|Abstract:||This paper aims at offering firsthand pieces of evidence that can support the moraic conception of geminates in Moroccan Arabic. The evidence presented comes from word minimality, compensatory lengthening and word formation. First, we argue that the existence of GV words (e.g. ddi 'to take'), to the exclusion of CV words (e.g. *di), in MA proves that a geminate is underlyingly moraic, in that a monovocalic word with an initial geminate is qualified as bimoraic. We also claim that the fact that MA involves a case of compensatory lengthening that leads to the creation of geminate consonants represents additional evidence in support of the moraic representation of geminates. This is because that compensatory lengthening has been shown to be mainly a moraic preserving process across many languages. The third type of evidence we introduce consists in demonstrating that a moraic approach to consonantal length proves to be the best option to account for morphological gemination in MA and elsewhere thanks to the cross-linguistic generalizations it allows.