What Are The Parameters Which Allow Distinction Between Genera?

[Ammojoe]

I think the title describes the question I am trying to ask. It's quite difficult to phrase, and so I hope I have got it right, and it is coherent. The other day I was thinking about the factors that allow us to distinguish one palaeospecies genus from another. I know usually we rely on morphology, but sometimes I think fossils can overlap morphologically, but perhaps be separated temporally and therefore classified as a separate genus - I think this is somewhat subjective though, and so I was wondering what other factors allow distinction.

For example: In Dorset, UK, Ammonites can be distinguished by morphology. But there are other methods you can use to distinguish them. E.g. You can use stratigraphy. Ammonites are index fossils and therefore they only occur in very specific beds of limestone. There is one genus of Ammonite named Coeloceras which only occurs in a bed that is about one inch in thickness. This Ammonite, at present, can only be found within this very specific horizon. It is possible that in the future new records of this Ammonite may be found in neighbouring beds, which could extend its stratigraphical range. However, it is impossible that it will be found in rocks that differ by say 20 million years, and therefore I would say this is a factor that could be used to distinguish fossils.

I came up with a few different factors, can you think of any others that I missed? I am going to relate all my factors to Ammonites, but in theory, they should be applicable to every fossil.

1) Morphology - fossils can be distinguished by their physical appearance. E.g. Different Ammonites have different sizes and therefore if you find a specimen that is 30 inches, that rules out some genera, allowing a distinction to be made.

2) Temporal - fossils only occur in limited strata. E.g. It is impossible to find a Cretaceous Ammonite in Jurassic rocks. You can use this to make distinctions.

3) Preservation / Rock Type - Similar to temporal, but I think it is slightly different. Some types of fossils only occur in specific rocks. E.g. In Dorset you won't find a Androgynoceras in a birchi nodule. If you know the rock type, you can rule out certain fossils.

4) Associations - Most fossils have a limited number of other fossils they can be found with - E.g. If you find an Ammonite that you know what it is, you may be able to rule out certain other unknown ammonites in the association. Again, similar to preservation and temporal.

5) Location - Most fossils only occur in certain locations. E.g. In Dorset, you won't find an insitu Promicroceras at Seatown. The fossils have a limited rock type, occurring in a limited location, allowing this to be used for distinction.

I suppose these 'factors' would almost be the criteria for defining the palaeospecies concept, quite different from the traditional things we think of when defining the biological species concept. What do you think?

Best wishes,

Joe

[Auspex]

All the factors named are salient, but all require some subjectivity in interpretation (especially for fossils, which do not have any surviving DNA). The Modern Synthesis introduced biological and ecological methodology into the mix, and Paleontology is still adapting after 40 years. It will never be cast in stone (no pun intended).

[Ammojoe]

Thanks, Chas. You are of course quite right about subjectivity being key in interpretation. Thank you for your input

[Diceros]

The fundamental building blocks of all taxonomy are the species, so genera are just helpful groups of species united by one or more shared-derived characters.

The critical issue is when is a population a valid species? To me, like a geological formation, a valid species must have discrete boundaries - morphological gaps on either side (as unconformities make a formation separate, and mappable, from the ones below and above it), distinguishing it from the species most closely related to it. Like members (or facies) in geology, subspecies, while not critical to taxonomy, are very useful groupings, especially in biostratigraphy. Because subspecies, like members, are intergradational with other close units (other subspecies or members), they can't be used in taxonomy (geology) in the same way discrete species (or formations) can.

So how do you distinguish between a useful subspecies and a useless one (to take an extreme position, it'd be pointless to name each specimen)? An American Museum ornithologist, Dean Amadon, once came up with what I think is an excellent rule for when a subspecies is useful, and it has the advantage of being just as useful for fossils. People call it the 75% Rule - Amadon wrote that if only 50% of a population of modern birds could be distinguished from the next closest population, then it was only a toss-up if a particular bird in the population could be distinguished from a member of the next subspecies (or an individual fossil from one level could be distinguished from one from the next lower or higher level). If, on the other hand, most (75%) of one population could be distinguished from most of the next closest population (and only 25% couldn't be distinguished), then the ornithologist/paleontologist was justified in artificially cutting up the continuum into meaningful (for modern taxa by geography, for fossils by time) subspecies. The point is, that you want subspecies which are useful.

[Ammojoe]

Thank you for your reply, Diceros, you have given me something to think about. The reason I used genera in my question was because the fossils my question is raised about are poorly classified to generic level, let alone specific level, and therefore for my purpose it would be more useful to distinguish genera apart than distinguishing the species within a given genus.

