Freshwater Clam Resource Assessment of the Ba River

By Esaroma Ledua ¹
Sione Vailala Matoto ¹
Apisai Sesewa ²
& Jovesa Korovulavula <sup>2

1</sup>South Pacific Commission, B.P.D5, Noumea, New Caledonia
²Fisheries Division, P.O.Box 358, Suva, Fiji.

September, 1996.

SPC Integrated Coastal Fisheries Management Project Reports Series #1

Executive Summary

1.0 Survey Methods

The Ba river survey was conducted between June 7 - 29, 1996 and a combination of survey methods were used so that appropriate data could be collected within that short period of time. The biomass and density of the freshwater clam, Batissa violacea or known in Fijian as kai waidranu, were estimated by transects and quadrats. The estimate of the fishing effort exerted on the resource was determined through the use of creel survey techniques. Water quality data were obtained from the Drainage and Irrigation Department data base. The rainfall data were obtained from the Fiji Meteorological Department and the freshwater clam market data were obtained from the Fisheries Department data base.

2.0 Results

The results of this survey showed that kai could be found between the lower end of Nailaga village and all the way up to the upper end of Kumukumu. It was discovered that rich kai beds occur between Station 4 (upper Nailaga) and Station 6 (Vaqia). The total distance of the kai bed was estimated to be approximately 17.8 kilometres in length. The total area of the kai bed was around 29,122,970 square metres. The overall mean density of kai for the Ba river was estimated to be 270.38 kai per square metre. The mean biomass was approximately 684.31grams per square metre. The estimated total population of kai for the Ba river was 787,608,829. The estimated total biomass was 1,993,374.5 kilogram. The annual average commercial catch estimated from the market data was approximately 124,511 kilogram. The total estimated catch for both subsistence and commercial fishing calculated from the results of the creel survey was 496,729 kilogram. Therefore commercial catch (kg) as a percentage of the standing stock will be approximately 6.24% only but for both commercial and subsistence would be 24.92%. The analysis of the kai size structure observed in this study, showed that 36.71% of the samples were above 25 mm in shell length (market size). Considering this as a reflection of the size structure of the whole population, then 36.71% of the whole population would be 731,767.77 kilogram of kai, and this would the marketable volume of wild stock, available for harvest in the Ba river.

3.0 Conclusion

The conclusion that was drawn from the results of this survey was that the kai stock in the Ba river was still in good shape and the current level of fishing is sustainable.

4.0 Management Recommendations

• There is a need for Fiji Fisheries to gather information on catch and effort data. This is quite important in the understanding of the accurate amount of kai being harvested from the river annually. It was observed in this survey that the market data was insufficient and was only reporting about 25% of the total catch from the river. This work could be easily be carried out, utilising the existing RA&D staff at the Ba office. Sufficient data could be collected utilising just one person for only two days per month, to conduct the creel survey and the count the number of people fishing for kai along the Ba river.
• The current dredging operation on the Ba river would definitely damage the rich kai beds. It was discovered during this survey that most of the kai live around shallow depth (<1.2metres). The aim of the Drainage and Irrigation Department is to dredge or deepen shallow areas and that is where most of the kai were found especially juveniles. It is therefore suggested, that the Fisheries Division work closely with the Drainage and Irrigation Department and to advise accordingly of necessary action to be taken so that rich kai beds are not severely damaged. An option would be to mark all rich kai beds properly and prohibit dredging in those areas. It was observed that kai were contagiously distributed, aggregated in just certain parts of the river which consisted of only a small portion of the river. Deepening of the other parts of the river should be sufficient to allow good water flow and boat navigation. Another option would be to restrict dredging to as close as possible to the centre of the river and prohibit dredging at distances of approximately 30 metres away from the low water mark on both banks, especially the lower reaches of the river. This would ensure the protection of the prime kai beds as well as avoid soil erosion on both banks.
• It is recommended that the current fishing method be maintained and any form of mechanised fishing or the use of SCUBA gears to be totally banned.

5.0 Acknowledgements

First and foremost we would like to sincerely thank ODA of the United Kingdom for funding this ICFMaP sub-project. We would also like to thank the Fiji Fisheries Division for its contribution to the project in terms of manpower and local costs involved. Our very special thanks to the Director of Fisheries, Mr. Maciu Lagibalavu, Officer In-charge at the Ba Fisheries office, Mr. Krishna Swamy, Mr. George Hazelman, RA&D Officer - Western, Mr. Anand Prasad, Fisheries Officer-Western, Mr. Suresh Chand and Mr. Anasa of Ba for all the support that was given to us. We are also grateful to Mr. Aduru Kuva, General Manager of the Ba FSC, for allowing us to access the report on the study of the Ba river water parameters which was conducted by USP which FSC funded. We would like to thank the people of Votua village for all the information that was supplied to us during this study. Last but not the least, we are grateful to Dr Tim Adams and Mr Paul Dalzell for their valuable contributions especially for editing the text.

1.0 Introduction

1.1 Background of the Fishery

The freshwater clam (Batissa violacea) called in Fijian kai waidranu, is an important item on the menu of many families in Fiji and may be purchased cheaply ($2/heap) in many of the local markets and at roadside stalls in the areas in which it is caught. The clam provides a source of protein on a subsistence basis but it is also the basis of a substantial commercial fishery operated mainly by Fijian women. It is the largest single domestic fishery in Fiji, producing approximately 1,300 metric tonnes annually (Richards, 1994), and the current value of the fishery is estimated to be around FJD$1million.

