We can get a little more clever about this. In this slide we use the same example, but the annotation is split into only two runs: the first above the family name, 林, and the second above the given name 和代. In other words, we have associated one run of annotations with multiple runs of base characters.

This is achieved by applying a method familiar to those who have created tables in HTML. The second and third rt elements become one, with the addition of an rbspan attribute whose value is, in this case, 2.

The markup is now as follows: