IMa Island Model Load Mode difficulty

[Clifford Cunningham]

Dear Jody,

I run IMa in Island model (-j4) .

In the Load trees mode model testing, it correctly gives T as very large, but for the alternate models, it seems to be estimating t, and the values of M are very strange.

This means I can't compare the log values either, correct?

Is there another instruction I should add to the load trees mode to do the model comparison correctly?  Thank you!

-i /home/abyss/cliff/IMaLo/LeFruBiEx.b1.s.txt -o /home/abyss/cliff/IMaLo/LeFruBiEx.b1.s.95.Eq.PST.out -r LeFruBiEx.b1.s.95.Eq.out -q1 100 -q2 100 -qA 100 -m1 100 -m2 100 -j4 -n 10 -c02 -b 20000 -L 100000 -p1234


Multidimensional Peak Location and Posterior Probability
--------------------------------------------------------
 Set   t       log(P)      q1         q2         m1         m2     
   0  1000000.0000  -1.1661     1.8538     1.1616     2.1456   0.000119
   1  1000000.0000  -1.2908     2.8910     1.2721     2.4471   0.000107
 ALL  1000000.0000  -1.3919     2.5014     1.1509     2.3627   0.000100


Nested Models Peak Location and Posterior Probability
------------------------------------------------------
Model               t          log(P)      q1          q2          m1          m2         df       2LLR
 AB_DD    0.009566    -4.6496     2.5014      1.1509  [   0.0001] [     -nan] [     -nan]    1        6.5153
 AA_DE    0.125329    -2.0768     2.5014  [   2.5014] [   0.0001] [   1.1509] [     -nan]    1        1.3698
 AA_DD    0.004498    -5.4042     2.5014  [   2.5014] [   0.0001] [   1.1509] [   1.1509]    2        8.0246