Joined: Jan. 2006
Generally, non-coding DNA is not so stable in amount or sequence, AFAIK. I remember the case of the fugu fish (I think) that seem to have lost a large amount of its non-coding DNA. Also, the amount of non-coding DNA in prokaryotes is rather low.
And non-coding sequences have a high mutation rate.
Also, some junk DNA is the result of genome duplications (common in plants) followed by gene silencing. The variation in genome size among angiosperms is pretty huge.
“Parasitic DNA” cannot expand indefinitely since it would certainly reduce the fitness of its owner at some point. Then you might ask “so, even a low level of transposable elements in a genome can be counter adaptive?”. It could be, but sometimes transposable elements can be adaptive; some of them carry resistance genes. Barbara Mclintock (Nobel price in the 80’s) demonstrated that transposable elements in maize are “activated” by abiotic stress. (I’m not an expert on this subject.)
Also, remember that transposable elements derive from viruses (and vice versa), so maybe they can move from “host” to host and replicate inside genomes even if their replications are counter adaptive for their hosts.
Actually, transposable elements cannot exactly be qualified as non-coding DNA, since they can be “activated”, but I think some of them can be completely deactivated at some point.