The animal is collected by groups of women wading in shallow water or diving from boats or rafts. Clams are located using hands and feet. Diving using diving glasses is rarely practised and kai fisher-women spent 2-4 hours per day harvesting and the next day selling. Catch per unit of effort varies from one river to the other. It has been reported that, catch per unit of effort on average is approximately 10 kg/person-hour (Anon., 1975; Richards, 1994). A fisher-women can earn up to FJ$30.00 per day from catching kai (Bibi, 1991).

Kai is an excellent simple fishery as the animals survives well out of water without refrigeration and may be stored for dispatch to market with minimal precautions for several days. In this case it is far superior to fin fish and prawns although its commercial value is far less than these.

It is a normal practice in Fijian homes that kai is immersed in clean water for at least a day before they are being cooked and consumed. This is to allow kai to defecate whereby sand and mud is excreted from the gut. If these solid particles in the guts are not removed, kai is almost inedible. Fijians have practised this simple depuration method for generations.

The Ba river contributes approximately 124 tonnes of kai annually, constituting about 9.5% of Fiji's total production. The majority of the commercial kai fishers in the Ba area comes from Nailaga, Votua and Nasolo villages. Subsistence fishers include both Indians and Fijians that are living within the Ba rural and town areas. Approximately 90% of the kai caught from the Ba river are sold at the Ba municipal market or roadside stalls at Nasolo and Nailaga villages.

Although the Fiji Fisheries Division has very good data on the approximate quantities of kai harvested for commercial purpose each year and the amounts sold in each market but there is no systematic attempt to monitor the fishery, especially the volume used for subsistence purposes. The Fisheries Division believed that some of the areas are over-fished, for example the Ba river, but this could only be confirmed through survey, both by direct field observation and interviews of local people. At present there are still considerable gaps in our knowledge of its biology, abundance, distribution and growth rate of kai in Fijian rivers, despite the importance of this fishery.

1.2 Distribution and Biology

Batissa violacea (kai) is a tropical freshwater corbiculid mollusc distributed throughout the western Pacific, i.e. Malaysia, Philippines, Papua New Guinea, North-west Australia and various Pacific areas (Morton, 1989). In Fiji, kai is present on eight major rivers (Rewa, Ba, Nadi, Sigatoka, Navua, Korovou, Dreketi, Labasa and Waikoro) of the two main islands, Viti Levu and Vanua Levu (see Figure 1).

Richards (1994) reported that Batissa larvae are partially incubated within the shell and are capable of only limited movements, so the influence of floods and tides on settlement is critical. This is contrary to Morton's (1989) report which described Batissa as dioecious and non-incubatory. Morton (1989) also reported of kai attaining a maximum shell length of 150mm, with numerous growth lines and probably long-lived. The growth rate of kai has been reported to be around 2 cm per year (Raj, 1981).

The osmolarity and ionic composition of Batissa blood indicates recent colonisation of freshwater (Raj and Fergusson, 1980). The shell of kai is equivalve and approximately equilateral. The periostracum is thick and dark and younger shells appear dark violet whereas the older ones are black. The older parts of the shell, around the umbones, are often eroded. The nacreous internal surface of the shell is varying shades of purple, particularly external to the pallial line. The anterior and posterior adductor muscles are deeply impressed into the shell of older individuals, just below the large hinge plate (Morton, 1989).

Raj (1981) reports that two ecomorphs (kai buli and kai bukuvula) of the species Batissa violacea are present in Fiji. It has been reported also that the occurrence of kai is restricted to the lower freshwater reaches of rivers, between the upper limit of tidal influence and the upper limit of salt penetration (Raj, 1981; & Lewis, 1985a). The clam is free living, burrowing to 10 cm in river beds and capable of substantial movements.

1.3 The Study Site

The Ba river is located on the north-western side of the main island, Viti Levu. Since the Ba river is on the lee side of Viti Levu, it experiences a dry environment with minimal rainfall. Valleys and hills along the Ba river are rich agriculture lands. The Ba district produces the largest volume of sugarcane annually in Fiji, and sugar is the main economic activity, followed by manufacturing, forestry and fisheries respectively. The average monthly temperature ranges between 24 - 27⁰C and rainfall ranges between 1440 - 1993 mm/yr. Ba town is the main business centre of the Ba district and is located on the western side of the river delta approximately 8 kilometres inland from the coast. The population of Ba is approximately 11,000 people. The majority of the people living in Ba town are Indian descent, while the majority of the population in rural areas are Fijians.

The total length of the main reach of the Ba river is approximately 28 kilometres long (mouth - Nabatolu) with a maximum water depth of 8 metres.

2.0 SPC Study

The South Pacific Commission was requested to assist in the study of the Ba river kai resource and to come up with a resource management plan for the Ba river. The SPC-Integrated Coastal Fisheries Management Project (SPC-ICFMaP) study, was directed at addressing the following concerns, which were raised by the Fiji Fisheries Division:-

1. The Fisheries Division was concerned about the drop in production from 169.9 tonnes in 1992 to 99.9 tonnes in 1993, coupled with the small sizes of kai, collected from the Ba river and sold at the Ba market, as well as on road side stalls in the Ba district, compared to kai collected from the Rewa river on the eastern side of the main island. Fiji Fisheries Division suspected over-fishing as a possible reason for the decline in catch and why smaller sizes of kai were brought to the Ba market.
2. Fiji was also concerned on possible health risks that may be associated with eating kai. The Ba river receives frequent wastes discharged from the sugar mill, cane fields and small industries within Ba town. It is regarded as one of the highly polluted rivers in Fiji. In the past, cases of massive fish kills had been reported from the Ba river.
3. Fiji Fisheries was also interested to know the socio-economic aspects of the kai fishery.

