Morphological and molecular characterisation of Scutellonema species from yam (Dioscorea spp.) and a key to the species of the genus

Summary The yam nematode, Scutellonema bradys, is a major threat to yam (Dioscorea spp.) production across yam-growing regions. In West Africa, this species cohabits with many morphologically similar congeners and, consequently, its accurate diagnosis is essential for control and for monitoring its movement. In the present study, 46 Scutellonema populations collected from yam rhizosphere and yam tubers in different agro-ecological zones in Ghana and Nigeria were characterised by their morphological features and by sequencing of the D2-D3 region of the 28S rDNA gene and the mitochondrial COI genes. Molecular phylogeny, molecular species delimitation and morphology revealed S. bradys, S. cavenessi, S. clathricaudatum and three undescribed species from yam rhizosphere. Only S. bradys was identified from yam tuber tissue, however. For barcoding and identifying Scutellonema spp., the most suitable marker used was the COI gene. Additionally, 99 new Scutellonema sequences were generated using populations obtained also from banana, carrot, maize and tomato, including the first for S. paralabiatum and S. clathricaudatum, enabling the development of a dichotomous key for identification of Scutellonema spp. The implications of these results are discussed.

Yam (Dioscorea spp.) is an important staple crop cultivated for its edible tubers in West Africa (Asiedu & Sartie, 2010). The plant-parasitic nematode Scutellonema bradys (Steiner & LeHew, 1933) Andrássy, 1958, or 'the yam nematode', is a migratory endoparasite that causes dry rot disease of yam tubers, creating a persistent decline of tuber quality and even total loss during storage (Bridge et al., 2005). Feeding by the nematode results in necrotic lesions beneath the outer skin. These lesions become yellow and gradually brown to black with progression of the disease. The outer skin may be intact, disguising the damage below, or it may become flaky or develop cracks, which serve to facilitate secondary infection by Scutellonema spp. are associated with roots of a wide range of crops (Sher, 1964;Van den Berg & Heyns, 1973;Knight, 2001;Agudelo & Harshman, 2011;Coyne et al., 2016). The genus Scutellonema was proposed by Andrássy (1958) and included all Rotylenchus members with large phasmids (scutella) located either opposite each other or nearly so and at the level of the anus or cloacal aperture. In a comprehensive review, Sher (1964) listed 11 species, a list later expanded to 45 valid Scutellonema species (Siddiqi, 2000). Just three new species have since been reported (Saha et al., 2000;Giribabu & Saha, 2002). Species-level identification has traditionally relied upon detailed morphological analysis, a lengthy and labour intensive task that requires considerable expertise (Coomans, 2000) given the morphological conservatism within a genus (Powers et al., 2011). Scutellonema spp. identification is based on the analysis of morphometrics and morphological characters, such as lip region morphology, lip region shape, number of lip region annuli, number of striations on the basal lip annulus, position of the hemizonid, secretory-excretory (S-E) pore and scutella, size of the scutella, structure of the female reproductive system, presence of 'vaginal glands' (conspicuous cuticular thickenings towards ends of vulva) and epiptygmata, and areolation at scutella level (Sher, 1964;Smit, 1971; Van den Berg & Heyns, 1973;Germani et al., 1985a;Baujard et al., 1990;Krall, 1990). However, given the lack of tangible morphological characters to distinguish important Scutellonema species, viz., S. bradys, S. cavenessi Sher, 1964and S. clathricaudatum Whitehead, 1959a, Baujard & Martiny (1995 grouped these three species into the "S. bradys complex".
DNA barcoding-based methods have proved invaluable for delineating species lacking contrasting morphological features. The DNA regions coding for ribosomal genetic markers (D2-D3 of 28S rRNA-, 18S-, and ITS-rRNA) have been commonly used to identify Scutellonema spp. (Chen et al, 2006;Subbotin et al., 2007;Van den Berg et al., 2013, 2017Tzortzakakis et al., 2016). The mitochondrial Cytochrome c Oxidase I gene (mtCOI) (Hebert et al., 2003), which is commonly used for barcoding, has also been explored for a limited number of nematode species (Palomares-Rius et al., 2014), including Scutellonema spp. (Van den Berg et al., 2013, 2017. The aims of this study were: i) to conduct species level characterisation of Scutellonema populations collected from yam tubers and yam rhizosphere in the main yam growing areas in Nigeria and Ghana, using morphological, morphometric and molecular data (D2-D3 expan-sion segments of 28S rDNA gene and Cytochrome c oxidase subunit 1 (COI); ii) to determine the phylogenetic interrelations to delimit species; and iii) to develop a morphological key for species of Scutellonema.

