Table Of ContentA PHYLOGENETIC ANALYSIS OFTHE ISOPODA WITH SOME CLASSIFICATORY
RECOMMENDATIONS
RICHARD C. BRUSCA AND GEORGE IXP. WILSON
Brusca, R.C. and Wilson, G.D.F. 199] 0901 Aphylogenetic analysisofthe Isopoda with
:
someelassificatorv recommendations, \femoirsoftheQueenslandMuseum 31. 143-204.
Brisbane. ISSN 0079-SN35.
The phylogenetic relationships of Ihe isopod crustacean suborders are assessed using
cladistic methodology. The monophyJy of ihe Flabellifera was tested by including all 15
component families separately in the analysts. Four other peracarid orders (Mysidacea,
Amphipoda. Mtctacea, and Tanaidarea) were used as multiple out-groups to root our
phylogeneticestimate:-- within the Isopoda. Abroad rangeofpossible characters foruse in
assessingisopodrelationships isdiscussedanda final data(character)matrixwasselected.
Thisdatamatrix,comprising29laxaanil92 Lharacters.wassubjectedlocomputer-assisled
analysis using lour different phylogenetic programs. HENNIG86. PAUP, PHYLIP, and
MacOade. Phylogenetic hypotheses from the literature(particular!)' Wagele, 1989a) ana
discussed and compared with ourown conclusions.
The following hypotheses are suggested by our analysis. The Isopoda constitutes a
monophyletic grojip. The Phreatoicidea is the earliest derived group of living isopods,
followed' b> an Asellota-Microcerberidea line, and nexl the Onls id§a Mbove the Onis-
cidca is a large clade of 'long-railed isopod taxa (Valvifera, Anihuridea. Flabellifera,
Epicarides, Gnalhiidea). The Microcerbtridea is the sister group of the Asellota, but
probablyshouldocm be included tnthe Asellota, The Onlscideaconstluji nophyletic
group, the monotypic laxon Calabo/oidea is either a primitive omscidean. or is a sister
groupnfthe Omscidea(Caliihozoa isnotan asellotan).Ourcladistic analysissuggeststhat
the primitive isopod body plan wasone in w,hieh well-developed lateral coxa] plateswere
lacking, the pleopods were, muliiarticulaie. ihe uropods arose on (he posterior margin of
thepleotelson, ihe telsonic region -.is not elongate,and the mandibularmolarprocesswas
a broad flat grinding structure. Extant taxa with this body plan (Phreatoiudca. Asellota,
Milrocerbcrideatoceui primarily in relictualhabitats. Oniscideaconform to thisbody plan
except in possessing lateral coxa! pknes.
The long-tailed isopod morphology (broad flat uropods, an elongate telsonic region, anil
well-developed lateral coxal plates) appears 10 be a derived coriijinon within the Isopoda.
Subordersand familieswith thisbody plan appeartobe mostspeciose, orto have had their
origin, in the Southern Hemisphere, The earidnid -like pleonal morphology of many
long-tailed isopods(Flabellifera. Onathiidea. Anihuridea) isthus secondarily derived and
genl [0 lb? condition SOetl 111 Ihfi myfiidaceanS and other true earidnid crustaee.-ins.
The broad,elongate tailfan ofihe long-(ailed isopod taxa isnotused foracaridoid-like tail
locomotor^ behaviour (e.g Eu i uidold escape reaction"), bul rather as a steering/stabil-
ising plane. The emergence ofthe long-tailed body plan seems to have coincided with a
shift in isopod habits from infaunal to more active swimming, epifaunal lifestyles.
Accompanying this transition was enlargement of the lateral coxal plates (perhaps to
increase hydrodynamic streamlining of the body) and a shift lo active carnivory and
predation, and eventually parasitism in several groups.
