Watermarking adds a unhearable signal to each sample e.g.: Add in frequency ranges which can not be heard some data.
If this data is different for any sold copy it is quite easy to find the person who put the samples to some file share systems and eact accordingly.
Also if someone publishes music with these stolen samples you can still figure it out and (assuming you get the identity) can sue these people.
Similar technologies are also used for movies to figure out in which cinema the movie has been stolen
In any case you need some pretty strong watermark which can not be deleted or gets removed after some format changing, so just the frequency thing is way to weak.
Details: http://en.wikipedia.org/wiki/Digital_watermarking
At the end it is kind of philosophy, if you invest time in protecting your things, or invest time in sueing guys who distribute the samples and and guys who use it. But it might also be impact on location (Most probably US courts bring more money then Austrian courts)