Shelxpro is lousy at dealing with the conversion of disordered residues from CNS-formatted .pdb files to Shelxl-formatted .ins files.  If you want to keep them out in initial Shelxl runs and add them in later, that's ok.  But, you've probably spent hours looking at maps and refining these alternate conformations in CNS, and you want to keep them from the get-go in Shelxl.

In short, you need to set the alternate comformations up manually in your .ins file with the PART instruction.  Below are the details of what you'll get initially and how to fix it.
 

If you have alternate conformations, such as two O positions in a Serine residue, you might define them in your CNS file something like this:

First you define the residue with segment ID's for the majority of the protein (AAAA) and for the first of the alternate conformations (AC1),

ATOM  473  N   SER   68   21.870  14.372  19.711  1.00  6.81   AAAA
ATOM  474  CA  SER   68   22.638  13.136  19.726  1.00  7.89   AAAA
ATOM  475  CB  SER   68   22.441  12.403  21.054  1.00 10.99   AAAA
ATOM  476  OG  SER   68   21.071  12.174  21.317  0.50 14.58   AC1
ATOM  477  C   SER   68   22.132  12.304  18.552  1.00  7.39   AAAA
ATOM  478  O   SER   68   20.943  11.986  18.465  1.00  7.94   AAAA

. . . and somewhere else in the .pdb file, you define the other alternate conformation (AC2),

ATOM 1639  OG  SER   68   21.537  11.965  21.876  0.50 13.45   AC2
 

Running this through Shelxpro, you would get this:

RESI   68   SER
N    3    0.456859    0.251183    0.246936    11.00000    0.08328
CA   1    0.457339    0.229514    0.247206    11.00000    0.08968
C    1    0.442274    0.214305    0.232672    11.00000    0.08838
O    4    0.421327    0.209382    0.231385    11.00000    0.09317
CB   1    0.449143    0.214084    0.263779    11.00000    0.30085
OG   4    0.427784    0.211529    0.269367    10.50000    0.21397

. . . and instead of the correct alternate conformation command, you would get the following somewhere farther down in the .ins file,

RESI   68   SER
OG   4    0.448548    0.198664    0.266939    10.50000    0.12765
 

This won't work.  Trust me.

Set up PARTs:

As the default in any .ins file, everything starts out in PART 0.

First, duplicate CB of the Ser and put it into PARTs 1 and 2.  Put the first OG position into PART 1, and the second OG into PART 2.  End the residue with PART 0.

Finally, get rid of the second   RESI   68   SER  instruction.  What you end up with for residue Ser68 should look like this:

RESI   68   SER
N    3    0.456859    0.251183    0.246936    11.00000    0.08328
CA   1    0.457339    0.229514    0.247206    11.00000    0.08968
C    1    0.442274    0.214305    0.232672    11.00000    0.08838
O    4    0.421327    0.209382    0.231385    11.00000    0.09317
PART    1
CB   1    0.449143    0.214084    0.263779    10.50000    0.30085
OG   4    0.427784    0.211529    0.269367    10.50000    0.21397
PART    2
CB   1    0.449143    0.214084    0.263779    10.50000    0.30085
OG   4    0.448548    0.198664    0.266939    10.50000    0.12765
PART     0
 

QUIRK!    Even though only the O is in an alternate conformation, you also need to have the C as part of PARTs 1 and 2.  I've tried, and it doesn't work when you ONLY have the OG there without the CB.