The Suborder Flabellifera (asitiscurrently recognised) isnota monophyletic(axon.Three
taxa usually ranked at the subordinal level (Anihuridea, Gnalhiidea and Epicaridea) have
their phylogeni within the lineage of families thai currently constitutes Ihe
' '
Flabellifera.The Protognalhiidae is nol closely related lo the Gnalhiidea. Prolognalhiidae
j lo Anuropidae and is par! \^(a clade culminatingin theparasitic
family Cymoihoidae. Wagele'S (1989a) recently proposed new classification of the
Isopoda, including his new suborders Sph&erOrnalitiea and Cyniothoida (sic), is not
Ited t>) 0U1 ph) logenetic malysis. Unambiguoussistergrouprelationshipscannot
be hypothesised for tl ill he cutreni data base. A new formal
classificationoftheorderIsopoda i aitbetterresolutionofthephytogenybasedupon
an expanded data set, tsopafa phylogetyj t tesslficeitoit. morphology, biogeography.
'
1*4 MEMOIRS UF He OUbBNSLAND MUSEUM
I
Menard C. firutca, San DiefB Statural History Miucum, PO flay i ?Wr San
California92!!2, VSA.;GeorgeD.F. Wihort,Australian Museum*6-4iCollegeSt.,Sydney
SouJh Sew South Wales2000,Avsuuttu; 20March, 100}
'Amtdsi thisprudent love 0/oh\curity\ (he (1905) considered a synonym of her 'FM
' ft
one feature ofmoral character which they lifera' (following San* to include the Acgidac,
possess in common is strong that Anthuridae, Cinolanidae, Corallanidae, Cyrnol-
allofthemmusthavesprang,fromacomtnott hoidac, Excorallanidac, Gnathiidae, Limnorii-
origin* dae, Serolidae. and Sphaeromidue). Mcnzies
The ReverendT.R.R.Stebbing(1893), Speak- (1962) considered »he Cirolanoidca to be a sub-
ingof isopods. tribe ofhis tribeFlabcllifera, synonymoustothe
Mosl 01 the isopod suborders were desci ibed Cymothoidcaofsome previous authors (includ-
anddelineated in Ihc early pari ofthe nineteenth ing the Anuropidae, Cirolanidae, Limnoriidac.
century, but tor the past 150 yearsclassification Sphaeromidae). Wagele (1989a) used Leach's
of these suborders and their families has been (1814) spelling of 'Cymoihoida', for his newly
unsettled, Until fairly recently many workers proposed suborder (for the Aegidae, Anuropi-
included the Tanaidacea within the Isopoda and dae, Bopyridae [=Ep»caridea] Cirolanidae,
includedcither(orboth)theGnathiidcaand An* Corallanidae, Cymothoidac, Gnat(hiidae, Phora-
thuridea within the Flabellifera (or Cy- topodidae, Protognathiidae,andTridentellidae).
mothoidea') (Bate and Westwood, 1863-68; In 1983 Van Lieshout erected a new mono-
Stcbbing, 1893; Sars, 1897, Richardson, I9Q5; typic suborder (Calabozoidca) for Culahozoa
SmithandWcldon. 1923; Hale, 1929; Nurstrasz petluciduyaground-waterisopod from Venezue-
and Schuurmans-Slekhov ii 1930; Menziex lan wells, and discussed its possible affinities to
1962; Naylor. I 9 >2). Hansen(1916) and Monod both the Oniscidca and the Asellota. Wagele
(1922)recognised (henecessityofseparatingthe (1989a) argued for placing the Calabozoidea
tanaidaceansfromthe isopods,andalsoremoved neat the Asellota, dc-puning these iwo suborders
the gnathiids and anthurideans from the Flabel- as sistergroups on his phylogenetic tree.
lifera Some authorities sought to establish a Recent summaries bv Bowman and Abele
fundamental split between the gnathiids and the (I9B2), Brusca and Iv'crson (1985), Schram
remaining Isopoda, Monod H922) called the (1986), and Brusca and Brusca (1990) took the
gnathiids Decernpedes (MO-footed'), and all conservative approach in recognizing 9 Subor-
other isopods the Quatuordecempedes (*M- ders (Table 1. Figs 1-3), maintaining separate
footed'). Following Latreillc (1804), Menkes subordinul status for ihe Mierocercridea, An-
(1962) used the name Tcuaccra for the non- thuridea, Gnathiidca, and Epiearidca.
