Euphilotes ancilla (Lycaenidae) in the Spring Mountains, Nevada: more than one species?

2008·8.9 MB·English

by  George T Austin

148 Journalofthe Lepidopterists’ Society
JournaloftheLepidopterists’Society
62(3),2008,148-160
EUPHILOTESANCILLA (LYCAENIDAE) IN THE SPRING MOUNTAINS, NEVADA:
MORE THAN ONE SPECIES?
George T. Austin
McGuireCenterforLepidopteraandBiodiversity,FloridaMuseumofNatural Ilistory. UniversityofFlorida, P.O. Box 112710,
Gainesville,Florida32611;email:[email protected]
Bruce M. Boyd
476 McBrideWay,Henderson,Nevada89015
AND
Dennis D. Murphy
DepartmentofBiology/314,UniversityofNevada, Reno,Nevada89557
ABSTRACT.TwoindependenttemporalcohortsofEuphilotesancilla(Lepidoptera: Lycaenidae)withdifferentlarvalhostplantsoccursym-
patricallyinportionsoftheSpringMountainsofsouthern Nevada. Theirdiapauseintensities(asdeterminedinthelaboratory)andflightsea-
sonsexhibitlittleornooverlap,butphenotypesofthecohortsappearidentical. Itisspeculatedthattheyarosefromchangesintherelativephe-
nologyoftheirlarvalhostplantsinresponsetoclimaticalterationssubsequenttothePleistocene. AlthoughtheseEuphilotesseemtobehave
asseparatebiologicalentities,theirtaxonomiclevelremainsequivocal. Untilmoreinformationisforthcoming,theyarerecognizedasseparate
subspecies:E. ancillapurpuraandE. ancillacryptican.ssp.
Additional keywords:conservation,newsubspecies,Polyommatinae,sympatricdivergence.
Euphilotes Mattoni, 1978, a genus ofpolyommatine confined to western North America (Reveal 1969,
Lycaenidae, exhibits an often baffling array of taxa at 1978). More than one taxon of Euphilotes may co-
both specificandinfraspecific levels. Not onlyhas their occur, either synchronously or not, but most co-
genus-level nomenclature experienced numerous occurring species use different host plants (Pratt &
upheavals over the years, but their species-level Ballmer 1986; Ballmer & Pratt 1988; Shields & Reveal
taxonomy has suffered through chaos dating nearly to 1988). Although there are exceptions, a single taxon of
their initial discovery. This taxonomyarguably has had Euphilotes uses but one species oflarval host plant at
less historical consensus than that of any other North any one site (Pratt & Ballmer 1986, Shields & Reveal
American genus of butterflies, with three to as manyas 1988). Theireclosionis closelyconsilientwith flowering
eleven species recognized in myriad combinations phenology of larval host plants, and nearly all
(Barnes & McDunnough 1917; McDunnough 1938; populations are univoltine (Pratt& Ballmer 1986, 1993;
Mattoni 1954a, 1954b, 1965, 1977, 1988; Downey 1961; but see Newcomer 1964; Langston 1974; Shields 1975,
dos Passos 1964; Shields 1974; Langston 1975; Miller& 1977; Pratt & Ballmer 1986; Pratt 1988, 1994; Pratt &
Brown 1981; Scott 1986; Pratt 1988, 1994; Shields & Emmel 1998; Davenport 2003). Pupae of some can
Reveal 1988; Pratt & Emmel 1998; Opler & Warren extend diapause (holdover) through more than one
2002; Warren 2005). This confusion originates from winter(Pratt& Balmer 1986). Phenologiesofbutterflies
veiy similar superficial appearances of the numerous may respond to elevational and latitudinal gradients
taxa; knowledge of larval morphology and host plants, tracking seasonal progression of larval host plants
adult genitalia, and geographical and temporal (Peterson 1997; Pratt & Ballmer 1993). Local
distributions are often necessaryforidentification. populations fly for no more than 4-8 weeks annually
The life cycles ofEuphilotes are closely coordinated (e.g., Langston & Comstock 1966; Arnold 1983a, b;
with those of their larval host plants, Eriogonum Arnold & Goins 1987; Peterson 1995b; Mattoni et al.
(Polygonaceae)(Langston 1963; Langston & Comstock 2001).
1966; Shields 1975, 1977; Arnold 1983a, b; Pratt & The generalities of the life history of Euphilotes
Ballmer 1986, 1993; Arnold & Goins 1987; Pratt 1988, obscure its complexity wherein members ofthe genus
1994; Mattoni 1990; Peterson 1997). That species-rich exploit nearly all possible combinations of spatial,
genus of plants, including spatially and/or temporally temporal, andlarvalhostplantusepatterns (Pratt 1988).
separatedvarieties, is widespread in and nearlyentirely These encompass avarietyofseasonal, elevational, and
Volume 62, Number 3 149
latitudinal replacements, irregular bivoltinism, and co- notable peak.
occurring overlaps in use of larval host plants. The tendency of males of Euphilotes ancilla in the
Numerous instances exist where two or more taxa are SpringMountains tocongregate on stream banks andat
more or less sympatric (either synchronic or seeps from late May to mid-June undoubtedly biased
allochronic), but these usually use different larval host earlyaccounts of distribution. The known andassumed
plants, andare distinguishablemorphologically(Pratt & onlylarvalhostplant (andprincipaladultnectarsource),
Ballmer 1986, 1993; Pratt 1988, 1994; Pratt & Emmel Eriogonum umbellatum Torrey var. subaridum S.
