On the validity of Heliconius tristero Brower and Heliconius melpomene mocoa Brower, with notes on species concepts in Heliconius kluk (Lepidoptera: Nymphalidae)

10 Pages·2000·5.3 MB·

by  A V Z Brower

PROC. ENTOMOL. SOC. WASH.
102(3), 2000, pp. 678-687
ON THE VALIDITY OF HELICONIUS TRISTERO BROWER AND
HELICONIUS MELPOMENE MOCOA BROWER, WITH NOTES ON SPECIES
CONCEPTS IN HELICONIUS KLUK (LEPIDOPTERA: NYMPHALIDAE)
Andrew V. Z. Brower
Department of Entomology, Oregon State University, Corvallis, OR 97331-2907,
U.S.A. (e-mail: [email protected])
—
Abstract. Lamas' (1998) criticisms of Brower's (1996a) taxonomy are shown to be
based on misinterpretations of the evidence and of the rules of nomenclature. The names
H. amaryllis amaryllis bellula Stichel and H. melpomene bellula Turner are unavailable.
f.
The name H. bellula Brower is available but invalid. The name H. melpomene mocoa
Brower is the valid name for the subspecies of H. melpomene from the Putumayo region
of southeastern Colombia. The Genotypic Cluster Species Concept is contrasted unfavor-
ably to the Phylogenetic Species Concept with respect to the aims of systematics in
general, and the resolution ofrelationships among geographically differentiated//^//cowm^
taxa in particular.
Key Words: Heliconius melpomene bellula, Heliconius tristero, nomenclature, circum-
scription, subspecies
Recently, Lamas (1998) published a crit- robust evidence stemming from three dif-
ical discussion of two new names applied ferent sources (mitochondrial DNA se-
to Heliconius butterflies from the upper Rio quences, genitalic morphology and wing
Putumayo basin in southeastern Colombia patterns), and that at least one of them re-
(Brower 1996a). The taxa in question are quired a new name is not at issue. Lamas'
Miillerian mimics: one {H. tristero) was de- criticisms of my paper focussed on the fol-
scribed as a species in the H. cydno Dou- lowing problems: (1) the availability of H.
bleday clade, while the other was named as bellula Stichel (1923); (2) the supposed
a subspecies, H. melpomene mocoa. Most synonymy ofH. melpomene mocoa Brower
members of the cydno clade exhibit blue- with bellula auctt.; (3) my interpretation of
and-white or blue-and-yellow wing patterns the International Code of Zoological No-
and mimic species in the sara-sapho group menclature (1985) with regard to hybrids;
(cf. Brown 1981, Brower 1994, and Brower and (4) the logic of my species concept.
and Egan 1997 for reviews ofphylogenetic First, I will address these problems in nar-
hypotheses among Heliconius species), so row relation to Lamas' critique ofmy paper
it was a surprise to discover a new species (1996a). My rebuttal of Lamas' criticisms
mimicking sympatric races of H. melpo- is followed by a briefdiscussion ofthe spe-
mene (L.) and H. erato (L.). cies problem in Heliconius. An English
That these formerly conflated taxa rep- translation ofthe relevant sectionofLamas'
resent two distinct entities is supported by paper is included as Appendix 1.
1
VOLUME 102, NUMBER 3 679
Reply to Lamas' Critique let (1993) and Brower (1996b) used the
name, but none ofthem provided a descrip-
Is Heliconius bellula Stichel (1923) an
available name? tion of the taxon, nor did any of them cite
Stichel's description (which would not per
As Lamas (1998) pointed out, Stichel se change its availability, in any case: Art.
(1923) applied the name bellula to a form lldii). All of those uses of bellula employ
&
of H. amaryllis amaryllis C. R. Felder it as a nomen nudum and none of them sat-
{H. amaryllis is now considered tobe a sub- isfies the criteria of availability (Art. 13a).
species of H. melpomene; cf. Ackery and Holzinger and Holzinger (1994) provided
Smiles, 1976). Lamas argued that Stichel's a description and illustration of H. melpo-
employment ofbellula was infrasubspecific mene bellula and cited Stichel (1923), but
and therefore formally unavailable. How- their book is not consistently binominal (it
ever, Neustetter (1929) cited Stichel's name uncritically employs numerous infrasubspe-
as a trinominal form of H. amaryllis, thus cific names), and its contents are thus un-
implying that the name was available available (Art.lie). Ironically, it appears
(ICZN Article 16). The subsequent employ- that my 1996 paper is the first to provide a
ment of the name for a geographical race description for H. bellula that would render
by more recent authors (e.g.. Turner 1971, it available under all the criteria of Art. 1
Brown 1979, Sheppard et al. 1985, Mallet (via illustrations, description of differenti-
1993, Brower 1996b) further suggests that ating features, identification of a holotype,
the name has been treated as available (Art. and citation of Stichel 1923). But because
45gii), and it was under that premise that I I circumscribed the concept narrowly, to re-
dealt with it in my paper (Brower 1996a)'. fer to a hybrid form typified by Stichel's
However, given that Stichel's holotype is original type specimen, the name is invalid.
not a representative ofthe melpomene clade Even so, my definition prevents subsequent
(see Brower 1996a and below). Lamas' usage of bellula for any other Heliconius
opinion that Stichel's name was quadrinom- taxon (Art. 23h; see hybrid section, below).
inal provides a convenient excuse to sink it Thus, the correct authorship of Heliconius
and avoid the complications described bellula is Brower (1996a), but the name
above. If its original designation is deemed does not refer to any valid taxon, regardless
infrasubspecific, a name remains unavail- of its repeated misapplication in recent
able until a description is provided (Art. works (including my own).
