G94 basis set problem

 Dear CCL,
 I got the following problem when using Gaussian 94 (rev. D.4). Doing some
 calculations with an in house defined basis set (for Hg) I got the
 following message (after one and half hour of CPU on our R10000 based
 POWER Challenge):
  **** Fatal Problem: The largest alpha MO coeffient is  0.11335235D+04
  Error termination via Lnk1e in /packages/run/g94_D4/g94/l801.exe.
 The calculation was single point SCF MP2=Full. As the guess is that there
 is some problem with the check on linear dependency of the defined basis,
 I would like to ask you for any option/parametr how to reduce the internal
 cutoff values for this case.
 Any help will be really appreciated.
 Thank you very much in advance.
 					Ludek Matyska, PhD
 Institute of Computer Science
 Masaryk University
 Botanicka 68a
 602 00 Brno				Fax: ++420-5-41212747
 Czech Republic				Phone: ++420-5-41512213
 e-mail: ludek |-at-| ics.muni.cz		       ++420-5-740267
   Dear Ludek,
   In the route line (#P), include IOP(8/11=1)
   This will give a warning about large eigenvalues but not crash
  the run.
     Best wishes,
  Dr. John Waite,                            e-mail:  chem8 |-at-| york.ac.uk
  The National Hellenic Research Foundation,   rosen |-at-|
  Organic and Pharaceutical Institute,       phone: ++30-1-7238958 (direct)
  Vas. Konstantinou 48,                      phone: ++30-1-7247913(secrtry. Mary)
  Athens 116-35,                             fax:   ++30-1-7247913
  NCRS "Democritos",                         phone: ++30-1-6513112-5
  c/o Dr. G.Kordas,                          e-mail john |-at-|
  Material Science Institute,
  Aghia Paraskevi,
  Athens 153-10,