The accepted normal range of the funereal duskywing skipper, Erynnis funeralis (Hesperiidae), is generally considered to be the southwestern United States, Mexico, Central America, and western South America from Colombia to northern Argentina and Chile. According to various records, however, adult funeralis have been observed in a variety of locations throughout the eastern half of central North America, sometimes thousands of kilometers outside its accepted range. These individuals are usually conceptualized as non-reproductive strays, but different lines of evidence (e.g. the observations of both males and females, the fresh condition of most adults, the recurrence of observations in the same locations at similar times of year in different years) suggest that funeralis regularly establishes seasonal breeding populations in the eastern half of central North America. To test this hypothesis, I examined reports of observations of funeralis throughout this region for evidence of 1) regular as opposed to random presence in the East, 2) eastern as opposed to western eclosion, 3) regular as opposed to random expansion throughout the East, and 4) reproductive activity. The results revealed evidence of all four phenomena. Consequently, I concluded that instead of being conceptualized as an irregular, non-reproducing stray, funeralis should be conceptualized as a regular seasonal immigrant that establishes temporary breeding populations in the eastern half of central North America during the warmer months of the year.
The accepted normal range of the funereal duskywing skipper, Erynnis funeralis (Scud. & Burg.) (Hesperiidae), is generally considered to be the southwestern United States, Mexico, Central America, and western South America from Colombia to northern Argentina and Chile (Burns 1964 pp. 174–175). In the southwestern United States, its accepted range is considered to be southern California, Arizona, southern New Mexico, and Texas. Individuals found somewhat outside this range (i.e. southern Nevada, Utah, and Colorado in the West; Oklahoma, Nebraska, and Kansas in the Plains; and western Louisiana and Arkansas in the South) are considered to be uncommon, but still within funeralis's accepted range.
According to various records, however, funeralis has been observed in a variety of locations throughout the eastern half of central North America, sometimes thousands of kilometers outside its accepted range. The record for such far-flung sightings appears to be one from Avondale Farm Preserve, Waverly (Washington Co.), Rhode Island (Koehn 2011 p. 145), some 3000 km from the focal point of funeralis' s typical U.S. range in the Southwest. Although observations of funeralis in the eastern half of central North America are uncommon, they are sufficiently numerous to provoke interest regarding their significance.
In the literature, funeralis is persistently described as a “stray” when discovered outside its accepted range (e.g. Hall 1936, Scott 1986 p. 492, Heitzman & Heitzman 1987 p. 32, Shull 1987 p. 41, Tuttle 2004 p. 49, Le Grand & Howard 2010, Koehn 2011 p. 145, Belth 2013 p. 82, Neese 2013). This characterization implies that its occurrence in the East is not only irregular, but also non-reproductive in nature. Exceptions to this perception include Bouseman et al. (2006 p. 56), who acknowledge that funeralis may occasionally establish temporary breeding population as far east as Illinois, but overwhelmingly, the presence of funeralis in the eastern half of central North America is viewed as an anomaly. The possibility that its occurrence there may be something more than accidental is generally not considered. Various lines of evidence, however, suggest that funeralis regularly establishes seasonal breeding populations in the East.
First, the data indicate that both male and female adults have been observed (e.g. Shull 1987 p. 41, Hanson et al. 2002, Bouseman et al. 2006 p. 57, Reese 2007, Arvin 2010, Bolin 2010, Brees 2010a, Brown 2010, Ennis 2010, Huf 2010, Legier et al. 2010, LeGrand & Howard 2010, Mann 2010, 2012, Rehm 2010a,b, Schulte 2010,“Stone Lark” 2010, Trently 2010a, Williams 2010, Yomeyer 2010, Cavasin 2012, Davis 2012, Geiser 2012, Lubahn 2012, Shively 2012, Spitzer 2012, Lotts & Naberhaus 2013a,c; Fig. 1). This suggests that individuals of one sex are not merely wandering outside their accepted range because of some sex-specific behavior. Instead, the appearance of both sexes suggests that funeralis's presence outside its accepted range is related to the behavior of the species in general. Both male and female adults are obviously required to establish a breeding population, so the presence of both sexes by itself suggests this possibility.
