This thesis focuses on fire resistance of 19th century cast iron framed structures.Based on material property data obtained from a comprehensive literature review, upper and lower bound relationships of the thermal and mechanical properties of 19th century fireproof floor construction materials have been derived. Because these materials have large variability, a sensitivity analysis has been undertaken to investigate the most effective ways of representing such variability. The sensitivity analysis has indicated that the elevated mechanical properties of cast iron should be reliably quantified. The thermal expansion of cast iron can be taken as equal to that of steel as in EN1993-1-2. Variabilities in other material properties have modest effects on fire resistance of cast iron structures and can be safely modeled according the Eurocode material models for similar modern materials (using thermal properties of modern steel for cast iron, using thermal properties of modern concrete for the insulation materials of cast iron structures). In order to resolve some of the uncertainties in mechanical properties of cast iron at elevated temperatures, a total of 135 elevated temperature tests have been performed, including tension and compression tests, transient state and steady state tests, tests after cooling down and thermal expansion tests. These test results have been used to establish the elevated temperature stress-strain-temperature relationships in tension and compression.Afterwards, calculation methods are developed to calculate the bending resistance of cast iron beams and compression resistance of cast iron columns at elevated temperatures. For cast iron beams, a fibre model has been developed to calculate elevated temperature moment capacity of cast iron beams in jack arch construction, taking into consideration non-uniform temperature distributions in the cross-section. The fibre model divides the cross section into a large number of fine layers and for a given curvature and neutral axis position calculates the strain, the temperature, the stress and the force of each layer. It has been found that under historically applied load, the fire resistance of such beams can be 60 minutes or higher. The Monte Carlo simulation method has been used to take into account the variabilities of important mechanical properties of cast iron at elevated temperatures; Young's modulus, 0.2% proof stress, ultimate strength, corresponding strain at ultimate strength and failure strain in tension and Young's modulus, proportional limit and 0.2% proof stress in compression. This has enabled material safety factors of 1.50, 2.50, 4.50 and 5.50 to be proposed for target failure probabilities of 10-1, 10-2, 10-3 and 10-4 respectively.For cast iron columns, a finite element model, built using the commercial software ABAQUS, has been used to examine the effects of changing different design parameters (column slenderness, member imperfection, cross section imperfection, degree of axial restraint, load factor and load eccentricity) on fire resistance of cast iron columns. Validation of the finite element model was by comparison of the simulation results against six fire resistance tests, three on unprotected and three on protected cast iron columns. The results of this numerical parametric study indicate that the fire resistance of cast iron columns is generally higher than that of modern steel columns because the applied loads on cast iron columns are lower and cast iron columns have thicker sections than modern steel columns. Comparison of the numerical parametric study results with the calculation results using the steel column design method in EN1993-1-2 has found that the EN 1993-1-2 calculation results are generally on the safe side.