Polydora and Dipolydora (Polychaeta: Spionidae) Associated with Molluscs on the South Coast of South Africa, with Descriptions of Two New Species

ABSTRACT Wild molluscs from five sites and farmed oysters and abalone from one site each along the south coast of South Africa were examined for shell-boring Polydora and Dipolydora. One new species each of Dipolydora and Polydora are described. D. keulderae sp. n. belongs to the D. barbilla / D. bidentata group and was present at all the sites sampled. P. dinthwanyana sp. n. belongs to the P. ciliata / P. websteri group and was recorded only at Haga Haga, the most easterly site sampled. In total, five Dipolydora species (D. cf. armata, D. capensis, D. cf. giardi, D. keulderae and D. normalis) and three Polydora species (P. cf. ciliata, P. dinthwanyana and P. cf. hoplura) were recorded. All but D. normalis bored into mollusc shells and while the rest of the worms showed no host preference, the larger, subtidal Haliotis spp. and Turbo sarmaticus usually hosted the largest number of species at each site. D. normalis and P. dinthwanyana were recorded only at Haga Haga, P. cf. ciliata only at Port Elizabeth and the rest at all the sites. This study extends the known distributions of D. normalis, D. cf. armata and P. cf. ciliata in South Africa.

Local interest in the polydorid worms has grown with the increase in abalone and oyster culture in South Africa over the last decade (Simon et al. 2006). This has led to the Abalone Farmers Association of South Africa, and later Marine and Coastal Management (now part of the Department of Agriculture, Forestry and Fisheries), worms associated with wild molluscs. This is the last of three papers describing the polydorid worms associated with molluscs along the south coast of South Africa (Simon 2009;Simon et al. 2010 (Day, 1957) Figs 2A-D
Pygidium cuff shaped with dorsal notch (Fig. 2D). Comparison: New material mostly corresponding with original description and paratype examined (Day 1957); in the paratype the last 12 chaetigers are abranchiate, falling within the range observed in the new material, posterior inferior companion chaetae are not mentioned in the description but are present in both the new material and the paratype, although they are present on fewer chaetigers in the latter. The specimens differ with respect to the length of the caruncle, which extends to the posterior edge of chaetiger 3 in the new material. In the original description it is described as extending to chaetiger chaetiger 4. The striated structure of posterior notochaetae described by Day (1957) was not observed.
spine; the paratype also has younger spines which are straight and pointed, as described for the new material. The concavities of older spines in the new material were never as pronounced as in the paratype. The differences observed here may be related to the smaller size of the new material examined. Distribution: Previously found along the north-east coast of southern Africa, with its southernmost distribution at Durban. In the current study the distribution range is extended further south, and it was found only at Haga Haga in the Eastern Cape Province. Habitat: Found in tubes on the surface of farmed Haliotis midae shells. Dipolydora keulderae sp. n.

Figs 3, 4
Etymology: This species is named for Felicia Keulder who helped with the sampling.
Pygidium reduced, with four lobes, dorsal pair smaller than ventral pair (Fig. 3E), in some specimens looks cuff-shaped. Comparison: D. keulderae sp. n. belongs to the D. bidentata / barbilla group (sensu Blake, 1996). The species resembles all other members of this group in having a notched or of the spine, and hooded hooks from chaetiger 7. D. keulderae most closely resembles D. barbilla Blake, 1980 andD. pilocollaris Blake &Kudenov, 1978. Generally, the length of the caruncle is the same in all three species. Common to D. keulderae and D. barbilla are the maximum number of hooded hooks and the loss of the second tooth in the hooded hooks; however, D. barbilla has heavy posterior notopodial spines (Blake 1980) that are lacking in D. keulderae. D. keulderae and D. pilocollaris are similar in the start of the branchiae and the absence of posterior notopodial spines, but differ with respect to the maximum number of hooded hooks per fascicle and the presence of unidentate hooks posteriorly (Blake & Kudenov 1978). D. keulderae differs from and companion chaetae: in both D. barbilla and D. pilocollaris the falcate spines have bristled collars, while the latter also has hastate companion chaetae. D. keulderae also has companion chaetae accompanying the hooded hooks for only the last two chaetigers (the presence of a winged chaeta with only the hooded hooks on the right neuropodia on chaetigers 7 and 8 in one specimen suggests that this is an anomaly), while the other two species have companion chaetae with the anterior hooded hooks. D. pilocollaris also lacks notochaetae on chaetiger; in D. keulderae these notochaetae are present but very inconspicuous.  Etymology: From isiXhosa dinthwa (spot) and -nyana the pigmented spots that resemble freckles. Description: Only two specimens, holotype complete, in two fragments, paratype anterior fragment. Holotype 7.5 mm for 70 chaetigers, 0.4 mm wide at chaetiger 5. Squat peristomium, width 0.2 mm, length 0.1 mm (Fig. 5A); body narrows posteriorly, penultimate chaetiger 0.2 mm wide (Fig. 5B). Prostomium rounded, entire; caruncle extending to end of chaetiger 2; no occipital tentacle; holotype with three eyes (Fig. 5A), paratype with four eyes, arranged in trapezoid. Pigmentation: prostomium, caruncle, peristomium and chaetigers 1-3 with yellow-brown spots; in holotype last three chaetigers without pigmentation, preceding 17 chaetigers spotted with yellow-brown pigment, but decreasing in intensity anteriorly (Figs 5A, 5B).

