Biology and Distribution.

Female Eupsilia lay eggs singly on twigs and buds in the early spring. Early instars of Eupsilia are believed to feed on a wide taxonomic range of woody plants in the spring, fashioning loose leaf shelters (Fig. 24) or entering those made by other Lepidoptera (Wagner et al. 2011). By the fourth or fifth instar Eupsilia abandon their shelters. Where they go is a mystery. In two decades of caterpillar collecting in Connecticut, we have never found a late instar on foliage during the day. Their dark coloration is suggestive that the larvae feed at night. It is our suspicion that the late instars are broadly polyphagous on woody and non-woody plants, and perhaps other organic matter on or near the ground (Wagner et al. 2011). We once found a last instar Eupsilia apparently feeding on dog food in a basement bulkhead. It is unclear if larvae ascend trees at night to feed, as is commonly the case with many climbing cutworms, e. g., Metaxaglaea Franclemont and related genera. The European Eupsilia transversa is recorded from forbs (e.g., dandelion), and a wide range of woody plants; additionally, it is reported to feed on the caterpillars of other Lepidoptera (Manley 2008). Eupsilia larvae mature in late spring, and then burrow underground and form a cell where they enter diapause for two to three months, before pupating (Wagner et al. 2011). Adults begin issuing in September in Connecticut.

Head capsule measurements for E. schweitzeri indicate that there are six larval instars (Fig. 31): 1st: = 0.3 mm (n=6, 0.3 mm); 2nd: = 0.5 mm (n=10, 0.4–0.5 mm); 3rd: = 0.8 mm (n=7, 0.8–0.9 mm); 4th: = 1.4 mm (n=8, 1.2–1.5 mm); 5th: = 2.2 mm (n=8, 2.2 mm); 6th: = 3.2 mm (n=16, 3.1–3.3 mm). The head capsule width ratios between successive instars were generally consistent with those observed by Dyar, which ranged from 1.3–1.7 (Dyar 1890, Berg and Merritt 2009). Observed ratios in E. schweitzeri ranged from 1.4 between the last two instars and 1.8 between the 3rd and 4th instars, with a mean, median, and mode of 1.6 across the six instars.

Ex ova lots have been reared on Carya glabra, Juglans cinerea, Malus sylvestris, Prunus serotina, and Quercus ilicifolia. Our wild larvae of E. schweitzeri, have been collected as middle instars in southern New England from Carpinus caroliniana, Cirsium sp., Hamamelis virginiana, Nyssa sylvatica, Prunus serotina, Quercus spp., and Vaccinium spp. We suspect that E. schweitzeri is a generalist on woody shrubs and trees in early and middle instars, but becomes even more generalized in diet breadth in late instars, consuming a wide variety of plants (e.g., Cirsium) and perhaps even other types of organic matter.

Because all Eupsilia (across both the Nearctic and Palearctic) are similar in size, we think it likely that head width measurements of any wild-collected Eupsilia larva would generally conform to those given in Figure 31. We would be interested to learn of wild larval collections of 4th, 5th, and 6th instars, and upon what substrates they were observed feeding.

Although all seven northeastern Eupsilia may cooccur (e.g., in southern Connecticut), E. sidus and E. schweitzeri are largely allopatric over much of their range. E. sidus is usually a denizen of “dry oak forests, savannas, woodlands, ridgetops, and barrens from Wisconsin, southern Ontario, and central New Hampshire south to northern Georgia and Texas” (Wagner et al. 2011) (this range may conflate records of E. schweitzeri, e.g., records of E. sidus from Ontario may in fact refer to the heretofore unrecognized E. schweitzeri). E. schweitzeri is more ecologically generalized. It is abundant in both mesic and dry oak woodlands, but also occurs in bottomlands, along the edges of wetlands, in upland woodlands and forests, ridgetops, and other forested communities. In most habitats it is only outnumbered by E. vinulenta and E. morrisoni.

Wing Scales. Eupsilia have derived forewing scales. We confirm the presence of three scale types: lamellar, excelsior, and hooked (Figs 17–22). Lamellar scales, common to Lepidoptera (Scoble 1992, Kristensen & Simonsen 2003), predominate in E. devia, E. morrisoni, and E. vinulenta (Fig. 22) and in the former species, virtually all scales are of the lamellar type. In addition, the lamellar scales of E. devia are wider than those found in the six other northeastern species. Excelsior scales are those where the outer teeth of forewing scales tend to be elongated and recurved; in many cases, scale edges are rolled upward to the extent that they almost enclose the dorsal scale surface. Forbes (1954) used the term “shredded wheat” effect to describe the wing surface appearance resulting from the overlap and interlocking of the modified scales.

Figs. 23–30.