1998; fide G. Pratt). Others have spatiallyapproximate Stokes, is widespread in these mountains, but is often
and apparently consubspecific populations using locally sparse. No butterflies were found at stands of
different larval hostplants that temporallyoverlap for a this plant during VIayand June, since flowers hadyet to
minorityoftheircollective flightseason (Arnold 983a). appear. Priorto 1998, femaleshad notbeenfounduntil
1
These situations suggest incipient speciation (Arnold late July when the then known host plant came into
1983a; butsee Pratt & Emmel 1998). Geneticexchange bloom. These records unfoundedly suggested that
was foundbetweenphenologicallydisjoinedpopulations males emerged a month or more before females and
in Washington (Peterson 1995b, 1996). Those often occurred at mud in large numbers early in their
populations usingthe same larval host plant and having flightseason. Maleswere infrequentlyseen at mudafter
overlapping diapause intensities, however, are not late June, althoughthis resourceiscontinuallyavailable.
sympatric, but elevationally disjunct, and gene How is Original observations on phenology of Euphilotes
thought to be in a stepping-stone fashion along an ancilla in the Spring Mountains were paradoxical for
elevational gradient tracking the phenology of larval several reasons. Euphilotes was not known to emerge
hostplants (Peterson 1995b). several weeks before host plants reach early bloom
During more than four decades ofinvestigations of (Langston 1963; Pratt & Ballmer 1986). Extreme
the butterfly fauna of southern Nevada, observations protandry was unknown among Euphilotes-, the lag of
were made on the endemic Euphilotes ancilla that female emergence had not been found to exceed eight
occurs as several apparently distinct populations at days (Arnold 1983a; Peterson 1995b). Males ofashort-
middle elevations of the Spring Mountains (Clark and lived butterflywith residence times oftwo to nine days
Nyecounties). ThisEuphilotes, referredtoas nearboth (Arnold 1983a, b)wouldnotbeexpectedtoeelose more
Euphilotes enoptes enoptes (Boisduval, 1852) and than a month before the first females emerge. It was
Euphilotes ancilla ancilla (Barnes & McDunnough, fortuitouslydiscoveredduring 1999 thattwovarieties of
1918), as a subspecies ofE. enoptes, by Shields (1977), Eriogonum umbellatum serve as larval host plants for
was considered as an undescribed endemic subspecies Euphilotes in the Spring Mountains: an early-flowering
(Austin &Austin 1980;Austin 1981, 1985). Arevisionof Eriogonum umbellatum Torreyvar.juniporinum Reveal
Euphilotes proposed recognition of several species and a late-flowering Eriogonum umbellatum var.
within the E. enoptes group, the Spring Mountains' subaridum. This suggestedthatperhaps thisEuphilotes
populations became a subspecies ofE. ancilla (Pratt & exhibited a simple bivoltine life history with seasonal
Emmel 1998), and this phenotype was subsequently replacement oflarval hostplants, astrategynotunusual
described as Euphilotes ancilla purpura by Austin among multivoltine butterflies. Since, however,
(1998). Its populations are located at the southern bivoltinism and seasonally alternate larval host plants
extent ofthe distribution ofE. ancilla and their flight are not common among members of Euphilotes,
period extends to the latest reported date for the investigations reported here were focused towards a
species. fuller understanding ofthe distribution and biology of
The first known records ofEuphilotes ancilla in the these butterflies.
Spring Mountains are represented by material at the
American Museum of Natural History taken in July Materials and Methods
1928. Subsequently, there hadbeen few reports (single Distribution andphenology. Spatialandtemporal
records in 1936, 1959, 1966, and 1972) until the late distributions of Euphilotes ancilla were determined
1970s when it was found to be locally common on from specimens, published accounts, field notes, and
occasion, flying from early June to mid-August at more recent surveys. These latter were facilitated by
elevations between 1860 and 2190m (Austin & Austin historical records of and searches for larval host plants,
1980). Later,Weissetal. (1997) hadrecords for 11 sites and observations at water sources where males are
between 1800 and 2500m with an overall flight season encountered at mud. Surveys along roads and trails in
from mid-May through mid-August. The majority of the Spring Mountains were undertaken from late April
records was from early June to early August with no through September 1998-2003includingthewestslope
.
