We tested theories of eye movem ent control in reading by lookin g at parafoveal processing. According to attention-processing theories, attention shifts towards word n+ 1 only when processing of the ®xated word n is ®nished, so that attended parafoveal processing does not start until the programming of the saccade program ming to word n+ 1 is initiated (Henderson & Ferreira, 1990;Morrison, 1984), or even later when the processing of word n takes too long (Henderson & Ferreira, 1990). Parafoveal preview bene®t should be constant whatever the foveal processing load (M orrison, 1984), or should decrease when processing word n outlasts an eye movement program ming deadline (Henderson & Ferreira, 1990). By manipulating the frequency and length of the foveal word n and the visibilit y of the parafoveal word n+ 1, we replicated the ®ndin g that the parafoveal preview bene®t is smaller with a low-frequency word in foveal vision. Detailed analyses, however, showed that the eye movement programming deadline hypothesis could not account for this ®ndin g which was due not to cases where the low-frequency words n had received a long ®xation, but to cases of a short ®xations less than 240 msec. In addition, there was a spill-ove r effect of word n to word n+ 1, and there was an element of parallel processing of both words. The results are more in line with parallel