Minto FUKUHARA Investigation for practical application of direct tensile test method for fragile solid materials such as soil Yutaka FUKUMOTO Tensile cracks in soil structures are mainly caused by tensile stress, which can have significant geotechnical and environmental consequences. Therefore, it is necessary to develop a method for quantifying tensile forces in soil materials. However, due to the fragile nature of soil, conventional direct tensile testing methods are difficult to apply. To address this, our research group has investigated a direct tensile testing method applicable to fragile solid materials such as soil. By utilizing vacuum technology, we demonstrated that this method could be applied to various soil samples. In this study, we conducted fundamental investigations to ensure the reproducibility of the testing method for practical use. Specifically, we examined test conditions, including vacuum pressure, membrane coverage length, compression force for verifying adhesion between the cap and specimen, membrane edge treatment, and loading rate. The results indicated that setting vacuum pressure higher than the specimen strength was effective, and an optimal pressure range was identified. A membrane coverage length of at least 20 mm for an 80 mm specimen height ensured stable results. Moreover, an appropriate range of compression force was necessary to prevent poor adhesion. Membrane edges wrapped around the specimen provided more consistent results, highlighting their impact on test outcomes. A loading rate of 1 mm/min or lower produced reproducible results. Additionally, the specimen properties affected the test results. The presence of surface and internal voids influenced the vacuum pressure effect, with higher void content leading to an expanded vacuum pressure area. This suggests that specimen characteristics affect vacuum formation.