West Central region of Brazil presents a well-defined dry season during the winter and a rainy summer season, therefore, forage alternatives such as pearl millet are evaluated to reduce the effects of the dry period. The objective of this study was to evaluate the morphological composition of millet at the south-east region of Goias State of Brazil. Plots were disposed in a completely randomized block design using a 3 × 2 factorial arrangement (3 sowing densities: 10, 15 and 20 kg.ha-¹ and 2 cutting heights: 50 and 70 cm) with 20 cm residual height and 6 replicates. The samples were collected at 3 points in the plots and separated and weighed in stem, leaves, dead/senescent material and panicles. Statistical analyzes were performed at 5% probability and means comparison evaluated by Tukey test. There was no significant effect of seeding density on the plant structural fractions nor interaction between sowing density and cutting height. Average values for leaf, stem, panicle and dead matter fractions were 27.24; 46.26; 19.63 and 6.42, respectively, showing a compensatory effect of the sowing rate on the plant components distribution, which allowed that the canopy structure was preserved even when sown changed the final population (53, 68 and 78 plants/m², respectively for 10, 15 and 20 kg seeds/ha densities). The effect of cutting height was significant for all fractions (Table 1). The higher cutting height decreased the leaves percentage and increased the stem percentage resulting in a lower leaf/stem ratio, probably due the longer period necessary to reach 70 cm height, with maturity, the plants also presented higher percentage of panicle and dead material. Considering that the grazing aims harvesting green leaves and a longer grazing period, it is recommended to use millet management at 50 cm. Table 1. Morphological composition of millet off-season on two cutting heights

Fraction (%)	Cutting Height (cm)		CV (%)1	P-value2
	50	70
Leaf	35.75	18.74	15.80	<0.01
Stem	44.63	47.89	4.24	<0.01
Panicle	14.54	24.72	7.09	<0.01
Dead material	4.67	8.39	27.12	<0.01

¹Coefficent of variation. ²F test.