The blue-fronted dancer, Argia apicalis Say (Odonata: Coenagrionidae), is an ecologically vagile species inhabiting both pond and stream environments of the eastern United States. Variation in color pattern in A. apicalis occurs between a southeastern United States morph and a south Florida morph. Southeastern populations often are described as “typical” with a predominantly bright blue pterothorax and narrow black humeral stripe, whereas the southern Florida populations are “atypical,” with a bright blue pterothorax and larger, wider black humeral stripes. Variability in color pattern has caused some researchers to question the true identity of the Florida morph. This study used color pattern and mitochondrial cytochromeb sequences to test the species identity of the 2 A. apicalis geographical color morphs. Mitochondrial cytochrome-b gene sequences showed that there is a single haplotype, showing no divergence between individuals, populations, or regions. This study is the first to test if color pattern variation is correlated with molecular characters within this species.
With only a few exceptions, odonates do not exhibit much intraspecific color pattern variation that is clearly correlated with geography despite distributions that encompass several ecological niches and wide areas. A noted exception is the Argia fumipennis (Odonata: Coenagrionidae) complex found in the southeastern region of the United States, which is composed of 3 sub-species: A. fumipennis fumipennis (Burmeister), A. fumipennis violacea (Hagen), and A. fumipennis atra Gloyd, and the subspecies are distinguished based only on wing color and distribution (Burmeister 1839; Hagen 1861; Gloyd 1968). Argia apicalis (Say) (Coenagrionidae: Odonata) is a species complex that shows a similar geographical color pattern variation.
Argia apicalis is an ecologically vagile species inhabiting both pond and stream environments. Say (1839) diagnosed A. apicalis as having a pearlaceous blue thoracic region and a hairline humeral and thin dorsal stripe (Fig. 1A). Other characters used to identify A. apicalis are the distinctive pointed and tooth-like cercus in males (Fig. 1B) and a distribution east of the Rocky Mountains and into Arizona (Fig. 1C) (Garrison 1994).
Bick & Bick (1965), Dunkle (1990), Johnson & Westfall (1970), and Johnson (1972) noted variation of the humeral stripe in southeastern populations of A. apicalis. The subpopulation of A. apicalis in the southeastern range (Suwannee County and Columbia County, Florida) have a broad humeral stripe across the length of the pterothorax. In his analysis of the geographical variation within A. apicalis in Florida, Johnson (1972) categorized both males and females into groupings based on the amount of black markings on the head, thorax, and abdomen, and defined “typical” A. apicalis as having humeral stripe values of 1 or 2, whereas “atypical” specimens had values of 4 or 5. Specimens with a value of 3 were deemed intermediates. Johnson (1972) then discussed the pattern variability in the southeastern distribution and noted that in this region specimens with the typical pattern were less than 5% and specimens with atypical thoracic patterns represented more than 95% of the population. In the summer of 2013, 2 authors herein (Smith-Herron & T. Cook) collected A. apicalis near the Suwannee River and noted that 100% of their specimens (122 specimens) fit Johnson's definition of atypical. This study is the first to combine color pattern and cytochrome-b gene sequences to document variation within an extensive portion of the distribution range of A. apicalis.
Fig. 1.
The 4 defining characters of Argia apicalis; (A) a pearlaceous blue pterothorax and a hairline humeral stripe; (B) a thin dorsal stripe; (C) males have a distinctive pointed and tooth-like cercus; (D) and their distribution east of the Rocky Mountains (shaded areas on map are the recorded distribution of A. apicalis).
Materials and Methods
INSECT COLLECTIONS
Argia apicalis adults were collected using aerial nets in May through Sep 2013 and 2014 from 10 localities in Texas, Louisiana, Florida, and Oklahoma (Table 1). Individuals were field preserved in 70 to 90% ethanol and subsequently dried and curated upon arrival at the laboratory. Specimens were photographed with a Canon EOS 70D mounted on an Olympus SZX12 micr oscope.