Second, the majority of adult funeralis collected, photographed, or otherwise reported in the East have been found in remarkably fresh condition (see previous paragraph for examples; in addition, see Ziebol & Homeyer 2007, Koehn 2011 p. 101, Lamond 2011, and Wormington 2012 for comments). Indeed, some observers have not only remarked about their immaculate state, but also marveled that they could remain in “mint” condition after having flown what are presumed to be great distances (Lamond 2011). The condition of these adults, however, suggests that they had recently eclosed and, consequently, would not have had the time required to travel hundreds, let alone thousands, of kilometers to the eastern locations where they were sighted. It is more logical to assume that most if not all of these adults were, in fact, the products of breeding populations that had become established much closer to where they were observed.
Fig. 1.
Erynnis funeralis in west central Indiana. Female (a) dorsal view, (b) ventral view, Terre Haute (Vigo Co.), Indiana, 15 September 2003, 1647 h., taking nectar from Perilla frutescens (L.) Britt. (Lamiaceae) at the authors home. This individual displayed little if any wing wear, suggesting it had recently eclosed and, therefore, came from a local breeding population. The majority of funeralis encountered in the eastern half of central North America have been found in similar condition. Species and sex identified by John M. Burns. Photos copyright © Roger Carpenter.
Third, various reports from the East indicate that not one but two or more adults have been observed in the same place at the same time. Examples of such sightings include one from Shaw Nature Reserve (Franklin Co.), MO on 10 July 2010 in which two adults were observed flying together (Homeyer 2010); one from the vicinity of Tamms (Alexander Co.), IL in which three males were collected between 18 and 20 September 1999 (Wiker, pers. com.); and one from Point Pelee National Park (Essex Co.), ON on 13 September 2010 in which three adults were observed flying together (Koehn 2011 p. 101). Because funeralis is not known to migrate en masse, the occurrence of two or more funeralis together so far outside its accepted range is consistent less with the idea that (presumably isolated) individuals have strayed outside their normal range than with the idea that groups of adults are eclosing in the same place at the same time.
Fourth, in similar fashion, various reports indicate that funeralis tends to be observed in the same locations at similar times of year in different years. For example, funeralis has been observed at Point Pelee National Park during the late summer-early fall of 1990, 1992, 2010, 2011, and 2012 (Layberry et al. 1998 p. 40, Koehn 2011 p. 101, 2012 p. 82, Cavasin 2012, Mann 2012, Wormington 2012). Other examples include Tippecanoe Co., IN during early/mid-Angust 2010 and 2012 (Arvin 2010, Belth pers. com.) and Dane Co., WI in mid-date August 2007 and 2010 (Reese 2007, Legier et al. 2010). This type of pattern is consistent less with random straying than with regular expansion that causes funeralis to arrive at specific eastern locations at similar times of year.
The purpose of the present study, then, was to determine whether funeralis is merely straying, in random fashion, outside its accepted range or regularly expanding its range and establishing breeding populations when conditions become favorable. I hypothesized that if the latter were true, then the funeralis observations from the East should display the following characteristics:
Observations should occur in many if not most years, suggesting regular as opposed to random presence in the East;
Observations should indicate that, in general, adults had eclosed relatively close to where they were observed, as suggested by a) excellent wing condition in many adults and b) observations of two or more adults in the same location in the same year in a species that does not appear to swarm;
Observations should indicate regular expansion throughout the eastern half of central North America during the warmer months of the year, as suggested by a) observations of adults in similar locations at similar times of year in different years, b) in general, a regular as opposed to random pattern of movement from west to east, c) a relatively constant rate of expansion from month to month, and d) similarities in the pattern of movement from year to year;
Observations should indicate the occurrence of breeding activity, as suggested by a) an overall sex ratio of 1:1, b) evidence of different stages of the life cycle, and c) regularly spaced increases in the number of adult observations, consistent with the production of successive generations.
Fig. 2.
Idealized range expansion pattern, based on data from all years, of E. funeralis throughout the eastern half of central North America from late spring to early fall. Polygons enclose outermost first observations of the year for specific counties reported through the end of five successive months: 5 = May; 6 = June; 7 = July; 8 = August; 9 = September. (New Hanover Co., NC and Escambia Co., FL observations occurred in October.)