There are three elements of this study. The health aspects was carried out by the Commission's Post-Harvest section, whereas the socio-economic study will be carried out, by the Women Fisheries Development section. This report will only highlight the resource assessment study that was carried out by the Resource Assessment team, of the SPC- ICFMaP.

3.0 Objectives of the ICFMaP Study

3.1 To carry out a study on the biomass and density of kai in the Ba river.

3.2 Make recommendations on management measures relating to the commercial and subsistence harvesting of the kai resource in the Ba river.

4.0 Expected Outcome

The expected outcome is a resource management plan. This management plan is aimed at securing a long term sustainability of the kai resource and its usage, taking into account the present and potential future developments in fishing and other income or food generating activities.

5.0 Survey Methods

1. Field activities involved the use of scuba gears to assess the density and biomass of kai, on chosen sites along the Ba river. A total of eight stations was chosen at random, from the U.K. admiralty chart No. 389 and scaled aerial photographs for the Ba river, available at the Fiji Lands Department. Four transects were laid at each station. Transects were placed parallel to the flow of the river. The four 50 x 1m transects were equally spaced across from one bank to the other in order to determine the distribution and density of kai at that station. This was to determine whether kai is evenly distributed across the river, at each station. On each transect, 8 grid samples was taken. The grid sizes of 50 x 50cm was made from 13mm galvanised steel. Using hand trowels, each diver dug sediments within the grid into a sieve and by shaking the sieve, all sand and silt are removed whereas kai remain in the sieve with other coarse substrate. All kai collected were measured using callipers as well as weighed using an electronic balance or hand scale. Work started from the lower parts of the river, approximately in line with the Drainage and Irrigation Department complex, at the lower end of Nailaga village, to determine the lower limit of the kai bed. The last Station was at upper Kumukumu, to determine the upper limit. A few more spots were chosen beyond this last Station at Kumukumu right up to Nabatolu to check the presence of kai. No transects or grids were taken but divers search for the presence of kai. Data collected were inputted in the computer using the excel spreadsheet.
2. An attempt was made to try and gauge the fishing pressure for kai in the river. The number of people that were fishing in the river per day plus catch and effort data were recorded. Catch and effort data were collected by intercepting some of the fishers and weighing their catch and interrogating them about the amount of time spent on fishing, fishing gears used, and questions about the kai stock.
3. Market samples of kai were also taken in-order to determine the range of sizes going into the market, and the weight and the number of kai per heap.
4. Rainfall data was obtained from the Fiji Meteorological Department office at Nadi. The water quality data were obtained from the Drainage and Irrigation Department as well as from the University of the South Pacific reports. The kai market data was obtained from the Fisheries Division data base.
   

Table1: This shows the summary of transects and grids taken at each station.

Figure 3. Arrangements of transect lines for kai sampling across the river.
(insert figure)

6.0 Equipment and Materials used

2 x 50m tapes, 4 by 50 x 50cm galvanised steel grids, 4 sets of SCUBA gears, four marker buoys with 4 anchors, admiralty chart of Ba, aerial photograph of the Ba river, glass thermometers, 4 stainless steel sieve, Nikonos camera, 2 sets of flower gardening hand tools, an 18 footer fibreglass skiff with 40 hp engine, 3 x 12 litres plastic buckets, plastic bags, Shimadshu electronic animal scale (LIBROR EB - 2800M), 2 stainless steel callipers (MITUTOYO), 1 x 30 kg Salter hanging scale and 2 laptop computers.

7.0 Results

7.10 Water Quality

During this survey, we attempted to collect as much information as possible on water quality and substrate types, hoping that this would assist us in describing the distribution pattern of kai along the Ba river. Fortunately water quality data were already available at USP and at the Drainage and Irrigation Department. Relevant data were extracted from files and reports from the two institutions. Despite the availability of these data, some estimates of water quality parameters, such as salinity and temperature were conducted, so that it could be crossed checked with available data. The data available at the Drainage and Irrigation Department were very reliable as most of the data were recorded on a 24 hour cycle. From the data extracted, we found out that, salt penetration could be traced as far as Nasolo during flood tides only. Therefore Nasolo was regarded as the upper limit of the salt wedge. Nasolo is approximately 10 kilometres from the river mouth. The salinity and pH profile of the Ba river could be seen on Figure 4. The graph on Figure 5 shows the temperature and DO profile of the Ba river. Water temperature decreased towards inland. The lowest mean temperature was recorded at the upper end of the river and the highest was recorded at the mouth of the river. The graph on figure 5 also shows that the dissolved oxygen level was quite low between Nailaga and Nasolo but it was observed that kai fishing was mainly carried out between these two areas. The dissolved oxygen level was the lowest at Namosau but the level increased towards the coast as well as towards the upper end of the river.

Figure 4. Salinity and pH values at different locations along the Ba river.

Figure 5. Temperature and dissolved oxygen at different locations along the Ba river.

7.11 Distribution of Kai

The results obtained from this survey showed that freshwater clams or kai waidranu, were distributed along river beds between the lower end of Nailaga village and right up to the upper end of Kumukumu. Although the total length of the main reach of the Ba river was around 28 kilometres (mouth - Nabatolu), kai was only found within a total distance of approximately 17.8 kilometres in length and that was the distance between the two areas specified above (Nailaga - Kumukumu). The lower limit of the kai bed (lower end of Nailaga village or Station 1) was approximately 4 kilometres inland from the mouth of the river. Salinity at this site ranged between 2.90 ppt (ebb tide) and 16.70 ppt (flood tide). Mangroves were still present at this site and the substrate was mainly composed of soft black mud which could be as thick as 120 centimetres in some places, especially the areas close to the banks on both sides. There was a thin line of sand along the mid section of the river. The kai that was found at this site was collected from the mid section of the river along the thin patch of sand. The size of the kai found at this site was around 80.2 mm in shell length.

