Ethical considerations

This study was carried out in strict accordance with the recommendations of the Guide for the National Council for Animal Experiments Control of Brazil. The experiment was approved by the Committee on Ethics of Animal Experiments of the Federal University of Mato Grosso do Sul, Mato Grosso do Sul State, Brazil (Protocol Number: 367/2011).

Location and experimental field of Brachiaria spp. cultivars

This study was carried out at the Federal University of Mato Grosso do Sul in partnership with the Biotechnology Laboratory Applied to Animal Nutrition at Dom Bosco Catholic University and Embrapa Beef Cattle. Brachiaria grasses were evaluated in experimental plots at Embrapa Beef Cattle (latitude 20°27'S, longitude 54°37'W and 530 m altitude, located in Campo Grande, MS, Brazil). The type of soil in the study area was Dystrophic Purple Latosol alic.

The climate according to Köppen & Geiger⁵ classification is rainy tropical, AW subtype, characterized by a well-defined occurrence of a dry period during the colder months of the year (April-September) and a rainy season during the summer months (October-March) with an average annual rainfall of 1,469 mm and an average annual temperature of 23 °C.

Forages were evaluated during two consecutive summers (December-February), which is the rainy season in the Brazilian Cerrado, due to the seasonality of the forages. The experimental area consisted of 20 plots (experimental units, four plots/cultivar), measuring 4.0 × 4.0 m (16 m²). The plots were cut at 5 cm above the ground in order to standardize them for evaluation.

Five cultivars of Brachiaria grasses were evaluated in each plot: B. decumbens cv. Basilisk; B. decumbens access D70; B. humidicola cv. Tupi; B. humidicola cv. Common and B. ruziziensis access R124. Each plot was divided into two parts, to be cut with 21 and 42 d of regrowth. Forage samples were collected by cutting at 5 cm from the ground inside a square 0.5 × 0.5 m (0.25 m²) and sampled with five repetitions from the four plots of each cultivar, obtaining a composite sample from each forage.

Forage mass and growth determination

After forages were sampled, each was wrapped in a plastic bag and identified. In the laboratory they were weighed and divided into two parts: one to be processed as total sample and the other separated into leaf blade, stem and dead material⁶. Forage mass was estimated by the square method, with the quantification of forage, on a dry matter basis, sampled inside the 0.5 × 0.5 m square converted to metric ton (1,000 kg) per hectare (t ha^-1).

The leaf blade/stem ratio was obtained by dividing the mass of the leaf blades by the mass of the stems⁷. The vegetative canopy growth was determined at six different points in each experimental unit, which were marked for measurement according to the different growth ages evaluated⁷.

Forage chemical composition

After sampling and separating into total sample, leaf blade, stem and dead material, materials were pre-dried at 55°C for 72 h and ground to 1 mm with a Wiley Mill (Tecnal, Piracicaba City, São Paulo, Brazil) then stored in hermetic plastic containers (ASS, Ribeirão Preto City, São Paulo, Brazil) until analysis. Dry matter (DM) content (Method 967.03, AOAC⁸) and CP content (Method 981.10, AOAC⁸) were determined. For the determination of the NDF and acid detergent fiber (ADF) content, the methodology of Van Soest et al³ was used with modifications proposed in the ANKOM device manual (ANKOM Technology Corporation, Macedon, New York, USA).

In vitro dry matter digestibility (ivDMD) and in vitro neutral detergent fiber digestibility (ivNDFD)

The total samples (all canopy structures) of the different varieties of Brachiaria grasses of two regrowth ages (21 and 42 d) were submitted to in vitro tests, incubated with bovine or ovine rumen liquid (inoculum). The bovine inoculum was collected from three Nellore × Angus crossbred cattle and the ovine inoculum from five Dorper × Suffolk crossbred sheep, already fitted with ruminal silicon cannula and adapted to forage diet.

The in vitro digestibility of nutrients was determined according to the methodology of Tilley and Terry⁹ adapted for the ANKOM Daisy system (ANKOM Technology Corp., Macedon, NY, USA) as described by Holden¹⁰. Bags of non-woven fabric containing samples of Brachiaria grasses were placed on jars (with a limit of 30 bags per jar, two of them blanks) containing approximately 1.6 L of buffer solution¹¹. Bovine or ovine ruminal liquid (400 ml) was then added and purged CO₂ in the jars. The jars remained incubated with shaking at a constant temperature of 39 °C for 48 h, after which 40 ml of HCl (6 N) and 8 g of pepsin were added to each jar and left for another 24 h. When the incubation was finished, the jars were drained and the bags washed with distilled water and dried at 105 ºC for 16 h. They were then weighed to determine the post-incubation DM and submitted to NDF analysis³ with adaptation of ANKOM device manual (ANKOM Technology Corp., Macedon, NY, USA). The in vitro digestibility (ivD) coefficients for DM (ivDMD) and NDF (ivNDFD) were obtained through the equation:

ivD (g/kg) = [(incubated nutrient, g) - (residual nutrient, g - blank, g)] / (incubated nutrient, g) × 1,000.

Statistical analysis

All data were analyzed using the statistical package SAS® (SAS® Inst. Inc., Cary, NC, USA), and the means were compared using Tukey’s test with a significance level of P<0.05, and tendency considered at P<0.10.

To evaluate bromatological composition, production and regrowth of grasses, the following statistical model was used:

where

µ is the general average,

I_i is the effect of the i-th age (21, 42), G_j is the effect of the j-th grass (1,…, 5),

IG_k is the effect of interaction of i-th age with j-th grass,

e_ijkl is the random error.