Methods and Materials
I analyzed observations of funeralis from the eastern half of North America derived from the following sources: Hall (1936), the Lepidopterists' Society Season Summary reports (1960–2013), Burns (1964), Shull (1987), Layberry et al. (1998), Belth (2013); the websites Bug Guide (VanDyk 2013), Butterflies and Moths of North America (Lotts & Naberhaus 2013a; Opler et al. 2013); Butterflies of North Carolina (LeGrand & Howard 2010), Butterflies of Ontario (Cavasin 2013), Illinois Butterfly Monitoring Network (2013), North American Butterfly Association (2013), North American Moth Photographers Group (Patterson 2013), and Wisconsin Butterflies (Reese 2013); the blogs ‘Burg Birder (Mann 2013) and Exploring the Remnants (Brees 2010); the online forums Illinois Birders’ Forum (Neise 2013) and Ontario Butterflies (2013); the photo sharing pages of Kevin Arvin (2010), Bruce Bolin (2010), Mark Brown (2010), Allan Trently (2010a), and Urs Geiser (2012); personal communication with Jeffrey Belth, John Burns, Myron Cole, and James Wiker and personal observation.
Because I focused on funeralis's movements outside its accepted range, I excluded data from the following states, based on the funeralis range map given in Burns (1964 Fig. 23): California, Arizona, New Mexico, Texas, and Oklahoma. In addition, I excluded data from states west of a line separating the Plains States from the eastern states as well as from Arkansas and Louisiana to control for the possibility of random movement of individuals within as opposed to outside the accepted funeralis range. The final data set included observations from the states of Alabama (1), Florida (12), Georgia (1), Illinois (24), Indiana (8), Iowa (6), Kentucky (4), Michigan (1), Mississippi (4), Missouri (16), North Carolina (2), Ohio (1), Pennsylvania (2), Rhode Island (1), South Carolina (1), Tennessee (7), West Virginia (1), and Wisconsin (7) as well as the province of Ontario (27) for a total of 126 observations. Data from the remaining eastern states and provinces were unavailable. Different subsets of this primary data set were used in various analyses depending on the type of information available for individual observations.
Observations by Year.
The number of reported observations of funeralis per year was totaled. In addition, the median number of observations per year, the median year of the observations, and the median day of the observations (corrected for multiple observations in a specific county in a specific year) were calculated.
Eclosion near Location Observed.
Wing condition. I evaluated the wing conditions of as many adults as possible as a proxy for adult age and, by inference, distance traveled during the adult stage. I examined photographs of live, unhandled adults in which the dorsal sides of all four wings were clearly visible to determine levels of wing wear. I rated the condition of each wing on a scale ranging from 0 (missing) to 5 (perfect) in 0.1 increments for finer discrimination. The ratings for the four wings were then averaged to yield total wing condition scores both for individuals and for wing type (RFW, RHW, LFW, LHW). One-way ANOVA was performed to determine if wing condition differed by wing type.
Same location/single year observations. I analyzed observations of funeralis in the East for instances of two or more observations in the same location in the same year. “Same location” was operationally defined as the same county or adjacent counties if the observations took place within 50 km of each other (the average length of a county in the United States).
Range Expansion.
Same location/different year observations. I analyzed observations of funeralis in the East for instances of observations in the same location in different years. “Same location” was operationally defined as it was in the same location/single year observation analysis.
General movement. I derived an idealized model of funeralis's expansion throughout the East by plotting the first observation of the year reported for a particular county on a county-level map of the eastern United States then drawing the smallest possible polygon around the outermost observations for successive 1-month periods beginning with all observations through the end of May and ending with those through the end of October. The data were analyzed for outlying values by dividing the distance of the observation in kilometers from the focal point of funeralis's range in the United States by the day of the year of the observation and converting the resulting values to z-scores. Because funeralis's perennial range in the United States is concentrated in the Southwest (see dot map in Burns 1964 Fig. 23), I operationally defined the focal point of its range as Phoenix, AZ (33.500°N, 112.083°W). Outlying values were eliminated from the expansion model.
