Morphogenic and structural characteristics of guinea grass pastures submitted to three frequencies and two defoliation severities

It was evaluated the morphogenic and structural characteristics of guinea grass under rotational at three grazing intervals and two defoliation intensities. Grazing intervals corresponded to the time needed by the forage canopy to reach 90, 95 or 100% of incident light interception during regrowth and they were evaluated combined to two defoliation severities (post-grazing conditions, 25 and 50 cm of height), being allocated to experimental units according to a complete randomized design, with three replicates and 3 × 2 factorial arrangement. The experiment was conducted from July 2003 to May 2004. For evaluation of morphogenetic and structural characteristics, ten tillers per experimental unit were selected. Morphogenetic and structural characteristics were strongly influenced by seasons of the year inasmuch as leaf elongation rate increased 3.5 fold from winter to summer. In addition to year season effect, there was also an effect of defoliation frequencies on tiller population density, which was greater in the defoliation period corresponding to 90% of light interception, especially if evaluated in relation to the interval corresponding to 100% of light interception. Defoliation frequency is determinant in expression of phenotypic plasticit, acting on the control of stem elongation.

Another important morphogenic component is stem elongation (Sbrissia & Da Silva, 2001).This characteristic increases forage yield (Santos et al., 2004) and interferes significantly with the canopy structure.Furthermore, it impedes and endangers grazing efficiency because of the alterations in the leaf blade to stem ratio.This ratio is directly related to the performance of the grazing animals (Euclides et al., 2000;Difante et al., 2008).
A database consisting of the morphogenic characteristics that determine the pasture structure is necessary to identify the conditions that maximize the efficiency of yield and harvest.This database, combined with the evidence of the effects of the canopy structure on forage intake and animal performance, lead to the development of management strategies based on the pasture conditions.The management goals have been defined by the canopy height (Hodgson, 1990) or forage mass (Matthews et al., 1999).
The objective of this study was to evaluate the morphogenic and structural characteristics of Tanzania grass (Panicum maximum Jacq.cv.Tanzânia) under rotational stocking.

