Wild bananas (Musa spp.) are, not to put too fine a point on it, nearly inedible. Their seeds are enormous: a wild banana fruit has only a thin layer of pulp between the peel and the huge hard seed. Obviously, something has changed between the wild type and the virtually-seedless, soft, sweet bananas sold in every American grocery store. In fact, it took decades of selective breeding to shrink the seeds that far. There are dozens of banana varieties now, most of them sweet and pulpy, with seeds that could easily dance on the head of a pin.
Seeds that small, unfortunately, are reproductively useless. They'll almost never sprout. Commercial banana growers have no desire to extract millions of minute seeds from their crop just to grow one or two new banana plants... so they don't. Bananas aren't grown from seed: they're propagated by cloning. To get a new banana plant, the farmer cuts a root sucker from the parent plant's corm and grows it to fruiting size in wet sand.
Because all commercial bananas are grown this way, commercial farms are monocultures -- huge masses of clones. The bananas found in American supermarkets, for instance, are all from the Cavendish variety, and every Cavendish banana is genetically identical to all the others. When it comes to producing uniform bananas and growing them all the same way, that's a good thing. When it comes to disease resistance, though, it's very, very bad. Since there's no genetic variation in these plants, they all have the same immune system: a disease that can kill one Cavendish banana plant can kill any other.
We had an object lesson in the problem once, not too long ago. Before 1960, the Cavendish banana was a nobody: commercial banana-growers all used clones of the Gros Michel, a variety with tastier fruit that was easier to ship. It worked out nicely, right until the outbreak of Panama disease. Caused by a soil fungus, the disease ripped through banana plantations worldwide, rendering the Gros Michel virtually extinct. Worse, since the fungus persists in soil, the Gros Michel plantations couldn't just be replanted: the new plants died before they could ever bear fruit. If you've ever heard the song "Yes, We Have No Bananas", you have some idea of the effect Panama disease had on the world banana trade.
The Cavendish banana's fruit is inferior to the Gros Michel in practically every way, but it has one thing its predecessor never will: it's resistant to Panama disease. With Gros Michel plants dropping like flies, the Cavendish came to the rescue. It may not be the perfect fruit, but it can actually be grown commercially even in a Panama-disease-haunted world. It has replaced the Gros Michel now; it's grown worldwide, with millions of its clones producing millions of pounds of bananas each year.
What's wrong with this picture, though? The Cavendish is still a clone. If another disease develops that can take down one Cavendish, it can kill them all. For as long as our bananas are all cloned, we're inches from a repeat of the Gros Michel disaster. Monoculture just doesn't work in the long run.