I've been looking at this and the example above is misleading.  It works for me in the nucleotide case, but when applying it to the codon situation (61x61) it provides a matrix with no likelihood.  I get the same likelihood whether I use matrix = {61,61}; as I do with matrix[i][j]:=mu; for each i!=j point in the matrix.

Thanks for the incredibly quick reply!  Below is the actual code I'm running and getting a flat likelihood.  I get the same likelihood if I insert a matrix designated as matrix = {61,61};


DataSet myData = ReadDataFile ("brown.nuc");
DataSetFilter codonFilter = CreateFilter (myData,3,"","","TAA,TAG,TGA");

HarvestFrequencies (obsFreqs, codonFilter, 3, 3, 1);

CFreqs = {61,1};
for(i=0; i<64; i=i+1)
{
count=0;
if(i!=48 && i!=50 && i!=56)
     {
     CFreqs[count][1] = obsFreqs[i][1];
     count = count+1;
     }
}

TMatrix = {61,61};
for(j=0; j<61; j=j+1)
{
for(z=0; z<61; z=z+1)
     {
     if(j!=z)
           {
           TMatrix[j][z]:=mu;
           }
     }
}

Model eval = (TMatrix , CFreqs);
Tree myTree = (((Human,Chimpanzee),Gorilla),(Orangutan,Gibbon));
LikelihoodFunction theLikFun = (codonFilter , myTree);
Optimize (MLEs, theLikFun);
fprintf (stdout, "\n","\n", MLEs[1][0], "\n","\n");

Thank you!  That clearly fixes it and I just ran it to check as well.
However, I will admit to you that I do not feel very intelligent at this moment

Sam