Biotic interactions recorded in shells of recent rhynchonelliform brachiopods from San Juan Island, USA.

ABSTRACT Biotic interactions between brachiopods and spionid polychaete worms, collected around San Juan Islands (USA), were documented using observations from live-collected individuals and traces of bioerosion found in dead brachiopod shells. Specimens of Terebratalia tranversa (Sowerby), Terebratulina unguicula (Carpenter), Laqueus californianus (Koch), and Hemithiris psittacea (Gmelin) were collected from rocky and muddy substrates, from sites ranging from 14.7-93.3 m in depth. Out of 1,131 specimens, 91 shells showed traces of bioerosion represented by horizontal tubes. Tubes are U-shaped, straight or slightly curved, sometimes branched, with both tube openings communicating externally. On internal surfaces of infested shells, blisters are observed. All brachiopod species yielded tubes, except for H. psittacea. Tubes are significantly more frequent on live specimens, and occur preferentially on larger, ventral valves. This pattern suggests selectivity by the infester rather than a taphonomic bias. Given the mode of life of studied brachiopods (epifaunal, sessile, attached to the substrate, lying on dorsal valve), ventral valves of living specimens should offer the most advantageous location for suspension-feeding infesters. Frequent infestation of brachiopods by parasitic spionids is ecologically and commercially noteworthy because farmed molluscs are also commonly infested by parasitic polychaetes. In addition, brachiopod shells are among the most common marine macroscopic fossils found in the Phanerozoic fossil record. From a paleontological perspective, spionid-infested brachiopod shells may be a prime target for studying parasite-host interactions over evolutionary time scales.

KEY WORDS: biotic interaction, commensalism, parasitism, bioerosion, rhynchonelliform brachiopods, spionid polychaetes, San Juan Islands

INTRODUCTION

Whereas biotic interactions involving benthic molluscs have been extensively studied in recent ecosystems (e.g., Gosling 2003), brachiopods, which inhabit many modern seafloors and share many ecological similarities with bivalves, have remained understudied. Even in the case of the better-studied molluscs, our knowledge is biased toward commercially important bivalves such as Mytilus or Crassostrea (Blake & Evans 1973, Wargo & Ford 1993, Sato-Okoshi & Okoshi 1997; 2000). However, in many recent marine benthic ecosystems, brachiopods are a remarkably abundant component of the regional biodiversity (Kowalewski et al. 2003, Ward et al. 2006). In fact, in some marine environments (e.g., Great Australian Bight), sessile, suspension feeding organisms such as brachiopods may comprise over 96% of the total biomass and 74% of the species present (Ward et al. 2006). In such ecosystems, those sessile organisms represent bulk of biotic and abiotic interactions. Modern brachiopod shells from the South Atlantic waters (Simees et al. 2004), for example, are a home to a variety of encrusting fauna, dominated by bryozoans and calcareous worm tubes (serpulids and spirorbids), followed by cemented bivalves, barnacles, foraminifera, and algae (Rodland et al. 2004, 2006). Moreover, brachiopods are a ubiquitous fossil group and thus the study of their modern ecology can greatly aid paleobiological research over evolutionary time scales. In particular, brachiopod shells may preserve records of biotic interactions and thus yield information on evolutionary trends in ecological processes.

Thanks to the research efforts of the last decade, biotic interactions, especially those of predatory nature, are increasingly well documented in the Phanerozoic fossil record of shelly benthos (Kowalewski & Kelley 2002, Kelley et al. 2003, Bassett et al. 2004, Harper 2006). The majority of these data come from recognizable prey damage (e.g., drill holes on shells and smashed shells by crushing predators) or even inferences of functional morphology of predators (Harper 2006).

Most previous actuopaleontological studies focused on drill holes and repair scars, which are especially true for brachiopods (Delance & Emig 2004, Simees et al. in press). Those traces are usually interpreted as predatory in origin, although a possible parasitic origin for drill holes has been noted in multiple studies on fossil brachiopods (e.g., Baumiller et al. 1999, Leighton 2001, Hoffmeister et al. 2004, Kowalewski et al. 2005). Other types of traces that can be found on brachiopod shells, those recording commensal and parasitic activities, are understudied in recent ecosystems and poorly documented in the fossil record. Even in the case of much better studied present-day molluscs, studies of parasitism focus primarily on shellfish, which are harvested commercially.

This paper focuses on biotic interactions recorded in shells of modern brachiopods (Fig. 1) that are common in benthic ecosystems found around the San Juan Islands, Washington (Fig. 2). Locally, brachiopod assemblages composed of shells and live specimens of 4 species, Terebratalia tranversa (Sowerby), Terebratulina unguicula Carpenter, Laqueus californianus (Koch), and Hemithiris psittacea (Gmelin) (Fig. 1) are abundant (e.g., Thayer 1975, Schumann 1991, and Kowalewski et al. 2003) and easily accessible for sampling. They offer a unique opportunity to investigate biotic interaction affecting present-day brachiopod communities. The interactions can be evaluated by studying traces of bioerosion left by infesting organisms on brachiopod shells as well as by direct observation of living parasites found in association with brachiopod shells. This study aims to: (a) identify traces of bioerosion left on brachiopod shells by infesting organisms; (b) recognize the bioeroder's identity; and (c) investigate the distribution of traces to better understand the relationships between bioeroders and infested brachiopods.

[FIGURES 1-2 OMITTED]

MATERIALS AND METHODS

All samples were collected by dredging during three boat trips (July 23, 29, and August 12, 2004). Sampled sites ranged from 14.7-93.3 m in depth, and included the following locations: (a) the San Juan Channel between Lopez Island and San Juan Island (Trip 1, 3 sites, Fig. 2); (b) Iceberg Point, ~0.4 km off SW end of Lopez Island and Rock Point, Lopez Island (Trip 2, 2 sites, Fig. 2); and (c) sites between Sentinel and Spieden Islands and Steep Point, Orcas Island (Trip 3, 2 sites, Fig. 2). These localities were targeted because of previously documented abundance of brachiopods in those areas (e.g., Thayer 1975, Schumann 1991, Kowalewski & LaBarbera 2004), including shells of living specimens of Terebratalia tranversa, Terebratulina unguicula, and Laqueus californianus (e.g., Kowalewski et al. 2003), as well as dead shells of Hemithiris psittacea.

