Chicken myoblasts and myotubes attach to, and elongate upon, solid substrates only in the presence of exogenous fibronectin derived from, for example, serum or fibroblasts ( M. Chiquet, E. C. Puri, and D. C. Turner, 1979, J. Biol. Chem. 254, 5475–5482). We have sought answers to two questions: (1) whether myoblasts depend on exogenous fibronectin for attachment because they themselves synthesize none or too little of this glycoprotein; and (2) whether interaction of myogenic cells with fibroblast-derived fibronectin might guide muscle morphogenesis. Under conditions in which fibroblasts synthesized fibronectin and released it into the medium, myoblasts and myotubes synthesized and accumulated only small amounts of fibronectin. Although proliferating myogenic cells treated with 5-bromodeoxyuridine accumulate fibronectin on their surfaces, other evidence suggests that normal myogenic precursor cells, like myoblasts and myotubes, synthesize little fibronectin: (1) the accumulation of newly synthesized fibronectin in untreated primary myogenic cultures is attributable, at least in part, to the fibroblasts present; and (2) 1-week-old myogenic clones fail to form a fibronectin network, in contrast to fibrogenic clones of the same age and cell density. In cryosections of chick embryos, less fibronectin was found in the premuscle masses than in the surrounding mesenchyme. The splitting of the muscle anlagen in the limb bud was correlated with the appearance of fibronectin-accumulating cells in the cleavage furrows. In vitro, myoblasts displayed a marked tendency to align themselves along oriented fibrils or streaks of purified fibronectin that had been deposited on the substrate; this myoblast alignment influenced shape and branching of the developing myotubes. These results suggest that formation of myotubes in a certain spatial arrangement during muscle morphogenesis may be regulated by a fibronectin-containing matrix produced by connective tissue cells.