( Robinson R (2006) Ciliate Genome Sequence Reveals Unique Features of a Model Eukaryote. PLoS Biol 4(9): e304 doi:10.1371/journal.pbio.0040304 )
Tetrahymena thermophila (fluorescently tagged, of course).
It's an aquatic cilliate -- a microscopic organism covered in little hairs (cilia) it uses for swimming.
Those creatures have two types of nuclei -- the micronucleus which is the germline, and the macronucleus which is used during the organism's everyday life (ie. somatically expressed). Upon conjugation, the old macronucleus is destroyed and the micronucleus is doubled, with one of them being modified into a macronucleus. The macronucleus genome is then duplicated multiple times -- up to 45 copies -- and the non-coding DNA is spliced off. (http://www.ciliate.org/genetics.shtml)
What a brilliant strategy: keep a copy of your genome safe and unmolested by transcription enzymes (in order to actually use it to make proteins), and have another copy amplified and optimised for everyday use.
Image you had your favourite film on a VHS tape (those ancient casette things, in case a reminder is needed!). Every time you play the tape, you damage it little-by-little, eventually ending up with loads of scratches and poor sound quality. Now you need to copy the film for a friend. The scratches will still be there when you copy it, since that particular data is permanently lost. You friend watches it few times, and the next copy is even more damaged. And so on.
So that the data is not completely lost, you can build some mechanism that edits and fixes the tape each time you play it. It notices the minute scratches, and patches those holes in the data while it can still guess what must be there. This mechanism would be fairly expensive, difficult to set up, and prone to malfunctions. This is what we animals do. Also, we have millions of cells, most of them not in the germline (ie. will not be passed on to the next generation). Mutations in those cells are usually benign and don't matter in the long run (even cancer itself is not transmitted to the next generation; although susceptability to it may be). We try to protect our germline DNA to the best of our ability, albeit in a very unnecessarily complicated, inefficient way.
The poor little Tetrahymena doesn't have a place for germline cells to be stored. Everything must be enclosed in one cell. And it's a bit too small to experiment with low efficiency complex strategies. Every joule of energy counts in its life.
When you first get your film, you could also make a copy of it right there, lock it away in a cupboard somewhere and use the other at your discretion. Hell, you can even keep it out of the box if you'd like -- if you damage it too much, you can always make a copy of the one in the cupboard. If a friend comes to you and asks for a copy of the film, you just make them a copy of the cupboard one, so the damages you've inflicted on your viewing copy do not propogate any further. Your friend also makes a viewing copy and a storage copy.
That's exactly what the Tetrahymena does. It leaves a copy of its genome for storage, locking it away in the micronucleus. The other copy is optimised for reading and is later destroyed upon conjugation.
In a way, that's kind of what we do, albeit in a rather complicated manner. Once the zygote is conceived, the male's somatic cells can all but disappear. We don't care if he dies. Oh, and after gestation and childrearing, the female can die too. Your gonads are micronuclei; the rest of your body consists of macronuclei.
Yet another example of convergent evolution between multicellular and subcellular structures?
Tetrahymena has more cool features, but I'll leave them for later. Since Tetrahymena is a convenient model organism, I'll probably return to it quite a few times.
RETRACTION(18.08.2009): The tangent about the 'adaptive advantages' of nuclear dimorphism is fundamentally flawed and therefore retracted. The reasoning is backwards, and ciliate nuclear dimorphism is in fact a testament to the generation of complexity by neutral forces. The micronucleus is almost completely packed with transposons and other toxic DNA that must be excised for proper nuclear function. Most other life simply takes better 'care' of their genomes...