Holomeric vs. meromeric segmentation: a tale of centipedes, leeches, and rhombomeres.

Explaining the origin and evolution of segmentation is central to understanding the body plan of major animal groups such as arthropods, annelids, and vertebrates. One major shortcoming of current views on segmentation is the failure to recognize the existence of two layers of segmentation. I distinguish here holomeric segmentation, involving the whole body axis (or the whole axis of an appendage) and producing "true" segments (eosegments); and meromeric segmentation, producing merosegments within one or more eosegment(s). In terms of developmental mechanisms, meromeric segmentation is probably the same as compartmentalization. This process follows two rules: (1) merosegments are formed from a stereotyped pattern of subdivisions, where only the merosegments in contact to the anterior or posterior boundary of the eosegment are allowed to divide; (2) contiguous eosegments undergoing meromeric segmentation generate merosegments according to identical lineage patterns apart from possible lineage truncation in one or a few terminal eosegments. The segmentation model proposed in this paper is mainly supported by evidence from comparative morphology, but it is compatible with known cellular and developmental mechanisms. The development of vertebrate rhombomeres, the annulation of leeches, the subdivision of the distal part of insect antenna into flagellomeres and the segmentation of centipedes are interpreted here in terms of meromeric segmentation. Some of these phenomena, like centipede segmentation, have thus far defied all attempts at an explanation, both in mechanistic (developmental) and phylogenetic terms. The model presented in this paper suggests a rich research agenda at all levels, from molecular and genetic to morphological and phylogenetic.