Parasitoid-Host Cophylogenetic Analyses
The relationships of the parasitoid and beetle Bayesian phylogenies used
in the cophylogenetic analyses were identical to those recovered in this
study and Peixoto et al. (2018) (SI Fig. 3 & 4). Using the default cost
model, ten isomorphic solutions were recovered (Fig. 4) with zero
cospeciation, three duplication, one duplication and host switch, 13
losses, and six failure-to-diverge events for a total cost of 24. The
second Jane analysis using an alternative cost model (cospeciation set
to -10, a much lower cost than all other events) were identical to the
default cost model (SI Fig. 5) with ten isomorphic solutions as listed
above and a total cost of 24. These results indicate that duplication,
and not cospeciation, is the dominate speciation event and that the lack
of cospeciation events inferred are robust.
The PACo distance-based analyses returned a sum of squares of 288.55
with p-value < 0.001 showing strong phylogenetic congruence.
This is an indication of cospeciation between the parasitoid and host
phylogenies, which contradicts the results of the Jane analyses. The
Procrustes residuals indicated that associations between E.
annleckieae and Pst. cloelia had a large amount of
cophylogenetic signal (residuals closer to 0) compared to all other
species of Eadya . Separating these associations from the rest and
comparing the residuals using a Welch’s t-test demonstrated that the
cophylogenetic signal between E. annleckieae and Pst.
cloelia is significantly greater than all other associations (t =
-6.48, df = 73.35, p < 0.001). This indicates that the
cophylogenetic signal observed between the parasitoids and their beetle
hosts is largely contained in associations between E. annleckieaeand Pst. cloelia , signifying sampling bias.