Nematology 2002 Vol 4(4) 489-504

Speciation in the Acugutturidae (Nematoda Aphelenchida)

Orville G MARTI JR 1curren Jean-Franccedilois SILVAIN 2 and Byron J ADAMS 3

1 United States Department of Agriculture Agricultural Research Service Crop Protection and Management ResearchLaboratory PO Box 748 Georgia Coastal Plain Experiment Station Tifton GA 31793 USA

2 Institut de Recherche pour le Deacuteveloppement co CNRS Laboratoire PopulationsGeacuteneacutetique et Evolution Bat 13BP1 91198 Gif-sur-Yvette Cedex France

3 University of Florida Institute of Food and Agricultural Sciences Entomology and Nematology DepartmentPO Box 110620 Gainesville FL 32611 USA

Received 25 March 2001 revised 26 February 2002Accepted for publication2 March 2002

Summary ndash Ectoparasitic nematodes (Acugutturus Vampyronema and Noctuidonema) of the family Acugutturidae from 13 species ofLepidoptera (moths) and one species of Blattodea (cockroaches) were collected measured and analysed by discriminant analysis inorder to identify conspeci cs and to determine which morphological characters were useful in de ning differences between nematodeson different host species The position of the excretory pore posterior to the metacorpus clearly differentiates Acugutturus parasiticusfrom Noctuidonema and Vampyronema Two populations of A parasiticus on Periplanetaamericana from the West Indian islands of StLucia and Guadeloupe are conspeci c Nematodes resembling Vampyronema collected from two genera of Sphingidae in Guadeloupeare characterised by extremely long stylets equal to ca 37 of the body length and may represent a new genus Noctuidonema fromfour species of Spodoptera (Noctuidae) are conspeci c However a somewhat different Noctuidonema from the noctuid Pseudaletiaunipuncta may represent a new species based primarily on its shorter stylet length Simple spicule structure a smaller V percentageand a shorter stylet in Vampyronema are characters that best de ne its differences from Noctuidonema Nematodes on the noctuidsAnicla infecta and Eulepidotis addens probably represent two new species of Vampyronema There was no clear differentiationbetweennematodes on the noctuids Mocis disseverans M latipes Agrapha oxygramma and Orthodes crenulata and they are all probablyconspeci c with V dibolia Thus far Noctuidonema has been found only on the tri ne noctuids Spodoptera and Pseudaletia whereasVampyronema has been found on both tri ne and quadri ne noctuids We found evidence for sequential evolution but not co-evolutionbetween these ectoparasitic nematodes and their lepidopteran hosts

Keywords ndash Acugutturus cockroaches ectoparasitic nematodes evolution Lepidoptera moths Noctuidonema Vampyronema

The nematode family Acugutturidaewas established byHunt (1980) to accommodate a newly described speciesAcugutturus parasiticus Hunt an ectoparasite of theAmerican cockroach Periplaneta americana L from theisland of St Lucia in the West Indies In establishingthis genus and family Hunt (1980) remarked that Aparasiticus appeared to have no close relatives amongthe Aphelenchoidoidea and was unique because of thegreat length of the stylet 50-60 sup1m in adults Otherdistinguishing features of A parasiticus include the longlength of the conus relative to the base and the absenceof basal swellings or knobs on the stylet the simple rose-thorn shaped spicules the presence of a gubernaculumand the position of the excretory pore posterior to themetacorpus (Hunt 1980) In 1996 A parasiticus was

curren Corresponding author e-mail omartitiftoncpespeachnetedu

found on P americana from the West Indian island ofGuadeloupe and on additional specimens collected fromSt Lucia (Marti et al 2000) However none was foundon cockroaches collected from mainland sites in FrenchGuiana and Mexico

In 1988 another ectoparasitic nematode Noctuido-nema guyanense Remillet amp Silvain was described onthe noctuid moths Spodoptera androgea S frugiperdaS latifascia S marima Anicla infecta and on severalspecies of Leucania from French Guiana (Remillet amp Sil-vain 1988) The total length of the stylet reported forthis species 135-165 sup1m was much longer than that ofA parasiticus and the authors also mentioned the pres-ence of basal knobs and an excretory pore located near thehead Anderson and Laumond (1990) while erroneously

copy Koninklijke Brill NV Leiden 2002 489Also available online - wwwbrillnl

OG Marti Jr et al

stating that the stylet of N guyanense lacks basal knobsprovided additionaldetails of the spicules redescribed themale and emended the diagnosis of the genus Noctuido-nema

Anderson and Laumond (1992) also described N dap-tria later redesignated as Vampyronema daptria by Hunt(1993) from the noctuid moth Lesmone porcia in Guade-loupe This species differs from A parasiticus in the ante-rior location of the excretory pore and in the shape of thespicules which have ventral arms separated from bladesand distinct elongated rostra The spicules of V daptriaare simpler in form than those of N guyanense and taperto ne points in contrast to the bluntly pointed spiculesof Vampyronema dibolia from Mocis latipes (Marti ampRogers 1995) The stylet however is only 59-69 sup1m infemales and is therefore much shorter than that of N guya-nense but differs from that of A parasiticus in havingdis-tinct knobs Vampyronema daptria was not included in thepresent study because we collected no host specimens andno nematodes were available for measurement Informa-tion about V daptria in this report was based on Andersonand Laumond (1992)

Hunt (1993) revised the taxonomyof the Acugutturidaeand established a new subfamily the Noctuidonematinaeto include only the acugutturid ectoparasites of mothsAcugutturus was retained in the Acugutturidae subfam-ily Acugutturinae as a monotypic genus Morphologicalcharacters distinguishing these subfamilies were primar-ily the small size and rose-thorn shape of the spicules ofAcugutturus the larger more complex spicules presentin Noctuidonema and the simpler spicules in the newlycreated genus Vampyronema which exhibited characterssuf ciently different from Noctuidonema to justify sepa-rate generic status

The only other species of Acugutturidae described todate is V dibolia (Marti amp Rogers) from the noctuidmoth Mocis latipes in south Georgia USA (Marti ampRogers 1995) This species has a stylet length of 83-114 sup1m in females an excretory pore located near thehead and simple spicules somewhat different in size andshape from those of V daptria In additionthe distributionof V dibolia over the body of its host is distinctlydifferent from that of N guyanense On M latipes alarge portion of the nematodes may frequently be foundon the anterior abdomen whereas on S frugiperda thenematodes are found primarily on the posterior abdomen(Rogers amp Marti 1992 1993 1994) In other species ofMocis (M disseverans M texana and M marcida) thenematodes were distributed in a manner similar to those

on M latipes Similarly in other species of Spodoptera(eg S dolichos S latifascia and S ornithogalli) thenematode distribution was similar to that found on Sfrugiperda (Marti unpubl)This difference in distributionon different host genera is quite striking and appears tobe a distinct biological difference between Noctuidonemaand Vampyronema

Acugutturid nematodes insert their stylet through thecuticle of the host and ingest haemolymph (Hunt 1980Marti et al 1990) Infestations of N guyanense onthe noctuid S frugiperda also reduce longevity and eggfertility (Simmons amp Rogers 1994 Marti amp Rogers2000) These studies indicate that the nematodes areparasitic and not merely phoretic or commensal

Since 1987 ectoparasitic nematodes have been col-lected from a variety of Lepidoptera from Fiji FrenchGuiana Guadeloupe and the southeastern United Statesof America (Rogers et al 1990a b 1997 Marti et al2000) Many of these nematode specimens have been ex-amined and measured Although Acugutturus is easily de- ned by its speci c suite of characters there are few dis-crete characters de ning differences between Noctuidone-matinae parasitising the Lepidoptera With the exceptionof obvious differences in spicule type between nematodesparasitising Spodopteraand Mocis spp all morphologicalcharacters examined appeared to be continuouslyvariablerather than discrete

We used discriminant analysis to determine which char-acters if any were important in de ning differences be-tween nematodes from different host species Of particu-lar interest was the possibility of determining whether anyof the nematodes from different host species represent in-dependently evolving lineages (species) according to theview ldquo that a species not be a speculative statementmade solely on the basis of a de nable lsquodifferencersquo butrather predictive based on evidence that suggests the twoprospective species are on independent trajectories rdquo(Adams 1998)

Materials and methods

Ectoparasitic nematodes measured in this study wereobtained from 13 species of Lepidoptera and from twopopulations of Periplaneta americana (Blattodea) col-lected from several different localities (Table 1) Periplan-eta americana from St Lucia were collected in 1997 byJeanette Williams All Lepidoptera and Blattodea spec-imens from Guadeloupe were collected in 1996 (Martiet al 2000) No male nematodes were found on Xylo-

490 Nematology

Speciation in the Acugutturidae

phanes chiron Lepidoptera specimens from Tifton Geor-gia USA were collected between 1988 and 1997 A totalof 15 groups of female nematodes and 14 groups of maleswere studied Nematodes from three of the 15 host specieshave been previously described These were Acugutturusparasiticus Hunt from P americana Noctuidonemaguya-nense Remillet amp Silvain from Spodoptera spp and sev-eral other noctuids and V dibolia (Marti amp Rogers) fromMocis latipes A fourth species not included in the presentanalysis was V daptria (Anderson amp Laumond)

Captured cockroaches were preserved whole in 10buffered formalin Lepidoptera were collected from Uni-versal traps baited with species-speci c pheromone luresor at night from a white sheet illuminated by a 125 or160W mercury vapour light powered by a portable elec-tric generator Abdomens were removed from moths andplaced into labelled vials containing 10 buffered for-malin Selected nematode specimens were later processedthrough an alcohol series to glycerin and mounted perma-nently on glass slides Specimens were studied by trans-mitted light differential interference contrast and phasecontrast optics Measurements of morphological charac-ters were obtained with a Nikon Optiphot photomicro-scope tted with a calibrated ocular micrometer All mea-surements are in sup1m Characters based on ratios or per-centages were calculated from appropriate original data

Statistica version 55 (StatSoft Inc Tulsa OK USA) wasused for all statistical analyses and graph production

STATISTICAL PROCEDURE

A total of 27 characters was measured from eachspecimen of each gender plus spicule and rostrum lengthin males and egg width and length in females From these29 characters per gender an additional three percentageand six ratio characters were calculated from the originaldata giving a total of 36 morphological variables in the nal dataset However not all characters could always befound or adequately measured in all specimens In the nal dataset therefore 18 morphologicalcharacters in themales and 16 in the females were used in addition to thesix ratios and three percentages The characters were asfollows maximum body width stylet length bulb lengthbulb width head width head length spicule length alongthe curve (males) gonad length tail length excretoryporeto anterior end anterior gonad to base of bulb anterior endto base of bulb body width at cloaca or vulva anterior endto cloaca or vulva body width at nerve ring anterior endto anterior gonad body length rostrum length (males) a(body lengthgreatest body width) b (body widthdistancefrom anterior end to base of bulb) c (body lengthtaillength) c0 (tail lengthbody width at vulva andor cloaca)V (cloaca or vulva to anterior end pound 100body length) G(reproductivesystem pound 100body length) gonad length as

Table 1 Host species family sub-family collection locality and numbers of ectoparasiticnematodes measured

Host species Code1 Family Subfamily Locality Nematodes

Periplaneta americana L pa Blattidae2 Blattinae St Lucia 5 13Periplaneta americana L pa2 Blattidae Blattinae Guadeloupe 7 18Spodoptera dolichos (Fabricius) sd Noctuidae3 Amphipyrinae GA USA 10 5Spodoptera frugiperda (JE Smith) sf Noctuidae Amphipyrinae GA USA 11 17Spodoptera latifascia (Walker) sl Noctuidae Amphipyrinae GA USA 6 12Spodoptera ornithogalli (Gueneacutee) so Noctuidae Amphipyrinae GA USA 10 10Eulepidotis addens (Walker) ea Noctuidae Catocalinae Guadeloupe 2 15Mocis disseverrans (Walker) md Noctuidae Catocalinae GA USA 4 9Mocis latipes Gueneacutee ml Noctuidae Catocalinae GA USA 16 39Orthodes crenulata (Butler) oc Noctuidae Hadeninae GA USA 4 9Pseudaletia unipuncta (Haworth) pu Noctuidae Hadeninae GA USA 9 18Anicla infecta (Ochsenheimer) ai Noctuidae Noctuinae GA USA 12 15Agrapha oxygramma (Geyer) ao Noctuidae Plusiinae GA USA 7 13Perigonia lusca Fabricius pl Sphingidae3 Macroglossinae Guadeloupe 4 19Xylophanes chiron Drury xc Sphingidae Macroglossinae Guadeloupe 0 18

1) These codes are used to identify host species in the gures2) Blattodea (cockroaches)For enumeration purposes P americana from St Lucia and Guadeloupe were counted as separate hosts3) Lepidoptera (moths)

Vol 4(4) 2002 491

OG Marti Jr et al

of body length stylet length as of body length andexcretory pore to anterior end as of body length

Male and female nematode untransformed data wereanalysed and plotted separately because different suitesof characters are involved for each gender Multivariateanalysis (MANOVA) was used to determine whether therewere differences between nematodes from different hostgroups based on the entire suite of characters for eachnematode gender Each character was then analysed withunivariateanalysis (ANOVA) to determine if the charactercontributedsigni cantly to differentiatingbetween at leasttwo host groups for each nematode gender The Bonfer-roni test was used to check against overestimation of sig-ni cance of univariate characters These procedures weredescribed by Footit and Sorensen (1992) and Sanmartiacutenand Martiacuten-Piera (1999)

Forward stepwise discriminant analysis was used toidentify which morphological characters made signi cantcontributions to de ning differences among nematodesfrom different host species Non-signi cant and redundantcharacters were identi ed and deleted from the modelsFour separate analyses were conducted for male and fe-male nematodes First nematodes from all host specieslisted in Table 1 were included (n D 15 female and n D14 male groups) Based on these results nematodes onBlattodea and Sphingidae were deleted from the secondstep which included only nematodes from noctuid hostsAgain morphological characters were ranked and non-signi cant (P gt 005) and redundant characters wereidenti ed and deleted from the model In the third steponly nematodes from groups having simple spicules simi-lar to those of V dibolia were included (n D 6 female andn D 6 male groups) The same procedure was followedin the last step which analysed only nematodes havingcomplex spicules similar to those of N guyanense (n D 5female and n D 5 male groups) This procedure producedeight models four for each nematode gender for discrim-inating between nematodes from all 15 host groups be-tween nematodes from noctuid hosts between nematodeshaving simple spicules (Vampyronema) and between ne-matodes having complex spicules (Noctuidonema)

Discriminant functions were used to calculate poste-rior probabilities that individual nematodes were classi- ed correctly into their respective groups Separate clas-si cations were obtained for male and female nematodesA high percentage of correct classi cation is a measure ofthe discriminatory power of the morphological characters(Footit amp Sorensen 1992) Discriminant functions wereevaluated for signi cance with chi-square and only signif-

icant (P lt 005) functions were used in data interpreta-tions Standard discriminant function coef cients the cor-relations of the morphological characters to the discrimi-nant functions were used to determine which morpholo-gical characters contributedmost to the differentiationbe-tween groups A large absolute value of the standard dis-criminant function coef cient indicates that a character isstrongly represented by the discriminant function

Two-dimensional graphs were produced that providedthe best differentiation between nematodes from differenthost species for each nematode gender Clearly de nedgroups were marked by 95 ellipses which delimit thearea in which a new observation is likely to occur Resultsfrom both nematode genders from each host species wereevaluated to arrive at a consensus regarding the status ofthe nematodes from each host group

The most important morphological characters identi- ed by discriminant analysis were used in cluster analysisto produce tree diagrams for each nematode gender Datafor each morphological character were standardised withmean D 0 and standard deviationD 1 Clusters were basedon Euclidean distance for the distance measure betweengroups and on complete linkage for the amalgamationrule Tree diagrams were scaled to (DlinkDmax)100andwere then evaluated and compared to results from dis-criminant analysis to arrive at a consensus regarding sim-ilarities between nematodes from different host species

EXAMINATION OF NEMATODES FROM ADDITIONAL

HOST SPECIES

We also examined additional nematodes from otherlepidopteran hosts from French Guiana and the FijiIslands These nematodes were not measured in detailbecause they were available only in small numberswere of uneven quality or both In all of the specimensfrom Noctuidae however we were able to make adetermination regarding their generic placement We wereable therefore to test our conclusions regarding thevalue of speci c characters in de ning the associationof nematode clades with clades and sub-families of theNoctuidae

Results

MANOVA showed differences regardless of nematodegender among nematodes from the 15 host speciesbased on all morphological characters All F values werehighly signi cant (P lt 0001 df D 13 91 for males

492 Nematology

Speciation in the Acugutturidae

and 14 113 for females) In addition ANOVA showedthat each morphological character contributed to thedifferentiation of nematodes of each gender from at leasttwo host species All F values were highly signi cant(P lt 0001 df D 11-13 83-93 for males and 14209-214 for females) Therefore all 18 characters sixratios and three percentages listed above were used insubsequentdiscriminant analysis to determine which weremost useful in distinguishing between nematodes fromdifferent hosts

STEP 1 15 HOST SPECIES (BLATTODEA AND

LEPIDOPTERA)

Discriminant analysis of female nematodes from 15host species revealed three separate groups i) fromthe two Periplaneta americana populations (Acugutturusparasiticus) ii) from the two species of Sphingidaeiii) all other host species (Fig 1A B) There were tensigni cant canonical roots accounting cumulatively for9978 of the explained variance between groups The rst two roots accounted for 9078 of the explainedvariance Examination of the factor structure matrixthe correlation of the morphological characters to thecanonical roots showed that the distance of the excretorypore from the anterior end the stylet length and V werethe characters most important in distinguishing betweenfemale nematodes from the 15 host species (Table 2)The other characters were of less importance althoughallwere signi cant for female nematodes

Discriminant analysis of male nematodes from 14 hostspecies (no male nematodes were found on Xylophaneschiron) revealed four distinct groups (Fig 1B Table 2)Canonical roots 1 (excretory pore) and 3 (stylet length)gave the best differentiation among males In additionto the differentiation of A parasiticus and the malenematodes on Perigonia lusca from each other and fromthe remaining host groups it was also found that malenematodes from Pseudaletia unipuncta were differentfrom those collected from the four species of SpodopteraThe two populations of A parasiticus did not differ butwere distinct from the other groups primarily due to thedifference in the position of the excretory pore

In the males there were eight signi cant (P lt 005)canonical roots accounting cumulatively for 9944 ofthe explained variance among males The rst threecanonical roots accounted for 9446 of the explainedvariance Canonical roots 1 and 3 are represented primar-ily by the excretory pore and the stylet length respec-

tively whereas root 2 (not shown) is represented primarilyby spicule length

STEP 2 11 HOST GROUPS (NEMATODES OF

LEPIDOPTERA NOCTUIDAE)

Analysis of morphological characters of nematodesparasitising only noctuids provides additional informa-tion differentiating several of these groups Among thefemales there were ten signi cant canonical roots ac-counting cumulatively for 100 of the explained vari-ance The rst three roots accounted cumulatively for806 of the explained variance with V stylet length andbulb width being the most important (Table 3) V was byfar the most important female character differentiatingne-matodes among the noctuid hosts and was more usefulthan tail length alone V provides clear separation of fe-male nematodes of Pseudaletia and Spodoptera from theother noctuid groups (Fig 1C) The former group is char-acterised by V gt 92 (ie short female tails) and by com-plex spicules and the latter group by V lt 92 (ie longerfemale tails) and simple spicules

Among male nematodes from noctuid hosts (Fig 1DTable 3) discriminant analysis provided clear differenti-ation in root 1 between males having complex spicules(nematodes from Pseudaletia and Spodoptera) vs thosehaving simple spicules (nematodes from Agrapha AniclaEulepidotis Mocis spp and Orthodes) There were ninesigni cant canonical roots which cumulativelyaccountedfor 9983 of the explained variance among groups The rst three roots accountingcumulatively for 920 of theexplained variance among groups were represented pri-marily by spicule length stylet length and stylet length asa percentage of body length respectively

Nematodes from Pseudaletia can be separated fromthose on Spodoptera by the shorter stylet length (Fig 1CD) Nematodes on Agrapha oxygramma can be separatedfrom M latipes primarily on the basis of the shorter styletlength (P lt 005) Similarly males on Anicla infectahave shorter spicules (P lt 005) than the nematodes onthe remaining host groups

Only two male nematodes from E addens were avail-able However these were distinct from nematodes on Mlatipes The female nematodes on E addens were alsowell differentiated from those on M latipes despite someoverlap (Fig 2A)

Vol 4(4) 2002 493

OG Marti Jr et al

Fig 1 Plot of discriminant functions for nematode morphometric characters Ellipses indicate 95 condence limits for selected hostspecies A Female nematodes from 15 host species (Blattodea and Lepidoptera) B Male nematodes from 14 host species (Blattodeaand Lepidoptera) C Female nematodes from 11 host species (Lepidoptera) D Male nematodes from 11 host species (Lepidoptera)Legend (Host Names) pa Periplaneta americana (St Lucia) pa2 P americana (Guadeloupe) ao Agrapha oxygramma ai Aniclainfecta ea Eulepidotis addens md Mocis disseverans ml M latipes oc Orthodes crenulata pl Perigonia lusca pu Pseudaletiaunipuncta sd Spodoptera dolichos sf S frugiperda sl S latifascia so S ornithogalli xc Xylophanes chiron

Table 2 Summary of forward stepwise discriminant analysis for male and female nematode morphological characters from 14 and 15host species respectively

Males ndash all groups (n D 14) Females ndash all groups (n D 15)

Step Character F (df1 df2) Character F (df1 df2)

1 Excretory pore 40313 (13 93) Excretory pore 61696 (14 214)2 Spicule length 20971 (13 92) Stylet length 16778 (14 213)3 Stylet length 5908 (13 91) V 15281 (14 212)4 Vulva or cloaca to ant end 1155 (13 90) Bulb width 1756 (14 211)5 Ant end to post bulb 905 (13 89) Excretory pore of body 1400 (14 210)6 Stylet percent of body 582 (13 88) a 1050 (14 209)7 B 702 (13 87) Ant end to post bulb 999 (14 208)8 G 640 (13 86) G 774 (14 207)9 Bulb width 447 (13 85) Vulva or cloaca to ant end 1023 (14 206)

10 Body width at nerve ring 419 (13 84) b 826 (14 205)

P lt 0001

494 Nematology

Speciation in the Acugutturidae

Table 3 Summary of forward stepwise discriminant analysis for male and female nematode morphological characters from 11 speciesof host Noctuidae (Lepidoptera)

Males ndash Lepidoptera groups1 (n D 11) Females ndash Lepidoptera groups1 (n D 11)

Step Character F (df1 df2) Character F (df1 df2)

1 Spicule length 24039 (10 80) V 10986 (10 150)2 Stylet length 2510 (10 79) Stylet length 6194 (10 149)3 Stylet of body length 1388 (10 78) Bulb width 1830 (10 148)4 Cloaca to anterior end 981 (10 77) Excretory pore of body 982 (10 147)5 Rostrum length 762 (10 76) Vulva to anterior end 935 (10 146)6 Ant end to ant gonad 700 (10 75) Body width at nerve ring 873 (10 145)7 G 795 (10 74) Stylet of body length 777 (10 144)8 c0 651 (10 73) Tail length 701 (10 143)9 Body width at nerve ring 615 (10 72) Head width 672 (10 142)

10 Bulb width 548 (10 71) Gonad of body length 595 (10 141)

1) Not including Perigonia lusca Xylophanes chiron or two populations from Periplaneta AmericanaP lt 0001

Fig 2 Plot of discriminant functions for nematode morphometric characters Ellipses indicate 95 condence limits for selected hostspecies A Female nematodes (Vampyronema) from six host species (Lepidoptera) B Male nematodes (Vampyronema) from six hostspecies (Lepidoptera)C Female nematodes (Noctuidonema) from ve host species (Lepidoptera)D Male nematodes (Noctuidonema)from ve host species (Lepidoptera) Legend (Host names) ao Agrapha oxygramma ai Anicla infecta ea Eulepidotis addens mdMocis disseverans ml M latipes oc Orthodes crenulata pu Pseudaletia unipuncta sd Spodoptera dolichos sf S frugiperda sl Slatifascia so S ornithogalli

Vol 4(4) 2002 495

OG Marti Jr et al

Table 4 Summary of forward stepwise discriminant analysis of morphological characters from male and female Vampyronema fromsix species of Noctuidae (Lepidoptera)

Males ndash Mocis group1 (n D 6) Females ndash Mocis group1 (n D 6)

Step Character F (df1 df2) Character F (df1 df2)

1 Stylet length 5807 (5 39) Stylet length 4573 (5 93)2 Rostrum length 1219 (5 38) Ant end to ant intestine 1879 (5 92)3 Bulb length 929 (5 37) Excretory pore of body 1803 (5 91)4 Ant end to ant gonad 837 (5 36) G 1500 (5 90)6 Pharyngeal gland width 511 (5 34) Stylet length of body 573 (5 89)7 Gonad length 438 (5 33) V 485 (5 88)8 Tail length 408 (5 32) Head length 476 (5 87)9 Bulb width 269 (5 31) Gonad length of body 445 (5 86)

10 Body width at nerve ring 355 (5 85)

1) Mocis group Agrapha oxygramma Anicla infecta Eulepidotis addens Mocis disseveransM latipes Orthodes crenulataP lt 005 P lt 001 P lt 0001

STEP 3 SIX HOST GROUPS (VAMPYRONEMANEMATODES WITH SIMPLE SPICULES)

The noctuidmoths Anicla infectaAgraphaoxygrammaEulepidotis addens Mocis disseverans M latipes andOrthodes crenulata are parasitised by Vampyronema inwhich the structure of the spicules is simpler and the tailof the females is longer than in Noctuidonema Analy-sis of morphometric characters of female nematodes fromthese hosts revealed ve signi cant canonical roots cu-mulatively accounting for 100 of the explainedvariancebetween groups The rst three roots representingprimar-ily stylet length distance from the anterior end to the baseof the bulb and the c ratio accounted cumulatively for9188 of the explained variance between groups Styletlength was the most important morphological characterdistinguishing female Vampyronema from different hosts(Fig 2A Table 4)

Analysis of results of male and female Vampyronemaon noctuid moths demonstrate the existence of fourdistinct clades One of these is V dibolia the speciesdescribed from M latipes The nematodes particularlythe males on M disseverans are very similar to V diboliaand we consider them to be conspeci c A second cladeis formed by nematodes from A infecta which are wellseparated from all the other groups (Fig 2A B) Styletlength is signi cantly shorter (P lt 001) in nematodeson A infecta than on those from the other hosts Athird Vampyronema clade is formed by nematodes onE addens which differ from V dibolia primarily in thedistance from the anterior end to the base of the bulb Thetwo male nematodes from E addens were also distinct

from other males an observation in agreement withresults from the females A fourth Vampyronema cladeis formed by nematodes from A oxygramma (Fig 2AB) Although the females appear to be intermediatebetween those on M latipes and E addens the malesare distinct based on differences in stylet and rostrumlength Nematodes on O crenulata are also intermediatebut males are very similar to V dibolia and we considerthem to be conspeci c

In the males there were ve signi cant canonical rootsrepresented primarily by stylet length spicule rostrumlength and bulb length respectively These accounted cu-mulatively for 929 of the explained variation betweenmales from these six host species Stylet length was themost important morphologicalcharacter in de ning males(Fig 2B) although rostrum length was also identi ed asan important character

STEP 4 FIVE HOST SPECIES (NOCTUIDONEMANEMATODES WITH COMPLEX SPICULES)

Discriminant analysis of female nematodes from fourSpodoptera species and Pseudaletia revealed four signif-icant (P lt 005) canonical roots The rst two rep-resenting primarily stylet length and body width at thevulva respectively accounted for a cumulative propor-tion of 8793 of the explained variance between groupsFemale nematodes from Pseudaletia were well separatedfrom those parasitising Spodoptera spp on the basis ofthe shorter stylet Nematodes from the four species ofSpodoptera were poorly de ned and showed a consider-able amount of overlap Even the groups that appeared to

496 Nematology

Speciation in the Acugutturidae

Table 5 Summary of forward stepwise discriminant analysis of morphological characters from male and female Noctuidonema from ve species of Spodoptera and Pseudaletia

Males ndash Spodoptera and Pseudaletia1 (n D 5) Females ndash Spodoptera and Pseudaletia1 (n D 5)

Step Character F (df1 df2) Character F (df1 df2)

1 Ant end to ant intestine 2524 (4 41) Stylet length 3914 (4 57)2 Body length 1491 (4 40) Body width at vulva 1250 (4 56)3 c0 1223 (4 39) Bulb width 591 (4 55)4 Bulb width 725 (4 38) a 498 (4 54)5 Body width at nerve ring 500 (4 37) Vulva to anterior end 380 (4 53)6 Spicule length 409 (4 36) Body width at nerve ring 579 (4 52)7 Maximum body width 265 (4 51)

1) Pseudaletia unipuncta Spodoptera dolichos S frugiperda S latifascia S ornithogalliP lt 005 P lt 001 P lt 0001

be the most different S frugiperda and S dolichos wereextensively overlapped (Fig 2D Table 5)

In the males there were three signi cant canonicalroots accounting for 9764 of the explained variancebetween groups The rst root represents the distancefrom the anterior end to the anterior intestine and V aboutequally whereas the second root represents primarily thebody width at the nerve ring (Fig 2D Table 5) Malenematodes from P unipuncta were well separated fromnematodes from Spodoptera which agrees with resultsfrom the females Males on S dolichosshowed a moderateamount of separation from males on the other threespecies of Spodoptera

DATA ANALYSIS

Eighteen morphological characters contribute to differ-entiation among male nematodes on 14 host species in-cluded in this study The most signi cant and useful ofthese are the distance of the excretory pore from the an-terior end spicule length and stylet length These threecharacters and V are the most useful in differentiating be-tween females (Table 2) Furthermore ANOVA showedthat the stylet length of nematodes on the two speciesof Sphingidae is signi cantly greater (P lt 0001) thanthe stylet length of nematodes on any of the other hostspecies The two populationsof nematodes from Sphingi-dae differ signi cantly in only eight morphological char-acters whereas each differs from V dibolia in about 20characters Females on P lusca differ from those on Xchiron in the longer length of the stylet greater length andwidth of the bulb greater distance of the excretory poreand bulb base from the anterior end greater body widthat the vulva and greater excretory pore distance as a per-

centage of body length However in the absence of malesfrom X chiron no conclusion can be reached on whethernematodes on Sphingidae are conspeci c The consistentseparation of nematodes parasitising Sphingidae from allother Acugutturidae suggests that they probably representat least one new genus Particularly remarkable is the longstylet which represents about 37 of the body length andaverages about 190 sup1m in length (Fig 3)

Discriminant analysis shows that the ve groups ofnematodes with complex spicules (Noctuidonema) arewell separated (Fig 1C D) from the six groups withsimple spicules (Vampyronema) Although the discretecharacter of spicule structure is suf cient to separate thesetwo genera particularly good separation can also be madebased on important continuously variable characters suchas stylet length distance of the excretory pore to theanterior end and the V ratio Furthermore within eachof these two genera separate groups can be detectedbased primarily on differences in stylet length Thusthe nematodes on Pseudaletia unipuncta differ fromthe nematodes on the four Spodoptera spp (Fig 1CD) and those on Anicla infecta Agrapha oxygrammaand Eulepidotis addens differ from V dibolia and itsconspeci c on M disseverans (Fig 2A B)

Female nematodes on P unipuncta differ (P lt 005)from those on S dolichos S frugiperda S latifasciaand S ornithogalli in eight 18 14 and 14 morpholo-gical characters respectivelyThese results providestrongsupport for separate taxonomic status of nematodes onP unipuncta as a new species of Noctuidonema Resultsfrom simple scatterplots prepared for each gender usingthe most importantmorphologicalcharacters identi ed bydiscriminant analysis agree with the preceding analysis

Vol 4(4) 2002 497

OG Marti Jr et al

Fig 3 Plot of stylet length of ectoparasitic nematodes from insect hosts (Blattodea and Lepidoptera) A Female nematodes from15 host species B Male nematodes from 14 host species Legend (Host names) ao Agrapha oxygramma ai Anicla infecta eaEulepidotis addens md Mocis disseverans ml M latipes oc Orthodes crenulata pa Periplaneta americana (St Lucia) pa2 Pamericana (Guadeloupe) pl Perigonia lusca pu Pseudaletia unipuncta sd Spodoptera dolichos sf S frugiperda sl S latifascia soS ornithogalli xc Xylophanes chiron

Table 6 Percent correct classication of male nematodes based on posterior probabilities that a case belongs to a particular hostspecies

Host species Percent correct classi cations ndash Percent correct classi cations ndashmale nematodes female nematodes

n Hosts Hosts Hosts Hosts n Hosts Hosts Hosts Hosts1-15 1-11 1-6 7-11 1-15 1-11 1-6 7-11

1 Agrapha oxygramma 7 100 100 100 ndash 12 917 100 917 ndash2 Anicla infecta 12 100 100 100 ndash 15 100 100 100 ndash3 Eulepidotis addens 2 100 100 100 ndash 15 100 100 100 ndash4 Mocis disseverans 4 100 75 75 ndash 9 889 889 889 ndash5 Mocis latipes 16 100 100 100 ndash 39 923 974 948 ndash6 Orthodes crenulata 4 100 100 100 ndash 9 889 889 889 ndash7 Spodoptera dolichos 10 100 100 ndash 90 5 100 100 ndash 808 Spodoptera frugiperda 11 909 909 ndash 100 17 765 824 ndash 7659 Spodoptera latifascia 6 667 833 ndash 50 12 75 833 ndash 667

10 Spodoptera ornithogalli 10 100 100 ndash 100 10 90 90 ndash 9011 Pseudaletia unipuncta 9 100 100 ndash 100 18 100 944 ndash 94412 Perigonia lusca 4 100 ndash ndash ndash 19 938 ndash ndash ndash13 Periplaneta americana (1) 7 100 ndash ndash ndash 18 882 ndash ndash ndash14 Periplaneta americana (2) 5 100 ndash ndash ndash 13 846 ndash ndash ndash15 Xylophanes chiron 0 ndash ndash ndash ndash 18 100 ndash ndash ndash

(1) St Lucia(2) Guadeloupe

498 Nematology

Speciation in the Acugutturidae

Fig 4 Tree diagram for ectoparasitic nematodes from insecthost species (Blattodea and Lepidoptera) The tree was pro-duced using Euclidean distance as the distance measure andcomplete linkage as the amalgamation rule Both graphs areplotted on an X-axis standardised by (DlinkDmax)100 whichrepresents the percentage of the range from the maximum tothe minimum distance in the data A Female nematodes datafor excretory pore stylet length V bulb width a anterior endto anterior intestine and vulva to anterior end were used BMale nematodes data for excretory pore spicule length andstylet length were used Legend (Host Names) pa Periplan-eta americana (St Lucia) pa2 P americana (Guadeloupe) aoAgrapha oxygramma ai Anicla infecta ea Eulepidotis addensmd Mocis disseverans ml M latipes oc Orthodes crenulatapl Perigonia lusca pu Pseudaletia unipuncta sd Spodopteradolichos sf S frugiperda sl S latifascia so S ornithogallixc Xylophanes chiron

Spicule length was not an important character in dis-tinguishing among male Noctuidonema Male nematodeson S dolichos and S frugiperda showed the most sepa-ration with moderate overlap whereas males on S or-nithogalli appeared to be most similar to those on S

dolichos and males on S latifascia appeared most similarto those on S frugiperda Results from both genders agreethat nematodes on Pseudaletia are different from thoseon Spodoptera and that the nematodes on S dolichos andS frugiperda are the most different with the nematodeson S latifascia and S ornithogallibeing intermediate be-tween these two Overall the results do not support theexistence of separate species of Noctuidonema on hostsof the genus Spodoptera (Fig 4 Table 6)

These results provide evidence that nematodes on Ainfecta A oxygramma and E addens are different fromeach other and from the remaining groups having simplespicules Nematodes on M disseverans and O crenulatawere not distinct from V dibolia the species parasitisingM latipes Nematodes on M disseverans differed signif-icantly from V dibolia in only six morphological charac-ters in the females and ten in the males The differenceswere small and can be attributed to adaptation of conspe-ci c nematodes to parasitism of closely related sympatrichost species

Comparisonof the description of V daptria to measure-ments of other Vampyronema in this study indicates that itdiffers in stylet length and spicule shape V daptria hasthe shortest stylet length observed in the genus rangingfrom 53-60 sup1m in males and 59-69 sup1m in females whichis shorter than the stylet length of nematodes on A in-fecta (Fig 3) In addition spicules of V daptria taper to ne points whereas in Vampyronema from other hosts thespicule tips are blunt V daptria on L porcia appears tobe a distinct species and we therefore do not assign any ofthe Vampyronema from this study to it

Nematodes on the St Lucia and Guadeloupe popula-tions of Periplaneta americana are not well differenti-ated from each other based on the suite of characters thatwere examined (Fig 1A B) Although there are a fewsigni cantly different characters between the two popula-tions the differences probably do not represent interspe-ci c variation In addition the characters of both popula-tions are in agreement with the published description ofAcugutturus parasiticus (Hunt 1980)

Examination of additional specimens of ectoparasiticnematodes from other species of Lepidoptera showed thatmost were Vampyronema whereas Noctuidonema wasfound only on moths from two genera Pseudaletia andSpodoptera (Table 7)

Vol 4(4) 2002 499

OG Marti Jr et al

Table 7 Association of quadrine and trine Noctuidae with Vampyronema and Noctuidonema

Host Taxon Hind Wing Venation1 Locality2 Nematodes Genus

NoctuidaeCatocalinae

Anomis oedema Gueneacutee quadri ne FG 4 1 VampyronemaBaniana sp quadri ne FG 1 0 VampyronemaBaniana ostia Druce quadri ne FG 7 0 VampyronemaEuglyphis sp quadri ne FG 1 0 VampyronemaEulepidotis addens (Walker) quadri ne GU 5 2 VampyronemaItomia opisthographa (Gueneacutee) quadri ne FG 5 1 VampyronemaLesmone formularis (Geyer) quadri ne FG 1 0 VampyronemaLesmone porcia (Stoll) quadri ne GU ndash3 Vampyronema3

Melipotis fasciolaris (Huumlbner) quadri ne FG 0 1 VampyronemaMetalectra sp quadri ne FG 4 10 VampyronemaMetria sp quadri ne FG 0 1 VampyronemaMocis dif uens (Gueneacutee) quadri ne FG 1 1 Vampyronema4

Mocis disseverans (Walker) quadri ne GA 9 4 Vampyronema4

Mocis latipes Gueneacutee quadri ne GA 9 16 Vampyronema4

Perasia ora (Cramer)5 quadri ne FG 10 0 VampyronemaSelenisa sp quadri ne FG 4 0 VampyronemaZale ctilis (Gueneacutee) quadri ne FG 4 1 Vampyronema

PlusiinaeAgrapha oxygramma (Geyer) quadri ne GA 13 7 VampyronemaArgyrogramma verruca (Fabricius) quadri ne FG 0 2 VampyronemaChrysodeixis chalcites (Esper) quadri ne Fiji 8 7 Vampyronema

AmphipyrinaeSpodoptera androgea (Stoll) tri ne FG 4 0 Noctuidonema6

Spodoptera dolichos (Fabricius) tri ne GA 5 10 Noctuidonema6

Spodoptera frugiperda (JE Smith) tri ne GA 17 11 Noctuidonema6

Spodoptera latifascia (Walker) tri ne GA 12 6 Noctuidonema6

Spodoptera litura (Fabricius) tri ne Fiji 15 7 Noctuidonema6

Spodoptera ornithogalli (Gueneacutee) tri ne GA 10 10 Noctuidonema6

HadeninaeOrthodes crenulata (Butler) tri ne GA 9 4 VampyronemaPseudaletia unipuncta (Haworth) tri ne GA 18 9 Noctuidonema

NoctuinaeAnicla infecta (Ochsenheimer) tri ne GA 15 12 VampyronemaTandilia rodea (Schaus) tri ne FG 2 3 Vampyronema

PyralidaePyraustinae

Lamprosema moccalis Schaus ndash FG 4 0 VampyronemaSphingidae

MacroglossinaeErinnys obscura (Fabricius) ndash FG 22 2 Vampyronema7

Perigonia lusca Fabricius ndash GU 19 4 Vampyronema7

Xylophanes chiron Drury ndash GU 8 0 Vampyronema7

1) Quadri ne and tri ne hindwing venation applies to Noctuidae only Pyralidae and Sphingidae are included for completeness thestatus of Plusiinae as quadri nes is ambiguous because it groups together mostly species having quadri ne wing venation yet is oftenincluded with chie y tri ne subfamilies (Kitching amp Rawlins 1998) 2 Fiji Fiji Islands FG French Guiana GU Guadeloupe GAGeorgia USA 3 V daptria Anderson amp Laumond 1992 designationbased on Hunt (1993) 4 V dibolia 5 Synonymous with Nymbisarcuata Walker (in Rogers et al 1990a and Simmons amp Rogers 1996) 6 N guyanense 7 Probably represents an undescribed genus

500 Nematology

Speciation in the Acugutturidae

Discussion

The position of the excretory pore posterior to thebulb the comparatively short length of the stylet thecomparatively long tail and the simple rose-thorn-shapedspicules are suf cient to clearly separate Acugutturus aparasite of Blattodea from Vampyronema and Noctuido-nema parasites of Lepidoptera Sampled populations re-veal four inferred monophyleticgroups Acugutturus par-asiticus from two West Indian populations of Periplan-eta americana a new genus on Sphingidae Noctuido-nema guyanense from Spodoptera spp with nematodeson P unipuncta as a probable new species of Noctuido-nema and Vampyronema spp on the remaining hosts inthis study with the nematodes on A infecta representinga new species of Vampyronema

Each inferred monophyletic group can be de ned byseveral unique combinations of characters quantitativeand qualitative that satisfy the minimal requirements forspecies diagnosis (Wheeler 1999 Wheeler amp Platnick2000) Such characters appear to represent lineage inde-pendence via autapomorphies hypotheses ideally testedvia outgroup comparison Such tests while more philo-sophically satisfying (Adams 1998 Wiley amp Mayden2000) are beyond the scope of the present paper

