Agronomic Behaviour and Chemical Composition of Three Varieties of Pennisetum sp in the Peruvian Tropics

purpureum × Pennisetum typhoides or King grass). Te experiment was carried out at the Estaci´on Experimental Agraria “Campo Verde,” Ucayali, Peru, under a randomized complete block design (RCBD) with a 3 ( Pennisetum sp. variety) × 3 (cutting age) factorial arrangement and 4 replicates each. Tirty-six subplots were installed, and treatments were randomly assigned, using vegetative seed with three buds each. Measurements of agronomic behaviour and chemical analysis were carried out in the 8th, 12th, and 16th week. Te data were subjected to ANOVA and multiple comparisons with the Duncan and Kruskal–Wallis tests ( p < 0.05) and correlations using Pearson coefcients ( p < 0.05). Te agronomic behaviour did not vary between Pennisetum sp. varieties, but according to cutting age, Cameroon grass produced greater plant height, biomass, dry matter, and Mg content at the 8th week. Te highest concentrations of N (2.38 ± 0.11%), P (0.18 ± 0.01%), and Ca (0.70 ± 0.03%) were recorded at the 8th week, Zn (27.33 ± 1.91mg/kg) at the 12th week, and Mg (0.26 ± 0.02%) at the 16th week. A negative correlation was found between biomass and the stem-leaf ratio at the 12th week because the frst leaves of plants died, but the stems increased in size. It is concluded that, at an early cutting age in varieties of Pennisetum sp., lower biomass and dry matter result, but higher content of N, P, Ca, and Fe; and at a late cutting age, it shows greater biomass, dry matter, Cu, and Mg content.


Introduction
In Peru, 70% of the deforested forests in the Amazon have been destined for the establishment of cultivated pastures for extensive livestock activity [1].Pastures are generally managed without adequate soil and pasture management techniques, resulting in overgrazing and erosion [2,3].Overgrazing causes negative efects on pastures, such as accelerated compaction, which reduces pore space and drainage capacity, hindering plant root development [4,5].
Grazed grasses are the main source of food for livestock in the tropics, as they provide nutrients at a low cost compared to concentrated feed [6,7].However, the dry season dramatically impacts the productivity and nutritional quality of these pastures, dependent on adequate humidity and light incidence [8][9][10], causing losses in livestock productivity [11].Cutting grasses can contribute to meeting feed needs during the dry season [12,13].Cutting grasses, such as those of the genus Pennisetum, were introduced from North America and tropical Africa and adapted to the tropical regions of Peru [14], exhibiting high biomass yields and nutritional content; however, agronomic behaviour can vary, depending on the type of soil, fertilization, as well as the frequency and intensity of cutting [15].
Pennisetum sp.o Maralfalfa is characterized by its adaptability up to 3000 meters above sea level and reaching a protein content of up to 16% in fertile soils and Ca, P, Mg, and K contents of 0.37%, 0.50%, 0.37%, and 6.28%, respectively [16,17].Pennisetum purpureum × Pennisetum typhoides or King grass reaches a protein content of 12-15%, lignin between 5.3 and 6.5% [18], and other varieties range between 3.85 and 6.70% [19].Tere are few reports of productivity of these species adapted to the Ucayali region, Peru; however, environmental conditions and crop management are relevant factors in the expression of agronomic behaviour and establishment of Pennisetum sp.Appropriate age and cutting height could infuence the growth rate, yield per hectare, nutritional quality, and persistence of the crop over time [20,21].Terefore, the objective of the study was to compare the agronomic behaviour and chemical composition of three varieties of Pennisetum sp subjected to diferent cutting ages.

Study Area and Location.
Te experiment was carried out at the Estación Experimental Agraria "Campo Verde" in Coronel Portillo province, Ucayali department, Peru.Te area is located between the coordinates of 08 °32′23.81″Ssouth latitude and 74 °52′48.86″Wwest longitude, at an altitude of 140-300 meters above sea level (Figure 1).Te average annual relative humidity is 84.2%, the average temperature is 31.60 °C, and the average precipitation is 2191.96mm.

