Two experiments were carried out to examine whether reading task influences the size of the attentional span during reading using the eye-movement contingent display change paradigm.In Experiment 1,thirty participants read sentences containing adjective-noun pairs while their eye movements were recorded.The size of the display change effect on the noun (its letters were initially replaced with random letters) was examined during proofreading vs.reading for comprehension.The results showed that the size of the change effect did not depend on the reading task.In Experiment 2,forty participants read a long expository text describing four different countries from one of two possible perspectives ("Imagine that you will have to move to Honduras / Pitcairn Island") or with the instructions to read for comprehension.The size of the change effect on target words within predefined perspective-relevant and–irrelevant sentences was examined.A reliable change effect as well as a parafoveal-on-foveal effect was obtained for during reading of irrelevant sentences.However,these effects were absent for relevant sentences.Experiment 2 demonstrates that readers attentional span is shrunk when reading materials highly relevantto the reading goal.Overall,the results demonstrate that different task instructions exert different effects on the eye movement control during reading:tasks that emphasize lexical-orthographic processing (e.g.,proofreading) induce a general slow-downin reading,whereas a task that increases readers attention on specific parts of the text modulates the engagement of attention.