Lack of Evidence for Cospeciation Between Retroelements and Their Hosts Sophia Kossida, 1, * Paul H. Harvey, 1 Paolo M. de A. Zanotto, 2, ² Michael A. Charleston 1 1 Department of Zoology, Oxford University, Oxford OX1 3PS, United Kingdom 2 Natural Environment Research Council, Institute of Virology and Environmental Microbiology, Oxford OX1 3SR, United Kingdom Received: 18 February 1999 / Accepted: 1 October 1999 Abstract. Some literature is available on cospeciation and on reconstructing the phylogenetic relationships of retroelements, but relatively little consideration has been given to whether there is cospeciation between retroele- ments and their hosts. Here we address this problem in detail. We conclude that there is no significant evidence for cospeciation between retroelements and their hosts. This conclusion was reached by noting that the branch- ing order of the two phylogenies was no more similar than would be expected by chance. Key words: Cospeciation — Phylogenetic relation- ships — Retroelements Introduction Coevolution is mutual evolution; for example, hosts evolve defese mechanisms, whereas parasites develop adaptations to these defenses (Briggs and Johal 1994; Parleviliet 1985). If a parasite speciates whenever its host speciates and there are no independent speciations or losses, then prefectly congruent host and parasite phy- logenies result (Barker 1994; Hafner and Nadler 1988; Page 1996; Reed and Hafner 1997). If the same rate of evolution is assumed, then approxi- mately the same amount of change should be expected for the associated host and parasite lineages (Hafner and Page 1995). Therefore, there should be agreement not only of the branching order but of the length of the branches as well (Hafner and Nadler 1988). On the other hand, independent speciation of hosts and parasites, extinction or loss of parasites in other ways, or transfer of parasites from one host to another results in incongruent phylogenies. The term “sorting events” is used to describe when parasites become ex- tinct or are differentially sorted among hosts. There are two types of association any species of parasite may have with a given host; association by de- scent and association by horizontal transfer from one host species to another. The latter gives rise to incongru- ent phylogenies. Congruent phylogenies form when co- speciation has taken place, and this is the event in which we are interested. With more cospeciation, the two phy- logenies become more congruent, so to test coevolution we investigate the maximum number of cospeciation events implied by the phylogenies. To support cospeciation for a particular host–parasite assemblage, two requirements should be met: first, that the host and parasite phylogenies are derived indepen- dently and, second, that the similarity of the branching order between the two phylogenies is significantly higher than that expected by chance (Hafner and Nadler 1990). Different approaches have been tried to unravel host– parasite phylogenies (Brooks 1981; Brooks 1988; Page 1990; Page 1994a). For example, Hennig’s parasitologi- cal method converts a parasite phylogenetic tree into a matrix of binary characters by additive binary coding (Brooks 1981). These characters provide the basis on * Present address: Molecular and Cellular Biology Department, Har- vard University, 16 Divinity Avenue, Cambridge, MA 02138, USA ² Present address: Dep. Microbiologia Imunologia e Parasitologia, UNIFESP, EPM, Rua Botucatu n° 862, 8° andar, Sao Paulo, SP, Brasil, CEP 04023-062 Correspondence to: Sophia Kossida; e-mail: skossida@fas.harvard.edu J Mol Evol (2000) 50:194–201 DOI: 10.1007/s002399910021 © Springer-Verlag New York Inc. 2000