Electrophoretic Study of Eastern North American Eurylophella (Ephemeroptera: ) with the Discovery of Morphologically Cryptic Species

DAVID H FUNK, BERNARD W SWEENEY, AND ROBIN L VANNOTE Stroud Water Research Center, Academy of Natural Sciences of Philadelphia, Avondale, Pennsylvania 19311 * Ann Entomol Soc Am 8112) 174-186 (19881

comparisons revealed significant geographic variation in allele frequencies, there were no fixed allelic differences between populations. In contrast, most interspecific comparisons were characterized by the lack of shared alleles at several loci. Thus, >99% of the individuals studied could be sorted to species by electromorph characteristics. The average Nei's genetic distance (D) was 0.02 (range = 0-0.12) between conspecific populations and 0.74 (range = 0.11-1.87) between species. A phenogram generated from D values shows five groups of species branching at D < 0.60: *E.uerisimilis, E. verisimilis-A, E. uerisimilis-B, E. verisimilis-C, E. &color, E. minimella, E, pnidentalis, E aestiua, and E. aestha-A; E. tefnporaks-A, E. temporalis-B, and E. temporalis-C; E. sp, 1; E funerahs, and E lutulenta InterspecificD values for Eurylophella were m the upper range of reported literature values for congeneric species, a fact that supports the recent elevation of Eurylophella from subgenenc to generic status Expected heterozygosities ranged from 0 03 to 0 21, those of most species averaged 0 10-0 12

KEY WORDS Insecta, biochemical genetics, , heterozygosity

THEGENUS Eurylophella comprises a small group western North America, and 3 from Europe Two of ephemerelhd whose larvae are often of the eastern North American species, E dons abundant members of the macroinvertebrate fauna (Traver) and E triltneata (Berner), subsequently of rivers, streams, and some lakes of eastern North were synonymized with E tenzporahs (Mc- America Although Eurylophella is easily recog- Dunnough) by Berner (1984), leaving a total of nized, its species are generally difficult to identify nine eastern species McCafterty (1978) included as larvae and often impossible to identify as adults Dannella bartonz (Allen) m Eurylophelln, but Al- (McDnnnough 1931, Allen & Edmunds 1963) In len (1979) disagreed and erected a new subgenus recent years enzyme electrophoresis has been used within Dannella to accommodate D. bartonz This to resolve species and clarify relationships among species was not included in the present study for congeners in many groups However, ex- lack of material We report here the results of an cept in two studies (Saura et a1 1979, Zurwerra et electrophoretic survey of over 2,000 adult Eury- a1 1986) multilocus enzvme electro~horesishas lophella, including all presently recognized eastern not been applied to ephemeropteran taxonomy North American species except E coxalis (Mc- We initiallv undertook an electroohoretic study of Dunnough) Our data show clear biochemical dif- two species (E funerals and E tktszrmhs) to as- ferences among all species We conclude that there ' sess geographic variation in population structure are at least 15 scenes of Eurulovhella m eastern and to verify our morphologically based concepts North America including 7 undeicribed species, A of these species (Sweenev et a1 1987). The discov- thorough morphological- treatment includina de- ery that E vensimihs is m fact a complex of several scnpti&s anda key is in preparation species led us to broaden our original- scope to in- clude the other eastern species Materials and Methods Eurylophella was considered a subgenus of Ephemerella until Allen (1979) elevated it to ge- larvae were collected from 40 locations nenc rank He recognized 15 species, 11 of which in eastern North America (Table 1; Fig 1) Larvae are known from eastern North America, 1 from were returned alive to our laboratory and reared

0013-8746/88/0174-0186902 0010 8 1988 Entomologictl Society of Amems

March 1988 FUNKET AL ELECTROPHORETIC STUDYOF Eurylophella 176 ANNALS OF THE ENTOMOLOGICALSOCIETY OF AMERICA Val 81. no 2

