The pattern of cleavage of the shoulder and arm muscles has been described only in a handful of taxa (urodeles, lizards, turtles, marsupials and birds); although the information from these seminal investigations has been used as a tool for understanding homologies across amniotes, that understanding is limited by restricted taxonomic scope and by the imaging technologies then available. Half a century later, we have new tools for studying and visualizing developing anatomy. We studied the development of closely spaced embryonic series of mammals, archosaurs, lizards and turtles — a sample comprising all major amniote clades — and studied the embryology of forelimb muscles, along with the nerves and skeleton, by using fluorescent immunostaining and confocal microscopy. Our data reveal that muscle division is much more conserved across amniotes than previously described. We tracked and followed the embryonic origin of each adult muscle, comparing it to supposed homologues across clades. Based on our results, we propose a reconsideration of some assumed homologies and provide new information regarding the development and evolution of amniote forelimb musculature. In addition, we compare this developmental pattern with that of amphibians and non-tetrapods, revealing drastically different modes of development at the origin of tetrapods and shedding light on the sequence of events in the evolution of paired appendages.