150 Journalofthe Lepidofterists’ Society
of therangefromWheelerPass southeastwardto Potosi with elastic nylon. Flowers of appropriate host plants
Mountain and the Red Rock area and the east slope were maintained in each cup in a plastic bud vial
from BigTimber Springto Harris Mountain (Fig. 1). provided with water that kept flowers fresh and
To further quantify phenology ofEnphilotes ancilla potentially more hydrated than under field conditions.
and its larvalhostplants, five transectswere established Larvae were kept at room temperature ca 21°C) and
(
nearWillowandCold creeks duringspring2002. Three ambient light. Containers and vials were cleaned daily
were within stands ol Eriogonum umbellatum var. and provided with fresh host plant as needed. Excess
juniporinum (two on a hillside above Willow Creek, host plantwas refrigerated at 4°C and replaced by new
1825m and 1850m in elevation, and one on a Hat along stock from the field every 3-5 days. Larvae were so
the roadfromWillowCreektoColdCreek, 1775m) and maintaineduntiltheypupated. Pupaewereplacedona
two in stands ofE. umbellatum var. subaridum (one at bed ofsterilized crushed limestone (collected from the
Cold Creek, 1825m, and the other adjacent to a seep same location as the larvae) in a ventilated plastic
between Willow and Cold creeks, 1775m). These container, separated by date ofpupation, and stored at
transectswerewalkedat 7-13 dayintervals during2002 room temperature and light regimens. On 1 Octoberof
and8-11 dayintervalsduring2004encompassingnearly each year, all pupae were refrigerated at 4°C. Their
the entire flowering season of Eriogonum. Stage of containerwas covered with paper towels (not touching
flower development was recorded for the first 100 the pupae) that were lightly misted every 7-10 days to
plants encountered as none, earlybud, late bud, flower prevent desiccation. Pupae were removed from the
(at least oneperinflorescence), andsenescent (e.g., see refrigerator the following 1 February and maintained
Peterson 1995b). The proportion of plants producing againataroom temperatureofabout21°C. Thesewere
flowers was the maximum in bud or flower, or that had monitored daily until pupae eclosed, died, or failed to
senesced on any one visit (Fig. 2 shows only those that break diapause. These latterwere then again subjected
were in flower). The presence of Euphilotes was also to the refrigeration protocol as above.
recorded.
Results and Discussion
Diapause intensity. Methods for determining
intensity ofdiapause followed those ofPratt & Ballmer General biology. Larvae of Euphilotes feed on
(1993). This, the mean number of days between reproductive parts of Eriogonum including sepals,
,
removal from refrigeration and eclosion, is a standard flowers, pollen, and young seeds (Arnold 1983a; Pratt
indicator offlight season in Euphilotes- its caveats were 1988, 1994; Mattoni 1990; Pratt & Ballmer 1993;
discussed by Pratt & Balmer (1993). Using these data, Peterson 1997). Those from the Spring Mountains are
the occurrence and intensity of diapause in different typical, feeding largely on developing fruit, although
populations may be compared. If two populations are one was recorded feeding on pollen. They remained
distinct, they will have different emergence patterns concealed within inflorescences throughout
that, in the field, should correlate with flowering development. No larval nests were constructed,
phenologies of theirrespective larval host plants. although they are in some populations of Euphilotes
Larvae of Euphilotes ancilla were obtained by (Pratt & Ballmer 1986). Pupation by Euphilotes is
examining larval host plants, with special attention to usually in the soil or among debris at the base ofthe
parts ofplants with ants (see also Arnold 1983a). These larval host plant, but may occur in flowerheads or near
parts and those on which larvae were found were basesofleafaxils (Arnold 1983a, Arnold& Goins 1987).
clipped and transported in plastic containers to the Since all pupae in this studywere on the floors oflarval
laboratory in Henderson, Nevada. In the Willow and containers, these populations are assumed to pupate in
Cold creeks area (1775-1825m in elevation; hereafter litter orsoil.
referred to as Willow Creek), 86 larvae were collected Many larvae in third and fourth instars during 2000
between 3 and 23 June 2000 from Eriogonum were attended by ants. These attendant ants included
umbellatum var. juniporinum-, eight were collected in five species, four associated with larvae from the first
earlyAugust2000, and31werecollectedbetween3and flight and two with the second; one ofthese occurred
22 August 2001 from E. umbellatum var. subaridum. during both flights (Table 1). Data for associations of
Searches forlarvae elsewhere in the Spring Mountains ants for the second flight mayhave been biased by the
proved fruitless during2000and2001, sincehostplants few larvae encountered. In contrast, no ants were seen
had apparentlybeen negativelyimpactedbycontinuing attending larvae during 2001. All species ofants were,
drought. as expected, those ofthe secretion-nectar feedingguild
Once at the laboratory, larvae from all samples were (e.g., Holldobler & Wilson 1990). Associations ofants
individually separated into small plastic cups covered with larval Euphilotes are facultative and seemingly
Volume 62, Number 3 151
® Euphilotesancillapurpura
©Euphilotesancilla cryptica 5 10 15 20 Kilometers
Both taxa
Fig. 1. Distributions ofEuphilotes ancilla in the Spring Mountains, Nevada. Identified sites are (1) Wheeler Pass, (2) Potosi
Mountain, (3) SwitchbackSpring, (4) BigTimberSpring, (5) Harris Mountain Road, (6)WillowCreek, and(7) ColdCreek.
