Pterosaur fossils have been known for over 200 years and yet we still argue about which way the fingers flexed. Undisturbed or minimally disturbed fossils give some clues. Fossil handprints do too.
Most pterosaur fossils are found crushed and the finger claws (unguals) are crushed along with the rest of the body (Figure 1). Tall and thin, like the raptorial, tree-clinging claws they were, pterosaur claws were most often preserved broad side down, tips pointing anteriorly, medial faces exposed. And that’s the way they are traditionally reconstructed, paying little attention to the cylinder-shaped interphalangeal joints or evolutionary precedent.
Figure 1. Click to enlarge. Right hand of Shenzhoupterus, dorsal view. Note the crushing of the ungual broadside down and disarticulation of the various finger joints. If you ignored these facts you might want to orient the claws anteriorly as shown.
There are Two Hypotheses at Work Here
In the traditional model (Bennett 2008) metacarpals 1-3 were stacked with #1 on top and all three were bound to metacarpal 4. See Figures 2 and 3. The fingers flexed anteriorly (in flight). Any specimens in which metacarpals 1-3 were found lined up anterior to metacarpal 4, Bennett (2008) ascribed to the effects of gravity on the bones after death.
Figure 2. Pterosaur hands, dorsal view, the two opposing hypotheses. See Figure 3 for anterior views to see how Bennett (2008) intended the fingers to stack with #1 on top. The Bennett configuration orients the fingers facing up when the hands are adducted (brought together), which would be unsuitable for tree climbing/clinging. The Peters configuration, like clapping hands, points the fingertips together, ideal for tree climbing.
In the heretical yet more conservative model (Peters 2002) metacarpals 1-3 lined up side-by-side anterior to metacarpal 4 and the fingers flexed ventrally (as in all other tetrapods). See Figures 2 and 3. In this configuration only metacarpal 4 twisted 90 degrees axially so the palmar side faced posteriorly (in flight) to facilitate wing folding. The three small fingers did not change their configuration or orientation. The palmar side of fingers 1-3 continued to point ventrally in flight.
Peters (2002) reported: “…in certain Cretaceous forms [the medial three digits] rotated into the vertical plane and became closely appressed to the much larger metacarpal IV (Bennett, 1991; 2000b). From this configuration, they flexed anteriorly in a subhorizontal plane when the wing was extended.” I apologize for this. Unfortunately I was influenced by Dr. Bennett at the time. Subsequent studies helped me realize the error of this statement.
Figure 3. Bennett (2008) and Peters (2002) pterosaur finger orientation configurations. See Figure 2 for dorsal views. Note: Bennett wanted digit 1 dorsal as shown here, not as in Figure 2.
The Wellnhofer (1991) Twist
Wellnhofer (1991) lined up the metacarpals anteriorly, but also twisted the unguals anteriorly. Of course, this could be a problem in pterosaurs with fingers of similar lengths and does not take into account the various disarticulations at several finger joints.
The Evolution of the Bennett (2008) Configuration
Bennett (2008) imagined the evolution of the pterosaur hand (Figure 4) based on an imaginary taxon. He started with the supination of the entire arm, which rotated all the palmar surfaces anteriorly. Metacarpal 4 became thicker than the others as it supported a lengthening wing finger. Overlooked by Bennett (2008), but implicit in his arguments, the next step involved migration of the metacarpals 1-3 as a unit down the anterior face of a much larger metacarpal 4. The supination of the hand envisioned by Bennett (2008) ultimately included reversing the flexion and extension of digit 4 such that hyperextension folded the wing in his view. No other tetrapod has ever done this. Also note there is no space for the large wing extensor between the attached metacarpals (1-3 back-to-front with 4) in the Bennett (2008) configuration. Bennett (2008) also envisioned the early disappearance of ungual 5 on the wing, which, due to supination, also faced anteriorly in this configuration. However, the wing ungual was retained. Bennett (2008) also imagined the loss of manual digit 5, but manual digit 5 was retained. He did not envision an origin for the preaxial carpal and pteroid, which occurred as far back as Cosesaurus (Peters 2009).
