The present study was designed to investigate the role of phonological coding in reading processes both of the Kanji-Kana mixed form sentences and of the Kana alone form sentences by using of a concurrent vocalization task. The following four factors were manipulated: (1) script types of sentences, (2) difficulty levels of story contents, (3) types of the reading comprehension question and (4) the ways of reading. The results indicated that (1) the concurrent vocalization task did not interfere with reading time regardless of the script types of stimulus sentences, (2) the intrasentential comprehension of a "difficult content" story was suffered from the concurrent vocalization task and that (3) the intersentential comprehension of a "difficult content" story, particularly when it was written in Kana script alone, was negatively affected by the concurrent vocalization task. These results were discussed with respect to the relationships among the necessity of phonological coding, script types of sentences, and the degree of comprehension demands.