According to Kitching and Rawlins (1998) differencesin hindwing venation broadly divide the Noctuidae intotwo phenetic groups tri nes and quadri nes Catocalinaeand Plusiinae are quadri nes (but see comment 1 in Ta-ble 7) whereas the other noctuid subfamilies mentionedin this study are tri nes Complex spicules probably rep-resent a derived condition found thus far only in nema-todes on two genera of tri ne noctuids Pseudaletia andSpodoptera while the simple spicules of Vampyronemaprobably represent a plesiomorphic state found in bothtri ne and quadri ne noctuids The extremely long styletsof nematodes from Sphingidae may have arisen in re-sponse to parasitisation of large moths

The Acugutturidae are associated with two very dis-tantly related insect orders Blattodea and Lepidoptera(Endopterygota) There is no record of occurrence ofthe Acugutturidae on any other insect order Among theLepidoptera the Acugutturidae and more precisely theNoctuidonematinae are known from only six familiesGeometridae Lasiocampidae Noctuidae NotodontidaePyralidae and Sphingidae (Marti et al 2000) most ofwhich are not closely related with the exceptionof Lasio-campidae and Sphingidae members of the Bombycoidea(and their relatives) clade (Lemaire amp Minet 1998) and

Notodontidaeand Noctuidae members of the Noctuoideasuperfamily (Kitching amp Rawlins 1998)

One explanation for such a surprising distribution ofthe Acugutturidae on their hosts may be that availabledata are too sparse due to insuf cient sampling of otherlepidopteran families While this may be true the studyconducted by Marti et al (2000) shows that there isa clear speci city in nematode host preference Martiet al (2000) updated the list of lepidopteran hosts ofNoctuidonematinae It seems clear that although manyLepidoptera species belonging to 11 families and 115genera have been examined for nematodes only membersof the six aforementioned obtectomeran (Kristensen ampSkalski 1998) families have been found to be positivefor Noctuidonematinae With the exception of Pyralidaeall these families are members of the Macrolepidopteraclade Moreover the positive Noctuidae species representnearly 90 of the total positive species found Thereforewe cannot state that the Noctuidonematinae are notpotentially parasites of most of the lepidopteran familiesbut we can state that at least in the Neotropical Regionthey parasitise the Macrolepidoptera preferentially andamong that clade mostly a single family the Noctuidae(Marti et al 2000)

The Aphelenchida include several different trophictypes with many species being mycetophagous or phy-toparasitic with about one third of known species be-ing entomoparasitic (Hunt 1993) Considering the list ofinsect hosts we cannot hypothesise that the Acugutturi-dae are derived from a single ancestor transferring froma plant to an insect host It seems more realistic to pro-pose several transfers from plant to insect hosts during thehistory of the Acugutturidae A transfer (Acugutturinae)occurred on the Blattodea and another one (Noctuidone-matinae) on the Lepidoptera Another transfer may haveoccurred on the Bombycoidea sensu lato If at least twosuch principal transfers occurred then it is possible thatthe Acugutturinae and Noctuidonematinae would not beconfamilial and that these two taxonomicunits deserve fa-milial rank Cladistic and molecularphylogeneticanalysiswill help to solve that problem The existenceof symbioticbacteria in these nematodes (Marti et al 1995) shouldprovide an independentmeans of analysing these relation-ships

Concerning the relationship between Lepidoptera andits ectoparasitic nematodes the present work highlightsthe following situation a new genus of Noctuidone-matinae parasitises Sphingidae a genuine Bombycoidea(Lemaire amp Minet 1998) Vampyronema parasitises the

Vol 4(4) 2002 501

OG Marti Jr et al

quadri ne noctuid subfamily Catocalinae (sensu lato in-cluding Ophiderinae) Plusiinae an Hadeninae and aNoctuinae (both typical tri ne noctuids) and Noctui-donema spp parasitises another Hadeninae and severalSpodoptera a tri ne noctuid genus previously includedin the Amphipyrinae but now of uncertain subfamilial af- liation (Poole 1995 Kitching amp Rawlins 1998) Exceptfor nematodes on Sphingidae which likely correspond toa well-separated clade whose phylogenetic relationshipto other Noctuidonematinae needs to be speci ed it ap-pears that Vampyronema and Noctuidonema both para-sitise Noctuidae but that Vampyronema parasitises bothquadri ne and tri ne noctuids Noctuidonema possibly aderived form of Noctuidonematinae is restricted to a fewgenera (Spodoptera and Pseudaletia) of tri ne noctuidsthat are not closely related (Mitchell et al 2000)

Recent work has questioned the monophylyof the Noc-tuidae and there is a growing feeling among moleculartaxonomists that the Noctuidae may be paraphyletic sev-eral quadri ne subfamilies including Catocalinae beingmore closely related to Arctiidae and Lymantriidae than totri ne noctuids (Weller et al 1994 Mitchell et al 1997Fang et al 2000 Mitchell et al 2000) The latter groupis considered as monophyleticClearly systematics of theNoctuidonematinaedo not match the tentative new classi- cation of the Noctuidae

This lack of congruence between parasite and hostsystematics leads us to consider that the relationshipsbetween Acugutturidae and their hosts have not beenguided by coevolutionary forces in the sense of Ehrlichand Raven (1964) but by horizontal transfers (Rogers ampMarti 1996) and sequential evolution (Jermy 1984 Fu-tuyma amp McCafferty 1990) The Acugutturidae and es-pecially the Noctuidonematinae have likely experiencedan adaptive radiation on higher Lepidoptera long aftertheir host diversi cation History of the Noctuidonemati-nae has been characterised by host transfer events ontofairly distantly related hosts followed by adaptation tonew hosts leading to speciation processes that now permitus to distinguish easily fairly species-speci c populationsor strains of Vampyronema or Noctuidonema In mothsthese host transfers may have occurred through contami-nation of nematode-free species by nematode eggs juve-niles or adults left on resting or feeding surfaces (leavesor owers) by individualsof infested species The successof such contamination may be linked to the availabilityand accessibility of host intersegmental membranes thinenough to be pierced by nematode stylets a factor that

may have physically and dramatically restricted the hostrange of all Acugutturidae

The present morphological analysis of acugutturid ne-matodes parasitising 13 species of Lepidoptera and onespecies of cockroach revealed a greater degree of diversitythan previously known in this group The genus Vampy-ronema grouping nematodes with simple spicules in-cludes two inferred new species in addition to V dibo-lia and V daptria already known from the LepidopteraThe genus Noctuidonema grouping nematodes of Lepi-doptera with complex spicules includes a new speciesand the nematodes found on Sphingidae probably belongto a new genus The relationshipbetween these nematodesand their hosts appears to be much more complex thanpreviously thought In addition to their bene ts as para-sites of pest Lepidoptera acugutturid nematodes are aninteresting model for parasite-host coevolutionarystudies

Acknowledgements

We thank David Hunt (CABI Bioscience UK) DanielLachaise (CNRS France) Alvin Simmons (USDA Char-leston SC USA) and S Patricia Stock (University ofCalifornia Davis CA USA) for reviewing earlier ver-sions of this paper We thank Jeanette Williams and JohnClayton for collecting cockroaches in St Lucia and mothsin Fiji respectively We thank the following individualsfor identifyingor con rming our identi cations of insectsDavid Nickle (Smithsonian Washington DC USA) forPeriplaneta americana Pierre Zagatti (INRA France)for Pyralidae and Hugo Kons (University of FloridaGainesville FL USA) for Orthodes crenulata

References

ADAMS BJ (1998) Species concepts and evolutionary para-digm in modern nematology Journal of Nematology 30 1-21

ANDERSON RV amp LAUMOND C (1990) NoctuidonemaguyanenseRemillet amp Silvain 1988 morphologie du spiculeredescription du macircle et diagnose amendeacutee du genre Noctui-donema Revue de Neacutematologie 13 433-436

ANDERSON RV amp LAUMOND C (1992) Noctuidonemadaptria n sp (Nematoda Aphelenchoididae)an ectoparasiteof the moth Lesmone porcia (Stoll) Journal of Nematology24 16-22

EHRLICH P amp RAVEN P (1964) Butter ies and plants astudy in evolution Evolution 18 586-608

FANG QQ MITCHELL A REGIER JC MITTER CFRIEDLANDER TP amp POOLE RW (2000) Phylogenetic

502 Nematology

Speciation in the Acugutturidae

phanes chiron Lepidoptera specimens from Tifton Geor-gia USA were collected between 1988 and 1997 A totalof 15 groups of female nematodes and 14 groups of maleswere studied Nematodes from three of the 15 host specieshave been previously described These were Acugutturusparasiticus Hunt from P americana Noctuidonemaguya-nense Remillet amp Silvain from Spodoptera spp and sev-eral other noctuids and V dibolia (Marti amp Rogers) fromMocis latipes A fourth species not included in the presentanalysis was V daptria (Anderson amp Laumond)

Captured cockroaches were preserved whole in 10buffered formalin Lepidoptera were collected from Uni-versal traps baited with species-speci c pheromone luresor at night from a white sheet illuminated by a 125 or160W mercury vapour light powered by a portable elec-tric generator Abdomens were removed from moths andplaced into labelled vials containing 10 buffered for-malin Selected nematode specimens were later processedthrough an alcohol series to glycerin and mounted perma-nently on glass slides Specimens were studied by trans-mitted light differential interference contrast and phasecontrast optics Measurements of morphological charac-ters were obtained with a Nikon Optiphot photomicro-scope tted with a calibrated ocular micrometer All mea-surements are in sup1m Characters based on ratios or per-centages were calculated from appropriate original data

Statistica version 55 (StatSoft Inc Tulsa OK USA) wasused for all statistical analyses and graph production

STATISTICAL PROCEDURE

A total of 27 characters was measured from eachspecimen of each gender plus spicule and rostrum lengthin males and egg width and length in females From these29 characters per gender an additional three percentageand six ratio characters were calculated from the originaldata giving a total of 36 morphological variables in the nal dataset However not all characters could always befound or adequately measured in all specimens In the nal dataset therefore 18 morphologicalcharacters in themales and 16 in the females were used in addition to thesix ratios and three percentages The characters were asfollows maximum body width stylet length bulb lengthbulb width head width head length spicule length alongthe curve (males) gonad length tail length excretoryporeto anterior end anterior gonad to base of bulb anterior endto base of bulb body width at cloaca or vulva anterior endto cloaca or vulva body width at nerve ring anterior endto anterior gonad body length rostrum length (males) a(body lengthgreatest body width) b (body widthdistancefrom anterior end to base of bulb) c (body lengthtaillength) c0 (tail lengthbody width at vulva andor cloaca)V (cloaca or vulva to anterior end pound 100body length) G(reproductivesystem pound 100body length) gonad length as

Table 1 Host species family sub-family collection locality and numbers of ectoparasiticnematodes measured

Host species Code1 Family Subfamily Locality Nematodes

Periplaneta americana L pa Blattidae2 Blattinae St Lucia 5 13Periplaneta americana L pa2 Blattidae Blattinae Guadeloupe 7 18Spodoptera dolichos (Fabricius) sd Noctuidae3 Amphipyrinae GA USA 10 5Spodoptera frugiperda (JE Smith) sf Noctuidae Amphipyrinae GA USA 11 17Spodoptera latifascia (Walker) sl Noctuidae Amphipyrinae GA USA 6 12Spodoptera ornithogalli (Gueneacutee) so Noctuidae Amphipyrinae GA USA 10 10Eulepidotis addens (Walker) ea Noctuidae Catocalinae Guadeloupe 2 15Mocis disseverrans (Walker) md Noctuidae Catocalinae GA USA 4 9Mocis latipes Gueneacutee ml Noctuidae Catocalinae GA USA 16 39Orthodes crenulata (Butler) oc Noctuidae Hadeninae GA USA 4 9Pseudaletia unipuncta (Haworth) pu Noctuidae Hadeninae GA USA 9 18Anicla infecta (Ochsenheimer) ai Noctuidae Noctuinae GA USA 12 15Agrapha oxygramma (Geyer) ao Noctuidae Plusiinae GA USA 7 13Perigonia lusca Fabricius pl Sphingidae3 Macroglossinae Guadeloupe 4 19Xylophanes chiron Drury xc Sphingidae Macroglossinae Guadeloupe 0 18

1) These codes are used to identify host species in the gures2) Blattodea (cockroaches)For enumeration purposes P americana from St Lucia and Guadeloupe were counted as separate hosts3) Lepidoptera (moths)

Vol 4(4) 2002 491

OG Marti Jr et al

of body length stylet length as of body length andexcretory pore to anterior end as of body length

Male and female nematode untransformed data wereanalysed and plotted separately because different suitesof characters are involved for each gender Multivariateanalysis (MANOVA) was used to determine whether therewere differences between nematodes from different hostgroups based on the entire suite of characters for eachnematode gender Each character was then analysed withunivariateanalysis (ANOVA) to determine if the charactercontributedsigni cantly to differentiatingbetween at leasttwo host groups for each nematode gender The Bonfer-roni test was used to check against overestimation of sig-ni cance of univariate characters These procedures weredescribed by Footit and Sorensen (1992) and Sanmartiacutenand Martiacuten-Piera (1999)

Forward stepwise discriminant analysis was used toidentify which morphological characters made signi cantcontributions to de ning differences among nematodesfrom different host species Non-signi cant and redundantcharacters were identi ed and deleted from the modelsFour separate analyses were conducted for male and fe-male nematodes First nematodes from all host specieslisted in Table 1 were included (n D 15 female and n D14 male groups) Based on these results nematodes onBlattodea and Sphingidae were deleted from the secondstep which included only nematodes from noctuid hostsAgain morphological characters were ranked and non-signi cant (P gt 005) and redundant characters wereidenti ed and deleted from the model In the third steponly nematodes from groups having simple spicules simi-lar to those of V dibolia were included (n D 6 female andn D 6 male groups) The same procedure was followedin the last step which analysed only nematodes havingcomplex spicules similar to those of N guyanense (n D 5female and n D 5 male groups) This procedure producedeight models four for each nematode gender for discrim-inating between nematodes from all 15 host groups be-tween nematodes from noctuid hosts between nematodeshaving simple spicules (Vampyronema) and between ne-matodes having complex spicules (Noctuidonema)

Discriminant functions were used to calculate poste-rior probabilities that individual nematodes were classi- ed correctly into their respective groups Separate clas-si cations were obtained for male and female nematodesA high percentage of correct classi cation is a measure ofthe discriminatory power of the morphological characters(Footit amp Sorensen 1992) Discriminant functions wereevaluated for signi cance with chi-square and only signif-

icant (P lt 005) functions were used in data interpreta-tions Standard discriminant function coef cients the cor-relations of the morphological characters to the discrimi-nant functions were used to determine which morpholo-gical characters contributedmost to the differentiationbe-tween groups A large absolute value of the standard dis-criminant function coef cient indicates that a character isstrongly represented by the discriminant function

Two-dimensional graphs were produced that providedthe best differentiation between nematodes from differenthost species for each nematode gender Clearly de nedgroups were marked by 95 ellipses which delimit thearea in which a new observation is likely to occur Resultsfrom both nematode genders from each host species wereevaluated to arrive at a consensus regarding the status ofthe nematodes from each host group

The most important morphological characters identi- ed by discriminant analysis were used in cluster analysisto produce tree diagrams for each nematode gender Datafor each morphological character were standardised withmean D 0 and standard deviationD 1 Clusters were basedon Euclidean distance for the distance measure betweengroups and on complete linkage for the amalgamationrule Tree diagrams were scaled to (DlinkDmax)100andwere then evaluated and compared to results from dis-criminant analysis to arrive at a consensus regarding sim-ilarities between nematodes from different host species

EXAMINATION OF NEMATODES FROM ADDITIONAL

HOST SPECIES

We also examined additional nematodes from otherlepidopteran hosts from French Guiana and the FijiIslands These nematodes were not measured in detailbecause they were available only in small numberswere of uneven quality or both In all of the specimensfrom Noctuidae however we were able to make adetermination regarding their generic placement We wereable therefore to test our conclusions regarding thevalue of speci c characters in de ning the associationof nematode clades with clades and sub-families of theNoctuidae

Results

MANOVA showed differences regardless of nematodegender among nematodes from the 15 host speciesbased on all morphological characters All F values werehighly signi cant (P lt 0001 df D 13 91 for males

492 Nematology

Speciation in the Acugutturidae

and 14 113 for females) In addition ANOVA showedthat each morphological character contributed to thedifferentiation of nematodes of each gender from at leasttwo host species All F values were highly signi cant(P lt 0001 df D 11-13 83-93 for males and 14209-214 for females) Therefore all 18 characters sixratios and three percentages listed above were used insubsequentdiscriminant analysis to determine which weremost useful in distinguishing between nematodes fromdifferent hosts

STEP 1 15 HOST SPECIES (BLATTODEA AND

LEPIDOPTERA)

Discriminant analysis of female nematodes from 15host species revealed three separate groups i) fromthe two Periplaneta americana populations (Acugutturusparasiticus) ii) from the two species of Sphingidaeiii) all other host species (Fig 1A B) There were tensigni cant canonical roots accounting cumulatively for9978 of the explained variance between groups The rst two roots accounted for 9078 of the explainedvariance Examination of the factor structure matrixthe correlation of the morphological characters to thecanonical roots showed that the distance of the excretorypore from the anterior end the stylet length and V werethe characters most important in distinguishing betweenfemale nematodes from the 15 host species (Table 2)The other characters were of less importance althoughallwere signi cant for female nematodes

Discriminant analysis of male nematodes from 14 hostspecies (no male nematodes were found on Xylophaneschiron) revealed four distinct groups (Fig 1B Table 2)Canonical roots 1 (excretory pore) and 3 (stylet length)gave the best differentiation among males In additionto the differentiation of A parasiticus and the malenematodes on Perigonia lusca from each other and fromthe remaining host groups it was also found that malenematodes from Pseudaletia unipuncta were differentfrom those collected from the four species of SpodopteraThe two populations of A parasiticus did not differ butwere distinct from the other groups primarily due to thedifference in the position of the excretory pore

In the males there were eight signi cant (P lt 005)canonical roots accounting cumulatively for 9944 ofthe explained variance among males The rst threecanonical roots accounted for 9446 of the explainedvariance Canonical roots 1 and 3 are represented primar-ily by the excretory pore and the stylet length respec-

tively whereas root 2 (not shown) is represented primarilyby spicule length

STEP 2 11 HOST GROUPS (NEMATODES OF

LEPIDOPTERA NOCTUIDAE)

Analysis of morphological characters of nematodesparasitising only noctuids provides additional informa-tion differentiating several of these groups Among thefemales there were ten signi cant canonical roots ac-counting cumulatively for 100 of the explained vari-ance The rst three roots accounted cumulatively for806 of the explained variance with V stylet length andbulb width being the most important (Table 3) V was byfar the most important female character differentiatingne-matodes among the noctuid hosts and was more usefulthan tail length alone V provides clear separation of fe-male nematodes of Pseudaletia and Spodoptera from theother noctuid groups (Fig 1C) The former group is char-acterised by V gt 92 (ie short female tails) and by com-plex spicules and the latter group by V lt 92 (ie longerfemale tails) and simple spicules

Among male nematodes from noctuid hosts (Fig 1DTable 3) discriminant analysis provided clear differenti-ation in root 1 between males having complex spicules(nematodes from Pseudaletia and Spodoptera) vs thosehaving simple spicules (nematodes from Agrapha AniclaEulepidotis Mocis spp and Orthodes) There were ninesigni cant canonical roots which cumulativelyaccountedfor 9983 of the explained variance among groups The rst three roots accountingcumulatively for 920 of theexplained variance among groups were represented pri-marily by spicule length stylet length and stylet length asa percentage of body length respectively

Nematodes from Pseudaletia can be separated fromthose on Spodoptera by the shorter stylet length (Fig 1CD) Nematodes on Agrapha oxygramma can be separatedfrom M latipes primarily on the basis of the shorter styletlength (P lt 005) Similarly males on Anicla infectahave shorter spicules (P lt 005) than the nematodes onthe remaining host groups

Only two male nematodes from E addens were avail-able However these were distinct from nematodes on Mlatipes The female nematodes on E addens were alsowell differentiated from those on M latipes despite someoverlap (Fig 2A)

Vol 4(4) 2002 493

OG Marti Jr et al

Fig 1 Plot of discriminant functions for nematode morphometric characters Ellipses indicate 95 condence limits for selected hostspecies A Female nematodes from 15 host species (Blattodea and Lepidoptera) B Male nematodes from 14 host species (Blattodeaand Lepidoptera) C Female nematodes from 11 host species (Lepidoptera) D Male nematodes from 11 host species (Lepidoptera)Legend (Host Names) pa Periplaneta americana (St Lucia) pa2 P americana (Guadeloupe) ao Agrapha oxygramma ai Aniclainfecta ea Eulepidotis addens md Mocis disseverans ml M latipes oc Orthodes crenulata pl Perigonia lusca pu Pseudaletiaunipuncta sd Spodoptera dolichos sf S frugiperda sl S latifascia so S ornithogalli xc Xylophanes chiron

Table 2 Summary of forward stepwise discriminant analysis for male and female nematode morphological characters from 14 and 15host species respectively

Males ndash all groups (n D 14) Females ndash all groups (n D 15)

Step Character F (df1 df2) Character F (df1 df2)

1 Excretory pore 40313 (13 93) Excretory pore 61696 (14 214)2 Spicule length 20971 (13 92) Stylet length 16778 (14 213)3 Stylet length 5908 (13 91) V 15281 (14 212)4 Vulva or cloaca to ant end 1155 (13 90) Bulb width 1756 (14 211)5 Ant end to post bulb 905 (13 89) Excretory pore of body 1400 (14 210)6 Stylet percent of body 582 (13 88) a 1050 (14 209)7 B 702 (13 87) Ant end to post bulb 999 (14 208)8 G 640 (13 86) G 774 (14 207)9 Bulb width 447 (13 85) Vulva or cloaca to ant end 1023 (14 206)

10 Body width at nerve ring 419 (13 84) b 826 (14 205)

P lt 0001

494 Nematology

Speciation in the Acugutturidae

Table 3 Summary of forward stepwise discriminant analysis for male and female nematode morphological characters from 11 speciesof host Noctuidae (Lepidoptera)

Males ndash Lepidoptera groups1 (n D 11) Females ndash Lepidoptera groups1 (n D 11)

Step Character F (df1 df2) Character F (df1 df2)

1 Spicule length 24039 (10 80) V 10986 (10 150)2 Stylet length 2510 (10 79) Stylet length 6194 (10 149)3 Stylet of body length 1388 (10 78) Bulb width 1830 (10 148)4 Cloaca to anterior end 981 (10 77) Excretory pore of body 982 (10 147)5 Rostrum length 762 (10 76) Vulva to anterior end 935 (10 146)6 Ant end to ant gonad 700 (10 75) Body width at nerve ring 873 (10 145)7 G 795 (10 74) Stylet of body length 777 (10 144)8 c0 651 (10 73) Tail length 701 (10 143)9 Body width at nerve ring 615 (10 72) Head width 672 (10 142)

10 Bulb width 548 (10 71) Gonad of body length 595 (10 141)

1) Not including Perigonia lusca Xylophanes chiron or two populations from Periplaneta AmericanaP lt 0001

Fig 2 Plot of discriminant functions for nematode morphometric characters Ellipses indicate 95 condence limits for selected hostspecies A Female nematodes (Vampyronema) from six host species (Lepidoptera) B Male nematodes (Vampyronema) from six hostspecies (Lepidoptera)C Female nematodes (Noctuidonema) from ve host species (Lepidoptera)D Male nematodes (Noctuidonema)from ve host species (Lepidoptera) Legend (Host names) ao Agrapha oxygramma ai Anicla infecta ea Eulepidotis addens mdMocis disseverans ml M latipes oc Orthodes crenulata pu Pseudaletia unipuncta sd Spodoptera dolichos sf S frugiperda sl Slatifascia so S ornithogalli

Vol 4(4) 2002 495

OG Marti Jr et al

Table 4 Summary of forward stepwise discriminant analysis of morphological characters from male and female Vampyronema fromsix species of Noctuidae (Lepidoptera)

Males ndash Mocis group1 (n D 6) Females ndash Mocis group1 (n D 6)

Step Character F (df1 df2) Character F (df1 df2)

1 Stylet length 5807 (5 39) Stylet length 4573 (5 93)2 Rostrum length 1219 (5 38) Ant end to ant intestine 1879 (5 92)3 Bulb length 929 (5 37) Excretory pore of body 1803 (5 91)4 Ant end to ant gonad 837 (5 36) G 1500 (5 90)6 Pharyngeal gland width 511 (5 34) Stylet length of body 573 (5 89)7 Gonad length 438 (5 33) V 485 (5 88)8 Tail length 408 (5 32) Head length 476 (5 87)9 Bulb width 269 (5 31) Gonad length of body 445 (5 86)

10 Body width at nerve ring 355 (5 85)

1) Mocis group Agrapha oxygramma Anicla infecta Eulepidotis addens Mocis disseveransM latipes Orthodes crenulataP lt 005 P lt 001 P lt 0001

STEP 3 SIX HOST GROUPS (VAMPYRONEMANEMATODES WITH SIMPLE SPICULES)

The noctuidmoths Anicla infectaAgraphaoxygrammaEulepidotis addens Mocis disseverans M latipes andOrthodes crenulata are parasitised by Vampyronema inwhich the structure of the spicules is simpler and the tailof the females is longer than in Noctuidonema Analy-sis of morphometric characters of female nematodes fromthese hosts revealed ve signi cant canonical roots cu-mulatively accounting for 100 of the explainedvariancebetween groups The rst three roots representingprimar-ily stylet length distance from the anterior end to the baseof the bulb and the c ratio accounted cumulatively for9188 of the explained variance between groups Styletlength was the most important morphological characterdistinguishing female Vampyronema from different hosts(Fig 2A Table 4)

Analysis of results of male and female Vampyronemaon noctuid moths demonstrate the existence of fourdistinct clades One of these is V dibolia the speciesdescribed from M latipes The nematodes particularlythe males on M disseverans are very similar to V diboliaand we consider them to be conspeci c A second cladeis formed by nematodes from A infecta which are wellseparated from all the other groups (Fig 2A B) Styletlength is signi cantly shorter (P lt 001) in nematodeson A infecta than on those from the other hosts Athird Vampyronema clade is formed by nematodes onE addens which differ from V dibolia primarily in thedistance from the anterior end to the base of the bulb Thetwo male nematodes from E addens were also distinct

from other males an observation in agreement withresults from the females A fourth Vampyronema cladeis formed by nematodes from A oxygramma (Fig 2AB) Although the females appear to be intermediatebetween those on M latipes and E addens the malesare distinct based on differences in stylet and rostrumlength Nematodes on O crenulata are also intermediatebut males are very similar to V dibolia and we considerthem to be conspeci c

In the males there were ve signi cant canonical rootsrepresented primarily by stylet length spicule rostrumlength and bulb length respectively These accounted cu-mulatively for 929 of the explained variation betweenmales from these six host species Stylet length was themost important morphologicalcharacter in de ning males(Fig 2B) although rostrum length was also identi ed asan important character

STEP 4 FIVE HOST SPECIES (NOCTUIDONEMANEMATODES WITH COMPLEX SPICULES)

Discriminant analysis of female nematodes from fourSpodoptera species and Pseudaletia revealed four signif-icant (P lt 005) canonical roots The rst two rep-resenting primarily stylet length and body width at thevulva respectively accounted for a cumulative propor-tion of 8793 of the explained variance between groupsFemale nematodes from Pseudaletia were well separatedfrom those parasitising Spodoptera spp on the basis ofthe shorter stylet Nematodes from the four species ofSpodoptera were poorly de ned and showed a consider-able amount of overlap Even the groups that appeared to

496 Nematology

Speciation in the Acugutturidae

Table 5 Summary of forward stepwise discriminant analysis of morphological characters from male and female Noctuidonema from ve species of Spodoptera and Pseudaletia

Males ndash Spodoptera and Pseudaletia1 (n D 5) Females ndash Spodoptera and Pseudaletia1 (n D 5)

Step Character F (df1 df2) Character F (df1 df2)

1 Ant end to ant intestine 2524 (4 41) Stylet length 3914 (4 57)2 Body length 1491 (4 40) Body width at vulva 1250 (4 56)3 c0 1223 (4 39) Bulb width 591 (4 55)4 Bulb width 725 (4 38) a 498 (4 54)5 Body width at nerve ring 500 (4 37) Vulva to anterior end 380 (4 53)6 Spicule length 409 (4 36) Body width at nerve ring 579 (4 52)7 Maximum body width 265 (4 51)

1) Pseudaletia unipuncta Spodoptera dolichos S frugiperda S latifascia S ornithogalliP lt 005 P lt 001 P lt 0001

be the most different S frugiperda and S dolichos wereextensively overlapped (Fig 2D Table 5)

In the males there were three signi cant canonicalroots accounting for 9764 of the explained variancebetween groups The rst root represents the distancefrom the anterior end to the anterior intestine and V aboutequally whereas the second root represents primarily thebody width at the nerve ring (Fig 2D Table 5) Malenematodes from P unipuncta were well separated fromnematodes from Spodoptera which agrees with resultsfrom the females Males on S dolichosshowed a moderateamount of separation from males on the other threespecies of Spodoptera

DATA ANALYSIS

Eighteen morphological characters contribute to differ-entiation among male nematodes on 14 host species in-cluded in this study The most signi cant and useful ofthese are the distance of the excretory pore from the an-terior end spicule length and stylet length These threecharacters and V are the most useful in differentiating be-tween females (Table 2) Furthermore ANOVA showedthat the stylet length of nematodes on the two speciesof Sphingidae is signi cantly greater (P lt 0001) thanthe stylet length of nematodes on any of the other hostspecies The two populationsof nematodes from Sphingi-dae differ signi cantly in only eight morphological char-acters whereas each differs from V dibolia in about 20characters Females on P lusca differ from those on Xchiron in the longer length of the stylet greater length andwidth of the bulb greater distance of the excretory poreand bulb base from the anterior end greater body widthat the vulva and greater excretory pore distance as a per-

centage of body length However in the absence of malesfrom X chiron no conclusion can be reached on whethernematodes on Sphingidae are conspeci c The consistentseparation of nematodes parasitising Sphingidae from allother Acugutturidae suggests that they probably representat least one new genus Particularly remarkable is the longstylet which represents about 37 of the body length andaverages about 190 sup1m in length (Fig 3)

Discriminant analysis shows that the ve groups ofnematodes with complex spicules (Noctuidonema) arewell separated (Fig 1C D) from the six groups withsimple spicules (Vampyronema) Although the discretecharacter of spicule structure is suf cient to separate thesetwo genera particularly good separation can also be madebased on important continuously variable characters suchas stylet length distance of the excretory pore to theanterior end and the V ratio Furthermore within eachof these two genera separate groups can be detectedbased primarily on differences in stylet length Thusthe nematodes on Pseudaletia unipuncta differ fromthe nematodes on the four Spodoptera spp (Fig 1CD) and those on Anicla infecta Agrapha oxygrammaand Eulepidotis addens differ from V dibolia and itsconspeci c on M disseverans (Fig 2A B)

Female nematodes on P unipuncta differ (P lt 005)from those on S dolichos S frugiperda S latifasciaand S ornithogalli in eight 18 14 and 14 morpholo-gical characters respectivelyThese results providestrongsupport for separate taxonomic status of nematodes onP unipuncta as a new species of Noctuidonema Resultsfrom simple scatterplots prepared for each gender usingthe most importantmorphologicalcharacters identi ed bydiscriminant analysis agree with the preceding analysis

Vol 4(4) 2002 497

OG Marti Jr et al

Fig 3 Plot of stylet length of ectoparasitic nematodes from insect hosts (Blattodea and Lepidoptera) A Female nematodes from15 host species B Male nematodes from 14 host species Legend (Host names) ao Agrapha oxygramma ai Anicla infecta eaEulepidotis addens md Mocis disseverans ml M latipes oc Orthodes crenulata pa Periplaneta americana (St Lucia) pa2 Pamericana (Guadeloupe) pl Perigonia lusca pu Pseudaletia unipuncta sd Spodoptera dolichos sf S frugiperda sl S latifascia soS ornithogalli xc Xylophanes chiron

Table 6 Percent correct classication of male nematodes based on posterior probabilities that a case belongs to a particular hostspecies

Host species Percent correct classi cations ndash Percent correct classi cations ndashmale nematodes female nematodes

n Hosts Hosts Hosts Hosts n Hosts Hosts Hosts Hosts1-15 1-11 1-6 7-11 1-15 1-11 1-6 7-11

1 Agrapha oxygramma 7 100 100 100 ndash 12 917 100 917 ndash2 Anicla infecta 12 100 100 100 ndash 15 100 100 100 ndash3 Eulepidotis addens 2 100 100 100 ndash 15 100 100 100 ndash4 Mocis disseverans 4 100 75 75 ndash 9 889 889 889 ndash5 Mocis latipes 16 100 100 100 ndash 39 923 974 948 ndash6 Orthodes crenulata 4 100 100 100 ndash 9 889 889 889 ndash7 Spodoptera dolichos 10 100 100 ndash 90 5 100 100 ndash 808 Spodoptera frugiperda 11 909 909 ndash 100 17 765 824 ndash 7659 Spodoptera latifascia 6 667 833 ndash 50 12 75 833 ndash 667

10 Spodoptera ornithogalli 10 100 100 ndash 100 10 90 90 ndash 9011 Pseudaletia unipuncta 9 100 100 ndash 100 18 100 944 ndash 94412 Perigonia lusca 4 100 ndash ndash ndash 19 938 ndash ndash ndash13 Periplaneta americana (1) 7 100 ndash ndash ndash 18 882 ndash ndash ndash14 Periplaneta americana (2) 5 100 ndash ndash ndash 13 846 ndash ndash ndash15 Xylophanes chiron 0 ndash ndash ndash ndash 18 100 ndash ndash ndash

(1) St Lucia(2) Guadeloupe

498 Nematology

Speciation in the Acugutturidae

Fig 4 Tree diagram for ectoparasitic nematodes from insecthost species (Blattodea and Lepidoptera) The tree was pro-duced using Euclidean distance as the distance measure andcomplete linkage as the amalgamation rule Both graphs areplotted on an X-axis standardised by (DlinkDmax)100 whichrepresents the percentage of the range from the maximum tothe minimum distance in the data A Female nematodes datafor excretory pore stylet length V bulb width a anterior endto anterior intestine and vulva to anterior end were used BMale nematodes data for excretory pore spicule length andstylet length were used Legend (Host Names) pa Periplan-eta americana (St Lucia) pa2 P americana (Guadeloupe) aoAgrapha oxygramma ai Anicla infecta ea Eulepidotis addensmd Mocis disseverans ml M latipes oc Orthodes crenulatapl Perigonia lusca pu Pseudaletia unipuncta sd Spodopteradolichos sf S frugiperda sl S latifascia so S ornithogallixc Xylophanes chiron

Spicule length was not an important character in dis-tinguishing among male Noctuidonema Male nematodeson S dolichos and S frugiperda showed the most sepa-ration with moderate overlap whereas males on S or-nithogalli appeared to be most similar to those on S

dolichos and males on S latifascia appeared most similarto those on S frugiperda Results from both genders agreethat nematodes on Pseudaletia are different from thoseon Spodoptera and that the nematodes on S dolichos andS frugiperda are the most different with the nematodeson S latifascia and S ornithogallibeing intermediate be-tween these two Overall the results do not support theexistence of separate species of Noctuidonema on hostsof the genus Spodoptera (Fig 4 Table 6)

These results provide evidence that nematodes on Ainfecta A oxygramma and E addens are different fromeach other and from the remaining groups having simplespicules Nematodes on M disseverans and O crenulatawere not distinct from V dibolia the species parasitisingM latipes Nematodes on M disseverans differed signif-icantly from V dibolia in only six morphological charac-ters in the females and ten in the males The differenceswere small and can be attributed to adaptation of conspe-ci c nematodes to parasitism of closely related sympatrichost species

Comparisonof the description of V daptria to measure-ments of other Vampyronema in this study indicates that itdiffers in stylet length and spicule shape V daptria hasthe shortest stylet length observed in the genus rangingfrom 53-60 sup1m in males and 59-69 sup1m in females whichis shorter than the stylet length of nematodes on A in-fecta (Fig 3) In addition spicules of V daptria taper to ne points whereas in Vampyronema from other hosts thespicule tips are blunt V daptria on L porcia appears tobe a distinct species and we therefore do not assign any ofthe Vampyronema from this study to it

Nematodes on the St Lucia and Guadeloupe popula-tions of Periplaneta americana are not well differenti-ated from each other based on the suite of characters thatwere examined (Fig 1A B) Although there are a fewsigni cantly different characters between the two popula-tions the differences probably do not represent interspe-ci c variation In addition the characters of both popula-tions are in agreement with the published description ofAcugutturus parasiticus (Hunt 1980)

Examination of additional specimens of ectoparasiticnematodes from other species of Lepidoptera showed thatmost were Vampyronema whereas Noctuidonema wasfound only on moths from two genera Pseudaletia andSpodoptera (Table 7)

Vol 4(4) 2002 499

OG Marti Jr et al

Table 7 Association of quadrine and trine Noctuidae with Vampyronema and Noctuidonema

Host Taxon Hind Wing Venation1 Locality2 Nematodes Genus

NoctuidaeCatocalinae

Anomis oedema Gueneacutee quadri ne FG 4 1 VampyronemaBaniana sp quadri ne FG 1 0 VampyronemaBaniana ostia Druce quadri ne FG 7 0 VampyronemaEuglyphis sp quadri ne FG 1 0 VampyronemaEulepidotis addens (Walker) quadri ne GU 5 2 VampyronemaItomia opisthographa (Gueneacutee) quadri ne FG 5 1 VampyronemaLesmone formularis (Geyer) quadri ne FG 1 0 VampyronemaLesmone porcia (Stoll) quadri ne GU ndash3 Vampyronema3

Melipotis fasciolaris (Huumlbner) quadri ne FG 0 1 VampyronemaMetalectra sp quadri ne FG 4 10 VampyronemaMetria sp quadri ne FG 0 1 VampyronemaMocis dif uens (Gueneacutee) quadri ne FG 1 1 Vampyronema4

Mocis disseverans (Walker) quadri ne GA 9 4 Vampyronema4

Mocis latipes Gueneacutee quadri ne GA 9 16 Vampyronema4

Perasia ora (Cramer)5 quadri ne FG 10 0 VampyronemaSelenisa sp quadri ne FG 4 0 VampyronemaZale ctilis (Gueneacutee) quadri ne FG 4 1 Vampyronema

PlusiinaeAgrapha oxygramma (Geyer) quadri ne GA 13 7 VampyronemaArgyrogramma verruca (Fabricius) quadri ne FG 0 2 VampyronemaChrysodeixis chalcites (Esper) quadri ne Fiji 8 7 Vampyronema

AmphipyrinaeSpodoptera androgea (Stoll) tri ne FG 4 0 Noctuidonema6

Spodoptera dolichos (Fabricius) tri ne GA 5 10 Noctuidonema6

Spodoptera frugiperda (JE Smith) tri ne GA 17 11 Noctuidonema6

Spodoptera latifascia (Walker) tri ne GA 12 6 Noctuidonema6

Spodoptera litura (Fabricius) tri ne Fiji 15 7 Noctuidonema6

Spodoptera ornithogalli (Gueneacutee) tri ne GA 10 10 Noctuidonema6

HadeninaeOrthodes crenulata (Butler) tri ne GA 9 4 VampyronemaPseudaletia unipuncta (Haworth) tri ne GA 18 9 Noctuidonema

NoctuinaeAnicla infecta (Ochsenheimer) tri ne GA 15 12 VampyronemaTandilia rodea (Schaus) tri ne FG 2 3 Vampyronema

PyralidaePyraustinae

Lamprosema moccalis Schaus ndash FG 4 0 VampyronemaSphingidae

MacroglossinaeErinnys obscura (Fabricius) ndash FG 22 2 Vampyronema7

Perigonia lusca Fabricius ndash GU 19 4 Vampyronema7

Xylophanes chiron Drury ndash GU 8 0 Vampyronema7

1) Quadri ne and tri ne hindwing venation applies to Noctuidae only Pyralidae and Sphingidae are included for completeness thestatus of Plusiinae as quadri nes is ambiguous because it groups together mostly species having quadri ne wing venation yet is oftenincluded with chie y tri ne subfamilies (Kitching amp Rawlins 1998) 2 Fiji Fiji Islands FG French Guiana GU Guadeloupe GAGeorgia USA 3 V daptria Anderson amp Laumond 1992 designationbased on Hunt (1993) 4 V dibolia 5 Synonymous with Nymbisarcuata Walker (in Rogers et al 1990a and Simmons amp Rogers 1996) 6 N guyanense 7 Probably represents an undescribed genus

500 Nematology

Speciation in the Acugutturidae

Discussion

The position of the excretory pore posterior to thebulb the comparatively short length of the stylet thecomparatively long tail and the simple rose-thorn-shapedspicules are suf cient to clearly separate Acugutturus aparasite of Blattodea from Vampyronema and Noctuido-nema parasites of Lepidoptera Sampled populations re-veal four inferred monophyleticgroups Acugutturus par-asiticus from two West Indian populations of Periplan-eta americana a new genus on Sphingidae Noctuido-nema guyanense from Spodoptera spp with nematodeson P unipuncta as a probable new species of Noctuido-nema and Vampyronema spp on the remaining hosts inthis study with the nematodes on A infecta representinga new species of Vampyronema

Each inferred monophyletic group can be de ned byseveral unique combinations of characters quantitativeand qualitative that satisfy the minimal requirements forspecies diagnosis (Wheeler 1999 Wheeler amp Platnick2000) Such characters appear to represent lineage inde-pendence via autapomorphies hypotheses ideally testedvia outgroup comparison Such tests while more philo-sophically satisfying (Adams 1998 Wiley amp Mayden2000) are beyond the scope of the present paper

According to Kitching and Rawlins (1998) differencesin hindwing venation broadly divide the Noctuidae intotwo phenetic groups tri nes and quadri nes Catocalinaeand Plusiinae are quadri nes (but see comment 1 in Ta-ble 7) whereas the other noctuid subfamilies mentionedin this study are tri nes Complex spicules probably rep-resent a derived condition found thus far only in nema-todes on two genera of tri ne noctuids Pseudaletia andSpodoptera while the simple spicules of Vampyronemaprobably represent a plesiomorphic state found in bothtri ne and quadri ne noctuids The extremely long styletsof nematodes from Sphingidae may have arisen in re-sponse to parasitisation of large moths

