To our knowledge, the earliest work concerning W decay was done by Flegenheimer et al. in 1963. They determined the half-life of W to be 11min. A maximum energy of around 1.4MeV for W was obtained. In 1965, Kauranen and Ihochi studied W through the Os(n, ) W reaction. Its half-life was determined to be 11:5 0:3min. An end-point energy of 2:5 0:2MeV for the -spectrum of W was obtained. Its ten rays with energies of 94, 130, 178, 222, 258, 360, 417, 550, 855, and 955 keV were found. A cascade of two rays at 222 and 258 keV was observed. In ref. 3, Mirzadeh et al. investigated W decay via the Os(n, ) W reaction. Through the measurement of singles spectra, they observed two rays with energies of 260.4 and 421.7 keV from W decay and got their relative intensities. In 1997, Yang et al. made a singles -ray spectrum measurement by using an HPGe detector. Twenty two rays which can be assigned to the decay of W were found. Eighteen of these rays were observed for the first time. Subsequently in 2003, Wu and Niu gave a sample decay scheme (Fig. 1) including two rays of W at 222.0 and 260.4 keV based on the -ray data of Yang et al. and the reaction data of Hirning et al. But until now, no precise coincidence measurement for W rays was made. The present work is to check the coincidence relations inferred from previous work. The measurement of rays from W decay was taken with X– and – coincidence methods in our experiments. The (n, ) and (n, ) reactions were proved to be efficient ways for producing W in previous work. So in this work, we selected the Os(n, ) W reaction to produce W. The present experiments were performed using 14MeV neutrons from the 600-kV Cockcroft–Walton accelerator at the Institute of Modern Physics, Chinese Academy of Sciences. 14MeV neutrons were produced by the reaction of TiT targets with deuterons. W activities were produced using irradiation of isotopically enriched Os metallic powder of 100mg/cm with 14MeV neutrons. Osmium targets each were irradiated for 30min to fit the 10.7min half-life of W. And then they were transported into a lead-shielded room by an improved rabbit system. The measurement started 20 s after the end of irradiation with a planar HPGe detector (for X-ray and low energy -ray measurement) and a clover detector which consists of four coaxial N-type Germanium detectors. The former has an energy resolution of 0.6 keV for the 122 keV line of Co, an active diameter of 32mm, and a sensitive depth of 10mm; each of the latter has a 25% efficiency and an energy resolution of 2.1 keV for the 1332 keV line of Co. The two detectors were placed face to face on both sides of the source in the lead-shielded room. The measurement lasted 30min to fit the half-life of W. The count rate of single detector was about 100/s in the experiment. The coincidence window width of 100 ns was chosen. Thus the infection of the chance coincidence was not significant. The procedure mentioned above was repeated many times to improve the counting statistics. (X)-ray singles events and three parameter coincidence (X)– –t were recorded with a Multi-Parameter Data Acquisition System, where t was the time of each event after the beginning of a counting period. During the irradiation, several radioactive isotopes of Os, Re, and W were produced by (n,2n), (n, ), (n,p) and (n, ) reactions, respectively. It was not possible to observe rays from W decay through singles spectra because of the lower W yield and the large backgrounds from other nuclides produced in the experiments. The X– coincidence measurement was made so as to obtain rays from W decay. A part of the -ray spectrum in coincidence with 61.1 keV Re K 1 and 59.7 keV Re K 2 X rays is presented in Fig. 2.