bead
A bead is one piece of page belonging to an article, threaded on it in the order it is read.
In more detail
Think of a newspaper. An article starts in a column on the front page and finishes three pages further on, in a column beside something else entirely. Each of those columns is a bead. Strung together in the order a person reads them, they are the article.
A bead is a rectangle on one page and nothing more: a corner, a width, a height. What makes it a bead rather than a rectangle is its place in the string — the one before it and the one after it.
What it does on a real document
Without the string, nothing on the sheet says which column carries on which: the person reading has to look, and on a page laid out in four columns they get it wrong. With it, the program jumps from the foot of one piece to the head of the next, on the right page, with nothing to hunt for.
The word belongs to the format and stays there. What you write is an article, a title, and the patches of page it runs over in order; what the file ends up holding is the string and its beads. Nobody has to say the word to use the thing.
Where it comes from, and why
The image is older than the format: a bead on a thread, one after another, in an order somebody chose. The people who wrote the specification borrowed it in 1993 and it has not been renamed since — which is why the word turns up in documentation, and only there.