If assumption of homogeneity of regression slopes is significant in *ANCOVA*analysis, what to do?

Any help would be much appreciated!