The bottom sediment at each collecting site (n = 7, Table 1) was sampled using mud and rock dredges. Whenever possible, the dredge haul was carried out parallel to isobaths to maintain reasonably constant bathymetry (see Table 1 for details). Collecting sites varied in substrate, with rocky substrate being slightly more frequent, and the muddy substrate restricted to shallower sites only (see Table 1).

For each collecting site, one or two buckets of 20 L were filled with the bulk sediment acquired by dredging (see Kowalewski et al. 2003, for more details about the sampling strategy). In addition, to increase the sample size, all live and dead brachiopods that could be recognized on the sorting table were hand picked and stored in zip lock bags.

The sampled bulk material was wet sieved through 12-mm and 4-mm mesh screens and air-dried at the Friday Harbor laboratories. After sieving, all shells were sorted to separate brachiopod specimens. During sample processing live specimens were kept in tanks with running seawater for 72 h. Subsequently, specimens were preserved in 70% ethanol, and stored separately for each collection site. All brachiopods killed during dredging and sieving were considered live-collected because of the presence of soft tissues.

The data collection involved exhaustive screening of all specimens for presence of traces of bioerosion (borings in the shell, see later for trace description), under 0.8x of magnification. All sampled brachiopod material was examined, including shell fragments. For this study, a fragmented shell was defined as a bioclastic particle representing less than 90% of a whole valve. In live specimens, the observations were restricted to the external shell surface only, but dead specimens were examined externally and internally.

Samples yielded shells and live individuals of Terebratalia tranversa, Terebratulina unguicula, and Laqueus californianus along with shells of Hemithiris psittacea (Fig. 1). All species were analyzed using the following protocol. Shell standard dimensions (length, width, height, and thickness) as well as anterior-posterior dimensions of all identifiable shell fragments were measured to the nearest 0.01 mm using a digital caliper.

The frequency of traces (TF) (i.e., proportion of specimens with at least one identifiable trace of a given type) was computed for pooled data and also separately for data grouped by: (a) brachiopod species; (b) collecting sites; (c) dead versus live-collected specimens by species and by collecting sites; (d) shell size; and (e) valve type. Statistical analyses of frequencies of traces were based on Fisher Exact (2-tailed) Test. Pearson correlation and binomial tests were also performed. All tests were executed using Statistical Analysis System (SAS, version 6), with a significance criterion of 5% ([alpha] = 0.05).

RESULTS

Out of seven collecting sites, five yielded brachiopods resulting in a total of 1,131 specimens (Table 1). Off these, 523 were live collected and 608 represented dead material. The most abundant and widespread brachiopod species across sampled sites was Terebratalia transversa (n = 536, [n.sub.live] = 265, and [n.sub.dead] = 271), followed by Terebratulina unguicula (n = 295, [n.sub.live] = 182, and [n.sub.dead] = 113), Laqueus californianus (n = 159, [n.sub.live] = 76, and [n.sub.dead] = 83), and Hemithiris psittacea ([n.sub.dead] = 141) (Table 1 and Table 2). All brachiopod species represent thin-shelled forms (Fig. 1), with the mean shell thickness of 0.3 mm (L. californianus), 0.5 mm (T. uniguicula), 0.6 mm (T. transversa), and 0.7 mm (H. psittacea), respectively. Externally, shells are smooth (L. californianus, H. psittacea) or ribbed (T. uniguicula, T. transversa).

Shells with traces of bioerosion were found in all sampling localities that yielded brachiopods. Out of 1,131 brachiopod shells, 91 showed horizontal borings represented by tubes that are parallel to the inner and outer shell surfaces. Dead (Fig. 3a to i) and live shells (Fig. 3j to m) show tubes of variable length, ranging from ~1.2 mm to ~17.7 mm. Both tube openings communicate with the outside of the shell, and are commonly located at the commissure edge (Fig. 3c). Tubes are U-shaped in cross-section and straight or slightly curved longitudinally, branching in some cases (Fig. 3d). A low central ridge can be seen only when tubes are exposed (Fig. 3d to e). In other words, the sides of the tube are deeper than the middle, resulting in a low central ridge. Multiple tubes (two or three) on the same shell are common (Fig. 3f). Tubes never penetrate the shell (Fig. 3g to i), which means that the borers never get direct access to soft tissues of the host. Moreover, on the internal surface of the infested brachiopod valves, blisters are observed extending along the entire length of the tube (Fig. 3g to i).

[FIGURE 3 OMITTED]

In the case of several live-collected specimens of the brachiopods T. transversa (Fig. 3j to m) and L. californianus, a living infesting parasite was found inside its domicile horizontal tube. Through the tube openings, a pair of long and coiled peristomial palps was observed, stretching out of the brachiopod shell.

Traces of bioerosion are nonrandomly distributed across brachiopod species. Out of 91 shells yielding tubes, 54 were found in T. tranversa, 26 in L. californianus, and 11 in T. unguicula. Tubes were not found in shells of H. psittacea (Table 2).

When all data are pooled (Table 2), the highest trace frequency (TF) is observed on shells of L. californianus (TF = 16.35%, [n.sub.total] = 159), followed by T. tranversa (TF = 10.07%, [n.sub.total] = 536), and T. unguicula (TF = 3.73%, [n.sub.total] = 295). All morphological types of shell surfaces (from smooth to ribbed) yielded tubes (Fig. 3).

Except for one site, the tubes are significantly more frequent on live specimens than on dead shells (P < 0.05; 2-tailed Fisher Exact Test; see also Table 3). However, this analysis pools all brachiopod species, and the brachiopod distribution is not homogeneous across the sites (Table 1). Infestations in T. transversa, T. unguicula, and L. californianus shells sampled at the same collecting site are observed only on sample of Trip2-Site1 (Table 2). When data are controlled for host species and sampling site, trace frequencies in live shells remain significantly higher than in dead ones for T. tranversa specimens (Fig. 4, p [much less than] 0.05, Fisher Exact Test). However, this pattern is not observed for T. unguicula (Fig. 4, P = 0.7538, Fisher Exact Test) and L. californianus (Fig. 4, P = 0.833, Fisher Exact Test) specimens.

[FIGURE 4 OMITTED]

Comparison of brachiopod shells with and without tubes in terms of shell size shows that there is a good agreement between shell size and the presence of tubes (Fig. 5). Tubes occur primarily in the largest size classes of shells, and this pattern is observed for all brachiopod species, for both dead shells and living specimens (Fig. 5). In addition, almost all size classes identified for living brachiopods are also represented as dead shells (Fig. 6).