NEMATODE SAMPLES
Nematode populations used in this study were isolated from yam rhizosphere and yam tubers taken from farmers' fields and experimental plots in different agro-ecological zones in Ghana and Nigeria during surveys conducted between 2012 and 2015 (Table 1). Nematode populations from soil, roots and tubers were isolated using the Whitehead and Hemming tray technique (see Hooper et al., 2005). Soil samples of 100 ml were used for nematode extraction. Yam roots retrieved from each soil sample were carefully washed, chopped into small pieces (0.5-2.0 cm) and processed separately from the soil. For tubers, three subsamples of 5 g were used for the extraction from yam peel (Coyne et al., 2006;Baimey et al., 2009). Nematode populations isolated from various substrates were collected on 28 μm sieves, washed, and divided into two parts for preservation for further analysis: one part was heat-killed and fixed in 4% formalin; the other was fixed directly in DESS solution (Yoder et al., 2006). Altogether, 120 rhizosphere and 84 tuber isolates were collected for species identification studies.

MORPHOLOGICAL CHARACTERISATION
Nematode specimens fixed in formalin were processed to anhydrous glycerin following the glycerin-ethanol method (Seinhorst, 1959) as modified by De Grisse (1969). Permanent slides were prepared and used to record morphometrics and morphological features (Sher, 1964;Germani et al., 1985a;Krall, 1990; Van den Berg et al., 2013) using an Olympus BX51 DIC microscope equipped with a Nikon digital camera. Additional morphological and morphometric data were recorded from temporary slides made from DESS fixed specimens, prior to DNA extraction. In addition, paratypes and other populations of the genus Scutellonema, available in the nematode collections in Ghent University Museum -Zoology Collections, Belgium (UGent), and in the Wageningen nematode collection, The Netherlands (WaNeCo), were included for comparison (viz., S. aberrans (Whitehead, 1959b) Sher, 1961;S. africanum Smit, 1971; S. brachyu-

KY639376
Host names followed by the letter T are samples from yam tuber; otherwise, samples are from rhizosphere. Sample codes are italicised if isolated from yam tubers; each soil sample is from a composite of the rhizosphere of four plants.

DNA extraction and PCR amplification of the 28S rDNA and COI gene
Following morphological identification, individual nematodes from temporary slides were picked and used for extraction of genomic DNA using a quick alkaline lysis protocol adapted from Schneider et al. (2015) (see Janssen et al., 2016). PCRs were performed following the protocol of the D2-D3 expansion segment of the large sub-unit (LSU) rDNA and the Cytochrome c oxidase subunits 1 (COI) as described in Van den Berg et al. (2013). The primer sets D2A (5 -ACA AGT ACC GTG AGG GAA AGT TG-3 ) and D3B (5 -TCG GAA GGA ACC AGC TAC TA-3 ) were used for the amplification of the D2-D3 expansion regions of 28S rDNA gene. The Cytochrome c oxidase subunit 1 (COI) gene fragment was amplified using the primer sets JB3 (5 -TTT TTT GGG CAT CCT GAG GTT TAT-3 ) and JB4 (5 -TAA AGA AAG AAC ATA ATG AAA ATG-3 ).
PCR products were separated by electrophoresis on a 1% agarose gel and stained with ethidium bromide. PCR products were purified as described in the manufacturer's instructions (Wizard ® SV Gel and PCR Clean-Up System Kit, Promega) and sequenced by Macrogen (Europe) in both forward and reverse directions. Consensus sequences were assembled using GENEIOUS 9.15 (Biomatters; http://www.geneious.com) and deposited in the NCBI GenBank (Table 1).