Expansion rate. I estimated the average rate of funeralis's expansion throughout the eastern half of central North America using the method that Davis and Howard (2005) employed to determine the rate at which Danaus plexippus that overwinter in Mexico recolonize the eastern half of central North America in the spring. After determining the extent of the species' range (in km2) for successive, equally spaced time periods, the net increase in range from one period to the next is calculated. The square root of the increase is then divided by the number of days in the period to estimate the average rate of expansion (in km/d). Periods of one month were used in the present study.
Single-year movement. Because 2010 appeared to be unusually rich in funeralis observations (n = 43), providing an adequate sample for detailed single-year analysis, I performed correlation and regression analyses to measure the relationship between the day of the year on which the 2010 observations occurred and the distance from Phoenix at which they occurred. To be counted in this analysis, an observation had to include 1) the location at which the observation occurred (to the county level or better) and 2) the specific date on which the observation occurred. For counties where two or more observations occurred, I included only the first observation for the year. I estimated the coordinates of each observation using the most accurate information available. This information included 1) specific locations (e.g. an observer's residence), 2) general locations (e.g. a city or a state park), or 3) county-level locations (e.g. Boone Co., IA). In cases 2 and 3, the geographical center of the location was determined and used. Then, using Google Earth (2010), I measured the distance in kilometers (and cardinal direction in degrees) from Phoenix to the location of the observation.
2010 vs. non-2010 comparison. To determine if the 2010 and non-2010 data significantly differed, I compared the mean distance of the observations from Phoenix, the mean cardinal direction of the observations from Phoenix, and the median day of the observations. For the distance and direction comparisons, I used t-tests, while for the median day comparison, I used the Mann-Whitney U-test. (All tests were two-tailed.)
Eastward and northward movement. I calculated the median easternmost longitude and the median northernmost latitude of the observations, respectively. In addition, I calculated the correlations between year and 1) easternmost longitude at which funeralis was observed and 2) northernmost latitude at which funeralis was observed. (All tests were two-tailed.)
Reproductive Behavior.
Sex ratio. I estimated the overall sex ratio of the funeralis that have been observed in the East by ascertaining the sexes of as many photographed and collected individuals as possible. I assumed that an adequate sample of such individuals would be representative of the proportion of males vs. females in the East, since eastern photographers and collectors appear to focus on acquiring images or specimens of the species rather than of a particular sex. The binomial test (two tailed), assuming a 0.5 probability, was used to determine if the proportion of the two sexes differed from the expected 1:1 ratio.
Life cycle observations. I examined the available records for observations of mating, oviposition, eggs, larvae, pupae, and eclosing adults as indicators of breeding activity.
Generational trends. I totaled the observations for successive 15-day periods beginning with the first dated observation and ending with the last to determine the presence or absence of regularly spaced peaks in adult observations suggestive of breeding activity.
Statistical analyses were performed using SPSS v20.0 software (IBM Corp. 2011).
Results
Observations by Year.
Observations reported before 1990 were sporadic with one observation reported in 1963, 1974, and 1979, respectively, two in 1975, and three in 1914. Hall (1936) reported the collection of a funeralis in Crawfordsville (Montgomery Co.), IN sometime in the early 1930s but did not give an exact date. Observations reported after 1990 were more constant with at least one observation reported each year between 1990 and 1992 as well as between 1998 and 2014 (except for 2013). In addition, at least one observation was reported as far east as the Illinois-Indiana border each year between 1998 and 2014 (except for 2013). The median number of observations per year (n = 112) was 1.5 and ranged from 1 to 43. The median year of the observations was 2010 and ranged from 1914 to 2014. The majority of the observations (92.9%) occurred in 1990 or after. The corrected median day of the observations (n = 80) was day 214 (= 2 August, 365-day calendar) and ranged from 20 April to 30 October.
Eclosion near Location Observed.
