The family Serranidae is represented by 92 genera and 579 valid species, with the genus Serranus Cuvier, 1816, containing 30 species. In this study, specimens of Butterfly-winged Comber, Serranus papilionaceus Valenciennes, 1832, were collected from the Canary Islands and compared morphologically and genetically to Painted Comber, Serranus scriba (Linnaeus, 1758), from the Mediterranean Sea. Morphological differences, especially in the colour banding pattern, were corroborated by genetic differences in mitochondrial (COI and ND2) and nuclear (Rhod and PTR) markers. The mitochondrial DNA markers revealed a high level of divergence and no shared haplotypes between the two species (interspecific divergence: COI 4.31%; ND2 8.68%), and a phylogenetic analysis showed that these two species are closely related sister species sharing common ancestry. This study is therefore offering to resurrect S. papilionaceus Valenciennes, 1832 as a valid species increasing the number of eastern Atlantic Serranus species to 11. This should direct new species-specific research, including its population conservation status assessment across its distribution.

Butterfly-winged Comber, Canary Islands, Mediterranean Sea, Painted Comber, phylogeny, taxonomy

The family Serranidae, which is composed of sea basses and groupers, has three subfamilies Anthiadinae, Epinepheline, and Serraninae, which are represented by 579 valid species in 72 genera (Parenti and Randall 2020). The genus Serranus (Perciformes, Serranidae) has 30 small reef-associated species (Iwamoto and Wirtz 2018; Horton et al. 2020), with the most recent additions to the genus Serranus being S. aliceae Carvalho Filho & Ferreira, 2013 from the Southwest Atlantic Ocean and S. pulcher Wirtz & Iwamoto, 2016, S. drewesi Iwamoto, 2018, and S. inexpectatus Wirtz & Iwamoto, 2018 from the Eastern Central Atlantic Ocean (Carvalho Filho and Ferreira 2013; Wirtz and Iwamoto 2016; Iwamoto and Wirtz 2018).

Most of the species belonging to this genus are recorded in the coastal regions of the Atlantic Ocean, with 14 species occurring in the Western Atlantic and 10 species within the Eastern Atlantic. In addition to these, another six species inhabit the Eastern Pacific Ocean while another two occur in the Indian Ocean (Iwamoto and Wirtz 2018). Both the Mediterranean Sea and the Canary Islands are known to host the same four Serranus species (Dooley et al. 1985; Bauchot 1987): the Blacktail Comber, S. atricauda Günther, 1874, which in the Mediterranean is restricted to the western basin; the Comber, S. cabrilla (Linnaeus, 1758); the Brown Comber, S. hepatus (Linnaeus, 1758); and the Painted Comber, S. scriba (Linnaeus, 1758). However, analysing photos of specimens commonly identified as S. scriba collected off the Canary Islands, we noticed distinct differences in colour pattern when compared to the specimens from the Mediterranean Sea, including Malta (Fig. 1), as was already observed by Iwamoto and Wirtz (2018). These notable differences led us to further investigate the species’ identity using morphological and genetic analyses, while reviewing first descriptions and museum specimens to accurately identify the species in question.

Based on the currently used literature and morphological keys (Bauchot 1987; Lloris and Rucabado 1998; Iwamoto and Wirtz 2018), all specimens collected and analysed during this study (Suppl. material 1: Table S1) fit the body shape description of S. scriba sensu lato; however, genetic data revealed divergence between the two forms. Based on the type specimens and descriptions evaluated in this study, we assigned the specimens collected from Malta to S. scriba in its traditional usage based on Cuvier and Valenciennes (1828). These specimens matched the museum specimens NRM 442 and MNHN-IC-0000-7129 (Suppl. material 1: Figs S1a, S2b). We assign the specimens collected from the Canary Islands to the resurrected taxon S. papilionaceus, as described by Valenciennes (1832) and matching the syntype specimens MNHN-IC-0000-7279 (Suppl. material 1: Fig. S2a) and museum specimens MNHN-IC-1896-0389 and MNHN-IC-1999-1053.

