Multigene phylogeny of the scyphozoan jellyfish family Pelagiidae reveals that the common U.S. Atlantic sea nettle comprises two distinct species (Chrysaora quinquecirrha and C. chesapeakei)

Introduction

Scyphozoan jellyfishes (Cnidaria, class Scyphozoa), which include the conspicuous moon, lion’s mane and sea nettle jellyfishes, exhibit significant and widespread economic and ecological impacts on a wide array of marine and estuarine communities. Jellyfish aggregations, blooms, and swarms damage economically important fisheries, close tourist beaches by stinging swimmers, clog intakes of coastal power and desalination plants, invade ecosystems, and can affect oxygen levels when mass numbers of carcasses are deposited (Arai, 1997; Purcell, Uye & Lo, 2007; Richardson et al., 2009; Bayha & Graham, 2014; Qu et al., 2015). On the other hand, jellyfish serve important roles as major prey items for some fish and sea turtles, in carbon capture and advection to the Deep Ocean, as important microhabitat for fish, invertebrates, and symbiotic algae, and as economic resources for humans (as food and therapeutic compounds) (Doyle et al., 2014, Omori & Nakano, 2001; Castro, Santiago & Santana-Ortega, 2002; Arai, 2005; Houghton et al., 2006; Lynam & Brierley, 2007; Ohta et al., 2009; Lebrato et al., 2012; Diaz Briz et al., 2017). Recent attention given to large medusae blooms has led to speculation that anthropogenic events are driving global increases in jellyfish bloom magnitudes, though long-term data sets are still equivocal on this point (Richardson et al., 2009; Brotz & Pauly, 2012; Condon et al., 2013).

Despite their importance, evolutionary and taxonomic relationships of even some of the most recognizable scyphozoan species remain unsettled, which can impede our abilities to effectively study, predict and mitigate the ecological and economic effects of these nuisance species. Recent systematics studies have directly challenged taxonomic relationships at all levels. A mitogenomic analysis recently challenged the placement of the order Coronatae, such as Periphylla, within Scyphozoa (Kayal et al., 2013; but see Kayal et al., 2017) and the new family Drymonematidae was created based on morphological, molecular, and life history data (Bayha & Dawson, 2010; Bayha et al., 2010). Studies employing molecular and/or morphological data have revealed novel species in multiple scyphozoan genera, including the moon jellyfish Aurelia (Dawson & Jacobs, 2001; Schroth et al., 2002; Dawson, 2003), the genus Drymonema (Bayha & Dawson, 2010), the upside down jellyfish Cassiopea (Holland et al., 2004), and the lion’s mane jellyfish Cyanea (Dawson, 2005; Kolbasova et al., 2015). Many of these studies have uncovered unrecognized jellyfish invasions and clarified evolutionary relationships in the group (from order to species level) vital to understanding their ecological and economic impacts, and elucidating the evolution of traits that permit these impacts.

The scyphozoan family Pelagiidae (Gegenbaur, 1856), currently made up of four genera (Pelagia, Chrysaora, Sanderia, and Mawia), contains some of the world’s most notorious blooming jellyfish. The geographically widespread mauve stinger (Pelagia noctiluca) forms dense aggregations that heavily impact aquaculture, fisheries, and tourism along the North Sea and Mediterranean Sea (Canepa et al., 2014). Recently, a species found for the first time in the Mediterranean was described and assigned first to the genus Pelagia (Piraino et al., 2014), but later to the novel genus Mawia, based on molecular and morphological data (Avian et al., 2016). Blooms of the jellyfish Chrysaora fulgida (previously identified as Chrysaora hysoscella) have increased over past decades in the Northern Benguela current on the west coast of Africa, coinciding with decreased fish catches and general breakdown of beneficial trophic interactions as compared to nearby ecosystems not jellyfish-dominated (Lynam et al., 2006; Flynn et al., 2012; Roux et al., 2013). Likewise, blooms of very large Chrysaora plocamia medusae form off the coast of Peru, interfering with fisheries, aquaculture, and power plants by clogging nets, seines, and water intakes (Mianzan et al., 2014).

A species of special note is the U.S. Atlantic sea nettle Chrysaora quinquecirrha (Desor, 1848), one of the most recognizable, well-studied, and ecologically important jellyfish along the U.S. Atlantic and Gulf of Mexico coasts (Mayer, 1910; Hedgepeth, 1954; Larson, 1976). Because its predation pressure shows ecosystem-wide, controlling influence on zooplankton dynamics (Feigenbaum & Kelly, 1984; Purcell, 1992; Purcell & Decker, 2005), Chrysaora quinquecirrha has been termed a keystone predator for the Chesapeake Bay ecosystem (Purcell & Decker, 2005). The jellyfish negatively impacts economically important fisheries by feeding on eggs and larvae (Duffy, Epifanio & Fuiman, 1997; Purcell, 1997) and blooms impact tourism by stinging swimmers (Cargo & Schultz, 1966; Schultz & Cargo, 1969; Cargo & King, 1990). As a result, a program was developed to predict both real-time occurrences of sea nettle blooms (Decker et al., 2007) and year-to-year bloom magnitudes using past data on environmental conditions (salinity, temperature, etc.) that favor jellyfish populations (Purcell et al., 1999; Purcell & Decker, 2005).

Generic definitions within what is currently accepted as family Pelagiidae (Gegenbaur, 1856) have been historically vague and genera have traditionally been differentiated, to a great extent, on a single morphological character (tentacle number). The generic names Pelagia and Chrysaora were originated by Péron & Lesueur (1810), though both included species not recognized today as pelagiids. Gegenbaur (1856) was the first to create a higher taxon, the family Pelagiidae, including all pelagiids known at the time, but which also included some jellyfish currently classified as coronates. Noting differences based on tentacle number between Chrysaora and Pelagia, Agassiz (1862) erected a new genus, Dactylometra, within the family. Among other characters, Agassiz (1862) classified genera based on tentacle and lappet numbers: Pelagia (eight tentacles, 16 marginal lappets), Chrysaora (24 tentacles, 32 marginal lappets), and Dactylometra (40 tentacles, 48 marginal lappets). Kishinouye (1902) subsequently described the genus Kuragea (56 tentacles, 64 marginal lappets) and Goette (1886) described Sanderia (16 tentacles, 32 lappets, and 16 rhopalia). To the genus Dactylometra, Agassiz (1862) added Pelagia quinquecirrha (Desor, 1848) from Nantucket Bay (MA) and Chrysaora lactea (Eschscholtz, 1829) from Rio de Janeiro. Based on established generic definitions, Piraino et al. (2014) placed an undescribed, presumably non-indigenous Mediterranean pelagiid, Pelagia benovici, in the genus Pelagia. However, Avian et al. (2016) created the novel genus Mawia for this new species (Mawia benovici) based on fine-scale morphological characters (tentacle, gonad, and basal pillar morphology) and molecular differences from other pelagiid genera included in a lightly sampled phylogenetic analysis of Pelagiidae.