Phylogenetic analysis
The D2-D3 of 28S rDNA and mtCOI sequence generated in this study and sequences available for genus Scutellonema in the GenBank were aligned using MUS-CLE (Edgar, 2004) with default settings. Outgroup taxa of each dataset were chosen based on previously published data (Van den Berg et al., 2013). The best-fit models of DNA evolution were estimated using the program jModeltest 0. 1.1 (Posada, 2008) under the Akaike information criterion (AIC). Bayesian phylogenetic analysis (BI) was done using MrBayes 3.2.6 (Huelsenbeck & Ronquist, 2001) for 5 × 10 6 generations with a general timereversible model with a gamma distribution for the remaining sites (GTR + I + G) for D2-D3 and COI. Two runs were performed for each analysis. After discarding burn-in samples and evaluating convergence, the remaining samples were used to generate a 50% majority rule consensus trees. Posterior probabilities (PP) were plotted and given on clades with >0.7 PP support. Pairwise divergences between taxa were computed as distance values and as percentage mean distance values based on the whole alignment, with adjustment for missing data using Geneious 9.15 (Kearse et al., 2012). To test distinctiveness of putative species, generated trees were imported into Geneious where the species delimitation plugin (Masters et al., 2011) was used to calculate Rosenberg's P AB , which tests the probability for reciprocal monophyly of the clusters (Rosenberg, 2007).

CHARACTERISATION
Using morphological and molecular data, the following taxa from yam tubers and yam rhizosphere were identified: S. bradys, S. cavenessi, S. clathricaudatum, Scutellonema sp. D sensu Van den Berg et al. (2013), and two unknown species: Scutellonema sp. 1 and Scutellonema sp. 2. Ninety-nine sequences of Scutellonema (45 D2-D3 and 54 of COI) were generated from 45 populations. Of the 99 sequences, 87 were from nematodes obtained from yam tubers or yam rhizosphere and 12 from Scutellonema species collected from other crops (banana, carrot, maize and tomato) ( Table 1).
The unknown species were considered different from all known species based on morphological differences, their unique phylogenetic position and molecular species delimitation. In addition, S. brachyurus was identified from banana, S. cavenessi from onion, and S. paralabiatum Siddiqi & Sharma, 1994 from banana, maize and onion rhizosphere (Table 1).  Table 2.

Female
Body straight to slightly curved ventrally after fixation. Lateral field areolated at anterior portion of body and at scutellum level, smooth to partially areolated at mid-body. Lip region, hemispherical, offset by slight to deep con-striction with seven (5-9) annuli and without longitudinal striations on basal lip annulus (observation from SEM). Labial disc rounded with small amphidial openings laterally. Stylet robust with knobs round to oval at base and in some cases with irregular anterior margin. Conus often shorter than shaft and knobs combined, m = 45.8 (38.5-54.5)%. Median bulb spherical to oblong. Pharyngeal gland lobes overlapping intestine dorsally. Excretory pore often located at posterior level of pharyngeal gland lobe, 137 (94-159) μm from anterior end. Hemizonid 0-3 annuli anterior to excretory pore. Spermatheca rounded to oval, filled with sperm cells. Often, with conspicuous 'vaginal glands' arranged around vulva (four in ventral     ± 0.79 ( 10.5-12.5) Median bulb valve length Lip region diam. Ant. end to S-E/pharynx length S-E = secretory/excretory pore position.
Vol. 19 (7), 2017 view and two in lateral view). Epiptygmata often absent, otherwise small and appearing double. Scutellum moderate sized with rounded shape, located opposite or slightly anterior or posterior to anus. Tail variable in shape, often tapering gradually with rounded end and striated terminus. Tail 1.1 (0.7-1.6) anal body diam. long and with 20 (13-25) annuli.

Male
Similar to female except for reproductive structures, bursa relatively narrow, not lobe-shaped with abrupt narrowing.

DIAGNOSIS AND RELATIONSHIPS
Scutellonema bradys is characterised by a straight to slightly ventrally curved female body. Lip region offset by a constriction with seven (5-9) lip annuli and lacking longitudinal striae on the basal lip annulus. Lateral field areolated at scutellum level. 'Vaginal glands' often present and well developed. Spermatheca present and filled with sperm cells.
Scutellonema bradys is similar to S. cavenessi from which it can be distinguished by its general habitus (slightly curved vs C-shaped), larger submedian lips, epiptygmata absent or very small vs long and protruding, and 'vaginal glands' conspicuous and very well developed vs not very well developed. Bursa relatively narrow vs lobe-shaped with abrupt narrowing.