To compare the values of ivD of DM and NDF carried out with different inoculums, the following statistical model was used:

where

µ is the general average,

I_i is the effect of the i-th rumen inoculum (1, 2),

G_j is the effect of the j-th cultivar (1,…, 5),

IG_k is the effect of interaction of i-th rumen inoculum with j-th cultivar,

e_ijkl is the random error.

Forage mass and growth determination

The forage mass and growth parameters are presented in Table 1. There was a significant effect (P<0.05) on forage mass from variety and age, leaf blade/stem ratio, leaf blade, stem, senescent material and growth. In addition, there was significant grass × age interaction (P≤0.0001) for all variables. The variety that presented the highest forage mass at 21 d was B. humidicola cv. Common, and at 42 d it was B. ruziziensis access R124 (8.86 and 12.81 g kg^-1, respectively; P=0.0001).

A lower leaf blade/stem ratio was observed in B. humidicola cv. Common at 21 and 42 d (0.51 and 0.30, respectively; P=0.0001) with similarity among the other grasses. However, the highest leaf blade/stem ratio at 21 d was observed in B. humidicola cv. Tupi (1.21 ratio; P=0.0001), whereas U. decumbens access D70 presented the highest ratio at 42 days of regrowth (1.51; P=0.0001). For stem results, B. humidicola cv. Common had significantly high values at 21 and 42 d of age (448.8 and 656.8 g kg^-1, respectively; P=0.0001).

In the production of senescent material in the studied varieties, the highest amount was observed for B. decumbens access D70 with 42 d of regrowth (684.8 g kg^-1; P=0.0001), and at 21 d B. decumbens cv Basilisk had numerically the highest value, close to B. decumbens access D70 (389.0 and 385.7 g kg^-1, respectively). In contrast, B. humidicola cv. Common and B. humidicola cv. Tupi had numerically the lowest senescent material at 21 d (220.2 and 276.3 g kg^-1, respectively).

The growth parameter was significant (P=0.00001) with highest growth shown by B. ruziziensis access R124 at 42 d (28.2 cm) and the lowest by B. humidicola cv. Common at 21d (11.6 cm).

Forage chemical composition

In relation to chemical composition, there was significant effect (P<0.05) from variety of grass and growth age on DM, CP, NDF and ADF in all sample types (total sample, leaf blade and stem; Table 2). Also, there was a significant grass × age interaction (P<0.05) for DM, CP, NDF and ADF in all samples, except DM and NDF from stem material, which showed no significant interaction (P>0.05), but a trend for NDF (P= 0.0985).

Comparing regrowth ages of total sample, forages at 21 d of regrowth showed lower values of DM, NDF and ADF (P<0.05); however, they presented higher CP values. As expected, CP from leaf blades was higher than from stem samples. The CP of leaf blades of B. ruziziensis R124 at 42 d of regrowth showed the highest amount (144.2 g kg^-1; P<0.05)

B. humidicola cv. Tupi presented higher NDF and ADF values for leaf blades at 42 d of regrowth (769.4 and 557.7 g kg^-1, respectively; P<0.05). The highest stem NDF values were also presented by B. humidicola cv. Tupi (884.6 g kg^-1; P<0.05). High ADF values were also observed for B. humidicola cv. Tupi and B. decumbens cv. Basilisk (384.7 and 367.6 g kg^-1, respectively; P<0.05).

In vitro dry matter digestibility (ivDMD) and in vitro neutral detergent fiber digestibility (ivNDFD)

There was no significant interaction (P>0.05) between inoculum and grasses. It was observed that ovine inoculum vs bovine inoculum resulted in higher ivDMD values for B. decumbens cv. Basilisk at 21 d (611.2 vs 571.9 g kg^-1; P=0.0200), B. humidicola cv. Tupi at 21 d (631.8 vs 568.4 g kg^-1; P=0.0370), and B. ruziziensis access R124 at 21 d (548.8 vs 613.4 g kg^-1; P=0.0420; Table 3). When evaluating only bovine inoculum, B. humidicola cv. Tupi and B. ruziziensis access R124 at 42 d of regrowth showed the lowest ivDMD values (544.3 and 542.0 g kg^-1, respectively; P=0.0078). When evaluating only ovine inoculum, B. ruziziensis access R124 at 42 d also presented the lowest ivDMD (535.0 g kg^-1; P=0.0184).

Regarding the ivNDFD, as presented on Table 4, ovine resulted in higher values than bovine inoculum on B. humidicola cv. Common at 21 d of regrowth (420.3 vs 369.5 g kg^-1; P=0.0040), and B. humidicola cv. Tupi at 42 d (490.8 vs 452.4 g kg^-1; P=0.0270). However, on B. decumbens cv. Basilisk at 42 d, the opposite was seen: ovine inoculum resulted in lower ivNDFD values than bovine (413.9 vs 472.4 g kg^-1; P=0.0150).

There was no effect (P>0.05) on ivNDFD from cultivar and regrowth age when the samples were incubated with ovine inoculum; however, the highest values were observed for B. humidicola cv. Tupi at 21 and 42 d (510.3 and 490.8 g kg^-1, respectively). When using bovine inoculum only the variety B. ruziziensis access R124 at 21 d presented ivNDFD values below the others (P=0.0044). At 21 d of regrowth the varieties that presented the higher values (P=0.0044) were B. decumbens cv. Basilisk, B. decumbens access D70, B. humidicola cv. Common and B. humidicola cv Tupi, B. decumbens access D70 being highlighted numerically, with the highest mean value. At 42 d B. decumbens cv. Basilisk, B. decumbens access D70, B. humidicola cv. Common, B. humidicola cv. Tupi and B. ruziziensis access R124 had statistically the highest values (P=0.0044), highlighting B. decumbens cv. Basilisk numerically with the highest mean.