Seasonal abundance of the shipworm Neoteredo reynei ( Bivalvia , Teredinidae ) in mangrove driftwood from a northern Brazilian beach

Shipworms are important decomposers of wood, especially in mangrove forests where productivity is high. However, little emphasis has been given to the activity of shipworms in relation to the export of nutrients from mangroves to adjacent coastal areas. As a first step to obtaining such information, the frequency of colonized mangrove driftwood as well as shipworm density and length were studied by collecting washed up logs during a year at Ajuruteua beach, state of Pará, northern Brazil. A single species, Neoteredo reynei (Bartsch, 1920), was found colonizing driftwood. Although large colonized logs were most common on the beach, shipworm density was higher in small logs, especially during the dry season. In general, however, density was higher during the wet season (January to April) and lowest in July. Overall shipworm mean length was 9.66cm. In large logs, mean length increased between the wet and dry seasons. However, there was no difference in length among log size categories. Mean shipworm length was similar throughout most of the year but tended to be greater in July. Although salinity varied between 10.9 and 40 during the year, no relationship was found between salinity and density or length. The results suggest that shipworm activity in driftwood logs is relatively constant throughout the year. Increased air humidity and rainfall may promote survival during the wet season. Large logs may take longer to colonize and thus have lower densities than small ones which are scarce probably because they are destroyed rapidly by shipworm activity. However, data on the disintegration of logs would be necessary to test this hypothesis. Larger size of shipworms in the dry season may be related to growth after an earlier recruitment period. Shipworms in large logs during the dry season may be better protected from dessication and high temperatures by the insulating properties of the larger volume of wood.

The wood-boring Teredinidae ("shipworms") are highly specialized boring bivalves (TURNER, 1966), most of which colonize wood during the free-swimming larval stage and have both economic (TURNER, 1984;TANAL & MATLIN, 1997) and ecological importance (TURNER, 1984;KOHLMEYER et al., 1995).A detailed account of Brazilian Teredinidae is given in MÜLLER & LANA (2004).Studies carried out in Brazil show that the burrowing activity of Teredinidae varies to differing degrees in a wide range of different types of wood (STILLNER & PEDROSO, 1977;JUNQUEIRA et al., 1991;REIS, 1995) and is responsible for damage to wharfs, boats and other man-made structures (FERNANDES & COSTA, 1967).This is especially important in coastal areas of developing countries where the use of wood for housing, fishing boats and fish traps is very common (BARLETTA et al., 1998;KRAUSE & GLASER, 2003).
On the other hand, shipworms cause the breakdown of wood into finer organic material thus contributing to nutrient recycling and reducing the buildup of woody debris in coastal areas and estuaries (TURNER, 1984).Cellulose degrading and nitrogen fixing bacteria associated with the gills of shipworms (WATERBURY et al., 1983) along with morphological evidence (TURNER, 1966) suggest that wood is used as a food by shipworms.However, although all shipworms bore into wood, the ability to use wood as a food source may vary among species (TURNER, 1966).The detritivorous action of shipworms is extremely important in mangroves (KOHLMEYER et al., 1995) where there is a high level of production of woody biomass (GONG & ONG, 1990).Within the mangrove, much of this biomass, composed of fallen trunks, branches, and even aerial roots (SANCHEZ-ALFEREZ & ALVAREZ- LEON, 2000), may be colonized in situ by shipworms.There is some evidence that shipworms may enter living trees (NAIR & SARASWATHY, 1971;REIS, 1995) but it is suggested that this may occur only after the bark has been removed and the wood initially colonized by bacteria and fungi (KOHLMEYER et al., 1995).However, the impact of live colonization on tree survival remains to be quantified.The colonized woody biomass may remain in the mangrove until it becomes completely decomposed.Experimental work in mangroves has shown that this may take up to two to three years (GONG & ONG, 1990;KOHLMEYER et al., 1995) and test panels (1" thick) may be destroyed by Teredinidae in less than six months in tropical waters (TURNER, 1947).ROJAS & SEVEREYN (2000) found that Psiloteredo healdi (Bartsch, 1931) was able to decompose up to 58% of the mass of pine test panels in five months at Lago Maracaibo, Venezuela.Alternatively, colonized woody biomass may be removed from the mangrove by tidal action and eventually be transported to the adjacent coastal area.Shipworms may remain in the floating logs until all the woody material is decomposed and/or eroded by wave action.However, the logs may be transported back towards the estuary and mangrove forest, as well as become washed up on nearby sandy beaches.In the case of the former habitats, shipworms may continue to grow and reproduce whereas in the latter, exposure to the sun and wind at the high tide mark may cause death of the shipworms relatively quickly.In controlled experiments in Rio de Janeiro, 50% mortality of Teredinidae occurs after 56 hours of exposure to the air (OMENA et al., 1990).Although shipworms are able to close off their burrow during low tide, some studies suggest that the wider the piece of wood, the better the protection from dessication and overheating (RIMMER et al., 1983).Nevertheless, eventually all such woody material along with shipworm biomass is decomposed and may provide an important source of organic material and nutrients for organisms in the coastal area food web (GONG & ONG, 1990).
At Ajuruteua beach, on the northeastern coast of the state of Pará, northern Brazil, large quantities of mangrove woody biomass containing shipworms are observed after high tide.The beach is surrounded by mangrove vegetation (KRAUSE & GLASER, 2003).Previous studies about the contribution of benthos to nutrient cycling/transport in northern Brazilian mangroves have emphasized the importance of the role of land crab Ucides cordatus Linnaeus, 1763 in processing leaf litter (WOLFF et al., 2000;KOCH & WOLFF, 2002).Teredinids are common in mangroves in the region (BEASLEY et al., 2005) and as a contribution towards knowledge of the fate of exported mangrove detritus to coastal areas, the present study aims to quantify variation in the frequency of colonized mangrove driftwood logs, and the density and size of shipworms colonizing these, during a one-year period at Ajuruteua beach.