Morphology

Meristic counts and measures

Meristic counts (Table 2) indicated a general overlap between the counts of the two species investigated, with only the anal soft fin rays counts showing a difference (7 in S. scriba and 8 in S. papilionaceus). Though the meristic measures (Table 3) overlapped between the two species, the specimens investigated also showed a longer average dorsal fin length in S. papilionaceus than in S. scriba (46.3% TL ± 1.1 and 41.4% TL ± 1.4, respectively). Extended morphometric studies on S. papilionaceus covering a wider distribution and life-stages are required to corroborate the significance of these differences. Consequently, the most important identifying morphological characteristic feature between the two species would be their body colouration patterns described below.

Table 2.

Download as

CSV

XLSX

Meristic counts for Serranus scriba, its synonyms¹ (Parenti and Randall 2020) and Serranus papilionaceus. Values in brackets represent the mean and SD for counts that have a range. ² Questionable locality (Fernholm and Wheller 1983).

Species name	References (sample size)	Meristic counts					Locality
		Dorsal	Pectoral	Ventral	Anal	Caudal
Perca scriba	Linnaeus, 1758	10 / 15	13	1 / 5	3 / 7	15	America (Linnaeus, 1764)2
Perca marina 1	Brünnich, 1768	10 / 16	13	1 / 5	3 / 8	15	Mediterranean
Holocentrus fasciatus 1	Bloch, 1790	10 / 15	13	1 / 5	3 / 7	16
Holocentrus marocannus 1	Bloch & Schneider, 1801	10 / 16	15	1 / 5	3 / 7	18	Morocco
Serranus scriba	Cuvier & Valenciennes, 1828	10 / 14	13	1 / 5	3 / 7	17	Mediterranean: France, Malta, Italy and Egypt
Serranus scriba	Zorica et al. 2010	10 / 14–17	12–16	1 / 4–6	3 / 7–8	15–18	Trogir, Turkey
	(n = 253)	(15.01 ±2.04)	(14.10 ±1.15)	(4.97 ±0.49)	(7.03 ±0.82)	(16.59 ±1.36)
Serranus scriba	current work	10 / 15	13–14	1 / 5	3 / 7	15–17	Malta
	(n = 15)		(13.27 ±0.46)			(16.67 ±0.62)
Serranus papilionaceus	Valenciennes, 1832	10 / 15	16	1 / 5	3 / 8	17	Gorée, Senegal
Serranus papilionaceus	current work	10 / 15–16	14–15	1 / 5	3 / 8	17	Canary Islands
	(n = 5)	(15.40 ±0.55)	(14.60 ±0.55)

Colour patterns

The S. scriba (Fig. 1) specimens analysed in this study had nine dorso-ventral brown bands. The first band was observed to originate dorsally over the head and fade ventrally just behind the preopercular area. The second band, starting in front of the origin of the dorsal fin, was found to touch the back of the operculum and end at the origin of the pectoral fin. The third, shorter band originating at the first hard spine of the dorsal fin, fades away ventrally. The fourth and fifth band originating at the membrane between the fourth and the seventh hard spine of the dorsal fin and fade away ventrally. They are usually paired up and at times are nearly overlapping. The sixth and seventh band originating at the membrane between the eighth hard spine and the first soft spine of the dorsal fin, may at times be fused, and fade away ventrally into a bluish-violet blotch located over the abdomen area. The eighth and ninth bands are much wider than the former bands, appearing as two fused bands, especially when they fork out ventrally. These two wider bands originate from the soft spines of the dorsal fin and fade away ventrally as they approach the anal fin. A brown banding pattern on the tail peduncle is not always present. The pectoral fins, pelvic fins, anal fin, and tail peduncle together with the tail, are brownish-orange. A faint bluish colouration on the outer side of the pelvic fins and anal fin membrane is present. The pelvic fins, anal fin, dorsal fin, tail peduncle, and tail have red-orange spots, while at the tip of each hard spine of the dorsal fin there is a small, red-orange-coloured membrane. Yellowish thin vertical striations are visible on the abdominal area but quickly fade away ventrally. The head area is reddish-brown with vermiculation. The head area below the eye is light coloured, while the area above the eye has a darker red-brown colouration. A longitudinal brown stripe runs through the eye area.