gnaihiid isopods. Menzies(1962)cbose toretain Ar) examination ofpreviously published stud-
Iheanthurideanswiihin the Flabellilaa fcrui later USconcerning isopod phytogeny revealsa fairly
removed them (MeftZieS and Glynn, (968); broad range of ideas (Fig. 4 ), Beginning with
Karaman(IQ33)BtticdMicrocerbemswtihthe Hansen (1905), however, two taxa have domi-
Anthuridca, and many subsequent worker-- ac- nated the literature as contenders for the title of
cepted Ihis placement (Reman- and Sieving, 'most primitive living isopods*, ihe Flabellifera
1953; ChappuiN and Pefatoare, 1954; Lang, and the Asello'a. Schultz (1969. |979) deviated
|96QjSchuIte, J979;Kussakfo, 197%However, markedly from this pattern, and his phytogeny
Lang (1961) created a new suborder tor this depicted the Gnathiidca as the most primitive
genus,the Microccrberidea,and Wagele(19B2b, living isopod group. Schrum (1974) appears to
1983b)argued against any fetal unship bei'.v have been the only person to have previously
tfie micrncerberids and anthurideans, instead specificallyespoused the Phreatoicideatobethe
suggesting that the former were highly special- earliestderived isopod suborder.
ized asellolans Supporters ofthe 'Asellota-arc-primitive' hy-
The name 'Cirotanoidca' has been used in potheses have included Hansen (1925) Moond
f
different waysby different workers Richardson ( 1922), Btrsrem ( 1951), Zenkcvich and Biretein
RO l Examples of "short-i^iled' isopod suborders. A7 Phreatoicidea fMesatnpkisopm depressus. alter
Nicholls, 1943). B, Asellota {lantfopsli motiUrtyensis, after Men/ies. 1952). C. Microccrberidea (Micro-
cerherus .sp , after Argann. l$88) D( Calatwnidea (Culahozoa rirltunda, after Van 1 jeshOttt, 1983), E,
Oniscidca (Artnaddbdtumvulgare, afterSutton. 1972),
PHYLOGENETIC ANALYSIS OFTHE ISOPODA 145
^^~^1
146 MEMOIRS OFTHE QUEENSLAND MUSEUM
TABLE 1.Taxa analysed in the present study potheses have included Racovltza (1912),
Stromberg (1972), Kussakin (1973, 1979),
OUT-GROLPS
Bruce (1981), and Wagele (19S9a) Among the
OrderMYSIDACEA Flabellilera, the Cirolanidae (especially Bathy-
OrderMICTACEA nonius) is usually chosen as the model for the
OrderTANAIDACEA arehtypical ancestral isopod. Kussakin (1979)
OrderAMPH1PODA refined his earlier views to present a phylogeny
IN-GROL'PS in which a *cirolanid-likeancestor' (but thatwas
not vet a 'true' flabelliferan) gave rise to an
OrderISOPODA Anthuridca Microcerberidea line as the most
Suborder Phreataicidea primitive living isopod group, followed by the
SuborderAsellota Oniseidcaand Valvifcra,with the extant Flabel-
SuborderMicrocerberidea lifera, Phreatoicidea, and Asellota being the
SuborderOniscidea mosthighly derived taxa. Kussakin (1979)came
IntraordcrTylornorpha to this conclusion despite his contention that the
fntraorderLigiamorpha mostprimitivearrangementofpereopodalcoxae
SuborderCalabozoidca occurs in the Asellota,agroup inwhich henoted,
SuborderValvifcra 'the coxopodite still looks like a normal seg-
SuborderEpicaridca ment'. Within the flabelliferan line, Kussakin
SuborderGnathiidea hypothesized threelineages. One lineage lead to
Suborder Anthuridcj
predacious/parasitic lifestyles (Cirolanidae,
SuborderFlabellilera
Aegidae, Cymothoidae, and ultimately the Epi-
Family Aegidac
caridca); the other two lines were said to have
Family Anuropidac
henthic herbivores and detritivores,
Family Bathynaialiidae
such as the Serolidae and Sphaeromatidae. He
FamilyCjrolan
allied the Anuropidac with the Valvifcra and
Family Corallai
nlly Cvniothoidae Oniseidca, rather than with the Flabellifera.
