Study area
The study was conducted at the Brazilian Agricultural Research Corporation – Embrapa Milho e Sorgo (Sete Lagoas, Minas Gerais, Brazil; 19º28′ S; 44º15′W, at 732 m altitude), where the Cratylia was planted for hay, and the Laboratory of Animal Metabolism and Calorimetry (LAMA/LACA) at the Veterinary School of the Federal University of Minas Gerais (EV-UFMG), Belo Horizonte, MG, where the lambs and infrastructure for the feeding experiment were located.
The unit where Cratylia was planted is located in a region with a Cwa climate type according to the Köppen classification system, i.e., savanna climate, with dry winters and humid-rainy summers. Cratylia was transplanted from a growth chamber to a research site on March 28th 2013, and several trials were conducted in the area until 2019 when the current research began. Cratylia’s shrubs were planted in an area of 105 m2, spaced at 1 m between rows and a total of 126 plants, or, 12,000 plants/ha− 1. The area size was determined based on the number of available Cratylia’s sprigs found available with research peers before planting, since commercial seeds are not available in Brazil. The soil test for the area where Cratylia was planted for hay is described in Table 1. No fertilization was made for Cratylia. Tifton 85 hay was purchased from a local dealer and presented a dry matter (DM) of 91.25%, CP 13.2% and NDF 68.98%, which was used to formulate the feed mixture. The agronomical accessions for the harvested Cratylia hay are presented in Table 2.
Table 1
Soil test result (00–20 cm) and planting date of Cratylia argentea.
Planting date | pH | NO3-N | P | K | OM |
----------------- (mg kg− 1) ------------------ | (dag kg− 1) |
28/03/13 | 5.5 | 18 | 2.18 | 39.26 | 3.44 |
Soil pH was measured using a 1:1 soil:water, NO3-N was measured using a 2M KCl extract, and plant available P and K were extracted using a Mehlich-3 solution (Mehlich, 1984).
Table 2
Agronomical accessions for the harvested Cratylia argentea hay.
| DM | CP | ADF | ADIN | NDF | NDIN |
| 93.2 | 19.6 | 41.6 | 0.492 | 58.4 | 2.28 |
sd | 1.28 | 3.45 | 5.52 | 0.100 | 5.40 | 1.19 |
DM, dry matter; CP, crude protein; ADF, acid detergent fiber; ADIN, acid detergent insoluble nitrogen, NDF, neutral detergent fiber; NDIN, neutral detergent insoluble nitrogen; sd, standard deviation. n = 530 |
Cratylia Hay And Feed Management
In order to establish a uniform regrowth, all Cratylia’s plots were mowed in January 2019. The first harvest for data collection happened in April of the same year and was repeated every other 90 days, until April 2020 (Table 3), leaving a 25 cm stubble height (Pinzón and Mariño, 2010), without the selection of less lignified material. During the haymaking period, temperatures were between 17.74 and 29.24°C, average rainfall was 885 mm, and relative humidity was 66.56% (NASA/POWER, 2022).
Table 3
Cratylia argentea harvest dates.
| 2019 | 2020 |
Harvest 1 | 04 − 02 | Harvest 4 | 01–09 |
Harvest 2 | 07–10 | Harvest 5 | 04–14 |
Harvest 3 | 10 − 07 | - | - |
In order to produce enough forage mass for the totality of animals, after each harvest, the produced hay was stored in raffia bags and packed in drums to avoid contamination, until the beginning of the feeding treatments at LAMA/LACA, which started in November 2020 and lasted until January 2020. All harvested material was cut through a forage chopper then spread in a 10-cm layer on a cemented area for drying in the sun, and turned frequently for drying to the point of hay, which happened when the Cratylia reached a DM content of 85%. The interval of exposure to the sun for the material to reach the point of hay varied between 48 and 72 h, depending on the weather conditions. During the night, or in case of rain, the material was covered with canvas to prevent moisture.
The Tifton 85 and Cratylia hay were ground to a particle size of 5 mm, homogenized, and then mixed with the other ingredients of the experimental diets according to the replacement levels (Table 4), as described by Lima (2016).
Table 4
Percentage and chemical composition of experimental diets.
| Replacement of Tifton 85 with C. argentea |
Variable | 0% | 20% | 40% | 100% |
Ingredient (g/kg DM) | | | | |
Cratylia argentea hay | - | 100.0 | 200.0 | 441.3 |
Tifton 85 hay | 549.9 | 425.4 | 300.7 | - |
Ground maize | 237.2 | 266.4 | 295.6 | 366.1 |
Soybean meal | 175.4 | 170.9 | 166.4 | 155.5 |
Mineral for sheep1 | 15.0 | 15.0 | 15.0 | 15.0 |
Limestone | 10.5 | 10.5 | 10.8 | 11.3 |
Dicalcium phosphate | 7.0 | 6.8 | 6.5 | 5.8 |
Sodium bicarbonate | 5.0 | 5.0 | 5.0 | 5.0 |
Chemical composition | | | | |
Dry matter (g/kg) | 908.4 | 913.2 | 903.3 | 903.0 |
Organic matter (g/kg DM) | 919.9 | 923.1 | 922.4 | 923.9 |
Ash (g/kg DM) | 80.1 | 76.9 | 77.6 | 76.1 |
Crude protein (g/kg DM) | 178.4 | 188.8 | 180.7 | 182.2 |
Ether extract (g/kg DM) | 24.9 | 13.0 | 16.3 | 17.7 |
Neutral detergent fiber (g/kg DM) | 433.4 | 426.0 | 390.1 | 352.5 |
Total carbohydrates (g/kg DM) | 716.6 | 721.4 | 725.4 | 724.0 |
Tortuga® (DSM) Mineral for sheep (content per kg of product): zinc − 3,800 mg; sodium − 147 g; manganese − 2,000 mg; cobalt − 15 mg; copper − 590 mg; sulfur − 18 g; selenium − 20 mg; iodine − 50 mg; chromium − 20 mg; molybdenum − 300 mg; calcium − 110 g; calcium (max.) − 135 g; fluorine (max.) − 870 mg; phosphorus − 87 g. DM – dry matter, NM – natural matter |
Feeding Trial
Twenty-four Lacaune lambs aged between five and six months (21.5 ± 3.38 kg) were divided into groups according to their initial weight (Table 5). The lambs were housed in individual metabolic cages (1.5 × 0.7 m) equipped with a feeding trough, a drinker, and a system for separate collection of feces and urine, inside a covered masonry shed with an exhaust system at LAMA/LACA laboratories.
