The quality of science fiction sequels will tend to vary inversely as the square of the distance from the original. (Somebody who knows math or something might want to straighten me out on how to express that.)
In other words, Larry Niven's Ringworld (1970) is a landmark, The Ringworld Engineers (1980) is tolerable, and The Ringworld Throne (1997) is a travesty, a mortal insult to anybody with a brain. The same thing happened much more rapidly with The Mote in God's Eye and The Gripping Hand.
This principle is also illustrated by Orson Scott Card's Ender's Game series, Philip Jose Farmer's Riverworld series, and Frank Herbert's Dune series. These things have become franchises. Hell, Larry Niven has rented out his Known Space universe to other writers, like Lassie or something. Ugh. Then there's Robert Silverberg's masterpiece Lord Valentine's Castle: Majipoor Chronicles worked because he explored unrelated parts of the same world, but from there on in it's been downhill.
If that's what it takes to put one's kids through college, I guess I can't object too much, but it just seems shameful and depressing when a good writer mutates into a hack.
This seems to happen when writers return to worlds and sets of characters that they just don't care about any more. Have you read Xenocide by Orson Scott Card? Card just didn't give a damn about Ender any more by the time he wrote that. He kept on writing about the guy not because he had to, but because he wanted to. The parts of Xenocide that happen on the OCD planet are vintage Card, because that was a new place with new characters and he was still exploring it all. He was excited about it.
Counterexamples include any series by Jack Vance, or anything at all by Jack Vance. Jack Vance is your lord and savior whether you know it or not. Does this apply to other genres? I'm not sure. Raymond Chandler never had this problem, and Ross MacDonald's quality seems to vary stochastically. Go figure.
Uberfetus: Clarke was one of those damned idea writers. I hate those. I mean, yeah, hey, fine, ideas are great and all, but must it all be medicine? Jesus, ideas are purely secondary in fiction. If it's not first and foremost an engaging story, it's not much. Heinlein got away with it because a) half his ideas were about sex, and b) in between ideas he'd blow stuff up.
Crux: William Gibson has earned the Wharfinger Seal of Approval. He's one of the few writers I'll buy in hardcover the minute anything appears. I never got through Mona Lisa Overdrive, but the last (current?) series (Virtual Light, Idoru, and All Tomorrow's Parties) has been consistently good throughout. He often seems to write the same damn book over and over, and most of his stuff is totally plotless, but he's intense and fun to read and that's good enough for me. The Law doesn't necessarily apply in all cases; Fritz Leiber and Edgar Rice Burroughs wrote great sequels too.