In the current fire safety design of steel structures in Japan, joints between steel beam ends and steel girders using friction type high strength bolts (this joint type is referred to as “pin joint”) are designed as pinned, therefore the fire resistance of the steel beams is determined only by the sagging moment resistance at the mid-span. On the other hand, research activities in Europe show that hogging moment resistance at pin joints of steel beams may be somewhat expected at elevated temperature. Moreover, in the case of steel - concrete composite beams, the reinforcing bars fixed in the concrete slab on the steel girders with headed studs work effectively so that the hogging moment resistance at the pin joints may be further improved. Furthermore, the sagging moment resistance at the mid-span is improved by the composite effect with concrete with small temperature rise. Considering these effects, there is a possibility that fire protections of the steel beams can be omitted or reduced.
 This paper presents results of load-bearing fire tests of unprotected fully composite beams pinned with steel girders conducted for the purpose of clarifying the deflection behavior and the fire resistance. 3 specimens were used, and the main test parameters were end boundary conditions (simply supported, pinned with the girders) and joint specifications (number of bolts, fire protection of joints). (Simply supported beam and beam pinned with girders are referred to as “simple-beam” and “pin-joint-beam” respectively.) The load level was a third of the yield moment, and the heating was in accordance with ISO 834 standard curve.
 The main conclusions from the test results were as follows:
 1) The bending stiffness of the pin-joint-beams at ambient temperature was close to that of the simple-beam. On the other hand, the deflection was largely suppressed by the restraining effect against rotation at the pin joint at elevated temperature. This effect was exerted because the rebar worked effectively by tying the steel girder and the concrete slab with the headed stud. As a result, the time to reach ISO limiting deflection of the simple-beam No. 1 was 24 min, whereas that of the pin-joint-beam No. 2 and No. 3 were 61 min and 57.5 min respectively, the fire resistance was greatly improved.
 2) Bending collapse was not observed for the pin-joint-beams. On the other hand, shear buckling of the web and local buckling of the lower flange occurred at the ends of the beam, but it is inferred that the load-bearing capacity was not lost immediately after the shear yield of the web.
 3) No significant damage (such as bolt failure and plate tear-out failure) and no deformation (such as shear deformation of the bolt and bearing deformation of plate) at the pin joints were observed during heating and cooling process. On the other hand, immediately after stopping heating, as the ends of the concrete slab was damaged due to cooling shrinkage of the unprotected steel beam, the deflection sharply increased temporarily.
 4) The full-plastic moment calculated using the strength obtained by the coupon test was able to accurately evaluate the sagging moment resistance for fully composite beam where the temperature of the lower flange reached 800 °C.
 5) The moment resistance of the pin-joint-beam was discussed by using the hogging moment resistance at pin joint considering the rebar inside the slab calculated by the equation (4). As a result, the effect on the total moment resistance was dominated by the sagging moment resistance at the mid-span at ambient temperature, whereas the influence of the hogging moment resistance at pin joint tended to increase at elevated temperature.