Family fl i te Kussakin described (but did not depict on his
Family Limnoriidae phylogenctic tree) the Asellota arising from a
Family Lynseiidae hypothetical ancestral cirulanid stem group, via
Family Phoratopodidac trie Phreatoicidea. Bruce (1981) supported Kus-
Family Piakarthrhdae sakin s (1979) views, and further hypothesised
Family Protognathiidae the Phoratopodidae to be the sister group ofthe
Family Serolidae Valvifera. Nicholts (1943, 1944). Dahl (1954),
FamilySphaeromatidae andStromberg 1972)alsoarguedthatthePhrea-
(
FamilvTridentellidae toicideaoriginated from an ancient Flabelliferan
(1961), Belyaev (1966), and most recently s-rock close to the modern Cirolanidae.
Sehmalfuss (1989), Although Schmalfuss' tree Wagele (1981) claimed that "general agree-
he appe.v dogram, it appears in cmeesntrtalexiissotspoadmboondgyisshoappoedawnodrkeexrtesrntahlatftehaetuarne-s
be an intuitive tree based on ad hoc assumptions
ofancestry. It used4 specificsynapomorphies to livin: gceCritarionlainoidhaaeve(tbheoeunghsimpielrahraptos tlhaocskeinogf
define 8 isopod suborders- Schmalfuss did not
coxal plates)," but later stated that the Cirolani-
iibe Jiis method of tree construction, tree ouldnot possiblybeconsideredasprimitive
selection, character analysis, or character polar- isopods and (hat they were the probable sister
ity assessment; did not calculate tree lengths or group of the Anthuridea. Still later Wagele
homoplasy Sri id nol desenberhe charac- (1989a) claimed that the (hypothetical) ancestor
ters he utilised, and. rooted his tree b&SCi of the lsopoda was cirolanid-like, even though
ambiguous statements regarding ad h ins 'Henuigjan' phylog&netic ai con-
pofhetical morphotypes rather ihan on methods firmed that the Cirolanidae was a highly derived
group or nntological analysis It group (Fig. 4D).
should be noted that tor B taxa there exist Stromberg ( I 972) counted the number of hy-
b6G\032 possible tree topologies (Fclsenstcin, pothesised pleSiotrtorphic Features occurring in
1978) each ofthe isopod suborders,concludingon this
Supporters of 'Flabellifera-arc-primitivc' hy- basis that the Flabellifera (notably the Ciro
PHYL0GENET1C ANALYSIS OFTHE1SOPODA
147
werethemostprimitive livinggroupand the number of possible trees quickly becomes
j
sicm group from which all other isopod sub- astronomical An analysis of the 1U nominate
orders were derived. He presented an argument isopod suborders alone requires assessmeni
for close alliance between the Flabellifera. the 282 million possible irecs. 34.5 millionofw
Epicaridea, and the Gaathiidea. are bifurcating trees (Felsenstein, 1978). The
All of the above hypotheses, except Waj present study analyses 29 taxa, for which I]
(1989a), consisted of ad hoc I tstruction are 8.7 X 1 3Gpassible bifurcating trees. Hence,
.voluiionury narratives in the traditional, or to select a single shortest tree with the highest
orthodox, sense. Each was based on a small set degree of parsimony and the lowest level of
elected characters that held sway ovi: norm p I v '$ ychalling the data' is difficult
others Mo$t relied on a mix of both primitive if not impossible. Nevertheless, Wagele's
and derived features to infer relationships. None (1989a) analysis was a very impoUam step for-
was based on a large data set oi' empirically ward in isopod phyiogenetics. and was the I
evaluated characters, and nunc usee) any strict published Study at the subordmal level to use b
analytical methodology Must, if not all. relied relatively large data set and provide lists Df
upon the (stated or unstated) adhoc selection of general synapomorphies that define putative
an extant group of isopods to represent a primi- monophylelic lines. For these reasons, we com-
tive ancestral morpholype. From these apriori- pare ouranalysisclosely to thatofWagele in the
Selected hypothetical ancestors, evolutionary ilt&CU! il ii ,ction at the end of thispaper.
