Species delimitation, discovery and conservation in a tiger beetle species complex despite discordant genetic data

Abstract

In an age of species declines, delineating and discovering biodiversity is critical for both taxonomic accuracy and conservation. In recent years, there has been a movement away from using exclusively morphological characters to delineate and describe taxa and an increase in the use of molecular markers to describe diversity or through integrative taxonomy, which employs traditional morphological characters, as well as genetic or other data. Tiger beetles are charismatic, of conservation concern, and much work has been done on the morphological delineation of species and subspecies, but few of these taxa have been tested with genetic analyses. In this study, we tested morphologically based taxonomic hypotheses of polymorphic tiger beetles in the Eunota circumpicta (LaFerté-Sénectère, 1841) species complex using multilocus genomic and mtDNA analyses. We find multiple cryptic species within the previous taxonomic concept of Eunota circumpicta, some of which were historically recognized as subspecies. We found that the mtDNA and genomic datasets did not identify the same taxonomic units and that the mtDNA was most at odds with all other genetic and morphological patterns. Overall, we describe new cryptic diversity, which raises important conservation concerns, and provide a working example for testing species and subspecies validity despite discordant data.

Figure 1

Morphological variants within the Eunota circumpicta species group. From left to right: E. c. pembina (North Dakota: Pembina County), E. c. johnsonii (Texas: Pecos County), E. mecocheila (Mexico: Coahuila), E. c. circumpicta (Texas: Kleberg County), E. c. circumpicta, unnamed variant from Houston area (Texas: Hardin County), described later in this paper.

Figure 2

Map of known E. circumpicta species group localities. Stars represent populations that were sampled for mtDNA and genomic analyses. Colors correspond to morphologically distinct taxa and variants. Purple = E. c. pembina, blue = E. c. johnsonii, green = E. mecocheila, red = E. c. circumpicta (typical), yellow = smaller, duller E. c. circumpicta from Houston area. The open circle represents a putative population from Grand Saline, TX that is based on a single worn specimen of uncertain affinity and was not available for DNA work. The arrow points to the mouth of the Colorado River (i.e. Matagorda Bay), a putative barrier (see “Discussion”). Figure created with QGIS 3.34 (https://qgis.org/).

Figure 3

Maximum-likelihood mtDNA genealogy inferred in IQ-TREE based on cytb gene. Taxon naming follows previous conventions^,,, and colors refer to the same entities as in Fig. 1. E. c. circumpicta (yellow) from Houston area have not been previously named as a distinct taxon. Asterisks indicate individuals that have been called E. c. salinae, generally regarded as a northern population of E. c. johnsonii^,.

Figure 4

TCS haplotype network generated with POPART 1.7 shows the evolutionary relationships of the cytb sequences in this study. Each hatch mark indicates a mutational step between adjacent alleles. The color of each circle corresponds to the geographic location of the sequence (see inset figure legend). The size of the circle is proportional to the haplotype frequency.

Figure 5

Maximum likelihood (RAxML) tree based on SNP dataset. Topology based on 125,895 total loci. Taxon naming follows previous naming conventions^,, with colors highlighting taxonomic groupings, as in Figs. 2 and 3. Asterisks denote individuals belonging to E. c. salinae, generally recognized as a synonym of E. c. johnsonii^,. Vertical bars indicate final taxonomic groups identified by the conclusion of this study, based on the plurality of results (see “Taxonomy” section).

Figure 6

Principal component analyses (PCA) of 4126 (A) and 930 (B) SNPs for the Gulf and Interior Clades, respectively. Loci were limited to those found in a minimum of 50% of individuals in each nominal taxonomic group and in 75% of individuals overall to produce a SNP matrix with relatively little missing data (17.70%; 17.25% for A and B respectively). Transparent points represent replicate analyses (N = 25) while opaque points represent the centroids of these replicates. (A) PCA of Gulf Clade individuals. Red points on the right-hand side of the graph corresponds to typical E. c. circumpicta individuals from west of Matagorda Bay. Gold points on the left correspond to morphologically atypical E. c. circumpicta individuals from East of Matagorda Bay. (B) PCA of Interior Clade individuals. Blue points correspond to E. c. johnsonii, purple points correspond to E. c. pembina.

Figure 7

STRUCTURE analyses of (A) the Gulf Clade based on 4106 SNPs. Shown is K = 2, a test of the hypothesis that the Gulf Clade represents two taxa, as predicted based on morphology. Additional K runs are reported in Supplemental Materials. (B) Interior Clade based on 882 SNPs. Shown is K = 2, a test of the hypothesis that the Gulf Clade represents two taxa, as predicted based on morphology and historical taxonomic treatments. Additional K runs are reported in Supplemental Materials.

Figure 8

Dorsal habitus of male and female Eunota houstoniana, sp. nov. from the type locality: Texas: 1.5 km north of Sour Lake.

Figure 9

Lateral habitus of E. houstoniana, sp. nov. male (top), female (bottom).

Figure 10

Frontal habitus of E. houstoniana, sp. nov. (A) male, (B) female.

Figure 11

Aedeagus of (A) E. houstoniana, sp. nov., (B) E. circumpicta.

Figure 12

Boxplot comparisons of morphological differences between E. circumpicta and E. houstoniana. A two-tailed student’s t-test was used to test for statistical differences. ***A P value of < 0.001.