), treated with a topical acaricide, ear tagged and placed under quarantine for 21 days. At the end of the quarantine period, and prior to starting P1, the lambs were randomly allocated to the experimental treatment groups and housed in two separated in-door housing units, one for the infected animals, while the other unit housed the two uninfected treatment groups. The housing provided individual pens with a feed trough and automatic drinker. Rubber mats were placed in each pen to cover the concrete floor. Feeding in pens was conducted twice daily, in two equal portions.

The lambs were housed in individual pens over Day 1 to 26 of each experimental period. On Day 27 they were transferred into metabolic crates for a seven-day period, during which faeces and urine were collected over the last five days (nitrogen and energy balance period). At the end of the collection period, the animals were transferred to the respiration chambers for the last three days of the study (Days 34–36). The lambs were then dewormed and transferred back to Kapiti Research Station.

The lambs were sourced from semi-arid grasslands and therefore had prior exposure to Haemonchus contortus. For deworming the lambs were orally dosed with 10% Albendazole (Albafas) at a dosage of 8 mg/kg LW on the first day of quarantine. Faecal samples taken after seven days of dosing revealed that the lambs still had parasite eggs. The lambs were then drenched with Levamisole at a dose of 7.5 mg/kg LW. Seven days after drenching with Levamisole and at the start of the trial the lambs were confirmed to be worm-free.

On their respective Day 1 of the experiment, the lambs in the infected experimental treatment group (I-adlib) were dosed orally with 1,000 L₃ stage H. contortus larvae suspended in distilled water. Dosing was carried out over four consecutive days to achieve a total larval dose of 4,000 larvae. At the same time, the respective lambs in the two uninfected treatments were dosed with 10 ml of clean water, as a placebo control.

The larvae used for infection were prepared a month prior to the experiment by infecting five adult Red Maasai ewes as donors managed in a pen as a single group. These donor ewes were first dewormed by orally dosing them with 10% Albendazole at a dosage of 8 mg/kg LW. Once confirmed to be worm-free through faecal examination, one ewe was dosed with 4,000 H. contortus L₃ stage larvae suspended in distilled water, while two ewes each were dosed with 6,000 and 8,000 H. contortus L₃ stage larvae, respectively. After a 21-day prepatent period, a faecal bag was put on the donor ewes for total faeces collection. The collected faeces were mixed with vermiculite and water to achieve a dough like consistency, the mixture was put in 1,000 ml glass jars to fill approximately a quarter of the jar height from the bottom. The jars were loosely closed to allow air to flow in and incubated at 27°C in a thermostatically controlled cabinet (Lovibond, Germany) for larvae production. Once the larvae hatched, they crawled up the side of the jar and were visible as white lines. Larvae were washed off the side of the jar with distilled water into 160 ml plastic cell culture flasks. Each flask was topped up with distilled water to a maximum of 60 ml and stored in a regular refrigerator at 4°C until use. The viability of the larvae was confirmed under a dissection microscope (Zeiss, Model Axiocam ERc 5s, Germany) every morning before dosing the lambs in the I-adlib treatment group.

The L₃ larvae doses were calculated by transferring 12 aliquots (0.1 ml each) of the distilled water containing the L₃ larvae onto a petri dish and counting the number of viable L₃ larvae in each aliquot and recorded. Then, the lowest and highest count aliquots were discarded, and the total of the remaining 10 aliquots calculated to get the number of L₃ larvae per 1 ml. Using this figure, the amount of larvae filled distilled water for dosing the required 1000 L₃ per animal was calculated. The donor ewes were not used for the experiment.

The experimental animals were fed solely on chaffed Rhodes grass (Chloris gayana) hay. The hay was sourced from a single plot on a commercial farm. The dry matter (DM) content of the hay was 92.1 ± 0.59%; whereas on a DM basis the hay contained 90.1 ± 0.74 9.9 ± 0.72% ash, 7.2 ± 0.56% crude protein (CP), 68.8 ± 1.74% neutral detergent fibre (NDF), 34.9 ± 1.29% acid detergent fibre (ADF), and 17.4 ± 0.22 MJ gross energy (GE) per kg DM (number of samples=12).

Lambs in the ad libitum feeding groups (I-adlib and Un-adlib) were offered 110% of voluntary feed intake (DM basis) using average feed intake per animal as measured during the acclimatization period, whereas the lambs within the restricted feeding level treatment (Un-restd) were offered 80% of their ad libitum feed on offer as measured during the same period. To avoid feed wastage through spillage, the lambs were fed in two equal portions, in the morning (08:30) and afternoon (16:30). Water was supplied ad libitum and a mineral block provided free choice.

Samples of feed on offer (pooled on a weekly basis) and within-animal pooled samples of feed refusals and faeces were oven-dried (Genlab oven, Model SDO/425/DIG, UK) at 50^°C for 72 h. The dried samples were ground through a five mm sieve using a clean dry Wiley mill (Model MF 10B, IKA Werke, Willmington, N.C., USA) and then passed through a hammer mill (MF 10 basic, IKA, Werke GmbH & CO. KG, Staufen, Germany) fitted with a 1 mm sieve. True DM was determined by drying the samples at 105^°C in an oven (Genlab oven, model SDO/425/TDIG, UK) for 24 h while ash was determined by combustion in a muffle furnace (Nabertherm GmbH, Germany) at 550^°C for six hours according to the methods of the Association of Official Analytical Chemists (AOAC, 1990 methods no. 924.05). Feed, refusals, and faecal samples were analyzed for NDF and ADF by the methods of Van Soest et al. (1991) using an Ankom 200 fibre analyzer, model A2001, USA. Total N content in urine samples was determined by micro-Kjeldhal procedure (AOAC 1990, method no. 988.05). Total N content in feed, refusals, and fecal samples was determined with an organic elemental analyzer (Vario Max, model vario SOLID Sampler, Germany)

