I'm not sure if this more properly belongs in "Geology and the Great Flood", but here goes.
Conodonts are small toothlike structures that are known from the Middle Cambrian to the Late Triassic; Though they are sometimes found in great abundance, enough to be very useful for stratigraphic correlation, the nature of their owners has been a subject of controversy for over a century:
Annelids Arthropods Chaetognaths (some obscure marine worms) Fish Plants
Conodonts were a mystery for a long time, because:
They were not an internal skeleton, like what vertebrates have.
They were not an external skeleton, like what arthropods and echinoderms have.
They did not resemble some distinctive part of some present-day organism, like shark teeth or mollusk or brachiopod shells; consider how Nicholas Steno had concluded in the 1600's that some "tongue stones" were really fossil shark teeth.
But in 1983, the first conodont-animal fossil was discovered; its only mineralized parts were its conodonts, ensuring that we see only those parts most of the time. It was followed by some others; they are most likely primitive vertebrates, like present-day lampreys and hagfish.
Fossils of soft parts are very rare; they form only under various exceptional circumstances. This explains why some animal phyla have little or no fossil representation -- they are mostly of soft-bodied animals.
And it's been seriously suggested that the Cambrian explosion is a simultaneous invention of mineralization provoked by some ecological change, like:
Some climatic or geochemical change The emergence of some early predators
Creationists often have unreasonable expectations of what the fossil record ought to contain; they sometimes fail to realize that soft parts are harder to find than hard parts.
Checking on UCMP's pages, I find that the oldest unambiguous diatom fossils are from the early Cretaceous; older diatoms may have been destroyed by recrystallization. I think that this question may eventually be settled by molecular techniques; the amount of divergence between various diatoms' genes may indicate when they started diversifying.
Also, teleost fish started appearing in the late Triassic, and conodonts are found up to the end of the Triassic, so the first teleosts and the last conodont animals had coexisted for ~20 million years. However, conodont animals had existed for ~300 million years before that, and teleosts would exist for ~200 million years afterward, becoming about half of present-day vertebrate species.
Which does not make Flood Geology any less preposterous.
Turning to another noted aquatic vertebrate, the creationists have enjoyed gloating over the discovery of the coelacanth Latimeria chalumnae. However, that fish lives in deep waters, where the sediments are not likely to be pushed above sea level. But such taphonomic details are lost on many creationists.