Proc. Natl. Acad. Sci. USAVol. 92, pp. 2017-2020, March 1995Evolution

A phylogeny of cockroaches and related insects based on DNAsequence of mitochondrial ribosomal RNA genes

(termites/mitochondrial DNA/molecular phylogenetics)

SRINIvAs KAMBHAMPATIDepartment of Entomology, Kansas State University, Manhattan, KS 66506

Communicated by Charles D. Michener, University of Kansas, Lawrence, KS, December 12, 1994 (received for review June 2, 1994)

ABSTRACT Cockroaches are among the most ancientwinged insects, the earliest fossils dating back to about 400million years. Several conflicting phylogenies for cockroachfamilies, subfamilies, and genera have been proposed in thepast. In addition, the relationship of Cryptocercidae to othercockroach families and the relationship between the cock-roach, Cryptocercus punctulatus, and the termite, Mastotermesdarwiniensis, have generated debate. In this paper, a phylogenyfor cockroaches, mantids, and termites based on DNA se-quence of the mitochondrial ribosomal RNA genes is pre-sented. The results indicated that cockroaches are a mono-phyletic group, whose sister group is Mantoidea. The inferredrelationship among cockroach families was in agreement withthe presently accepted phylogeny. However, there was onlypartial congruence at the subfamil and the generic levels. Thephylogeny inferred here does not support a close relationshipbetween C. punctulatus and M. darwiniensis. The apparentsynapomorphies of these two species are likely a manifestationof convergent evolution because there are similarities inbiology and habitat.

Cockroaches (order: Dictyoptera; suborder: Blattaria) are amongthe oldest winged insects known, dating back to the Carbonifer-ous (1). About 4000 species of cockroaches have been described(2). A number of conflicting classifications exist for cockroaches,the most widely accepted ofwhich is that of McKittrick (1), basedon morphological characters. She considered the order Dic-tyoptera to include cockroaches, mantids, and termites, each withits own suborder. She divided the suborder Blattaria into twosuperfamilies, Blaberoidea and Blattoidea, and five families,Polyphagidae, Blattellidae, and Blaberidae (all Blaberoidea), andBlattidae and Cryptocercidae (both Blattoidea). Three othermajor cockroach classifications, based on morphological charac-ters, have been published during the past four decades (3-5).

In addition to the overall phylogenetic relationships amongcockroaches, two other issues have generated debate. The firstis the relationship of Cryptocercidae to other cockroachfamilies. Cryptocercidae consists of one genus (Cryptocercus)and three species (6) and is generally considered a sister groupof Blattidae (1). However, it was recently proposed thatCryptocercidae be merged with Polyphagidae (7). The secondissue concerns the relationship among cockroaches, mantids,and termites. Three major schemes have been proposed:Blattaria and Mantoidea are sister groups and Isoptera is asister group of the Blattaria-Mantoidea complex (8), cock-roaches and termites belong to the order Blattodea andmantids are a sister group to that order (9), and all threegroups belong to Dictyoptera (1, 10). Of particular interest isthe presumed close phylogenetic relationship between Cryp-tocercus and the termite, Mastotermes darwiniensis. M. darwini-ensis has been considered the most archaic living termitespecies and the "missing link" between cockroaches and

termites (11). It is the sole living representative of the familyMastotermitidae and is found in northern Australia (12). Theclose relationship between M. darwiniensis and Cryptocercus isbased on several apparent synapomorphies (13-16).A phylogeny of mantids (one species), cockroaches (two

species), and termites (three species) based on previouslypublished morphological characters has recently been pro-posed (6). A conclusion of that study was that Cryptocercuspunctulatus is not closely related to M. darwiniensis but is a partof Blattodea (= Blattaria), which is a sister group to Man-toidea. A phylogenetic study (12) that included four termitespecies and one each of cockroaches and mantids, and wasbased on DNA sequence of a portion of the nuclear 18S rRNAgene, indicated a sister group affinity of M. darwiniensis toother termites. In contrast, a study (17) that included C.punctulatus, M. darwiniensis, Blatta orientalis, and Reticulo-termes flavipes and was based on DNA sequence of the entire18S rRNA gene suggested that C. punctulatus and M. darwini-ensis are closely related. This led the author to conclude that"Mastotermitidae is considered to belong to Blattodea, insteadof Isoptera" (ref. 17, p. 132). The conflicting conclusions of theabove studies (6, 12, 17) suggest a need to verify their findingsby including a more diverse range of cockroach taxa andemploying a DNA sequence from a different gene because ofthe issue of gene trees and species trees (18, 19). Thus, theprimary objective of this study was to infer a phylogeny forcockroaches, mantids, and termites based on the DNA se-quence of mitochondrial large (16S rRNA) and small ribo-somal (12S rRNA) subunit genes. The specific objectives wereto (i) compare the molecular phylogeny with that proposed byMcKittrick (1), (ii) infer the relationship between Cryptocer-cidae and other cockroach families, and (iii) infer the rela-tionship between C. punctulatus and M. darwiniensis.*