REMARKS
Scutellonema bradys was the only Scutellonema species retrieved from yam tubers. Adults from tubers are relatively large compared with those from the rhizosphere ( Table 2). Sher, 1964 (Fig. 5) Four populations were obtained from yam rhizosphere from separate locations in Nigeria.

MEASUREMENTS
See Table 3.

Female
Body curved ventrally, inverted comma to C-shaped after fixation and tapering slightly towards anterior end. Cuticle at mid-body with annuli 2.1 μm wide. Lateral field areolated at anterior portion of body and at scutellum level, smooth to partially areolated at mid-body, comprising one-fifth diam. of mid-body. Lip region, hemispherical, offset by slight to deep constriction with seven (5)(6)(7)(8) annuli and lacking longitudinal striations on basal lip annulus (observation from SEM). Labial disc rounded with small amphidial openings laterally. Stylet well developed with knobs oval at base and slightly indented anteriorly. Conus often shorter than shaft and knobs combined, m = 45.5 (38.3-53.1)%. Median bulb spherical to oblong. Pharyngeal gland lobe overlapping intestine dorsally. Excretory pore often located at pharyngeal gland lobe level, 104 (83-129) μm from anterior end. Hemizonid immediately anterior to excretory pore, 1-2 annuli long. Spermatheca rounded and filled with sperm cells. Vagina with nondeveloped to well-developed 'vaginal glands' arranged around vulva (seen as four in ventral view and two in lateral view). Epiptygmata often present and single, double in some rare cases. Scutellum rounded, moderate to large in size, located at level of anus. Tail with rounded end and striated terminus, slightly ventrally curved. Tail 0.75 (0.52-0.95) anal body diam. long and with 13 (8-17) annuli.

Male
Similar to female except for reproductive structures, with wide and broadly enveloping bursa.

DIAGNOSIS AND RELATIONSHIPS
Scutellonema cavenessi is similar to Scutellonema sp. D sensu Van den Berg et al. (2013), Scutellonema sp. 1, Scutellonema sp. 2, S. clathricaudatum and S. bradys with respect to the lack of striation at the basal lip  3. Phylogenetic relationships within Scutellonema. Bayesian 50% majority rule consensus tree as inferred from the analysis of the D2-D3 expansion segments of 28S rDNA sequence alignment under a GTR + I model. Newly obtained sequences are indicated in bold. Posterior probabilities equal or more than 0.7 are given. Intraspecific variation of a clade indicated by a bar is given to the left of the bar, nucleotide differences between sister clades is provided right to the bars. Thick bars are clades that are supported in both analyses and by significant Rosenberg's species delimitation probabilities. Species that are supported as distinct taxonomic identities with significant Rosenberg's probabilities are indicated by a star.  4. Phylogenetic relationships within Scutellonema. Bayesian 50% majority rule consensus tree as inferred from the analysis of the COI mtDNA sequence alignment under a GTR + I + G model. Newly obtained sequences are indicated in bold. Posterior probabilities equal or more than 0.7 are given. Intraspecific variation of a clade indicated by a bar is given to the left of the bar, nucleotide differences between sister clades is provided right to the bars. Thick bars are clades that are supported in both analyses and by significant Rosenberg's species delimitation probabilities. Species that are supported as distinct taxonomic identities with significant Rosenberg's probabilities are indicated by a star.    30.0 ± 5.5 (25.6-36) Diam. at anus 24.7 ± 3.6 (20. 6-27.1)   Median bulb diam. 12.8 ± 1.3 (11.5-14.0) 11.5 11.5 11.5 11.9 ± 1.1 (10.0-13.5) -13.3 ± 1.8 (12-14.5) 10.7 ± 1.3 (9.5-12.0) Median bulb valve length Lip region diam. 10.8 ± 0.31 (10.5-11.0 Scutellonema cavenessi sequences formed a highly supported clade (PP = 0.97) with an intraspecific variation of 1-6 bp (0.1-0.7%) and 0-20 bp (0-5.9%) for D2-D3 and COI respectively (Figs 3, 4). Molecular divergence between S. cavenessi and its sister taxon according to D2-D3, Scutellonema sp. 1 and Scutellonema sp. 2 is 13-14 bp (1.6-2.1%) (Fig. 3) and according to COI, 49-70 bp (16.4-19.6%) and 69-81 bp (18.5-21.4%) for Scutellonema sp. 1 and Scutellonema sp. 2 respectively (Fig. 4).

