Twenty-five-centimeter pre-grazing canopy height in palisade grass and forage peanut

Short-term grazing behavior variables are sensitive to the canopy structure and have an impact on daily forage intake. This study evaluated the effect of preand post-grazing canopy heights on the forage harvesting process at a patch scale in a mixture of Brachiaria brizantha (Hochst. ex A. Rich.) Stapf. syn. Urochloa brizantha R.D. Webster cv. Marandu (palisade grass) and Arachis pintoi Krapov. & W.C. Greg. cv. Belomonte (forage peanut). Treatments were allocated to a split-plot arrangement in a completely randomized design. The plots, in their entirety, consisted of two pre-grazing canopy heights: 25 cm (CH25) and 35 cm (CH35); subplots consisted of three levels of defoliation severity: no defoliation (DS0); 20% depletion of pre-grazing canopy height (DS20); and 40% depletion of pre-grazing canopy height (DS40), with eight replications. Heifers were allowed to graze the patches (0.7 × 0.7 m) and their grazing behavior was recorded. Canopy structure measurements were taken both before and after grazing. Patches from CH35 presented greater stem mass for grass (p = 0.001) and legume (p = 0.002) than did patches from CH25. Bite rate, bite mass and instantaneous intake rate were greater for CH25 than for CH35 (p < 0.001, p = 0.068, and p = 0.074), and bite mass and instantaneous intake rate were lower for DS20 compared to DS0 (p = 0.032 and p = 0.016). Greater stem mass in the grazing strata negatively influenced the instantaneous intake rate.


Introduction
The canopy structure is continuously changing as a result of plant growth, defoliation, and senescence (Mezzalira et al., 2014). It modulates forage harvesting during the grazing process (Benvenutti et al., 2016). Ungrazed canopies facilitate forage harvesting; however, forage prehension by grazing livestock throughout the occupation period mainly in tropical grass gradually becomes increasingly difficult. Short-term grazing behavior variables are sensitive to the canopy structure and have a pronounced impact on daily forage intake (Carvalho et al., 2015). In rotational stocking, there is typically a decrease in leaf/ stem ratio throughout the stocking period (Fonseca et al., 2013). Consequently, there is a progressive reduction in short-term intake rate and bite volume (Fonseca et al., 2012;Fonseca et al., 2013), which can result in a reduction in daily herbage intake.
In grass monoculture canopies, a reduction in the short-term intake rate was observed when grazing down was higher than 40 % of the pre-grazing canopy height, negatively affecting the forage intake (Fonseca et al., 2013). As leaf blades deplete with grazing, the animals will spend more time searching for green leaf blades within a canopy filled with stem and dead material, which results in a decrease in bite rate and intake rate (Fonseca et al., 2013). A better understanding of the forage harvesting process in the short-term may reveal opportunities for manipulation so as to achieve greater harvesting efficiency.
In mixed pastures of grass and legume, animal preferences also influence the forage harvest process (Tsutsumi et al., 2016). Thus, the pre-grazing canopy height has an even more important effect on animal behavior because it affects not only the canopy structure, but also the botanical distribution between grass and legume. For instance, pastures of palisade grass and forage peanut managing pre-grazing canopies less than 25 cm in height have shown a greater proportion of legumes than pastures under 35 cm pre-grazing canopy height (Gomes et al., 2018). We hypothesized that animals grazing mixed canopies with greater pre-grazing heights would have their instantaneous intake rate reduced by an increased presence of stems. Additionally, more severe grazing intensity would also reduce the instantaneous intake rate due to a negative impact on bite mass and rate. Therefore, the objective of this study was to evaluate the effect of a different canopy structure, achieved by altering pre-grazing canopy heights and defoliation severity, on short-term ingestive behavior at a patch scale, in mixed grass-legume pastures.