10c). Lamas claimed that Turner (1971)
employed the name bellula to refer to a Subjective synonymy ofbellula and
subspecific entity, thereby becoming its au- mocoa?
thor (Arts. 23j and 50c). However, Turner's Lamas (1998) argued that H. melpomene
tentative use of the name in a figure legend mocoa Brower is a junior subjective syno-
(followed by a question mark) was not ac- nym of H. melpomene bellula auctt. Since
companied by a description. Likewise, bellula is not a valid name, this is a moot
Brown (1979), Sheppard et al. (1985), Mal- point, but it raises another important issue:
Lamas' claim shows that has been deceived
' I examined and discussed Stichel's holotypes of by Miillerian mimicry! It is the holotype of
bellula, permira and degener. Lamas suggested that I bellula and the holotype of H. tristero that
was unawareoffouradditionalformsdescribedby Sti- are closely-related, while the holotype ofH.
chel from the Mocoa region. Rather, these were delib- melpomene mocoa represents a different
erately omitted from discussion, because it was clear clade. Had Lamas examined the relevant
from the original descriptions that they represent var- characters, he would have seen that bellula
ious hybrid forms with recombinant wing patternsthat
are even less similarto the modem ""bellula" concept and mocoa are not the same under any rea-
than the two other forms I discussed. sonable circumscription H. bellula (what-
680 PROCEEDINGS OF THE ENTOMOLOGICAL SOCIETY OF WASHINGTON
ever its status) is not an example of the H. heurippa and H. tristero, both of which I
melpomene species group; instead, it is explicitly claimed were species in that pa-
close to H. thstero in the H. cydno species per. Under such circumstances ICZN article
group. To include bellula and mocoa in the 23h specifically applies-^.
same species, one would have to view H. Whether or not the bellula holotype ac-
cydno Doubleday as a subjective junior tually represents a hybrid or not (and which
synonym ofH. melpomene (L.)! I described parental species might have spawned it) is
this problem explicitly in Brower (1996a), difficult to determine in retrospect and
and it was precisely the desire to avoid such without extensive data from experimental
confusion that prompted me to coin new genetic crosses. That additional specimens
names for both the cydno relative (//. ths- displaying the same pattern have not been
tero) and the melpomene relative (//. mel- collected suggests that the holotype exhibits
pomene mocoa) in the first place. a rare recombinant wing pattern (typical of
Even if bellula and mocoa were closely the phenotypic diversity found in otherHel-
related, it is clear from Stichel's (1923) de- iconius hybrid zones). By contrast, there are
scriptions (Appendix 2) that he felt that bel- multiple specimens from multiple localities
lula was different enough from the local that appear very similar to the holotype of
"nominate form" to warrant a separate H. melpomene mocoa, and illustrations of
name. Stichel believed (in error) that his "//. melpomene bellula'' in recent works
three "typical" specimens were H. amaryl- (e.g.. Turner 1971, Brown 1979, Holzinger
lis amaryllis, today recognized as a distinct and Holzinger 1994) all lack the yellow
H. melpomene subspecies with an allopatric spots that are present in the bellula holo-
distribution in the Huallaga valley of Peru type. My interpretation that the specimen
(Sheppard et al., 1985). Although ideas does not represent a "pure" geographical
about what entities within Heliconius de- race is complementary to Stichel's (1923)
serve names have changed as knowledge of opinion that the bellula holotype is transi-
phenotypic variation and geographical dis- tional, as shown by his description of it as
tribution has grown, efforts to redefine an distinct and separate from his series of
old name of dubious availability in specific specimens of the local "nominate" form.
contradiction to its original author's intent
seem procrustean. A more reasonable view Logical Consistency
imsentheatfrthoem "Mnoocmoianaatnedfeonrvmi"ronosf Hw.asmealnpoon-- choIsematodeovesrelvoeroaklinshtiastecmoennctlsustihoantthLaatmmays
taxonomic argumentation (Brower 1996a)
ymous until I (Brower 1996a) named it. "no tiene sustento logico al ser refutable."
Hybrids and the ICZN First, I did not argue that the current clas-
Lamas (1998) also criticizedaserroneous sification of Heliconius is not illogical. In-
my invocation of Articles lb and 23h to deed, I stated that I chose the ranks I em-
ployed to preserve nomenclatorial stability.
curtail the current usage of bellula because
its holotype is a hybrid (Brower 1996a). He
pointed out that because the definition of -Lamas objected my use of the term "forbid"
"hybrid" in the ICZN glossary states that t(otrtahneslIatCeZdNasiTi"lperoonhinbaem"esinaphipslipeadpetor)inwtietrshperceiffeircehnyc-e
offspring of crosses between conspecific brids. Article 23h says, "A species-group name estab-
subspecies are not hybrids, bellula cannot lishedforananimal laterfoundtobeahybrid . . . must
be considered a hybrid taxon. Re-exami- not be used as a valid name for either of the parental
nation of Brower (1996a) shows Lamas' species, even ifit has priority over all other available
point to be technically incorrect: I hypoth- lneaamveesitfotrotthheem,rebaudterittmoacyonetnetmeprlaintteowhhoemtohneyrm"ym.u"stI
esized that the bellula holotype specimen not be used" and "forbids the use of" mean the same
was the result of hybridization between H. thing, or not.
