How do I correct a threshold warning message in S-plus?

["Ryder, Anthony" <ARyder@sph.uth.tmc.edu>]

I am using the window version of S-plus6.1. 
My data set named “shore” has the following characteristics: 
  
Summary of missing values 
81 numeric variables, 546 observations, 166 patterns of missing values 
78 numeric variables (96%) have at least one missing value 
228 observations (42%) have at least one missing value 
  
In order to perform a multiple imputation, I have used S + MISSING DATA and the 
following commands: 
  
>library(missing) 
>First <- function() {library(missing) } 
>M <- miss(shore) 
>M 
  
Then I have used the following preprocessing function to prepare the data set 
for an EM algorithm 
>shore.pre <- preGauss(shore). 
   
What is the proper S-plus syntax statement I should use to fit correctly a 
Gaussian model to the shore 
data set using the EM algorithm?I have tried the following statements and the 
error message blocking my 
process from moving forward is that the reciprocal condition number falls below 
threshold in 
emGauss(shore.pre, prior = "ml").I thought that tolerance meant threshold and 
by changing the 
tolerance default of 0.000001 to a tolerance of 0.01 I would not have any more 
threshold warning message, 
and therefore I would be able to process ahead in the convergence assessment.
 
> control = emGauss.control(maxit = 200, tolerance = 0.01) 
control = list(maxit = 200, tolerance = 0.01) >shore.EM <- emGauss(shore.pre, 
prior = "ml") 
  
Iterations of EM: 
Iteration   ParChange 
      1        1.1846 
Warning messages: 
  reciprocal condition number falls below threshold in: emGauss(shore.pre, 
prior = "ml") 
  
Thanks. 
  
Anthony B. Ryder