Hippocampuswhitei Bleeker, 1855, a senior synonym of the southern Queensland seahorse H.procerus Kuiter, 2001: molecular and morphological evidence (Teleostei, Syngnathidae)

Abstract The taxonomic status of the seahorse Hippocampusprocerus Kuiter, 2001, type locality Hervey Bay, QLD, Australia, was re-examined based on its strong morphological similarity and geographical proximity to its congener H.whitei Bleeker, 1855, a species recorded in ten estuaries of New South Wales, Australia. Kuiter (2001) distinguished H.procerus from H.whitei by a taller coronet, marginally lower meristics, and spinier physiognomy. Meristic, morphometric, and key diagnostic morphological character comparisons from vouchered specimens of the two purported species collected from Sydney Harbour, Nelson Bay, Port Stephens, NSW and Hervey Bay, Bundaberg, and Moreton Bay, QLD did not show diagnostic differences to support species-level classification of H.procerus. Furthermore, partial mitochondrial COI sequence data from specimens sampled from known geographical distributions in NSW and Southport, QLD failed to discriminate between populations as a result of shared haplotypes, and revealed an average intraspecific divergence of 0.002%. Hippocampusprocerus is hereby placed in the synonymy of H.whitei; a redescription is provided, with a revised record of its range across eastern Australia.

type specimens of H. whitei corresponded closely with the examined non-type specimens, paratypes, and the holotype specimen of H. procerus, including: coronet height, absence of neck spines, indiscernible or small parietal spine, the numbers and positions of cleithral ring and subdorsal ridge spines, and overall spine physiognomy. Partial mitochondrial COI sequence data generated from specimens sampled from known geographical distributions in NSW and from Southport, QLD failed to discriminate between populations as a result of shared haplotypes, and revealed an average intraspecific divergence of 0.002%. Hippocampus whitei Bleeker, 1855, is herein formally redescribed as a senior synonym of H. procerus. This estuarine species is apparently endemic to estuaries of central NSW, the Tweed River, and southern QLD.

Materials and methods
Four individuals referred to as H. procerus, based on known locality of this species (Kuiter 2001), were collected from Southport, Gold Coast Harbour, QLD in 2014 by seine in seagrass beds in 1-2 m depth or by hand nets while scuba diving in less than 8 m depth (Figure 1), from which tissue was sampled from the caudal tip of the tails and preserved in a NaCl-saturated DMSO solution for genetic analyses. Similarly, thirtyone individuals of H. whitei were tissue sampled from the caudal tip of the tail at seven localities along the species' known geographic range in New South Wales, Australia (Table 1, Figure 1) from [2007][2008][2009]. DNA extraction, PCR amplification, alignment,  and analysis of partial mitochondrial cytochrome c oxidase subunit I (COI) sequences was performed following standard protocols described in Hamilton et al. (2017). Genetic distances (uncorrected p-distances) were calculated and neighbour-joining (NJ) trees constructed with confidence levels assessed using 1000 bootstrap replications based on partial COI using MEGA v. 7.0.26 (Kumar et al. 2017). Proportional measurements and counts based on eight morphometric and six meristic variables (Tables 2, 3), including 17 diagnostic morphological characters, were per- formed on dried or ethanol-preserved specimens and high-resolution digital images of specimens using ImageJ (Rasband et al. 1997) to the nearest 0.1 mm following Lourie and Randall (2003) and Lourie and Kuiter (2008). External morphological characters were documented using a dissecting microscope or on high-resolution digital images of specimens. The holotype specimen of H. whitei is unknown. The original description of H. whitei Bleeker, 1855 is based on an artistic and non-informative rendering (White 1790: 264, plate 50) from Sydney Harbour, NSW. Morphometric measurements were recorded for two non-type specimens of H. whitei from Nelson Bay, NSW, two non-type specimens of H. procerus from Southport, Gold Coast Harbour, QLD, two non-type specimens of H. procerus from Moreton Bay, QLD, and one non-type specimen of H. procerus from Bundaberg and Mackay, QLD, respectively (Table 2). These data were compared to morphometric data from the holotype specimen of H. procerus from Hervey Bay, QLD (Table 2). Meristic counts and diagnostic morphological characters were recorded for 12 non-type specimens of H. whitei, including two from Sydney Harbour, six from Pittwater, and four from Nelson Bay, NSW, and 13 type and non-type specimens of H. procerus, including four from Southport, Gold Coast Harbour, QLD, two from Moreton Bay, one from Hervey Bay, QLD, one from Elliot Bay, four from Bundaberg, and one from Mackay, QLD (Table 3). These data were compared with similar morphological data (Table 3) for the holotype specimen of H. procerus from Hervey Bay, QLD.   Diagnosis. Hippocampus whitei differs from its congeners by the following combination of characters: trunk rings 11; tail rings 34-35; dorsal fin rays 17-18; pectoral fin rays 16; subdorsal rings three; subdorsal spines four, superior trunk ridge ending with three enlarged spines, superior tail ridge commencing with one enlarged spine (3/0,1,0); cleithral ring spines three, one small spine at each end of pectoral-fin base but none at gill-opening, large single or double spine at ventral extent of head; small lateral head spines, two, one directly posterior of eye, one anterodorsally of operculum and ventral of coronet; distinct snout spine; parietal spine, diminutive or absent; single eye spine, large, protruding dorsally; small single or double spine, rugose, posteroventrally of eye; coronet, distinct and tall, protruding anteriorly in juveniles, angled dorsoposteriorly in adults, five small spines present on apex in a star-like arrangement; superior trunk with enlarged spines on 1 st and 8 th tail ridges.
