Check if time/no. nodes between colonisation of a super- and sub-island clade can force the colonisations to be separate

[joshwlambert]

Lygodactylus genus in the squamate tree had a colonisation to the island, then a subgroup of the clade back-colonises the mainland, and then a subgroup of the mainland clade recolonises the island. There should be a mechanism in extract_island_species() to force these to be separate island colonisation events.

Two criteria for separation would be time (millions of years) off the island, or number of nodes between original colonisation and re-colonisation events.