I've heard from a couple of people who don't believe that the economics mentioned in Part One work for textbook prices.
I look at the biology textbooks I've used -- which have been massive, sturdy hardbacks with lots and lots of illustrations and photographs (pro photographers expect to get paid) and color ink and slick paper -- and I see pretty high production/printing cost right off the bat.
Specialist nonfiction of any kind pays much better than fiction, and publishers have to pay more to interest a professor in producing a textbook that will take a lot of his or her energy and time away from teaching and research (in some cases, choosing to write a textbook may actually harm a prof's career because a textbook doesn't "count" the same as other scholarly publications that may take much less time to write). The publisher might, for instance, have to recoup an advance of $40,000 or so across 5,000 copies, and I don't think it's greedy to expect $40K for authoring a book that takes a lot of expertise and several years to write. And finally, distribution will still be expensive no matter what kind of book you're producing.
According to the National Association of College Stores (NACS), Collegiate Retailing Industry, Higher Education Retailing Market, the breakdown of each $1 of an average new textbook's price goes like this:
Paper and printing: 32.1 cents
Distribution: 22.9 cents
Marketing: 15.3 cents
Author's income: 11.5 cents
Shipping: 1.3 cents
Publisher's operations: 9.9 cents
Publisher's income: 7.0 cents
So in the case of textbooks, printing costs more than distribution, and marketing and the author get the other big hunks of the cheese. The publisher ends up making about $5 profit from each copy of a $70 textbook, which costs about $22.50 to print. Percentage-wise that's not hugely different from what you get with a mass-market paperback.