As Well As Rocky Mountain Wild Flowers

As usual, variation abounds

Spatial Distribution

Though both W. otho and W. egeremet are widespread in eastern North America, otho is decidedly more southern. According to Evans (1955), it extends far to the south, in various differentiated forms, through the West Indies and Central and South America to Argentina. According to MacNeill (1975), it goes only as far as Costa Rica and does not inhabit the Caribbean islands. Within the United States, otho occurs commonly in the Gulf States, ranging northward (1) in the Piedmont and especially the Atlantic Coastal Plain to the Baltimore-Washington area and (2) in the Mississippi-Ohio Valley to the vicinity of Chicago (Figure 63). It shuns the main mass of the Appalachian Mountains and becomes increasingly scarce toward the north.

By contrast, egeremet occurs from the Gulf States to southern Canada and penetrates the Appalachian Mountains (without reaching high elevations); and it is more generally distributed and more common in the northern than in the southern United States (Figure 64). To the south, egeremet stops in central peninsular Florida and, with the exception noted below, in eastern Texas (short of the Mexican border), whereas otho ranges throughout Florida (including the Keys) and west into central Texas. The western limit of egeremet (South Dakota to Texas) varies from about 96° to 97°W longitude (except for four old dateless specimens from Blanco and Burnet counties in central Texas); the western limit of otho (Oklahoma and Texas) is about 99°.

Despite major differences in distribution and abundance, otho and egeremet are sympatric over a large area—shaped something like a fat U—extending from the south end of Lake Michigan southwest to eastern Texas, east to central Florida, and northeast to eastern Maryland. Since otho becomes rare to the north, sympatry is most evident in the lower part of the U. These species coexist at the level of the local population and often fly together.

Nothing elicits additional locality records so well as newly published distribution maps. Because mine (Figures 63 and 64) derive strictly from material examined, they are less than complete (but completely reliable as far as they go). Many of the blanks in the maps are trivial: it is clearly not species of Wallengrenia but collectors of them that tend to avoid, say, Alabama. More critical are distributional limits which, in spite of some effort, remain inexact. For example, with respect to egeremet in the north, I have seen just one specimen from Quebec (Gatineau County) and three from Minnesota (Wabasha County), while this skipper is known to occur (e.g., Macy and Shepard, 1941; Duffy and Garland, 1978) in southern New Brunswick and southern Maine, in and around Montreal, Quebec, and so widely in Minnesota that I suspect it enters extreme eastern North Dakota.

It does not go farther west, however, Lindsey's (1921:84), Lindsey, Bell, and Williams's (1931: 108), and Freeman's (1950) claim of “west to the Rocky Mountains” and Evans's (1955:332) citation of “2 1 Montana” notwithstanding. Evans misinterpreted the opaque abbreviation for Missouri: all three of the so-called “Montana” specimens of egeremet in the British Museum (Natural History) are labelled St. Louis, Mo. Both males are egeremet all right, but the female is Pompeius verna (Edwards). Evans's (1955:332) “Winnipeg” record of egeremet is wrong, too: the Manitoba material filed under egeremet in BMNH comprises four females of Euphyes vestris (Boisduval), of which two are labelled “The Pas” and two, “Swan River. 350 mls. N. of Winnipeg.”

Except for the female from California (another E. vestris), the BMNH specimens on which Evans (1955:333) based the following records of egeremet (“3 2 Florida. 2 1 California. 1 1 Texas”) are correctly determined; but some are incorrectly labelled. I can dismiss without comment the two males marked simply “Californie.” The Texas pair supposedly hails from “Shovel Mt,” which I place around the Blanco/Burnet County line. Because two other old specimens (in the USNM collections) labelled Blanco County reinforce this record, I use it in Figure 64, although it lies appreciably to the west of other known Texas localities. Modern confirmation is desirable. One male and two females marked simply “Florida” remind me phenotypically of egeremet from that state, but two males labelled “Palm Beach, Fla.” do not; and, as this locality is sensibly south of all others, I question it and omit it from Figure 64.

As for Evans's (1955:333) unusual records of otho from the United States (“2 Ohio…3 2 Illinois. 1 1 Alabama…2 1 New York”), the only “supporting” material filed under otho in the BMNH actually consists of one male, one female of otho labelled simply “Ohio”; one male of otho and one male of egeremet, each labelled simply “Illinois”; one male of otho labelled “Chickasaw, Mobile County, Alabama, U.S.A.” and one female of Polites themistocles (Latreille) labelled “Mobile, Alabama, U.S.A.”; and one female of egeremet labelled simply “New York, U.S.A.” In perusing the BMNH holdings of otho and egeremet (as well as of various other skippers), I often found that Evans's published counts of sexed specimens from different geographic areas are off.

