Phylogenetic Relationships Among Eighteen Neotropical Culicini Species

Journal of the American Mosquito Control Association, 16(2):75-.g5,2OOO Copyright @ 20OO by the American Mosquito Control Association, Inc.

PHYLOGENETIC RELATIONSHIPS AMONG EIGHTEEN NEOTROPICAL CULICINI SPECIES

JUAN-CARLOS NAVARRO aNp JONATHAN LIRIA

Instituto de Zoologia Tropical, Facultad de Ciencias, Laboratorio de Biologia de Vectores, Universidad Central de Venezuela, Apartado 47058, Caracas I04I_A, Venezuela

ABSTRACT. The subgeneric relationships among 8 generic and infrageneric taxa of the Culicini rribe were examined by cladistic analysis based on 30 larval mouthpart characters (maxillae and mandibles) of7 subgenera of Culex L. and I of Deinocerites Theobald. We analyzed 18 ingroup species as well as Deinocerites melanophylum Dyar and Knab as a sister group, and Aedes taeniorhynchus Wiedemann as an outgroup. A parsimony analysis using the Nona program resulted in 2 trees each of lO9 steps (consistency and retention indices : 0.88j. The topologies obtained were similar to the current classification of the tribe, based on nonexplicit methods mainly including adult characters, with 2 exceptions. In the present work, the monophyly of the tribe Culicini (Culex * Deinoceites) was supported by 4 synapomorphies. The subgenus Lutzia Theobald formed the most basal clade in the tribe Culicini and the placement of Phenacomyra Harbach and Peyton as a subgenus was validated by its location as a sister group of the subgenus Culex and other subgenera. The subgenus Carrollia Lutz was the most robust taxon, supported by 5 synapomorphies, and was congruent with the infragroups of the current classification. The relationships arnong Deinocerites, Anoedioporpa Dyar, Microculex Theobald, and Melanoconion Theobald were unresolved, but were placed in the most internal clade of the tribe. The lst exception to the accepted classification was the poorly resolved boundary between Anoedioporpa and Microculex The 2nd was the strong support (with 1 I synapomorphies) for the inclusion of Deinocerites as a subgenus of Culex in the Culicini, which is proposed here.

KEY \ilORDS cladistics, Culex, Cnlicidae, Deinocerites, larval mouthparts, mandibles, maxillae, phylogeny

INTRODUCTION natural classification of the Culicidae. Cladistics us- Since the pioneering work of Edwards (1932) ing molecular characters has been used by Pape (1992) and Dyar (1923, 1928) the taxonomy and classifi- with chromosomal characters in Anopheles (Cellia): (1992) cation of the family Culicidae has been based on Wesson et al. with Aedesa Besansky (1994) classical adult and immature morphologic charac- et al. with Anopheles gambiae, and Miller (1996) ters. Although the use of holomorphology in sys- et al. with the Pipiens Complex. All of these tematics inference (Hennig 1966) is preferred, re- studies have used ribosomal DNA. However, Be- (1997) cent authors have found that larval mouthparts sansky and Fahey estimated phylogenetic provide good morphologic characters for alpha tax- relationships among 14 Culicidae species using the gene. onomy (Harbach and Peyton 1993) and their use in white-eye phylogenetics should be explored. Using classical morphologic characters and cla- (1996) Snodgrass (1959) was the lst to describe the distic analyses, Judd studied the tribe Sa- bethini and (1998) morphology of some mosquito mouthparts and Harbach and Kitching analyzed 34 genera Knight (1971) showed the structural diversity in the of Culicidae. However, despite conclu- mandibles of several genera. Later, Harbach and sions about the higher relationships in the family, (infrageneric) Knight (1977) reported a variety of maxillary struc- the internal relationships remain tures and shapes that offer additional diagnostic largely unresolved. characters for identifi cation. We used larval mouthpart characters to infer Harbach and Peyton (1993) resumed studies with phylogenetic relationships among taxa within the some species of the tribe Sabethini and recognized tribe Culicini. We examined species belonging to 7 the importance of these structures for the identifi- of 8 subgenera reported from Venezuela, and 13 cation of generic taxa; they suggested their use to that occur in the Neotropics. We also evaluated the power achieve more natural classifications. Recently, Pe- and importance of larval mouthparts in the rez and Navarro (1996) reported diagnostic char- cladistic analysis in obtaining a natural classifica- acters for 3 subgenera of Anopheles Meigen, and tion for this medically important taxon. concluded that morphology of mouthparts is an additional tool for identification of these taxa. MATERIALS AND METHODS Despite extensive alpha taxonomy studies of Cu- licidae, phylogenetic relationships have not been Source of specimens.' We used 4th-stage-larvae intensively studied. The large size of the family belonging to the Mosquito Collection of the La- (3,000 species) (Knight and Stone 1977) represents boratorio de Biologfa de Vectores, Museo de Bio- a challenge. The new taxonomy described by Mun- logfa of the Universidad Central de Venezuela stermann and Conn (1997) represents a powerful (LBY after Guimardes 1997). These specimens approach toward obtaining a more objective and came from a variety of different breeding sites (Ta- JoURNALor rns AuerrcAN MosQUno Coxrnol AssocmrroN Vol. 16,No.2

Table l. List of species examined with data of locality in Venezuela, date of collection, and breeding site.

