Neurocordulia (Collins & McIntyre, 2017)

Geographic distributions of Neurocordulia View in CoL

The genus Neurocordulia exists with two main distribution patterns: while some species span only the eastern-most parts of the USA, and Ontario and Quebec in Canada, N. alabamensis , N. virginiensis , N. yamaskanensis, and N. molesta are distributed on both sides of the Mississippi River and N. xanthosoma is found only west of the Mississippi River. Neurocordulia molesta is recovered as sister to the rest of the Neurocordulia , perhaps suggesting that the original distribution was more broad, extending from eastern Canada south to Florida and west to Texas/New Mexico. Notably, most Neurocordulia are commonly collected at rivers, while N. yamaskanensis is commonly found at lakes/reservoirs as well as rivers; it is possible that other Neurocordulia can dwell in lakes but more behavioural observations are needed. The ecological niche model (ENM) analysis we performed had two aims. First, to understand areas of suitable habitat at the community level that represent hotspots of species richness for Neurocordulia , and second, to estimate the suitable habitat available individually for each species. The stacked pSSDM model shows suitable habitat across a range that covers much of the eastern part of the United States and Canada. Species richness based on this suitability was highest at mid latitudes and low elevations of the eastern United States. In the individual species models, the patchy distribution observed in N. michaeli is perhaps an artifact of sampling. Notably,this is a rare species that wasn’t described until the 21 st century. In contrast, observations of N. yamaskanensis, one of the most widely distributed species of Neurocordulia , is highly abundant and may have wide variation across its range reflecting their population biology. The individual model for this species shows a broad distribution compared to most other species in the genus. Conservation assessments such as NatureServe (NatureServe.org) suggest that several species are critically imperiled in some parts of their range, but overall a species may be listed as secure. Neurocordulia yamaskanensis , for example, is considered to have a G5 rating of “secure” but it is considered status unknown, vulnerable, imperiled or critically imperiled across its range except in Canada, New York and Illinois where it is “apparently secure”. Our estimates of species richness and individual models of habitat suitability could inform these assessments by providing insight into potential areas of distribution where population assessments could be directed. The patterns described above are causes for caution—while a species may be considered of low conservation risk across its range, if local populations decrease in size or local extirpation occurs, the species itself may be at increasing risk

Species maintain separation in areas of overlap

Our PCA, admixture, and haplotype network analyses showed clear genetic separation between N. yamaskanensis and N. michaeli , suggesting little to no hybridization occurring between these two species in the localities we sampled, although our small sample sizes, particularly for N. michaeli , are not adequate to exclude low levels of hybridization. While hybridization may occur within this genus elsewhere, more sampling across species ranges would be needed to assess this.

Neurocordulia genomic relationships supported by morphological patterns

The only other study to evaluate Neurocordulia in a molecular phylogenetic context was limited to N. obsoleta and N. xanthosoma ( Ware et al., 2007) , so there is not a prevailing phylogenetic hypothesis to which we could compare our tree. Neurocordulia virginiensis , N. yamaskanensis, and N. michaeli form a highly supported clade. Neurocordulia michaeli and N. yamaskanensis have similar nymphal features, although N. michaeli are generally smaller ( Catling et al., 2004; Tennessen, 2019).

Genitalic characters can be used to distinguish the species in this genus ( Figure 1 View Figure 1 ). Previous assessments looked at some, but not all of the present species; Byers (1937) noted similarities in the genitalia of N. yamaskanensis and N. virginiensis , in which the female vulvar lamina has a medium rectangular notch for these species. In males, Byers (1937) noted that these species share the trait of a distinct, deep, square/rectangular notch on the inferior margin of the superior appendages. Dunkle (2000) described the shallow notch found in N. molesta , while noting that N. obsoleta females have a smaller, less pronounced notch ( Dunkle, 2000). Neurocordulia alabamensis females have been described as having a small concave crevasse in their vulvar lamina, while N. xanthosoma and N. michaeli had not been previously described in detail. Our work shows that while the notch varies across species, likely in a way that facilitates species identification, what is perhaps more useful is the swollen or flat area just proximal to the vulvar lamina. In N. michaeli these swollen areas have a very distinctive shape that is pointed and almost triangular while these humps are two semicircular flap-like projections in N. alabamensis . The vulvar notch and swollen proximal area are unique to a species in Neurocordulia , and can be considered synapomorphies among populations of a given species.

A key to the species of Neurocordulia View in CoL

1. Wing amber in colour; only two rows of veins in the space between the Cu vein in the hindwing and the trailing edge of the wing; male cerci with ventromedial spine; nymphs with pronounced dorsal hooks: Yes .............. N. xanthosoma

1’. No ........................................................................... 2

2. Presence of a cross vein in the mid basal space; mesotibia and protibia with a long keel that runs half the length of the tibia; male cerci without ventral tubercle; large erect dorsal hooks in nymph: Yes .............................. N. obsoleta

2’.No ........................................................................... 3

3. More than 2 rows of cells in space between the

Cu vein and trailing edge of hindwing and swollen area anterior to the vulvar lamina with pointed, triangular humps; Male cerci with small ventromedial spine: Yes, and nymph with short, blunt dorsal hooks ........................................... N. michaeli 3’.More than 2 rows of cells in space between the

Cu vein and trailing edge of hindwing and swollen area anterior to the vulvar lamina with rounded humps, male cerci with small tubercule ventromedially or no tubercule ........................................ 4

4. Wing with dark spot at the base of the wing extending across part of the anal veins and with a deep rectangular notch in vulvar lamina of female; Yes ............................................................. 5

4’. No, and with fewer than 4 cells afer the pterostigma ............................................................ 6

5. Dark patch in hind wing extending to triangle; Yes, and no ventral tubercle on male cerci, very blunt dorsal hooks that are very short .. N. yamaskanensis

5’. No, and small ventral tubercle on male cerci right adjacent to the distal swelling,very pointed dorsal hooks in nymph ................................. N. virginiensis

6. Swollen humps anterior to vulvar lamina almost as big as flaps of the vulvar lamina; female cerci not longer than 1.6 mm, male cerci with small ventromedial tubercule;large,very hooked dorsal hooks in nymph; Yes ........................ N. alabamensis

6’. No, and male cerci lacking ventral spine or tubercule, female cerci at least 2.4 mm long; nymph with frontal horn on head, dorsal hooks pronounced and midsized ............................ N. molesta