Fig. 1. Geographic distribution of several eastern North American species of Eurylophella as delimited in Allen & Edmunds (1963) and the location of populations (symbols)tested electrophoretically in this study to subimagmes in flow-through polypropylene trays Lap-1, Ldh, Me, and Xdh. Adk was easily scored (23 by 45 by 22 cm deep), which were checked but was not "discovered" until late in the study. several times per day during the emergence period The remaining 18 loci (Mdh-1, Mdh-2, a-Gpdh, Subimagmes were then reared to imagines (here- Hex, Sod-1, Sod-2, Gpi, Pgm, 6pgd, GQpdh, Isdh-1, after called adults) in small (1 liter) freezer con- Mpi, Acp, Aid, GSpdh, Lap-2, Est-5, Est-4) could tamers with screening on top All associated larval be scored unambiguously in all 15 species, and all exuviae were preserved in 80% EtOH and refer- were polymorphic in at least 1 species enced individually (whenever possible) to frozen Expected heterozygosity, or average gene di- adults Genitalia were removed from many of the versity (Hexp), was defined and calculated as de- males and preserved with exuviae When two or scribed by Nei (1978) Unbiased estimates of ge- more individuals emerged together, exnviae were netic distance (D) and identity (I)were calculated referenced by groups When mixtures of species according to Nei (1978) and used to construct pbe- were present, only adults individually associated nograms by the unweighted pair-group method of with exuviae were electrophoresed All of the pre- cluster analysis (UPGMA) (Sneath & Sokal 1973) served material is deposited at the Stroud Water Research Center of the Academy of Natural Sci- Results ences of Philadelphia Discovery of New -60aC Adults were stored individually at until (or Presently Unrecognized) Species electrophoresed Allozymes were separated by hor- izontal starch gel electrophoresis using methods de- F.urylophe1la reriMwitlis. Our ii,iiiipling ol h.'. scribed in Sweeney et al (1986, 1987). A total of c~iiii~iiilibitt i irln li.it.-.ilior.'itin the iiopi~Delii~iirv 34 enzyme systems was screened Of these, 24 were River in New knrk revealed strikingly different scorable in at least one species and yielded data on electromomhs in one population. . and a severe de- 32 presumptive gene loci (see Table 5 in Sweeney ici(-iiq 01 livter~zyg~ttciat rcrlaiii loci Alioin 25 i' et a1 [1987] for full enzyme names and enzyme 21out of St of iii(!ivi

Table 2. Allele frequencies for 15 species of Eu?ylapheIla averaged by species, including 1,426 individuals from 52 populations

spec'es EV EVA EVB EVC EM EB EF ETB ETA ETC EP EA EAA EL ESl Nopols563 47 143 34 14 77 55 113 78 7 106 97 54 37 I 16251341 4 2 2 45111 1 ,,"!,C

Mdh 2 100 0 50

Sod 1 133 130 1 25 100 0 96 0 72 0 48 178 ANNALSOF THE ENTOMOLOGICALSOCIETY OF AMERICA Vol 81, no 2

Table 2. Continued

Species" EV EVA EVB EVC EM EB ETB ETA ETC EP EA EAA EL ES1 563 47 143 34 14 77 113 78 7 106 97 54 37 1 pops 16 422 45111 LOCUS 2 5 1 3 4

0.89 0.88 0.87 0.83 0.78 0.75 0.71 0.68 0.65 6pgd 1.30 1.00 0.55 0.54 0.52 0.30 0.25 0.08 CGpdh 1.54 1.17 1.00 0.94 0.88 0.81 0.67 0.62 Isdh-J 1.00 0.97 0.84 0.75 0.68 0.51 Mpi 1.29 1.19 1.16 1.12 1.10 1.06 1.02 1.00 0.95 0.94 0.92 0.89 0.85 0.81 0.78 0.74 0.72 Acs 2.33 1.75 1.38 1.00 0.81 0.65 0.57 0.36 Aid -1.05 March 1988 FUNK ET AL ELECTROPHORETICSTUDY OF Eurylophella 179

Table 2. Continued

Species" EV EVA EVB EVC EM EB EF ETB ETA ETC EP EA EAA EL ES1 N~ pops 563 47 143 34 14 77 55 113 78 7 106 97 54 37 1 Locus 16 2 51341 422 45111

