Palæos:		Unit 140: Sarcopterygii
The Vertebrates		230: Rhabdoderma

Sarcopterygii: Actinistia: Rhabdoderma

The Actinistian Dermal Skull

TELEOSTOMI
|--NEOPTERYGII
`--Sarcopterygii
   |--+--Onychodontiformes 
   |  `--Actinistia   
   |     |--Miguashaia   
   |     `--+--Diplocercides  
   |        `--+--Hadronector 
   |           `--+--Caridosuctor  
   |              `--+--Rhabdoderma  
   |                 `--Coelacanthiformes   
   `--Rhipidistia
      |--Dipnomorpha
      `--+--Rhizodontiformes  
         `--Osteolepiformes 
            |--Tristichopteridae 
            `--Elpistostegalia
               |--Panderichthys 
               `--TETRAPODA

Rhabdoderma and the Actinistian Dermal Skull

A few words are required to explain our modifications to Forey's (1981) reconstruction of the braincase, rostrum and skull table -- or perhaps the lack of modification is what needs an explanation.  The figure legend refers to "changes to nomenclature."  In fact, there are none.  Originally we had made a significant changes to the osteology; but we hastily changed almost everything back, hoping that no one would notice our short-lived departure from convention.  

The reason for this change of heart is quickly stated.  One salubrious side-effect of the modest success of Palaeos has been its tendency to suppress any lingering tendency toward principled independence.  The intimidating realization that persons expert in these matters might lurk among the readers of Palaeos has done wonders to restore us to a happy state of quivering sycophancy.  Specifically, we will stick closely to the system of osteological nomenclature described by Forey (1998), on the theory that anyone who writes a 400 page book on actinistian paleontology is likely to be the guy calling the shots on that particular subject for a decade or two.  

As the result of rigorous training and long professional experience, we are likewise adept at crafting facile rationalizations for even the most corrupt and disturbing behaviors.  Thus, within minutes of completing our un-revision of these images, we had already concocted a remarkably plausible line of argument to justify our cowardly retreat to the comforts of orthodox nomenclature.  After touching on a few, trivial matters of detail (most of which actually deal with the neurocranium), we will turn to a brief description of the actinistian dermal skull and, finally, to this post hoc excuse for adopting an attitude of robotic conformity.  

Trivial Matters of Detail

Forey's (1981) original Figure 1 (above) includes an irregular, somewhat elongate bone with two foramina near the ventral margin of the snout.  This bone is unlabelled in the original.  This is most likely the antorbital (= preorbital), in which case the lateral rostral is not shown.  By way of explanation, the lateral rostral is, likewise, an irregular, somewhat elongate bone with two foramina near the ventral margin of the snout; and, to our knowledge, it has never yet been recovered in articulation with a suitable anatomical label.  Accordingly, we have slipped a lateral rostral into the figure, in order to better indicate the margins of the nares.  

Very primitive sarcopterygians, such as Psarolepis and Achoania were unknown in when Forey wrote in 1981, and our expectations of the orbital region have changed a bit.  Note the presence of a basisphenoid pillar close to the double foramina for the main optic nerve in Rhabdoderma.  As discussed in connection with the introductory material on the eye, and in connection with Psarolepis, this is likely an indicator of an eyestalk.  

Finally, when the orbital area is marked out in this way, the processus connectens looks suspiciously like an enlarged and distorted basipterygoid process.  This turns out to not to be a case of simple homology, since Diplocercides has both structures.  However, the structure in Diplocercides confirms that the processus connectens and basipterygoid process are formed from the same location and suggests that the processus connectens may result from duplication of the basipterygoid.  

Brief Description of The Actinistian Dermal Skull

Actinistians almost all have a small premaxilla; but the homey familiarity of this bone is fraudulent.  It serves only to lull us into a false sense of security.  In fact, we are plunged almost immediately into an alien wilderness of tiny rostral bones whose names, positions, and relationships are quite variable, and are subject to additional variation among individuals -- individual paleontologists as well as individual fishes.  Commonly, there are one or more median unpaired rostrals.  In some derived coelacanths, as in derived Dipnoi, the fishes themselves become frustrated with trying to keep track of all these insignificant little slivers of bone, and they wisely fuse the entire rostrum into a solid mass.  Above the rostrum, the skull table of the sphenethmoid region is made up of a variable number of paired nasals, which meet on the midline (except for the first pair, which may be separated by a median internasal).  The nasals are followed by (usually) two pairs of parietals.  The anterior parietals normally overlap the posterior parietals, or they may meet with an interdigitating suture.  In either case, the contact between the anterior and posterior parietals is complex.  Forey (1998).  

