Fruitafossor: now a Late Jurassic echidna from Colorado

While reviewing the terrestrial descendants of tree shrews
yesterday, the Late Jurassic Fruitafossor (Figs. 1, 2) stuck out as a chronological misfit as it nested in the otherwise Tertiary edentates (= Xenarthrans).

Here is the problem,
and the solution.

A Jurassic edentate? No.
Fruitafossor windscheffeli (Luo and Wible 2005) used to nest in the LRT with digging edentates, like the armadillo-mimic, Peltephilus (Miocene), and for good reason…

Wikipedia reports,
“The teeth of Fruitafossor bear a striking resemblance to modern armadillos and aardvarks. Its vertebral column is also very similar to armadillos, sloths, and anteaters (order Xenarthra). It had extra points of contact among similar to the xenarthrous process that are only known in these modern forms.”

By contrast, Wikipedia concludes,
“Its shoulder-girdle is similar to a platypus or reptile, but many other features are more similar to most other modern mammals.”

What would Larry Martin say?
Run a complete analysis. Don’t rely on one, two or a dozen traits. And the Late Jurassic is so early in mammal evolution that it becomes important, too. There were fewer mammal clades back then. Edentates had not yet arrived.

Figure 1. Several drawings from Zhou and Wible that one must trust for accuracy. The verification data is too fuzzy to validate.

So is Fruitafossor a Late Jurassic edentate?
Or an edentate-mimic in the Late Jurassic?
With current scoring in the LRT, shifting Fruitafossor from the edentates to the base of the Monotremata adds 23 steps. Shifting to Early Cretaceous Lactodens within the Monotremata adds just 17 steps, the lowest number I could find. Lactodens has typical differentiated teeth and five fingers with small, sharp claws, traits not shared with Fruitafossor + edentates. Lactodens nests with the echidnas, Tachyglossus (extant, Figs. 3–5) and Cifelliodon (Early Cretaceous; Fig. 3). The latter has simple blunt teeth and the former is a known digger.

Figure 2. Fruitafossor in situ from Digimorph.org and used with permission and here colorized to an uncertain extent.

So let’s reexamine scored traits… and solve this conundrum.
Has the LRT met its match? Very few skull traits are known from Fruitafossor. Even so, earlier I overlooked or mis-scored the following that gain importance in hindsight:

Fruitafossor:

1. orbit contacts the maxilla
2. 4 rather than 5 sacrals,
3. coracoid present
4. I could not score hind limb length without a pes and estimates won’t do
5. proximal sesamoid of fibula present
6. fibula diameter greater than half of tibia
7. dorsal osteoderms absent (I misinterpreted scattered elements at Digimorph.org)

Tachyglossus:

1. retroarticular process present as in Fruitafossor
2. metacarpal 1 and 2 are the longest as in Fruitafossor
3. longest manual digit 3 as in Fruitafossor
4. manual digit 4 narrower than 3 as in Fruitafossor

Cifelliodon:

1. three molars, as in Fruitafossor

Figure 3. Early Cretaceous Cifelliodon is ancestral to the living echidna, Tachyglossus according to the LRT. The lack of teeth here led to toothlessness in living echidnas. The skull of Tachyglossus is largely fused together, lacks teeth and retains only a tiny lateral temporal fenestra (because the jaws don’t move much in this anteater. Compared to Cifelliodon the braincase is greatly expanded, the lateral arches are expanded and the two elements fuse, unlike most mammals.

Figure 4. Tachyglossus skeleton, manus and x-rays. Note the perforated pelvis.

Figure 5. The echidna (genus: Tachyglossus) in life. This slow-moving spine-covered anteater has digging claws.

Results (as you might imagine, given these changes):
Fruitafossor is an edentate-mimic nesting basal to Cifellidon and Tachyglossus as a Late Jurassic echidna and monotreme in the LRT. Glad to get rid of that problem!

In their original description of Fruitafossor,
Luo and Wible 2005 nested their discovery between a monotreme clade and a clade with the mammal-mimic, Gobiconodon at its base, then a clade with another egg-laying mammal, Tinodon at its base, then a pangolin ancestor, Zhangheotherium, then a rabbit ancestor Henkelotherium, then two other monotremes, Dryolestes, Amphitherium and the carnivorous marsupial, Vincelestes.  Luo and Wible tested Tachyglossus, but not Cifelliodon, which was published in 2018. Note the simple, blunt teeth in Cifelliodon, nearly matching those in Fruitafossor. Given that the only fossil of Fruitafossor is a bit jumbled, it is possible that it, too, had five fingers in vivo, like other monotremes. With only four fingers (Fig. 1) Fruitafossor had a good excuse for pretending to be an edentate.

So, yes, the LRT was up to the challenge.
But it took insight, lacking until now, to provide the correct matrix scoring. I’m happy to announce that the twenty or so corrections made yesterday were added to the 120,000 or so corrections made over the past ten years. With these corrections the LRT gets better and stronger every week. Minimizing taxon exclusion maximizes the opportunity to correctly nest new and enigma taxa with old and established taxa, even if the new and old specimens are incomplete or scattered about.

The earlier August 2017 blogpost for Fruitafossor
was updated yesterday to erase old errors and enter the corrections.

---

References
Huttenlocker AD, Grossnickle DM, Kirkland JI, Schultz JA and Luo Z-X 2018. Late-surviving stem mammal links the lowermost Cretaceous of North America and Gondwana. Nature Letters  Link to Nature
Luo Z-X and Wible JR 2005. A late Jurassic digging mammal and early mammal diversification. Science 308:103–107.
Shaw G 1792. Musei Leveriani explicatio, anglica et latina.

wiki/Fruitafossor
digimorph.org/specimens/Fruitafossor_windscheffeli/