Random Roll: DMG, p. 125

 At the start of the "explanations and descriptions of magic items" on page 125 of the AD&D Dungeon Masters Guide, there is a section on potions.

Potions are typically found in ceramic, crystal, glass, or metal flasks in enough quantity to provide one person with one complete dose so as to be able to achieve the effects which are given hereafter for each type of potion. Potion containers can be other than described at your option.

That's straightforward enough, though I recall many referees in my youth allowing a "half-dose" of a potion, with commensurately lesser effects. I don't know how widespread such a house rule was outside of the circles in which I moved, however. Likewise, I remember one referee who had replaced potions with what he called "lozenges," little pills like cough drops that had to be eaten to achieve the appropriate effect. I never understood why he'd made this change, but but it stuck with me decades later.

As a general rule they should bear no identifying marks, so that the players must sample from each container in order to determine the nature of the liquid. However, even a small taste should suffice to identify a potion in some way – even if just a slight urge.

Once again, we see the conflation of "player" and "player character" that is quite commonplace in the early years of the hobby (Empire of the Petal Throne does this often, for example). While I understand why, from a game perspective, potions are not labeled, it's one of those things that doesn't make much sense to me from a setting perspective. Unless the manufacturers of potions have a better memory than I do, it seems like it would be all too easy to forget whether this phial contains a potion of extra-healing or a potion of gaseous form. Drawing once again on my early experiences, I knew of a referee whose potions could be identified by color and taste. Over time, characters learned which were associated with which potions and that made it easier to determine which was which. Of course, he also had greatly expanded the roster of available potions, so it wasn't quite as easy as it might sounds.

As Dungeon Master, you should add a few different sorts of potions, both helpful and harmful, of such nature as to cause difficulties in identification. In addition, the same type of potion, when derived form different sources, might smell, taste, and look differently.

 I don't mean to keep harping on this, because I know opinions differ strongly about the matter. Nevertheless, I am not in favor of this approach as a general rule. I think it's important that a setting, even in its magical aspects, have some degree of intelligibility. Players and their characters should be rewarded for learning over time and it seems to me that shifting the correspondence between color, taste, scent, etc. and effect is intended to undermine the value of that. Clearly, Gygax felt differently.

Unless otherwise stated, the effects of a potion will last for 4 complete turns plus 1–4 additional turns (d4). If half a potion is quaffed, the effect will last one-half as long in some cases. Potions take effect 2–5 segments after they are imbibed.

I find it intriguing that Gygax's rules for "half-doses" are more lenient than those of the referees of old I encountered, who halved not just the duration but the overall efficacy.

While potions can be compounded by magic-user/alchemist teams at a relatively low cost, they must have an actual potion to obtain the formula for each type. Furthermore, the ingredient are always rare and/or hard to come by. 

This is where most fantasy RPGs falter a bit. If magic items are capable of being manufactured, why then aren't they more readily available? The simplest answer is the one Gygax offers: they're hard and expensive to make. That's a fair answer, but, if that's the case, then why do they seem so easy to come by in dungeons and similar places? The simplest answer to that is to limit the presence of magic items across the board, making them quite rare. Again, that's a good answer, but how often was it ever observed, even by Gygax in his published adventure modules? I actually don't know the answer to that question, honestly. Perhaps someday I should do an analysis of the prevalence of magic items in published modules (assuming someone else hasn't already done it).

Not from the DMG but a favorite of mine nonetheless