Not long after Agassiz erected Dactylometra, Dactylometra quinquecirrha served to cast doubt on pelagiid generic discrimination. Bigelow (1880) recognized that some brackish water (e.g., Chesapeake Bay) Dactylometra quinquecirrha matured at 24 tentacles (a character of Chrysaora) rather than 40 (a character of Dactylometra), something Mayer (1910), saw as the “Chrysaora” stage in their development to the “Dactylometra” stage. Stiasny (1930) also cast doubt on the ability to effectively differentiate Chrysaora and Dactylometra. As a result, Kramp (1955) reasoned Dactylometra and Kuragea to be merely developmental stages and subsumed both within the genus Chrysaora (Eschscholtz, 1829), since it has taxonomic priority. Calder (1972) determined that Chrysaora quinquecirrha went through stages of one to more than seven tentacles per octant, often in the same geographic region, supporting the contentions of Mayer (1910) and Kramp (1955). A morphology-based phylogeny of the Pelagiidae (Gershwin & Collins, 2002) indicated two groups coinciding with the previous genera Dactylometra and Chrysaora, but noted that the weak phylogenetic support would make resurrecting the genus Dactylometra premature. Another morphology-based phylogeny (Morandini & Marques, 2010) found support for a Dactylometra clade based on tentacle and lappet number, but noted that this would require many Chrysaora species to have their own genera. A robust phylogenetic hypothesis of relationships within Pelagiidae based on comprehensive taxon sampling is an important step toward removing taxonomic confusion at the genus and species-levels, including assessing the taxonomic status of the new genus Mawia (Avian et al., 2016) and clarifying taxonomic questions related to Chrysaora quinquecirrha.

In order to examine evolutionary relationships and taxonomic boundaries in the family Pelagiidae, with special focus on the genus Chrysaora and the species Chrysaora quinquecirrha, we collected nuclear (large subunit ribosomal rDNA) and mitochondrial (cytochrome c oxidase I and large subunit ribosomal rDNA) sequence data from individuals representing all four extant genera (Chrysaora, Mawia, Pelagia, and Sanderia), including 11 currently recognized species of Chrysaora and one species each of Mawia (Mawia benovici), Pelagia (P. noctiluca), and Sanderia (S. malayensis). To further examine the taxonomy of the U.S. Atlantic sea nettle Chrysaora quinquecirrha, specimens were included from its entire range along the U.S. Atlantic and Gulf of Mexico coasts (estuarine and coastal), taking care to sample all recognized morphotypes, with representatives also examined morphologically (macromorphology and cnidome).

Materials and Methods

Sample collection

Specimens were collected in the field or at public aquaria husbandry facilities, either by the authors or others with extensive knowledge of Scyphozoa, in an effort to collect as many species of Chrysaora as possible, as well as representative species of Pelagia, Mawia, and Sanderia (Table 1; Fig. 1). An unknown and unidentified pelagiid specimen was collected from Dakar, Senegal and was accompanied by a photograph that did not allow for specific identification (Fig. S1). For Chrysaora quinquecirrha, samples were collected from 10 different sites along the Atlantic and Gulf of Mexico coasts (Table 1; Fig. 2), covering both coastal and estuarine environments, with the intention of capturing as many structural and color morphotypes as possible (Fig. 3). Both white (Table 1: NF1–NF3) and red-striped (Table 1: NF4–NF5) color morphs (Figs. 3C and 3D) were collected from Norfolk, VA (NF). In all cases, a small piece of gonad, tentacle or oral arm tissue was excised and preserved in 80–99% ethanol or DMSO-NaCl solution (Dawson, Raskoff & Jacobs, 1998). Where possible for some sites (Table S1), individuals were also preserved in 4% buffered formalin and seawater for later morphological analyses. Additional published pelagiid sequences were included in the final data set (Table 2).

Table 1:

Geographic source regions of samples used for molecular analyses in this study, identified by taxon (original, morphologically based identification) and molecular ID (identification after molecular analyses).

Original ID	Final ID	Location	Code	n
				COI	16S	28S
Chrysaora achlyos	C. achlyos	Monterey Bay Aquarium*	MBA	1	1	1
Chrysaora africana	C. africana	Coastal Namibia	NAM	2	2	2
Chrysaora chinensis	C. chinensis	Monterey Bay Aquarium^	MBA	2	2	2
Chrysaora colorata	C. colorata	Aquarium of the Americas+	AQA	1	1	1
Chrysaora fulgida	C. fulgida	Coastal Namibia	NAM	5	5	2
Chrysaora fuscescens	C. fuscescens	Aquarium of the Americas+	AQA	1	1	HM194868
Chrysaora hysoscella	C. hysoscella	Cork, Ireland	IRE	3	3	3
Chrysaora lactea	Chrysaora c.f. chesapeakei	Kingston, Jamaica	JAM	5	5	2
Chrysaora lactea	C. lactea	Rio de la Plata, Argentina	ARG	1	1	1
Chrysaora melanaster	C. melanaster	Bering Sea	BER	–	1	AY920780
Chrysaora melanaster	C. pacifica	Monterey Bay Aquarium	MBA	1	1	HM194864
Chrysaora plocamia	C. plocamia	Puerto Madryn, Argentina	PMA	2	2	2
Chrysaora quinquecirrha	C. quinquecirrha	Buzzard’s Bay, MA (USA)	MA	1	1	1
Chrysaora quinquecirrha	C. quinquecirrha	Cape Henlopen, DE (USA)	CHP	3	3	2
Chrysaora quinquecirrha	C. quinquecirrha	Offshore South Carolina (USA) (32.60 N, 79.21 W)	OSC	2	2	1
Chrysaora quinquecirrha	C. chesapeakei	Charlestown Pond, RI (USA)	RI	4	4	–
Chrysaora quinquecirrha	C. chesapeakei	Tom’s River Harbor, NJ (USA)	NJ	3	3	1
Chrysaora quinquecirrha	C. chesapeakei	Rehoboth Bay, DE (USA)	RB	3	3	–
Chrysaora quinquecirrha	C. chesapeakei	Norfolk, VA (USA)	NF	5	5	–
Chrysaora quinquecirrha	C. chesapeakei	Pamlico Sound, NS (USA)	PAM	3	3	–
Chrysaora quinquecirrha	C. chesapeakei	St. Simon’s Island, GA (USA)	GA	3	3	1
Chrysaora quinquecirrha	C. chesapeakei	Perdido Pass, AL (USA)	AL	3	3	1
Pelagia noctiluca	P. noctiluca	Offshore Virginia (USA) (37.81 N, 73.91 W)	OVA	1	1	HM194865
Sanderia malayensis	S. malayensis	Monterey Bay Aquarium	MBA	1	1	HM194861
Unknown Pelagiidae	M. benovici	Dakar, Senegal	SEN	2	2	1
Cyanea capillata	C. capillata	Blomsterdalen, Norway	BLO	1	1	HM194873

DOI: 10.7717/peerj.3863/table-1

Notes:

For six individuals, 28S sequences from those individuals were published previously. For S. malayensis, 16S/COI and 28S sequences came from the same culture, but two different individuals. For some aquarium specimens, the geographic source region for the culture is known: *near Los Angeles, CA (USA); ^{^}Northern Malaysia; ⁺near Monterey Bay, CA (USA).

