| Abstract: | <p>The dorylomorph g roup of a nts comprises the thr ee subfam ilies of a rmy<br />
ants (Aenictinae , Dor ylinae , E ci tonina e) together with the subfam ilies A e nic tog itoninae , Cer apachyinae , a nd L eptanilloidinae . W e d es cribe ne w morpholog ical charac ters and sy nthes ize d ata from the lite rature i n o rder to pres en t the first hy po thesisof ph ylo ge ne tic rela ti on sh ips amo n g a ll d or yl o m or ph g en era . T he se da ta inc lud e t hefirst available character information from the newly discovered male caste of Leptanilloidina e. We use d a nt tax a from Lept ani llina e, My rme ciinae , a nd the pone romo rph ( Po ne ri na e sensu lato) subfam ilies A mblyoponinae , E cta tom mi na e, andParaponerinae as outgroups. We scored a total of 126 characters from twenty-twoterminal taxa and used these data to conduct maximum parsimony and bootstrapanaly ses. T he s ing le most-parsim oniou s tre e and bootstra p res ults support a singleor igin of a rmy ants. T he Old W orld ar my ant g en us Dorylus forms a monophy leticgroup w ith the enigma tic ge nus Aenictogiton, w hich is curre ntly known only fr ommale s; the s econd Old W or ld arm y ant g en us Ae nictu s is siste r to this c lade. Thisres ult ge ne rate s the predicti on that fem ale s o f Aenic tog it on, w hen d isc ov ere d, will beobserved to possess the army ant syndrome of behavioural and reproductive traits.Th e monophy ly of the New W orl d arm y ants (Ecitoninae ) is s upporte d ve rystrong ly, and within this gr ou p the ge ne ra Ecito n, Noma myrm ex,and Labidusfor m a robust cla de. T he monophyly o f L eptanilloidina e is al so upheld. The subfamily Cerapachyinae appears paraphyletic, although this conclusion is not suppo rte d by str ong bootstr ap results. R elati onships a mong gener a o f C era pa chyinaesimila rl y are not re solv ed robustly , a lthoug h pa rsim on y results s ug ge st c la de s c onsisting of (Ac an th os ti ch us þ Cylindromyrmex)and (Cerapachys þ S p hi nc to my rm ex).We test ed for the effe ct of incompl etely know n taxa by conducting a secondaryanalys is in whi ch the two genera contai ning 50 % missing character data ( Aenictogiton and Asphinct anilloides ) wer e remove d. The strict consensus of the seventeen most-pars imoni ous trees from this seconda ry an alysis is poor ly resol vedoutsid e the army ants an d contain s no clades con flicting with the primary analys is.The position of Leptani lla shif ts from form ing the sister group to Lept anilloidi nae(without high bootst rap suppo rt) in the prim ary analysis, to falling within apolytom y at the base of the root of the dor ylomor phs when incompl etel y knowntaxa are remove d. This inst ability suggests that the placem en t of Leptani lla withinthe dor ylomor phs in our prim ary an alysis may be spurious .</p>
|