Experimental Design.
Te experiment was carried out in a randomized complete block design (RCBD) with a 3 × 3 factorial arrangement (Pennisetum sp variety × cutting age).Four blocks per treatment were used, considering the blocking criteria to irregular soil topography.In an area of 3,072 m 2 (128 × 24 m), four plots or blocks of 945 m 2 each (32 × 24 m) were divided.In each block (B1, B2, B3, and B4), 3 main plots were established, with an efective area of 210 m 2 each (32 × 6 m), making a total of 12 plots.Within each plot, 3 secondary plots or subplots were randomly arranged, with an efective sampling area of 50 m 2 (10 × 5 m) each, and this excludes the edge efect (Figure 2).Te experimental unit consisted of 175 plants planted at a distance of 1 m between rows and 0.5 m between plants.Tree varieties of Pennisetum sp. were installed: Maralfalfa (Pennisetum sp.), King grass (Pennisetum purpureum × Pennisetum typhoides), and Cameroon (Pennisetum purpureum Schum cv.Cameroon).

Pasture Installation and Fertilization.
Te soil characteristics at the time of the experiment were as follows: sandy clay loam texture, pH 5.12, N 0.08%, P 2.66 mg/kg, and K 0.22 Cmol/kg.Te preparation of the land was performed manually, and phosphate rock and dolomite were incorporated by broadcast in a ratio of 100 kg of each per plot.Sowing was carried out using vegetative seed, with cuttings that had three buds.

Agronomic Behaviour.
Te variables such as plant height (cm), coverage (%), number of tillers, and stem/leaf ratio were recorded at three cutting ages at the 8th, 12th, and 16th week of growth.Plant height was measured with a metric ruler and was considered from the base of the soil to the highest part of the plant without stretching or counting inforescence.Coverage was recorded as a percentage per square meter and was estimated based on the apparent proportion in which the pasture covered each area of the quadrat.Te number of tillers was counted in an area of 1 m 2 , and the number of seedlings within the square meter was considered.Te stem/leaf ratio consisted of the quotient of the weight of all the leaves between the weight of all the stems.For the biomass yield, a vegetative material of 1 m 2 was cut and weighed, taking into account the established cutting frequencies.Te cut was made at a height of 15 cm from the soil, later extrapolating the value to one hectare.Te dry matter was obtained from the fresh weight and dry weight of the sample in grams, and 250 g of biomass was used to submit to 60 °C in an oven (Memmert, Germany) until a constant weight was obtained.

Chemical Composition.
A total of 16 samples were analysed, the samples were weighed and ground in a hammer mill for further analysis, and to determine the nitrogen (N) content, micro-Kjeldahl digestion was used [22]; phosphorus (P) content was determined by colorimetry [23], and potassium (K), calcium (Ca), magnesium (Mg), iron (Fe), copper (Cu), zinc (Zn), and manganese (Mn) contents were determined by the Mehlich method using an atomic absorption spectrophotometer [24].
2.6.Data Analysis.Te normality of the data of quantitative variables was verifed using the Shapiro-Wilk test (p > 0.05), and homogeneity of variances was verifed with the Levene test (p > 0.05).Ten, analysis of variance was performed, and if the diference between groups was signifcant, multiple comparisons were performed with Duncan's parametric test (p < 0.05) for the agronomic variables and for N, P, K, Mg, Fe, and Mn.Te concentration of Ca, Cu, and Zn was also