Material and Methods
The experiment was conducted in Embrapa Gado de Corte -CNPGC, in Campo Grande/MS.The geographic coordinates of this site are 20º27'S and 54º37'W, and the altitude is 530 m.According to the Köppen classification, the climate is tropical wet savannah type, subtype Aw.The area is characterized by the irregular annual distribution of rain, with a well-defined dry period during the colder months and a rainy period during the summer months (Figures 1 and 2).
The climate data were recorded at the Embrapa Gado de Corte (Beef Cattle Embrapa) Weather Station, approximately 800 m from the experiment site.
Sowing was performed by using 2 kg/ha of viable pure seeds in January 1995.The seeds were incorporated into the soil by light harrowing, followed by light compaction by rolling tires.The area has been kept under grazing since August 1995.The experiment lasted a total of 434 days (May 8 th 2003 to May 15 th 2004), with data collection from July 11 th 2003 to May 15 th 2004, totaling 309 days.The period from May 8 th 2003 to July 10 th 2003 was used to control invading plants, to collect soil samples for analysis and to standardize grazing of the experimental area.
All of the areas with 90% and 95% light interception (LI) received four applications of 50 kg/ha N. The area with 100% LI and 50 cm of residue received one application of 50 kg/ha N, followed by two applications of 70 kg/ha N and a final application of 10 kg/ha N. The area with 100% LI and 25 cm of residue received one application of 50 kg/ha N followed by one application of 70 kg/ha N and a final application of 80 kg/ha N.  The experimental plot was subdivided into three blocks of 1.5 hectares with six paddocks (experimental units) of 0.25 hectare.Nelore steers at18 months of age and with an average weight of 310 kg were used for the grazings.When the animals were not needed in the study, they grazed in an area consisting of 3.0 ha of Mombaça grass, 3.0 ha of Massai grass and 1.5 ha of Tanzania grass.
It was used a completely randomized 3×2 factorial block design with three intervals and two grazing severities, with three replicates, yielding a total of 18 experimental units.The grazing intervals corresponded to the time period necessary for the forage canopy to reach 90%, 95% or 100% of interception of incident light during regrowth.The grazing intensities corresponded to the residue heights of 25 and 50 cm.
The monitoring of light interception by the canopy was performed under the post-grazing condition (25 and 50 cm) and at variable time intervals during regrowth for each season.In the winter, the evaluation interval was 15 days for the 90% and 95% paddocks of LI and 20 days for the 100% paddocks of LI.During the rest of the year, the interval was reduced to 7 days in the treatments of 90% and 95% LI and 10 days for the treatments of 100% LI.When the levels of light interception were close to the pre-grazing levels, the monitoring frequency was increased with evaluations every three days.In the evaluations, the canopy was analyzed with an AccuPAR Linear PAR/LAI ceptometer, Model PAR 80(Decagon Devices, EUA).Readings were collected at six sampling points per paddock (Carnevalli et al., 2006).At each point, three readings above the forage canopy and three readings close to the ground were performed.Each reading was the mean of five instantaneous readings.
In the early part of each regrowth period, ten tillers were randomly tagged for evaluation of the morphogenic and structural characteristics of the canopy during the rest period of the pastures.The tillers were identified with plastic rings, and a pole with numbered tags was fixed beside each tiller for better visualization in the field.The evaluations were performed every 15 days in the winter and every 7 days during the other seasons.The length of the leaf blades and the height of the ligule of the last expanded leaf were measured during these evaluations.Furthermore, the number of new leaves appearing on each of the tillers at each evaluation was recorded.These data were used to calculate the following variables.
1) Leaf appearance rate (LAR), which is the number of leaves appearing per tiller, divided by the number of days of the evaluation period (leaves/tiller.day) 2) Phyllochron (PHY), which is the opposite of leaf appearance rate (days/leaf.tiller) 3) Leaf elongation rate (LER), which is the summation of all elongation of the leaf per tiller divided by the number of days of the evaluation period (cm/tiller.day)4) Final length of the leaf blade (FLLB), which is the average length of all the leaves of the tiller, measured from the leaf apex to the ligule (cm) 5) Stem elongation rate (SER), which is the summation of all the stem/pseudo-stem elongations per tiller divided by the number of days of the evaluation period (cm/tiller.day)and 6) Accumulation of stem per tiller, which is calculated from the difference between the height of the ligule of the last expanded leaf on the last evaluation day and the height of the ligule of the first leaf tagged at the beginning of the evaluation (cm).The variables were grouped by season (winter, spring, summer and autumn) due to the varying nature of the intervals and the number of grazing cycles in the evaluated treatments.The data were submitted to the analysis of variance, utilizing the GLM procedure of the SAS statistical package (SAS Institute, 1993).It was used a split-plot arrangement in which the treatments (Res/LI) constituted the plots and the season constituted the subplots.The RANDOM and TEST commands were used for identification and performing the suitable tests.The levels of LI were regarded as the discrete variable because of the small amount of levels evaluated.The comparison of means was performed by appropriate contrasts for the main effects, the effects of residue (RES), the effects of LI and for the nested effects of RES (LI) and LI (RES), using Tukey's test at 10% probability.