Female
Body arcuate, C-shaped when relaxed, annuli ca 2.1 (2-3 μm) wide at mid-body, lateral fields areolated anteriorly and at level of scutellum, often smooth at mid-body. Lip region hemispherical to conical, slightly to flattened anteriorly, not offset, slightly offset, to well offset by constriction, with seven (6-8) annuli. Basal lip annulus lacking longitudinal striations, stylet well developed with rounded to oval basal knobs posteriorly and with irregular anterior surface. Excretory pore at level of pharyngeal gland lobe, 109 (80-142) μm from anterior end. Hemizonid 0-1 annulus anterior to excretory pore and 1-4 annuli long. Genital tract often not seen in detail. Spermatheca not developed. Intestine slightly overlapping rectum. Epiptygmata usually present, single or double. Scutellum crescent to rounded in shape, located opposite or slightly anterior or posterior to anus. Tail conoid, round to squarish and ventrally curved, 0.94 (0.67-1.3) anal body diam. long and with 16 (11-21) annuli, terminus of variable shape.

REMARKS
The 15 populations showed considerable morphological and molecular variation, which could be assigned into four groups A, B, C and D, based on minor morphological and morphometric differences associated with molecular clades (for at least one of the markers). The four types all fit within S. clathricaudatum as defined by Germani et al. 3)); and iv) S. clathricaudatum type D (one population) characterised by having a broader lip width, lateral field areolated at tail level and its tail length and shape (conoid).

DIAGNOSIS AND RELATIONSHIPS
Scutellonema clathricaudatum is similar to S. conicephalum with respect to the absence of males and the spermatheca, the lack of longitudinal striae on the basal lip annulus and with areolation at scutellum level. However, S. clathricaudatum can be distinguished by having 6-8 lip annuli vs three in S. conicephalum.
Although some molecular clades within S. clathricaudatum were found associated with some minor morphological differences, species delimitation did not appoint distinct taxonomic identities within S. clathricaudatum (no significant Rosenberg's P AB values).

Female
Body slightly ventrally curved to spiral. Cuticle at midbody with 1.8 μm annuli wide. Lateral fields one-sixth diam. of mid-body, completely areolated at anterior portion of body and at tail level (from anterior region of scutella to tail end) and partially to completely areolated at mid-body. Lip region broadly rounded, slightly flattened anteriorly and slightly offset from body, with 6-7 annuli. Basal lip annulus without longitudinal striations. Stylet well developed. Stylet knobs, rounded posteriorly, flattened and slightly indented anteriorly, 3.0 (2.7-3.6) μm wide and 3.0 (2.2-5.2) μm high. Conus shorter than shaft and knobs, m = 43. 8 (40-45)%. Median bulb spherical to oblong. Pharyngeal gland lobe overlapping intestine dorsally. Excretory pore situated at nerve ring level, 107 (99-117) μm from anterior end. Hemizonid two annuli long, situated opposite excretory pore. Spermatheca thick-walled and either empty or filled with sperm cells. Vagina with not well developed 'vaginal glands'. Epiptygmata single to double, not observed in some cases (double in original description). Intestine not overlapping rectum. Scutellum moderate to large size, crescent to rounded in shape, situated opposite anus to posterior to anus. Tail straight to ventrally curved, 0.86 (0.7-0.9) anal body diam. long and with 15 (13-19) annuli.

Male
Similar to female except for reproductive structures with a broadly enveloping bursa.

DIAGNOSIS AND RELATIONSHIPS
Scutellonema sp. D is similar to S. clathricaudatum and S. cavenessi. It is distinguished from S. clathricaudatum by the presence of the spermatheca and males and from S. cavenessi by the areolation of the lateral field at mid-body (partially areolated vs partially striated) and spermatheca obscure and reduced in Scutellonema sp. D vs generally developed and filled with sperm cells in S. cavenessi.
Vagina often with obscure 'vaginal glands', epiptygmata often present and single, double or not observed in rare cases. Tail rounded to gradually tapering towards tail tip, 0.87 anal body diam. long with 10-17 annuli, terminus variably shaped.

Male
Similar to female except for reproductive structures. Bursa narrow and not lobe-shaped with abrupt narrowing.