Redescription. General body shape as in Figs 2-11. Morphometric and meristic characters are listed in Table 2. Coronet distinct and tall, coronet height 44.8-47.89% in HL, protruding anteriad in juveniles, angled dorsoposteriorly in adults; bilateral gill-openings ventral of coronet; dorsal fin rays 17-18; pectoral fin rays 16; subdorsal rings three; dorsal fin base starting immediately posterior to ninth trunk ring and ending immediately posterior to first tail ring; trunk rings 11; tail rings 34-35. Body spines: coronet with five small spines present on apex in a star-like arrangement; neck spines absent; prominent spine dorsally of eye, small single or double spine, rugose, ventroposteriorly of eye; small lateral head spines, two, one directly posterior of eye,  one anterodorsally of operculum and ventral of coronet; cleithral ring spines three, one small spine at each end of pectoral-fin base but none at gill-opening, large single or double spine at ventral extent of head; distinct snout spine on midline between eyes; parietal spine, diminutive or absent in adults, present in juveniles and subadults;  subdorsal spines four, superior trunk ridge ending with three enlarged spines, superior tail ridge commencing with one enlarged spine (3/0,1,0); superior trunk with enlarged spines on 1 st and 8 th tail ridges observed in adults, on 1 st , 7 th , 8 th , and 9 th tail ridges observed in subadults; lateral trunk ridge with small spines on 2 nd -11 th trunk rings; inferior trunk ridge with small spines beginning on 5 th trunk ring and ending on 11 th trunk ring; superior tail ridge spines well developed anteriorly, becoming smaller posteriorly, with enlarged spines on 1 st -12 th tail rings; inferior tail ridge spines well developed anteriorly, becoming smaller posteriorly, with enlarged spines on 1 st -8 th tail rings; caudal fin absent in juveniles and adults. Morphological remarks. In his original description, Kuiter (2001:328-329) erected H. procerus based on several observations on its distinguishing characters: "Previously confused with Hippocampus tristis and H. whitei, H. procerus is more similar to the latter, differing from it in having a taller and spinier coronet, higher fin-ray       Based on the material examined, we found minor variation in coronet height in proportion to the head (45.5-46.6% in H. whitei from Nelson Bay, NSW vs. 45.1-47.8% in H. procerus from Gold Coast Harbour, QLD, 48.9-52.9% in the paratypes from Moreton Bay, QLD, 44.8% in the holotype from Hervey Bay, QLD, 50.8% in the paratype from Bundaberg, QLD and 46.7% from Mackay, QLD). The nontype specimens are comprised of juveniles, subadults, and adults, all of which exhibit distinct and tall coronets. However, we noted that in juveniles the coronet protrudes anteriad whereas in subadults and adults it is strongly angled dorsoposteriad. Similarly, dorsal fin ray counts exhibited marginal differences (17 in non-type specimens of H. whitei vs. 18 in all the specimens of H. procerus from Queensland), which may  reflect north-south clinal variation. We did not observe an overall spinier physiognomy in the majority of adult specimens of H. procerus relative to H. whitei. However, a spinier physiognomy was present in one juvenile specimen from Mackay, and Port Curtis, QLD (Figs 10, 11), respectively, and one adult specimen from Burnett River (Fig. 9) and Waddy Point, QLD (Fig. 12), respectively, on all principal trunk and tail ridges and head. It has been observed that in juvenile and subadult H. whitei from NSW (<8 cm Total Length) that spines are more pronounced, but as they increase in size the spines disappear, with specimens > 12 cm TL being much smoother and spines not obvious. The adult specimens from Burnett River and Waddy Point, QLD, are an exception to these observations and appear to reflect variation in spine morphology similarly observed in juvenile H. whitei.
We also observed across the majority of examined adult specimens the following key diagnostic morphological characters (Table 3): the absence of true neck spines between the coronet and the 1 st superior trunk ring spines (small neck spines were detected in juvenile specimens; variation in neck ridge topology is often mistaken for true neck spines in adult specimens), indiscernible or small parietal spine, three cleithral ring spines with the uppermost spine at the dorsal level of the pectoral fin base, presence of a distinct snout spine, four subdorsal ridge spines (3/0,1,0), and superior trunk ridge with enlarged spines on 1 st and 8 th tail ridges. Based on these findings, we find that spine physiognomy, neck spines present or absent in juveniles and adults, respectively, and subtle differences in meristics, are unreliable diagnostic characters, and that key and informative morphological characters are congruent across all specimens, which conform to the diagnosis of H. whitei. Therefore, it can be concluded that the species-level classification of H. procerus is unsupported, and that H. procerus may be treated as a junior synonym of H. whitei.