Gross energy (GE) contents of feed on offer, feed refusal and faeces were determined according to the methods by Harris (1970) using a bomb calorimeter (Parr 6200, model A1290DDEE, UK). Gross energy contents of urine samples were determined by a method described by Zhao et al. (2016) where the urine was soaked in a filter paper whose GE was predetermined, the soaked filter paper was dried, weighed and its GE determined in a bomb calorimeter. To get the urine GE, the GE value of the filter paper was subtracted from the GE value of the urine-soaked filter paper.

Feed intakes were calculated for the periods when animals were in individual pens, metabolic crates, and respiration chambers. Daily feed intakes were expressed on daily basis (g/d) as well as a g/kg metabolic LW (g/kg LW^0.75)

Apparent total tract digestibility (%) of feed DM and its constituents (OM, N, NDF, ADF, GE; DM basis) were calculated from the ingested and faecal excretions of each item during the feed digestibility and balance period (Days 29–33). For the same period, the nitrogen balance (retained N) was calculated by subtracting the amounts excreted in faeces and urine from the amount ingested. Nitrogen balance was expressed as g N/d as well as % of the N ingested.

Daily metabolizable energy (ME) intake was calculated by subtracting the energy outputs in faeces, and urine (determined for the balance period) and the energy lost in CH₄ calculated using the energy conversion factor (Ym, MJ CH₄/MJ GEI) determined during the chamber period.

Average daily liveweight gain (LWG, g/d) was calculated by subtracting the LW measured on Day 1 from that on Day 27 (morning) and dividing the value by 26 to represent the days in individual pen, while the whole period LWG was calculated by subtracting the LW measured on Day 1 from that on Day 37 (morning) and dividing the value by 36.

Measurements of daily CH₄ emissions were conducted over approximately 23.5 h, and mathematically extrapolated to 24 h. Given that in each period each animal was measured thrice (one day per each chamber), the average daily emissions from each animal were calculated averaging the emissions over the three days. Calculations of the daily emissions in each chamber took into account the ventilation rate and net concentrations at each data point, the gas conditions, and the gas recovery rate. Daily emission of CH₄ was expressed on an absolute basis (g/d), CH₄ yield (g/kg dry matter intake, DMI; g/kg organic matter intake, OMI; g/kg digestible DMI, DDMI and g/kg digestible OMI, DOMI), as well as Ym (CH₄ energy/GEI).

Data were analysed using a linear mixed model in R 4.1.2 (R Development Core Team, USA). The dependent variables measured were first tested for normality by plotting normal quantile-quantile plots in R and visual inspection of the plots for normal distribution of data conducted. The model for each variable was fitted using “lme4” package in R with Breed (B), Experimental treatment (T), Period (P), and interactions of B×T, P×B, P×T and P×B×T as fixed effects and animal identity and batch number as random effect. ANOVA type 3 analysis was used to test for significance of the fixed factors. Least square means were calculated using package “lsmeans” package and Tukey’s built in “Multicompview” package in R was used to separate the means per treatment. The significance level was determined at (P<0.05).

Statistical analysis had two limitations caused by an imbalance in the crossover design. First, from Period 1 of the study (P1), lambs in treatment groups were not equally re-distributed to the alternate treatments for Period 2 (P2); i.e., four lambs in I-adlib group during P1 would ideally have two lambs moved to Un-adlib, and two to Un-restd for P2, but this was not the case except for Dorper Un-adlib treatment in P1. Secondly, one Dorper lamb was mistakenly allocated to the Un-restd treatment group in both periods. These two limitations mean that statistical estimation of treatment × period effect (carry-over effect) could not be efficiently measured. Where the carry-over effect was significant in the models the paper reports in the results and discusses why it should not alter the outcomes and conclusions of the experiment.

At the start of the study (Day 1), the packed cell volume (PCV Day 1) did not differ between experimental treatments (P>0.05). At this time, Dorper lambs had lower PCV than their Red Maasai counterparts, the difference approaching statistical significance (P=0.06) (Table 1). There was a significant period effect (P<0.05) on PCV Day 1, with lambs at start of P2 having higher PCV than at start of P1 (31.2 vs 27.5%). At the end of the study (Day 36) the PCV of infected lambs (I-adlib) was significantly lower (P<0.01) than those of uninfected lambs (Un-adlib and Un-restd) (Table 1) and Dorper lambs continued having lower PCV than Red Maasai (P<0.05). There was a significant Experimental treatment×Period interaction (P<0.05) on PCV Day 36, where the uninfected lambs had higher PCV in P2 than P1 (Un-adlib 29.8 vs 25.7%; Un-restd 30.7 vs 25.9%), whereas the infected (I-adlib) lambs had similar values of PCV Day 36 (P>0.05) in both P2 and P1 (22.5 vs 21.0%).

None of the experimental animals had parasite eggs in faeces on Day 1 of the experiment (FEC Day 1). As expected, experimental treatments differed in FEC at Day 36 (P<0.01). Uninfected lambs (Un-adlib and Un-restd) had no parasite eggs in their faeces. The faecal egg count (FEC) of infected lambs (I-adlib) did not differ with breed (P>0.05) at any point of the measurement period. On Day 36, infected Dorper lambs had only a numerically higher (P>0.05) FEC than their Red Maasai counterparts (Table 1).