Wing condition. The mean wing condition rating for all individuals examined (n = 36) was 4.97 (SD = 0.08). Twenty-nine individuals (80.6%) received a rating of 5.0, while the remaining seven (19.4%) received ratings that ranged from 4.58 to 4.98. Mean ratings for each wing were RFW = 4.98 (SD = 0.05), RHW = 4.97 (SD = 0.07), LFW = 4.98 (SD = 0.05), and LHW = 4.92 (SD = 0.25). The lower LHW mean and higher standard deviation resulted primarily from one individual (Lotts & Naberhaus 2013c) that received a LHW score of 3.5 because of a large linear tear extending from the LHW outer margin to the discal area. When this individuals LHW score was removed, the LHW mean/standard deviation (n = 35) became more similar to those for the other three wings (mean = 4.98, SD = 0.05). The difference in wing condition by wing type was not significant, F(3, 105) = 0.72, P = .542.
Same location/single year observations. The results of the same-location/single-year observation analysis are presented in Table 1. Thirteen one-county/two-county units reported two or more observations of funeralis in the same year.
Range Expansion.
Same location/different year observations. The results of the same-location/different-year observation analysis are presented in Table 2. Eleven one-county/two-county units reported at least one observation of funeralis per year in two or more years.
General movement. From an initial subsample of 66 observations, one observation from Clark State Forest (Clark Co.), IN on 20 April 1991 (Belth 2013 p. 82, pers. com.) was eliminated because it was revealed to be an outlier (distance/day, z = 3.96; second highest, z = 2.27). The resulting data set, then, contained 65 observations. Because there was only one remaining observation for April (Benton Co., MO, 27 April 1975, Lotts & Naberhaus 2013b), and this observation occurred late in the month, I included this value in the polygon for May. A dot map of these observations is presented in Fig. 2. The mean coordinates for the observations (n = 65) were 38.88°N, 88.02°W (SE Jasper Co., IL).
Table 1.
Erynnis funeralis Observations in the Same Location in the Same Year
Expansion rate. The expansion rates for the five months analyzed (June-October) were June = 15.9 km/d, July = 26.4 km/d, August = 25.3 km/d, September = 24.8 km/d, and October = 11.9 km/d, resulting in a median expansion rate of 24.8 km/d.
Single-year movement (2010). Before performing analyses, I eliminated one observation from Will Co., IL, 26 September (already seen 5 September); one from Dane Co., WI, 30 October (already seen 27 August); one from Lauderdale, Co., TN, 16 August (already seen 31 July); and nine from Point Pelee National Park, ON that occurred after the initial 5 September observation. In addition, I counted the observation of two funeralis at Shaw Nature Reserve in Franklin Co., MO on 10 July as one observation. The resulting data set, then, contained 30 observations.
Correlation analysis revealed a positive correlation between day of observation and distance of observation from Phoenix, r = .655, df = 28, P <.0001. Regression analysis revealed that a linear model provided the best fit to the data. The equation of the best-fitting line was y = 10.55x — 77.73, where x = day and y = distance from Phoenix in km. A scatterplot of the 2010 data is presented in Fig. 3.
2010 and non-2010 comparison. Mean distances of the observations from Phoenix for the 2010 and non-2010 data were 2263.19 km (SD = 491.03) and 2417.08 km (SD = 404.49), respectively. The difference was not significant, t = 1.526, df = 79, P = .131. The mean cardinal directions of the observations from Phoenix were 63,38° (SD = 8.49) and 67.77° (SD = 11.04) or roughly ENE (= 67.50°). The difference was not significant, t = 1.875, df = 79, P = .065. The median days of the observations were days 214.5 (range = 177-281) and 217.0 (range 117-294), respectively (2 August and 5 August, respectively, 365-day calendar). The difference was not significant, U = 743.0, z = —0.210, P = .834.
Eastward and northward movement. The median easternmost longitude of the observations was 82.32°W (range = 71.84°-87.97°W). The correlation between year (1990–1992, 1998–2012, 2014) and easternmost longitude at which funeralis was observed was not significant, r = .089, df = 17, P = .717. The median northernmost latitude of the observations was 42.51°N (range = 30.77°-46.77°N). The correlation between year (1990–1992, 1998–2012, 2014) and northernmost latitude at which funeralis was observed was not significant, r = —.005, df= 17, P = .984.
Reproductive Behavior.
