Marsupial teeth: Polyprotodonty to Diprotodonty

Defintion: Polyprotodont
“any marsupial of the group Polyprotodontia, characterized by four or more upper incisor teeth on each side of the jaw: includes the opossums and bandicoots.” “defined by having multiple pairs of small, similar lower incisors.”

This is the primitive or plesiomorphic condition retained by early placentals and derived from primitive pre-mammals and monotremes.

Defintion: Diprotodont
“any marsupial of the group or suborder Diprotodontia, including kangaroos, phalangers, and wombats, having fewer than three upper incisor teeth on each side of the jaw.” “diprotodont” (meaning “two front teeth”), they have a pair of large, procumbent incisors on the lower jaw, a common feature of many early groups of mammals and mammaliforms.”

This is the derived or apomorphic condition restricted to diprotodontid marsupials, their placental descendants (e.g. rodents and rabbits) and any convergent marsupials.

The following marsupial skulls
(Fig 1) more or less show one transition from polyprotodonty to diprotodonty in the large reptile tree (LRT, 2340 taxa). There are others.

In this sequence
(Fig 1) you can see all four canines were absent prior to the evolution of this line of taxa.

The four lower incisors
of Glironia either became enlarged together and laminated to create one larger incisor – or only the medial lower incisor was greatly enlarged. The taxa that demonstrate that transition are not yet known, but will look like Glironia and/or Petaurus (Fig 1).

Re: the last extant ancestor of the marsupial line, Thylacoleo
The diet of Pseudocheirus (the ringtail possum, Fig 1) “consists mainly of eucalyptus leaves, flowers, and soft plant shoots.” Not other animals. If Thylacoleo was a predator, that happened sometime around Wakaleo (Fig 1), after Pseudocheirus.

Genetic cladograms cannot and do not demonstrate
the transition from polyprotodonty to diprotodonty. They don’t operate in the realm of trait transition. Invading viruses corrupt genes. Genomic cladograms too often recover odd and untenable interrelationships in deep time studies. Recent genes are not corrupt. It takes deep time to corrupt genes.

This is the value of trait-based phenomic cladograms like the LRT. You can see traits. You can see the traits change from taxon to taxon, which is the whole point! That’s where the fun is!

This appears to be a novel hypothesis on phylogenetic interrelationships.
If not, please provide a citation so I can promote it here.

PS
Re: Thylacoleo, the marsupial line (Fig 1). The diet of Pseudocheirus (the ringtail possum, Fig 1) “consists mainly of eucalyptus leaves, flowers, and soft plant shoots.” – Not other animals. If Thylacoleo was a predator, which all experts agree on, that happened sometime around Wakaleo (Fig 1), following Pseudocheirus with few evolutionary changes.

References
wiki/Diprotodontia
wiki/Marsupial