Species Locality Date Breeding site

Culex (Carrollia) bihaicolus Loma de Hierro, Aragua State Feb. 1992 Heliconia aurea Cx. (Car.) rausseoi Loma de Hierro, Aragua State Feb.1992 Palm spathes Euterpe sp. Cx. (Car.) iridescens Hacienda Rfo Claro, Zulia State Aug. 1995 Palms spathes Euterpe sp. Cx. (Car.) urichi La Azulita, M€rida State Dec. 1981 Xanthosoma sp. Cx. (P he nac omyia) c orni ge r Panaquire, Miranda State Dec.198l Cacao husks Cx. (Andoedioporpa) bam- Panaquire, Miranda State Feb. 1986 Bamboo internodes borum Cx. (Culex) dolosus La Azulita, Mdrida State March 1995 Discarded tire/rockhole Cx. (Cux.) nigripalpus El Jobo (Sinamaica), Zulia State Ocr. 1997 Tiee hole at ground level Cx. (Cux.) coronator Instituto de Zoologia Tropical UCV, Aug. 1997 Artificial container Caracas, D.E Cx. (Cux.) quinquefosciatus Cementerio del Sur, Caracas, D.E Jan.1981 Artificial container Cx. (Microculex) microphy- Guanay Tepui, Amazonas State Feb. 1995 Brocchinia tatei lus Cx. (Mcx.) chryselatus Sierra de San Luis, Falc6n State Aug. 1993 Guzmania mucronata Cx. (Mcx.) pleuristriatus Cerro Santa Ana. Fa1c6n State Aug. 1993 Aechmea aquilega Cx. (Melanoconion) albi- Hacienda Las Nubes, Zulia State Oct. 1997 Ground pool with aquatic plants nens$ "grapo Cx. (Mel.) atratus" Haciendas Las Nubes/Rfo Claro, Oct. 1997 Ground pools Zulia State Cx. (Mel.) nicceriensis Haciendas Las Nubes/Rfo Claro, Oct. 1997 Ground pools Zulia State Cx. (Lutzia) bigoti Sierra San Luis, Falc6n State April 1994 Discarded tire Deinoce rite s me lanophylum Cayo Borracho, Falc6n State Jan.1983 Crab hole Aedes taeniorhynchus Guajira-Paraguaipoa, Zulia State Oct. 1986 Ground pools in mangroves

ble l). A total of 19 species was examined. These species), and also Aedes taeniorhynchus Wiedmann belonged to the subgenera Anoedioporpa Dyar (l and Deinocerites melanophylum Dyar and Knab. species), Microculex Theobald (3 species), Lutaia Mounting techniques and terminology: Larvae Theobald (l species), Culex L. (4 species), Melan- were stored in 8OVaethanol and clarified in a lOTo oconion Theobald (3 species), Carrollia Lltz (4 KOH solution, then processed as described by Har- species), and Phenacomyia Harbach and Peyton (1 bach and Peyton (1993), with the modifications of Perez and Navarro (1996). A total of 30 characters was scored. These consisted of 16 characters associated with mandibles and 14 characters of the maxillae (Appendix 1). Characters were scored based on the nomenclature of Harbach and Knight (1980). The generic and subgeneric abbreviations followed those of Reinert (1975). Drawings of the morphometric characters used are shown in Fig. 1. A schematic of the general morphology from each supraspeciflc taxon examined is shown in Figs. 2a (Lutzia), 2b (Phenacomyia), 3a (Culex), 3b (Car- rollia), 4a (Deinocerites), 4b (Melanoconion), 5a (Anoedioporpa), and 5b (Microculex). Selection of characters and cladistics analysis: Characters were selected based upon the results of previous papers that also used mouthpart structures (Knight 1971, Harbach and Knight 1977, Harbach and Peyton 1993, Perez and Navarro 1996). We also included new characters (character [ch.] 13, PMnL, and ch. 15, ppMAdA) not previously studied based on structural diversity (Appendix 1). Polymorphic (multistate characters) were also included in the data set. Characters not determined in one or more taxa were treated as missing. Fig. 1. Morphometric measurements for some man- The genus Aedes Meigen (Ae. taeniorhynchus) dibular and maxillary structures. was chosen as outgroup to root trees but without JUNE2000 PHyr-ocnNsrrc RpLarroNsrnpsrr..i CwrcrNr 77 ,ffi,

Fig.2. a. Ventral (left) and dorsal (right) view of Cu- lex (Lutzia) bigoti. Mandibles (above) and maxillae (below) are shown. b. Ventral (left) and dorsal (righr) view of Culex (Phenacomyia) corniger. Mandibles (above) and maxillae (below) are shown.

Fig. 3. a. Ventral (left) and dorsal (right) view of Cu- the intention of exploring the sister relationships of lex (Culex) dolosus. Mandibles (above) and maxillae (be- Culicini. Choice of this taxon as outgroup is sup- low) are shown. b. Ventral (left) and dorsal (right) view ported by the sister relationships of Culex reported of Culex (Carrollia) bihaicolus. Mandibles (above) and by Miller et al. (1997), Pawlowsky et d. (1996), maxillae6elow) are shown. Besansky and Fahey (1997), and Harbach and Kitching (1998). Deinocerites melanophylum was used as sister group of Culex in agreement with penter 1993). The Nona program (Goloboff 1996) current and accepted classifications. was used to search for the most parsimonious clad- All characters were coded as either binarv or ograms using the Tiee Bisection Reconnection multistate (Appendix l) and all characters were (TBR) heuristic algorithm. Fifty random additions treated as unordered (Appendix 2) (Nixon and Car- were completed using the MULT*5O option. The 78 Jor,nNnr- oF THE AMERICAN Moseurro Col.Irnol AssoctATIoN Vol. 16,No.2 a

