Many researchers working on the phyloge? netic analysis of nucleotide or amino acid se? quences find it necessary, at some time, to build upon prior work. When this involves importing an alignment (and masks) that have been des? cribed and sometimes illustrated in a published paper but for which the sequences have not been deposited in a public database, there may be no alternative but to painstakingly retype the data, with all the possibilities that this offers for error. More commonly, the un? aligned sequences may be available from a se? quence database, but the awkward task of man? ual alignment reconstruction remains. Most conveniently, alignments may be available as computer files (1) on disk, by application to the corresponding author, (2) from an ftp or http site maintained by the corresponding author or his institution, or (3) from an alignment data? base. However, when received by these routes, an alignment (and masks) may well be in a file format that requires tedipus, error-prone wordprocessing adjustment to make it readable by a sequence alignment editor or phylogenetic analysis program. Moreover, of the three approaches, the first two are potentially eva? nescent and the third suffers from underuse and lack of file-format standardization (Stoesser, pers. comm.). Masks and weighting informa? tion, when presented PHYLIP-style (Felsen? stein, 1993) as a string of digits, offer a parti? cular difficulty; few alignment editors or phy? logenetic analysis programs that we have encountered can import them successfully. Given these difficulties, we propose that it should become standard practice in molecular phylogenetics for alignments, masks, and sup? plementary information to be made available in NEXUS file format by use of the EMBL align? ment database. NEXUS file format is an ASCII text-file for? mat that was designed to make sharing of data b tween programs as easy and flexible as pos? sible (Maddison and Maddison, 1992: p. 145; Maddison et al., 1997), Importantly, NEXUS file format provides unlimited flexibility, by the use of character partition and assumption blocks, for the identification of special cate? g ries of data such as masks, weights, second? ary structure features, etc., and comments can be freely added. NEXUS files are written auto? matically by some alignment editors, including GDE (Smith et al., 1994), MacClade (Maddison nd Maddison, 1992) and SeqApp (Gilbert, 1993), and most such programs also provide file-format interchange facilities. Alternative? ly, NEXUS files may readily be constructed hand from published descriptions (Swof? ford, 1993; Maddison et al., 1997). Many phylogeneticists already use or have access to pro? grams that read and write NEXUS files, and many more are likely to do so after publication of the new multicapable PAUP* (Swofford, pers. comm.). The existence of a sequence alignment data? base separate from the GenBank (Benson et al., 1998), EMBL (Stoesser et al, 1998), or other d tabases for individual sequences is little known, but would certainly be advantageous if were widely used and if its use were stan-