scenarios were inferred, and re con-
METHODS
structedbasedupon these scenarios. Becausethe
phylogcnctic scenarios cued above were not
derived from empirical an ' I the data, nor Out-Croups
utilizedanyrepeatablemethodology,itwouldbe The questionsofperacarid monophyly andthe
-.ir(anddifficult)to compare them dhrcci I phylogenetic sequence ofappearance ofthepeT-
the present study. It is interesting to note that, acarid Orders have long been favorite subjectof
tte the fact "that the Phreatoicidea have the debate among carcinojogists. Nearly every im-
;
oldest known fossil record (Pcnnsylvanian; aginable topology ofphylogenetic relationships
Scbram, L$7Q, 1974), none of the above pro- among the In 19SI pcracaridahasbeenproposed
posalshypothesised thisgroup(or a phrcaioicid- another.There is no needtoreview
:
like morphology) to represent the ancestral Dlis debate here (DahL 1977, Walling, i
isopod type. 3; Schram, 1981, 1986; Dahl and ffes
The only previousattempt to undertake a phy- 1982; Hessler, 1983; BftIS 34), Howe
uetic analysisofthe Isopotia basedonalarge most published ideas over the years have sug-
data set and a specific methodology ister group of the Isopo
Wagele's (1989a) recent study (rig" 4D). either the Amphipoda or the Tanaidacea. The
WageIc proposed a sweeping reorganisation of recently described Mietacea may also bi I
isnpodclassification. Some ofthe manvchanges related to the isopods (Schram. 19S6). Because
he proposed included the complete elimination ofthisuncertainty,we usefourout-groupsinour
ofthe SuborderFlabellifera,and the rcduco analysis: Mysidacea. Amphipoda.Mietacea,and
family status of the suborders Gnathiidea and Tanaidacea.The increasedaccuracyofchara
F.picaridea (reducing the families of the latter to ssment and tree resolution that can
subfamiliesandeliminatingthe name Epicaridea be achieved by use o( the multiple nut-group
altogether). However, even though V method has been explained by Maddison et ai.
study wasbasedona largerset icters than (1984) 2nd o e basic premise being that
anyprevious analysts, itwas still based on an ud cladograms shouldbe globally parsimonious.
/toehypothetical ancestral morphotype, the phy-
logenetic tree was computed by hand, arid In-Gri
-Ktempt was made to achieve either global or Our in-group includes all 10 nominate isopod
in-group parsimony or utilise any strict criteria suborders(Table I), plusthe 15 nominate flabel-
oftree construction or tree selection. Wagele's liferan families. The relationshipsofthefact
classification schemewasnot strictlycladistic in included within the Flabellifera have been con-
that itdid not recognise the sistergroup trove i i id ii has been frequently suggested
cladogram. ii- : Flabellifera is a non-monophyletic
In data sets with more than a fie the a. Kussakin (1979), Bruce (1981), and
MEMOIRS OFTHE QUEENSLAND MUSEUM
148
^
FIG. 2. Examples of various 'long-tailed' isopod suborders. A, Epicaridea (Argeia pugettensis). B-C,
Gnathiidea(B,Gnathiatridensfemale;C, Gnathiatridensmale).D,Valvifera,Idoteidae(Idoteametallica).
E,Valvifera,Arcturidae(Jdarcturushedgpethi). F, Anthuridea,Anthuridae (Haliophasmageminatamale).
G, Anthuridea, Paranthuridae (Paranthura elegans).
