Difficulties in the interpretation of a recent IPAC result for the g-factor of the 589 keV state of ~ ~ 7In prompted us to remeasure its half-life. The value, T1/2 = 192(16) p s quoted in the literature, which is the average of three independent and well agreeing measurements, was not confirmed. The half-life was found to be at least one order of magnitude smaller. This result solves the puzzle of the g-factor.PACS: 27.60; 23.20.C; 21.10.KThe recently determined g-factor of the 3/2-state of 117in at 589 keV g3/2(589 keV) = + 0.068 (39) [1] led to severe difficulties when an interpretation was tried within the assumed nuclear structure. Our result favoured the interpretation of the state as the first rotational state built up on the single proton [301] 1/2-Nilsson orbit in a strongly deformed core of prolate shape. K. Heyde found out, however, that the spectroscopic factor calculated within this picture is not compatible with the existing experimental data [2]. Therefore this interpretation cannot be correct. Our quoted result for the g-factor relies on a value for the half-life of the state which was taken from the literature: T1/2(llVIn, 589 keV)= 192(16) ps.[3, 4].In spite of the fact that three independent measurements of this half-life exist which agree within the error bars given, some doubts arose finally about their correctness and we decided to remeasure this half-life. The now available fast BaF2 detectors should allow a very precise direct slope measurement by delayed coincidences with the 1,303 -273 keV 7-7 cascade populated in the decay of 1 lVCd"For the detection of the 1,303 keV y-quanta a large BaF2 scintillator of dimension 38mm x 25 mm length was installed whereas a smaller crystal, 15 mm ~x 10mm length, was used for the 273 keV radiation. A 14 mm lead shield in front of the large detector eliminates perturbances by backscattering. The single channel analysers were set with small windows on the photopeaks of the two lines. The energy resolution of both BaF2 detectors was 10% FWHM for 662 keV and thus comparable to that of NaI(T1) detectors. During the measurements both coincidence spectra were registered simultaneously in order to allow a precise background correction. The measured coincidence spectra looked very similar to those shown in Figs. 3 and 4 of [1]. Especially the coincidence spectrum gated with the 1,303 keV window exhibits a clean photopeak of the 273 keV 7-line. We derived from this spectrum that the background of unwanted coincidences amounts~ to 15_+ 3%. This background is essentially contributed by the prompt 564--1,433 keV cascade in the decay of la7mCd by the Compton background of the 564 keV line under the 273 keV photopeak.In Fig. 1 the time spectrum of the coincidences is shown together with a prompt curve which was measured in the same geometry without changing the windows of the single channel analysers by use of a 6~ source instead of 117Cd" Apparently our measurement does not confirm the half-life reported in the literature. The spectrum shows essentially a prompt curv...