The banding pattern of the S. papilionaceus (Fig. 1) specimens examined in this study was composed of two wide dorso-ventral brown bands, at times each appearing as composed of multiple fused bands. The first anterior brown band, occurring between the head area and the membrane of the eighth hard spine of the dorsal fin, fades ventrally at the abdomen and was found to be longer than the bands noted in S. scriba. This brown band is followed by a lighter cream-coloured band originating between the last few hard spines and the first few soft spines of the dorsal fin. This light-coloured band ends with a blue-violet patch at the abdomen. The latter was not present in all specimens we examined, and when present, it was less conspicuous than that found on S. scriba. This light-coloured band is followed by another dorso-ventral, wide, brown band originating from the soft spines of the dorsal fin and which appears as being composed of multiple fused bands. A brown banding pattern is also visible on the tail peduncle. The pectoral fins, pelvic fins, and tail, are brownish-orange. The anal fin and the membrane of the soft-rays of the dorsal fin have red-orange spots that are more pronounced than those of S. scriba and are separated by bluish lines. The tip of each hard spine of the dorsal fin has a small, red-orange-coloured membrane. The head area of S. papilionaceus is brownish as opposed to the red-brown of S. scriba. In S. papilionaceus, the head area below the eye is lighter in colour than the dorsal area of the head. The vermiculation on the head is more pronounced in S. papilionaceus than in S. scriba, and as for the former species, the lighter patterns contrast more against the darker brown colouration. In S. papilionaceus, the brown longitudinal stripe that runs through the eye area is not as conspicuous as that of S. scriba.

Most of the vivid colourations noted on both S. scriba and S. papilionaceus are mostly visible on live specimens, and some of the details are lost once the individuals die; however, the main brown bands tend to remain visible for a longer time and remain persistent in some old museum preserved specimens (Suppl. material 1: Fig. S2).

Table 3.

Download as

CSV

XLSX

External measures of Serranus papilionaceus and Serranus scriba expressed as a percentage of the total length.

Meristic measures	Serranus papilionaceus			Serranus scriba			Serranus scriba
	Canary Islands			Malta			Turkey
	current study (n = 5)			current study (n = 15)			Zorica et al. 2010 (n = 253)
	range	mean	SD	range	mean	SD	range	mean	SD
Total length (mm)	187–245	216.6	±26.3	101–205	131.8	±30.9	71–200	110	±17.0
Standard length	82.9–85.7	84.5	±1.2	82.6–88.3	85.3	±1.6	74.7–94.3	84.3	±1.8
Head length	28.9–33.9	31.6	±2.0	31.8–35.6	33.8	±1.1	16.8–40.0	33.5	±1.6
Preocular head length	8.9–10.0	9.4	±0.4	8.0–10.2	9.4	±0.6
Eye diameter	4.8–5.3	5.1	±0.3	4.9–6.8	6.2	±0.6
Postocular head length	14.3–17.6	16.0	±1.5	15.8–18.8	17.1	±0.8
Predorsal distance	28.6–32.9	30.7	±1.6	31.2–35.9	33.7	±1.3	28.9–43.5	33.6	±1.9
Preanal distance	52.4–56.7	55.2	±1.6	53.0–59.2	56.4	±1.9	48.3–64.8	55.3	±1.9
Maximum body height	24.2–25.8	25.1	±0.6	22.9–28.7	25.5	±1.5	16.8–47.1	27.9	±2.6
Minimum body height	10.7–11.7	11.3	±0.4	9.0–10.2	9.7	±0.3
Length of dorsal fin	45.3–47.6	46.3	±1.1	39.8–44.1	41.4	±1.4	31.0–47.4	40.9	±1.4
Length of anal fin	12.7–15.3	14.0	±1.1	12.0–14.2	12.9	±0.7	7.5–15.5	12.2	±1.3
Length of pectoral fin	18.4–23.0	21.1	±1.8	21.1–23.8	22.0	±0.8	17.4–27.8	22.5	±1.4
Length of ventral fin	17.6–21.6	19.3	±1.6	15.6–17.4	16.6	±0.5	10.3–21.0	17.6	±1.4
Length of caudal fin	16.3–18.6	17.0	±0.9	11.7–17.4	14.7	±1.6	5.7–18.4	15.7	±1.4

Figure 1. 

