The paper describes the stable isotopic composition of water within the catchment area of Thingvallavatn. The lake is about 90% fed by groundwater springs, mainly entering the lake by the northern shore. These springs have stable isotopic composition considerably lower than the local precipitation, but similar to the present precipitation falling on the southernmost part of the Langjokull ice cap. It is therefore suggested that the spring water originates from Langjokull ice cap but is slightly enriched in the heavier isotopes due to mixing with local precipitation on the 50 km way from the recharge area to the lake. At the southwestern shoreline of the lake several springs have been observed that are luke warm (7-11-degrees-C) due to the geothermal activity at the Nesjavellir field. The oxygen isotopic composition of this water is slightly heavier than observed elsewhere within the catchment area due to contamination by the Nesjavellir thermal fluid, which has become considerably enriched in the heavy isotope O-18 due to oxygen isotopic exchange with the bedrock. In the paper it is shown that the thermal water at Nesjavellir also originates from the Langjokull ice cap. The delta-O-18 - delta-D relation for the waters in open fissures. springs and wells within the Thingvellir National Park follows a line with a similar slope to that of the meteoric line (i.e. 8), but most of the data points lie above the meteoric line. The deuterium excess which is 10 parts per thousand for the meteoric line is 10 to 13 parts per thousand for the Thingvellir water. Seasonal variation in the delta-O-18 value was studied for three fissure springs within thc National Park. For the fissures with the least rainwater component (the Vellankatla springs) the variation is usually very small, whereas the spring Silfra, that has higher local rainwater component, shows more variation. Stable isotope data suggests that 75% of the water in the Vellankatla springs originates in the Langjokull ice cap. Furthermore, it is shown that precipitation within the 5 previous years amounts to about 9% of the Vellankatla flow and that the 25% of the Vellankatla water, which consists of recent precipitation, is younger than 10 vr. Radiocarbon analyses of the water give, on the other hand, an anomalously high age due to influx of CO2 from the mantle. It is shown how evaporation from lakes and ponds can be modelled and studied quantitatively by measuring the deuterium excess. The results demonstrate the importance of studying simultaneously the oxygen and hydrogen isotopes and suggest that if the deuterium excess is < 10 parts per thousand the water is isotopically different from its original isotopic composition due to evaporation. This is particularly important to bear in mind when the geographical distribution of delta values in precipitation is being studied and the delta-D values are used to trace the origin of a spring water.