I think you are right that a species should have morphological gaps on either side of it - however I have found sometimes fossils separated by a few million years overlap morphologically. This is problematic, and presumably, they should be classified under one palaeospecies (accepting the reality that there may be numerous biological species). Without the luxuries of DNA testing and other cutting-edge methods; classification can become tricky. I think my question concerns other methods that could be used for distinguishing extinct organisms, where we don't have access to DNA and alike. This would have to be in reference to Linnaean classification instead of phylogenetic nomenclature.

I think the 75% rule is interesting. What is the factor used to distinguish one population from its next closest one (or what would be an example, within say Palaeontology, where this would be applied)? I think it makes a lot of sense, but for extant animals, I think a better approach is using genetics and the phylogenetic system and determining evolutionary relationships. What do you think?

Once again, thanks for your input, I really appreciate. You have certainly given me 'food for thought'.

Best wishes,

Joe

[Auspex]

Carl Linnaeus bequeathed upon the world a system by which we might begin to understand the relationships between living things. It is at once powerful, and compromised, by its elegant simplicity. Our best attempt to empirically quantify things of unimaginable wonder, we cannot hope to ever claim complete knowledge of what simply is.

[Diceros]

Let's see, examples of modern and fossil species vs subspecies. Say you've got a modern bird that's gray on one side of the state, and red on the other. If, toward the middle of the range, there's a muddle of grayish-red ones, then the form is intergradational, and there are two useful geographic subspecies (assuming that in the middle population, only 25% of the birds can't be distinguished), which intergrade across space. If, on the other hand, there's a sharp morphological gap between the red and gray one, then you have two valid species.

OK, fossil example. Say there's a bony fish in the Turonian to lt. Santonian early part of the Lt. Cretaceous whose large, anterior teeth have an oval X-section, with no carinae (ant. & post. cutting edges) or longitudinal facets (like Xiphactinus audax audax); and then a lt. Santonian to lt. Campanian one from the later Lt. Cret., which has both carinae and facets (like X. audax vetus), and only a few teeth in the lt. Santonian sample have the beginnings of the carinae and facets. Those would be (and are) fossil subspecies which intergrade across time (some call them chrono-subspecies, but really, it's the same thing). Until the lt. Sant. pop. was found, they were separate species, X. audax and X. vetus.

Ammojoe - I hope that helps.

[verydeadthings]

Short answer is, it's arbitrary. Hopefully higher level classification like genera reflects some sort of close evolutionary relationship, but these are really working hypotheses. There is a conflict between hierarchical Linnaean classification and cladistics, even if we understood the evolutionary relationship of fossil organisms perfectly. I won't get into it here, but there's a very good paper on the subject that is free online (http://depts.washington.edu/phylo/LabMeetingReadings/Week1/BrummittParaphyly2.pdf).

I'm skeptical of using factors other than morphology (or perhaps chemical/isotopic signatures) to identify fossils. This is partly because I work with nannofossils, which can easily be reworked into younger age sediments. You can never assume that the fossil is the same age as the sediment. This is of course less of an issue with macroscopic fossils, but it's still an issue. However, if two species are morphologically indistinguishable, and are only separated because they had different geographic or temporal ranges, I don't think there is generally justification for splitting the fossils into different species or genera. When dealing with an anagenetic lineage, for practical reasons paleontologists divide the lineage into temporal species or chronospecies. Discoaster quinqueramus is one example of this (http://ina.tmsoc.org/Nannotax3/index.php?dir=Coccolithophores/Discoasterales/Discoasteraceae/Discoaster/D.%20quinqueramus%20group/Discoaster%20quinqueramus). The division is arbitrary, but stratigraphically useful.

Great question, by the way. Taxonomy is one of the most fundamental issues in paleontology.

[FossilDAWG]

Interesting question.