0 34 ------041003- * - - - - 0 32 007 - 0,0001------0 30 042 079 - - 002 073 008 - - 095043 - - - - 0 29 ------023 005 - - 0 25 002 - 028 040 - 004014 - - - - 003 - - - 0 23 033 010 - - - 004 008 - - 005 042 020 024 - - 0 22 ------om--- 0 21 *------002 - - - 0 20 - - 001052------0 17 k---- 001 - - - - 005 026 052 - - 0 13 ---007--~~---~--- 0 10 ------010 013 - - Est 4 0 65 ------001--*-_-_ 0 bl - - - - A - - 012------0 59 ------002010 ------0 57 ------ofi4---ÑÑ 0 55 *---- 001 - - *0.-i7-ÑÑ 0 53 ------036--ÑÑ-- 0 52 001 - - 003 - * - 020090043 - - - - - 0 51 005 - - 013 ------003 028 - - 0 50 020 002 003 010 - 007~~~~~-005 - - 0 48 028001 ------0 47 004 - - 036 - 034 - - - - - 008 001 10 - 0 45 045 046 072 - 10 043 - - - - - * 004 - - 0 43 ------036~----~-_ 0 42 009 - - 023 - 013 - - - - 003 022 012 - - 0 41 ------009---- 0 40 010 004 023 - - 001" -- --002 001 - - 0 39 ------008 - - - - 0 37 005 - - 012------075 061 041 - - 0 35 - 043001------0 34 ------005 002 - - - 0 33 * - *003----ÑÑÑ- 0 31 ------002 - - - 0 30 - 005--Ñ------0 25 ------008 - - 0 10 ------10

EV,E vensimilis,EVA, E verisimiiu A, EVB, E oensimtlis-B, EVC, E wrisimlis-C, EM E minimelfa, EB, E bicolor, EF, E funeralis, ETA, E temporalis A, ETB E temporahs-B, ETC, E temporalis-C, EP E prudenfalu, EA E aestwa, EAA, E ae8ttm A, EL, E lutulenta, ES1, E sp 1 "*,

Table 3. Observed and expected C6pdh genotype frequencies for Eurylophella from DE3 and the Delaware River"

- -- Observed Expected -Ñ,ÑÑÑ.,ÑÑpà Observed Expected Observed Expected

Departure from Hardy Weinberg eou~l~bnumi~? NSh DlwRerpopulationsNY1-7and PA6pooled for thisanalysis Delaware Riverpopulationsareshown two ways as"wrisimilis"

'Not significant ignated types Unless or until morphological char- erozygote deficiencies, in this case a mixture of acters are found to distinguish reliably all members individuals from two species that had significantly of the E verzstmalis complex in the adult stage (see different allele frequencies (referred to as the Wah- Discussion), we cannot be certain of this associa- lund effect) caused the apparent deficiency No tion Lacking evidence to the contrary, we assume significant deficiencies of heterozygotes were ob- McDunnoueh (1930) was correct and that these served at the C6vdh locus once the two species

v~nszmtlis-Aat four other sites (Table 1) differences in allele frenue~ciesat three other loci After removing E verismlis-A individuals from (Mpz, Acp, Est-5) give additional support to our the data set for all sites on the Delaware River, conclusions (Table 2) Cornoared with E van- there was still a severe deficiency of heterozygotes sds-A, E. verzszmtlzs-B was nearly fixed for al- at the CGvdh locus for E vensmihs (vl test. P < ternate alleles at four loci (I^Idh-l, Pgm, Gpgd, 0005, see Table 3) This was especially notable CGpdh) and shared no common alleles at two others because none of 12 other E venszmilas populations (Sod-1. Est-5) Althoueh larvae of E. verzsmlas-A distributed throughout eastern North America have relatively ~on~ersubmediantubercles on ab- (Georgia to Quebec) exhibited significant depar- dominal tereites five through seven than E. veri- nificantly from site to site (Sweeuey et a1 1987) noth her cryptic species resembling E. v&similis Examination of the preserved larval exnviae as- was discovered in Sinking Creek, Va. (VA5). and is here referred to as E. verisimilis-C. The common allele for E. veriszmilis was absent in this popula- tion at six loci (Sod-1, Pgm, G6pdh, Aid, Est-5, (hereafter referred to as E vmsmlas-B) than in- and Est-4). and there were large differences in dividuals homozveous for allele 1 00 or heterozv- allele frequency at another four loci (Sod-2, Mpi, gnus for alleles 100and 0 81 Subsequent tests on CSpdh, and Lap-2; Table 2). As with E. verisimi- individuals sorted according to head tubercle size lis-A and E. verisimilis-B, E. verisimilis-C larvae prior to electrophoresis confirmed this correlation could be distinguished morphologically from E. for the Delaware River nonulation as well as DOD- verisimilis on the basis of their smaller head tu- . . A illations in Maine, Vermont, and Pennsylvania (sites bercles Compared with E. v~risimihs-A,E. vera- ME4, VT3, and PA2, respectively). Thus, 96% of szmtlas-C had fixed allelic differences at three loci E. verisimilis-B individuals were homozygous for (Gpgd, G6pdh, Aid), shared no common alleles at the 0.81 allele compared with less than 1%of the three loci (Pm, Est-5, Est-4). and had large dif- E. verisimilis individuals found ~~mpatrically.Of ferences in frequency at four others (sod-[ Acp, 171 E. verisimilis-B individuals tested, none was G3pdh. Lad Compared with E verzsimihs-B, homozygous for the common E, verisimilis allele E venstmtlas-C was fixed for an alternate allele at (1.00) at the C6vdh locus. one locus (Aid),shared no common alleles at three loci (Sod-1, Pgm, Est-4). and had large frequency differences at two loci (C3pdh, Lap-2) E uerzsm- March 1988 FUNKET AL ELECTROPHORETIC STUDYOF Eurylophella 181