Table 2:

Geographic source regions of previously published sequences used in in this study identified by taxon (previous identification) and molecular ID (identification after molecular analyses).

Original ID	Final ID	Location	Code	n
				COI	16S	28S
Chrysaora melanaster	C. melanaster	Bering Sea	BER1	KJ026191	–	–
Chrysaora melanaster	C. melanaster	Bering Sea	BER2	KJ026212	–	–
Chrysaora melanaster	C. melanaster	Bering Sea	BER3	KJ026256	–	–
Chrysaora sp.	Chrysaora c.f. chesapeakei	Bocas del Toro, Panama	PAN	JN700941*	JN700941*	AY920779*
Chrysaora pacifica	Chrysaora pacifica	Kyoto, Japan	KYO	LC191577	–	–
Chrysaora quinquecirrha	C. pacifica	Geoje-do, Korea	KOR	HQ694730	HQ694730	–
Chrysaora sp.	Chrysaora sp. 1	Noosa Heads, Australia	AUS	DQ083524	–	–
Chrysaora sp.	C. chinensis	Malaysia	MAL1	–	JN184784	–
Chrysaora sp.	C. chinensis	Malaysia	MAL2	–	JN184785	–
Chrysaora sp.	C. chinensis	Malaysia	MAL3	–	JN184786	–
Pelagia benovici	P. benovici	Northern Adriatic Sea	ADR1	KJ573409	–	KJ573396
Pelagia benovici	P. benovici	Northern Adriatic Sea	ADR2	KJ573410	–	KJ573397
Pelagia benovici	P. benovici	Northern Adriatic Sea	ADR3	KJ573412	–	KJ573401
Pelagia noctiluca	P. noctiluca	Southern Tyrrhenian Sea, Italy	TYR	KJ573419	–	KJ573408
Pelagia noctiluca	P. noctiluca	Cape Town, South Africa	SA	JQ697961	–	–
Pelagia noctiluca	P. noctiluca	Dispensa Island, Costa Rica	CR1	JX235441	–	–
Pelagia noctiluca	P. noctiluca	Dispensa Island, Costa Rica	CR2	–	JX235404	–
Pelagia noctiluca	P. noctiluca	Dispensa Island, Costa Rica	CR3	–	JX235405	–
Pelagia c.f. panopyra	Pelagia c.f. panopyra	Papua, New Guinea	PNG	KJ573422	–	–

DOI: 10.7717/peerj.3863/table-2

Note:

* Sequences came from the same individual.

Results

Sequence data characteristics and phylogenetic inference

The COI dataset consisted of 73 sequences, 59 of which are new. All sequences were 616 bp in length. The 16S data set was made up of 67 sequences, including 60 new sequences and 7 published sequences. New complete sequences varied in length from 598 base pairs (bp) for Chrysaora lactea to 608 bp (Chrysaora chinensis). The MAAFT-aligned data set (included published sequences) was 628 bp, but the dataset was truncated to 582 bp (95.7%) after treatment with GBlocks. The 28S dataset included 35 sequences, including 24 new sequences and 11 published sequences. New sequences ranged in size from 998 (Chrysaora chinensis) to 1,018 bp (Chrysaora africana). The MAFFT alignment (which included published sequences) was 1,027 bp, but the final data set was 1,015 bp (98.8%) after removal of regions via GBlocks.

All phylogenetic analyses (COI, 16S, 28S, and combined) revealed similar terminal clades, but they differed in the resolution of relationships among them. The combined analysis provided the best resolution (smallest proportion of polytomous nodes) and highest support values for evolutionary relationships (Figs. 4–7). In all analyses, Chrysaora is revealed as paraphyletic with respect to species of Sanderia, Pelagia, and Mawia. In the combined analyses, Mawia benovici is most closely related to S. malayensis (Bayesian support 100/ML support 100), with these two species forming a close relationship with Chrysaora africana and Chrysaora pacifica in the combined (88/67) and 28S trees (100/61) (Figs. 6 and 7). Except for the COI tree, P. noctiluca formed a close relationship with a clade of Pacific jellies (Chrysaora achlyos, Chrysaora colorata, Chrysaora fuscescens, and Chrysaora melanaster) with high support values (combined: 100/99; 16S: 100/92; 28S: 82/58) (Figs. 5–7). For the combined analyses (100/100) and 28S (100/100), a highly supported clade was composed of Atlantic species, including Chrysaora quinquecirrha, Chrysaora lactea, Chrysaora plocamia, Chrysaora fulgida, Chrysaora hysoscella, Chrysaora chesapeakei [see Discussion], and the Caribbean Chrysaora, while this clade was less supported for COI (100/61) and 16S (75/60) (Figs. 4–7). Chrysaora fulgida (NAM), Chrysaora plocamia (PMA), and Chrysaora hysoscella (IRE) formed a closely related group in all analyses with high support values (combined: 100/100; 28S: 100/99; COI: 100/94; 16S: 100/83). For sequences taken from Piraino et al. (2014) only, nuclear 28S sequences for Mawia benovici from the Mediterranean (ADR) occurred in the distantly related clade for P. noctiluca from the Atlantic (OVA), and a P. noctiluca from the Mediterranean (TYR) occurred in the distantly related clade for Mawia benovici from the Mediterranean (ADR) (Fig. 6).

At the species level, our analyses highlighted multiple species boundaries, and showed the samples identified as Chrysaora quinquecirrha to be polyphyletic. In all analyses, Chrysaora quinquecirrha sequences fell into two distinct, highly diverged clades (Figs. 4–7; Tables S3–S5), with one clade (Chrysaora chesapeakei—see “Discussion” and “Systematics”) made up of animals from U.S. Atlantic estuaries and the Gulf of Mexico animals and another (Chrysaora quinquecirrha—see “Discussion” and “Systematics”) made up of U.S. coastal Atlantic animals. Caribbean Chrysaora (Jamaica and Panama) formed a clade closely related to Chrysaora chesapeakei in all analyses (Figs. 4–7). Aquarium animals previously identified as Chrysaora melanaster (AQA) were genetically distinct from Chrysaora melanaster collected from the Bering Sea (BER) in all analyses where both were included (Figs. 4–6) and formed a clade with Chrysaora pacifica collected from South Korea (KOR) and Japan (KYO) for COI and/or 16S. While aquarium collected Chrysaora chinensis formed a well-supported clade with field collected Chrysaora chinensis (MAL), analyses differed in where this species fell out in the trees (Figs. 4–7). The unknown pelagiid collected from the Western African coast (SEN) was nearly identical to the newly described Mawia benovici from the Mediterranean for COI (0.0–0.3% difference) and 28S (0.0–0.2% difference) (Figs. 4 and 6).

