Four sampling stations were selected for this study as shown in Figure 1. The southwest coast of Marudu Bay is covered with 9550 ha of mangrove forest where it is known to harbor at least 6 mollusks species [6] . East coast of Marudu Bay is relatively more developed than the west coast, where many light and small medium industries can be found along the coastal areas.

Systematic surveys were carried out from 8th to 14th May 2017 during low tide

according to Garza et al. [7] . Clams were collected at the mangrove swamps which covered an area of 500 m² (50 m × 10 m) per site. Surface sediment of 15 cm depth was removed to obtain the buried clams. The clam specimens were then placed in labeled plastic bag and stored at 4˚C, and then transported to the laboratory for analysis within 24 h. In the laboratory, the samples were sorted and washed to remove all adhering organisms and other debris.

All clam samples were counted and identified to species level. Individual specimens were then measured for its shell length, shell width and shell height using a vernier caliper to the nearest 0.1 mm. The maximum dimension of the anterior-posterior axis was recorded as shell length, the maximum lateral axis as shell width and the maximum distance between the valves when they are closed was considered as height. Few representative individuals (n = 5) of different groups of clams were preserved in 70% alcohol and further identified in the UMS laboratory according to Morris [8] , Keen [9] , Skoglund [10] and Carpenter and Niem [11] .

Surface sediment of 100 g was collected and stored at 4˚C. In laboratory, sediment subsample was air dried at room temperature, grinded and mixed thoroughly. The sediment particle size and clay-silt percentage were determined by a laser diffraction particle size analyzer (Sequola, Canada) according to Agrawal and Pottsmith [12] .

Statistical analyses were performed using the SPSS Windows Statistical Package (version 21). Tests were considered significant at p < 0.05. Prior to analyses, all variables were tested for normality and homogeneity of variances. One-way ANOVA was used to test for significant differences among sites for sediment clay-silk composition and shell length. The shell length distribution of mud clam in each station was illustrated in a histogram, whereas skewness and kurosis were calculated according to Groenveld and Meeden [13] .

The silt-clay composition of the sediment of the mangrove swamps in the sampling sites was found to range from 62.22% to 88.73% (Figure 2). However, no significant difference (p > 0.05) was observed in the silt-clay composition among the sites.

There were only two species of marsh clam identified in sampling sites namely P. erosa and P. expansa. The occurrence of P. erosa and P. expansa were 100% and 50%, respectively. Bivalve density (Figure 3) in station A (n = 32) and station B (n = 140) was significantly higher and lower (p < 0.05), respectively than that in other stations.

P. erosa in sites B and D recorded significantly higher (p < 0.05) and lower (p < 0.05) length/width ratio (1.10 to 1.14 vs 1.03 to 1.07, respectively) and length/depth ratio (1.96 to 2.03 vs 1.76 to 1.77, respectively), respectively than that in sites A and C (Figure 4). However, no significant difference (p > 0.05) was recorded in length/width or length/ depth ratio of the P. expansa collected from all the sites.

Size distribution of P. erosa at the sampling sites is illustrated in Figure 5. The shell length of P. erosa ranged from 2.1 to 8.9 cm. The size distribution for P. erosa in all sites showed a bell shape distribution pattern with minor degrees of skewness and kurtosis. In site A, P. erosa population showed a clear leptokurtic and right skewed distribution. P. erosa in sites B and C showed an extreme left skewed distribution, where P. erosa in site D was moderately platykurtic and left skewed. Juvenile P. erosa in sites A, B, C and D was accounted for 0%, 15.7%, 0% and 7.6%, respectively. On the other hand, size distribution for the other species, P. expansa was not included in current study due to its low occurrence and less abundance.

Silt-clay composition of the sampling sites was not statistically different, despite the observation that the silt-clay composition in sites A and C were relatively

higher than that in sites B and D. Since the sample size in current study was relatively small, therefore we did not rule out the possibility of sediment grain size distribution could be influenced by the relatively small sample size.

Two marsh clam species, P. erosa and P. expansa were found to inhabit the sampling sites. Occurrence of P. erosa in the mangrove swamp of Marudu Bay had previously been reported by Zakaria and Rajpar [6] . The occurrence and distribution of many marine benthic organisms are known to associate with substrate types [14] . In current study, the occurrence of P. expansa was relatively lower compared to P. erosa, where the former has only been observed in the sites with silt-clay content of more than 80%. This observation was in agreement with the finding of Dolorosa and Galon [15] , where distribution of marsh clam can vary between patches of mangroves due to the influence of environmental conditions. Active habitat choice for a specific sedimentary environment, in the form of habitat selection by settling larvae, may explain the lower distribution of P. expansa in Marudu Bay.

It is also interesting to note that there were no other bivalve species found inhabiting the sampling stations. Mangrove sediment is known to contain high organic level which may not be suitable to many organisms [16] , this makes the area uninhabitable by other bivalve species. The abundance of marsh clam in site A (32 ind/500m²) was very low. Site A is located at a river mouth which may provide an easy access to bivalve collectors. Observation during sampling also strengthen the fact that site A is a site whereby frequently be visited by bivalve collectors hence experience high exploitation. Intensification in term of frequency and effort of bivalve fishing over a long period of time in a particular site can be harmful to the population dynamic of the bivalve in that habitat. Temporally site closure management strategies have been reported to be an effective way to increase the marsh clams population in Philippine [17] . Hence, such a strategy can be adopted to site A in the present study to avoid further population depletion. On the other hand, site B which is located relatively far from the rivers and away from fishermen reach was observed to experience low level of exploitation. In this connection, site B could be a good alternative marsh calm fishing ground when temporally site closure management strategy has been implemented in site A. Despite the increase fishing cost due to more fuel is required to reach site B, it can be compensated by more catch with equal or lower fishing effort.

Studying bivalve growth and establishing allometric relationships are essential for generating useful information for managing resources and understanding changing environmental condition and pollution [18] . The length-width and length-height ratio of P. erosa in sites B and D were significantly higher and lower, respectively compared to the other sites. This means that the increase in shell length is superior to increase in shell width and height. Sediment grain size has been reported to influence the shell morphology of bivalve [19] . Coincidently, the silt-clay composition in sites B and D were relatively lower than that in sites A and C, which suggest the shell of P. erosa at sites B and D become progressively longer to smoother movement in grosser sediment [20] . Although the relationship between allometric ratios (length/width and length/depth ratios) and silt-clay composition were not statistically different, a link between allometric ratios and sediment grain size is visible in the current study.

The shell length range of P. erosa in Marudu Bay (2.1 to 8.9 cm) was comparable to those reported in Butuan bay, Philippines (2.9 to 8 cm) [21] . The distribution of P. erosa in current study was varied among sampling stations. This observation was in agreement with the finding of Clemente and Ingole [14] , where non-random, patchy distribution of P. erosa has been reported in the Chorao mangrove swamp, Goa. Higher number of juvenile P. erosa was observed in site B which was located far from river. Similar result was also been reported by Elvira and Jumawan [21] , where areas nearby river mouth of Agusan river, Philippines has significantly fewer juvenile marsh clam (12%) compared to those located at 1 km away from Agusan river mouth (33%) [21] . Moreover, Clemente and Ingole [14] also reported the majority of the juveniles were confined to the seaward zone while adults were found conspicuously towards the landward zone.