THE PURPOSE OF THIS ANALYSIS WAS TO DISCOVER CERTAIN ASPECTS OF STYLE (BOTH LEXICAL AND GRAMMATICAL) WHICH COULD BE COUNTED AND WHICH WOULD, WHEN COMPARED, DIFFERENTIATE BETWEEN THE WRITTEN PROSE OF TWO AUTHORS. THE SUBJECT SELECTED FOR ANALYSIS WAS "THE O'RUDDY," BEGUN BY STEPHEN CRANE AND COMPLETED BY ROBERT BARR. SINCE THERE WAS NO RECORD OF THE POINT OF TRANSITION, THE AUTHORSHIP OF VARIOUS CHAPTERS WAS IN DOUBT. THE RESULTS OF A PILOT STUDY, DESIGNED TO GENERATE TECHNIQUES FOR SUCH WORK, INDICATED THAT NOT ONLY SINGLE ITEMS, BUT ALSO PATTERNS OF INTERRELATIONSHIPS WOULD DISCRIMINATE. AFTER PRELIMINARY ANALYSIS TO GENERATE 18 WEIGHTED VARIABLES, AND A SUBSEQUENT CHECK ON THE VALIDITY OF THOSE VARIABLES AND THEIR WEIGHTING, THE NEW TECHNIQUE WAS APPLIED TO PARAGRAPHS, BLOCKS OF 1,000 WORDS, AND WHOLE CHAPTERS OF "THE O'RUDDY," RESULTING IN THE ASSIGNMENT OF CHAPTERS 1-24 TO CRANE, CHAPTER 25 AS TRANSITIONAL, AND CHAPTERS 26-33 TO BARR. THIS RESEARCH METHOD OFFERS A WAY OF ANALYZING NOT ONLY THE PATTERN OF STUDENT COMPOSITION, BUT ALSO THE PATTERN AND SPECIFIC STYLISTIC DEVICES OF ACCEPTED AUTHORS. THE RESULTS OF THIS STUDY HIGHLIGHT THE IMPORTANCE OF THE VARIABLES IN THE COMPOSING PROCESS. (THIS ESSAY APPEARED IN "THE COMPUTER AND LITERARY STYLE," KENT STUDIES IN ENGLISH, NUMBER 2, KENT STATE UNIVERSITY PRESS, OHIO, 1966. IT IS BASED ON A DOCTORAL DISSERTATION COMPLETED AT HARVARD UNIVERSITY, 1963.) (MM)