Macromorphological and nematocyst analyses

A total of 120 medusae (57 field collected and 63 museum specimens) (Table S1) previously identified as Chrysaora quinquecirrha s.l. were examined for 19 quantitative and qualitative macromorphological characters taken from Gershwin & Collins (2002) and one new to this study (maximum oral arm length) (Table 3). Overall, three macromorphological characters differed significantly: tentacle number, lappet number, and maximum oral arm length vs. bell diameter (Table 3). Animals collected from the estuarine Atlantic and all Gulf of Mexico sites (Table S1) had an average of [3.07 ± 0.07] 95% CI tentacles per octant, excluding two aberrant individuals (6 and 4.625—see “Discussion”) (Fig. 8A; Table 3). In all instances when there were more than three tentacles per octant (excluding aberrant individuals above), the additional tentacle(s) occurred between the secondary tentacles and the rhopalia (i.e., 3-2-1-2-3 octant tentacle orientation) and were typically undeveloped, being of similar size to nearby lappets. Animals collected from coastal regions along the U.S. Atlantic (Table S1) had an average of [5.28 ± 0.48] 95% CI tentacles per octant (Fig. 8A; Table 3). Animals collected from the estuarine Atlantic and all Gulf of Mexico sites (Table S1) had oral arms that were on average 3.00 ± 0.39 (95% CI) times as long as the bell diameter (Fig. 8B; Table 3). Animals collected from coastal regions of the U.S. Atlantic (Table S1) had oral arms that were on average [1.24 ± 0.27] 95% CI times as long as bell diameter (Fig. 8B; Table 3). Of the animals that were examined morphologically, a total of 38 individuals were also sequenced for 16S to see which Chrysaora clade they fell into (K2P sequence divergence <1.5%). Medusae examined morphologically that fell into the Chrysaora chesapeakei phylogenetic clade had an average of 2.99 ± 0.03 tentacles per octant and oral arms that were [2.80 ± 0.78] 95% CI times as long as bell diameter on average, while all those that fell in the Chrysaora quinquecirrha clade had an average of 5.63 ± 0.78 tentacles per octant and oral arms that were on average 0.93 ± 0.18 (95% CI) times as long as bell diameter on average (Figs. 8A and 8B).

We also studied the cnidome of medusae identified as Chrysaora quinquecirrha, examining the measurements of individual nematocyst types (Fig. 8C; Fig. S1), as well as the representation of each type overall. Nematocyst measurements indicated significant grouping for holotrichous A-isorhizas, but not for other types. A-isorhiza measurements (length vs. width) showed two distinct groups, with one group containing only animals from U.S. Atlantic estuaries and the Gulf of Mexico and the other containing coastal Atlantic animals (Fig. 8C). All sequenced animals in the smaller group (coastal Atlantic) were genetically similar to Chrysaora quinquecirrha for 16S, while all jellyfish from the larger group (estuarine Atlantic and Gulf of Mexico) that were sequenced for 16S were genetically similar to Chrysaora chesapeakei (Fig. 8C). For animals identified as Chrysaora chesapeakei (based on habitat, macromorphology, and/or genetics), LDA analysis indicated that individual A-isorhiza measurements correctly identified species 97.8% of the time (2.2% of the time, they were incorrectly identified at Chrysaora quinquecirrha), while they were correctly identified 100% of the time using the mean of 10 nematocyst measurements. For animals previously identified as Chrysaora quinquecirrha (based on habitat, macromorphology, and/or genetics), LDA correctly identified them 100% of the time, whether one or 10 nematocysts were used. Figure S2 (A–C) shows measurement graphs for a-isorhiza, O-isorhiza, and heterotrichous microbasic rhopaloids, all of which indicate no significant groupings of measurements.

Nematocysts from proximal, medial, and distal regions were typed and counted (200 total) for 10 individuals originally identified as Chrysaora quinquecirrha, chosen based on their previous molecular and macromorphological groupings (five from each group). All in all, heterotrichous microbasic rhopaloids were most frequent ([62.1 ± 9.8%] 95% CI), followed by O-isorhizas ([13.4 ± 5.0%] 95% CI), a-isorhizas ([12.4 ± 2.8%] 95% CI) and A-isorhizas ([12.2 ± 3.7%] 95% CI). As pilot studies indicated, nematocyst type proportions were different for different tentacles regions. While A-isorhizas and a-isorhizas were consistent over the entire tentacle, O-isorhizas were overrepresented in proximal regions and heterotrichous microbasic rhopaloids were overrepresented in the medial and distal regions (Fig. S3A). Individuals varied considerably in proportions of nematocyst types (Fig. S3B). Individuals collected from coastal Atlantic regions (circles) were generally clustered, including those genetically similar to Chrysaora quinquecirrha, while those from estuarine Atlantic and Gulf of Mexico regions (squares) were much more dispersed, as were those genetically similar to Chrysaora chesapeakei (Fig. S3B). LDA was moderately effective in distinguishing species using overall nematocyst proportions (four of five Chrysaora quinquecirrha and three of five Chrysaora chesapeakei correctly classified) and this was almost entirely due to different proportions of A-isorhiza nematocysts. A-isorhiza proportions were significantly different (t = 3.623, p = 0.0068), with Chrysaora chesapeakei individuals averaging 16.5 ± 3.4% for A-isorhiza and Chrysaora quinquecirrha cnidomes averaging 7.8 ± 3.4%.

