Evans and semantic/spreaker's reference

Hi
I'm still working with Evans' 'The Causal Theory of Names'. It's rather difficult text. That's why I have another question.
Donnellan [and Kripke as well] distinguishes between spreaker's reference and semantic reference.
Evans argued that names can change theirs references [Madagascar case]. But the way he explain the problem suggests that this is spearker's reference change.
Is that way of thinking right? Yes? No? and why.
If it's necessary I can provide more details of this issue [it would be nice exercise for me to put the toughts in a right order].
Best regards