The Acugutturidae are associated with two very dis-tantly related insect orders Blattodea and Lepidoptera(Endopterygota) There is no record of occurrence ofthe Acugutturidae on any other insect order Among theLepidoptera the Acugutturidae and more precisely theNoctuidonematinae are known from only six familiesGeometridae Lasiocampidae Noctuidae NotodontidaePyralidae and Sphingidae (Marti et al 2000) most ofwhich are not closely related with the exceptionof Lasio-campidae and Sphingidae members of the Bombycoidea(and their relatives) clade (Lemaire amp Minet 1998) and

Notodontidaeand Noctuidae members of the Noctuoideasuperfamily (Kitching amp Rawlins 1998)

One explanation for such a surprising distribution ofthe Acugutturidae on their hosts may be that availabledata are too sparse due to insuf cient sampling of otherlepidopteran families While this may be true the studyconducted by Marti et al (2000) shows that there isa clear speci city in nematode host preference Martiet al (2000) updated the list of lepidopteran hosts ofNoctuidonematinae It seems clear that although manyLepidoptera species belonging to 11 families and 115genera have been examined for nematodes only membersof the six aforementioned obtectomeran (Kristensen ampSkalski 1998) families have been found to be positivefor Noctuidonematinae With the exception of Pyralidaeall these families are members of the Macrolepidopteraclade Moreover the positive Noctuidae species representnearly 90 of the total positive species found Thereforewe cannot state that the Noctuidonematinae are notpotentially parasites of most of the lepidopteran familiesbut we can state that at least in the Neotropical Regionthey parasitise the Macrolepidoptera preferentially andamong that clade mostly a single family the Noctuidae(Marti et al 2000)

The Aphelenchida include several different trophictypes with many species being mycetophagous or phy-toparasitic with about one third of known species be-ing entomoparasitic (Hunt 1993) Considering the list ofinsect hosts we cannot hypothesise that the Acugutturi-dae are derived from a single ancestor transferring froma plant to an insect host It seems more realistic to pro-pose several transfers from plant to insect hosts during thehistory of the Acugutturidae A transfer (Acugutturinae)occurred on the Blattodea and another one (Noctuidone-matinae) on the Lepidoptera Another transfer may haveoccurred on the Bombycoidea sensu lato If at least twosuch principal transfers occurred then it is possible thatthe Acugutturinae and Noctuidonematinae would not beconfamilial and that these two taxonomicunits deserve fa-milial rank Cladistic and molecularphylogeneticanalysiswill help to solve that problem The existenceof symbioticbacteria in these nematodes (Marti et al 1995) shouldprovide an independentmeans of analysing these relation-ships

Concerning the relationship between Lepidoptera andits ectoparasitic nematodes the present work highlightsthe following situation a new genus of Noctuidone-matinae parasitises Sphingidae a genuine Bombycoidea(Lemaire amp Minet 1998) Vampyronema parasitises the

Vol 4(4) 2002 501

OG Marti Jr et al

quadri ne noctuid subfamily Catocalinae (sensu lato in-cluding Ophiderinae) Plusiinae an Hadeninae and aNoctuinae (both typical tri ne noctuids) and Noctui-donema spp parasitises another Hadeninae and severalSpodoptera a tri ne noctuid genus previously includedin the Amphipyrinae but now of uncertain subfamilial af- liation (Poole 1995 Kitching amp Rawlins 1998) Exceptfor nematodes on Sphingidae which likely correspond toa well-separated clade whose phylogenetic relationshipto other Noctuidonematinae needs to be speci ed it ap-pears that Vampyronema and Noctuidonema both para-sitise Noctuidae but that Vampyronema parasitises bothquadri ne and tri ne noctuids Noctuidonema possibly aderived form of Noctuidonematinae is restricted to a fewgenera (Spodoptera and Pseudaletia) of tri ne noctuidsthat are not closely related (Mitchell et al 2000)

Recent work has questioned the monophylyof the Noc-tuidae and there is a growing feeling among moleculartaxonomists that the Noctuidae may be paraphyletic sev-eral quadri ne subfamilies including Catocalinae beingmore closely related to Arctiidae and Lymantriidae than totri ne noctuids (Weller et al 1994 Mitchell et al 1997Fang et al 2000 Mitchell et al 2000) The latter groupis considered as monophyleticClearly systematics of theNoctuidonematinaedo not match the tentative new classi- cation of the Noctuidae

This lack of congruence between parasite and hostsystematics leads us to consider that the relationshipsbetween Acugutturidae and their hosts have not beenguided by coevolutionary forces in the sense of Ehrlichand Raven (1964) but by horizontal transfers (Rogers ampMarti 1996) and sequential evolution (Jermy 1984 Fu-tuyma amp McCafferty 1990) The Acugutturidae and es-pecially the Noctuidonematinae have likely experiencedan adaptive radiation on higher Lepidoptera long aftertheir host diversi cation History of the Noctuidonemati-nae has been characterised by host transfer events ontofairly distantly related hosts followed by adaptation tonew hosts leading to speciation processes that now permitus to distinguish easily fairly species-speci c populationsor strains of Vampyronema or Noctuidonema In mothsthese host transfers may have occurred through contami-nation of nematode-free species by nematode eggs juve-niles or adults left on resting or feeding surfaces (leavesor owers) by individualsof infested species The successof such contamination may be linked to the availabilityand accessibility of host intersegmental membranes thinenough to be pierced by nematode stylets a factor that

may have physically and dramatically restricted the hostrange of all Acugutturidae

The present morphological analysis of acugutturid ne-matodes parasitising 13 species of Lepidoptera and onespecies of cockroach revealed a greater degree of diversitythan previously known in this group The genus Vampy-ronema grouping nematodes with simple spicules in-cludes two inferred new species in addition to V dibo-lia and V daptria already known from the LepidopteraThe genus Noctuidonema grouping nematodes of Lepi-doptera with complex spicules includes a new speciesand the nematodes found on Sphingidae probably belongto a new genus The relationshipbetween these nematodesand their hosts appears to be much more complex thanpreviously thought In addition to their bene ts as para-sites of pest Lepidoptera acugutturid nematodes are aninteresting model for parasite-host coevolutionarystudies

Acknowledgements

We thank David Hunt (CABI Bioscience UK) DanielLachaise (CNRS France) Alvin Simmons (USDA Char-leston SC USA) and S Patricia Stock (University ofCalifornia Davis CA USA) for reviewing earlier ver-sions of this paper We thank Jeanette Williams and JohnClayton for collecting cockroaches in St Lucia and mothsin Fiji respectively We thank the following individualsfor identifyingor con rming our identi cations of insectsDavid Nickle (Smithsonian Washington DC USA) forPeriplaneta americana Pierre Zagatti (INRA France)for Pyralidae and Hugo Kons (University of FloridaGainesville FL USA) for Orthodes crenulata

References

ADAMS BJ (1998) Species concepts and evolutionary para-digm in modern nematology Journal of Nematology 30 1-21

ANDERSON RV amp LAUMOND C (1990) NoctuidonemaguyanenseRemillet amp Silvain 1988 morphologie du spiculeredescription du macircle et diagnose amendeacutee du genre Noctui-donema Revue de Neacutematologie 13 433-436

ANDERSON RV amp LAUMOND C (1992) Noctuidonemadaptria n sp (Nematoda Aphelenchoididae)an ectoparasiteof the moth Lesmone porcia (Stoll) Journal of Nematology24 16-22

EHRLICH P amp RAVEN P (1964) Butter ies and plants astudy in evolution Evolution 18 586-608

FANG QQ MITCHELL A REGIER JC MITTER CFRIEDLANDER TP amp POOLE RW (2000) Phylogenetic

502 Nematology

OG Marti Jr et al

of body length stylet length as of body length andexcretory pore to anterior end as of body length

Male and female nematode untransformed data wereanalysed and plotted separately because different suitesof characters are involved for each gender Multivariateanalysis (MANOVA) was used to determine whether therewere differences between nematodes from different hostgroups based on the entire suite of characters for eachnematode gender Each character was then analysed withunivariateanalysis (ANOVA) to determine if the charactercontributedsigni cantly to differentiatingbetween at leasttwo host groups for each nematode gender The Bonfer-roni test was used to check against overestimation of sig-ni cance of univariate characters These procedures weredescribed by Footit and Sorensen (1992) and Sanmartiacutenand Martiacuten-Piera (1999)

Forward stepwise discriminant analysis was used toidentify which morphological characters made signi cantcontributions to de ning differences among nematodesfrom different host species Non-signi cant and redundantcharacters were identi ed and deleted from the modelsFour separate analyses were conducted for male and fe-male nematodes First nematodes from all host specieslisted in Table 1 were included (n D 15 female and n D14 male groups) Based on these results nematodes onBlattodea and Sphingidae were deleted from the secondstep which included only nematodes from noctuid hostsAgain morphological characters were ranked and non-signi cant (P gt 005) and redundant characters wereidenti ed and deleted from the model In the third steponly nematodes from groups having simple spicules simi-lar to those of V dibolia were included (n D 6 female andn D 6 male groups) The same procedure was followedin the last step which analysed only nematodes havingcomplex spicules similar to those of N guyanense (n D 5female and n D 5 male groups) This procedure producedeight models four for each nematode gender for discrim-inating between nematodes from all 15 host groups be-tween nematodes from noctuid hosts between nematodeshaving simple spicules (Vampyronema) and between ne-matodes having complex spicules (Noctuidonema)

Discriminant functions were used to calculate poste-rior probabilities that individual nematodes were classi- ed correctly into their respective groups Separate clas-si cations were obtained for male and female nematodesA high percentage of correct classi cation is a measure ofthe discriminatory power of the morphological characters(Footit amp Sorensen 1992) Discriminant functions wereevaluated for signi cance with chi-square and only signif-

icant (P lt 005) functions were used in data interpreta-tions Standard discriminant function coef cients the cor-relations of the morphological characters to the discrimi-nant functions were used to determine which morpholo-gical characters contributedmost to the differentiationbe-tween groups A large absolute value of the standard dis-criminant function coef cient indicates that a character isstrongly represented by the discriminant function

Two-dimensional graphs were produced that providedthe best differentiation between nematodes from differenthost species for each nematode gender Clearly de nedgroups were marked by 95 ellipses which delimit thearea in which a new observation is likely to occur Resultsfrom both nematode genders from each host species wereevaluated to arrive at a consensus regarding the status ofthe nematodes from each host group

The most important morphological characters identi- ed by discriminant analysis were used in cluster analysisto produce tree diagrams for each nematode gender Datafor each morphological character were standardised withmean D 0 and standard deviationD 1 Clusters were basedon Euclidean distance for the distance measure betweengroups and on complete linkage for the amalgamationrule Tree diagrams were scaled to (DlinkDmax)100andwere then evaluated and compared to results from dis-criminant analysis to arrive at a consensus regarding sim-ilarities between nematodes from different host species

EXAMINATION OF NEMATODES FROM ADDITIONAL

HOST SPECIES

We also examined additional nematodes from otherlepidopteran hosts from French Guiana and the FijiIslands These nematodes were not measured in detailbecause they were available only in small numberswere of uneven quality or both In all of the specimensfrom Noctuidae however we were able to make adetermination regarding their generic placement We wereable therefore to test our conclusions regarding thevalue of speci c characters in de ning the associationof nematode clades with clades and sub-families of theNoctuidae

Results

MANOVA showed differences regardless of nematodegender among nematodes from the 15 host speciesbased on all morphological characters All F values werehighly signi cant (P lt 0001 df D 13 91 for males

492 Nematology

Speciation in the Acugutturidae

and 14 113 for females) In addition ANOVA showedthat each morphological character contributed to thedifferentiation of nematodes of each gender from at leasttwo host species All F values were highly signi cant(P lt 0001 df D 11-13 83-93 for males and 14209-214 for females) Therefore all 18 characters sixratios and three percentages listed above were used insubsequentdiscriminant analysis to determine which weremost useful in distinguishing between nematodes fromdifferent hosts

STEP 1 15 HOST SPECIES (BLATTODEA AND

LEPIDOPTERA)

Discriminant analysis of female nematodes from 15host species revealed three separate groups i) fromthe two Periplaneta americana populations (Acugutturusparasiticus) ii) from the two species of Sphingidaeiii) all other host species (Fig 1A B) There were tensigni cant canonical roots accounting cumulatively for9978 of the explained variance between groups The rst two roots accounted for 9078 of the explainedvariance Examination of the factor structure matrixthe correlation of the morphological characters to thecanonical roots showed that the distance of the excretorypore from the anterior end the stylet length and V werethe characters most important in distinguishing betweenfemale nematodes from the 15 host species (Table 2)The other characters were of less importance althoughallwere signi cant for female nematodes

Discriminant analysis of male nematodes from 14 hostspecies (no male nematodes were found on Xylophaneschiron) revealed four distinct groups (Fig 1B Table 2)Canonical roots 1 (excretory pore) and 3 (stylet length)gave the best differentiation among males In additionto the differentiation of A parasiticus and the malenematodes on Perigonia lusca from each other and fromthe remaining host groups it was also found that malenematodes from Pseudaletia unipuncta were differentfrom those collected from the four species of SpodopteraThe two populations of A parasiticus did not differ butwere distinct from the other groups primarily due to thedifference in the position of the excretory pore

In the males there were eight signi cant (P lt 005)canonical roots accounting cumulatively for 9944 ofthe explained variance among males The rst threecanonical roots accounted for 9446 of the explainedvariance Canonical roots 1 and 3 are represented primar-ily by the excretory pore and the stylet length respec-

tively whereas root 2 (not shown) is represented primarilyby spicule length

STEP 2 11 HOST GROUPS (NEMATODES OF

LEPIDOPTERA NOCTUIDAE)

Analysis of morphological characters of nematodesparasitising only noctuids provides additional informa-tion differentiating several of these groups Among thefemales there were ten signi cant canonical roots ac-counting cumulatively for 100 of the explained vari-ance The rst three roots accounted cumulatively for806 of the explained variance with V stylet length andbulb width being the most important (Table 3) V was byfar the most important female character differentiatingne-matodes among the noctuid hosts and was more usefulthan tail length alone V provides clear separation of fe-male nematodes of Pseudaletia and Spodoptera from theother noctuid groups (Fig 1C) The former group is char-acterised by V gt 92 (ie short female tails) and by com-plex spicules and the latter group by V lt 92 (ie longerfemale tails) and simple spicules

Among male nematodes from noctuid hosts (Fig 1DTable 3) discriminant analysis provided clear differenti-ation in root 1 between males having complex spicules(nematodes from Pseudaletia and Spodoptera) vs thosehaving simple spicules (nematodes from Agrapha AniclaEulepidotis Mocis spp and Orthodes) There were ninesigni cant canonical roots which cumulativelyaccountedfor 9983 of the explained variance among groups The rst three roots accountingcumulatively for 920 of theexplained variance among groups were represented pri-marily by spicule length stylet length and stylet length asa percentage of body length respectively

Nematodes from Pseudaletia can be separated fromthose on Spodoptera by the shorter stylet length (Fig 1CD) Nematodes on Agrapha oxygramma can be separatedfrom M latipes primarily on the basis of the shorter styletlength (P lt 005) Similarly males on Anicla infectahave shorter spicules (P lt 005) than the nematodes onthe remaining host groups

Only two male nematodes from E addens were avail-able However these were distinct from nematodes on Mlatipes The female nematodes on E addens were alsowell differentiated from those on M latipes despite someoverlap (Fig 2A)

Vol 4(4) 2002 493

OG Marti Jr et al

Fig 1 Plot of discriminant functions for nematode morphometric characters Ellipses indicate 95 condence limits for selected hostspecies A Female nematodes from 15 host species (Blattodea and Lepidoptera) B Male nematodes from 14 host species (Blattodeaand Lepidoptera) C Female nematodes from 11 host species (Lepidoptera) D Male nematodes from 11 host species (Lepidoptera)Legend (Host Names) pa Periplaneta americana (St Lucia) pa2 P americana (Guadeloupe) ao Agrapha oxygramma ai Aniclainfecta ea Eulepidotis addens md Mocis disseverans ml M latipes oc Orthodes crenulata pl Perigonia lusca pu Pseudaletiaunipuncta sd Spodoptera dolichos sf S frugiperda sl S latifascia so S ornithogalli xc Xylophanes chiron

Table 2 Summary of forward stepwise discriminant analysis for male and female nematode morphological characters from 14 and 15host species respectively

Males ndash all groups (n D 14) Females ndash all groups (n D 15)

Step Character F (df1 df2) Character F (df1 df2)

1 Excretory pore 40313 (13 93) Excretory pore 61696 (14 214)2 Spicule length 20971 (13 92) Stylet length 16778 (14 213)3 Stylet length 5908 (13 91) V 15281 (14 212)4 Vulva or cloaca to ant end 1155 (13 90) Bulb width 1756 (14 211)5 Ant end to post bulb 905 (13 89) Excretory pore of body 1400 (14 210)6 Stylet percent of body 582 (13 88) a 1050 (14 209)7 B 702 (13 87) Ant end to post bulb 999 (14 208)8 G 640 (13 86) G 774 (14 207)9 Bulb width 447 (13 85) Vulva or cloaca to ant end 1023 (14 206)

10 Body width at nerve ring 419 (13 84) b 826 (14 205)

P lt 0001

494 Nematology

Speciation in the Acugutturidae

Table 3 Summary of forward stepwise discriminant analysis for male and female nematode morphological characters from 11 speciesof host Noctuidae (Lepidoptera)

Males ndash Lepidoptera groups1 (n D 11) Females ndash Lepidoptera groups1 (n D 11)

Step Character F (df1 df2) Character F (df1 df2)

1 Spicule length 24039 (10 80) V 10986 (10 150)2 Stylet length 2510 (10 79) Stylet length 6194 (10 149)3 Stylet of body length 1388 (10 78) Bulb width 1830 (10 148)4 Cloaca to anterior end 981 (10 77) Excretory pore of body 982 (10 147)5 Rostrum length 762 (10 76) Vulva to anterior end 935 (10 146)6 Ant end to ant gonad 700 (10 75) Body width at nerve ring 873 (10 145)7 G 795 (10 74) Stylet of body length 777 (10 144)8 c0 651 (10 73) Tail length 701 (10 143)9 Body width at nerve ring 615 (10 72) Head width 672 (10 142)

10 Bulb width 548 (10 71) Gonad of body length 595 (10 141)

1) Not including Perigonia lusca Xylophanes chiron or two populations from Periplaneta AmericanaP lt 0001

Fig 2 Plot of discriminant functions for nematode morphometric characters Ellipses indicate 95 condence limits for selected hostspecies A Female nematodes (Vampyronema) from six host species (Lepidoptera) B Male nematodes (Vampyronema) from six hostspecies (Lepidoptera)C Female nematodes (Noctuidonema) from ve host species (Lepidoptera)D Male nematodes (Noctuidonema)from ve host species (Lepidoptera) Legend (Host names) ao Agrapha oxygramma ai Anicla infecta ea Eulepidotis addens mdMocis disseverans ml M latipes oc Orthodes crenulata pu Pseudaletia unipuncta sd Spodoptera dolichos sf S frugiperda sl Slatifascia so S ornithogalli

Vol 4(4) 2002 495

OG Marti Jr et al

Table 4 Summary of forward stepwise discriminant analysis of morphological characters from male and female Vampyronema fromsix species of Noctuidae (Lepidoptera)

Males ndash Mocis group1 (n D 6) Females ndash Mocis group1 (n D 6)

Step Character F (df1 df2) Character F (df1 df2)

1 Stylet length 5807 (5 39) Stylet length 4573 (5 93)2 Rostrum length 1219 (5 38) Ant end to ant intestine 1879 (5 92)3 Bulb length 929 (5 37) Excretory pore of body 1803 (5 91)4 Ant end to ant gonad 837 (5 36) G 1500 (5 90)6 Pharyngeal gland width 511 (5 34) Stylet length of body 573 (5 89)7 Gonad length 438 (5 33) V 485 (5 88)8 Tail length 408 (5 32) Head length 476 (5 87)9 Bulb width 269 (5 31) Gonad length of body 445 (5 86)

10 Body width at nerve ring 355 (5 85)

1) Mocis group Agrapha oxygramma Anicla infecta Eulepidotis addens Mocis disseveransM latipes Orthodes crenulataP lt 005 P lt 001 P lt 0001

STEP 3 SIX HOST GROUPS (VAMPYRONEMANEMATODES WITH SIMPLE SPICULES)

The noctuidmoths Anicla infectaAgraphaoxygrammaEulepidotis addens Mocis disseverans M latipes andOrthodes crenulata are parasitised by Vampyronema inwhich the structure of the spicules is simpler and the tailof the females is longer than in Noctuidonema Analy-sis of morphometric characters of female nematodes fromthese hosts revealed ve signi cant canonical roots cu-mulatively accounting for 100 of the explainedvariancebetween groups The rst three roots representingprimar-ily stylet length distance from the anterior end to the baseof the bulb and the c ratio accounted cumulatively for9188 of the explained variance between groups Styletlength was the most important morphological characterdistinguishing female Vampyronema from different hosts(Fig 2A Table 4)

Analysis of results of male and female Vampyronemaon noctuid moths demonstrate the existence of fourdistinct clades One of these is V dibolia the speciesdescribed from M latipes The nematodes particularlythe males on M disseverans are very similar to V diboliaand we consider them to be conspeci c A second cladeis formed by nematodes from A infecta which are wellseparated from all the other groups (Fig 2A B) Styletlength is signi cantly shorter (P lt 001) in nematodeson A infecta than on those from the other hosts Athird Vampyronema clade is formed by nematodes onE addens which differ from V dibolia primarily in thedistance from the anterior end to the base of the bulb Thetwo male nematodes from E addens were also distinct

from other males an observation in agreement withresults from the females A fourth Vampyronema cladeis formed by nematodes from A oxygramma (Fig 2AB) Although the females appear to be intermediatebetween those on M latipes and E addens the malesare distinct based on differences in stylet and rostrumlength Nematodes on O crenulata are also intermediatebut males are very similar to V dibolia and we considerthem to be conspeci c

In the males there were ve signi cant canonical rootsrepresented primarily by stylet length spicule rostrumlength and bulb length respectively These accounted cu-mulatively for 929 of the explained variation betweenmales from these six host species Stylet length was themost important morphologicalcharacter in de ning males(Fig 2B) although rostrum length was also identi ed asan important character

STEP 4 FIVE HOST SPECIES (NOCTUIDONEMANEMATODES WITH COMPLEX SPICULES)

Discriminant analysis of female nematodes from fourSpodoptera species and Pseudaletia revealed four signif-icant (P lt 005) canonical roots The rst two rep-resenting primarily stylet length and body width at thevulva respectively accounted for a cumulative propor-tion of 8793 of the explained variance between groupsFemale nematodes from Pseudaletia were well separatedfrom those parasitising Spodoptera spp on the basis ofthe shorter stylet Nematodes from the four species ofSpodoptera were poorly de ned and showed a consider-able amount of overlap Even the groups that appeared to

496 Nematology

Speciation in the Acugutturidae

Table 5 Summary of forward stepwise discriminant analysis of morphological characters from male and female Noctuidonema from ve species of Spodoptera and Pseudaletia

Males ndash Spodoptera and Pseudaletia1 (n D 5) Females ndash Spodoptera and Pseudaletia1 (n D 5)

Step Character F (df1 df2) Character F (df1 df2)

1 Ant end to ant intestine 2524 (4 41) Stylet length 3914 (4 57)2 Body length 1491 (4 40) Body width at vulva 1250 (4 56)3 c0 1223 (4 39) Bulb width 591 (4 55)4 Bulb width 725 (4 38) a 498 (4 54)5 Body width at nerve ring 500 (4 37) Vulva to anterior end 380 (4 53)6 Spicule length 409 (4 36) Body width at nerve ring 579 (4 52)7 Maximum body width 265 (4 51)

1) Pseudaletia unipuncta Spodoptera dolichos S frugiperda S latifascia S ornithogalliP lt 005 P lt 001 P lt 0001

be the most different S frugiperda and S dolichos wereextensively overlapped (Fig 2D Table 5)

In the males there were three signi cant canonicalroots accounting for 9764 of the explained variancebetween groups The rst root represents the distancefrom the anterior end to the anterior intestine and V aboutequally whereas the second root represents primarily thebody width at the nerve ring (Fig 2D Table 5) Malenematodes from P unipuncta were well separated fromnematodes from Spodoptera which agrees with resultsfrom the females Males on S dolichosshowed a moderateamount of separation from males on the other threespecies of Spodoptera

DATA ANALYSIS

Eighteen morphological characters contribute to differ-entiation among male nematodes on 14 host species in-cluded in this study The most signi cant and useful ofthese are the distance of the excretory pore from the an-terior end spicule length and stylet length These threecharacters and V are the most useful in differentiating be-tween females (Table 2) Furthermore ANOVA showedthat the stylet length of nematodes on the two speciesof Sphingidae is signi cantly greater (P lt 0001) thanthe stylet length of nematodes on any of the other hostspecies The two populationsof nematodes from Sphingi-dae differ signi cantly in only eight morphological char-acters whereas each differs from V dibolia in about 20characters Females on P lusca differ from those on Xchiron in the longer length of the stylet greater length andwidth of the bulb greater distance of the excretory poreand bulb base from the anterior end greater body widthat the vulva and greater excretory pore distance as a per-

centage of body length However in the absence of malesfrom X chiron no conclusion can be reached on whethernematodes on Sphingidae are conspeci c The consistentseparation of nematodes parasitising Sphingidae from allother Acugutturidae suggests that they probably representat least one new genus Particularly remarkable is the longstylet which represents about 37 of the body length andaverages about 190 sup1m in length (Fig 3)

Discriminant analysis shows that the ve groups ofnematodes with complex spicules (Noctuidonema) arewell separated (Fig 1C D) from the six groups withsimple spicules (Vampyronema) Although the discretecharacter of spicule structure is suf cient to separate thesetwo genera particularly good separation can also be madebased on important continuously variable characters suchas stylet length distance of the excretory pore to theanterior end and the V ratio Furthermore within eachof these two genera separate groups can be detectedbased primarily on differences in stylet length Thusthe nematodes on Pseudaletia unipuncta differ fromthe nematodes on the four Spodoptera spp (Fig 1CD) and those on Anicla infecta Agrapha oxygrammaand Eulepidotis addens differ from V dibolia and itsconspeci c on M disseverans (Fig 2A B)

Female nematodes on P unipuncta differ (P lt 005)from those on S dolichos S frugiperda S latifasciaand S ornithogalli in eight 18 14 and 14 morpholo-gical characters respectivelyThese results providestrongsupport for separate taxonomic status of nematodes onP unipuncta as a new species of Noctuidonema Resultsfrom simple scatterplots prepared for each gender usingthe most importantmorphologicalcharacters identi ed bydiscriminant analysis agree with the preceding analysis

Vol 4(4) 2002 497

OG Marti Jr et al

Fig 3 Plot of stylet length of ectoparasitic nematodes from insect hosts (Blattodea and Lepidoptera) A Female nematodes from15 host species B Male nematodes from 14 host species Legend (Host names) ao Agrapha oxygramma ai Anicla infecta eaEulepidotis addens md Mocis disseverans ml M latipes oc Orthodes crenulata pa Periplaneta americana (St Lucia) pa2 Pamericana (Guadeloupe) pl Perigonia lusca pu Pseudaletia unipuncta sd Spodoptera dolichos sf S frugiperda sl S latifascia soS ornithogalli xc Xylophanes chiron

Table 6 Percent correct classication of male nematodes based on posterior probabilities that a case belongs to a particular hostspecies

Host species Percent correct classi cations ndash Percent correct classi cations ndashmale nematodes female nematodes

n Hosts Hosts Hosts Hosts n Hosts Hosts Hosts Hosts1-15 1-11 1-6 7-11 1-15 1-11 1-6 7-11

1 Agrapha oxygramma 7 100 100 100 ndash 12 917 100 917 ndash2 Anicla infecta 12 100 100 100 ndash 15 100 100 100 ndash3 Eulepidotis addens 2 100 100 100 ndash 15 100 100 100 ndash4 Mocis disseverans 4 100 75 75 ndash 9 889 889 889 ndash5 Mocis latipes 16 100 100 100 ndash 39 923 974 948 ndash6 Orthodes crenulata 4 100 100 100 ndash 9 889 889 889 ndash7 Spodoptera dolichos 10 100 100 ndash 90 5 100 100 ndash 808 Spodoptera frugiperda 11 909 909 ndash 100 17 765 824 ndash 7659 Spodoptera latifascia 6 667 833 ndash 50 12 75 833 ndash 667

10 Spodoptera ornithogalli 10 100 100 ndash 100 10 90 90 ndash 9011 Pseudaletia unipuncta 9 100 100 ndash 100 18 100 944 ndash 94412 Perigonia lusca 4 100 ndash ndash ndash 19 938 ndash ndash ndash13 Periplaneta americana (1) 7 100 ndash ndash ndash 18 882 ndash ndash ndash14 Periplaneta americana (2) 5 100 ndash ndash ndash 13 846 ndash ndash ndash15 Xylophanes chiron 0 ndash ndash ndash ndash 18 100 ndash ndash ndash

(1) St Lucia(2) Guadeloupe

498 Nematology

Speciation in the Acugutturidae

Fig 4 Tree diagram for ectoparasitic nematodes from insecthost species (Blattodea and Lepidoptera) The tree was pro-duced using Euclidean distance as the distance measure andcomplete linkage as the amalgamation rule Both graphs areplotted on an X-axis standardised by (DlinkDmax)100 whichrepresents the percentage of the range from the maximum tothe minimum distance in the data A Female nematodes datafor excretory pore stylet length V bulb width a anterior endto anterior intestine and vulva to anterior end were used BMale nematodes data for excretory pore spicule length andstylet length were used Legend (Host Names) pa Periplan-eta americana (St Lucia) pa2 P americana (Guadeloupe) aoAgrapha oxygramma ai Anicla infecta ea Eulepidotis addensmd Mocis disseverans ml M latipes oc Orthodes crenulatapl Perigonia lusca pu Pseudaletia unipuncta sd Spodopteradolichos sf S frugiperda sl S latifascia so S ornithogallixc Xylophanes chiron

Spicule length was not an important character in dis-tinguishing among male Noctuidonema Male nematodeson S dolichos and S frugiperda showed the most sepa-ration with moderate overlap whereas males on S or-nithogalli appeared to be most similar to those on S

dolichos and males on S latifascia appeared most similarto those on S frugiperda Results from both genders agreethat nematodes on Pseudaletia are different from thoseon Spodoptera and that the nematodes on S dolichos andS frugiperda are the most different with the nematodeson S latifascia and S ornithogallibeing intermediate be-tween these two Overall the results do not support theexistence of separate species of Noctuidonema on hostsof the genus Spodoptera (Fig 4 Table 6)

These results provide evidence that nematodes on Ainfecta A oxygramma and E addens are different fromeach other and from the remaining groups having simplespicules Nematodes on M disseverans and O crenulatawere not distinct from V dibolia the species parasitisingM latipes Nematodes on M disseverans differed signif-icantly from V dibolia in only six morphological charac-ters in the females and ten in the males The differenceswere small and can be attributed to adaptation of conspe-ci c nematodes to parasitism of closely related sympatrichost species

Comparisonof the description of V daptria to measure-ments of other Vampyronema in this study indicates that itdiffers in stylet length and spicule shape V daptria hasthe shortest stylet length observed in the genus rangingfrom 53-60 sup1m in males and 59-69 sup1m in females whichis shorter than the stylet length of nematodes on A in-fecta (Fig 3) In addition spicules of V daptria taper to ne points whereas in Vampyronema from other hosts thespicule tips are blunt V daptria on L porcia appears tobe a distinct species and we therefore do not assign any ofthe Vampyronema from this study to it

Nematodes on the St Lucia and Guadeloupe popula-tions of Periplaneta americana are not well differenti-ated from each other based on the suite of characters thatwere examined (Fig 1A B) Although there are a fewsigni cantly different characters between the two popula-tions the differences probably do not represent interspe-ci c variation In addition the characters of both popula-tions are in agreement with the published description ofAcugutturus parasiticus (Hunt 1980)

Examination of additional specimens of ectoparasiticnematodes from other species of Lepidoptera showed thatmost were Vampyronema whereas Noctuidonema wasfound only on moths from two genera Pseudaletia andSpodoptera (Table 7)

Vol 4(4) 2002 499

OG Marti Jr et al

Table 7 Association of quadrine and trine Noctuidae with Vampyronema and Noctuidonema

Host Taxon Hind Wing Venation1 Locality2 Nematodes Genus

NoctuidaeCatocalinae

Anomis oedema Gueneacutee quadri ne FG 4 1 VampyronemaBaniana sp quadri ne FG 1 0 VampyronemaBaniana ostia Druce quadri ne FG 7 0 VampyronemaEuglyphis sp quadri ne FG 1 0 VampyronemaEulepidotis addens (Walker) quadri ne GU 5 2 VampyronemaItomia opisthographa (Gueneacutee) quadri ne FG 5 1 VampyronemaLesmone formularis (Geyer) quadri ne FG 1 0 VampyronemaLesmone porcia (Stoll) quadri ne GU ndash3 Vampyronema3

Melipotis fasciolaris (Huumlbner) quadri ne FG 0 1 VampyronemaMetalectra sp quadri ne FG 4 10 VampyronemaMetria sp quadri ne FG 0 1 VampyronemaMocis dif uens (Gueneacutee) quadri ne FG 1 1 Vampyronema4

Mocis disseverans (Walker) quadri ne GA 9 4 Vampyronema4

Mocis latipes Gueneacutee quadri ne GA 9 16 Vampyronema4

Perasia ora (Cramer)5 quadri ne FG 10 0 VampyronemaSelenisa sp quadri ne FG 4 0 VampyronemaZale ctilis (Gueneacutee) quadri ne FG 4 1 Vampyronema

PlusiinaeAgrapha oxygramma (Geyer) quadri ne GA 13 7 VampyronemaArgyrogramma verruca (Fabricius) quadri ne FG 0 2 VampyronemaChrysodeixis chalcites (Esper) quadri ne Fiji 8 7 Vampyronema

AmphipyrinaeSpodoptera androgea (Stoll) tri ne FG 4 0 Noctuidonema6

Spodoptera dolichos (Fabricius) tri ne GA 5 10 Noctuidonema6

Spodoptera frugiperda (JE Smith) tri ne GA 17 11 Noctuidonema6

Spodoptera latifascia (Walker) tri ne GA 12 6 Noctuidonema6

Spodoptera litura (Fabricius) tri ne Fiji 15 7 Noctuidonema6

Spodoptera ornithogalli (Gueneacutee) tri ne GA 10 10 Noctuidonema6

HadeninaeOrthodes crenulata (Butler) tri ne GA 9 4 VampyronemaPseudaletia unipuncta (Haworth) tri ne GA 18 9 Noctuidonema

NoctuinaeAnicla infecta (Ochsenheimer) tri ne GA 15 12 VampyronemaTandilia rodea (Schaus) tri ne FG 2 3 Vampyronema

PyralidaePyraustinae

Lamprosema moccalis Schaus ndash FG 4 0 VampyronemaSphingidae

MacroglossinaeErinnys obscura (Fabricius) ndash FG 22 2 Vampyronema7

Perigonia lusca Fabricius ndash GU 19 4 Vampyronema7

Xylophanes chiron Drury ndash GU 8 0 Vampyronema7

1) Quadri ne and tri ne hindwing venation applies to Noctuidae only Pyralidae and Sphingidae are included for completeness thestatus of Plusiinae as quadri nes is ambiguous because it groups together mostly species having quadri ne wing venation yet is oftenincluded with chie y tri ne subfamilies (Kitching amp Rawlins 1998) 2 Fiji Fiji Islands FG French Guiana GU Guadeloupe GAGeorgia USA 3 V daptria Anderson amp Laumond 1992 designationbased on Hunt (1993) 4 V dibolia 5 Synonymous with Nymbisarcuata Walker (in Rogers et al 1990a and Simmons amp Rogers 1996) 6 N guyanense 7 Probably represents an undescribed genus

500 Nematology

Speciation in the Acugutturidae

Discussion

The position of the excretory pore posterior to thebulb the comparatively short length of the stylet thecomparatively long tail and the simple rose-thorn-shapedspicules are suf cient to clearly separate Acugutturus aparasite of Blattodea from Vampyronema and Noctuido-nema parasites of Lepidoptera Sampled populations re-veal four inferred monophyleticgroups Acugutturus par-asiticus from two West Indian populations of Periplan-eta americana a new genus on Sphingidae Noctuido-nema guyanense from Spodoptera spp with nematodeson P unipuncta as a probable new species of Noctuido-nema and Vampyronema spp on the remaining hosts inthis study with the nematodes on A infecta representinga new species of Vampyronema

Each inferred monophyletic group can be de ned byseveral unique combinations of characters quantitativeand qualitative that satisfy the minimal requirements forspecies diagnosis (Wheeler 1999 Wheeler amp Platnick2000) Such characters appear to represent lineage inde-pendence via autapomorphies hypotheses ideally testedvia outgroup comparison Such tests while more philo-sophically satisfying (Adams 1998 Wiley amp Mayden2000) are beyond the scope of the present paper

According to Kitching and Rawlins (1998) differencesin hindwing venation broadly divide the Noctuidae intotwo phenetic groups tri nes and quadri nes Catocalinaeand Plusiinae are quadri nes (but see comment 1 in Ta-ble 7) whereas the other noctuid subfamilies mentionedin this study are tri nes Complex spicules probably rep-resent a derived condition found thus far only in nema-todes on two genera of tri ne noctuids Pseudaletia andSpodoptera while the simple spicules of Vampyronemaprobably represent a plesiomorphic state found in bothtri ne and quadri ne noctuids The extremely long styletsof nematodes from Sphingidae may have arisen in re-sponse to parasitisation of large moths

The Acugutturidae are associated with two very dis-tantly related insect orders Blattodea and Lepidoptera(Endopterygota) There is no record of occurrence ofthe Acugutturidae on any other insect order Among theLepidoptera the Acugutturidae and more precisely theNoctuidonematinae are known from only six familiesGeometridae Lasiocampidae Noctuidae NotodontidaePyralidae and Sphingidae (Marti et al 2000) most ofwhich are not closely related with the exceptionof Lasio-campidae and Sphingidae members of the Bombycoidea(and their relatives) clade (Lemaire amp Minet 1998) and

Notodontidaeand Noctuidae members of the Noctuoideasuperfamily (Kitching amp Rawlins 1998)

One explanation for such a surprising distribution ofthe Acugutturidae on their hosts may be that availabledata are too sparse due to insuf cient sampling of otherlepidopteran families While this may be true the studyconducted by Marti et al (2000) shows that there isa clear speci city in nematode host preference Martiet al (2000) updated the list of lepidopteran hosts ofNoctuidonematinae It seems clear that although manyLepidoptera species belonging to 11 families and 115genera have been examined for nematodes only membersof the six aforementioned obtectomeran (Kristensen ampSkalski 1998) families have been found to be positivefor Noctuidonematinae With the exception of Pyralidaeall these families are members of the Macrolepidopteraclade Moreover the positive Noctuidae species representnearly 90 of the total positive species found Thereforewe cannot state that the Noctuidonematinae are notpotentially parasites of most of the lepidopteran familiesbut we can state that at least in the Neotropical Regionthey parasitise the Macrolepidoptera preferentially andamong that clade mostly a single family the Noctuidae(Marti et al 2000)

The Aphelenchida include several different trophictypes with many species being mycetophagous or phy-toparasitic with about one third of known species be-ing entomoparasitic (Hunt 1993) Considering the list ofinsect hosts we cannot hypothesise that the Acugutturi-dae are derived from a single ancestor transferring froma plant to an insect host It seems more realistic to pro-pose several transfers from plant to insect hosts during thehistory of the Acugutturidae A transfer (Acugutturinae)occurred on the Blattodea and another one (Noctuidone-matinae) on the Lepidoptera Another transfer may haveoccurred on the Bombycoidea sensu lato If at least twosuch principal transfers occurred then it is possible thatthe Acugutturinae and Noctuidonematinae would not beconfamilial and that these two taxonomicunits deserve fa-milial rank Cladistic and molecularphylogeneticanalysiswill help to solve that problem The existenceof symbioticbacteria in these nematodes (Marti et al 1995) shouldprovide an independentmeans of analysing these relation-ships

Concerning the relationship between Lepidoptera andits ectoparasitic nematodes the present work highlightsthe following situation a new genus of Noctuidone-matinae parasitises Sphingidae a genuine Bombycoidea(Lemaire amp Minet 1998) Vampyronema parasitises the

Vol 4(4) 2002 501

OG Marti Jr et al

quadri ne noctuid subfamily Catocalinae (sensu lato in-cluding Ophiderinae) Plusiinae an Hadeninae and aNoctuinae (both typical tri ne noctuids) and Noctui-donema spp parasitises another Hadeninae and severalSpodoptera a tri ne noctuid genus previously includedin the Amphipyrinae but now of uncertain subfamilial af- liation (Poole 1995 Kitching amp Rawlins 1998) Exceptfor nematodes on Sphingidae which likely correspond toa well-separated clade whose phylogenetic relationshipto other Noctuidonematinae needs to be speci ed it ap-pears that Vampyronema and Noctuidonema both para-sitise Noctuidae but that Vampyronema parasitises bothquadri ne and tri ne noctuids Noctuidonema possibly aderived form of Noctuidonematinae is restricted to a fewgenera (Spodoptera and Pseudaletia) of tri ne noctuidsthat are not closely related (Mitchell et al 2000)

Recent work has questioned the monophylyof the Noc-tuidae and there is a growing feeling among moleculartaxonomists that the Noctuidae may be paraphyletic sev-eral quadri ne subfamilies including Catocalinae beingmore closely related to Arctiidae and Lymantriidae than totri ne noctuids (Weller et al 1994 Mitchell et al 1997Fang et al 2000 Mitchell et al 2000) The latter groupis considered as monophyleticClearly systematics of theNoctuidonematinaedo not match the tentative new classi- cation of the Noctuidae

This lack of congruence between parasite and hostsystematics leads us to consider that the relationshipsbetween Acugutturidae and their hosts have not beenguided by coevolutionary forces in the sense of Ehrlichand Raven (1964) but by horizontal transfers (Rogers ampMarti 1996) and sequential evolution (Jermy 1984 Fu-tuyma amp McCafferty 1990) The Acugutturidae and es-pecially the Noctuidonematinae have likely experiencedan adaptive radiation on higher Lepidoptera long aftertheir host diversi cation History of the Noctuidonemati-nae has been characterised by host transfer events ontofairly distantly related hosts followed by adaptation tonew hosts leading to speciation processes that now permitus to distinguish easily fairly species-speci c populationsor strains of Vampyronema or Noctuidonema In mothsthese host transfers may have occurred through contami-nation of nematode-free species by nematode eggs juve-niles or adults left on resting or feeding surfaces (leavesor owers) by individualsof infested species The successof such contamination may be linked to the availabilityand accessibility of host intersegmental membranes thinenough to be pierced by nematode stylets a factor that

may have physically and dramatically restricted the hostrange of all Acugutturidae

