Rob Prester (Pf, New York, NY) opened an E-mail dialog within the ACMP Board and Council members by posing this question: "Does ACMP have any available info about standardized guidelines for measure numbering?"
Board Member Susie Ikeda (Vn, Cambridge, MA) responds: It may seems silly, but consistent measure-numbering is actually a meaningful and not-always-obvious issue, particularly when people show up at a workshop with their own parts numbered using one convention and others show up having used different conventions. (Do pickups count as a bar? How do you count first endings?) Being off by even one bar prohibits the group from being able to start at exactly the most efficient place.
I'm active at the Bennington Chamber Music Conference (cmceast.org). For decades we've used a generally successful measure numbering convention that seems to match most pre-printed numbering conventions. Thanks to Eve and Don Cohen we also have a vast listing of measure number totals (by movement), The Cohens meticulously and brilliantly compiled this list, manage, and generously share it. You'll find the numbering guidelines, and within that you'll find a link to the measure number totals page by clicking on the link http://don-eve.dyndns.org/musicparties/barnum.html Of course, if one never has to stop, and never has to rehearse anything, then the numbers are indeed pointless!
Advisory Council Member Missy Goldberg (Vn/Va, Chevy Chase, MD) concurs: If you haven't tried numbering measures, you may not appreciate the value of The Cohen's reference list of measure number totals. Even if you follow the conventions, it's very easy to make a mistake — either by missing some subtle oddity of the music or, more likely, by simply miscounting. Knowing the actual number of measures in the movement makes it easy to see if you have counted correctly.
Advisory Council Member Ted Rust (Ob, EH, Berkeley, CA) writes: According to Brian Blood, www.dolmetsch.com/musictheory3.htm#barnumbers, bar lines date from the seventeenth century, and rehearsal letters began appearing in orchestral scores and parts in the early 19th century. My own guess is that they were picked up by publishers from conductors' own markings jotted on manuscripts during rehearsals when they needed to tell their players where to start. He doesn't say if bar numbers followed soon after.
Letters certainly become less useful and measure numbers more so as compositional style has moved away from the clear, regular divisions of classical and early romantic music and a more detailed navigation system was needed. Debussy's Pelléas et Mélisande, for example, is hell to rehearse with its rehearsal NUMBERS instead of letters ("Violins, play what you have nineteen bars after eighteen in Part III." Huh??
My Dover facsimile of an "early" edition of Dvorák's Serenade for Winds, Opus 44 (1878) has measure numbers AND rehearsal letters whereas his Serenade for Strings, Opus 22 (1875) has neither! (Maybe wind players needed more help, or maybe the practice was just being introduced.) Measure numbering was still not commonplace before the mid-20th century. For example, the Peters Bach editions from the 1930s have rehearsal letters but no measure numbers, whereas Henle editions from the 1970s have numbers only.
I'd love to dig into a good music library !
Advisory Council Member Jerry Fischbach (Vn, Glen Dale, MD) writes: This is indeed an interesting and deceptively complex question. The music publishing industry has certainly been dealing with this question for at least a century. The Bennington numbering rules are outstanding! Models of concision and thoroughness. Just for the record, I explored two of my publishing sources as follows:
Composer/arranger George A. Speckert, an editor at Baerenreiter, (and former student of mine) defines the Baerenreiter conventions
The first full measure is "1" - , do not count upbeats.
The first and second ending brackets are numbered "a" and "b" ( i.e. 16a, 16b.) This will only be printed if the bracket is at the beginning of a line. No special treatment for repeats and jumps — even if the repeat is in the middle of a measure.
In cases where one voice is notated with a repeat and another without (i.e. different phrase endings), both numbers are noted, like 17 (25).