152 Journalofthe Lepidopterists’ Society
2002
Date
Fig. 2a. Phenology ofEriogonum umbellatum near Willow Fig. 2b. Phenology ofEriogonum umbellatum nearWillow
Creek, Spring Mountains, Nevadaduring2002 (encircleddata- Creek, SpringMountains, Nevadaduring2004 (encircleddata-
points indicate presence of adult Euphilotes; see text for loca- points indicate presence ofadult Euphilotes-. see text forloca-
tionsoftransects). tionsoftransects).
unpredictable (Ballmer & Pratt 1991, Peterson 1995a; 1983a); Shields (1975) also found equal numbers of
see also Shields 1973); it is therefore not surprisingthat males and females.
nonewas foundduring2001. These dataon attendance Distribution and phenology oflarval host plants.
by ants, although likely incomplete, represent the first Eriogonum umbellatum var. juniporinum was
reports forthese populations. encountered in the Spring Mountains only in the
Of the 86 larvae collected in June 2000, 81 pupated northeastern portion ofthe range at elevations of 1775
between 8 June and 1 July, the eightcollectedin August to 1950m. At these sites, its dispersion is patchyon diy
2000 pupated between 14 and 24 August, and30 ofthe slopes in sparsepinon-juniperwoodland and in areas of
31 collectedinAugust2001 pupatedbetween 11 August disturbance (especially old burns), with loose soils of
and3 September. Larvae from Eriogonum umbellatum highlimestonecontent. Itbloomsfrom lateApriltolate
var.juniporinum pupated overaperiod of 24 days at an June. This phenotype of Eriogonum umbellatum, with
average of12.8 days (SEM = + 0.55, variance = 38.6%) cream-colored flowers and a rather prostrate growth
after the first pupation; those on E. umbellatum var. form, was described relatively recently and reported
subaridum also pupated over a 24 day period at an from White Pine and Lincoln counties of Nevada
average of 16.6 days (SEM = ±0.84, variance 6.3%). (Reveal 1985a, b). In the Spring Mountains, this is
These means are significantly different (t = 1.982, apparentlytheplant previouslyidentifiedas Eriogonum
df=109). umbellatum var. versicolor S. Stokes (Beatley 1976;
Ofthe 125immatures collected, onedied inthelarval Kartesz 1987). That plant has also been recorded in
stage andfive diedaspupae, all ofunknowncauses, and upperClarkCanyon (Beatley 1976), butthat recordwas
five larvaewereintentionallysacrificedforpreservation.
None was parasitized. The absence ofparasitism was Table 1. Ants associatedwith larvaeofEuphilotesancilla in
the SpringMountains, Nevadaduring2000.
unexpected, although Shields (1973) also reported no
instances ofparasitism. High incidences ofparasitism Species Number Date
by tachnids (Diptera)(42-60%) and braconids
MYRMICINAE
(Hymenoptera)(20%), however, were recorded among
CrematogastermormonumEmery 14 3,8,9, 17June;
Euphilotes in California (Arnold 1983a; Mattoni 1990); 3August
likewise, parasitism by Hymenoptera and Diptera
approaching 60% occurred in Washington (Peterson Monomoriumminimum (Buckley) 9 3August
DOLICHODERINAE
1997).
The reared sample from the Spring Mountains was
Foreliuspruinosum (Roger) 3 3,8,9June
%
female biased (43:65, 39.8 males), althoughthis is not
FORMICIMAE
a significant deviation from equality (chi square =
1.871). In two species of Euphilotes reared from Camponotushyatti Emery 13 3,8,9June
Californianpopulations, thesexratiowas nearly 1:1 with FormicalaevicepsCreighton 6 3,8,9June
males slightly outnumbering females (52.3%; Arnold
Volume 62, Number 3 153
not reverifietl since parts ofClark Canyon are privately percentage produced flowers and seeds. It was first
owned and inaccessible. Most Eriogonum umbellatum seen in bloom during early July and flowered at some
var.juniporinum hadonetoseveralflowerheads during sites into early September. On two transects studied in
2000. It was in flower (10-20% ofthe plants) on the 2002, theplantexhibiteddistinctlycontrastingflowering
first visit to Willow Creek on 20 May, with bloom patterns. Noplantsproducedinflorescences atonesite,
continuing through 28 June; these largely appeared to while 58% of those at the other did. These latter
produce seeds. During 2001, the majority ot plants bloomed from mid-July to beyond mid-August, with a
again had one to several flower heads and also largely peak in late July (Fig. 2a). In 2004, both populations
appeared to produce seeds. The flowering season producedflowers (33-86% of the plants) between early
extended from 9 May through 30 June 2002 (Fig. 2a); Julyand earlyAugustwith apeakin lateJuly(Fig. 2b).
senescence was rapid after mid-June. Plants at two of Distribution and phenology of Euphilotes.
