Dear Natalie,

I would recommend using the alternative uranium furnace method where a 2-5
microliter sample is passed into the hot (750C) furnace and the gas sample
is collected with a Toepler pump and introduced into the IRMS.  See
Bigeleisen et al. (1952) Anayl. Chem 24: 1356-1357 or Friedman and
Harcastle (1970) GCA 34: 125-126.  It may mean that you will have to send
the samples to someone who is already set up to use this method (e.g. CSIRO
Water Resources, Glen Osmond, SA, Australia - FAX - 61-08-303-8750).  Also
the the on-line, continuous flow method, now being used on the EUROPA GEO
20-20 IRMS might work but you should contact them directly.  Again the
CSIRO group may know about this since they have this machine and may know
about the method.

I worry about the 10 per mil difference and wonder if you are not getting
incomplete combustion/reduction  - can you rule this out?  I am not sure if
pH should actually affect the combustion reaction is such a way as to
produce the results in the way you see.  If I had to guess, I would not
have guessed that pH would do this.  Perhaps someone on this ISOGEOCHEM
line can provide you with some feedback on this.

Todd Dawson
Cornell Stable Isotope Group.

**********************************************
Todd Dawson
Section of Ecology and Systematics
E409 Corson Hall
Ithaca, NY 14853  USA

(607) 254-4234 - telephone
(607) 255-8088 - FAX

*********************************************