1s-C showed minor morphological differences from and female adults collected in southern Quebec both E verastmdts-A and E oeriszrmlts-B, with Larvae have never been definitely associated, but closer similarities to the former, but so far we have McDunnough (1931) tentatively assigned a single found no reliable morphological characters to sep- larva collected in Ontario to this species We col- arate these three species from each other lected larvae frnm two localities in North Carolina Eurylophella aesziva. E aestzva was sampled and South Carolina that appear mnrphol~gicall~to from seven locations (Table 1, Fig 1) Allele fre- be conspecific with McDunnough's (1931) speci- quencies were reasonably uniform among popu- men Because adults reared from these larvae clear- lations except at PA4 This population, hereafter ly do not fit McDunnough's (1926) description of referred to as E aestiva-A, shared no alleles with E coxalzs, we refer to these specimens here as E other populations of E aestiva at two loci (G6pdh sp 1 Our electrophoretic characterization of this and Isdh-1) and had unique alleles that were fairly species is based on a single individual from site common (0 19)at two other loci (Mdh-2 and Lap-2, NC3 (Fig 1) As with morphological data, conclu- Table 2) Morphologically, E aestzua A larvae from sions based on electrophoresis of a single specimen PA4 appear identical to E aestiva from the other must be considered tentative However, a reason- trophire

1McDunnoueh)u, bv Burner (1984) Mornholoeicalu and PAS, the populations generally do not sort out ly, E temporalzs-B individuals bear a closer resem- according to geographic proximity. However, the blance to McDunnoueh'su 11931) material than do most electrophoretically distinctive population was E temporalis-A or E temporahs-B However, it DE3 located on the Delmarva Peninsula. This site is unclear which, if any, of the three taxa in this is somewhat isolated from streams in the mountain study represents McDunnough's (1924) onginal E and piedmont areas of eastern North America where temvoralts E. verisimilis is most commonlv found (Fig. 11 182 ANNALS OF THE ENTOMOLOGICALSOCIETY OF AMERICA Vol 81, no 2

Table 4. Intraspecific genetic distances (Nei 1978) and expected heterozygosity for 14 species of Eurylophella. Only samples >five individuals were used to calculate average helerozygosilies

Net's distance Heterozygosity Species No of No of X Range i Range populations populahois 16 0 032 0-0 119 16 0 12 0 10-0 17 2 0 019 2 0 11 0 10-0 12 5 ODOR 0-0 014 5 0 13 0 12-0 14

Population from PA2 only Populations from PA2 and PA6 only cPopulation from PA1 only dPopulation from PAS only Populations from PA6, PA2, VA2 and DE2 only

