Abrupt Geographical Transition between Aposematic Color Forms in the Spittlebug Prosapia ignipectus (Fitch) (Hemiptera: Cercopidae)

1Office of the President, Metropolitan College of New York, 60 West Street, New York, NY 10006, USA 2Department of Entomology, Division of Invertebrate Zoology, American Museum of Natural History, Central Park West at 79th Street, New York, NY 10024, USA 3Laboratório de Entomologia, Faculdade de Biociências, Programa de Pós-Graduação em Zoologia, Pontif́ıcia Universidade Católica do Rio Grande do Sul, Av. Ipiranga 6681, 90619-900 Porto Alegre, RS, Brazil


Introduction
Many organisms have conspicuous coloration that warns potential predators of strong defenses, such as toxic chemical compounds, potent venom delivery systems, or particularly effective escape mechanisms [1,2].This aposematic, or warning, coloration signals to potential predators that a particular species is not a desirable subject for pursuit.Individual predators learn to avoid aposematic prey through initial unpleasant or unfruitful encounters, sparing subsequently encountered individuals with similar warning coloration from predation.This results in positive frequency-dependent selection for conspicuous coloration.As a result, local populations should become monomorphic for a single color form [1][2][3] and, by extension, closely adjacent populations should over time become monomorphic for the same color form through displacement of one form by another [4][5][6].Here we detail the existence of local populations of the spittlebug Prosapia ignipectus (Fitch) that violate these expectations.

Area of present study
Figure 1: Approximate species ranges of Prosapia ignipectus (in black) and Prosapia bicincta (in orange).The populations examined in this study lie in the far northeastern portion of the P. ignipectus range in the area of the state of Maine outlined in red (see Figure 6).Yellow dots represent the occurrence of populations with mixed color forms, nominally all P. ignipectus, towards the southern part of the P. ignipectus range (see text for details).
3(b)), a pasture pest in the southeastern USA [11].The two species are morphologically similar.P. ignipectus was originally described by Fitch [12] as Monecphora ignipecta.It was synonymized as a variety of Monecphora bicincta by Van Duzee [13].Fennah designated P. bicincta as the type species of the newly created genus Prosapia [14] and later included the form ignipecta as a geographical subspecies of P. bicincta [15].Hamilton reinstated P. ignipectus as a distinct species [7].
P. ignipectus is distributed across the northeastern USA from Minnesota to Maine and in the southernmost portion of Ontario, Canada.P. bicincta is distributed in the southeastern USA from Oklahoma to Florida (Figure 1).The two species are distinguished by geographical distribution, by subtle differences in the male genitalia [7], and by differences in host plant associations.P. ignipectus is monophagous on the late season C4 perennial grass Schizachyrium scoparium (Michx.)Nash, Little Bluestem [16,17].Nymphs live on Little Bluestem roots.Adults emerge in late July or early August and feed on stems of the host [18,19].Although the range of Little Bluestem broadly overlaps the range of P. bicincta (compare Figure 1 with [20]), P. bicincta has never been reported on Little Bluestem and appears to be absent from Little Bluestem in New Jersey (VT observations) and North Carolina (J.Urban, pers.comm.).P. bicincta is polyphagous on a wide variety of C4 grasses [11,17].Nymphs live on the roots of host grasses and adults on the upper parts of the same hosts.Adult males of P. bicincta feed on ornamental hollies (Ilex spp.) in addition to C4 grasses [21].
Adults of both species exhibit reflex bleeding ( [22] and Figures 4(a) and 4(b)).When molested they release an odiferous yellow liquid from their tarsi.This behavior deters some predators and is probably the primary basis for the colorful patterns widespread among the Cercopidae [22] (see examples of striking coloration in [9,10]).Consistent with this behavior, both P. bicincta and P. ignipectus have dorsal coloration that is conspicuous in natural settings.P. bicincta individuals have a black dorsal surface traversed by a single narrow lateral orange line across the widest part of the pronotum and a pair of narrow orange lines across the elytra (Figure 3 In 1920, while collecting in preparation for a report on the Orthoptera of Maine, Albert Morse stumbled across a Prosapia population in Norridgewock, Maine, which consisted entirely of individuals with P. bicincta-like dorsal coloration [18,23].Morse interpreted these specimens as a disjunct New England occurrence of P. bicincta.With the exception of one piece of contemporary work [19], Morse's observation of lined Prosapia in Maine went unremarked and uninvestigated for several decades.Many, including the authors, assumed that the report of lined specimens in Maine, about 500 km northeast of the known range of P. bicincta, was in error.However, reexamination of Maine populations has vindicated Morse's original observation, though not his interpretation, and disclosed an unusual case of polymorphic aposematism rivaling, on a smaller scale, some of the paradigmatic cases observed in the tropics [24].