MATERIALS AND METHODSInsects. The species included in this study are as follows:

Blaberidae: Archimandrita tessellata, Blaberus atropos, Blabe-rus craniifer, Blaberus discoidalis, Blaberus giganteus, Byrsotriafumigata, Diploptera punctata, Epilampra azteca, Eublaberusposticus, Gromphadorhina portentosa, Nauphoeta cinerea, Pan-chlora nivea, Phoetalia pallida, Phortioeca phoraspoides, Pyc-noscelus surinamensis, Rhyparobia maderae, Schultesia lampy-ridiformis; Blattellidae: Blattella vaga, Nahublattella fraterna,Nahublattella nahu, Nyctibora azteca, Nyctibora lutzi, Parco-blatta pensylvanica, Symploce pallens; Blattidae: Blatta orien-talis, Melanozosteria soror, Periplaneta americana, Periplanetaaustralasiae, Periplaneta brunnea, Periplaneta fuliginosa, Shel-fordella lateralis; Cryptocercidae: C. punctulatus; Mantidae:Mantis religiosa; Rhinotermitidae: Coptotermes formosanus,Reticulotermesflavipes; Mastotermitidae: M. darwiniensis. Oneor more live specimens or DNA of the organism (M. darwini-ensis and M. religiosa) were obtained from colleagues. In most

*The sequences reported in this paper have been deposited in theGenBank data base (accession nos. U17761-U17832).

The publication costs of this article were defrayed in part by page chargepayment. This article must therefore be hereby marked "advertisement" inaccordance with 18 U.S.C. §1734 solely to indicate this fact.

2017

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 15

, 202

1

Proc. Natl Acad. Sci USA 92 (1995)

cases, at least one individual was preserved as a voucherspecimen.DNA Extraction, PCR, and DNA Sequencing. DNA was

extracted from a small portion of the fat body of frozenspecimens and PCR was set up as described (20). The PCRconditions were an initial denaturation step of 94°C for 3 minfollowed by 35 cycles of 94°C for 30 sec, 50°C for 1 min, and72°C for 1.5 min. The amplification product was electropho-resed on a 2% low-melting-point agarose gel and purified usingminicolumns (Wizard PCRpreps, Promega). DNA sequencewas obtained directly from 3 ,u of the purified PCR productusing the cycle sequencing method (fmol Sequencing System,Promega). The reaction mixtures were electrophoresed on 6%polyacrylamide denaturing gels. Both strands of the PCRproduct were sequenced.

Oligonucleotide Primers. The primers for the amplificationof a 415-bp fragment of the 16S rRNA gene were forward,5'-TTA CGC TGT TAT CCC TTA-3' (positions 13,000-13,017 of Drosophila yakuba), and reverse 5'-CGC CTG TTTATC AAA AAC AT-3' (13,396-13,415 of D. yakuba). Theprimers for amplification of a 431-bp fragment of the 12SrRNA gene were forward, 5'-TAC TAT GTT ACG ACTTAT-3' (14,182-14,199 of D. yakuba), and reverse, 5'-AAACTA GGA TTA GAT ACC C-3' (14,594-14,612 of D.yakuba). The primers were derived from previously publishedinsect mitochondrial sequences (21-24). Both sets of primersresult in the amplification of a homologous fragment from awide range of insects. Internal primers (16S rRNA: 5'-TCTATA GGG TCT TCT CGT C-3' and its reverse complement;12S rRNA: 5'-TGC ACC TTGACCTGA A-3' and its reversecomplement) were used to obtain the sequence on the ends ofthe fragments.