The presence of H. contortus eggs in faeces and drop in PCV in the infected lambs (I-adlib) indicate that parasitic infection was achieved, while lack of eggs in the faeces of uninfected lambs indicates that they were worm-free.

Table 2 shows feed dry matter intakes (DMI) during the individual pen period (Days 1−26). Infection did not affect DMI (P>0.05). As expected, daily feed intakes (expressed either as g/d or per kg metabolic LW) of lambs fed ad libitum (I-adlib and Un-adlib) were higher than those of restricted-fed lambs (Un-rest) (P<0.01). There was no breed effect on feed DMI (P>0.05). However, there was a Breed×Period interaction effect on daily DMI (g/d), with Dorper lambs having higher intakes in P1 than in P2 of the study (744 vs 672 g/d; P<0.05), whereas DMI by Red Maasai were maintained across the two periods. The Breed×Period effect on feed DMI per kg of metabolic LW had similar trend as for when expressed as g/d and approached statistical significance (P=0.06).

The liveweight of lambs at the start of the study (LW Day 1) and the subsequent LWs (LW Day 26 and LW Day 36) did not differ between breeds (P>0.05) (Table 2). Experimental treatments did not differ in LW measured at Day 1 (P>0.05), however, there was a significant treatment difference on Day 26 (LW Day 26) and on Day 36 (LW Day 36), with ad libitum fed lambs (I-adlib and Un-adlib) being significantly heavier (P<0.05) than lambs restricted-fed (Un-restd) (Table 2).

The experimental treatments differed (P<0.01) in LWG by lambs over the first 26 days of study (LWG 1−26) as well as over the entire period of study (LWG 1−36) (Table 2). Over Days 1−26 (LWG 1−26), restricted-fed lambs (Un-rest) had negative LWG, whereas ad libitum fed lambs (I-adlib and Un-adlib) had positive and similar LWG (P>0.05). Over Days 1−36, the LWG of lambs in all the experimental treatments differed from each other (P<0.05). The restricted-fed lambs continued losing LW. Over this same period infected lambs had negative LWG, whereas Un-adlib lambs had a smaller positive LWG.

here were significant period effects (P<0.01) on LW and LWG at all the measuring times (Table 2). Liveweight of lambs were higher in P2 than in P1 (26.9 vs 25.9 kg, 27.3 vs 25.6 kg, and 26.7 vs 25.0 kg, for LW Day 1, LW Day 26 and LW Day 36, respectively). Similarly, LWG in P2 were higher (P<0.01) than in P1 (12.1 vs -12.7 g/d and -6.88 vs -27.11 g/d for LWG 1−26 and LWG 1−36, respectively) (Table 2).

Feed dry matter intake (DMI, g/day and g/kg LW^0.75) during the balance period (Days 29−33) (Table 3) followed a similar pattern observed during the previous period (Days 1−26) (Table 2); i.e., experimental treatments differed significantly in DMI (P<0.01), whereas breeds did not (P>0.05). The daily intakes of feed chemical constituents (OM, CP, NDF and ADF) followed the pattern of DMI (P<0.01). As expected, the restricted-fed lambs (Un-restd) had lower intakes of DMI and chemical constituents (g/d) than their ad libitum fed counterparts (I-adlib and Un-adlib) (P<0.01). Overall, infection did not influence feed intake during the balance period (P>0.05) (Table 3). Breeds did not differ (P>0.05) in daily intakes of chemical constituents of feed (Table 3). However, there was a significant period effect on DMI per kg of metabolic liveweight (g/kg LW^0.75), with lambs in P1 having larger DMI intakes than in P2 (60.4 vs 55.2 g/kg LW^0.75). The effect of period on DMI (g/d) approached statistical significance (P=0.06). There was a significant effect of period (P<0.01) on the intake of crude protein (CPI), being this larger in P1 than in P2 (53.9 vs 49.5 g/d). There were significant Breed×Period effects (P<0.05) on intakes (g/d) of NDF (NDFI) and ADF (ADFI) (Table 3), with Dorper lambs having larger NDF (483 vs 444 g/d) and ADF (236 vs 227 g/d) intakes in P1 than in P2, whereas Red Maasai maintained similar intakes in both periods.

Neither experimental treatments nor breeds differed in the digestibilities of feed dry matter (DM) or its chemical constituents (OM, CP, NDF, ADF and GE) (P>0.05) (Table 4). However, there were significant period effects (P<0.01) on digestibilities of DM (DMD), CP (CPD) and ADF (ADFD) (Table 4), these values being higher in P2 than in P1 (68.7 vs 63.5%, 56.2 vs 51.4% and 67.2 vs 63.5%, for DMD, CPD and NDFD, respectively). Nevertheless, for the CPD there was Experimental treatment×Period interaction effect (P=0.01), with lambs in the ad libitum feeding treatments having higher CPD in P1 than in P2 (59.6 vs 50.2% and 57.3 vs 49.3%, for I-adlib and Un-adlib, respectively) (P<0.05), whereas the corresponding CPD for lambs within the Un-rest treatments were only numerically different between periods (51.7 vs 54.4%, respectively). Experimental treatment×Period interaction effect on gross energy digestibility (GED, %) approached statistical significance (P=0.06), showing a trend similar to that of CPD (Table 4).

