SPELLCASTING DURING MELEE
Believe it or not, here (along with "breaking off from melee," about 5 pages down the line) we see the roots of attacks of opportunity, all the way back in first edition.
In later editions, with their 6-second melee rounds (equivalent to a segment in AD&D), it was sometimes hard to reconcile spell disruption if you really took the time to think about it. After all, if a spell took one action to cast (and you could even move after), how can someone hit you and disrupt it?
In AD&D, most spells higher than third level had a casting time of at least one round, which meant they took a full minute of ritual to get off (Spells of third level and below, incidentally, mostly had casting times equal to their level in segments, which, as a side note, is why I use a d10 for initiative; lets me know exactly what segment a character acts on).
A great deal goes on in that minute's worth of time--and there's plenty of time for someone to whack you with a club or shoot you with a bow while casting. Also, there was no concentration check for disruption. You got hit and took damage, your spell fizzled. Period. This may seem harsh, but consider the sheer power of higher level wizards in first edition. The ability to disrupt their spells was a powerful balancing factor in the game.
This means that while in melee combat, a wizard could feasibly and safely cast a first level spell, so long as no enemies act on the same segment as he. A second level spell could be safely used so long as nobody acts within two segments, and a third, so long as he has three segments' worth of free movement.
A fourth level or higher spell, however, he'd have problems with unless his enemy missed him on a melee attack.
The book gives us five guidelines for spell casting in melee.
1. Casters must note what spell they intend to cast (and its casting time) before initiative dice are rolled. This makes sense on some level, particularly if he may be casting a higher level spell requiring him to begin casting immediately. On another level, if he's casting a lower level spell requiring only one or two segments to complete, there's no reason (in my opinion) that the DM shouldn't allow him to change his mind if, for example, he's attacked before his turn.
This brings up another interesting interpretive point. Do magic users cast on their initiative segment or no? The book indicates that they attack along with ranged attacks, which would mean that spells, like missile fire, are always resolved first...or at least, are begun first. That would seem to jibe with the requirement above that casters have to declare their spell before initiative is rolled. The book remains difficult on this, however; do missile weapons act on a completely separate initiative sequence? Generally speaking, I have everyone roll for initiative once. I then resolve missile fire in the order in which their dice would otherwise indicate (basically ignoring the initiative scores of those who are engaging in melee combat until this is done), then move to melee. Magic users casting spells would, I presume, work the same way. However, while they'd begin casting along with ranged troops, they would complete casting on the appropriate segment of melee combat as indicated by their spell's casting time. Thus, all spell casters begin casting on segment 1 of melee combat, regardless of their initiative die. First level spells would then activate on segment 2, second level on segment 3, and third level on segment 4. This leaves an opening for a caster using a first level spell to be disrupted, if an enemy has rolled a 10 (or higher) for initiative (on my house initiative system), allowing him to act on the first segment (basically just after missile fire goes off).
2. The by-the-book system for initiative isn't much different; it's just that all attacks are begun on segments 1-6 of combat (p. 65, DMG). The second point for handling spell casting in melee--ignoring my interpretation using my initiative system above--states that all attacks against the spell caster come on the segment indicated by the winning side's initiative die, even if the PCs' side is the winner. I'm having a hard time working this bit out, but I presume the idea is that it's a balancing factor to allow enemies a chance to disrupt the spell.
3. Intelligent creatures will, lacking other specific motivations, always attack a spell caster first, as all intelligent creatures can recognize the danger of spells and magic users.
4. The spell caster loses his or her Dexterity Reaction Bonus to AC when casting in melee combat.
5. Any successful attack or non-saved-against attack disrupts the spell. This is a cause of much debate amongst (shall I say it?) munchkins. For years there were players that would argue, "if I poke him with a pointy stick, even though it doesn't really damage him, that's an attack so it should ruin his spell!" To that I give a resounding NO. To qualify as a successful attack, it has to do real damage or cause some other quantifiable negative effect (such as an enemy caster throwing off a sleep spell before the PC caster finishes his spell).
Now, here's an interesting tidbit. Unless I am mistaken, activating a magic item such as a wand in 3.x provoked an AoO (and I could be mistaken; it's been a good while since I played 3.x). In AD&D, magic items are single attacks that cannot be disrupted. It does clarify that rods, wands, staffs, etc., attack with missile discharge.
Effects of Cover on Spells and Spell-like powers: Cover provides its normal AC bonus to saving throws against area effect or damaging spells. Further, if 90% cover is available, a successful "half-damage" save means no damage will be suffered. Good stuff.
Monster Charm Power: This alters the "charm person or monster" spell-like ability of monsters such as nixies and vampires. Unlike the spell of the same name, monster charm powers are more like domination; not only is the subject favorably inclined towards the monster, but he or she is, until the power wears off, subject to commands from his or her "master." Telepathic communication is possible within 60' or line of sight (whichever is greater), and the subject will obey any command that isn't physically harmful, including giving away worldly possessions, betraying friends, etc, but not committing suicide or attacking an obviously superior foe (a vampire couldn't make a third level thrall attack a Type VI demon, for example).
We're all pretty well familiar with how this works. Even in 3.x the cleric made his check and referenced a chart to see the success of the action. This is one of the very few elements of D&D that didn't change much up through third edition (I'm not going to compare 4th, as it's the first version of the game, IMO, where the rules disconnect is so great as to make it a wholly different game from prior editions--and I'd ask we not discuss the good or bad of that here. There's a preponderance of 4e discussion threads all over the D&D boards for that).
Clerics, rather than having a set number of turn attempts per day as in later editions, can attempt to turn any given type of undead once per encounter. In effect, you get one shot at it, and if it fails you can't try again next round...though if there's a second cleric in your group, she can give it a shot.
That being said, it is specific that it's a type of undead. So long as you're successful in your attempt, you can turn a second type of undead in the same encounter. That means if you turn a group of zombies (and are successful), you can then try and turn the skeletons with them next round. If that's successful, you can move on to the mummy that's with them, etc. The book clarifies that the DM may opt to disallow turning for the most powerful among the undead. So if the aforementioned group of zombies, skeletons and a mummy is led by a vampire, the DM could grant "turn immunity" to the vampire for this encounter, though he should inform the cleric's player of that.
Evil Clerics: Just as in later editions, evil clerics can compel undead to obey them, rather than turning the creatures. If the cleric gains a "Turn" result on his attempt against such creatures, they will remain neutral or serve the cleric for up to 24 hours (24 hours minus the minimum number needed to turn the creature).
It's interesting that the book discusses creatures being neutral or serving, but gives no advice on how to adjudicate neutrality vs. servitude. The implication is that very powerful undead like liches or demons and devils (yes, in first edition, the turning ability works against creatures from the "lower" planes as well as on corporeal undead) will be neutral, while those of ghost or vampire status or less will serve, but it's not 100% clear.
Hostile action against undead cancels the cleric's control over them.
Counter-affecting: enemy clerics can try to cancel a cleric's turn or control ability by making his own turn check. Straightforward, but a bit easy as it doesn't take the difference in power between the two clerics into account. I'd probably rule that to do this the counter-effect has to take place via turning an undead with hit dice equivalent to the enemy cleric (i.e. you're not re-turning the undead in question; you're attempting to turn the power of the other cleric).
Evil Areas: Some areas can be designated "evil," inflicting a penalty to attempts to turn undead in that area at the DM's discretion. This is similar to Turn Resistance in later editions of the game, but associated with an area rather than a creature.