Rapods during their emergence from water to land. We therefore suggest that the 117793 neural crest population found in mouse rather reflects a secondary broadening of the neural crest diversity that occurred in mammals. New shoulder elements such as the endochondral clavicle, a part of the scapular spine, and the sternal manubrium appear, which represent apomorphic characteristics of the Theria [3,11], and which mosaically evolved in primitive mammals [21,22]. These anatomical mammalian innovations could receive new contribution from neural crest rather than ML240 site co-opting cells from the former dermal skeleton. This idea supports the view that the neural crest proper is an evolving entity and that the number of derived cell types may change during the evolution of vertebrates, with some cell types appearing de novo and some disappearing in particular lineages [15,23]. For example, the population of neural crest cells, which gave rise to the cleithrum and other dermal bones of the primitive shoulder girdle, has disappeared completely in the axolotl, i.e., neural crest cells were neither found as separate dermal bones nor as cartilage or connective tissue derivatives at the muscle attachment sites. At the same time an evolutionarily younger population of neural crest cells, which later gave rise toLack of Neural Crest in the Axolotl ShoulderLack of Neural Crest in the Axolotl ShoulderFigure 2. Results of grafting one short left neural fold fragments. a, Schematics demonstrating orthotopical grafting of a short left GFP+ neural fold fragment (including neural crest) into a white (d/d) host. The graft is extirpated from a GFP+ neurula (green, stage 16) and extends from a prospective posterior head to an anterior trunk region. It is implanted into a white host where a similarly sized fragment was extirpated previously. b and c, left flank of white hosts 1 day (b) and 3 days (c) after the operation. In vivo visualization of GFP+ neural crest cells at an anterior trunk level where they migrate laterally from the top of the neural tube; arrows show the main direction of migration. d , two months old juvenile carrying a short GFP+ neural fold fragment. No neural crest cells were present in the scapula, or elsewhere in the shoulder girdle. However, all other neural crest derivatives located at this level were GFP+. d, left side of operated juvenile where cranial and ventral margins of the GFP negative shoulder girdle are visible through the transparent skin. Girdle cartilage is outlined with a dashed line. e, ventral aspect of the juvenile. Gills, nerve fibres in the limb, pigment cells, heart and enteric ganglia are clearly GFP+, while the ventral halves of the cartilaginous coracoid plates (indicated with the dashed line) are GFP negative. f, enlarged area of the scapula framed in (d). Only spinal nerves of the brachial plexus appear GFP+. The cranial margin of the scapula is marked with white arrowheads. No GFP+ cells are detectable along its cranial margin, where muscles exist that attach it to the skull. g, h, transverse sections through the juvenile (sectioning planes see (f)) with GFP+ spinal nerves but GFP negative scapular cartilage and connective tissue. i , sagittal sections through the shoulder girdle region in a 1.5 month old juvenile from dorso-medial (i, scapula tip as in h) to ventro-lateral (l, glenoid region). Anti-Myosin heavy chain-rhodamine immunostaining only in i, for better visualization of GFP+ cells. Note GFP+ staining in all secti.Rapods during their emergence from water to land. We therefore suggest that the neural crest population found in mouse rather reflects a secondary broadening of the neural crest diversity that occurred in mammals. New shoulder elements such as the endochondral clavicle, a part of the scapular spine, and the sternal manubrium appear, which represent apomorphic characteristics of the Theria [3,11], and which mosaically evolved in primitive mammals [21,22]. These anatomical mammalian innovations could receive new contribution from neural crest rather than co-opting cells from the former dermal skeleton. This idea supports the view that the neural crest proper is an evolving entity and that the number of derived cell types may change during the evolution of vertebrates, with some cell types appearing de novo and some disappearing in particular lineages [15,23]. For example, the population of neural crest cells, which gave rise to the cleithrum and other dermal bones of the primitive shoulder girdle, has disappeared completely in the axolotl, i.e., neural crest cells were neither found as separate dermal bones nor as cartilage or connective tissue derivatives at the muscle attachment sites. At the same time an evolutionarily younger population of neural crest cells, which later gave rise toLack of Neural Crest in the Axolotl ShoulderLack of Neural Crest in the Axolotl ShoulderFigure 2. Results of grafting one short left neural fold fragments. a, Schematics demonstrating orthotopical grafting of a short left GFP+ neural fold fragment (including neural crest) into a white (d/d) host. The graft is extirpated from a GFP+ neurula (green, stage 16) and extends from a prospective posterior head to an anterior trunk region. It is implanted into a white host where a similarly sized fragment was extirpated previously. b and c, left flank of white hosts 1 day (b) and 3 days (c) after the operation. In vivo visualization of GFP+ neural crest cells at an anterior trunk level where they migrate laterally from the top of the neural tube; arrows show the main direction of migration. d , two months old juvenile carrying a short GFP+ neural fold fragment. No neural crest cells were present in the scapula, or elsewhere in the shoulder girdle. However, all other neural crest derivatives located at this level were GFP+. d, left side of operated juvenile where cranial and ventral margins of the GFP negative shoulder girdle are visible through the transparent skin. Girdle cartilage is outlined with a dashed line. e, ventral aspect of the juvenile. Gills, nerve fibres in the limb, pigment cells, heart and enteric ganglia are clearly GFP+, while the ventral halves of the cartilaginous coracoid plates (indicated with the dashed line) are GFP negative. f, enlarged area of the scapula framed in (d). Only spinal nerves of the brachial plexus appear GFP+. The cranial margin of the scapula is marked with white arrowheads. No GFP+ cells are detectable along its cranial margin, where muscles exist that attach it to the skull. g, h, transverse sections through the juvenile (sectioning planes see (f)) with GFP+ spinal nerves but GFP negative scapular cartilage and connective tissue. i , sagittal sections through the shoulder girdle region in a 1.5 month old juvenile from dorso-medial (i, scapula tip as in h) to ventro-lateral (l, glenoid region). Anti-Myosin heavy chain-rhodamine immunostaining only in i, for better visualization of GFP+ cells. Note GFP+ staining in all secti.