Recently, adhesive bonding methods have been applied for joining of several structural members. To extend the application of adhesively bonded joints, it is necessary to define fracture criteria for the joints. In particular, fatigue failure criteria are crucial for adhesive joints as mechanical parts in cyclic loads via vibration and transmission of power. Furthermore, most of the bonded joints have complex multiaxial stresses in the adhesive layer; hence fatigue fracture criteria for the joints should be based on the fatigue strength under multiaxial stress conditions.In a previous study [1], we proposed a method for evaluating the fatigue strength under multiaxial stress condition using adhesively bonded butt, scarf and butterfly type butt joints, where the stress multiaxiality in the adhesive layer was varied in a wide range. These joints with thin adhesive layer thickness have considerable uniform stress distributions except at the free end of the adhesive layer. However, an increase in the adhesive layer thickness enlarges the range having nonuniform stress distributions, resulting in lowering of the adhesive strength. To obtain critical fatigue strength of the joints under multiaxial stress conditions, the effect of adhesive layer thickness on the fatigue strength must be quantified.Some studies have been published on the effect on the static and fatigue strength of the butt joint [2][3][4][5]; however, there are few studies on the scarf and butterfly type joints [6,7]. Furthermore, the effect of adhesive layer thickness on the fatigue strength of these joints has not been investigated. In the present study, fatigue tests under multiaxial stress conditions were conducted for the three types of butt, scarf and butterfly type butt joints in a variety of scarf angle and adhesive layer thickness.Int ~u r n o¢ Fracture 80 (1996) R70 An 0.5% C Carbon steel (JIS.S55C) was used as the adherend for butt, scarf and butterfly type butt joints, and the adhesive was a thermosetting epoxy adhesive (Toyoda Gosei: EA9432NA).The shape and sizes of the butt and scarf joints used in this study are shown in Fig. 1, where the scarf angle 0 was of four values: 300,450,60 ° and 90 ° (butt joint). As shown in this figure, the width of the adherend was varied in accordance with the scarf angle 0 so as to maintain a constant adhesive area of 190 mmh Figure 2 illustrates the details of the butterfly type butt joint together with its loading frame. This basic geometry was proposed by Arcan et al. [8]; the adhesive area was also the same as that of the scarf joints. The butterfly type butt joint was assembled into the loading frame by six pins. Then, the assembled loading frame was fixed with two pins into a fatigue testing machine. In this study, the fatigue tests were conducted under conditions where the center of the adhesive layer was located on the loading line (0---0°).The adhesively bonded joints were prepared as follows: the bonding surfaces of the adherends were polished with emery paper of grade 180 mesh under a dry condi...