Eupsilia larvae. 23, E. schweitzeri, early instar. 24, E. schweitzeri middle instar making shelter. 25, E. schweitzeri, 5th instar. 26, E. schweitzeri, 6th instar. 27, E. cirripalea, late instar venter. 28, E. schweitzeri, late instar venter. 29, E. sidus, late instar venter. 30, E. tristigmata, late instar venter.

Excelsior scales occur in E. cirripalea, E. morrisoni, E. schweitzeri, E. sidus, and E. tristigmata, (Figs. 17–21), and in lower frequency and with an intermediate degree of recurvature of the outer teeth in E. colorado and E. vinulenta. E. tristigmata best represents the scales we describe as hooked; i.e., just the outer (two) teeth are recurved, sometimes as much as 180º (Fig. 21). The outer teeth also tend to be much darker in coloration in E. tristigmata.

Eupsilia walkeri (Grote, 1864). With the possibility that Eupsilia walkeri might refer to our new species, we compared our type series to Grote's description, which he based on a single specimen from Philadelphia. The type of E. walkeri is believed to be lost (Franclemont 1952). Grote's description of E. walkeri makes no mention of purple or related violaceous colors (yet he used purplish twice as a descriptor in his description of E. vinulenta, which appeared in the same publication, immediately following the account for E. walkeri). Violaceous scaling on the forewings is among the most reliable features for the recognition of E. schweitzeri adults, and is usually present basad to the antemedial line. Grote's characterization of the terminal line in E. walkeri as “very clearly defined, semi-lunulate,” further eliminates E. schweitzeri as a possibility (as well as E. cirripalea or E. sidus and others, a point we return to below). Grote uses ochrecous to describe the forewing color of E. walkeri and “rich dull red” to describe the forewing of E. vinulenta. Ironically, the only Eupsilia in the Northeast that is accurately characterized as ochraceous are the pale (overwintered) forms of E. vinulenta (Fig. 9), the most phenotypically variable Nearctic Eupsilia. Both E. schweitzeri and E. sidus are a richer red than Grote's rendering of E. walkeri.

In his 1952 treatment of Eupsilia, Franclemont regarded E. walkeri (Grote) to be a synonym of E. sidus:

“ As identified here, walkeri Grote is a synonym, and judging from the figure and description it was based on a worn hibernated specimen. Grote's type is also lost, but was probably also from the vicinity of Philadelphia since it was in the collection of the Entomological Society of Philadelphia.”

Forbes (1954) and subsequent workers have followed Franclemont. However, based on Grote's original description, remarks made above, and other arguments, we disagree with Franclemont's decision, and treat E. walkeri as synonym of E. vinulenta (Grote). As noted above, Grote states that E. walkeri possessed a clearly defined, semi-lunulate terminal line. Such is absent not only from E. schweitzeri but also E. cirripalea and E. sidus, and is commonly present in E. vinulenta. Grote's description of E. walkeri states: “transverse anterior [antemedial] line distinct, darker shaded, straight, composed of two narrow lines with a central light shade.” Yet in E. sidus the antemedial line is poorly differentiated, especially relative to that of E. vinulenta. In addition to the character evidence, we note that E. vinulenta is among the most common Eupsilia in eastern Pennsylvania. E. sidus is an oak barrens species, at least at the latitude of Philadelphia, that we associate with scrub oak and low-bush blueberry communities, a habitat absent from the vicinity of present-day Philadelphia. In sum, we are convinced that Grote based his description of E. walkeri on a flown (perhaps overwintered) individual of E. vinulenta, and could not have been either E. schweitzeri or E. sidus.

