A　new　method　and　instrument　were　developed　for　in　situ　measurement　of　the　gas　diffusion　coefficient　(D-g)　and　gas　coefficient　of　permeability　(k(g))　of　unsaturated　soil.　Air　containing　tracer　gas　of　volumetric　concentration　of　5%-10%　was　continuously　pumped　into　a　point　source　at　a　constant　flux　rate　(q(0))　until　approaching　steady　state.　Then,　the　pseudo-steady-state　tracer　gas　concentration　(C-r0)　and　gas　gauge　pressure　(Delta　Pr0)　were　recorded　at　a　distance　(r0)　away　from　the　point　source.　The　Dg　and　kg　were　determined　through　the　Cr0-q0　and　Delta　Pr0-q0　relationships,　respectively.　Soil　column　tests　(silty　sand)　and　field　tests　(sandy　lean　clay)　were　conducted　to　validate　the　new　method.　The　in　situ　measured　Dg　and　kg　were　comparable　with　those　of　undisturbed　specimens　measured　by　a　two-chamber　apparatus.　The　in　situ　measurement　of　Dg　and　kg　ranged　from　2x10-8　m2　s-1　to　4x10-6　m2　s-1　and　1.9x10-8　ms-1　to　3.9x10-7　ms-1,　respectively.　The　measurement　accuracy　increased　at　deeper　depth,　due　to　reduced　boundary　effect　at　soil　surface.　The　measurement　of　in　situ　Dg　was　more　susceptible　to　the　boundary　effects　than　that　of　kg　at　shallower　depth　(i.e.,　20　cm　deep)　in　general.　The　field　test　results　showed　that　the　ratio　of　in　situ　Dg　(or　kg)　to　that　measured　by　the　element　tests　was　generally　between　1.2　and　3.3　at　or　below　depth　of　0.30　m,　due　to　the　in　situ　heterogeneity　of　soil.