Theory Biosci. (2002) 121: 297±320 Ó Urban & Fischer Verlag http://www.urbanfischer.de/journals/theorybiosc From Haeckel to event-pairing: the evolution of developmental sequences Olaf R. P. Bininda-Emonds 1,2 Jonathan E. Jeffery 1 , Michael I. Coates 3 , and Michael K. Richardson 1 1 Institute of Evolutionary and Ecological Sciences, Kaiserstraat 63, P.O. Box 9516, Leiden University, 2300 RA Leiden, The Netherlands 2 Current address: Lehrstuhl fu Èr Tierzucht, Technical University of Munich, Alte Aka- demie 12, 85354 Freising-Weihenstephan, Germany 3 Department of Organismal Biology & Anatomy, University of Chicago, 1027 East 57th Street, Chicago, Illinois 60637-1508, USA Address for correspondence: Olaf R. P. Bininda-Emonds, Lehrstuhl fu Èr Tierzucht, Technical University of Munich, Alte Akademie 12, D-85354 Freising-Weihenstephan, Germany, e-mail: olaf.bininda@tierzucht.tum.de, phone: +49(0)8161713741; fax: +49(0)8161713107 Received: May 5, 2002; accepted: June 21, 2002 Key words: Developmental sequence, Haeckel, event-pairing, Normal Table, phylo- geny Summary: Development involves a series of developmental events, separated by trans- formations, that follow a particular order or developmental sequence. The sequence may in turn be arbitrarily subdivided into contiguous segments (developmental stages). We discuss the properties of developmental sequences. We also examine the differing analytical approaches that have been used to analyse developmental se- quences in an evolutionary context. Ernst Haeckel was a pioneer in this field. His ap- proach was evolutionary and he introduced the idea of sequence heterochrony (evo- lutionary changes in the sequence of developmental events). Despite the availability of detailed developmental data (e.g. Franz Keibel's `Normal Tables'), Haeckel was un- able to undertake a quantitative analysis of developmental data. This is now possible through computer-based analytical techniques such as event-pairing, which can extract important biological information from developmental sequences by mapping them onto established phylogenies. It may also yield data that can be used in phylogeny re- construction, although the inherent `non-independence' of the data may make this in- valid. In future, the methods discussed here may be applied to the analysis of patterns of gene expression in embryos, or adapted to studying gene order on chromosomes. Introduction Embryonic development is a continuous process; the embryo possesses an ever-changing morphology. This continuum may be analysed by treating it as a series of discrete developmental events, where each event is the first 1431-7613/02/121/03±297/$ 15.00/0