Results and Discussion
3.1.Agronomic Behaviour.Te average values of the agronomic behaviour of the three varieties of Pennisetum sp and according to their cutting age are detailed in Table 1.No signifcant diferences (p > 0.05) were found between varieties for any of the agronomic variables.Plant height, the number of tillers, biomass, and dry matter content were afected by cutting age (p < 0.05).Te tallest plants were observed at the 16th week, but the highest number of tillers was observed at the 8th and 12th week, with no signifcant diferences between them (p > 0.05).Biomass production was higher in the 16th week, and dry matter was higher in the 12th and 16th week (p < 0.01).
Te results found in plant height are lower than that of Pennisetum sp reported by Hermitaño et al. [13], with 257 cm length at 75 days.Calzada-Marín et al. [25] also reported higher heights of Pennisetum sp in warm subhumid climates (230 cm at 150 days).Variations in plant height can be explained by the size of the leaf area, since with a greater leaf area, there will be greater competition for light between plants.Anten and Hirose [26] maintained that the height of the plant decreases with a greater leaf area, since the penetration of solar rays to other plants is difcult.Plant height is strongly linked to the length of the photoperiod, temperature, humidity, and soil fertility.If these environmental variables are suitable for the physiology of grasses, they will accelerate their growth [27,28].Te height and other morphophysiological changes of plants are observed when 95% of light interception is exceeded, this being the optimal time for grass harvesting [29,30].
A greater number of tillers were found in 8th and 12th weeks of cutting age.According to Tinini and Limache [31], the number of tillers in pasture for cutting, especially in Pennisetum sp, varies from 14 to 55 tillers.Pennisetum sp is characterized by developing high capacity to generate numerous tillers [13,32].Te growth dynamics of the morphological components of Pennisetum, such as tillering, leaves, stems, and dead material, constantly increase up to 151 days [25].For this reason, Portugal et al. [33] argued that tropical pastures need more efcient handling, because they present high rates of pasture mass accumulation due to the high number of tillers, but quality could deteriorate more quickly.
Biomass production and dry matter content varied according to cutting age.Tese fndings coincide with the reports by Cifuentes et al. [34], who found biomass productions of 30.66, 47.88, and 51.11 t DM/ha in Maralfalfa.In the case of Maralfalfa, Hermitaño et al. [13] reported that biomass did not difer between 45, 60, and 75 days of cutting age.Te dry matter content in this research varied according to cutting age, from 28 to 51 g DM/100 g MV.In this regard, Gurrola et al. [35] reported approximately 22 t DM/ha in Maralfalfa (Pennisetum sp) at 90 days of age, and Araya and Boschini [36] found 15.2 t DM/ha in King grass.Te production of biomass and dry matter content were due to the contribution of leaves and stems and to the phenological state of the plant [37,38].
In the stem/leaf ratio, no signifcant diferences were found between varieties of Pennisetum sp and cutting age.However, Calzada-Marín et al. [25] mentioned that the ratio tends to decrease as the age of the pasture increases.Tis is supported by the increase in biomass of the stems, and the leaves begin to be senescent [39].No signifcant efect of the interaction between variety and cutting age was found for any variable evaluated for agronomic behaviour.

Chemical Composition.
Te chemical composition of the three varieties of Pennisetum sp is detailed in Table 2. Te content of N, P, K, Ca, Fe, Cu, Zn, and Mn did not difer signifcantly (p > 0.05) between varieties, but Mg concentration was higher (p < 0.05) in the Cameroon variety (0.27 ± 0.02%) than in Maralfalfa (0.22 ± 0.01%) and King grass (0.19 ± 0.01%).Cutting age had a signifcant efect (p < 0.05) on the content of N, P, Ca, Mg, Fe, and Zn, where at the 8th week, the highest content of N, P, Ca, and Fe was observed, and Zn and Mg were higher at the 12th and 16th week, respectively.
Te concentration of macroelements in pastures is important because minerals establish a vital factor and imbalances or defciencies thereof can cause serious pathologies, which will generate productive and economic losses for producers [40].For example, the initial postpartum stage is where the metabolic-nutritional axes that control Ca, P, and Mg levels are most compromised, and the association of endocrine systems that control metabolism and regulate the immune system is evident [41].
Te N content varied according to cutting age, which coincides with Ramos-Trejo et al. [42] and Herrera et al. [43].Te N content reduces as plants become more adult and mature, and cellulose content will increase [19].Te reduction in protein content as the grass ages is due to the reduction in the metabolic activity of the plant, so that as the grass is harvested at an older age, the synthesis of protein compounds in the plant is lower.Terefore, at older age, the N values will be lower [44].For this reason, determining an optimal grass cutting moment is important to obtain grass with the highest protein content [45].P in animal feed is directly related to the productive and reproductive efciency of cows [46,47].In soils rich in P, grasses will have a high concentration of this mineral, but if P is not usable, the grass will be defcient in this mineral [48], generating the need for mineral supplementation.In this study, P concentrations in Pennisetum sp.varied between 0.10% and 0.18%.For dairy cattle, a P concentration of 0.31%-0.38% is recommended in the diets of moderate-to-high producing cows [49].

