Well, Cau found it as a basal spinosaurid instead of a baryonychine like Allain et al. (2012) found, so this wasn't totally unexpected. So I constrained it to be a spinosaurid and found it takes five extra steps, which isn't too bad. Next I constrained it to be a non-allosaurian carnosaur in Allain et al.'s matrix and it takes nine extra steps, which is getting into improbable territory. Let's look at the details.
|It's a sauropod pubis... er no wait.... Ichthyovenator's ischium (after Allain et al., 2012).|
In Allain et al.'s matrix...
- Ichthyovenator is a megalosauroid due to having a vertical ridge anterior to the hyposphene in the dorsal vertebra. But this is miscoded and is actually absent.
- It is a spinosaurid due to the tall dorsal neural spine, but these are also present in Concavenator, which was not included in their analysis.
- It is placed in Baryonychinae due to four characters- the basally webbed dorsal neural spine.
- ... An accessory centrodiapophyseal lamina on the dorsal, which is also present in the unincluded Concavenator.
- ... A mediolaterally expanded pubic boot.
- ... And a posterior pubic boot "reduced to a small flange". But this is correlated with the last character, as a reduced posterior pubic boot (with insignificant anterior boot) leads to a transversely expanded distal pubis, and indeed the only taxa coded as having the first state are also the only ones coded as having the second.
- It is excluded from Avetheropoda due to the small pubic boot as well.
- It is excluded from Allosauroidea because the posterior dorsal neural spine is not anteriorly inclined, but this is also true in Concavenator).
- It's also excluded from Allosauroidea because the ischial obturator foramen is closed. Yet the two coelurosaurs are miscoded as lacking an obturator notch; when that's corrected, Ichthyovenator's closed foramen becomes an autapomorphy.
So of the nine characters, three are due to Allain et al. not including Concavenator in their matrix. Two are due to miscodings. The two pubic boot characters are correlated, so are really only one. Yet this character was not used by Carrano et al.. Finally, the neural spine webbing is valid and used by Carrano et al..
In Carrano et al.'s matrix...
- Ichthyovenator is sister to Concavenator due to three characters. The tall dorsal neural spine (also in spinosaurids).
- ... The accessory centrodiapophyseal lamina (also in baryonychines).
- ... And a peg-and-socket ilioischial articulation (not included by Allain et al.).
- It is an avetheropod because of four characters. It has a narrow brevis fossa (coded more strictly in Allain et al.).
- ... An m. cuppedicus shelf (not included by Allain et al.).
- ... An open pubic obturator notch. However, Suchomimus and Baryonyx are miscoded by Carrano et al. (Rauhut, 2003; Charig and Milner, 1997).
- ... And a large and oval pubic obturator notch/foramen (not included by Allain et al.).
- It is excluded from Megalosauroidea due to lacking a vertical ridge anterior to the hyposphene (miscoded by Allain et al.).
- ... And having a vertical ilial ridge (not included by Allain et al.).
So of these nine characters, two are correctly coded as also present in baryonychines, and another is miscoded as being absent in baryonychines. Five others aren't used by Allain et al., and another was miscoded by Allain et al.
In Allain et al.'s matrix, after recoding Ichthyovenator's hyposphene ridge and the coelurosaurs' obturator notches, and excluding the correlated pubic boot character, Ichthyovenator is still a baryonychine. Constraining it to be a non-allosauroid carnosaur takes five extra steps, which is down from the nine it took before (though it was slightly more nested in Carnosauria before, but only by two steps). If we add the five characters Allain et al. didn't use that support this position (the brevis fossa expansion counts too since it is an additional state of a character), then either alternative is equally parsimonious.
In Carrano et al.'s matrix, after recoding Baryonyx and Suchomimus for their obturator notches, Ichthyovenator is now a non-allosauroid carnosaur and Concavenator is back to Carcharodontosauridae (in case you're wondering, making them sister taxa like before is now one step longer). It now takes three more steps to make Ichthyovenator a spinosaurid, so that's dropped from five. If we add on the transversely wide pubic boot character that they didn't include, we could say it only takes two more steps to make Ichthyovenator a spinosaurid.
Amazingly, after only six corrections and taking into account the unused characters in each, both matrices give almost the same answer- Ichthyovenator is either equally likely or two steps more likely to be a basal carnosaur instead of a baryonychine. The main flaw was that Allain et al. didn't include several useful characters, which is expected considering their total character number is 51% of Carrano et al.'s. I'd put it in Orionides incertae sedis for now.
References- Charig and Milner, 1997. Baryonyx walkeri, a fish-eating dinosaur from the Wealden of Surrey. Bulletin of the Natural History Museum of London (Geology). 53, 11-70.
Rauhut, 2003. The interrelationships and evolution of basal theropod dinosaurs. Special Papers in Palaeontology. 69, 1-213.
Allain, Xaisanavong, Richir and Khentavong, 2012. The first definitive Asian spinosaurid (Dinosauria: Theropoda) from the Early Cretaceous of Laos. Naturwissenschaften. 99(5), 369-377.
Carrano, Benson and Sampson, 2012. The phylogeny of Tetanurae (Dinosauria: Theropoda). Journal of Systematic Palaeontology. 10(2), 211-300.