Surprised to find this:
Acheloma (Cope 1882; Dilkes and Reisz 1987; Early Permian, 275 mya; aka Trematops), a trematopsid amphibamid lepospondyl basal tetrapod had a confluent antorbital fenestra and naris. Bolt 1974 considered this a “very elongate external naris” and then considered two hypotheses for its origin and use:
- as a nasal salt gland (rather improbable, but still possible, according to Bolt)
- to transfer of forces away from the antorbital bar (Bolt’s preferred hypothesis)
Bolt also noted
that earlier papers referred this morphology to a confluent antorbital vacuity, but dismissed the notion by saying, “There is no evidence that any labyrinthodont, including the ancestors of trematopsids, possessed such an [completely separate] antorbital vacuity.” IMHO, this convergent trait need not have been completely separate to qualify as an antorbital vacuity/fenestra. As Bolt noted, in nearly every case, there is a slight constriction in this vacuity marking the end of the naris and the beginning of the antorbital vacuity (Fig. 1). A nasal flange descends inside the vacuity.
Earlier
we looked at the antorbital fenestra in other tetrapods here.

Figure 1. Acheloma dunni skull with a confluent antorbital fenestra and naris. Scale bar = 5 cm.
Perhaps of interest to this discussion
is the relatively large diameter palatal teeth on the vomers, palatines and ectopterygoids (Fig. 1). Bolt also found evidence for a nasal flange in related Doleserpeton and Tersomius, but not in unrelated Seymouria and Eryops.
Olson 1941 had this odd explanation:
The anterior part was for smelling, the longer posterior part was for respiration and the reason for this was the internal naris lies beneath only the posterior part. Bolt noted the shortest route was not always the only route in tetrapods. Air passages can be quite complicated.
The odd otic notch
that likely housed an eardrum in related taxa, is long and narrow in Acheloma. Dilkes and Reisz (1987) noted, “The shape of the otic notch, however, argues against an impedence-matching hearing system because the vibrational properties of the postulated tympanum would be profoundly different from one with the same surface area but circular in outline.”
Acheloma cumminsi was originally considered a temnospondyl, but here nests between Dendrerpeton and Cacops within the lepospondyls with many traits convergent with temnospondyls, like that large wide skull and large overall size. The related Acheloma dunni (Fig. 1) had giant palatal teeth.
As promised earlier:
lepospondyl traits of Acheloma and Cacops not present in temnospondyls from the character list of the LRT. Let me know if you see errors here:
- Ventral naris chiefly maxilla in lateral view
- Prefrontal separate from postfrontal
- Preorbital length of skull sub-equal to postorbital length of skull
- Naris shape in lateral view < 2x longer than tall
- Palatine exposure on the external skull below orbit.
- Squamosal posterior rim is a ‘big curve’
- Squamosal descends to ventral skull
- Mandible tip straight, does not rise
- Cervical centrum longer than tall
- Cervical neural spines not taller than centra
- Pleurocentra larger than intercentra
- Two sacral vertebrae
- Sacral spines not > acetabulum depth
- Anterior chevron shapes, not wider proximally
- Anterior caudal neural spines not higher than centra
- Clavicle shorter than scapula
- Humerus not ‘L’-shaped
- Manual metacarpals 1-3 align
- Longest metacarpals: 2, 3 and sometimes 4
- Longest manual digit: three and four
- Manual unguals sharp pointed
- Metacarpal 5 absent – except in Cacops. Acheloma has 5 carpals.
- Posterior ilium not longer than anterior ilium
- Pubic apron wide
- Longest metatarsals: 3 and 4
- Pedal 3.1 not > p2.1
- Overall size not > 60 cm in length
Shifting
Acheloma, Broilleus and Cacops to Eryops adds 24 steps at present. Shifting those three + Dendrepeton and Tersomius adds 17 steps at present. Shfting those five + the three members of the Amphibamus clade adds 35 steps at present.
On a side note:
Having a fifth finger on basal tetrapods (no matter how you count them, 1-4 or 2-5) is rare after Acanthostega partly because a complete manus is rare in basal tetrapods and partly because many taxa have only four fingers. Proterogyrinus, Seymouria, Cacops and basal reptiles all have five fingers preserved. Presently that’s a discontinuous list, but those five fingers could be homologous. If you know of any other related examples, let me know. I need that data.
References
Bolt JR 1974. Osteology, function, and evolution of the trematopsid (Amphibia: Labyrinthodontia) nasal region. Fieldiana: Geology 33(2): 11-30.
