Making a tree using morphological characters is the basis of phylogeny. What do you think people did before they had molecular data! Though, if it is publishable in 2014, I doubt it! Why don't you try to extract DNA from the sample to run PCR, you'll find lots of DNA formalin extraction protocols by simply using google.

regards

Hi, of course both molecular and morphological characters are useful for phylogenetic researches. in my opinion molecular data are more trustful because the genome change slowly. morphological properties will change faster and by a change in climate we can get a biased results on phylogenetic tree which isn't really correct.

It depends if you have good characters that clearly define species and potential clades, if little molecular data is availeble to you a morphological phylogeny could at least give some support to morphoanatomic theories. If you can follow it up with formaline extraction and do a molecular analysis in colaboration with the morphological that would of course be the best as Orr and Eslami suggests. Morphology based phylogenies are still being published when moelcular data is unavaileble, and can be quite usefull for future work on obscure species see obscure goastropod examples:

Systematics and phylogeny of Philine (Gastropoda: Opisthobranchia), with emphasis on the Philine aperta clade Price et al. 2011

Article Towards a phylogeny and evolution of Acochlidia (Mollusca: G...

Of course! Before genetic revolution phylogenetic relationships were constructed based in morphological characters. For example, even with the amount of genetic data available today, one of the most used paper of skates phylogenetic was only with morphological observations:

McEachran, J. D., & Dunn, K. A. (1998). Phylogenetic Analysis of Skates, a Morphologically Conservative Clade of Elasmobranchs (Chondrichthyes: Rajidae). Copeia, 1998(2), 271–290.

My samples are too old about 2-3 years but i will try to extract DNA and hope PCR works.

https://www.researchgate.net/post/Can_anyone_suggest_a_protocol_for_extracting_DNA_from_formalin_fixed_wax_embedded_tissue1

As a palaeontologist, the answer to this one is rather obvious.... it's hard to get DNA out of a 500-million-year-old sponge in a completely extinct lineage. When describing the evolutionary past, we use whatever tools are at our disposal; genetics is one of them, morphology another. There are even some other approaches, such as stratophenetics, which have been developed in order to make use of the data available, whatever they may be. We would do well to remember that all of these tools are just that; the truth is an abstract that we try to approach by using them.

In a word yes. Molecular data is just one of the tools we have available to define species and develop hypotheses about the relationship between them. In plants molecular data has reinforced the relationships discovered using morphology but also allowed us to discover new and often surprising relationships where the morphology had mislead us.

I suggest molecular data since morphological data cannot be trustable which can be reflect environmental adaptations, etc. Molecular, however, is more reliable. Your materials were preserved in formalin, so you can work with small parts of genes at the beginninng.

Yes, you can do phylogenetic (cladistic) analysis with any type of data. However, there is a fundamental difference. With morphological data, what you obtain is a hypothesis about the underlying phylogeny, whereas with molecular data, you can test phylogenetic hypotheses. This is because molecular data per se are not inherently "better", but you have much more characters at hand (orders of magnitude), the degree of homoplasy is generally much lower, and there is a natural code already present. Therefore, a phylogeny based on phenotype data can never be considered a "trusted" phylogeny, but it may give very good insight into patterns of character state distributions and it certainly will provide an excellent base for subsequent molecular approaches.

Robert,

Your three points: 1) more avaible data, 2) more easily comparable data and 3) less homoplastic data, do not justify the dichotomy you proposed. Maybe they are arguments in favor of using molecular data in other aspects, but not justify you assertion:: "With morphological data, what you obtain is a hypothesis about the underlying phylogeny, whereas with molecular data, you can test phylogenetic hypotheses". Could you explain what you mean by that?

As far as I know, the metodology of hypotesis proposal (characters coding/sequence allignment) and characters tests of homology (congruence given parsimany or parametric tests) are present in both aproachs. So both methods have hypotesis proposal and test (which, in turn, produce only a more robust hypotesis).