[FIGURES 5-6 OMITTED]

Trace frequency was also investigated according to the brachiopod valve. When data are pooled across all species, ventral valves are preferentially infested (Fig. 7a, T[F.sub.ventral valves] = 62.6%, P = 0.011, Binomial Test). Moreover, this pattern persists in almost all brachiopod species, when each of them is examined separately. In the case of living specimens of T. tranversa, the infestation preference on ventral valves is nearly significant (Fig. 7b, T[F.sub.ventral valves] = 64.3%, P = 0.058, Binomial Test). The same is also true for living specimens of L. californianus, for which the preference for the ventral valve (Fig. 7c, T[F.sub.ventral valves] = 76.9%, P = 0.036, Binomial Test) is high. The exception is T. unguicula (Fig. 7d, T[F.sub.ventral valves] = 50.0%, P > 0.05, Binomial Test), for which no infestation preferential was observed.

Trace location on brachiopod valve was also taken into account. The tubes appear to be preferentially located (85.6%, n = 83, p [much less than] 0.05, Binomial Test) in the median deflection region of the shells. For L. californianus, this pattern is even more significant (n = 30, p [much less than] 0.05), with 90% (n = 27) of the tubes situated in the median deflection.

DISCUSSION

Identity of the Bioeroder

According to the descriptions of borers in calcareous substrates, horizontal, simple U-shaped tubes or complex branching burrows, parallel to the surface of the shell typically represent activities of annelid worms (Blake & Evans 1973, Zottoli & Carriker 1974, Bromley & D'Alessandro 1983, Bromley 1994, Martin & Britayev 1998). The u-shape tube morphology observed here is similar to that typically produced by polychaete annelids of the family Spionidae (Blake & Evans 1973, Bromley & D'Alessandro 1983, Bromley 1994). Spionidae (e.g., Polydora) are known to bore into bivalve molluscs (e.g., Crassostrea virginica), gastropod shells (e.g., Littorina littorea), corals (e.g., Leptastrea purpurea), and coralline algae (e.g., Lithothamnion sp.) (Blake & Evans 1973), among others.

Recently, boring traces made by Polydora sp., a spionid polychaete with long and coiled peristomial palps, and large setae on both sides of the fifth segment, were identified on shells of Bouchardia rosea from the Brazilian shelf (Rodrigues et al. 2005). The borings in B. rosea shells form distinct traces, comparable to the trace fossil Caulostrepsis (Rodrigues et al. 2005), with morphology similar to that observed here. The main difference is the common presence of internal blisters, which were found in this study on every examined dead shell and also on a living infested brachiopod that was dissected. The brachiopod response to the worm infestation is to create blisters of calcareous shell layer to wall them off. Hence, such structures seem to be related to thin shells of the brachiopods, because blisters are not present in the thicker shells of B. rosea from the Brazilian shelf.

Similarly to the infested living bouchardiids from Brazil (Rodrigues et al. 2005), live worms were often found in horizontal tubes of living specimens of T. transversa (Fig. 3j to m) and L. californianus. Through the tube openings, a pair of long and coiled peristomial palps was observed, stretching out of the brachiopod shell. This feature agrees well with the feeding apparatus of most spionid worms (e.g., Polydora sp., Martin & Britayev 1998). Additionally, the morphology of the peristomial palps is similar to that observed on the shell borers in B. rosea specimens (Rodrigues et al. 2005).

Although this biotic interaction can be host-specific and the shapes and locations of shell borings can be characteristic of a species, the infester is tentatively identified as belonging to the Polydora-group of species, often referred to as mudworms. This assumption is based on the similarities in anatomy of peristomial palps and virtually identical trace morphologies (see Rodrigues et al. 2005). Unfortunately, most of the diagnostic anatomical features of the worm were altered or lost after the animal had been fixed. In particular, diagnostic characters of the Polydora worm (e.g., modified setae of the fifth segment) were not well preserved or missing entirely.

Stereotypy of Infestation

Given that the highest trace frequency was not observed for the most abundant brachiopod species, even considering a sampling site where brachiopod species are well represented, except for H. psittasea (e.g., Trip2-Site 1, Tables 1 and 2, see also Fig. 4), infestation appears to be species selective. In addition, infestation seems to occur on shells of a particular size, that is, borers preferably infest larger shells (see Fig. 5). This is not a taphonomic artifact because the studied assemblages are not biased toward a particular size class (see Fig. 6). In other words, host size does matter for the shell borer and, probably, is a limiting factor or even may reflect the chance of colonization between hosts and spionids.

The infestation observed here can be compared with traces of bioerosion identified in the fossil record. Vinn (2005) described borings produced by worms on Ordovician brachiopod shells from North Estonia. Out of 21 brachiopod genera studied, nine showed borings, with frequencies varying from 6.5% in Bekkerina to 51% in Estlandia (Vinn 2005). No common morphological feature other than shell size and valve thickness distinguished brachiopods with borings from those without them (Vinn 2005). The worm preferred large hosts with thick lamellose shells (>0.5 mm thick) (Vinn 2005). These patterns agree only partly with the results reported here. Similarly to Vinn's data, larger shells are more frequently infested by the spionid worms (Fig. 5). However, all infested brachiopod are thin-shelled, with shell thickness varying from 0.3 mm to 0.6 mm thick. Moreover, the highest trace frequency was observed for a species with the thinnest shell. Consequently, shell thickness need not be a limiting factor for the borers of recent brachiopods from the San Juan Islands. As highlighted by Vinn (2005), ornamentation may have also influenced the borer larva's choice for substrate. However, whereas worm borings on Ordovician brachiopods from Estonia occur preferentially on strongly to moderately costate shells (Vinn 2005), the highest boring frequency on Recent brachiopods from the San Juan Islands is observed on smooth shells (i.e., L. californianus). Shells of B. rosea from Brazil are also smooth, and yet trace frequency can reach 18 % (Rodrigues et al. 2005). The data presented here and the ones available for B. rosea indicate that thickness and relief of brachiopod shells need not be decisive factors in determining substrate preference by boring worms. Possibly, ecological factors and abiotic conditions may play important roles in the biotic interactions between spionid worms and brachiopods. Determining the role of all those factors is a difficult task, which is beyond the scope of this study. However, it may be worth noting in this context, that, in the study area, the biotic relation does not seem to be highly constrained in terms of water depth and substrate, as borings are observed over a substantial bathymetric range (from 18.3 93.3 m) and for various bottom types (from muddy to rocky substrates).

