This study aimed to evaluate performance and behaviour of Nellore steers (Bos taurus indicus) in integrated systems. Treatments consisted of three different systems: livestock; crop-livestock; and livestock-forestry, with triple rows of eucalyptus trees. The experiment was carried out in Sinop, MT, Brazil, in a randomised complete block design with four replicates, during two seasons (dry-rainy transition and rainy). The experimental period was from 19 September 2015 to 10 January 2016. Nellore steers used in the experiment had an average initial weight of 344 kg ± 11 kg, in continuous stocking and variable stocking rate. Forage mass accumulation rate was evaluated and, independently of the system, had the highest values recorded during the rainy season. Microclimate was evaluated through temperature and humidity index, black globe temperature and humidity index and radiant heat load. Highest values were obtained from 11:00 a.m. to 3:00 p.m. at the dry-rainy transition period for all systems. Animal behaviour was assessed every 10 min from 7:00 a.m. to 5:00 p.m., and averages were calculated for every 2 h. The longest grazing period was recorded from 9:00 a.m. to 1:00 p.m., and then after 3:00 p.m., regardless of the system. Rumination occurred more frequently from 1:00 p.m. to 3:00 p.m. during the rainy season, and also in the shade for the livestock-forestry system. Highest weight gains per area were recorded during the rainy season and for the crop-livestock system. It is concluded that the highest yields can be achieved with the inclusion of the crop component in livestock production systems, and that Nellore steers in silvopastoral systems prefer to ruminate and stay idle in the shade.

Adoption of silvopasture is increasing, but proper design and management of pasture–tree associations is required to sustain pasture productivity. Our objective was to compare agronomic and physiological characteristics of Brachiaria brizantha (Hochst. Ex A.Rich.) Stapf. (palisade grass) cv. Marandu growing in monoculture, or in silvopasture with Eucalyptus urograndis, in the Brazilian Amazon biome. Trees were planted in triple rows (intra-row spacing 3 m, inter-row spacing 3.5 m), referred to as a grove, with three groves per 2-ha silvopasture experimental unit. Space between groves was 30 m, and tree density was 270 ha–1. From October 2015 to September 2016, all experimental units were stocked continuously with cattle by using a variable stocking rate to maintain a Marandu canopy height of 30 cm. In the silvopasture, sampling occurred at 4, 7.5 and 15 m into the non-treed area from the outer tree row in a grove. Photosynthetically active radiation (PAR) reaching the Marandu canopy averaged 21% less for silvopasture than for grass monoculture; however, forage accumulation was not different between systems, nor were Marandu stomatal conductance, water-use efficiency (WUE), transpiration rate, transpiration ratio and leaf temperature. In silvopasture, PAR and forage accumulation were 33% and 29% less, respectively, at the 4-m distance than the 15-m distance from the edge row of trees, associated with lower transpiration rate and transpiration rate and greater WUE. We conclude that Marandu palisade grass can tolerate PAR reduction in silvopasture up to ~20% without reduction in forage accumulation, supporting its potential use in agroforestry systems for cattle production in the Brazilian Amazon Biome.