Lungfish fingers and bird fingers

[David Marjanovic <david.marjanovic@gmx.at>]

An interesting paper came out last year in a journal that probably not many 
paleontologists read:

Zerina Johanson, Jean Joss, Catherine A. Boisvert, Rolf Ericsson, Margareta 
Sutija & Per E. Ahlberg (2007): Fish Fingers: Digit Homologues in 
Sarcopterygian Fish Fins, Journal of Experimental Biology (Mol Dev Evol) 
308B, 757 -- 768 (published online 11 September 2007)

Abstract:
A defining feature of tetrapod evolutionary origins is the transition from 
fish fins to tetrapod limbs. A major change during this transition is the 
appearance of the autopod (hands, feet), which comprises two distinct 
regions, the wrist/ankle and the digits. When the autopod first appeared in 
Late Devonian fossil tetrapods, it was incomplete: digits evolved before the 
full complement of wrist/ankle bones. Early tetrapod wrists/ankles, 
including those with a full complement of bones, also show a sharp pattern 
discontinuity between proximal elements and distal elements. This suggests 
the presence of a discontinuity in the proximal-distal sequence of 
development. Such a discontinuity occurs in living urodeles, where digits 
form before completion of the wrist/ankle, implying developmental 
independence of the digits from wrist/ankle elements. We have observed 
comparable independent development of pectoral fin radials in the lungfish 
*Neoceratodus* (Osteichthyes: Sarcopterygii), relative to homologues of the 
tetrapod limb and proximal wrist elements in the main fin axis. Moreover, in 
the *Neoceratodus* fin, expression of Hoxd13 closely matches late expression 
patterns observed in the tetrapod autopod. This evidence suggests that 
*Neoceratodus* fin radials and tetrapod digits may be patterned by shared 
mechanisms distinct from those patterning the proximal fin/limb elements, 
and in that sense are homologous. The presence of independently developing 
radials in the distal part of the pectoral (and pelvic) fin may be a general 
feature of the Sarcopterygii.

The core of the paper is development genetics of the pectoral fins of the 
Australian lungfish.

One conclusion is that three parts can be distinguished in the 
cartilaginous/endochondral skeleton of the pectoral extremities of any 
sarcopterygian (the dermal skeleton being fin rays and scales):
- the (metapterygial) axis: humerus, ulna, ulnare, and 0 to at least 11 more 
elements (0 in rhizodonts and tetrapods, 1 in *Latimeria* and 
*Eusthenopteron*, 2 in *Tiktaalik*, 11 in the figured specimen of 
*Neoceratodus*). This forms as a continuous spike of prechondral 
(pre-cartilage) cells in *Neoceratodus* that segments itself while it grows. 
In *Neoceratodus* (and, judging from morphology, other lungfishes and 
porolepiforms) it reaches all the way to the tip of the fin.
- 1 to 3 "nonindependent preaxial radials" (perhaps homologous to part of 
the mesopterygium of non-sarcopterygian gnathostomes, but that's just my 
guess): one articulating with the humerus and consisting of the radius, 
radiale, and presumably centrale 1, distal precarpal, and prepollex; one 
(absent in *Neoceratodus*) articulating with the ulna and consisting of the 
intermedium (and presumably centralia 4 and 2 & 3); and one that articulates 
with the ulnare and is only present in *Eusthenopteron* among the taxa 
Johanson et al. mention (unless it is centrale 3, but that's unlikely and 
was just my guess in the first place). One might think that the axis simply 
branches, so that radius and ulna are formed at the same time, but that's 
not how it works in *Neoceratodus*: the radius appears, without a connection 
to humerus or ulna, after the humerus, ulna, and ulnare have already started 
to form. The only "nonindependent preaxial radial" *Neoceratodus* has 
(radius and so on) forms, like the axis, as a continuous spike that segments 
while it grows.
- the "independently developing radials" which are found on both sides of 
the axis in taxa with a long axis (lungfishes, porolepiforms, to a lesser 
extent *Tiktaalik*) and mostly distal to the axis in taxa with a short axis 
(coelacanths, rhizodonts, *Eusthenopteron*, tetrapods). They are very 
variable (between taxa) in number, size, shape, and arrangement and do not 
have a consistent relationship with elements of the axis of the 
nonindependent preaxial radials. In *Neoceratodus* they start forming after 
the axis has grown to at least 6 elements beyond the ulnare, and, just 
before they appear, Hoxd13 is expressed, which is exactly what happens just 
before the appearance of digits in tetrapods; it is not expressed in the 
axis. After appearing, they grow and segment (like the axis and the 
nonindependent preaxial radials), and then they form joints with the nearest 
elements of the axis.