Daniel, the reason I said this is because you can propose a hypothesis based on any set of characters but you need statistical support to test a hypothesis. Usually, with morphological characters you do not get such support. But this depends on the group. In organisms with complex morphology and anatomy, such as vertebrates, you might get quite a high number of shared synapomorphies, similarly so in vascular plants, which is why higher-level phylogenies in these groups have been predicted rather well. Still, even there, hardly anybody would accept morphological characters for hypothesis testing (of course in fossils you have no alternative). But at the lower level (family, genus, species), you run out of potentially synapomorphic characters because of the higher degree of evolutionary plasticity. In contrast, for example in a multi-gene dataset with thousands of bases, you can easily get dozens to even hundreds of synapomorphic sites for shared lineages, even if many sites might also be homoplastic. This provides sufficient statistical resolution for testing a hypothesis. But for example if you have 3-5 morphological synapomorphies for a presumed sister-group relationship, you can use these only for proposing the hypothesis, but not for testing it. This has been shown many times, that organisms may look extremely similar even to the detail but turn out not to be closely related. On the other hand, if you get strong support for a relationship based on molecular data, especially when using more than one gene, in rare cases this relationship is later shown to be incorrect. Good examples are found in every group of organisms, from fungi to flowering plants to birds. In fact, the general tendency now is to say: two things that look alike (i.e. with presumed morphological synapomorphies)... experience says they are perhaps not closely related.

Thanks Robert, it's clearer now.

Molecular and morphological data both have issues with homology assessment. Morphological data is every bit as reliable as molecular data and there are many studies using real and simulated data that clearly show this (e.g., see the work of J.J. Wiens). However, I would not recommend using morphometric/measurement data to build a tree (i.e., phenetics). Those data are better suited to mapping onto a phylogeney constructed using molecular or morphological data or the combination of them (i.e., a combnied analysis). 

David J Cantrill's statement above on plants applies 100% to Mollusca, for instance.

Using morphology alone is not the best choice, but we have to rely on it when nothing better is availabe. Especially when fossil species have to be included in classification, for which biological data will ever be wanting. Most Mollusca families with recent representatives include a lot of fossil species, known only from their shell, and ranging from Holocene to Upper Cretaceous, if not more.

All biological methods, which are the main basis for a modern classification, are applied to recent species. When soft part are available for study, what is far from always being the case. Then shell morphology remains the only clue to include the huge amount of related fossil forms, and to confirm the global philogeny of the considered group.

Of course, devising a phylogeny for soft-bodied animals (Solenogastra, Nudibranchia, etc.) where fossil forms are not known is a bit more tricky.

In most of the hymenopteran groups I work on, there are generally only old museum specimens (often 50++ years) available and no fresh or freshly preserved specimens we can use to extract DNA. We often have to rely on morphology.

Leaping into this discussion a couple of years late, but perhaps still pertinent.

Comments in response to Robert Lücking’s observations

“But at the lower level (family, genus, species), you run out of potentially synapomorphic characters because of the higher degree of evolutionary plasticity.”

If one run’s out or cannot find morphological characters then using molecular characters certainly provides additional information

“In contrast, for example in a multi-gene dataset with thousands of bases, you can easily get dozens to even hundreds of synapomorphic sites for shared lineages, even if many sites might also be homoplastic. “

Problem here is that the homologies are not necessarily synapomorphic as they are often constructed by similarity algorithms that try to balance gaps and substitutions according to some overall measure of similarity – really a phenetic methodology. The homologies created by such methods do not exist in nature.

“This provides sufficient statistical resolution for testing a hypothesis. “

Statistics is not necessarily the proof of a method. Pheneticists used a lot of statistics.

“ This has been shown many times, that organisms may look extremely similar even to the detail but turn out not to be closely related. “

If this is because molecular trees did not get the same result then all one can say is that there is incongruence between the morphological and molecular trees, not that one is necessarily the test (falsification) of the other. A classic example is for human-great ape relationships. DNA sequence trees give a human-chimp clade, morphological trees give a human-orangutan clade. The latter is congruent with fossil hominids which also show specifically orangutan apomorphies.

So, is it worth making phylogenetic trees in the absence of molecular data? Yes, but if molecular data are incongruent with that tree then both sources have to be considered as potentially falsified, not just one.