VOLUME NUMBER
102, 3 681
even though I knew that the species delim- named his new Heliconius subspecies (de-
itations among geographically polymorphic scribed in the same paper) not "//. timareta
Heliconius butterflies were effectively ar- timoratus, " but "//. heurippa timoratus''
bitrary. The final points of my paper were (if timareta and heurippa are conspecific,
that, then the former is either a junior synonym
"The mitochondrial DNA data suggest or a subspecies of the latter). Under the in-
terpretation Lamas claimed that he prefers,
that the degrees ofrelationship among the
timoratus would seem to be an infrasubspe-
H. melpomene races and among the H.
cydno races are similar, and that diver- cific form of the subspecies H. heurippa ti-
maretal The complexity of sorting out
gence times within each group are also
names, ranks and relationships among geo-
approximately the same, implying that
graphic races in Heliconius obviously pre-
the taxonomic rank of each group should
also be the same. The entire species-level sents a challenge that neither Lamas nor I
have yet resolved in a fully satisfactory
classification of the genus will probably
manner.
require revision as additional data be-
come available." Synonymies
Ignoring these qualifying remarks, Lamas Heliconius melpomene mocoa Brower,
offered (without providing new data or ex- 1996a.
amining the characters discussed in Brower Heliconius melpomene beHula auctt. (Turn-
[1996a]) a hypothesis of circumscription er 1971, Brown 1979, Sheppard et al.
that he considered "more valid" than my 1985, Mallet 1993, Brower 1994, Holzin-
view: that H. timareta Hewitson, H. heiirip- ger & Holzinger 1994) [misidentifica-
pa Hewitson andH. tristero are conspecific. tions; unavailable name].
He based this notion on these entities' geo- Heliconius melpomene mocoa Brower
graphical proximity on the northeastern 1996a: 328. Holotype: Colombia, Dpto.
slope of the Andes, evidence of hybridiza- Putumayo, 1-3 km N. Mocoa on rd. to
tion among them (which he denied on the Pitalito, 25 March 1992 leg. AVZ Brow-
following page), and their hypothesized er. Deposited in Cornell University Insect
close relationship (based on studies of Collection (examined).
mtDNA: Brower 1996a, 1996b). However,
both the mtDNA studies and data from a Heliconius heurippa x Heliconiustristero^
more comprehensive recent analysis (Brow- natural hybrid
er and Egan 1997) imply that H. timareta, Heliconius amaryllis amaryllis f. bellula
H. heurippa and H. tristero are no more Stichel, 1923: 262. Original type speci-
closely related to one another than any of men: Colombia, Rio PutumaWyo, Rio Gua-
them is to H. cydno or H. pachiniis Salvin. yuyaco, 7 July 1921 leg. Hopp (see
Therefore, if one wanted to lump diagnos- Brower 1996a for transcript of original
ably different taxa based on the symple- labels).
siomorphy of potential interbreeding, the Heliconius amaryllis f. bellula Stichel;
logically consistent choice would be to col- Neustetter 1929: 58.
lapse all of these taxa under the oldest spe- Heliconius amaryllis f. bellula Stichel;
cies-group name (//. cydno). I said as much Brower 1996a [identified as hybrid H.
in 1996a. heurippa X H. tristero].
If Lamas (1998) were as concerned with
Notes on Species Concepts in Heliconius
the consistent assignment of taxa to the ap-
propriate rank as his criticism of Brower I suspect that the root of Lamas' criti-
(1996a) implies, then according to his cisms ofmy concepts oftristero and mocoa
"more valid" hypothesis, he should have lies in his concept of species in Heliconius,
682 PROCEEDINGS OF THE ENTOMOLOGICAL SOCIETY OF WASHINGTON
which differs from mine. The species prob- species; otherwise, they are collapsed into
lem in Heliconius has been controversial a single species. That the cladistic approach
for many years (e.g., Eltringham 1916 vs. may yield finer resolution of the hierarchi-
Kaye 1916); indeed, it is the very complex- cal pattern of diversity than alternative
ity of geographical diversification within methods is considered by many to be an
the genus that has led it to become a model asset. In short, the guidelines of the PSC
system for the study ofthe evolutionary ge- are clear: it is Mallet's understanding of
netics of mimicry (Sheppard 1960, Emsley them that seems cloudy.
