Re: Now that I have time... genuine new papers, part 1

["Ken Carpenter" <Kcarpenter@dmns.org>]

Uh, Neues Jahrbuch IS peer reviewed. I just got back a manuscript from 
reviewers.
Ken

>>> "David Marjanovic" <david.marjanovic@gmx.at> 04/03/04 14:16 PM >>>
"New" as of December 2003. :-(

First of all, the *Schowalteria* paper, which has been mentioned yesterday.
It contains weirdnesses. These may be due to the apparent fact that Neues
Jahrbuch is not peer-reviewed (e. g. Wolf-Ernst Reif's series on theoretical
aspects of cladistics contains glaring mistakes in its English).

Most of what I've snipped are references.

If you aren't interested in the general shortcomings of the paper, scroll to
the bottom...

p. 394: "Palaeoryctid insectivorans, including the Paleocene *Procerberus*
[...] and, more remotely, the Late Cretaceous *Cimolestes* [...], are often
cited in regards [of] taeniodont origins [...]". Insectivorans? Those
beasties are among the first that were removed from Insectivora, resulting
in its division in Proteutheria (heavily paraphyletic) and Lipotyphla
(apparently diphyletic)! The authors don't mention this problem, nor do they
mention or cite any evidence for insectivoran monophyly. Very confusing.

p. 395:
"T y p e   l o c a l i t y :  Locality KUA-1, [...] Red Deer River Valley,
Alberta [...]. KUA-1 is located in flat-lying continental siltstones
approximately 24 m below the Nevis coal seam as exposed nearby in the same
section. The Nevis seam is a field marker for the Cretaceous-Tertiary (K-T)
boundary in the region (Lerbekmo et al. 1979a, b)."

Though... it's dangerous to use the lowest coal seam as the K-T boundary; it
is below the K-T boundary in some places and above it in others. I don't
know about the Nevis seam, which

"is closely associated with the iridium anomaly nd boundary clay that are
signatures of the end-of-Cretaceous bolide impact, and with palynofloral
changes that are diagnostic of the boundary".

Sounds fair enough -- but what does "closely associated" mean? It isn't
possible that there was faster sedimentation in that place so that the
boundary is 24 or more m below the coal seam, is it???

"The mammalian fossils at KUA-1 come from 'a well-indurated clay and silt
layer which varies laterally from less than 1 foot in thickness to about 3
feet' (LILLEGRAVEN 1969: 12); there is no evidence that this layer is part
of a cut-and-fill channel system from stratigraphically younger (Paleocene)
horizons above."

Sounds good.

"The fossils at KUA-1 are part of the Trochu local fauna (ARCHIBALD 1982:
4), and comprise an important and well-documented Lancian (latest
Cretaceous) mammalian assemblage that includes ptilodontoid and
'taeniolabidoid' multituberculates, deltatheridiids, didelphoid marsupials,
and palaeoryctid and leptictid eutherians [...]."

This just about convinces me that *Schowalteria* is indeed Cretaceous.
Deltatherid-your-favorite-ending-here are not known from the Cenozoic.
Interestingly, they're known both from other (older) sites in North America
and from Mongolia (e. g. *Deltatheridium* itself), but not from the Hell
Creek Fm according to
http://www.dinosauria.com/jdp/misc/hellcreek.html#mammals. This may well
mean that KUA-1 is the only place that shows that these basal metatheres
reached the K-T boundary.

However, I'm willing to bet that the "didelphoid marsupials" are
alphadontine peradectid basal metatherians far away from the crown group
Marsupialia which includes Didelphidae (the opossums).

p. 396:
"Dinosaurs living at the time of deposition of the mammal-containing stratum
at KUA-1 include *Tyrannosaurus* and *Triceratops*. Both are particularly
prominent members of the late pre-extinction communities preserved in the
Scollard Formation, and specimens of each have been found nearby KUA-1, a
horizons clearly above it stratigraphically (LERBEKMO et al. 1979a, b,
LERBEKMO 1985)."