Both the Lined and Black Maine Forms Are P. ignipectus, Which Is Distinct from P. bicincta. Morse believed that his
Norridgewock specimens represented a geographically disjunct population of P. bicincta, a judgment apparently based solely on the presence of the lined orange pattern characteristic of P. bicincta.GC dissected genitalia of representative males of southern P. bicincta and black and lined specimens from the Maine populations in question.Male genitalia of the black and lined Maine forms are indistinguishable.Both differ from the genitalia of P. bicincta consistent with the subtle interspecific differences in the apices of the styles illustrated by Hamilton [7].We conclude that both color forms belong to P. ignipectus and, in concurrence with Hamilton [7], that P. ignipectus is distinct from P. bicincta.The specific difference is also supported by subtle differences in shape between the species (P.ignipectus more slender and P. bicincta more robust), by the consistent difference in host plant associations noted above, and by the allopatric distribution of the taxa (Figure 1).DNA bar code analysis, based on sequences of a single mitochondrial gene, indicates that the species are closely related [25], probably reflecting recent speciation.

Populations Monomorphic for the Lined Form Are
Uniquely Clustered at the Northeastern Species Margin and Separated from Populations Monomorphic for the Black Form by Less Than 10 km.Table 1 reports the color forms encountered in 32 P. ignipectus collections made in 2003, 2004, 2006, and 2012 between the New Hampshire border and Norridgewock.These populations represent the northeasternmost part of the species range.To the immediate south, between Norridgewock and the Atlantic coast, Little Bluestem appears to be absent.In some areas east and northeast of Norridgewock, for example, in parts of the Penobscot River Valley, Little Bluestem is abundant but P. ignipectus is absent.In contrast, P. ignipectus is common on Little Bluestem from Norridgewock north almost to Bingham in the Kennebec River Valley and along some of its tributaries to the west as well as in a portion of the Androscoggin River Valley centering on the city of Rumford.North and west of these populations, occasional Little Bluestem patches without P. ignipectus extend up small river valleys into the Longfellow Mountains.The absence of P. ignipectus on Little Bluestem to the north and east indicates that factors other than host plant distribution limit its range.
Populations monomorphic for the lined form (Figures 5(a) and 5(b)) lie in a discrete cluster at the extreme northeastern edge of the species distribution (Figure 6).They encompass an area of about 675 km 2 bounded roughly by the towns of Norridgewock, Bingham, Carrabassett, and New Portland.Less than 10 km to the southwest in New Vineyard they are replaced by populations essentially monomorphic for the black form (with the exception of a single male individual taken in New Vineyard in 2004 that had a partially obscured single orange line across the pronotum) (Figure 6 and Table 1).These populations are linked to other monomorphic black populations to the southwest by a monomorphic black population in Wilton.The area in which the collections were taken varies in elevation from about 75 to 250 m and consists of low forest cover, broken by bogs, streams, human habitations, and farmland.There are no significant barriers to P. ignipectus dispersal and the roadside habitats in which P. ignipectus occurs are similar throughout the area.All of the monomorphic lined populations lie in the Kennebec River drainage, but this drainage also includes the monomorphic black populations in New Vineyard and Wilton.Monomorphic black populations to the immediate southwest lie in the Androscoggin River drainage (Figure 6).one in Rumford Center and two in nearby South Rumford on the opposite side of the river, exhibit a mixture of black, lined, and partially lined individuals (Figure 6 and Table 1).These populations are 45 kilometers from the nearest monomorphic lined population to the northeast.In each case the lined individuals are phenotypically variable: some fully lined and some partially lined, most of the partially lined specimens with a single line across the pronotum.All populations sampled in Maine to the south and southwest of Center Rumford as well as in adjacent parts of New Hampshire were monomorphic for the black color form (Figure 1 and VT unpublished observations), consistent with Morse's observations in Maine almost a century earlier [18].In 1993-2000 James Bess collected short series of P. ignipectus specimens, some lined and some black, from Little Bluestem in fens near the Michigan-Indiana border, close to what appears to be the boundary between the P. ignipectus and P. bicincta distributions in the central Midwest (see [26] for background on the habitats and insect associations).At Liberty Fen in Jackson Co., MI, Bess collected a lined male and a black male.At Sawmill Fen in LaGrange Co., IN, he collected a lined male and a black female.At Indian Bowl Fen and nearby in Berrien Co., MI, he collected a lined male and a black male.At Brant Road Fen in Oakland Co., MI, he collected a lined male.