DE3 from those of populations elsewhere (see Ta- evident for at least one locus for all pairwise in- ble 10 in Sweeney et a1 [1987] Although there terspecific comparisons except between E. veri- were no clear macrogeographic clines with regard similis and E. verisimilis-B. In general, a single to allele frequencies (Sweeney et a1 1987). there individual of any species can be positively identi- does appear to he a microgeographic cline at the fied by its electromorphic phenotype. Hexand Est-5 loci between DE3 and the mainland The phylogenetic relationship among Eury- The 103 allele at the Hex locus has a frequency lophella species was assessed by comparing allele of 77% at DE3 This allele is absent from 12 pop- frequencies at 18 loci for 52 populations of 15 ulations, rare (<3%)at two others, but present at species. For this analysis, E. funeralis was repre- 18% at DE2, the nearest site to DE3 (distance = sented by a single population (PAI); data for nine 44 km) A similar pattern is evident for the 0 30 other populations of E. funeralas could not be in- allele at Est-5, which averages 47% (SE = 4 2) for cluded because some of the 18 loci were not stud- the 14 populations from the mainland, but de- ied, The average interspecific genetic distance (D) creases to 7% at DE2 and 3% at DE3. for Eurylophella was 0.74 (range = 0.11-1.87; Ta- ble 5). A UPGMA phenogram generated from the Interspecific Relationships Nei's (1978) genetic distances for 52 populations is All 18 loci used for this analysis varied signifi- presented in Fig. 3. Individual populations are not cantly among species A lack of shared alleles was " verialmilis verlsimllls-B bicolor verlslmilis-C verlsimllis-A mlnlmella prudontalis aesliwa aesllva-A lemporalls-B L temporalls-A lemporalis-C IP. 1 runeralls lututenfa 0 075 0 050 0 025 0 El'% DISTANCE 1.5 1.0 0.5 0 Nairs Distance Fig. 2. Phenogram of 16 populations of E vertstrnt- 14s based on a cluster analysis (UPGMA) of Neis (1978) Fig. 3. Phenogram of 15 Euri/lopheUa species based genetic distance (D) coefficients Site descriptions for the on a cluster analysis (UPGMA) of Nei's (1978) genetic population codes are given in Table 1 distance (D)coefficients March 1988 FUNK ET AL ELECTROPHORETICSTUDY OF Eurylophella 183

Table 5. Averaee Nei (1978) distance (above diagonal) and idenlily (below diagonal) among 15 species of Eury- lophella

EV EVA EVE EVC EM EB EF ETA ETB ETC EP EA~ EAA~~~~~ FI-- -.-FSI E misirnilis (EV) - 0.43 0.11 0.36 0.56 0.16 0.81 0.48 0.70 0.67 0.42 0.58 0.61 1.24 0.84 E. iimilis-A (EVA) 0.65 - 0.46 0.61 0.50 0.29 0.97 0.62 0.58 0.75 0.58 0.52 0.43 1.24 0.74 E. verisimilis-B (EVE) 0.90 0.63 - 0.33 0.60 0.28 0.99 0.44 0.64 0.69 0.47 0.65 0.58 1.45 0.75 E. censimitis-C (EVCI 0.70 0.55 0.72 - 0.87 0.36 0.83 0.47 0.61 0.66 0.56 0.83 0.73 1,59 0.80 E. minimella (EM) ' 0.57 0.60 0.55 0.42 - 0.46 0.60 0.86 0.99 0.87 0.42 0.51 0.62 1.49 0.79 E. tricolor (EB) 0.85 0.75 0.76 0.70 0.63 - 0.73 0.48 0.60 0.58 0.36 0.44 0.47 1.12 0.74 E. funeralto (EF) 0.45 0.38 0.37 0.44 0.55 0.48 - 1.04 1.01 0.96 0.61 0.68 0.94 1.71 0.82

~-E. lmnrnlis-A~ IETAI 0.162 0.54 0.64 0.63 0.42 0.62 0.35 - 0.28 0.29 0.74 1.03 0.93 1.43~~ 078~ - . ~ ,~ , E. imr~poralis-B(ETB) 0.49 0.56 0.53 0.54 0.37 0.55 0.36 0.75 - 0.37 0.73 0.85 0.75 1.20 0.74 E.temporala-C(ETC) 0.51 0.47 0.50 0.52 0.42 0.56 0.38 0.75 0.69 - 0.63 1.02 0.93 1.87 0.96 E. prudentalis (EP) 0.66 0.56 0.62 0.57 0.66 0.70 0.54 0.48 0.48 0.53 - 0.35 0.56 1.38 0.81 E, aestiva (EA) 0.56 0.60 0.52 0.44 0.60 0.64 0.50 0.36 0.43 0.36 0.70 - 0.15 1.20 0.85 E. aestica-A (EAA) 0.55 0.65 0.56 0.48 0.54 0.63 0.39 0.39 0.47 0.40 0.57 0.86 - 1.55 0.91 E. lutulenta (EL) 0.29 0.29 0.23 0.20 0.23 0.33 0.18 0.24 0.30 0.15 0.25 0.30 0.21 - 1.25 E. sp. 1 (ES1) 0.43 0.48 0.47 0.45 0.45 0.48 0.44 0.46 0.48 0.38 0.45 0.43 0.40 0.29 - shown because all intraspecific lengths were less (mainland) populations (namely, E verzsamzlzs from than the smallest interspecific branch length DE3, E temporalts-B and E aestwa from DE2, (range = 0-0 07) and were generally too short to E prudentalts from DE4, and E funeralis from resolve at this scale The species cluster into the following Gve groups, all of which branch at D > 0 6 -E venszmzlis, E. verisimtlts-A, E vensimalis-B, E verisamtlzs-C, E. bicolor, E minzmella, E pru- our survey of a substantially higher number- of dentalis, E aestiva, E aestiva-A, individuals and populations of E. vensimzhs com- *E. temporahs-A, E. temporahs-B, E temporal- pared with the other species (i e, about 800 indi- is c, viduals from 25 localities ranging- - from Ouebec. to *E sp 1, Georgia) Nevertheless, the geographic variation in E funeralas, and allele frequencies reuorted for E vensamzlas (Swee- E lutulenta. ney et al1987, able 10) probably represents the typical range of variation one might expect within Discussion a species of Eurylophella Although most compar- isons between conspecific populations of Eury- A previous study of geographic variation in the lophella revealed significant allelic variation, there population genetic structure of five ephemerellid were never any fixed allelic differences between species, including E. verisimilis and E. funeralis, them In contrast, almost all interspecific compar- revealed significant genetic differentiation among isons revealed at least one and usually several fixed conspecific populations but no geographic clines in allelic differences Similarly, there was virtually no allele frequencies (Sweeney et al. 1987). In this overlap in the range of genetic distance values mea- study, we compared the levels of genetic differ- sured between conspecific populations (range = 0- entiation among local populations, morphological- 0 119, Table 4) and the range of average values ly cryptic species, and morphologically distinct estimated between congeners (range = 0 11-1 87, congeners of Eurylophella mayflies. Again we found Table 5) Our data suggest that although gene flow no evidence of macrogeographic clines of allele mav be relativelv low amonc consnecific mavflv frequencies among conspecific populations. How-