Correlations.
Correlations between the agronomic behaviour variables according to age are shown in Figure 3. Biomass and dry matter were positively correlated with plant height (p < 0.05); however, at the 8th week, a negative 4 International Journal of Agronomy correlation between biomass and dry matter was observed.Tis is explained by the fact that at the 8th week, the plant still contains a greater volume of water in its cells and that as the plant becomes an adult, the water levels drop and dry matter increases [36].Te negative correlation between biomass and the stem/leaf ratio at the 12th week is due to the frst leaves of plants dying, but the stems increase in size [39].Te correlations of agronomic behaviour according to Pennisetum sp varieties are shown in Figure 4. We found signifcant correlations between plant height and biomass and dry matter (p < 0.05).Plant height of King grass was positively correlated with biomass and dry matter (p < 0.05).In this regard, Contreras et al. [19] indicated that the higher the altitude, the greater the production of biomass and dry matter.
Te correlations of chemical composition according to age of Pennisetum sp varieties are shown in Figure 5. N is positively correlated with P and Ca (p < 0.05), and P is positively correlated with K, Ca, and Cu (p< 0.05).Sterling and Guerra [50] mentioned that Ca and P contents are generally similar between Pennisetum sp varieties.Tis would indicate that defciency of Ca in soil, therefore relatively high acidity, and the agronomic development of the crop could be afected by this variable.Maleko et al. [51] indicated that Ca and P concentrations in Napier pastures generally coincide with the values reported in this research, where P ranged between 0.13 and 0.19% and Ca ranged between 0.21 and 0.29%.According to NRC [52], minerals are very essential for ruminant reproduction (conception, gestation, and parturition), growth, maintenance, and production (e.g., milk and meat).However, the P and Ca contributions of Pennisetum sp.varieties are below the recommended levels in cattle of 0.36 and 0.43%, respectively [52].Te concentration of minerals in pastures is infuenced by edaphic factors, seasons, and proportion of dry matter of biomass [53].
Te correlations of the chemical composition according to Pennisetum sp varieties are shown in Figure 6.Correa [15] found Ca, P, Mg, and K contents of 0.37%, 0.50%, 0.37%, and 6.28%, respectively, in Maralfalfa at 56 days.However, in the present research, we found a higher level of Ca and lower levels of P, Mg, and K. Likewise, Andino and Pérez [54] reported values of 1.0% N, 0.6% P, 1.6% K, 0.2% Ca, and 0.15% Mg in Pennisetum sp.In the present research, we found higher levels of N, Ca, and Mg and lower levels of P and K. Te high Ca level in this research could be related to the application of dolomite and phosphate rock at the time of planting.
Maleko et al. [51] and NRC [52] indicated that minerals are very essential for ruminant reproduction (conception, gestation, and parturition), growth, maintenance, and production of milk, meat, and wool.Te reports of Maleko   8

International Journal of Agronomy
International Journal of Agronomy International Journal of Agronomy et al. [51] are even further below the recommended concentrations of 0.36% and 0.43% for P and Ca, respectively, by NRC [52], while Louhaichi et al. [53] indicated that the content of minerals and pasture ash is infuenced by edaphic factors, seasons, and proportion of dry matter of biomass.
Te N, P, K, Ca, Mg, and Fe contents are negatively related to biomass, dry matter, and age of the plant (Table 3).As the yields and age of the plant increase, the concentrations of these elements decrease.Otherwise happens with the Cu concentration, which is positively related to the stem/leaf ratio.Tis is due to the most of Cu translocated towards the fowering part of the plant [54].

Conclusions
Te agronomic behaviour did not vary signifcantly between the Pennisetum sp.varieties evaluated or in the interaction between variety and cutting age.Te Cameroon variety presented a higher Mg content and higher plant height, biomass, and dry matter content, although not signifcantly.Cutting at the 8th week allows obtaining a higher content of N, P, Ca, and Fe in the forage, Zn at the 12th week, and Mg at the 16th week.At an early cutting age, in varieties of Pennisetum sp., it shows lower biomass and dry matter result, and higher content of N, P, Ca, and Fe; but at a late cutting age, it shows greater biomass, dry matter, and Cu and Mg content.

Table 1 :
Agronomic behaviour of three varieties of Pennisetum sp according to cutting age and interaction variety × cutting age 1 .), at the p < 0.01 level ( * * ), or at the p < 0.001 level ( * * * ) by multiple comparisons with Duncan's test.
1 Mean values ± standard error of the mean are presented.Diferent superscript letters in columns (a, b, c) represent signifcant diferences at the p < 0.05 level ( *

Table 2 :
Chemical composition of three varieties of Pennisetum sp according to cutting age and interaction variety ×

Table 3 :
Correlation values between chemical composition and agronomic behaviour according to age and Pennisetum sp.variety 1 .Signifcant correlation according to Pearson's test (p < 0.05). *