\\l i\ ! ,\ i Hr\\ \ 'f=- *q.,h I d- *t ilP,l\1lr r&.l.> t\\ r t\\ =J\--"--J

1,,.. I

I ?.f i h\ Ir\\ I iI ""N\ _---J Fig. 4. a. Ventral (left) and dorsal (right) view of Ventral (left) (right) view of Deinocerites melanophylum. Mandibles (above) and max- Fig. 5. a. and dorsal (Anoedioporpa) bamborura. Mandibles (above) and illae (below) are shown. b. Ventral (left) and dorsal (right) Culex maxillae (below) are shown. b. Ventral (left) and dorsal view of Culex (Melanoconion) nicceiensis. Mandibles (right) view (Microculex) Mandibles (above) and maxillae (below) are shown. of Culex chryselatus. (above) and maxillae (below) are shown. JuNr 2000 PnvlocrNprrc RlLerroxsgIps w Cur-rcrNr

Ae. (O ch.) E en lo h yn c h u t Node a: Culicini Cx. (Lut.) blEotl Cx. (Phc.) comlg.r The tribe Culicini has never been in doubt. since Cx. (C ux.) qu laq uef.a c latut Edwards (1932) included it as I of the 3 mosquito Cx. (Cux) doloeus tribes (Anophelini, Megarhini [Toxorhynchitini], Cx. (Cux.) ntgrtp.tpu. and Culicini), and Belkin (1962) listed it as 1 of lO Cr, (Cux.t corcntu. tribes. However, the internal arrangements of this Cx. (Cer.) blh.tcolus tribe have not been well studied. Cx- (Cer.) r.uss.ol The monophyly of Culicini (Culex and Deino-

Cx. (C.r.) utlct l cerites) is supported by 4 synapomorphies (Fig. 7): -+ Cx. (Cer-) lrldscens posterior dorsal teeth (ch. 1: 1 0, changing from De. melanophylum occulTence with 1 tooth and 3 smaller accessories

Cx. (And-) b.mbqum to subsequent absence or presence with different -+ Cx. (M.1.) dunDl shapes), mandibular seta no. I (ch. 4: I 0, with Cx. (Mel.) .lblnerals loss of this character), mandibular spiculose area

Cx. (i,,'!et.) n lcce.tdttt (ch.24:0 -+ 3, change from 5 spiculae to 9, 13,

Cx- Ucx.) mtcrephytus loss in Culex quinquefasciatus Say, and indepen-

Cx. (Ncr. ) pleu rltU.lut dent occurrence in Culex nigripalpus T"heobald), Cx. (ilcx.) chryacttu. and galeastipital stem (ch. 29: 0 -+ I, with loss of this character). Culex bigoti Bellardi (subgenus Lutzia), whicll

Bdsmp v€lueshigher thsn 50% are shwn has enlarged maxillary setae 8, 9, 10, and 11 (ch. 22: O -+ 1), is the basal taxon and sister to the other Fig. parsimonious 6. Strict consensus tree of 2 most taxa (node b). This arrangement agrees with the solutions. traditional classification sensu Belkin (1962) ot Lutzia as a specialized lineage that shares similarity and hence ancestry with members of the subgenus JUMP*l,2 option was used to perform branch Culex. T\is result is also in concordance with the swapping on all trees with a difference of I or 2 topology of Miller et al. (1996) who used sequence steps relative to the most parsimonious tree. Sup- variation in the internal transcribed spacer of ribo- port for individual derived branches was evaluated somal sequences. Both results and the geographic by bootstrapping with l,@O replications, after de- distribution of the subgenus imply that Lutzia rep- leting the autapomorphic characters (Carpenter resents an ancient Gondwanian lineage. 1996). The options hold* and mult*20 were spec- ified such that each tree was searched for by ran- Node f dom additions and all of the most parsimonious trees were retained in memorv. Among the internal clades, Culex corniger Theo- bald (subgenus Phenacomyia) formed a sister group of node f (ch. l0: | -+ 2, occurrence of labula de- RESULTS AND DISCUSSION veloped, and ch. 28: O --> 1, maxillary palpus normal) that is located at the node c of Culex (slb- Two trees, each with lO9 steps, were detected. genus) species: (Cx. quinquefasciatus + Cx. Both trees had consistency indices (CIs) reten- and dolosus Aribalzaga t Cx. nigripalpus -t Cx. co- tion indices (RIs) of 0.88 (887o). Permutation of ronator Dyar and Knab). Within this node c (ch. l: suboptimal trees and random addition did not iden- 3 -+ 2, sequence and size of the Mn ventral teeth; tify additional cladograms. The strict consensus and ch. 23: O -+ 1, shape of maxillary body, with (with cladogram bootstrapping values) and I of the reversion in Carrollia, Deinocerites, and Culex most parsimonious trees (with synapomorphic char- pleuristriatus Theobald) was found the following acters) are shown in Figs. 6 and 7, respectively. arrangement: Cx. quinquefasciatus (widespread The strict consensus tree (Fig. 6) derived from species) formed a sister clade to 3 neotropically re- the 2 most parsimonious cladograms had an unre- stricted species (node d), with C.r. dolosus regrre- solved node a with species of Melanoconion, Mi- senting the most related species to Cx. nigripalpus croculex (both internally resolved), Anoedioporpa, and Cx. coronator (node e). Although the Neotrop- and Deinocerites. Tbre remaining basal taxa oc- ical species node is not well sul4rorted, the place- curred within resolved clades. Bootstrap values ment of Cx. quinqucfasciatus afid Cx. nigripalpus > 5O7o are also shown in Fig. 6. The highest boot- is also in agreement with the topology of Miller et strap score was found for the subgenus Carrollia al. (1996), suggesting the validity of both cladistic (997") and its internal groups, followed by Melan- hypotheses and also suggesting the Gondwanian or- oconion (59Vo) and the node that included the sub- igin of the Pipiens Complex. genera of Culex (Carrollia lDe. + And. * Mel. + On the other hand, the basal position of Phena- Mcx.l) (527o). comyia in the tree supports the proposal of Harbach 80 JounNrl oF THE AMERTCaNMoseuno CoNrnol AssocrerroN

