An experiment was carried out in Garca-SP to evaluate the effects of application using a direct sprayer with and without "Napoleons's hat" protection for total action herbicide tank mixtures on dwarf internode castor bean crop during the 2004/2005 agricultural year, with Lyra hybrid in a 1.0 x 0.5 m spacing. The experimental design was randomized complete blocks, with treatments in a factorial scheme 2 x 10, with four replications, the first factor being direct sprayer application with and without "Napoleon's hat" protection and the second application of the herbicides glyphosate (0.72 kg ha-1); glyphosate + 2,4-D (0.72 + 0.35 kg ha¹); glyphosate + flumioxazin (0.72 + 0.025 kg ha-1); glyphosate + carfentrazone-ethyl (0.72 + 0.016 kg ha-1); glyphosate + diuron (0.72 + 0.75 kg ha-1); MSMA + diuron (1.44+ 0.75 kg ha-1); paraquat + diuron (0.9 kg ha-1); paraquat + diquat (0.20 + 0.20 kg ha-1); paraquat + bentazon (0.40 + 0.48 kg ha-1) and hand weeded check. Herbicide application was accomplished in weed post-emergence, between crop lines,using a carbon dioxide-pressurized backpack sprayer, with one XR 8002-VS nozzle for "Napoleon's hat" protection and two XR 11002-VS nozzles,50 cm spaced , without protection, both using 200 L ha-1 of water. Glyphosate and paraquat + bentazon, glyphosate + 2,4-D and paraquat + diquat tank mixtures provided the highest yield levels for Lyra castor bean hybrid, when applied using a direct sprayer with "Napoleon's hat" protection. Paraquat + bentazon (0..4 + 0.48 kg ha-1) and paraquat + diquat (0.20 + 0.20 kg ha-1) tank mixtures were the most indicated for direct sprayer with nozzle protection.
Ricinus communis; weed; selectivity; yield
