A single XML document may contain multiple out-of-line extended links. However, the current XLink specification is relatively silent on exactly what the format of such a compound document should look like. About all it says is that such a document must be a well-formed XML document. An XLink processor would presumably read the entire document an extract any extended links that indicate connections to or from the current document.
A browser or other application that's reading the individual pages needs to be informed that there is a separate linkbase elsewhere that it should read and parse so that it can show the links to the user.
Ideally it would be handled through some external mechanism like HTTP headers.
The only currently defined way to do this
is to add an extended link inside the documents the out-of-line
link connects. This extra extended link has the xlink:role
xlink:external-linkset and one or more locator elements that
point to linkbases.
<LINKBASE xlink:type="xlink:extended"
xmlns:xlink="http://www.w3.org/1999/xlink"
role="xlink:external-linkset">
<xlink:locator href="courselinks.xml"/>
</LINKBASE>