As I don't know of evidence to the contrary, I'll believe this. :-)

"the opossum-like marsupial *Alphadon marshi* [...] (Didelphidae)"

See above. These animals seem to share nothing with opossums except
plesiomorphies. The authors don't mention or cite any evidence that they do;
this implies that they've been sleeping for the last 10 years, which is
quite hard to believe. ~:-|

"the marsupial *Didelphodon vorax* [...] (Stagodontidae)" is a bit closer to
Marsupialia than *Alphadon*... OK, maybe the authors just don't use the
node-based definition of Marsupialia, but they don't mention any
nomenclatural problem. Throughout the text they use "eutherian" and
"marsupial".

p. 408:

"Comparisons of *Schowalteria* with taeniodonts"

No phylogenetic analysis is conducted, but *S.* is found to share 12 derived
character states with taeniodonts. Character polarities are established by
comparison with *Prokennalestes*, *Murtoilestes*, *Montanalestes*,
*Asioryctes*, *Daulestes*, *Kennalestes*, "and *Cimolestes* spp.", for one
character also with *Eomaia*. Now it's easy to demand more taxa in an
analysis, but IMHO it's still worth mention that except for *Cimolestes* the
used taxa are _very_ basal eutherians. It might have been interesting to
compare relatives of *Cimolestes* as well as leptictids. However, all of the
characters that I understand sound convincing, and I couldn't argue that
*S.* were not a taeniodont. So, yes, we do have a Cretaceous taeniodont
here, unless some quite strange things have happened.

On p. 409f, *S.* is found to share 5 derived character states with
Stylinodontidae (the more derived of the two clades) than with Conoryctidae.

p. 410:
"However, RTMP 93.90.01 ( = *S.*) differs in several characters that are
otherwise found throughout the Taeniodonta, including *Onychodectes* [the
basalmost conoryctid], resembling more closely basal Cretaceous eutheians
in these regards" -- this seems to mean that the following 5 character
states argue for *S.* being the basalmost taeniodont at all. Throughout the
rest of the paper, however, the authors firmly regard *S.* as a
stylinodontid, without justifying why they choose one equally parsimonious
alternative over another or even mentioning the problem.

p. 410f:
"Discussion
The discovery of *Schowalteria clemensi* and recognition that it is a basal
stylinodontid [see?] leads to further issues: 1) What are the relationships
of taeniodonts to other mammals? 2) Why have no Late Cretaceous taeniodonts
been found within the geographic range of Paleocene species but only much
further [sic!] north? [600 km north of the northernmost previously known
one, they write earlier.] We briefly consider these issues below.

Relationships of Taeniodonts

Traditionally, the origin of taeniodonts from among Late Cretaceous
eutherians has been sought in the palaeoryctid insectivoran [see above]
*Cimolestes* (see SCHOCH 1986: 172 and references therein) via its presumed
descendant *Procerberus*. [...] In considering possible descendants of
*Procerberus*, LILLEGRAVEN (1969: 69) cited an undescribed species of that
genus from the Puercan [earliest Paleocene] Mantua Lentil (Fort Union
Formation, Wyoming) as a 'likely candidate for the origin of stylinodontine
taeniodonts <sic> .... although the proposed phylogeny is far from certain."

The "<sic>" is a "[sic]" in the original; I wonder what it's doing here...
probably it alludes to the use of "stylinodontine" instead of "-id" and
should therefore have been placed elsewhere.

p. 411:
"Whereas LILLEGRAVEN's (1969) views concerning taeniodont relationships have
received wide approval since (e.g. KIELAN-JAWOROWSKA et al. 1979, MIDDLETON
1983, SCHOCH 1986) [not later? ~:-| ], the most recent and only rigorous
analysis of the relationship of palaeoryctids with Taeniodonta has been by
EBERLE (1999). She described a new 'cimolestid didelphodont' (= [ic]
Palaeoryctidae of LUCAS et al. 1998 and this paper), *Alveugena
carbonensis*, from the middle Puercan [...], and characterized it (p. 936)
as both temporally and morphologically intermediate between the 'presumed
ancestor' *Procerberus* and [the] 'plausible descendant', the basal
taeniodont *Onychodectes* (Family Conoryctidae). Indeed, in discussing the
age of the beds in which *A. carbonensis* occurs, EBERLE (1999: 940) wrote
that this taxon '... is found right where one might expect an evolutionary
intermediate between cimolestids and conoryctid taeniodonts.'"