A–C Serranus papilionaceus D–F Serranus scriba. Photos include details of specimens representing both species (underwater photography by Mr. P. Schembri, BICREF NGO). Scale bars: 1 cm (A–C, E, F).

Figure 2. 

A map showing: the sampling locations for this study [1 – Malta, 2 – Canary Islands]; and the locations from where Serranus papilionaceus was recorded [2 – Canary Islands (Barker-Webb and Berthelot 1836; current study); 3 – Mauritania (MNHN-IC-1999-1053); 4 – Gorée, Senegal (Valenciennes, 1832; MNHN-IC-0000-7279; MNHN-IC-1896-0389)].

Genetic investigations

Genetic divergence

The COI and ND2 data produced three and six haplotypes for the specimens collected from the Canary Islands (n = 12), while five and nine haplotypes were identified amongst the specimens collected from Malta (n = 15) for each respective gene (Table 4; Fig. 3). The intraspecific divergence for COI of the analysed specimens from Malta was 0.09%, while on incorporating similar publicly available data for S. scriba from other areas of the Mediterranean (n = 41; Suppl. material 1: Table S3), the intraspecific divergence increased to 1.12%, with all sequences clustering together on the same branch of the phylogenetic tree (Fig. 4). For the specimens collected from the Canary Islands, the intraspecific divergence on COI was 0.11%. The genetic divergence between the two collections was found to be 4.31% (Table 5), that is four times larger than the intraspecific divergence, exceeding the species boundary delimitation of fish species (Ward et al. 2005). This led to the confirmation that specimens belonged to two different species. This large interspecific genetic difference was also noted on the ND2 gene where the intraspecific divergence was 0.45% and 0.31% for S. scriba and S. papilionaceus, respectively, while the interspecific divergence was 8.68% (Table 5), again surpassing the intraspecific divergence differences noted on the ND2 gene in other fish species by Naylor et al. (2012) and Vella et al. (2017). The haplotype networks of this mtDNA data show that the two species form two distinct clades, with no overlapping haplotypes (Fig. 3).

These results were further corroborated by both nuclear markers, where species specific SNPs were noted (Table 6). This data therefore confirms that genetically the specimens collected from the Canary Islands belong to a different species from those collected from Malta.

Table 4.

Download as

CSV

XLSX

The number of haplotypes noted (hap.), the haplotype diversity (h) and the nucleotide diversity (π) at each gene for Serranus papilionaceus and Serranus scriba.

mtDNA gene	Serranus papilionaceus			Serranus scriba
	hap	h	π	hap	h	π
COI (621 bp)	3	0.621 ±0.087	0.0011 ±0.0010	5	0.476 ±0.155	0.0009 ±0.0009
ND2 (846 bp)	6	0.758 ±0.122	0.0031 ±0.0020	9	0.924 ±0.044	0.0045 ±0.0027
concatenated (1467 bp)	7	0.833 ±0.100	0.0023 ±0.0014	11	0.952 ±0.040	0.0029 ±0.0017

Table 5.

Download as

CSV

XLSX

Intraspecific and intraspecific average number of pairwise differences (percentage differences) for the mtDNA loci of Serranus papilionaceus and Serranus scriba.

mtDNA gene	Intraspecific divergence		Interspecific divergence
	Serranus papilionaceus	Serranus scriba
COI (621 bp)	0.70 (0.11%)	0.53 (0.09%)	26.77 (4.31%)
ND2 (846 bp)	2.62 (0.31%)	3.79 (0.45%)	73.43 (8.68%)
concatenated (1467 bp)	3.32 (0.23)	4.33 (0.30%)	100.20 (6.83%)

Table 6.