Before this, in the lengthy introduction, Johanson et al. compare the ankles 
of *Acanthostega*, *Ichthyostega*, and the Carboniferous anthracosaur 
*Proterogyrinus* (which already has the usual 5 toes). Attention is directed 
to the fact that digits + metapodials seem to have evolved before the 
centralia were complete: *Acanthostega* has 3 metatarsals articulating 
directly with the fibulare; it has an intermedium, a tibiale, and distal 
tarsals for 4 or 5 of the remaining 5 metatarsals, and a hole in between 
that may have held 1 or 2 unossified centralia, but not 4. *Ichthyostega* 
has a completely ossified ankle with a single centrale (and 2 metatarsals 
articulating directly with the fibulare, while the remaining 5 metatarsals 
have distal tarsals). *Proterogyrinus* has 3 small centralia and no 
metatarsals articulating directly with the fibulare. They also point out 
that in salamanders "digits develop as individual buds, before the 
condensation of the mesopodium [carpus/tarsus]. For example, in 
*Desmognathus aenus*, metacarpal II is the first element of the autopodium 
[metapodials + phalanges] to condense and chondrify [become cartilaginous] 
and is well separated from the only other limb condensations [that exist at 
that time], the radius and ulna" (p. 759); in *Ambystoma mexicanum*, 
metacarpal I and digit I appear first. Johanson et al. propose that -- see 
above -- the digital rays (digits, metapodials, distal carpals/tarsals) are, 
in ontogeny and phylogeny, independent of the more proximal elements, that 
the distal carpals/tarsals are (in salamanders and stem-tetrapods) proximal 
outgrowths of the metapodials, and that the centralia are evolutionarily 
more recent outgrowths of the intermedium and radiale/tibiale.

In amniotes and frogs, the limb chondrifies strictly from proximal to 
distal, while in salamanders this is not the case (see above). The 
traditional assumption is therefore that the salamander condition is 
derived. Johanson et al. turn this on its head, based on the evidence from 
the development of lungfish fins and from the morphology of tetrapods in 
general and *Acanthostega* and *Ichthyostega* in particular: the salamander 
condition is plesiomorphic, and the frog + amniote condition is (twice 
independently) derived. I wonder if that happened because most frogs and 
amniotes use their limbs for terrestrial locomotion soon after the limbs are 
formed, so there's no point in regulating a complicated sequence of 
chondrification, while salamander larvae grow limbs early and leave the 
water (if at all) late...
       From pp. 765f.: "The earliest phase of tetrapod evolution shows a 
gradual elaboration of the mesopodium after the establishment of the digital 
arch [...]. Furthermore, this elaboration involves the addition of new 
centrals and distal carpals/tarsals, elements that lie immediately on either 
side of the transverse 'discontinuity line' [...] across which it is 
impossible to trace one-to-one relationships between elements. This suggests 
that the elaboration may have comprised two mirror-image processes of 
terminal addition of new elements, one at the distal end of the limb proper, 
producing centrals, and one at the proximal ends of the digits [or rather 
metapodials], producing distal carpals/tarsals [...]. The fossil evidence 
thus implies a degree of digit independence in early tetrapods. We argue 
that the early establishment of digits during urodele limb development is a 
retained primitive character that has been lost independently in anurans and 
amniotes. We note, however, that Cohn et al. [...] questioned the occurrence 
of branching mechanisms in the more distal elements of the amniote limb 
skeleton. A piece of foil placed within a developing chick wing directly in 
front of the most recently condensing element failed to prevent more distal 
elements from forming, suggesting some degree of independence of the latter 
from the former. Independence of the digits may thus be a general feature, 
masked by the proximal-to-distal chondrification sequence within the amniote 
paddle." (In amniotes and frogs the limb bud assumes a paddle shape while 
the hand/foot forms; that's not the case in salamanders.)
       Of course, as Johanson et al. point out, this moreover fits the 
well-known fact that the digits ossify before the carpus/tarsus in a wide 
variety of tetrapods, including many amniotes as well as fossil 
temnospondyls, seymouriamorphs and lepospondyls.
       Johanson et al. go on to cite Fröbisch et al. (2007), who report 
that the digits of the temnospondyl *Apateon* ossified in the salamander 
sequence: first 1 and 2, then 3 and 4, and then (in the foot) 5. This was 
originally taken as evidence of close relationship; it may be plesiomorphic 
instead.
       "These observations imply that the autopod [digits + metapodials] 
evolved before the origin of tetrapods, represented by the more distal 
region of the sarcopterygian fin, with the 'origin of digits' simply 
representing a modest repatterning of this region rather than the origin of 
a new structure." (p. 766) The authors also mention the correlation between 
late-phase Hoxd13 expression and "radial formation" in the basal 
actinopterygian *Polyodon* (Davis et al. 2007); that would mean that, rather 
than us gaining digits, the teleosts have lost them.