We also obtained about 200 specimens on loan from the following museum collections: National Museum of Natural History (NMNH), International Odonata Research Institute (IORI), Georgia Museum of Natural History (GMNH), and Sam Houston State Entomology Collection (SHSUEC), which allowed us to augment our coverage of the distributional range of A. apicalis. The combination of our field-collected samples with existing museum specimens represent about 59% of the reported distribution of A. apicalis in the United States (Fig. 2) (Westfall & May 2006).
MORPHOLOGICAL CHARACTERS
Since its description in 1839, researchers have used hairline humeral stripe found on the thorax to identify and distinguish A. apicalis from its congeners (Johnson 1972). We examined the width of the humeral stripe of the thorax and the pattern and coloration of the head and thorax. For the humeral stripe, the width of the stripe, along the entire length of the thorax, was evaluated and whether or how much it narrowed along its length (Fig. 3). Johnson (1972) reported that individuals with large black patterning on the head usually had hairline humeral stripes whereas individuals with small patterning had wider humeral stripes. To test this observation, we examined head and abdomen of specimens for distinct patterns between the southeastern and the northwestern populations. We also evaluated variation in the morphology of the male caudal appendages and female mesostigmal plates (Garrison 1994; Westfall & May 2006).
MOLECULAR ANALYSES
We extracted genomic DNA from legs for a subset of 29 individuals (Table 2) following the standard protocol instructions of Zymo Research's Quick-gDNATM Mini Prep extraction kit (Zymo Research, Irvine, California). Individuals selected for DNA extraction were taken from field-collected specimens as most museum specimens included in the morphological analyses were collected over 20 yr ago. Special care was taken to select individuals from populations throughout the geographic distribution of A. apicalis (i.e., Florida, Kansas, Louisiana, Oklahoma, and Texas). Although the mitochondrial cytochrome oxidase I (COI) gene has been used widely to detect cryptic diversity and inter-population differentiation in odonates (Brown et al. 2000) and other hexapods (Folmer et al. 1994; Marcus et al. 2009), repeated attempts to amplify this gene fragment in our samples using various previously published primers [e.g. HCO/LCO (Folmer et al. 1994), EVA/ JERRY (Blum et al. 2003)] failed to produce positive amplicons. We thus decided to PCR amplify a 361 bp fragment of the mitochondrial gene cytochrome-b (cyt-b) using primers (151F: 5′-TGTGGRGCNACYGTW-3′; 270R: 5′-AANAGGAARTAYCAYTCNGGYTG-3′) and conditions published by Merritt et al. (1998). The usefulness of this gene fragment in detecting cryptic diversity and population structure in insects and other arthropods is well established (Simmons & Weller 2001; Santamaria et al. 2013). Positive amplicons were cleaned and sequenced at Genewiz (South Plainfield, New Jersey), with resulting sequences assembled and edited (e.g., removal of primer regions) using Geneious R 8.0.2 (Biomatters Ltd.). Dried specimens from which DNA was extracted were labeled as DNA vouchers and deposited into the SHSUEC. DNA template vouchers are stored at −40 °C in the Sam Houston State Natural History Collections (SHSNHC) housed at the Texas Research Institute for Environmental Studies (TRIES) facility, Sam Houston State University.
Fig. 2.
(A) Distribution map of specimens examined (this accounts for about 59% of the reported distribution of Argia apicalis). Dark gray states with black dots represent collected specimen localities, and light gray states with no dots represent areas where no specimens were collected. (B) Distribution of color morphs of A. apicalis in Florida; dark gray represents counties with typical A. apicalis, light gray represents counties with atypical A. apicalis, and the striped area represents the county in which both typical and atypical A. apicalis morphs are present.
Results
ANALYSIS OF COLOR PATTERN
Specimens from the northwestern distribution of A. apicalis (north of the Florida Panhandle) displayed a “typical” hair-line humeral stripe. The humeral stripe in this distribution range varied from almost absent in some specimens to very narrow at most, and rarely extended more than a third of the length of the pterothorax. Although the thickness of the humeral stripe varied in the northwestern populations, the variation was not as pronounced as that between northwestern and southeastern populations (Figs. 4–6). Southeastern populations had an “atypical” humeral stripe that was much wider than the “typical” variation and generally extended beyond the anterior third of the pterothorax and might extend the entire length of the pterothorax. Patterns on the head and abdomen were variable among all individuals and were not correlated with geography. After examining all specimens, we noted that the black patterns on the head and abdomen were variable across all specimens and all geographic regions.