PHYLOULShHC ANALYSISOFTHE fSOPODA
Wagele(1989a)depictedthisgroupparaphylcti- study we argue that protognathiids share no
iiy on their trees oi the tsopouai. W&geU uniquesynapomorphieswithgnutttiids,although
(1989a) recommended a reorganisation of llie somesuperficial similaritiesarepresent.Wagclc
Isopodathatwouldeliminatethree currently rec- (1983b, 1989a) has argued that the Mien
i
ognized suborders, the Flabellifera, Epicaridca, beridea are membersoftheasellotc supcrfamily
and Gnathiidea. Although Wagclc's tree and Aselloidca- Although the microeerherids
I
classificationarcnotcorroboratedbythe present several features typically viewed as asellotan
study, the Flabellifera as it is currently recog- to-articulate antennularpeduncle; pleonites3-5
nized is almost certainly not a monophyletic fused with the plcotclson; females lacking first
taxon.W^gcle reorganized the above suborders pair of pieopods; male second pleopod with en-
into two new groups, which he called the Cy- dopod transformed into a complex gonopod)>
molhoida (sic) and the Sphaeromatoidea, sub- lhey lack other features generally also regarded
suming the Gnathiidea, Epicaridea, and sevrtal as definilivt* synapomorphies of the Asellota
flahclliferan families into the former. (Note that (e.g. antenna! peduncle with a scale; female
Wagele's CymothuKJa is noi fbe equivalent of pleopod 2 iiniianious; exopods of male second
Cymothoidea ofRichardson. 1905. and others). pieopods modified to work with the elongate
In the. present simlv, we. lest the monophvlyOf geniculate endopods in sperm transfer; and,
the Flabelliferaby includingal! oi itscomponent possibly, the unique asellotan spcrmathecal
families in the analysiswith the other suborders duct). Forthesereasonswe treatthe.Asellotaand
ofthe Isopoda.Werecognizethefollowingnom- Microcerbcridca as separate groups (OTLTs) in
inate families of Flabellifera: Acgidae Dana. ouranalysis
1853; Anuropidae Stcbbing, 1893; Bathy-
nataliidae Kensley, 97S; Cirolanidiie Dana Data Sources
1
1853;CoTallaiudae Hansen, 1890; Cvmothoidac Specimens were examined for all taxa treated
Leach, 1818: Kcuphyliidae Bruce, 1980; Lim- except Protognathiidae. Material was examined
noriidae White 1850; Lynseiidae PoOTt, 19S7; iot loan from 8 variety of Institutions, and during
Phoralopodidae Hale, 1925; Plakarthriidac Ri- visits to ihc U.S. National Museum of Natural
chardson, 1904; Protognathiidae Wage e and History. Smithsonian institution (USNM). Los
I
Brandt. 19SS; Serolidae Dana, IKS V Sphaem Angeles County Museum of Natural History
maodae Burmeister, 1834; and, Tridcntellidae (LACM). Zoologisch Museum, Amsterdam
Bruce, 1984. (ZMA). Australian Museum. Sydney (AM),
The two infraordcrs of Oniscidea Latreille, Queensland Museum, Brisbane (QM). Victoria
1803 (Tylomorpha Vandel. 1943 and I.ii-iamor- Museum, Melbourne (VM), San Diego Natural
pba Vandel, 1943; see Holdich el ak< LOT4) are History Museum (SDNHM). and Scripps Insti-
also analysed separately because opinion bas tution of Oceanography (SIO). In addition to
been divided on whether or not the Tylidae arc examiningspecimens,theoriginalliteraturewas
trueouiscideans(Kussakin, 1979; HoldicheM/„ extensively perused.
