Segments in glossiphoniid leeches, such as Helobdella triserialis , are the products of stereotyped cell lineages that yield identifiable cells from first cleavage. Cell lines generating segmental tissues are separated from those generating prostomial tissues early in development. Segments arise from five bilateral pairs of longitudinal columns of primary blast cells that are generated by five bilateral pairs of embryonic stem cells called teloblasts . There are four ectodermal cell lines (N, O, P and Q) and one mesodermal cell line (M) on each side of the embryo. In normal development, each cell line generates a segmentally iterated set of identified definitive progeny comprising a mixture of cell types. In the M, O and P cell lines, each blast cell generates one segment's worth of definitive progeny (segmental complement). But the clones of blast cells in each of these three cell lines interdigitate longitudinally with cells of the adjacent clones from the same line, so that the clone of an individual m, o and p blast cell is distributed across more than one segment. Thus, there is no simple clonal basis for morphologically defined segments. In the N and Q cell lines, two blast cells are required to produce one segmental complement of definitive progeny; in each of these two cell lines, two classes of blast cells (nf and ns, qf and qs) are produced in exact alternation. Primary n and q blast cells are about the same size and are produced at the same rate as blast cells for the o and p bandlets, but the longitudinal extent of their clones is roughly half that of the o and p blast cells' clones. During division of the blast cells, the n and q bandlets become compressed relative to the o and p bandlets, so that the segmental complements of the different cell lines can come into register. This compression movement is manifest as a movement of n and q bandlets relative to o and p bandlets in the posterior portion of the germinal band. The number of true segments in leech is fixed at 32; the counting mechanism is not known, but several hypotheses have been disproved. Segmentation in annelids and arthropods differs extensively at the cellular level, yet these phyla are presumed to share a common segmented ancestor. One strategy to identify homologous processes in annelid and arthropod segmentation is to compare the patterns of expression of evolutionarily conserved, developmentally important genes. Preliminary observations using a cross-reacting antibody that is thought to recognize a highly conserved region of a Drosophila segmentation gene, engrailed , labels nuclei of some blast cells early in development and, later, some neurones in the differentiating suboesophageal ganglion.