Higher frequencies of borings in live specimens may be potentially explained by several hypotheses or by combination of them: (1) Traces on brachiopod shells result from infestation by spionid worms only on living hosts. Infestation on living brachiopods seems advantageous, because living hosts would provide better protection from sedimentation, overturning, breakage or abrasion than empty shells could offer (Pickerill 1976). Note here that brachiopods were sampled at sites characterized by high-energy bottom currents (e.g., San Juan Channel, Messina & Labarbera 2004). In addition, all four brachiopod species are epifaunal, sessile forms strongly attached to the substrate. Preferential infestation of ventral valves is also consistent with this hypothesis. This pattern is unlikely to be because of taphonomic bias because high infestation on ventral valves was also observed when data were restricted to living specimens only. Also, the concentration of traces in larger size classes of the host may reflect increasing chance of colonization with ontogentic age. In addition, live brachiopods revealed infestation by living spionids (Fig. 3j to m), indicating that this biotic interaction begins when the host is still alive. The fact that traces are more frequent on live specimens indicates strong preference for larvae to settle on living shells. This is because many dead shells stay for prolonged time intervals around the sediment-water interface (e.g., Davies et al. 1989, Olszewski 2004), and thus would be available for infestation for many generations of larvae. It is noteworthy that recent AMS radiocarbon dating of brachiopod shells from the study area indicates the presence of shells as old as 154 y BP (Kidwell 2005). Finally, none of the traces open inward, suggesting that the spionids always entered the shells from the outside and all traces found on dead shells also show blisters, suggesting that infestation occurred when those brachiopods were still alive. (2) Traces could reflect a very recent biological interaction, between T. tranversa, T. unguicula, L. californianus, and Polydora sp. that is not documented by the dead shells yet. For example, if Polydora invaded the study area in the last several years, the high frequency of infestation in live specimens could have resulted. Even with low frequency, the occurrence of a few traces on dead shells (30 of 608) undermines this hypothesis, but further investigation is necessary for two reasons. There are no data available about the structure of time-averaging (for the only brachiopod-focused case study, see Carroll et al. 2003) of the studied dead assemblages. Only a few shells of H. psittacea were dated, indicating the presence of slightly older shells (154 y BP) in the studied thanathocoenosis (Kidwel12005). (3) Infestation is temporary and traces on shells can be healed after infestation but before the brachiopod dies, making the recognition of infested dead shells difficult. However, given that all infested shells showed distinct blisters (including all dead shells), it is difficult to support this hypothesis. (4) Trace frequencies on dead shells may be underestimated. Trace identification on dead shells depends on the stage of infestation. Early stages of infestation result in short traces restricted primarily to the commissure edge and/or the median deflection of the shell. Actually, the commissure is the thinnest part of the shell and, consequently, more prone to fragmentation. Thus, the preservational potential of such traces in dead shells may be low, making it difficult to recognize infested dead shells. On the other hand, traces of late stages of infestation are easily recognizable, especially if the shell is abraded and/or corroded, exposing the tubes (Fig. 3d to e). Moreover, when infested shells are subjected to taphonomic alterations, shells tend to break along the surface defined by the worm traces. Exposed tube walls and shell fragments that broke off along the traces were also observed on infested shells of B. rosea, from the Brazilian shelf (Rodrigues et al. 2005). Similar to Rodrigues et al. (2005), fragments with distinctive, biotic source-dependent breakage patterns ("biologically-facilitated fragmentation" of Rodrigues et al. 2005) that can be easily identified were also recognized in the material studied here. If tubes cause shells to break more readily, it would be predicted that in a high energy environment (e.g., San Juan Channel), infested shells should fragment more than noninfested ones. Although infestation may weaken the living shells, traces of bioerosion can be easily identified in the dead assemblages, even in a fragmented shell, as demonstrated by distinct "biologically-facilitated" breakage patterns. However, differences in preservational potential between the infested and noninfested shells would be observed at a taphocoenosis-level, especially if those shells stay in taphonomically active zone for a prolonged time. In this situation, there likely is a taphonomic bias operating against preserving the infested shells. Thus, this hypothesis predicts that in time-averaged death assemblage shells with traces tend to be under represented relative to shells without traces, in the fossil record.

Ecological and Evolutionary Significance of Traces

The biotic interaction reported here between modern brachiopods and spionids may represent either commensalisms or parasitism. The interaction is clearly beneficial to spionids because the host provides an elevated, immobile, and strongly attached substrate in areas dominated by high-energy currents. Furthermore, the preferential location of tubes around the median deflection of the shell minimizes contact with brachiopod metabolic wastes while maximizing access to feeding resources. This indicates that spionids may benefit from brachiopod-induced currents, but need not be exclusively dependent on their hosts in getting their food. The effect of this interaction on brachiopod hosts is more difficult to evaluate because it can either be neutral (commensalism) or negative (parasitism). Blisters found on infested brachiopods indicate that the host responds biologically to the infester. The energetic cost associated with blisters may be detrimental to brachiopods, but the quantitative value of this cost is not known at this point so its significance is difficult to evaluate. In addition, the feeding activity of worms may result in a significant loss of nutrients to brachiopods (kleptoparasitism, see Iyengar 2002, for an example of kleptoparasitism by gastropods on polychaetes). In sum, traces on the brachiopod shells can either represent commensalism or parasitism. Further studies are needed to determine which of those two interactions is dominant.

