Scopus hit counts shrinking for same query

[Peter Jacso]

I spotted Scopus producing very different (mostly shrinking number of) hit counts for identical queries when repeated minutes later, i.e. 
     hawaii 115,295 then when repeated 
     hawaii 112,798
I spotted this Sunday evening (after the MLA exhibit hall closed here). I e-mailed the Amsterdam office as it was Monday morning there, but I received no reply.
I would not be much concerned about a temporary dementia but I have done quite a number  of searches the days before on Scopus for a research paper, and I wonder if anyone has spotted this problem already last week, making my earlier results invalid. 
Thanks
peter jacso
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mail.asis.org/pipermail/sigmetrics/attachments/20090518/db5b3609/attachment.html>