The other two models, density-dependent logistic (DDL + E) and density-dependent exponential (DDE + E), assume diversity-dependence and thus quantify diversification rates as functions of changes in accumulating diversity over time (while accounting for extinctions, E).
Our results provide evidence that the New Caledonian diversifications follow a process of diversity-dependence.
The third model (BOTHVAR) uses linear diversity-dependence for both speciation and extinction rates [ 24].
It was originally suggested that a signature of diversity-dependence could only be observed if extinction was zero [ 73].
Even if we imagine that the LiMe ratio was sufficiently high to produce a pattern of diversity-dependence, another issue remains.
Recent simulation studies have demonstrated that in the case of a diversity-dependence process with high LiMe ratio, there is only a short temporal window where the γ statistics can detect a diversification slowdown.
Cuban and Hispaniolan taxa analyzed separately did not have constant diversification rates; rates changed possibly due to increased speciation, diversity-dependence processes, or a combination of both (Akaike Weights were unable to discern among models).
We will use the term diversity-dependence [ 28] rather than density-dependence [ 31, 32] because this process refers to the density of species (diversity) rather than the density of individuals [ 33].
Although statistically significant, diversity-dependence remains moderate, and affects the transposition rate only marginally (the current transposition rate for all families is more than 85% of the estimated initial transposition rate when one copy only was present in the genome).