These midline elements are flanked on either side by a paramedian series of tectals.  Those tectals which border the parietals are called "supraorbitals," but they are simply tectals with delusions of homology.  Lund & Lund (1984, 1985) report two, or even three, series of tectals in several Bear Gulch coelacanths, but Forey and Cloutier disagree in each case.  Forey (1998) asserts that coelacanths fall into two categories: those with relatively few tectals (< 7) and those with many (> 10).  

By contrast, the dermal skull of the otoccipital region has few pretensions.  For the most part it is formed by two enormous postparietals.  In most actinistians, these are supplemented by a pair of modest supratemporals, and perhaps spiracular bones.  Posteriorly, the skull is typically finished off by a median extrascapular and 1-2 pairs of lateral extrascapulars, which may be tightly or loosely connected to the postparietals.  All of these skull table bones, of both cranial regions, tend to be relatively substantial and frequently bear a thick ornament of ridges or closely spaced tubercles.  

The anterior skull roof (parietonasal shield) is always clearly separated from the roof of the otoccipital region (postparietal shield).  The postparietal shield is always shorter, but the range of proportions is large.  Generally, the more derived forms have shorter postparietal shields.  This is somewhat odd, since one long-term trend in actinistian evolution has been to concentrate all of the brain in the otoccipital compartment.  The dermal joint between the two cranial regions may be straight and simple, or complex and interdigitating.  There does not seem to be any simple functional explanation for these variations.  Forey (1998).  

The lateral skull is full of holes.  There are holes for the nares. There are holes for the rostral organ. There are holes for sensory lines, and there more holes -- apparently just for hell of it.  Specifically, the premaxilla bears the anterior pores of the rostral organ (in Paleozoic forms), the lateral rostral is emarginated to form parts of both anterior and posterior narial openings, with a ventral process braced on the lateral ethmoid (see discussion at Diplocercides).  An antorbital or preorbital normally includes both posterior foramina for the rostral organ.  As for the rest, there is an enormous literature on the forms and courses of the sensory lines in actinistians which, as usual, we will not take up [1]. 

The cheek region is simplified by comparison with most osteichthyans.  Again, there may be individual variation and some exceptions, but the general rule is six bones: three operculars, the squamosal, a postorbital, and any posterior process of the lacrimojugal.  There are no maxillae, quadratojugals, separate jugals, suborbitals, etc.  Generally, the preopercular overlaps the subopercular, which overlaps the opercular (and both of the last two overlap the shoulder girdle).  Forey (1998).  The cheek bones are typically less ornamented, and are frequently thin and overlapping.  In more derived actinistians, the cheek bones may not even articulate directly, leaving large gaps between them.  In this respect, Rhabdoderma is more like a Mesozoic actinistian.

Post Hoc Excuse for Adopting an Attitude of Robotic Conformity

The nomenclature of actinistian dermal bones has been the subject of much uncertainty in the past.  The coelacanth dermal skull is so different from the typical rhipidistian or tetrapod model, that questions of homology are not easy to address.   Historically, actinistian dermal skull bones were first named based on an implicit actinopterygian model, then, successively, on porolepiform and tetrapod models.  Now that we know of Psarolepis, Achoania, and, especially, Styloichthys, it seems increasingly unlikely that most dermal bones of the skull had stabilized before the Actinistia diverged from the sarcopterygian stem.  Zhu & Yu (2002). 

Triangulating back to the stem lineage from these new forms, the following bones seem to have clear homologues throughout the Sarcopterygii: the premaxilla, postparietal, postorbital and (where present) the maxilla.  The premaxilla, primitively, does not seem to have had a strong posteroventral process, and the very short snout of the earliest Sarcopterygii would not have required or accommodated such a structure (Achoania, Psarolepis) (but contra Forey, 1998: 88).  Thus the development of a weak version of the posterodorsal (or facial) process, with lengthening of the rostrum, might very well have developed independently in Actinistia and Rhipidistia.  Where present, the facial process of the premaxilla is pierced or emarginated for the anterior opening of the rostral organ and/or the anterior naris.  Without the stabilizing influence of a strong, unbroken medial buttress, as in osteolepiforms, the rest of the rostrum probably evolved in quite a different direction from the osteolepiform type.  Accordingly, none of the other actinistian rostral bones may be homologous with those of osteolepiforms.  