Sequence Alignments and Phylogenetic Inference. The se-quences were read manually from autoradiographs into acomputer. They were aligned using CLUSTAL V (25) and thenoptimally aligned manually. The alignment parameters werek-tuple score = 1, gap penalty = 3, and window size = 5(pairwise alignments); fixed and floating gap penalties = 10(multiple alignments). Phylogenetic analysis was carried out inPAUP 3.1.1 (26) using the multiple equally parsimonious heu-ristic search option with tree bisection-reconnection. The dataset was too large to be used with the exhaustive or the branchand bound algorithms. Gaps were treated as a fifth base. Thesequences of the two genes were analyzed as a single data set(27) without character weighting. The data set was boot-strapped for 1000 replications using PAUP. The DNA sequenceof the 16S rRNA gene ofLocusta migratoria (24) was includedas the outgroup. The DNA sequence of the small ribosomalsubunit gene of L. migratoria was not available; thus the 12SrRNA data for this taxon were designated as missing.

RESULTSDNA Sequences of rRNA Genes. The sequences for taxa in

this study can be obtained directly from GenBank or from theauthor. The average size of the sequenced portion of the 16SrRNA gene was 415.3 ± 0.75 bp (mean ± SE) and that of the12S rRNA gene was 431.4 ± 0.72 bp. Summary statistics for thetwo genes are presented in Table 1.

Phylogenetic Inference. The alignment of the sequencesresulted in a total of 923 characters, including gaps. Unam-biguous alignment was possible for most regions; six regionstotaling 140 characters were relatively more difficult to align.When the data were analyzed without these regions, thetopology of the tree identified by PAUP was nearly identical tothat of the tree generated with the full data set. Therefore, allfurther analyses were carried out with the full data set.PAUP identified a single most parsimonious tree of 3750

steps (Fig. 1). The taxa in the various orders, suborders, andfamilies formed distinct clades. Four major clades were iden-

Table 1. Summary statistics for the DNA sequences of themitochondrial rRNA genes

Parameter 16S rRNA 12S rRNA

Base composition (mean % ± SE)Adenine 39.1 ± 0.42 39.6 ± 0.39Cytosine 17.9 ± 0.27 17.8 ± 0.21Guanine 10.6 ± 0.15 10.9 ± 0.13Thymine 32.4 ± 0.50 31.7 ± 0.46

Transition rate (%)Overall 7.9 11.3C ++ T 65.5 66.2A * G 34.5 33.8

Transversion rate (%)Overall 22.0 24.1A +-> C 23.4 22.3A <-> T 70.8 69.6G C 1.7 3.1G T 4.1 5.0

CharactersTotal (including gaps) 455 468Variable 276 287Invariable 179 181Parsimony informative 237 242

The statistics represent the means for 36 taxa, excluding L. migra-toria.

tified among cockroach taxa corresponding to the four familiesin this study. Taxa within Blaberidae were subdivided into twosubclades. One consisted of genera in Blaberinae, Oxyhaloi-nae, Panchlorinae, and Diplopterinae and the second con-sisted of genera in Zetoborinae, Epilamprinae, and Pycnos-celinae. Within Blattellidae, Bla. vaga and Par. pensylvanicawere shown to be sister taxa and joined to Sy. pallens, followedby the joining of Nyctibora spp. and Nahublattella spp. to theabove three genera. Within Blattidae, the four species ofPeriplaneta were found to be paraphyletic. The Periplaneta-Shelfordella clade was first joined to B. orientalis followed byMe. soror.At the family level, Blattellidae and Blaberidae were shown

to be more closely related to each other than either was toBlattidae. The sole representative of Cryptocercidae, C.punctulatus, formed a separate clade that was joined to theabove three families. This was followed by the joining ofMreligiosa to the cockroach clade. Among the three species oftermites included in this study, C. formosanus and R. flavipeswere shown to be sister taxa. Next, M. darwiniensis was joinedto the above two taxa. Most of the relationships shown in Fig.1 were supported in 70-100% of the replications in a treederived from bootstrap analysis (Fig. 2).

