Larva and pupa of Amyna axis ( Guenée , 1852 ) and affirmation of its taxonomic placement in Bagisarinae ( Lepidoptera , Noctuidae )

Th e larva and pupa of Amyna axis (Guenée, 1852) are described and illustrated, and observations are provided on the insect’s life history and larval biology. Larval, adult, and life history characters support the transfer of Amyna Guenée from Acontiinae Guenée, 1841 to Bagisarinae Crumb, 1956. Th e phylogenetic placement of the Bagisarinae is enigmatic; some adult and larval features indicate that the subfamily is a basal trifi d proximate to Acontiinae, whereas other larval and life history characters suggest an association with Scoliopteryginae, a basal quadrifi d group. Larvae exhibit a green-to-black color polyphenism presumably linked to larval density, with darker phenotypes occurring during outbreak densities. Parallel color shifts that occur in other Lepidoptera are briefl y discussed.


Introduction
Th e genus Amyna Guenée, 1852, contains about 35 recognized species distributed worldwide (Savela 2009). Th e majority of the species occur in Asia. Only three are known from the New World; the pan-tropical Amyna axis Guenée, 1852[= A. octo (Guenée, 1852], A. bullula (Grote, 1873), and A. amplifi cans (Walker, 1858). Forbes (1954) placed Amyna in his concept of the Acontiinae, but noted that it was "A curious genus, probably not really Acontiine though with proper M of hindwing.…" He also mentioned the presence of a blunt appendiculate tooth on the larval crochets ( Fig. 7) that suggested the genus was related to "the Malvaceous group" of his Acontiine [Bagisara presumably]. Franclemont and Todd (1983) left Amyna in the Acontiinae, grouping it in their concept of the Eustrotiini, a heterogeneous assemblage of taxa that includes genera now classifi ed in the Acontiinae, Acronictinae, Condicinae, Eubleminae, Plusiinae, Xyleninae, etc. Rawlins 1998, Fibiger andLafontaine 2005, Lafontaine andSchmidt (in press). Poole (1989) left Amyna in the Acontiinae. Kitching and Rawlins (1998) mentioned Amyna in their synopses of both Eustrotiinae and Bagisarinae, noting that the studies of Jeremy Holloway indicated the genus was better placed in the latter subfamily. Holloway (2009) tentatively moved the genus into Bagisarinae in his latest volume on the Moths of Borneo.
Amyna axis, the type-species for Guenée's genus, is the focus of this paper and the basis for the placement of the genus in Bagisarinae. Below we describe the larva and pupa, extending the eff orts of Gardner (1941Gardner ( , 1946 and provide information on Amyna's larval biology. We discuss morphological, behavioral, and biological data that support the taxonomic assignment of Amyna in Bagisarinae, and conclude the paper with a discussion of the green-to-black color polyphenism that occurs in Amyna and other Lepidoptera.

Materials examined and methods
Larvae of Amyna axis were obtained from Amaranthus L. in Scissors, Hidalgo County Texas in October 2006 (DLW) and October 2009 (Berry Nall), and Croton L. in Starr County, Texas in October 2008 (DLW and Henning von Schmeling).
One larva from Scissors was prepared for SEM study by running it through a series of ethanol baths (70%, 80%, 90%, 95%, 100%) before it was dehydrated with hexamethyldisilazane. Th e caterpillar was then coated with gold palladium for three minutes in a Polaron E 5100 sputter coater. Images were obtained with a Zeiss DSM-982 Gemini FE SEM at 3 kV. Larval, pupal, adult, fi lm, and digital vouchers have been deposited at the University of Connecticut. 12). Elongate, approximately 10 × longer than wide, shallowly constricted between segments, especially rearward. Ground color glaucous green, sea green, or emerald green to nearly black, often with black supraspiracular spots. Th in, white, broken addorsal, subdorsal, and supraspiracular stripes run length of body. Spiracular stripe continuous and roughly 2 × thickness of others, weakening rearward of A8. Caudal segments somewhat fl attened with anal prolegs elongated, splayed outward and held nearly horizontal behind body. Head shiny, pale green to orange, with setae borne from dark spots.
Proboscis, meso-and metatarsi, and antennae all ending together between wing tips. Antennal segmentation vague. Small black pimplelike spot anterodorsal to spiracle on A5-A7. Cremaster consisting of single pair of spines, each shallowly hooked and slightly sinuate at apex.