the three transect sites flowered more or less Surveys since 1998 indicated a broader spatial
synchronously peaking in early June; those at the distribution of Euphilotes ancilla in the Spring
remaining site exhibited a peak in mid-June (Fig. 2a). Mountains than previously known and clarified
The proportion of plants producing flowers differed knowledge of its temporal distribution. Euphilotes
between sites with maxima of38 to 64%. In 2004, the ancilla is nowknown from anumberofsites distributed
plants were in flower from 7 May to after 13 June. across much of the range on both slopes from Big
Those at two sites again peaked simultaneously, but in Timber Spring to Switchback Spring in the Red Rock
late May, and the other peaked in early June (Fig. 2b). Canyon areaandon Potosi Mountain between 1775 and
The proportion of plants that produced flowers 2750m (Fig. 1). Its spatial and temporal distributions
(63-90%) exceededthatin 2002. are a subset ofthose ofEriogonum umbellatum. Since
The distribution of the more apparent, brightly all species of Euphilotes fly only during the flowering
yellow-flowered, and erect Eriogonum umbellatum var. period oftheir host plants and do not occur far from
subaridum in the Spring Mountains has been better them, the perceived distributions of butterfly and plant
documented both historically (Clokey 1951; Beatley reflect reality, at least in the more readily accessible
1976) and through more recent surveys. It occurs as portions of the Spring Mountains. No butterflies,
scatteredpopulations across much ofthe range on both however, have been found at numerous other sites that
slopesbetween about 1800and3000m and flowers from support larval host plants. At some ofthose localities,
July through September. Throughout the Spring Eriogonum umbellatum var. subaridum seems too
Mountains, Eriogonum umbellatum var. subaridum had sparse to support Euphilotes dense and apparent
;
a poor flowering year in 2002 with only 5-10% populations of Eriogonum are preferred (Shields &
producing flowers, these mostly in shaded situations. Reveal 1988). Othersites appearsuitable and maywell
Many ofthe flowers dried before they produced seeds support the butterfly, but will require visits over several
and, atWillowCreek,were heavilygrazedbyungulates, years to confirm recorded absences (e.g., see Shapiro
severelyreducingthe numberolflowers availabletoany 2006). Adults in populations of Euphilotes includingin
,
Euphilotes present. It was first seen in bloom in early the Spring Mountains, appearabsentorveryrare during
Julyandhad essentially senesced bythe end ofAugust. dry years suggesting holdover pupae (e.g., Pratt &
The plant flowered profusely in 2004 when a large Balmer 1986; Shields & Reveal 1988). Their absence
Fig. 3a. PhenologyofEuphilotesancilla in theWillowCreek Fig.3b. PhenologyofEuphilotesancillaintheSpringMoun-
area. SpringMountains, Nevada. tains, Nevada, awayfromtheWillowCreekarea.
1
154 Journalofthe Lepidopterists’ Society
from a population of Eriogonum, therefore, may be
more indicative of that year's local weather than the 40
absence of the butterfly. Flowering by Eriogonum is 35
&
related to age of the plant (Arnold Goins 1987;
30
Arnold 1990) and thus age structure must also be
I 25
accountedforinconsideringdistributionsofEuphilotes.
« 20
Historical and recent phenological data from Willow 3
TJ
Creek reveal two flight periods that could be z0 15
interpreted as indicating allochronic sympatry. This z13 10
discoveryoftwocohorts ofEuphilotes separatedintime 5
and with distinct larval host plants having disparate 0
flowering seasons solved the originally perceived
enigma. Cumulative records of adults for this site
extend from 9 May to 25 June and from 11 }ulv to 19 Fig. 4. Diapause intensity of Euphilotes ancilla from the
Augustindicatingpeakflightsin earlyand mid-Juneand Spring Mountains, Nevada.
in mid- andlateJuly,with asingle recordon3 fuly (Fig. associated with Eriogonum umbellatumvar. subaridum
3a). These temporal data from a period of31 years do appear to be chronicallysmall, and adults are often not
not account for annual variation in weather, detected during some years. Records for these
overestimating season length that may occur in any populations extend from 18 June to 14 August and
individualyear, and underestimating intervals between suggestasingle flightpeakingin earlyandmid-Julythat
flight periods. For Euphilotes, initiation, peak, and correspondslargelywith thelate-flyingcohortatWillow
apparentlengthofflight seasons canvaryannuallyupto Creek (Fig. 3b). These records, however, not onlyspan
about three weeks, but dispersion of emergence times 85 years, but are from a variety ofelevations and slope
shows littlevariability(Mattonietal. 2001). Inaddition, exposures that, when combined, obscure local
prolonged rainy periods and high soil moisture may phenological patterns. Adults were frequently seen
extend flowering times ofEriogonum and drought may during 1998 and 1999 along Harris Mountain Road, in
curtail them; both have consequent impact on the lower Kyle Canyon, and in the vicinity ofDeer Creek.
eclosionofEuphilotes (Pratt & Ballmer 1986; Shields & Along Harris Mountain Road, few adults (1-4
Reveal 1988; see also Langston 1974). The virtual individuals) were seen between 19 June and 13 July
absence of records at Willow Creek during late June 2000. Neither adults nor larvae were encountered at
and early July (the three records between 25 June and known sites in lower Kyle Canyon from 12 July to 30
11 July were in one year, 1995) is clarified when data August 2000, or at Deer Creek from 29 June to 13
fromindividualyears areconsideredwhen thetwo flight August. During July2001, adults were encountered in
periods are separated by more than four to perhaps as fair numbers in the Wheeler Pass area, but none was
manyas sevenweeks. Thus, noEuphiloteswere seen at encountered furthersouth in the Spring Mountains.