related speciesL This pattern has also been observed for horseflies (Sofield et a1 1984),pme beetles (Stock tance flight by mayflies to be rare or nonexistent et a1 1984), crickets (Howard 1983), and other because of their short adult life (<48 h for most insect species (see review by Brussard et al r19851) species), we hypothesize that gene flow between We found no electrophbretic evidence for hi- peninsular and mainland populations involves step- bridization among any of the 15 E~lophella wise oveilaud processesrather than direct dispersal species. Although the possibility of some hybrid- across the Chesapeake Bay Thus, these local clines ization between E. verisimilis and E. verisimilis-B probably reflect genetic drift associated with low cannot be ruled out, the fact that E. verisimilis-B gene flow rather than a selectional gradient Con- was always found coexisting with E. uerisimilis sistent with this hypothesis is the fact that popu- suggests that reproductive barriers are effective and lations of all five Eurylophella species sampled from that hybridization is negligible. In fact, many of the Delmarva Peninsula had the highest average the most closely related Eurylophella species are genetic distances from their respective conspecific sympatric in the narrowest sense (i.e., present in 184 ANNALS OF THE ENTOMOLOGICALSOCIETY OF AMERICA Vol 81, no 2 the same habitat at the same localitv. see Table 1) with adults that were comnared electronhoretical-

