There was always tremendous confusion over the prevalence of magic items in 1st edition. If you read the DMG, magic items were clearly supposed to be rare and wonderful. So "a single book of spells," if I remember correctly, was in one place described as "a benison beyond belief."
However, if you actually used the NPC tables in the DMG, the random treasure tables in the Monster Manual, or ran the "Official" TSR modules as written, you came away with a drastically different sense of how common magic items were in the game.
David Godwin wrote a very insightful article in Dragon 99 called "History of a Game that Failed" that very clearly describes this very phenomenon. The DMG (and countless anti-Monte Haul Dragon and Polyhedron articles) said one thing, but all official examples of actual play (published modules) showed a new DM something completely different.
Godwin's conclusion was that the TSR modules were vastly overstocked with magic -- probably because the designers assumed that (1) no PC party would ever find ALL the magic items, certainly not that Staff of the Magi hidden in the barrel of pickles in the kobold lair, and (2) DMs would freely alter the treasure lists on the fly.
Clearly the designers were wrong on both counts: one should never underestimate the ingenuity, determination, or rapacity of PCs looking for magic items, and most young DMs naively believed that the modules as written were the gold standard -- and besides, with no hard guidelines as to what level of magic was appropriate, what yardstick would have been used for those DMs willing to depart from the module as written?
Certainly in my own first campaign I ran nothing but Official TSR modules (tm) and yet still ended up with a group of PCs loaded down with magic goodies.
3e just took the reality that was already on the ground and codified it into reasonable guidelines. I think that in the process they actually made magic items more sensibly distributed -- no more spheres of annihilation hidden in the jam closets of 0-level peasants.