02569naa a2200289 a 450000100080000000500110000800800410001902400330006010000170009324501190011026000090022950002330023852014600047165300220193165300120195365300260196565300220199165300230201365300260203665300340206265300230209670000160211970000160213570000130215170000150216477301000217910611062021-02-22       2020    bl uuuu     u00u1 u    #d7 aDOI: 10.1002/agj2.202372DOI1 aDO CARMO, M.  aNumber of samples for accurate visual estimation of mean herbage mass in Campos grasslands.h[electronic resource]  c2020  aArticle history: Manuscript received: 23 October 2019. Manuscript revised: 23 March 2020. Manuscript accepted: 01 April 2020. Accepted manuscript online: 09 April 2020. Version of Record online: 19 May 2020. Published July 2020.  aThe number of samples is a major issue when estimating the mean herbage mass of grazed paddocks. The aim of this study was to assess the number of samples required for accurate visual estimation of mean herbage mass in relation to the herbage mass heterogeneity and size of paddocks. Data were collected across scales of space and time (273 sampling events) from paddocks on Campos grasslands in Uruguay, using the visual estimation technique. The mean herbage mass of the paddocks ranged from 270 to 6350 kg DM ha&#8722;1 with coefficient of variation (CV) of 0.13 to 1.26. Twenty-four events representing four levels of herbage mass heterogeneity (CV = 0.3, 0.5, 0.7 and 1.0) × three levels of paddock size (small, 5?13 ha; medium, 41?67 ha; large, 100?140 ha) were chosen (two replicates per group), and analyzed for the probability that the estimation error exceeded 10% of the mean (10% error probability) using the bootstrap technique. The number of samples required for controlling the 10% error probability below 0.1 increased gradually from 50 to 150 per paddock as the CV increased from 0.3 to 0.7, then sharply to 350 until the CV increased to 1.0, with no effect of paddock size. Taking account of the distribution of CV (< 0.7 in nearly 80% of the events), we propose a general recommendation to take a minimum of 150 samples per paddock for accurate estimation of mean herbage mass in Campos grasslands irrespective of the size of paddocks.  aANIMAL PRODUCTION  aE-TANIN  aESTIMACION DE BIOMASA  aFIELD IMPROVEMENT  aLOTUS PEDUNCULATUS  aMEJORAMIENTO DE CAMPO  aMEJORAMIENTO DE CAMPO NATURAL  aPRODUCCIÓN ANIMAL1 aCARDOZO, G.1 aMECATTI, F.1 aSOCA, P.1 aHIRATA, M.  tAgronomy Journal, July 2020, Vol. 112(4), p. 2734-2740. Doi: https://doi.org/10.1002/agj2.20237