The first excited state of BIDP was shown in a previous communication to exhibit an ultrafast decay in fluid nonpolar, nonprotic solutions due to the presence of a S-1/S-0 conical intersection (CI). In frozen polar and nonpolar glasses a strong fluorescence was observed, rationalized by the hindering of the internal torsion required to reach the geometry of the Cl. Complete analysis of the data was hampered by son-le unusual observations in nonpolar glasses. In this paper we show that they can be explained by assuming dimer formation, with a formation constant of K-eq = (4 +/- 3) x 10(5) M-1 at 83 K and Delta H-dim = 7 +/- 2 kcal/mol. A complete analysis of the spectra is presented, and fluorescence quantum yields of the monomer and dimer are reported. Efficient self-quenching is found, with a Stern Volmer constant, K-SV = (1.5 +/- 0.1) x 10(6) M-1, assigned to static quenching. The dimer absorption spectrum was extracted from the data and is compared to Kasha's exciton model and to quantum chemical (QC) calculations. The basic features of exciton splitting are reproduced by quantum mechanical calculations, but complete quantitative agreement of the QC computations with the experimental results is not attained. The previous analysis of the monomer spectra using the displaced harmonic oscillator model is extended to the more demanding conditions prevailing at cryogenic temperatures. The derived Delta H-dim is in good agreement with other dimers formation enthalpies and with the quantum mechanical calculation presented. The new analysis corrects tau(f) in MCHIP to 2.9 x 10(-13) s, somewhat smaller than the value reported in polar solvent in a previous communication, thereby strengthening the assumption that polarity can reduce the efficiency of Cl.