Life history
Host records include Amaranthus L., Celosia L., and Digera Forssk.  (Ferguson et al. 1991, Robinson et al. 2001, 2002. Our southern Texas collections came from Amaranthus and Croton L., but not from any of the dozen or more other forbs and low-growing plants (e.g., Abutilon Mill., Ambrosia L., Eupatorium L., Helianthus L., Heterotheca Cass., Lantana L., Malvastrum A. Gray, Melolochia L., Parthenium L., Physalis L., Sida L., etc.) sampled with a beating sheet over the course of general (larval) collecting eff orts in October 2006October , 2007October , and 2008. Nearly all of our Amyna caterpillars were collected from mature amaranth plants, especially those that were into seed-set, turning red, and senescing. None came from lush, actively growing pre-fl owering plants. Amyna caterpillars were so abundant in stands of amaranthus at Scissors, Texas in October 2006 that they had conspicuously damaged the plants.
Larvae perch on the undersides of leaves and along stems and petioles. At rest, the abdominal segments are often looped upward (Fig. 9). When alarmed the larva essentially jumps from the host and continues to wreathe and wriggle wildly. A QuickTime® movie of the alarm response of Amyna is mounted with this paper on http://www.eeb. uconn.edu/people/wagner/.
Prepupal larvae take on a pinkish cast (Fig. 14). Pupation occurs in a silken cocoon below (usually) or at the soil surface; sand and/or plant debris are interwoven into the cocoon wall. Adults occur throughout the growing season in southern Texas, but numbers peak in September and October during the wet season (Ed Knudson, personal communication).
At low densities caterpillars are pale green (n > 20). Where we encountered caterpillars in high density and in our rearing containers, we saw higher frequencies of melanistic phenotypes. Th e degree of blackening varied widely, ranging from individuals with simple subdorsal-lateral patches (Fig. 9) to those that were mostly black (Fig. 10). In the fi eld, melanized individuals (common in 2006) were visible from many meters away.