In their last paragraph Johanson et al. discuss the "primary axis" 
hypothesis, whose latest incarnation is Wagner & Larsson (2007). This 
hypothesis says that digit IV is the tip of the (metapterygial) axis which 
has assumed digit shape because it has grown into digit territory, and that 
it is repeated on both sides to initiate the other digits. Wagner & Larsson 
go so far as to suggest that digits I and/or II of salamanders, which (see 
above) appear first, are in fact digit IV, with the lower-numbered digits 
lost and the higher-numbered ones lacking homologues in frogs and 
amniotes -- a frameshift much more massive than the one they suggest for 
birds. Without mentioning this extreme, Johanson et al. disagree, "because 
it does not seem to fit well with the palaeontological or molecular data" 
(p. 767) explained above.

To recapitulate, there are as of 2007 three hypotheses on the origin of 
digits, and Johanson et al. consider the first two more or less falsified:
- The digital arch: the metapterygial axis runs (forelimb example) 
humerus-ulna-ulnare-distal carpal 4-dc3-dc2-dc1. The radius-radiale-who 
knows what else branches off preaxially, and so does the intermedium, which 
(if I understand it correctly) continues into centralia and digits I and II 
of salamanders; dcV is a postaxial branch, and so are all metacarpals + 
digits. Illustrated here 
http://www.press.uchicago.edu/books/gee/shubin3.jpeg (ignore *Panderichthys* 
and the gene expression patterns).
- The primary axis: the metapterygial axis runs 
humerus-ulna-ulnare-dc4-mcIV-digit IV; its distal part is repeated to form 
the other digits.
- Digits as "radials": the metapterygial axis runs humerus-ulna-ulnare and 
then stops in tetrapods and rhizodonts, while it continues in other 
sarcopterygians with elements that lack tetrapod homologs; the 
radius-radiale-unspecified centralia and the intermedium-unspecified 
centralia form independently; and so do the "radials".

Shubin (2002) discusses three modes of reducing the number of elements in 
the hands and feet of salamanders:
- Truncation by paedomorphosis: elements are lost because they never form. 
This applies to the absence of digits V (foot only), V + IV, or V + IV + 
III, and probably to the rare appearance of centrale 3 (traditionally called 
"mediale" in salamanders).
- "Fusion" by paedomorphosis: elements that form from a single cluster of 
prechondral cells never separate. This applies to the "fusion" of distal 
tarsals 4 and 5 (commonly seen in salamanders as inter- and intraspecific 
variation), the "fusion" of distal tarsals 3 + 4 + 5 (same, but rarer), to 
the rare separation of centrale 1 (traditionally called "element y" in 
salamanders) and the distal tarsal of the prehallux, and to the (outside of 
hynobiids) rare separation of the intermedium and centrale 4. One cannot 
help but suspect that this also accounts for the "basale commune", the fact 
that distal carpals/tarsals 1 and 2 are not known to ever separate in any 
salamander. "The phylogenetic outcome of this is that patterns of 
consolidation of elements involve neighboring elements that lie within one 
of the three series of the mesopodium: the preaxial column, the central 
series, and the digital arch. Condensations in these zones largely remain 
distinct from one another and rarely fuse" (p. 25). In the terms of Johanson 
et al., these are the first "nonindependent preaxial radial" (radius, 
radiale, etc.), the second "nonindependent preaxial radial" (intermedium 
etc.), and the "radials" -- with the only exception that Johanson et al. do 
not postulate a connection between the distal carpals/tarsals, but adding 
one does not seem to contradict the rest of their model.
- Actual fusion: elementsfirst form and then fuse. Apparently invariably, 
they are from different "series". Centrale 2 and dc4 apparently fuse fairly 
often, intermedium and ulnare rarely, and that basically was it.

