Originally Posted by JPP
I think you're sort of confusing lines with morphs. Some people frown on line-breeding, which usually refers to breeding related frogs (like siblings) to each other. But if you get nominal leucs from "breeder X", and pair them with some from "breeder Y", hopefully they aren't closely related and therefore you're maintaining some genetic diversity. But they're still the same morph.
Sort of; 'line breeding' has two different meanings. The one you mention is accurate, but is not what is at issue when someone recommends against crossing lines of frogs.
A 'line' of frogs is all the frogs and their offspring from one import event, or one breeder's group. An example: Ranitomeya imitator "Green" is a morph (also called 'standard' and (incorrectly) 'nominate'). There are at least three lines of this morph: Uhern, Sens, and Understory Enterprises* -- all named after the person/organization who first introduced that distinct line of that morph to captivity.
The reason not to cross lines is that they each represent distinct genetic backgrounds that are worth preserving for that reason alone -- so as not to muddle all the 'green' imis into one homogenized mass.
It may (with a very high probability) turn out that these frogs are not related in the way we think they are
; animals that all look alike may in fact turn out to be two distinct populations or even species that differ in e.g. their call. R. imitator
is a great example of this, of course, since various morphs of imitator imitate the appearance of at least three other species really well.
These issues have made mutts of quite a few animals in the larger herp hobby: nearly all captive leopard geckos are a mix of subspecies, African house snakes are likely a complex of species, most of which have been crossed into captive populations. The taxonomic understanding of many captive herps is still very primitive, and the more we keep lines pure the more likely we will preserve animals that aren't mutts.
*Also called 'UE' (not to be confused with EU=European) -- this line has locality data and so while a line of the 'green' morph is more accurately referred to as "Cainarachi Valley")