DISCUSSIONA phylogeny of cockroaches inferred from DNA sequence ofthe mitochondrial rRNA genes indicated that cockroaches aremonophyletic. This is-in contrast to the suggestion (17) that thegroup is perhaps paraphyletic. The relationship among familiesinferred here was in agreement with the presently acceptedphylogeny (1). However, there were significant differences atthe subfamily and generic levels between McKittrick's (1) andthe molecular phylogenies. Of particular note were taxa inBlaberidae, within which relationships among genera did notalways reflect the subfamily designations of McKittrick. Forexample, D. punctata and Pho. pallida were placed in Diplop-terinae by McKittrick (1) and in Diplopterinae and Epilam-prinae, respectively, by Princis (4). In this study, Pho. pallidawas shown to be more closely related to Sc. lampyridiformis ofZetoborinae than to D. punctata.Two subfamilies of Blattellidae, Blattellinae and Nyctibori-

nae, were represented in this study. Blattella, Parcoblatta, and

2018 Evolution: Kambhampati

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 15

, 202

1

Proc. Natt Acad Sci USA 92 (1995) 2019

3 I Phor. phoraspoides 128 Pho. pallida 2

Sc. lampyridiformis 131 Py. surinamensis 3

90Ep. azteca 244 A. tessellata 4

31 0 BI. discoidalls 447 1 Bl. atropos 4

66 46 37 1 Bl. cranilfer 46162 BI. giganteus 4

34 _ 868 By. fumigata 434 86 ~~~Eu. postlcus 441 108 D. punctata 5

68 Nau. cinerea 651 42 32 69R. maderae 6

G. portentosa 682 Pn. nivea 7

32 85 BIa. vaga 836 61 Par. pensylvanica 8

48 Sy. pallens 871 Ny. lutzi 9

58 41 Ny. azteca 9118 10- - Na. naha 8

Na. fraterna 868 B. orientalis 10

46 4X30 Sh. lateralis 1061 46

21 62 Pe. brunnea 10

2 4 Pe. americana 10 II22 Pe. fuliginosa 10

52 Pe.fuliginoPe.australasiae 1 072 Me. soror 11

1 1 5 C. punctulatus 1 2 V_ 132 M. religiosa 13 V

73 85 R. flavipes 1455 C. formosanus 15 VI

106 1 74 M. darwiniensis 16L. migratoria OG

FIG. 1. Single most parsimonious tree for DNA sequence ofmitochondrial 16S rRNA and 12S rRNA genes of cockroaches,termites, and mantid, rooted by the outgroup, L. migratoria. Treelength, 3750 steps; consistency index, 0.35; retention index, 0.54.Numbers above the branches are branch lengths. Numbers accompa-nying scientific names are family or subfamily designations: 1, Zeto-borinae; 2, Epilamprinae; 3, Pycnoscelinae; 4, Blaberinae; 5, Diplop-terinae; 6, Oxyhaloinae; 7, Panchlorinae; 8, Blattellinae; 9, Nyctibori-nae; 10, Blattinae; 11, Polyzosteriinae; 12, Cryptocercinae; 13,Mantidae; 14, Heterotermitinae; 15, Coptotermitinae; 16, Mastoter-mitidae. The roman numerals indicate ordinal, subordinal, or familydesignations: I, Blaberidae; II, Blattellidae; III, Blattidae; IV, Cryp-tocercidae; V, Mantoidea; VI, Isoptera; OG, outgroup.

Symploce, all Blattellinae, were inferred to be closely related toone another as previously suggested (1). However, because ofthe placement of Nahublattella relative to the above genera,Blattellinae as presently recognized is paraphyletic. Nyctibori-nae has been inferred to be a sister group of Blattellinae (1),a relationship confirmed by this analysis.Two subfamilies of Blattidae were included in this study:

Blattinae and Polyzosteriinae. As expected, all genera inBlattinae were sister taxa to one another and Polyzosteriinaewas shown to be the sister group of Blattinae. Within Blattinae,Sh. lateralis was found to be a sister taxon of Pe. brunnea andthus the genus Periplaneta is paraphyletic according to myanalysis. No consensus on the generic status of Sh. lateralis isapparent. Historically, this species has been successively placedin the genera Periplaneta, Blatta, and Shelfordella. Walker (28)originally described this species as Periplaneta lateralis. Therelationship inferred here suggests that this species shouldperhaps be placed in the genus Periplaneta as originallyproposed rather than Blatta (4, 29) or Shelfordella as presentlyrecognized.The results of my analysis indicated that Cryptocercidae is

most closely related to Blattidae. As mentioned, however, aproposal to merge Cryptocercidae and Polyphagidae has been