1964, Turner 1971, Brown et al. 1974, Mal- Mallet's (1995) alternative, the GCSC,
let 1993). Many of the Heliconius papers views species as "identifiable genotypic
from the post-typological period applied the clusters" recognizable by "a deficit of in-
biological species concept (BSC; Mayr termediates" at single and multiple loci,
1940)), which unites allopatric taxa based and views speciation as "the production of
on their potential to interbreed. In recent divergent populations that can coexist in
years, the BSC has been criticized on the- sympatry." There are anumberofproblems
oretical and practical grounds, and numer- with this approach. First, there are no ex-
ous alternative species definitions have plicit criteria for identification of genotypic
been proposed that offer more operational clusters. That Mallet employed "deficit"
criteria for species delimitation (Eldredge instead of "absence" implies that he be-
and Cracraft 1980, Mishler and Donoghue lieves that the clusters need not be fixed for
1982, de Queiroz and Donoghue 1988, Nix- alternate alleles, but merely differ in allele
on and Wheeler 1990, Baum and Shaw frequency by some "significant" amount.
1995, Mallet 1995). Lamas' critique im- Further ambiguities include how many
plied that he favors the "genotypic cluster" markers should be sampled (Mallet [1996]
species concept (GCSC; Mallet 1995), suggested multiple loci), and what should
while I (Brower 1996a) applied the phylo- be done if one locus suggests continuity,
genetic species concept (PSC) as elaborated while another suggests distinction. The
by Nixon and Wheeler (1990). across-locus averages that Mallet (1995)
employed in the hybrid indices he presented
Evidence and Criteria ofSpecific would disguise heterogeneity among loci. A
Distinctness more sophisticated and logically consistent,
Mallet (1995) argued that the PSC fails but equally labor-intensive approach to
to provide "clear guidelines" for dividing multilocus species discrimination was pre-
species because it recognizes groups based sented by Doyle (1995). In practice, how-
on apomorphy: "With detailed morphology ever, methods like these that rely on ex-
and modern molecular techniques," he as- haustive characterization of gene pools are
serted, "one can find apomophies for al- rarely employed in the study of biological
most every individual," which he suggested diversity, because they entail intensive,
would result in rampant splitting and pro- quantitative sampling that is not feasible in
liferation of species. This simplistic cari- most circumstances.
cature ofthe cladistic method misrepresents Another diffi—culty with the GCSC is non-
the procedure of species delimitation, dimensionality. Mallet's concept is only
which has been explored in depth by Davis useful for contemporaneous taxa in sym-
and Nixon (1992), Doyle (1995) and Brow- patry or parapatry. To cover everything
er (1999). Cladists identify populations of else. Mallet (1995) suggested that "closely
organisms that they hypothesize to be dis- related allopatric forms should mostly be
tinct, and seek discrete differences between considered conspecific." But how are we to
them. Ifthey discover such differences, cla- determine that allopatric forms are "closely
dists consider the populations to be separate related?" If allopatric populations are
VOLUME NUMBER
102, 3 683
"identifiable," then they logically satisfy they are perhaps the most thoroughly-di-
Mallet's GCSC criteria and are distinct spe- agnosed Heliconius taxa that have been
cies by definition; if they are not identifi- published, the differentiating characters
able, then why would we hypothesize that having been drawn from the results of cla-
they were different in the first place? When distic analyses ofmtDNA, and corroborated
the hypothetical taxa under investigation with diagnostic characters from external
occur in allopatry, researchers are forced to and internal morphology (Brower 1996a)
turn to the empirical comparison offeatures and data from a nuclear gene (Brower and
of organisms to draw inferences about their Egan 1997).
taxonomic relationships. The PSC recogniz-
es separate species only when fixed differ- Recognition and Circumscription of
ences are discovered between hypothesized Subspecies
groups, which will always result in the rec- Although the genus Heliconius contains
ognition of a minimal number of taxa in numerous diagnosably different populations
comparison to methods based on frequency that bear valid species-group names, most
differences. The claim that the PSC over- of these are considered to refer to intraspe-
splits taxa relative to the GCSC is simply cific variations. Since the rejection of ram-
false. pant typological splitting in the early 20""
In summary. Mallet's GCSC is a meth- Century (e.g., Riffarth 1902), Heliconius
odologically explicit version of the nondi- species have been circumscribed primarily
mensional BSC (Mayr 1963). It is labor- by the BSC criterion of interbreeding. Un-
intensive, depending upon sampling ofmul- der that criterion, otherwise uniform para-
tiple individuals at multiple loci to provide patric populations that hybridize where they
empirical evidence of the absence of inter- abut have been considered conspecific.
