_Compsognathus_ minutiae

["Nicholas J. Pharris" <pharrinj@PLU.edu>]

OK.  Here goes.  I finally have the _Dinosauria_ blurb on _Compsognathus_ 
here in front of me, along with a good photo of the original specimen 
(_The Ultimate Dinosaur Book_, by David Lambert, page 81).

After examining the material at hand, I have come to three conclusions:

1)  that the traditional, often-reproduced reconstruction of the pes of 
_Compsognathus_ (after Ostrom, 1978) is seriously in error;

2)  that the same is true of Ostrom's reconstruction of the manus; and

3)  that _Compsognathus_ is apparently an arctometatarsalian.


_Compsognathus_'s Feet:

Ostrom's reconstruction of the metatarsus of _Compsognathus longipes_ 
shows metatarsals II-IV of about equal width and length, with parallel 
borders along their entire legths.  It almost looks like Ostrom drew the 
lower two-thirds or so of the _Compsognathus_ pes and capped it off with 
the upper third of the pes of _Rioarribasaurus_.

In fact, according to the photos I have, the proximal portions of 
metatarsals II and IV are markedly flared, contacting each other and 
covering the proximal end of metatarsal III.  The three metatarsals are 
also much more closely appressed to one another than indicated by 
Ostrom.  There are no gaps visible between any of the metatarsals, unlike 
the situation in the reconstructed pes.


_Compsognathus_'s Hand:

The manus as reconstructed by Ostrom has some serious defects.  According 
to the reconstruction, metacarpal I is significantly longer and more 
robust than metacarpal II.  This is not the case in any theropod with a 
well-preserved manus.  

Ostrom has also reconstructed the animal's thumb 
with metacarpal I about twice the length of phalanx I-1.  This is *never* 
the case; in fact the situation is generally closer to the reverse.
 
Finally, digit II is reconstructed with only two phalanges, not the case 
in any theropod (in fact in hardly any tetrapod) of which I am aware.  In 
fact, Ostrom's interpretation of two phalanges per finger cannot possibly 
be correct, as there are at least *5* discernible non-ungual phalanges 
preserved on the slab, along with 4 claws.

The left manus is better preserved than the right and clearly shows three 
metacarpals:  one fairly slender and straight, another rather more 
robust and flared at either end, and the third very slender and bowed.  
The most robust is most likely metacarpal II, the smaller, straight bone 
metatarsal I, and the bowed bone metacarpal III.  The positions of the 
bones are also more in line with this interpretation than with Ostrom's, 
particularly when it is considered that it is the dorsal surface of the 
left manus that we are seeing.  The manus appears to indeed be didactyl.

All the disarticulated phalangial bones in the specimen appear to have 
drifted ventrocaudally before fossilization.  This being the case, the 
single articulated digit present appears to be associated with metacarpal I 
and looks like a fairly typical, long theropod thumb.

The only phalanx clearly associated with the second metacarpal is the 
very short element (about half the length of metacarpal II) which Ostrom 
interpreted as phalanx I-1.  In view of the above reassignment of the 
metacarpal bones, however, it is phalanx II-1.  Phalanges II-2 and II-3 
of the left hand are apparently not preserved.

Although the right manus is not as well preserved, there appear to be 
three associated non-ungual phalanges.  One is quite long and apparently 
identical to the thumb phalanx of the left manus.  There is also a short 
phalanx present, as in the left hand.  Finally, there is a phalanx with 
its ungual lying across the right metacarpus.  This is slightly shorter 
than the thumb phalanx and bears a slighly smaller claw.  My best guess 
is that this is phalanx II-2 of the right manus and the associated ungual.


A brief, informal revised description of the manus of _Compsognathus 
longipes_ follows:

Metacarpal I approximately 3/5 the length of the radius, straight, and 
tapering slightly distally.  phalanx I-1 subequal to metacarpal I in 
length, straight, and tapering slightly distally.  Ungual I deep but not 
particularly large or curved.

Metacarpal II approximately 3/4 the length of the radius, straight, 
flared proximally and to a lesser extent distally.  Phalanx II-1 
approximately 1/2 the length of metacarpal II and slender.  Phalanx II-2 
approximately 1 1/2 times the length of phalanx II-1 and slender.  Ungual 
II deep but not particularly curved and slightly smaller than ungual I.

Metacarpal III subequal in length to metacarpal I, very slender, and 
bowed.  No phalanges from manual digit III are preserved.

If the above reassessment of the manual material for _Compsognathus 
longipes_ proves correct, then this animal will have the distinction of 
having the longest hands in relation to forearm length of any theropod.  
If I have correctly assigned the phalangeal material correctly, then the 
manus is fully *twice* as long as the radius.  This may be a function of 
age, as "_Jurapteryx recurva_", possibly (possibly not) a juvenile 
_Archaeopteryx_, and juvenile tyrannosaurs also tend to show long manus 
in relation to forearm length.

No carpal material is preserved on the slab, but the great distance 
between the forearm and hand implies to me that the carpus was sizeable 
and not degenerate as in tyrannosaurs.


Conclusion:

The reexamination of the material available to me of _Compsognathus 
longipes_ leads me to believe that this animal is most likely referable 
to the Arctometatarsalia Holtz, 1994.  Several features indicate this to 
be the case:

Metatarsals II and IV in contact over the dorsal surface of metatarsal 
III at the ankle;

metatarsal III very narrow in plantar view but broad in dorsal view, 
indicating that it is wedge-shaped in cross-section;

phalanx II-1 significantly shorter than either metacarpal II or phalanx II-2;

dorsal margin of ilium curved dorsally.

Other features in the scapula, hand, foot, ilium, and ischium also link 
this animal variously with tyrannosaurs, troodonts, and ornithomimosaurs.

If this interpretation proves correct, then _Compsognathus_ is the earliest 
clear member of the Arctometatarsalia.  I do not agree with George 
Olshevsky that this animal is clearly referable to the Tyrannosauria; it 
looks fairly generalized to me and is probably quite close to the common 
arctometatarsalian ancestor.

At the very least, I believe that this indicates that more study is 
necessary on this very interesting little beastie (more technical jargon, 
I know).

As always, questions and comments greatly appreciated.  Tom, are you out 
there? :-)

Nick Pharris
Pacific Lutheran University
Tacoma, WA 98447
(206)535-8204
PharriNJ@PLU.edu

"If you can't convince them, confuse them." -- Harry S Truman