Some researchers have argued that low-achieving students may never acquire sufficient levels of academic vocabulary to be successful in school without some form of explicit vocabulary instruction (e.g. Snow, Lawrence, & White, 2009). In this paper, I summarize the available data on the efficiency, in words learned per minute of instruction, of explicitly teaching academic vocabulary. I also examine another possible source for academic vocabulary knowledge: pleasure reading, or what Krashen (2004) refers to as "free voluntary reading." A large corpus of popular, young adult fiction is analyzed to assess the likelihood that academic words can be acquired at least in part through reading. Comparing the relative efficiency of direct instruction and free reading, I found that reading is between two and six times more efficient than explicit teaching of academic vocabulary.