Evidence for non-Gaussianity in the COBE DMR Four Year Sky Maps

We introduce and study the distribution of an estimator for the normalized bispectrum of the Cosmic Microwave Background (CMB) anisotropy. We use it to construct a goodness of fit statistic to test the coadded 53 and 90 GHz COBE-DMR 4 year maps for non-Gaussianity. Our results indicate that Gaussianity is ruled out at the confidence level in excess of 98%. This value is a lower bound, given all the investigated systematics. The dominant non-Gaussian contribution is found near the multipole of order $\ell=16$. Our attempts to explain this effect as caused by the diffuse foreground emission from the Galaxy have failed. We conclude that unless there exists a microwave foreground emission which spatially correlates neither with the DIRBE nor Haslam maps, the cosmological CMB anisotropy is genuinely non-Gaussian.