I'm afraid this depends entirely on what you're look at! Of course, there will be an average (mean, median, etc.), but it will be meaningless because it is measuring different things in different groups.

Of course, a genus is an entirely artificial concept - a genus of a well-studied group of birds is nothing like the same as a genus of nematodes. Recognising genera is the next problem: the features used to define them may not be recognisable in the fossil record. So, for example, we get Lingula going back to the Ordovician, although the soft tissue anatomy is likely to have changed dramatically.

Then there are apparently very long-lived genera such as the pterobranch Rhabdopleura, which appears in near-identical form (including soft tissues, from what we can see) in the Cambrian. Is that because it was morphologically very stable, or because we don't properly understand it?

I know that's not very helpful, though, so here's my perspective on fossil sponges. In most cases, fossil genera appear to last for a few tens of millions, up to around 100 million years. There are, however, many genera recorded from only a single site, so we can't say how long they lasted, but this broad range also covers most of the molecular clock predictions for generic divergence dates. However, there are also some remarkably long-ranging fossil sponges like Nucha (300 million years, apparently)... and we don't yet have a clue what was going on in the deep sea, where the very long-lived sponges tend to thrive (but watch this space...).

As a final thought, I have seen a tendency of some palaeontologists to name new genera for fossils that are out the 'expected' time range--despite morphological similarity. This, of course, will result in shorter ranges in the fossil record than the the true ages based on divergence of living species within extant genera. So many problems!

I'm afraid this depends entirely on what you're look at! Of course, there will be an average (mean, median, etc.), but it will be meaningless because it is measuring different things in different groups.

Of course, a genus is an entirely artificial concept - a genus of a well-studied group of birds is nothing like the same as a genus of nematodes. Recognising genera is the next problem: the features used to define them may not be recognisable in the fossil record. So, for example, we get Lingula going back to the Ordovician, although the soft tissue anatomy is likely to have changed dramatically.

Then there are apparently very long-lived genera such as the pterobranch Rhabdopleura, which appears in near-identical form (including soft tissues, from what we can see) in the Cambrian. Is that because it was morphologically very stable, or because we don't properly understand it?

I know that's not very helpful, though, so here's my perspective on fossil sponges. In most cases, fossil genera appear to last for a few tens of millions, up to around 100 million years. There are, however, many genera recorded from only a single site, so we can't say how long they lasted, but this broad range also covers most of the molecular clock predictions for generic divergence dates. However, there are also some remarkably long-ranging fossil sponges like Nucha (300 million years, apparently)... and we don't yet have a clue what was going on in the deep sea, where the very long-lived sponges tend to thrive (but watch this space...).

As a final thought, I have seen a tendency of some palaeontologists to name new genera for fossils that are out the 'expected' time range--despite morphological similarity. This, of course, will result in shorter ranges in the fossil record than the the true ages based on divergence of living species within extant genera. So many problems!

Hi Jan, the two papers below might give you a good introduction to the topic.

Cheers, Steffen

Sepkoski, J.J. (1998) Rates of speciation in the fossil record. Philosophical Transactions of the Royal Society of London B, 353, 315-326.

Foote, M. (2007) Symmetric waxing and waning of marine invertebrate genera. Paleobiology, 33, 517–529.

As Joseph Botting said, you have to be VERY cautious to look at comperable datasets. You cannot directly compare fossil with modern data--they use different definitions of "species", which makes them incomperable.

Within the fossil record, I would advise you to start with index fossils such as amonites. These tend to be well-studied and (with amonites nayway) relatively common, so you have a decent chance of getting a real answer. That said, you run into the question of whether the names are real or not--a few years back I recall some informal discusisons about whether amonite genra were real, or an artifact of overzealous naming of new taxa. Still, it's a starting point, with solid data sets. If you can find index fossils within various clades (say, ammonites, bison, forams, fish, etc) it will give you a preliminary assessment.

Rodney Feldman recently uploaded a paper on lobster evolution to this site; it will have bearing on this question. Decapods are fairly common, being hard-shelled nearly ubiquitous creatures. Again, it may be a good place to start looking.

One final word of caution: Origination and extinction aren't shown by the fossil record. What we have are the first appearance datum (FAD) and last appearance datum (LAD). Even in a well sampled speceis, the actual timespan of the species may be longer than its fossil record shows. There are a variety of statistical methods for dealing with this, but it's still one of the frustraiting aspects of the field.

I know Paleontologia Electronica has been focusing on molecular clock and other calibration techniques recently. It's open-access, so it wouldn't hurt to look and see what they have. They also have some papers on defining amonite genra that could help you out.

"I am aware of the subjectivity of higher taxonomic categories, but there must be some time-span in which the genus is being found in the paleontological record."

Within paleontology, too, different people classify in very different ways. The timespans depend on the quality of the fossil record, and on who classified the fossils. It's all completely inconsistent – there's no definition for "genus" after all.

In many branches of vertebrate paleontology, the genus is basically the unit of biodiversity: if populations – or individuals, if there aren't enough to speculate about populations – can be distinguished, different genera are named for them, and then a species name is appended to honor some colleague.

"Similarly, using molecular clocks calibrated with fossils, we may assume the age of extant genera."

Here, too, these ages are all over the place and depend on who did the classifications!

I'm sorry, your question really doesn't have an answer. Genera are not comparable and not countable.

Well, just claiming that everything is completely inconsistent isn’t really helpful and certainly does not bring science forward. Here is another paper coming out of the ‘Chicago group’ that compared the composition of genera based on morphological vs. molecular characters, for mollusks and mammals. The consistency rate was about 80%, which certainly is better than “completely inconsistent”:

Jablonski, D., Finarelli, J.A., 2009. Congruence of morphologically-defined genera with molecular phylogenies. Proceedings of the National Academy of Sciences of the USA, 106: 8262-8266.

Regarding the original question; chances are indeed good that people working on different phyla have different approaches and traditions of defining genera, which could result in artificially different geologic lifespans of genera. I’m not aware of any study that has systematically addressed this issue, for example by comparing rates of molecular evolution between, say, brachiopods and mollusks (which have a good fossil record due to their shells made of calcium carbonate). If anyone knows about such study, I’d be happy to know about it!

Cheers, Steffen

That's about whether the genera are monophyletic, not about whether they're comparable to each other.

And no, it's not only between different phyla, it's often different people working on the same taxon, or the same person working on the same taxon at different times.

I agree with all other answers, but probably it would be better to have some actual numbers. Prothero (2013) wrote in his textbook "Bringing fossils to life" that Cambrian trilobite zones have average 3 m.y. duration, and Jurassic ammonites for ca. 600 k.y. duration (see his Fig. 10.19). A few million years would be a good answer in general.