11 4f 160 zf 24o xr (Och.) tacnlorhynchus

o3 s2 tol ttl tl1 tz4 tat p2 ztl Cx. (Lut) bigoti i3 r5o Cx. (Phc.) comlger

Cx. (Cux. ) q,llnguefascJafus

Cx. (Cux.) dofosus

Cx. (Cux.) nlgrlpaFus

Cx. (Cux.) @ronator

Cx. (Car.) brhaicolus

Cx. (Car.) rausseoi

Cx. (Car.) urichi

to z+3 Cx. (Car.) irtdtscens l z2 l2 sI e s6 toi tt3 tz6 r2tt3 tg2 t Dc. nrelanophytum

193 Cx. (Mel.) dunnt

03 n3 l9l f51 p1 Cx. (Mel.) albrnensis

Cx. (Mct.) n ccenbnsis

173 I Synapomorphlcs z?2 nl Cx. (And.) bemborum r Autapomorphies Cx. (Mal) microphylus ll Homoplaslcs 17s5 rsoztl 10- q6iladqlState 3{ n3 n2 Cx. (Mcx. ) pleuristn'atus

a,b, c,otc.. Nodes Cx.(Mcx.) chryseratus

Fig. 7. One of the most parsimonious trees, showing the synapomorphies, some autapomorphies, and homoplasies.

and Peyton (1992) to split these species (Cx. cor- Nodesj and k niger * Culex lactator Dyar and Knab and Culex airozai Lane) from the subgenus Culex and srug- The subgenus Culex node is a sister clade of the gests a possible close affinity with Old World spe- completely Neotropical node j (ch. 15:. 2 -+ l, re- cies not examined in this work. However, the au- duction of posterior projection of aductor apodeme tapomorphic characters ch. l0 (0) and 28 (O) not Mn; and clr..2O:O -+ 1, position of seta 4-Mx with reported by these authors should be considered in subsequent change to state 2 in the Urichi Group the future for subgenus diagnosis. and reversion in node k) where subgenus Carrollia JUNE 2OO0 PHyLocENETrc RELATIoNSHTPS rN CuLrcnir 8l