MATERIAL AND METHODS
Between January and December 2003, searches for mangrove driftwood logs were carried out each week during low tide over a 4km stretch of Ajuruteua beach (00 o 50'S, 46 o 36'W), municipality of Bragança, Pará.The mangrove forest is the overwhelmingly predominant vegetation type in the region and immediately surrounds the beach.Mangrove poles are used to construct fishtraps and, although there are few in the vicinity of the beach, it is possible that some of the driftwood may have come from such constructions.Logs washed ashore on the day that sampling took place remain along the tide mark and these were collected and sorted into three size categories regarding the circumference: small (up to 10cm), medium (10-20cm) and large (20-50cm).All logs were sawn to a standard 40cm length.The logs were numbered and five logs in each size category, a total of 60 logs per month, were randomly selected for examination for Teredinidae.
The logs were initially opened using a hammer and a wood-chisel, and a forceps was used from then on, in order to avoid damaging shipworms that might be inside.The bivalves were removed carefully with the forceps and in many cases, larger specimens could be removed without damage by inverting the log and allowing the animal to slide out through the gallery tube.
Shipworms were identified using TURNER (1966), RIOS (1994) and MORAES & LOPES (2003).The volume of each log was calculated using the formula v=π.r 2 .h,where r is the radius of the log and h is the length of the log.The number of individuals was counted per log and density was expressed as the number of individuals per cm 3 .Shipworm length (cm) was measured from the shell to the pair of pallets at the posterior end of the body.Salinity was measured weekly with an optical refractometer, using the Practical Salinity Scale.
Associations between the presence or absence of shipworms and log size and season were tested using Chi-square (χ 2 ) (ZAR, 1999).Two-way analysis of variance (ANOVA) was used to determine differences in mean density (in colonized logs only) and mean length of shipworms between seasons and log size categories.Density and length data were tested for homogeneity of variances using the Cochran C test (UNDERWOOD, 1997) and if significant differences occurred, the data were transformed appropriately.Where a significant difference was detected by ANOVA, Tukey tests were carried out, a posteriori, to assess pairwise differences between means (ZAR, 1999).Due to significant differences in variances after transformation and unbalanced sample sizes (different numbers of colonized logs per month) it was not possible to carry out a robust test (UNDERWOOD, 1997) for differences in density and length between months of the year.Instead, these data were presented graphically using box-and-whiskers plots (DALGAARD, 2002).The relationship between salinity and density and length was analyzed using Spearman Rank correlation.Data are presented as means ± standard error (S.E.) unless otherwise stated.All analyzes used a critical level of significance of α=0.05 and were carried out using the R statistical package (IHAKA & GENTLEMAN, 1996).