The distinct preservable traces left by spionids may potentially provide a rich fossil record of parasitic interactions. The fossil record of parasitism is still poorly documented and spionid-brachiopod system may provide an important model for studying the long-term evolutionary history of parasite-host (and/or commensal) interactions. Also, specifically in the case of brachiopods, biotic interactions with spionid worms may provide important insights into the history of this important metazoan group with an extensive fossil record. In the brachiopod fossil record only a limited number of possible examples of parasitism are known (see Baumiller & Gahn 2002, for literature review). However, these records span back all the way to Lower Cambrian. A specimen of the lingulate brachiopod Linnarssonia constans, from the late Lower Cambrian, Shabakty Group of the Malyi Karatau Range in Kazakhstan, Central Asia, preserves evidence of infestation within the mantle cavity by a vermiform animal, which is recorded by internal tubular protuberance (Bassett et al. 2004). In addition, some Devonian spiriferids from Australia show changes in the outer mantle cavity that were apparently caused by the presence of a small, filter-feeding organism (Chatterton 1975). Notably, all the six illustrated specimens had horizontal tubes that are very consistent in their morphology. This is also the case of the brachiopods from the San Juan Islands, which all infested shells of L. californianus, T. transversa, and T. unguicula specimens yielded similar U-shaped tubes. Also, similar to the tubes described here, the Devonian tubes were located in the median deflection of brachiopod shells (Chatterton 1975) and were interpreted as commensal and/or parasitic in origin, possibly caused by activity of polychaetes, tunicates, cnidarians, ctenophores, or arthropods.

Implications of Biotic Interaction for Conservation Paleobiology 

Spionid polychaetes are ubiquitous shell-borers in many mollusc species and their interactions have been studied, with a particular focus on commercially important oyster species (Blake & Evans 1973, Skeel 1979, Lauckner 1983). Also, scallops, abalones, and mussels, among other shellfishes, often have a number of unwanted spionid polychaetes living on their shells. Some are harmless, and none actually eat their hosts, but shell-boring spionids can do severe damage, especially considering the economic importance of mollusc cultivation worldwide. The problem is that worm infestation triggers shell malformations, sometimes resulting in an altered shell shape, damaged adductor muscles, energy wasted on shell repair, increased vulnerability to serious pathogens, reduction in shell strength, and internal shell blisters (Read 2004). Besides weakening the shell, blisters also detrimentally affect the mollusc flesh by releasing anaerobic metabolites including hydrogen sulphide (e.g., Handley 1995). This biotic interaction between marine shells and shell-borer spionids has an obvious relevance to shellfish industry, because the end product can be unsaleable to consumers (Read 2004).

As noted by Wargo & Ford (1993), Handley (1997), and Caceres-Martinez et al. (1999), spionid outbreaks have previously been associated with oyster mortality in Australia. On the other hand, spionid infestation may have had no significant adverse health impact on molluscs in other cases (e.g., Clavier 1989, Handley & Bergquist 1997, Caceres-Martinez et al. 1998). These contrasting conclusions may be because of differences in the severity of infestation (measured by spionid count or blister damage), the size of the host, host species, and environmental factors such as food availability (Lleonarta et al. 2003). Otherwise, the infestation by shell-boring polychaetes on molluscs are believed to affect mostly thick shells, but one exception was the infestation observed on thin shells of farmed green-lipped mussels (Perna canaliculus) from New Zealand (Read & Handley 2004).

Shell thickness seems to be an important factor for spionid infestation on molluscs (Read & Handley 2004). For other shelly invertebrates, such as brachiopods, shell thickness does not seem to be a limiting factor. Spionid infestations can affect brachiopods such as the extant brachiopod Bouchardia rosea from the Brazilian shelf (Rodrigues et al. 2005), a thick-shelled species (see Brunton 1996). However, all infested brachiopod species documented in this study have thin, smooth to ribbed shells. In both cases, the brachiopod shells are heavily infested, with infestation frequencies reaching 18% (Rodrigues et al. 2005) and 16.35%, respectively. Multiple traces may be present in a single infested shell, but no more than three traces in the same shell are observed in the studied material, and no severe damage could be directly identified. Thus, the impact of spionid infestation on the health of living brachiopods needs further investigations. Apart from that, the data raise some interesting questions. If the infestation traces reflect a very recent biological interaction between brachiopods and spionids, as inferred by the lower trace frequencies on dead accumulations (see Table 2, Fig. 4), at least for T. tranversa shells, this biotic interaction could be a result of recent changes in local epifaunal assemblages as well as in the abiotic environmental conditions (e.g., contamination by ship ballast water, eutrophization). For example, studies on the sewage impact on the composition and distribution of polychaetes conducted for Southwestern Atlantic settings (Elias et al. 2003) revealed positive and significant correlation between spionid worms and total content of organic carbon. In addition, some spionids have largely been used as indicators of organically enriched sediments (e.g., Pocklington & Well 1992). Unfortunately, neither the precise identity of the infester, nor the ecological historical data for San Juan Islands benthos are available, at this point.

CONCLUSIONS

The most abundant and widespread brachiopod species sampled is Terebratalia transversa (n = 536, [n.sub.live] : 265, and [n.sub.dead] = 271), followed by Terebratulina unguicula (n = 295, [n.sub.live] = 182, and [n.sub.dead] = 113), Laqueus californianus (n = 159, [n.sub.live] = 76, and [n.sub.dead] = 83), and Hemithiris psittacea ([n.sub.dead] = 141).

Brachiopod shells with traces of bioerosion were found in all sampling sites that yielded brachiopods. The infested shells occur over a wide depth range (18.3-93.3 m) and are present on muddy to rocky bottoms, but mostly on rocky substrates.

Traces of bioerosion are represented by tubes that are parallel to the inner and outer shell surfaces, and never penetrate across the shell into the brachiopod's body cavity. Both dead and live shells show tubes of variable length, ranging from ~1.2 mm to ~17.7 mm. Tubes are U-shaped, straight or slightly curved, sometimes branched. Both tube openings communicate with the outside of the shell, and are commonly located at the commissure edge. Blisters on the inner side of the infested valve are always observed along the entire length of the tube. Multiple tubes (2 or 3) on the same shell are common.

Tubes are nonrandomly distributed across infested species. The highest trace frequency (TF) is observed on shells of L. californianus (TF = 16.35%, [n.sub.total] = 159), followed by T. tranversa (TF = 10.07%, [n.sub.total] = 536), and T. unguicula (TF = 3.73%, [n.sub.total] = 295). Tubes were not found in shells of H. psittacea. These patterns suggest that bioeroders are species selective in choosing their brachiopod hosts.

Smooth and ribbed, thin-shelled brachiopods show tubes, indicating that shell relief and thickness are not limiting factors for the bioeroders.

Traces are significantly more frequent on live-collected specimens and ventral valves. Also, tubes are more frequent in larger shells, and preferentially situated in the median deflection region of the shells. These patterns suggest site selectivity by the infesting worms.