Fig. 3.
Geographical color variation: individuals from the (A) northwestern range of the distribution have a typical (= narrow) humeral stripe, whereas individuals from the (B) southeastern part of the range have an atypical (= wide) humeral stripe.
Table 2.
List of Argia apicalis specimens processed for molecular analysis.
MOLECULAR ANALYSIS
All sequences produced in this study were deposited in GenBank under accession numbers KP770147 to KP770165. We successfully sequenced the target cyt-b gene fragment for 19 A. apicalis individuals from throughout its range in the United States: 8 from 5 populations in Texas, 7 from a single population in Florida, 3 from 2 populations in Oklahoma, and 1 from a single population in Kansas. All 19 individuals harbored a single haplotype for the cyt-b gene, indicating no divergence between individuals, populations, or regions. The remaining 11 DNA extractions (4 from Texas, 1 from Florida, 1 from Oklahoma, 2 from Louisiana, 2 from Kansas, and 1 from Nebraska) failed to produce positive amplicons despite repeated attempts at PCR amplification and re-extraction of DNA. Given our findings of non-existent genetic divergence in this gene, we decided against further efforts to produce sequences for these samples.
Fig. 4.
Variation in width and extension of humeral stripes in the northwest region of the distribution: (A) Dallas County, Texas; (B) Fairfax County, Virginia; (C) Hunterdon County, New Jersey; (D) Iowa; (E) Missouri County, Oregon; (F) Washington Parish, Louisiana; (G) Holmes County, Florida; (H) Wharton County, Texas; and (I) Washington D.C.
Discussion
The 2 color forms of A. apicalis were described briefly by Johnson (1972) and Dunkle (1990). Descriptions of A. apicalis have characterized it as having a pterothorax with a narrow dorsal stripe and either no or at most thin humeral stripe, except for populations from Florida in which the humeral stripe is wider and extends more than half the length of the pterothorax. The geographical divide of the 2 color forms is the Suwannee River and reflects the population disjunctions of the sub-species of A. fumipennis in which A. f. atra occurs east of the Suwannee River, A. f. fumipennis west of the Suwannee River, and A. f. violacea north/northwest of Florida. Where the color morphs overlap, they are intermediate in the color pattern and cannot be identified further.
The 2 color forms of A. apicalis were once documented to coexist near the Suwannee River but their co-existence may no longer be true. As we found no differences in cyt-b gene sequences between the 2 color morphs throughout their distribution range, their recognition as separate species is not justified. Although no mechanism or cause is known for the variation seen in both the A. fumipennis and A. apicalis complexes, Johnson (1972) suggested allopatric processes associated with sea level changes during the Pleistocene (e.g., fragmentation of Florida into islands and formation of the Suwannee Straits) may be responsible for the observed differences.
Acknowledgments
The authors would like to thank Karen Pruitt and Cowboy Joe Herron for aiding in the collection of specimens from west and south Texas and Lenora Reid for granting permission to collect on her property. We would also like to extend our appreciation to Daniel Haarmann, Sibyl R. Bucheli, Brent C. Rahlwes, Soheyla Bayat, and Ashley R. Morgan for their assistance in the identification of specimens and discussion of molecular techniques. We are truly grateful to Christopher P. Randle, who allowed the use of his research lab to conduct the molecular portion of this research. The authors would like to acknowledge the Georgia Museum of Natural History, the International Odonata Research Institute, the National Museum of Natural History, and the Sam Houston State University Natural History Collections for providing loan specimens that aided in the overall research. Collection permits were granted by the Florida Department of Environmental Protection Division of Recreation and Parks (#05291310) and the Louisiana Department of Wildlife and Fisheries. This research was made possible by funding provided by a Sam Houston State University Enhancement Research Grant (#290041). Lastly, we would like to thank the several anonymous reviewers for their careful edits and helpful comments that greatly improved the manuscript.
