That question sort of bothered me - I have no idea how big a tree is, and even if I saw one, I have no reference for how many feet it is without doing some fairly exhaustive mathematics (and at the scale of "largest redwood", I'd likely be wrong). Given some information about redwood trees, of course people are going to use that information in the subsequent guess. They're not going to imagine a redwood tree, then imagine a building next to it, then count the floors and estimate the height. Or estimate the girth, then guess a height/girth ratio that makes sense given the composition of a tree, and then estimate the height. They're lazy.

If I asked you, "Will a frooble fit in my pocket/empire states building?", and then asked you to estimate the average size of a frooble, you'd certainly take into account my earlier question.

See http://lesswrong.com/lw/k3/priming_and_contamination/ for some better examples. IMO, the more insidious form of anchoring is contamination (vs sliding adjustment).