Experimental area
The experiment was conducted at the research farm of the Federal University of Lavras, Brazil (21°14'08.94" S, 44°58'06.96" W; altitude of 923 m), during the rainy season (spring and summer) of 2016-2017. The experimental area consisted of mixed pastures of palisade grass and forage peanut. In Nov of 2016, the experimental area was standardized by mechanical cutting at 10 cm of stubble height and the forage

Animal Science and Pastures
Research Article Short-term intake on mixed pasture Sci. Agric. v.79, n.2, e20200090, 2022 removed from the experimental area. Subsequently, the experimental area was divided into 16 whole plots of 3.5 × 3.0 m. Each whole plot was subsequently divided into three subplots of 3.5 × 1.0 m. Each subplot was further divided into three patches of 0.7 × 0.7 m, creating ninepatches per whole plot (Figure 1).

Treatments and experimental design
The experimental design was a completely randomized design with a split-plot arrangement (2 × 3). Two pre-grazing canopy heights (CH) were studied in the whole plot: 25 (CH25) or 35 cm (CH35). These pregrazing heights were chosen drawing on the results of Gomes et al. (2018), in which CH25 corresponded to light interception between 90 and 95 %, and CH35 corresponded to maximum light interception. This difference in canopy heights generated contrasting canopy structures. Three levels of defoliation severity (DS) corresponding to phases of the stocking period were investigated in the subplots: 1) no defoliation (DS0); 2) 20% depletion of pre-grazing canopy height (DS20); and 3) 40% depletion of pre-grazing canopy height (DS40).

Preparation of patches
When the whole plot reached the pre-grazing height, the subplots were randomized, followed by the preparation of grazing patches. In each subplot there were three patches, delineated with frames measuring 0.7 × 0.7 m. All biomass located outside the delineated areas was cut at ground level and removed from the subplots. After the patch preparation, all with similar canopy height within each whole plot, the defoliation severity was established.
Thus, the forage of each patch was defoliated manually before the grazing (i.e., hand-plucking; Vries, 1995) to the desired canopy heights of DS20 and DS40 (Figure 1).
After all the patch preparation, three corridors were created within each whole plot, using fences. In each corridor, there were three patches with different DS in random order. In two corridors grazing animals had unique access to each patch (Grazed; Figure 1). To facilitate an evaluation of the canopy structure before and after grazing and the forage removed by grazing animals in each treatment, there was no grazing in the third corridor (Ungrazed). The patch-grazing data were considered as the average of the two corridors where grazing occurred. After the preparation of all patches, the ingestive behavior was evaluated.

Canopy structure measurements
Both before and after grazing, the canopy heights were measured at eight points within each patch using the canopy surface height method (Braga et al., 2009). After grazing in two corridors, herbage mass was measured in all patches (grazed and ungrazed). In each patch, a frame was placed for canopy stratification. The upper stratum was the 0 to 20 % forage of the initial canopy height. The second and third sward strata were between the layers of 20 to 40 % and 40 to 60 % of the initial canopy height, respectively. The fourth stratum was considered as forage less than 60 % of the initial canopy height up to ground level (Fonseca et al., 2012). All the forage contained in each stratum was cut using the pruning shears. After harvesting the forage, botanical and morphological separations were performed. Grass samples were separated into stem (stem + sheath), leaf (leaf blade), and dead material. Legume samples were separated into stolon, leaf (stipule + petiole + leaflet), and dead material. Forage samples were oven-dried at 55 °C for 72 h to constant weight. The forage mass (g m -2 ) of the botanical and morphological components was calculated for each stratum. Total herbage, grass, and legume masses were the sum of the masses in each stratum. The masses of patches that were not grazed were considered as the pre-grazing mass. The herbage, leaf, and stem bulk density above 40 % sward strata were calculated by dividing the respective mass above 40 % canopy strata (g m -2 ) by its height (m).