breeding. It bases the decision of specific Likewise, distinct populations that hybrid-
distinctness on an arbitrary and unstated ize with each of two otherwise allopatric
level of phenetic bimodality in a histogram neighbors provide a transitive link that has
of average genetic scores. Although Mallet allowed lumping of chains of populations
and colleagues have made a rather convinc- into single, geographically extensive "bio-
ing case for the distinctness of Heliconius logical species." These species' component
erato cyrbia Godart and H. himera Hewit- "geographical races" are diagnosably dif-
son on the basis of this method (Jiggins et ferent (i.e., they display heritable characters
al. 1996), that work represents a laborious that allow their unambiguous determina-
multi-year, multi-authored effort to corrob- tion), and have been considered by many
orate a conclusion about a single pair of researchers (e.g.. Brown et al. 1974, Shep-
taxa that had already been hypothesized by pard et al. 1985, Brower 1996a, 1996b) to
systematists years before (Descimon and represent historically distinct entities. That
Mast-de Maeght 1984). How Mallet's con- their names are in common use in the lit-
cept could be useful in the best of circum- erature and in museum classification
stances to a museum taxonomist working schemes is a de facto acknowledgement of
with qualitative samples of preserved, dead their recognition as taxa, even by those bi-
specimens on pins is not clear. What is ologists who would emphatically deny their
quite evident is that Lamas has never em- specific status. According to the PSC, diag-
ployed the GCSC in any of his published nosably distinct groups that it is useful to
taxonomic work, including the descriptions name are considered separate species, and
of new subspecies in Lamas (1998). Given every recognized "geographical race" in
this lack of consistency and rigor, it is es- Heliconius should be a phylogenetic spe-
pecially ironic that my (1996a) names cies.
should be subject to such scrutiny, when The ICZN (Art. 45a) considers names at
684 PROCEEDINGS OF THE ENTOMOLOGICAL SOCIETY OF WASHINGTON
both specific and subspecific levels to be the arbitrarily determined ranks that partic-
labels for taxa ofa single category, the Spe- ular taxa do or do not represent. H. mel-
cies Group. The concept of subspecies is pomene mocoa and H. tristero are both
simply a convenient label for taxa at one of diagnosably distinct taxa of the Species
the potentially many hierarchical levels Group.
nested within the genus, as revealed by cla-
distic studies. The Code is sensibly silent Acknowledgments
on the problems ofdefinition and boundary Thanks to D. D. Judd and F. C. Thomp-
determination of concepts associated with
son for advice on the ICZN, A. Warren for
name-bearing type specimens; such deci- checking my translation ofLamas' Spanish,
sions are considered subjective, and left to
and A. Kretzer for help with Stichel's Ger-
the judgement of the describer and subse-
man. Thanks to R. I. Vane-Wright for his
quent employers of the name. In most in-
thorough and candid reviews, and to Ger-
stances, the older Heliconius names were ardo Lamas for making me play by the
described from one or a few dead speci-
rules.
mens in European collections, and were not
accompanied by any discussion of the cir- Literature Cited
cumscription of the associated concept, be-
yond designation of an intraspecific level Ackery, P. R. and R. L. Smiles. 1976. An illustrated
(e.g., subspecies, form, aberration, etc.). A list of the type-specimens of the Heliconiinae
good example of such a perfunctory de- (Lepidoptera: Nymphalidae) in the British Muse-
um (Natural Hi.story). Bulletin ofthe British Mu-
scription is SticheFs original diagnosis seum (Natural History) Entomology 32: 171-214.
(1923) of H. amaryllis amaryllisf. bellula Baum, D. A. and M. J. Donoghue. 1995. Choosing
(Appendix 2). among alternative "phylogenetic" species con-
Minimal original descriptions leave a cepts. Systematic Botany 20: 560-573.
great deal of latitude for subsequent inter- Baum, D. A. and K. L. Shaw. 1995. Genealogical per-
spectives on the species problem, pp. 289-303. In
pretation. Such interpretations should strive Hoch, P. C. andA. G. Stephenson,eds.. Molecular
to maintain nomenclatorial stability, but and Experimental Approaches to Plant Biosyste-
only when the names preserved are precise- matics. Missouri Botanical Garden., St. Louis.
ly and accurately associated with empiri- Brower, A. V. Z. 1994. Phylogeny of Heliconius but-
DNA
cally supported concepts. If a concept di- terflies inferred from mitochondrial se-
quences (Lepidoptera: Nymphalidae). Molecular
verges from the description due to the ac-
Phylogenetics and Evolution 3: 159-174.
quisition of new specimens and data to the 1996a. A new mimetic species ofHeliconius
.
point that the description is no longer ade- (Lepidoptera: Nymphalidae), from southeastern
quate, some action is called for, ranging Colombia, as revealed by cladistic analysis ofmi-
from redescription of the holotype to the tochondrial DNA sequences. Zoological Journal
ofthe Linnean Society 116: 317-332.
separate description of differentiated con-
cepts as distinct taxa. The currently-recog- luti.on19o9f6bm.imPiacrralyleilnHrealciecofnoirumsatbiuottneraflnides:thAe pehvyo--
nized "geographical races" (= phylogenet- logenetic hypothesis from mitochondrial DNA se-
ic species) of Heliconius are taxa corrobo- quences. Evolution 50: 195-221.
rated by a century of empirical research in cies. w1i9t9h9.DTNhAesdeeqluiemnicteatsi:onAocfritpihqyuleogoefnDeatviicssapned-
laboratories, museums, and the field. They Nixon's population aggregation analysis. System-
bear names originally applied to the single atic Biology 48: 199-213.
holotype or short type series, and to which Brower, A. V. Z. and M. G. Egan. 1997. Cladistics of
the current concepts correspond. The focus Heliconiusbutterflies andrelatives(Nymphalidae:
of modem systematic effort should be on Heliconiiti): The phylogenetic position ofEueides
based on sequences from mtDNA and a nuclear
the empirical diagnosis ofsuch taxa and the gene. Proceedingsofthe Royal SocietyofLondon
inference of hierarchical relationships B 264: 969-977.