Discussion
Taxonomic matters. Crumb (1956) was so perplexed by the taxonomic affi nities of Bagisara that he placed the genus into its own subfamily in his monograph, noting that "All of the principal characters of the larvae are primitive but the moths are rather stout and stubby winged, showing none of the superfi cial characters students of the moths have come to associate with lower Phalaenidae [Quadrifi nae]." Kitching and Rawlins (1998) concurred with Crumb's assessment and recognized Bagisarinae as a distinct subfamily of trifi ne noctuids. Th e subfamily's position remains essentially unchanged in Fibiger and Lafontaine (2005), Lafontaine and Fibiger (2006), and Lafontaine and Schmidt (in press), i.e., at the base of the trifi nes in the vicinity Plusiinae, Eustrotiinae, and Acontiinae.
Character evidence that argues for a close phylogenetic association between Amyna and Bagisara include the (1) elongate body, with shallow incisures between segments and tapered caudal segments (Figs 8-13); (2) prolegs absent or greatly reduced on A3 and A4; (3) at least some crochets with a subapical tooth (Fig. 7); (4) spiracle on A1 1.2 × that on A2 and that on A8 2 × height of spiracle on A7, yellowish; (5) spiracles on A2-A4 shifted dorsad; (6) L3 shifted anteriorad below L2 on A7; (7) setae long, especially those on T1 and A8-A10 (Figs 12, 13); (8) SD1 on A10 elongate and thickened; (9) long labial palpus, >3 × basal diameter; and (10) pink to reddish prepupae (Fig. 14). Behavioral characters suggesting that Amyna is a bagisarine include (11) a resting posture with the mid-abdomen looped above the substrate (Figs 9, 11); (12) an animated and prolonged alarm response (QuickTime® movies of the alarm response of both genera are mounted on http://www.eeb.uconn.edu/people/wagner/); and (13) the fl icking of the excreta. Th e appendiculate crochets and shared spiracular "formula" (i.e., with the height of the spiracle on A1 being 1.2 × than that of those on A2-A7, that on A8 being >2 × that on A7, as well as the dorsal shift of the spiracles on segments A2-A4), strike us as compelling evidence of (recent) common ancestry, and support Holloway's (2009) suggestion that Amyna should be transferred to Bagisarinae. Conversely, numerous structural diff erences separate the two genera. Bagisara (and some Xanthodes Guenée) are unusual among noctuids in their possession of two SV setae on A7 (Crumb 1956, Kitching andRawlins 1998), whereas Amyna axis has only a single SV seta on A7. Th e prolegs are represented on A3 and A4 in Amyna axis by minute nibs; there is no evidence of the prolegs on these segments in the Bagisara that we examined. Th e basal tooth on the crochet of Amyna is much less developed than that seen in Bagisara species, although some Indian members of the genus evidently have a well-developed subapical tooth (Gardener 1946). Bagisara possess short, stout, upcurved D2 setae on A10 (Fig. 13) that presumably play an important role in fecal fl icking (see Weiss 2006); the D2 setae are straight and much longer in Amyna. Bagisara are dietary specialists with much of their species radiation tied to the Malvaceae, whereas Amyna axis has been reared from a wide array of forbs.
Like Crumb (1956), we confess to being confused about the phylogenetic affi nities of the Bagisarinae. Bagisara was long classifi ed in the Acontiinae (e.g., Forbes 1954, Franclemont andTodd 1983) and most recently in its own subfamily proximate to the Acontiinae and Eustrotinae (e.g., Kitching and Rawlins 1998, Fibiger and Lafontaine 2005, Lafontaine and Fibiger 2006, Holloway 2009). In form and habit, the larvae of Bagisara (with Amyna) share similarities with Acontiinae (such as Tarachidia Hampson): e.g., trifi d venation in adults, frass fl icking habit, absence of prolegs on A3 and A4, and a simple cremaster. But other features, especially in the larvae, suggest linkages with erebids and especially the genus Anomis Hübner (Scoliopteryginae): in both the SV group is trisetose on A1 and A2 and the crochets are appendiculate. In addition, Anomis and Bagisara are Malvaceae specialists. Still other traits are shared with the Hypeninae and Scoliopteryginae taken together: all have some reduction of the prolegs on A3 and A4, fl ick their frass, have a sustained thrashing response when alarmed, and a spiracular stripe that continues on to the anal plate.
Melanism matters. Amyna axis larvae are green at low densities in the wild, but at high larval densities -several per sq. m -we encountered larvae with varying amounts of dark pigmentation (Fig. 10). We have also induced blackening in our rearing containers; green, unmarked larvae reared together usually darken within days of their collection. Homologous green-to-black polymorphism occurs in more than two dozen other lepidopteran lineages (Wagner in prep.). In a fashion similar to that seen in Amyna, caterpillars of the elm spanworm, Ennomos subsignarius (Hübner), darken as a result of physical contact with other larvae (Drooz 1966). Green-to-black color changes have also been linked to crowding in other noctuids: e.g., Plusia gamma L. (Long 1953) and Mamestra brassicae (L.) (Goulson 1994). In addition to crowding, temperature can induce color polyphenisms in caterpillars. Arizona populations of Battus philenor (L.) raised in high temperatures display a red (cooler core) larval phenotype, whereas those raised at lower temperatures displayed a black (warmer core) phenotype, suggesting that color changes could be a way of "beating the heat" (Nice and Fordyce 2005). Th e eastern black swallowtail, Papilio polyxenes (Stoll), has a dark, coolseason phenotype that has thermoregulatory advantages over the lighter mid-summer morphs (Hazel 2002). Long (1953) reported that the degree of blackening in Plusia gamma was also infl uenced by a caterpillar's diet. In Amyna, the production of dark forms appears to be linked to crowding as such were observed at high fi eld densities and were induce in rearing containers with multiple larvae. Whether or not the greento-black shift in Amyna and other caterpillars represents an adaptive response and the selective consequences of a shift to darker coloration change remain to be elucidated.