DIAGNOSIS AND RELATIONSHIPS
Scutellonema sp. 1 is similar to S. cavenessi, Scutellonema sp. D, and Scutellonema sp. 2 with respect to the presence of males and absence of longitudinal striae on the basal lip annulus. Scutellonema sp. 1 differs from S. cavenessi in having the spermatheca relatively smaller vs well developed, tail less broadly rounded vs rounded, and the S-E ratio (anterior end to S-E/pharynx length = 0.9 (0. 8-1.1) in Scutellonema sp. 1 vs 1.1 (1.0-1.3) in S. cavenessi). Scutellonema sp. 1 differs from Scutellonema sp. 2 in having non-projecting epiptygmata vs projecting in Scutellonema sp. 2, and with the hemizonid observed at pharyngeal gland level vs hemizonid at pharyngointestinal junction and nerve ring level).

Scutellonema sp. 2
( Fig. 8I-N) Two populations of this species were collected from yam rhizosphere from two locations in Nigeria. Table 7.

Female
Body arcuate, C-shaped to spiral when relaxed, annuli ca 1.4 μm wide at mid-body, lateral fields areolated anteriorly and at level of scutellum, in some cases areolated in additional regions. Lip region hemispherical, slightly flattened anteriorly, usually slightly offset, occasionally well offset, with seven (6-7) annuli. Basal lip annulus without longitudinal striations (SEM), stylet well developed with rounded to oval basal knobs and an irregular anterior surface. Excretory pore at level of pharyngeal gland lobe, 103 (98-105) μm from anterior end. Hemizonid one annulus long and situated from directly anterior to two annuli anterior to excretory pore. Spermatheca not developed to spherical and of small size when visible. Vagina often with obscure 'vaginal glands', projecting epiptygmata often present, single, double or not observed in some rare cases. Tail rounded to gradually tapering towards tail tip, 0.97 (0.74-1.2) anal body diam. long with 15 (12-20) annuli, terminus variably shaped.

Male
Similar to female except for reproductive structures. Bursa lobe-shaped with abrupt narrowing.  13.5 (11.0-17.0) μm in Scutellonema sp. 2 vs 11.5 (9.5-13.5) μm in Scutellonema sp. 1 and 11.3 (8.5-15.0) μm in Scutellonema sp. D. Scutellonema sp. 2 is readily distinguished from Scutellonema sp. D and Scutellonema sp. 1 by the hemizonid located at the posterior level of the pharyngeal gland lobe compared with at the anterior level of the pharyngeal gland lobe.

MOLECULAR PHYLOGENY OF Scutellonema
The tree topologies based on D2-D3 and COI are largely similar and do not show inconsistencies, except for the positions of clades, which are not well supported according to the COI-based tree topology (see below). However, the COI-based tree topology was better resolved (Figs 3, 4). The D2-D3 of 28S rDNA gene sequence alignment was 677 bp long and contained 73 Scutellonema sequences and three outgroup taxa. The COI gene sequence alignment was 390 bp in length and contained 82 sequences of Scutellonema and two outgroup taxa. Intra-and interspecific variation are given at the species description section and on the respective trees (Figs 3, 4). The