Blister occurrences on every infested dead shell and also on a living infested brachiopod, that was dissected, indicate that worm infestations are restrict to living hosts, and so these traces of bioerosion represent a truly biotic interaction, between living hosts and living infestors.

Observations on live infesting organisms found on some live-collected brachiopods revealed a pair of long and coiled peristomial palps stretching out of the brachiopod shell. Based on the similarities in both peristomial palps and morphology of dwelling tubes to previously documented worms infesting brachiopods from the Brazilian shelf (Rodrigues et al. 2005), the traces documented here are tentatively attributed to spionid polychaetes (Polydora sp.).

Tubes on brachiopod shells may record commensal or parasitic relation between brachiopods and spionids. Because these tubes cart be readily preserved and identified in the fossil record, they offer a promising window into the evolutionary history of parasitic/commensal interactions between brachiopod hosts and their infestors.

ACKNOWLEDGMENTS

This study was conducted as a project in the scope of the Summer Course on the Predator-Prey Interactions, conducted at Friday Harbor Laboratories, University of Washington, July to August 2004. The author thanks Richard Krause (Virginia Polytechnic Institute and State University at Blacksburg, VA, USA), classmates and FHL staff for all help; Michal Kowalewski and Lindsey Leighton for their supervision, enthusiasm, and careful revision of the manuscript; Allan P. Hoffmeister for his thoughtful insights; and two anonymous reviewers for their detailed and helpful revisions. The author also extends special thanks to Marcello Guimaraes Simoes (Sao Paulo State University at Botucatu, SP, Brazil) who encouraged and helped her to attend the course, and for his suggestions on the earlier drafts of this manuscript. Financial support was provided by The University of Washington, and State of Sao Paulo Research Foundation (FAPESP, Proc. 02/13552-7).

LITERATURE CITED

Bassett, M. G., L. E. Popov & L. E. Holmer. 2004. The oldest known metazoan parasite? J. Paleontol. 78:1214-1216.

Baumiller, T. K. & F. J. Gahn. 2002. Fossil record of parasitism on marine invertebrates with special emphasis on the platyceratid-crinoid interaction. In: M. Kowalewski & P. H. Kelley, editors. The fossil record of predation, Paleontological Society Special Papers, 8. New Haven: Paleontological Society, Yale Printing Service. pp. 195-210.

Baumiller, T. K., L. R. Leighton & D. L. Thompson. 1999. Boreholes in Mississsipian brachiopods and their implications for Paleozoic gastropod drilling. Palaeogeogr. Palaeoclimatol. Palaeoecol. 147:283-289.

Blake, J. A. & J. W. Evans. 1973. Polydora and related genera as borers in mollusc shells and others calcareous substrates. Veliger 15:235-249.

Bromley, R. G. 1994. The palaecology of bioerosion. In: S. K. Donovan, editor. The palaeobiology of trace fossils. Chichester: John Wiley & Spons. pp. 134-1154.

Bromley, R. G. & A. D'Alessandro. 1983. Bioerosion in the Pleistocene of southern Italy: ichnogenera Caulostrepsis and Meandropolydora. Riv. Ital. Paleontol. Strat. 90:227 296.

Brunton, C. H. C. 1996. The functional Morphology of the Recent brachiopod Bouchardia rosea. Acta Zool-Stockholm. 77:233-240.

Caceres-Martinez, J., P. M. M. De Oca & R. Vasquez-Yeomans. 1998. Polydora sp. infestation and health of the Pacific Oyster Crassostrea gigas cultured in Baja California, NW Mexico. J. Shellfish Res. 17:259-264.

Caceres-Martinez, J., G. D. Tinico, M. L. Bustamente & I. M. Gomez-Humaran. 1999. Relationship between the burrowing worm Polydora sp. and the black clam Chione fluctifraga Sowerby. J. Shellfish Res. 19:85-89.

Carroll, M., M. Kowalewski, M. G. Simoes & G. A. Goodfriend. 2003. Quantitative estimates of time-averaging in brachiopod shell accumulations from a modern tropical shelf. Paleobiology 29:382-403.

Chatterton, B. D. E. 1975. A commensal relationship between a small filter feeding organism and Australian Devonian spiriferid brachiopods. Paleobiology 1:371-378.

Clavier, J. 1989. Infestation of Haliotis tuberculata shells by Cliona celata and Polydora species. In: S. A. Shepherd, M. J. Tegner & S. A. Guzman del Proo, editors. Abalone of the world, biology, fisheries and culture--supplementary papers. Proceedings of the 1st International Symposium on Abalone, La Pez, Mexico. November 21-25. pp. 16-20.

Davies, D. J., E. N. Powell & R. J. Jr. Stanton. 1989. Taphonomic signature as a function of environmental process: shells and shell beds in a hurricane-influenced inlet on the Texas coast. Palaeogeogr. Palaeoclimatol. Palaeoecol. 72:317-352.

Delance, J. H. & C. C. Emig. 2004. Drilling predation on Gryphus vitreus (Brachiopoda) off the French Mediterranean coasts. Palaeogeogr. Palaeoclimatol. Palaeoecol. 208:23-30.

Elias, R., M. S. Rivero & E. A. Vallarino. 2003. Sewage impact in the composition and distribution of Polychaeta associated to intertidal mussel beds of the Mar Del Plata rocky shore, Argentina. lheringia (SOr. Zool.) 93:309-318.

Gosling, E. 2003. Bivalve molluscs. Biology, ecology and culture. Oxford: Fishing News Books, Blackwell. 441 pp.

Handley, S. J. 1995. Spionid polychaetes in Pacific oysters, Crassostrea gigas (Thunberg) from Admiralty Bay, Marlborough Sounds, New Zealand. N. Z. J. Mar. Freshwater Res. 29:305-309.

Handley, S. J. 1997. Spionid polychaete worm infestations of the Pacific oyster Crassostrea gigas in New Zealand. PhD Thesis, University of Auckland, June 1997. 151 pp.

Handley, S. J. & P. R. Bergquist. 1997. Spionid polychaete infestations of intertidal pacific oysters Crassostrea gigas (Thunberg), Mahurangi Harbour, northern New Zealand. Aquaculture 153:191-205.

Harper, E. P. 2006. Dissecting post-Palaeozoic arms races. Palaeogeogr. Palaeoclimatol. Palaeoecol. 232:322-343.