Download as

CSV

XLSX

SNPs noted for the nDNA loci of Serranus papilionaceus and Serranus scriba (^a species-specific differences).

nDNA gene	Rhod							PTR
Position	30	204	225a	232	273	628a	690a	160a	205a
Serranus papilionaceus	C (n = 11)	C (n = 15)	G (n = 12)	T (n = 9)	C (n = 9)	A (n = 15)	C (n = 15)	T (n = 12)	G (n = 12)
	C/T (n = 1)			C/T (n = 3)	C/T (n = 3)
Serranus scriba	C (n = 15)	C (n = 14)	T (n = 15)	C (n = 15)	C (n = 15)	G (n = 15)	T (n = 15)	C (n = 15)	A (n = 15)
		C/T (n = 1)

Figure 3. 

Two parsimony haplotype networks representing the genetic relationship between Serranus papilionaceus (Canary Islands) and Serranus scriba (Malta) for the COI and ND2 data (The haplotype frequencies are proportional with the area of the circle. The numbers in the brackets represent the number of interspecific substitutions, while the black circles represent inferred putative haplotypes within species that were not observed during this study).

Phylogeny

All the species of Serranus found in the Mediterranean and the Canary Islands (S. atricauda, S. cabrilla, S. hepatus, S. papilionaceus, and S. scriba) share the same ancestral branch of the phylogenetic tree (Fig. 4), with each species forming its own distinct clade. This phylogenetic tree placed S. scriba and S. papilionaceus as sister species to each other.

Figure 4. 

A neighbour-joining phylogenetic tree using p-distance. 186 COI sequences (578 bp each) representing 16 different Serranus species were utilized (Suppl. material 1: Table S3). The numbers near nodes represent the bootstrap values.

Mitochondrial DNA data show that the genetic distance between the Mediterranean specimens of S. scriba and the ones collected from waters of the Canary Islands exceeded the intraspecific genetic distance (COI: 4.31%; ND2: 8.68%; Figs 3, 4) for both genes (Ward et al. 2005; Naylor et al. 2012; Vella et al. 2017), confirming that the specimens from these two different sampling sites belong to two different species. The morphological colour and banding pattern difference between these two species support the genetic distinction found.

Serranus scriba was first described by Linnaeus (1758: 292) as Perca scriba. This description is fairly brief and lacks the locality; however, specimen NRM 442 from the Swedish Museum of Natural History (Suppl. material 1: Fig. S1a) is believed to be the holotype for P. scriba (Fernholm and Wheller 1983; Parenti and Randall 2020). In Museum Adolphi Friderici volume 2, Linnaeus (1764: 86) updated the description of Perca scriba and included some more details on the colouration of the species, namely reticulations and undulations over the head and brownish transverse bands, relating it to the Mediterranean S. scriba. However, in this account Linnaeus (1764: 86) indicated America as its area of distribution (Fernholm and Wheller 1983). Cuvier and Valenciennes (1828) made a detailed description of S. scriba, providing an illustration of the species (plate 28; Suppl. material 1: Fig. S1b), as well as indicating its distribution as including the coast of Provence (France), Malta, Naples (Italy), and Alexandria (Egypt). This description matches the two museum specimens MNHN-IC-0000-7129 (Suppl. material 1: Fig. S2b) and the specimens collected from Malta for the current study (Fig. 1). We therefore keep the widely accepted S. scriba nomenclature here in reference to the Mediterranean species sampled and genetically investigated in this study.