Scopelsoma colorado (Smith, 1903), rev. stat. In his 1952 treatment of Eupsilia, Franclemont synonymized Scopelsoma colorado with E. sidus. The species is known only from the type series evidently collected by William H. Barnes in March and April, 1893–1895, in Glenwood Springs, Colorado. We were able to locate and examine 1 (of the 2 formerly present) specimens in the AMNH, including the lectotype male and its associated genitalic slide; 13 specimens in the USNM; and 2 specimens in the CNC. After examining the above, as well as three additional male genitalic dissections (Franclemont #7574, USNM 38903, USNM 118685) and three female dissections (CNC 16801, USNM 38901 [bearing the notation: (From spem. labeled “♂ cotype”) [sic]], USNM 38902), we conclude that Scopelsoma colorado is a valid species. The ground color ranges from tan or ochraceous (most) to reddish brown (few); violet scales are absent and any hint of red or orange typical of E. sidus is subdued; the antemedial line in E. sidus and to a lesser extent E. schweitzeri is nearly absent in E. colorado (Fig. 4). Most notably the reniform and associated satellite spots common to E. schweitzeri, E. sidus, and E. vinulenta are vague and illdefined in E. colorado save for the dark scales along the lower edge of the reniform spot (Figs 1–9). The forewing margin is less crenulate than that of E. schweitzeri or E. cirripalea; and especially that of E. vinulenta. The forewing scales are intermediate between lamellar and excelsior, and most similar in character and distribution to E. vinulenta. They lack the “shredded wheat” aspect that we (and Forbes) would attribute to E. schweitzeri, E. sidus, and E. cirripalea. Forewing scales in E. colorado distad to the postmedial line are lamellar, whereas great numbers of excelsior scales are present in this area in E. cirripalea, E. schweitzeri, and E. sidus. The male genitalia of E. colorado, based on dissections from Franclemont (#7574), the USNM material (#38,903), and J. Donald Lafontaine (#118685), are close to those of both E. sidus and E. schweitzeri, but with important differences. The coronal spines are stouter and fewer in number (14–17) than in other North American Eupsilia (Figs 10, 12, 14). The setal rosette at midlength on the vesica (Fig. 13) is like that of E. schweitzeri (Fig. 15) with 25–30 fine spines, and but differs from that of E. sidus (Fig. 11), whose vesica bears a run of fewer (ca. 20) stouter setae, and these are drawn out along the vesica, rather than pulled into a single rosette. Nor does E. colorado have the six small basal pouches, arranged in two groups, as seen in E. sidus and E. schweitzeri. In our preparations, the most basal set of pouches are much enlarged and give the basal portion of the vesica considerable girth. We note that the lectotype preparation from the AMNH, made by Smith, is badly damaged and distorted.

Fig. 31.

Head capsule width of Eupsilia schweitzeri larvae. Because all Eupsilia (across both the Nearctic and Palearctic) are similar in size, we think it likely that head width measurements of other wild-collected Eupsilia larva would fall into these same size classes.

Eupsilia sidus (Guenée, 1852). The type of E. sidus is missing (Franclemont 1952). Given the difficulty of adult identifications, even with dissection, and the taxonomic confusion within the genus, to promote nomenclatural stability we anchor the concept of the name as understood by Franclemont (1952), Forbes (1954), current on-line resources, and institutional collections, by hereby designating a neotype of E. sidus from New Jersey, the presumed type locality. (According to Franclemont (1952) “From what we know about the source of Boisduval's material described by Guenée, it seems safe to assume that the type of sidus was collected by Say in New Jersey, given to LeConte and sent by LeConte to Boisduval.”) The identity of the neotype (Fig. 5) was confirmed by larval characters, the male was dissected, and a leg was submitted for DNA barcoding. Data on the neotype labels are as follows: New Jersey, Cumberland Co., Millville, Poplar Drive f coll. at bait III.11.2010, Dale F. Schweitzer // bred ex ova on Prunus serotina at Windham, CT, by Eric L. Quinter, emerge[d]: IX.26-X.15.2010 // Eupsila sidus (Guenée), Det. Eric L. Quinter // CNC # ♂ 17072 // Specimen ID CNCLEP 00116372 // Barcode of Life Project // Leg removed, DNA extracted // Photographed J. D. Gill. The neotype has been deposited at the CNC, Ottawa, Ontario, Canada.

Eupsilia is a difficult taxon given the uniformity of the genitalia across species. Likely the new species would have gone undiscovered until the advent of barcoding had not the larval characters revealed that the name E. sidus was being applied to two different species. Given the conservative nature of phenotypic evolution in adult Eupsilia, it would not be surprising to learn that still other cryptic species of Eupsilia remain to be recognized and described from North America. Barcodes are suggestive that E. tristigmata represents more than a single entity and perhaps as many as three (Don Lafontaine pers comm.).

Eupsilia colorado has not been collected since 1895—and awaits rediscovery. Efforts to barcode specimens were unsuccessful. The type locality, Glenwood Springs, is noteworthy for its exceptional complex of moths that occur sparingly or nowhere else on the west slope of the Rockies (Don Lafontaine in litt, Terhune Dickel pers. comm.). Other notable lepidopterans known from the area include Acronicta exempta Dyar, 1922, Heterocampa rufinans (Dyar, 1921), Hyparpax venus Neumoegen, 1892, Orthosia flaviannula (Smith, 1899), and Phoberia ingenua (Walker, 1858). What is special about the conditions in the vicinity of Glenwood Springs remains a mystery; perhaps they relate to a unique paleoclimatological history, specialized edaphic factors, or unique aspects of its flora. Many of the above are oak feeders, and Quercus is a known host for many eastern Eupsilia (Wagner et al. 2011). Eupsilia colorado flies in March and April, when few moth collectors are afoot in the Rockies and weather conditions are especially challenging for nocturnal collecting. Larval sampling, especially on oak in the spring, might prove to be the easiest means to survey for the moth.