The process of compiling the Henry James tales in collections index on this site has revealed the occasional unreliability of anthology editors. I have therefore compiled a page tabulating sample differences between James’s different versions of his tales, which started out with comparing first book texts with the New York edition but is now (August 2010) as comprehensive as I can make it, including variants from the majority of the early, magazine, versions also. I hope that, with the benefit of these tables, I will eventually be able to check up on the entries where the editor’s notes were originally taken on trust for inclusion in the index. The page will also be of use if you want to identify the version of an unlabelled text (for example one found in a multi-author anthology). Here are some tips on how to go about this.

how to use the sample differences

My tables provide sample ‘substantive’ differences from the beginning and ending of each text. Because interventionist editors and printers sometimes feel free to change spellings (for example British to American) and punctuation (at least within sentences), a ‘substantive’ difference is defined as one where the actual word or words have been changed by James. Occasionally, in a very lightly revised text, I have included a change of sentence break as substantive. Examples are taken from near the start and end of a tale to (a) make it easier to find them, and (b) provide some check that an editor hasn’t switched sources midway through (this does happen: see for example the Collected stories selected by John Bayley).

The locations of the samples are given by citing the chapter (where relevant), paragraph and sentence number, so you don’t have to read the whole block of text or find a copy of a particular edition to check page and line numbers. When there are only a few differences the given examples may be some way into a tale, from either end, but you are advised to check as many as possible if you want to be sure which text you have. In the tables, the text which changes between adjacent columns in shown in bold font.


This is now very nearly complete. Apart from the 19 tales which have only one version anyway, there are only 5 tales with no table and a further 5 where the table is missing a version (usually that appearing in a newspaper). A headnote with the table mentions any ‘missing’ source.

Included in the variants’ page therefore are all 55 tales which James selected for his New York edition (NYE, 1907-1909), comparing these final thoughts with their first book publications and with original magazine appearances (where I have access to those sources). This covers such staples as : Daisy Miller, The Aspern papers, The turn of the screw and The beast in the jungle.

The five post-NYE tales are also analysed, comparing their first, magazine text, with that of the 1910 book editions of The finer grain. In addition, the page includes variants for the 1896 tale Glasses, which James revised separately in 1915.

Finally, although quite a number of them are early tales and therefore appear first in the list, I have included as many of the non-NYE magazine to book comparisons that I have been able to make, given the availability of sources. In some cases I have relied on Aziz (see the next section) for the early tales, but many American nineteenth century periodicals are available online at the Cornell University Making of America site. The most difficult items to get hold of are the tales which first appeared in newspapers.

other published lists of variants

This section summarizes the current situation for published, printed lists of variants, which provide alternatives to my samples.

All the tales published by James up to The diary of a man of fifty in 1879 (thirty-six in number, see my chronological index of tales for a listing) are covered by the complete lists in the three-volume Tales of Henry James edited by Maqbool Aziz (OUP, 1973-1984). Full details of substantive variant readings, for those of the 36 that have them, are given in appendices to that edition, although I find Aziz at fault when he claims that there is only one substantive variant in The diary of a man of fifty (I make it at least five), so that I am not sure how reliable his lists are in the early tales for which I have not attempted a complete comparison! All the same, it’s a great pity that this set wasn’t completed (Aziz, a Professor at McMaster University, Canada, died in 2000, although the series seems to have come to a stop before that).

An earlier collection, listing variants in footnotes for just three tales, was compiled by Herbert Ruhm. His somewhat quirky selection of tales arose because they were ‘not readily available in recent editions’ and the variants are therefore equally disparate : Lady Barberina compares the book and New York edition texts, Benvolio the magazine and book texts, and Glasses the book and the late, Secker 1916, edition, which, as already mentioned, James revised especially (his last work on his series of tales).

As far as I can establish, just three paperbacks containing tales have lists of some significant textual variants. They are three volumes in the OUP World’s Classics series : The Aspern papers and other stories (1983), Daisy Miller and other stories (1985, 1998) and A London life, and The Reverberator (1989). To save you following each of those links, the tales covered by these volumes are: The Aspern papers, Daisy Miller, The death of the lion, Four meetings, A London life, The middle years, Pandora, The Patagonia and The private life. In all these cases the comparison is between the New York edition text printed in the volume and the first book edition text; the magazine originals are ignored. The only other easily available list of variants is that for The turn of the screw in the Norton critical edition of that work.


In compiling the extended tables of variants in 2010, access to original appearances of James’s tales, where not available online, has been essential and I thank those librarians providing this material, in particular the staff of the former Central Reference Library in Birmingham (now replaced by the Library of Birmingham), based on its floor 4, who retrieved many heavy and dusty volumes of source periodicals from the dark and chaotic stores on floors 3 and 6!

errors and corrections

If you discover misprints in the variants’ tables, discrepancies between texts there or errors in the reported text used in any particular reprint edition on the index, please feel free to contact me with all the relevant details. Together we may be able to build up a more comprehensive online resource. Hours of painstaking fun await you!

Adrian Dover – September 2013

you can go to the menu of tales’ index pages