(very "Stream of consciousness" comments here.. I'm just jotting everything down as it comes to me)
I might be a tad more relaxed on the "crossing voices" part than Ron, but otherwise, his comments are all pretty much spot on.
1. Give singers time to breath
2. Be careful of large or particularly dissonant intervals within a single melodic line (they should generally be well-supported in the other parts)
3. You MUST respect the prosody of the text you are setting. which means taking the text aside and marking off all the tonic and secondary accents in the lyrics you intend on setting.
4. Respect the vocal ranges you are using. Try to keep "extremes" to a minimum.
5. Remember that range extremes have different uses: high notes are harder to sing softly, low notes are harder to sing loudly.
6. Avoid "closed" or "tight" vowels on very high notes (oo, i, ee)
7. Give the text time to breath as well.
8. Strive to make EACH and every voice a uniquely singable experience for the performers. There's nothing more boring that having a part that's just "filler" while the other section gets to sing all the melodies.
9. Text doesn't HAVE to be treated in a monophonic fashion (every syllable in every voice matching the exact same rhythm)
10. Where text is NOT monophonic, give important lines the chance to shine through so that the text CAN be understood.
11. Treat registers carefully... if everyone is in a comfortable median register and you give a melodically un-important part to a voice that is in a very high register, you WILL hear the un-important part more than the others.
12. Likewise, avoid giving the most musically important material to a voice in a weak register, while the rest of the chorus is in a very powerful register.
13. distance between voices, as Ron said, should avoid intervals greater than an 8ve between soprano and alto, and between alto and tenor (two voices that CAN overlap without deleterious effect). You can go considerably wider between basses and tenors, 8ve and a half is not unusual.
14. "Voice leading" is considerably more important in vocal music than in any other type. Dissonances should be prepared and resolved in the right voice.
Vocal music is, by its very nature, more conducive to less-avantgarde music.
Unless accompanied, a chorus has only its collective ear to use as (I always have trouble finding the right word for this) a "roadsign". They don't have open strings to rely on. They can't just blow into the instrument and get a specific sound.
A capella music tends to gradually go flat the longer a piece of music lasts. Your chorus might start out singing "In C" but end up in Bb or worse by the final cadence.
There are different vocal effects that can be achieved, depending on context and the writing of the music.
For example "sotto voce" ("subtle voice") which is a very soft effect, subdued, almost a whisper, a beautiful effect in so many circumstances.
"Falsetto" is usable by men's voices, though be warned, many men are actually incapable of producing a falsetto tone. This CAN be used to cheat high notes in soft dynamics.
The "richest" and more powerful sound you can achieve with a chorus will more than likely be when the music is set in relatively close harmony.
A 4-part choir with enough members can easily divide each part in two, giving you an effective 8-part choir. The more amateur the choir, the more singers it takes to achieve this. A good professional chorus can handle divisi starting at as few as 12-16 singers. While an amateur chorus might have trouble even with 20 people on each part.
Be wary of non-harmonic notes (notes outside your harmonic framework) that need to be attacked without preparation.
For larger and more qualified groups,
clusters, when properly prepared, are a gorgeous effect.