The trial was conducted to evaluate the use of sorghum silage, maniçoba hay, and two concentrate levels (10.0% and 15.0% DM) in diets based on spineless cactus of dairy cows. Eight Girolando cows were used with average body weight and milk yield of 463 and 9.65kg, respectively, with 95 days of lactation. The experimental design was a 4 x 4 Latin Square in a 2 x 2 factorial arrangement (maniçoba hay and sorghum silage x 10.0% and 15.0% of concentrate), two simultaneous Latin Square were used. Each experimental period lasted 21 days; 14 days for adaptation and seven for data and sample collection. There was no interaction between roughage and concentrate level for any of the variables. Thereby the factors were analyzed and discussed separately. The intake of crude protein, neutral detergent fiber and total digestible nutrients differ according to roughage used with lower intake for maniçoba hay (1.50; 4.10; 6.17kg/day, respectively). The apparent digestibility of dry matter, organic matter, crude protein and neutral detergent fiber differed according to the roughage, and lower values were found for maniçoba hay (51.70; 53.91 and 20.74%, respectively). The milk yield with and without adjustment to 4.0% of fat (9.65 and 9.1kg/day, respectively) and the milk composition did not differ between roughages or between concentrate levels. The efficiency of microbial protein synthesis differs between roughages and was lower in the treatment with sorghum (99.46g/kg TDN). The concentration of plasma urea differed between the ratios of concentrate with lower amounts in the diet with 15.0% concentrate (1.41mg/dL). To meet the nutritional requirements and cost reduction with supplementation a supply with sorghum silage as roughage and 100% of concentrate for diet of lactation cows is suggested.
digestibility; milk yield; semiarid; microbial protein; roughage