RESULTS
A single species, Neoteredo reynei (Bartsch, 1920), was found in driftwood logs during the entire study.Of the total of 720 logs examined, only 87 (12.08%) contained N. reynei, with a total of 744 specimens.
Seasonal abundance of the shipworm Neoteredo reynei (Bivalvia...There was a significant association between the presence or absence of shipworms and log size category.There were fewer small logs colonized and more large logs colonized than expected (χ 2 =65.3, d.f.=2, p<0.001;Tab.I).Large colonized logs were found throughout the year, medium sized colonized logs were less frequent, whereas small colonized logs were rare (Tab.II).In terms of season, there was no difference in the overall frequency of colonized logs between the wet and dry seasons (χ 2 =2.5, d.f.=1, n.s.; Tab.I).
The density of N. reynei in driftwood logs varied between 0.000160 cm -3 and 0.025650 cm -3 , with an overall mean of 0.003607 cm -3 .Mean density was not significantly different between seasons (Tab.III; wet season: 0.00402 ± 0.00053 cm -3 , dry season: 0.00302 ± 0.00079 cm -3 ).Between log size categories there were highly significant differences in mean density (Tab.III).The mean density in small logs (0.01101 ± 0.00383 cm -3 ) was significantly higher than in both medium (0.00439 ± 0.00075 cm -3 ) and large (0.00273 ± 0.00043 cm -3 ) logs (Tukey, p<0.05); these last two categories were not significantly different from each other (Tukey, p<0.05).Interaction between season and log size category was significant (Tab.III, Fig. 1) with density in small logs increasing between the wet and dry seasons whereas density decreased in the same period in both medium and large logs.Density was highest at the beginning of the wet season (January to April) but decreased during the transition period (May to July) between the wet and dry seasons.Shipworm densities were lowest in July (Fig. 2), increasing again slightly between August and October.
The length of N. reynei varied between 1 and 113cm with a mean of 9.66cm.Mean length was significantly different between seasons (wet season: 9.17 ± 6.07 cm; dry season: 11.28 ± 10.28 cm) but not among log size categories (small: 7.57 ± 7.61 cm; medium: 10.72 ± 7.10 cm; large: 9.45 ± 7.45 cm; Tab.IV).However, interaction between both factors was significant (Tab.IV, Fig. 3) so there was no general seasonal effect on length.Mean length of N. reynei in large logs increased between the wet and dry seasons whereas in medium and small logs, mean length decreased (Fig. 3).Shipworm length was relatively similar during months of the year but was highest in July and lowest in February (Fig. 4).
Mean monthly salinity varied between 10.9 and 40 during the study (Fig. 5).However, there was no significant relationship between either salinity and shipworm density (r s =-0.2697, d.f.=12, n.s.) or salinity and shipworm length (r s =0.412, d.f.=12, n.s.).Mean densities of N. reynei in mangrove habitat in the Ajuruteua Peninsula varied between 0.004 and 0.022 individuals per cm 3 .At Ilha Canela beach, densities were lower and ranged between 0.0015-0.0051individuals per cm 3 (BEASLEY et al., 2005) and these values are very similar to those at Ajuruteua beach, possibly indicating the harsher environmental conditions of beach habitats.High variation in density among individual logs occurred in our study showing that shipworms are very patchily distributed.Densities of shipworms ranged from 0.006 to 2.12 individuals per cm 2 at locations along the Rio de Janeiro coast and factors such as the wood supply, especially that of wood already colonized by shipworms, the abundance and diversity of the encrusting fauna as well as salinity are responsible for the variation in shipworm infestation within and between localities (JUNQUEIRA et al., 1989).
Only a small proportion of the mangrove driftwood examined in the present study was colonized by Teredinidae and only one species, N. reynei, was found.This result contrasts with our observation that pieces of wood in the mangrove forest are always colonized by shipworms, predominantly N. reynei.It may be that much of this driftwood is exported from the mangrove before shipworms are able to colonize it.As there is extensive removal of mangrove wood for housing, fish-traps and firewood in the region (KRAUSE & GLASER, 2003), debris from such logging may be rapidly exported by the tides and washed up, uncolonized, on nearby beaches.
The availability of wood is a fundamental requirement for the establishment of shipworm populations (TURNER, 1966).Mangrove wood appears to be a suitable substrate for shipworm species (LEONEL et al., 2002(LEONEL et al., , 2006) ) and although abundant at Ajuruteua, much of it is free of shipworms and therefore other factors may affect their ability to colonize driftwood.RIMMER et al. (1983) found that wider stumps of wood were colonized to a greater height by Bankia australis (Calman, 1920) and suggested that the larger volume of wood afforded greater protection through insulation against dessication and heat stress during low tide.JUNQUEIRA et al. (1991) registered highest densities of shipworms in solid collecting panels than in thin sheet panels.In contrast, the density of N. reynei was significantly lower in large logs in the present study, possibly because the wood in large driftwood logs may take longer to colonize than in small ones.LOPES & NARCHI (1997) found that settlement of shipworm larvae on 25cm long mangrove wood collectors was greater and occurred more rapidly on thin (0.3cm) sheets than on thick (2.4cm) cylinders.The ability to colonize wood may be dependent on its moisture content and permeability and also on the formation of a microbial coating (NAIR & SARASWATHY, 1971), which, because of the smaller surface-area to volume ratio, may be lower in large logs.Smaller logs may absorb water more easily and thus may be appropriate for colonization by shipworm larvae after a relatively shorter period of time in the water than larger logs.Once colonized, small logs may soon be destroyed by shipworm activity, thus explaining the relatively low number of small colonized logs found in this study.