Ingestive behavior
Two Bos indicus (Tabapuã) heifers with 450 ± 13 kg of body weight were allowed to graze the patches (one in each corridor). The heifers had been previously conditioned to the experimental conditions, and trained for grazing in corridors. The animals were used as tools for forage harvesting. Heifers had no access to feed for 12 h before grazing. During the grazing of all patches, the animals were individually recorded using a digital video camera. All the videos of each animal were analyzed. In each patch, the total number of bites was measured. Grazing time in each patch was considered from the moment of the first bite in the patch until the animal moved to the next patch. A feeding station was defined as the patch area directly in front of and on each side of the animal when its front feet are stationary (Laca et al., 1992). The bite rate (bites min -1 ) was calculated by dividing the number of bites by the respective grazing time in each patch. The feeding-station per min (FS min -1 ) was calculated by dividing the number of feeding-stations by the time spent grazing. The number of bites per feeding-station was calculated by dividing the number of bites in each patch by the number of feeding-stations (Spalinger and Hobbs, 1992).
The ingested forage mass was calculated by the difference between forage mass in the ungrazed patch and the forage mass of the patch after grazing. The bite mass was calculated by dividing the ingested forage mass by the number of bites in each patch. The instantaneous intake rate was calculated by the ingested forage mass divided by the time spent by the animal in each patch (Benvenutti et al., 2006;Benvenutti et al., 2008). Bite mass, and instantaneous intake rate were not measured in the DS40 patches because the animals spent minimal time grazing this treatment.

Statistical analysis
Data were analyzed using the mixed model method (Littell et al., 2000), using the MIXED procedure of SAS (SAS Institute, version 9.0). The effects of pregrazing height, defoliation severity and their interaction were considered fixed. All variance components were estimated using the restricted maximum likelihood method. The treatment averages were estimated using the Least-Squares Means (LSMEANS) statement and compared using Fisher's protected least significant difference (LSD) with p ≤ 0.10. The statistical model for data analysis was as follows: where Y ij = value observed in the ith CH of the jth DS; μ = overall average; CH i = fixed effect associated with the ith pre-grazing heights, i = 1, 2; e i = random error associated with the ith pre-grazing heights and repetitions; DS j = fixed effect associated with the jth levels of defoliation severity, j = 1, 2, 3; e i = random error associated with the ith pre-grazing heights, jth levels of defoliation severity and repetitions. The fixed effect of strata was included in the statistical model to run canopy strata data. Strata interactions with other variables were also considered in the model. The ungrazed data were not evaluated statistically, since the data corresponding to the initial canopy condition were gathered before grazing.

Canopy structure
The canopy structure data in the ungrazed and grazed patches are shown in Table 1. The canopy height in the ungrazed patches was established within the target height by hand-plucking. There was interaction between CH and DS at the canopy height of grazed patches (p = 0.032). The canopy height in grazed patches was greater in CH35 than in CH25 in all DS. The percentage of the height removed by grazing was similar in the CH25 and CH35 patches (on average, 11.3 % of the initial height). The canopy height in the grazed patches was greater in DS0 and reduced with increasing DS. The percentages of the height removed by grazing were on average 18.9, 9.6, and 2 % of the initial height in the DS0, DS20, and DS40, respectively.
Total herbage mass, grass leaf mass, and legume leaf mass in the grazed patches were influenced by the interaction between CH and DS (p = 0.066, p = 0.040 and p = 0.066, respectively). In DS0, total herbage, grass leaf, and legume leaf masses were greater in CH35 than CH25. However, there was no effect of CH on DS20 and DS40. In the CH25 target, total herbage and leaf legume masses in grazed patches were less in DS40 than in DS0 and DS20. In CH35, total herbage and leaf legume masses decreased progressively in the greater DS. Grass leaf mass in the grazed patches decreased progressively in the greater DS in both CH.
There was an interaction between CH and DS in the stem grass mass (p = 0.099). In all DS, greater stem mass was observed in CH35 than CH25. Post-grazing stem mass in CH35 was 156 %, 102 % and 80 % in DS0, DS20 and DS40, respectively, then in CH25. In CH25, there was no difference in stem mass between the different DS treatments. In CH35, stem mass was higher in DS0 than in DS20 and DS40. Legume stem mass was influenced by both CH (p = 0.002) and DS (p = 0.026). The greatest stem legume mass was observed in CH35, while DS0 and DS20 were greater than DS40. The botanical composition (proportion of legume in the canopy) was not influenced by either CH and or DS (p = 0.633 and p = 0.372, respectively). The mean botanical composition was 57 % of forage peanut on herbage mass.
Herbage bulk density above 40 % sward strata was influenced by CH (p = 0.063) and DS (p = 0.018). The greatest herbage bulk density above 40 % sward strata was found for CH35 and DS0. There was no CH effect (p = 0.867) on leaf herbage bulk density above the 40 % sward strata. However, the highest value of leaf herbage bulk density above 40 % sward strata was found for DS0 (p = 0.001). The highest stem herbage bulk density above 40 % sward strata was observed in CH35 (p < 0.001). No differences were found in stem herbage bulk density above the 40 % sward strata for DS (p = 0.944). Short-term intake on mixed pasture Sci. Agric. v.79, n.2, e20200090, 2022

