Removing the cause of a decline does not automatically reverse it. Below a certain size a population faces a second set of problems that arise from smallness itself, and they operate whether or not the original pressure has gone.
Chance starts to dominate
In a large population, births and deaths average out. Individual bad luck is absorbed because thousands of other outcomes are happening simultaneously.
In a population of a few dozen, a run of male-biased births, a bad breeding season or the loss of the only experienced female can move the whole trajectory.
This randomness has no relationship to habitat quality. A well-protected small population can still decline simply because unfavourable outcomes clustered together.
Inbreeding raises the cost of every mating
Related individuals share more of their genetic material, so their offspring are more likely to inherit two damaged copies of the same gene rather than one working copy.
The consequences appear as reduced fertility, lower birth weights, weaker disease resistance and higher juvenile mortality, none of which is obvious from a headcount.
The effect compounds because it reduces the number of successful breeders, which further shrinks the effective population and accelerates the next round of relatedness.
Not every adult contributes equally
Genetic health depends on the effective population size, which counts only individuals actually reproducing and weights them by how much they contribute.
Where a few dominant males father most offspring, or where sexes are unbalanced, the effective size can be a small fraction of the number of animals present.
This is why a census figure can look reassuring while the genetic situation deteriorates, and why managers track parentage rather than only numbers.
Some species stop functioning when sparse
Certain behaviours require a minimum density. Colonial nesters need neighbours to trigger breeding, and cooperative hunters need enough adults to bring down prey.
Species with dispersed individuals may simply fail to find mates, particularly where the remaining animals are separated by habitat they will not cross.
Where these effects apply, decline becomes self-reinforcing below a threshold, and the population falls faster as it gets smaller rather than levelling off.
What management does about it
The standard intervention is moving individuals between isolated groups so that unrelated genes enter a population that cannot reach them naturally.
Studbooks track ancestry so that pairings minimise relatedness, an approach developed in captivity and now applied to intensively managed wild populations.
None of this substitutes for habitat. Genetic management buys time for a population to grow past the size where chance dominates, which is where recovery actually becomes possible.