Additional Natural History Observations. Little
Bluestem, a grass of nutrient-deficient soils, is an attractive host for spittlebugs, probably because it exhibits associative nitrogen-fixation [17] Psyche 7 speculation by Morse [18] and Boring [19] and a statement by Hamilton [16] that the nymphs are root feeders.Also, while surveying the Little Bluestem associated populations recorded in Table 1, we collected four other spittlebug species, all of the family Aphrophoridae: Philaenus spumarius and Neophilaenus lineatus (both introduced from Eurasia) and the native spittlebugs Lepyronia quadrangularis and Philaenarcys killa.Only L. quadrangularis occurred in sizable numbers at multiple sites.It is clearly associated with Little Bluestem [17] but has a much broader host range [16].P. killa, in contrast, is a Little Bluestem specialist [17].In this study it cooccurred with P. ignipectus only in West Bethel, the westernmost Maine site sampled.It cooccurs frequently with P. ignipectus on Little Bluestem in Central New Hampshire (VT observations).P. spumarius occurs primarily on broadleaved herbs [16] and was likely associated with herbaceous dicots mingled with Little Bluestem.N. lineatus lives primarily on cool season C3 grasses [17] and was likely associated with C3 grass species mingled with Little Bluestem.

Discussion
The observation that P. ignipectus has populations monomorphic for alternative aposematic color forms sharply separated by only a few kilometers seems to be unique among polymorphic aposematic organisms.The closest parallels that have come to our attention form part of a complex Müllerian mimicry system involving the millipede genera Apheloria and Brachoria in the Appalachian Mountains of the USA.Marek and Bond [27] present detailed data suggesting that, in three of the seven species they investigated, populations monomorphic or nearly monomorphic for contrasting light and dark color forms occur in close geographical proximity.In the taxon Apheloria clade B a series of apparently monomorphic "chess" color form populations is separated from an apparently monomorphic "striped" color form population by only about 20 km.Similarly, in Brachoria mendota a series of populations that appear to be monomorphic for the "taillight" color form is separated by only 8-10 kilometers from populations that appear to be monomorphic for the contrasting "caution/road" color forms.Neither of these cases involves separation by a major mountain ridge, but a high ridge does separate a single population of Brachoria dentata monomorphic for the "headlight" color form from populations apparently monomorphic for the contrasting "caution/road" color form about 10 km distant.In many of these localities the sample sizes are small, but in each case the distributions suggest a sharp boundary of less than 10-20 km separating populations monomorphic or nearly monomorphic for contrasting aposematic forms.
In other reported cases, populations monomorphic for alternative warning color forms are clearly linked by mixed populations in the intervening territory.For example, in the bumblebee Bombus melanopygus, California populations monomorphic for red abdominal color are separated from Washington State populations monomorphic for black abdominal color by a 600 km cline of populations polymorphic for both [28,29].In Peru, populations of the poison-dart frog Ranitomeya imitator lying just 15 km apart are monomorphic for alternative aposematic color forms but are linked by intermediate mixed populations [30,31].In the well-studied Müllerian mimicry system involving the Neotropical butterflies Heliconius erato and Heliconius melpomene, populations of each species monomorphic for parallel mimetic forms are linked by hybrid clines [4,32].At one location in Peru, populations monomorphic for the "postman" and "ray" forms of the these butterflies are separated by a zone as narrow as 10 km [3], but in this and similar cases hybrid linking populations are the rule (see Figure 1 of [33]).
Sharp contact zones imply strong natural selection; the narrower the zone the stronger the selection [3].The continued separation for at least 90 years of Maine lined and black populations in close proximity suggests that strong natural selection is maintaining homogeneity for the alternative color forms.Though there are no direct data bearing on their dispersal ability, P. ignipectus adults are strong fliers, making substantial gene flow over the short distance between monomorphic black and monomorphic lined populations not only possible but likely.In addition, there are indications that contact zones between aposematic forms may be temporally unstable and subject to a sweep of one form by another [4][5][6].This raises two questions concerning the Maine populations: what special selective forces, if any, maintain the lined color form and why is the division between the forms so sharp and, apparently, stable?We say apparently stable recognizing that Morse seems to have sampled only one location in the area covered in this study.He reported observing black specimens in unspecified locations between Norway, Maine, and Norridgewock [18], indicating that the separation in 1920 was, at most, shorter than the 80 km separating those locations.
Part of the answer may lie in the migratory avian predators that P. ignipectus encounters in its Maine habitats.Insectivorous birds migrating to and from the southeastern USA would encounter P. bicincta in abundance in many grassy habitats, including cattle pastures.Likewise, birds migrating to and from Mexico and Central America would encounter Neotropical pasture spittlebugs with similar lined color patterns, including the phenotypically similar species Prosapia simulans, Aeneolamia contigua, and Aeneolamia albofasciata [34,35], which constitute parts of a putative Müllerian mimicry ring in pastures and sugar cane plantations from Mexico to Costa Rica (VT observations).
