|MadSci Network: Evolution|
Your question addresses several issues about how evolution works and how we know about it from the fossil record.
First, let's clearly distinguish between coelacanth fossils, which go back a long way in time, and modern coelacanths - the living fish.
The fossil record of the coelacanth is particularly interesting. The fossils are found in geological layers laid down from the Devonian to the Cretaceous periods (roughly from 400 million to 65 million years ago), but there are no known fossils more recent than that. There were many variations of coelacanth in the fossil record. The most recent fossil coelacanths are remarkably similar to the living ones. I will return to this shortly.
The gap in the fossil record led many researchers to conclude that the coelacanths were all extinct from the end of the Cretaceous period onward. Thus it was a major surprise when a living coelacanth was discovered in the late 1930s.
Nowadays, the gap in the fossil record has a different explanation, which is probably better. The conditions for forming fossils are uncommon, and it is extremely unlikely that any particular dead animal will end up in the right place to become a fossil. The ancient (fossilized) coelacanths probably lived in shallow waters, though there may also have been deep water branches of the coelacanth family. Fossils formed in shallow water are much easier to discover, and it would also be rare for deep-water fossils to be raised up to elevations where they could be subsequently discovered. Therefore the fossil record may only show us part of the range of coelacanth habitat. This would, of course, be true for many other animals as well. Modern coelacanths are all deep-water fish.
As an aside, it is perfectly reasonable that scientific explanations can change (improve, we hope) with time. All explanations are provisional, meaning that we understand that better information may come along. When that happens, we often see that our explanations are too simple or are simply plain wrong!
The main part of your question deals with the similarity between (some of) the coelacanths from the fossil record and the living animals. This should not be a surprise when you understand more about how species are formed, a process called speciation.
There are a number of ways that speciation can take place, and there is considerable debate and discussion about which ways are most important under different circumstances. Instead of going into the details, let me discuss human evolution, which is rather different than that of the coelacanth.
The last common ancestor between humans and chimpanzees lived only about 6-8 million years ago. There is no fossil record of such an animal, so we have no idea what it looked like, but people speculate that it may have resembled current chimps more than current humans. Our line, therefore, has changed fairly rapidly in only the last few million years.
You should not think, though, that a single group of animals started out as (something like) chimps and ended up turning smoothly into humans. Speciation happens when a part of a group - probably a small part - branches off and lives in a new habitat. The new conditions require different sorts of things for success. In humans, a group branched off and lived under conditions where walking upright was an advantage. Later, there was another branching and the group that split lived under conditions where creatures with bigger brains had an advantage. The unique thing about human evolution is that all of the creatures that were part of the other branches are now dead, so it might appear like we all turned into humans. Instead, just a tiny group split off (and split and split ...) and experienced different conditions over time.
If the other branches were still around, we would be able to see how the splitting happened in more detail. Some of these branches might have stayed living in relatively constant conditions where the same sort of animal was successful for many millions of years. In that case, one would see (if the fossil record were complete) that in some branhes the ancient and modern animals were quite similar. In other branches, there would be more change with time. We're just one of the many branches of human evolution through history.
Back to the coelacanths. The similarity between the old and recent coelacanths merely shows that one branch of the family kept living in the deep ocean all this time. The coelacanths were successful then (they could eat and reproduce and avoid being eaten into extinction) and the same conditions exist there today. Thus the living and fossil coelacanths would therefore resemble each other very closely.
Try the links in the MadSci Library for more information on Evolution.