What do you have filled in your conf file at the heading "ratehetmodel"?

Send that, and I may be able to help!

I had partitioned data set, and had gamma for 3 models and one gammafixed . It really was something weird as I just removed one taxa, run the modeltests again, and based on that changed the gammafixed to gamma - otherwise identical .conf files. With the -1 taxa dataset the program run just nicely.

oh, and I triple checked the file with other people also just to be sure I did not miss something...

I think what the problem is is that if you input "gammafixed" you need to provide an alpha parameter value. For the most part, I would just run them all as gamma.

Glad you sorted it out!

Weston, you´re probably right. Have to remember that next time! Thanks for help.