It is likely that some of the avian predators that encounter P. ignipectus in Maine have already learned an aversion to lined color patterns at points south and are predisposed to select in favor of lined color patterns in Maine.However, this in itself does not explain why the lined form has achieved monomorphism in one small geographical area over a protracted period of time.Work with the arctiid moth Parasemia plantaginis in Finland suggests that the composition of local avian predator communities can influence the relative fitness of polymorphic aposematic color forms [36] and local populations of the same avian predator can Psyche vary in their response to aposematic prey [37].This suggests that detailed study of the foraging and migration patterns of Maine insectivorous birds might shed light on P. ignipectus color form selection.There are hints in the literature that some Neotropical insectivorous birds specialize in jumping Hemiptera, including Cercopidae [38,39], though none of the species studied are migratory to Maine.
Independent of the selective factors that might maintain the two distinct aposematic color forms in close proximity, the sharp boundary, apparently without intermediate polymorphic populations, itself poses interesting evolutionary questions.How is it possible to maintain such a sharp boundary for a strong flier in a region with no significant barriers to dispersal?One possibility is that factors other than aposematic selection reinforce the separation of populations.For example, Lis et al. [40] suggest that the endosymbiont bacterium Wolbachia may play a role in maintaining genetic barriers at contact zones between mitochondrial linages of the spittlebug Philaenus spumarius in Eastern Europe.Likewise, in the absence of definitive genetic evidence to the contrary, it is possible that the lined and unlined Maine populations represent an undetected case of cryptic speciation.If so, similar questions regarding lack of long-term selection for a single color form would still apply, but in the context of two species, not one.
The existence of three Rumford area populations polymorphic for aposematic forms violates the expectation that local populations of aposematic organisms will be monomorphic, but there is ample precedent for polymorphic warning coloration in the Cercopidae.For example, P. simulans, an abundant and widespread pasture grass pest in Mexico and Central America, has females polymorphic for a lined form and a dark form [34,35]. Color polymorphism is also present in populations of the Costa Rican pasture pests Prosapia plagiata and Zulia vilior [35], the Costa Rican Heliconia spittlebug Mahanarva costaricensis [41], and the South American sugar cane pest Mahanarva bipars [42], among others.Consequently, whether the Rumford area local polymorphism proves to be stable or fleeting, it is consistent with the observation that the Cercopidae as a group are susceptible to color polymorphism [10].
Interestingly, local polymorphisms for aposematic color forms also occur in many other taxa, despite ample theoretical expectation that they should be rare and transient [1,2,6,36,[43][44][45][46].These "almost embarrassingly frequent" occurrences of polymorphism [6,47] may be due to simultaneous multiple predator selection for warning coloration and rarity [48]; to spatially varying selection pressures in complex Müllerian mimicry systems [6]; to complexities of predator sampling strategy [47]; to balancing selection for warning coloration, sexual signaling, and thermal melanism [2]; or to a mélange of factors [43,49].Additionally, it has been suggested that relatively weak defenses [43,[50][51][52] or defenses that vary in effectiveness against mixed predators [53] might be especially conducive to polymorphism for aposematic forms.This hypothesis may be applicable to spittlebugs as a group, including some species in the Aphrophoridae and Clastopteridae that lack the reflex bleeding characteristic of the Cercopidae (see [54] and Figures 7 and 42-45 in [16]).
The rare cooccurrence of black and lined forms of P. ignipectus in populations close to the bicincta-ignipectus species boundary raises another set of questions, unlikely to be resolved in the absence of molecular genetic analysis.These populations might reflect introgression through hybridization, analogous to the hybrid zones between parallel color forms of Heliconius erato and H. melpomene [3].Alternatively, they could reflect incomplete selection for homogeneity of the black form following an earlier speciation event in which P. bicincta gave rise to P. ignipectus; genetic reversion to a lined phenotype only recently lost as suggested by Boring [19]; or altered avian selection regimes due to closer proximity to P. bicincta populations to the south.
In his initial report of lined Maine spittlebugs, Morse [18] put the question succinctly: "Why the banded form alone should be found at the Norridgewock locality and only there . . .at the northern limit of the distribution of the species . . . is as yet an unsolved problem."The problem is still unsolved but perhaps now within reach of solution, given recognition that the color forms in question are aposematic and the ongoing development of comprehensive tools for the analysis of selection for warning coloration [1,6,44,47].Because they involve a single species, without the complicating presence of local Müllerian mimics, Maine P. ignipectus populations may present an especially tractable case for analysis of the causes and context of polymorphism for warning coloration.
) and 2(b)), the northernmost representative of the family in the New World, and Prosapia bicincta (Say) (Figures3(a) and Psyche Black: Prosapia ignipectus us Orange: Prosapia bicincta (a)).With exceptions described below, P. ignipectus individuals have uniformly black dorsal coloration (Figure 2(a)).Both species stand out against the grass stems on which they characteristically sit prominently exposed (Figures 5(a) and 5(b)).The abdomen of P. ignipectus, which is visible while flying, is predominantly red or reddish orange (Figures 2(b) and 5(b)), as are the legs, although there is variation in the intensity of the red and in its combination with darker tones on the legs and on the red and black patterned ventral surface.In contrast, red markings on the abdomen and legs of P. bicincta are more subdued (Figure 3(b)).

