Compilation of a list of the most common phrases used in reading was begun with the rationale that the quick recognition of phrases would facilitate reading comprehension. These first efforts showed that categorizing phrases by parts of speech did not provide acceptable levels of accuracy. The system that was effective, however, used a computer program that recorded every consecutive two- and three-word sequence in the text sample and determined which of these word strings recurred most frequently. The computer program makes possible samplings of large amounts of text-50,000 words or more-thus eliminating the idiosyncrasies of text sampling. The researchers who developed this system believe that the common phrases it identifies should be taught in much the same manner as common words are now taught in beginning reading instruction. (RL)