This paper explored the effects of the amount of extensive reading (ER) on writing ability. Participants were 157 first and second-year non-English majors at a private university in Japan who took a writing test in class. Some of them were reading extensively, while others had no experience in ER. The outcomes of Pearson's correlation indicated that ER correlated with total writing scores positively. A one-way analysis of variance with a planned comparison was performed with the total number of words read as the independent variable and total scores and scores for four areas of writing, task achievement, coherence and cohesion, lexical resource and language use, as dependent variables. The results showed that the amount of reading made a statistically significant difference in the mean scores of vocabulary and grammar. Moreover, findings suggested that participants who had accomplished reading more than 108,000 words statistically significantly wrote better lexically and grammatically.