The present morphological analysis of acugutturid ne-matodes parasitising 13 species of Lepidoptera and onespecies of cockroach revealed a greater degree of diversitythan previously known in this group The genus Vampy-ronema grouping nematodes with simple spicules in-cludes two inferred new species in addition to V dibo-lia and V daptria already known from the LepidopteraThe genus Noctuidonema grouping nematodes of Lepi-doptera with complex spicules includes a new speciesand the nematodes found on Sphingidae probably belongto a new genus The relationshipbetween these nematodesand their hosts appears to be much more complex thanpreviously thought In addition to their bene ts as para-sites of pest Lepidoptera acugutturid nematodes are aninteresting model for parasite-host coevolutionarystudies

Acknowledgements

We thank David Hunt (CABI Bioscience UK) DanielLachaise (CNRS France) Alvin Simmons (USDA Char-leston SC USA) and S Patricia Stock (University ofCalifornia Davis CA USA) for reviewing earlier ver-sions of this paper We thank Jeanette Williams and JohnClayton for collecting cockroaches in St Lucia and mothsin Fiji respectively We thank the following individualsfor identifyingor con rming our identi cations of insectsDavid Nickle (Smithsonian Washington DC USA) forPeriplaneta americana Pierre Zagatti (INRA France)for Pyralidae and Hugo Kons (University of FloridaGainesville FL USA) for Orthodes crenulata

References

ADAMS BJ (1998) Species concepts and evolutionary para-digm in modern nematology Journal of Nematology 30 1-21

ANDERSON RV amp LAUMOND C (1990) NoctuidonemaguyanenseRemillet amp Silvain 1988 morphologie du spiculeredescription du macircle et diagnose amendeacutee du genre Noctui-donema Revue de Neacutematologie 13 433-436

ANDERSON RV amp LAUMOND C (1992) Noctuidonemadaptria n sp (Nematoda Aphelenchoididae)an ectoparasiteof the moth Lesmone porcia (Stoll) Journal of Nematology24 16-22

EHRLICH P amp RAVEN P (1964) Butter ies and plants astudy in evolution Evolution 18 586-608

FANG QQ MITCHELL A REGIER JC MITTER CFRIEDLANDER TP amp POOLE RW (2000) Phylogenetic

502 Nematology

Speciation in the Acugutturidae

and 14 113 for females) In addition ANOVA showedthat each morphological character contributed to thedifferentiation of nematodes of each gender from at leasttwo host species All F values were highly signi cant(P lt 0001 df D 11-13 83-93 for males and 14209-214 for females) Therefore all 18 characters sixratios and three percentages listed above were used insubsequentdiscriminant analysis to determine which weremost useful in distinguishing between nematodes fromdifferent hosts

STEP 1 15 HOST SPECIES (BLATTODEA AND

LEPIDOPTERA)

Discriminant analysis of female nematodes from 15host species revealed three separate groups i) fromthe two Periplaneta americana populations (Acugutturusparasiticus) ii) from the two species of Sphingidaeiii) all other host species (Fig 1A B) There were tensigni cant canonical roots accounting cumulatively for9978 of the explained variance between groups The rst two roots accounted for 9078 of the explainedvariance Examination of the factor structure matrixthe correlation of the morphological characters to thecanonical roots showed that the distance of the excretorypore from the anterior end the stylet length and V werethe characters most important in distinguishing betweenfemale nematodes from the 15 host species (Table 2)The other characters were of less importance althoughallwere signi cant for female nematodes

Discriminant analysis of male nematodes from 14 hostspecies (no male nematodes were found on Xylophaneschiron) revealed four distinct groups (Fig 1B Table 2)Canonical roots 1 (excretory pore) and 3 (stylet length)gave the best differentiation among males In additionto the differentiation of A parasiticus and the malenematodes on Perigonia lusca from each other and fromthe remaining host groups it was also found that malenematodes from Pseudaletia unipuncta were differentfrom those collected from the four species of SpodopteraThe two populations of A parasiticus did not differ butwere distinct from the other groups primarily due to thedifference in the position of the excretory pore

In the males there were eight signi cant (P lt 005)canonical roots accounting cumulatively for 9944 ofthe explained variance among males The rst threecanonical roots accounted for 9446 of the explainedvariance Canonical roots 1 and 3 are represented primar-ily by the excretory pore and the stylet length respec-

tively whereas root 2 (not shown) is represented primarilyby spicule length

STEP 2 11 HOST GROUPS (NEMATODES OF

LEPIDOPTERA NOCTUIDAE)

Analysis of morphological characters of nematodesparasitising only noctuids provides additional informa-tion differentiating several of these groups Among thefemales there were ten signi cant canonical roots ac-counting cumulatively for 100 of the explained vari-ance The rst three roots accounted cumulatively for806 of the explained variance with V stylet length andbulb width being the most important (Table 3) V was byfar the most important female character differentiatingne-matodes among the noctuid hosts and was more usefulthan tail length alone V provides clear separation of fe-male nematodes of Pseudaletia and Spodoptera from theother noctuid groups (Fig 1C) The former group is char-acterised by V gt 92 (ie short female tails) and by com-plex spicules and the latter group by V lt 92 (ie longerfemale tails) and simple spicules

Among male nematodes from noctuid hosts (Fig 1DTable 3) discriminant analysis provided clear differenti-ation in root 1 between males having complex spicules(nematodes from Pseudaletia and Spodoptera) vs thosehaving simple spicules (nematodes from Agrapha AniclaEulepidotis Mocis spp and Orthodes) There were ninesigni cant canonical roots which cumulativelyaccountedfor 9983 of the explained variance among groups The rst three roots accountingcumulatively for 920 of theexplained variance among groups were represented pri-marily by spicule length stylet length and stylet length asa percentage of body length respectively

Nematodes from Pseudaletia can be separated fromthose on Spodoptera by the shorter stylet length (Fig 1CD) Nematodes on Agrapha oxygramma can be separatedfrom M latipes primarily on the basis of the shorter styletlength (P lt 005) Similarly males on Anicla infectahave shorter spicules (P lt 005) than the nematodes onthe remaining host groups

Only two male nematodes from E addens were avail-able However these were distinct from nematodes on Mlatipes The female nematodes on E addens were alsowell differentiated from those on M latipes despite someoverlap (Fig 2A)

Vol 4(4) 2002 493

OG Marti Jr et al

Fig 1 Plot of discriminant functions for nematode morphometric characters Ellipses indicate 95 condence limits for selected hostspecies A Female nematodes from 15 host species (Blattodea and Lepidoptera) B Male nematodes from 14 host species (Blattodeaand Lepidoptera) C Female nematodes from 11 host species (Lepidoptera) D Male nematodes from 11 host species (Lepidoptera)Legend (Host Names) pa Periplaneta americana (St Lucia) pa2 P americana (Guadeloupe) ao Agrapha oxygramma ai Aniclainfecta ea Eulepidotis addens md Mocis disseverans ml M latipes oc Orthodes crenulata pl Perigonia lusca pu Pseudaletiaunipuncta sd Spodoptera dolichos sf S frugiperda sl S latifascia so S ornithogalli xc Xylophanes chiron

Table 2 Summary of forward stepwise discriminant analysis for male and female nematode morphological characters from 14 and 15host species respectively

Males ndash all groups (n D 14) Females ndash all groups (n D 15)

Step Character F (df1 df2) Character F (df1 df2)

1 Excretory pore 40313 (13 93) Excretory pore 61696 (14 214)2 Spicule length 20971 (13 92) Stylet length 16778 (14 213)3 Stylet length 5908 (13 91) V 15281 (14 212)4 Vulva or cloaca to ant end 1155 (13 90) Bulb width 1756 (14 211)5 Ant end to post bulb 905 (13 89) Excretory pore of body 1400 (14 210)6 Stylet percent of body 582 (13 88) a 1050 (14 209)7 B 702 (13 87) Ant end to post bulb 999 (14 208)8 G 640 (13 86) G 774 (14 207)9 Bulb width 447 (13 85) Vulva or cloaca to ant end 1023 (14 206)

10 Body width at nerve ring 419 (13 84) b 826 (14 205)

P lt 0001

494 Nematology

Speciation in the Acugutturidae

Table 3 Summary of forward stepwise discriminant analysis for male and female nematode morphological characters from 11 speciesof host Noctuidae (Lepidoptera)

Males ndash Lepidoptera groups1 (n D 11) Females ndash Lepidoptera groups1 (n D 11)

Step Character F (df1 df2) Character F (df1 df2)

1 Spicule length 24039 (10 80) V 10986 (10 150)2 Stylet length 2510 (10 79) Stylet length 6194 (10 149)3 Stylet of body length 1388 (10 78) Bulb width 1830 (10 148)4 Cloaca to anterior end 981 (10 77) Excretory pore of body 982 (10 147)5 Rostrum length 762 (10 76) Vulva to anterior end 935 (10 146)6 Ant end to ant gonad 700 (10 75) Body width at nerve ring 873 (10 145)7 G 795 (10 74) Stylet of body length 777 (10 144)8 c0 651 (10 73) Tail length 701 (10 143)9 Body width at nerve ring 615 (10 72) Head width 672 (10 142)

10 Bulb width 548 (10 71) Gonad of body length 595 (10 141)

1) Not including Perigonia lusca Xylophanes chiron or two populations from Periplaneta AmericanaP lt 0001

Fig 2 Plot of discriminant functions for nematode morphometric characters Ellipses indicate 95 condence limits for selected hostspecies A Female nematodes (Vampyronema) from six host species (Lepidoptera) B Male nematodes (Vampyronema) from six hostspecies (Lepidoptera)C Female nematodes (Noctuidonema) from ve host species (Lepidoptera)D Male nematodes (Noctuidonema)from ve host species (Lepidoptera) Legend (Host names) ao Agrapha oxygramma ai Anicla infecta ea Eulepidotis addens mdMocis disseverans ml M latipes oc Orthodes crenulata pu Pseudaletia unipuncta sd Spodoptera dolichos sf S frugiperda sl Slatifascia so S ornithogalli

Vol 4(4) 2002 495

OG Marti Jr et al

Table 4 Summary of forward stepwise discriminant analysis of morphological characters from male and female Vampyronema fromsix species of Noctuidae (Lepidoptera)

Males ndash Mocis group1 (n D 6) Females ndash Mocis group1 (n D 6)

Step Character F (df1 df2) Character F (df1 df2)

1 Stylet length 5807 (5 39) Stylet length 4573 (5 93)2 Rostrum length 1219 (5 38) Ant end to ant intestine 1879 (5 92)3 Bulb length 929 (5 37) Excretory pore of body 1803 (5 91)4 Ant end to ant gonad 837 (5 36) G 1500 (5 90)6 Pharyngeal gland width 511 (5 34) Stylet length of body 573 (5 89)7 Gonad length 438 (5 33) V 485 (5 88)8 Tail length 408 (5 32) Head length 476 (5 87)9 Bulb width 269 (5 31) Gonad length of body 445 (5 86)

10 Body width at nerve ring 355 (5 85)

1) Mocis group Agrapha oxygramma Anicla infecta Eulepidotis addens Mocis disseveransM latipes Orthodes crenulataP lt 005 P lt 001 P lt 0001

STEP 3 SIX HOST GROUPS (VAMPYRONEMANEMATODES WITH SIMPLE SPICULES)

The noctuidmoths Anicla infectaAgraphaoxygrammaEulepidotis addens Mocis disseverans M latipes andOrthodes crenulata are parasitised by Vampyronema inwhich the structure of the spicules is simpler and the tailof the females is longer than in Noctuidonema Analy-sis of morphometric characters of female nematodes fromthese hosts revealed ve signi cant canonical roots cu-mulatively accounting for 100 of the explainedvariancebetween groups The rst three roots representingprimar-ily stylet length distance from the anterior end to the baseof the bulb and the c ratio accounted cumulatively for9188 of the explained variance between groups Styletlength was the most important morphological characterdistinguishing female Vampyronema from different hosts(Fig 2A Table 4)

Analysis of results of male and female Vampyronemaon noctuid moths demonstrate the existence of fourdistinct clades One of these is V dibolia the speciesdescribed from M latipes The nematodes particularlythe males on M disseverans are very similar to V diboliaand we consider them to be conspeci c A second cladeis formed by nematodes from A infecta which are wellseparated from all the other groups (Fig 2A B) Styletlength is signi cantly shorter (P lt 001) in nematodeson A infecta than on those from the other hosts Athird Vampyronema clade is formed by nematodes onE addens which differ from V dibolia primarily in thedistance from the anterior end to the base of the bulb Thetwo male nematodes from E addens were also distinct

from other males an observation in agreement withresults from the females A fourth Vampyronema cladeis formed by nematodes from A oxygramma (Fig 2AB) Although the females appear to be intermediatebetween those on M latipes and E addens the malesare distinct based on differences in stylet and rostrumlength Nematodes on O crenulata are also intermediatebut males are very similar to V dibolia and we considerthem to be conspeci c

In the males there were ve signi cant canonical rootsrepresented primarily by stylet length spicule rostrumlength and bulb length respectively These accounted cu-mulatively for 929 of the explained variation betweenmales from these six host species Stylet length was themost important morphologicalcharacter in de ning males(Fig 2B) although rostrum length was also identi ed asan important character

STEP 4 FIVE HOST SPECIES (NOCTUIDONEMANEMATODES WITH COMPLEX SPICULES)

Discriminant analysis of female nematodes from fourSpodoptera species and Pseudaletia revealed four signif-icant (P lt 005) canonical roots The rst two rep-resenting primarily stylet length and body width at thevulva respectively accounted for a cumulative propor-tion of 8793 of the explained variance between groupsFemale nematodes from Pseudaletia were well separatedfrom those parasitising Spodoptera spp on the basis ofthe shorter stylet Nematodes from the four species ofSpodoptera were poorly de ned and showed a consider-able amount of overlap Even the groups that appeared to

496 Nematology

Speciation in the Acugutturidae

Table 5 Summary of forward stepwise discriminant analysis of morphological characters from male and female Noctuidonema from ve species of Spodoptera and Pseudaletia

Males ndash Spodoptera and Pseudaletia1 (n D 5) Females ndash Spodoptera and Pseudaletia1 (n D 5)

Step Character F (df1 df2) Character F (df1 df2)

1 Ant end to ant intestine 2524 (4 41) Stylet length 3914 (4 57)2 Body length 1491 (4 40) Body width at vulva 1250 (4 56)3 c0 1223 (4 39) Bulb width 591 (4 55)4 Bulb width 725 (4 38) a 498 (4 54)5 Body width at nerve ring 500 (4 37) Vulva to anterior end 380 (4 53)6 Spicule length 409 (4 36) Body width at nerve ring 579 (4 52)7 Maximum body width 265 (4 51)

1) Pseudaletia unipuncta Spodoptera dolichos S frugiperda S latifascia S ornithogalliP lt 005 P lt 001 P lt 0001

be the most different S frugiperda and S dolichos wereextensively overlapped (Fig 2D Table 5)

In the males there were three signi cant canonicalroots accounting for 9764 of the explained variancebetween groups The rst root represents the distancefrom the anterior end to the anterior intestine and V aboutequally whereas the second root represents primarily thebody width at the nerve ring (Fig 2D Table 5) Malenematodes from P unipuncta were well separated fromnematodes from Spodoptera which agrees with resultsfrom the females Males on S dolichosshowed a moderateamount of separation from males on the other threespecies of Spodoptera

DATA ANALYSIS

Eighteen morphological characters contribute to differ-entiation among male nematodes on 14 host species in-cluded in this study The most signi cant and useful ofthese are the distance of the excretory pore from the an-terior end spicule length and stylet length These threecharacters and V are the most useful in differentiating be-tween females (Table 2) Furthermore ANOVA showedthat the stylet length of nematodes on the two speciesof Sphingidae is signi cantly greater (P lt 0001) thanthe stylet length of nematodes on any of the other hostspecies The two populationsof nematodes from Sphingi-dae differ signi cantly in only eight morphological char-acters whereas each differs from V dibolia in about 20characters Females on P lusca differ from those on Xchiron in the longer length of the stylet greater length andwidth of the bulb greater distance of the excretory poreand bulb base from the anterior end greater body widthat the vulva and greater excretory pore distance as a per-

centage of body length However in the absence of malesfrom X chiron no conclusion can be reached on whethernematodes on Sphingidae are conspeci c The consistentseparation of nematodes parasitising Sphingidae from allother Acugutturidae suggests that they probably representat least one new genus Particularly remarkable is the longstylet which represents about 37 of the body length andaverages about 190 sup1m in length (Fig 3)

Discriminant analysis shows that the ve groups ofnematodes with complex spicules (Noctuidonema) arewell separated (Fig 1C D) from the six groups withsimple spicules (Vampyronema) Although the discretecharacter of spicule structure is suf cient to separate thesetwo genera particularly good separation can also be madebased on important continuously variable characters suchas stylet length distance of the excretory pore to theanterior end and the V ratio Furthermore within eachof these two genera separate groups can be detectedbased primarily on differences in stylet length Thusthe nematodes on Pseudaletia unipuncta differ fromthe nematodes on the four Spodoptera spp (Fig 1CD) and those on Anicla infecta Agrapha oxygrammaand Eulepidotis addens differ from V dibolia and itsconspeci c on M disseverans (Fig 2A B)

Female nematodes on P unipuncta differ (P lt 005)from those on S dolichos S frugiperda S latifasciaand S ornithogalli in eight 18 14 and 14 morpholo-gical characters respectivelyThese results providestrongsupport for separate taxonomic status of nematodes onP unipuncta as a new species of Noctuidonema Resultsfrom simple scatterplots prepared for each gender usingthe most importantmorphologicalcharacters identi ed bydiscriminant analysis agree with the preceding analysis

Vol 4(4) 2002 497

OG Marti Jr et al

Fig 3 Plot of stylet length of ectoparasitic nematodes from insect hosts (Blattodea and Lepidoptera) A Female nematodes from15 host species B Male nematodes from 14 host species Legend (Host names) ao Agrapha oxygramma ai Anicla infecta eaEulepidotis addens md Mocis disseverans ml M latipes oc Orthodes crenulata pa Periplaneta americana (St Lucia) pa2 Pamericana (Guadeloupe) pl Perigonia lusca pu Pseudaletia unipuncta sd Spodoptera dolichos sf S frugiperda sl S latifascia soS ornithogalli xc Xylophanes chiron

Table 6 Percent correct classication of male nematodes based on posterior probabilities that a case belongs to a particular hostspecies

Host species Percent correct classi cations ndash Percent correct classi cations ndashmale nematodes female nematodes

n Hosts Hosts Hosts Hosts n Hosts Hosts Hosts Hosts1-15 1-11 1-6 7-11 1-15 1-11 1-6 7-11

1 Agrapha oxygramma 7 100 100 100 ndash 12 917 100 917 ndash2 Anicla infecta 12 100 100 100 ndash 15 100 100 100 ndash3 Eulepidotis addens 2 100 100 100 ndash 15 100 100 100 ndash4 Mocis disseverans 4 100 75 75 ndash 9 889 889 889 ndash5 Mocis latipes 16 100 100 100 ndash 39 923 974 948 ndash6 Orthodes crenulata 4 100 100 100 ndash 9 889 889 889 ndash7 Spodoptera dolichos 10 100 100 ndash 90 5 100 100 ndash 808 Spodoptera frugiperda 11 909 909 ndash 100 17 765 824 ndash 7659 Spodoptera latifascia 6 667 833 ndash 50 12 75 833 ndash 667

10 Spodoptera ornithogalli 10 100 100 ndash 100 10 90 90 ndash 9011 Pseudaletia unipuncta 9 100 100 ndash 100 18 100 944 ndash 94412 Perigonia lusca 4 100 ndash ndash ndash 19 938 ndash ndash ndash13 Periplaneta americana (1) 7 100 ndash ndash ndash 18 882 ndash ndash ndash14 Periplaneta americana (2) 5 100 ndash ndash ndash 13 846 ndash ndash ndash15 Xylophanes chiron 0 ndash ndash ndash ndash 18 100 ndash ndash ndash

(1) St Lucia(2) Guadeloupe

498 Nematology

Speciation in the Acugutturidae

Fig 4 Tree diagram for ectoparasitic nematodes from insecthost species (Blattodea and Lepidoptera) The tree was pro-duced using Euclidean distance as the distance measure andcomplete linkage as the amalgamation rule Both graphs areplotted on an X-axis standardised by (DlinkDmax)100 whichrepresents the percentage of the range from the maximum tothe minimum distance in the data A Female nematodes datafor excretory pore stylet length V bulb width a anterior endto anterior intestine and vulva to anterior end were used BMale nematodes data for excretory pore spicule length andstylet length were used Legend (Host Names) pa Periplan-eta americana (St Lucia) pa2 P americana (Guadeloupe) aoAgrapha oxygramma ai Anicla infecta ea Eulepidotis addensmd Mocis disseverans ml M latipes oc Orthodes crenulatapl Perigonia lusca pu Pseudaletia unipuncta sd Spodopteradolichos sf S frugiperda sl S latifascia so S ornithogallixc Xylophanes chiron

Spicule length was not an important character in dis-tinguishing among male Noctuidonema Male nematodeson S dolichos and S frugiperda showed the most sepa-ration with moderate overlap whereas males on S or-nithogalli appeared to be most similar to those on S

dolichos and males on S latifascia appeared most similarto those on S frugiperda Results from both genders agreethat nematodes on Pseudaletia are different from thoseon Spodoptera and that the nematodes on S dolichos andS frugiperda are the most different with the nematodeson S latifascia and S ornithogallibeing intermediate be-tween these two Overall the results do not support theexistence of separate species of Noctuidonema on hostsof the genus Spodoptera (Fig 4 Table 6)

These results provide evidence that nematodes on Ainfecta A oxygramma and E addens are different fromeach other and from the remaining groups having simplespicules Nematodes on M disseverans and O crenulatawere not distinct from V dibolia the species parasitisingM latipes Nematodes on M disseverans differed signif-icantly from V dibolia in only six morphological charac-ters in the females and ten in the males The differenceswere small and can be attributed to adaptation of conspe-ci c nematodes to parasitism of closely related sympatrichost species

Comparisonof the description of V daptria to measure-ments of other Vampyronema in this study indicates that itdiffers in stylet length and spicule shape V daptria hasthe shortest stylet length observed in the genus rangingfrom 53-60 sup1m in males and 59-69 sup1m in females whichis shorter than the stylet length of nematodes on A in-fecta (Fig 3) In addition spicules of V daptria taper to ne points whereas in Vampyronema from other hosts thespicule tips are blunt V daptria on L porcia appears tobe a distinct species and we therefore do not assign any ofthe Vampyronema from this study to it

Nematodes on the St Lucia and Guadeloupe popula-tions of Periplaneta americana are not well differenti-ated from each other based on the suite of characters thatwere examined (Fig 1A B) Although there are a fewsigni cantly different characters between the two popula-tions the differences probably do not represent interspe-ci c variation In addition the characters of both popula-tions are in agreement with the published description ofAcugutturus parasiticus (Hunt 1980)

Examination of additional specimens of ectoparasiticnematodes from other species of Lepidoptera showed thatmost were Vampyronema whereas Noctuidonema wasfound only on moths from two genera Pseudaletia andSpodoptera (Table 7)

Vol 4(4) 2002 499

OG Marti Jr et al

Table 7 Association of quadrine and trine Noctuidae with Vampyronema and Noctuidonema

Host Taxon Hind Wing Venation1 Locality2 Nematodes Genus

NoctuidaeCatocalinae

Anomis oedema Gueneacutee quadri ne FG 4 1 VampyronemaBaniana sp quadri ne FG 1 0 VampyronemaBaniana ostia Druce quadri ne FG 7 0 VampyronemaEuglyphis sp quadri ne FG 1 0 VampyronemaEulepidotis addens (Walker) quadri ne GU 5 2 VampyronemaItomia opisthographa (Gueneacutee) quadri ne FG 5 1 VampyronemaLesmone formularis (Geyer) quadri ne FG 1 0 VampyronemaLesmone porcia (Stoll) quadri ne GU ndash3 Vampyronema3

Melipotis fasciolaris (Huumlbner) quadri ne FG 0 1 VampyronemaMetalectra sp quadri ne FG 4 10 VampyronemaMetria sp quadri ne FG 0 1 VampyronemaMocis dif uens (Gueneacutee) quadri ne FG 1 1 Vampyronema4

Mocis disseverans (Walker) quadri ne GA 9 4 Vampyronema4

Mocis latipes Gueneacutee quadri ne GA 9 16 Vampyronema4

Perasia ora (Cramer)5 quadri ne FG 10 0 VampyronemaSelenisa sp quadri ne FG 4 0 VampyronemaZale ctilis (Gueneacutee) quadri ne FG 4 1 Vampyronema

PlusiinaeAgrapha oxygramma (Geyer) quadri ne GA 13 7 VampyronemaArgyrogramma verruca (Fabricius) quadri ne FG 0 2 VampyronemaChrysodeixis chalcites (Esper) quadri ne Fiji 8 7 Vampyronema

AmphipyrinaeSpodoptera androgea (Stoll) tri ne FG 4 0 Noctuidonema6

Spodoptera dolichos (Fabricius) tri ne GA 5 10 Noctuidonema6

Spodoptera frugiperda (JE Smith) tri ne GA 17 11 Noctuidonema6

Spodoptera latifascia (Walker) tri ne GA 12 6 Noctuidonema6

Spodoptera litura (Fabricius) tri ne Fiji 15 7 Noctuidonema6

Spodoptera ornithogalli (Gueneacutee) tri ne GA 10 10 Noctuidonema6

HadeninaeOrthodes crenulata (Butler) tri ne GA 9 4 VampyronemaPseudaletia unipuncta (Haworth) tri ne GA 18 9 Noctuidonema

NoctuinaeAnicla infecta (Ochsenheimer) tri ne GA 15 12 VampyronemaTandilia rodea (Schaus) tri ne FG 2 3 Vampyronema

PyralidaePyraustinae

Lamprosema moccalis Schaus ndash FG 4 0 VampyronemaSphingidae

MacroglossinaeErinnys obscura (Fabricius) ndash FG 22 2 Vampyronema7

Perigonia lusca Fabricius ndash GU 19 4 Vampyronema7

Xylophanes chiron Drury ndash GU 8 0 Vampyronema7

1) Quadri ne and tri ne hindwing venation applies to Noctuidae only Pyralidae and Sphingidae are included for completeness thestatus of Plusiinae as quadri nes is ambiguous because it groups together mostly species having quadri ne wing venation yet is oftenincluded with chie y tri ne subfamilies (Kitching amp Rawlins 1998) 2 Fiji Fiji Islands FG French Guiana GU Guadeloupe GAGeorgia USA 3 V daptria Anderson amp Laumond 1992 designationbased on Hunt (1993) 4 V dibolia 5 Synonymous with Nymbisarcuata Walker (in Rogers et al 1990a and Simmons amp Rogers 1996) 6 N guyanense 7 Probably represents an undescribed genus

500 Nematology

Speciation in the Acugutturidae

Discussion

The position of the excretory pore posterior to thebulb the comparatively short length of the stylet thecomparatively long tail and the simple rose-thorn-shapedspicules are suf cient to clearly separate Acugutturus aparasite of Blattodea from Vampyronema and Noctuido-nema parasites of Lepidoptera Sampled populations re-veal four inferred monophyleticgroups Acugutturus par-asiticus from two West Indian populations of Periplan-eta americana a new genus on Sphingidae Noctuido-nema guyanense from Spodoptera spp with nematodeson P unipuncta as a probable new species of Noctuido-nema and Vampyronema spp on the remaining hosts inthis study with the nematodes on A infecta representinga new species of Vampyronema

Each inferred monophyletic group can be de ned byseveral unique combinations of characters quantitativeand qualitative that satisfy the minimal requirements forspecies diagnosis (Wheeler 1999 Wheeler amp Platnick2000) Such characters appear to represent lineage inde-pendence via autapomorphies hypotheses ideally testedvia outgroup comparison Such tests while more philo-sophically satisfying (Adams 1998 Wiley amp Mayden2000) are beyond the scope of the present paper

According to Kitching and Rawlins (1998) differencesin hindwing venation broadly divide the Noctuidae intotwo phenetic groups tri nes and quadri nes Catocalinaeand Plusiinae are quadri nes (but see comment 1 in Ta-ble 7) whereas the other noctuid subfamilies mentionedin this study are tri nes Complex spicules probably rep-resent a derived condition found thus far only in nema-todes on two genera of tri ne noctuids Pseudaletia andSpodoptera while the simple spicules of Vampyronemaprobably represent a plesiomorphic state found in bothtri ne and quadri ne noctuids The extremely long styletsof nematodes from Sphingidae may have arisen in re-sponse to parasitisation of large moths

The Acugutturidae are associated with two very dis-tantly related insect orders Blattodea and Lepidoptera(Endopterygota) There is no record of occurrence ofthe Acugutturidae on any other insect order Among theLepidoptera the Acugutturidae and more precisely theNoctuidonematinae are known from only six familiesGeometridae Lasiocampidae Noctuidae NotodontidaePyralidae and Sphingidae (Marti et al 2000) most ofwhich are not closely related with the exceptionof Lasio-campidae and Sphingidae members of the Bombycoidea(and their relatives) clade (Lemaire amp Minet 1998) and

Notodontidaeand Noctuidae members of the Noctuoideasuperfamily (Kitching amp Rawlins 1998)

One explanation for such a surprising distribution ofthe Acugutturidae on their hosts may be that availabledata are too sparse due to insuf cient sampling of otherlepidopteran families While this may be true the studyconducted by Marti et al (2000) shows that there isa clear speci city in nematode host preference Martiet al (2000) updated the list of lepidopteran hosts ofNoctuidonematinae It seems clear that although manyLepidoptera species belonging to 11 families and 115genera have been examined for nematodes only membersof the six aforementioned obtectomeran (Kristensen ampSkalski 1998) families have been found to be positivefor Noctuidonematinae With the exception of Pyralidaeall these families are members of the Macrolepidopteraclade Moreover the positive Noctuidae species representnearly 90 of the total positive species found Thereforewe cannot state that the Noctuidonematinae are notpotentially parasites of most of the lepidopteran familiesbut we can state that at least in the Neotropical Regionthey parasitise the Macrolepidoptera preferentially andamong that clade mostly a single family the Noctuidae(Marti et al 2000)

The Aphelenchida include several different trophictypes with many species being mycetophagous or phy-toparasitic with about one third of known species be-ing entomoparasitic (Hunt 1993) Considering the list ofinsect hosts we cannot hypothesise that the Acugutturi-dae are derived from a single ancestor transferring froma plant to an insect host It seems more realistic to pro-pose several transfers from plant to insect hosts during thehistory of the Acugutturidae A transfer (Acugutturinae)occurred on the Blattodea and another one (Noctuidone-matinae) on the Lepidoptera Another transfer may haveoccurred on the Bombycoidea sensu lato If at least twosuch principal transfers occurred then it is possible thatthe Acugutturinae and Noctuidonematinae would not beconfamilial and that these two taxonomicunits deserve fa-milial rank Cladistic and molecularphylogeneticanalysiswill help to solve that problem The existenceof symbioticbacteria in these nematodes (Marti et al 1995) shouldprovide an independentmeans of analysing these relation-ships

Concerning the relationship between Lepidoptera andits ectoparasitic nematodes the present work highlightsthe following situation a new genus of Noctuidone-matinae parasitises Sphingidae a genuine Bombycoidea(Lemaire amp Minet 1998) Vampyronema parasitises the

Vol 4(4) 2002 501

OG Marti Jr et al

quadri ne noctuid subfamily Catocalinae (sensu lato in-cluding Ophiderinae) Plusiinae an Hadeninae and aNoctuinae (both typical tri ne noctuids) and Noctui-donema spp parasitises another Hadeninae and severalSpodoptera a tri ne noctuid genus previously includedin the Amphipyrinae but now of uncertain subfamilial af- liation (Poole 1995 Kitching amp Rawlins 1998) Exceptfor nematodes on Sphingidae which likely correspond toa well-separated clade whose phylogenetic relationshipto other Noctuidonematinae needs to be speci ed it ap-pears that Vampyronema and Noctuidonema both para-sitise Noctuidae but that Vampyronema parasitises bothquadri ne and tri ne noctuids Noctuidonema possibly aderived form of Noctuidonematinae is restricted to a fewgenera (Spodoptera and Pseudaletia) of tri ne noctuidsthat are not closely related (Mitchell et al 2000)

Recent work has questioned the monophylyof the Noc-tuidae and there is a growing feeling among moleculartaxonomists that the Noctuidae may be paraphyletic sev-eral quadri ne subfamilies including Catocalinae beingmore closely related to Arctiidae and Lymantriidae than totri ne noctuids (Weller et al 1994 Mitchell et al 1997Fang et al 2000 Mitchell et al 2000) The latter groupis considered as monophyleticClearly systematics of theNoctuidonematinaedo not match the tentative new classi- cation of the Noctuidae

This lack of congruence between parasite and hostsystematics leads us to consider that the relationshipsbetween Acugutturidae and their hosts have not beenguided by coevolutionary forces in the sense of Ehrlichand Raven (1964) but by horizontal transfers (Rogers ampMarti 1996) and sequential evolution (Jermy 1984 Fu-tuyma amp McCafferty 1990) The Acugutturidae and es-pecially the Noctuidonematinae have likely experiencedan adaptive radiation on higher Lepidoptera long aftertheir host diversi cation History of the Noctuidonemati-nae has been characterised by host transfer events ontofairly distantly related hosts followed by adaptation tonew hosts leading to speciation processes that now permitus to distinguish easily fairly species-speci c populationsor strains of Vampyronema or Noctuidonema In mothsthese host transfers may have occurred through contami-nation of nematode-free species by nematode eggs juve-niles or adults left on resting or feeding surfaces (leavesor owers) by individualsof infested species The successof such contamination may be linked to the availabilityand accessibility of host intersegmental membranes thinenough to be pierced by nematode stylets a factor that

may have physically and dramatically restricted the hostrange of all Acugutturidae

The present morphological analysis of acugutturid ne-matodes parasitising 13 species of Lepidoptera and onespecies of cockroach revealed a greater degree of diversitythan previously known in this group The genus Vampy-ronema grouping nematodes with simple spicules in-cludes two inferred new species in addition to V dibo-lia and V daptria already known from the LepidopteraThe genus Noctuidonema grouping nematodes of Lepi-doptera with complex spicules includes a new speciesand the nematodes found on Sphingidae probably belongto a new genus The relationshipbetween these nematodesand their hosts appears to be much more complex thanpreviously thought In addition to their bene ts as para-sites of pest Lepidoptera acugutturid nematodes are aninteresting model for parasite-host coevolutionarystudies

Acknowledgements

We thank David Hunt (CABI Bioscience UK) DanielLachaise (CNRS France) Alvin Simmons (USDA Char-leston SC USA) and S Patricia Stock (University ofCalifornia Davis CA USA) for reviewing earlier ver-sions of this paper We thank Jeanette Williams and JohnClayton for collecting cockroaches in St Lucia and mothsin Fiji respectively We thank the following individualsfor identifyingor con rming our identi cations of insectsDavid Nickle (Smithsonian Washington DC USA) forPeriplaneta americana Pierre Zagatti (INRA France)for Pyralidae and Hugo Kons (University of FloridaGainesville FL USA) for Orthodes crenulata

References

ADAMS BJ (1998) Species concepts and evolutionary para-digm in modern nematology Journal of Nematology 30 1-21

ANDERSON RV amp LAUMOND C (1990) NoctuidonemaguyanenseRemillet amp Silvain 1988 morphologie du spiculeredescription du macircle et diagnose amendeacutee du genre Noctui-donema Revue de Neacutematologie 13 433-436

ANDERSON RV amp LAUMOND C (1992) Noctuidonemadaptria n sp (Nematoda Aphelenchoididae)an ectoparasiteof the moth Lesmone porcia (Stoll) Journal of Nematology24 16-22

EHRLICH P amp RAVEN P (1964) Butter ies and plants astudy in evolution Evolution 18 586-608

FANG QQ MITCHELL A REGIER JC MITTER CFRIEDLANDER TP amp POOLE RW (2000) Phylogenetic

502 Nematology

OG Marti Jr et al

Fig 1 Plot of discriminant functions for nematode morphometric characters Ellipses indicate 95 condence limits for selected hostspecies A Female nematodes from 15 host species (Blattodea and Lepidoptera) B Male nematodes from 14 host species (Blattodeaand Lepidoptera) C Female nematodes from 11 host species (Lepidoptera) D Male nematodes from 11 host species (Lepidoptera)Legend (Host Names) pa Periplaneta americana (St Lucia) pa2 P americana (Guadeloupe) ao Agrapha oxygramma ai Aniclainfecta ea Eulepidotis addens md Mocis disseverans ml M latipes oc Orthodes crenulata pl Perigonia lusca pu Pseudaletiaunipuncta sd Spodoptera dolichos sf S frugiperda sl S latifascia so S ornithogalli xc Xylophanes chiron

Table 2 Summary of forward stepwise discriminant analysis for male and female nematode morphological characters from 14 and 15host species respectively

Males ndash all groups (n D 14) Females ndash all groups (n D 15)

Step Character F (df1 df2) Character F (df1 df2)

1 Excretory pore 40313 (13 93) Excretory pore 61696 (14 214)2 Spicule length 20971 (13 92) Stylet length 16778 (14 213)3 Stylet length 5908 (13 91) V 15281 (14 212)4 Vulva or cloaca to ant end 1155 (13 90) Bulb width 1756 (14 211)5 Ant end to post bulb 905 (13 89) Excretory pore of body 1400 (14 210)6 Stylet percent of body 582 (13 88) a 1050 (14 209)7 B 702 (13 87) Ant end to post bulb 999 (14 208)8 G 640 (13 86) G 774 (14 207)9 Bulb width 447 (13 85) Vulva or cloaca to ant end 1023 (14 206)

10 Body width at nerve ring 419 (13 84) b 826 (14 205)

P lt 0001

494 Nematology

Speciation in the Acugutturidae

Table 3 Summary of forward stepwise discriminant analysis for male and female nematode morphological characters from 11 speciesof host Noctuidae (Lepidoptera)

Males ndash Lepidoptera groups1 (n D 11) Females ndash Lepidoptera groups1 (n D 11)

Step Character F (df1 df2) Character F (df1 df2)

1 Spicule length 24039 (10 80) V 10986 (10 150)2 Stylet length 2510 (10 79) Stylet length 6194 (10 149)3 Stylet of body length 1388 (10 78) Bulb width 1830 (10 148)4 Cloaca to anterior end 981 (10 77) Excretory pore of body 982 (10 147)5 Rostrum length 762 (10 76) Vulva to anterior end 935 (10 146)6 Ant end to ant gonad 700 (10 75) Body width at nerve ring 873 (10 145)7 G 795 (10 74) Stylet of body length 777 (10 144)8 c0 651 (10 73) Tail length 701 (10 143)9 Body width at nerve ring 615 (10 72) Head width 672 (10 142)

10 Bulb width 548 (10 71) Gonad of body length 595 (10 141)

1) Not including Perigonia lusca Xylophanes chiron or two populations from Periplaneta AmericanaP lt 0001

Fig 2 Plot of discriminant functions for nematode morphometric characters Ellipses indicate 95 condence limits for selected hostspecies A Female nematodes (Vampyronema) from six host species (Lepidoptera) B Male nematodes (Vampyronema) from six hostspecies (Lepidoptera)C Female nematodes (Noctuidonema) from ve host species (Lepidoptera)D Male nematodes (Noctuidonema)from ve host species (Lepidoptera) Legend (Host names) ao Agrapha oxygramma ai Anicla infecta ea Eulepidotis addens mdMocis disseverans ml M latipes oc Orthodes crenulata pu Pseudaletia unipuncta sd Spodoptera dolichos sf S frugiperda sl Slatifascia so S ornithogalli

Vol 4(4) 2002 495

OG Marti Jr et al

Table 4 Summary of forward stepwise discriminant analysis of morphological characters from male and female Vampyronema fromsix species of Noctuidae (Lepidoptera)

Males ndash Mocis group1 (n D 6) Females ndash Mocis group1 (n D 6)

Step Character F (df1 df2) Character F (df1 df2)

1 Stylet length 5807 (5 39) Stylet length 4573 (5 93)2 Rostrum length 1219 (5 38) Ant end to ant intestine 1879 (5 92)3 Bulb length 929 (5 37) Excretory pore of body 1803 (5 91)4 Ant end to ant gonad 837 (5 36) G 1500 (5 90)6 Pharyngeal gland width 511 (5 34) Stylet length of body 573 (5 89)7 Gonad length 438 (5 33) V 485 (5 88)8 Tail length 408 (5 32) Head length 476 (5 87)9 Bulb width 269 (5 31) Gonad length of body 445 (5 86)

10 Body width at nerve ring 355 (5 85)

1) Mocis group Agrapha oxygramma Anicla infecta Eulepidotis addens Mocis disseveransM latipes Orthodes crenulataP lt 005 P lt 001 P lt 0001

STEP 3 SIX HOST GROUPS (VAMPYRONEMANEMATODES WITH SIMPLE SPICULES)

The noctuidmoths Anicla infectaAgraphaoxygrammaEulepidotis addens Mocis disseverans M latipes andOrthodes crenulata are parasitised by Vampyronema inwhich the structure of the spicules is simpler and the tailof the females is longer than in Noctuidonema Analy-sis of morphometric characters of female nematodes fromthese hosts revealed ve signi cant canonical roots cu-mulatively accounting for 100 of the explainedvariancebetween groups The rst three roots representingprimar-ily stylet length distance from the anterior end to the baseof the bulb and the c ratio accounted cumulatively for9188 of the explained variance between groups Styletlength was the most important morphological characterdistinguishing female Vampyronema from different hosts(Fig 2A Table 4)

