This paper presents the observations of the study on arsenic removal from a contaminated ground water (simulated) by adsorption onto Fe3+ impregnated granular activated carbon (GAC-Fe). Fe2+, Fe3+ and Mn2+ have also been considered along with arsenic species in the water sample. Similar study has also been done with untreated granular activated carbon (GAC) for comparison. The effects of adsorbent dose, particle size of adsorbent and initial arsenic concentration on the removal of As(T), As(III), As(V), Fe2+, Fe3+ and Mn2+ have been discussed. Under the experimental conditions, the optimum adsorbent doses for GAC-Fe and GAC have been found to be 8g/l and 24g/l, respectively with an agitation time of 15h. Particle size of the adsorbents (both GAC and GAC-Fe) has shown negligible effect on the removal of arsenic and Fe species. However, for Mn removal the effect of adsorbent particle size is comparatively more. Percentage removal of As(T), As(V) and As(III) increase with the decrease in initial arsenic concentration (As0). However, the increase in percentage removal of all the arsenic species with decrease in As0 are less for higher value of As0 (3000-500ppb) than those of the lower value of As0 (500-10ppb). The % removal of As(T), As(III), As(V), Fe, and Mn were 95%, 92.4%, 97.6%, 99% and 41.2%, respectively when 8g/l GAC-Fe was used at the As0 value of 200ppb. However, for GAC these values were 55.5%, 44%, 71%, 98% and 97%. The pH and temperature of the study were 7+/-0.1 and 30+/-1 deg C, respectively.