among them, not on empty disputes over Brown, K. S., Jr. 1979. Ecologia Geografica e Evolu-
VOLUME NUMBER
102, 3 685
9ao nas Florestas Neotropicais. Universidade Es- Lamas, G. 1998. Comentarios taxonomicos y nomen-
tadual de Campinas, Campinas, Sao Paulo, Brasil. claturales sobre Heliconiini neotropicales con de-
. 1981. The biology of Heliconius and related signacionde lectotiposydescripciondecuatrosu-
genera. Annual Reviews ofEntomology 26: 427- bespecies nuevas (Lepidoptera: Nymphalidae:
456. Heliconiinae). Revista Peruana de Entomologia
Brown, K. S., Jr., R M. Sheppard and J. R. G. Turner. 40: 111-125.
1974. Quaternary refugia in tropical America: Ev- Nixon, K. C. and Q. D. Wheeler. 1990. An amplifi-
idence from race formation in Heliconius butter- cation of the phylogenetic species concept. Cla-
flies. Proceedings ofthe Royal Society ofLondon distics 6: 211-223.
B 187: 369-378. Mallet, J. 1993. Speciation, raciation and colorpattern
Davis, J. I. and K. C. Nixon. 1992. Populations, ge- evolution in Heliconius butterflies: evidence from
netic variation, and the delimitation of phyloge- hybrid zones, pp. 226-260. In Harrison, R. G. ed..
netic species. Systematic Biology 41: 421—435. Hybrid Zones and the Evolutionary Process. Ox-
de Queiroz, K. and M. J. Donoghue. 1988. Phyloge- ford University Press, Oxford.
netic systematics and the species problem. Cladis- . 1995. A species definition for the Modern
tics 4: 317-338. Synthesis. Trends in Ecology and Evolution 10:
Descimon, H. andJ. M. de Maeght. 1984. Semispecies 294-298.
relationships between Heliconius erato cyrbia . 1996. ReplyfromJ. Mallet.TrendsinEcology
Godt. and H. himera Hew. in southwestern Ec- and Evolution 11:174-175.
uador. Journal ofResearchon the Lepidoptera 22: Mayr, E. 1940. Speciation phenomena in birds. The
229-237. American Naturalist 74: 249-278.
Doyle, J. J. 1995. The irrelevance ofallele tree topol- . 1963. Animal species and evolution. Belknap
ogies for species delimitation, and a non-topolog- Press, Cambridge, Massachusetts.
ical alternative. Systematic Botany 20: 574-588. Mishler, B. D. andM. J. Donoghue. 1982. Speciescon-
Eldredge, N. and J. Cracraft. 1980. Phylogenetic Pat- cepts: A case for pluralism. Systematic Zoology
terns and the Evolutionary Process. Columbia 31: 491-503.
University Press, New York. Neustetter, H. 1929. Nymphalididae: Subfam. Helicon-
Eltringham, H. 1916. On specific and mimetic rela- iinae, pp. 1-136. In Strand, E. ed.. Lepidoptero-
tionships in the genus Heliconius. L. Transactions rum Catalogus, part 36. W. Junk, Berlin.
of the Entomological Society of London 1916: Riffarth, H. 1902. Nochmals Ch. Oberthur's Etudes
101-148. d'Entomologie, Vol. 21. BerlinerEntomologische
Emsley, M. G. 1964. The geographical distribution of Zeitschrift 47: 157-166.
the color-pattern components ofHeliconius erato Sheppard, P. M. 1960. Natural Selection and Heredity.
and Heliconius melpomene with genetical evi- Harper Torchbooks, New York.
dence for the systematic relationship between the Sheppard, P M., J. R. G. Turner, K. S. Brown, W. W.
two species. Zoologica NY 49: 245-286. Benson, and M. C. Singer. 1985. Geneticsandthe
Hewitson, W. C. 1854. Illustrations ofNew Speciesof evolution of Muellerian mimicry in Heliconius
Exotic Butterflies, V butterflies. Philosophical Transactions ofthe Roy-
. 1867. Descriptions of some new species of Sticheall,SoHci.ety19o2f3.LoKnodloonmbBis3c0h8:e 4H3e3l-i6c1o3n.ius. (Lep.,
diurnal Lepidoptera. Transactions ofthe Entomo-
Rhop.). Deutsches Entomologische Zeitschrift
logical Society of London, Third Series, 5: 561-
1923: 260-270.
566.
Holzinger, H. and R. Holzinger. 1994. Heliconius and Turner, J. R. G. 1971. Studies of Miillerian mimicry
and its evolution in burnet moths and heliconid
related genera. Sciences Nat, Venette, France.
butterflies, pp. 224-260. In Creed, R. ed.. Ecolog-
International Commission on Zoological Nomencla-
ical Genetics and Evolution. Blackwell Scientific
ture. 1985. International Code ofZoological No- Publications, Oxford and Edinburgh.
menclature. InternationalTrustforZoological No-
menclature and the British Museum (Natural His- Appendix 1
tory), London.
Translation of Lamas' (1998: 119-120) critique.