Willow Creek for 31 days between 24 June and 25 July Diapause intensity. Ofthe 81 pupae from larvae
1998, 35 days from 16 June to 21 July 1999, 45 days collected on Eriogonum umbellatum var.juniporinum,
from 3 June to IS July2000, 45 days from 27 Mayto 1 five (6.8%; 1 male, 4 females) emerged without
July2001, 49 days from 24 Mayto 12 July2002, and 46 apparent diapause between 26 June and 26 September
days between 29 May and 14 July 2003, and 39 days after an average pupal period of 47.8 days (range
between4 June and 13 July2004. Counts ofEuphilotes 14-106 days, SEM = +18.09). The remaining pupae
along transects during 2002 and 2004 were, at best, were refrigerated. One died before and one died
marginally successful probably due to an extended duringrefrigeration. The 74viablepupae (28 males,46
drought. In all instances, however, adults were females) emergedbetween2 March and6April 2001 at
observed during the early or peak stages of flowering an average of46.9 days afterremovalfrom refrigeration
when inflorescenceswere often still largelyin bud (Fig. (range 30-65 days, SEM = +0.54; Fig. 4). Ofthe eight
2, see also Peterson 1997). It is of interest that the pupae from larvae on Eriogonum umbellatum var.
early-flying cohort seems more abundant at mud than subaridum in 2000, seven (4 males, 3 females) emerged
the late-flving cohort; water sources are often spatially between 13 May and 29 June 2001 (one remained as a
closer to the larval host plant used later in the season viable pupa), an average of 116.9 days after removal
than tothatusedearlier. from refrigeration (range 102-149 days, SEM = +6.25;
Away from Willow Creek, populations ofEuphilotes Fig. 4). Of31 larvae collected from that plant in 2001,
t
Volume 62, Number 3 155
one was preserved. The remaining larvae pupated and similar, but yet slightly higher (24%). Variance ofthe
were refrigerated, along with the holdover pupa from emergence dates for the few pupae without apparent
2000. Ofthese 31 pupae, two died, one disappeared, diapause was extreme (89%), perhaps suggesting that
and 15 from larvae collected in 2001 (7 males, 8 this is an abnormal response. The lack ofdiapause is
females) eclosed between 3 April and 17 June 2002 at likely an artifact of experimental protocol (fide G.
an average of 96.8 days afterremoval from refrigeration Ballmer) and notan indication of bivoltinism.
(range62-136days, SEM = ±5.14; Fig. 4); theholdover The seeminglylongerflightperiod ofthe early-flying
from 2000, a female, emerged on day 90. The cohort may be a function of accumulation of degree
remaining 12 pupae continued in diapause and were days that vary annually and with microhabitats of
again refrigerated for four months. One died, six (3 individual pupae since, under controlled conditions,
males, 3 females) eclosed between 28 April and 19 July there was little variance in diapause intensity (82%
2003 at an average of 132.2 days after their return to eclosed within a ten day period); this agrees with a
room temperature (range 87-169, SEM = ±10.58; Fig. seeminglylongeroverallfloweringperiodofEriogonum
4). The remainder, still viable, was again refrigerated. umbellatum var. juniporinum (see Table 2). The
Two of these died in 2004 after removal from extended length of the emergence period of the late-
refrigeration and one eclosed after 76 days. The flying cohort, with its higher variance, may reflect
remaining two pupae remain viable through the end of adaptation to a potentially irregular flowering of the
2005. The eclosion of all 30 pupae from the second larval host plant due to annualvariability in timing and
Bight averaged 109.0 days after removal from amount ol summer precipitation and its effects on soil
refrigeration (SEM = ±4.59). The mean diapause moisture. Holdover pupae may act as a hedge against
intensitiesofpupaefrom the firstandsecondflights and drought (e.g., Nakamura & Ae 1977; Waklbauer 1978;
emerging after one year of overwintering were Shapiro 1980; Sims 1983). Pratt's (1988) data also
significantlydifferent ( = 19.61, df=87). indicated a greater spread of emergence dates for
Emergence ofpupae from the first flight (excluding Euphilotes eclosing late in the season, but perhaps no
thosethatdidnotenterdiapause) extendedover26 days significantseasonaltrend in its variance.
with a variance of 10%, and over 75 days with a Taxonomic considerations. The revelation of two
significantly different variance of21% (F = 18.35) cohorts ofEuphilotes in the Spring Mountains spawns
ig 74
forthe sample from thelate flight collected during2001 uncertainties on their eonspecificity. As noted above,
that emerged without holding over (Table 2). The nearly all Euphilotes are univoltine; bivoltinism occurs
diapauseintensities of males andfemaleswereidentical in few populations and not in all years (Langston 1974;
for first flight individuals, but second flight males Pratt & Ballmer 1986). The existence oftwo ostensibly
preceded females by an average of eight days obligate and site-specific strategies ofvoltinism within
(combined2000 and2001 individuals that emerged the one gene pool seems unlikely. Despite low average
first yeai' after pupation. Table 2). The differences vagility of Euphilotes individual dispersal may exceed
,
between the sexes in theirtime ofemergencearewithin 1000m (Arnold 1983a; Peterson 1997). Consequently,
the range ofreported lag times for Euphrates (Arnold there is no reason to consider that spatially separated
1983a; Peterson 1995b). populations in the Spring Mountains exist as closed
The diapause responses ofthese Euphrates serve to genepools (see alsoPeterson 1995b, 1996). Someother
elaborate the existence oftwo cohorts. The difference butterflies, however, exhibit life histories with split
in mean diapause intensity of 62 days is essentiallythe generations, where some individuals develop directly
same as the differences between first dates that adults and others of the same generation enter diapause or
have been recorded (63 days; 9 May, 11 July) and develop more slowly, in part as a function ofhost plant
between the median date that adults ofeach flight have quality or temperature (e.g., Lees & Archer 1980;
been seen in the field (55 days; 6 June, 31 July). Also, Wiklund et al. 1983; Nylin et al. 1989; Nylin 1992;
and perhaps not coincidently, the differences in first Wedell et al. 1997; Schonrogge et al. 2000; Fischer &
emergence dates of the two cohorts from pupae that Fiedler 2001). Genetic discontinuity of the Spring
entered diapause (32 days) and in the median dates of Mountains' Euphilotes is more likely along a temporal
emergence for each cohort (52 days) are nearly axis.