After reviewing the possible junior synonyms for S. scriba (Fernholm and Wheller 1983; Smith-Vaniz 2015; Horton et al. 2020; Parenti and Randall 2020), it was noted that our specimens from the Canary Islands matched the description of S. papilionaceus (Valenciennes, in Cuvier and Valenciennes 1832: 471). Serranus papilionaceus was first collected from Gorée Island, Senegal, by M. Rang and described by Valenciennes (1832: 471), and it is represented by two extant syntypes in the Muséum National d’Histoire Naturelle (MNHN-IC-0000-7279; Suppl. material 1: Fig. S2a). Our findings corroborate the second account for this species as described by Barker-Webb and Berthelot (1836), who indicated that S. papilionaceus occurs in the Canary Islands. In the current study, we also reviewed the description of H. marocannus Bloch & Schneider, 1801, as Peters (1865) suggested that this is synonymous to S. papilionaceus Valenciennes, 1832. However, given that the original description (Bloch and Schneider 1801: 320) and the quality of the type specimen (ZMB5531) lack the morphological details required to synonymise the two, we consider H. marocannus to be a junior synonym of S. scriba as noted by Cuvier and Valenciennes (1828), Gunther (1859), and Parenti and Randall (2020).

Historic records show that S. papilionaceus was accepted as a valid species name for a couple of decades (Barker-Webb and Berthelot 1836; Gunther 1859). However, by 1890, studies on specimens from the Canary Islands were considered as the variant S. scriba var. papilionaceus (Dooley et al. 1985). As a result, the Butterfly-winged Comber had lost its status as a distinct species. The similarity between the two species was reported since the first description of S. papilionaceus, with Barker-Webb and Berthelot (1836) stating that the two species, considered here, were very similar and could be easily misidentified. Nonetheless, the banding pattern, especially of live specimens is the most distinctive morphological feature to differentiate between the two species. Additionally, the results of our genetic analyses confirm the genetic divergence between these two species.

The resurrection of S. papilionaceus adds another species to the genus Serranus, genetically sister to S. scriba (Fig. 4). This calls for new studies that look into the geographical distribution of both S. papilionaceus and S. scriba. Online searches for S. scriba indicate that the morphological details belonging to S. papilionaceus are present in photos that have originated from different coastal regions along the North-Eastern Atlantic Ocean, and we cannot exclude the possibility of potential range overlap of the two species in certain geographical areas.

Though Serranus species have been listed as Least Concern (Smith-Vaniz 2015; IUCN 2020), the conservation status of the now recognised distinct species, S. papilionaceus, needs to be evaluated. This is especially so when one considers that the Canary Islands had already listed their Serranus species as threatened some time ago (Bonnet and Rodriguez 1992; Tuset et al. 2005) and fishing pressures have not decreased.

Genetic diversity investigation results and different body colour patterns observed have led us to resurrect S. papilionaceus Valenciennes 1832 as a species distinct from S. scriba (Linnaeus, 1758). Phylogenetic analysis confirmed that these two species are sister species within the genus Serranus and that their interspecific genetic divergence is four times larger than the usual intraspecific divergence in fish species. The resurrection of S. papilionaceus also emphasizes the required new research to understand the distinct distributions, threats, and conservation management of these species.

The research disclosed in this publication has been funded through the BioCon_Innovate Research Excellence Grant from the University of Malta awarded to AV. The authors would like to thank several fishers from Malta and the Canary Islands, who have supported this research and assisted in specimen collection. A special thanks goes to Jonathan Pfliger from the Muséum National d’Histoire Naturelle (France), Edda Aßel from Museum für Naturkunde – Leibniz Institute for Evolution and Biodiversity Research (Berlin, Germany), and Bo Delling and Anders Silfvergrip from the Swedish Museum of Natural History (Stockholm, Sweden) for providing the authors with the required information and photos of museum specimens. Additionally, thanks are also due to the scuba diver Patrick Schembri from the NGO BICREF for providing photography of live specimens from Malta. The three scientific reviewers of this manuscript are appreciated for their valuable feedback.

This study did not require ethical approval as it made use of dead specimens and tissue samples collected from specimens caught by local fishermen during regular fishing activities.

DNA sequences generated in this study have been deposited in GenBank. Details of the specimens and the associated accession numbers are included in the Suppl. material 1: Table S2.