Study area and housing

The study was done at the sheep experimental unit of the National Institute of Forestry, Agricultural and Livestock Research (Instituto Nacional de Investigaciones Forestales Agrícolas y Pecuarias - INIFAP), in Huimanguillo, Tabasco, Mexico (17°50’ N, 93°23’ W). Regional climate is warm humid with year-round rains [Af (m)] and a 27.8 °C average annual temperature¹⁶. During the study period, minimum and maximum temperature was measured daily at 0800 h (24 hours) with a Six’s type thermometer. Weekly and general averages were calculated from this data: average minimum temperature was 23.0 ± 1.1 and average maximum was 35.5 ± 2.1 °C. The E. americana foliage was fed to sheep in individual pens with a 2.4 m² area, a concrete floor, drinker, feed trough and asbestos sheet roof.

Animal handling

Experimental animals were nine non-gestating, non-lactating Blackbelly x Pelibuey ewes distributed into two groups based on live weight category (LWC) and age. Four animals of two years of age were included in the light LWC (22.2 ± 1.2 kg LW), while five animals of three years of age were included in the heavy LWC (34.4 ± 1.1 kg LW). Prior to beginning the experiment each animal was injected with 1 ml ADE vitamins (Vigantol Bayer^®) (500,000 IU vitamin A, 75,000 IU vitamin D, 50 mg vitamin E per ml). Animal management practices complied with established institutional guidelines (Reglamento para el Uso y Cuidado de Animales Destinados a la Investigación en el Colegio de Postgraduados, CP-02.11.16).

The experimental period was 42 days, including 14 d of adaptation to the diet and 28 d for the feeding trial. Initially, all the sheep were allowed to graze Cynodon plectostachyus pasture from 0800 to 1300 h after which each group was placed in a pen with free access to E. americana foliage (300 g sheep^-1 d^-1), water and mineral salts (Magnophoscal^®, phosphorus 17.5 g; calcium 6.5 g; sodium 10.5 g; magnesium 4.5 g; sulfur 2.0 g). The grazing period was gradually reduced at a rate of one hour every two days until the sheep spent all day in the pen. The amount of E. americana foliage provided was increased by 100 g sheep^-1 day^-1 until it became the sole feed source. During the 28-day feeding trial each animal was housed in an individual pen and provided free access to water and E. americana foliage from 0800 to 1800 h, supplied at a rate that resulted in at least 10% rejected feed. When the feeding trial ended all the light LWC sheep were housed in one pen and all the heavy LWC sheep in another. They were kept there overnight for safety, with free access to water and mineral salts.

Collection and chemical analysis of E. americana foliage

Foliage (leaves and petioles) collection was done during the late northwinds season and early dry season (February-March 2017). Collections were taken from trees with no history of pruning and which formed part of live fences marking sheep pastures. Pruning shears and a machete were used to remove branches and the foliage then removed from them. The collected foliage was dried by spreading in layers no thicker than 3 cm on a concrete floor under a roof at room temperature (28.2 ± 1.3 °C) for 72 h. It was turned over and mixed twice daily to improve drying.

Evaluated variables

Data on chemical composition of and secondary metabolites in the foliage of E. americana was collected from foliage samples taken every week during the experimental period. Duplicate analyses were done of dry matter content (DM); ash; organic matter (OM) and crude protein (CP) using AOAC methods¹⁷. Other techniques were used to quantify neutral detergent fiber (NDF) and acid detergent fiber (ADF)¹⁸; in situ DM degradation (IDMD)¹⁹, metabolizable energy (ME, Mcal kg^-1 DM)²⁰, total polyphenols (g kg^-1 DM), non-tannin phenols (g kg^-1 DM)²¹, condensed tannins (CT; g kg^-1 DM), hydrolysable tannins (g kg^-1 DM) and total tannins (g kg^-1 DM)^22,23,24.

The nylon bag technique used to quantify foliage IDMD¹⁹, was implemented with three male cattle (Bos indicus x Bos taurus)(average LW = 500 ± 20 kg), that had been castrated and fitted with a rumen cannula. They were grazed in a pasture containing predominantly white gramalote grass (Paspalum fasciculatum) which provided 22.04 % DM, 6.73 % CP, 78.55 % NDF and 53.8 % ADF. This was supplemented with 2 kg of feed consisting of 70 % chickpea, 20 % rice polish and 10 % molasses (83.87% DM, 20.80 % CP, 28.43 % NDF and 7.42 % ADF). For the IDMD trial 5 g dry and ground E. americana foliage (Thomas-Willey mill, model 4 Laboratory Mill) was incubated in each rumen cannula with a 2 mm sieve, in polysilk bags (10 x 20 cm, 45µm porosity), in duplicate for 24 h. The bags were removed, washed with running water, and dried in a forced air oven at 105 °C for 72 h. Dry matter degradation was calculated with the formula: (g initial DM - g residual DM) / (g initial DM) X 100.