encompassedbythenumberofdays the specieswas not The origin of and selective agents leading to two
seen at Willow Creek in each of several years (31-49 cohorts ofEuphilotes ancilla in the Spring Mountains
days; see above). The variance in diapause intensity of are at best conjectural. These Euphilotes have likely
individuals ofthe second flight (21%) was greater than been isolated from populations elsewhere since at least
that ofthe first (10%) and that ofholdover pupae was the termination of the Pleistocene as probably have
.
156 Journalofthe Lepidopterists’ Society
other taxa of butterflies there (e.g., Emmel & Austin Devender & Spaulding 1979; Van Devender et al.
1998). Their constancy in phenotype and genital 1987), it is feasiblethatfloweringseasonsofthetwotaxa
morphology across space and time suggests a single of Eriogonum diverged, each blooming at a more
origin. During the Pleistocene and perhaps early favorable season for its respective development (e.g.,
Holocene, they mayhave used one or both varieties of Pratt & Ballmer 1993; see also Shields & Reveal 1988).
Eriogonum umbellatum, although earlierfloweringtaxa This inturnalloweddivergenceofthe butterflyintotwo
seem usualwithin the Euphilotesenoptes species group cohorts having somewhat different diapause intensities
(Pratt & Ballmer 1986; Pratt & Emmel 1998). These with a selection against genetic configurations outside
Eriogonum likely had overlapping or completely the optimum imposedbythephenologies oftheirlarval
synchronous floweringperiods (see arguments by Pratt host plants. Thewarm and dry altithermal (7000-4500
& Ballmer 1993; Pratt 1994) constrained by a cooler BP)(Antevs 1938, 1948; Baumhoff& Ileizer 1965), may
climate (e.g., Spalding & Graumlich 1986; Van well have been the coup degrace fortotallyallochronie
Devenderet ul. 1987; Wharton et al. 1990). It follows flowering periods of Eriogonum and forced the
that the butterfly had a single flight throughout the selection fortwo seasonal cohorts ofEuphilotes
region as is presently usual among Euphilotes ancilla The presumptive seasonal isolation of and perhaps
elsewhere. With subsequent climatic warming (Van absent genetic continuity between these cohorts of
Table2. ComparisonoftheseasonalcohortsofEuphilotesancilla atWillowandColdcreeks. SpringMountains, Nevada.
trait earlyseasoncohort lateseasoncohort
PLANT
larvalhostplant Eriogonumumbellatum Eriogonumumbellatum
var.juniporinum var.subaridum
floweringperiod lateApril-lateJune mid-|uly-earlySeptember
BUTTERFLY
flightseason1 earlyMay-earlyJuly mid-July-mid-August
lengthofflightseason1 55days 39days
visitationtomud commontoabundant infrequent
lengthofpupationperiod2 24days(n=81) 24days(n=30)3
meanlengthofpupationperiod 12.8days 16.6days
varianceofpupationdate 38.6% 6.3%
diapauseintensity 46.9days(range39-65) 109.0days(range62-169)4
varianceofdiapauseintensity 9.9% 22.7%
emergencespan 26days(n=74)5 75days(n=15)6
meanlengthofemergenceperiod7 16.9days 35.3days
emergencetimelag(male-female)8 -0.3days(n=28m,46f) 8.0days(n=11m, Ilf)
non-diapausepupae 5(n=81) 0(n=38)
holdoverpupae 0(n=79) 13(n=35)
1overallfrommanyseasons
2timefromfirsttolastlarvatopupate
3fromlarvaecollectedin2001;thosecollectedin2000pupatedovera10dayperiod(n=8)
4includesholdoverpupae;thiswas96.8daysforthoseemergingthefirstyearafterpupation(range62-149days,n=22)and132.2daysafter
holdingoverforoneyear(range87-169days,n=6)
5onlythosepupaeoflarvaecollectedin2001 thatemergedthefirstyearafterpupation;emergencespanwas47daysforthosecollectedin
2000andnotholdingover(n=7)and82daysforpupaeholdingoverforoneyear(n=6)
6meanofsummationofemergencedaysafterfirstadulteclosed
' non-holdoverpupaeonly
8onlythosepupaethatentereddiapause;apparentnon-diapausepupaehadanemergencespanof92days(n=5)
Volume 62, Number 3 157
Euphrates tempt the consideration of two taxa (the Eriogonum facilitates this phenomenon through its
holotype ol E. ancilla purpura is from the first flight), species richness and biological diversity, where any one
yet the conundrum oftaxonomic level appears. They site may be inhabited by several taxa with an array ol
maybe adjudgedas hostplant andtemporal subspecies actual orpotential phenological specializations.