The areas between the upper end of Nailaga (Station 4) and Vaqia (Station 6) contained the largest population of kai (see Figure 3). The population density decreased towards down river from station 4 and decreased towards upriver after station 6. The last patches of mangrove trees growing along river banks ended at station 4 (refer to Figure 2). Station 4 was approximately 2 kilometres above the lower limit of the kai bed (station 1). It was observed that, between stations 4 and 6, large numbers of kai were found where the substrate was made up of a mixture of fine sand, fine gravel and sticky bluish mud. Very few kai were found where soft mud occur (see Table 1). It was observed that most of the kai were found in shallow waters (0.2 - 1.4 metres), especially the juveniles. Although kai found in the deeper sides of the river were mainly larger sized individuals but the densities were consistently low.

The substrate at the upper limit of the kai bed (upper Kumukumu) was mainly composed of large gravel, rocks, brown mud and thin deposits of bluish mud. The minimum size of kai found at the upper limit of the kai bed was 20 mm in shell length. The current at this site was quite strong. Further inland, the gravel sizes, mud and sand increased. Mud turned brownish in colour and the particles were very coarse compared to the lower end of the river.

It was observed that juvenile kai occurred in large numbers between Stations 4 and 6. It decreased rapidly in numbers towards inland from Station 6 and towards the coast from Station 4.

Table 2. Density of kai on different types of substrate.

7.12 Population Description

A total of 17304 kai were collected at the 8 study sites. The sizes in terms of shell length ranged from 3.5 - 80.2 mm (Figure 6). The population mean was 23 mm in shell length and the median was also 23 mm. The histogram on Figure 6 shows that it closely conforms to a normal distribution pattern. The total size structure of kai observed in this study, showed that 63.29% were below 25 mm and 36.71% were above or around 25 mm (marketable size) in shell length (Figure 7). Approximately 96.3% of the kai observed, were below 40 mm in shell length, and only 2.7% were above 40 mm in shell length (Figures 6 & 7). Despite the large volume of data, it was quite difficult to see any pattern of different cohort from the histogram. Different cohorts overlap each other too much and there is difficulty involved in cohort separation. It could also be seen on Figure 6 that the spread of the curve was small and approximately 90% of the samples were within 10 mm from the population mean. The graph may suggest slow growth or more than one spawning a year or a combination of both. The relationship between shell length and shell weight shows that, y=0.0002x^2.9948 and R² =0.9893 (Figure 8). The graph on Figure 9 shows the strong positive linear correlation between shell length versus shell height and shell width.

Figure 6. Size frequency of kai observed in this study.

Figure 7. Size frequency of kai in the Ba River grouped in 5 mm size classes. Approximately 35% of the population falls within the marketable size range (>25 mm).

Figure 8. Length versus weight relationship for Ba River kai.

Figure 9. Shell height and shell width versus shell length for Ba River kai.

7.13 Population Density and Biomass

The estimated total kai population on the Ba river was calculated using the mean transect density and the total area of the kai bed which was worked out from the aerial photograph as well as from the U.K. admiralty chart.

Summary of results from kai abundance estimation

a = Total Area of kai bed (m² ) = 2,912,970
d = Mean Density (n) per m² = 270.38
b = Mean Biomass (g) per m² = 684.31

TEP = Total Estimated Population of kai (n) = 787,608,829
TEB = Total Estimated Biomass (kg) = 1,993,374.5

The overall mean density for the Ba river was 270.38 kai per m² (95% CI = 229.2) and the mean biomass was 684 grams per m² (95% CI = 585.82). The individual weight ranged from 0.05 - 110 grams per kai. The total estimated population of kai in the Ba river was 787,608,829 kai. The total estimated biomass for the Ba river was 1,993,374.5 kilograms or 1,993.3745 metric tonnes (see table 4).

It was observed that there was a wide variation in population density within each study site, as well as from one station to the other. The mean density per station along the Ba river ranged from 0.13 - 1002.38 kai per m² (Station 1 - 8). It was also observed that, there was a high variation in density from one transect to the other, within a station. At station 4, kai density ranged between 223 - 1194.5 kai per m² , and station 5 ranges from 1.5 - 867.5 per m² , and station 6 ranged from 173.5 - 3543 per m² (Table 4). On Figure 11, the vertical bars shows the variability of density on a 95% Confidence Interval.

Table 4. This table shows the density of kai on the four transects at each station. It shows that there is a wide variation in population density across the river.

Table 5. This table shows the mean density and biomass of kai at the 8 study sites as well as the Ba River in total.

Figure 10. Biomass and density of Batissa at the 8 study sites in the Ba River.

Figure 11. Kai density at each sampling station on the Ba River. The vertical bars represent the 95% confidence interval.

Fishing Activity

A total of 14 fisher-women were intercepted along the Ba river while fishing for kai. The individual catches were measured and fishing time recorded. The catch rate per person ranged from 3.7 - 38.0 kg/person-hour. The average catch per unit of effort was estimated to be approximately 11.7 kg/person-hour (Table 6). At the lower half of the river (Nailaga - Nasolo), kai fishing was mainly onducted at low tide whereas fishing at the upper half (Vaqia - Kumukumu), could be carried on throughout the day. The time spent on fishing ranged from 1.35 - 3.45 hours per person. On average, the fishing period was approximately 2.43 hours per person per day.