Evans did not have an eye for Wallengrenia females. Besides those specifically cited above in connection with geographically questionable records, I noted the following misdetermined females (generally with geographically acceptable data) among otho: one Polites themistocles and one Euphyes conspicua (Edwards); and among egeremet: eight P. origenes (Fabricius), four Pompeius verna, and three E. vestris. All told, 21 out of 55 supposed females of “egeremet” in BMNH were misidentified, and four out of 29 females of “otho.” The males were in good order.

Temporal Distribution

In eastern North America, W. otho is multivoltine throughout its (essentially southern) range—at least trivoltine in peninsular Florida and the Keys and bivoltine to the north (Figure 65). The more northern species, W. egeremet, is univoltine at higher latitudes (and altitudes) and becomes bivoltine at lower ones (Figure 66).

Because univoltine egeremet so neatly fits the gap between the two generations of bivoltine otho, cursory comparison of Figures 65 and 66 might suggest temporal displacement between closely related species. Nothing could be further from the truth. The staggered pattern comes from comparing allopatric populations. Where otho and egeremet are regularly sympatric (see the histograms for the region from South Carolina through the Gulf States in Figures 65 and 66), they fly at similar times. For the most part, their two flight periods are May/June and August/ September. Tendencies to shift toward April and spill over into October are most pronounced in otho from eastern Texas. The flight period in univoltine populations of egeremet embraces July.

Various fragmentary data seem to indicate that egeremet starts developing a second generation where it approaches the main range of otho, or, in other words, that egeremet tends to respond like otho under similar conditions (until, of course, it reaches some limiting level farther south). The histograms in Figure 66 that include (1) Maryland and Virginia and (2) North Carolina and Tennessee are noticeably broader than the unimodal histograms above them, probably owing to incipient bivoltinism in the skipper as well as to great topographic diversity in the sample areas.

Size

METHODS AND CAUTIONS

As a rule, the wings of Lepidoptera reliably reflect adult size. Once expanded and hardened, they hold their dimensions (save for wear and tear), whereas the body, with its flexible and telescoping abdominal segments, does not. To assess size variation, the right primary of spread specimens was measured from base of costa to apex (including fringe) with a vernier caliper calibrated to tenths of a millimeter. All measuring was done by one person, and each specimen was measured on two different days. Repeated measurements were often the same; when they differed, it was usually by just 0.1 mm or 0.2 mm. (To be exact, 41.8% of first and second measurements agreed; 41.4% differed by 0.1 mm; 15.5%, by 0.2 mm; 1.2%, by 0.3 mm; and 0.1%, by 0.4 mm.) Nonidentical measurements were averaged before calculation of sample means and related statistics.

Samples were defined by species, sex, and geographic source—and, at first, in the multivoltine W. otho, by time of flight. After temporal distribution had been worked out, individuals of generation one were segregated from those of generation two (or two plus three) in an effort to extract some of the nongenetic components of size. Similar analysis of size variation in multivoltine species of the pyrgine genus Erynnis revealed a significant increase in the average winglength of individuals between first and later generations (Burns, 1964). Be that as it may, otho showed nothing but an inconsistent tendency toward slightly smaller average winglength in generation two (or two plus three). When the significance of the difference between first and subsequent generation sample means was tested using Student's t, P varied from a low between .10 and .20 to a high of more than .90 and was usually between .40 and .50 (except for the male and female samples from east Texas, in which P was less than .001 and between .001 and .01). Accordingly, data were pooled with respect to time (and calculations were done over).

The areas defining samples are unequal. Both species of Wallengrenia in eastern North America have reasonably continuous spatial distributions that can be divided somewhat arbitrarily. Areas were selected to yield samples of comparable size, at least with respect to males (the commoner sex in collections). For convenience, state boundaries were followed, but with an eye to ecologic as well as areal continuity when two or more states were lumped. A glance at the maps and the lists of material examined will show that, as might be expected, no sample is drawn from its sample area uniformly.

Other sources of sample heterogeneity include weather, which may significantly affect adult size. Specimens of W. egeremet taken at the end of a prolonged drought in the coastal plain of South Carolina (2 4, Plantersville, Georgetown County, 25 August 1974, J.M. and S.N. Burns) were much smaller than average: the females (13.3, 13.3, 13.4, 13.4 mm) were at the lower limit of the observed range in female winglength (13.3–16.0 mm) in the entire southeastern sample (Table 2); and the males (12.6, 12.8 mm) were just above the lower limit of the observed range in male winglength (12.5–15.5 mm). (Winglengths of a dozen specimens of otho taken at the same time and place were more or less average except for two small males, both 12.1 mm.) Numerous specimens of otho taken during the favorably moist and mild spring of 1967 at Austin, Texas, in connection with a study of mating frequency (Burns, 1968) were much larger than average (Table 1). The Austin sample was kept separate (rather than lumped into east Texas) not only because the exceptionally large specimens from 1967 make up 89% of the sample but also because the sample represents a point in space and little more than a point in time. By contrast, all other samples of both species comprise small contributions from many different places and many different times. Such pooling tends to even out the kind of environmental bias under discussion. As a result, means are probably not distorted seriously, although variation within samples is increased.