"Rigorous analysis"? I haven't seen the paper, but "presumed ancestor" and
"plausible descendant" sound somewhat sickening to me... :-S Are they really
all metataxa?

"The occurrence of the stylinodontid *Schowalteria* in the Late Cretaceous
and its possession of dental features that are more primitive than known in
other taeniodonts, including conoryctids, is clearly inconsistent with
EBERLE's (1999) account of taeniodont ancestry."

It is implied here that those features are more plesiomorphic than those
found in *Alveugena*, but to find this out one has to read the next page.

"However, even if a Cretaceous taeniodont had remained undiscovered,
EBERLE's (1999) contention that *Alveugena* is temporally intermediate
between *Procerberus* and *Onychodectes* is unfounded" because it's coeval
with the latter and 2 other taeniodonts (at least one of which is a
stylinodontid).

"The core of EBERLE's (1999) evidence as to the intermediate position of
*Alveugena* between *Procerberus* and taeniodonts was morphological,
however, not temporal. EBERLE (1999) undertook a computer-assisted cladistic
analysis employing 15 dental features of *A. carbonensis*, *O[nychodectes]
tisonensis*, and two species of *Procerberus*, with *Cimolestes* spp. as the
outgroup."

Works only, needless to say, if taeniodonts are a priori assumed to be
cimolestids/palaeoryctids.

p. 412: "The intermediate position of *Alveugena*between *Procerberus* and
*Onychodectes* rests on characters from only the upper dentition, with the
lower dentition of *Alveugena* still to be discovered [...]. Hence, possible
comparisons with RTMP 93.90.01 are relatively few, but the specimen from
KUA-1 nonetheless preserves novel dental information from a
stratigraphically older and structurally more primitive taeniodont than
known before. Fossils having such qualities can be an asset in
distinguishing homologous from homoplastic resemblances among the
purportedly descendant taxa to be compared [...]. It is in this context that
RTMP 93.90.01 brings into question the dental evidence for the transition
from 'cimolestid didelphodonts' to taeniodonts.
    Although *Schowalteria* is a stylinodontid taeniodont [see above], it is
nonetheless more primitive than even the 'cimolestid' *Alveugena* in regards
[of] the latter's 'taeniodont-like' postcanine morphology;"

Then 4 characters are mentioned for which *S.* and "basal Cretaceous
eutherians" share one state and *A.* and Tertiary taeniodonts the other.

"In our view, the primitive postcanine morphology of *Schowalteria* as
established above implies that the more derived, *Onychodectes*-like
postcanine dentition of *A. carbonensis* is only convergent on that of *O.
tisonensis* itself; accordingly, the 'cimolestid' *A. carbonensis* is
neither transitional to Taeniodonta nor is [sic] its sister taxon."

I have to agree with this.

p. 413:
"Our doubts concerning EBERLE's account of the transition to taeniodonts
extend to the species of *Procerberus* and *Cimolestes*. In her cladograms,
these species arise at successively more remote nodes from *A. carbonensis*
and[,] hence, share correspondingly fewer dental specializations with
*Onychodectes* [...]. EBERLE [...] cited '*Cimolestes* spp.' as the outgroup
to *Procerberus* + *Alveugena* + *Onychodectes*, although of the presently
recognized species of *Cimolestes*, only *C. cerberoides* is of interest in
this regard."

Wh?

"Occurring at KUA-1 and hence a contemporary of *Schowalteria clemensi*, *C.
cerberoides* has traditionally been deemed the palaeoryctid phylogenetically
closest to taeniodonts via its links with *Procerberus* (LILLEGRAVEN 1969:
69);"

That's no answer to the question.