Figure 4 :Figure 5 :
Figure 4: Reflex bleeding in Prosapia bicincta: (a) droplets of yellow liquid emanating from tarsi, photograph by Sam Houston; (b) bleeding in response to attack by spider (see droplet of yellow liquid on front tarsus to left in picture and another droplet caught in the spider web in the background), photograph by Andy Williams.

Figure 6 :
Figure 6: Distribution of Prosapia ignipectus populations with lined, black, and mixed dorsal color forms at the northeastern extremity of the species range in Maine.Some closely spaced localities sampled are represented by a single color symbol.In New Vineyard, a single partially lined specimen was observed among 114 specimens collected; the rest were black.
Populations near the P. bicincta-P.ignipectus Species Boundary Exhibit Mixed Color Forms.The AMNH collection includes one black specimen and one lined specimen (two elytral lines but no pronotal line) of P. ignipectus collected in Tuxedo Park, NY, in 1928.It also includes two lined specimens (one male and one female) collected along with two black specimens (one male and one female) on Staten Island, NY, in 1970.A sweep of Little Bluestem 20 km from Tuxedo Park by VT in August 2014 yielded 42 P. ignipectus specimens, all black.The Staten Island site appears to have been obliterated by residential development.Several other AMNH historical collections from northern New Jersey and Long Island, NY, consist entirely of specimens of the black color form.

Table 1 :
. During our sampling we searched for and found Prosapia ignipectus nymphs living slightly below ground level in Little Bluestem clumps, confirming Psyche Prosapia ignipectus collections in Maine, September 2003-August 2012.All specimens were swept from patches of Little Bluestem (Schizachyrium scoparium (Michx.))growing on C: female; D: male.