SEXUAL DIMORPHISM

In both species of Wallengrenia, females average significantly larger than males. In the various geographic samples of W. otho (Table 1), the female edge in mean winglength varies from 0.38 mm to 0.87 mm; in W. egeremet (Table 2), from 0.41 mm to 0.76 mm. For both species, the average female edge is about 0.6 mm. In t-tests of the significance of the difference between male and female sample means, P was usually less than .001, ranging as high as .02–.05 only in two cases involving small sample size.

GEOGRAPHIC VARIATION

Both W. otho and W. egeremet increase significantly in size to the south and to the west, but the parallel cannot be pushed much beyond this level of abstraction. Smallest through the bulk of the Southeast, otho becomes notably large in Florida and, again, in the westernmost states of its range (Tables 1 and 3). Wallengrenia egeremet gets significantly bigger from north to south—in the interior (from Michigan to Tennessee) as well as along the coast (from Massachusetts to Virginia)—but reverses this trend in the southern tier of states (South Carolina to Arkansas and Texas) before becoming extra large in Florida; it also gets extra large to the west, but only in the northern part of its range (Illinois to Nebraska and Kansas; Tables 2 and 4).

The seemingly anomalous north-to-south size decrease in egeremet—which occurs across the entire east-west width of its range—correlates with the shift from a univoltine to a bivoltine condition and probably stems directly therefrom: although the season is mild longer and the overall larval food supply greater, the increase in available time or food may not be quite enough to permit two generations within one year to produce adults as large as the single-brooded ones immediately to the north.

INTERSPECIFIC DIFFERENCE

With due regard for sexual dimorphism, geographic variation, and differences in spatial distribution, comparison of Tables 1 and 2 shows that, despite extensive overlap in winglength, W. egeremet is essentially a larger species than W. otho in eastern North America. Northeastern egeremet averages significantly larger than northeastern otho; northwestern egeremet, than northwestern otho; and Floridian egeremet, than Floridian otho: in these comparisons, the difference in mean winglength approaches or equals one millimeter. In their main area of sympatry—the southern coastal states exclusive of Florida—egeremet again averages about a millimeter longer than otho, except to the west where, especially in Texas, otho comes to equal or even exceed egeremet in mean winglength.

Material Examined

Prolonged and widespread confusion of W. otho and W. egeremet calls for detailed lists of specimens examined. Know the system of organization and condensation before entering the lists and fighting through. It stems from the one used for Erynnis (Burns, 1964:19–20) but differs in minor ways.

Each entry comprises (1) locality, (2) date, and (3) number of males and/or females; it also includes (4) collector(s), when known, and (5) any museum(s) holding some or all of the material cited. Names of museums (abbreviated in parentheses) are spelled out on page 3. I do not cite private collections because sooner or later specimens are given away, exchanged, sold, discarded, eaten, or otherwise destroyed—and owners die. For each species, I arrange entries primarily by locality and secondarily by date, as elaborated below. Besides numbers of males and females from each particular place and time, I give total numbers examined for countries and for states (or provinces).

Entries are ordered alphabetically by (1) country; (2) state (including District of Columbia) or province; (3) county, parish, independent city, district, or municipal region; and (4) specific locality. When a locality is no more specific than county, the entry comes right after the name of the county, ahead of the alphabetic sequence of specific localities. Localities that are elusive because they are extinct, too local, or too politically or geomorphically broad are relegated to County undetermined at the end of the alphabetic sequence of counties for the appropriate state.

Original label data are adjusted as required: localities are carefully placed to county (or the equivalent) whenever county is lacking; misspellings and other errors are corrected; information is occasionally added to locality designations so minutely local as to convey almost nothing or deleted from those that are overly detailed; measures of distance and elevation are converted to the metric system. N, S, E, and W, which are used alone and together (e.g., SW) for cardinal directions and their hybrids, are to be read “north of,” “southwest of,” and so forth.

At the end of an entry, a semicolon in place of a period signals that the previous locality designation still applies; it is not repeated. An elliptic phrase such as “3 km W” following a semicolon means “3 kilometers west of the last-named locality.”

