"Total" evidence is better than partial evidence simply by its being "total", i.e. taking into consideration all available information, not pre- (or post-)selected according to the researcher's preferences. Of course the term "total" is somewhat confusing - we never have really total knowledge, and even part of what we know is often impossible to use and/or reliably interprete - but if we wish to maximally approximate the truth, we should try to make the basic evidence as total as possible.

With fossils we can compare any of the features that are preserved in the fossils, such as number and shape of teeth, ratio of femur length to tibia length etc. And these can be compared to other fossils and to living organisms. But we can't say much about body temperature or color of the fur etc from bone fossils.

With physical characteristics such as length of femur bone, it can be difficult with fossils to know the age of the organism when it died, so it may be difficult to know if a small femur bone is from a juvenile of a large species or an adult of a smaller species, unless we have a lot of fossils of that species. Also, there can be a lot of variation within one species (think about miniature horses, small and large breeds of dogs, different body sizes and shapes within humans, etc) so that can cause some problems with fossil data.

Molecular data (DNA sequences etc) can have its own set of problems too, but it has become very easy to generate massive amounts of DNA sequence data, and it has proven to be very helpful in the study of evolution.

Brian is correct. The phylogenetic relationships of fossil organisms may be inferred the same as for extant organisms, except that those inferences are usually based on a subset of available characters (e. g., no DNA in most cases). For this reason, fossils are often discussed in the context of the problems of "missing data." See, for example, chapters in Novacek and Wheeler, "Extinction and Phylogeny" (1992) ISBN-13: 978-0231074384

and articles by Colin Patterson and John Wiens, among others.

Another problem with the comparability of fossils occurs when living taxa are defined by molecular similarity so there becomes a disjunction between the comparison of morphological similarity used for fossils and molecular similarity used for living taxa. The problem is especially acute when morphology and molecules do not give the same answer for living taxa. In which case the usual choice is to say that morphology is unreliable as a phylogenetic indicator. But then those very same people will look at fossils and claim to know what they are and their relationships with living taxa. The absurdity of all this is exemplified in human-ape phylogeny where we are supposed to be the sister taxon to chimps, and yet the australopithicenes have all the features that would place them closer to orangutans. Nuts.

Don't rely on a text from 1977 to learn about the current state of phylogenetics! Phylogenetic analysis was only introduced 20 years later in many disciplines.

DNA and molecular inferences of phylogenetic relationships among extant and fossil taxa are increasingly becoming under scrutiny. While DNA data/information has been a huge step forward in systematic taxonomy, it is noteworthy to remember that what Charles Darwin would call, “Selective Pressure,” a phenomenon that alters the behavior and fitness of living organisms within a given environment, it is the driving force of evolution and natural selection. If you don’t evolve and adapt, you’ll stagnate and may became replaced/extinct. None of this is captured in the present phylogenetic analyses, molecular/DNA or otherwise.

Of course there are other current views about evolution that do not assume that selective pressure is the driving force of evolution (or even that there is a 'force' pushing evolution along). Certainly under changing conditions if a species does not have the range of variation to survive it will become extinct - but that does not make selective pressure a driving force of evolution. All phylogenetic analysis does is to identify relationships of different taxa (at whatever level).

They DO apply to ALL ORGANISMS both fossils and Extant ! This is basic in student teaching !! It should be !!! Against many molecular biologists' opinions !!! And some paeontologists's ……… Total evidence MUST be favored in ALL cases !!!!! Ask experts as e.g. Pascal Tassy : Pascal TASSY

Alain, please explain what it is that you were referring to that 'apply to all organisms'

Also, what is your argument about total evidence?

Lelièvre et al was interesting to read, particularly for the view that age of taxa should be considered in the context of the sister group age (a classic example is the relative age of the oldest angiosperm fossils vs the age of gymnosperms, former placing origins in the Cretaceous, latter in the Triassic or earlier), but not clear about relevance to your overall comment.

YES all organisms, both Extant and Fossil must be included in analysis !! That's it. With of course coding problems that can be partially solved … Total evidence has been shown as "better", that's it !! Years ago. Go to Cladistics and equivalent journals.

Alain

"Total" evidence is better than partial evidence simply by its being "total", i.e. taking into consideration all available information, not pre- (or post-)selected according to the researcher's preferences. Of course the term "total" is somewhat confusing - we never have really total knowledge, and even part of what we know is often impossible to use and/or reliably interprete - but if we wish to maximally approximate the truth, we should try to make the basic evidence as total as possible.