species that had previous& been found to orient to the area delineated in our Fig 1 All of our col- different (and predictable) swarm markers even lections of E verisimi-lis-A, E verisimtl~s-B,and after being captured and transported to another E verzsimzhs-C are from that area Of the specific area E veriszmihs and E verisirmhs-B are the localities they listed, two in particular are likely to most closely related pair of Eurylophella species have included what we consider E. verastmihs-B based on electrophoretic evidence, and at the seven and possibly E. verisimilis-A- "Beaverkill" (New localities where they coexist (Table l),their emer- York) nrobablv corresnnnds with our site NY3. gence periods appeared to be identical We hy- where E verzsmhs-B is slightly more abundant nothesize that thev too mav avoid interbreedmz- than E. veriszimhs (Table 1). and "Scranton" by differences in swarming behavior, whether by (Pennsylvania) is near our sites PA2 and PAX, where onentine to different swarm markers or swarminff- E vertsmlis-B and E veriszimlis-A are common at different times of day. (Table 1) It seems likely that their concept of E Of the 15 electrophoretically distinct species re- vertsmlis was based on a mixture of at least two, ported here, 7 are morphologically distinct and and possibly all four, of the species we regard as separable from one another as full-grown larvae, the E verisimilts complex These species include E. verisimilis, E. bicolor, E. In comparing McDunnoughSs(1938) E hcolor- minimella, E. prudentalis, E. funeralis, E. sp. 1, aides with E veriszmzhs, Allen & Edmunds (1963) and E. lutulenta. The other 8 species fall into three found only one distinguishing feature, a lack of morphologically distinct groups: head tubercles in the male larvae from Mc- E. aestiva and E. aestioa-A; Dunnough's E hicolorotdes type series They con- E. temporalis-A, E. temporalis-B, and E. tem- cluded that "havmg regard to geographic distri- poralis-C; and bution and morphological characters in all stages, rn E. uerisimilis-A, E. verisimik-B, and E. veri- we consider the nominal E hcoloroides to he a sirnilis-C. junior svnonvm of E verisimihs " But thev SUE- ough morphological study may reveal characters head tubercle character was important enough to to distinguish these cryptic species We found as merit specific or subspecific status This argument- many as five species of Euvylophella to be com- assumes that such a series (i.e., a group of similar mon at a given locality (e g , PA2, Table l), some- individuals collected at the same time and place) times including more than one representative of would consist of a single species. Our data show the second or third group above, and all from the this assumption may be risky; we have found E. same habitat For these reasons, accurate identifi- verisimilis, E. verisimilis-A, and E. verisimilis-B cation of Eurylophella from benthic samples may (all of which would be included in their concept be difficult of E. verisimilis) at the same locality. Thus, a large According to Allen & Edmunds (1963), many series from the type locality of E. bicoloroides could Eurylophella species are morphologically variable, include several snecies. However. based on our especially in the size and spacing of tuhercles on present knowledge of this group, we believe that the head and abdomen However. our examination either E. verisimilis-A or E. verisimilis-B of this

L, u variation to be interspecific. For example, the E. been proposed to date. Allen & ~dkunds(1963) verisimilis complex, shown here to contain at least considered four species, can be divided into two groups based E hicolor and E minimella to be "near cog- on the size of head tubercles on larvae. One group nate" and to form a "complex of closely related consisting of E. verisimilis has distinct, well-de- species with E aestwa and E verisds"; 'eloped head tuhercles in both sexes. In the other E. prudentulis to be similar to E vermmihs group, male larvae have only small roughened areas and E aestiva in some respects, but quite distinct or lack head tubercles altogether, and female larvae overall, have relatively small head tubercles. This group E. temporalis (including E. dons and E. trili- includes three species that are referred to here as neata, which they considered to be of "qnestion- E. verisimilis-A, E. vwisimilis-B, and E. veri- able taxonomic status") and E funeralis to each similis-C. We examined more than 1,000 individ- be quite distinct, and ual larval exuviae from 25 localities ranging from E. lutulenta and "E. coxah?" to be similar to Quebec to Georgia and South Carolina (associated each other based on larval characters. March 1988 FUNK ET AL ELECTROPHORETICSTUDY OF Eurylophella 185

Our phenngram also groups E. verasamalis, E bicolor. E. aestiva. and E mmmella. hut nairs them differently (E verisimzlis with E bicolorr and E aestwa with E manimella) and includes E vru- dentalas with the E aestava and E m&ella References Cited subgroup E. temporalas, E funeralis, and E lu- tulenta represent distinct groups in our pheno- Allen, R. K. 1979. Geographic distribution and re- gram, but E sp l does not group with E lutulenta classification of the subfamily Ephemerellmae as would be indicated by morphological evidence (Ephemeroptera Ephemerelhdae), pp 71-91 In J (assuming it to he conspecific with the larva ten- F Flannagan & K E Marshall [eds 1, Advances in tatively assigned to E coxahs) Ephemeroptera biology Plenum, New York Brussard et al (1985) reviewed genetic similarity Allen, R. K. & G. F. Edmunds, Jr. 1963. A revision of the genus Ephemerella (Ephemeroptera values from 14 studies of various insect taxa I) Ephemerelhdae) VII The subgenus Eurylophella with regard to different levels of evolutionary di- Can Entomol 96 597-623 vergence The average literature value at the local Berner, L. 1984. Eurylophella temporahs (Ephem- population level corresponds to a genetic distance eroptera Ephemerelhdae), a case of synonymy Fla (D) of 0 03 (where D = -In I), which compares Entnmnl 67 567 well with our data for Eurylophella (average D = Rrumard, P. F., P. R. Ehrl~ch,D. D. Murphy, B. A. 0 02, Table 4) Average literature values for sibling Wilcox & J. Wright. 1985. Genetic distances and speciesrauged from 0 06 to 0 57, and for nonsihling the taxonomy of checkerspot butterflies (Nymphali- dae Nymphalmae) J Kans Entomol Soc 58 403- species from 0 15 to 104 Eurylophella species in A19 A*- our phenogram range from 0 11 to 132, and the Howard. D. 1. 1983. Electrovhntet~csurvev of eastern average value was 0 74 (Table 5) E lutulenta has a branch length of 1 32, which is outside the range of literature values for nonsihling species, and would be considered representative of genera within subfamilies (reported literature range = 0 84-1.66) MeCafferty, W. P. 1978. A natural suhgenenc clas If we exclude E lutulenta, our range of D values sificatron of Ephemerella bartom and related species for nnnsibline svecies was 0 11-0 82 faveraee from (Ephemeroptera Ephemerelhdae) Great Lakes Entomol 11 137-138 Table 5 = 0 Nevertheless, the average distance 64) McDunnough, J. 1924. New Ephemendae from New between Eurylophella species is in the upper range England Occas Pap Boston Soc Nat Hist 5 73- of literature values for nonsihlmg species We be- 76 lieve Allen's (1979) elevation of the subeenus Eu- 1926. Notes on North American Ephemeroptera with descriptions of new species Can Entomol 58 184-

