The static relaxation program DEVIL has been updated to use N-body Finnis-Sinclair potentials. Initial calculations of self-interstitial and monovacancy formation energies confirm that the modified program is working correctly. An extra repulsive pair potential (constructed to leave the original fitting unaltered) overcomes some deficiencies in the published Finnis-Sinclair potentials. The modified potentials are used to calculate interstitial energies and relaxations in the b.c.c. transition metals vanadium, niobium, tantalum, molybdenum and tungsten. Further adaptation enables DEVIL to model dislocations running parallel to any lattice vector. Periodic boundary conditions are applied in the direction of the dislocation line, giving an infinite straight dislocation. The energies per unit length of two different dislocations are compared with experiment. A study of migration of point defects in the perfect lattice provides information on the mobility of interstitials and vacancies. The possible reorientation of split dumbbell interstitials in a migration step comes under scrutiny. The total energy needed to form and migrate an interstitial is compared with that required for a vacancy. The interaction between point defects and dislocations is studied in detail. Binding energies for both sclf-intcrstitials and monovacancies at edge dislocations are calculated for the five metals mentioned above. Formation energies of the point defects in the neighbourhood of the edge dislocation are calculated for niobium, and the extent of the regions from which the defects are spontaneously absorbed are found. For split dumbbell interstilials, the size and shape of the absorption region depends on the orientation of the dumbbell. Migration of both interstitials and vacancies into the absorption zone is studied; the presence of the dislocation has a particularly dramatic effect on vacancy migration. The results on absorption zones are related to the dislocation sink strengths vital in radiation damage theory.