Analysis of results of male and female Vampyronemaon noctuid moths demonstrate the existence of fourdistinct clades One of these is V dibolia the speciesdescribed from M latipes The nematodes particularlythe males on M disseverans are very similar to V diboliaand we consider them to be conspeci c A second cladeis formed by nematodes from A infecta which are wellseparated from all the other groups (Fig 2A B) Styletlength is signi cantly shorter (P lt 001) in nematodeson A infecta than on those from the other hosts Athird Vampyronema clade is formed by nematodes onE addens which differ from V dibolia primarily in thedistance from the anterior end to the base of the bulb Thetwo male nematodes from E addens were also distinct

from other males an observation in agreement withresults from the females A fourth Vampyronema cladeis formed by nematodes from A oxygramma (Fig 2AB) Although the females appear to be intermediatebetween those on M latipes and E addens the malesare distinct based on differences in stylet and rostrumlength Nematodes on O crenulata are also intermediatebut males are very similar to V dibolia and we considerthem to be conspeci c

In the males there were ve signi cant canonical rootsrepresented primarily by stylet length spicule rostrumlength and bulb length respectively These accounted cu-mulatively for 929 of the explained variation betweenmales from these six host species Stylet length was themost important morphologicalcharacter in de ning males(Fig 2B) although rostrum length was also identi ed asan important character

STEP 4 FIVE HOST SPECIES (NOCTUIDONEMANEMATODES WITH COMPLEX SPICULES)

Discriminant analysis of female nematodes from fourSpodoptera species and Pseudaletia revealed four signif-icant (P lt 005) canonical roots The rst two rep-resenting primarily stylet length and body width at thevulva respectively accounted for a cumulative propor-tion of 8793 of the explained variance between groupsFemale nematodes from Pseudaletia were well separatedfrom those parasitising Spodoptera spp on the basis ofthe shorter stylet Nematodes from the four species ofSpodoptera were poorly de ned and showed a consider-able amount of overlap Even the groups that appeared to

496 Nematology

Speciation in the Acugutturidae

Table 5 Summary of forward stepwise discriminant analysis of morphological characters from male and female Noctuidonema from ve species of Spodoptera and Pseudaletia

Males ndash Spodoptera and Pseudaletia1 (n D 5) Females ndash Spodoptera and Pseudaletia1 (n D 5)

Step Character F (df1 df2) Character F (df1 df2)

1 Ant end to ant intestine 2524 (4 41) Stylet length 3914 (4 57)2 Body length 1491 (4 40) Body width at vulva 1250 (4 56)3 c0 1223 (4 39) Bulb width 591 (4 55)4 Bulb width 725 (4 38) a 498 (4 54)5 Body width at nerve ring 500 (4 37) Vulva to anterior end 380 (4 53)6 Spicule length 409 (4 36) Body width at nerve ring 579 (4 52)7 Maximum body width 265 (4 51)

1) Pseudaletia unipuncta Spodoptera dolichos S frugiperda S latifascia S ornithogalliP lt 005 P lt 001 P lt 0001

be the most different S frugiperda and S dolichos wereextensively overlapped (Fig 2D Table 5)

In the males there were three signi cant canonicalroots accounting for 9764 of the explained variancebetween groups The rst root represents the distancefrom the anterior end to the anterior intestine and V aboutequally whereas the second root represents primarily thebody width at the nerve ring (Fig 2D Table 5) Malenematodes from P unipuncta were well separated fromnematodes from Spodoptera which agrees with resultsfrom the females Males on S dolichosshowed a moderateamount of separation from males on the other threespecies of Spodoptera

DATA ANALYSIS

Eighteen morphological characters contribute to differ-entiation among male nematodes on 14 host species in-cluded in this study The most signi cant and useful ofthese are the distance of the excretory pore from the an-terior end spicule length and stylet length These threecharacters and V are the most useful in differentiating be-tween females (Table 2) Furthermore ANOVA showedthat the stylet length of nematodes on the two speciesof Sphingidae is signi cantly greater (P lt 0001) thanthe stylet length of nematodes on any of the other hostspecies The two populationsof nematodes from Sphingi-dae differ signi cantly in only eight morphological char-acters whereas each differs from V dibolia in about 20characters Females on P lusca differ from those on Xchiron in the longer length of the stylet greater length andwidth of the bulb greater distance of the excretory poreand bulb base from the anterior end greater body widthat the vulva and greater excretory pore distance as a per-

centage of body length However in the absence of malesfrom X chiron no conclusion can be reached on whethernematodes on Sphingidae are conspeci c The consistentseparation of nematodes parasitising Sphingidae from allother Acugutturidae suggests that they probably representat least one new genus Particularly remarkable is the longstylet which represents about 37 of the body length andaverages about 190 sup1m in length (Fig 3)

Discriminant analysis shows that the ve groups ofnematodes with complex spicules (Noctuidonema) arewell separated (Fig 1C D) from the six groups withsimple spicules (Vampyronema) Although the discretecharacter of spicule structure is suf cient to separate thesetwo genera particularly good separation can also be madebased on important continuously variable characters suchas stylet length distance of the excretory pore to theanterior end and the V ratio Furthermore within eachof these two genera separate groups can be detectedbased primarily on differences in stylet length Thusthe nematodes on Pseudaletia unipuncta differ fromthe nematodes on the four Spodoptera spp (Fig 1CD) and those on Anicla infecta Agrapha oxygrammaand Eulepidotis addens differ from V dibolia and itsconspeci c on M disseverans (Fig 2A B)

Female nematodes on P unipuncta differ (P lt 005)from those on S dolichos S frugiperda S latifasciaand S ornithogalli in eight 18 14 and 14 morpholo-gical characters respectivelyThese results providestrongsupport for separate taxonomic status of nematodes onP unipuncta as a new species of Noctuidonema Resultsfrom simple scatterplots prepared for each gender usingthe most importantmorphologicalcharacters identi ed bydiscriminant analysis agree with the preceding analysis

Vol 4(4) 2002 497

OG Marti Jr et al

Fig 3 Plot of stylet length of ectoparasitic nematodes from insect hosts (Blattodea and Lepidoptera) A Female nematodes from15 host species B Male nematodes from 14 host species Legend (Host names) ao Agrapha oxygramma ai Anicla infecta eaEulepidotis addens md Mocis disseverans ml M latipes oc Orthodes crenulata pa Periplaneta americana (St Lucia) pa2 Pamericana (Guadeloupe) pl Perigonia lusca pu Pseudaletia unipuncta sd Spodoptera dolichos sf S frugiperda sl S latifascia soS ornithogalli xc Xylophanes chiron

Table 6 Percent correct classication of male nematodes based on posterior probabilities that a case belongs to a particular hostspecies

Host species Percent correct classi cations ndash Percent correct classi cations ndashmale nematodes female nematodes

n Hosts Hosts Hosts Hosts n Hosts Hosts Hosts Hosts1-15 1-11 1-6 7-11 1-15 1-11 1-6 7-11

1 Agrapha oxygramma 7 100 100 100 ndash 12 917 100 917 ndash2 Anicla infecta 12 100 100 100 ndash 15 100 100 100 ndash3 Eulepidotis addens 2 100 100 100 ndash 15 100 100 100 ndash4 Mocis disseverans 4 100 75 75 ndash 9 889 889 889 ndash5 Mocis latipes 16 100 100 100 ndash 39 923 974 948 ndash6 Orthodes crenulata 4 100 100 100 ndash 9 889 889 889 ndash7 Spodoptera dolichos 10 100 100 ndash 90 5 100 100 ndash 808 Spodoptera frugiperda 11 909 909 ndash 100 17 765 824 ndash 7659 Spodoptera latifascia 6 667 833 ndash 50 12 75 833 ndash 667

10 Spodoptera ornithogalli 10 100 100 ndash 100 10 90 90 ndash 9011 Pseudaletia unipuncta 9 100 100 ndash 100 18 100 944 ndash 94412 Perigonia lusca 4 100 ndash ndash ndash 19 938 ndash ndash ndash13 Periplaneta americana (1) 7 100 ndash ndash ndash 18 882 ndash ndash ndash14 Periplaneta americana (2) 5 100 ndash ndash ndash 13 846 ndash ndash ndash15 Xylophanes chiron 0 ndash ndash ndash ndash 18 100 ndash ndash ndash

(1) St Lucia(2) Guadeloupe

498 Nematology

Speciation in the Acugutturidae

Fig 4 Tree diagram for ectoparasitic nematodes from insecthost species (Blattodea and Lepidoptera) The tree was pro-duced using Euclidean distance as the distance measure andcomplete linkage as the amalgamation rule Both graphs areplotted on an X-axis standardised by (DlinkDmax)100 whichrepresents the percentage of the range from the maximum tothe minimum distance in the data A Female nematodes datafor excretory pore stylet length V bulb width a anterior endto anterior intestine and vulva to anterior end were used BMale nematodes data for excretory pore spicule length andstylet length were used Legend (Host Names) pa Periplan-eta americana (St Lucia) pa2 P americana (Guadeloupe) aoAgrapha oxygramma ai Anicla infecta ea Eulepidotis addensmd Mocis disseverans ml M latipes oc Orthodes crenulatapl Perigonia lusca pu Pseudaletia unipuncta sd Spodopteradolichos sf S frugiperda sl S latifascia so S ornithogallixc Xylophanes chiron

Spicule length was not an important character in dis-tinguishing among male Noctuidonema Male nematodeson S dolichos and S frugiperda showed the most sepa-ration with moderate overlap whereas males on S or-nithogalli appeared to be most similar to those on S

dolichos and males on S latifascia appeared most similarto those on S frugiperda Results from both genders agreethat nematodes on Pseudaletia are different from thoseon Spodoptera and that the nematodes on S dolichos andS frugiperda are the most different with the nematodeson S latifascia and S ornithogallibeing intermediate be-tween these two Overall the results do not support theexistence of separate species of Noctuidonema on hostsof the genus Spodoptera (Fig 4 Table 6)

These results provide evidence that nematodes on Ainfecta A oxygramma and E addens are different fromeach other and from the remaining groups having simplespicules Nematodes on M disseverans and O crenulatawere not distinct from V dibolia the species parasitisingM latipes Nematodes on M disseverans differed signif-icantly from V dibolia in only six morphological charac-ters in the females and ten in the males The differenceswere small and can be attributed to adaptation of conspe-ci c nematodes to parasitism of closely related sympatrichost species

Comparisonof the description of V daptria to measure-ments of other Vampyronema in this study indicates that itdiffers in stylet length and spicule shape V daptria hasthe shortest stylet length observed in the genus rangingfrom 53-60 sup1m in males and 59-69 sup1m in females whichis shorter than the stylet length of nematodes on A in-fecta (Fig 3) In addition spicules of V daptria taper to ne points whereas in Vampyronema from other hosts thespicule tips are blunt V daptria on L porcia appears tobe a distinct species and we therefore do not assign any ofthe Vampyronema from this study to it

Nematodes on the St Lucia and Guadeloupe popula-tions of Periplaneta americana are not well differenti-ated from each other based on the suite of characters thatwere examined (Fig 1A B) Although there are a fewsigni cantly different characters between the two popula-tions the differences probably do not represent interspe-ci c variation In addition the characters of both popula-tions are in agreement with the published description ofAcugutturus parasiticus (Hunt 1980)

Examination of additional specimens of ectoparasiticnematodes from other species of Lepidoptera showed thatmost were Vampyronema whereas Noctuidonema wasfound only on moths from two genera Pseudaletia andSpodoptera (Table 7)

Vol 4(4) 2002 499

OG Marti Jr et al

Table 7 Association of quadrine and trine Noctuidae with Vampyronema and Noctuidonema

Host Taxon Hind Wing Venation1 Locality2 Nematodes Genus

NoctuidaeCatocalinae

Anomis oedema Gueneacutee quadri ne FG 4 1 VampyronemaBaniana sp quadri ne FG 1 0 VampyronemaBaniana ostia Druce quadri ne FG 7 0 VampyronemaEuglyphis sp quadri ne FG 1 0 VampyronemaEulepidotis addens (Walker) quadri ne GU 5 2 VampyronemaItomia opisthographa (Gueneacutee) quadri ne FG 5 1 VampyronemaLesmone formularis (Geyer) quadri ne FG 1 0 VampyronemaLesmone porcia (Stoll) quadri ne GU ndash3 Vampyronema3

Melipotis fasciolaris (Huumlbner) quadri ne FG 0 1 VampyronemaMetalectra sp quadri ne FG 4 10 VampyronemaMetria sp quadri ne FG 0 1 VampyronemaMocis dif uens (Gueneacutee) quadri ne FG 1 1 Vampyronema4

Mocis disseverans (Walker) quadri ne GA 9 4 Vampyronema4

Mocis latipes Gueneacutee quadri ne GA 9 16 Vampyronema4

Perasia ora (Cramer)5 quadri ne FG 10 0 VampyronemaSelenisa sp quadri ne FG 4 0 VampyronemaZale ctilis (Gueneacutee) quadri ne FG 4 1 Vampyronema

PlusiinaeAgrapha oxygramma (Geyer) quadri ne GA 13 7 VampyronemaArgyrogramma verruca (Fabricius) quadri ne FG 0 2 VampyronemaChrysodeixis chalcites (Esper) quadri ne Fiji 8 7 Vampyronema

AmphipyrinaeSpodoptera androgea (Stoll) tri ne FG 4 0 Noctuidonema6

Spodoptera dolichos (Fabricius) tri ne GA 5 10 Noctuidonema6

Spodoptera frugiperda (JE Smith) tri ne GA 17 11 Noctuidonema6

Spodoptera latifascia (Walker) tri ne GA 12 6 Noctuidonema6

Spodoptera litura (Fabricius) tri ne Fiji 15 7 Noctuidonema6

Spodoptera ornithogalli (Gueneacutee) tri ne GA 10 10 Noctuidonema6

HadeninaeOrthodes crenulata (Butler) tri ne GA 9 4 VampyronemaPseudaletia unipuncta (Haworth) tri ne GA 18 9 Noctuidonema

NoctuinaeAnicla infecta (Ochsenheimer) tri ne GA 15 12 VampyronemaTandilia rodea (Schaus) tri ne FG 2 3 Vampyronema

PyralidaePyraustinae

Lamprosema moccalis Schaus ndash FG 4 0 VampyronemaSphingidae

MacroglossinaeErinnys obscura (Fabricius) ndash FG 22 2 Vampyronema7

Perigonia lusca Fabricius ndash GU 19 4 Vampyronema7

Xylophanes chiron Drury ndash GU 8 0 Vampyronema7

1) Quadri ne and tri ne hindwing venation applies to Noctuidae only Pyralidae and Sphingidae are included for completeness thestatus of Plusiinae as quadri nes is ambiguous because it groups together mostly species having quadri ne wing venation yet is oftenincluded with chie y tri ne subfamilies (Kitching amp Rawlins 1998) 2 Fiji Fiji Islands FG French Guiana GU Guadeloupe GAGeorgia USA 3 V daptria Anderson amp Laumond 1992 designationbased on Hunt (1993) 4 V dibolia 5 Synonymous with Nymbisarcuata Walker (in Rogers et al 1990a and Simmons amp Rogers 1996) 6 N guyanense 7 Probably represents an undescribed genus

500 Nematology

Speciation in the Acugutturidae

Discussion

The position of the excretory pore posterior to thebulb the comparatively short length of the stylet thecomparatively long tail and the simple rose-thorn-shapedspicules are suf cient to clearly separate Acugutturus aparasite of Blattodea from Vampyronema and Noctuido-nema parasites of Lepidoptera Sampled populations re-veal four inferred monophyleticgroups Acugutturus par-asiticus from two West Indian populations of Periplan-eta americana a new genus on Sphingidae Noctuido-nema guyanense from Spodoptera spp with nematodeson P unipuncta as a probable new species of Noctuido-nema and Vampyronema spp on the remaining hosts inthis study with the nematodes on A infecta representinga new species of Vampyronema

Each inferred monophyletic group can be de ned byseveral unique combinations of characters quantitativeand qualitative that satisfy the minimal requirements forspecies diagnosis (Wheeler 1999 Wheeler amp Platnick2000) Such characters appear to represent lineage inde-pendence via autapomorphies hypotheses ideally testedvia outgroup comparison Such tests while more philo-sophically satisfying (Adams 1998 Wiley amp Mayden2000) are beyond the scope of the present paper

According to Kitching and Rawlins (1998) differencesin hindwing venation broadly divide the Noctuidae intotwo phenetic groups tri nes and quadri nes Catocalinaeand Plusiinae are quadri nes (but see comment 1 in Ta-ble 7) whereas the other noctuid subfamilies mentionedin this study are tri nes Complex spicules probably rep-resent a derived condition found thus far only in nema-todes on two genera of tri ne noctuids Pseudaletia andSpodoptera while the simple spicules of Vampyronemaprobably represent a plesiomorphic state found in bothtri ne and quadri ne noctuids The extremely long styletsof nematodes from Sphingidae may have arisen in re-sponse to parasitisation of large moths

The Acugutturidae are associated with two very dis-tantly related insect orders Blattodea and Lepidoptera(Endopterygota) There is no record of occurrence ofthe Acugutturidae on any other insect order Among theLepidoptera the Acugutturidae and more precisely theNoctuidonematinae are known from only six familiesGeometridae Lasiocampidae Noctuidae NotodontidaePyralidae and Sphingidae (Marti et al 2000) most ofwhich are not closely related with the exceptionof Lasio-campidae and Sphingidae members of the Bombycoidea(and their relatives) clade (Lemaire amp Minet 1998) and

Notodontidaeand Noctuidae members of the Noctuoideasuperfamily (Kitching amp Rawlins 1998)

One explanation for such a surprising distribution ofthe Acugutturidae on their hosts may be that availabledata are too sparse due to insuf cient sampling of otherlepidopteran families While this may be true the studyconducted by Marti et al (2000) shows that there isa clear speci city in nematode host preference Martiet al (2000) updated the list of lepidopteran hosts ofNoctuidonematinae It seems clear that although manyLepidoptera species belonging to 11 families and 115genera have been examined for nematodes only membersof the six aforementioned obtectomeran (Kristensen ampSkalski 1998) families have been found to be positivefor Noctuidonematinae With the exception of Pyralidaeall these families are members of the Macrolepidopteraclade Moreover the positive Noctuidae species representnearly 90 of the total positive species found Thereforewe cannot state that the Noctuidonematinae are notpotentially parasites of most of the lepidopteran familiesbut we can state that at least in the Neotropical Regionthey parasitise the Macrolepidoptera preferentially andamong that clade mostly a single family the Noctuidae(Marti et al 2000)

The Aphelenchida include several different trophictypes with many species being mycetophagous or phy-toparasitic with about one third of known species be-ing entomoparasitic (Hunt 1993) Considering the list ofinsect hosts we cannot hypothesise that the Acugutturi-dae are derived from a single ancestor transferring froma plant to an insect host It seems more realistic to pro-pose several transfers from plant to insect hosts during thehistory of the Acugutturidae A transfer (Acugutturinae)occurred on the Blattodea and another one (Noctuidone-matinae) on the Lepidoptera Another transfer may haveoccurred on the Bombycoidea sensu lato If at least twosuch principal transfers occurred then it is possible thatthe Acugutturinae and Noctuidonematinae would not beconfamilial and that these two taxonomicunits deserve fa-milial rank Cladistic and molecularphylogeneticanalysiswill help to solve that problem The existenceof symbioticbacteria in these nematodes (Marti et al 1995) shouldprovide an independentmeans of analysing these relation-ships

Concerning the relationship between Lepidoptera andits ectoparasitic nematodes the present work highlightsthe following situation a new genus of Noctuidone-matinae parasitises Sphingidae a genuine Bombycoidea(Lemaire amp Minet 1998) Vampyronema parasitises the

Vol 4(4) 2002 501

OG Marti Jr et al

quadri ne noctuid subfamily Catocalinae (sensu lato in-cluding Ophiderinae) Plusiinae an Hadeninae and aNoctuinae (both typical tri ne noctuids) and Noctui-donema spp parasitises another Hadeninae and severalSpodoptera a tri ne noctuid genus previously includedin the Amphipyrinae but now of uncertain subfamilial af- liation (Poole 1995 Kitching amp Rawlins 1998) Exceptfor nematodes on Sphingidae which likely correspond toa well-separated clade whose phylogenetic relationshipto other Noctuidonematinae needs to be speci ed it ap-pears that Vampyronema and Noctuidonema both para-sitise Noctuidae but that Vampyronema parasitises bothquadri ne and tri ne noctuids Noctuidonema possibly aderived form of Noctuidonematinae is restricted to a fewgenera (Spodoptera and Pseudaletia) of tri ne noctuidsthat are not closely related (Mitchell et al 2000)

Recent work has questioned the monophylyof the Noc-tuidae and there is a growing feeling among moleculartaxonomists that the Noctuidae may be paraphyletic sev-eral quadri ne subfamilies including Catocalinae beingmore closely related to Arctiidae and Lymantriidae than totri ne noctuids (Weller et al 1994 Mitchell et al 1997Fang et al 2000 Mitchell et al 2000) The latter groupis considered as monophyleticClearly systematics of theNoctuidonematinaedo not match the tentative new classi- cation of the Noctuidae

This lack of congruence between parasite and hostsystematics leads us to consider that the relationshipsbetween Acugutturidae and their hosts have not beenguided by coevolutionary forces in the sense of Ehrlichand Raven (1964) but by horizontal transfers (Rogers ampMarti 1996) and sequential evolution (Jermy 1984 Fu-tuyma amp McCafferty 1990) The Acugutturidae and es-pecially the Noctuidonematinae have likely experiencedan adaptive radiation on higher Lepidoptera long aftertheir host diversi cation History of the Noctuidonemati-nae has been characterised by host transfer events ontofairly distantly related hosts followed by adaptation tonew hosts leading to speciation processes that now permitus to distinguish easily fairly species-speci c populationsor strains of Vampyronema or Noctuidonema In mothsthese host transfers may have occurred through contami-nation of nematode-free species by nematode eggs juve-niles or adults left on resting or feeding surfaces (leavesor owers) by individualsof infested species The successof such contamination may be linked to the availabilityand accessibility of host intersegmental membranes thinenough to be pierced by nematode stylets a factor that

may have physically and dramatically restricted the hostrange of all Acugutturidae

The present morphological analysis of acugutturid ne-matodes parasitising 13 species of Lepidoptera and onespecies of cockroach revealed a greater degree of diversitythan previously known in this group The genus Vampy-ronema grouping nematodes with simple spicules in-cludes two inferred new species in addition to V dibo-lia and V daptria already known from the LepidopteraThe genus Noctuidonema grouping nematodes of Lepi-doptera with complex spicules includes a new speciesand the nematodes found on Sphingidae probably belongto a new genus The relationshipbetween these nematodesand their hosts appears to be much more complex thanpreviously thought In addition to their bene ts as para-sites of pest Lepidoptera acugutturid nematodes are aninteresting model for parasite-host coevolutionarystudies

Acknowledgements

We thank David Hunt (CABI Bioscience UK) DanielLachaise (CNRS France) Alvin Simmons (USDA Char-leston SC USA) and S Patricia Stock (University ofCalifornia Davis CA USA) for reviewing earlier ver-sions of this paper We thank Jeanette Williams and JohnClayton for collecting cockroaches in St Lucia and mothsin Fiji respectively We thank the following individualsfor identifyingor con rming our identi cations of insectsDavid Nickle (Smithsonian Washington DC USA) forPeriplaneta americana Pierre Zagatti (INRA France)for Pyralidae and Hugo Kons (University of FloridaGainesville FL USA) for Orthodes crenulata

References

ADAMS BJ (1998) Species concepts and evolutionary para-digm in modern nematology Journal of Nematology 30 1-21

ANDERSON RV amp LAUMOND C (1990) NoctuidonemaguyanenseRemillet amp Silvain 1988 morphologie du spiculeredescription du macircle et diagnose amendeacutee du genre Noctui-donema Revue de Neacutematologie 13 433-436

ANDERSON RV amp LAUMOND C (1992) Noctuidonemadaptria n sp (Nematoda Aphelenchoididae)an ectoparasiteof the moth Lesmone porcia (Stoll) Journal of Nematology24 16-22

EHRLICH P amp RAVEN P (1964) Butter ies and plants astudy in evolution Evolution 18 586-608

FANG QQ MITCHELL A REGIER JC MITTER CFRIEDLANDER TP amp POOLE RW (2000) Phylogenetic

502 Nematology

Speciation in the Acugutturidae

Table 3 Summary of forward stepwise discriminant analysis for male and female nematode morphological characters from 11 speciesof host Noctuidae (Lepidoptera)

Males ndash Lepidoptera groups1 (n D 11) Females ndash Lepidoptera groups1 (n D 11)

Step Character F (df1 df2) Character F (df1 df2)

1 Spicule length 24039 (10 80) V 10986 (10 150)2 Stylet length 2510 (10 79) Stylet length 6194 (10 149)3 Stylet of body length 1388 (10 78) Bulb width 1830 (10 148)4 Cloaca to anterior end 981 (10 77) Excretory pore of body 982 (10 147)5 Rostrum length 762 (10 76) Vulva to anterior end 935 (10 146)6 Ant end to ant gonad 700 (10 75) Body width at nerve ring 873 (10 145)7 G 795 (10 74) Stylet of body length 777 (10 144)8 c0 651 (10 73) Tail length 701 (10 143)9 Body width at nerve ring 615 (10 72) Head width 672 (10 142)

10 Bulb width 548 (10 71) Gonad of body length 595 (10 141)

1) Not including Perigonia lusca Xylophanes chiron or two populations from Periplaneta AmericanaP lt 0001

Fig 2 Plot of discriminant functions for nematode morphometric characters Ellipses indicate 95 condence limits for selected hostspecies A Female nematodes (Vampyronema) from six host species (Lepidoptera) B Male nematodes (Vampyronema) from six hostspecies (Lepidoptera)C Female nematodes (Noctuidonema) from ve host species (Lepidoptera)D Male nematodes (Noctuidonema)from ve host species (Lepidoptera) Legend (Host names) ao Agrapha oxygramma ai Anicla infecta ea Eulepidotis addens mdMocis disseverans ml M latipes oc Orthodes crenulata pu Pseudaletia unipuncta sd Spodoptera dolichos sf S frugiperda sl Slatifascia so S ornithogalli

Vol 4(4) 2002 495

OG Marti Jr et al

Table 4 Summary of forward stepwise discriminant analysis of morphological characters from male and female Vampyronema fromsix species of Noctuidae (Lepidoptera)

Males ndash Mocis group1 (n D 6) Females ndash Mocis group1 (n D 6)

Step Character F (df1 df2) Character F (df1 df2)

1 Stylet length 5807 (5 39) Stylet length 4573 (5 93)2 Rostrum length 1219 (5 38) Ant end to ant intestine 1879 (5 92)3 Bulb length 929 (5 37) Excretory pore of body 1803 (5 91)4 Ant end to ant gonad 837 (5 36) G 1500 (5 90)6 Pharyngeal gland width 511 (5 34) Stylet length of body 573 (5 89)7 Gonad length 438 (5 33) V 485 (5 88)8 Tail length 408 (5 32) Head length 476 (5 87)9 Bulb width 269 (5 31) Gonad length of body 445 (5 86)

10 Body width at nerve ring 355 (5 85)

1) Mocis group Agrapha oxygramma Anicla infecta Eulepidotis addens Mocis disseveransM latipes Orthodes crenulataP lt 005 P lt 001 P lt 0001

STEP 3 SIX HOST GROUPS (VAMPYRONEMANEMATODES WITH SIMPLE SPICULES)

The noctuidmoths Anicla infectaAgraphaoxygrammaEulepidotis addens Mocis disseverans M latipes andOrthodes crenulata are parasitised by Vampyronema inwhich the structure of the spicules is simpler and the tailof the females is longer than in Noctuidonema Analy-sis of morphometric characters of female nematodes fromthese hosts revealed ve signi cant canonical roots cu-mulatively accounting for 100 of the explainedvariancebetween groups The rst three roots representingprimar-ily stylet length distance from the anterior end to the baseof the bulb and the c ratio accounted cumulatively for9188 of the explained variance between groups Styletlength was the most important morphological characterdistinguishing female Vampyronema from different hosts(Fig 2A Table 4)

Analysis of results of male and female Vampyronemaon noctuid moths demonstrate the existence of fourdistinct clades One of these is V dibolia the speciesdescribed from M latipes The nematodes particularlythe males on M disseverans are very similar to V diboliaand we consider them to be conspeci c A second cladeis formed by nematodes from A infecta which are wellseparated from all the other groups (Fig 2A B) Styletlength is signi cantly shorter (P lt 001) in nematodeson A infecta than on those from the other hosts Athird Vampyronema clade is formed by nematodes onE addens which differ from V dibolia primarily in thedistance from the anterior end to the base of the bulb Thetwo male nematodes from E addens were also distinct

from other males an observation in agreement withresults from the females A fourth Vampyronema cladeis formed by nematodes from A oxygramma (Fig 2AB) Although the females appear to be intermediatebetween those on M latipes and E addens the malesare distinct based on differences in stylet and rostrumlength Nematodes on O crenulata are also intermediatebut males are very similar to V dibolia and we considerthem to be conspeci c

In the males there were ve signi cant canonical rootsrepresented primarily by stylet length spicule rostrumlength and bulb length respectively These accounted cu-mulatively for 929 of the explained variation betweenmales from these six host species Stylet length was themost important morphologicalcharacter in de ning males(Fig 2B) although rostrum length was also identi ed asan important character

STEP 4 FIVE HOST SPECIES (NOCTUIDONEMANEMATODES WITH COMPLEX SPICULES)

Discriminant analysis of female nematodes from fourSpodoptera species and Pseudaletia revealed four signif-icant (P lt 005) canonical roots The rst two rep-resenting primarily stylet length and body width at thevulva respectively accounted for a cumulative propor-tion of 8793 of the explained variance between groupsFemale nematodes from Pseudaletia were well separatedfrom those parasitising Spodoptera spp on the basis ofthe shorter stylet Nematodes from the four species ofSpodoptera were poorly de ned and showed a consider-able amount of overlap Even the groups that appeared to

496 Nematology

Speciation in the Acugutturidae

Table 5 Summary of forward stepwise discriminant analysis of morphological characters from male and female Noctuidonema from ve species of Spodoptera and Pseudaletia

Males ndash Spodoptera and Pseudaletia1 (n D 5) Females ndash Spodoptera and Pseudaletia1 (n D 5)

Step Character F (df1 df2) Character F (df1 df2)

1 Ant end to ant intestine 2524 (4 41) Stylet length 3914 (4 57)2 Body length 1491 (4 40) Body width at vulva 1250 (4 56)3 c0 1223 (4 39) Bulb width 591 (4 55)4 Bulb width 725 (4 38) a 498 (4 54)5 Body width at nerve ring 500 (4 37) Vulva to anterior end 380 (4 53)6 Spicule length 409 (4 36) Body width at nerve ring 579 (4 52)7 Maximum body width 265 (4 51)

1) Pseudaletia unipuncta Spodoptera dolichos S frugiperda S latifascia S ornithogalliP lt 005 P lt 001 P lt 0001

be the most different S frugiperda and S dolichos wereextensively overlapped (Fig 2D Table 5)

In the males there were three signi cant canonicalroots accounting for 9764 of the explained variancebetween groups The rst root represents the distancefrom the anterior end to the anterior intestine and V aboutequally whereas the second root represents primarily thebody width at the nerve ring (Fig 2D Table 5) Malenematodes from P unipuncta were well separated fromnematodes from Spodoptera which agrees with resultsfrom the females Males on S dolichosshowed a moderateamount of separation from males on the other threespecies of Spodoptera

DATA ANALYSIS

Eighteen morphological characters contribute to differ-entiation among male nematodes on 14 host species in-cluded in this study The most signi cant and useful ofthese are the distance of the excretory pore from the an-terior end spicule length and stylet length These threecharacters and V are the most useful in differentiating be-tween females (Table 2) Furthermore ANOVA showedthat the stylet length of nematodes on the two speciesof Sphingidae is signi cantly greater (P lt 0001) thanthe stylet length of nematodes on any of the other hostspecies The two populationsof nematodes from Sphingi-dae differ signi cantly in only eight morphological char-acters whereas each differs from V dibolia in about 20characters Females on P lusca differ from those on Xchiron in the longer length of the stylet greater length andwidth of the bulb greater distance of the excretory poreand bulb base from the anterior end greater body widthat the vulva and greater excretory pore distance as a per-

centage of body length However in the absence of malesfrom X chiron no conclusion can be reached on whethernematodes on Sphingidae are conspeci c The consistentseparation of nematodes parasitising Sphingidae from allother Acugutturidae suggests that they probably representat least one new genus Particularly remarkable is the longstylet which represents about 37 of the body length andaverages about 190 sup1m in length (Fig 3)

Discriminant analysis shows that the ve groups ofnematodes with complex spicules (Noctuidonema) arewell separated (Fig 1C D) from the six groups withsimple spicules (Vampyronema) Although the discretecharacter of spicule structure is suf cient to separate thesetwo genera particularly good separation can also be madebased on important continuously variable characters suchas stylet length distance of the excretory pore to theanterior end and the V ratio Furthermore within eachof these two genera separate groups can be detectedbased primarily on differences in stylet length Thusthe nematodes on Pseudaletia unipuncta differ fromthe nematodes on the four Spodoptera spp (Fig 1CD) and those on Anicla infecta Agrapha oxygrammaand Eulepidotis addens differ from V dibolia and itsconspeci c on M disseverans (Fig 2A B)

Female nematodes on P unipuncta differ (P lt 005)from those on S dolichos S frugiperda S latifasciaand S ornithogalli in eight 18 14 and 14 morpholo-gical characters respectivelyThese results providestrongsupport for separate taxonomic status of nematodes onP unipuncta as a new species of Noctuidonema Resultsfrom simple scatterplots prepared for each gender usingthe most importantmorphologicalcharacters identi ed bydiscriminant analysis agree with the preceding analysis

Vol 4(4) 2002 497

OG Marti Jr et al

Fig 3 Plot of stylet length of ectoparasitic nematodes from insect hosts (Blattodea and Lepidoptera) A Female nematodes from15 host species B Male nematodes from 14 host species Legend (Host names) ao Agrapha oxygramma ai Anicla infecta eaEulepidotis addens md Mocis disseverans ml M latipes oc Orthodes crenulata pa Periplaneta americana (St Lucia) pa2 Pamericana (Guadeloupe) pl Perigonia lusca pu Pseudaletia unipuncta sd Spodoptera dolichos sf S frugiperda sl S latifascia soS ornithogalli xc Xylophanes chiron

Table 6 Percent correct classication of male nematodes based on posterior probabilities that a case belongs to a particular hostspecies

Host species Percent correct classi cations ndash Percent correct classi cations ndashmale nematodes female nematodes

n Hosts Hosts Hosts Hosts n Hosts Hosts Hosts Hosts1-15 1-11 1-6 7-11 1-15 1-11 1-6 7-11

1 Agrapha oxygramma 7 100 100 100 ndash 12 917 100 917 ndash2 Anicla infecta 12 100 100 100 ndash 15 100 100 100 ndash3 Eulepidotis addens 2 100 100 100 ndash 15 100 100 100 ndash4 Mocis disseverans 4 100 75 75 ndash 9 889 889 889 ndash5 Mocis latipes 16 100 100 100 ndash 39 923 974 948 ndash6 Orthodes crenulata 4 100 100 100 ndash 9 889 889 889 ndash7 Spodoptera dolichos 10 100 100 ndash 90 5 100 100 ndash 808 Spodoptera frugiperda 11 909 909 ndash 100 17 765 824 ndash 7659 Spodoptera latifascia 6 667 833 ndash 50 12 75 833 ndash 667

10 Spodoptera ornithogalli 10 100 100 ndash 100 10 90 90 ndash 9011 Pseudaletia unipuncta 9 100 100 ndash 100 18 100 944 ndash 94412 Perigonia lusca 4 100 ndash ndash ndash 19 938 ndash ndash ndash13 Periplaneta americana (1) 7 100 ndash ndash ndash 18 882 ndash ndash ndash14 Periplaneta americana (2) 5 100 ndash ndash ndash 13 846 ndash ndash ndash15 Xylophanes chiron 0 ndash ndash ndash ndash 18 100 ndash ndash ndash

(1) St Lucia(2) Guadeloupe

498 Nematology

Speciation in the Acugutturidae

Fig 4 Tree diagram for ectoparasitic nematodes from insecthost species (Blattodea and Lepidoptera) The tree was pro-duced using Euclidean distance as the distance measure andcomplete linkage as the amalgamation rule Both graphs areplotted on an X-axis standardised by (DlinkDmax)100 whichrepresents the percentage of the range from the maximum tothe minimum distance in the data A Female nematodes datafor excretory pore stylet length V bulb width a anterior endto anterior intestine and vulva to anterior end were used BMale nematodes data for excretory pore spicule length andstylet length were used Legend (Host Names) pa Periplan-eta americana (St Lucia) pa2 P americana (Guadeloupe) aoAgrapha oxygramma ai Anicla infecta ea Eulepidotis addensmd Mocis disseverans ml M latipes oc Orthodes crenulatapl Perigonia lusca pu Pseudaletia unipuncta sd Spodopteradolichos sf S frugiperda sl S latifascia so S ornithogallixc Xylophanes chiron

Spicule length was not an important character in dis-tinguishing among male Noctuidonema Male nematodeson S dolichos and S frugiperda showed the most sepa-ration with moderate overlap whereas males on S or-nithogalli appeared to be most similar to those on S

dolichos and males on S latifascia appeared most similarto those on S frugiperda Results from both genders agreethat nematodes on Pseudaletia are different from thoseon Spodoptera and that the nematodes on S dolichos andS frugiperda are the most different with the nematodeson S latifascia and S ornithogallibeing intermediate be-tween these two Overall the results do not support theexistence of separate species of Noctuidonema on hostsof the genus Spodoptera (Fig 4 Table 6)

These results provide evidence that nematodes on Ainfecta A oxygramma and E addens are different fromeach other and from the remaining groups having simplespicules Nematodes on M disseverans and O crenulatawere not distinct from V dibolia the species parasitisingM latipes Nematodes on M disseverans differed signif-icantly from V dibolia in only six morphological charac-ters in the females and ten in the males The differenceswere small and can be attributed to adaptation of conspe-ci c nematodes to parasitism of closely related sympatrichost species

Comparisonof the description of V daptria to measure-ments of other Vampyronema in this study indicates that itdiffers in stylet length and spicule shape V daptria hasthe shortest stylet length observed in the genus rangingfrom 53-60 sup1m in males and 59-69 sup1m in females whichis shorter than the stylet length of nematodes on A in-fecta (Fig 3) In addition spicules of V daptria taper to ne points whereas in Vampyronema from other hosts thespicule tips are blunt V daptria on L porcia appears tobe a distinct species and we therefore do not assign any ofthe Vampyronema from this study to it

Nematodes on the St Lucia and Guadeloupe popula-tions of Periplaneta americana are not well differenti-ated from each other based on the suite of characters thatwere examined (Fig 1A B) Although there are a fewsigni cantly different characters between the two popula-tions the differences probably do not represent interspe-ci c variation In addition the characters of both popula-tions are in agreement with the published description ofAcugutturus parasiticus (Hunt 1980)

Examination of additional specimens of ectoparasiticnematodes from other species of Lepidoptera showed thatmost were Vampyronema whereas Noctuidonema wasfound only on moths from two genera Pseudaletia andSpodoptera (Table 7)

Vol 4(4) 2002 499

OG Marti Jr et al

Table 7 Association of quadrine and trine Noctuidae with Vampyronema and Noctuidonema

Host Taxon Hind Wing Venation1 Locality2 Nematodes Genus

NoctuidaeCatocalinae

Anomis oedema Gueneacutee quadri ne FG 4 1 VampyronemaBaniana sp quadri ne FG 1 0 VampyronemaBaniana ostia Druce quadri ne FG 7 0 VampyronemaEuglyphis sp quadri ne FG 1 0 VampyronemaEulepidotis addens (Walker) quadri ne GU 5 2 VampyronemaItomia opisthographa (Gueneacutee) quadri ne FG 5 1 VampyronemaLesmone formularis (Geyer) quadri ne FG 1 0 VampyronemaLesmone porcia (Stoll) quadri ne GU ndash3 Vampyronema3

Melipotis fasciolaris (Huumlbner) quadri ne FG 0 1 VampyronemaMetalectra sp quadri ne FG 4 10 VampyronemaMetria sp quadri ne FG 0 1 VampyronemaMocis dif uens (Gueneacutee) quadri ne FG 1 1 Vampyronema4

Mocis disseverans (Walker) quadri ne GA 9 4 Vampyronema4

Mocis latipes Gueneacutee quadri ne GA 9 16 Vampyronema4

Perasia ora (Cramer)5 quadri ne FG 10 0 VampyronemaSelenisa sp quadri ne FG 4 0 VampyronemaZale ctilis (Gueneacutee) quadri ne FG 4 1 Vampyronema

PlusiinaeAgrapha oxygramma (Geyer) quadri ne GA 13 7 VampyronemaArgyrogramma verruca (Fabricius) quadri ne FG 0 2 VampyronemaChrysodeixis chalcites (Esper) quadri ne Fiji 8 7 Vampyronema

AmphipyrinaeSpodoptera androgea (Stoll) tri ne FG 4 0 Noctuidonema6

Spodoptera dolichos (Fabricius) tri ne GA 5 10 Noctuidonema6

Spodoptera frugiperda (JE Smith) tri ne GA 17 11 Noctuidonema6

Spodoptera latifascia (Walker) tri ne GA 12 6 Noctuidonema6

Spodoptera litura (Fabricius) tri ne Fiji 15 7 Noctuidonema6

Spodoptera ornithogalli (Gueneacutee) tri ne GA 10 10 Noctuidonema6

HadeninaeOrthodes crenulata (Butler) tri ne GA 9 4 VampyronemaPseudaletia unipuncta (Haworth) tri ne GA 18 9 Noctuidonema

NoctuinaeAnicla infecta (Ochsenheimer) tri ne GA 15 12 VampyronemaTandilia rodea (Schaus) tri ne FG 2 3 Vampyronema

PyralidaePyraustinae

Lamprosema moccalis Schaus ndash FG 4 0 VampyronemaSphingidae

MacroglossinaeErinnys obscura (Fabricius) ndash FG 22 2 Vampyronema7

Perigonia lusca Fabricius ndash GU 19 4 Vampyronema7

Xylophanes chiron Drury ndash GU 8 0 Vampyronema7

1) Quadri ne and tri ne hindwing venation applies to Noctuidae only Pyralidae and Sphingidae are included for completeness thestatus of Plusiinae as quadri nes is ambiguous because it groups together mostly species having quadri ne wing venation yet is oftenincluded with chie y tri ne subfamilies (Kitching amp Rawlins 1998) 2 Fiji Fiji Islands FG French Guiana GU Guadeloupe GAGeorgia USA 3 V daptria Anderson amp Laumond 1992 designationbased on Hunt (1993) 4 V dibolia 5 Synonymous with Nymbisarcuata Walker (in Rogers et al 1990a and Simmons amp Rogers 1996) 6 N guyanense 7 Probably represents an undescribed genus