All sarcopterygians, with the possible exceptions of Achoania and Psarolepis, minimally have a dermal bone covering the posteroventral quadrant of the orbital area.  If this is the same bone in each case, as it appears to be, this is the jugal.  The only uncertainty comes from the actinistians themselves, since they possess the characteristic lacrimojugal with its twisting antorbital process.  However, if we continue to suppose that there is no necessary homology among the rostral bones, then nothing prevents the actinistian lacrimojugal from being just another jugal with antorbital ambitions. 

One would think that the squamosal and opercular  would be easy cases, but they are not.  We think of the opercular as the bone, almost always mobile, which covers the gill chamber.  However, in the  Acanthodii, this is not necessarily the case.  The opercular is more often made up of one or more immobile bones which channel the excurrent from gills with individual covers. We have insufficient information on the most basal Sarcopterygians to say exactly what is going on.  However, Grossius and the Onychodontiformes (sister to the Actinistia) don't seem to have a traditional opercular, the bone of that name being a relatively small, dorsally located plate which looks more like a slightly displaced extrascapular.  That may also be the case with Styloichthys, from the limited information available.  

Not only is the opercular series intermingled with the extrascapular series, it also mixes with the squamosal.   Something often lies behind the jugal and accepts the posterior branch of the sensory line.  However, the variability of this bone is considerable.  In Grossius, there are two possible squamosals.  Both appear to be in series with the small, dorsal opercular, and neither one of the squamosals is  known to contain a sensory line.  In Styloichthys, the squamosal is a massive bone which excludes all others from the cheek.  In the Onychodontiformes, as in Actinistians, the squamosal is usually paired with a bone variously known as the preopercular (in Actinistia), the squamosal 2, or the quadratojugal.  This bone continues the sensory line from the squamosal; and thus it may well develop in series with the squamosal.  Finally, Forey (1998) has developed a strong case that the actinistian squamosal represents the ontogenetic fusion of an anamestic, dorsal plate, with a more ventral bone which develops around the sensory line.  

The presence of a more-or-less complete tectal/supraorbital series appears to be primitive for the Sarcopterygii, and is probably associated with the dorsal sensory line system.  The individual ossifications are too variable to speak of any one-to-one homology between actinistians and others.  However, almost all Sarcopterygian lineages tend to develop an additional bone between the squamosal and the supraorbital series (or the postparietal if the supraorbitals have been eliminated).  This is variously known as the postspiracular (Actinistia), the extratemporal (Onychodontiformes, Grossius), or it is artfully left unlabelled.  See, e.g., our images of Eusthenopteron and Styloichthys.  It does seem to be associated with the spiracular groove; but such an ossification is also present in, for example, Cladistia.  Thus, we aren't quite sure what to make of it.  

In short, most bones of the actinistian dermal skull either lack homology, or have very uncertain homology, with the dermal skull bones of the groups to which they have historically been compared.  In that case, one naming convention is probably as good as another.  The alternatives are to give these bones unique names (postrostralfenestroid, Clément's Scrap of Rostral Bone, Bubba), or names indicating doubt (squamosalish, Forey's fake frontal, operculothingy).  While these approaches would provide us with a unique opportunity to add to the world's bulging arsenal of anatomical nomenclature, they are unlikely to attract a meaningful number of adherents.  Thus, we elect the path of least resistance and follow Forey (1998).  ATW060109

[1] We almost never discuss this subject (a) because, frankly, it is one we don't understand well enough and (b) because it seems to have been at the root of a substantial number of phylogenetic errors.  The developmental connection between sensory lines and dermal skull bones is a deep one.  In fishes, the dermal bones very frequently originate as tubular structures around the sensory lines during embryonic development.   Yet, somehow, the same bones generally appear when the sensory line is changed, moved, or eliminated in the course of evolution.  

checked ATW060320