Revision of the “ Aemilia ” ambigua ( Strecker ) species-group ( Noctuidae , Arctiinae )

Th e New World taxa related to Aemilia ambigua (Strecker) are revised and transferred to the genus Pseudohemihyalea Rego-Barros, resulting in the following nomenclatorial changes: Pseudohemihyalea ambigua comb. n., P. fallaciosa (Toulgoët) comb. n., and P. syracosia (Druce) stat. rev., comb n. Two Mexican species are newly described, P. sonorosa sp. n., and P. potosi sp. n. “Aemilia” carmen Schaus is also closely related to the ambigua-group, and is transferred to Pseudohemihyalea, comb. n. Like most species of Pseudohemihyalea, the ambigua species-group is restricted to southwestern North America and Central America. Th e forewing pattern and habitat association of the ambigua-group are likely the result of a novel larval host switch from broadleaf trees to pines (Pinus spp., Pinaceae). Adults and genitalia of all species are illustrated, except the female of P. potosi which is unknown. A key to the species of the ambigua-group is provided.


Introduction
Th e genus Aemilia Kirby currently includes six neotropical species, with an additional nine species misplaced in the genus (Watson and Goodger 1986;Toulgoët 1997) including the North American taxon A. ambigua (Strecker).Aemilia (sensu stricto) was associated with the Euchaetes Harris generic group by Watson and Goodger (1986) within the phaegopterines (Arctiini: Phaegopterina; family-group names used herein follow the changes proposed by Lafontaine and Fibiger 2006).Aemilia ambigua was correctly associated with Hemihyalea Hampson by Franclemont (1983), who placed A. ambigua prior to Apocrisias Franclemont and Hemihyalea in the North American checklist.Th e generic group placement of Aemilia within the Phaegopterina remains uncertain; based on the male genitalia of the type species, Aemilia rubriplaga (Walker, 1855), Aemilia does not appear to belong to the Euchaetes group as it lacks the characteristic complex structure of the uncus (Weller et al. 2008), nor does it belong to the Amastus-group, here defi ned as including Amastus Walker, Pseudohemihylaea Rego-Barros, Apocrisias Franclemont, and Praeamastus Toulgoët.
As discussed below, the ambigua-group species exhibit structural characters unambiguously placing them in Pseudohemihyalea Rego-Barros, although the wing pattern of the ambigua-group is atypical of Pseudohemihyalea.Th e forewing pattern of longitudinal striae along the wing veins (Figs.1-5), appears to mimic the dried pine needles common in the habitat of these moths, whereas the typical wing pattern of other Pseudohemihyalea consists of a weakly defi ned transverse pattern (see for example Toulgoët 1994).
Although Pseudohemihyalea has received more study than most neotropical arctiines, the dissimilarity in wing pattern between the ambigua-group and other Pseudohemihyalea has likely resulted in the oversight of this group's true generic affi nity.
Th e purpose of this paper is to revise the generic placement and provide a taxonomic review of the "Aemilia" ambigua species-group and the closely related "Aemilia" carmen (Schaus).Two new species are described, and one name is raised from synonymy.

Methods and materials
Adult genitalia were prepared following the methods detailed by Lafontaine (2004).Line drawings were prepared from genitalia suspended in 30 % ethanol, using a camera lucida mounted to a Leica M-165C dissecting microscope.Approximately 500 specimens were examined from the following collections: CNC, CSU, CMNH, DEB and USNM.Museum), Washington, DC, USA.