500 Nematology

Speciation in the Acugutturidae

Discussion

The position of the excretory pore posterior to thebulb the comparatively short length of the stylet thecomparatively long tail and the simple rose-thorn-shapedspicules are suf cient to clearly separate Acugutturus aparasite of Blattodea from Vampyronema and Noctuido-nema parasites of Lepidoptera Sampled populations re-veal four inferred monophyleticgroups Acugutturus par-asiticus from two West Indian populations of Periplan-eta americana a new genus on Sphingidae Noctuido-nema guyanense from Spodoptera spp with nematodeson P unipuncta as a probable new species of Noctuido-nema and Vampyronema spp on the remaining hosts inthis study with the nematodes on A infecta representinga new species of Vampyronema

Each inferred monophyletic group can be de ned byseveral unique combinations of characters quantitativeand qualitative that satisfy the minimal requirements forspecies diagnosis (Wheeler 1999 Wheeler amp Platnick2000) Such characters appear to represent lineage inde-pendence via autapomorphies hypotheses ideally testedvia outgroup comparison Such tests while more philo-sophically satisfying (Adams 1998 Wiley amp Mayden2000) are beyond the scope of the present paper

According to Kitching and Rawlins (1998) differencesin hindwing venation broadly divide the Noctuidae intotwo phenetic groups tri nes and quadri nes Catocalinaeand Plusiinae are quadri nes (but see comment 1 in Ta-ble 7) whereas the other noctuid subfamilies mentionedin this study are tri nes Complex spicules probably rep-resent a derived condition found thus far only in nema-todes on two genera of tri ne noctuids Pseudaletia andSpodoptera while the simple spicules of Vampyronemaprobably represent a plesiomorphic state found in bothtri ne and quadri ne noctuids The extremely long styletsof nematodes from Sphingidae may have arisen in re-sponse to parasitisation of large moths

The Acugutturidae are associated with two very dis-tantly related insect orders Blattodea and Lepidoptera(Endopterygota) There is no record of occurrence ofthe Acugutturidae on any other insect order Among theLepidoptera the Acugutturidae and more precisely theNoctuidonematinae are known from only six familiesGeometridae Lasiocampidae Noctuidae NotodontidaePyralidae and Sphingidae (Marti et al 2000) most ofwhich are not closely related with the exceptionof Lasio-campidae and Sphingidae members of the Bombycoidea(and their relatives) clade (Lemaire amp Minet 1998) and

Notodontidaeand Noctuidae members of the Noctuoideasuperfamily (Kitching amp Rawlins 1998)

One explanation for such a surprising distribution ofthe Acugutturidae on their hosts may be that availabledata are too sparse due to insuf cient sampling of otherlepidopteran families While this may be true the studyconducted by Marti et al (2000) shows that there isa clear speci city in nematode host preference Martiet al (2000) updated the list of lepidopteran hosts ofNoctuidonematinae It seems clear that although manyLepidoptera species belonging to 11 families and 115genera have been examined for nematodes only membersof the six aforementioned obtectomeran (Kristensen ampSkalski 1998) families have been found to be positivefor Noctuidonematinae With the exception of Pyralidaeall these families are members of the Macrolepidopteraclade Moreover the positive Noctuidae species representnearly 90 of the total positive species found Thereforewe cannot state that the Noctuidonematinae are notpotentially parasites of most of the lepidopteran familiesbut we can state that at least in the Neotropical Regionthey parasitise the Macrolepidoptera preferentially andamong that clade mostly a single family the Noctuidae(Marti et al 2000)

The Aphelenchida include several different trophictypes with many species being mycetophagous or phy-toparasitic with about one third of known species be-ing entomoparasitic (Hunt 1993) Considering the list ofinsect hosts we cannot hypothesise that the Acugutturi-dae are derived from a single ancestor transferring froma plant to an insect host It seems more realistic to pro-pose several transfers from plant to insect hosts during thehistory of the Acugutturidae A transfer (Acugutturinae)occurred on the Blattodea and another one (Noctuidone-matinae) on the Lepidoptera Another transfer may haveoccurred on the Bombycoidea sensu lato If at least twosuch principal transfers occurred then it is possible thatthe Acugutturinae and Noctuidonematinae would not beconfamilial and that these two taxonomicunits deserve fa-milial rank Cladistic and molecularphylogeneticanalysiswill help to solve that problem The existenceof symbioticbacteria in these nematodes (Marti et al 1995) shouldprovide an independentmeans of analysing these relation-ships

Concerning the relationship between Lepidoptera andits ectoparasitic nematodes the present work highlightsthe following situation a new genus of Noctuidone-matinae parasitises Sphingidae a genuine Bombycoidea(Lemaire amp Minet 1998) Vampyronema parasitises the

Vol 4(4) 2002 501

OG Marti Jr et al

quadri ne noctuid subfamily Catocalinae (sensu lato in-cluding Ophiderinae) Plusiinae an Hadeninae and aNoctuinae (both typical tri ne noctuids) and Noctui-donema spp parasitises another Hadeninae and severalSpodoptera a tri ne noctuid genus previously includedin the Amphipyrinae but now of uncertain subfamilial af- liation (Poole 1995 Kitching amp Rawlins 1998) Exceptfor nematodes on Sphingidae which likely correspond toa well-separated clade whose phylogenetic relationshipto other Noctuidonematinae needs to be speci ed it ap-pears that Vampyronema and Noctuidonema both para-sitise Noctuidae but that Vampyronema parasitises bothquadri ne and tri ne noctuids Noctuidonema possibly aderived form of Noctuidonematinae is restricted to a fewgenera (Spodoptera and Pseudaletia) of tri ne noctuidsthat are not closely related (Mitchell et al 2000)

Recent work has questioned the monophylyof the Noc-tuidae and there is a growing feeling among moleculartaxonomists that the Noctuidae may be paraphyletic sev-eral quadri ne subfamilies including Catocalinae beingmore closely related to Arctiidae and Lymantriidae than totri ne noctuids (Weller et al 1994 Mitchell et al 1997Fang et al 2000 Mitchell et al 2000) The latter groupis considered as monophyleticClearly systematics of theNoctuidonematinaedo not match the tentative new classi- cation of the Noctuidae

This lack of congruence between parasite and hostsystematics leads us to consider that the relationshipsbetween Acugutturidae and their hosts have not beenguided by coevolutionary forces in the sense of Ehrlichand Raven (1964) but by horizontal transfers (Rogers ampMarti 1996) and sequential evolution (Jermy 1984 Fu-tuyma amp McCafferty 1990) The Acugutturidae and es-pecially the Noctuidonematinae have likely experiencedan adaptive radiation on higher Lepidoptera long aftertheir host diversi cation History of the Noctuidonemati-nae has been characterised by host transfer events ontofairly distantly related hosts followed by adaptation tonew hosts leading to speciation processes that now permitus to distinguish easily fairly species-speci c populationsor strains of Vampyronema or Noctuidonema In mothsthese host transfers may have occurred through contami-nation of nematode-free species by nematode eggs juve-niles or adults left on resting or feeding surfaces (leavesor owers) by individualsof infested species The successof such contamination may be linked to the availabilityand accessibility of host intersegmental membranes thinenough to be pierced by nematode stylets a factor that

may have physically and dramatically restricted the hostrange of all Acugutturidae

The present morphological analysis of acugutturid ne-matodes parasitising 13 species of Lepidoptera and onespecies of cockroach revealed a greater degree of diversitythan previously known in this group The genus Vampy-ronema grouping nematodes with simple spicules in-cludes two inferred new species in addition to V dibo-lia and V daptria already known from the LepidopteraThe genus Noctuidonema grouping nematodes of Lepi-doptera with complex spicules includes a new speciesand the nematodes found on Sphingidae probably belongto a new genus The relationshipbetween these nematodesand their hosts appears to be much more complex thanpreviously thought In addition to their bene ts as para-sites of pest Lepidoptera acugutturid nematodes are aninteresting model for parasite-host coevolutionarystudies

Acknowledgements

We thank David Hunt (CABI Bioscience UK) DanielLachaise (CNRS France) Alvin Simmons (USDA Char-leston SC USA) and S Patricia Stock (University ofCalifornia Davis CA USA) for reviewing earlier ver-sions of this paper We thank Jeanette Williams and JohnClayton for collecting cockroaches in St Lucia and mothsin Fiji respectively We thank the following individualsfor identifyingor con rming our identi cations of insectsDavid Nickle (Smithsonian Washington DC USA) forPeriplaneta americana Pierre Zagatti (INRA France)for Pyralidae and Hugo Kons (University of FloridaGainesville FL USA) for Orthodes crenulata

References

ADAMS BJ (1998) Species concepts and evolutionary para-digm in modern nematology Journal of Nematology 30 1-21

ANDERSON RV amp LAUMOND C (1990) NoctuidonemaguyanenseRemillet amp Silvain 1988 morphologie du spiculeredescription du macircle et diagnose amendeacutee du genre Noctui-donema Revue de Neacutematologie 13 433-436

ANDERSON RV amp LAUMOND C (1992) Noctuidonemadaptria n sp (Nematoda Aphelenchoididae)an ectoparasiteof the moth Lesmone porcia (Stoll) Journal of Nematology24 16-22

EHRLICH P amp RAVEN P (1964) Butter ies and plants astudy in evolution Evolution 18 586-608

FANG QQ MITCHELL A REGIER JC MITTER CFRIEDLANDER TP amp POOLE RW (2000) Phylogenetic

502 Nematology

OG Marti Jr et al

Table 4 Summary of forward stepwise discriminant analysis of morphological characters from male and female Vampyronema fromsix species of Noctuidae (Lepidoptera)

Males ndash Mocis group1 (n D 6) Females ndash Mocis group1 (n D 6)

Step Character F (df1 df2) Character F (df1 df2)

1 Stylet length 5807 (5 39) Stylet length 4573 (5 93)2 Rostrum length 1219 (5 38) Ant end to ant intestine 1879 (5 92)3 Bulb length 929 (5 37) Excretory pore of body 1803 (5 91)4 Ant end to ant gonad 837 (5 36) G 1500 (5 90)6 Pharyngeal gland width 511 (5 34) Stylet length of body 573 (5 89)7 Gonad length 438 (5 33) V 485 (5 88)8 Tail length 408 (5 32) Head length 476 (5 87)9 Bulb width 269 (5 31) Gonad length of body 445 (5 86)

10 Body width at nerve ring 355 (5 85)

1) Mocis group Agrapha oxygramma Anicla infecta Eulepidotis addens Mocis disseveransM latipes Orthodes crenulataP lt 005 P lt 001 P lt 0001

STEP 3 SIX HOST GROUPS (VAMPYRONEMANEMATODES WITH SIMPLE SPICULES)

The noctuidmoths Anicla infectaAgraphaoxygrammaEulepidotis addens Mocis disseverans M latipes andOrthodes crenulata are parasitised by Vampyronema inwhich the structure of the spicules is simpler and the tailof the females is longer than in Noctuidonema Analy-sis of morphometric characters of female nematodes fromthese hosts revealed ve signi cant canonical roots cu-mulatively accounting for 100 of the explainedvariancebetween groups The rst three roots representingprimar-ily stylet length distance from the anterior end to the baseof the bulb and the c ratio accounted cumulatively for9188 of the explained variance between groups Styletlength was the most important morphological characterdistinguishing female Vampyronema from different hosts(Fig 2A Table 4)

Analysis of results of male and female Vampyronemaon noctuid moths demonstrate the existence of fourdistinct clades One of these is V dibolia the speciesdescribed from M latipes The nematodes particularlythe males on M disseverans are very similar to V diboliaand we consider them to be conspeci c A second cladeis formed by nematodes from A infecta which are wellseparated from all the other groups (Fig 2A B) Styletlength is signi cantly shorter (P lt 001) in nematodeson A infecta than on those from the other hosts Athird Vampyronema clade is formed by nematodes onE addens which differ from V dibolia primarily in thedistance from the anterior end to the base of the bulb Thetwo male nematodes from E addens were also distinct

from other males an observation in agreement withresults from the females A fourth Vampyronema cladeis formed by nematodes from A oxygramma (Fig 2AB) Although the females appear to be intermediatebetween those on M latipes and E addens the malesare distinct based on differences in stylet and rostrumlength Nematodes on O crenulata are also intermediatebut males are very similar to V dibolia and we considerthem to be conspeci c

In the males there were ve signi cant canonical rootsrepresented primarily by stylet length spicule rostrumlength and bulb length respectively These accounted cu-mulatively for 929 of the explained variation betweenmales from these six host species Stylet length was themost important morphologicalcharacter in de ning males(Fig 2B) although rostrum length was also identi ed asan important character

STEP 4 FIVE HOST SPECIES (NOCTUIDONEMANEMATODES WITH COMPLEX SPICULES)

Discriminant analysis of female nematodes from fourSpodoptera species and Pseudaletia revealed four signif-icant (P lt 005) canonical roots The rst two rep-resenting primarily stylet length and body width at thevulva respectively accounted for a cumulative propor-tion of 8793 of the explained variance between groupsFemale nematodes from Pseudaletia were well separatedfrom those parasitising Spodoptera spp on the basis ofthe shorter stylet Nematodes from the four species ofSpodoptera were poorly de ned and showed a consider-able amount of overlap Even the groups that appeared to

496 Nematology

Speciation in the Acugutturidae

Table 5 Summary of forward stepwise discriminant analysis of morphological characters from male and female Noctuidonema from ve species of Spodoptera and Pseudaletia

Males ndash Spodoptera and Pseudaletia1 (n D 5) Females ndash Spodoptera and Pseudaletia1 (n D 5)

Step Character F (df1 df2) Character F (df1 df2)

1 Ant end to ant intestine 2524 (4 41) Stylet length 3914 (4 57)2 Body length 1491 (4 40) Body width at vulva 1250 (4 56)3 c0 1223 (4 39) Bulb width 591 (4 55)4 Bulb width 725 (4 38) a 498 (4 54)5 Body width at nerve ring 500 (4 37) Vulva to anterior end 380 (4 53)6 Spicule length 409 (4 36) Body width at nerve ring 579 (4 52)7 Maximum body width 265 (4 51)

1) Pseudaletia unipuncta Spodoptera dolichos S frugiperda S latifascia S ornithogalliP lt 005 P lt 001 P lt 0001

be the most different S frugiperda and S dolichos wereextensively overlapped (Fig 2D Table 5)

In the males there were three signi cant canonicalroots accounting for 9764 of the explained variancebetween groups The rst root represents the distancefrom the anterior end to the anterior intestine and V aboutequally whereas the second root represents primarily thebody width at the nerve ring (Fig 2D Table 5) Malenematodes from P unipuncta were well separated fromnematodes from Spodoptera which agrees with resultsfrom the females Males on S dolichosshowed a moderateamount of separation from males on the other threespecies of Spodoptera

DATA ANALYSIS

Eighteen morphological characters contribute to differ-entiation among male nematodes on 14 host species in-cluded in this study The most signi cant and useful ofthese are the distance of the excretory pore from the an-terior end spicule length and stylet length These threecharacters and V are the most useful in differentiating be-tween females (Table 2) Furthermore ANOVA showedthat the stylet length of nematodes on the two speciesof Sphingidae is signi cantly greater (P lt 0001) thanthe stylet length of nematodes on any of the other hostspecies The two populationsof nematodes from Sphingi-dae differ signi cantly in only eight morphological char-acters whereas each differs from V dibolia in about 20characters Females on P lusca differ from those on Xchiron in the longer length of the stylet greater length andwidth of the bulb greater distance of the excretory poreand bulb base from the anterior end greater body widthat the vulva and greater excretory pore distance as a per-

centage of body length However in the absence of malesfrom X chiron no conclusion can be reached on whethernematodes on Sphingidae are conspeci c The consistentseparation of nematodes parasitising Sphingidae from allother Acugutturidae suggests that they probably representat least one new genus Particularly remarkable is the longstylet which represents about 37 of the body length andaverages about 190 sup1m in length (Fig 3)

Discriminant analysis shows that the ve groups ofnematodes with complex spicules (Noctuidonema) arewell separated (Fig 1C D) from the six groups withsimple spicules (Vampyronema) Although the discretecharacter of spicule structure is suf cient to separate thesetwo genera particularly good separation can also be madebased on important continuously variable characters suchas stylet length distance of the excretory pore to theanterior end and the V ratio Furthermore within eachof these two genera separate groups can be detectedbased primarily on differences in stylet length Thusthe nematodes on Pseudaletia unipuncta differ fromthe nematodes on the four Spodoptera spp (Fig 1CD) and those on Anicla infecta Agrapha oxygrammaand Eulepidotis addens differ from V dibolia and itsconspeci c on M disseverans (Fig 2A B)

Female nematodes on P unipuncta differ (P lt 005)from those on S dolichos S frugiperda S latifasciaand S ornithogalli in eight 18 14 and 14 morpholo-gical characters respectivelyThese results providestrongsupport for separate taxonomic status of nematodes onP unipuncta as a new species of Noctuidonema Resultsfrom simple scatterplots prepared for each gender usingthe most importantmorphologicalcharacters identi ed bydiscriminant analysis agree with the preceding analysis

Vol 4(4) 2002 497

OG Marti Jr et al

Fig 3 Plot of stylet length of ectoparasitic nematodes from insect hosts (Blattodea and Lepidoptera) A Female nematodes from15 host species B Male nematodes from 14 host species Legend (Host names) ao Agrapha oxygramma ai Anicla infecta eaEulepidotis addens md Mocis disseverans ml M latipes oc Orthodes crenulata pa Periplaneta americana (St Lucia) pa2 Pamericana (Guadeloupe) pl Perigonia lusca pu Pseudaletia unipuncta sd Spodoptera dolichos sf S frugiperda sl S latifascia soS ornithogalli xc Xylophanes chiron

Table 6 Percent correct classication of male nematodes based on posterior probabilities that a case belongs to a particular hostspecies

Host species Percent correct classi cations ndash Percent correct classi cations ndashmale nematodes female nematodes

n Hosts Hosts Hosts Hosts n Hosts Hosts Hosts Hosts1-15 1-11 1-6 7-11 1-15 1-11 1-6 7-11

1 Agrapha oxygramma 7 100 100 100 ndash 12 917 100 917 ndash2 Anicla infecta 12 100 100 100 ndash 15 100 100 100 ndash3 Eulepidotis addens 2 100 100 100 ndash 15 100 100 100 ndash4 Mocis disseverans 4 100 75 75 ndash 9 889 889 889 ndash5 Mocis latipes 16 100 100 100 ndash 39 923 974 948 ndash6 Orthodes crenulata 4 100 100 100 ndash 9 889 889 889 ndash7 Spodoptera dolichos 10 100 100 ndash 90 5 100 100 ndash 808 Spodoptera frugiperda 11 909 909 ndash 100 17 765 824 ndash 7659 Spodoptera latifascia 6 667 833 ndash 50 12 75 833 ndash 667

10 Spodoptera ornithogalli 10 100 100 ndash 100 10 90 90 ndash 9011 Pseudaletia unipuncta 9 100 100 ndash 100 18 100 944 ndash 94412 Perigonia lusca 4 100 ndash ndash ndash 19 938 ndash ndash ndash13 Periplaneta americana (1) 7 100 ndash ndash ndash 18 882 ndash ndash ndash14 Periplaneta americana (2) 5 100 ndash ndash ndash 13 846 ndash ndash ndash15 Xylophanes chiron 0 ndash ndash ndash ndash 18 100 ndash ndash ndash

(1) St Lucia(2) Guadeloupe

498 Nematology

Speciation in the Acugutturidae

Fig 4 Tree diagram for ectoparasitic nematodes from insecthost species (Blattodea and Lepidoptera) The tree was pro-duced using Euclidean distance as the distance measure andcomplete linkage as the amalgamation rule Both graphs areplotted on an X-axis standardised by (DlinkDmax)100 whichrepresents the percentage of the range from the maximum tothe minimum distance in the data A Female nematodes datafor excretory pore stylet length V bulb width a anterior endto anterior intestine and vulva to anterior end were used BMale nematodes data for excretory pore spicule length andstylet length were used Legend (Host Names) pa Periplan-eta americana (St Lucia) pa2 P americana (Guadeloupe) aoAgrapha oxygramma ai Anicla infecta ea Eulepidotis addensmd Mocis disseverans ml M latipes oc Orthodes crenulatapl Perigonia lusca pu Pseudaletia unipuncta sd Spodopteradolichos sf S frugiperda sl S latifascia so S ornithogallixc Xylophanes chiron

Spicule length was not an important character in dis-tinguishing among male Noctuidonema Male nematodeson S dolichos and S frugiperda showed the most sepa-ration with moderate overlap whereas males on S or-nithogalli appeared to be most similar to those on S

dolichos and males on S latifascia appeared most similarto those on S frugiperda Results from both genders agreethat nematodes on Pseudaletia are different from thoseon Spodoptera and that the nematodes on S dolichos andS frugiperda are the most different with the nematodeson S latifascia and S ornithogallibeing intermediate be-tween these two Overall the results do not support theexistence of separate species of Noctuidonema on hostsof the genus Spodoptera (Fig 4 Table 6)

These results provide evidence that nematodes on Ainfecta A oxygramma and E addens are different fromeach other and from the remaining groups having simplespicules Nematodes on M disseverans and O crenulatawere not distinct from V dibolia the species parasitisingM latipes Nematodes on M disseverans differed signif-icantly from V dibolia in only six morphological charac-ters in the females and ten in the males The differenceswere small and can be attributed to adaptation of conspe-ci c nematodes to parasitism of closely related sympatrichost species

Comparisonof the description of V daptria to measure-ments of other Vampyronema in this study indicates that itdiffers in stylet length and spicule shape V daptria hasthe shortest stylet length observed in the genus rangingfrom 53-60 sup1m in males and 59-69 sup1m in females whichis shorter than the stylet length of nematodes on A in-fecta (Fig 3) In addition spicules of V daptria taper to ne points whereas in Vampyronema from other hosts thespicule tips are blunt V daptria on L porcia appears tobe a distinct species and we therefore do not assign any ofthe Vampyronema from this study to it

Nematodes on the St Lucia and Guadeloupe popula-tions of Periplaneta americana are not well differenti-ated from each other based on the suite of characters thatwere examined (Fig 1A B) Although there are a fewsigni cantly different characters between the two popula-tions the differences probably do not represent interspe-ci c variation In addition the characters of both popula-tions are in agreement with the published description ofAcugutturus parasiticus (Hunt 1980)

Examination of additional specimens of ectoparasiticnematodes from other species of Lepidoptera showed thatmost were Vampyronema whereas Noctuidonema wasfound only on moths from two genera Pseudaletia andSpodoptera (Table 7)

Vol 4(4) 2002 499

OG Marti Jr et al

Table 7 Association of quadrine and trine Noctuidae with Vampyronema and Noctuidonema

Host Taxon Hind Wing Venation1 Locality2 Nematodes Genus

NoctuidaeCatocalinae

Anomis oedema Gueneacutee quadri ne FG 4 1 VampyronemaBaniana sp quadri ne FG 1 0 VampyronemaBaniana ostia Druce quadri ne FG 7 0 VampyronemaEuglyphis sp quadri ne FG 1 0 VampyronemaEulepidotis addens (Walker) quadri ne GU 5 2 VampyronemaItomia opisthographa (Gueneacutee) quadri ne FG 5 1 VampyronemaLesmone formularis (Geyer) quadri ne FG 1 0 VampyronemaLesmone porcia (Stoll) quadri ne GU ndash3 Vampyronema3

Melipotis fasciolaris (Huumlbner) quadri ne FG 0 1 VampyronemaMetalectra sp quadri ne FG 4 10 VampyronemaMetria sp quadri ne FG 0 1 VampyronemaMocis dif uens (Gueneacutee) quadri ne FG 1 1 Vampyronema4

Mocis disseverans (Walker) quadri ne GA 9 4 Vampyronema4

Mocis latipes Gueneacutee quadri ne GA 9 16 Vampyronema4

Perasia ora (Cramer)5 quadri ne FG 10 0 VampyronemaSelenisa sp quadri ne FG 4 0 VampyronemaZale ctilis (Gueneacutee) quadri ne FG 4 1 Vampyronema

PlusiinaeAgrapha oxygramma (Geyer) quadri ne GA 13 7 VampyronemaArgyrogramma verruca (Fabricius) quadri ne FG 0 2 VampyronemaChrysodeixis chalcites (Esper) quadri ne Fiji 8 7 Vampyronema

AmphipyrinaeSpodoptera androgea (Stoll) tri ne FG 4 0 Noctuidonema6

Spodoptera dolichos (Fabricius) tri ne GA 5 10 Noctuidonema6

Spodoptera frugiperda (JE Smith) tri ne GA 17 11 Noctuidonema6

Spodoptera latifascia (Walker) tri ne GA 12 6 Noctuidonema6

Spodoptera litura (Fabricius) tri ne Fiji 15 7 Noctuidonema6

Spodoptera ornithogalli (Gueneacutee) tri ne GA 10 10 Noctuidonema6

HadeninaeOrthodes crenulata (Butler) tri ne GA 9 4 VampyronemaPseudaletia unipuncta (Haworth) tri ne GA 18 9 Noctuidonema

NoctuinaeAnicla infecta (Ochsenheimer) tri ne GA 15 12 VampyronemaTandilia rodea (Schaus) tri ne FG 2 3 Vampyronema

PyralidaePyraustinae

Lamprosema moccalis Schaus ndash FG 4 0 VampyronemaSphingidae

MacroglossinaeErinnys obscura (Fabricius) ndash FG 22 2 Vampyronema7

Perigonia lusca Fabricius ndash GU 19 4 Vampyronema7

Xylophanes chiron Drury ndash GU 8 0 Vampyronema7

1) Quadri ne and tri ne hindwing venation applies to Noctuidae only Pyralidae and Sphingidae are included for completeness thestatus of Plusiinae as quadri nes is ambiguous because it groups together mostly species having quadri ne wing venation yet is oftenincluded with chie y tri ne subfamilies (Kitching amp Rawlins 1998) 2 Fiji Fiji Islands FG French Guiana GU Guadeloupe GAGeorgia USA 3 V daptria Anderson amp Laumond 1992 designationbased on Hunt (1993) 4 V dibolia 5 Synonymous with Nymbisarcuata Walker (in Rogers et al 1990a and Simmons amp Rogers 1996) 6 N guyanense 7 Probably represents an undescribed genus

500 Nematology

Speciation in the Acugutturidae

Discussion

The position of the excretory pore posterior to thebulb the comparatively short length of the stylet thecomparatively long tail and the simple rose-thorn-shapedspicules are suf cient to clearly separate Acugutturus aparasite of Blattodea from Vampyronema and Noctuido-nema parasites of Lepidoptera Sampled populations re-veal four inferred monophyleticgroups Acugutturus par-asiticus from two West Indian populations of Periplan-eta americana a new genus on Sphingidae Noctuido-nema guyanense from Spodoptera spp with nematodeson P unipuncta as a probable new species of Noctuido-nema and Vampyronema spp on the remaining hosts inthis study with the nematodes on A infecta representinga new species of Vampyronema

Each inferred monophyletic group can be de ned byseveral unique combinations of characters quantitativeand qualitative that satisfy the minimal requirements forspecies diagnosis (Wheeler 1999 Wheeler amp Platnick2000) Such characters appear to represent lineage inde-pendence via autapomorphies hypotheses ideally testedvia outgroup comparison Such tests while more philo-sophically satisfying (Adams 1998 Wiley amp Mayden2000) are beyond the scope of the present paper

According to Kitching and Rawlins (1998) differencesin hindwing venation broadly divide the Noctuidae intotwo phenetic groups tri nes and quadri nes Catocalinaeand Plusiinae are quadri nes (but see comment 1 in Ta-ble 7) whereas the other noctuid subfamilies mentionedin this study are tri nes Complex spicules probably rep-resent a derived condition found thus far only in nema-todes on two genera of tri ne noctuids Pseudaletia andSpodoptera while the simple spicules of Vampyronemaprobably represent a plesiomorphic state found in bothtri ne and quadri ne noctuids The extremely long styletsof nematodes from Sphingidae may have arisen in re-sponse to parasitisation of large moths

The Acugutturidae are associated with two very dis-tantly related insect orders Blattodea and Lepidoptera(Endopterygota) There is no record of occurrence ofthe Acugutturidae on any other insect order Among theLepidoptera the Acugutturidae and more precisely theNoctuidonematinae are known from only six familiesGeometridae Lasiocampidae Noctuidae NotodontidaePyralidae and Sphingidae (Marti et al 2000) most ofwhich are not closely related with the exceptionof Lasio-campidae and Sphingidae members of the Bombycoidea(and their relatives) clade (Lemaire amp Minet 1998) and

Notodontidaeand Noctuidae members of the Noctuoideasuperfamily (Kitching amp Rawlins 1998)

One explanation for such a surprising distribution ofthe Acugutturidae on their hosts may be that availabledata are too sparse due to insuf cient sampling of otherlepidopteran families While this may be true the studyconducted by Marti et al (2000) shows that there isa clear speci city in nematode host preference Martiet al (2000) updated the list of lepidopteran hosts ofNoctuidonematinae It seems clear that although manyLepidoptera species belonging to 11 families and 115genera have been examined for nematodes only membersof the six aforementioned obtectomeran (Kristensen ampSkalski 1998) families have been found to be positivefor Noctuidonematinae With the exception of Pyralidaeall these families are members of the Macrolepidopteraclade Moreover the positive Noctuidae species representnearly 90 of the total positive species found Thereforewe cannot state that the Noctuidonematinae are notpotentially parasites of most of the lepidopteran familiesbut we can state that at least in the Neotropical Regionthey parasitise the Macrolepidoptera preferentially andamong that clade mostly a single family the Noctuidae(Marti et al 2000)

The Aphelenchida include several different trophictypes with many species being mycetophagous or phy-toparasitic with about one third of known species be-ing entomoparasitic (Hunt 1993) Considering the list ofinsect hosts we cannot hypothesise that the Acugutturi-dae are derived from a single ancestor transferring froma plant to an insect host It seems more realistic to pro-pose several transfers from plant to insect hosts during thehistory of the Acugutturidae A transfer (Acugutturinae)occurred on the Blattodea and another one (Noctuidone-matinae) on the Lepidoptera Another transfer may haveoccurred on the Bombycoidea sensu lato If at least twosuch principal transfers occurred then it is possible thatthe Acugutturinae and Noctuidonematinae would not beconfamilial and that these two taxonomicunits deserve fa-milial rank Cladistic and molecularphylogeneticanalysiswill help to solve that problem The existenceof symbioticbacteria in these nematodes (Marti et al 1995) shouldprovide an independentmeans of analysing these relation-ships

Concerning the relationship between Lepidoptera andits ectoparasitic nematodes the present work highlightsthe following situation a new genus of Noctuidone-matinae parasitises Sphingidae a genuine Bombycoidea(Lemaire amp Minet 1998) Vampyronema parasitises the

Vol 4(4) 2002 501

OG Marti Jr et al

quadri ne noctuid subfamily Catocalinae (sensu lato in-cluding Ophiderinae) Plusiinae an Hadeninae and aNoctuinae (both typical tri ne noctuids) and Noctui-donema spp parasitises another Hadeninae and severalSpodoptera a tri ne noctuid genus previously includedin the Amphipyrinae but now of uncertain subfamilial af- liation (Poole 1995 Kitching amp Rawlins 1998) Exceptfor nematodes on Sphingidae which likely correspond toa well-separated clade whose phylogenetic relationshipto other Noctuidonematinae needs to be speci ed it ap-pears that Vampyronema and Noctuidonema both para-sitise Noctuidae but that Vampyronema parasitises bothquadri ne and tri ne noctuids Noctuidonema possibly aderived form of Noctuidonematinae is restricted to a fewgenera (Spodoptera and Pseudaletia) of tri ne noctuidsthat are not closely related (Mitchell et al 2000)

Recent work has questioned the monophylyof the Noc-tuidae and there is a growing feeling among moleculartaxonomists that the Noctuidae may be paraphyletic sev-eral quadri ne subfamilies including Catocalinae beingmore closely related to Arctiidae and Lymantriidae than totri ne noctuids (Weller et al 1994 Mitchell et al 1997Fang et al 2000 Mitchell et al 2000) The latter groupis considered as monophyleticClearly systematics of theNoctuidonematinaedo not match the tentative new classi- cation of the Noctuidae

This lack of congruence between parasite and hostsystematics leads us to consider that the relationshipsbetween Acugutturidae and their hosts have not beenguided by coevolutionary forces in the sense of Ehrlichand Raven (1964) but by horizontal transfers (Rogers ampMarti 1996) and sequential evolution (Jermy 1984 Fu-tuyma amp McCafferty 1990) The Acugutturidae and es-pecially the Noctuidonematinae have likely experiencedan adaptive radiation on higher Lepidoptera long aftertheir host diversi cation History of the Noctuidonemati-nae has been characterised by host transfer events ontofairly distantly related hosts followed by adaptation tonew hosts leading to speciation processes that now permitus to distinguish easily fairly species-speci c populationsor strains of Vampyronema or Noctuidonema In mothsthese host transfers may have occurred through contami-nation of nematode-free species by nematode eggs juve-niles or adults left on resting or feeding surfaces (leavesor owers) by individualsof infested species The successof such contamination may be linked to the availabilityand accessibility of host intersegmental membranes thinenough to be pierced by nematode stylets a factor that

may have physically and dramatically restricted the hostrange of all Acugutturidae

The present morphological analysis of acugutturid ne-matodes parasitising 13 species of Lepidoptera and onespecies of cockroach revealed a greater degree of diversitythan previously known in this group The genus Vampy-ronema grouping nematodes with simple spicules in-cludes two inferred new species in addition to V dibo-lia and V daptria already known from the LepidopteraThe genus Noctuidonema grouping nematodes of Lepi-doptera with complex spicules includes a new speciesand the nematodes found on Sphingidae probably belongto a new genus The relationshipbetween these nematodesand their hosts appears to be much more complex thanpreviously thought In addition to their bene ts as para-sites of pest Lepidoptera acugutturid nematodes are aninteresting model for parasite-host coevolutionarystudies

Acknowledgements

We thank David Hunt (CABI Bioscience UK) DanielLachaise (CNRS France) Alvin Simmons (USDA Char-leston SC USA) and S Patricia Stock (University ofCalifornia Davis CA USA) for reviewing earlier ver-sions of this paper We thank Jeanette Williams and JohnClayton for collecting cockroaches in St Lucia and mothsin Fiji respectively We thank the following individualsfor identifyingor con rming our identi cations of insectsDavid Nickle (Smithsonian Washington DC USA) forPeriplaneta americana Pierre Zagatti (INRA France)for Pyralidae and Hugo Kons (University of FloridaGainesville FL USA) for Orthodes crenulata

References

ADAMS BJ (1998) Species concepts and evolutionary para-digm in modern nematology Journal of Nematology 30 1-21

ANDERSON RV amp LAUMOND C (1990) NoctuidonemaguyanenseRemillet amp Silvain 1988 morphologie du spiculeredescription du macircle et diagnose amendeacutee du genre Noctui-donema Revue de Neacutematologie 13 433-436

ANDERSON RV amp LAUMOND C (1992) Noctuidonemadaptria n sp (Nematoda Aphelenchoididae)an ectoparasiteof the moth Lesmone porcia (Stoll) Journal of Nematology24 16-22

EHRLICH P amp RAVEN P (1964) Butter ies and plants astudy in evolution Evolution 18 586-608

FANG QQ MITCHELL A REGIER JC MITTER CFRIEDLANDER TP amp POOLE RW (2000) Phylogenetic

502 Nematology

Speciation in the Acugutturidae

Table 5 Summary of forward stepwise discriminant analysis of morphological characters from male and female Noctuidonema from ve species of Spodoptera and Pseudaletia

Males ndash Spodoptera and Pseudaletia1 (n D 5) Females ndash Spodoptera and Pseudaletia1 (n D 5)

Step Character F (df1 df2) Character F (df1 df2)

1 Ant end to ant intestine 2524 (4 41) Stylet length 3914 (4 57)2 Body length 1491 (4 40) Body width at vulva 1250 (4 56)3 c0 1223 (4 39) Bulb width 591 (4 55)4 Bulb width 725 (4 38) a 498 (4 54)5 Body width at nerve ring 500 (4 37) Vulva to anterior end 380 (4 53)6 Spicule length 409 (4 36) Body width at nerve ring 579 (4 52)7 Maximum body width 265 (4 51)

1) Pseudaletia unipuncta Spodoptera dolichos S frugiperda S latifascia S ornithogalliP lt 005 P lt 001 P lt 0001

be the most different S frugiperda and S dolichos wereextensively overlapped (Fig 2D Table 5)

In the males there were three signi cant canonicalroots accounting for 9764 of the explained variancebetween groups The rst root represents the distancefrom the anterior end to the anterior intestine and V aboutequally whereas the second root represents primarily thebody width at the nerve ring (Fig 2D Table 5) Malenematodes from P unipuncta were well separated fromnematodes from Spodoptera which agrees with resultsfrom the females Males on S dolichosshowed a moderateamount of separation from males on the other threespecies of Spodoptera

DATA ANALYSIS

Eighteen morphological characters contribute to differ-entiation among male nematodes on 14 host species in-cluded in this study The most signi cant and useful ofthese are the distance of the excretory pore from the an-terior end spicule length and stylet length These threecharacters and V are the most useful in differentiating be-tween females (Table 2) Furthermore ANOVA showedthat the stylet length of nematodes on the two speciesof Sphingidae is signi cantly greater (P lt 0001) thanthe stylet length of nematodes on any of the other hostspecies The two populationsof nematodes from Sphingi-dae differ signi cantly in only eight morphological char-acters whereas each differs from V dibolia in about 20characters Females on P lusca differ from those on Xchiron in the longer length of the stylet greater length andwidth of the bulb greater distance of the excretory poreand bulb base from the anterior end greater body widthat the vulva and greater excretory pore distance as a per-

centage of body length However in the absence of malesfrom X chiron no conclusion can be reached on whethernematodes on Sphingidae are conspeci c The consistentseparation of nematodes parasitising Sphingidae from allother Acugutturidae suggests that they probably representat least one new genus Particularly remarkable is the longstylet which represents about 37 of the body length andaverages about 190 sup1m in length (Fig 3)

Discriminant analysis shows that the ve groups ofnematodes with complex spicules (Noctuidonema) arewell separated (Fig 1C D) from the six groups withsimple spicules (Vampyronema) Although the discretecharacter of spicule structure is suf cient to separate thesetwo genera particularly good separation can also be madebased on important continuously variable characters suchas stylet length distance of the excretory pore to theanterior end and the V ratio Furthermore within eachof these two genera separate groups can be detectedbased primarily on differences in stylet length Thusthe nematodes on Pseudaletia unipuncta differ fromthe nematodes on the four Spodoptera spp (Fig 1CD) and those on Anicla infecta Agrapha oxygrammaand Eulepidotis addens differ from V dibolia and itsconspeci c on M disseverans (Fig 2A B)

Female nematodes on P unipuncta differ (P lt 005)from those on S dolichos S frugiperda S latifasciaand S ornithogalli in eight 18 14 and 14 morpholo-gical characters respectivelyThese results providestrongsupport for separate taxonomic status of nematodes onP unipuncta as a new species of Noctuidonema Resultsfrom simple scatterplots prepared for each gender usingthe most importantmorphologicalcharacters identi ed bydiscriminant analysis agree with the preceding analysis

Vol 4(4) 2002 497

OG Marti Jr et al

Fig 3 Plot of stylet length of ectoparasitic nematodes from insect hosts (Blattodea and Lepidoptera) A Female nematodes from15 host species B Male nematodes from 14 host species Legend (Host names) ao Agrapha oxygramma ai Anicla infecta eaEulepidotis addens md Mocis disseverans ml M latipes oc Orthodes crenulata pa Periplaneta americana (St Lucia) pa2 Pamericana (Guadeloupe) pl Perigonia lusca pu Pseudaletia unipuncta sd Spodoptera dolichos sf S frugiperda sl S latifascia soS ornithogalli xc Xylophanes chiron

Table 6 Percent correct classication of male nematodes based on posterior probabilities that a case belongs to a particular hostspecies

Host species Percent correct classi cations ndash Percent correct classi cations ndashmale nematodes female nematodes

n Hosts Hosts Hosts Hosts n Hosts Hosts Hosts Hosts1-15 1-11 1-6 7-11 1-15 1-11 1-6 7-11

1 Agrapha oxygramma 7 100 100 100 ndash 12 917 100 917 ndash2 Anicla infecta 12 100 100 100 ndash 15 100 100 100 ndash3 Eulepidotis addens 2 100 100 100 ndash 15 100 100 100 ndash4 Mocis disseverans 4 100 75 75 ndash 9 889 889 889 ndash5 Mocis latipes 16 100 100 100 ndash 39 923 974 948 ndash6 Orthodes crenulata 4 100 100 100 ndash 9 889 889 889 ndash7 Spodoptera dolichos 10 100 100 ndash 90 5 100 100 ndash 808 Spodoptera frugiperda 11 909 909 ndash 100 17 765 824 ndash 7659 Spodoptera latifascia 6 667 833 ndash 50 12 75 833 ndash 667

10 Spodoptera ornithogalli 10 100 100 ndash 100 10 90 90 ndash 9011 Pseudaletia unipuncta 9 100 100 ndash 100 18 100 944 ndash 94412 Perigonia lusca 4 100 ndash ndash ndash 19 938 ndash ndash ndash13 Periplaneta americana (1) 7 100 ndash ndash ndash 18 882 ndash ndash ndash14 Periplaneta americana (2) 5 100 ndash ndash ndash 13 846 ndash ndash ndash15 Xylophanes chiron 0 ndash ndash ndash ndash 18 100 ndash ndash ndash

