Another word on arachnid phylogeny

Shultz (2007) has just published a new paper on arachnid phylogenetics, based on morphology. As you can see if you scan my profile blurb, I'm currently working on arachnids - specifically on harvestmen (Opiliones), so I'm always happy to see work on them. Shultz (2007) sits in couterpoint to the most recent other publication on arachnid high-level phylogeny, Giribet et al. (2002), which used combined molecular and morphological data.

I do feel the need to make a few comments on character coding. One of the issues with high-level morphological phylogenetics is that it becomes increasingly difficult to code characters without interpretative bias. One example of this that I can spot in Shultz (2007) is his coding for the tracheal system (character 126). In the past, presence or absence of a tracheal system has generally been treated as a single character. Shultz argues on the basis of differences in the layout of the tracheal system that it has probably evolved independently a number of times (an idea that does gain some support from the definitely independent evolution of a tracheal system in some spiders), and codes the different tracheal systems as different character states. While Shultz is likely to be correct that the different tracheal systems have evolved independently, his coding of the systems separately a priori excludes the possibility of their homology. Also, in his comments on presence of a penis (character 160), a synapomorphy of Opiliones, Shultz notes that 'A clearly homologous structure is present in Cyphophthalmi (Opiliones) and apparently functions in depositing a spermatophore in the female’s genital chamber', then follows with 'The ‘penis’ in Oribatida is really a spermatopositor; it functions in construction of a spermatophore'. Is it really justifiable without prior phylogenetic expectations to code the cyphophthalmid structure as a 'penis', but not the oribatid structure?

Shultz uses a few exemplars from each of the living arachnid orders, as well as a fossil exemplars of a number of them (he includes more fossil taxa than Giribet et al.) plus the fossil Eurypterida, Trigonotarbida and Plesiosiro. An analysis is run without fossil taxa, then one with. At first glance, the inclusion or exclusion of fossil taxa has a significant impact on topology. Without fossils, and using Xiphosura (horseshoe crabs) as an outgroup, the recovered topology is (Palpigradi ((Ricinulei (Anactinotrichida, Actinotrichida)) ((Araneae (Amblypygi, Uropygi)) ((Scorpiones, Opiliones) (Pseudoscorpiones, Solifugae))))). However, through in a few fossil taxa and you get ((Scorpiones, Opiliones) (Palpigradi, (Actinotrichida (Ricinulei, Anactinotrichida), ((Pseudoscorpiones, Solifugae) (Araneae (Amblypygi, Uropygi)))))*.

*I wanted to use actual trees for this section, if only in ASCII format, because that would be one hell of a lot easier to read, but it looks like they won't show up properly in the final page. if anyone knows of a way I can put in trees on this site, I'd be ever so grateful to hear it).

Actually these two topologies are nowhere near as different as they appear - the support values for most supraordinal clades are ghastly. If we collapse all nodes in the final tree with less than 50% support, we get (Scorpiones, Opiliones, Palpigradi, (Ricinulei, Anactinotrichida, Actinotrichida), Solifugae, Pseudoscorpiones, (Trigonotarbida, Araneae (Plesiosiro (Amblypygi, Uropygi)))). To somewhat mitigate the drawbacks of this low support, however, Shultz does test his results against other past theories.

Shultz runs his neontological data set through a number of analyses constrained to recover particular clades. Clades suggested in the past that appear in trees only one step longer than Shultz's most parsimonious tree are Ricinulei + Anactinotrichida, Megoperculata (Palpigradi + Araneae + Amblypygi + Uropygi) and Rostrosomata (all arachnids except Scorpiones and Opiliones). The last one, notably, is what is recovered when palaeontological data are included. Scorpions sister to all other arachnids is only two steps longer, as is Micrura (arachnids except Scoropiones, Opiliones, Pseudoscorpiones and Solifugae). It is a little disappointing that these comparisons are run on the neontological data set alone rather than the complete data set, considering that the fossil taxa are not without influence on the result. Two comparisons are made using the full data set, testing a scorpion + eurypterid clade and a trigonotarbid + ricinuleid clade (the latter possibility was found by Giribet et al., 2002). Both possibilities are noticeably longer than the preferred tree.

A notable absence from Shultz (2007) is the Pycnogonida. Pycnogonids or 'sea spiders' are patently bizarre marine animals of very obscure relationships. Traditionallly they have been regarded as basal chelicerates owing to their possession of chelate pre-oral appendages, and many authors still support this view. Other authors regard pycnogonids as the sister group to all other living arthropods. When pycnogonids were included in the analysis of Giribet et al. (2002), they appeared in a completely unexpected position as sister to Palpigradi, within Arachnida. My impression on reading Giribet et al. is that the authors themselves are extremely sceptical of this result, and seem more inclined to attribute it to the high level of autapomorphy in pycnogonids. While it would have been interesting to see Shultz test the position of pycnogonids, it is possible that said degree of autapomorphy may have simply blown Shultz's analysis out of the water.

REFERENCES

Giribet, G., G. D. Edgecombe, W. C. Wheeler & C. Babbitt. 2002. Phylogeny and systematic position of Opiliones: a combined analysis of chelicerate relationships using morphological and molecular data. Cladistics 18: 5-70.

Shultz, J. W. 2007. A phylogenetic analysis of the arachnid orders based on morphological characters. Zoological Journal of the Linnean Society 150 (2): 221-265.