Hoffmeister, A. P., M. Kowalewski, R. K. Bambach & T. K. Baumiller. 2004. Drilling predation in brachiopods and bivalve molluscs from the Permian Strata of the Glass Mountains, West Texas. Acta. Palaeontol. Polonica 49:443-459.

Iyengar, E. V. 2002. Sneaky snails and wasted worms: kleptoparasitism by Trichotropis cancellata (Mollusca, Gastropoda) on Serpula columbiana (Annelida, Polychaeta). Mar. Ecol. Prog. Ser. 244:153-162.

Kelley, P. H., M. Kowalewski & T. A. Hansen. 2003. Predator-prey interactions in the fossil record. New York: Kluwer Academic/Plenum Press. 464 pp.

Kidwell, S. M. 2005. Brachiopod versus bivalve radiocarbon ages: implications for Phanerozoic trends in time-averaging and productivity. Geol. Soc. Amer. (Abstract with Programs) 37:117.

Kowalewski, M. & P. H. Kelley. 2002. The fossil record of predation. Paleontological Society Special Papers 8. New Haven: Paleontological Society, Yale Printing Service. 398 pp.

Kowalewski, M. & M. LaBarbera. 2004. Actualistic taphonomy: death, decay, and desistegration in comtemporary settings. Palaios 19: 423-427.

Kowalewski, M., M. Carroll, L. Casazza, N. S. Gupta, B. Hannisdal, A. Hendy, R. A. Krause, M. LaBarbera, D. G. Lazo, C. Messina, S. Puchalski, T. A. Rothfus, J. Salgeback, J. Stempein, A. C. Terry & A. Tomasovych. 2003. Quantitative fidelity of brachiopod-mollusc assemblages from modern subtidal environments of San Juan Islands, USA. J. Taphonomy 1:43-65.

Kowalewski, M., A. P. Hoffmeister, T. K. Baumiller & R. K. Bambach. 2005. Secondary evolutionary escalation between brachiopods and enemies of other prey. Science 308:1774-1777.

Lauckner, G. 1983. Diseases of mollusca: bivalvia, agents: annelida. In: O. Kinne, editor. Diseases of marine animals, 2. Hamburg: Biologische Anstalt Helgoland. pp. 805-816.

Leighton, L. R. 2001. New example of Devonian predatory boreholes and the influence of brachiopod spines on predator success. Palaeogeogr. Palaeoclhnatol. Palaeoecol. 165:53-69.

Lleonarta, B. M., J. Handlingera & M. Powell. 2003. Spionid mudworm infestation of farmed abalone (Haliotis spp.). Aquaculture 221:85-96.

Martin, D. & T. A. Britayev. 1998. Symbiotic polychaetes: review of known species. Oceanogr. Mar. Biol. Annu. Rev. 36:217-340.

Messina, C. & M. Labarbera. 2004. Hydrodynamic behavior of brachiopod shells: experimental estimates and field observations. Palaios 19:441-450.

Olszewski, T. D. 2004. Modeling the influence of taphonomic destruction, reworking, and burial on time-averaging in fossil accumulations. Palaios 19:39-50.

Pickerill, R. K. 1976. Vermiforichnus borings from the Ordovician of Central Wales. Geol. Mag. Cambridge 113:159-164.

Pocklington, P. & P. G. Well. 1992. Polychaetes. Key taxa for marine environmental quality monitoring. Mar. Pollut. Bull. 24:593-598.

Read, G. 2004. Compiler. Guide to New Zealand shell polychaetes. Web publication. <http://biocollections.org/pub/worms/nz/Polychaeta/ ShellsPolychaeta.htm>

Read, G. & S. Handley. 2004. New alien mudworm now becoming a pest in longtine mussels. Mar. Biosecurity 12:30-31.

Rodland, D., M. Kowalewski, M. Carroll & M. G. Simres. 2004. Colonization of a "Lost World": encrustation patterns in modern subtropical brachiopod assemblages. Palaios 19:384-399.

Rodland, D., M. Kowalewski, M. Carroll & M. G. Simres. 2006. The temporal resolution of epibiont assemblages: are they ecological snapshots or overexposures? J. Geol. 114:313-324.

Rodrigues, S. C., M. G. Simres & M. Kowalewski. 2005. Fragmentation and bioerosion as taphonomic records of biotic interactions: a case study of extant brachiopods (Bouchardiidae) from Brazil. In: 2nd International Meeting Taphos'05. Abstracts: pp. 135-136.

Sato-Okoshi, W. & K. Okoshi. 1997. Survey of the genera Polydora, Boccardiella and Boccardia (Polychaeta, Spionidae) in Barkley Sound (Vancouver Island, Canada), with special reference to boring activity. Bull. Mar. Sci. 60:482-493.

Sato-Okoshi, W. & K. Okoshi. 2000. Structural characteristics of self-excavated burrouws by boring polydorid species (Polychaeta, Spionidae). Bull. Mar. Sci. 67:235-248.

Schumann, D. 1991. Hydrodynamic influences in brachiopod shell morphology of Terebratalia transversa from the San Juan Islands, USA. In: D. I. MacKinnon, D. E. Lee & J. D. Campbell, editors. Brachiopods through time: proceedings of the International Brachiopod Congress 2, Dunedin, New Zealand: A. A. BGalkema, Rotterdam. pp. 265-271.

Simoes, M. G., M. Kowalewski, L. H. C. Mello, D. L. Rodland & M. Carroll. 2004. Present-day terebratulid brachiopods from the southern Brazilian Shelf: paleontological and biogeographic implications. Palaeontology 47:515-533.

Simoes, M. G., S. C. Rodrigues & M. Kowalewski. Comparative analysis of drilling frequencies in recent brachiopod-mollusc associations from the Southern Brazilian Shelf. Palaios (in press).

Skeel, M. E. 1979. Shell-boring worms (Spionidae: Polychaeta) infecting cultivated bivalve molluscs in Australia. Proc. Worm Maric. Soc. 10:529-533.

Thayer, C. W. 1975. Strength of pedicle attachment in articulate brachiopods: ecologic and paleoecologic significance. Paleobiology 1:388-399.

Vinn, O. 2005. The distribution of worm borings in brachiopod shells from the Caradoc Oil Shale of Estonia. Carnets Geol., Not. Geology 3:1-11.

Wargo, R. N. & S. E. Ford. 1993. The effect of shell infestation by Polydora sp. and infection by Haplosporidium nelsoni (MSX) on the tissue conditions of oysters, Crassostrea virginica. Estuaries 16:229-234.