Taking all evidence into consideration is certainly a laudable goal. But the discerning question is how that evidence is considered. In some views total evidence means combining all evidence, as if all the evidence is equal in its validity or significance. One sees this in systematic studies where morphological and molecular similarities are combined as if they were equal. Maybe that is true, but not necessarily so. This is the problem in hominid systematics where molecular and morphological evidence actually give contradictory answers about the sister taxon of humans. Of course there are many more sequence similarities involved so they can always swamp out any contradictory morphological similarities. And in so doing this creates another enigma for the study of fossils where morphology is seen to give the wrong answer for living taxa, but miraculously can be relied on for classification of fossils and their relationships to living taxa.

"where morphology is seen to give the wrong answer for living taxa"

That, actually, hasn't been tested.

There is no phylogenetic analysis of morphological data on the phylogeny of crown-group birds or placental mammals, for example, that has anywhere near enough characters or anywhere near enough fossils; you can tell this on their own terms, without comparing them to molecular datasets. And that's before we even get to such common issues as redundant characters...

Morphological phylogenetics, done right, is a lot of work. With molecular data you can do in a few weeks, or months at most, what takes a PhD thesis or three with morphological data.

Rather than giving the wrong answer, it often hasn't given any answer at all yet. We don't know that it still gets things wrong when it's done right.

The point I was making is that all the 'authorities' on hominid systematics view molecular similarities as giving the correct answer about sister taxa relationships for humans. According to these authorities morphology is unreliable because it does not support the same sister taxon relationship. These same people then tie themselves in knots because they treat morphology as reliable when studying fossils.

To say that difference between the results of morphological and molecular analyses means that the morphological "gives the wrong answer" is nowadays common but is no more than a foundless preconception: in reality such difference means that either morphological or molecular data (or both...) are misleading (or had been misinterpreted). As to the "total evidence", of course it should be weighted (if we have some information as to the appropriate "weights") or (preferably) various (groups of) data should be analysed separately, the results compared, and the differences (if any) evaluated (e.g. based on some "external evidence" - biogeography, ecology &c.).

The idea that DNA sequences swamp out morphological characters simply by virtue of their greater number of characters is not necessarily true. DNA is almost always very homoplastic, and if there is a good morphological signal in the data, it usually prevails. In my experience, morphology and molecules for my group (butterflies) are usually in agreement with one another. See

Simonsen, T. J., N. Wahlberg, A. V. Z. Brower, and R. de Jong. 2006. Morphology, molecules and fritillaries: approaching a stable phylogeny for Argynnini (Lepidoptera: Nymphalidae). Insect Syst. Evol. 37:405-418.

Wahlberg, N., M. F. Braby, A. V. Z. Brower, R. de Jong, M.-M. Lee, S. Nylin, N. E. Pierce, F. A. H. Sperling, R. Vila, A. D. Warren, and E. Zakharov. 2005. Synergistic effects of combining morphological and molecular data in resolving the phylogeny of butterflies and skippers. Proc. R. Soc. Lond. B 272:1577-1586.

Brower, A. V. Z., N. Wahlberg, J. R. Ogawa, M. Boppre, and R. I. Vane-Wright. 2010. Phylogenetic relationships among genera of danaine butterflies (Lepidoptera: Nymphalidae) as implied by morphology and DNA sequences. Systematics and Biodiversity 8:75-89.

I haven't followed the hominid literature, but so far I haven't seen a well-made phylogenetic analysis of morphological data in that field either...

OK. So what is a 'well made' phylogenetic analysis. Just because one has not seen one does not mean that one does not exist. Without any criteria given its not possible to empirically address. So I would assert that the attached paper is the most 'well made' morphological study ever made. But of course it went against the grain and so had been largely ignored (which is standard scientific practice and so beyond reproach).

Thank you very much for your paper, I'll read it and report back.

In short, a "well made" phylogenetic analysis contains enough characters (what that means depends mostly on the number of taxa) that are not redundant or, ideally, otherwise correlated (heterochrony is a frequently overlooked cause for such problems), and it isn't riddled with typos and other accidental mistakes as so many published matrices are. Most of my publications so far are about these issues.

"it went against the grain and so had been largely ignored (which is standard scientific practice and so beyond reproach)" - ???