During the balance period, the daily intakes of nitrogen (NI) by lambs fed ad libitum (I-adlib and Un-adlib) were larger than those of restricted-fed (Un-Restd) lambs (P<0.05), but no effect of infection was observed (P>0.05) on NI; i.e., there were no differences between lambs in I-adlib and Un-adlib treatments (Table 5). Breeds did not differ in NI or N partitioning during the balance period (P>0.05) (Table 5).

Experimental treatments differed in faecal N output (P<0.05) but not in urinary N (P>0.05) (Table 5), with ad libitum fed lambs having larger faecal N outputs than restricted fed lambs (P<0.05); whereas faecal N output was not affected by infection status (P>0.05). The effect of experimental treatments on retained N (g/d and g/g NI) followed the same pattern as for NI and faecal N output, with no effect of infection status (P>0.05) (Table 5). There was a significant period effect (P<0.01) on NI, with intakes being higher in P1 than in P2 (8.62 vs 7.92 g/d). Similarly, retained N in P1 also was higher (P<0.01) in P1 than in P2 (2.34 vs 1.53 g/d and 0.33 vs 0.26 g/g NI). Most importantly, there was Experimental treatment×Period interaction effect (P<0.05) on retained N (g/d), with lambs within the I-adlib treatment having higher retained N in P1 than in P2 (3.01 vs 1.60 g/d). This pattern in I-adlib lambs also approached statistical significance (P=0.06) when retained N was expressed as g/g NI (0.32 vs 0.19 g/g NI for P1 and P2, respectively). Lambs in the Un-adlib treatment also had similar pattern of difference between periods in retained N as those observed for I-adlib lambs (P<0.05) (2.88 vs 1.77 g/d and 0.30 vs 0.21 g/g NI), whereas retention of N by lambs fed at restricted level remained unchanged.

The daily intakes of gross energy (GEI, MJ/d) by lambs fed ad libitum (I-adlib and Un-adlib) were larger than those of restricted-fed (Un-Restd) lambs (P<0.05), but again no effect of infection was observed (P>0.05) on GEI (Table 5). Breeds did not differ (P>0.05) in GEI or energy partitioning, digestibility or metabolisability, except that the calculated CH₄ energy output was higher (P=0.05) in Dorper than in Red Maasai (Table 5). Experimental treatments differed (P<0.01) in partitioning (MJ/d) of GE into faecal energy (FE), calculated CH₄ energy, digestible energy (DE) and metabolizable energy (ME), but not in urinary energy (UE) or metabolisability of energy (MEI/GEI) (Table 5). Overall, restricted fed lambs (Un-restd) had lower partitioning (P<0.05) than ad libitum fed lambs (I-adlib and Un-adlib), with no effect (P>0.05) of infection (Table 5).

There was a significant period effect (P<0.05) on DE and ME partition (Table 5), with values in P1 being higher than in P2 (7.23 vs 6.77 MJ/d and 5.71 vs 522 MJ/d for DE and ME, respectively). There was as a significant Breed×Period interaction effect (P<0.05) on the calculated CH₄ energy output (Table 5, with Dorper lambs having higher outputs in P1 than in P2 (0.99 vs 0.93 MJ/d).

There was effect of experimental treatments (P<0.01) on daily CH₄ emissions (g/d), with emission from ad libitum fed lambs (I-adlib and Un-adlib) being higher than those from restricted fed lambs (Un-restd), but there was no effect (P>0.05) of infection on daily CH₄ emission (g/d) (Table 6). On the other hand, treatments did not differ (P>0.05) in CH₄ yield (emissions per kg of intake of dry matter (DMI), organic matter (OMI), digestible DMI (DDMI), digestible organic matter intake (DOMI)) or CH₄ conversion factor (Ym, MJ CH₄/MJ GEI) (Table 6). Breeds differed (P<0.05) in daily CH₄ emissions (g/d) and CH₄ yields per kg DMI and OMI, and Ym, with Dorper lambs having higher emissions of CH₄ than their Red Maasai counterparts (Table 6). Dorper lambs also tended to have higher CH₄ yields (P<0.08), when emissions were expressed per kg of digestible DMI (DDMI) and digestible OMI (DOMI) (Table 6).

There was a significant Breed×Period effect on daily CH₄ emissions (g/d) (P<0.05), with Dorper lambs having higher emissions in P1 than in P2 (18.0 vs 16.8 g/d), while emissions from Red Maasai were similar between periods (14.4 vs 14.8 g/d). There were period effects on CH₄ yields expressed per kg of DDMI and DOMI. Emissions being higher in P1 than in P2 (40.5 vs 37.0 g/kg DDMI and 39.4 vs 35.8 g/kg DOMI).

There was significant Experimental treatment×Breed×Period interaction effects (P<0.05) on CH₄ yields (emissions per kg DMI, OMI, DDMI, DOMI) and CH₄ conversion factor (Ym). Daily CH₄ yields per unit of DMI and OMI by infected (I-adlib) Red Maasai lambs in P1 (20.7 g/kg DMI and 20.8 g/kg OMI) were significantly lower than those of Infected (I-adlib) Dorper lambs in P2, Un-adlib Dorper lambs in P2 and Un-restd Dorper lambs in both P1 and P2 (27.0 and 27.3, 26.6 and 26.9, 27.1 and 27.4, and 26.5 and 26.5 g/kg DMI and OMI, respectively). In the same line, Daily CH₄ yields per unit of DDMI and DOMI by infected (I-adlib) Red Maasai lambs in P1 (29.2 g/kg DDMI and 28.5 g/kg DOMI) were significantly lower than those of Un-restd Red Maasai lambs in P1, Un-adlib Dorper lambs in P1 and Un-restd Dorper lambs in P1 (43.8 and 42.0, 41.5 and 40.7, and 44.0 and 42.6 g/kg DDMI and g/kg DOMI, respectively). The CH₄ conversion factor (Ym) by I-adlib Red Maasai lambs in P1 (0.063) was also significantly higher than the Ym of I-adlib Dorper lambs in P2 and Un-adlib Dorper lambs in P2 (0.083 and 0.081, respectively).