!984;Wagele, 1989a; Schmalfuss, [3S9),
Three taxa that are included in our analysis ScoringorCharacters
require brief comment. The Calabofcoidea is a One of the advantages oi the available com-
monoiypic groun4»we|ei (freshwater) taxon so puter-assisted numerical techniques(seebelow)
far known only from Venezuela. In her original is that they treat each character independently.
description Van Lieshout (1983) suggested Thus, if the state of a particular character is
possibleaffinitiesofCalabozoatoboththe Asel- unknown, inapplicable, orwe have simply been
lotaandtheOniscidea. Wehaveexaminedsped* unable to resolve it to our satisfaction, wc have
mens of Calabozoa and found Van I.ieshouis scored it as 'missing data' (indicated by a kV in
illustrations and description misleading; newil- the data matrix). In preliminary analyses, char-
lustrations of the male pieopods I and 2 are acters forwhich noclearpolarity couldbe estab-
provided mi Fig. 10. Calabozoa appears to lished were nol coded in any primitive-derived
possess no asellotan synapomorphies. WMgele sequence,butwerelefttochange inanydirection
and Brandt (1988) created the Protognathiidae such that simple parsimony (fewest changes)
based upon rbe.ir examination of a single, ap- was the arbiter These unpolunsed (nonadditivc
parently manca-stagc. individual Wagele or unordered) characters are indicated in the
(1989a) concluded that this new family was the :-h;irac-ter discussions below. These anal
sister group of the Gnathiidea. In the present proved useful inassessingcharacterhomonlasy.
150 MEMOIRS OFTHE QUEENSLAND MUSEUM
PHYLOGENETIC ANALYSIS OFTHE ISOPODA
151
Kussakln (1973)
FIG.4.Someevolutionarytreesfrompreviousstudies,byKussakin(1979),Bruce(1981),Schmalfuss(1989),
and Wagele (1989a).
For the final analyses, however, we decided to phic for that character, but is scored plesiomor-
analyse the data with all characters left un- phic. Initially polarized characters were scored
ordered (nonadditive). as indicated in theorderingofthecharacterstate
Ifa characterstatejudged tobe plesiomorphic numbers: = plesiomorphic, 1 = apomorphic, 2
ispresentforonly somemembersofthetaxon in = more apomorphicthan 1, etc. Homology deci-
question, e.g. 'accessoryflagellum on antennule sionswere madeonthebasisofontogeneticdata
in most gammaridean amphipods', it is scored and comparative morphology (positional data
present in the data matrix for the entire taxon and anatomical similarity).
unlessotherwisestated,i.e.thederivedcondition
is presumed to define a subset within the taxon. PhylogeneticAnalysis
Conversely, ofcourse, ifan apomorphic state is The character state data were analysed with
present in only some members of the taxon in four numerical cladistic analysis packages:
question, the entire taxon is not scored apomor- HENNIG86 (version 1.5), PHYLIP (version
FIG. 3. Examples of various isopod families and genera of the suborder Flabellifera. A, Cirolanidae
{Metacirolanajoanneae, SDNHM). B, Tridentellidae (Tridentellaglutacantha, from Delaney and Brusca,
1985). C, Aegidae(Aegaplebeia, from Brusca, 1983). D, Cymothoidae (Ceratothoagilberti, from Brusca,
SDNHM
1981). E, Limnoriidae (Limnoria quadripunctaia). F, Serolidae (Serolis carinata, A.0114). G,
Anuropidae (Anuropus bathypelagicus). H, Sphaeromatidae (Gnorimosphaeroma insulare). I, Sphaero-
matidae(Exosphaeromaamplicauda).J,Sphaeromatidae(Bathycopeadaltonae). K,Sphaeromatidae(Par-
aleptosphaeromaglynni).
|$2 tOFRS nr THE QUEENSLAND MUSEUM
3,2), PAUP (version 3.0.. afVd MacClade (ver- suborder or family (Appendix 111), However.
sion 2.1). HENNIG86 is advantageous because because we were concerned in this study with
of Us speed, successive weighting algorithm. identifying sister group relationships within the
ability to depict polytornous tree brunches, and Isopoda.we did notmakean effort to identifyall
ability to store many equal-length trees in Of the unique synapomorphies thai define only
mcmoTy. The successive weighting program individual taxa (suborders or families). Some
(Farris, 1969 1989) is useful in reducing the characters that proved to define only terminal
r
impact of homoplasous characters on tree to- taxainour final trees were early-onsuspectedto
pology. Despite Plamick's(1989) rccommenda- be useful in distinguishing larger sister groups.
