This paper provides a discussion of the results of a cross-sectional examination of linguistic variables that are predicted to influence L2 reading fluency. This study is part of a larger, longitudinal mixed-methods study into reading fluency development using online Timed Reading (TR) with participants from a mid-to-high level private university in western Japan. The larger study will also be briefly explained to provide some background. For the study presented in this paper, participants read two, short, equivalent, graded, online texts against a clock and then answered comprehension questions without recourse to the texts. A custom-made web application was used to administer the readings. The texts used were all 300-words long and graded using tools available on "The Compleat Lexical Tutor" website (Cobb, 2003) to be within the first 1,000 words of English. Data from a battery of tests that included a paper-based vocabulary size test (Nation & Beglar, 2007), and a computer-based word-recognition reaction time test were recorded. These data were triangulated with graded text reading performance data to assess the relative importance of the components assessed. The relationship between the measured variables is explained using correlation and regression analysis, providing insight for reading researchers and teachers. [For full proceedings, see ED565044.]