The pH of rumen fluid was found in a narrow range (6.8-6.9) with no difference between treatments and breeds.

The PCV and FEC are preferred indicators of gastrointestinal nematode (GIN) infection (Xiang et al., 2021). In this study, the lower PCV and higher FEC observed in infected lambs (I-adlib) compared to those of uninfected counterparts (Un-adlib and Un-restd) indicates clearly that parasite infection was achieved. However, at the beginning of the experiment infected Dorper lambs exhibited lower PCV levels compared to their Red Maasai counterparts, and this pattern was maintained throughout the study. In addition, following the two weeks post-infection, FEC increased at a faster rates in infected Dorper lambs than in their Red Maasai counterparts, resulting in 45% higher FEC in Dorper lambs at the end of the study (Day 36). However, this difference was not statistically significant. The above is in line with findings from previous studies (Mugambi et al., 1996; Mugambi et al., 1997; Wanyangu et al., 1997), which reported Red Maasai sustaining higher PCV and lower FEC than Dorper. Nevertheless, breeds did not differ in feed intake, LWG feed digestibility, N retention or metabolisability of feed energy.

According to Alba-Hurtado and Muñoz-Guzmán (2013), sheep can be susceptible, resistant or resilient to H. contortus infection and animals can be found anywhere in the combination of resilience (low–high) × resistance (low–high) quadrants (Blackburn et al., 2013). Resistant animals, based on their innate and acquired immunological capabilities, have lower parasitic load than susceptible counterparts; whereas resilient animals are able to perform in the face of parasite infection, compensating for the negative consequences of infection (Alba-Hurtado and Muñoz-Guzmán, 2013; Blackburn et al., 2013). Improved feed efficiency and protein synthesis are the mechanisms by which susceptible animals display resilience (Doyle et al., 2011). There is general agreement (Alba-Hurtado and Muñoz-Guzmán, 2013; Albuquerque et al., 2019) that native breeds are resistant due to their indirect selection in the prevailing GIN endemic environment, with no treatments against helminths. This is supported by Mugambi et al. (1997) who found that Red Maasai sheep are resistant to H. contortus infection, whereas Dorper sheep were susceptible.

The damage caused by haemonchosis is associated with the parasites blood-feeding activity, a destruction of the abomasal mucosa and increased mucus production (Ramos-Bruno et al., 2021). These events affect feed digestibility and the absorption of nutrients, hence reducing feed conversion efficiency (Hoste et al., 2016). Gastrointestinal parasitism, such as haemonchosis, is also associated with anorexia (Kyriazakis, 2014), which degree of presentation can be affected by the level of infection, species of parasite, site of infection, and animal breed, age and resistance (Sahoo and Khan, 2016). Blood loss, depressed feed intake, digestion and absorption, all affect animal performance. However, further reductions in performance due to H. contortus infection are associated to the nutritional cost of infection; i.e., the use of macro- and micro-nutrients required for the repair of damaged tissues and the setting of the immune response (Sahoo and Khan, 2016; Méndez-Ortíz et al., 2019; Ramos-Bruno et al., 2021). A recent study conducted by Xiang et al. (2021) concluded that lambs infected by H. contortus had decreased feed intake and digestibilities (especially of protein and lipids), reduced LWG, disturbed protein absorption, altered plasma amino acid profiles, and changes in the gastrointestinal microbial community composition. In the present study, however, no effect of parasite infection on feed intake, feed digestibility, nitrogen retention or diet metabolisability was found; neither was the rumen pH altered (6.8–6.9).

Unlike other GIN infections, such as Teladorsagia (T.) circumcincta (Fox et al., 2018), that induce a persistent reduction in voluntary feed intake, the effect of H. contortus on feed intake seems much less pronounced and transient (Sykes and Coop, 1977; Coop et al., 1982; Kyriazakis et al., 1996; Hoste et al., 2016). Furthermore, contradicting results on the effect of H. contortus infection have been reported elsewhere. For instance, Doyle et al. (2014) reported that Merino lambs trickle infected with approximately 9,000 H. contortus L₃ in an 8-week experimental period did not show depression in feed intake. Similarly, studies with creole kids infected with 10,000 H. contortus L₃ in two 7-week phases (a primary and a secondary infection) (Bambou et al., 2009; Bambou et al., 2013) revealed that a transient feed intake depression was only observed during the 3-week period post the primary infection. In line with the above, a meta-analyses study conducted by Méndez-Ortíz et al. (2019) on the effects of gastrointestinal nematodes on feed intake and LWG by lambs revealed that 41% and 27% of the studies did not observe negative effect of GIN on feed intake and LWG, respectively.