The chemical content of some tree species as well as the presence of bark, which contains high concentrations of chemical components (NAIR & SARASWATHY, 1971), may cause mortality in larvae attempting to penetrate certain types of wood (TURNER, 1976).However, LOPES & NARCHI (1997) found almost no shipworm settlement in experiments with artificial collectors made of pine, whereas settlement was greater in those collectors made of native mangrove wood.
Prior to settlement, changes in planktonic food supply (MANN & GALLAGER, 1984a), temperature or salinity may delay or stop development of free-swimming larvae (TURNER & JOHNSON, 1971;NAIR & SARASWATHY, 1971) and larvae may therefore be unable to penetrate the wood.Shipworm density was greater in the months of the wet season, when humidity and precipitation are greater, and may be related to ease of colonization as well as higher survivorship in logs out of the water.Neoteredo reynei is an euryhaline species found in waters that range in salinity between 0.6 to 14 in mangroves in Pará, northern Brazil (REIS, 1995) and between 2 and 35 in mangroves from Paraíba (LEONEL et al., 2002).LOPES & NARCHI (1993) registered higher densities of N. reynei in the upper section of the Rio Comprido estuary (São Paulo) corresponding to a salinity of 0 to 5. Thus, it is unlikely that the much lower densities observed during the transition period between the wet and dry seasons (May and July) are related to mortality caused by a reduction in salinity.Lower salinity between March and April may be linked to a reduction in reproductive output in N. reynei perhaps as a result of higher levels of mortality during the planktonic stage, and consequently lower numbers of recruits in the following months.However, to date, nothing is known of the reproductive cycle of N. reynei in northern Brazil.LEONEL et al. (2006) reported highest densities of shipworms immediately following the period of lowest salinity in Paraíba.Data on Teredinidae in Paraíba show that reproduction may be influenced by seasonal changes; spawning occurs during periods of low salinity and lower temperatures in N. fusticula and in periods of high salinity and warmer temperatures in B. fimbriatula (LEONEL et al., 2002).Many of the Australian Teredinidae can breed and settle throughout the year but, even in warmer latitudes, seasonal changes in shipworm reproductive activity occur (IBRAHIM, 1981).The experiments of SARASWATHY & NAIR (1974) show that N. hedleyi Schepman, 1919 (from Cochin Harbour, India, where salinity varies between 0.5 and 38.3) cannot survive in salinities higher than 29.However, although some individuals in the natural environment were able to tolerate higher salinities, this species reproduces only during the rainy period when salinity is optimal (11.2 to 14.5) for egg development (NAIR & SARASWATHY, 1971).Adults and larvae of two incubatory species of Teredinidae from a tropical estuary in Papua New Guinea can tolerate experimental salinities different from those found in the estuary (RAYNER, 1979).This author suggested that the most important factor affecting shipworm distribution may be the salinity tolerance range of settling and metamorphosing larvae.
Neoteredo reynei is one of the largest species of shipworm in Brazil and has been known to attain 1.80m in length (MÜLLER & LANA, 2004).Although the largest specimen measured in the present study was 1.13m, the mean length of N. reynei in mangrove driftwood at Ajuruteua Beach was much smaller.Larger size of shipworms in the dry season may be the result of growth following an earlier settlement event in the wet season.Shipworms in large logs were longer in the dry season than in the wet one, those in medium and small logs were shorter.Insulating properties of the larger volume of wood may favor survivorship in beached driftwood and in turn allow growth and, perhaps after a high tide, reproduction may even occur.MANN & GALLAGER (1984b) showed that Teredo navalis Linnaeus, 1758 and B. gouldi could grow and reproduce in an experimental wooden substrate in the absence of a phytoplankton supplement.In experiments carried out by OMENA et al. (1990), 100% survival was recorded after 72 hours out of the water for Nototeredo knoxi (Bartsch, 1917) and appears to be related to adaptations in the pallets that better seal the burrow, as well as gas exchange via the mantle.Neoteredo reynei is known to have morphological adaptations that promote its survival out of water; a pair of highly vascularized dorsal lappets occur at the posterior end of the body near the burrow opening and are probably involved in cutaneous gas exchange (MORAES & LOPES, 2003).Modifications of the appendix, anal canal, ctenidia and labial palps, as well as food grooves, indicate that N. reynei is less dependent on phytoplankton and feeds predominantly on wood (TURNER, 1966;LOPES et al., 2000;MORAES & LOPES, 2003).Thus, it may be possible for N. reynei to resist dehydration and overheating and continue feeding in upper intertidal zones of mangroves, estuaries and beaches, for some time, especially if colonizing a large log.Indeed, our own observations (BEASLEY et al., 2005) and those of MORAES & LOPES (2003) and LOPES & NARCHI (1993) show that N. reynei is able to thrive in mangrove habitat with a low frequency of tidal inundation.