As far as "species" are concerned, paleontology differs from biology in that we are stuck with morphospecies (species defined by morphology without information about genetics/DNA, behavior, or other non-fossilizable traits). Of course, the only morphology we can observe is in hard parts such as shell or bone. In addition, we are often limited in our information about ontogeny-related changes (how morphology changes with development) and in knowledge of how morphology is influenced by environment. For example, many modern molluscs have shells that are thicker in individuals living in high-energy environments, and features such as number and strength of spines is also plastic. Unless we have a large sample size of specimens collected from the same stratum (so a narrow interval of time) that also records a gradient in environment (deep water below wave base to shallow high energy, for example) we have no way of knowing if a shell with subdued spines from one locality is or is not the same as a more spiny shell from a second environmentally different locality. The usual tendency would be to describe these as separate species, a trend carried to an extreme by some workers (cough cough Petuch cough cough). In general, low sample sizes tends to lead to excessive splitting of species. On the other hand, if we really had a continuous fossil record we would see many more examples of one species gradually grading into another, such that the "boundary" between species is essentially arbitrary. For example, think about the last common ancestor between chimpanzees and humans, and now imagine the ancestors of each lineage just a few generations this side of the split. Odds are there would be no significant morphological difference you could use to tell one lineage from the other at that point; nor could you tell either from the last common ancestor. Almost certainly the important difference is that the two populations had become reproductively isolated, most likely by simple geographic separation. We are able to distinguish many fossil "species" because the fossil record is discontinuous, so we get what amount to "snapshots" of lineages at intervals that may be separated by thousands of generations.

It is interesting to think about "ring species". These are species that vary in a continuous fashion in some feature along a series of connected environments that wrap around and overlap at the ends (hence the "ring"); each population can interbreed with the ones adjacent to it but the populations at the morphological extremes at the ends of the ring cannot interbreed. The classic example is the Herring Gull in the North Atlantic, which can interbreed with gulls in arctic North America, which can interbreed with gulls in arctic Siberia, which can interbreed with gulls in northern Europe. As you go from east to west the gulls gradually change in color, so the Herring Gull smoothly grades into the Black-Backed Gull. Remember, each population throughout this gradually changing morphological gradient can interbreed with populations next to it on the gradient. Here is the amazing thing, though: the beginning (Herring Gull) and end (Black-Backed Gull) of the continuum overlap in northern Europe, where Herring Gulls do not interbreed with Black-Backed Gulls. So where they overlap they behave as reproductively isolate good biological species, even though there are no breaks in the east-west gradient where reproductive isolation suddenly appears.

Speciation over time must be a lot like this. Each of us would certainly have been interfertile with people from a few generations ago, and indeed with people from hundreds or even a few thousand generations ago. However if you go back far enough there would be enough accumulated genetic differences to make fertile pairings impossible. There is a gradual change from generation to generation (similar to the small differences between adjacent populations of gulls), not enough to create reproductive barriers (=speciation) until enough distance (in space or in time) has accumulated.

When it comes to fossils, distinguishing species is not too hard if all we have are the end points (Herring vs Black-Backed Gulls), or at least samples of widely separated populations. It becomes a lot more arbitrary when we have lots of closely spaced populations.

Personally, although I recognize the utility of the "holotype" concept (designating a single specimen as the perfect exemplar of the species), I think we would be better served by basing species concepts on a series of specimens that encompass the morphological variability of the species. Of course when it comes to fossils sometimes all we have is a single specimen, which by default becomes the holotype, even it later is discovered that that specimen was actually an extreme morphological variant. Also if we can't compare DNA or other features, when we have a series of specimens it becomes a circular argument to decide what ones are similar enough to include in our species, and what ones vary too much to include. Often it comes down to a matter of personal philosophy about "how different is too different" to be the same species.

In biology, a genus is a group of species that are judged to share a common ancestor (base on a series of shared traits). Sometimes, of course, a species is so distinctive that it is not close to any other species, in which case it is put in its own genus. There is no set amount of difference, either in terms of morphology or genetics, that is universally agreed to be "just right" for a genus, for example in comparison to the amount of distance needed to distinguish species on the one hand or families (groups of genera that share a common ancestor) on the other hand. In the past what was enough to make a genus was pretty much arbitrary and depended on the taxonomic philosophy of the scientist doing the study. It's a bit more objective these days, at least you would have to do some sort of statistical analysis to reconstruct likely phylogenetic trees and show your genus is composed of species that form a natural group (adjacent twigs that root to the same spot on your tree).

Since paleontological genera are based only on morphology, they may be more arbitrary, groupings of species based on some researchers idea of what features are most important in indicating relationships. I suspect certain groups, such as the ammonites, have been split to an excessive degree based on fine and intergrading morphological features. When species get split based on minute differences, I suspect genera get split too. As a hypothetical example, if a single variable biological species is split into 10 morphospecies based on tiny variations, those 10 species will form a cluster of closely related "species" and that cluster will be named as a genus. I wonder about this with regard to ammonites, especially many of the Jurassic ammonites, because you will find so many genera that are represented by numerous named species in a single layer at a single locality. I suppose it is possible that there were a dozen species of the same genus, all differing by very subtle characters, living side by side, but that seems unlikely to me, in part because those species would have had to exploit different niches (different food, or feeding in different parts of the water column), or else be in direct competition with one another. I think such species concepts should at least be supported by statistical analysis of large sample sizes, showing the "species" can be distinguished at the 95% level (95% of the specimens unambiguously assigned to one or the other species), but this is rarely done.

