From fin rays to DNA: supplementary morphological and molecular data to identify Mormyrus subundulatus Roberts, 1989 (Pisces: Mormyridae) from the Bandama River in Côte d'Ivoire

ABSTRACT The genus Mormyrus Linnaeus, 1758 has been well studied in western Africa where five species are currently recognized. Mormyrus subundulatus Roberts, 1989 was described in 1989, and, although morphologically and genetically close to Mormyrus rume Valenciennes, 1847, is still considered by the most recent studies as a distinct species. It is endemic to the Côte d'Ivoire and Ghana River systems. It was historically recorded from only three locations but is today very likely extirpated from one of them and threatened in the remaining two. It is categorized as “Vulnerable B1ab(iii)” by the IUCN Red List (Dankwa 2020). Although a few papers describe precisely the morphology and electric signal of the species, they are based on a very limited number of samples. In this paper, we present new data on the morphological characteristics of the species (extending the dorsal fin rays count range, which is the main morphometric criterium to distinguish M. subundulatus and M. rume) based on Museum samples and samples collected recently a few kilometers away from the type locality. Determination of the newly collected material was confirmed by the analysis of a fragment of the Cytochrome b gene, and we produced the first sequence fragment of the cytochrome oxydase subunit 1 (COI) for this species.


INTRODUCTION
Among the more than 200 known mormyrid fish species of Africa, the genus Mormyrus Linnaeus, 1758 comprises 22 species (according to Froese & Pauly 2018), five of which occur in western Africa.
The two short-snouted species, Mormyrus hasselquistii Valenciennes, 1847 and M. macrophtalmus Günther, 1866, have relatively wide distributions in Africa. M. hasselquistii is known from Senegal to Ethiopia, in the Sahelo-Sudanese basins, rivers Géba, Bandama, Comoé, Mono and lagoons Ebrié and Aguien. Mormyrus macrophthalmus is patchily distributed from Guinea, in the Niger and Volta basins, to Cameroon, in the Cross and Sanaga Rivers.
Mormyrus subundultaus Roberts, 1989 from the Bandama River, in the Lamto Nature Reserve was described in relation to the sympatric M. rume. Its description is based on anatomical features, mainly the higher number of dorsal fin rays count (DFRC), higher range of both anal fin rays and scales in lateral series. Subsequent studies from Crawford & Hopkins (1989) confirmed the distinction of M. rume and M. subundulatus by analyzing and describing M. subundulatus specific electric discharge. M. subundulatus is currently known only from three locations according to Entsua-Mensah & Lalèyè (2010): two in the Bandama River in Côte d'Ivoire and one in the Tano River in Ghana (restricted to two locations only according to Dankwa 2020).
While the available literature precisely describes the morphology (Lévêque & Bigorne 1985; Kramer 2013) and electric signal (Crawford & Hopkins 1989;Kramer 2013) of the species, these studies are based on a very limited number of samples and only one DNA fragment sequence is available (Lavoué et al. 2000). We here report on the observation of five specimens collected close to the type locality in the Bandama River, present the results of DNA analyzes for the genes COI and Cyt b and discuss the morphological traits of M. subundulatus. Although still relying on a small number of specimens, this additional data is valuable as the species is threatened in and around its type locality, as well as in its entire range, in particular, by pollution, placer mining and hydroelectric power plants.

