Question

How reliable are morphological and molecular data in evolutionary studies? What are their weakness? How can they complement each other? Pls provide a reference like a journal article. thanks

Answer 1

The major difference lies in the space of possible states and in the number of relevant levels of granularity and frames of reference. Whereas in molecular data at the nucleotide level the space is limited to only 4 different states (the 4 different nucleotide types), in morphology there is practically no such limitation. Moreover, whereas in molecular data you usually analyse the data either on the granularity level of individual nucleotides or on the level of amino acids. Different frames of reference are usually not used, but sometimes gene arrangement and 3d structure is also considered. In morphology, all different kinds of levels of granularity are considered, from the molecular level to the gross anatomy level. Moreover, different frames of reference can be considered, as for instance a purely spatio-structureal frame as opposed to a functional or a developmental frame.

All this affects the descriptive level as well as the analytical level in phylogenetic investigations. Molecular data can usually be described as a sequence of letters using a 4-letter code. This lends itself to very formalized descriptions of sequences. In morphology, we are dealing with a lack of such formalism. Instead, we are facing different terminological traditions that are often taxon-dependent, where the meaning of terms can vary between different authors and through time and where also homology considerations influenced terminology. No formalized standard exists for morphological descriptions.

In the step of character identification and homology assessment, using molecular data we must align sequences and then each nucleotide position in the sequence is a putative phylogenetic character with a set of 4 different possible character states. This can be easily formalized for phylogenetic algorithms to analyze the data. This is straight forward. In morphology, this step is tricky due to the space of possible states. The delimitation of characters is non-trivial and the different levels of granularity and frames of reference make it even more complicated. We have discussed these differences between morphological and molecular data in some detail in our 'The Linguistic Problem of Morphology' paper (https://www.researchgate.net/publication/215827731_The_linguistic_problem_of_morphology_Structure_versus_homology_and_the_standardization_of_morphological_data).