Changes in live weight (LW) were quantified by weighing the sheep on two consecutive days at 14-d intervals during the feeding trial. A platform scale (Oken^®) was used with 200 g accuracy.

Daily weight gain (DWG) was calculated by the difference of final weight minus initial weight divided by the number of days in the experimental period. Foliage offered and rejected was weighed per pen weekly over three consecutive days.

Weekly intake per pen was calculated by subtracting the quantity offered and that rejected. The foliage consumption index (%) was calculated at seven-day intervals for four periods by multiplying total DM intake by 100 and dividing by animal LW.

Mineral salt intake was measured by weighing the salt offered and rejected per pen weekly over three consecutive days. Weekly intake per pen was calculated by subtracting the quantity offered and that rejected.

Crude protein (CP), metabolizable energy (ME) and condensed tannins (CT) intake were calculated by multiplying total DM intake by foliage nutrient content and dividing by 100.

Body condition (BC) was calculated at 14-day intervals during the feeding trial using a one-to-five scale²⁵.

FAMACHA^© monitoring was done based on ocular mucosa color as determined by the five-point FAMACHA^© card color scale²⁶; intense red corresponds to a value of 1 (healthy) while white corresponds to a value of 5 (anemia = heavily infected). This evaluation was performed by the same person at the beginning and end of the feeding trial.

Blood variables were measured by taking blood samples with the jugular vein puncture technique and collecting the blood in 4 ml vacutainer tubes containing EDTA. Blood collection was done at 0800 h on days 1 (beginning) 14 and 28 of the feeding trial. Blood samples were transferred to the laboratory for analysis in an automated hematology device (Medonic CA 620/530). The components analyzed included red blood cells (x10¹² L), hemoglobin (g dl^-1), hematocrit (%), mean red blood cell corpuscular volume (x 10¹⁵ L), white blood cells (x 10⁹ L), lymphocytes (x 10⁹ L) and granulocytes (x 10⁹ L).

Gastrointestinal nematode egg counts were done be taking fecal samples from each sheep at 0700 h at 14-day intervals throughout the experimental period. Each sample was collected directly from the rectum of each animal using a plastic bag and 2 g processed with the McMaster technique to determine egg count per gram of feces (EGF)²⁷.

Experimental design and statistical analysis

During the feeding trial (28 d) measurements were taken of the same animals at 7- and 14-d intervals, therefore evaluation period (EP) was considered an independent variable. An experimental two-factor design was used with repeated measures in one factor²⁸. The first factor was sheep LWC (light and heavy). The second factor was the EP (two 14-d periods to evaluate changes in LW, and four 7-d periods to evaluate changes in nutrient intake). The experimental unit was one sheep.

Statistical analyzes were done with the SAS statistical package²⁹. Descriptive statistics (mean ± standard deviation) were used to describe the values of E. americana foliage chemical composition, IDMD, DM, phenols and their fractions, as well as daily mineral salt intake. The Shapiro-Wilk test was applied to the remaining data to verify normal data distribution and the Levene test to confirm variance homogeneity. The EGF results were transformed with the natural logarithm (Log EGF +1) to provide them a normal distribution. The experimental unit was one sheep. Statistical analyzes of LW and total DWG were run with the GLM PROC. The PROC MIXED³⁰ was applied to identify the influence of EP, LWC and their interaction on DM, CP, ME and CT intakes, DWG, IDMD, EGF, and the blood variables. The means were compared with a Student t test and least mean squares using the pdiff option in SAS. The body condition (BC) and FAMACHA variables were analyzed with the Wilcoxon rank sum test for unpaired data³¹.

Dry matter and nutrient intake

No interaction (P>0.05) was found between the studied factors. The light LWC sheep exhibited higher DM, CP, ME and CT intakes than the heavy LWC sheep (Table 2). This higher nutrient intake in the light sheep can be attributed to their not yet having reached mature weight^37,38. Concentrations of CT greater than 5% in diet DM and the high effectiveness of CT in forming complexes with saliva proteins can reduce DM intake in sheep³³. However, no reduction in DM intake was observed, probably because the E. americana used in the present study had a low CT level.

The IDMD, and DM, nutrient and CT intakes increased up to the third week (Figure 1), after which nutrient intake remained constant. In a study of male Blackbelly sheep fed E. poeppigiana forage intake index was 3.5 %¹¹, which is higher than in the present results. Differences in IDMD between studies can be attributed to differences in Erythrina species chemical quality, sheep sex and breed^11,39. Another consideration is that when sheep are fed diets containing CT, feed intake may decline due to an astringent reaction to the feed associated with formation of CT-protein complexes and reduction in IDMD³³. This could at least partially explain the lower DM intake observed in the present study during the first two weeks of the trial. This response suggests that the sheep and their rumen reticulum microbes required a period of two weeks to adapt to the new diet. When sheep eat a diet containing CT their salivary glands produce proteins that bind to both CT and hydrolysable tannins, thus making them more tolerable^33,40. In addition, ruminants exposed to diets with CT can develop microbe populations with the ability to alter and degrade CT, thereby preventing the animal from experiencing reductions in DM intake and/or IDMD^33,41.