Discussion

Genus-level systematic inference

Our most robust phylogenetic hypothesis for Pelagiidae (Fig. 7), based on the combined data set, directly contradicts current generic definitions, as well as earlier morphological-based phylogenies of the Pelagiidae. Both Gershwin & Collins (2002) and Morandini & Marques (2010) considered Chrysaora to be reciprocally monophyletic with respect to both Sanderia and Pelagia, with Sanderia in a basal position (Figs. 9A and 9B). In contrast, our analyses indicate that Chrysaora is paraphyletic with respect to Pelagia, Sanderia, and the newly erected Mawia (Figs. 4–7 and 9E). Mediterranean Mawia benovici is not in the combined analysis, but our Senegal pelagiid (SEN) can be treated as Mawia benovici, based on COI (Fig. 4) and 28S (Fig. 6) phylogenies (see below). Paraphyly of Chrysaora is not supported in morphological or genetic analyses in Avian et al. (2016) (Figs. 9C and 9D), but this is almost certainly a result of incomplete taxon sampling. For example, their analysis based on combined morphological and genetic data only included Chrysaora hysoscella (Mediterranean), while the 28S dataset included a subset of sequences published at the time, (Chrysaora hysoscella, Chrysaora lactea, and Chrysaora c.f. chesapeakei [see below]), all of which occur in a single clade in our analysis (Figs. 7 and 9E). Including fewer published sequences gave the appearance of Chrysaora monophyly, which may have biased the establishment of Mawia. For instance, throughout Avian et al. (2016), Chrysaora is often used as a singular entity (i.e., monophyletic), such as an entire section that examines characters at the “genus level”. This more readily allows for the conclusion of a novel genus Mawia, as it sidesteps the difficult taxonomic questions raised by the paraphyly of Chrysaora. That notwithstanding, in agreement with both Piraino et al. (2014) and Avian et al. (2016), our analyses show Mawia benovici to be a close relative of S. malayensis (Figs. 4–7). Given the stark morphological differences between Sanderia and Mawia (Piraino et al., 2014; Avian et al., 2016), this relationship is more than a bit surprising.

Our working hypothesis for the relationships within Pelagiidae (Figs. 7 and 9), especially the paraphyletic Chrysaora, raises serious systematic questions for the genus level. To accept the validity of Mawia, as well as previously established Pelagia and Sanderia, each of which can be easily distinguished morphologically from those currently classified as Chrysaora, additional genera would have to be erected within Pelagiidae in order to maintain monophyly of these generic groupings. An initial matter would be to which clade should the genus Chrysaora should be limited. Because the type species of Chrysaora is Chrysaora hysoscella, the genus would best be limited to the clade containing Chrysaora hysoscella, Chrysaora fulgida, Chrysaora lactea, Chrysaora plocamia, Chrysaora quinquecirrha, and Chrysaora chesapeakei (see below). This then would leave three other lineages in need of new genera: (1) Chrysaora africana plus Chrysaora melanaster; (2) Chrysaora chinensis; and (3) Chrysaora achlyos, Chrysaora colorata, and Chrysaora fuscescens. The latter grouping (Chrysaora achlyos, Chrysaora colorata, and Chrysaora fuscescens) has a close relationship to P. noctiluca (except for COI) and there is genetic support for generic designation. Unfortunately, none of the morphological characters employed in this study clearly diagnose this clade or other Chrysaora lineages, as has been the case in other studies seeking to reconcile jellyfish taxonomy based on morphology and molecular data. (Dawson & Martin, 2001; Dawson, 2003; Bayha & Dawson, 2010). Future study will benefit from more detailed morphological analyses to identify additional characters that could then be mapped onto molecular phylogenies (e.g., Fig. 7), as well as greater taxonomic sampling (e.g., two additional Chrysaora species accepted and two declared nomen dubium in Morandini & Marques (2010), more geographic samples of Pelagia and Sanderia). Both would allow for better resolution to define genera and better explain their evolutionary relationships.

Systematics

• Chrysaora quinquecirrha Desor, 1848

• Figs. 3A, 3B, 4–9; Figs. S1 and S2.

Pelagia quinquecirrha-Desor (1848): p. 76 (original description—Nantucket Sound, MA).

Dactylometra quinquecirrha: Agassiz (1862): 126, 166 [tentacle number]. Agassiz (1865): 48, 49 [tentacle number; Naushon, MA]. Fewkes (1881): 173, Pl. VIII Fig. 14 [tentacle number, drawing]. Brooks (1882): 137 [tentacles, drawing in Mayer, 1910; southern variety outside Beaufort Inlet]. Agassiz & Mayer (1898): 1–6, Plate I [tentacles, oral arms, drawing]. Fish (1925): 128, 130 [Vineyard Sound, MA; Nonamesset, MA; Lackeys Bay, MA]. Mayer (1910): 585–588, Pl. 64A [tentacles, drawing].

Chrysaora quinquecirrha: Kramp (1961): 327–328 [description fits both Chrysaora quinquecirrha and Chrysaora chesapeakei]. Calder (1972): 40–43, Figures 5–6 [mouth of Broadkill River, DE]. Kraeuter & Setzler (1975): 69, Figures 1–2 [offshore samples, Sea Buoy]. Calder (2009): 24–28 [offshore animals collected on continental shelf possibly Chrysaora quinquecirrha].

Diagnosis: Living medusae up to 40 cm (observed 59.0–176.0 mm) (Figs. 3A and 3B); tentacles typically 40 or more; 5.28 ± 0.45 (95% CI) tentacles/octant on average (Table 3; Fig. 8A); lappets rounded typically 48 or more; 6.26 ± 0.46 lappets/octant on average; rhopalar lappets slightly larger than tentacular lappets; can be differentiated from Chrysaora chesapeakei based on (1) smaller size of holotrichous A-isorhiza nematocysts: average: 20.25 [±0.38] × 11.27 [±0.37] μm (Table 3; Fig. 8C); (2) larger tentacle number (more than five per octant); and (3) typically shorter maximum oral arm length (average: 1.24 ± 0.27 time bell diameter).

Material examined: USNM 24496 (n = 1; Buzzard’s Bay, MA), USNM 53860 (n = 1; Assateague Island, VA), USNM 53861 (n = 1; Assateague Island, VA), USNM 54511 (n = 2; Cape Henlopen, DE), USNM 56702 (n = 1; Cape Henlopen, DE), USNM 1454776–USNM 1454778, KMBCDE2, KMBCDE4 (n = 5; Cape Henlopen, DE).

Description of holotype: No holotype located, no neotype designated.

Description of specimens: Bell diameter up to approximately 40 cm (observed 59.0–176.0 mm), almost hemispherical. Exumbrella finely granulated with small, inconspicuous marks (papillae); exumbrellar color varies from entirely transparent white to white with inconspicuous radial markings. Tentacle number approximately five tentacles per octant, but can be more (average 5.28 ± 0.48) (Table 3; Fig. 8A); lappets rounded typically 48 or more (average 6.26 ± 0.46 per octant); tentacle clefts of varied depth with primary clefts deeper than secondary clefts. Radial and ring musculature not obvious. Brachial disc circular. Pillars evident. No quadralinga. Subgenital ostia rounded, approximately 1/8 of bell diameter. Oral arms v-shaped with frills emanating from tube-like structure; straight without spiral; curved, frilled edges taper toward distal end of oral arms. Oral arms short, approximately the same length as bell diameter (average 1.24 ± 0.27 times bell diameter). Oral arms typically transparent white. Four semi-circular gonads, white, pinkish, or slightly orange, well developed within pouch outlining gastric filaments. About 16 stomach pouches bounded by 16 septae. Septae bent at 45° angle distally toward the rhopalia terminating near tentacle in rhopalar lappet, resulting in tentacular pouches being somewhat larger than rhopalar pouches distally.

