what is row migration?
On theory,
Say you have a record 'SCOTT' store somewhere ( in a block).
and You have someother records too, like this in three blocks .
( assume every line as block and start and end of block (size) is marked by 3 ***.).

[b]
***a,b,c 	    ***
***Tiger,SCOTT,Frank***
***sName,oname,aname***
[/b]

Now you update SCOTT to somethingelse..say SCOTTNandy.
And You may think the row is placed like this.!
Acutally NO!

[b]
***a,b,c 	         ***
***Tiger,SCOTTNandy,Frank***
***sName,oname,aname     ***
[/b]
It   depends.
If there is enough room for updated record , it will be placed there.
The above update does not fit it there ( see the 2 block is bigg).
SO the row is migrated to another block!. In order maintain things (integrity,index) oracle leaves
a kind of forwarding address in the old place that refers to new place.
[b]
***a,b,c            ***
***Tiger,go_4,Frank ***
***sName,oname,aname***
***SCOTTNandy       ***

[/b]
So this row is called as an Migrated Row.

what is a chained row?

Now what if you have updated SCOTT to 'SCOTTWilliamAchenbach' or something BIG enough that will
NOT EVEN FIT In the fourth block ( or the updated row is BIGGER than the block size. what if the blocksize is
4k and rowsize 8k?)?
In that case a the row is split between the blocks ( and an internal mechanism CHAINS them).
Somthing like
[b]
***a,b,c            ***
***Tiger,go_4,Frank ***
***sName,oname,aname***
***SCOTTWilliam_go_5***
***from_4_Achenbach *** 
[/b]
So if you look further into this you can see that,
a migrated row can be a chained row.
But a chained row may or may not be a migrated row.

All above discussed to the greatest possible extent in server concepts manual.