Cope ED 1882. Third contribution to the history of the vertebrata ofthe Permian Formation of Texas. Proc. Phil. Soc., 20: 447-461.
Dilkes DW and Reisz R 1987. Trematops milleri identified as a junior synonym of Acheloma cumminsi with a revision of the genus. American Museum Novitates 2902.
Olson EC 1941. The family Trematopidae. Journal of Geology 49:149-176.
That’s all true, but I’m not sure if an antorbital fenestra as in archosauriforms or (quite differently) in chroniosuchids should be scored as present. If you can get a hold of Schoch’s 2012 paper on dissorophids, you’ll find some with a transitional morphology: the naris is enlarged to only about twice its normal size, and yet the enlargement is set off by a constriction.
About the function of this phenomenon I can only speculate. It’s practically obvious, though, that the living animal’s nostril was where it normally is: at the rostroventral end of the bony naris. The caudal extension must have housed something novel. A salt gland would make sense, given the fact that adult trematopids were terrestrial.
There are things I can’t look up at half past 1 am on a Sunday. If you have further questions, I’ll look them up later. For now:
Artefact of counting the “antorbital fenestra” part of the naris. Also artefact of size: small animals have proportionally larger nostrils (and eyes and pineal organs and ear capsules and spiracles and and and).
Come on. There’s not a single temnospondyl where the prefrontal and the postfrontal have fused, to each other or actually to anything else (as long as I don’t speculate about the intertemporal).
If that’s not how you’ve scored your matrix, the trees you get are hollow inside.
In fact, there’s scarcely any reason to think any lepospondyl had this strange condition either… which is why it’s never been suggested in the literature.
You’re counting the extension again, I think.
Um. Lepospondyls don’t have lateral exposures of the palatine. Ever.
Never found in lepospondyls unless Triadobatrachus is one. Admittedly found in Solenodonsaurus, separately in the diadectids, and in cute baby seymouriamorphs but not in adults.
How close to the ventral edge? Anyway, take a look at your own fig. 1: the squamosal of Acheloma is unusually far away from the ventral edge of the skull.
Are you sure tons of temnospondyls don’t also have that? Certainly Trimerorhachis does.
A very widespread condition.
That’s indeed rare among non-dissorophoid temnospondyls.
Two? Are you sure? Few lepospondyls have that even, it’s mostly an amniote thing.
Size- and or terrestriality-related?
Not sufficiently independent from cervical and sacral neural spine height.
Terrestrial adaptation.
How do you define L-shaped? The way everyone else defines it, it’s a condition found in Tulerpeton, colosteids and some but not all anthracosaurs, and in no temnospondyls whatsoever.
…Which others would ever be longest? 1 or 5???
Which others would ever be longest???
I don’t think pointed unguals ever occur outside Amniota. Even Tuditanus and Casineria don’t have this condition. (A photo I took of manual ungual IV-5 of Casineria is in the preprint.)
…Are you trying to say Cacops has five metacarpals??? No known temnospondyl has five metacarpals, unless Casineria is one.
Which others would ever?
This, too, is actually correlated with terrestriality. That holds for extant tetrapods as well: aquatic ones are systematically larger than their closest terrestrial relatives.
Dave, thanks for your thorough report. Let’s remember that what you think are temnospondyls (Cacops and kin) nest with Lepospondyls in the LRT.
re: Cacops having five metacarpals, take a look at my data: http://www.reptileevolution.com/cacops.htm. I’m hoping that the old illustration I used for data is in error as no basal taxa have five fingers. However, Acheloma has five distal carpals, according to the data here: http://reptileevolution.com/acheloma.htm
Finger count is puzzling. I’ll lay out the pattern in an upcoming blogpost.
Antorbital fenestra or extended naris? The circumnarial bones are the same either way. Given the pinch at mid length or thereabouts, I opted for the former. Either way it’s not going to change the topology.
Prefrontal/postfrontal fusion: You are correct. That was scored that way only once for Brachydectes. A reexamination revealed my mistake and that has been rescored.
I realize that the topology I present is not the topology you recovered. I’ll try to get back to the logic behind this topology in a future blog now that several pertinent taxa from your tree have now been added to the LRT and several other taxa have been rescored as my experience with basal tetrapods grows.
It’s a reconstruction drawing, not a specimen drawing. It’s over a hundred years old, so people assumed the living animal would as likely as not have had five fingers and reconstructed five fingers.
Try it.
That’s not logic. It’s redundant characters, misinterpretations of the literature, and sheer lack of knowledge about anatomy and ontogeny.