Fig. 1 .
Fig. 1.Mean density ± S.E.(number of individuals per cm 3 ) of Neoteredo reynei (Bartsch, 1920) in small, medium and large mangrove driftwood logs at Ajuruteua beach, northern Brazil, during the Wet (January to June) and Dry (July to December) seasons in 2003.

Fig. 3 .
Fig. 3. Mean length ± S.E.(cm) of Neoteredo reynei (Bartsch, 1920) in small, medium and large mangrove driftwood logs at Ajuruteua beach, northern Brazil, during the Wet (January to June) and Dry (July to December) seasons in 2003.

Fig. 5 .
Fig. 5. Mean salinity ± S.E. at Ajuruteua beach, northern Brazil, in each month of the study period (January to December 2003).

Table III .
ANOVA summary of the effects of season and log size category on mean density of Neoteredo reynei (Bartsch, 1920) (untransformed data) in driftwood logs, Ajuruteua beach, northern Brazil in 2003.No differences in variances were detected prior to ANOVA (Cochran C=0.32; d.f.=6,4; n.s.).

Table IV .
ANOVA summary of the effects of season and log size category on mean length of Neoteredo reynei (Bartsch, 1920) (untransformed data) in driftwood logs, Ajuruteua beach, northern Brazil in 2003.No differences in variances were detected prior to ANOVA (Cochran C=0.396; d.f.=6,7; n.s.).