In sum, more support for the model of Johanson et al..

And now I have finally waffled enough about lungfish and salamanders and can 
go on to the birds: to the implications of that model on the frameshift 
hypothesis (and, of course, at the same time on the BAND hypothesis).

If digit IV is not part of the metapterygial axis, we cannot safely use the 
spatial relations between any digits on the one hand and the radius, ulna, 
or any proximal or central carpals on the other hand as evidence for the 
identity of any digit. Nor can we safely assume that digit IV must be the 
first to chondrify; in amniotes that retain 5 digits as adults, it is in all 
of the few cases that have been checked, but in salamanders it isn't. In the 
absence of more convincing evidence, the most parsimonious option is of 
course that the first digit to chondrify in the bird wing is digit IV, but 
we can no longer assume that this alone, and/or its position vaguely 
opposite the ulna, is stronger evidence than morphology.

I conclude that the only evidence for the frameshift hypothesis so far is, 
surprisingly, fossil. It was presented at this year's SVP meeting; I'll post 
the abstract later (this e-mail is long enough, and I should go to bed for a 
change). But, while suggestive, this, too, is not a clincher (I think I'm 
allowed to mention that I talked to the presenter, Jim Clarke, and we agree 
it's all very confusing). If the digits really are independent of the 
metapterygial axis and the "primary axis" hypothesis is wrong, the 
frameshift hypothesis may simply be unnecessary, for birds and for 
salamanders, and we can stop feeling guilty about calling the fingers of 
birds I, II and III.

Ironically, however, Johanson et al. confirm that an axis can grow into 
"radial" territory and have its tip modified into the shape of a "radial": 
as mentioned way above, the most proximal preaxial radial of *Neoceratodus* 
really does form by segmentation of the radius-radiale-etc. axis (in other 
words, *Neoceratodus* has a prepollex...), unlike all other "radials". It's 
just not the metapterygial axis, but the first "nonindependent preaxial 
radial".

Unsurprisingly, questions remain. Shubin (2002) casually mentions 
(especially in fig. 5) that plenty of (at least cryptobranchoid) salamanders 
have an extra distal tarsal called the postminimus. What is that? More 
evidence that the digits don't care about the rest of the limb when deciding 
where to form? Or is it the fourth element of the metapterygial axis, 
homologous to the last one in *Latimeria* and *Eusthenopteron*? Shubin 
reproduces the only drawing of one of about two decently known fully 
ossified temnospondyl feet, that of *Acheloma* (...he uses the junior 
synonym *Trematops*). Dt 4 is very large, mt V articulates _between_ dt4 and 
dt5, and dt5 is a rounded isosceles triangle, with the tip lying postaxial 
to mtV. Isn't that strange? Both dt4 and dt5 lie distal to the fibulare. 
Perhaps "dt4" is actually the "fusion" product of dt4 and dt5, and "dt5" is 
actually the postminimus? And what is the pisiform (an extra proximal carpal 
widespread in mammals, but also captorhinids, diadectomorphs, and many 
others) -- an ancient sesamoid, the postminimus, or what? And what about the 
toes in *Acanthostega* and *Ichthyostega* that articulate directly with the 
fibulare? Not to mention the finger of *Tulerpeton* that articulates 
directly with the ulna -- not the ulnare, but the ulna? (Illustrated here 
http://www.press.uchicago.edu/books/gee/shubin3.jpeg and 
http://www.nature.com/nature/journal/v440/n7085/fig_tab/nature04637_F4.html 
here.) To what extent are digits of Devonian tetrapods homologous to digits 
of pentadactyl tetrapods or to those of other Devonian tetrapods -- is it 
all just an unspecified mess of "radials" that grew wherever there was space 
for them? And to what extent are _bird_ fingers homologous to anything, 
then...?

All that makes me tired. It's a quarter to three at night. I'm going to bed.

==== References ====

Davis C, Dahn RD, Shubin NH. 2007. An autopodial-like pattern of _Hox_ 
expression in the fins of a basal actinopterygian fish. Nature 447: 437 --  
476.
Fröbisch NB, Carroll RC, Schoch RR. 2007. Limb ossification in the Paleozoic 
branchiosaurid *Apateon* (Temnospondyli) and the early evolution of preaxial 
dominance in tetrapod limb development. Evolution & Development 9: 69 -- 75.
Shubin NH. 2002. Origin of evolutionary novelty: examples from limbs. 
Journal of Morphology 252: 15 -- 28.