Cnidome (tentacle): Average dimensions (length ± 95% CI × width ± 95% CI)

• Holotrichous A-isorhizas: 20.15 ± 0.33 × 11.13 ± 0.24 μm;

• Holotrichous a-isorhizas: 8.27 ± 0.49 × 4.22 ± 0.07 μm;

• Holotrichous O-isorhizas: 21.63 ± 0.39 × 18.91 ± 0.78 μm;

• Heterotrichous microbasic rhopaloids: 13.58 ± 0.19 × 8.09 ± 0.09 μm;

Type locality: Nantucket Bay, Nantucket Island, Massachusetts, East Coast of USA.

Habitat: Medusae are found in open coastal waters on the U.S. Atlantic coast.

Distribution: Western North Atlantic, east coast of the USA south of Cape Cod in coastal Atlantic waters at least as far south as Georgia/Northern Florida.

DNA sequence: Mitochondrial COI and 16S and nuclear 28S sequence data are available in NCBI GenBank under accession numbers MF141552–MF141556, MF141608, MF141613–MF141614, MF141628, MF141635, MF141642–MF141646, MF141688–MF141689, MF141697.

Phylogeny: Chrysaora quinquecirrha and Chrysaora chesapeakei sequences form reciprocally monophyletic groups for 16S, COI, 28S, and combined analyses (Figs. 4–7). Minimum sequence divergences between Chrysaora quinquecirrha and Chrysaora chesapeakei clades (COI: 12.1%, 16S: 8.4%, 28S: 2.4%) were much larger than the maximum within clades for Chrysaora quinquecirrha (COI: 0.2%, 16S: 0.1%, 28S: 0.0%) or Chrysaora chesapeakei (COI: 0.7%, 16S: 0.6%, 28S: 0.4%). Chrysaora quinquecirrha sequences did not form monophyletic groups with any other species (Figs. 4–7).

Biological data: Although the name Chrysaora quinquecirrha applies to the U.S. coastal Atlantic species, almost no ecological studies have been done on the coastal species, apart from (Kraeuter & Setzler, 1975), which found the largest Chrysaora quinquecirrha individual was found in a coastal area about 90 km offshore in full seawater (salinity >30).

Notes: Since this species retains the scientific name Chrysaora quinquecirrha, we advocate it retaining the common name “U.S. Atlantic sea nettle”, since it is also a coastal and open ocean species.

• Chrysaora chesapeakei Papenfuss, 1936

• Figs. 3C–3E and 4–9; Figs. S1 and S2

Dactylometra quinquecirrha: Bigelow (1880): 66 [white colored morph, Chesapeake Bay]. Brooks (1882): 137 [Chesapeake Bay—USA]. Agassiz & Mayer (1898): 48–49 [upper Narragansett Bay (RI)]. Mayer (1910): 585–588, Pl.63–64 [24 tentacle morph, white, red/brown striped morph, Tampa Bay (FL), Hampton Roads (VA), St. Mary’s (MD)]. Papenfuss (1936): 14–17, Figures 7, 11, 16, 20 [lower Chesapeake Bay; red-striped morph based on A-isorhiza measurements]. Littleford & Truitt (1937): 91 [Chesapeake Bay]. Littleford (1939): 368–381, Pls. I–III [Chesapeake Bay]. Hedgepeth (1954): 277–278 [Tampa Bay (FL), Gulf of Mexico].

Dactylometra quinquecirrha var. chesapeakei: Papenfuss (1936): 14–17, Figures 12, 21 [Chesapeake Bay; white colored morph based on A-isorhiza measurements].