separated by those and other biological divergences As noted above, the use ofa seasonal progression of
(Table 2, see below). These differences in sympatry, larval host plants is not unusual among multivoltine
however, are potentiallyeffective isolating mechanisms butterflies. In the SpringMountains, thetwoostensibly
thatcouldwell describesiblingspecies, despiteidentical phenotypically identical cohorts ol Euphilotes using
phenotypes. Arguments for two species allied to E. different seasonally available larval host plants appear
ancilla in the Spring Mountains appear most likely superficially to use a simple and comparatively
(Table 2), yet potential for gene flow exists via the few uninterestingbivoltine strategy. It would, however, lie
earlyflight individuals that may not enterdiapause (but unusualin thatEuphilotes are notusuallybivoltine, and
see caveat above) and through the overlap (by one it is apparently unique in that the earlier-flying
individual) in diapause intensity. Low levels of gene “generation” does not give rise to the later-flying one.
flow, however, do not discount species-level Whatever the level of differentiation between these
differentiation (e.g., Sperling 1993). A continuum of cohorts, the Euphilotes in the Spring Mountains
differentiation exists among taxa and, although some represent a heretofore unknown step within the
may not necessarily possess the range of criteria to evolutionary sequence proposed for the genus. The
consider them lull species, they exhibit sufficient question posed in the title of this paper, however, yet
differentiation that does not permit inclusion within a remainswith anequivocal answer.
single species (e.g., Martin et al. 2002). Mallet's (1995, Subspecies Description
see also Sperling 2003) proposal that evaluation of Since the two cohorts ofthe Euphilotes in the Spring
discontinuities in any of many genetic, ecological, Mountains are obviously different taxa feeding on
behavioral, and morphological traits to provide useful differenttaxaofplants, theyat least qualifyas hostplant
templates lor taxon-level inquiries has merit among races. Onlyone ofthese, that flyingin MayandJune (as
these Euphilotes. Species-level recognition, however, noted above), has been named. The later-flying cohort
imposes more questions (e.g., which, if either, is ishere namedanddescribed.
Euphilotes ancilla) and requires information on gene
flowandgenetic distance. Euphilotes ancilla cnjptica Austin & Boyd,
The system, whatever it may be in the Spring new subspecies
Mountains, strongly supports the evolutionaryscenario Diagnosis. Euphilotes ancilla cnjptica is
proposed by Pratt & Ballmer (1993) and Pratt (1994) distinguished from E. ancilla purpura by several
whereby speciation processes in Euphilotes are biological traits (see above. Table 2) including larvae
propelled by opportunistic colonizations of alternately feeding on Eriogonum umbellatum var. subaridum (vs.
available and seasonally disjunct larval host plants E. umbellatum var. juniporinum), flight season in July
concomitant with modification ofdiapause intensities. and August (vs. May and June), and diapause intensity
These differentiations were probably effected in many ol 109 days (vs. 47 days).
instances by climatic perturbations during the Description. Size,wingpattern,andgenitalmorphologyap-
Pleistocene and Holocene modifying distributions anti parentlyidenticalwithE. ancillapurpura (seeAustin 1998),but
phenologies of larval host plants (see also Shields & cwiiltlhacdnijfpfteirecnatibsidoilsotgiincgaluicshhaerdacftreroimstoitchse(rTtaabxlaeo2f).E.Eaunpchiillloatbesyatnh-e
Reveal 1988; Seriber & Ording 2005). The uses of samecharactersasisE. ancillapurpura (seeAustin 1998).
alternative host plants, often with temporal variables, Types. Holotype: Male- NEVADA: ClarkCo.; Spring Mts.,
appear as key events leading to divergence, potentially ColdCreek,20July 1978,leg. G. T. Austin. Allotype: Female—
same data as holotype. Paratypes: (all NEVADA: Clark Co.;
in sympatry, not onlyfor Euphilotes (Pratt 1988, 1994) SpringMountains,leg. G.T.Austin,includingsomeparatypesof
and other butterflies (e.g.. Brown & Heineman 1961; E. ancillapurpura) - same dataas holotype (10m, 2f); same lo-
Carde et al. 1970; Pratt & Ballmer 1991; Scott 1998), cation as holotype, 28 July 1977 (5m, If), 19 August 1977 (If);
but also among other insects (e.g., Tauber & Tauber WalilldleopwosCirteeedk,at2t0heJuMlycGiu.9i7r7e(1Cemn)t,e2r0fo|rulLye1p9i7d8op(t7emr)a.aTnydpeBisoadrie-
1978; Smith 1988; Bush 1994; Feder 1998;Abrahamson versity, Florida Museum of Natural Histoiy, University of
et al. 2001; Emelianov et al. 2001; Berlocher & Feder Florida, Gainesville, Florida.
2002; see also Kankare et al. 2005). Differentiation Type locality. NEVADA: ClarkCounty; SpringMountains,
through allochronic isolation maybe rapid on the order Cold Creek, 1825m in elevation, towards the northern end of
thesemountains.
ol a few centuries in some Lepidoptera (Groman & Distribution. The taxon is known from several sites
Pellmyr 2000; Thomas et al. 2003). The nature of