1993 0-7734-9276-3 This study attempts three tasks: 1) to suggest a procedure for ascertaining statistically meaningful relationships among manuscripts of a given work by the use of common statistic procedures never used before in textual study -- principally linear and multiple correlation tests, cluster analysis, and factor analysis; 2) to apply that method to the "landmark" manuscripts of the Canterbury Tales, i.e., those published within ten years of the death of Chaucer which establish independent lines of transmission; and 3) to draw from that testing a series of limited conclusions and inferences. Statistical procedures are explained, and where general conjectures concerning debatable matters are necessary, they are stated hypothetically in the text and argued briefly in the notes.