Kai fishing was regularly carried out from Monday to Saturday. It was observed during this survey that the number of people fishing for kai, both for commercial and subsistence purpose, along the river on a daily basis ranged from 0 - 196 per day. The fishing activities were dictated by multiple factors such as the weather conditions, public holidays, festivals or the state of the kai gonad. During the study period, the highest number (196) was recorded on a public holiday and no one was seen fishing when the river was flooding. The number of people fishing between Monday and Saturday, on average, was around 56 per day. Most of the commercial kai fishers fish hard on Monday to Wednesday and spend Thursday and Friday selling at the market.

Using the above figures to calculate the amount of kai extracted from the river per year, the following formulae was used.

Where TC = total estimated catch, c = average catch per person per day, n = average number of fishers fishing per day, d = number of days of fishing per year.

This would mean that the total estimated catch for both subsistence and commercial purpose per year for the Ba river would be approximately 496,729 kg or 496.73 tonnes.

Table 6. This table contains the catch and effort data on Kai fishing in the Ba River. The data were collected using the creel survey technique.

Kai Sold at the Ba Market

Analysis of the market samples showed that the average weight per heap of kai sold at the Ba market was approximately 3.99 kg and the average number of individual kai per heap was 378 kai. Samples taken during this study period, showed that the weight of kai per heap, ranged between 3.41 - 4.57 kg. The number ranged from 230 - 526 individual kai per heap depending on the sizes of individuals.

The sizes of kai sold at the Ba Market, in terms of individual length, ranged from 24 - 84 mm and shell weight ranged from 2.2 - 110.3 grams. Mean shell length was 37.9 mm and mean shell weight was 10.6 grams (Figures 12 - 14 below). Figure 12 below, shows that a high proportion of kai sold at the Ba market were below 40 mm in shell length and Figure 14 shows that majority of the individual kai were below 10 grams in shell weight. Figure 15 shows a positively curved relationship between length and weight of kai sold at the Ba market. A comparison between sizes of kai sold at the Ba and Rewa market is shown in Figure 13. The largest kai measured from the Rewa kai was 120 mm in shell length whereas the Ba river was only 84 mm.

Figure 12. Size frequency of kai sold at the Ba market.

Figure 13. Shows the range of sizes of kai from the Ba and Rewa River sold at the market.

Figure 14. Range of individual shell weight of kai sold at the Ba market.

Figure 15. Length - weight relationship for kai from the Ba market.

Historical Data.

According to the information that were obtained from the kai fishers during this survey, the harvesting of kai was controlled by seasonal changes. It was reported that during the cooler months, less kai were harvested but in warmer months, more kai were harvested from the river due to the gonad content. The majority of people preferred eating kai when the gonads were ripe. Annual kai production from the Ba river is shown on Figure 16, along with the annual rainfall data obtained from the Fiji Meteorological Department in Nadi. The highest kai production in 1992 occurred during the year with the least rainfall. Figure 17 shows that there is a significant negative correlation between annual kai production and total annual rainfall.

Figure 16. Annual rainfall and annual kai production at the Ba river.

Figure 17. Relationship between rainfall and kai production in the Ba river.

8. Discussion

8.1 Limits of the Kai bed

Little is known about the factors that determine the upper and lower limit of the kai bed and further research is needed to elucidate this. Although Raj (1981), reported that the upper and lower limits of the natural distribution of kai, are generally governed by the tidal influence, there was no data to support this statement. Raj (1981), further mentioned that the lower limit is determined by the salinity at high tide and the clam does not inhabit the river bed below the salt wedge. Although this statement is useful as a basis for further investigation, unfortunately it is rather vague. Our interpretation of Raj's statement was that, the lower limit of the kai bed would be the point where the last traces of salt could be recorded at high tide.

The SPC-ICFMaP study showed that, kai were present on river beds between Nailaga and Kumukumu. The average salinity at Nailaga (lower limit) was approximately 8 ppt but this site experiences a salinity range between 2.90 ppt (ebb tide) and 16.70 ppt (flood tide). Our study showed that kai were present or inhabit river beds within the salinity range of 0 - 8 ppt (see Figure 4). Nailaga is approximately 4 kilometres from the mouth of the river. Salt water does penetrate beyond Nailaga and the last traces of salt could be detected as far as Nasolo, which is approximately 10 kilometres from the river mouth. Therefore, Nasolo would be the end point of the salt wedge. The lower limit of the kai bed (Nailaga) was approximately 6 kilometres below Nasolo. This result showed that kai can tolerate and dwell on river beds with average salinity level of 8 ppt. A question that should be posed is, should kai be still called a freshwater species? According to Redeke's classification of saline waters (Roberts, 1989), kai inhabit mesohaline (1.84 - 10 ppt), oligohaline (0.21 - 1.84 ppt) and freshwater (>0.21 ppt) areas of the river. Therefore kai may not be strictly classified as a freshwater species alone due to its presence in the mesohaline, oligohaline waters as well. Kai seemed to have reproduced under oligohaline and mesohaline conditions as well. It was observed that the distribution of juveniles, were not restricted to the freshwater zones alone but were present in large numbers from upper Nailaga (Station 4) and all the way up to upper Vaqia. It may still be appropriate at the moment for kai to be called a freshwater species but it should be noted that they have the ability to survive and live comfortably in oligohaline and mesohaline environment. Raj et al. (1979), supported this argument, and reported that the osmolarity and ionic composition of Batissa violacea blood indicates recent colonisation of freshwater. There is still a high possibility, that kai was a freshwater species and was pushed to the lower regions of the river through floods but through acclimation, kai has managed to survive well in mesohaline and oligohaline conditions.