Parasite-induced anorexia is viewed as the result of the interaction of the unavoidable consequence of infection and the coping mechanism of the animal to deal with infection (Kyriazakis, 2014). In the present study, daily feed intake was measured over the entire experimental period. Considering that the housing change of the animals from individual pens to metabolic crates and then to respiration chambers may have had implied some stress response. Feed intake and LWG was calculated for the individual pen period (Days 1–26) as well as for the entire experimental period (Days 1–36). These calculations did show effects of feed restriction on feed intake and LWG (restricted-fed lambs having lower feed intakes and LWG than ad libitum fed lambs). On Day 1-26 the uninfected and infected ad libitum fed lambs had similar LWG. However, when LWG was calculated for the entire experimental period (Days 1-36), infected ad libitum fed lambs lost weight while their uninfected counterparts had a marginal LWG, this is in line with other studies where infected sheep had an increased weight depression in the later stages of the experiment (three to five weeks post infection) compared to the first three weeks post infection (Albers et al., 1989). Pen studies allow for full control of the GIT worm infection, however, this also obscures the normal effect of infection on sheep production as opposed to field conditions where the effects of anemia and the resultant lethargy would exert more pronounced negative effects on feed intake as the sheep search for sufficient feed and water (Albers et al., 1989). This could explain why infected lambs in the present study did not have a significant drop in feed intake and LWG in earlier stages of infection, but the difference became more evident by a drop in LWG in the later weeks of infection.

Various studies (Retama-Flores et al., 2012; Méndez-Ortíz et al., 2019; Ramos-Bruno et al., 2021) have tried to assess the nutritional cost of GIN infection. The most used approach is assessing the N retention, metabolisability of feed energy and LWG of infected and non-infected animals. Ramos-Bruno et al. (2021) have suggested that the nutritional cost of infection would depend on the animal’s ability to withstand the parasitic impact (resilience) or the ability to mount an immune response (resistance). In line with the latter, Blackburn et al. (2013) have suggested that mounting the immune response can be of a higher cost than displaying resilience, and host immune response rather than parasite damage seems the major cause of loss in productivity. Sahoo and Khan (2016) have stated that resistance and animal performance compete each other for nutrients, especially protein.

Infection with H. contortus in the present study not only did not affect apparent nutrient digestibility, but also there was no evidence of nutritional cost judged from the nitrogen retention and metabolisability of the diet by infected and uninfected lambs. However, regarding energy cost, it should be pointed out that the best indicator of it would be ratio of heat production to metabolizable energy; i.e., the higher the ratio, the higher the energy cost of infection. However, heat production was not measured in this study.

The nutritional status of the host animal is an important factor influencing the host-parasite relationship and the pathogenesis of infection (Sahoo and Khan, 2016). Positive energy balance and improved body condition score are important in setting up the immune response (Lord, 2002; Valderrabano et al., 2006). In the same line, improved protein nutrition is important for the reduction in the establishment of infection by improving the immune response as well as improving the resistance and resilience to infection (Houdijk et al., 2006; Louvandini et al., 2006), as well as the repair of the damage in the gastrointestinal tract (Khan et al., 2011; Ramanan et al., 2016) caused by GIT worms. In this study, the diet was poor both in crude protein and metabolizable energy contents and the animals fairly maintained their body condition score around 3 (scale 1 to 5), but the negative impact of infection on animal performance was mild.

Reasons for the lack of effect of parasite infection on feed intake, N retention and diet metabolisability can only be hypothesized to be due to: (1) the level of infection achieved was not severe enough to produce feed intake depression and hence LW loss during the pen period. Further, at Day 36 (when infection peaked) the impact of parasitism on LWG was still marginal (2) both sheep breeds have high resistance to parasite infection (especially Red Maasai), as discussed earlier; i.e., Red Maasai being indigenous to the environment and Dorper having an indigenous genetic component (Blackhead Persian). (3) In the present experiment, the lambs were sourced from a production system based on the utilization of rangelands and therefore had prior exposure to H. contortus. It is therefore plausible that due to the prior exposure, the lambs had acquired a degree of resistance that could explain the lack of pronounced effects of infection on feed intake and hence LWG. (4) both sheep breeds have adaptation to the low feeding value of the diets, allowing them to respond to infection and sustain the immune response against parasite infection (Ramos-Bruno et al., 2021). The mechanisms by which nutritional status of the animal enhances resilience and resistance to parasites are still unknown (Sahoo and Khan, 2016). However, the low N and ME contents of the diet used in the study makes it difficult to sustain the latter hypothesis.

Daily enteric methane emissions from ruminants are closely related to the amount of feed consumed (e.g., Buddle et al., 2011; Hammond et al., 2013; Charmley et al., 2015) and the results of the present study confirm the above relationship, with daily methane emission (g/d) from restricted-fed lambs being lower than those from their ad libitum fed counterparts.