(1) St Lucia(2) Guadeloupe

498 Nematology

Speciation in the Acugutturidae

Fig 4 Tree diagram for ectoparasitic nematodes from insecthost species (Blattodea and Lepidoptera) The tree was pro-duced using Euclidean distance as the distance measure andcomplete linkage as the amalgamation rule Both graphs areplotted on an X-axis standardised by (DlinkDmax)100 whichrepresents the percentage of the range from the maximum tothe minimum distance in the data A Female nematodes datafor excretory pore stylet length V bulb width a anterior endto anterior intestine and vulva to anterior end were used BMale nematodes data for excretory pore spicule length andstylet length were used Legend (Host Names) pa Periplan-eta americana (St Lucia) pa2 P americana (Guadeloupe) aoAgrapha oxygramma ai Anicla infecta ea Eulepidotis addensmd Mocis disseverans ml M latipes oc Orthodes crenulatapl Perigonia lusca pu Pseudaletia unipuncta sd Spodopteradolichos sf S frugiperda sl S latifascia so S ornithogallixc Xylophanes chiron

Spicule length was not an important character in dis-tinguishing among male Noctuidonema Male nematodeson S dolichos and S frugiperda showed the most sepa-ration with moderate overlap whereas males on S or-nithogalli appeared to be most similar to those on S

dolichos and males on S latifascia appeared most similarto those on S frugiperda Results from both genders agreethat nematodes on Pseudaletia are different from thoseon Spodoptera and that the nematodes on S dolichos andS frugiperda are the most different with the nematodeson S latifascia and S ornithogallibeing intermediate be-tween these two Overall the results do not support theexistence of separate species of Noctuidonema on hostsof the genus Spodoptera (Fig 4 Table 6)

These results provide evidence that nematodes on Ainfecta A oxygramma and E addens are different fromeach other and from the remaining groups having simplespicules Nematodes on M disseverans and O crenulatawere not distinct from V dibolia the species parasitisingM latipes Nematodes on M disseverans differed signif-icantly from V dibolia in only six morphological charac-ters in the females and ten in the males The differenceswere small and can be attributed to adaptation of conspe-ci c nematodes to parasitism of closely related sympatrichost species

Comparisonof the description of V daptria to measure-ments of other Vampyronema in this study indicates that itdiffers in stylet length and spicule shape V daptria hasthe shortest stylet length observed in the genus rangingfrom 53-60 sup1m in males and 59-69 sup1m in females whichis shorter than the stylet length of nematodes on A in-fecta (Fig 3) In addition spicules of V daptria taper to ne points whereas in Vampyronema from other hosts thespicule tips are blunt V daptria on L porcia appears tobe a distinct species and we therefore do not assign any ofthe Vampyronema from this study to it

Nematodes on the St Lucia and Guadeloupe popula-tions of Periplaneta americana are not well differenti-ated from each other based on the suite of characters thatwere examined (Fig 1A B) Although there are a fewsigni cantly different characters between the two popula-tions the differences probably do not represent interspe-ci c variation In addition the characters of both popula-tions are in agreement with the published description ofAcugutturus parasiticus (Hunt 1980)

Examination of additional specimens of ectoparasiticnematodes from other species of Lepidoptera showed thatmost were Vampyronema whereas Noctuidonema wasfound only on moths from two genera Pseudaletia andSpodoptera (Table 7)

Vol 4(4) 2002 499

OG Marti Jr et al

Table 7 Association of quadrine and trine Noctuidae with Vampyronema and Noctuidonema

Host Taxon Hind Wing Venation1 Locality2 Nematodes Genus

NoctuidaeCatocalinae

Anomis oedema Gueneacutee quadri ne FG 4 1 VampyronemaBaniana sp quadri ne FG 1 0 VampyronemaBaniana ostia Druce quadri ne FG 7 0 VampyronemaEuglyphis sp quadri ne FG 1 0 VampyronemaEulepidotis addens (Walker) quadri ne GU 5 2 VampyronemaItomia opisthographa (Gueneacutee) quadri ne FG 5 1 VampyronemaLesmone formularis (Geyer) quadri ne FG 1 0 VampyronemaLesmone porcia (Stoll) quadri ne GU ndash3 Vampyronema3

Melipotis fasciolaris (Huumlbner) quadri ne FG 0 1 VampyronemaMetalectra sp quadri ne FG 4 10 VampyronemaMetria sp quadri ne FG 0 1 VampyronemaMocis dif uens (Gueneacutee) quadri ne FG 1 1 Vampyronema4

Mocis disseverans (Walker) quadri ne GA 9 4 Vampyronema4

Mocis latipes Gueneacutee quadri ne GA 9 16 Vampyronema4

Perasia ora (Cramer)5 quadri ne FG 10 0 VampyronemaSelenisa sp quadri ne FG 4 0 VampyronemaZale ctilis (Gueneacutee) quadri ne FG 4 1 Vampyronema

PlusiinaeAgrapha oxygramma (Geyer) quadri ne GA 13 7 VampyronemaArgyrogramma verruca (Fabricius) quadri ne FG 0 2 VampyronemaChrysodeixis chalcites (Esper) quadri ne Fiji 8 7 Vampyronema

AmphipyrinaeSpodoptera androgea (Stoll) tri ne FG 4 0 Noctuidonema6

Spodoptera dolichos (Fabricius) tri ne GA 5 10 Noctuidonema6

Spodoptera frugiperda (JE Smith) tri ne GA 17 11 Noctuidonema6

Spodoptera latifascia (Walker) tri ne GA 12 6 Noctuidonema6

Spodoptera litura (Fabricius) tri ne Fiji 15 7 Noctuidonema6

Spodoptera ornithogalli (Gueneacutee) tri ne GA 10 10 Noctuidonema6

HadeninaeOrthodes crenulata (Butler) tri ne GA 9 4 VampyronemaPseudaletia unipuncta (Haworth) tri ne GA 18 9 Noctuidonema

NoctuinaeAnicla infecta (Ochsenheimer) tri ne GA 15 12 VampyronemaTandilia rodea (Schaus) tri ne FG 2 3 Vampyronema

PyralidaePyraustinae

Lamprosema moccalis Schaus ndash FG 4 0 VampyronemaSphingidae

MacroglossinaeErinnys obscura (Fabricius) ndash FG 22 2 Vampyronema7

Perigonia lusca Fabricius ndash GU 19 4 Vampyronema7

Xylophanes chiron Drury ndash GU 8 0 Vampyronema7

1) Quadri ne and tri ne hindwing venation applies to Noctuidae only Pyralidae and Sphingidae are included for completeness thestatus of Plusiinae as quadri nes is ambiguous because it groups together mostly species having quadri ne wing venation yet is oftenincluded with chie y tri ne subfamilies (Kitching amp Rawlins 1998) 2 Fiji Fiji Islands FG French Guiana GU Guadeloupe GAGeorgia USA 3 V daptria Anderson amp Laumond 1992 designationbased on Hunt (1993) 4 V dibolia 5 Synonymous with Nymbisarcuata Walker (in Rogers et al 1990a and Simmons amp Rogers 1996) 6 N guyanense 7 Probably represents an undescribed genus

500 Nematology

Speciation in the Acugutturidae

Discussion

The position of the excretory pore posterior to thebulb the comparatively short length of the stylet thecomparatively long tail and the simple rose-thorn-shapedspicules are suf cient to clearly separate Acugutturus aparasite of Blattodea from Vampyronema and Noctuido-nema parasites of Lepidoptera Sampled populations re-veal four inferred monophyleticgroups Acugutturus par-asiticus from two West Indian populations of Periplan-eta americana a new genus on Sphingidae Noctuido-nema guyanense from Spodoptera spp with nematodeson P unipuncta as a probable new species of Noctuido-nema and Vampyronema spp on the remaining hosts inthis study with the nematodes on A infecta representinga new species of Vampyronema

Each inferred monophyletic group can be de ned byseveral unique combinations of characters quantitativeand qualitative that satisfy the minimal requirements forspecies diagnosis (Wheeler 1999 Wheeler amp Platnick2000) Such characters appear to represent lineage inde-pendence via autapomorphies hypotheses ideally testedvia outgroup comparison Such tests while more philo-sophically satisfying (Adams 1998 Wiley amp Mayden2000) are beyond the scope of the present paper

According to Kitching and Rawlins (1998) differencesin hindwing venation broadly divide the Noctuidae intotwo phenetic groups tri nes and quadri nes Catocalinaeand Plusiinae are quadri nes (but see comment 1 in Ta-ble 7) whereas the other noctuid subfamilies mentionedin this study are tri nes Complex spicules probably rep-resent a derived condition found thus far only in nema-todes on two genera of tri ne noctuids Pseudaletia andSpodoptera while the simple spicules of Vampyronemaprobably represent a plesiomorphic state found in bothtri ne and quadri ne noctuids The extremely long styletsof nematodes from Sphingidae may have arisen in re-sponse to parasitisation of large moths

The Acugutturidae are associated with two very dis-tantly related insect orders Blattodea and Lepidoptera(Endopterygota) There is no record of occurrence ofthe Acugutturidae on any other insect order Among theLepidoptera the Acugutturidae and more precisely theNoctuidonematinae are known from only six familiesGeometridae Lasiocampidae Noctuidae NotodontidaePyralidae and Sphingidae (Marti et al 2000) most ofwhich are not closely related with the exceptionof Lasio-campidae and Sphingidae members of the Bombycoidea(and their relatives) clade (Lemaire amp Minet 1998) and

Notodontidaeand Noctuidae members of the Noctuoideasuperfamily (Kitching amp Rawlins 1998)

One explanation for such a surprising distribution ofthe Acugutturidae on their hosts may be that availabledata are too sparse due to insuf cient sampling of otherlepidopteran families While this may be true the studyconducted by Marti et al (2000) shows that there isa clear speci city in nematode host preference Martiet al (2000) updated the list of lepidopteran hosts ofNoctuidonematinae It seems clear that although manyLepidoptera species belonging to 11 families and 115genera have been examined for nematodes only membersof the six aforementioned obtectomeran (Kristensen ampSkalski 1998) families have been found to be positivefor Noctuidonematinae With the exception of Pyralidaeall these families are members of the Macrolepidopteraclade Moreover the positive Noctuidae species representnearly 90 of the total positive species found Thereforewe cannot state that the Noctuidonematinae are notpotentially parasites of most of the lepidopteran familiesbut we can state that at least in the Neotropical Regionthey parasitise the Macrolepidoptera preferentially andamong that clade mostly a single family the Noctuidae(Marti et al 2000)

The Aphelenchida include several different trophictypes with many species being mycetophagous or phy-toparasitic with about one third of known species be-ing entomoparasitic (Hunt 1993) Considering the list ofinsect hosts we cannot hypothesise that the Acugutturi-dae are derived from a single ancestor transferring froma plant to an insect host It seems more realistic to pro-pose several transfers from plant to insect hosts during thehistory of the Acugutturidae A transfer (Acugutturinae)occurred on the Blattodea and another one (Noctuidone-matinae) on the Lepidoptera Another transfer may haveoccurred on the Bombycoidea sensu lato If at least twosuch principal transfers occurred then it is possible thatthe Acugutturinae and Noctuidonematinae would not beconfamilial and that these two taxonomicunits deserve fa-milial rank Cladistic and molecularphylogeneticanalysiswill help to solve that problem The existenceof symbioticbacteria in these nematodes (Marti et al 1995) shouldprovide an independentmeans of analysing these relation-ships

Concerning the relationship between Lepidoptera andits ectoparasitic nematodes the present work highlightsthe following situation a new genus of Noctuidone-matinae parasitises Sphingidae a genuine Bombycoidea(Lemaire amp Minet 1998) Vampyronema parasitises the

Vol 4(4) 2002 501

OG Marti Jr et al

quadri ne noctuid subfamily Catocalinae (sensu lato in-cluding Ophiderinae) Plusiinae an Hadeninae and aNoctuinae (both typical tri ne noctuids) and Noctui-donema spp parasitises another Hadeninae and severalSpodoptera a tri ne noctuid genus previously includedin the Amphipyrinae but now of uncertain subfamilial af- liation (Poole 1995 Kitching amp Rawlins 1998) Exceptfor nematodes on Sphingidae which likely correspond toa well-separated clade whose phylogenetic relationshipto other Noctuidonematinae needs to be speci ed it ap-pears that Vampyronema and Noctuidonema both para-sitise Noctuidae but that Vampyronema parasitises bothquadri ne and tri ne noctuids Noctuidonema possibly aderived form of Noctuidonematinae is restricted to a fewgenera (Spodoptera and Pseudaletia) of tri ne noctuidsthat are not closely related (Mitchell et al 2000)

Recent work has questioned the monophylyof the Noc-tuidae and there is a growing feeling among moleculartaxonomists that the Noctuidae may be paraphyletic sev-eral quadri ne subfamilies including Catocalinae beingmore closely related to Arctiidae and Lymantriidae than totri ne noctuids (Weller et al 1994 Mitchell et al 1997Fang et al 2000 Mitchell et al 2000) The latter groupis considered as monophyleticClearly systematics of theNoctuidonematinaedo not match the tentative new classi- cation of the Noctuidae

This lack of congruence between parasite and hostsystematics leads us to consider that the relationshipsbetween Acugutturidae and their hosts have not beenguided by coevolutionary forces in the sense of Ehrlichand Raven (1964) but by horizontal transfers (Rogers ampMarti 1996) and sequential evolution (Jermy 1984 Fu-tuyma amp McCafferty 1990) The Acugutturidae and es-pecially the Noctuidonematinae have likely experiencedan adaptive radiation on higher Lepidoptera long aftertheir host diversi cation History of the Noctuidonemati-nae has been characterised by host transfer events ontofairly distantly related hosts followed by adaptation tonew hosts leading to speciation processes that now permitus to distinguish easily fairly species-speci c populationsor strains of Vampyronema or Noctuidonema In mothsthese host transfers may have occurred through contami-nation of nematode-free species by nematode eggs juve-niles or adults left on resting or feeding surfaces (leavesor owers) by individualsof infested species The successof such contamination may be linked to the availabilityand accessibility of host intersegmental membranes thinenough to be pierced by nematode stylets a factor that

may have physically and dramatically restricted the hostrange of all Acugutturidae

The present morphological analysis of acugutturid ne-matodes parasitising 13 species of Lepidoptera and onespecies of cockroach revealed a greater degree of diversitythan previously known in this group The genus Vampy-ronema grouping nematodes with simple spicules in-cludes two inferred new species in addition to V dibo-lia and V daptria already known from the LepidopteraThe genus Noctuidonema grouping nematodes of Lepi-doptera with complex spicules includes a new speciesand the nematodes found on Sphingidae probably belongto a new genus The relationshipbetween these nematodesand their hosts appears to be much more complex thanpreviously thought In addition to their bene ts as para-sites of pest Lepidoptera acugutturid nematodes are aninteresting model for parasite-host coevolutionarystudies

Acknowledgements

We thank David Hunt (CABI Bioscience UK) DanielLachaise (CNRS France) Alvin Simmons (USDA Char-leston SC USA) and S Patricia Stock (University ofCalifornia Davis CA USA) for reviewing earlier ver-sions of this paper We thank Jeanette Williams and JohnClayton for collecting cockroaches in St Lucia and mothsin Fiji respectively We thank the following individualsfor identifyingor con rming our identi cations of insectsDavid Nickle (Smithsonian Washington DC USA) forPeriplaneta americana Pierre Zagatti (INRA France)for Pyralidae and Hugo Kons (University of FloridaGainesville FL USA) for Orthodes crenulata

References

ADAMS BJ (1998) Species concepts and evolutionary para-digm in modern nematology Journal of Nematology 30 1-21

ANDERSON RV amp LAUMOND C (1990) NoctuidonemaguyanenseRemillet amp Silvain 1988 morphologie du spiculeredescription du macircle et diagnose amendeacutee du genre Noctui-donema Revue de Neacutematologie 13 433-436

ANDERSON RV amp LAUMOND C (1992) Noctuidonemadaptria n sp (Nematoda Aphelenchoididae)an ectoparasiteof the moth Lesmone porcia (Stoll) Journal of Nematology24 16-22

EHRLICH P amp RAVEN P (1964) Butter ies and plants astudy in evolution Evolution 18 586-608

FANG QQ MITCHELL A REGIER JC MITTER CFRIEDLANDER TP amp POOLE RW (2000) Phylogenetic

502 Nematology

OG Marti Jr et al

Fig 3 Plot of stylet length of ectoparasitic nematodes from insect hosts (Blattodea and Lepidoptera) A Female nematodes from15 host species B Male nematodes from 14 host species Legend (Host names) ao Agrapha oxygramma ai Anicla infecta eaEulepidotis addens md Mocis disseverans ml M latipes oc Orthodes crenulata pa Periplaneta americana (St Lucia) pa2 Pamericana (Guadeloupe) pl Perigonia lusca pu Pseudaletia unipuncta sd Spodoptera dolichos sf S frugiperda sl S latifascia soS ornithogalli xc Xylophanes chiron

Table 6 Percent correct classication of male nematodes based on posterior probabilities that a case belongs to a particular hostspecies

Host species Percent correct classi cations ndash Percent correct classi cations ndashmale nematodes female nematodes

n Hosts Hosts Hosts Hosts n Hosts Hosts Hosts Hosts1-15 1-11 1-6 7-11 1-15 1-11 1-6 7-11

1 Agrapha oxygramma 7 100 100 100 ndash 12 917 100 917 ndash2 Anicla infecta 12 100 100 100 ndash 15 100 100 100 ndash3 Eulepidotis addens 2 100 100 100 ndash 15 100 100 100 ndash4 Mocis disseverans 4 100 75 75 ndash 9 889 889 889 ndash5 Mocis latipes 16 100 100 100 ndash 39 923 974 948 ndash6 Orthodes crenulata 4 100 100 100 ndash 9 889 889 889 ndash7 Spodoptera dolichos 10 100 100 ndash 90 5 100 100 ndash 808 Spodoptera frugiperda 11 909 909 ndash 100 17 765 824 ndash 7659 Spodoptera latifascia 6 667 833 ndash 50 12 75 833 ndash 667

10 Spodoptera ornithogalli 10 100 100 ndash 100 10 90 90 ndash 9011 Pseudaletia unipuncta 9 100 100 ndash 100 18 100 944 ndash 94412 Perigonia lusca 4 100 ndash ndash ndash 19 938 ndash ndash ndash13 Periplaneta americana (1) 7 100 ndash ndash ndash 18 882 ndash ndash ndash14 Periplaneta americana (2) 5 100 ndash ndash ndash 13 846 ndash ndash ndash15 Xylophanes chiron 0 ndash ndash ndash ndash 18 100 ndash ndash ndash

(1) St Lucia(2) Guadeloupe

498 Nematology

Speciation in the Acugutturidae

Fig 4 Tree diagram for ectoparasitic nematodes from insecthost species (Blattodea and Lepidoptera) The tree was pro-duced using Euclidean distance as the distance measure andcomplete linkage as the amalgamation rule Both graphs areplotted on an X-axis standardised by (DlinkDmax)100 whichrepresents the percentage of the range from the maximum tothe minimum distance in the data A Female nematodes datafor excretory pore stylet length V bulb width a anterior endto anterior intestine and vulva to anterior end were used BMale nematodes data for excretory pore spicule length andstylet length were used Legend (Host Names) pa Periplan-eta americana (St Lucia) pa2 P americana (Guadeloupe) aoAgrapha oxygramma ai Anicla infecta ea Eulepidotis addensmd Mocis disseverans ml M latipes oc Orthodes crenulatapl Perigonia lusca pu Pseudaletia unipuncta sd Spodopteradolichos sf S frugiperda sl S latifascia so S ornithogallixc Xylophanes chiron

Spicule length was not an important character in dis-tinguishing among male Noctuidonema Male nematodeson S dolichos and S frugiperda showed the most sepa-ration with moderate overlap whereas males on S or-nithogalli appeared to be most similar to those on S

dolichos and males on S latifascia appeared most similarto those on S frugiperda Results from both genders agreethat nematodes on Pseudaletia are different from thoseon Spodoptera and that the nematodes on S dolichos andS frugiperda are the most different with the nematodeson S latifascia and S ornithogallibeing intermediate be-tween these two Overall the results do not support theexistence of separate species of Noctuidonema on hostsof the genus Spodoptera (Fig 4 Table 6)

These results provide evidence that nematodes on Ainfecta A oxygramma and E addens are different fromeach other and from the remaining groups having simplespicules Nematodes on M disseverans and O crenulatawere not distinct from V dibolia the species parasitisingM latipes Nematodes on M disseverans differed signif-icantly from V dibolia in only six morphological charac-ters in the females and ten in the males The differenceswere small and can be attributed to adaptation of conspe-ci c nematodes to parasitism of closely related sympatrichost species

Comparisonof the description of V daptria to measure-ments of other Vampyronema in this study indicates that itdiffers in stylet length and spicule shape V daptria hasthe shortest stylet length observed in the genus rangingfrom 53-60 sup1m in males and 59-69 sup1m in females whichis shorter than the stylet length of nematodes on A in-fecta (Fig 3) In addition spicules of V daptria taper to ne points whereas in Vampyronema from other hosts thespicule tips are blunt V daptria on L porcia appears tobe a distinct species and we therefore do not assign any ofthe Vampyronema from this study to it

Nematodes on the St Lucia and Guadeloupe popula-tions of Periplaneta americana are not well differenti-ated from each other based on the suite of characters thatwere examined (Fig 1A B) Although there are a fewsigni cantly different characters between the two popula-tions the differences probably do not represent interspe-ci c variation In addition the characters of both popula-tions are in agreement with the published description ofAcugutturus parasiticus (Hunt 1980)

Examination of additional specimens of ectoparasiticnematodes from other species of Lepidoptera showed thatmost were Vampyronema whereas Noctuidonema wasfound only on moths from two genera Pseudaletia andSpodoptera (Table 7)

Vol 4(4) 2002 499

OG Marti Jr et al

Table 7 Association of quadrine and trine Noctuidae with Vampyronema and Noctuidonema

Host Taxon Hind Wing Venation1 Locality2 Nematodes Genus

NoctuidaeCatocalinae

Anomis oedema Gueneacutee quadri ne FG 4 1 VampyronemaBaniana sp quadri ne FG 1 0 VampyronemaBaniana ostia Druce quadri ne FG 7 0 VampyronemaEuglyphis sp quadri ne FG 1 0 VampyronemaEulepidotis addens (Walker) quadri ne GU 5 2 VampyronemaItomia opisthographa (Gueneacutee) quadri ne FG 5 1 VampyronemaLesmone formularis (Geyer) quadri ne FG 1 0 VampyronemaLesmone porcia (Stoll) quadri ne GU ndash3 Vampyronema3

Melipotis fasciolaris (Huumlbner) quadri ne FG 0 1 VampyronemaMetalectra sp quadri ne FG 4 10 VampyronemaMetria sp quadri ne FG 0 1 VampyronemaMocis dif uens (Gueneacutee) quadri ne FG 1 1 Vampyronema4

Mocis disseverans (Walker) quadri ne GA 9 4 Vampyronema4

Mocis latipes Gueneacutee quadri ne GA 9 16 Vampyronema4

Perasia ora (Cramer)5 quadri ne FG 10 0 VampyronemaSelenisa sp quadri ne FG 4 0 VampyronemaZale ctilis (Gueneacutee) quadri ne FG 4 1 Vampyronema

PlusiinaeAgrapha oxygramma (Geyer) quadri ne GA 13 7 VampyronemaArgyrogramma verruca (Fabricius) quadri ne FG 0 2 VampyronemaChrysodeixis chalcites (Esper) quadri ne Fiji 8 7 Vampyronema

AmphipyrinaeSpodoptera androgea (Stoll) tri ne FG 4 0 Noctuidonema6

Spodoptera dolichos (Fabricius) tri ne GA 5 10 Noctuidonema6

Spodoptera frugiperda (JE Smith) tri ne GA 17 11 Noctuidonema6

Spodoptera latifascia (Walker) tri ne GA 12 6 Noctuidonema6

Spodoptera litura (Fabricius) tri ne Fiji 15 7 Noctuidonema6

Spodoptera ornithogalli (Gueneacutee) tri ne GA 10 10 Noctuidonema6

HadeninaeOrthodes crenulata (Butler) tri ne GA 9 4 VampyronemaPseudaletia unipuncta (Haworth) tri ne GA 18 9 Noctuidonema

NoctuinaeAnicla infecta (Ochsenheimer) tri ne GA 15 12 VampyronemaTandilia rodea (Schaus) tri ne FG 2 3 Vampyronema

PyralidaePyraustinae

Lamprosema moccalis Schaus ndash FG 4 0 VampyronemaSphingidae

MacroglossinaeErinnys obscura (Fabricius) ndash FG 22 2 Vampyronema7

Perigonia lusca Fabricius ndash GU 19 4 Vampyronema7

Xylophanes chiron Drury ndash GU 8 0 Vampyronema7

1) Quadri ne and tri ne hindwing venation applies to Noctuidae only Pyralidae and Sphingidae are included for completeness thestatus of Plusiinae as quadri nes is ambiguous because it groups together mostly species having quadri ne wing venation yet is oftenincluded with chie y tri ne subfamilies (Kitching amp Rawlins 1998) 2 Fiji Fiji Islands FG French Guiana GU Guadeloupe GAGeorgia USA 3 V daptria Anderson amp Laumond 1992 designationbased on Hunt (1993) 4 V dibolia 5 Synonymous with Nymbisarcuata Walker (in Rogers et al 1990a and Simmons amp Rogers 1996) 6 N guyanense 7 Probably represents an undescribed genus

500 Nematology

Speciation in the Acugutturidae

Discussion

The position of the excretory pore posterior to thebulb the comparatively short length of the stylet thecomparatively long tail and the simple rose-thorn-shapedspicules are suf cient to clearly separate Acugutturus aparasite of Blattodea from Vampyronema and Noctuido-nema parasites of Lepidoptera Sampled populations re-veal four inferred monophyleticgroups Acugutturus par-asiticus from two West Indian populations of Periplan-eta americana a new genus on Sphingidae Noctuido-nema guyanense from Spodoptera spp with nematodeson P unipuncta as a probable new species of Noctuido-nema and Vampyronema spp on the remaining hosts inthis study with the nematodes on A infecta representinga new species of Vampyronema

Each inferred monophyletic group can be de ned byseveral unique combinations of characters quantitativeand qualitative that satisfy the minimal requirements forspecies diagnosis (Wheeler 1999 Wheeler amp Platnick2000) Such characters appear to represent lineage inde-pendence via autapomorphies hypotheses ideally testedvia outgroup comparison Such tests while more philo-sophically satisfying (Adams 1998 Wiley amp Mayden2000) are beyond the scope of the present paper

According to Kitching and Rawlins (1998) differencesin hindwing venation broadly divide the Noctuidae intotwo phenetic groups tri nes and quadri nes Catocalinaeand Plusiinae are quadri nes (but see comment 1 in Ta-ble 7) whereas the other noctuid subfamilies mentionedin this study are tri nes Complex spicules probably rep-resent a derived condition found thus far only in nema-todes on two genera of tri ne noctuids Pseudaletia andSpodoptera while the simple spicules of Vampyronemaprobably represent a plesiomorphic state found in bothtri ne and quadri ne noctuids The extremely long styletsof nematodes from Sphingidae may have arisen in re-sponse to parasitisation of large moths

The Acugutturidae are associated with two very dis-tantly related insect orders Blattodea and Lepidoptera(Endopterygota) There is no record of occurrence ofthe Acugutturidae on any other insect order Among theLepidoptera the Acugutturidae and more precisely theNoctuidonematinae are known from only six familiesGeometridae Lasiocampidae Noctuidae NotodontidaePyralidae and Sphingidae (Marti et al 2000) most ofwhich are not closely related with the exceptionof Lasio-campidae and Sphingidae members of the Bombycoidea(and their relatives) clade (Lemaire amp Minet 1998) and

Notodontidaeand Noctuidae members of the Noctuoideasuperfamily (Kitching amp Rawlins 1998)

One explanation for such a surprising distribution ofthe Acugutturidae on their hosts may be that availabledata are too sparse due to insuf cient sampling of otherlepidopteran families While this may be true the studyconducted by Marti et al (2000) shows that there isa clear speci city in nematode host preference Martiet al (2000) updated the list of lepidopteran hosts ofNoctuidonematinae It seems clear that although manyLepidoptera species belonging to 11 families and 115genera have been examined for nematodes only membersof the six aforementioned obtectomeran (Kristensen ampSkalski 1998) families have been found to be positivefor Noctuidonematinae With the exception of Pyralidaeall these families are members of the Macrolepidopteraclade Moreover the positive Noctuidae species representnearly 90 of the total positive species found Thereforewe cannot state that the Noctuidonematinae are notpotentially parasites of most of the lepidopteran familiesbut we can state that at least in the Neotropical Regionthey parasitise the Macrolepidoptera preferentially andamong that clade mostly a single family the Noctuidae(Marti et al 2000)

The Aphelenchida include several different trophictypes with many species being mycetophagous or phy-toparasitic with about one third of known species be-ing entomoparasitic (Hunt 1993) Considering the list ofinsect hosts we cannot hypothesise that the Acugutturi-dae are derived from a single ancestor transferring froma plant to an insect host It seems more realistic to pro-pose several transfers from plant to insect hosts during thehistory of the Acugutturidae A transfer (Acugutturinae)occurred on the Blattodea and another one (Noctuidone-matinae) on the Lepidoptera Another transfer may haveoccurred on the Bombycoidea sensu lato If at least twosuch principal transfers occurred then it is possible thatthe Acugutturinae and Noctuidonematinae would not beconfamilial and that these two taxonomicunits deserve fa-milial rank Cladistic and molecularphylogeneticanalysiswill help to solve that problem The existenceof symbioticbacteria in these nematodes (Marti et al 1995) shouldprovide an independentmeans of analysing these relation-ships

Concerning the relationship between Lepidoptera andits ectoparasitic nematodes the present work highlightsthe following situation a new genus of Noctuidone-matinae parasitises Sphingidae a genuine Bombycoidea(Lemaire amp Minet 1998) Vampyronema parasitises the

Vol 4(4) 2002 501

OG Marti Jr et al

quadri ne noctuid subfamily Catocalinae (sensu lato in-cluding Ophiderinae) Plusiinae an Hadeninae and aNoctuinae (both typical tri ne noctuids) and Noctui-donema spp parasitises another Hadeninae and severalSpodoptera a tri ne noctuid genus previously includedin the Amphipyrinae but now of uncertain subfamilial af- liation (Poole 1995 Kitching amp Rawlins 1998) Exceptfor nematodes on Sphingidae which likely correspond toa well-separated clade whose phylogenetic relationshipto other Noctuidonematinae needs to be speci ed it ap-pears that Vampyronema and Noctuidonema both para-sitise Noctuidae but that Vampyronema parasitises bothquadri ne and tri ne noctuids Noctuidonema possibly aderived form of Noctuidonematinae is restricted to a fewgenera (Spodoptera and Pseudaletia) of tri ne noctuidsthat are not closely related (Mitchell et al 2000)

Recent work has questioned the monophylyof the Noc-tuidae and there is a growing feeling among moleculartaxonomists that the Noctuidae may be paraphyletic sev-eral quadri ne subfamilies including Catocalinae beingmore closely related to Arctiidae and Lymantriidae than totri ne noctuids (Weller et al 1994 Mitchell et al 1997Fang et al 2000 Mitchell et al 2000) The latter groupis considered as monophyleticClearly systematics of theNoctuidonematinaedo not match the tentative new classi- cation of the Noctuidae

This lack of congruence between parasite and hostsystematics leads us to consider that the relationshipsbetween Acugutturidae and their hosts have not beenguided by coevolutionary forces in the sense of Ehrlichand Raven (1964) but by horizontal transfers (Rogers ampMarti 1996) and sequential evolution (Jermy 1984 Fu-tuyma amp McCafferty 1990) The Acugutturidae and es-pecially the Noctuidonematinae have likely experiencedan adaptive radiation on higher Lepidoptera long aftertheir host diversi cation History of the Noctuidonemati-nae has been characterised by host transfer events ontofairly distantly related hosts followed by adaptation tonew hosts leading to speciation processes that now permitus to distinguish easily fairly species-speci c populationsor strains of Vampyronema or Noctuidonema In mothsthese host transfers may have occurred through contami-nation of nematode-free species by nematode eggs juve-niles or adults left on resting or feeding surfaces (leavesor owers) by individualsof infested species The successof such contamination may be linked to the availabilityand accessibility of host intersegmental membranes thinenough to be pierced by nematode stylets a factor that

may have physically and dramatically restricted the hostrange of all Acugutturidae

The present morphological analysis of acugutturid ne-matodes parasitising 13 species of Lepidoptera and onespecies of cockroach revealed a greater degree of diversitythan previously known in this group The genus Vampy-ronema grouping nematodes with simple spicules in-cludes two inferred new species in addition to V dibo-lia and V daptria already known from the LepidopteraThe genus Noctuidonema grouping nematodes of Lepi-doptera with complex spicules includes a new speciesand the nematodes found on Sphingidae probably belongto a new genus The relationshipbetween these nematodesand their hosts appears to be much more complex thanpreviously thought In addition to their bene ts as para-sites of pest Lepidoptera acugutturid nematodes are aninteresting model for parasite-host coevolutionarystudies

Acknowledgements

We thank David Hunt (CABI Bioscience UK) DanielLachaise (CNRS France) Alvin Simmons (USDA Char-leston SC USA) and S Patricia Stock (University ofCalifornia Davis CA USA) for reviewing earlier ver-sions of this paper We thank Jeanette Williams and JohnClayton for collecting cockroaches in St Lucia and mothsin Fiji respectively We thank the following individualsfor identifyingor con rming our identi cations of insectsDavid Nickle (Smithsonian Washington DC USA) forPeriplaneta americana Pierre Zagatti (INRA France)for Pyralidae and Hugo Kons (University of FloridaGainesville FL USA) for Orthodes crenulata

References

ADAMS BJ (1998) Species concepts and evolutionary para-digm in modern nematology Journal of Nematology 30 1-21

ANDERSON RV amp LAUMOND C (1990) NoctuidonemaguyanenseRemillet amp Silvain 1988 morphologie du spiculeredescription du macircle et diagnose amendeacutee du genre Noctui-donema Revue de Neacutematologie 13 433-436

ANDERSON RV amp LAUMOND C (1992) Noctuidonemadaptria n sp (Nematoda Aphelenchoididae)an ectoparasiteof the moth Lesmone porcia (Stoll) Journal of Nematology24 16-22

EHRLICH P amp RAVEN P (1964) Butter ies and plants astudy in evolution Evolution 18 586-608

FANG QQ MITCHELL A REGIER JC MITTER CFRIEDLANDER TP amp POOLE RW (2000) Phylogenetic

502 Nematology

Speciation in the Acugutturidae

Fig 4 Tree diagram for ectoparasitic nematodes from insecthost species (Blattodea and Lepidoptera) The tree was pro-duced using Euclidean distance as the distance measure andcomplete linkage as the amalgamation rule Both graphs areplotted on an X-axis standardised by (DlinkDmax)100 whichrepresents the percentage of the range from the maximum tothe minimum distance in the data A Female nematodes datafor excretory pore stylet length V bulb width a anterior endto anterior intestine and vulva to anterior end were used BMale nematodes data for excretory pore spicule length andstylet length were used Legend (Host Names) pa Periplan-eta americana (St Lucia) pa2 P americana (Guadeloupe) aoAgrapha oxygramma ai Anicla infecta ea Eulepidotis addensmd Mocis disseverans ml M latipes oc Orthodes crenulatapl Perigonia lusca pu Pseudaletia unipuncta sd Spodopteradolichos sf S frugiperda sl S latifascia so S ornithogallixc Xylophanes chiron

Spicule length was not an important character in dis-tinguishing among male Noctuidonema Male nematodeson S dolichos and S frugiperda showed the most sepa-ration with moderate overlap whereas males on S or-nithogalli appeared to be most similar to those on S

dolichos and males on S latifascia appeared most similarto those on S frugiperda Results from both genders agreethat nematodes on Pseudaletia are different from thoseon Spodoptera and that the nematodes on S dolichos andS frugiperda are the most different with the nematodeson S latifascia and S ornithogallibeing intermediate be-tween these two Overall the results do not support theexistence of separate species of Noctuidonema on hostsof the genus Spodoptera (Fig 4 Table 6)

These results provide evidence that nematodes on Ainfecta A oxygramma and E addens are different fromeach other and from the remaining groups having simplespicules Nematodes on M disseverans and O crenulatawere not distinct from V dibolia the species parasitisingM latipes Nematodes on M disseverans differed signif-icantly from V dibolia in only six morphological charac-ters in the females and ten in the males The differenceswere small and can be attributed to adaptation of conspe-ci c nematodes to parasitism of closely related sympatrichost species

Comparisonof the description of V daptria to measure-ments of other Vampyronema in this study indicates that itdiffers in stylet length and spicule shape V daptria hasthe shortest stylet length observed in the genus rangingfrom 53-60 sup1m in males and 59-69 sup1m in females whichis shorter than the stylet length of nematodes on A in-fecta (Fig 3) In addition spicules of V daptria taper to ne points whereas in Vampyronema from other hosts thespicule tips are blunt V daptria on L porcia appears tobe a distinct species and we therefore do not assign any ofthe Vampyronema from this study to it

Nematodes on the St Lucia and Guadeloupe popula-tions of Periplaneta americana are not well differenti-ated from each other based on the suite of characters thatwere examined (Fig 1A B) Although there are a fewsigni cantly different characters between the two popula-tions the differences probably do not represent interspe-ci c variation In addition the characters of both popula-tions are in agreement with the published description ofAcugutturus parasiticus (Hunt 1980)

Examination of additional specimens of ectoparasiticnematodes from other species of Lepidoptera showed thatmost were Vampyronema whereas Noctuidonema wasfound only on moths from two genera Pseudaletia andSpodoptera (Table 7)

Vol 4(4) 2002 499

OG Marti Jr et al

Table 7 Association of quadrine and trine Noctuidae with Vampyronema and Noctuidonema

Host Taxon Hind Wing Venation1 Locality2 Nematodes Genus

NoctuidaeCatocalinae

Anomis oedema Gueneacutee quadri ne FG 4 1 VampyronemaBaniana sp quadri ne FG 1 0 VampyronemaBaniana ostia Druce quadri ne FG 7 0 VampyronemaEuglyphis sp quadri ne FG 1 0 VampyronemaEulepidotis addens (Walker) quadri ne GU 5 2 VampyronemaItomia opisthographa (Gueneacutee) quadri ne FG 5 1 VampyronemaLesmone formularis (Geyer) quadri ne FG 1 0 VampyronemaLesmone porcia (Stoll) quadri ne GU ndash3 Vampyronema3

Melipotis fasciolaris (Huumlbner) quadri ne FG 0 1 VampyronemaMetalectra sp quadri ne FG 4 10 VampyronemaMetria sp quadri ne FG 0 1 VampyronemaMocis dif uens (Gueneacutee) quadri ne FG 1 1 Vampyronema4

Mocis disseverans (Walker) quadri ne GA 9 4 Vampyronema4

Mocis latipes Gueneacutee quadri ne GA 9 16 Vampyronema4

Perasia ora (Cramer)5 quadri ne FG 10 0 VampyronemaSelenisa sp quadri ne FG 4 0 VampyronemaZale ctilis (Gueneacutee) quadri ne FG 4 1 Vampyronema

PlusiinaeAgrapha oxygramma (Geyer) quadri ne GA 13 7 VampyronemaArgyrogramma verruca (Fabricius) quadri ne FG 0 2 VampyronemaChrysodeixis chalcites (Esper) quadri ne Fiji 8 7 Vampyronema

AmphipyrinaeSpodoptera androgea (Stoll) tri ne FG 4 0 Noctuidonema6

Spodoptera dolichos (Fabricius) tri ne GA 5 10 Noctuidonema6

Spodoptera frugiperda (JE Smith) tri ne GA 17 11 Noctuidonema6

Spodoptera latifascia (Walker) tri ne GA 12 6 Noctuidonema6

Spodoptera litura (Fabricius) tri ne Fiji 15 7 Noctuidonema6

Spodoptera ornithogalli (Gueneacutee) tri ne GA 10 10 Noctuidonema6

HadeninaeOrthodes crenulata (Butler) tri ne GA 9 4 VampyronemaPseudaletia unipuncta (Haworth) tri ne GA 18 9 Noctuidonema

NoctuinaeAnicla infecta (Ochsenheimer) tri ne GA 15 12 VampyronemaTandilia rodea (Schaus) tri ne FG 2 3 Vampyronema

PyralidaePyraustinae

Lamprosema moccalis Schaus ndash FG 4 0 VampyronemaSphingidae

MacroglossinaeErinnys obscura (Fabricius) ndash FG 22 2 Vampyronema7

Perigonia lusca Fabricius ndash GU 19 4 Vampyronema7

Xylophanes chiron Drury ndash GU 8 0 Vampyronema7

1) Quadri ne and tri ne hindwing venation applies to Noctuidae only Pyralidae and Sphingidae are included for completeness thestatus of Plusiinae as quadri nes is ambiguous because it groups together mostly species having quadri ne wing venation yet is oftenincluded with chie y tri ne subfamilies (Kitching amp Rawlins 1998) 2 Fiji Fiji Islands FG French Guiana GU Guadeloupe GAGeorgia USA 3 V daptria Anderson amp Laumond 1992 designationbased on Hunt (1993) 4 V dibolia 5 Synonymous with Nymbisarcuata Walker (in Rogers et al 1990a and Simmons amp Rogers 1996) 6 N guyanense 7 Probably represents an undescribed genus

500 Nematology

Speciation in the Acugutturidae

Discussion

The position of the excretory pore posterior to thebulb the comparatively short length of the stylet thecomparatively long tail and the simple rose-thorn-shapedspicules are suf cient to clearly separate Acugutturus aparasite of Blattodea from Vampyronema and Noctuido-nema parasites of Lepidoptera Sampled populations re-veal four inferred monophyleticgroups Acugutturus par-asiticus from two West Indian populations of Periplan-eta americana a new genus on Sphingidae Noctuido-nema guyanense from Spodoptera spp with nematodeson P unipuncta as a probable new species of Noctuido-nema and Vampyronema spp on the remaining hosts inthis study with the nematodes on A infecta representinga new species of Vampyronema

Each inferred monophyletic group can be de ned byseveral unique combinations of characters quantitativeand qualitative that satisfy the minimal requirements forspecies diagnosis (Wheeler 1999 Wheeler amp Platnick2000) Such characters appear to represent lineage inde-pendence via autapomorphies hypotheses ideally testedvia outgroup comparison Such tests while more philo-sophically satisfying (Adams 1998 Wiley amp Mayden2000) are beyond the scope of the present paper

