I’ve noticed that, for example, the, uh, “preferred” display of the abbreviations for “kilowatt-hour” and “milli-ampere-hour” are kWh and mAh, respectively. What is the reference source for indicating the mixed case as opposed to all upper or all lower?
I suspect it’s just the SI abbreviation for the kilo- prefix is ‘k’, the abbreviation for watts is W, and the abbreviation for hours is ‘h’. Smash it all together and you get kWh, etc.
I have only an incomplete answer, and it doesn’t appear to be consistent with your observations.
When in school, I was taught to always use lower-case letters to represent prefixes with negative exponents (e.g., “m” for milli- and “p” for pico-) and upper-case letters for prefixes with positive exponents (e.g., “M” for mega- and “P” for peta-). We always used a lower-case “h” for hours.
This explains mAh (milli-amp-hours), as it distinguishes it from MAh (mega-amp-hours).
As for the lower-case “k” in kWh, I got nothing.
Ampere and Watt are units named after people, so perhaps that’s why they are capitalized.
I think the Caps are proper names - A for Ampere, W for Watt, etc.
Correct. It is customary to capaitalize Watts, Amperes, Coulombs, Farads, Henrys, Ohms and the like.
Basically, you get the correct abbreviation for a derived abbreviation by concatenating the correct abbreviations for prefixes and units - which happen to be k m, W, A and h respectively.
BTW, how would you know if a MAH is a milli- or a megaamperehour?
Small m for milli, capital M for Mega. The h should be lower case.
Uh-huh. Apparently dictionary editors haven’t received that memo.
Well, it isn’t a hard and fast rule
Dictionary editors have been known to have foilbles.
No. According to SI rules, the full names of the units are NOT capitalised: watt, coulomb, volt, farad, ohm, etc. Only the symbol sometimes uses a capital latter: W, C, V, F. But also note Pa for pascal (a two-letter symbol), and capital omega for ohm.
And you should indicate multiplication by a middle dot:
So, kilowatt-hour is kW·h.
(deleted)
Thanks for then corrections.
You would look four posts up from yours in the thread.
It should work that way, but it doesn’t. The prefixes k (10[sup]3[/sup]), h (10[sup]2[/sup]), and da (10[sup]1[/sup]) are in lower case.
If they were in upper case, we could have K, H, and D (instead of da) to balance out m, c, and d, and we wouldn’t have to deal with the oddity of da for deka.
You’re quite welcome! 
NIST list of prefixes and symbols: http://physics.nist.gov/cuu/Units/prefixes.html
Same source, list of style rules: SI Unit rules and style conventions checklist
Look around, it’s a great site for understanding the nature of measurement in science.
Tris
The use of lowercase prefixes for larger-than-one multipliers isn’t universally preferred. For example, the excellent unit conversion routine in Hewlett-Packard’s 28S calculator also accepted “K” for “kilo”.
Note that the W isn’t an abbreviation for watt, it’s a symbol for it, and likewise for all SI units. IIRC “h” for “hour” is an abbreviation and can take a period after it. The hour isn’t an SI unit, though its mixing in is practically accepted.
Sorry to nitpick, but I don’t see how this acceptance is an example of the correct case not being preferred. I just see it as good, non-enforcing design. Did it output K for kilo?
The hour is not only practically accepted - it is one of the “Units accepted for use with the International System” and therefore is on the same level as the liter and the tonne. And h is its symbol, no period. Germans, for example, write their Kilometer pro Stunde as km/h, not km/St. or such.
You probably meant German scientists, and not Germans in general. Because they’ll happily use km/st, just as us Norwegians use km/t.