% of chaetigers branchiate.
Pygidium small, rounded (Fig. 5B). Glands in chaetigers 7-10 in paratype, with large sacs (Fig. 5E). Comparison: P. dinthwanyana sp. n. belongs to the P. ciliata / P. websteri group (Blake 1996). It resembles P. woodwicki Kudenov, 1978 andWilliams, 2001 with respect to the shape of the prostomium and peristomium, length of the caruncle, All three species are shell-borers. P. dinthwanyana further resembles P. woodwicki in and in having a small pygidium, although the shape differs. Differences include the arrangement of anterior notochaetae in two tiers and the presence of superior dorsal fascicle of geniculate chaetae on chaetiger 5 in P. woodwicki (Blake & Kudenov 1978). P. dinthwanyana resembles in the arrangement of anterior notochaetae in Habitat: Both specimens found boring into mollusc shells (H. midae and T. sarmaticus).

RESULTS AND DISCUSSION
Five Dipolydora species (D. cf. armata, D. capensis, D. cf. giardi, D. keulderae sp. n. and D. normalis) and three Polydora species (P. cf. ciliata, P. dinthwanyana sp. n. and P. cf. hoplura) were recorded (Table 1). All but D. normalis, which was found in tubes on the shell surface, bored into the host shell. Species richness appears to decrease Grootbank and four each at Mossel Bay and Struisbaai. The farmed oysters at Port Elizabeth had only three species. Two species, D. normalis and P. dinthwanyana, were recorded only at Haga Haga, while P. cf. ciliata was found only on the farmed oysters in Port Elizabeth. The rest of the species were found at most sites sampled, suggesting that they do not have discrete distributions. The distribution of D. capensis and D. cf. giardi in the current study corresponds with that given by Day (1967), while the records of P. cf. hoplura in Haga Haga correspond with earlier distribution records (Day 1967;Nel et al. 1996;Simon et al. 2006). In contrast, the records of D. normalis at Haga Haga and P. ciliata at Port Elizabeth represent a southward expansion of the ranges of these species. Day (1967) considered the single record of D. cf. armata, from an unknown locality, as dubious, yet it occurred commonly in the wild hosts sampled here, and was present on farmed abalone in Haga Haga. The absence of D. cf. and D. maculata in the current study is not surprising; the former species inhabits sandy sediments while the latter is associated with hermit crabs (Day 1967(Day ). et al. 2006Sato-Okoshi et al. 2008), and the results of the current study support this. Polydora cf. hoplura and D. capensis were the most catholic with respect to their host also the most abundant. The other species were recorded on one to three hosts, and seldom in great numbers. The two Haliotis species and Turbo sarmaticus hosted the most species; one or the other would host all the species recorded at each site compared to the other hosts. These results suggest that infestation of a host is a function of its size and habitat -the Haliotis species and Turbo sarmaticus are larger, and possibly longerlived, than the other host species sampled and occur in the subtidal where they could be exposed to potential settling larvae almost continuously. In older or longer-lived animals the diversity of shell-borers can also increase due to the presence of secondary borers (Evans 1969). As the shell ages, it is eroded by the boring activity of primary borers. When these polychaetes die, they can be replaced by secondary borers which will settle in the abandoned burrows, often enlarging them as they grow. and oysters as far west as Saldanha Bay and Jakobsbaai: common to both are Polydora cf. hoplura and Boccardia proboscidea, while the abalone are often also infested by Dipolydora capensis (Nel et al. 1996;Simon et al. 2006;Simon & Booth 2007;Haupt et al. 2010). In the current study farmed oysters were also infested by P. cf. ciliata, a common borer of molluscs in other locations (e.g., Radashevsky et al. 2006;Walker 2011), and a few individuals of Dipolydora keulderae, while farmed abalone were also infested by low numbers of Dipolydora cf. armata and D. normalis. Thus the species that have been recorded as problematic, or which have the potential to become problematic, are either cryptogenic (P. cf. ciliata and P. cf. hoplura 1950s, although the latter species is considered invasive (Day 1955(Day , 1957Haupt et al. 2010)) or a recent invader (B. proboscidea, Simon et al. 2009). In South Africa, cultured oysters are often moved between farms and geographical zones (Dr K. Christison, Marine Aquaculture Diseases, Department of Agriculture, Forestry and Fisheries, pers. comm.), which could inadvertently spread the worm, as occurred with B. proboscidea on abalone ). The presence of P. cf. ciliata in Port Elizabeth and B. proboscidea in Saldanha Bay (Haupt et al. 2010) therefore suggests that caution should be exercised with continued movement of oysters. Detailed epidemiological surveys of internal parasites and polydorid worms associated with oysters and abalone from different zones in South Africa are currently underway. The results should provide more information towards formulating policies governing the movement of oysters.
Of the eight species recorded in this study, three species are considered cosmopolitan (D. cf. armata, P. cf. ciliata and P. cf. hoplura D. cf. giardi of cosmopolitan species may in fact be erroneous. For example, P. cf. ciliata has been demonstrated to be a complex of several morphologically similar species (Manchenko & Radashevsky 1998), while P. cf. hoplura and P. uncinata are easily confused (cf. Day 1967;Blake & Kudenov 1978;Sato-Okoshi 1998). Similarly, the cosmopolitan distribution of D. cf. armata has also been questioned (Delgado-Blas & Salazar-Silva 2011). A systematic review, including detailed morphological descriptions and molecular and invasive status. D. capensis suggest that there may in fact be several morphologically similar species. This, too, will be addressed later.