of HENNIG86 as the program of choice. Thesemay beviewedas 'uninformative*charac-
PAUP. MacClade. and the PHYLIP program ters inthefinaltreesbysomeworkers.However,
package remain useful tot comparative and ana- were imporinm incomparativeanalysesand
lytical purposes (Sanderson, 1990). PAUP is by tree testing, and as additional taxa and data arc
far the most uscr-fricndly, is useful to check described some of these characters may no
differentcharacter optimisations (a feature cur- longer remain unique to a single terminal tavon
iL-nily absent from HENNIG86) on the final For these reasons, we felt it was important to
trees,andtoobtain detailedcomputationsofC.L leave them in the data matrix, thus allowing
(consistency index), character changes, and others to use our data set as a starting point for
OTU apomorphy lists. The program MacClade further tree testing The data set is available on
3.0 was used (on a Macintosh Computer) to diskette on rcqc<
branchswap on the final set of trees, in ortfe
..are changes in tree length, homoplasy DISCUSSION OFCHARACTERS
levels, and character placement on selected al-
ternative trees, including those of Schmalfuss kedEybs
(1989), Wagele (1989a), and others. MacClade Mysidaceans and mictaceans have compound
and PAUP arc extremely useful in their user- eyes set on short, mo. estalks (although
friendly ability to generate graphic repre- eyegtalKs are &bSeftt in the mictacear Hirsutia),
sentationsofcharacter traces on trees, although Inamphipods,a *rudimentaryeyestalk* hasbeen
MacClade is seriously hindered by its inability reported from ingolfieilids. Dahl (1977) and
to depictmuJtifurcations. Lowry and Poore (1989) have argued that this
Theprincipal statistics used in tree evaluation small process in ingolfieilids is no: a true cye-
were overall tree length (step length) and con- stalk, but rather is a cuticular process or scale.
sistency index (C.L). Consistency and retention Lowry and Poorc's argument hinged on the ob-
indices for each individual character were also servation that unequivocal eye stalks in other
computedand used to evaluate theiroverall ho- peracarids have 'an attitude and position vcrv
moplasy levels. different1 than seen in he mgulru'llids Oahl %
Carpenter(1988) recently arguedthat consen- argument was based onithe absence of 'dioptric
sus treesshould not be used to construct clado- and nervouselements' tn this structure. The first
grarns However, we agree with Anderberg and argument is not particularly strong because the
Tehlcr(1990)thaistrici consensustrees ~rc both position and altitudeofperacarideye stalks vary
useful and informative because they reduce the greatly. A positional change in the ingolfieilids
conclusions toonly thosecomponentswhichall could have been caused by a lateral rotation of
equaJ-Iength shortest Irees have in common. In the entire cyc-antcnnular-antcnnal complex.
face, they are probably a necessity when high DahTsargument isstronger,although itrelicson
levels of homoplasy invest a data set Even if reductions rather than homologies. Among tan-
successive weighting (i.e rhe successive ap- aidaccans, articulated eye-lobes occur in some
proximations character weighting method of Apseudomorpha and Tanaidomorpha, including
Farris, 1969) is used, multiple equally parsi- those with eyes in a variety ofpositions ranging
moniuustreesmayderive frorr u high in from that seen in the Mictaceatothatseen in the
homoplasy Thus, we believe thai when numer- ingolfieilids. In amphipodsand isopodsthce
ous equally parsimonious trees exist, a strict arc entirely sessile, although they may be ele-
consensustreeshouldbe presented. vated on lobes of varying sizes in some species
In order to distinguish between some closely of Phreatoicidea, Gnathiidea, Valvifcra, and
related taxa. we included some characters that AseIlota Al the level of the Pcracarida most
are currently known io ^c unique to a given workers might regard motile stalkedeye*as |he