Salinity seems to be the main factor responsible for determining the lower limit of the kai bed. However, it may be possible that salinity is not the only factor that dictates the lower limit of the kai bed. On Figure 4, the pH profile shows that, inland or towards the coast from Nasolo, the pH increased. The areas that hold the largest population of kai occur within the pH range of 7.0 - 8.0. Further research is needed to investigate the tolerance of kai to certain levels of salinity and pH. This may provide better information on the factors that determine the distribution limits as well as the growth of kai.

Our knowledge of the factors that dictate the upper limit of the kai bed is limited. Raj (1981), reported that the upper limit of the kai bed is determined by the velocity of water current. He stated that the highest point of distribution in a river is limited to a locality where the water current is reversed or nullified at high spring tide. While we agree that the current velocity may play some part in determining the upper limit of the kai bed, we felt that there are other factors involved. Further research on factors such as pH, temperature, substrate type, nutrients and dissolved gases should be conducted in-order to determine their influence on the upriver distribution of kai. On Figure 4, pH level increase further inland. It could be seen from the graph that pH at Nasolo was around 7.19 and was approximately 9.0 at Nabatolu. It was observed that kai were no longer present, at approximately 10 km below Nabatolu. The range of pH values in freshwater is wide but fish commonly live in the range 5.0 - 9.5 (Roberts, 1989). It has been observed in fish that, although they may tolerate and even reproduce in environments with a wide range of pH values, their optimum performance, defined in terms of fast growth or maximal reproductive capacity may well be restricted to within a much narrower pH. It is possible that kai may not survive at higher pH and this could be one of the factors that determines the upper limit of the kai bed. Since the pH increased as we move inland, it is possible to cause changes to the levels of factors such as dissolved carbon dioxide, ammonia and hydrogen sulphide. An increase in pH would mean a decrease in dissolved carbon dioxide, which is essential for aquatic plant life in terms of photosynthesis. The natural plant productivity provides the basis for fish and kai food production. Kai is a filter feeder and relies on the productivity of the water column.

All natural waters contain a certain amount of naturally occurring suspended solids. During flooding, suspended solid levels may rise considerably but kai usually survive these episodes quite well, although mechanical gill damage may have occurred. The tolerance of kai to different sizes of particulate materials and to certain exposure periods need to be investigated. It was noticed during our study that soil, gravel and sand particle sizes increased as we move further inland. Since kai are filter feeders, the gills may only tolerate certain sizes of particles. Due to high current velocity on the upper end of the river, large particles are consistently flushed down river. It is possible that kai gills may not have the ability to tolerate continuous exposure to large particles due to the possibilities of mechanical gill damage.

It is also important that the behaviour of kai to different light intensity is understood. We saw in this study that water clarity increased inland. There is a possibility that kai may only prefer certain level of light intensity. This may explain why they burrow deep so as to avoid high light intensities.

Sizes and Growth of Kai in Ba

It has been reported that Batissa can grow to a maximum shell length of 150 mm and are are slow growing (Morton, 1989). The result of this survey tend to agree with Morton's suggestion that kai are possibly slow growing. The histogram on Figure 6 shows that growth overlap was evident due to the absence of cohort separation pattern. It was also be seen from the graph that the spread of the curve was small and around 90% of the population were within plus or minus 10 mm around the population mean. Raj (1981), suggested that growth rate for kai was approximately 2 cm per year but unfortunately there was no data to support this. If this calculation was accurate, it would mean that approximately 90% of the kai in the Ba river were below 2 years of age. It would also mean that it would take more than a year for kai to reach the minimum market size of 25 mm in shell length. Despite the absence of supportive data, we believe that the growth rate of kai in the Ba river is lower than 2 cm per year.

Samples the of Rewa river kai, which were obtained from the Nausori market by the SPC-ICFMaP team in March, 1996, showed that the maximum shell length was approximately 120 mm. It was observed during this study, that the maximum shell length of kai obtained from Ba river was only 84 mm. This was quite small compared to the length of 150 mm that was reported by Morton as well as the samples obtained from the Rewa river. Figure 8 shows that the y and R² values for the length/weight relationship of kai in the Ba river were y= 0.0002x^2.9948 and R² =0.9893, whereas the values for the Rewa river were, y=0.0133x^2.0309 and R² =0.9492 (Appendix IV). There is a significant difference in the y values of the Ba and Rewa river kai.

From our observation, we found that 96.3% of kai collected, were below 40 mm in shell length and 63.3% were below 25 mm. Therefore, the majority of the kai were of small sizes. Since the Ba kai were smaller in sizes compared to the kai caught from the Rewa river, many questions have been raised concerning these differences. Prior to this study, it was suggested by the Fiji Fisheries Division that over-fishing was the reason for the small size of kai appearing at the Ba market. This study showed that only 24.92% of the standing stock was removed annually and the Ba river kai stock was still in good shape.

The reasons why the freshwater clams in Ba remain small in sizes, are not known yet. Normally animal growth is controlled by three main factors namely, genetics, environmental and social factors. Social factors such as competition for limited food sources or over-crowding should not be ruled out. It was observed during our study that the kai density was as high as 3,543 kai per m² on some parts of the river.