Figure 4. Pterosaur finger orientation in lateral view, the two hypotheses. There are several problems with the Bennett (2008) hypothesis, least of all it leaves no room for the big wing extensor tendon.
The Evolution of the Peters (2002) Configuration
Peters (2002) discussed the evolution of his pterosaur hand configuration (Figure 4) based on actual taxa, including Longisquama insignis. Between Longisquama and the first pterosaur manual digit 4 was rotated axially so that the palmar (flexor) surface became the new posterior surface to facilitate wing folding in the plane of the wing with hyperflexion. Digit 5, already reduced, became a vestige. It revolved, along with metacarpal 4, to the new dorsal side of the metacarpus. During the rotation, metacarpals 1-3 shifted to the ventral rim of metacarpal 4 while retaining their configuration. This left plenty of room for a large extensor tendon (Figure 4). Contra Bennett (2008), flexion remained flexion in all the fingers. There was no reversal of function. The preaxial carpal and pteroid first appeared in the fenestrasaur and pterosaur precursor, Cosesaurus (Peters 2009) having migrated to the medial wrist from the central carpus where they were identified as the two centralia seen in Sphenodon.
Here’s Where the Trouble Started
In many crushed fossils, like Shenzhoupterus (Figure 1) and the left hand of Eudimorphodon (Figure 5), the claws point anteriorly because they are crushed broadside down. In order to do this, some finger joints must disarticulate and this is always observable.
But look what happens in the right hand of Eudimorphodon
In the same Eudimorphodon (Figure 5) the right hand has the palmar surface exposed. The metacarpals were lifted and flipped over the palmar surface of metacarpal 4, but metacarpal 3 remained attached to metacarpal 4. Moreover the unguals pointed posteromedially. According to Bennett (2008) this should not have been possible if metacarpals 1-3 were bound to the anterior face of metacarpal 4 and pointed anteriorly. Digit 1 should have been buried first and deepest, but it was not.
Figure 5. Eudimorphodon hands. The right hand preserves the metacarpals lined up anteriorly with only metacarpal 3 attached to metacarpal 4. The claws are disarticulated due to crushing. The right hand, preserved with its palmar side exposed shows what happens when the lighter digits 1-3 drift as a unit with metacarpal 3 moving the least because it was attached to metacarpal 4.
Figure 6. Click to enlarge. Santanadactylus hand with metacarpals preserved at a 45 degree angle to the anterior face of metacarpal 4.
Even in 3D Fossils There Can be Some Confusion
Here, in this 3D Santanadactylus hand, the metacarpals have been raised like a drawbridge, far from their original orientation (palmar side down) and close to being pressed against the large metacarpal 4, palmar side anterior. Such a configuration permits no space for the big tendon between the appressed surfaces of metacarpal 4 and the three small metacarpals. This is an excellent example of taphonomic lifting on the hinge at the metacarpal 3-4 interface from an origin with the palmar side down for metacarpals 1-3. It is the only configuration that permits the big extensor tendon of metacarpal 4 to run unimpeded dorsal to the three small metacarpals and their extensor tendons (Figure 4). When the extensor tendon rots or pops, there is nothing to prevent metacarpals 1-3 from rising and falling like an airplane elevator in the drifting sea currents.
Figure 7. The left manus of Pteranodon KUVP 49400 in a rare anterior burial. Here the unguals were crushed in their natural orientation, palmar side down. Thanks to M. Everhart at OceansofKansas.com for this image.
YPM 49400 – a Rare Anterior Burial and Posterior Exposure in Pteranodon
Here in a Pteranodon specimen YPM 49400 metacarpus was preserved anterior face down. This is a very rare burial. The manus was preserved intact and in its natural orientation, fingers 1-3 palmar side down and metacarpal 4 palmar (flexor) side now posterior for wing folding. The claws here pointed ventrally as in other tetrapods. The metacarpals lined up as in other tetrapods. Buried like this there was no chance for them to wave around or become disarticulated in the bottom currents. The proximal wing phalanx would have stood vertically erect (essentially the Z-axis) in this configuration, but it rotted, disarticulated and fell on its dorsum, exposing its ventral face.