Discussion
In 1964, six of the 11 species revised and described by Sher were recorded from Nigeria alone, with over 60% of all valid species of Scutellonema reported from Africa (Siddiqi, 2000), demonstrating the high diversity of the genus on the continent. Based on morphology and morphometrics, we identified four morphospecies of Scutellonema from yam tubers and yam rhizosphere in Ghana and in Nigeria: S. bradys, S. cavenessi, S. clathricaudatum, and Scutellonema sp. D. However, phylogenetic analysis based on COI and D2-D3 sequences, in combination with a molecular species delimitation method, revealed two additional unknown species, namely Scutellonema sp. 1 and Scutellonema sp. 2. This indicates a much wider diversity of the genus Scutellonema than previously recognised, confirming the need for more robust and accurate diagnostics of the genus.
The species found in the present study are mainly characterised by their large number of lip region annuli (>5), the presence of areolation at scutella level, the absence of longitudinal striations on the basal lip annulus, and their relatively small stylets (rarely exceeding 30 μm). They all belong to either the amphimictic or parthenogenetic group. The parthenogenetic populations were all categorised within S. clathricaudatum sensu lato, which is known for showing large variability in size and tail shape, and with the lip region "rounded and distinctly offset, to truncate and slightly offset or truncate and distinctly offset" (Sher, 1964;Ali et al., 1973;Germani et al., 1985a;Baujard & Martiny, 1995). Based on a combination of morphological and molecular data, four putative clusters (types A, B, C and D) could be observed which all fit morphologically within S. clathricaudatum. However, molecular species delimitation could not confirm the taxonomic distinctness of these lineages. The topologies based on both rDNA and mtCOI are in agreement with Van den Berg et al. (2013, 2017, especially in respect to the three major clades (I, II and III) displayed.
However, while Clade III was not well resolved based on rDNA topology, some relationships that were not well supported based on D2-D3 received close to maximal support based on COI (viz., Scutellonema sp. 1 with Scutellonema sp. 2; S. clathricaudatum type A with S. clathricaudatum type D). Nevertheless, all taxa could be identified independently of the used marker. Hence, our results confirm that both the D2-D3 expansion segments and COI are useful markers for Scutellonema species delimitation. However, the uni-parental inheritance and the high mutation rate in the mitochondrial sequences provide a better differentiation of closely related species (Janssen et al., 2016). This is especially important for the identification and description of hybrid or cryptic species (Powers, 2004;Kanzaki & Giblin-Davis, 2012;Palomares-Rius et al., 2014). A better phylogenetic resolution of COI is already well known, based on Hoplolaimus (Holguin et al., 2015), Rotylenchus (Cantalapiedra-Navarrete et al., 2013) and Rhabditidae (Fonseca et al., 2008), for example. Furthermore, the higher PCR success rate for COI compared with D2-D3 experienced here (70 vs 40%) identifies COI as a preferred and superior marker for Scutellonema.
The current study revealed a wide diversity of Scutellonema species occurring in the yam rhizosphere in Ghana and Nigeria, although only S. bradys was recovered from yam tuber tissue, a finding which is of clear biological interest. Comparing parasitism genes of S. bradys and congeners could provide insights into the evolution of endoparasitism in Scutellonema and improve our understanding of the molecular basis of host-parasite interactions and endoparasitism in Scutellonema. This could be tackled by comparing the transcriptome analyses of S. bradys and well selected species from both Clade I and Clade III.
As the only species that appears able to enter and damage tubers, the need for a precise identification of S. bradys among its diverse congeners is clear in order to select appropriate management strategies against the yam nematode and to enable accurate monitoring of its distribution aimed at preventing its spread. Extracting nematodes from clean yam peels, without adhering soil, however, could be advised in order to detect only S. bradys and rule out other Scutellonema spp. not causing damage on yam. However, while S. cavenessi and S. clathricaudatum occur on most crops across West Africa (Caveness, 1967;Baujard & Martiny, 1995), they can cause significant damage to groundnut (Arachis hypogaea L.) production (Germani et al., 1985b;Sharma et al., 1992). Conse-quently, the ability to readily and accurately differentiate these three closely related species would be very useful.
Using only morphological-based identification, the margin of error for misidentifying as S. bradys the cohabiting Scutellonema species from the rhizosphere of yam is high. However, by combining molecular and morphological data, S. bradys appears to be a well defined monophyletic group with its morphological and morphometric characters aligning clearly with the available data (Sher, 1964;Van den Berg, 1973;Germani et al., 1985a;Humphreys-Pereira et al., 2014).
To facilitate the morphological identification of Scutellonema a dichotomous key is proposed. This key is based on the new information of current study and species descriptions by Germani et al. (1985a), Siddiqi (2000), and those described since 2000 (S. bamboosae Saha, Lal, Singh, Kaushal & Sharma, 2000;S. himachalensis Saha, Lal, Singh, Kaushal & Sharma, 2000;S. coffeae Giribabu & Saha, 2002;S. clavicaudatum). However, compared with Siddiqi (2000), S. mabelei Van den Berg & De Waele, 1990 was not included as this species has a pore-like amphid and was originally described as Rotylenchus mabelei Van den Berg & De Waele, 1990. It was listed as a species of Scutellonema by Siddiqi (2000), although not designated as a new combination, and was therefore probably a lapsus. Scutellonema southeyi Williams, 1986 andS. hoabinhiensis Nguyen &Nguyen, 1993 were not included in Siddiqi (2000) but are added in the current key.