A""I Qh ican E lodi or the ~uro~eanE karelica, but ac- 1930. The Ephemeroptera of the North Shore of the cording to Allen & Edmunds (1963) they are mor- Gulf of St Lawrence Can Entomol 62 54-62 phologically similar to E. lutulenta These three 1931. The bicolor group of the genus Eplzem~rella species may represent distinct lineage within the with particular reference to the nymphal stages a (Ephemeroptera) Can Entomol 63 30-42. 61-68 Eurylophella worthy of subgenenc status 1938. New species of North America Ephemeroptera This study represents the first of its kind on a with cntical notes Can Entom01 70 23-34 -eenus of mayflies In the case of Eurulovhella.". elec- Nei, M. 1978. Estimation of average heterozygosity trophoretic techniques have proven to hemore sen- and genetic distance from a small number of indi- sitive than classical morphological methods, en- viduals Genetics 89 583-590 abling the resolution of what appear to be 15 species Sara, A,, .I.Loklii & E. Savolainen. 1979. Etholog- included in what is recognized presently as 8 or 9 ical isolation and genetic diversity Aquilo Ser Zoo1 Similar studies on other groups of Ephemeroptera 20 13-16 are clearly needed, but our results suggest that the Sneath, P. H. A. & R. R. Sokal. 1973. Principles of numerical taxonomy Freeman, San Francisco diversity of Ephemeroptera (or at least the Sofield. R. K., N. E. Bnroker. E. J. Hansens & R. C. Ephemerelhdae) has been underestimated Vrijenhoek. 1984. Genetic diversity within and between sibling species of salt-marsh horseflies (Dip- Acknowledgment tera Tahamdae) Ann Entomol Soc Am 77 663- 668 We thank G M Davis, C Hesterman, andT M Uzzell Stock, M. W., G. D. Amman & P. K. Hitby. 1984. for valuable assistance m developing electrophoretic Genetic variation among mountain pine beetle (Den- techmques for mayflies J D Newhold and T W Con- droctonus ponderosae) (Coleoptera Scolyhdae) Ann don helped considerably with data analysis We thank Entomol Soc Am 77 760-764 B C Kondratieff,R W Lake, and D I Rehuck for help Sweeney, B. W. & R. L. Vannote. 1987. Geographical in locating and collecting specimens and J E H Martin parthenogenesis in the stream mayfly Eurylophella and L L Pechuman for the loan of tvoe soecmens We funeralis in eastern North America Holarctic Ecol 10 52-59 Sweeney, B. W., D. H. Funk & R. L. Vannote. 1986. Population genetic structure of two mayflies 186 ANNALSOF THE ENTOMOLOGICALSOCIETY OF AMERICA vol 81, no 2

(Ephemerella subvana, Ewylwhella verts~mtlzs)m phological and enzyme electrophoretic studies on the the Delaware River drainage basin J N Am Ben- relationships of the European Epeorus species Ihol Soc 5 253-262 (Ephemeroptera Heptagenndae) Syst Entomol 11 1987. Genehc variation in stream mayfly (Insecta 255-266 Ephemeroptera) populations of eastern North Amer- a Ann Entornnl Soc Am 80 600-612 Received for publication 5 March 1987; accepted 23 Zurwerra, A,, I. Tomka & G. Lampel. 1986. Mor- September 1987