However the practice of introducing plants from tanks with very different fish has proved useful. When the guppy was removed, there were annulatus fry and some pairs did grow up there. There were no hybrids.

There are times when it seems like fish are "spontaneously generated," though I hear that the concept has been out of fashion for some centuries. ;)

Was baby sitting about 23 tanks for a friend who took his wife on a much deserved vacation. When they returned home a couple weeks later, there were 27 tanks, bowls and a bucket containing fish. The situation I was most proud of was that waste water bucket. Some peat moss had been dumped there. I saw fry, pulled the bucket off to the side and fed it freshly hatched baby b.s. It turns out that they were late hatching annuals, and one of the more treasured of the annual killies, Nothobranchius rachovii. If you take a look at http://images.google.com/images?q=Nothobranchius+rachovii&hl=en&btnG=Search+Images
you will see why I was pleased to leave them for my friend. :)

That did start a tradition of nurturing mystery fish once in a while. However as a rule I will rotate plants and tanks between quite different species or genus or even families. So a livebearer tank may become home to a rainbowfish and then a small killie and later home to a large predatory killie which looks quite different. Fish are often much less tolerant of the fry of another species, than of their own. That keeps most of those anomalies from even coming to my attention. On the rare occasions when something does survive the rotation, they are pretty easily identified.

Different guppy or gardneri or Chromaphyosemion strains are raised on different sides of the room. That way "adventurers" have little chance of messing up another strain, despite their travels.