Dates are written day/month/year, with the month reduced to its first three letters and the year to its last two digits (e.g., 14 Aug 54)—except that nineteenth-century years keep all four digits. Omissions and imprecisions in original data are reflected in the citations: 1–9 Sep = the first to ninth of September, 10 Jun = the tenth of June, Jun 10 =June 1910, 00 = 1900, and “no date” = no date. Multiple entries from the same locality are ordered chronologically by (1) year, (2) month, and (3) day, any entries that do not include year coming after those that do.

I routinely put my machine-printed, hand-dated determination label on each specimen examined, which is a little like a dog urinating on the plants along its beat.

As suffix:

0

Unattached to axial skeleton

1

Attached to neural arch or spine

2

Attached to centrum

3

Attached to parapophysis

4

Attached to rib

5

Attached to haemal arch or spine

E

Attached to epipleural bone

N

Attached to epineural bone

P

Attached to lateral process of vertebra

EDITOR'S NOTE: For ease of use, Table 3 appears on page 85.

Characters (* = run unordered):

0. ANA (accessory neural arch) present (0) or absent (1)

1*. Baudelot's ligament originates on V1 (0), or on V1 and V2 (1), or on V2-4 (2), or is ossified (3), or absent (4)

2*. Epineurals all independent (0), or the first five or fewer fused with the neural arch (1), or the first six or more fused with the neural arch (2), or the majority fused with the centrum (3)

3*. Epineurals all originate on the neural arch (0), or those on –V5–15 originate on the centrum or parapophysis (1), or all beyond V5 originate on centrum (2), or all originate on centrum (3)

4*. Epineurals all attached to the axial skeleton (0), or a few unattached epineurals posteriorly (1), or the majority unattached (2), or the unattached epineurals represent only the anteroventral fork and posterior body of bone (notation V0 in Table 5) (3)

5. First epineural “descended” (notation b1, in Table 5) (1) or none descended (0)

6. First two or three epineurals “descended” (1) or one or none descended (0)

7. Some epineurals forked proximally (1) or none forked (0) (autapomorphous condition in Gigantura coded (0)

8*. All epicentrals in ligament (0), or all in bone from VI (1), or all in bone from V3 (2), or epicentrals absent (3)

9. Epicentral on VI in ligament (0), or in bone (1), or absent (2)

10. Epicentral series begins anteriorly at VI (0), or more posteriorly (1), or absent (2)

11. Epipleurals on VI and V2 fused with centrum (1) or free/absent (0)

12. Anterior epipleurals partially in (1) or entirely beneath (0) the horizontal septum

13*. Transition between epipleurals in and beneath the horizontal septum absent (none are in the septum) (0), or gradual (1), or abrupt (2)

14. Number of epipleurals with distal end in horizontal septum zero (0), one (1), or more (2)

15. One or more epipleurals forked distally at the transition between epipleurals in and beneath the horizontal septum (1), or no forking/no epipleurals in horizontal septum (0)

16. Epipleural series confined to the middle region of the body (0), or extends anteriorly to V2 (1), or to VI (2)

17. Epipleural fused with VI and V2 (1) or unfused (0)

18*. All epipleurals attached to the axial skeleton (0), or a few unattached epipleurals posteriorly (1), or the majority unattached (2), or the unattached epineurals represent only the anterodorsal fork and posterior body of the bone (notation D0 in Table 5) (3)

19. Some epipleurals forked proximally (1), or all either unforked or forked only distally (0) (autapomorphous condition in Gigantura coded 0)

20. A brushlike posterodorsal outgrowth of the first neural arch present (1) or absent (0)

21*. All ribs ossified in cartilage (0), or some ossified in membrane bone (1), or all in membrane bone (2), or ribs absent (3)

22*. First rib originates on V3 (0), or V4 (1), or V5 (2). or V2 (3), or VI (4), or ribs absent (5)

23. Ribs on VI and V2 fused to centrum (1), or absent/free (0)

24*. Ratio of number of abdominal vertebrae to number of caudal vertebrae 0.25–1.75 (0), <0.25 (1), >1.75 (2)

25. Median caudal cartilages (CMC of Kafuku, 1990) present (0) or absent (1)

Appendix 1

Alphabetical List of Neopterygian Genera Cited in Text and Tables

(Names in bold italic indicate genera for which we have studied the intermusculars; those in italic indicate genera studied by others. References to Tables 1–5, 7, and 8 are given so that the list provides an alphabetical index to those tables. Family names are given except where a genus is the type of a family. Names of extinct genera are preceded by a dagger (†), and names of extinct paraphyletic stem-group families are placed in quotation marks.)

Abudefduf, Pomacentridae

Acanthopsoides, Cobitidae

Table 4

Adioryx, Holocentridae

Agonostomus, Mugilidae

Table 8

Ahliesaurus, Notosudidae

Table 5