The giant fiber system (GFS) of Drosophila melanogaster provides a convenient system in which to study neural development. It mediates escape behaviour through a small number of neurons, including the giant fibers (GFs), to innervate the tergotrochantral jump muscle (TTM) and the dorsal longitudinal flight muscles. The GFS has been intensively studied physiologically in both wild-type and mutant flies, and is often used as a system to study the effects of neural mutations on the physiology of the adult nervous system. Recently, much information has been gleaned as to how and when synaptogenesis, with its major target neurons, is achieved. However, little is known of the earlier development of this neuron. Here we have used an enhancer-trap, marking parts of the GFS, in conjunction with BrdU labelling, to attempt to follow the birth, axonogenesis, and the early morphological meeting of the GFs with their target neurons. From these anatomical observations we propose that the GF cell is not born during the larval or pupal stages and, therefore. appears to be a persistent embryonic cell. The axons of the GFs develop during the third instar. During the early pupal stages the GFs contact other identified neurons of the GFS. In addition, we see some aberrant development of the network, with some flies carrying only one GF, and yet others with extended axons. We present a model for the initial joining of the GFs and tergotrochanteral motorneurons (TTMns).