In recent decades, reexamination of described Cretaceous ammonites by workers such as Kennedy has resulted in many previously recognized species being synonymized (as well as some new ones being recognized). The result is more natural species and genus concepts, with an improvement in the utility of ammonites for global correlation as certain species are recognized as being very widespread, instead of having each population from different geographical areas being given a different name.

Don

[Auspex]

The bugbear is in acknowledging convergence in the evolution of a morphological trait. Under conditions of environmental change, populations will acquire and shed traits that become useful or harmful. That which can persist, might; that which cannot, will not.

[Doctor Mud]

Carl Linnaeus bequeathed upon the world a system by which we might begin to understand the relationships between living things. It is at once powerful, and compromised, by its elegant simplicity. Our best attempt to empirically quantify things of unimaginable wonder, we cannot hope to ever claim complete knowledge of what simply is.

Well put Auspex!

Taxonomy these days is far more than "stamp collecting", but is in fact an attempt to understand the "inter-relatedness" of all life on earth. Where did organisms come from? How do they interact? What is their ecology?

As has already been pointed out here, there is often a large morphological variation within what might be defined as a species.

The challenge comes in the fossil record since it is difficult to examine how a fossil organism interacted with close relatives (could it breed and produce viable offspring, this is one definition of a species) is its ecology or physiology different? (altered tolerance to environment, changed diet, behaviour).

I work with fossils of animals that are often still extant. I work a lot with fossil chironomids or non-biting midges in lake sediments. I see a lot of morphological variation within what is defined as a species, and sometimes there are individuals that are quite different, but could still be defined morphologically as a described species. In this case I tend to call these morphotypes until the status as a species can be defined. This is an unusual situation in the fossil record though as often we will never have the chance to test if an organism is a species by any more than morphology.

I work a lot with Holocene sediments (Last ~ 12,000 years) so sometimes I still come across extinct species though and have the common fossil record problem. Is the morphology different enough to define a new species.

This comes back to what Auspex said and why it resonated with me. I really support taxonomic research and think it is really important. To me though as a paleoecologist, often I use species as proxies for environmental change. So the question for me is, is the behaviour or physiology of the morphotype different?

I am also interested in biogeography. Why are organisms located when and where they are. The species is a useful unit to talk about organisms, but it does have its limits. It is a pigeonhole system and nature is more elastic.

Very interesting topic here and a fundamental question in paleontology.

[RichW9090]

There is also the problem with species allocations in fossils, which are often, in faunal papers, done on the basis of geography rather than on the basis of a specdimen having a diagnostic character (apomorphy) for that species. A good example is the Pleistocene distribution of Antilocapra, the pronghorn "antelope". There are, in the literature, about 90 Pleistocene records for Antilocapra in western and central North America. However, when I went through the literature carefully, and looked at most of the specimens in person, only 3 of them are definitely identifiable as Antilocapra, all three horn cores. The rest are all fragmentary postcranials, and could just as easily be identified as Tetrameryx, which overlaps greatly in size with Antilocapra, or even be Stockoceros, which has a slight overlap with Antilocapra. So they are most correctly listed as Antilocapridae.

Eventually, we may be able to identify some of the postcranial elements - but we lack any associated cranial and postcranial remains of Tetrameryx, and there are no population samples of Tetrameryx like we have for Antilocapra (Natural Trap Cave) and Stockoceros (Papago Springs Cave and San Josecito Cave). While the remains of about 50 individuals of Tetrameryx have been found at Irvington, California, the specimens are all horn cores - only a few - perhaps 6 - postcranials were collected there.

Most of the identifications of Antilocapra in the literature should never be used for zoogeographic evidence. The identifications were made on the basis of supposed age and geographical location.

[ashcraft]

There is a trackway located in a cave in Perry county Missouri, where I used to teach. It's published identity is a jaguar trackway. ID is a bit iffy in my mind. First of all, the size of the tracks fall into the range of the Mtn. Lion, jaguar, and some sabertooths. They pinned the jaguar ID on it because a very small mandible section was also found in the cave, which fell into the jaguar size range.

I keep this in mind every time I read a journal.

Brent Ashcraft