Every edition of D&D has put its own stamp on the cleric, in the process rendering the class more incoherent than it was to start. Gary Gygax contributed further to this mess in his article "Clerics Live by Other Rules," which appeared in issue #92 (December 1984) of Dragon. To be fair, Gary's article is actually pretty good, but it laid the seeds for much mischief later. His intention was to suggest that individual referees, for the purposes of fleshing out their campaign settings, could change the rules under which clerics (and, by extension, druids) operate, either restricting their opportunities or expanding them (or, preferably, both).
In the article, Gygax gives an example of a sect worshiping the woodland deity Ehlonna, from his Greyhawk campaign setting. Owing to tragic events in the past, this sect operates differently than others of its kind, having a unique selection of spells, armor, and weapons, in addition to having certain ritual taboos placed upon them. Thus, for example, they're not allowed to use fire-based spells of any kind, but clerics, after proving themselves -- gaining levels -- can wield broadswords and druids can wear elfin chain.
Normally, I loathe this kind of stuff, in large part because I think it contributes further to the dilution of what the cleric class is -- and it's already pretty diluted as it is. What makes Gygax's approach work, though, is that a) it's solidly grounded in the setting and b) he's limiting these changes to a particular sect, not establishing it as a baseline. That's how I think things like this ought to be done. Unfortunately, players (and later designers) didn't care for these nuances, instead using them as a template for "fixing" the cleric class. This led, in my opinion, to a variety of changes over the years that have rendered the cleric one of the least coherent classes, both mechanically and as an archetype. But I'm pretty sure I'm in the minority in feeling this way.
Interestingly, Gygax himself warns against taking what he wrote in this article as an official, universal change. In his concluding paragraph, he says some very sensible things:
Now when you hear someone, DM or player, mentioning something about "unknown" cleric spells or similar difficulties, don't panic. It could well be a cleverly planned campaign where difference and the unexpected are desirable -- and who can fault that?! Perhaps you might wish to try it in your own campaign, too. A cautionary word is necessary, however, for there is a problem with such variations. Unless the full and complete details of the differences are known to other DMs, they might well not wish to have clerics or druids of such nature participating in their games. This is their right, and skepticism on their part is justified. Players of these clerics and druids must be forewarned that such characters might be "one-campaign-only" adventurers who are not welcome elsewhereIt's good advice, but it's also, I think, advice rooted in an older style of play that was already on its way to dying out by the time this article was published. Campaign hopping of the sort Gary envisions was already rare in the early '80s when I was most deeply immersed in the hobby and I have a hard time imagining that it was any more widespread on the cusp of 1985. Ironically, the advent of the Net and online play make well lead to a resurgence of the Old Ways in this regard, in which case Gary's advice might well prove useful again.