The effect of environmental factors such as water temperature, carbonate alkalinity, water hardness, natural acidity and pH on the growth of kai remains to be investigated. Optimal environmental conditions suitable for kai growth and reproduction have to established. All aquatic animals have upper and lower tolerance limits and optimum temperatures for growth, egg incubation, food conversion and resistance to specific diseases. These optima may all be different and may change according to the parameter conditions such as oxygen tension and water pH. Water temperature in the Rewa river may be lower than Ba river and it is possible that kai grow well at lower temperature. It has been observed in prawns and other aquatic animals that certain levels of pH and water hardness suppress growth and reproduction. Since Ba is on the dry side and Rewa is on the wet side of Fiji, it is possible that environmental conditions in Ba and Rewa are different and that could be the reason why growth performances are different.

The effects of genetics on growth cannot be ruled out. Raj, 1981 reported that two ecomorphs of kai are present in Fiji. These are kai bukuvula and kai buli. He described kai bukuvula as thin, relatively elongated with eroded white patch at the shell hinge or umbo of the shell. Kai buli was described by Raj as relatively fat and short specimens. Our study of the Ba river showed that only one type of kai was present. All the kai observed matched Raj's description of kai bukuvula. It was observed that all the kai in the Ba river have eroded white patch at the umbo. We did not see any kai conforming to Raj's description of kai buli.

Since kai are present on several rivers in Fiji, it is also possible that populations might have been separated for considerable periods. This separation may have some impact on the differences in sizes observed between kai in the Ba and the Rewa river. The difference between the population of kai in Ba and the Rewa river could be examined using available modern techniques such as electrophoresis.

8.3 Distribution, Biomass and Density of kai

The mean density of kai on the Ba river was estimated to be within the range of 0.13 - 1002.38 kai per m² and biomass was estimated to be within the range 10.92 - 2017.39 grams per m² . Swami (1994), reported that the estimated biomass of the lower portion of the Ba river to be around 340 grams per m². This estimate was much lower than our observation. In our study, the estimated population biomass and density were 748.83 grams per m² and 214 kai per m² respectively for the lower portion of the Ba river. The differences in the population estimates could be mainly due to the differences in the survey methods used. When comparing the density between the Ba and Rewa rivers, Ba was much higher than the estimate for the Rewa river which was reported to be within the range of 6 - 36 individuals per m² and 98 - 805 grams per m² for density and biomass respectively (Raj, 1981).

It was observed in this survey that kai was contagiously distributed along or across the river. On some sites, transects were only about 5 - 10 metres apart and large differences in population densities were observed. Kai tend to follow the contagious distribution pattern where individuals clumped in certain areas and near zero in other areas. The density obtained from transects at station 5 ranged from 1.5 - 867.5 per m² . The wide difference in population density across the river was consistently observed at the other stations. Appendix III shows that the variance of the means at all locations were significantly higher than the means. Substrate type was seen as the regulating factor. It was observed that freshwater clams or kai were in abundance on areas where the substrate was made up of a mixture of fine sand, fine gravel and firm bluish black mud. Low numbers of kai were observed where thick layers of soft mud were present. Kai were absent on hard soil or rocky substrate.

An interesting piece of information was obtained from the Votua villagers, who reported that kai beds normally shift from one place to the other. According to the people of Votua, kai beds shift either upriver or down the river. It was reported that kai do not remain at one particular location all the time. This information was partly supported by Raj, 1981, who reported that in heavy floods, kai beds shift downstream. The upriver movement of kai is yet to be explained. No research has been done on the movement of kai and it would be interesting to tag some specimens to determine the movement patterns.

Although it was observed in this study that substrate type regulated the density and distribution of kai, a follow up survey would be necessary to consolidate this finding. It would be interesting to find out whether kai remain dense on one particular type of substrate all throughout the year or not. If kai do move up and down the river, it would be interesting to find whether they move to similar substrate or not. It is also possible that they may prefer different type of substrate at different times of the year.

Catch Per Unit of Effort

The catch per unit of effort on the Ba river ranged from 3.7 - 38 kg/person-hour, with a mean of 11.7 kg/person-hour. This figure was only slightly higher than the estimate reported by Richards (1994), which was 10 kg/person-hour. Catch ranged from 3.7 - 38 kg/man-hour. The difference in catch rate per fisher-woman, heavily depended on where fishing was carried out. Some of the fisher-women live far away from the prime fishing grounds, especially the people of Votua village. Normally the fisher-women would try hard to reach those good fishing grounds before low tide. When fisher-women are not able to reach good fishing grounds before low tide, fishing was carried out where ever was convenient but this may affected the days catch.

It was observed during this study that, the average time spent by the individual fisher-women on fishing per day was approximately 2.43 hours. Compared to reef fishing, the time spent on kai fishing was quite short. The possible reasons why the fishing period for kai was shorter when compared to reef fishing, could be due to the difference in water temperature coupled with the differences in the type of fishing gears used. In the tropics, seawater is normally warmer than freshwater and therefore fishers can fish comfortably for longer periods on the reef. The graph on Figure 5 shows that temperature decreases as you move inland from the river mouth. Another possible reason why the time spent on kai fishing was shorter, could be due to the simple fishing method used by the kai fishers. It was observed that the kai fishers in the Ba river were not using goggles, or fins for swimming, or wet suit to protect them from the cold. Swimming against the current without swimming fins is tiring and this may have contributed to the shortening of the fishing time. The only gears used were, flour bags for collecting kai, tyre inner tubes which were used as buoyancy support for carrying the bags of kai as well as for assisting the fishers while swimming along, especially crossing the deeper parts of the river, and a short rope that is used for tying the bag to the tyre inner tube.

