The ability to target specific cytoskeletal components in axons for disruption within intact developing embryos would provide a valuable tool for studying neuronal development. Neurofilaments are an attractive target for such an approach, because they are neuron specific and are expressed late in embryogenesis principally beginning during axon outgrowth. No pharmacological agents are currently available that disrupt neurofilaments without also affecting general development. One approach that has been used successfully to affect proteins in vivo is to inject specific antibodies into living cells. We employed this approach in Xenopus laevis embryos by injecting two antibodies directed against the middle molecular weight neurofilament protein (NF-M) into a single blastomere of a two-cell stage embryo. Injected antibodies could be detected for as long as 3.5 days in cells descended from the injected blastomere. Only cell bodies of neurons descended from anti-NF-M-injected blastomeres contained abnormal accumulations of intermediate filament proteins, and peripheral nerve development was unilaterally retarded in these neurofilament antibody-injected tadpoles. Such accumulations and peripheral nerve defects were not seen in neurons derived from uninjected blastomeres or from blastomeres injected with control antibodies. These data demonstrate the usefulness of specific antibodies to perturb neuronal development in intact frog embryos and, in addition, suggest a role for neurofilaments in axon elongation.