1 INTRODUCTION The paper presents an investigation on the retention properties of artificial bonded and destructured soils with values of void ratio equal to 0.6 or 0.9. Artifi- cial bonded and destructured soil samples were pre- pared in the laboratory by using a technique devel- oped by Ali Rahman (2004). For each type of soil, the water retention curve was obtained in the labora- tory and interpreted in the light of the corresponding soil fabric as observed using a standard plane polar- ized light microscope. Water retention curves in terms of total and ma- tric suction were obtained by using different meth- ods of suction measurement. Matric suction was measured using the axis-translation and the contact filter paper techniques whereas total suction was measured using the transistor psychrometer and the non-contact filter paper techniques. Such techniques have various limitations with regard to the range of measurement, equilibration time and cost (Bulut et al. 2000). The present study also offers an oppor- tunity of comparing measurements of soil suction using different experimental methods. Due to space limitation, only the measurements of matric suction, using the axis-translation and con- tact filter paper techniques, are presented for all four types of artificial soil investigated in this study. To- tal suction measurements from psychrometer and non-contact filter paper tests are presented for a sin- gle soil type in order to highlight the importance of the calibration curve chosen for the interpretation of the non-contact filter paper results. 2 METHODS OF SUCTION MEASUREMENT 2.1 Axis-translation Technique A modified standard triaxial cell was used for the experiments using the axis-translation technique. The modification of the cell consisted in fitting a high air entry porous disc into the cell pedestal. The high air entry porous disc was initially saturated and a small water back-pressure was subsequently ap- plied. An oven-dry sample was placed on top of the porous disc with a small weight over it so that a good contact was achieved between the soil and the porous disc. The cell was then closed and filled with little water at the bottom (not in contact with the sample) to ensure that the atmosphere around the soil was close to 100% relative humidity. Air pres- sure was applied inside the cell to give progressively decreasing values of suction as the sample was sub- jected to a wetting path. By monitoring the water volume flowing into the sample, the corresponding change in water content could be calculated. Once the water content had equilibrated under a given suc- tion, the air pressure was reduced to the next level and the sample was left to equilibrate under the new suction. Paper published in: S.C. Walker, D. Gallipoli, D.G. Toll (2005). The effect of structure on the water retention of soil tested using different methods of suction measurement. In Proceedings Advanced Experimental Unsaturated Soil Mechanics (Advanced Experimental Unsaturated Soil Mechanics): 33-39. Trento, Italy, 27-29 June 2005, Balkema Publishers, ISBN: 978-0-415-38337-4 The effect of structure on the water retention of soil tested using different methods of suction measurement S.C. Walker, D. Gallipoli & D.G. Toll University of Durham, School of Engineering, Durham. ABSTRACT: The paper presents an investigation into the retention properties of artificial bonded and de- structured soil samples of different void ratios using different methods of suction measurement. The axis- translation technique and the contact filter paper method are used for measuring matric suction whereas the non-contact filter paper and the transistor psychrometer are used for measurements of total suction. The main objective is the investigation of the effect of soil structure and inter-particle bonding on the soil water reten- tion properties of a granular material. An additional objective is the study of different experimental techniques for soil suction measurement and the effect of these different techniques on the definition of soil water reten- tion relationship. Most of the results presented in the paper refer to measurements of matric suction. However, some results from total suction measurements are also presented to show the importance of the choice of the calibration curve for non-contact filter paper tests.