Presently there is no simple, or generally acknowledged, way to apply fracture mechanics to low-strength, "structural", steels. These materials are used in heavy structures as ships, boiler vessels, off-shore drilling rigs, and pipe lines. Hence, the need for fracture analyses is pressing, and considerable research is going on extending more or less well-founded non-linear fracture mechanics theories to very ductile materials and a wide range of structure sizes. In many cases the efforts to give experimental support to the theories are made with specimens and materials where conditions and properties deviate only slightly from those for valid standard tests, or where the test conditions are highly artifical. In the present work it has then been thought important to perform a few series of fracture mechanics tests on an everyday material and with specimens simulating a real structure on one hand and representing the group of laboratory test specimens on the other hand. Then ambiguities and inconsistencies can be revealed and food for thoughts about improved procedures can be obtained. The work comprises tests on large center-cracked specimens and on standard three-point bend specimens. The tests are evaluated with theories presented in the literature during the last decade using the J-integral concept with some of its extended interpretations. The report comprises three separate parts of which this is the first one, dealing with the series of large scale tests. The second part gives results from the three-point bend tests, and the last part is a theoretical study of the J-concept in fracture mechanics in which the tests results are discussed critically.