In Tiger, unlike Panther, one must explicitly mount drives to make them available at the point in /etc/rc where virtual memory is set up. The catch-22 here is that one can only use BSD device names of the form /dev/disk1s10 at this early stage of the boot process, but BSD device names are not consistent across boots.
An idea from Markus Hitter is to use pdisk to inspect properties of unmounted drives by BSD device name. I refined this idea by using md5 to fingerprint the drive:
swaphash=`pdisk /dev/disk${swapdisk} -dump 2>/dev/null | \
grep -v '/dev/disk' | md5 -q`
Still, two drives might be identical twins intended for different purposes, so one can check for a hidden file .enablevm as confirmation. This issue can be circumvented by choosing a unique volume size to be used only for swap partitions. I have updated my web page,
Moving the swapfile in Mac OS X 10.4 (Tiger), to explain all of this. I also provide a Perl script which automates this system modification. This script has been 100% reliable for me.
Here are a bunch of resources on the topic of moving swap:
Mac OS X Hints
http://hints.macworld.com/article.php?story=20060206013801960