Generic placement of the "Aemilia" ambigua-group
Th e unnatural placement of Aemilia ambigua within Aemilia was recognized by Watson and Goodger (1986), who placed this taxon within Aemilia "sensu lato."An additional Aemilia species, A. carmen Schaus, was inadvertently omitted by these same authors.While investigating the generic placement of A. ambigua, it became apparent that the male genitalic structure of A. ambigua is virtually identical in gross morphology to that of A. carmen (illustrated in Watson 1971), despite marked diff erences in wing pattern."Aemilia" carmen and the ambigua-group are very similar structurally to Pseudohemihyalea daraba (Druce) and P. anapheoides (Rothschild) and to a lesser extent also to P. testacea (Rothschild) and P. ochracea (Rothschild).Furthermore, the phenotype of P. anapheoides shows characters transitional between those of the typical banded Pseudohemihyalea pattern and that of the ambigua-group in that the forewing banding is highly reduced, the ground colour is whitish yellow, and the veins are outlined in rusty brown.Th e pink colouration of the dorsal abdomen is shared among P. ambigua, P. schausi, P. testacea, P. daraba and P. anapheoides.Th e following structural characters are shared between A. carmen, the ambigua-group, P. daraba and P. anapheoides: base of uncus broad and lobe-like (deeply excavated in P. schausi and the P. edwardsii group), apex of uncus tapering to a point (bifi d in edwardsii group), process of transtilla relatively small and scobinate (with large cornuti in some Pseudohemihylea species; transtilla elongate, large and fi nger-like in Amastus); valve fl attened and lobate overall, divided into two lobes beyond apical third or less (deeply divided and/or with a third, costal process in some Pseudohemihylaea and most Amastus species).Wing venation, palp structure and structure of the spines on the legs are fairly constant across Amastus and Pseudohemihylaea, and are consistent with those of the ambigua-group and A. carmen.Th e close relationship between ambigua and carmen as indicated by male genitalic morphology (female carmen were unavailable for study) is also supported by molecular data (mtDNA cox1 gene), with the two clustering together (4 % divergence) in a neighbour-joining tree containing representatives of most Central and North American arctiine genera (C.Schmidt, M. Laguerre, B. Vincent, unpubl.data).Both ambigua and carmen group within the current concept of Pseudohemihyalea, including the type species, P. schausi Rothschild.Based on this morphological and molecular evidence, Pseudohemihylea ambigua comb.n. and P. carmen comb.n. are accordingly transferred to Pseudohemihyalea.Pseudohemihyalea carmen and P. daraba are possibly the sister group to the ambigua-group, as suggested by morphology and mtDNA sequence data (COI barcode fragment).Th e striate forewing pattern of the ambiguagroup appears to be a derived trait, likely linked with a larval host shift to conifers from the broad-leaved trees utilized by other Pseudohemihylaea species (e.g., P. edwardsii on Quercus species; McFarland 1975).Larvae of P. ambigua feed on Ponderosa pine (Pinus ponderosa Dougl.ex Lawson) (R. Nagle, pers.comm.), and it is probable that other species of the ambigua-group also feed on pines.Th e striate white-and-tan wing pattern (mimicking dead pine needles) is an interesting example of convergent evolution in cryptic colouration in pine-feeding Lepidoptera, as a similar pattern occurs also in such unrelated groups as the Geometridae, such as the Caripeta piniata (Pack.) group and particularly Sabulodes niveostriata (Cockerell, 1894), which often occurs in strict sympatry with P. ambigua.A parallel (but evolutionarily independent) host switch has occurred (possibly multiple times) in Lophocampa Harris, where two lineages (L.roseata-group and L. argentata-group) feed on conifers, compared to deciduous trees for most congeners.
Pseudohemihyalea has a long and confusing taxonomic history, primarily a result of confusion with the genus Amastus.In describing Hemihyalea, Hampson (1901) distinguished this genus from Amastus by diff erences in the branching of the forewing radial veins.As noted by Dyar (1914), the branching pattern of the radial vein is highly variable in Hemihyalea and Amastus (as it is in a number of other arctiine genera such as Grammia, Apantesis and Phragmatobia), so this character is not diagnostic.In reviewing Hemihyalea, Rego-Barros ( 1956) recognized two additional genera, Machadoia Rego-Barros and Pseudohemihyalea, although only six species were examined.Pseudohemihyalea was later placed into synonymy under Hemihyalea by Watson and Goodger (1986).Toulgoët (1988) considered the type species of Hemihyalea (Phaegoptera cornea Herrich-Schäff er, 1853) to be congeneric with Amastus and accordingly synonymized the two, subsequently raising Pseudohemihylaea from synonymy (Toulgoët 1992) and providing a review of the genus (Toulgoët 1994), consisting primarily of those species treated as Hemihyalea by Watson and Goodger (1986).To complicate matters further, when describing Pseudohemihyalea, Rego-Barros believed the type-species to be Phaegoptera rhoda Druce, 1894 but the species he diagnoses and illustrates is Hemihyalea schausi Rothschild, 1909 (Toulgoët 1994).Following Toulgoët's (1994) review, several additional taxa were described or revised (see Toulgoët 1996Toulgoët , 2001)).
In summary, Pseudohemihyalea as currently defi ned is a relatively small group of about 30 species restricted to Central America and southern North America.Toulgoët (1994) arranged 20 species of Pseudohemihyalea into three groups, but provided no diagnostic characters or synapomorphies for the genus or for the three species-groups.Genitalic structure is very diverse in the mansueta-group (Toulgoët 1994) which may prove to be a polyphyletic assemblage.In addition, at least one species likely does not belong to Pseudohemihyalea (Toulgoët 1994), so much work is still needed to establish relationships among Pseudohemihyalea and related genera, and to objectively defi ne the generic limits of the genus.

