Mixtures of African breadfruit (Treculia africanaDecne), corn, and defatted soybean were extruded in a singlescrew Brabender laboratory extruder at process variables derived from a second-order central composite design. The variables consisted of feed composition (0-100% breadfruit, 0-55% soybean, and 0 or 5% corn); fed moisture (15-27%), and screw speed (100-180 rpm). Effects of these variables on residence time distribution were investigated using Congo red as tracer. The extrudate spent longer time in the extruder as feed moisture or screw speed was decreased from 27% to 15% or 180 to 100 rpm, respectively, thereby increasing the residence time distribution characteristics. At 70% feed composition and screw speed of 140 rpm, mean residence time ( t ) increased from 40 to 50 s, whereas extrudate total collection time (t c ) increased from 65 to 70 s. At screw speed of 100-180 rpm, mean residence time ( t ) decreased from 55 to 35 s, whereas extrudate total collection time (t c ) decreased from 75 to 65 s. At 100% African breadfruit composition, these time values decreased to 35 and 70 s, respectively, at the same screw speed, indicating the significant influence of feed composition and soybean addition to the mixture on residence time distribution. Residence time distribution curves indicated an early breakthrough time of 20 s at maximum screw speed (180 rpm), minimum tail of 65 s, and a plug flow pattern of the extrudates.tracer concentration appearing at the outlet at time, t T time (s) t mean residence time (s) t c total collection time (tail) (s) θ normalized time, a dimensionless parameter (t/t) E(t) exit age distribution. E(t) curves represent the variation of the dye tracer concentration at the exit F(θ) F(t)=∫E(t). F(θ) curves represent the accumulated quantity of tracer at the exit at a given time. These two curves {E(t) and F(θ)} are generally used to quantify the residence time distribution (RTD) t m minimum residence time (s) T extrudate temperature (°C) Y dependent process variables (T, RTD, t m, tand t c ) k number of independent process variables (three in this case) X 0 estimated regression coefficient β 0 intercept (constants and regression coefficients of the model) ɛ random error term β 1 X 1 -β 3 X 3 linear (first)-order effect b 1 X 2 1 À b 3 X 2 3 quadratic order effect β 12 X 1 X 2 -β 23 X 2 X 3 cross product (interactive) effect