Re: Caenagnathus species (even longer!!)

[TWILLIAMS@canr1.cag.uconn.edu]

I pulled the following summary of the Caenagnathus/Chirostenotes/Elmisaurus 
confusion from an old file of mine (which I just updated):  


Tim




The similarities between Chirostenotes and Elmisaurus prompted Osmolska 
(1981) to a erect a new family of theropods, the Elmisauridae, to 
accommodate these two genera.  As described by Sues (1997), the specimen ROM 
43250 provided compelling evidence for synonymising Caenagnathus with 
Chirostenotes, and Elmisauridae with Caenagnathidae.  However, packaging the 
caenagnathid (including "elmisaurid") material into individual genera has 
proven to be difficult.

Currie and Russell (1988) originally pooled all the "elmisaurid" material 
then described from North America into C. pergracilis, and sorted the 
material into two distinct morphs based on the morphology of the manus and 
pes: a "robust" morph (CMN 2367 [type of Chirostenotes pergracils], CMN 8538 
[type of Macrophalangia canadensis]) and a "slender" morph (ROM 781 [type of 
Ornithomimus elegans], RTMP 79.20.1).  Currie and Russell suggested that the 
two morphs may represent the two sexes of a single dimorphic species.  They 
also speculated that the mandibles named Caenagnathus collinsi and 
Caenagnathus sternbergi (which also differ in their relative slenderness) 
may also be referrable to these robust and slender morphs of C. pergracilis, 
respectively.  

Currie (1989, 1990) made ROM 781 the holotype of his new combination, 
Elmisaurus elegans.  This specimen, an incomplete hindlimb, had at first 
been identified as an ornithomimid, which Parks named Ornithomimus elegans.  
Russell (1972) referred the material to Macrophalangia.  Currie (1989) noted 
similarities between ROM and theropod hindlimb bones from Mongolia named 
Elmisaurus rarus.  These similarities included the degree of coossification 
of the tarsometatarsus (complete fusion of metatarsals II-IV with adjoing 
tarsals III and IV), a prominent proximolateral process on tarsal IV, and 
the proximal part of metatarsal III being triangular in section (not 
diamond-shaped, as in C. pergracilis).  Currie (1989) consequently removed 
ROM 781 from C. pergracilis and made it the holotype of a second species of 
Elmisaurus, E. elegans.  Currie referred to this species two hitherto 
undescribed tarsometatarsi from Alberta which also showed this same pattern 
of coossification (RTMP 82.39.4, ROM 37163).  

Sues did not include the degree of tarsometatarsal fusion as a defining 
character for Chirostenotes (and its included species), but instead employed 
much the same criteria that Currie and Russell (1988) had earlier used to 
distinguish two separate morphs of C. pergracilis.  In light of ROM 43250, 
Sues (1997) could confidently refer CMN 8776 to C. pergracilis, thereby 
sinking Caenagnathus collinsi as a junior synonym of C. pergracilis.  Sues 
(1997) returned ROM 781 to the genus Chirostenotes as the holotype of a 
second species, C. elegans.  Sues also assumed that C. elegans and 
Caenagnathus sternbergi represented the same species.

According to Sues (1997), C. pergracilis can be distinguished from C. 
elegans by its more elongate and shallow dentary; a proportionately longer 
mandibular symphysis; and the presence of a median ridge on the dorsal edge 
of this symphysis.  C. pergracilis also appears to be a larger and (judging 
by the construction of the hand and foot bones) less gracile species than C. 
elegans.  

As the holotype of Caenagnathus collinsi had been referred by Sues to C. 
pergracilis, Sues proposed that CMN 2690 ([type of Caenagnathus sternbergi] 
which shows marked differences from the holotype of Caenagnathus collinsi) 
belongs to C. elegans.  As defined by Sues (1997), C. elegans is a smaller 
and more delicate species than C. pergracilis, with a shorter and deeper 
lower jaw.  However, any diagnosis of C. elegans that includes cranial 
characters must be considered very tentative.  Unlike C. pergracilis, there 
is no direct evidence that CMN 2690, or any other jaw material referred to 
Ca. sternbergi, belongs to C. elegans.  Sues (1997) refers CMN 2690 to C. 
pergracilis on the presumption that only two caenagnathid/elmisaurid species 
are represented in Alberta, and his diagnoses of C. elegans and C. 
pergracilis are incompatible with the characters Currie (1989) used to 
diagnose E. elegans and C. pergracilis. 

Basically, there are two sets of criteria for separating caenagnathid 
species.  The first method divides the material based largely upon the 
degree of coossification of the tarsometarsus.  This was the method employed 
by Osmolska (1981) and Currie (1989, 1990).  Under this scenario, Elmisaurus 
has a coossified tarsometatarsus, Chirostenotes does not.  (Elmisaurus can 
also be distinguished from Chirostenotes by additional characters outlined 
above.)  Based on these criteria, Elmisaurus is found in both North America 
and Mongolia.  Currie (1989, 1990) recognised three, two from North America 
(C. pergracilis, E. elegans) and one from Mongolia (E. rarus).

The second method divides the material on the basis of the construction of 
the limb elements - gracile versus robust.  This was the method used by 
Currie and Russell (1988) and Sues (1997).  The latter author provided 
cranial characters for the diagnosis of his gracile and robust species, 
courtesy of ROM 43250 and the Caenagnathus jaw material.  Sues recognised 
two North American species: C. pergracilis and C. elegans.  Sues doubted 
that the tarsometatarsal characters observed by Osmolska (1981) and Currie 
(1989, 1990) were of any diagnostic value, and regarded Elmisaurus as a 
suspect genus.