Chrysaora quinquecirrha: Kramp (1961): 327–328 [parts of description covers both Chrysaora quinquecirrha and Chrysaora chesapeakei]. Rice & Powell (1970): 180–186 [Chesapeake Bay]. Burke (1976): 20, 22–28 [Mississippi Sound, Gulf of Mexico]. Calder (1971): 270–274 [Gloucester Point (VA)—Chesapeake Bay]. Calder (1972): 40–43, Figures 1–4 [Chesapeake Bay, Pamlico Sound, Gulf of Mexico]. Loeb (1972): 279–291 [Chesapeake Bay]. Loeb (1973): 144–147 [Chesapeake Bay]. Loeb & Blanquet (1973): 150–157 [Chesapeake Bay]. Calder (1974b): 326–333 [Chesapeake Bay]. Loeb (1974): 423–432 [Chesapeake Bay]. Blanquet & Wetzel (1975): 181–192 [Chesapeake Bay]. Cargo (1975): 145–154 [Chesapeake Bay]. Kraeuter & Setzler (1975): 69, Figures 1–2 [Doboy Sound (GA)]. Loeb & Gordon (1975): 37–41 [Chesapeake Bay]. Lin & Zubkoff (1976): 37–41 [Chesapeake Bay]. Calder (1977): 13–19 [Gloucester Point, MD—Chesapeake Bay]. Clifford & Cargo (1978): 58–60 [Patuxent River, MD—Chesapeake Bay]. Cargo (1979): 279–286 [Chesapeake Bay]. Cargo & Rabenold (1980): 20–26 [Patuxent River (MD)]. Hutton et al. (1986): 154–155 [Chesapeake Bay]. Cargo & King (1990): 486–491 [Chesapeake Bay]. Purcell et al. (1991): 103–111 [Choptank River, MD—Chesapeake Bay]. Nemazie, Purcell & Glibert (1993): 451–458 [Chesapeake Bay]. Purcell, White & Roman (1994): 263–278 [Chesapeake Bay]. Burnett et al. (1996): 1377–1383 [Chesapeake Bay]; Houck et al. (1996): 771–778 [St. Margaret’s, MD—Chesapeake Bay]. Olesen, Purcell & Stoecker (1996): 149–158 [Broad Creek (MD)—Chesapeake Bay]. Ford et al. (1997): 355–361 (Choptank River (MD)—Chesapeake Bay]. Kreps, Purcell & Heidelberg (1997): 441–446 [Choptank River (MD)—Chesapeake Bay]. Wright & Purcell (1997): 332–338 [Patuxent River (MD)—Chesapeake Bay]. Suchman & Sullivan (1998): 237–244 [Green Hill Pond (RI)]. Purcell, Malej & Benović (1999): 241–263 [Chesapeake Bay]. Purcell et al. (1999): 187–196 [Choptank River (MD)—Chesapeake Bay]. Bloom, Radwan & Burnett (2001): 75–90 [St. Mary’s (MD)—Chesapeake Bay]. Condon, Decker & Purcell (2001): 89–95 [Choptank River (MD)—Chesapeake Bay]. Graham (2001): 97–111 [Gulf of Mexico]. Johnson, Perry & Burke (2001): 213–221 [Gulf of Mexico]. Matanoski, Hood & Purcell (2001): 191–200 [Choptank River (MD)—Chesapeake Bay]. Segura-Puertas, Suárez-Morales & Celis (2003): 9 [Gulf of Mexico]. Ishikawa et al. (2004): 895–899 [Gibson Island (MD)—Chesapeake Bay]. Grove & Breitburg (2005): 185–198 [Patuxent River (MD)—Chesapeake Bay]. Purcell & Decker (2005): 376–385 [Chesapeake Bay]. Thuesen et al. (2005): 2475–2482 [Chesapeake Bay]. Breitburg & Fulford (2006): 776–784 [Solomon’s Island [MD]—Chesapeake Bay]. Kimmel, Roman & Zhang (2006): 131–141 [mid to upper Chesapeake Bay]. Decker et al. (2007): 99–113 [Chesapeake Bay]. Condon & Steinberg (2008): 153–168 [York River (VA)—Chesapeake Bay]. Calder (2009): 24–28 [estuarine animals]. Matanoski & Hood (2006): 595–608 [Choptank River (MD)—Chesapeake Bay]. Purcell (2007): 184, 190–192 [Chesapeake Bay]. Purcell (2009): 23–50 [Chesapeake Bay]. Duffy, Epifanio & Fuiman (1997): 123–131 [Port Aransas (TX)—Gulf of Mexico]. Bayha & Graham (2009): 217–228 [Rhode Island, New Jersey, Chesapeake Bay, Georgia, Alabama]. Sexton et al. (2010): 125–133 [Choptank River (MD)—Chesapeake Bay]. Birsa, Verity & Lee (2010): 426–430 [Skidaway River (GA), Wassow Sound (GA)]. Condon, Steinberg & Bronk (2010): 153–170 [York River (VA)—Chesapeake Bay]. Condon et al. (2011): 10225–10230 [Chesapeake Bay]. Frost et al. (2012): 247–256 [Steinhatchee River (FL)—Gulf of Mexico]. Duarte et al. (2012): 91–97 [St. Leonard’s (MD)—Chesapeake Bay]. Kimmel, Boynton & Roman (2012): 76–85 [Solomon’s Island (MD)—Chesapeake Bay]. Sexton (2012): 1–153 [Chesapeake Bay]. Brown et al. (2013): 113–125 [Chesapeake Bay]. Robinson & Graham (2013): 235–253 [Gulf of Mexico]. Breitburg & Burrell (2014): 183–200 [Patuxent River (MD)—Chesapeake Bay]. Kaneshiro-Pineiro & Kimmel (2015): 1965–1975 [Pamlico Sound (NC). Meredith, Gaynor & Bologna (2016): 6248–6266 [Barnegat Bay (NJ)]. Tay & Hood (2017): 227–242 [Choptank River (MD), Chesapeake Bay].

Diagnosis: Living medusae up to 20 cm (observed 17.0–175.0 mm; average: 63.0 mm); tentacles typically number 24 or 3 per octant (average 3.07 ± 0.07); primary tentacle central and secondary tentacles lateral (2-1-2); rarely, additional tentacles arise lateral to secondary tentacles (3-2-1-2-3) and are typically undeveloped; marginal lappets rounded and typically 32 or 4 per octant (average 4.08 ± 0.06); rhopalar lappets are typically about the same size as tentacular lappets; can be differentiated from Chrysaora quinquecirrha based on (1) larger size of holotrichous A-isorhiza nematocysts: 26.21 [±0.50] × 19.74 [±0.55] μm; (2) smaller tentacle number (∼3 tentacles per octant); and (3) larger maximum oral arm length (average: 3.00 ± 0.39 times bell diameter).

Material examined: Neotype: USNM 1454948—(Gloucester Point, MD—Chesapeake Bay). Other comparative specimens: USNM 57925 (n = 9; Orange Inlet, NC), USNM 56758 (n = 5; Charlestown Pond, RI), USNM 33456 (n = 4; Plum Point, MD), USNM 49733 (n = 1; Alligator Harbor, FL), USNM 53826 (n = 2; Timbalier Bay, LA), USNM 56703 (n = 2; Chesapeake Bay 37.23 N 76.04 W), USNM 56704 (n = 4; Chesapeake Bay 37.23 N 76.04 W), USNM 53870 (n = 3; Beaufort, NC), USNM 53828 (n = 2; Drum Point, MD), USNM 33458 (n = 3; Plum Point, MD), USNM 33457 (n = 4; Plum Point, MD), USNM 55621 (n = 6; near Chesapeake Beach, MD), USNM 53867 (n = 1; Arundel on the Bay, MD), USNM 54404 (n = 1; Chesapeake Bay 37.23 N 76.04 W), USNM 33121 (n = 6; Arundel on the Bay, MD), USNM 42155 (n = 2; Louisiana, Gulf of Mexico), USNM 54372 (n = 1; Lake Pontchartrain, LA); USNM 1454941–USNM 1454943, KMBCSC1, KMBCSC4–KMBCSC5, KMBCSC7 (n = 7; Charleston Harbor, SC), USNM 1454944–USNM 1454951, KMBGVA1, KMBGVA5, KMBGVA7, KMBGVA10 (n = 12; Gloucester Point, VA), KMBCRI1–KMBCRI14 (n = 14; Charlestown Pond, RI), KMBRDE1–KMBRDE16 (n = 16; Rehoboth Bay, DE), USNM 1454956, KMBDIA2–KMBDIA3 (n = 3; Dauphin Island, AL).

Description of neotype specimen: USNM 1454948. Bell diameter 110.4 mm, almost hemispherical. Exumbrella white/clear with granulated surface of small white marks. Eight rhopalia. No ocelli. Deep rhopalar clefts; deep sensory pits. Marginal lappets rounded, 32 total or 4 per octant made up of two rhopalar lappets and two tentacular lappets. Lappet size barely heterogeneous, with rhopalar lappets about the same width as tentacular lappets but longer. Tentacle number 24 or 3 per octant, with primary tentacle surrounded by two secondary tentacles (2-1-2), primary tentacle longer than secondary tentacles, up to 3–4 times bell diameter. Tentacles are white, slightly pinkish. Tentacle clefts of varied depth with primary clefts deeper than secondary clefts. Radial and ring musculature not obvious. Brachial disc circular. Pillars evident. No quadralinga. Subgenital ostia rounded, approximately 1/10 of bell diameter. Oral arms white, v-shaped with frills emanating from tube-like structure. Oral arms straight without spiral curved, frilled edges taper toward distal end of oral arms. Orals arms long, approximately five (4.98) times bell diameter. Four semi-circular gonads, white (a bit orange), well developed within pouch outlining gastric filaments. About 16 stomach pouches bounded by 16 septae. Septae bent at 45° angle distally toward the rhopalia terminating near tentacle in rhopalar lappet, resulting in tentacular pouches being somewhat larger than rhopalar pouches distally.

