popSize posterior problems

[Sam Passmore]

Hi,

I am trying to decipher why some parameters in my Bayesian Extended Skyline model are not converging. I thought a sensible approach is to look at the taxa that might be influencing the poorly estimated parameters (e.g. popSizes.alltrees.19 in the photo).

I have so far been assuming that  '19' refers to the node of the phylogeny. 

However, after creating an MCCT of the posterior, I found that the node of the tree labelled 19 by read.nexus / read.beast is a taxa.

So that assumption seems incorrect.

Could someone offer some direction on where to look for why these popSize parameters are so wild?

Thanks in advance,

Sam