Jiggins, C. D., W. O. McMillan, W. Neukirchen, and
J. Mallet. 1996. What can hybrid zones tell us Twrearneslattrainosnlaitseadsfprreeecliysetoasimpopsrsoivbele,cobmuptreshoenmseibiidliiotmys.
about speciation? The case of Heliconius erato
For complete literature citations, see Lamas (1998)
and H. himera (Lepidoptera: Nymphalidae). Bio-
logical Journal of the Linnean Society 59: 221- Status of Heliconius tristero Brower and H. melpo-
242. mene mocoa Brower
Kaye, W. J. 1916. A reply to Dr. Eltringham's paper Recently, Brower (1996) has described two new
on the genus Heliconius. Transactions ofthe En- taxa of Heliconius from Putumayo, southeastern Co-
tomological Society ofLondon 1916: 149-155. lombia, based on preliminary data from analysis of
686 PROCEEDINGS OF THE ENTOMOLOGICAL SOCIETY OF WASHINGTON
mitochondrial DNA sequences. Both names are dis- Brower examined the male .specimen upon which Sti-
cussed separately here. chel based his infrasubspecific name bellula (and
Heliconiustristerowasdiagnosedby Brower(1996) which constitutes the holotype of bellula Turner) and
as a taxon of the species group, belonging to the H. decided that it represented an individual belonging to
cydno Doubleday clade, and narrowly separated from the cydno clade, and not to melpomene. Brower also
H. heurippa Hewitson and H. timareta Hewitson. Al- examined the "holotypes" of two other infrasubspe-
though he did ntn mention it explicitly, Brower seems cific names proposed by Stichel (1923), permira and
to imply that heurippa, tristero and timareta could degener (which have never been elevated to the sub-
constitute a subclade of the cydno group. The three specific level and thus are nomenclaturally unavail-
taxa exhibit an allopatric distribution, heurippa occur- able), suggesting that they might represent hybridsbe-
ring in the central eastern region ofColombia, tristero tween H. heurippa and H. tristero. What is more, he
in southeastern Colombia and timareta in eastern Ec- assumed that the type of bellula appeared to be a re-
uador (and northern Peru, vide infra); while heurippa combinant backcross between these hybrids and tris-
and tristero (so far as is known) are phenotypically tero. Arguing that the ICZN "forbids" the use ofspe-
monomorphic,timaretaispolymorphic.Brower(1996: cies-level names that are based on hybrids, he decided
330) concluded that ". . . tristero is considered a spe- to set aside the name bellula and establish new names
ciesonly because its geographically adjacentcloserel- for the member of the cydno clade (tristero) and the
atives, H. heurippa and H. timareta. have been tradi- subspecies ofmelpomene present in the upper Ri'o Pu-
tionally considered species as well." This conclusion tumayo (mocoa).
is not so logical astobe irrefutable. A more valid(and Brower does not seem to have realized that Stichel
refutable) hypothesis would be to consider the three (1923), in addition to bellula, permira and degener,
taxa conspecific, based on Brower's own molecular described four other "forms" of amaryllis from the
and morphological analyses (which would show their region ofthe upper Putumayo (anacreontica, perrara,
narrow evolutionary relationship), their allopatric dis- rufata andaglaspis), and thatotherauthors introduced
tributions (and geographical proximity), and the sup- sixadditionalnamesforspecimensfromthesamezone
posed existence of transitional forms ("hybrids") be- (parva Neustetter, tenuifasciata Neustetter,aurofascia-
tween heurippa and tristero (represented by the ta Neustetter, paula Neustetter, pau—lina Niepelt and
"type" ofbellula Stichel, but vide infra), which might carminata Niep—elt). In my opinion and contrary to
suggest that, as Mallet (1995) put it nicely in his def- Brower's view these 13 names pertain to examples
inition of species as "genotypic aggregations:" ". . . of melpomene, and represent transitional forms ("hy-
closely related allopatric forms should mostly be con- brids") between the subspecies bellula Turner and
sidered conspecific." Naturally, even though it is well malleti Lamas. All those specimens were captured
known that in the absence ofevidence ofspecific dis- around Mocoa (01°09'N, 76°37'W) by collectors who
tinctness in sympatry the conspecificity of allopatric worked for Werner Hopp, and were very likely ob-
taxaisarbitrary,sucharbitrarinessiscertainlylessthan tained in the hybrid zone near Villa Garzon (= Villa
to assign a taxon to a particular taxonomic level fol- Amazonica; 0r05'N, 76°35'W, 420 m), a site studied
lowing a "tradition." by Mallet (1993), a few km. East of Mocoa. The ho-
In synthesis, there are two opposing taxonomic hy- lotype ofbellula is labeled as having beenobtainedon
potheses: 1) tristero is a taxon at the species level, the Rio Guayuyaco (written "Guagzayaco" on the la-
evolutionarily independent of heurippa and timareta, bel), possibly very near the village of Guayuyaco (=
with its own historical destiny; or2) heurippa, tristero Napoles; 01°04'N, 76°26'W). The other specimens
and timareta constitute a polytypic species with three bear the localities Rio Mulato (OTOS'N, 76°36'W),
geographical races (subspecies) which have the same 500 m; Mocoa, 530 m; orsimple "Mocoa" (foragen-
historical destiny. With the scarce taxonomic and ge- eral description ofthe area, see Salazar 1995). Mallet
netic information available at the moment, it is impos- (1993) referred to the hybrids found nearVillaGarzon
sible to decide which of these is closer to the truth. as the product of the transition between bellula and
The case of Heliconius melpomene mocoa seems aglaope C. & R. Felder; the correct name ofthe latter
much simpler, and here I offerthe hypothesis thatmo- subspecies is malleti Lamas, since the subspecies
coa is no more than a new synonym of H. m. bellula aglaope is limited to the lower Rio Maranon, lower
Turner, 1971, as I will demonstrate below. The name Rio Huallaga and the Rio Ucayali basin in Peru (La-
bellulawasintroducedforthefirsttimeintheliterature mas 1988).
