Nearly 200 analyses of meltwaters, ice and snow from three alpine glac
ial sites in carbonate terrain are summarized and discussed in terms o
f sources of solutes and kinetic controls on the progress of weatherin
g reactions. Most data derive from the Swiss Glacier de Tsanfleuron wh
ich is based on Cretaceous and Tertiary pure and impure limestones. Tw
o other sites (Marmolada, Italian Dolomites and the Saskatchewan Glaci
er, Alberta) are based on a mixed calcitic-dolomitic substrate. Most s
olutes originate from carbonate dissolution; moreover, where pyrite is
present its oxidation supplies significant sulphate and forces more d
issolution of carbonate. The ratios Sr2+/Ca2+ and Mg2+/Ca2+ are much h
igher in Tsanfleuron meltwaters than local bedrock, a phenomenon that
can be reproduced in the laboratory at small percentages of dissolutio
n. These anomalous ratios are attributed to incongruent dissolution of
traces of the metastable carbonates Mg-calcite and aragonite. These p
hases also provide Na+ to solution. K+ is argued to originate mainly b
y ion-exchange on clays with solute Ca2+. Quartz and very minor feldsp
ar dissolution are also inferred. Locally enhanced input from atmosphe
ric sources is recognized by high Cl- and associated Na+. The progress
of weathering reactions has been evaluated by the trends in the data,
computer modelling and some simple laboratory experiments. The most d
ilute samples show a trend towards removal of CO2 to low partial press
ures (c.10(-5.5) atmospheres), reflecting initially rapid carbonate di
ssolution and relatively slow dissolution of gaseous CO2. Later additi
on of atmospheric CO2 or acid from pyrite oxidation allows further car
bonate dissolution, but solutions show a wide range of saturations, an
d CO2 pressures as high as 10(-2.2) where pyrite oxidation is importan
t. In a carbonate terrain, measurement of electroconductivity (correct
ed to 25 degrees C) and alkalinity in the field allows the following p
reliminary deductions (where meq stands for milliequivalents): meq(Ca2
+ + Mg2+) = 0.011 electroconductivity (in units of mu S cm(-1)) meq su
lphate (from pyrite oxidation where gypsum absent) = meq(Ca2+ + Mg2+)-
meq alkalinity meq(Ca2+ + Mg2+)- meq sulphate = S where S is the mini
mum meq(Ca2+ + Mg2+) produced by simple dissolution of carbonate uncon
nected with pyrite oxidation. As with any proxy method, these deductio
ns do not remove the need for chemical analysis of waters in a given s
tudy area.