Two hypotheses proposed to explain the variance in second language tests are investigated. Hypothesis 1 (H1) claims that language skill is separable into components either related to linguistically defined categories (e.g., listening, speaking, reading, and writing). Another possibility (H2) is that second language ability may be a more unitary factor, so that once the common variance on a variety of language tasks is explained, essentially no meaningful unique variance attributable to separate components will remain. Previous studies have provided rather convincing support for H2. Data from 159 Iranian subjects at the University of Tehran, Iran who took a cloze test, a dictation, and the five subparts of the Test of English as a Foreign Language also support H2 in this report. However, when an oral interview task is included, the picture is less clear. Data from 106 foreign students (from mixed language backgrounds) at the Center for English as a Second Language at Southern Illinois University suggest the possibility of unique variances associated with separate skills. (Author/CFM)