Pseudohemihyalea ambigua species-group
Diagnosis.Members of the ambigua species-group can be immediately recognized by the simple, striate forewing pattern (Figs.1-5).Internally, the male uncus in dorsal profi le is characteristically shaped like a bicycle saddle , with the apex slightly down-turned, pointed and beak-like (Fig. 6).Th e posterior portion of the uncus consists of two heavily setose, globose or slightly fl attened lobes (Fig. 6).Th e male valve is relatively simple and bipartite (tripartite in most Amastus and a few Pseudohemihyalea), with the apical 1/3 to ¼ divided into costal and saccullar processes (fi gs.6-10).Th e transtilla is low, hump-like and fi nely scobinate (coarsely spinose or scobinate in most other Pseudohemihyalea, long, prong-like and variously scobinate in Amastus).Female genitalia are relatively simple across the whole Pseudohemihylaea-Amastus group, and no characters were found that distinguish the ambigua-group from other Pseudohemihyalea, although the shape of the lamella antevaginalis may prove useful in a more in-depth review of the group.Th e pine-feeding habits of the larvae (P.ambigua) are unique within the genus.Vincent, pers. comm.).Given the distinctive original description and absence of species similar to P. ambigua in the USA, bolteri is retained as a junior synonym of ambigua.

Key to species of the
Diagnosis.Pseudohemihyalea ambigua is very similar to P. syracosia externally, but the two can usually be separated without dissection by the slightly larger size, broader forewing striae (Figs. 1 and 2) and more northerly distribution (Fig. 25) of P. ambigua.Internally, both the saccular process (cf.Figs. 6 and 7) and uncus (cf.Figs.11 and 12) are shorter and wider compared to P. syracosia (mean length to width ratio of uncus 1.9 in P. ambigua, 2.4 in P. syracosia).Th e coecum of the aedeagus is longer and more conical in P. ambigua (cf. Figs. 16 and 17).In females, the antevaginal plate is less fl ared laterally with a shallower distal indentation compared to P. syracosia (cf.Figs.21 and 23).
Biology and distribution.Collection dates indicate the peak fl ight is in July, with extreme dates ranging from mid June to early August, presumably representing a single annual brood.Pseudohemihyalea ambigua is the most widespread of the ambiguagroup, occurring from southern Wyoming (Ferguson et al. 2000) to Durango, Mexico (Fig. 25).Ferguson et al. (2000) show P. ambigua as occurring in Tamaulipas, Mexico, but no specimens from Tamaulipas could be located and their record may refer to another species, possibly P. potosi.
Diagnosis.Pseudohemihyalea syracosia is most similar to and was long confused with P. ambigua.Th ey can usually be distinguished without dissection, since P. syracosia is slightly smaller and less robust with narrower rust-brown forewing striae on average and less rust brown overall.Th is species also has a more southern distribution than P. ambigua.Sexual size dimorphism is less pronounced in P. syracosia than in Remarks.Pseudohemihyalea syracosia has long been treated as a subjective synonym of P. ambigua, but the characters given above in the diagnosis and species key show it is distinct from P. ambigua, with a more southerly and apparently allopatric distribution.Th e diagnosis and illustrations of "P.ambigua" given by Toulgoët (1997) apply to P. syraciosa.
Biology and distribution.Th e immature stages are unknown.Pseudohemihyalea sonorosa fl ies in upper elevation pine-oak forests of the Sierra Madre Occidental, Mexico during September.It is known only from the type locality, in the state of Sonora.
American Museum of Natural History, New York, New York, USA.
CMNH Carnegie Museum of Natural History, Pittsburgh, Pennsylvania, USA.DEB Personal collection of Don E. Bowman, Pueblo West, Colorado, USA.FMNH Field Museum of Natural History, Chicago, Illinois, USA.MNHN Muséum National d'Histoire Naturelle, Paris, France.BMNH Natural History Museum, London, UK.USNM National Museum of Natural History (formerly United States National