Results and Discussion
The morphogenic and structural characteristics varied significantly among frequencies, intensities of defoliation and seasons of the year.Structural characteristics are ruled by the morphogenic characteristics (Chapman & Lemaire, 1993).
The LER was the highest during summer (4.16 cm/tiller/day) and the lowest in the winter (1.16 cm/tiller.day).These results can be explained by the greater availability of growth factors such as water, temperature and light during the summer.According to Ludlow & Ng (1977), leaf expansion is one of the physiological processes most sensitive to water deficit.This is because division and cell growth are extremely sensitive to cell turgor, causing leaf and root elongation to be interrupted long before the processes of photosynthesis and cell division are affected (Ludlow & Ng, 1977).
Leaf appearance rate (LAR) was influenced by several factors including residue (P=0.0162),light interception (P=0.0001),season (P=0.0001),residue × light interception interactions (P=0.0710),residue × season interactions (P=0.0019),light interception × season interactions (P=0.0001) and residue × light interception × season interactions (P=0.0006).Similar to LER, LAR was the highest during the summer (0.09 leaf/tiller.day)and lowest during the winter (0.04 leaf/tiller.day).This can be explained by differences in the availability of the growth factors during the different seasons (Table 2).
The effects of residue and light interception differed significantly between the seasons.Grazing intervals with high defoliation intensities caused an increase in LAR.According to Nabinger (1997) and Bos & Neutboon (1998) (Frank & Hofman, 1994).More frequent defoliations enabled greater removal of leaf tissues and, as a result, increased leaf production.The increased leaf production was probably caused by the greater light penetration at the canopy base.The behavior of phyllochron was similar to the behavior of LAR.The phyllochron was affected by light interception (P=0.0001),season (P=0.0001),residue × interception interactions (P=0.0022),season × residue interactions (P=0.0013),season × light interception interactions (P=0.0001) and residue × light interception × season interactions (P=0.0001).The highest value of phyllochron was recorded during the winter (27 days), and the lowest was recorded during the summer (12 days).Intermediate values were recorded during the spring and autumn.
The final length of the leaf blade (FLLB) was affected by the residue (P=0.0001),light interception (P=0.0002),season (P=0.0001),residue × light interception interactions (P=0.0121),light interception × season interactions (P=0.0033) and residue × light interception × season interactions (P=0.0001).The highest values of leaf blade length were recorded during the spring (48.2 cm) and summer (48.9 cm).These results can be explained by the more favorable climate conditions for the growth of tillers during these seasons.The lowest value was recorded during the winter (32.5 cm).The paddocks highest values of final length of leaf blade were observed in the paddocks with a residue of 50 cm (36.6 cm to 52.6 cm; Table 3).
The lower values of FLLB for the paddocks with a 25 cm residue were not affected by the season and ranged from 28.4 cm to 46.7 cm.Furthermore, the paddocks with the paddocks with the highest defoliation intensity yielded shorter length of the leaf sheath.This could have been caused by a reduction of the cell multiplication phase, leading to a shorter final length of the blade, as demonstrated by Duru & Ducrocq (2000).According to Lemaire & Chapman (1996), the length of the leaf blade is a plastic characteristic responsive to defoliation severity and it is regarded as a morphologic strategy of the plant to escape grazing.
The 90% frequency of LI resulted in shorter lengths of leaf blades during each season with the exception of winter, when no differences were detected.This was probably a result of the high LAR.Furthermore, the shortest length of the sheath for the 95% and 100% frequencies of LI resulted in lower values of final lengths of leaf blades.This might have been caused by the poor accumulation of stems on the tiller (Barbosa et al., 2007).
The results for the final length of leaf blades were similar to the results for the LER.There was a positive correlation between the increases in the LER and FLLB.During the winter and spring, the highest values of the LER for the 50-cm residues were1.28cm/tiller.dayand 3.68 cm/tiller.day,respectively.This caused the final length of leaf blades to increase.The highest LER in the same paddocks were observed during the times of low temperatures and precipitation.These results can be explained by the larger mass of residue forage, which may have provided greater nitrogen mobilization from older leaves to the younger leaves (Nabinger & Pontes, 2001).
The morphogenic characteristic determining the population density of tillers is the leaf appearance rate (LAR) (Davies, 1974;Mazzanti et al., 1994).However, even with the high LAR during the summer, the paddock population density of tillers remained smaller in the paddocks with 100% LI compared to the paddocks with 90% and 95% LI.
The reason for the differences in the population density of tillers was the light interception.The lowest population densities were found in the paddocks with 100% LI, and they were between 324 tillers/m 2 and 524 tillers/m 2 .This can be explained by the processes of intra-species competition by growth factors such as light.This gives rise to a process known as "compensation size/population density of tillers" (Matthew et al., 1995).Thus, when rotational stocking is used, the regrowth of the grazing land starts with an increased number of larger tillers.This process begins soon after grazing.During the 95% interception of incident light, both the quantity and quality of light is limited.As a result, tillering is reduced, and the death of tillers begins to occur.Increases in the size of the tillers compensate for reductions in the number of tillers.Furthermore, during situations of intense competition for light such as 100% LI, the composition of the light spectrum in the interior canopy changes, reducing to blue light and of the ratio red light: infrared light (Gautier et al., 1999).This alteration in the light spectrum reduces the tillering and the population density of tillers.According to Davies et al. (1983), plants submitted to shading prioritize the allocation of assimilates to the growth of existing tillers to the detriment of the formation of new tillers.Thus, the young tillers would die first as a result of the overlapping and shading by the older tillers (Ong et al., 1978).Furthermore, some buds are probably aborted even before the emergence of new tillers in shaded grazing lands (Sbrissia & Da Silva, 2001).Barbosa et al. (2002) demonstrated that in Tanzania grass under rotational stocking, the number of basilar and aerial tillers decreased linearly during the regrowth period.However, the LAI increased.
In general, the treatments of 90% LI resulted in a high population density of tillers, varying from 470 tillers/m 2 to 744 tillers/m 2 .This was promoted by high LAR and by the decreased effect of shading which is inherent to the plant community at the end of the regrowth period (Korte, 1986).However, those tillers were smaller in size.This was most evident for the 25 cm residue, which was also where the lowest values of FLLB were observed.According to Sbrissia & Da Silva (2001), the population density of tillers is the component of the LAI which allows increased flexibility of the plant and different regimes of defoliation.Furthermore, LAI can be optimized in grazing lands by means of a high population density of small tillers.However, this does not necessarily mean that the population is stable as grazing lands can be kept stable even with low values of population density of tillers (Sbrissia et al., 2010).
Higher values of population density of tillers were recorded during the summer (621tillers/m 2 ), followed by the autumn, spring and winter (407 tillers/m 2 ).This behavior reflects the behavior observed for LAR.Furthermore, under normal conditions, seasonal fluctuations in the amount of light energy received by the plants caused seasonal variations in the leaf area and, consequently, seasonal alterations in the population density of tillers (Matthew et al., 1999).High population densities during the summer were also observed by Carnevalli et al. (2006) for guinea grass cv.Mombaça in a similar experiment.The authors attributed these results to the intense tiller renewal during that season.The effect of residue on the population density of tillers was only verified during the winter.During this season, the greatest population density was recorded for the 50 cm residue treatment (435 tillers/m 2 ).
Values of SER observed during the autumn flowering time were 0.335 cm/tiller.day.During the winter, there were no differences in SER among the defoliation intensities.During the summer and spring, values of SER recorded for the area subjected to 100% LI were 0.125 and 0.576 cm/tiller.day,respectively.These values were higher than the SER values of the areas receiving 90% and 95% LI.During the autumn, SER was larger for the areas with residue of 25 cm (0.419 cm/tiller.day),and the largest SER values were observed in the pre-grazing condition of 100% LI.The stem accumulation in the tiller was affected by light interception (P=0.0001),season (P=0.0001) and light interception × season interactions (P=0.0001).The accumulation of stems in the tiller did not differ between the residues of 25 cm and 50 cm.However, there were significant differences between the levels of light interception and the seasons (Table 6).
The greatest accumulation of stems per tiller was recorded in autumn.Stem accumulation on the tiller increased as the levels of light interception were raised from 90% to 100% in every season except for winter.During the winter, there was no correlation between stem accumulation and the level of light interception.The areas with 100% light interception had the highest stem elongation, which was mainly observed during the summer and autumn.The stem is an important morphological component of grass.It has supported functions in the spatial arrangement of the plant and assimilates translocation to the leaves (Fagundes et al., 2006).The accumulation of stems increased during the spring as the water deficit decreased.During the autumn flowering, stem accumulation on the tiller became more noticeable as a consequence of the high SER.The highest stem accumulation was recorded during the combination of 25 cm of post-grazing residue and 100% of light interception at pre-grazing.During the autumn, the accumulation of stems in the areas with 100% LI was 4.25 times larger than in the areas with 90% LI and 2.0 times larger than in the areas with 95% LI.These results indicate that long periods of rest during the autumn can endanger the canopy structure.