"accordingly, we compared RTMP 93.90.01 with [...] *C. cerberoides* [...].
Because of the poor preservation of RTMP 93.90.01, our comparisons were
limited, but nonetheless reveal that several [ = 5] dental features [of
*S.*] having important phylogenetic implications are more primitive (i.e.
more like those in basal Cretaceous eutherians) than their counterparts in
*C. cerberoides* [...]. Hence, although already a stylinodontid [see
above!], *Schowalteria clemensi* is more primitive in key dental features
than the 'cimolestid' that allegedly is at the base of the transition from
'didelphodonts' to all taeniodonts".

Fine -- though still no answer to the question.

p. 413f:
"In contrast to the dental evidence, non-dental evidence supporting
'cimolestid'-taeniodont relationships has never been critically assessed.
For example, although LUCAS et al. (1998) and EBERLE (1999) cited earlier
workers [...] for resemblances in structure of the 'ear region' and tarsus
as evidence [...] favoring 'cimolestid'-taeniodont relationships, in fact,
none of the middle ear, inner ear, basicranium, or tarsal anatomy is known
for any species of *Cimolestes* or *Procerberus*; the necessary articulated
specimens in which these parts are preserved in association with diagnostic
dentitions have yet to be discovered. Moreover, Leptictimorpha, an order
named by SZALAY (1977) to unite Leptictidae (in his view, [sic] containing
*Cimolestes*, *Procerberus* and *Gypsonictops* [...]), Pantolestidae and
Taeniodonta on the basis of tarsal structure (the 'leptictimorph
astragalo-calcaneal morphotype' of LUCAS et al. 1998), has not been
validated by other studies concerning either the content of Leptictidae or
the relationships of letictids to other mammals (compare, e.g.,
KIELAN-JAWOROWSKA et al. 1979, NOVACEK 1986, MCKENNA & BELL 1997)."

Quite depressing, eh?
Pantolestidae is a group of, well, "insectivores" from the Paleocene and
Eocene; *Gypsonictops* is from the Hell Creek
(http://www.dinosauria.com/jdp/misc/hellcreek.html#mammals) and noteworthy
for its retention of five premolars, seemingly common among Cretaceous
eutherians, but seemingly repeatedly lost, and seemingly not found in
Cenozoic mammals altogether.

p. 414:
"[...] we agree with LUCAS et al. (1998) that the identification of the
nearest relatives of taeniodonts is a problem that is at present unresolved.
We suggest that its solution can only come with a richer fossil record, one
that includes Cretaceous taeniodonts more basal than *Onychodectes
tisonensis* or *Schowalteria clemensi*."

"Nothing precise ain't known"
-- Old Austrian proverb

It goes without saying that this calls for a cladistic analysis with _many_
taxa. This includes _all_ the Paleogene weirdos, not just Cimolestidae,
Leptictidae and Pantolestidae!!! But no. Fox & Naylor don't indicate that
anyone were preparing such an analysis. Understandable -- it would probably
be nothing smaller than an analysis of eutherian phylogeny in general.

No, I'm not going to do it. ;-|

Now the interesting things start:

"Northern Origin of North American Paleocene Mammals in the Late Cretaceous

p. 414f:
The discovery of *S[...] at KUA-1, Alberta, extends the range of taeniodonts
stratigraphically [...]. This is the first extension of the taeniodont
record to earlier horizons in over 100 years, since COPE's (1888)
description of *Onychodectes*. The discovery of *S. [...]* at KUA-1 also
extends the geographic range of taeniodonts to 1500 kms [kilometers times
seconds?!? 1 km, 2 km!!!] north of their next youngest records, in the
Paleocene of the San Juan Basin, northwestern New Mexico [...], and 600 kms
[see above] north of their northernmost Paleocene occurrences, in te Crazy
Mountain Basin, southern Montana. Late Cretaceous taeniodonts are
undiscovered in the United States, even though rocks of this age are widely
distributed in the Western Interior and have yielded some of the richest
Late Cretaceous mammalian assemblages in the world [...]; significantly, it
is in this same region where most specimens of Tertiary taeniodonts have
been found, as well [...]. Now that the origin of taeniodonts is known to
have occurred before the K-T boundary extinction event, the question arises
as why Late Cretaceous taeniodonts have not been discovered in the United
States. Is it owing to the low probability of encountering the remains of
animals that were always rare in communities sampled by the record, as is
likely for Tertiary taeniodonts [...]? Or might it be owing to major
differences in the continental distribution of early taeniodonts between
Late Cretaceous and Paleocene time? In other words, perhaps Late Cretaceous
taeniodonts are undiscovered in the United States not because of sampling
error but because they were never there."