Phor. phoraspoidesPho. pallidaSc. lampyridiformisPy. surinamensisEp. aztecaA. tessellataBI. discoidalisBI. atroposBI. craniiferBI. giganteusBy. fumigataEu. posticusPa. niveaNau. cinereaR. maderaeG. portentosaD. punctataBla. vagaPar. pensylvanicaSy. pall ensNy. lutziNy. aztecaNa. nahaNa. fraternaB. orientalisSh. lateralisPe. brunneaPe. americanaPe. ful i gi nosaPe. australasiaeMe. sororC. punctulatusM. religiosaR. flavipesC. formosanusM. darwiniensisL. migratoria

FIG. 2. Bootstrap parsimony tree (1000 replications) based onDNA sequence of mitochondrial 16S rRNA and 12S rRNA genes ofcockroaches, termites, and mantid, rooted by the outgroup, L. migra-toria. Tree length, 3912 steps; consistency index, 0.34; retention index,0.51. Numbers above the branches indicate percent times the branchwas recovered.

made (7). Unfortunately, I could not include Polyphagidae totest its relatedness to Cryptocercidae. Further work is neces-sary to infer the relationship between Polyphagidae andCryptocercidae.The results of this study suggested that Mantoidea is a sister

group of Blattaria and that termites are a sister group of thecockroach-mantid complex. Although it seems appropriatethat cockroaches and mantids be retained in Dictyoptera witheach group being assigned a subordinal status (8), futurestudies that include a broader range of mantids and termitesare needed to confirm the relationships inferred here.The results of my analysis indicated that M. darwiniensis and

C. punctulatus are not closely related. The sister group affinityof M. darwiniensis was clearly with other termites and that ofC. punctulatus was with other cockroaches. At present, twoopposing views exist concerning the phylogenetic relationshipof these two taxa. One states that they are no more closelyrelated to each other than cockroaches in general are relatedto termites and any apparent synapomorphies of Cryptocercusand Mastotermes together are the result of convergent evolu-tion (6, 9). The alternate view (1, 17) holds that the synapo-morphies of the two taxa are a result of common ancestry.Specifically, the conclusions of morphological analysis (6) andearlier analyses of DNA sequences (17) are at odds concerningthe relationship of C. punctulatus with M. darwiniensis. Thepresent study, also based on DNA sequence analysis, but ofdifferent genes and with an extensive sampling of cockroachtaxa, clearly supported the conclusions of the morphologicalanalysis (6). In both trees (Figs. 1 and 2), M. darwiniensis wasa sister taxon of the other termites. In the parsimony tree (Fig.1) the affinity of C. punctulatus was clearly with other cock-roaches.

Evolution: Kambhampati

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 15

, 202

1

Proc. Natl. Acad Sci. USA 92 (1995)

A significant synapomorphy on which a close phylogeneticrelationship of Cryptocercus with Mastotermes has been pro-posed is the cellulolytic gut fauna that is present in both taxabut not in other cockroaches (1, 13-16). Previous studies (1,30-33) and the present study indicate that C. punctulatusbelongs to a primitive phyletic line from which the othercockroaches have descended. This suggests two alternatescenarios for the presence of cellulolytic symbionts in Crypto-cercus but not in other cockroaches. First, M. darwiniensis andCryptocercus share gut fauna due to common ancestry. In thiscase, it must be assumed that all other cockroaches (andmantids) that descended from a Cryptocercus-like ancestorhave secondarily lost gut fauna (6). Alternatively, Cryptocercusacquired the symbionts through some mechanism (see below)after other cockroaches arose. In this scenario, a secondaryloss need not be invoked for the non-cryptocercid cockroaches.