Comparison of emissions across experiments is more relevant when emissions are expressed in terms of CH₄ yield (e.g., g/kg DMI, g/kg DOMI, and on energy basis, MJ CH₄/MJ GEI). The CH₄ yield found in this study (23.3–26.4 g/kg DMI) is within the range found by other studies in tropical environments with tropical low-quality forages (Hunter, 2007; Archimède et al., 2018; Costa et al., 2020; Gere et al., 2022). For example, Archimède et al. (2018), reported CH₄ yields in the range of 15–29 g/kg DMI from sheep fed low quality C4 grasses; whereas Hunter (2007) and Gere et al. (2022) reported even a higher CH₄ yield (31 g/kg DMI each) from cattle fed on Angleton grass and sheep fed on Rhodes grass hay (a forage similar to that used in the present study), respectively. These values are higher than those reported from studies in temperate regions with temperate forages. For example, numerous studies conducted in New Zealand (Muetzel and Clark, 2015; Swainson et al., 2016) with temperate pasture forages reported a mean CH₄ yield of 23.5 g/kg DMI, with no difference between cattle and sheep. Ruminants fed tropical forages usually produce more enteric CH₄ (yield) than ruminants fed temperate forages and this fact is related to the more fibrous nature of C4 grasses compared to C3, adaptation of animals to low quality C4 grasses (e.g., longer retention time of lower quality feed) and rumen environment more favorable to acetogenic fermentation and methanogenesis (Beauchemin et al., 2009; Archimède et al., 2011; Kennedy and Charmley, 2012). For example, Rira et al. (2016) reported more abundant total ruminal bacteria and methanogens in animals kept in tropical environments than in temperate environments. Nevertheless, much lower CH₄ yields were reported from other studies with sheep (Amaral et al., 2016; Lima et al., 2019; Corrêa et al., 2021; Fernandes et al., 2022). For example, the study of Amaral et al. (2016) reported CH₄ yield of 13 g/kg DMI from sheep fed on high quality (N fertilized) C4 annual grasses. In the present study, the lambs were fed on a low-quality tropical grass with high fibre but low crude protein contents. A study using Boran steers fed poor quality tropical forages providing sub-optimal maintenance energy requirements showed that although the methane production was reduced due to reduced feed intake, both CH₄ yield and CH₄ conversion factor increased as intake fell (Goopy et al., 2020).

One important finding from the present study regarding enteric CH₄ emissions was that H. contortus infection did not affect the absolute daily enteric CH₄ emissions (g/d) or CH₄ yield (g/kg DMI). This is in contrast with findings from other studies (Fox et al., 2018; Lima et al., 2019; Corrêa et al., 2021; Fernandes et al., 2022), which showed not only increased CH₄ yields from GIN infected sheep, but also depressed feed intake and LWG. The research by Lima et al. (2019); Corrêa et al. (2021) and Fernandes et al. (2022) were all conducted with parasite naive lambs of a tropical sheep breed (Santa Ines), fed either on Cynodon dactylon hay as a sole diet (Corrêa et al., 2021) or a mixed diet containing 70% (Lima et al., 2019) or 60% (Fernandes et al., 2022) of this forage plus concentrates. Corrêa et al. (2021) reported increase in CH₄ yield of 62 and 99%, respectively when the lambs were separately infected with H.contortus and T. colubriformes. In turn, Fernandes et al. (2022) reported a 59% increase in CH₄ yield from lambs infected with both H. contortus and T. colubriformes, whereas for a similar mixed infection, Lima et al. (2019) reported an up to 250% increase in CH₄ yield. In a similar way, from a research work with parasite naïve lambs infected with T. circumcincta and fed on grass pellets (DM digestibility of 55.4–58.4%), Fox et al. (2018) reported 33% increase in CH₄ yield. In the present study, the lack of effect of parasite infection on CH₄ yield is in line with the lack of effect on feed intake and LWG.

Infection by GIN in small ruminants alters substantially the gastrointestinal tract and understanding of the effects of infection on microbiome diversity and function, and ultimately on rate of methanogenesis is still in early development (Li et al., 2016; El-Ashram et al., 2017; Correa et al., 2020; Corrêa et al., 2021; Xiang et al., 2021). These studies showed alterations in the ruminal and abomasal microbial functional structure and composition as consequence of GIN infection. Corrêa et al. (2021) hypothesized that the increased CH₄ yield from lambs infected with GIN might be related to the increase in methanogenic archaea population in the rumen, as well as to the higher abundance of genes related to methane metabolism and key enzymes involved in methanogenesis (e.g., methyl-coenzyme M and methanopterin). In the present study no measurements of rumen function and microbiome were carried out, except for rumen pH, which was observed not to be affected by infection.

To our knowledge, this is the first study that compared CH₄ emissions from the two most common breeds in Sub-Saharan Africa, whereas fed on a low-quality tropical forage, that mimicked the predominant diet quality found in production systems. Under these conditions, Dorper had higher daily methane emission and methane yields than their Red Maasai counterparts. It is difficult to ascertain about the animal, dietary and environmental (e.g., health stress) factors that are likely involved in the observed difference in CH₄ emission between these breeds. The only other observed measured difference between these two breeds was the health status (i.e., PCV and FEC) resulting from infection with H. contortus. It is unknown whether rumen microbial population structure and function (e.g., Corrêa et al., 2021) and digesta retention time in rumen (Pinares-Patiño et al., 2003) were altered in Dorper breed in response to infection and hence leading to increased CH₄ emission (absolute and CH₄ yield). Although, in this study, feed digestibility was not affected by breed or infection, the effect of breed on emissions per unit of digestible feed intake (g/kg DDMI, g/kg DOMI) approached significance, with Dorper breed having higher CH₄ yields than Red Maasai. Pinares-Patiño et al. (2013) from extensive measurements of CH₄ yield from sheep fed on temperate forages in respiration chambers reported that CH₄ yield is heritable and a major factor for differences between low and high CH₄ yield animals is the rumen retention time of the particulate fraction of digesta. In the same line, a study by Goopy et al. (2014) comparing high and low methane yield producing sheep reported that sheep that had a low methane yield also had lower mean rumen retention time and lower rumen volume and fill compared to high methane yield sheep despite having similar DMI and DM digestibility. These results could provide a plausible explanation for the differences observed between the Red Maasai and Dorper sheep in the present study, and further investigation would be beneficial in understanding these results. In this respect, a study conducted by Archimède et al. (2018) where a tropical indigenous (Blackbelly) and a highly selected temperate (Texel) sheep breeds were compared at tropical and temperate sites while fed on good (N fertilized) and poor quality of both tropical (C4) and temperate (C3) forages, revealed that CH₄ yield (g/kg DMI) was not affected by experimental site, whereas there was a breed × site interaction on CH₄ yield. In the tropical site, the CH₄ yield was consistently (all forage types and qualities) lower for the tropical breed than for the temperate breed, whereas the opposite was observed in the temperate site. The latter suggests that CH₄ emission is the result of complex interactions of factors such as diet, animal genetics and general environment.