Table 5
Block distribution of lambs.
Block | Category | | Mean weight (kg) | Standard deviation | Replicates per treatment |
1 | Light | | 15.9 | 0.580 | 1 |
2 | Medium-light | | 20.6 | 0.970 | 2 |
3 | Medium-heavy | | 23.5 | 0.640 | 2 |
4 | Heavy | | 26.0 | 1.21 | 1 |
The total experimental period was 64 days, the first 21 of which were used as a period of adaptation to the experimental diets and cage management, followed by 43 days of intake and weight gain (WG) measurements. Digestibility trials were conducted in two periods, here entitled “initial” and “final”. The initial period consisted of 7 days of total collection of feces and urine, which began right after the adaptation period (5 to 6 months, 26.8 ± 3.26 kg), and the final period began in the last 7 days of the experiment (6 to 7 months, 37.5 ± 3.26 kg).
Feed was supplied twice daily (08h00 and 14h00) for ad libitum intake (200 g/kg orts, as-fed basis). Supplied feed and orts were weighed daily on an electronic scale. Each of the diets was composed of different levels of Cratylia hay to replace Tifton 85 hay (Table 4). The diets were formulated to meet the requirements for finishing lambs with a DM intake of 4% BW and an average daily gain (ADG) of 200 g, thus, a 16% CP and total digestible nutrients (TDN) of 75%, following the NRC (2007). All diets were supplied as a total mixture ratio and adjusted daily allowing up to 20% orts.
Live weight was measured weekly from the beginning of the experimental period, in the morning period before the feed was supplied, using a platform scale. Initial live weight (ILW) was determined immediately after the end of the adaptation period, and final live weight (FLW) at 43 days of the experiment. The ADG was determined by the following equation: ADG = ((FLW– ILW)/(Number of days in the feedlot). Dry matter intake (DMI) was individually calculated as the difference between the amount of feed supplied and daily orts.
The digestibility trial involved total collection of feces and urine on two occasions: (1) in the first week and (2) in the last week of the experimental period. Samples of supplied feed, orts, and feces were analyzed for the dry matter (DM), ash, and nitrogen (N) contents according to AOAC (1990) and the neutral detergent fiber (NDF) content according to Van Soest et al. (1991). The total tract apparent digestibility coefficient (ADC) was determined as follows: ADC = ([OF – Intake – CF]/OF – Intake) × 100, which was proposed by Maynard et al. (1984), where AD = apparent digestibility; OF = offered feed [(offered feed amount in kg DM) x (offered nutrient content in % DM)/100]; Intake = orts [(removed orts feed in kg DM) x (orts nutrient content in % DM)/100]; and CF = fecal output [(amount of collected feces in kg DM) x (collected feces content in % DM)/100].
Urine was collected in plastic containers containing a hydrochloric acid solution (6N HCl) to keep the urine pH below 2 and prevent the loss of nitrogen by volatilization. The total collected material was quantified, homogenized, and filtered through cotton gauze and a 10% sample was taken and stored at -18°C. The nitrogen content of the sample was determined by the micro Kjeldahl method (Campos et al., 2004). Nitrogen balance was calculated according to Gentil et al. (2011).
Statistical Analyses
The 24 lambs were allocated in a randomized block design with four blocks, where each block was randomly assigned with a dietary treatment of 0, 20, 40, and 100% Cratylia hay replacing Tifton 85 hay in the diet, and six replicates. The effect of treatments was submitted to an analysis of variance (ANOVA), with the following general model:
$$Yijkl=\mu +\tau i+\gamma k+eikl+\beta j+\left(\tau \beta \right)ij+ϵijkl$$
,
where Yijkl = value observed at the Tifton 85 replacement level for Cratylia hay level I, in block k, replicate l, and subplot j; µ = mean overall effect; τi = effect of Tifton 85 replacement level for Cratylia hay i; γk = effect of block k; ϵikl = error attributed to the plot Yikl (ϵik ~ N [µ, σ2]); βj = effect of collection period j; (τβ)ij = interaction between Tifton 85 replacement level for Cratylia hay i and collection period j; and ϵijkl = error attributed to subplot Yijk (ϵijkl ~ N [µ, σ2]).
For performance variables, the fits of polynomial regression models were tested to obtain the equation as a function of the Tifton 85 replacement levels for Cratylia hay or collection periods.
Analyses were carried out using R statistical analysis software (Core Team, 2021). For all analyses, the statistical assumptions of normality and homoscedasticity were evaluated by the Shapiro-Wilk and Bartlett tests, respectively. An effect was considered significant when the α error was less than 5%.