Beta-decay properties of neutron-rich 3=Br isotopes confirm the predicted smooth onset of quadrupole deformation for Z<37 already below N=60. The observed increase of the energy of the first 2 + state in the N=56 nucleus ~2Kr may indicate octupole softness.During the last decade, extensive studies have been devoted to the level structures of neutronrich nuclei around A=IO0.For ee-isotopes in this region, already the behaviour of the first 2 + levels (see Fig. i) indicates the well-known transition from spherical to deformed shapes as neutrons are added beyond N=58, with an extremely sharp transition for 4oZr and ~aSr (see, e.g. [1,2]). The suddeness of this transition has been interpreted as an effect of the spherical Z=40 shell gap which can locally reinforce the N=56 spherical gap (see the 2 + peak for ~Zr) to delay the shape transition until the Z=38, N=60 gaps at large deformation can reinforce each other. As the proton number moves away from 40, the influence of the spherical N=56 subshell should quickly disappear. While for the heavier elements (4zMo to 46Pd) the E(2+), indeed, decrease more gradually indicating a smooth onset of deformation, so far no experimental data were available for lighter elements (34Se, s6Kr) with N>54. The region around Z=36, N=56 is of particular interest because of the predicted octupole softness [2][3][4]. With the halogen beams available at the mass separators OSTIS (ILL-Grenoble) and ISOLDE (tERNGeneva), we have started a systematic investigation of ~-decay properties of heavy Br isotopes in order to check the above predictions for Z<37 nuclei. At both facilities UC-graphite targets were used which were connected to negative-surface ion sources with a Lab6 ionizer [5,6]. Spectroscopic information was obtained from multiscaling of delayed neutrons and B-particles to derive T~/2 and P~ values, and from Y-singles and YT-coincidence measurements to establish partial decay schemes. The most surprising result is the observation of an increase of E(2 +) in 92Kr by about 60 keV relative to ~~ indicating a gap at N=56 for Z=36 (see Fig. i). On the one hand, this increase appears small compared to the 830 keV in the double semi-magic 96Zr; on the other hand, however, the effect in 92Kr is remarkable when regarding the E(2 § trends in the isotones with Z=(40• and Z=(40+4). Already in 98Mo and ~4Sr the increase of the E(2 § has with 9 keV, respectively 22 keV become very small; and in the 4~Ru isotopes, one observes a decrease of the E(2 § by about ii0 keV when going from N=54 to 56. Based on this behaviour the effective N=56 gap in 92Kr has to be considered even stronger than the apparent increase of E(2+). One is now forced to question why a subshell gap exists in ~2Kr while it has more or less disappeared in the other N=56 isotones around 9~Zr. For the double semi-magic Q6Zr there is consistent interpretation that the spherical gaps in both the proton and neutron systems give a particularly low shell energy for the spherical shape [1,2]. With the proton shell structure domin...