Se expuso a tres palomas privadas de comida a sesiones experimentales de 50 ciclos de 64 s cada uno (ciclos T) dentro de los que siempre se presentó un dispensador de comida durante 3 s (ER1). El dispensador se podía presentar nuevamente (ER2) cuando terminaba el ciclo si el sujeto se abstenía de aproximarse al ER1, de lo contrario, este se retiraba y se cancelaba la entrega del ER2. En cinco fases experimentales se presentó el ER1 3, 32, 16, 8, y 3 s antes del final del ciclo T. Además, conforme a un diseño ABA, un cambio de color en una tecla de respuesta señaló (A1 y A2) o no (B) las presentaciones del ER1 sirviendo como un operando distractor. Se encontró que el número de ER2 obtenidos disminuyó conforme el ER1 se presentó más cerca del inicio del ciclo T. Así mismo, se observó que únicamente durante las condiciones A de la primera fase experimental, el número de ER1 interrumpidos fue relativamente bajo y el número de picotazos a la tecla fue comparativamente mayor. Estos hallazgos reprodujeron y extendieron los resultados de otras investigaciones en las cuales se estudió la conducta autocontrolada de comer.Self-controlled eating behavior was studied in three food-deprived pigeons exposed to a procedure consisting on fifty 64-s time cycles (T cycles). In each cycle a food-hopper was always presented once for 3 s (SR1) and could be presented for other 3 s (SR2) after the T cycle elapsed, according to the following contingency. If the subject approached to SR1 it was withdrawn and SR 2 delivery was cancelled; otherwise SR2 was presented when the T cycle elapsed. In five experimental phases SR1 was presented 3, 32, 16, 8, and 3 s before the end of the T cycle. Additionally, according to an ABA-design, a color change of a response key signaled (conditions A1 and A2) or not (condition B) SR1 presentations and served as a distracting operandum. It was found that the number of SR2 obtained was lower as SR 1 was presented closer to the beginning of each T cycle, regardless of signaling SR1 presentations. It was also found that only in the A conditions of the first experimental phase (3 s) SR1 interruptions were relatively low and R>0 of key pecks was comparatively high. Globally, these results replicated and extended those of previous studies in which self-controlled eating behavior was explored.