Covariance of beta and lambda in geoR likfit

Greetings comRades:
Has anyone worked out a procedure to calculate the estimated covariance 
between the beta and lambda estimates in the likelihood-based 
geostatistical model implemented in package geoR? If so, please help me 
out by letting me know.
Do I have reasons to expect that this covariance term would be big or 
perhaps more happily, that it could be discarded without much effect?
Thanks.
Rub?n