ABSTRACT. The forage palm is one of the main forages of ruminants in semiarid regions. Measurements of leaf area are required in agronomic studies because they are one of the main parameters used to evaluate plant growth. The objective of this study was to validate and define the best models for estimating the area and weight of Giant Sweet clone (Nopalea cochenillifera) forage cladodes in a non-destructive way based on the linear dimensions of length, width and thickness. There were 432 randomly measured cladodes at 550 days after planting. The length, width and thickness of each cladode were measured using a digital calliper. The cladodes were weighed individually. The cladode area was calculated by the gravimetric method. The power regression model was the most efficient method to explain the cladode area as a function of the product of length by width, while the gamma model was the most efficient method to explain the weight of cladodes as a function of the product of length by width and thickness. The power model, R C A ̂ = L W 0.982, and gamma model, W C ̂ = 0.536 T + 0.028 L W, were used to determine the area and weight of Nopalea cochenillifera Giant Sweet clone cladodes, respectively, based on the values of linear dimensions measured independently of the order of the cladode.

ABSTRACT. Agricultural production in semi-arid regions is limited due to water availability. In addition, the water quality available for irrigation is often compromised due to the high salt content present. Millet is a forage species considered tolerant to water deficit and moderately salt tolerant. In view of the above, the objective was to evaluate the growth of millet under water and saline stress associates. The experiment was carried out in a randomized block design, in a 4x3 factorial scheme, composed of four levels of water replacement, based on crop evapotranspiration (ETc): 25%.ETc, 50%.ETc, 75%.ETc and 100%.ETc and three levels of water salinity (0.03, 2.0 and 4.0 dS m-1). With 25%.ETc independent of salinity, all morphological characteristics of millet were affected, occurring death of plants in the initial growth phase. In the absence of salt and greater availability of water, greater plant growth occurred. With respect to salinity, there was a reduction in the increment of all variables evaluated, with the highest reduction at the highest saline level (4.0 dS m-1). Water and salt stresses, when associated, reduce the growth of millet, since concentrations above 2.0 dS m-1 and less than 50%.ETc compromise its full development, providing declines in yield.

Resumo Estudos com pastagens necessitam de informações sobre a área foliar, por ser um dos principais parâmetros de avaliação do crescimento das plantas. Desse modo, objetivou-se estimar a área da lâmina foliar do capim-pangolão (Digitaria pentzii Stent.), utilizando métodos não destrutivos por meio de análise de modelos de regressão. O delineamento utilizado foi em blocos casualizados, com três alturas de corte (10, 15 e 20 cm) e quatro repetições. Foram coletadas aleatoriamente 300 lâminas foliares do capim-pangolão e determinados os seus respectivos comprimentos (C) e larguras (L), com uso de paquímetro digital. A área da lâmina foliar do capim-pangolão foi estimada pelo método gravimétrico, sendo utilizados os modelos de regressão linear e potência para explicar a área das lâminas foliares em função do produto do comprimento e máxima largura. A área da lâmina foliar real apresentou valor médio de 18,64 cm2, variando de 4,29 a 45,95 cm2. A área da lâmina foliar do capim-pangolão, independentemente da altura de corte, foi estimada com melhor acurácia pelo modelo potência. O modelo potência, Ŷ=CL1,007, pode ser usado para estimar a área da lâmina foliar do capim-pangolão com base nos valores de comprimento e largura da lâmina foliar dessa espécie.

Abstract Pasture studies require information on leaf area, as it is one of the main parameters for evaluation of plant growth. Thus, the objective of this study was to estimate the leaf blade area of pangolão grass (Digitaria pentzii Stent.) using non-destructive methods by regression model analysis. The experimental design consisted of randomized blocks, with three cutting heights (10, 15, and 20 cm) and four replications. Three hundred leaf blades of pangolão grass were randomly collected, and their respective lengths (L) and widths (W) determined using a digital caliper. The leaf blade area of pangolão grass was estimated by the gravimetric method, using linear and power regression models to explain the leaf blade area as a function of the product of L and maximum W. The real leaf blade area presented an average value of 18.64 cm2, ranging from 4.29 to 45.95 cm2. The leaf blade area of pangolão grass, regardless of cutting height, was estimated with greater accuracy by the power model. The power model, Ŷ=LW1.007, can be used to estimate the leaf blade area of pangolão grass based on leaf blade L and W values.

ABSTRACT. Leaf area measurements are of the main parameters used in agronomic studies to evaluate plant growth. The current study used a non-destructive method based on linear leaf dimensions (length and width) to select the regression model to estimate millet (Pennisetum glaucum) leaf area. For two millet genotype (IPA BULK 1 BF and ADR 300) 128 randomly-chosen leaves were measured at different vegetative growth stages. Measures of length and width of each leaf were made using digital calipers. Leaf area was measured using a gravimetric method. The best-fit leaf area estimation model was selected via linear, potential and gamma regression models. Leaf area values varied from 3.02 to 209.21 cm2. The average value was 95.31 cm2. The potential regression model exhibited lower residual sum of squares and Akaike's information criterion and similar determination coefficient and Willmott index. Thus, potential regression was more efficient in explaining the leaf area of millet, independent of the genotype, when compared to other models evaluated in this research. Length (L) and width (W) could be used in the following potential regression model Y ̂ = 0.879 L W 0.971to estimate millet leaf blade.