Canopy stratification
Total herbage mass, grass mass, and legume mass per stratum of the ungrazed and grazed canopies in CH25 are shown in Figures 2A, B, C, 3A, B, C, and 4A, B, C. In DS0, total herbage mass and legume mass were higher in the ungrazed patch than in the grazed one, in the 0-20, 20-40 and 40-60 % of the sward strata. The grass mass was lower in grazed patches than in ungrazed ones in the 0-20 and 20-40 % of the sward strata. For DS20, total herbage mass and legume mass decreased in the 20-40 and 40-60 % of the sward strata in the grazed plot. Grass mass was no different in either grazed or ungrazed patches of DS20. In DS40, no difference was observed in total herbage, grass, and legume masses (p = 0.678, p = 0.562 and p = 0.789, respectively) between grazed and ungrazed patches.
Total herbage mass, grass mass, and legume mass per stratum of the ungrazed and grazed canopies in CH35 are shown in Figure 2D, E, F, 3D, E, F, and 4D, E, F. In the DS0, total herbage mass, grass mass, and legume mass decreased in grazed patches relative to ungrazed ones in the 0-20 and 20-40 % of the sward strata. In DS20, the total forage mass and grass mass were lower in the grazed patches than in the ungrazed ones in 20-40 % of the sward stratum. Legume mass was no different in either grazed or ungrazed patches of DS20 (p = 0.213). In the DS40, there was no difference in total herbage, grass, and legume masses (p = 0.843, p = 0.723 and p = 0.298, respectively) in either grazed or ungrazed patches.

Short-term ingestive behavior
Data on short-term ingestive behavior are presented in Table 2. The total number of bites was not influenced by CH (p = 0.519), with an average of 44 bites patch -1 . In both CH treatments, the total number of bites was progressively reduced in line with increases in defoliation severity (p < 0.001). Interaction of CH and DS in time grazing in each patch (p = 0.001) was verified. In DS0, total grazing time was longer in CH35    than in CH25. In the other DS treatments, there was no difference between CH35 and CH25. In both CH treatments, total grazing time was longer in DS0 and progressively reduced in line with increases in DS. The number of feeding stations was longer in CH25 than CH35 (p = 0.006). The total feeding station time was longer in DS0 and progressively reduced with increases in DS (p < 0.001). Short-term intake on mixed pasture Sci. Agric. v.79, n.2, e20200090, 2022 Bite rate and feeding station per min were influenced by CH (p < 0.001 and p = 0.020, respectively) and DS (p = 0.026 and p < 0.001, respectively). The maximum bite rate was observed for CH25 (56.8 % more bites than CH35). The same bite rate behavior in the CH was observed in feeding stations per min. As regards DS, there was no difference in CH25 in bite rate. In CH35, the bite rate was the same for DS0 and DS20, but both differed from DS40. In the CH25, all DS were different, and DS40 presented the highest feeding station per min. The same was observed for CH35, where DS40 obtained the highest feeding station per min.
There was interaction between CH and DS (p = 0.081) for bites/feeding station. Between CH, there was only a difference in DS0, in which CH35 was 18.8 % greater than CH25. As regards DS, the same behavior was observed in bites/feeding stations in both CHs, where all were different and the highest value found in DS0. For the bite mass and instantaneous intake rate, there were effects of CH (p = 0.068 and p = 0.074, respectively) and DS (p = 0.032 and p = 0.016, respectively), with no interaction. For CH, greater bite mass and instantaneous intake rate were obtained in the CH25 (increases of 50 % and 128 %, respectively). As regards DS, lower bite mass and instantaneous intake rate were found for DS20 compared to DS0 (reduction of 20 % in CH25 and 40 % in CH35 for bite mass, respectively; and decreases of 18 % in CH25 and 52 % in CH35 for instantaneous intake rates, respectively).

