Pretty much as title says - wondering if evolution can go from mammal back to reptile - either in theory or practice.
Secondary ectothermy has been suggested in living crocodylians by Seymour et al. (2004). Their rationale is based on:
- Presence of a 4 chambered heart (otherwise only found in endotherms)
- Post-cardiac shunting via the foramen of Panizza and "cog-tooth" valve, which develop secondarily in crocodylians
- Lung structure and ventilation during locomotion
Thus ectothermy in extant crocodylians is associated with being aquatic or semiaquatic. These authors suggest that endothermy may have been ancestral for dinosaurs as well.
Seymour RS, Bennett-Stamper CL, Johnston SD, Carrier DR, Grigg GC. 2004. Evidence for endothermic ancestors of crocodiles at the stem of archosaur evolution. Physiol Biochem Zool 77:1051–1067.
There is no evolutionary "ratchet" that prevents a life form from evolving "backwards". However, it is statistically unlikely because it would pretty much require all of the environmental factors that changed during the transition from "cold blooded" to "warm-blooded" to change back in the reverse order, because all of those factors contributed to the selection pressures that caused the "forward" evolution.
It would be analogous to getting a set of moving billiard ball to reverse direction and re-form an earlier pattern: reversing any one ball's velocity would be fairly easy, but getting all of the balls to cooperate to re-form a static pattern that ends up shooting out the queue ball is so unlikely as to be effectively impossible.
That said, for a modern form (B) to evolve into another form (C) that resembles an archaic form (A) without actually being that archaic form would "only" require a sequence of environmental changes that leads to the evolution of (B) from (A). It just wouldn't actually be the archaic form. It would be a process analogous to convergent evolution.