SpecimenS examined
The type material of M. subundulatus is held at the Muséum national d'Histoire naturelle in Paris (MNHN). The type series consists in one holotype (MNHN-IC-1987-1610 and a batch of 16 paratypes (MNHN-IC-1987-1611, nine adults and seven juveniles) preserved in 90° ethanol, all coming from the type locality, the Lamto Nature Reserve in the Bandama River (Côte d'Ivoire). The original publication also mentions as paratype a specimen from the Tano River in Ghana, held in the Stanford University collection in the California Academy of Sciences (CAS -SU 63507). This specimen was pictured and radiograph for dorsal fin rays count (Fig. 1A).
An additional specimen (MNHN-IC-1963-0241) collected in 1961 in Beoumi, upper Bandama was examined in the MNHN collections. Four specimens are preserved in the Institut royal des Sciences naturelles de Belgique in Bruxelles: two specimens were collected by G. Teugels on the upper Bandama River near Marabadiassa in 1985 (88-055-P-0027 and 88-055-P-0028), and another two specimens by the same author, further upstream where the Bandama River crosses the road between Niakaramandougou and Tortiva (87-018-P-0040 and 87-018-P-0041).
These museum specimens represent the only distribution data available for the species, except for a doubtful record from Lévêque & Bigorne (1985) in the Sassandra River, west of the Bandama (see Roberts 1989 and Discussion).
Three specimens were sampled alive by us in the Bandama River for morphometric and genetic analyses ( For all the specimens examined, a precise DFRC was performed on photographs (available upon request).  All the available sequences of Mormyrus spp. for the mitochondrial genes cytochrome oxydase subunit 1 (COI) and cytochrome b (Cytb) were mined in GenBank (Table 1).
dna extraction, amplification, and Sequencing A small flesh sample was collected in the field and placed directly into 90% ethanol. The whole genomic DNA was extracted from the tissue samples using the DNeasy blood and tissue kit (Qiagen) according to the manufacturer's protocol.
A fragment of 710 bp of the mitochondrial cytochrome c oxidase subunit I (COI) gene and the complete cytochrome b (1100 bp) gene were amplified by Polymerase Chain Reaction (PCR) using the primers LCO1490 and HCO2198 (Folmer et al. 1994) for COI and L14724 (GAC TTG AAA AAC CAC CGT TG) and H15915 (CTC CGA TCT CCG GAT TAC AAG AC) (Schmidt & Gold 1993;Brito et al. 1997) for Cyt b. Amplifications were performed in 25 μl total volume including 0.5 μl of gDNA, 1 × GoTaq Green reaction buffer (Promega), 200 μM of dNTPs (Promega), 0.5 μM of both primers and 0.1 U of GoTaq DNA polymerase (Promega). PCR conditions were: 4 min at 94 °C followed by 30 cycles of 45 s at 94 °C, 45 s at annealing temperature (45 °C for COI and 48 °C for Cyt b) and 45 s at 72 °C, and then a final extension of 10 min at 72 °C. PCR products were visualized on 1% agarose gel and then purified and sequenced with each universal primer on an automated ABI3730XL Genetic Analyzer. The new sequences obtained in this study were submitted to GenBank (Table 1).

dna analySeS
Cytb sequences were successfully amplified for the three specimens sampled. They all shared the same haplotype as the sequence produced by Lavoué et al. (2000), supporting the correct identification of the specimens collected ( Fig. 2A). COI sequences were successfully amplified for the three specimens sampled, resulting in a single haplotype. This haplotype

dorSal fin rayS count
The number of dorsal fin rays is the most reliable morphological character to distinguish M. subundulatus from M. rume. Dorsal fin rays were counted for the 18 specimens of the type series held in MNHN, the paratype held in CAS, the four specimens held in MRAC and the five specimens observed by us (Fig. 3). DFRC ranged for M. subundulatus from 57 (specimens 88-055-P-0027 and 88-055-P-0028 collected upstream the Bandama River) to 73 (specimen MNHN-IC-2018-0558 collected by us downstream the Bandama River).
other morphological characterS According to the original description, anal fin ray count should be between16 and 18. Our specimens had respectively 18 (MNHN-IC-2018-0558) and 15 (MNHN-IC-2018-0559) anal fin rays. Interestingly, the paratype SU 63507 also has 15 anal fin rays (Fig. 1A). Depth of caudal peduncle was about 17 times in standard length for both specimens observed, in accordance with the original description. The color of the live specimens observed in the Bandama was almost uniformly black (specimen MNHN-IC-2018-0558, fig. 1B) and seem to turn paler on the belly and throat after death (specimen MNHN-IC-2018-0559, fig. 1 C).