Genetic remarks. Meristic, morphometric, and key diagnostic morphological characters in our study did not support the separation of H. procerus from H. whitei into two distinct species. Here we further confirm the synonymization of H. procerus with H. whitei based on partial mitochondrial COI (655 bp) data. This analysis is based on sequences generated from 31 H. whitei individuals sampled from Empire Bay, Forster (Wallis Lake), Port Hacking, Nelson Bay, Sydney Harbour, and Tuggerah Lake, NSW, and from 4 specimens referred to as H. procerus from Southport, Gold Coast Seaway, QLD. Alignment of sequence data detected 23 variable sites without any indels, resulting in 14 haplotypes: one in Sydney, four in Nelson Bay, five in Forster, and four in Gold Coast Seaway (Suppl. material 1, 2). Three haplotypes are shared between Forster and Gold Coast Seaway whereas no other haplotypes are shared between collection localities. One unique haplotype was obtained in Sydney, four in Nelson Bay, two in Forster, and one in Gold Coast Seaway (Suppl. material 1, 2). Fig. 12 shows a neighbour-joining tree based on the same mtDNA COI data, which recovered H. procerus as paraphyletic with respect to H. whitei. Hippocampus procerus clustered among individuals of H. whitei from Forster, NSW in one subclade and with individuals from several localities in NSW in another subclade. Additionally, genetic distance analysis (uncorrected p distances) of the same data failed to discriminate H. procerus from H. whitei (Suppl. material 3), which revealed an average intraspecific divergence of 0.002%, further confirming lack of support of species status for H. procerus.
Distribution and habitat. Hippocampus whitei is known to occur in coastal estuaries and embayments of central NSW and southern QLD, Australia. In central NSW it has been recorded, from south to north, in St. Georges Basin, Port Hacking, Botany Bay, Sydney Harbour, Hawkesbury River, Tuggerah Lake, Lake Macquarie, Port Stephens, Wallis Lake -Forster, and Tweed River. The record from St Georges Basin was based on a recent sighting and photograph of a small juvenile in January 2018 that was logged through REDMAP (http://www.redmap.org.au/sightings/3379/) and therefore extends the range reported by Harasti et al. (2012) southwards by 70 km. A previous 1903 Australian Museum record from Lake Illawarra cannot be confirmed as the local-ity information is likely erroneous (Mark McGrouther, pers. comm.) and whilst it is possible that H. whitei could occur in Lake Illawarra, at this stage there is no definitive evidence. Museum records indicate the species has been recorded in QLD within the Gold Coast Seaway, at various locations around Moreton Bay estuary, Hervey Bay, Waddy Point, Elliot Heads, Bundaberg, Port Curtis, and Mackay. The synonymization of H. procerus extends the northward range significantly by approximately 1,000 kilometres for H. whitei to Mackay QLD; as of now, this is the most northern location with confirmed H. whitei specimens. However, its current occurrence in the Mackay region, Port Curtis, Burnett River, and Bundaberg, is unknown as it has not been recorded in those locales since 1939, 1929, and 1938, respectively. The most recent northern records are from Elliot Heads in 1982 and Waddy Point in 2005.
Additionally, museum records claim species occurrences of H. whitei outside its geographic range, in South Australia, Victoria Australia, Papua New Guinea, South Africa, Solomon Islands, and Vanuatu (Kuiter 2009;Lourie et al. 2016). However, these specimens have subsequently been re-identified by the authors as H. breviceps, H. camelopardalis, H. kelloggi, or H. spinosissimus (see Table 6). The specimens originally identified as H. whitei from Port Moresby, Papua New Guinea and the Solomon Islands are no longer accessible and therefore cannot be re-identified; however, we consider them highly unlikely to be H. whitei since it is markedly outside the range for this species. We now consider that the species is constrained to estuaries and embayments along the east coast of Australia from Hervey Bay, QLD, in the north to St Georges Basin, NSW, in the south.
Hippocampus whitei occurs in a variety of habitats including seagrasses, soft corals, sponge gardens and artificial structures to depths of 12 m (Hellyer et al. 2011;Harasti et al. 2014a;Manning et al. 2018), and is known to display strong site fidelity and monogamous behaviour (Vincent and Sadler 1995;Vincent et al. 2005;  Gladstone 2013). The locations with the largest recorded populations are found within Sydney Harbour and Port Stephens (Harasti et al. 2012;Harasti et al. 2014b), beyond which there is very little information about the occurrence, habitat use, and population numbers in QLD as the species is not known to be regularly found in any QLD locations and is seldom seen or collected.
We introduced this paper with a quote from John White (1736-1832) who was under the assumption that the Mediterranean and North Atlantic seahorse H. hippocampus and H. whitei from Australia were conspecific due to highly similar morphology. Seahorse taxonomy has been in a state of confusion since its inception. While comprehensive revisions of the genus have greatly advanced our understanding of how many species of seahorses exist (Lourie et al. 2016), much further work remains to answer this most fundamental question about one of the world's most extraordinary fish.