If secondary loss is not invoked, the shared gut fauna ofCryptocercus and Mastotermes implies convergent evolution (9)and perhaps interspecific transfaunation (34). It is not knownwhich group first acquired the cellulolytic gut fauna. However,a scenario in which transfaunation could occur through inter-specific predation has been proposed (34). Since Cryptocercusand Mastotermes were sympatric in the Americas before thelatter became extinct there (9, 12), the possibility exists that thegut fauna was obtained through interspecific predation.Thorne (34) concluded that "Cryptocercus shares no unam-biguous morphological, behavioral or symbiotic synapomor-phies with the Isoptera that are not shared with other cock-roaches as well. Cryptocercidae are clearly primitive roaches,but it is more likely that the family is a sister group to othercockroach taxa rather than a sister group to modern termites."Cogent arguments against the scenario in which transfaunationcould occur have been presented (35). Thus, the question ofhow Cryptocercus and Mastotermes came to share gut symbi-onts is yet to be resolved. Further studies, directly on the DNAof the gut fauna, may help resolve the issue of whether the gutsymbionts of Cryptocercus and Mastotermes are more closelyrelated to each other than a random pair of symbionts fromxylophagous insects. With PCR, it is feasible to study symbiont-specific genes as has been done for other insect symbionts (36,37).As with other insect mtDNA studied to date (21-24), the

base composition of sequences in this study was strongly biasedtoward adenine and thymine, which comprised 72% of thetotal. The observed transition and transversion rates werecomparable to those reported for other insect groups and therate of transversions was significantly greater than that oftransitions. A greater transversion rate relative to transitionrate has also been observed in these mtDNA genes: 16S rRNAof leafhoppers (38) and black flies (39), cytochrome oxidase IIof 10 insect orders (40), and NADH 1 and 16S rRNA ofDrosophila spp. (41). The bias toward transversions in insectmtDNA is in contrast to the mtDNA of primates in which 92%of the substitutions were transitions (42). It has been suggestedthat the bias may be due to deficient mtDNA repair mecha-nisms and a tautomeric base pairing chemistry (43). As withother insects (references above), >70% of transversions incockroaches, termites, and mantid were A *-+ T transversions.

In summary, the significant findings of this study are asfollows. The molecular phylogeny was congruent at the familylevel with the morphological phylogeny proposed by McKit-trick (1). However, significant differences were observed at thesubfamily and generic levels. DNA sequences from morespecies need to be obtained to further delineate the relation-ships at the subfamily level. The results of this study suggestedthat Cryptocercidae is most closely related to Blattidae.Whether it should be placed in the Polyphagidae needs to beresolved using appropriate taxa. My analysis did not lendsupport to the hypothesis that C. punctulatus and M. darwini-ensis are closely related. Thus, the suggestion that M. darwini-

ensis belongs to Blattaria instead of Isoptera (17) can beignored for the present. Finally, my analysis indicated thatcockroaches are monophyletic and are a sister group tomantids. The issue of whether Isoptera is a sister group of theBlattaria-Mantoidea complex needs to be resolved by includ-ing other closely related taxa such as Zoraptera and represen-tatives of diverse termite families.

I thank the following: R. D. Bowling, J. Y. Bradfield, R. DeSalle, S.Gatti, D. E. Mullins, C. A. Nalepa, R. S. Patterson, L. M. Roth, S.Starkey, B. Stay, and L. Vawter for insects/DNA; D. W. Alsop, W. J.Bell, W. C. Black IV, R. DeSalle, T. L. Hopkins, and L. M. Roth foradvice; W. J. Bell, C. D. Michener, and two anonymous reviewers forcomments on the manuscript; and A. L. Nus for technical assistance.This research was supported by a seed grant from the Department ofEntomology, Kansas State University. This is journal article no.94-540-J of the Kansas Agricultural Experiment Station.

1. McKittrick, F. A. (1964) Mem-N. Y Agric. Exp. Stn. (Ithaca) 389,1-191.

2. Bell, W. J. (1990) in Cockroaches as Models for Neurobiology, eds.Huber, I., Masler, E. P. & Rao, B. R. (CRC, Boca Raton, FL), Vol.1, pp. 7-12.

3. Rehn, J. W. H. (1951) Am. Entomol. Soc. Mem. 14, 1-134.4. Princis, K. (1960) Eos (Madrid) 36, 429-449.5. Huber, I. (1974) Univ. Kans. Sci. Bull. 50, 233-332.6. Thorne, B. L. & Carpenter, J. M. (1992) Syst. Entomol. 17, 253-268.7. Grandcolas, P. (1992) Syst. Entomol. 19, 145-158.8. Boudreaux, H. B. (1979)Arthropod Phylogeny with Special Reference to

Insects (Wiley, New York).9. Hennig, W. (1981) Insect Phylogeny (Wiley, New York).

