I want to study the orchestration of Mozart's "Credo" mass (k.257). There seem to be two editions available. One is available free on IMSLP; the other is fairly expensive to purchase. The free edition is a 1878 publication edited by Espagne and Nottebohm. The second is the Barenreiter edition.
Are there significant differences between the two editions that would affect my study?
Bärenreiter is a publishing house, so we should note that we're making generalizations about an entire company that publishes many different kinds of things. They're also a century old this year, so sometimes some older Barenreiter products might not fit the generalizations.
BUT yes, we can make such generalizations. Barenreiter tends to produce "Urtext" editions. Even this term is a generalization, but the "Ur-" means "original"; it tends to indicate an edition that takes a more scholarly approach, that cares about an early artifact like the composer's manuscript or first edition, and that makes minimal changes to this source, or makes any changes clearly documented and distinguishable from the composer's original.
In contrast, editions from some other houses like Schirmer or International often produce one individual editor's "version" of a piece. For instance, a Mozart violin concerto still sold today is also old enough to be on IMSLP, as the 1907 "version" of Eduard Herrmann. Herrmann edition completely alters bowing, adds articulation marks wherever and however he feels appropriate, adds dynamics and expressive markings that suit his interpretation, and even alters notes that he thinks might have been mistakes, or when he likes others better. Such an editor might translate period practices into contemporary equivalents, like writing out grace notes that would typically come with certain trills, or adding trills entirely. Of course they freeze in place a snapshot of the scholarship or opinion current at their time about what was "normally done" in the composer's period.
So in short, I would generally avoid such an edition. For some 19th-century works, such an edition often is the "Urtext"—Brahms's violin concerto is incomplete without Joachim's alterations. But for a work from an earlier century, like Mozart's, a 19th-century edition can be an important artifact in the reception history of the work, but I don't want to perform from it. I'd rather start from the composer's material and add my own interpretation than have to "weed out" someone else's. And I've often found Barenreiter editions (and those by Henle, a similar mostly-Urtext publisher) simply to be less error-prone and better typeset than some other publishers'.
Urtext editions can have their disadvantages—for instance, they assume that you do know how to realize ornaments or other unwritten practices, and they might leave some inconsistencies of the composer unchanged, when some alterations might seem obvious. They do require you to be inventive when adding expression and dynamics. Some Urtext editions can focus more on scholarship—adding nothing, which can leave a lot of work for the performer—and some can lean more toward "performing editions," making more changes. And sometimes the whole "Ur-" aspect of Urtext can be short-sighted or quixotic, and leave us with an incomplete picture. What of Vivaldi's Four Seasons, for which we have a handful of various manuscripts in various hands, as well as the first printed edition? And some of the manuscripts appeared after the printed edition, but are perhaps copies of an earlier version? What of all the changes and revisions Handel made to his Messiah within his lifetime? Is the earliest artifact really the "original" text with a "capital U," and is it really that much holier than these other early sources?
But a good Urtext can at least address these dilemmas. My Barenreiter Four Seasons has a helpful scholarly introduction about these sources, acknowledges that it is deriving primarily from the Manchester partbooks rather than the Amsterdam first edition that has informed most other editions, and marks important differences within the score with "M" for Manchester or "A" for Amsterdam.
I assume, you are asking for KV 257, since the name is not entirely unique.
This seems to be the difference between a pretty standard printed edition (appearing more than 100 years after the first performance it can't claim to be contemporary) and a well-edited edition revised using research result of nearly 150 years, which might include information from better copies found in the mean-time. The German wikpedia links to this Revision report (German) for an idea of the differences to expect. If Wikipedia is correct, the copy used for the first performance was identified 2007.
If studying orchestration patterns (a pretty abstracted level of information)I don't assume the difference is significant, while for any performance it may easily become important.