Yes, perhaps. Could anyone quantify the probability? The authors don't.

p. 415:
"This asymmetric distribution of early taeniodonts -- with the earliest
species occurring at a northern, Late Cretaceous site and more southerly
species first appearing in the Paleocene -- is not unique among mammals from
these intervals (RUSSELL 1975, FOX 1997). Otehr examples are furnished by
the neoplagiaulacid multituberculate *Parectypodus* [...] and by
Condylarthra (archaic ungulates or hoofed mammals), suggesting an
evolutionarily more meaningful pattern than if illustrated by taeniodonts
alone."

Er... pardon? Cretaceous "condylarths"?

"*Parectypodus* first apears in the Lancian Frenchman Formation at the Gryde
locality, southwestern Saskatchewan [...]. The genus has long been known
from the Paleocene/Eocene of the United States, where it is a common member
of mammalian local faunas, but it has no Cretaceous record there;like that
of taeniodonts, its first American occurrence is in the middle Puercan of
the San Juan Basin [...]. Similarly, the earliest condylarths have been
found at Lancian horizons in the Frenchman Formation of Saskatchewan, where
at least two lineages, Arctocyonidae and Periptychidae, but possibly
Hyopsodontidae, as well, are present (JOHNSTON 1980, JOHNSTON & FOX 1984,
FOX 1989); this diversity implies an origin of the order still earlier in
the Cretaceous. However, no arctocyonids, periptychids, or hyopsodontids are
known from indisputably Cretaceous rocks in the United States; instead, they
are first known from the early Puercan of Wyoming and Montana [...]."

Something is rotten in the state of Denmark. As rotten as pterosaur
cartilage. Have I been sleeping, and much of the world with me? Or have the
abovementioned fossils been reinterpreted or redated (I just say Bug Creek)?
Or what is going OOOOONNN!?!

The morphological analysis that finds whales as artiodactyls (JVP of
December) does not include Periptychidae, but finds the other two in the
following positions (all 4 node names taken from molecular cladograms!):

+--Leptictidae
`--Placentalia
     |--Afrotheria
     |    `--*Orycteropus* (aardvark)
     `--+--+--*HYOPSODUS*
        |  `--Phenacodontidae (traditionally close to perissodactyls)
        `--Boreo(eu)theria
             |--*Rattus*
             `--Laurasiatheria
                  |==Carnivora (paraphyletic)
                  `--+--ARCTOCYONIDAE
                     `--+--Mesonychia
                        `--+--*Andrewsarchus*
                           |--Perissodactyla
                           `--Cetartiodactyla

In other words, if those two groups reach back to the Cretaceous, then lots
of other crown-group clades were already around at the same time.

"Given the Lancian occurrence of *Schowalteria*, it may have been only after
the K-T boundary event and the ecologic disruption which it caused that
taeniodonts moved southward and entered nto the rich Paleocene mammalian
record of the United States. Perhaps they were parallelled in their
southward expansion by other mammals, as the distributions of *Parectypodus*
and early condylarths imply."

I appreciate any feedback on those Cretaceous "condylarths". The references
mentioned above are:

P. A. Johnston (1980): First record of Mesozoic mammals from Saskatchewan,
CJES 17, 512 -- 519
P. A. Johnston & R. C. Fox (1984): Paleocene and Late Cretaceous mammals
from Saskatchewan, Canada, Palaeontographica A 242, 163 -- 222
R. C. Fox (1989): The Wounded Knee local fauna and mammalian evolution near
the Cretaceous-Tertiary boundary, Saskatchewan, Canada, Palaeontographica A
208, 1 -- 59

I haven't read any of these, and probably can't get most or all of them.
Don't ask me how Pal. A 208 can be younger than Pal. A 242.