Serial sections of 99 human embryos from Carnegie stages 8–23 were investigated and 38 graphic reconstructions were evaluated. At stage 9 somite 1 is of appreciable size and is separated from the otic disc, as also in the next several stages by rhombomeres and pharyngeal arches 3 and 4, thereby differing from the chick. At stage 10 somite 1 begins to differentiate into sclerotome and dermatomyotome. At stage 11 spinal neural crest begins to develop. At stage 12 parts of somites 1–4 are being transformed into the hypoglossal cell cord. It is stressed that the numbers of somites present at stages 9–12 are part of the definition of those stages. At stage 13 dense and loose zones begin to be detectable rostrally in the sclerotomes and also, although out of phase, in the perinotochord. Spinal ganglia begin to develop in phase with the somites. At stages 14–16 the maximum number of somites observed was 38–39 rather than 42–44, as usually given. Moreover, they did not extend to the tapered end of the trunk, which is not a (vertebrated) ‘tail’. At stages 17–23 the maximum number of centra was 38–39, including coccygeal vertebrae 4–5. Although most of the somites appear during primary development, all of the spinal ganglia develop during secondary development (stages 13–18). The number of ganglia was at a maximum of 35 at stage 18, but was reduced to 32 already by stage 23. Important points confirmed in this study are that the number of occipital somites in the human is four, and that the level of final closure of the caudal neuropore is future somite 31, which represents approximately future sacral vertebra 2. The interpretation of relevant neural tube defects is discussed in the light of the findings. The ascensus of the conus medullaris during the fetal period is well established, but a concomitant ascent of the situs neuroporicus is proposed here, and has implications for defects that involve tethering of the spinal cord. The main results are integrated in comprehensive graphic representations of the levels and the interrelationships of (a) somites and centra, and (b) somites, neural crest, and spinal ganglia. These may aid in the elucidation of some frequently occurring anomalous conditions.