Figure 8. Pteraichnus nipponensis (Lee et al. 2009) along with a matching trackmaker, n23 of Wellnhofer 1970.
The Evidence of Ichnites
Pterosaur handprints (Figure 8) commonly preserve digits 1 and 2 laterally and digit 3 posteriorly. Sometimes digit 1 extends anteriorly (Lee et al. 2009). In the Bennett (2008) configuration, all three fingers would have extended posteriorly when quadrupedal because the arms were supinated. In the Peters (2002) configuration, all three fingers would have extended laterally when quadrupedal because the arms were neither supinated nor pronated, but in the neutral position.
Digit 3 Goes the Opposite Way
The key to orienting digit 3 posteriorly (and sometimes digit 1 anteriorly) goes back to the lizard ancestry of pterosaurs. The metacarpophalangeal joint of digit 3 is different than digit 2. Shaped more like a hemisphere than a cylinder, it permits digit orientation in several directions. With digit 3 directed posteriorly, digit 4 never touched the substrate. This evidence is in direct contrast with the configuration envisioned for the difficult to support wing launch hypothesis currently in favor also illustrated here.
On a Side Note
Bennett (2008) connects the extensor and flexors tendons to the proximal tips of the first wing phalanx. This is wrong. In lizards these tendons split in half, bypassing the nearest points and inserting further down the bone. As shown here, this permits complete wing folding, something the Bennett (2008) attachment is unable to do. We just learned about the evolution of the manus hand and pteroid from a Sphenodon-like ancestor here.
Bennett (2008) reported, “The reconstruction of the long extensor and flexor of thewingfinger suggests that there was no rotation of Mc IV about its long axis. If there had been a rotation, the tendon of m. flexor digiti quarti would have had to spiral posteriorly under the metacarpus to insert on the posterior process of wing phalanx 1 and the tendon of m. extensor digiti quarti longus to digit.” As mentioned above, it is a mistake to attach the flexor to the posterior process of wing phalanx 1, as Bennett (2008) proposes. Rather the insertion is further down the phalanx shaft, as in lizards, and in order to complete wing folding. Thus there would have been no problematic spiraling if metacarpal 4 rotated axially.
In Summary
All the evidence points to a configuration in which the hand of pterosaurs was configured the same as in all other tetrapods with the exception that the big wing finger was axially rotated so that the old palmar surface became the new posterior surface. This configuration is based on actual predecessor taxa, not figments of Bennett’s imagination. The new configuration creates a large channel for the massive extensor tendon that is missing from the Bennett configuration. The new configuration does not require the flexor side of the wing finger to become the extensor side and vice versa.
As always, I encourage readers to see specimens, make observations and come to your own conclusions. Test. Test. And test again.
Evidence and support in the form of nexus, pdf and jpeg files will be sent to all who request additional data.
References:
Bennett SC 2008. Morphological evolution of the wing of pterosaurs: myology and function. Zitteliana B28:127-141.
Lee YN, Azuma Y, Lee H-J, Shibata M, Lu J 2009. The first pterosaur trackways from Japan. Cretaceous Research 31, 263–267.
Peters D 2002. A New Model for the Evolution of the Pterosaur Wing – with a twist. – Historical Biology 15: 277–301.
Peters D 2009. A reinterpretation of pteroid articulation in pterosaurs. Journal of Vertebrate Paleontology 29:1327-1330.
Wellnhofer P 1991. The Illustrated Encyclopedia of Pterosaurs. Salamander Books, Limited, London, 192 pp.
The Pterosaur Heresies 
Thank you very much, David, for your comment pointing out the assembly problem in my pterosaur skeleton model, and for such a detailed description in this article.