Sex ratio. The examination of the available records in which sex was or could be determined (n = 39) revealed 24 males and 15 females. The resulting sex ratio (M:F) was 1.60:1. This ratio did not differ significantly from the expected 1:1 sex ratio, P = .200.
Life cycle observations. An examination of the available records revealed no observations of mated pairs, eggs, larvae, pupae, or eclosing adults. One photograph by Spitzer (2012) taken on 5 May 2012 at Lake Bloomington (McLean Co.), IL, however, appeared to show a female funeralis ovipositing on Lathyrus sp. (Fabaceae), which is in the same family as funeralis's primary larval hosts.
Generational trends. A bar graph of the number of observations during successive 15-day periods (n = 107) is presented in Fig. 4. The graph revealed two primary peaks in adult observations occurring around early/mid-July and mid-/late September and a secondary peak occurring around mid-August.
Discussion
The results of the present study suggest that E. funeralis is not randomly straying into the eastern half of central North America, but regularly expanding its range during the warmer months and establishing breeding populations in that region. Several facts argue in favor of this interpretation:
Adults were reported as far east as the Illinois-Indiana border every year except one between 1998 and 2014, suggesting regular as opposed to random presence in the East;
The majority of adults encountered in the East have been in excellent condition, suggesting that they had not traveled the long distances they are usually assumed to travel. In addition, the multiple observations of two or more adults in the same location in the same year in a species that does not appear to swarm further suggests that these adults had eclosed relatively close to where they were observed;
Data from all years suggest regular as opposed to random expansion throughout the East, as indicated by a) the appearance of adults in the same locations at similar times of year in different years (Table 2 and Fig. 2), b) the relatively constant expansion rate of ∼25 km/d between July and September, and c) the similarities between the 2010 and non-2010 data;
Evidence of reproductive behavior in the East, as indicated by a) an overall sex ratio of 1:1, consistent with that of breeding populations, b) an observation of a female funeralis ovipositing on a legume, implying more widespread reproductive activity, and c) regularly spaced increases in adult observations, consistent with the production of at least two eastern generations per year.
Given these facts, E. funeralis should be conceptualized as a regular seasonal immigrant that establishes temporary breeding grounds in the East rather than as the irregular, non-reproducing stray that it is usually considered to be.
As far as the low number of records of funeralis in the East is concerned, several factors may be involved. First, the data suggest that funeralis may penetrate into the eastern half of central North America to varying degrees each year. The dot map (Fig. 2) reveals a disproportionate number of observations in the more western states (e.g. Illinois, Iowa, Missouri, Tennessee, Wisconsin), suggesting that funeralis does not typically penetrate into the more eastern states and provinces. Therefore, the low number of reports from the more eastern reaches of central North America may reflect actual absence in many years.
Second, eastern populations of funeralis may typically be smaller or more diffuse than those of other seasonal immigrants, which means that observers would be unlikely to encounter it most years in spite of its presence. For example, in some years funeralis has been reported exclusively from more eastern states or provinces (e.g. Ontario in 1990, 1992, 2006, and 2008; North Carolina in 2000; Florida in 2005), suggesting it was also present farther west in those years but remained undetected.
Table 2.
Erynnis funeralis Observations in the Same Location in Different Years
Fig. 3.
Scatterplot of day of year (365-day calendar) of observations of E. funeralis vs. distance (km) of observations from Phoenix, AZ (33.500°N, 112.083°W).
Fig. 4.
Number of observations of E. funeralis in the East during successive 15-day intervals (365-day calendar, n = 107). Days given are median days for ranges.
Third, funeralis is not a “showy” species, being small and having a stereotypieally “moth-like” appearance, so the average (and even knowledgeable) person may tend to overlook it. In addition, if the white hindwing fringe that distinguishes funeralis from all other Erynnis species found in the East is missing, then funeralis can easily be mistaken for one of the dark-fringed species expected for the region. Consequently, an unknown number of funeralis may have been observed or collected in the East but misidentified as dark-fringed members of the genus.