Using the catch per unit of effort data to calculate the total catch for both the subsistence and commercial fishing, it was found that a total of 496.729 tonnes was harvested annually. The market figures extracted from the Fiji Fisheries Division showed that approximately 124 tonnes of kai are sold in Ba annually. This would mean that only approximately 25% of the kai harvested each year were sold in the markets and around 75% are used for subsistence needs. It should be noted that this data were collected within a short period of time of three weeks and further work and long term monitoring is needed to properly estimate the catch and effort data for the Ba river. There is a possibility of bias due to the limited catch and effort data collected during this survey period. It is also important to know that using the market data alone to calculate the amount of kai removed from the river annually would not be reliable and would greatly underestimate the true kai harvest.

Historical Data

The Ba market data from 1992 - 1995, obtained from the Fiji Fisheries database showed that the mean annual commercial kai production for Ba was approximately 124.51 metric tonnes. The highest production was recorded in 1992, with 169.9 metric tonnes and the lowest was in 1993 with 99.95 metric tonnes. Production again increased in 1994 and 1995. The big increase in production in 1992, followed by the big drop in 1993 could be wrongly interpreted by many. The Fisheries Division viewed it as a possible sign of over-fishing. Our study, as outlined earlier, showed that the kai resource in the Ba river was still in good shape.

During this survey, an attempt was made to try and provide some logical reasons for the sudden decline in production in 1993 compared to 1992. The kai fishers gave us some very good information on factors that regulate the harvesting of kai. According to our information sources during this study, the harvesting of kai was controlled by the weather condition and the maturity state of the gonad. It was reported that more people would go out fishing when the weather was fine or warm or when the river was not flooding. During the warmer months, fishing pressure were normally high because the gonads of the kai were full or ripe and consumers prefer to eat kai at that stage. It was also reported that the warm seasons were extended in some years due to low levels of rainfall. As a result, the kai gonads continue to be ripe for a longer period.

Appendix II showed that, kai production for 1992 was the highest but rainfall for the same year was the least, compared to the other 3 years (1993 - 1995) (see Figures 11 & 12). The production in 1993 (Figure 7) was the lowest but rainfall was the highest compared to the other years (see Figure 11). In summary, the increase in production in 1992 was mainly due to the extended period of drought that was experienced in Ba during the year. This extended period of dry weather and low level of rainfall, possibly extended the fishing period as well as the maturation period of the kai gonad and may have resulted in continuous high demand for kai from the consumers.

Appendix II shows that during normal years (1993 - 1995) , production were usually low during the months of June - September, but 1992 was an exceptional year. Production continue to be high during those months. This increase in production in 1992 was not only restricted to the Ba river, but throughout Fiji. Market data from 1986 - 1995 showed that, the total annual kai production in Fiji fluctuated around 1,300 metric tonnes, but in 1992 it went right up to 1,800 metric tonnes (Appendix IV). In 1992 , the whole of Fiji experienced a long period of drought and low rainfall.

9.0 References Cited

Anderson, E. 1994. Preliminary results, Ba river survey, 1994. Institute of Applied Science,

University of the South Pacific. Water quality assessment carried out by USP on behalf of the Fiji Sugar Corporation at Ba.

Bibi, H. 1991. Women and the kai fishery. Prepared for Qitawa, Fiji Fisheries Division newsletter.

Lewis, A.D. (ed.) 1985a. Fishery Resource Profiles: information for development planning.

Fisheries Division, Ministry of Primary Industries, Suva, Fiji: 90p. (partially updated in 1988).

Morton, B. 1989. The functional morphology of the organs of the mantle cavity of Batissa violacea (Lamark, 1797) (Bivalvia: Corbiculacea), Department of Zoology, The University of Hong Kong, American Malacological Bulletin, No. 7(1) (1989): 73 - 79.

Morton, B. 1984. A review of the Polymesoda (Gelonia) Gray 1842 (Bivalavia: Corbiculacea) from Indo - Pacific mangrove. Department of Zoology, The University of Hong Kong. Asia Marine Biology. (1984), 1: pp, 77 - 86.

Richards, R. 1994. Fiji Resource Profiles. Research Co-ordination Unit, Forum Fisheries

Agency & Fiji Fisheries Division, Ministry of Agriculture, Fisheries and Forest. FFA Report No. 94/4: pp 140 - 144.

Raj, U. 1981. Kai a Freshwater clam. University of the South Pacific, Paper re-presented at the kai workshop on the 8^th of June, 1991: pp, 10 -16.

Raj, U. & Fergusson, J.E. 1979. Osmotic and ionic composition of a tropical freshwater mussel Batissa violacea Lamarck (Lamellibranchia: Sphaeridae). School of Natural Resources, University of the South Pacific, Suva, Fiji. New Zealand Journal of Science, 1980, Vol. 23, 199-204.

Roberts, J. R. 1989. Fish Pathology. Second Edition. Institute of Aquaculture, University of Stirling, Stirling, Scotland. Bailliere Tindall, 1989.

Swamy, K. 1994. Kai Distribution in the Ba river. A survey carried out to map kai beds and to estimate biomass of kai in the lower part of the Ba river: pp 1 - 8.

10.0 Appendices

Appendix I. This table shows the density, biomass, and the number of transects that was taken at each location.

Appendix II. This table shows the amount of kai sold at the Ba market on each month in the years 1992 - 1995. The highest production was recorded in 1992 and the lowest was in 1993.

Appendix IV. Fiji's total commercial kai sales.

Appendix III. Shows the observations, location means and the variance of the overall mean.

Appendix IV. Relationship between Length and Weight of kai from the Rewa river. The value of (y ) is larger than the value for the Ba River (Figure 15).