▲ Each value is the least mean square (± SE) of nine animals.

^abc Different lowercase letters indicate significant difference (Student t test), P<0.05.

Figure 1 Erythrina americana foliage intake (g DM kg^-1 LW) during feeding trial in Blackbelly x Pelibuey sheep 

Mean (± SE) daily mineral salt intake was 23.7 ± 1.1 g animal^-1 in the light LWC sheep and 17.6 ± 2.1 g animal^-1 in the heavy LWC sheep. Mineral salt intake stabilized in both groups between the second and fourth week (Figure 2). Ash content (<10 %) in the E. americana foliage (Table 1) was within ranges reported for E. americana foliage of different regrowth ages^2,13,32. Foliage from this tree legume has a lower ash content than tropical grasses such as C. nlemfuensis and Panicum maximum^42,43. Consequently, it is important to provide mineral supplementation to sheep fed E. americana. In addition, the CT present in legumes can form complexes with some minerals, reducing their availability³³. There are no studies to date on the mineral requirements of hair sheep under different feeding scenarios (grazing, penned)⁴⁴, but sheep fed E. americana exhibit higher mineral salt intake than hair sheep in a grazing system under a different feed supplementation regime^38,45,46. Differences between studies can be attributed at least in part to sheep LW, feed supplement level and composition, mineral salt composition and season.

▲ Each value is the average (standard errors) of four sheep.

● Each value is the average (standard errors) of five sheep.

Figure 2 Mineral salt intake (g d^-1 sheep^-1) by live weight category in Blackbelly x Pelibuey sheep fed Erythrina americana 

Changes in live weight

Sheep LWC and the LCW/EP interaction did not affect (P>0.05) DWG. However, EP alone did affect (P<0.01) DWG, since it was lower at fourteen days than at 28 d (Figure 3). The negative DWG observed at 14 d was probably due to lower nutrient intake (Figure 1). At 28 d, however, DWG increased substantially, a response that can be attributed to higher nutrient intake and compensatory growth^47,48.

▲ Each value is the mean square (± standard errors) of nine sheep.

^ab Different letters on the same line, indicate difference (test of "t"), P<0.01.

Figure 3 Daily weight gain over time in Blackbelly x Pelibuey sheep fed Erythrina americana foliage 

Total DWG, BC and FAMACHA index values were unaffected by LWC (Table 3). Total DWG in the present results was positive and higher than reported in male Pelibuey lambs fed Pennisetum purpureum and E. poeppigiana¹² but lower in growing Blackbelly male lambs fed only E. poeppigiana foliage. These differences in DWG between studies can be attributed to animal age, sex and breed¹¹.

Blood components

No interaction (P>0.05) was found between the studied blood components. Day number (i.e. 1, 14 and 28 d) only affected (P<0.05) hemoglobin and hematocrit levels. The general mean (± SD) was 8.1 ± 1.5 (x 10¹² L) for red blood cells; 10.2 ± 1.6 (g dl^-1) for hemoglobin; 25.1 ± 4.1 (%) for hematocrit; 31.2 ± 1.9 (x 10¹⁵ L) for mean red blood cell corpuscular volume; 10.2 ± 2.8 (x 10⁹ L) for white blood cells; 7.3 ± 2.2 (x 10⁹ L) for lymphocytes; and 1.0 ± 0.3 (x 10⁹ L) for granulocytes. The least mean square (± SE) for hemoglobin (g dl^-1) was 11.2^a ± 0.5 at d 1, 10.1^b ± 0.5 at d 14 and 9.4^b ± 0.5 at d 28. For hematocrit they were 22.5^b ± 1.3 at d 1, 26.6^a ± 1.3 at d 14 and 26.9^a ± 1.3 at d 28. The blood variable values observed in the present results are within the ranges reported for grazing hair sheep in tropical regions⁴⁹. Erythrina americana foliage intake levels during the four-week feeding trial maintained blood variables at levels appropriate for non-gestating, non-lactating sheep.

Gastrointestinal nematode eggs

None of the studied factors affected EGF (P>0.05), and the overall mean (± SD; unprocessed data) for EGF was 264 ± 670. Consumption of CT in the diet (15 % CT from Acasia molissima, based on DM) can reduce EGF in small ruminants^15,33, but the CT intake level in the present study was insufficient to detect any changes in EGF that could be attributed to LWC or EP. In addition, in the present study the sheep were fed E. americana in pens and were thus prevented from sustaining a natural nematode infection level, explaining in part the low EGF values.