Due to their long life span, changing climatic conditions are of particular importance for trees. Climate changes will affect the water balance, which can become an important limiting factor for photosynthesis and growth. Long-distance water transport in trees is directly related to the transpiration stream and very sensitive to changes in the soil-plant-atmosphere water continuum. Therefore the study of long distance transport gives information about tree response to changing climatic conditions. Here the dynamic behaviour of water transport processes in trees has been studied by the MRI method, which is a direct and non-invasive tool.. MRI flow imaging has been applied to diffuse- and ring-porous trees to study drought stress and the occurrence of xylem vessel cavitations. World-wide unique dedicated MRI hardware is described that allows imaging of sap flow in intact trees with a maximal trunk diameter of 4 cm and height of several meters. This setup is used to investigate xylem and phloem flow in an intact tree quantitatively. Flow is quantified in terms of (averaged) velocity, volume flow (flux) and flow conducting area, either in imaging mode or resolved on the level of annual rings. Results obtained for the same tree, imaged at two different field strengths (0.7 and 3 T), are compared. An overall shortening of observed T2 values is manifest going from 0.7 to 3 T. Although susceptibility artefacts may be present at 3 T, the results are still reliable and the gain in sensitivity due to the higher magnetic field strength results in shorter measurement time (or a better spatial resolution or a higher signal to noise ratio) with respect to the 0.7 T system. By use of such dedicated hardware xylem and phloem flow, and its mutual interaction, can be studied in intact trees in relation to the water balance and in response to environmental (stress) conditions (Chapter 2). To further investigate the effect of susceptibility artefacts on MRI flow imaging by PFG-STE MRI on 3 T, water flow was studied in a number of model porous media with or without surface relaxation, internal magnetic field inhomogeneities (susceptibility artefacts) and exchange with stagnant water pools, mimicking the tree situation (Chapter 3). In such situations a clear dependence of the flow characteristics on the observation time is demonstrated. The most reliable results are obtained at relatively short observation times. This limits the observation of low flow velocities and the discrimination between flowing and non-flowing water. It is shown that correlated displacement-T2 measurements are available to improve the discrimination of flowing and non-flowing water and can be of help to decide about the functional activity of xylem conduits (Chapter 4). A method that reveals exchange between the flowing and stagnant fractions in the system is presented. Further it is demonstrated how this exchange can be quantified (Chapter 3). Xylem flow, flow conducting area and water content in the storage pools of sapwood and cambial zone were investigated simultaneously and non-destructively by MRI in diffuse-porous laurel (Laurus nobilis) and viburnum (Viburnum tinus) trees during a drydown period and recovery after watering (Chapter 4). The development of the drought stress was detected by the decrease in average velocity, volume flow and flow conducting area as observed by MRI flow imaging. A decrease in flow conducting area was observed with a delay of one day in comparison to the observed reduction in average velocity and volume flow. The re-watering of the plants resulted in the fast restore of the flow conducting area to the value observed under well watered conditions, demonstrating that if cavitations had been induced they refilled quite fast. In addition, a significant increase in the average velocity and volume flow was observed, but still lower than the original values. Imaging water content in the cambial zone indicated a gradual decrease of the water content, which speeded up during the drought stress. The rate of decrease was dependent of day/night conditions. Watering resulted in the partial restore of water content in this zone. Water content in sapwood showed a clear diurnal variation. The water storage pool in sapwood depleted quickly upon switching on the light, gradually restoring in the afternoon. Drought stress did not change the character of diurnal variation of water content significantly, but it increased the amplitude of the diurnal variation. Re-watering of the tree resulted in a 10% water loss in sapwood. Thus, for the first time the coupling between water floe in xylem vessels and water content in storage pools was demonstrated. The oldest annual ring was rather inactive in long distance water transport. We found that the transport activity of this ring was not sensitive to any environmental change and that the variation of water content in sapwood was uniform in all annual rings Non-destructive measurements of cavitation were made with MRI to test whether large earlywood vessels of ring-porous xylem are as vulnerable as some standard methods have suggested (Chapter 5). Potted, 3-4 year old Quercus robur L. trees were droughted to water potentials measured with temperature-corrected stem psychrometers. Imaging of (vessel) water content indicated that earlywood cavitation in trunks was not detectible until water potentials dropped below -3 MPa. Most earlywood vessels were cavitated below -4 MPa. Dye perfusions through excised branch segments gave comparable results. Imaging of flow conducting area (FCA) indicated a gradual decline in trunk water conduction that was not solely associated with cavitation, but probably resulted from stomatal closure and too low velocities to be discriminated from non-flowing water. Dye perfusion and FCA indicated a significant portion of earlywood vessels were non-conducting even at the most favorable water potentials. No refilling of embolized vessels was detected in rewatering experiments. Contradictory to the MRI results, standard centrifuge and air-injection methods on Q. robur stem segments indicated complete cavitation at xylem pressures at or below -1 MPa. An artifact in these destructive methods was revealed by experiments on the related species Q. gambelii Nutt. When earlywood vessels became air-filled during collection prior to being refilled in the lab, they became much more vulnerable to cavitation. Residual bubbles left behind in the refilled vessels may be responsible. These results suggest revised protocols for measuring vulnerability curves by destructive methods. An about linear correlation between water potential and decrease of water content in cambial zone of oak (Quercus robur L.) was observed (Chapter 5).