In spite of rare reports of funeralis in the East by midspring (e.g. Clark Co., IN, 20 April 1991), the available evidence suggests that in the United States, funeralis does not overwinter outside the Southwest. Erynnis species overwinter during the larval stage (Burns 1964 p. 14), and although some members of the genus are coldadapted (e.g. horatius, juvenalis, baptisiae etc.), not all Erynnis species are (Burns, pers. com.). The fact that funeralis is typically found in tropical and subtropical environments suggests that it may not be able to survive the northern winters. In addition, the fact that funeralis adults do not appear throughout the northern and eastern United States within a short period of time in the spring further suggests they are not overwintering there.
Extrapolating backward using the expansion rate estimate, one scenario suggests that funeralis begins to disperse from its perennial range in the Southwest sometime in mid- to late April. This time frame is consistent with the onset of expansion of various warmadapted butterfly species into the more northern regions of central North America (e.g. D. plexippus). Spreading out from its perennial range in the Southwest, a funeralis migration wave moving ENE at an average rate of 25 km/d could easily reach the western edge of the Midwest by June (when records there generally start appearing). Of course, funeralis originating farther east, traveling faster, or both could arrive earlier, possibly explaining some eastern observations that have occurred earlier than expected (e.g. Clark Co., IN, 20 April 1991). E. funeralis appears to reach the ultimate extent of its expansion by late September-early October, stopped by natural barriers (e.g. the Appalachian Mountains, the Atlantic Ocean) in the east and probably by developing thermal barriers to the north.
The 2010 expansion of funeralis appears to have attracted special attention (e.g. LeGrand & Howard 2010), but the data presented here suggest that it was different in quantity, not quality. E. funeralis appeared to be especially numerous that year, presumably causing observers to take special note of it. The 2010 and non-2010 comparisons suggest, however, that funeralis behaves roughly the same way in different years.
One question that the present study cannot fully answer is that of how far individual adult funeralis can and do travel within their lifetimes. How the 25-km/d expansion rate estimate relates to the vagility of individual adult funeralis is unclear. This figure includes the lag time required for stationary pre-adult development, which suggests that adult funeralis travel faster and farther than the 25-km/d estimate indicates. Only mark-release-recapture studies assessing the vagility of adult funeralis will settle the issue.
Another question that the present study cannot fully address is the extent to which climate change is affecting funeralis's northward movements. Apart from one observation from Whitefish Point (Chippewa Co.), MI (46.77°N) on 18 August 2001 (Tuttle 2002 p. 48), funeralis has not been reported north of 44°N. In general, the northern limit of funeralis's summer range is about 43°N, a figure that has remained fairly constant for at least 50 years. To answer the question of the possible effects of climate change on funeralis's migratory behavior, data with better coverage of the northern United States and southern Canada would be required.
The primary peaks in adult observations that occur in early/mid-July and mid-date September suggest the production of at least two eastern generations per year. E. funeralis is multivoltine, producing three generations per year (Burns 1964 p. 175 & Fig. 24). The July and September increases may represent the emergence of adults that are the first- and second-generation offspring, respectively, of adults that began migrating northeastward in late spring. Consequently, the midsummer and early fall observations in the East, respectively, appear to be those of second- and thirdflight individuals, not first-flight ones. The secondary peak in mid-August may be an artifact of combining data from different years, which is suggested by the greater number of observations in July and September 2010 than in mid-August of that year (see Fig. 3).
The results of the present study are tentative and much additional data need to be gathered to determine the full extent and nature of funeralis's movements throughout the eastern half of central North America during the warmer months. The implication of the results, however, is that funeralis's presence in that region has been misunderstood. Far from being the occasional vagrant it is usually presumed to be, funeralis appears in fact to be a regular seasonal resident of the eastern reaches of central North America whose presence in that region during the warmer months is much less random than has traditionally been considered.
Acknowledgements
First and foremost, I would like to thank John Burns of the Department of Entomology, National Museum of Natural History, Smithsonian Institution for answering my numerous questions about Erynnis funeralis. In addition, I would like to thank both John Burns and an anonymous reviewer for JLS for their comments on the original version of this paper, which greatly improved the quality of the final version. Finally, I would like to thank Jeffrey Belth, James Wiker, and Myron Cole for sharing their information regarding observations of E. funeralis in Indiana and Illinois.
