The NAT tag is used to wrap implementation (in our case Pico, 'real' Pico) functions such that they can be used by MC-eval. A NAT tree is simply a tag together with the associated native function. Whenever nn is a tree representing such a primitive function, then the function can be retrieved using nn[NAT_NAT_idx]. The name can be accessed through nn[NAT_NAM_idx]. Of course, in our case where primitive functions our implemented in Pico, this name will be the same af the name of the function in the underlying Pico. In the C version, this is not necessarily the case.