According to Kitching and Rawlins (1998) differencesin hindwing venation broadly divide the Noctuidae intotwo phenetic groups tri nes and quadri nes Catocalinaeand Plusiinae are quadri nes (but see comment 1 in Ta-ble 7) whereas the other noctuid subfamilies mentionedin this study are tri nes Complex spicules probably rep-resent a derived condition found thus far only in nema-todes on two genera of tri ne noctuids Pseudaletia andSpodoptera while the simple spicules of Vampyronemaprobably represent a plesiomorphic state found in bothtri ne and quadri ne noctuids The extremely long styletsof nematodes from Sphingidae may have arisen in re-sponse to parasitisation of large moths

The Acugutturidae are associated with two very dis-tantly related insect orders Blattodea and Lepidoptera(Endopterygota) There is no record of occurrence ofthe Acugutturidae on any other insect order Among theLepidoptera the Acugutturidae and more precisely theNoctuidonematinae are known from only six familiesGeometridae Lasiocampidae Noctuidae NotodontidaePyralidae and Sphingidae (Marti et al 2000) most ofwhich are not closely related with the exceptionof Lasio-campidae and Sphingidae members of the Bombycoidea(and their relatives) clade (Lemaire amp Minet 1998) and

Notodontidaeand Noctuidae members of the Noctuoideasuperfamily (Kitching amp Rawlins 1998)

One explanation for such a surprising distribution ofthe Acugutturidae on their hosts may be that availabledata are too sparse due to insuf cient sampling of otherlepidopteran families While this may be true the studyconducted by Marti et al (2000) shows that there isa clear speci city in nematode host preference Martiet al (2000) updated the list of lepidopteran hosts ofNoctuidonematinae It seems clear that although manyLepidoptera species belonging to 11 families and 115genera have been examined for nematodes only membersof the six aforementioned obtectomeran (Kristensen ampSkalski 1998) families have been found to be positivefor Noctuidonematinae With the exception of Pyralidaeall these families are members of the Macrolepidopteraclade Moreover the positive Noctuidae species representnearly 90 of the total positive species found Thereforewe cannot state that the Noctuidonematinae are notpotentially parasites of most of the lepidopteran familiesbut we can state that at least in the Neotropical Regionthey parasitise the Macrolepidoptera preferentially andamong that clade mostly a single family the Noctuidae(Marti et al 2000)

The Aphelenchida include several different trophictypes with many species being mycetophagous or phy-toparasitic with about one third of known species be-ing entomoparasitic (Hunt 1993) Considering the list ofinsect hosts we cannot hypothesise that the Acugutturi-dae are derived from a single ancestor transferring froma plant to an insect host It seems more realistic to pro-pose several transfers from plant to insect hosts during thehistory of the Acugutturidae A transfer (Acugutturinae)occurred on the Blattodea and another one (Noctuidone-matinae) on the Lepidoptera Another transfer may haveoccurred on the Bombycoidea sensu lato If at least twosuch principal transfers occurred then it is possible thatthe Acugutturinae and Noctuidonematinae would not beconfamilial and that these two taxonomicunits deserve fa-milial rank Cladistic and molecularphylogeneticanalysiswill help to solve that problem The existenceof symbioticbacteria in these nematodes (Marti et al 1995) shouldprovide an independentmeans of analysing these relation-ships

Concerning the relationship between Lepidoptera andits ectoparasitic nematodes the present work highlightsthe following situation a new genus of Noctuidone-matinae parasitises Sphingidae a genuine Bombycoidea(Lemaire amp Minet 1998) Vampyronema parasitises the

Vol 4(4) 2002 501

OG Marti Jr et al

quadri ne noctuid subfamily Catocalinae (sensu lato in-cluding Ophiderinae) Plusiinae an Hadeninae and aNoctuinae (both typical tri ne noctuids) and Noctui-donema spp parasitises another Hadeninae and severalSpodoptera a tri ne noctuid genus previously includedin the Amphipyrinae but now of uncertain subfamilial af- liation (Poole 1995 Kitching amp Rawlins 1998) Exceptfor nematodes on Sphingidae which likely correspond toa well-separated clade whose phylogenetic relationshipto other Noctuidonematinae needs to be speci ed it ap-pears that Vampyronema and Noctuidonema both para-sitise Noctuidae but that Vampyronema parasitises bothquadri ne and tri ne noctuids Noctuidonema possibly aderived form of Noctuidonematinae is restricted to a fewgenera (Spodoptera and Pseudaletia) of tri ne noctuidsthat are not closely related (Mitchell et al 2000)

Recent work has questioned the monophylyof the Noc-tuidae and there is a growing feeling among moleculartaxonomists that the Noctuidae may be paraphyletic sev-eral quadri ne subfamilies including Catocalinae beingmore closely related to Arctiidae and Lymantriidae than totri ne noctuids (Weller et al 1994 Mitchell et al 1997Fang et al 2000 Mitchell et al 2000) The latter groupis considered as monophyleticClearly systematics of theNoctuidonematinaedo not match the tentative new classi- cation of the Noctuidae

This lack of congruence between parasite and hostsystematics leads us to consider that the relationshipsbetween Acugutturidae and their hosts have not beenguided by coevolutionary forces in the sense of Ehrlichand Raven (1964) but by horizontal transfers (Rogers ampMarti 1996) and sequential evolution (Jermy 1984 Fu-tuyma amp McCafferty 1990) The Acugutturidae and es-pecially the Noctuidonematinae have likely experiencedan adaptive radiation on higher Lepidoptera long aftertheir host diversi cation History of the Noctuidonemati-nae has been characterised by host transfer events ontofairly distantly related hosts followed by adaptation tonew hosts leading to speciation processes that now permitus to distinguish easily fairly species-speci c populationsor strains of Vampyronema or Noctuidonema In mothsthese host transfers may have occurred through contami-nation of nematode-free species by nematode eggs juve-niles or adults left on resting or feeding surfaces (leavesor owers) by individualsof infested species The successof such contamination may be linked to the availabilityand accessibility of host intersegmental membranes thinenough to be pierced by nematode stylets a factor that

may have physically and dramatically restricted the hostrange of all Acugutturidae

The present morphological analysis of acugutturid ne-matodes parasitising 13 species of Lepidoptera and onespecies of cockroach revealed a greater degree of diversitythan previously known in this group The genus Vampy-ronema grouping nematodes with simple spicules in-cludes two inferred new species in addition to V dibo-lia and V daptria already known from the LepidopteraThe genus Noctuidonema grouping nematodes of Lepi-doptera with complex spicules includes a new speciesand the nematodes found on Sphingidae probably belongto a new genus The relationshipbetween these nematodesand their hosts appears to be much more complex thanpreviously thought In addition to their bene ts as para-sites of pest Lepidoptera acugutturid nematodes are aninteresting model for parasite-host coevolutionarystudies

Acknowledgements

We thank David Hunt (CABI Bioscience UK) DanielLachaise (CNRS France) Alvin Simmons (USDA Char-leston SC USA) and S Patricia Stock (University ofCalifornia Davis CA USA) for reviewing earlier ver-sions of this paper We thank Jeanette Williams and JohnClayton for collecting cockroaches in St Lucia and mothsin Fiji respectively We thank the following individualsfor identifyingor con rming our identi cations of insectsDavid Nickle (Smithsonian Washington DC USA) forPeriplaneta americana Pierre Zagatti (INRA France)for Pyralidae and Hugo Kons (University of FloridaGainesville FL USA) for Orthodes crenulata

References

ADAMS BJ (1998) Species concepts and evolutionary para-digm in modern nematology Journal of Nematology 30 1-21

ANDERSON RV amp LAUMOND C (1990) NoctuidonemaguyanenseRemillet amp Silvain 1988 morphologie du spiculeredescription du macircle et diagnose amendeacutee du genre Noctui-donema Revue de Neacutematologie 13 433-436

ANDERSON RV amp LAUMOND C (1992) Noctuidonemadaptria n sp (Nematoda Aphelenchoididae)an ectoparasiteof the moth Lesmone porcia (Stoll) Journal of Nematology24 16-22

EHRLICH P amp RAVEN P (1964) Butter ies and plants astudy in evolution Evolution 18 586-608

FANG QQ MITCHELL A REGIER JC MITTER CFRIEDLANDER TP amp POOLE RW (2000) Phylogenetic

502 Nematology

OG Marti Jr et al

Table 7 Association of quadrine and trine Noctuidae with Vampyronema and Noctuidonema

Host Taxon Hind Wing Venation1 Locality2 Nematodes Genus

NoctuidaeCatocalinae

Anomis oedema Gueneacutee quadri ne FG 4 1 VampyronemaBaniana sp quadri ne FG 1 0 VampyronemaBaniana ostia Druce quadri ne FG 7 0 VampyronemaEuglyphis sp quadri ne FG 1 0 VampyronemaEulepidotis addens (Walker) quadri ne GU 5 2 VampyronemaItomia opisthographa (Gueneacutee) quadri ne FG 5 1 VampyronemaLesmone formularis (Geyer) quadri ne FG 1 0 VampyronemaLesmone porcia (Stoll) quadri ne GU ndash3 Vampyronema3

Melipotis fasciolaris (Huumlbner) quadri ne FG 0 1 VampyronemaMetalectra sp quadri ne FG 4 10 VampyronemaMetria sp quadri ne FG 0 1 VampyronemaMocis dif uens (Gueneacutee) quadri ne FG 1 1 Vampyronema4

Mocis disseverans (Walker) quadri ne GA 9 4 Vampyronema4

Mocis latipes Gueneacutee quadri ne GA 9 16 Vampyronema4

Perasia ora (Cramer)5 quadri ne FG 10 0 VampyronemaSelenisa sp quadri ne FG 4 0 VampyronemaZale ctilis (Gueneacutee) quadri ne FG 4 1 Vampyronema

PlusiinaeAgrapha oxygramma (Geyer) quadri ne GA 13 7 VampyronemaArgyrogramma verruca (Fabricius) quadri ne FG 0 2 VampyronemaChrysodeixis chalcites (Esper) quadri ne Fiji 8 7 Vampyronema

AmphipyrinaeSpodoptera androgea (Stoll) tri ne FG 4 0 Noctuidonema6

Spodoptera dolichos (Fabricius) tri ne GA 5 10 Noctuidonema6

Spodoptera frugiperda (JE Smith) tri ne GA 17 11 Noctuidonema6

Spodoptera latifascia (Walker) tri ne GA 12 6 Noctuidonema6

Spodoptera litura (Fabricius) tri ne Fiji 15 7 Noctuidonema6

Spodoptera ornithogalli (Gueneacutee) tri ne GA 10 10 Noctuidonema6

HadeninaeOrthodes crenulata (Butler) tri ne GA 9 4 VampyronemaPseudaletia unipuncta (Haworth) tri ne GA 18 9 Noctuidonema

NoctuinaeAnicla infecta (Ochsenheimer) tri ne GA 15 12 VampyronemaTandilia rodea (Schaus) tri ne FG 2 3 Vampyronema

PyralidaePyraustinae

Lamprosema moccalis Schaus ndash FG 4 0 VampyronemaSphingidae

MacroglossinaeErinnys obscura (Fabricius) ndash FG 22 2 Vampyronema7

Perigonia lusca Fabricius ndash GU 19 4 Vampyronema7

Xylophanes chiron Drury ndash GU 8 0 Vampyronema7

1) Quadri ne and tri ne hindwing venation applies to Noctuidae only Pyralidae and Sphingidae are included for completeness thestatus of Plusiinae as quadri nes is ambiguous because it groups together mostly species having quadri ne wing venation yet is oftenincluded with chie y tri ne subfamilies (Kitching amp Rawlins 1998) 2 Fiji Fiji Islands FG French Guiana GU Guadeloupe GAGeorgia USA 3 V daptria Anderson amp Laumond 1992 designationbased on Hunt (1993) 4 V dibolia 5 Synonymous with Nymbisarcuata Walker (in Rogers et al 1990a and Simmons amp Rogers 1996) 6 N guyanense 7 Probably represents an undescribed genus

500 Nematology

Speciation in the Acugutturidae

Discussion

The position of the excretory pore posterior to thebulb the comparatively short length of the stylet thecomparatively long tail and the simple rose-thorn-shapedspicules are suf cient to clearly separate Acugutturus aparasite of Blattodea from Vampyronema and Noctuido-nema parasites of Lepidoptera Sampled populations re-veal four inferred monophyleticgroups Acugutturus par-asiticus from two West Indian populations of Periplan-eta americana a new genus on Sphingidae Noctuido-nema guyanense from Spodoptera spp with nematodeson P unipuncta as a probable new species of Noctuido-nema and Vampyronema spp on the remaining hosts inthis study with the nematodes on A infecta representinga new species of Vampyronema

Each inferred monophyletic group can be de ned byseveral unique combinations of characters quantitativeand qualitative that satisfy the minimal requirements forspecies diagnosis (Wheeler 1999 Wheeler amp Platnick2000) Such characters appear to represent lineage inde-pendence via autapomorphies hypotheses ideally testedvia outgroup comparison Such tests while more philo-sophically satisfying (Adams 1998 Wiley amp Mayden2000) are beyond the scope of the present paper

According to Kitching and Rawlins (1998) differencesin hindwing venation broadly divide the Noctuidae intotwo phenetic groups tri nes and quadri nes Catocalinaeand Plusiinae are quadri nes (but see comment 1 in Ta-ble 7) whereas the other noctuid subfamilies mentionedin this study are tri nes Complex spicules probably rep-resent a derived condition found thus far only in nema-todes on two genera of tri ne noctuids Pseudaletia andSpodoptera while the simple spicules of Vampyronemaprobably represent a plesiomorphic state found in bothtri ne and quadri ne noctuids The extremely long styletsof nematodes from Sphingidae may have arisen in re-sponse to parasitisation of large moths

The Acugutturidae are associated with two very dis-tantly related insect orders Blattodea and Lepidoptera(Endopterygota) There is no record of occurrence ofthe Acugutturidae on any other insect order Among theLepidoptera the Acugutturidae and more precisely theNoctuidonematinae are known from only six familiesGeometridae Lasiocampidae Noctuidae NotodontidaePyralidae and Sphingidae (Marti et al 2000) most ofwhich are not closely related with the exceptionof Lasio-campidae and Sphingidae members of the Bombycoidea(and their relatives) clade (Lemaire amp Minet 1998) and

Notodontidaeand Noctuidae members of the Noctuoideasuperfamily (Kitching amp Rawlins 1998)

One explanation for such a surprising distribution ofthe Acugutturidae on their hosts may be that availabledata are too sparse due to insuf cient sampling of otherlepidopteran families While this may be true the studyconducted by Marti et al (2000) shows that there isa clear speci city in nematode host preference Martiet al (2000) updated the list of lepidopteran hosts ofNoctuidonematinae It seems clear that although manyLepidoptera species belonging to 11 families and 115genera have been examined for nematodes only membersof the six aforementioned obtectomeran (Kristensen ampSkalski 1998) families have been found to be positivefor Noctuidonematinae With the exception of Pyralidaeall these families are members of the Macrolepidopteraclade Moreover the positive Noctuidae species representnearly 90 of the total positive species found Thereforewe cannot state that the Noctuidonematinae are notpotentially parasites of most of the lepidopteran familiesbut we can state that at least in the Neotropical Regionthey parasitise the Macrolepidoptera preferentially andamong that clade mostly a single family the Noctuidae(Marti et al 2000)

The Aphelenchida include several different trophictypes with many species being mycetophagous or phy-toparasitic with about one third of known species be-ing entomoparasitic (Hunt 1993) Considering the list ofinsect hosts we cannot hypothesise that the Acugutturi-dae are derived from a single ancestor transferring froma plant to an insect host It seems more realistic to pro-pose several transfers from plant to insect hosts during thehistory of the Acugutturidae A transfer (Acugutturinae)occurred on the Blattodea and another one (Noctuidone-matinae) on the Lepidoptera Another transfer may haveoccurred on the Bombycoidea sensu lato If at least twosuch principal transfers occurred then it is possible thatthe Acugutturinae and Noctuidonematinae would not beconfamilial and that these two taxonomicunits deserve fa-milial rank Cladistic and molecularphylogeneticanalysiswill help to solve that problem The existenceof symbioticbacteria in these nematodes (Marti et al 1995) shouldprovide an independentmeans of analysing these relation-ships

Concerning the relationship between Lepidoptera andits ectoparasitic nematodes the present work highlightsthe following situation a new genus of Noctuidone-matinae parasitises Sphingidae a genuine Bombycoidea(Lemaire amp Minet 1998) Vampyronema parasitises the

Vol 4(4) 2002 501

OG Marti Jr et al

quadri ne noctuid subfamily Catocalinae (sensu lato in-cluding Ophiderinae) Plusiinae an Hadeninae and aNoctuinae (both typical tri ne noctuids) and Noctui-donema spp parasitises another Hadeninae and severalSpodoptera a tri ne noctuid genus previously includedin the Amphipyrinae but now of uncertain subfamilial af- liation (Poole 1995 Kitching amp Rawlins 1998) Exceptfor nematodes on Sphingidae which likely correspond toa well-separated clade whose phylogenetic relationshipto other Noctuidonematinae needs to be speci ed it ap-pears that Vampyronema and Noctuidonema both para-sitise Noctuidae but that Vampyronema parasitises bothquadri ne and tri ne noctuids Noctuidonema possibly aderived form of Noctuidonematinae is restricted to a fewgenera (Spodoptera and Pseudaletia) of tri ne noctuidsthat are not closely related (Mitchell et al 2000)

Recent work has questioned the monophylyof the Noc-tuidae and there is a growing feeling among moleculartaxonomists that the Noctuidae may be paraphyletic sev-eral quadri ne subfamilies including Catocalinae beingmore closely related to Arctiidae and Lymantriidae than totri ne noctuids (Weller et al 1994 Mitchell et al 1997Fang et al 2000 Mitchell et al 2000) The latter groupis considered as monophyleticClearly systematics of theNoctuidonematinaedo not match the tentative new classi- cation of the Noctuidae

This lack of congruence between parasite and hostsystematics leads us to consider that the relationshipsbetween Acugutturidae and their hosts have not beenguided by coevolutionary forces in the sense of Ehrlichand Raven (1964) but by horizontal transfers (Rogers ampMarti 1996) and sequential evolution (Jermy 1984 Fu-tuyma amp McCafferty 1990) The Acugutturidae and es-pecially the Noctuidonematinae have likely experiencedan adaptive radiation on higher Lepidoptera long aftertheir host diversi cation History of the Noctuidonemati-nae has been characterised by host transfer events ontofairly distantly related hosts followed by adaptation tonew hosts leading to speciation processes that now permitus to distinguish easily fairly species-speci c populationsor strains of Vampyronema or Noctuidonema In mothsthese host transfers may have occurred through contami-nation of nematode-free species by nematode eggs juve-niles or adults left on resting or feeding surfaces (leavesor owers) by individualsof infested species The successof such contamination may be linked to the availabilityand accessibility of host intersegmental membranes thinenough to be pierced by nematode stylets a factor that

may have physically and dramatically restricted the hostrange of all Acugutturidae

The present morphological analysis of acugutturid ne-matodes parasitising 13 species of Lepidoptera and onespecies of cockroach revealed a greater degree of diversitythan previously known in this group The genus Vampy-ronema grouping nematodes with simple spicules in-cludes two inferred new species in addition to V dibo-lia and V daptria already known from the LepidopteraThe genus Noctuidonema grouping nematodes of Lepi-doptera with complex spicules includes a new speciesand the nematodes found on Sphingidae probably belongto a new genus The relationshipbetween these nematodesand their hosts appears to be much more complex thanpreviously thought In addition to their bene ts as para-sites of pest Lepidoptera acugutturid nematodes are aninteresting model for parasite-host coevolutionarystudies

Acknowledgements

We thank David Hunt (CABI Bioscience UK) DanielLachaise (CNRS France) Alvin Simmons (USDA Char-leston SC USA) and S Patricia Stock (University ofCalifornia Davis CA USA) for reviewing earlier ver-sions of this paper We thank Jeanette Williams and JohnClayton for collecting cockroaches in St Lucia and mothsin Fiji respectively We thank the following individualsfor identifyingor con rming our identi cations of insectsDavid Nickle (Smithsonian Washington DC USA) forPeriplaneta americana Pierre Zagatti (INRA France)for Pyralidae and Hugo Kons (University of FloridaGainesville FL USA) for Orthodes crenulata

References

ADAMS BJ (1998) Species concepts and evolutionary para-digm in modern nematology Journal of Nematology 30 1-21

ANDERSON RV amp LAUMOND C (1990) NoctuidonemaguyanenseRemillet amp Silvain 1988 morphologie du spiculeredescription du macircle et diagnose amendeacutee du genre Noctui-donema Revue de Neacutematologie 13 433-436

ANDERSON RV amp LAUMOND C (1992) Noctuidonemadaptria n sp (Nematoda Aphelenchoididae)an ectoparasiteof the moth Lesmone porcia (Stoll) Journal of Nematology24 16-22

EHRLICH P amp RAVEN P (1964) Butter ies and plants astudy in evolution Evolution 18 586-608

FANG QQ MITCHELL A REGIER JC MITTER CFRIEDLANDER TP amp POOLE RW (2000) Phylogenetic

502 Nematology

Speciation in the Acugutturidae

Discussion

The position of the excretory pore posterior to thebulb the comparatively short length of the stylet thecomparatively long tail and the simple rose-thorn-shapedspicules are suf cient to clearly separate Acugutturus aparasite of Blattodea from Vampyronema and Noctuido-nema parasites of Lepidoptera Sampled populations re-veal four inferred monophyleticgroups Acugutturus par-asiticus from two West Indian populations of Periplan-eta americana a new genus on Sphingidae Noctuido-nema guyanense from Spodoptera spp with nematodeson P unipuncta as a probable new species of Noctuido-nema and Vampyronema spp on the remaining hosts inthis study with the nematodes on A infecta representinga new species of Vampyronema

Each inferred monophyletic group can be de ned byseveral unique combinations of characters quantitativeand qualitative that satisfy the minimal requirements forspecies diagnosis (Wheeler 1999 Wheeler amp Platnick2000) Such characters appear to represent lineage inde-pendence via autapomorphies hypotheses ideally testedvia outgroup comparison Such tests while more philo-sophically satisfying (Adams 1998 Wiley amp Mayden2000) are beyond the scope of the present paper

According to Kitching and Rawlins (1998) differencesin hindwing venation broadly divide the Noctuidae intotwo phenetic groups tri nes and quadri nes Catocalinaeand Plusiinae are quadri nes (but see comment 1 in Ta-ble 7) whereas the other noctuid subfamilies mentionedin this study are tri nes Complex spicules probably rep-resent a derived condition found thus far only in nema-todes on two genera of tri ne noctuids Pseudaletia andSpodoptera while the simple spicules of Vampyronemaprobably represent a plesiomorphic state found in bothtri ne and quadri ne noctuids The extremely long styletsof nematodes from Sphingidae may have arisen in re-sponse to parasitisation of large moths

The Acugutturidae are associated with two very dis-tantly related insect orders Blattodea and Lepidoptera(Endopterygota) There is no record of occurrence ofthe Acugutturidae on any other insect order Among theLepidoptera the Acugutturidae and more precisely theNoctuidonematinae are known from only six familiesGeometridae Lasiocampidae Noctuidae NotodontidaePyralidae and Sphingidae (Marti et al 2000) most ofwhich are not closely related with the exceptionof Lasio-campidae and Sphingidae members of the Bombycoidea(and their relatives) clade (Lemaire amp Minet 1998) and

Notodontidaeand Noctuidae members of the Noctuoideasuperfamily (Kitching amp Rawlins 1998)

One explanation for such a surprising distribution ofthe Acugutturidae on their hosts may be that availabledata are too sparse due to insuf cient sampling of otherlepidopteran families While this may be true the studyconducted by Marti et al (2000) shows that there isa clear speci city in nematode host preference Martiet al (2000) updated the list of lepidopteran hosts ofNoctuidonematinae It seems clear that although manyLepidoptera species belonging to 11 families and 115genera have been examined for nematodes only membersof the six aforementioned obtectomeran (Kristensen ampSkalski 1998) families have been found to be positivefor Noctuidonematinae With the exception of Pyralidaeall these families are members of the Macrolepidopteraclade Moreover the positive Noctuidae species representnearly 90 of the total positive species found Thereforewe cannot state that the Noctuidonematinae are notpotentially parasites of most of the lepidopteran familiesbut we can state that at least in the Neotropical Regionthey parasitise the Macrolepidoptera preferentially andamong that clade mostly a single family the Noctuidae(Marti et al 2000)

The Aphelenchida include several different trophictypes with many species being mycetophagous or phy-toparasitic with about one third of known species be-ing entomoparasitic (Hunt 1993) Considering the list ofinsect hosts we cannot hypothesise that the Acugutturi-dae are derived from a single ancestor transferring froma plant to an insect host It seems more realistic to pro-pose several transfers from plant to insect hosts during thehistory of the Acugutturidae A transfer (Acugutturinae)occurred on the Blattodea and another one (Noctuidone-matinae) on the Lepidoptera Another transfer may haveoccurred on the Bombycoidea sensu lato If at least twosuch principal transfers occurred then it is possible thatthe Acugutturinae and Noctuidonematinae would not beconfamilial and that these two taxonomicunits deserve fa-milial rank Cladistic and molecularphylogeneticanalysiswill help to solve that problem The existenceof symbioticbacteria in these nematodes (Marti et al 1995) shouldprovide an independentmeans of analysing these relation-ships

Concerning the relationship between Lepidoptera andits ectoparasitic nematodes the present work highlightsthe following situation a new genus of Noctuidone-matinae parasitises Sphingidae a genuine Bombycoidea(Lemaire amp Minet 1998) Vampyronema parasitises the

Vol 4(4) 2002 501

OG Marti Jr et al

quadri ne noctuid subfamily Catocalinae (sensu lato in-cluding Ophiderinae) Plusiinae an Hadeninae and aNoctuinae (both typical tri ne noctuids) and Noctui-donema spp parasitises another Hadeninae and severalSpodoptera a tri ne noctuid genus previously includedin the Amphipyrinae but now of uncertain subfamilial af- liation (Poole 1995 Kitching amp Rawlins 1998) Exceptfor nematodes on Sphingidae which likely correspond toa well-separated clade whose phylogenetic relationshipto other Noctuidonematinae needs to be speci ed it ap-pears that Vampyronema and Noctuidonema both para-sitise Noctuidae but that Vampyronema parasitises bothquadri ne and tri ne noctuids Noctuidonema possibly aderived form of Noctuidonematinae is restricted to a fewgenera (Spodoptera and Pseudaletia) of tri ne noctuidsthat are not closely related (Mitchell et al 2000)

Recent work has questioned the monophylyof the Noc-tuidae and there is a growing feeling among moleculartaxonomists that the Noctuidae may be paraphyletic sev-eral quadri ne subfamilies including Catocalinae beingmore closely related to Arctiidae and Lymantriidae than totri ne noctuids (Weller et al 1994 Mitchell et al 1997Fang et al 2000 Mitchell et al 2000) The latter groupis considered as monophyleticClearly systematics of theNoctuidonematinaedo not match the tentative new classi- cation of the Noctuidae

This lack of congruence between parasite and hostsystematics leads us to consider that the relationshipsbetween Acugutturidae and their hosts have not beenguided by coevolutionary forces in the sense of Ehrlichand Raven (1964) but by horizontal transfers (Rogers ampMarti 1996) and sequential evolution (Jermy 1984 Fu-tuyma amp McCafferty 1990) The Acugutturidae and es-pecially the Noctuidonematinae have likely experiencedan adaptive radiation on higher Lepidoptera long aftertheir host diversi cation History of the Noctuidonemati-nae has been characterised by host transfer events ontofairly distantly related hosts followed by adaptation tonew hosts leading to speciation processes that now permitus to distinguish easily fairly species-speci c populationsor strains of Vampyronema or Noctuidonema In mothsthese host transfers may have occurred through contami-nation of nematode-free species by nematode eggs juve-niles or adults left on resting or feeding surfaces (leavesor owers) by individualsof infested species The successof such contamination may be linked to the availabilityand accessibility of host intersegmental membranes thinenough to be pierced by nematode stylets a factor that

may have physically and dramatically restricted the hostrange of all Acugutturidae

The present morphological analysis of acugutturid ne-matodes parasitising 13 species of Lepidoptera and onespecies of cockroach revealed a greater degree of diversitythan previously known in this group The genus Vampy-ronema grouping nematodes with simple spicules in-cludes two inferred new species in addition to V dibo-lia and V daptria already known from the LepidopteraThe genus Noctuidonema grouping nematodes of Lepi-doptera with complex spicules includes a new speciesand the nematodes found on Sphingidae probably belongto a new genus The relationshipbetween these nematodesand their hosts appears to be much more complex thanpreviously thought In addition to their bene ts as para-sites of pest Lepidoptera acugutturid nematodes are aninteresting model for parasite-host coevolutionarystudies

Acknowledgements

We thank David Hunt (CABI Bioscience UK) DanielLachaise (CNRS France) Alvin Simmons (USDA Char-leston SC USA) and S Patricia Stock (University ofCalifornia Davis CA USA) for reviewing earlier ver-sions of this paper We thank Jeanette Williams and JohnClayton for collecting cockroaches in St Lucia and mothsin Fiji respectively We thank the following individualsfor identifyingor con rming our identi cations of insectsDavid Nickle (Smithsonian Washington DC USA) forPeriplaneta americana Pierre Zagatti (INRA France)for Pyralidae and Hugo Kons (University of FloridaGainesville FL USA) for Orthodes crenulata

References

ADAMS BJ (1998) Species concepts and evolutionary para-digm in modern nematology Journal of Nematology 30 1-21

ANDERSON RV amp LAUMOND C (1990) NoctuidonemaguyanenseRemillet amp Silvain 1988 morphologie du spiculeredescription du macircle et diagnose amendeacutee du genre Noctui-donema Revue de Neacutematologie 13 433-436

ANDERSON RV amp LAUMOND C (1992) Noctuidonemadaptria n sp (Nematoda Aphelenchoididae)an ectoparasiteof the moth Lesmone porcia (Stoll) Journal of Nematology24 16-22

EHRLICH P amp RAVEN P (1964) Butter ies and plants astudy in evolution Evolution 18 586-608

FANG QQ MITCHELL A REGIER JC MITTER CFRIEDLANDER TP amp POOLE RW (2000) Phylogenetic

502 Nematology

OG Marti Jr et al

quadri ne noctuid subfamily Catocalinae (sensu lato in-cluding Ophiderinae) Plusiinae an Hadeninae and aNoctuinae (both typical tri ne noctuids) and Noctui-donema spp parasitises another Hadeninae and severalSpodoptera a tri ne noctuid genus previously includedin the Amphipyrinae but now of uncertain subfamilial af- liation (Poole 1995 Kitching amp Rawlins 1998) Exceptfor nematodes on Sphingidae which likely correspond toa well-separated clade whose phylogenetic relationshipto other Noctuidonematinae needs to be speci ed it ap-pears that Vampyronema and Noctuidonema both para-sitise Noctuidae but that Vampyronema parasitises bothquadri ne and tri ne noctuids Noctuidonema possibly aderived form of Noctuidonematinae is restricted to a fewgenera (Spodoptera and Pseudaletia) of tri ne noctuidsthat are not closely related (Mitchell et al 2000)

Recent work has questioned the monophylyof the Noc-tuidae and there is a growing feeling among moleculartaxonomists that the Noctuidae may be paraphyletic sev-eral quadri ne subfamilies including Catocalinae beingmore closely related to Arctiidae and Lymantriidae than totri ne noctuids (Weller et al 1994 Mitchell et al 1997Fang et al 2000 Mitchell et al 2000) The latter groupis considered as monophyleticClearly systematics of theNoctuidonematinaedo not match the tentative new classi- cation of the Noctuidae

This lack of congruence between parasite and hostsystematics leads us to consider that the relationshipsbetween Acugutturidae and their hosts have not beenguided by coevolutionary forces in the sense of Ehrlichand Raven (1964) but by horizontal transfers (Rogers ampMarti 1996) and sequential evolution (Jermy 1984 Fu-tuyma amp McCafferty 1990) The Acugutturidae and es-pecially the Noctuidonematinae have likely experiencedan adaptive radiation on higher Lepidoptera long aftertheir host diversi cation History of the Noctuidonemati-nae has been characterised by host transfer events ontofairly distantly related hosts followed by adaptation tonew hosts leading to speciation processes that now permitus to distinguish easily fairly species-speci c populationsor strains of Vampyronema or Noctuidonema In mothsthese host transfers may have occurred through contami-nation of nematode-free species by nematode eggs juve-niles or adults left on resting or feeding surfaces (leavesor owers) by individualsof infested species The successof such contamination may be linked to the availabilityand accessibility of host intersegmental membranes thinenough to be pierced by nematode stylets a factor that

may have physically and dramatically restricted the hostrange of all Acugutturidae

The present morphological analysis of acugutturid ne-matodes parasitising 13 species of Lepidoptera and onespecies of cockroach revealed a greater degree of diversitythan previously known in this group The genus Vampy-ronema grouping nematodes with simple spicules in-cludes two inferred new species in addition to V dibo-lia and V daptria already known from the LepidopteraThe genus Noctuidonema grouping nematodes of Lepi-doptera with complex spicules includes a new speciesand the nematodes found on Sphingidae probably belongto a new genus The relationshipbetween these nematodesand their hosts appears to be much more complex thanpreviously thought In addition to their bene ts as para-sites of pest Lepidoptera acugutturid nematodes are aninteresting model for parasite-host coevolutionarystudies

Acknowledgements

We thank David Hunt (CABI Bioscience UK) DanielLachaise (CNRS France) Alvin Simmons (USDA Char-leston SC USA) and S Patricia Stock (University ofCalifornia Davis CA USA) for reviewing earlier ver-sions of this paper We thank Jeanette Williams and JohnClayton for collecting cockroaches in St Lucia and mothsin Fiji respectively We thank the following individualsfor identifyingor con rming our identi cations of insectsDavid Nickle (Smithsonian Washington DC USA) forPeriplaneta americana Pierre Zagatti (INRA France)for Pyralidae and Hugo Kons (University of FloridaGainesville FL USA) for Orthodes crenulata

References

ADAMS BJ (1998) Species concepts and evolutionary para-digm in modern nematology Journal of Nematology 30 1-21

ANDERSON RV amp LAUMOND C (1990) NoctuidonemaguyanenseRemillet amp Silvain 1988 morphologie du spiculeredescription du macircle et diagnose amendeacutee du genre Noctui-donema Revue de Neacutematologie 13 433-436

ANDERSON RV amp LAUMOND C (1992) Noctuidonemadaptria n sp (Nematoda Aphelenchoididae)an ectoparasiteof the moth Lesmone porcia (Stoll) Journal of Nematology24 16-22

EHRLICH P amp RAVEN P (1964) Butter ies and plants astudy in evolution Evolution 18 586-608

FANG QQ MITCHELL A REGIER JC MITTER CFRIEDLANDER TP amp POOLE RW (2000) Phylogenetic

502 Nematology

Speciation in the Acugutturidae

utility of the nuclear gene dopa decarboxylase in noctuoidmoths (Insecta Lepidoptera Noctuoidea) Molecular Phylo-genetics and Evolution 15 473-486

FOOTIT RG amp SORENSEN JT (1992) Ordination methodstheir contrast to clustering and cladistic techniques InSorensen JT amp Footit RG (Eds) Ordination in thestudy of morphology evolution and systematics of insectsapplications and quantitative genetic rationals AmsterdamThe Netherlands Elsevier Science Publishers pp 1-10

FUTUYMA D amp MCCAFFERTY S (1990) Phylogeny and theevolution of host plant associations in the leaf beetle genusOphraella (Coleoptera Chrysomelidae) Evolution 44 1885-1913

HUNT D (1980) Acugutturus parasiticus n g n sp a re-markable ectoparasitic aphelenchoid nematode from Peri-planeta americana (L) with proposal of Acugutturinae nsubf Systematic Parasitology 1 167-170

HUNT D (1993) Aphelenchida Longidoridae and Trichodori-dae their systematics and bionomics WallingfordUK CABInternational 352 pp

JERMY T (1984) Evolution of insecthost plant relationshipsThe American Naturalist 124 609-630

KITCHING IJ amp RAWLINS JE (1998) The NoctuoideaIn Kristensen NP (Ed) Vol 1 Evolution systematics andbiogeography Part 35 Lepidoptera moths and butter iesHandbuch der Zoologie Berlin Germany Walter de Gruyterpp 355-401

KRISTENSEN NP amp SKALSKI AW (1998) Phylogeny andpaleontology In Kristensen NP (Ed) Vol 1 Evolutionsystematics and biogeography Part 35 Lepidoptera mothsand butteries Handbuch der Zoologie Berlin GermanyWalter de Gruyter pp 7-25

LEMAIRE C amp MINET J (1998) The Bombycoidea andtheir relatives In Kristensen NP (Ed) Vol 1 Evolutionsystematics and biogeography Part 35 Lepidoptera mothsand butteries Handbuch der Zoologie Berlin GermanyWalter de Gruyter pp 321-353

MARTI OG JR amp ROGERS CE (1995) Noctuidonemadibolia n sp (AphelenchidaAcugutturidae)an ectoparasiteof the moth Mocis latipes (Lepidoptera Noctuidae) Journalof Nematology 27 387-394

MARTI OG JR amp ROGERS CE (2000) Effect of Noctuido-nema guyanense (Nematoda Acugutturidae) infection on thelongevity of feral male Spodoptera frugiperda (LepidopteraNoctuidae) moths Journal of Entomological Science 35 259-266

MARTI OG ROGERS CE SILVAIN J-F amp SIMMONSAM (1990) Pathological effects of an ectoparasitic ne-matode Noctuidonema guyanense (Nematoda Aphelenchoi-didae) on adults of the fall armyworm (Lepidoptera Noc-tuidae) Annals of the Entomological Society of America 83956-960

MARTI OG JR ROGERS CE amp STYER EL (1995)Report of an intracellular bacterial symbiont in Noctuido-

nema guyanense an ectoparasitic nematode of Spodopterafrugiperda Journal of InvertebratePathology 66 94-96

MARTI OG JR LALANNE-CASSOU B SILVAIN J-FKERMARREC A amp SIMMONS AM (2000) Ectopara-sitic nematodes (AphelenchoidoideaAcugutturidae) of Lep-idoptera and Blattodea in Guadeloupe Nematology 2 669-684

MITCHELL A CHO S REGIER JC MITTER C POOLERW amp MATTHEWS M (1997) Phylogenetic utility ofelongation factor-1 in Noctuoidea (Insecta Lepidoptera) thelimits of synonymous substitution Molecular Biology andEvolution 14 381-390

MITCHELL A MITTER C amp REGIER JC (2000) Moretaxa or more characters revisited combining data fromnuclear protein-encoding genes for phylogenetic analyses ofNoctuoidea (Insecta Lepidoptera) Systematic Biology 49202-224

POOLE RW (1995)NoctuoideaNoctuidae (part) CucullinaeStiriinae Pasaphidinae (part) In The moths of Americanorth of Mexico includingGreenland WashingtonDC USAThe Wedge EntomologicalResearch Foundation Fascicle 261246 pp

REMILLET M amp SILVAIN J-F (1988) Noctuidonema guya-nense n g n sp (Nematoda Aphelenchoididae) ectopara-site de noctuelles du genre Spodoptera (Lepidoptera Noctui-dae) Revue de Neacutematologie 11 21-24

ROGERS CE amp MARTI OG JR (1992) Noctuidonemaguyanense (Nematoda Aphelenchoididae) Population pro- les on male and female fall armyworm moths Journal ofEntomological Science 27 354-360

ROGERS CE amp MARTI OG JR (1993) Infestation dynam-ics and distribution of Noctuidonema guyanense (NematodaAphelenchoididae) on adults of Spodoptera frugiperda andMocis latipes (Lepidoptera Noctuidae) Florida Entomolo-gist 76 326-333

ROGERS CE amp MARTI OG JR (1994) Population struc-ture and transfer success of Noctuidonema guyanense (Nema-toda Aphelenchoididae) on moths of Spodoptera frugiperda(Lepidoptera Noctuidae) Journal of the Entomological So-ciety of America 87 327-330

ROGERS CE amp MARTI OG JR (1996) Beet armyworm(Spodoptera exigua) as a host for the ectoparasitic nematodeNoctuidonema guyanense Journal of Agricultural Entomo-logy 13 81-86

ROGERS CE MARTI OG JR SIMMONS AM amp SIL-VAIN J-F (1990a) Host range of Noctuidonema guyanense(Nematoda Aphelenchoididae) an ectoparasite of moths inFrench Guiana Environmental Entomology 19 795-798

ROGERS CE SIMMONS AM amp MARTI OG JR (1990b)Parasitism of Lepidoptera adults by Noctuidonema guya-nense Remillet and Silvain (Nematoda Aphelenchoididae) insoutheastern United States Journal of Agricultural Entomo-logy 7 242-245

Vol 4(4) 2002 503

OG Marti Jr et al

ROGERS CE MARTI OG JR amp CLAYTON JA (1997)Report of the ectoparasitic nematode Noctuidonema guya-nense (Acugutturidae) infesting Lepidoptera in the Fiji Is-lands Nematologica 43 505-506

SANMARTIacuteN I amp MARTIacuteN-PIERA F (1999) A morpho-metric approach to the taxonomy of the genus Ceramida(Coleoptera Scarabaeoidea Melolonthidae) The CanadianEntomologist 131 573-592

SIMMONS AM amp ROGERS CE (1994) Effect of an ec-toparasitic nematode Noctuidonema guyanense on adultlongevity and egg fertility in Spodoptera frugiperda (Lepi-doptera Noctuidae) Biological Control 4 285-289

SIMMONS AM amp ROGERS CE (1996) Ectoparasiticacugutturid nematodes of adult Lepidoptera Journal of Ne-matology 28 1-7

WELLER SJ PASHLEY DP MARTIN JA amp CONSTA-BLE JL (1994) Phylogeny of noctuoidmoths and the utilityof combining independent nuclear and mitochondrial genesSystematic Biology 43 194-211

WHEELER QD (1999) Why the phylogenetic species con-cept mdash Elementary Journal of Nematology 31 134-141

WHEELER QD amp PLATNICK NI (2000) The phylogeneticspecies concept (sensu Wheeler and Platnick) In WheelerQD amp Meier R (Eds) Species concepts and phylogenetictheory a debate New York NY USA Columbia UniversityPress pp 55-69

W ILEY EO amp MAYDEN RL (2000) The evolutionaryspecies concept In Wheeler QD amp Meier R (Eds) Speciesconcepts and phylogenetic theory a debate New York NYUSA Columbia University Press pp 70-89

504 Nematology