Cnidome (tentacle): Average dimensions (length ± 95% CI × width ± 95% CI)

• Holotrichous A-isorhizas: 25.66 ± 0.83 × 19.16 ± 0.54 μm;

• Holotrichous a-isorhizas: 7.77 ± 0.20 × 4.17 ± 0.10 μm;

• Holotrichous O-isorhizas: 22.02 ± 0.30 × 19.95 ± 0.24 μm;

• Heterotrichous microbasic rhopaloids: 12.35 ± 0.47 × 8.55 ± 0.55 μm.

Description of other specimens: Bell diameter up to approximately 20 cm (observed 17.0–175.0 mm), almost hemispherical but flattened in small individuals. Exumbrella finely granulated with small, inconspicuous marks (papillae); exumbrellar color varies considerably, varying from all white to a completely brown or red colored bell, to a bell with radial lines of red/brown with a spot in the center of the bell. Radial lines may be relatively inconspicuous without a noticeable spot in the center. Tentacles typically number 24 or 3 per octant (average 3.07 ± 0.07), with primary tentacle surrounded by two secondary tentacles (2-1-2), primary tentacle longer than secondary tentacles, up to 3–4 times bell diameter. In some rare cases, small tentacles may occur laterally to secondary tentacle, occurring between the secondary tentacle and rhopalium. In almost all cases, this tentacle is similar in size to or smaller than the lappets surrounding it. In very rare cases (twice observed), about five or more tentacles per octant have been seen, though these medusae had aberrant tentacle patterns overall (e.g., more than one tentacle emerging from same spot, tentacles emerging below lappet). Tentacles are white, slightly pinkish. Marginal lappets rounded and typically 32 or 4 per octant (average 4.08 ± 0.06). Tentacle clefts of varied depth with primary clefts deeper than secondary clefts, which are deeper than rare tertiary clefts. Radial and ring musculature not obvious. Brachial disc circular. Pillars evident. No quadralinga. Subgenital ostia rounded, approximately 1/10 of bell diameter. Oral arms v-shaped with frills emanating from tube-like structure; straight without spiral; curved, frilled edges taper toward proximal end of oral arms. Oral arms long, approximately three times bell diameter on average (as much as 5.6 times bell diameter). Oral arms vary in color, from transparent white, to red or brown colored tubule surrounded by pinkish frilled edges. Four semi-circular gonads, white, pinkish or slightly orange, well developed within pouch outlining gastric filaments. About 16 stomach pouches bounded by 16 septae. Septae bent at 45° angle distally toward the rhopalia terminating near tentacle in rhopalar lappet, resulting in tentacular pouches being somewhat larger than rhopalar pouches distally.

Cnidome (tentacle): Average dimensions (length ± 95% CI × width ± 95% CI)

• Holotrichous A-isorhizas: 26.21 ± 0.50 × 19.74 ± 0.55 μm;

• Holotrichous a-isorhizas: 7.88 ± 0.13 × 4.29 ± 0.07 μm;

• Holotrichous O-isorhizas: 23.10 ± 0.43 × 20.75 ± 0.62 μm;

• Heterotrichous microbasic rhopaloids: 12.73 ± 0.22 × 8.29 ± 0.13 μm.

Type locality: Gloucester Point (VA), Chesapeake Bay, east coast of USA.

Habitat: Medusae are found in estuarine waters on the U.S. Atlantic coast and estuarine and nearshore waters of the Gulf of Mexico.

Distribution: Western North Atlantic, east coast of the USA south of New England to the Gulf of Mexico, restricted to estuarine waters on the Atlantic coast, known to exist outside of estuaries in the Gulf of Mexico.

Notes: Since Chrysaora chesapeakei is commonly found in estuarine waters, we advocate the common name “Atlantic bay nettle” to distinguish it from the “U.S. Atlantic sea nettle” (Chrysaora quinquecirrha). The specific name chesapeakei originates from Dactylometra quinquecirrha var. chesapeakei of Papenfuss (1936). For Papenfuss (1936), it is clear that: (1) the manuscript likely compared nematocyst measurements between two color morphs of Chrysaora chesapeakei and did not include Chrysaora quinquecirrha s. str. (see Discussion; Fig. 8C); and (2) differences invoked for holotrichous a-isorhizas are in question, since the nematocysts are small (∼1.5 μm), making identifying differences difficult even with more precise, modern instruments, and the data are not accompanied by any statistics or measurement error. Regardless, based on Article 35.6.4 of the International Code of Zoological Nomenclature 4th Edition (ICZN, 1999), the specific name chesapeakei has taxonomic priority and Chrysaora chesapeakei applies to the Chesapeake Bay animals, as well as estuarine Atlantic and Gulf of Mexico animals that are genetically similar, and have similar macromorphological and cnidome characteristics (Figs. 4–9). Papenfuss (1936) did not designate a type specimen for Dactylometra (=Chrysaora) quinquecirrha var. chesapeakei. We designate the specimen USNM 1454948 as a neotype specimen so that a physical specimen, along with preserved tissue for genetic analysis, will be available to objectively define Chrysaora chesapeakei [see Article 75 of the International Code for Zoological Nomenclature (ICZN, 1999)], which will be necessary given the close genetic relationship between this species and specimens from the Caribbean (see below). Our neotype specimen originates from Gloucester Bay (VA), within the Chesapeake Bay, where Papenfuss (1936) hypothesized Dactylometra (=Chrysaora) quinquecirrha var. chesapeakei to be confined.

DNA sequence: Mitochondrial COI and 16S and nuclear 28S sequence data are available in GenBank under accession numbers MF141564–MF141587, MF141615–MF141617, MF141637–MF141639, MF141649–MF141669, MF141699–MF141718, MF167556–MF167568.

Phylogeny: Chrysaora chesapeakei and Chrysaora quinquecirrha sequences form reciprocally monophyletic groups for 16S, COI, 28S, and combined analyses (Figs. 4–7). Minimum sequence divergences between Chrysaora chesapeakei and Chrysaora quinquecirrha clades (COI: 12.1%, 16S: 8.4%, 28S: 2.5%) were much larger than the maximum within clades for Chrysaora quinquecirrha (COI: 0.3%, 16S: 0.1%, 28S: 0.0%) or Chrysaora chesapeakei (COI: 2.2%, 16S: 1.9%, 28S: 0.7%). Chrysaora chesapeakei sequences do not form monophyletic groups with any other species (Figs. 4–7).

Supplemental Information