In this study PCV at Day 1 (PCV Day 1) and PCV at Day 36 (PCV Day 36) differed with experimental periods (P), being higher in P1 than P2. It is difficult to ascertain the reasons for these observations. The depression in PCV from P1 to P2 may be an artifact of a continuous feeding on a low-quality diet rather than a more severe infection in P2 than in P1. The higher PCV Day 36 observed in uninfected lambs in P2 than in P1 and no change in PCV Day 36 of infected lambs support the above hypothesis.

There were period (P) effects on LW, LWG, digestibilities of dry matter (DMD), crude protein (CPD) and of acid detergent fibre (ADFD), being these values higher in P2 than in P1. The above may indicate that the improved LW and LWG in P2 was probably due to improved digestibilities rather than to improved feed intake. In fact, DMI during the first 26 days (pen period) and over the balance period (Days 29–33) were higher in P1 than in P2. In the same way, intakes of crude protein (CPI), N intake and retention, and intakes of digestible energy (DE) and metabolizable energy (ME) were higher in P1 than in P2. So, determining the causal reasons for improved LWG observed in P2 is not straight forward.

There were Breed×Period interaction effects on intakes of DM (DMI), NDF (NDFI) and ADF (ADFI), being these higher for Dorper in P1 than in P2. Accordingly, Dorper had higher daily CH₄ emissions (g/d) in P1 than in P2. However, there were complex Experimental treatment×Breed×Period interaction effects on CH₄ yields and Ym, with emissions being higher for Dorper, especially in P2; whereas emissions were the lowest for infected (I-adlib) Red Maasai in P2. Therefore, it can be postulated that the higher CH₄ (g/d) emissions observed in Dorper compared to Red Maasai was derived mostly from P1 and this was due to higher DMI and NDFI observed in this period. On the other hand, the higher CH₄ yields observed in Dorper, compared to Red Maasai arose principally from higher yields in P2 observed in Dorper. The feed used in this study was sourced from a single crop plot and was relatively uniform in quality.

The above-described interactions cannot be attributed to GIN infection. Firstly, as stated earlier, effects of parasite infection on feed intake, LWG, feed digestibility, nitrogen retention, diet metabolisability or methane emission were not evident. Secondly, the lambs in the infected group (I-adlib) in Period 1 had recovered from anemia and their PCV returned to normal after anthelmintic treatment and washout period. Regarding H. contortus infection, Albers et al. (1989) concluded that sheep recover rapidly from anemia after treatment, but effects of anemia on growth may continue for a few days post treatment, while compensatory growth and refilling of the gastrointestinal will also take place. Ascertaining the reasons for the above-described interactions are beyond the scope of this manuscript and it can only be hypothesized that it was a result from an artefact related to the low quality of feed as well as the contrasting adaptabilities of the sheep breeds.

The feed used in the present study was of a very poor quality (7.2% crude protein and 8.0 MJ ME/kg DM), whereas the lambs were in a physiological stage with high demand for nutrients. The feed was chosen to represent poor nutritional qualities of the feed commonly found in farming systems in SSA. Due to long-term natural selection under environmental stressors at specific environments, indigenous breeds of ruminants can cope with periods of nutritional stress (e.g., shortage of feed and quality deterioration) (De Tarso et al., 2016; Zhou et al., 2019). The latter may explain the inability of Dorper (introduced breed) to sustain consistent intakes of the low-quality diet over the extended experimental period (114 days including the quarantine period), as indicated by the drop in DMI from Period 1 to Period 2, whereas Red Maasai (indigenous) was able to cope.

It is generally accepted that a decrease in feed intake results in an increase in diet digestibility, due to a longer retention time of feed particles in the rumen, leading to a more efficient reduction in feed particles by comminution, hence favoring microbial digestion (Galyean and Owens, 1991). However, the above mechanism seems less marked with forage diets than with concentrate-rich diets (Grimaud and Doreau, 1995). Nevertheless, longer retention times are conducive to enhanced methanogenesis (Pinares-Patiño et al., 2003). The latter seems to be in line with the observed higher CH₄ emissions by Dorper lambs than by Red Maasai, but similar feed digestibilities between breeds.

Microbial protein is the main protein source for ruminants fed on diets low in nitrogen, with energy and N availability in the rumen being the most important limiting factors for microbial growth and hence rumen fermentation (Firkins, 1996). It is difficult to ascertain whether the drop in DMI and NDFI from Period 1 to Period 2 observed in Dorper lambs had an influential effect on the observed reduction in CPD and retained N from Period 1 to Period 2 in the ad libitum fed lambs. The rates of endogenous N loss and hindgut microbial protein synthesis could also have been altered by nutritional status (Giráldez et al., 1997), hence leading to decreased apparent CPD and retention of N from Period 1 to 2

Decreased nitrogen excretion by the kidney and recycling of urea to the rumen during undernutrition seems to be the mechanism by which some breeds and species of ruminants maintain digestion of poor-quality forages low in N content (Reynolds and Kristensen, 2008). Whether Red Maasai were better equipped than Dorper to maintain rumen fermentation from Period 1 to Period 2 is unknown.