Many thanks to Sharon, Barbara and Vic for replying to my question about which "electronic" versions of Twain's books might be suitable for computer-based text analysis. The consensus so far seems to be the the Gutenberg versions may not be academically reliable. There may be, according to Vic, some 1st edition electronic transcriptions and I will look into this. Perhaps someone on the list knows where they might be? Of course if I set out to transcribe from my own 1st editions (at least very early printings) I still have the same dilemma, namely accuracy and reliability. Ultimately it may come to choosing between the lessor of the evils. Vic points out that he found some of the Gutenberg files to be transcribed from the 4th American version and are thus less than desirable. My inquiry is primarily of a content analysis nature. So crossed t's and dotted I's, and proper illustrations and so on, the kinds of things that people look to for early 1st additions, are not that important, especially when looking across several thousand pages (three books) of content for trends in what is being said. I took Barbara's suggestion and contacted Dr. Hirst at the Mark Twain Papers in Berkeley to see what plans they may have to bring early editions online. But in the absence of something ideal I may have to fall back on what is available and acknowledge this in the study. Thanks again, Steve Crawford Jyväskylä Finland