(node g) is the most basally related to the unre- clude that the subgenus Anoedioporpa should be solved node (k) with MeL 'f Mcx. 4 And. -t Dei- studied further to clarify the relationships with Mi- nocerites. Within the Carrollia node. 2 sister taxa croculex species in search of the natural classifi- form the node h (Culex bihaicolus Dyar and Nunez- cation of the genus. Tovar * Culex rausseai Cova-Garcia, Sutil, and The 2nd and the most important exception to the Pulido) and the node i (Culex urichi Coqtillett + cunent classification is the paraphyletic position of Culex iridescens (Lutz)) as a monophyletic group Deinocerites melanophylum, with internal place- and a sister clade of the remaining taxa (node k). ment in the analysis and suppolted by 11 synapo- The monophyly of Canol/ia is strongly supported morphies suggesting that it is a subgenus within the by 5 synapomorphies(ch. 1, 3,8, 12,25) and a gents Culex (sensu lato). These results are in agree- 99Vo bootstrap value, possibly representing the ment with Mallampalli (1995), who placed Deino- more recent ancestor of the fully Neotropical Calex cerites cancer Theobald and Galindomyia leii subgenera. Stone and Barreto-Reyes in the most internal clade The node k is supported by 2 synapomorphies of tl;le Culex topology, using mainly adult charac- (ch. 22: | -+ 2, enlargement of maxillary setae 8, ters. 9, 10, ll; and ch. 23: | -+ 2, shape of maxillary The hierarchic position of Deinocerires as a dif- body, with a regression on Cx. pleuristriatus). ferent genus in the tribe Culicini and family Culic- Deinocerites is the taxon most related to the sub- idae had been largely based upon autapomorphic genera Anoedioporpa, Microculex, and Melanocon- characters (9 included in the present study) (Theo- ion, but the analysis did not find any synapomor- bald 1901, Belkin and Hogue 1959, Adames 1971) phies to establish its sister relationships. Despite the that do not seem to represent synapomorphies and unclear relationships among these taxa, the internal therefore are misleading in estimating evolutionary relationships among Melanoconion and, Anoedi- relationships among other members of the tribe oporpa wit}r Microculex are supported by 2 and 3 Culicini or the Culicidae sensu Belkin (1962). "opinions synapomorphies, respectively. Belkin and Hogue (1959) commented re- In general, the most parsimonious cladograms in- garding the relationships of Deinocerites have var dicate very few homoplasic characters (Fig. 7) with ied widely," including it as separate subfamily, CI and RI : 0.88 (88Vo). These homoplastic char- subgenus, or the current genus level, by Mitchell, acters are 16, 17, 19,23, and 24, and the most Dyar, and Edwards, respectively. Additionally, homoplasic character change occurs in seta l-Mx Belkin and Hogue, aware of the close relationships (ch. 17), which occurs 4 times independently in the between Culex and Deinocerites, also said "this trees (CI : 0.66; RI : 0.71). This means that ch. lack of comparative studies has been a great hand- 17 is uninformative to explain evolutionary trends, icap in our attempt to determine the relationships and for species or taxa diagnosis. of Deinocerites" and finally accepted the genus status of this taxon. Later, in the last review of the genus, Adames Congruence with the current classification "Deinocerites (1971) concluded that is undoubt- With 2 exceptions, the tree topologies found are edly a member of the tribe Culicini" and reported congruent with the current classification, including 3 characters of adults not shared with Culex to sup- monophyletic clades for Lutzia, Phenacomyia, Cu- port the generic status without considering possibly lex, Carrollia, and Melanoconion. Also, the mono- homologous characters. In our case, Deinocerites phyletic Culex and Carrollia clades are consistent has 9 autapomorphic characters but also has 1 I syn- with the intuitive internal groups such as the basal apomorphic characters that place it within the most Pipiens Group in Culex, and Valencia's groups Bi- internal clade within Culex. haicolus (bihaicolus and rausseoi) and Urichi (iri- The strictly coastal distribution of Deinocerites descens and urichi) (Valencia 1973). However, 2 associated with crab holes and the internal position incongruences with current and accepted classifi- in the cladogram supported by ll synapomorphies "a cations occur in the clade k showing Anoedioporpa suggests a recent origin from stock of Culex "from in the same clade with Microculex and the genus subgenus," and not a common ancestral Deinocerites at the most internal branches of the stock which separated very early from the stock Culicini clade. that gave rise to the dominant genus Culex" sensu Anoedioporpa, a subgenus proposed by Dyar Adames (1971\. (1923), had later hierarchic taxonomic changes and Finally, no evidence from our analysis supports was subsequently included in the subgenera Melan- the generic status for Deinocerites in the tribe Culi- oconion and Tinolestes Coquillett. In the last revi- cini, at least in the current classification. In contrast, sion, Belkin (1968) raised Anoedioporpa to a sub- our phylogenetic evidence recognizes this taxon as genus comprised of 12 species (Berlin and Belkin another member of Culex sensu lato. Based on l98O), among them Culex restrictor Dyar and comments of Zavortink (1990) and Harbach and Knab, formerly included in the subgents Micro- Kitching (1998) and the studies of Judd (1996, cale.r. Nevertheless, our analysis did not permit us 1998a), a natural classification of Culicidae may be to propose a hypothesis about this group. We con- achieved by recognizing subordinate infrageneric 82 Jotnxel or rur AvenrceN Mosqurro CoNrnol AssocnnoN taxa as valid genera. This could be reached in the additional suggestions. This research was part of future using cladistic methods. Nevertheless, until multiple studies supported by CONICIT grant RP- the boundaries for the nominal subgenus are clearly VII-240060. We also thank CONICIT (Consejo Na- defined we propose the reduction of Deinocerites cional de Investigaciones Cientificas y Tecnologi- to subgenus status (Dei.). This proposal will be cor- cas-Recursos Humanos), Consejo de Desarrollo roborated by additional phylogenetic studies using Cientlfico y Humanfstico-RH-UC{ and the John D. molecular characters. and Catherine T. McArthur Foundation for support. Using a new methodological approach, with 30 larval mouthpart characters we obtained similar to- REFERENCES CITED pological cladograms to Mallampalli (1995), who used 67 mainly adult characters. We have demon- Adames AJ. 1971. Mosquito studies XXIV. A revision of strated that the larval structures and particulary the the crabhole mosquitoes of the genus Deinocerites. morphology of mouthparts are useful for the for- Contrib Am Entomol Inst (Ann Arbor) 7(2):l-154. mulation of evolutionary hypotheses. Belkin JN. 1962. The mosquitoes of the Sauth Pacific (Diptera: Culicidae) Volumes I and 2. Los Angeles, Judd (1998b) suggested that data from a single "will CA: University of California Press. life stage sometimes produce an erroneous Belkin JN. 1968. Mosquito studies (Diptera: Culicidae) arrangement of the taxa," in agreement with other IX. The type specimens of the New World Mosquitoes authors who followed phenetic methods (Rohlf in European Museums. Contrib Am Entomol Inst (Ann 1963). In our case for Culicini, the trees obtained Arbor) 3(4): 1-69. "congruent" are with those of Mallampalli (1995) Belkin JN, Hogue CL. 1959. A review of crabhole mos- and molecular data (unpublished). quitoes of the genus Deinocerites (Diptera: Culicidae). Calif Univ Publ Entomol 14:411-458. Berlin OGW, Belkin JN. 1980. Mosquito studies (Diptera: Evolutionary trends of breeding sites Culicidae) XXXVI. Subgenera Aedinus, Tinolestes, An- oedioporpa of Culex. Contrib Am Entomol Inst (Ann Our resolved phylogeny suggests that the ances- Arbor) 17(2):l-1O4. tral member of this group was a ground-pool-breed- Besansky NJ, Fahey T. 1997. Utility of the white gene in ing, filter-browsing organism that gave rise to 2lin- estimating phylogenetic relationships among mosqui- eages. Members of the lst lineage evolved toes (Diptera: Culicidae). Mol Biol Evol 14:442-454. structural reductions in the mandibular brush and Besansky NJ, Powell JR, Caccone A, Hamm DM, Scott development of mandibular teeth and a maxillar JA, Collins FH. 1994. Molecular phylogeny of the palpus that enabled predation (Lutzia type). The Anopheles gambiae complex suggests genetic intrc- 2nd lineage retained symplesiomorphic characters gression between principal malaria vector. Proc Natl allowing the independent colonization of special- Acad Sci USA 91:6885-6888. Cld- ized breeding places with mono- and dicotyledo- Carpenter JM. 1996. Uninformative bootstrapping. distics 12:177-181. neous phytotelmata in the most apomorphic envi- Dyar HG. 1923. On some of American subgenera of Czlex (e.g., The latter ronments Car. and Mcx.). (Diptera: Culicidae). Insecutor Inscitiae Mens ll:18'7- morphologic modifications could explain the large 190. radiation and diversification of some lineages (sub- Dyar HG. 1928. The mosquitoes of the Americas. Car- genera) with species-speciflc breeding site associ- negie Inst Wash Publ 387:l-616. ations, for example, Microculex and bromeliads Edwards FW. 1932. Genera insectorum. Diptera. Fam. (Frank 1983, Navarro et al. 1995). Culicidae Fascicle 194. Brussels, Belgium: Desmet- Despite the complexity of Culex, due to its large Verteneuil. and taxonomically problematic groups that have Frank JH. 1983. Bromeliad phytotelmata and their biota, Frank Lounibos LR eds. been poorly studied, this study addresses questions especially mosquitoes. In: JH, Phytotelmata: terrestrial plants as hosts aquatic in- about the monophyly of current Neotropical genera for sect communities Medford, NJ: Plexus Publishing. p (based on intuitive analysis), and in- and subgenera 101-128. dicates that the use of additional taxa and morpho- Goloboff PA. 1996. Nona. Noname ver. 1,80 (a bastard logic characters may clarify poorly supported son of Pee-Wee) 32-bit version Program and documen- groups and improve current classifications. tation. Computer program distributed by J. M. Carpen- ter. Department of Entomology, American Museum of Natural History, New York. ACKNOWLEDGMENTS Guimardes JH. 1997. Systematic database of Diptera of We thank Pablo Goloboff for advice on the im- the Americas south of the United States, family Culic' plementation of the Nona and PeeWee programs, idae 56o Paulo, Brazil: Fundacao de Amparo a Pesquisa do Estado de Sao Paulo/Ed. Pleide. and for his suggestions for the data analysis. Harbach RE, Kitching IJ. 1998. Phylogeny and classifi- Wilmer our faithful Thanks especially to M6ndez, cation of the Culicidae. Syst Entomol 23:327-370. collaborator in the field collecting trips for the LBV. Harbach RE, Knight KL. I9'l^1. Maxillae of fourth stage We are grateful to Antonio Machado-Allison and mosquito f arvae (Diptera: Culicidae). Mosq Syst9:455- Roberto Barrera for critiques and valuable improve- 477. ments in the I st version of the manuscript, and Harbach RE, Knight KL. 1980. Taxonomists' glossary of Scott C. Weaver and Robert Tesh for review and mosquito anqtomy Marlton, NJ: Plexus Publishing, Inc' Jure 20O0 PHyLocENETrc RELATIoNsHIpsIN Culrctr\r 83