Discussion
Grazing management controlling the regrowth of tropical grass affects the canopy structure, influencing plant, and consequently, animal responses. In the regrowth period, when light is limited in the lower canopy stratum (i.e., light interception above 95 %), tropical grasses increase stem elongation rather than leaf elongation (Carnevalli et al., 2006;Silveira et al., 2016). In mixed palisade grass-forage peanut canopies, pregrazing canopy height ranged from 24 to 30 cm results in low competition for light between plants (Gomes et al., 2018) which is accentuated/maximized in pregrazing canopy heights in excess of 35 cm (Gomes et al., 2018). Thus, in the present study, the total herbage mass of CH35 and no defoliation (DS0) was higher (15.7 %) than in CH25. This increase in herbage mass was due to increased stem grass mass, which was 161 % greater in CH35 than CH25 (Table 1). Moreover, the increase in grass stem mass contributed to promote stem elongation up to the top of the canopy (0-20 % sward stratum, Figure 3D and 4D). In CH35 without defoliation, 20.1 % of the stem mass was located in the canopy top strata. In CH25, only 1 % of the stem mass was located in the canopy top layer. Furthermore, CH35 had an increase of 50 % in the stem herbage bulk density above the 40 % sward strata.
Canopy structure modifies livestock short-term ingestion capacity (Carvalho et al., 2015;Guzatti et al., 2017). The highest instantaneous intake rate in CH25/ DS0 was due to the increased bite mass associated with a faster bite rate, which can be explained by the ease of selection, apprehension, and forage intake since the upper canopy stratum was characterized by a predominance of leaves. Thus, the bite removed leaf up to the 20-40 % sward stratum. This is due to a deep bite of the high leaf bulk density in that stratum (Table 1).
In CH35/DS0, the instantaneous intake rate was lower than in CH25 because of the lower bite mass and rate. The bite mass was lower due to decreases in the leaf bulk density in the grazed strata, and possibly, to the animal ingesting a smaller volume with each bite. The bite volume is defined by the bite depth and bite area (Stobbs, 1973). We can infer that in CH35/DS0 the bite depth (the difference between canopy height of the ungrazed and grazed patches) was greater than in CH25/ DS0. However, the percentage removed from the height in the DS0 in both CH was 20 % relative to the height of the ungrazed patches. Thus, the length removed in CH35 exceeded that in CH25 (7.6 vs. 4.1 cm, respectively; Table 1), and the smaller bite volume probably resulted from a reduction in the bite area which is related to the pasture structure (Benvenutti et al., 2006;Drescher et al., 2006). In pastures with predominant upper strata of leaves, the bite area is maximized due to the benefits of tongue movement to grasp forage (Benvenutti et al., 2006;Benvenutti et al., 2008). In canopies with greater presence of stem at the top strata, the bite area is limited, since the shearing force of the stem is elevated (Barrett et al., 2001;Baumont et al., 2004;Gregorini et al., 2011). Thus, the stems act as a barrier interfering with the tongue movements leading the cattle to exclude the stems from the bite by reducing the reach of the tongue sweeps (Benvenutti et al., 2006). Grazing animals in patches CH35/DS0 had significantly more bites in each feeding station and remained longer at each feeding station than with the combination of CH25/DS0 (Table  2). This is indicative of a likely reduction in bite area in this treatment combination. A greater presence of stem in the top canopy layer not only reduces bite area but also causes the animal to spend more time in the selection and apprehension of the leaf (Benvenutti et al., 2006). This may explain the lower bite rate in CH35/DS0 than in CH25/DS0. Thus, our data support the hypothesis that animals grazing mixed canopies with greater pre-grazing heights have their instantaneous intake rate reduced by a greater presence of stems.
There was a presence of legumes at the top of the canopy in both CHs, even at the 37.5 cm canopy height. In taller canopies, forage peanut grows up in a vertical direction as a light competition strategy (Pereira et al., 2017;Tamele et al., 2018). In canopies where light interception does not exceed 95 %, only leaves of the forage peanut are located at the top of the canopy. On the other hand, when the canopy height exceeds 30 cm, there Short-term intake on mixed pasture Sci. Agric. v.79, n.2, e20200090, 2022 is a more significant proportion of forage peanut stolon in the top canopy strata (Gomes et al., 2018). This pattern of response was observed in the present study ( Figure  4D and E). The vertical stolon elongation of the forage peanut observed in CH35 leaves the apical meristems more exposed to grazing. This is a possible explanation of a three-fold greater disappearance of legume stem in CH35/DS0 than in CH25/DS0 (9 vs. 27 g m -2 ; Table 1). Greater decapitation of forage peanut stolon may impact plant response on subsequent regrowth, which, in turn, compromises legume stability in the canopy (Black et al., 2009;Gomes et al., 2018).
