However, even those particles and fragment sent “up” - at one side of their orbit - would be forced to a lower orbit at the other side of the new orbit. Likewise, the chunks of satellite debris forced “down” on this side of the orbit (at the collision point) go higher on the other side. So, eventually, all parts hit more air resistance on at least half of the new orbits, and so the new orbits degrade faster, and the fragments are removed faster than the original satellites.
Of course, in the meaantime, the fragments are a greater threat: they form a much larger "cloud" of potential secondary collision targets.
But it seems that - unlike the hype about high-speed collisions on a recent show (Discovery or History Channel) very, very few collisions would have a “relative speed” of extremely high velocities: Only the polar orbiting satellites would have a high speed difference with a satellite more normal equatorial orbit. And, since the three most frequent launch station have different latitudes (US, France, and Soviet Union), the three resulting “normal” orbit patterns intersect each other even less often.
So it would appear that most collisions are going to happen with a speed difference of "tens of miles per hour" not "tens of thousands of miles per hour"
That may very well be true, but I saw a photo once, in "NASA Tech Briefs" if I recall correctly, of a small divot out of a Space Shuttle windshield which was determined to have been caused by an orbiting speck of paint.