Harbach RE, Peyton EL. 1992. A new subgenus of Culex quito disease vectors (Diptera: Culicidae): impact of in the Neotropical region (Diptera: Culicidae). Mosq molecular biology and cladistic analysis. Annu Rev En- Syst 24:242-252. tomol 42'.351-69. Harbach RE, Peyton E,L. 1993. Morphology and evolution Navarro JC, Ingunza J, Ferndndez Z, Barera R. 1995. of the larval maxilla and its importance in the classifi- Mosquitoes and bromeliads: species-specificselectivity cation of the Sabethini (Diptera: Culicidae). Mosq Syst patterns on the northern coast and southern Guiana 25:l-16. Shields in Venezuela. J Am Mosq Control Assoc ll: Hennig W. 1966. Phylogenetic systematics Urbana, IL: 345-346. University of Illinois Press. Nixon KC, Carpenter JM. 1993. On outgroups. Cladistics Judd DD. 1996. Review of the systematics and phyloge- 9:413-426. netic relationships of the Sabethini (Diptera: Culicidae). Pape T. (1992). Cladistic analysis of mosquito chromo- Syst Entomol 2l:129-15O. some data in Anopheles subgenus Cellia (Diptera; Cl- Judd, DD. 1998a. Review of a bromeliad-ovipositing lin- licidae). Mosq Syst 24:I-ll. eage in Wyeomyia and the resurrection of Hystatomyia Pawlowsky J, Szadziewski R, Kmieciak D, Fahrni J, Bittar (Diptera: Culicidae). Ann Entomol Soc Am9l:572-589. G. 1996. Phylogeny of infraorder Culicomorpha (Dip- Judd DD. 1998b. Exploring component stability using tera: Culicidae) based on 28S RNA gene sequences. life-stage concordance in sabethine mosquitoes (Dip- Syst Entomol 21:167-118. tera: Culicidae). Cladistics 14:63-93. Perez BX Navarro IC. 1996. Morfologia bucal de larvas (Diptera: Knight KL. l97l. Comparative anatomy of the mandible de Anopheles Culicidae) como herramienta of the fourth instar mosquito larva (Diptera: Culicidae). taxon6mica a nivel subgen6ico. Acta Zool Mex 68:13- J Med Entomol 8(D:189-2O5. 26. Reinert JE 1975. Mosquito generic and subgeneric Knight KL, Stone A. 1977. A catalog of mosquitoes of abbreviations (Diptera: Culicidae). Mosq Syst 7:105-1 the world (Diptera: Culicidae). Thomas Say Found En- 10. Rohlf FJ. 1963. Congruence tomol Soc Am 6:l-677. of larval and adult classifications in Aedes. (Diptera: Culicidae). Syst Zool 12:97- Mallampalli V. 1995. Phylogenetic relationships among 117. taxa in the genus Culex, and phylogenetic relationships Snodgrass RE. (1959). The anatomical life of the mos- of vectors of the world alphaviruses in the subgenus quiio. Smithson Misc Coll 139(8):l-87. Melanoconion (Diptera: Culicidae). Ph.D. dissertation. Theobald FV. 1901. The classification of mosquitoes. J University of Maryland, College Park, p. MD. 250 Trop Med 4:229-235. Miller BR, Crabtree MB, Savage HM. 1996. Phylogeny Valencia I. 1973. A revision of the subgenus Carrollia of of fourteen Culex mosquito species, including the Culex Culex XXXI. Contrib Am Entomol Inst (Ann Arbor) pipiens complex, inferred from the internal transcribed 9@):l-134. spacers of ribosomal DNA. lnsea Mol Biol 5(2):93- Wesson DM, Porter CH, Collins FH. 1992. Sequence and ro7. secondary structure comparisons of ITS rDNA in mos- Miller BR, Crabtree MB, Savage HM. 1997. Phylogenetic quitoes (Diptera: Culicidae). MoI Phylogenet Evol l: relationships of the Culicomorpha inferred from l8S 253-269. and 5.8S ribosomal DNA sequences (Diptera: Culici- Zavortink TJ. (1990). Classical taxonomy of mosqui- d,ae).Insect Mol Biol 6(2):105-114. toes-a memorial to John N. Belkin. J Am Mosq Con- Munstermann LM, Conn IE. 1997. Systematics of mos- trol Assoc 6:593-599.