The reduction in canopy height after grazing using DS0 in both CH was approximately 20 %, similar to that performed manually in DS20 (Table 1). Thus, the short-term ingestive behavior responses in DS20 were equivalent to second-layer grazing during the lowering process. In CH35 with no defoliation, approximately 50 % of the leaves were located in the 0-20 % stratum of the canopy height. In CH25 without defoliation, only approximately 30 % of the leaves were located in the same stratum. Thus, proportionally, leaf depletion in DS20 was greater in CH35 than in CH25. This response, associated with greater stem proportion at the top of the CH35 canopy, increased the difference in the canopy structure compared to CH25 in the DS20. This difference was reflected in the instantaneous intake rate, which was more than three times greater in the CH25/DS20 than CH35/DS20. The reduction in the instantaneous intake rate in DS20 compared to DS0 was due to the lower bite mass since the bite rates were similar ( Table  2). The bite mass was reduced mainly due to a lower bite depth; the canopy height was reduced by 9.2 % (1.8 cm) and 10 % (3.0 cm) in the CH25/DS20 and CH35/DS20, respectively, compared to the ungrazed patches (Table 1). The reduction in bite mass in CH35/DS20 compared to CH35/DS0 was more pronounced than that observed in CH25/DS20. The bite mass and bite rate were lower than in CH35/DS20 than the observed CH25/DS20 (Table 2). This fact is linked to a greater stem proportion in the 20-40 % stratum in CH35/DS20 than in CH25/DS20 (12.1 vs. 33.8 %, respectively) and lower leaf herbage bulk density  (Table 1). Therefore, the pre-grazing height of 35 cm should not be used. However, in the case of a pre-grazing canopy height above 25 cm, the increase in defoliation severity causes negative impacts on the instantaneous intake rate. Therefore, the time spent in each feeding station and the number of bites in each feeding station decreased as the defoliation became more severe, indicating that the animal was going through the patches faster as a grazing strategy. In DS40 patches, this strategy became so evident that the low number of bites in each patch resulted in difficulty in estimating bite mass and instantaneous intake rate. This behavior explains the linear decrease in intake rates observed by Fonseca et al. (2012) when the animals grazed more than 40 % of pre-grazing canopy height. This linear decrease is linked to the large-scale increase in the stem proportion in the available grazed stratum ( Figure  3C-F and 4C-F). Consequently, the number of mandibular movements per unit of ingested dry matter increased linearly beyond the 40 % reduction of pre-grazing height (Fonseca et al., 2012). Additionally, in the DS40 patches in both CH treatments, there was a greater number of feeding stations per min and a lower number of bites per feeding station both of which characterize a search by the animal for a patch with a better canopy structure. In rotational stocking with defoliation severity near 40 %, animals stopped grazing when the sward structure became a limiting factor (e.g., the proportion of leaves at the end of the grazing period; Amaral et al., 2013). This effect has been described by Ribeiro Filho et al. (2003) as a change in the disposition of the animal waiting to enter a new plot, which leads to reduced total herbage intake. Thus, our data support the hypothesis that more severe grazing intensity reduces the instantaneous intake rate due to a negative impact on bite mass and rate.

Conclusion
The canopy structure in terms of the leaf and stem presence in the upper stratum is the main factor which influences the ingestive behavior characteristics in the short-term. Thus, greater stem mass in the grazing strata negatively influenced the instantaneous intake rate. Under rotational stocking, a pre-grazing canopy height of 25 cm should be used in mixed pastures of palisade grass and forage peanut. Instantaneous intake rate has a proportionately inverse relationship with defoliation severity.