Jordan Journal of Agricultural Sciences, Volume 16, No.3 2020 DOI: https://doi.org/10.35516/jjas.v16i3.61 -71- Corresponding author E-mail: a.s.alkhtib@gmail.com © 2020 DSR Publishers/The University of Jordan. All Rights Reserved. Eye-balling and heart girth models for live weight estimation of highly admixed Sudani Shorthorn Zebu Cattle for Precise Production and Veterinary Services Ashraf Alkhtib 1 , Obaida Almasri 2 , Emily Burton 1 , Jane Wamatu 3 1 Nottingham Trent University, School of Animal, Rural and Environmental Sciences, Southwell, Nottingham, UK. 2 General Commission for Scientific Agricultural Research, Damascus, Syria. 3 International Centre for Agricultural Research in Dry Areas, PO Box 5689, Addis Ababa, Ethiopia. Received on 7/2/2020 and Accepted for Publication on 25/2/2021. ABSTRACT Cattle production is a key pillar of food security in Africa. The majority of African cattle are highly admixed with unknown breed composition. Accurate estimation of the live weight (LW) of these cattle would improve the precision of feeding, veterinary services, and pricing resulting in an improvement in profitability. This study assessed estimating LW of admixed Sudani zebu cattle using eyeballing and heart girth (HG) models. Live weight and HG of 432 Baggara cattle, an admixed Sudani breed, were measured. Three models (a simple linear, a simple linear with box-cox transformed LW, and a quadratic) were generated using 382 heads and validated using 50 heads. A published model (LW (kg) = 3.54*HG (cm) - 322.63) was validated using the data of this study. The error of LW estimation by a breeder and five cattlemen were recorded. All constructed models had high R2 (0.725 - 0.728). However, the 95th percentile of the prediction error of the constructed and published models was higher than 20%. The 95th percentile of LW estimation error of all participants was high (>20%). Accordingly, HG models and eyeballing are not suitable methods to determine the LW of highly admixed zebu cattle for production, veterinary, and marketing purposes as they are prone to a high rate of error. Keywords: Indigenous, cattle, linear, non-linear, prediction error INTRODUCTION Cattle in Sub-Saharan Africa play a key role in the livelihoods of farmers since they are the main source of drought power, manure, food (6.5 million tons of red meat and 35.6 million tons of milk (FAOSTAT, 2018)) and cash (Rege, Kahi, & Okomo-Adhiambo, 2001). Furthermore, cattle have social and political values that impact the social life of farmers in Africa (Ghaffar & Ahmed, 2014). The majority of cattle in Africa are admixed with unknown breed composition due to uncontrolled crossbreeding and arbitrary mating which resulted in high variability in appearance and body conformation. Precision in agriculture is now widely regarded as a key route to optimal use of global resources in food production, but often focuses on the application of modern technologies (Fuglie, 2016). This focus overlooks the importance of generating simpler data such as correct estimates of livestock weight in developing countries to ensure livestock are optimally maintained and used. Live weight (LW) of cattle is closely related to nutrient requirements (Kearl, 1982), milk production (Kanuya et