APPENDIX I Character used in phylogenetic analysrs. (0) VT: Ventral teeth Size: VTO > VTl = VT2 = VT3 > VT4, teeth l, 2, and3 normal (0) Size: VTO > VTI = VT2 = VT3 > VT4, reerh l,2,and3 thin (t) Size: VTO = VTI > VT2 : VT3 > VT4, teeth l, 2, and 3 slightly sclerotized (2) Size: VTO > VTl > VT2 > VT3 > VT4, teeth 1, 2, and 3 (3) not serrated (3) Size: VTO > VT2 > VTl : VT3 > VT4, teeth 1, 2, and3 normal (4) Size: VTO > VT3 > VTl : VT2 > VT4, teeth l, 2, and 3 normal (5) Size: VTO = VT3 > VTI = VT2 > VT4, teeth l, 2, and3 normal (6) (1) PDT Posterior dorsal tooth Absent (0); Simple serrate, and tree accessory smaller (1); Simple not serrate, and 2 accessories (external bigger than) (2); Simple not serrate, and tree accessory (the medial is smallest) (3); Simple not serrate, and tree accessories (all same size) (4); Simple not serrate, and tree accessory (l over a spheric surface) (5); Simple not serrate, and 4 accessory (the 2 intemal smaller) (6); Simple not serrate, and several small accessories (all of these over spheric surface) (7); Simple not serrate, and 2 accessories over spheric surface (8) (2) ADT: Anterior dorsal tooth Simple serrate (0); Simple not serrate (l); Absent (2) (3) MnC: Mandibular comb Absent (0); Large: 1 spicule per each insertion (1); Large: several spicules in each insertion (2); Latge: I spicule in central position surrounded by short spicules in each insertion (3); Large: I central and branched spicule surrounded by short ones in each insertion (4); Short spicules (5); Short spicules over conspicuous protuberanc;s (6) 84 JounNer-op rnr Avnnrcen Moseuno Covrnol AssocrerroN VoL. 16,No.2