dna analySeS
The correct identification of the sampled specimens was supported by the Cyt b gene fragment amplified, as those specimens had the same haplotype as the one produced by Lavoué et al. (2000). We here provide the first COI reference sequence for M. subundulatus for barcoding analyses.
The COI marker confirms that M. subundulatus is genetically divergent from M. rume, as identified by Lavoué et al. (2012). The genetic distances are rather low, but the fact that both species live in the same river stretches support their distinction as different species.   dorSal fin rayS count and field identification Intermediate between the two morphologically similar species M. rume and M. tapirus, M. subundulatus was described by Roberts (1989) based on morphological characters. Mormyrus subundulatus differs from M. tapirus in having a "larger eye, more slender caudal peduncle, more anal fin rays (21-29) and fewer circumpeduncular scales (14-19)". However, the author notices that M. tapirus "agrees with M. subundulatus in DFRC (60-74) and number of scales in lateral series (83-98)". According to Roberts (1989), M. subundulatus superficially resembles M. rume but differs in the DFRC which ranges from 60 to 71, while M. rume's counts range from 72 to 95. This gap between DFRC allows the author to identify both species in a batch of specimens collected in the Sassandra River by Lévêque & Bigorne (1985). However, the DFRC in the supposed M. subundulatus of the Sassandra River would range from 68 (one specimen) to 74 (three specimens) (data from Lévêque & Bigorne 1985: 327).
We here report on a larger range of DFRC than in the original description, with barcoded specimens having 73 dorsal fin rays. These data agree with the fin ray count observed in the Sassandra-provided the specimens in question are really M. subundulatus and really come from the Sassandra. Indeed, the population of M. subundulatus purportedly identified in the Sassandra River by Roberts (1989) is enigmatic from two points of views: first because it is unclear which of the mate-rial deposit in the MNHN by Lévêque & Bigorne (1985) comes from the Sassandra, as locality labels indicate both the Sassandra and Bandama River, and second because DFRC is slightly different from that of the Bandama population. The fact that fin ray count is significantly different in this batch supports the hypothesis that it comes from a different basin, as genetic isolation and genetic drift or selection could have produced morphological divergences.
As a conclusion, reliable identification of both species in the field based only on DFRC is difficult because they both overlap, at least between 72 and 74 dorsal fin rays. Only 39 specimens were examined for M. subundulatus. It is therefore questionable whether our data reflect the actual variability of this character. The fin ray count range could exceed the upper value of 74 given here. However, on average, Mormyrus subundulatus has fewer fin rays than M. rume: 15-18 vs 16-21 for the anal fin, 60-74 vs 72-95 for the dorsal fin (Fig. 3).
According to Roberts (1989), local fishermen in the Tano River in Ghana were able to tell the two species apart because M. rume does not give perceptible electric shocks when handled, while M. subundulatus does produce a slight electric shock. During our field trip, local fishermen warned us about the fact that Mormyrus species could give a perceptible electric shock but did not specify if this could be a differentiating feature between M. subundulatus and M. rume. Such an electric shock was experienced by one of us (BA) while handling the fish MNHN-IC-2018-0558 (Fig. 1B) from a tank to another. This characteristic is highly subjective and may not be considered diagnostic for field identification. Indeed, apart from this unique experience, the specimens we caught and handled in the Bandama River have not given any other electric shocks, 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95

actual diStribution and conServation StatuS
The distribution of Mormyrus subundulatus is currently known only from museum-preserved specimens. Its natural distribution seems to be restricted to the Tano River in Ghana, the Bandama River and possibly the Sassandra River in Côte d'Ivoire. However, reliable data is very scarce (Fig. 4A) and the species is likely to also live in other rivers between the Tano and the Sassandra Rivers such as the Bia or Comoe Rivers basins.
In the type locality, the Taabo dam has changed the flow regime from fast-flowing waters and rapids to muddy substrate with an important drawdown (Fig. 4B). The site is no longer suitable for M. subundulatus. We have caught specimens downstream of the type locality in fast-flowing river stretches (Fig. 4C). These are the southernmost known localities for the species in the Bandama River. But this area is also threatened by another dam under construction in 2020 and we can assume that all the river downstream Beoumi will soon be threatened by river management and will no longer host the species. Probably, the only remaining preserved habitats in the Bandama are located upstream the village of Beoumi. However, the river is quite narrow upstream, which makes it more vulnerable to human impact in this very populated area, pollution due to intensive agriculture, draughts and climate change effects.
From an aerial view, the Tano River in Ghana seems to be mostly slow flowing, with very few stream habitats (Fig. 4D, E). While the riversides are still forested, vast areas of mining have probably led to a significant buildup of soil and mud in the river and increased turbidity, leaving very few suitable habitats for sustainable populations of M. subundulatus.
Studies on the Sassandra River to confirm the occurrence of a population of M. subundulatus are pending. Aerial photos suggest it hosts similar habitats to those of the Bandama River, apparently well preserved, apart from the few hundreds of kilometers affected by the huge Irebouo dam. If a population of the species exists in the Sassandra River, the extent of occurrence (EOO of the IUCN) of M. subundulatus would be of about 170 000 km². If absent from this basin, the EOO would only be about 46 000 km². Considering the possible extirpation of the downstream populations in the Bandama and Tano Rivers, the EOO of the remaining healthy population of the species in the upstream Bandama River will soon be restricted to about 3000 km².
Despite this alarming suggestion, we must acknowledge that African rivers are poorly studied and the possibility remains that unnoticed populations are still to be discovered. On the other hand, the overwhelming demographic growth of African countries provides reason for concern about the future