Marcie is talking about multi-author books, with a particularly extreme case. I was never really a fan of large multi-author books, even though I've participated in a few. 4 to 6 authors is manageable if you can work well together. Wrox Professional ASP.NET is a good example, 5 authors, but we spent a day all locked in a room together (one of the labs at Microsoft) working out the spec down to level 5 headings. That way we all knew exactly what the other was writing about. And we'd all written together before so pretty much knew each others style.
This one person per chapter occasionaly happens when authors drop out or can't complete the amount of work they committed to. The book is half written, booked into schedules (ie editors, indexers, the printers, etc), so becomes expensive to withdraw or postpone. Therefore authors are found to take on chapters to make up the difference. This is where there is no flow, and did happen several times to Wrox books. It's no excuse but might go some way to explain things.
Another area where this happens is the uber-book; those 1000 page tomes that cover everything. There are so many topics that you need more people to cover them, not from the time standpoint (though that is a factor), but from the knowledge. People are drafted in to cover just their area: ie "Please write us a chapter about xxx". The problem lies when the author doesn't look at or question the spec, or doesn't read other chapters. Their chapter, that may be excellent, stands-alone. The flow is gone.
Now you might argue that you don't like large books like this, but the reason they are still produced is that they are still bought. Figures prove this. You don't think publishers put out large books in the hope they sell; they have extremely good figures as to what sells, where from and when. Large books sell well. The fact Marcie bought this book proves that. But then again she buys a zillion books a month. I think she's addicted 8)