TITLE: GCN GRB OBSERVATION REPORT NUMBER: 3915 SUBJECT: GRB 050904: Environmental Constraints DATE: 05/09/04 14:36:07 GMT FROM: Daniel E. Reichart at U.North Carolina D. Reichart reports on behalf of the UNC team of the FUN GRB Collaboration. We have computed preliminary error bars on the temporal and spectral indexes of Haislip et al. (GCN 3914). The temporal index is -1.20 +/- 0.17 and the spectral index is -0.35 +/- 0.33. This disfavors the ISM-RED and WIND-RED cases at the 2.2 sigma confidence level, the ISM-BLUE case at the 1.3 sigma confidence level, and the WIND-BLUE case only at the 0.3 sigma confidence level (e.g., Sari, Piran & Narayan 1998; Chevalier & Li 2000). If the WIND-BLUE case is borne out by further observations, this suggests an electron-energy distribution index of p = 1.9 +/- 0.2, as well as a massive-star origin. If the NIR afterglow is strongly extinguished, and the intrinsic spectrum is even shallower than -0.35, or even positive, then none of the standard cases hold. This also suggests that the afterglow is probably not strongly extinguished, and supports the conclusion of Haislip et al. (GCN 3914) that the sharp spectral break between the J and i bands is likely due to dropout. Furthermore, given that the K to J spectral index is so shallow, and the J to i spectral index is so steep, the redshift range is probably narrower than that implied by the J and i filter centers (5.3 < z < 9.0), and is probably between 6 and 8. Careful modeling will better constrain this, but z, Z and/or Y observations, preferably contemporaneous with addition NIR observations would be most useful.