Conclusions
Defoliation frequency is a determining factor in the expression of phenotypic plasticity.This is mainly caused by variations in the population density of tillers, which affects the control of stem elongation.

Figure 1 -
Figure1-Distribution of rainfall during the experimental period and in the last 30 years.

Figure 2 -
Figure 2 -Average temperatures ( o C) during the experimental period and in the last 30 years.
, environmental factors such as light and temperature can Means with the same lower case letter in the column and capital letter in the row within season of the year do not significantly differ from each other (P>0.10)by Tukey's test.Means followed by capital letter followed by (') compare seasons of the year (P>0.10)by Tukey's test.

Table 2 -
Leaf appearance rate (leaf/tiller.day) of guinea grass cv.Tanzania swards at various defoliation intensities and frequencies during the seasons Means with the same lower case letter in the column and capital letter in the row within season of the year do not significantly differ from each other (P>0.10)by Tukey's test.Means followed by capital letter followed by (') compare seasons of the year (P>0.10)by Tukey's test.

Table 1 -
Leaf elongation rate (cm/tiller.day) of guinea grass cv.Tanzania swards at various defoliation intensities and frequencies during the seasons influence LAR.These factors are related to the upper part of the canopy being shaded by leaves and other structures, inhibiting the activity of buds and endangering the formation of new leaves and new tillers

Table 3 -
Final length of leaf blade (cm) of guinea grass cv.Means with the same lower case letter in the column and capital letter in the row within the season of the year do not significantly differ from each other (P>0.10)by Tukey's test.Means followed of capital letter followed by (') compare seasons of the year (P>0.10)by Tukey's test.
Means with the same lower case letter in the column and capital letter in the row within season of the year do not significantly differ from each other (P>0.10)by Tukey's test.Means followed by capital letter followed by (') compare seasons of the year (P>0.10)by Tukey's test.

Table 4 -
Population density of tillers (tillers/m 2 ) of guinea grass cv.Tanzania swards at various defoliation intensities and frequencies during the seasons