That's an excellent question, and you're thinking along the right lines to ask it. To use your example, how would some ancestral bat species with smaller wings "know" that it had to evolve a little more to become an evolutionarily successful bat species with longer wings? And how would the species survive if its small wings were less than evolutionarily successful? Get ready, because my answer is long.

There are two ways to think about evolutionary change. You can picture it as a gradual process, where a species undergoes small changes very slowly, taking enormously long periods of time to evolve into a new species. This is sort of the classical view of evolution. (Remember that individual animals don't evolve, populations and species do, based on which individuals reproduce better and what traits they pass on to future generations.) If this gradual change was what happened with your bats, it would have to be because the intermediate forms (those with shorter wings), even though they were less successful than current bats with longer wings, were still more successful than previous forms (probably some mouse-like ancestor without wings). So, longer wings lead to better reproductive success, and EVEN longer wings lead to EVEN better reproductive success.

If you follow your question backward through evolutionary time, you would next have to ask why it was advantageous for wings to evolve in the first place? In all cases, according to this gradualist view of evolution, no trait will persist through evolutionary time unless it is in some way advantageous to an animal's reproductive success. Animals don't go through evolutionarily disadvantageous forms in order to get to the advantageous forms.

There's another way to look at evolution, though. You could picture evolution doing very little for long periods of time, and then bam!--a sudden, huge change that results in a new species in a very short time. This second viewpoint is often called "punctuated equilibrium," and more and more scientists think that it is a better description of the way things work.

According to the punctuated equilibrium viewpoint, tiny evolutionary changes actually happen very quickly. To switch examples, imagine a species of fish that attracts mates with its bright colors, and you suddenly make their pond murky so that it's hard for them to see each other. In a surprisingly small number of generations, the fish population will have adapted to the new conditions, thus evolving. (Remember: this is not a conscious "decision" on their part, it's just a matter of which individual fish happen to reproduce better in the new environment.)

Different traits in the population that were sort of neutral in terms of reproductive success may suddenly become advantageous. These traits would suddenly become more common in the population--that's evolutionary change. The fish population may become even brighter, or the fish may start attracting each other by smell instead of color. Either way, it will happen very quickly, and it will depend entirely on the random collection of traits in the population and which traits lead to increased reproductive success under different environmental conditions. The speed of this change is the "punctuated" part of punctuated equilibrium.

The "equilibrium" part refers to the fact that, before you showed up and muddied their water, the fishes' color was in perfect balance with the clarity of the water. Maybe if it rained a lot for a year or two and runoff made the water just a little murkier, brighter fish would breed better and the fish population would get just a little brighter overall. In drier years, then, maybe duller fish would breed better and the fish population would get a little duller again. This is an equilibrium with their environment.

Most of us would probably never notice such tiny evolutionary changes happening right in front of us because, despite its ups and downs, the water has an average level of murkiness, so the fish population has an average level of brightness, and we think nothing is happening. Then, when something big suddenly happens (maybe an earthquake that makes the pond permanently muddier) we notice a big, sudden, evolutionary change in the fish population: they now attract each other with smell instead of color. This is an example of punctuated equilibrium without any intermediates between the old form and the new form.

So, I would say there are two possible answers to your question. Either intermediate forms are themselves advantageous (even if less so then current forms), or intermediate forms are rare, because evolutionary changes are big and fast instead of slow and gradual. The fossil record is pretty sketchy, by the way--I think sometimes people see fossils of mice with big webbed feet and later fossils of bats, and they assume that there must have been short-winged bats in between. Maybe there weren't, though.

Whew! That was a long answer. Thanks for the question.