APPENDIX I Continued

(4) l-Mn: Mandibular seta no. 1 Absent (0); Present: behind the mandibular teeth (1) (5) 2-Mn: Mandibular seta no. 2 Several groups of setae (2a, 2b,2c, .. . , etc) (0); Several groups of setae (2 of them serrate) (l); Only 1 reduced and behind the MnT (2) (6) MnS: Mandibular sweeper MnSl with 7 filaments and MnS2 with 5 filaments (0); MnSl 6 filaments, MnS2 4 filaments, all with no branched apex ( I ); MnS I and MnS2 with the same number of filaments (2); MnS I with 8 filaments and MnS2 6 filaments (3); MnSl and MnS2, absent (4) (7) MRB: Mandibular rake blade Simple serrate and curved towards the MnB (O); Simple serrate and very developed (1); Simple serrate and reduced (2); Simple serrate normal (3); Simple lightly serrate (4); Double: 1 lightly serrate and the other I normal (5); Double (both lightly serrate) (6); Double serrate: I very thin (7); Double serrate (8) (8) MSA: Mandibular spiculose area Dorsal spicules, large and few (0); large and abundant spicules (l); short and few spicules (2); short and thin (3); inconspicuous (4); Dorsals and ventrals large (5); Dorsals very developed (6); Dorsals, large and blunt (7) (9) MnR: Mandibular rake Number of filaments: Five not pectinate (0); Seven or 8 large and pectinate (1); Five to 7 large pectinate (2); Seven or 8 short pectinate (3); Five or 6 short pectinate (4); Three or 4 short pectinate (5); Absent (6) (10) L: Labula Lightly developed (0); Absent (l); Developed (2); Strongly developed (3) (11) MnL: Mandibular lobe One and one-half the width of base (0); reduced (l); One and one-half the width of base, and trilobulate (2); Four times width of base, with spicules (3) (12) Poa: Postartis Rectangular shape and short (0); Rectangular and large (l); Bilobulate: Both lobes equal in size (2); Bilobulate: Posterior lobe larger than anterior one (3); U-shaped, short (4); U-shaped, large (5) (13) PMnL: Posterior mandibular lobe Absent (O); Normal (l); Well developed (2) (14) MnB: Mandibular brush Normal (0); Reduced, toward MnT (l) (15) ppMAdA: Posterior projection of mandibular aductor apodeme Very short: 0.013-0.019 mm (0); Short: 0.025-0.038 mm (1); Large: 0.051-0.057 mm (2); Large: 0.051-O.O57 mm, hooklike (3) (16) MxB: Maxillary brush Short:0.140-4.178 mm (0); Short 0.152 mm, with little serrate spicules (1); Very short 0.102 mm (2); Large: 0.305--0.406 mm (3) (17) l-Mx: Maxillary seta no. I Large: 0.064-O.076 mm, thin at beginning of DMxS (0); Large: 0.064-0.076 mm, thin and beyond end of DMxS (l); Short: 0.038-0.051 mm, thin and medial to DMxS (2);Yery short 0.025 mm, thin and medial to DMxS (3); Very short 0.O25 mm, thick and far apart DMxS (4) (18) 3-Mx: Maxillary seta no. 3 Present (0); Absent (1) (19) 4-Mx: Maxillary seta no. 4 Very short: 0.038-0.064 mm, normal (0); Very short: 0.038-{.O64 mm and sclerotized (1); Short: 0.O89-O.O95 mm (2); Short: O.O89-O.O95mm and sclerotized(3); Large:0.1O2-0.114 mm (4); Very large:0.127-O.159 mm (5); Very laryq 0.1274.159 mm and sclerotized (6) (2O) Position of 4-Mx Anterior, beyond palp (0); Posterior, beyond LRI (l) medial to MxBo (2) (21) 7-Mx: Maxillary seta no. 7 Absent (0); Present (1) (22) 8-,9-, l0-, and 1l-Mx (:MS): Maxillary setaenos.8,9, 10, and 11 JUNE 2000 Pnvr-ocsNrtIc RELATIoNsHTps tN Cur-rctNt 85

APPENDIX I Continued

Very short (O); Short (l); Large (2) (23) MxBo: Maxillary body Approximately as large as wide, square shape (0); twice its width, rectangular shape (1); more than twice its width (2) (24) MSpA: Maxillary spiculose area Five dorsolateral spicules (0); Nine to 13 dorsolateral spicules (l); Fourteen dorsolateral rounded spicules (2); Absent (3) (25) DMxS: Dorsal maxillary suture Large and vertical (0); Large and diagonal (l); Short and diagonal (2) (26) LRI: Laciniarastrum no. 1 Large spicules (0); Nine to 13 short spicules (1) (27) LR2: Laciniarastrum no. 2 Present (0); Absent (l) (28) MPIp: Maxillary palpus Prominent (0); Normal (l) (29) GSS: Galeastpial stem Absent (0); Present (1)

APPENDIX 2 Data matrix used for cladistic analysis.

Charactersr I l l l l l l1112222222222 Taxa o1234567890 1234567890123456789

Aedes taeniorhynchus 0lo11000oo0000020 0 0 0000 000000 I Deinocerite s melanophylum 20220117641352033 2 0 2ll 23000I o Culex (Lutzia) bigoti 302oo2427611?Ot22 1 I 2000 030010 0 Cx. (Carrollia\ bihaicolus 08 30028 520400 o 0 510 o32ro 0 Cx. (Car.) rausseoi 08 30028 520400 0 0 510 o32ro 0 Cx. (Car.) urichi o8 30028 520400 o 0 620 03200 o Cx. (Car.) iridescens 08 30028 520400 0 0 620 o3200 o Cx. (Culex) dolosus o2 soo234 1202002 J 0 000 13000 0 Cx. (Cux.) nigripalpus o2 500334 1202002 J 0 000 10000 0 Cx. (Cux.) quinquefasciatus o2 soo234 120200z J 2 0 000 11000 0 Cx. (Cux.) coronator 42 500334 t20200 z 3 1 0 000 13000 0 (Anoedioporpa) A Cx. bamborum 15 40025 432200 I J 0 010 .l 2 23000 0 Cx. (M ic roc ulex) p le uri st riatus 57 400285 2322rO o 2 0 4to2 00r00 o (Mcx.) Cx. chryselatus 54 400264 23221OI J 3 0 4to2 22000 0 (Mcx.) Cx. microphylus l4 400244 432200 I J 2 0 4IO2 23000 0 Cx. (Melanoconion) albinensis 66 500253 320300 I J 2 0 2to ? 23000 0 (Mel.) Cx. dunni 66 50025J 420300I -) 2 0 3102 23000 o (Mel.) Cx. nicceriensis 66 50025J 320300 I -) 2 0 0102 23000 o Cx. (Phenacomyia) corniger 03 600212 1001000 I 2 0 o0010300000 '?, Could not be determined.