Ward, T. M., S. J. Sorokin, D. R. Currie, P. J. Rogers & L. J. McLeay. 2006. Epifaunal assemblages of the eastern Great Australian Bight: effectiveness of a benthic protection zone in representing regional biodiversity. Cont. Shelf Res. 26:25-40.

Zottoli, A. & M. R. Carriker. 1974. Burrow morphology, tube formation, and microarchitecture of shell dissolution by the Spionid Polychaete Polydora websteri. Mar. Biol. 27:307-316.

Figure 7. Frequencies of traces on dorsal and ventral valves. (a) all
infested species, including live and dead specimens; (b) live infested
 specimens of Terebratalia transversa; (c) live infested specimens of
Laqueuss californianus; (d) live infested specimens of Terebraulina
unguicula. Note the preferential distribution of traces on ventral
valves. See the numbers of infested live specimens in Table 2.

     ventral    dorsal

a     62.6%     37.4%
b     64.3%     35.7%
c     76.9%     23.1%
d     50.0%     50.0%

Note: Table made from pie chart.

TABLE 1.

Summary of Sampling Sites. The total number of brachiopod shells
reported includes live-collected specimens, empty shells, and shell
fragments.

                          Brachiopods

 Sample ID     Total    Tt     Tu     Lc     Hp

Trip1-Site1       0      --     --     --    --
Trip1-Site2     151     134      0      0    17
Trip1-Site3       0      --     --     --    --
Trip2-Site1     463      24    283    151     5
Trip2-Site2     225     122      7      2    94
Trip3-Site1     181     154      4      5    18
Trip3-Site2     111     102      1      1     7

                         Geographic Location

 Sample ID        Location          Starting Point

Trip1-Site1    Near Friday       48[degrees] 32.56'N
                 Harbor          123[degrees] 00.76'W
                 Laboratories
Trip1-Site2    San Juan          48[degrees] 29.33'N
                 Channel         122[degrees] 57.03'W
Trip1-Site3    Flat Point,       48[degrees] 32.98'N
                 Lopez Island    122[degrees] 35.37'W
Trip2-Site1    Iceberg Point,    48[degrees] 24.821'N
                 Lopez Island    122[degrees] 52.674'W
Trip2-Site2    Rock Point,       48[degrees] 29.800'N
                 Lopez Island    122[degrees] 57.163'W
Trip3-Site1    Between           48[degrees] 38.56'N
                 Sentinel/       122[degrees] 9.09'W
                 Spieden Is.
Trip3-Site2    Steep Point,      48[degrees] 36.66'N
                 Orcas Island    123[degrees] 1.77'W

               Geographic Location
                                         Depth
 Sample ID         Ending Point        (meters)       Bottom sed.

Trip1-Site1    48[degrees] 32.65'N     17.4-32.9    Muddy substrate
               123[degrees] 00.50'W
Trip1-Site2    48[degrees] 29.65'N     73.2-76.8    Rocky substrate
               122[degrees] 56.96'W
Trip1-Site3    48[degrees] 32.71'N     18.3-62.2    Muddy substrate
               122[degrees] 55.74'W
Trip2-Site1    48[degrees] 24.885'N    93.3         Rocky substrate
               122[degrees] 52.97'W
Trip2-Site2    48[degrees] 29.599'N    89.6         Rocky substrate
               122[degrees] 57.05'W
Trip3-Site1    48[degrees] 38.47'N     36.6-54.9    Rocky substrate
               122[degrees] 8.72'W
Trip3-Site2    48[degrees] 36.49'N     29.3-24.7    Muddy substrate
               123[degrees] 1.47'W

Tt = Terehratalia tranversa, Tu = Terebratulina unguicula, Lc = Laqucus
californianus, Hp = Hemithiris psittacea.

TABLE 2.

Number of infested and noninfested, live and dead shells per species
and per sampling sites. Frequency of infestation traces (TF) is
estimated by species, pooling data across all sampling sites.

                                Brachiopod Shells (n)

                      L. californianus           T. transversa

                     Live         Dead        Live           Dead

Collecting Site    I     NI     I     NI     I     NI      I     NI

  Trip1-Site2       0     0      0     0    16     39       7    72
  Trip2-Site1      13    62     13    63     6      9       0     9
  Trip2-Site2       0     1      0     1    10     59       4    49
  Trip3-Site1       0     0      0     5     6     75       1    72
  Trip3-Site2       0     0      0     1     4     41       0    57
  Total               76           83          265265     271
  TF (%)                   16.35                     10.07

                             Brachiopod Shells (n)

                       T. unguicula           H. psittacea

                     Live        Dead       Live       Dead

Collecting Site    I     NI    I     NI    I    NI    I    NI

  Trip1-Site2      0      0    0      0    0     0    0    17
  Trip2-Site1      6    169    5    103    0     0    0     5
  Trip2-Site2      0      5    0      2    0     0    0    94
  Trip3-Site1      0      2    0      2    0     0    0    18
  Trip3-Site2      0      0    0      1    0     0    0     7
  Total              182         113          0  0      141
  TF (%)                   3.73                     0

I = Infested, NI = Noninfested.

TABLE 3.

Number of infested and noninfested shells per collecting sites (data
pooled across all brachiopod species for both live and dead
brachiopods). Infestation is significantly more frequent on live than
on dead specimens, according to Fisher Test (2-tailed), except for
Trip2-Site1 (see text for explanation).

                             Live                        Dead

Collecting Site    Infested    Non-Infested    Infested    Non-Infested

  Trip1-Site2         16            39             7            89
  Trip2-Site1         25           240            18           180
  Trip2-Site2         10            65             4           146
  Trip3-Site1          6            77             1            97
  Trip3-Site2          4            41             0            66

                       Total
                                   Fisher's Test
Collecting Site    Live    Dead    (two- tailed)

  Trip1-Site2       55      96      p = 0.00006
  Trip2-Site1      265     198      p = 1
  Trip2-Site2       75     150      p = 0.003
  Trip3-Site1       83      98      p = 0.049
  Trip3-Site2       45      66      p = 0.025

SABRINA COELHO RODRIGUES

Instituto de Biociencias, Universidade Estadual Paulista, Distrito de Rubiao Junior, CP. 510, 18.610-000, Botucatu, SP, Brazil

E-mail: [e-mail]