10. Kristensen, N. P. (1981) Annu. Rev. Entomol. 26, 135-157.11. Hennig, W. (1966) Phylogenetic Systematics (Univ. of Illinois Press,

Urbana, IL).12. DeSalle, R., Gatesy, J., Wheeler, W. & Grimaldi, D. (1992) Science

257, 1933-1935.13. Crampton, G. C. (1923) Brooklyn Entomol. Soc. Bull. 18, 85-93.14. Tilliyard, R. J. (1936) Nature (London) 137, 655.15. Hill, G. F. (1925) R. Soc. Victoria Proc. 37, 119-124.16. Cleveland, L. R., Hall, S. R., Sanders, E. P. & Collier, J. (1934)

Brooklyn Entomol. Soc. Bull. 13, 49-68.17. Vawter, L. (1991) Ph.D. Thesis (Univ. of Michigan, Ann Arbor).18. Nei, M. (1987) Molecular Evolutionary Genetics (Columbia Univ.

Press, New York).19. Li, W.-H. & Graur, D. (1990) Fundamentals of Molecular Evolution

(Sinauer, Sunderland, MA).20. Kambhampati, S., Black, W. C., IV, & Rai, K. S. (1992) J. Med.

Entomol. 29, 939-945.21. Crozier, R. H. & Crozier, Y. C. (1993) Genetics 133, 97-117.22. Clary, D. 0. & Wolstenholme, D. R. (1985) J. Mol. Evol. 22,252-271.23. Mitchell, S. E., Cockburn, A. F. & Seawright, J. A. (1993) Genome 36,

1058-1073.24. Uhlenbusch, I., McCracken, M. & Gellissen, G. (1987) Curr. Genet.

11, 631-638.25. Higgins, D. M. & Sharp, P. M. (1989) Comput. Appl. Biol. Sci. 5,

151-153.26. Swofford, D. L. (1993) PAU.: Phylogenetic Analysis Using Parsimony

(Smithsonian Inst., Washington, DC), Version 3.1.1.27. Miyamoto, M. M., Kraus, F. & Ryder, 0. A. (1990) Proc. Natl. Acad.

Sci. USA 87, 6127-6131.28. Walker, F. (1868) British Museum Catalog (British Museum, London).29. Gurney, A. B. (1978) Plant Pest Rep. 3, 295.30. Roth, L. M. & Willis, E. R. (1958) Trans. Am. Entomol. Soc. 83,

221-238.31. Roth, L. M. (1967) Ann. Entomol. Soc. Am. 61, 83-111.32. Roth, L. M. (1982) Proc. Entomol. Soc. Wash. 84, 277-280.33. Roth, L. M. (1989) Proc. Entomol. Soc. Wash. 91, 441-451.34. Thorne, B. L. (1990) Proc. R. Soc. London B 241, 37-41.35. Nalepa, C. A. (1991) Proc. R. Soc. London B 246, 185-189.36. O'Neill, S. L., Giordano, R., Colbert, A. M. E., Karr, T. L. & Rob-

ertson, H. M. (1992) Proc. Natl. Acad. Sci. USA 89, 2699-2705.37. Kambhampati, S., Rai, K. S. & Burgun, S. J. (1993) Evolution 47,

673-677.38. Fang, Q., Black, W. C., IV, Blocker, D. & Whitcomb, R. F. (1993)

Mol. Phylogenet. Evol. 2, 119-131.39. Xiong, B. & Kocher, T. D. (1991) Genome 34, 306-311.40. Liu, H. & Beckenbach, A. T. (1992) Mol. Phylogenet. Evol. 1, 41-52.41. DeSalle, R. (1992) Mol. Phylogenet. Evol. 1, 31-40.42. Brown, W. M., Prager, E. M., Wang, A. & Wilson, A. C. (1982)J. Mol.

Evol. 18, 225-239.43. Topal, M. D. & Fresco, J. R. (1976) Nature (London) 263, 285-289.

2020 Evolution: Kambhampati

Dow

nloa

ded

by g

uest

on

Feb

ruar

y 15

, 202

1