by Stichel (1923), who proposed it as a form of Hel- Of the 14 names applied to bellula X malleti hy-
iconius amaryllis amaryllis C. & R. Felder, thus con- brids, the "purest" example corresponds to the holo-
stituting an infrasubspecific name (excluded by the type ofbellula, which only shows tiny yellow spots at
ICZN). The name bellula recently became available the costal inner edge ofthe discal red forewing band,
when Turner(1971) elevated it tothe subspecific level appearing almost identical to the holotype of mocoa
(as a subspecies of H. melpomene (L.), making the (which lacks the yellow spots). Brower based his
name attributable toTurner(ICZN Arts. 10c, 23j, 50c). judgement on the presence ofthese yellow spots, and
VOLUME NUMBER
102, 3 687
the morphology of the genitalia of the belhda holo- //. amaryllis amaryllis Feld.
type, to assert that that specimen was a heurippa X Forma typ. —
tristero "hybrid," (the same aspermira and degener). Two males. Forewing with a somewhat variable
But, in the first place, heurippa (one of the supposed broad red discal patch, somewhat like the figure of//.
parents) is not known from the Mocoa region, only amaryllis euryades Riff, in Gen. Ins. v. 112, plate 3,
from much further north, from the Meta and Guaya- fig. 10, but with a more blurred distal margin, and
bero basins, while surely tristero should occur in the which is somewhat indented below the forewing me-
basin ofthe Rio Orteguaza, since belhda has been re- dian vein. The yellow transverse band ofthe hindwing
ported from there. (It wouldbe very interesting todis- fragmented near the base by thin black lines, and the
coverifheurippa, ortristero, orboth, occurinthe Rio veins cutting across the band are more or less black.
Caguan basin, between the Guayabero and the Orte- Mocoa (Put[umayo]), September, October.
guaza). Further, even though Brower admitted that Among other things, a character of this species is
"male genitalia of the taxa examined are similar, and saidtobe the absenceofredbasal streaksonthecostal
display substantial intra-racial variability in form," he margin ofthe forewing underside. This feature is pre-
concluded that the genital morphology of the bellula sent in all ofthe observed nominate and closely allied
holotype corresponded to the cydno clade, based on forms, but based on the other specific characters, par-
the examination ofthat one specimen. ticularly the position of the hindwing band, they can
Finally, Browerarguedthatthe name bellula should only be considered forms of amaryllis. This band is
be set aside because it is applied to a hybrid (which, somewhat variable, but it is always positioned such
as indicated by the discussion above, I consider com- that the posteriorborderlies outside the posterioredge
pletelyerroneous)andbecausetheICZN "forbids" the of the discal cell. In a third male specimen (Mocoa,
use ofnames applied to hybrids. In the glossary ofthe May), the forewing band is somewhat smaller, so that
Code (1985: 256) it is clearly indicated that "The it scarcely touches the end ofthe cell, the red is faded
progeny oftwo individuals belonging to different sub- and greasy, which is often seen as a pathological fea-
species of same species are not hybrids." The Code ture ofred-banded Heliconius.
does not "forbid" the use ofnames given to interspe-
cific hybrids, it simply excludes them from nomencla- Forma bellul—a f. nov.
ture, exceptforthe principle ofhomonymy. Therefore, One male. Most similar to the nominate form, the
if the name bellula does not correspond to an inter- crimson forewing patch more ragged on the margin,
specific hybrid, its use todesignate a subspecifictaxon partly dusted with black scales, the cell almost com-
is perfectly valid, and in consequence,Heliconiusmel- pletely free ofred, an additional character is a sulfur-
pomene mocoa Brower, 1996 is a subjective junior yellow spotproximaltothe subcostal vein. Theyellow
synonym of Heliconius melpomene bellula Turner, transverse band on the hindwing is broad, extending
1971. nearly 7 mm. from the apex to the trailing edge.
Hindwing beneath very similar to //. amaryllis rosina
Appendix 2 Bsd. With red basal streaks on the leading edge ofthe
Translation ofStichel's (1923)descriptionof//, am- forewing and three red basal spots in the hindwing.
aryllis amaryllis and //. amaryllis amaryllis f. bellu- Forewing length 41 mm. Rio Guaqzayaco
la.from the upper Putumayo ofColombia. (Put[umayo]).