Fits of theoretical spectra from Monte Carlo radiation-transfer calculations to dips at « 20 and 40 keV in a spectrum of the "classical" /-ray burst GB 880205 give best-fit values and 68%-confidence intervals £-(1.71±0.07)xl0 12 G, A^e-(L2±0.6)xl0 21 electrons/cm 2 , and ^ ~cos0-0.31 ±0.05, where 6 is the viewing angle relative to the field. Physical self-consistency fixes the temperature and Tc^5.3-M keV. These results suggest that this /-ray burst and many others which exhibit a lowenergy dip originate from strongly magnetic neutron stars and are galactic in origin.PACS numbers: 98.70.Rz, 95.30.Jx, 95.85.Qx, 97.60.Jd y-ray bursts were discovered over twenty years ago, 1 yet their nature and origin have remained a profound mystery. 2 Mazets et al. 3 reported single low-energy (E « 50 keV) dips in the spectra of more than 20% of the bursts observed with the Konus experiment on Venera 11-14. Hueter 4 reported single low-energy dips in several bursts observed with the A4 experiment on HEAO-1. These features have been interpreted as due to absorption at the cyclotron first harmonic (fundamental), h (OB « 11.6Z?i2 keV, where Bn is the magnetic field strength in units of 10 12 G. They constitute the strongest evidence in favor of a magnetic-neutron-star origin for y-ray bursts. 2 The statistical significance of these low-energy dips is low. Moreover, low-energy dips can be an artifact of the spectral deconvolution and their properties depend sensitively on the assumed form of the continuum spectrum. 5 Consequently, the existence and interpretation of these dips, and therefore the magnetic-neutron-star model of y-ray bursts, have remained controversial.Low-energy dips have recently been seen during certain time intervals in the spectra of two of the y-ray bursts detected with the burst detector on the Ginga satellite. 6 These dips are better resolved than previously: They span six or eight detector energy channels, not just one, as in the Konus data. 3 Furthermore, the spectra of both bursts show two dips-not just one, as in the Konus (Ref. 3) and A4 (Ref. 4) data-and the dips are nearly harmonically spaced at « 20 and 40 keV, as is expected for cyclotron features. Ginga could confidently identify line features with energies as low as 15 keV, 6 lower than in previous y-ray observations, making it ideal for detecting first-harmonic cyclotron lines for /?« 10 12 G. The single features at 40-60 keV reported by Konus (sensitive to E^30 keV) could be second-harmonic lines, analogous to the lines at «40 keV seen in GB 870303 and GB 880205. Finally, the continuum spectra of both bursts are characteristic of "classical" y-ray bursts. In the case of GB 880205, PVO data show that there is substantial emission above 1 MeV, 7 as is typical of classical bursts.Two aspects of the dips might seem puzzling. First, the dips are narrow, indicating that the scattering medium is much cooler than the typical photon energy E r^\ MeV in the continuum spectrum. Second, the strengths of the dips at the first and second harmonics are similar, d...