During survey effort in May 2000 and February 2001 covering 545 km in the inner and middle portions of the lake (Thale Luang), north of Papayurn Island, Beasley et al. (2002) recorded only four sightings and calculated a sighting rate of 0.03 dolphins/linear km (mean group size = 4.3 dolphins, S.D. = 2.9, range = 1?8). All sightings were made in the inner portion, which is the deepest part of the lake (2.1?2.5 m). Due to shallow water and the extremely high density of fixed fishing gear, Beasley et al. (2002) proposed that dolphins were probably absent from the outer portion of the lake (Thale Sap) and the southern part of Thale Luang. If their supposition is correct, it means that dolphins do not move out of the lake into the Gulf of Thailand, or vice-versa, making the Songkhla population permanently and completely isolated. The same researchers presented additional evidence of the population?s geographic isolation based on interviews with 86 local fishermen. All the fishermen who stated that they had seen dolphins in the lake (69%) fished in the inner portion of Thale Luang (where the sightings were made during the 2000 and 2001 surveys), whereas the remaining fishermen, who had not seen dolphins, fished either to the north or south. Geographic isolation of the population is also supported by observations made during the line-transect survey described below.

In September 2003 a line-transect survey, covering 234 km in 21 hours of search effort, was conducted in Songkhla Lake (B. Smith, unpublished). Sighting conditions were good but no dolphins were seen. The intention had been to search the entire lake, following parallel track lines spaced 3.5 km apart. However, shallow water, dense sea grass and an extremely high density of fixed fishing gear prevented the survey vessel from following the designed track lines in the southern portion of Thale Luang and Thale Sap. Search effort was therefore effectively limited to the same area covered by Beasley et al. (2002). The 2003 survey was conducted using a similar vessel/observation platform and the same methods and number of observers as had been used for line-transect surveys for Irrawaddy dolphins in Malampaya Sound, Philippines (see Smith et al. in press). If the sighting rate in Songkhla Lake had been equivalent to that recorded in Malampaya Sound (0.0865 sightings/km; a subpopulation also proposed to be listed as Critically Endangered), 20 dolphin groups should have been detected. Even accounting for the difference in size between the area occupied by dolphins in Malampaya Sound (ca. 134 km²) and the area searched during the 2003 Songkhla Lake survey (ca. 755 km²), if the two water bodies supported approximately the same number of groups and individuals, 5?6 groups consisting of 27?32 individuals (based on the mean group size of 5.3 dolphins recorded in Malampaya Sound) should have been detected.

Currently, there is only a single connecting channel between the lake and the Gulf of Thailand. This channel is located at the southern tip of the lake and almost certainly inaccessible to dolphins due to the extremely high density of fixed fishing weirs (which remain set throughout the year) and gillnets (some of which are set throughout the year, others of which are removed when some fishermen practice other occupations during the monsoon season ? May through December, peaking from September through November). Previously a second smaller connecting channel existed at the northern tip of the lake. This has been blocked by a closure dam constructed to prevent saline inputs, which supplies irrigation water to intensive agriculture practiced in the fields surrounding the lake. The published records of Irrawaddy dolphins geographically closest to the southern connecting channel are of two specimens in 1901 at Pattani (ca. 100 km to the south; Bonhote 1903 ? not seen; cited in Pilleri and Gihr 1974) and one stranding in 1994 at Surat Thani (ca. 300 km to the north; Chantrapornsyl et al. 1996). The absence of records near the connecting channel (although this may also be explained by a lack of survey effort), the extremely high density of fixed fishing weirs that probably constitute a physical barrier to dolphin movement, and interview surveys indicating that dolphins do not occur in the southern portion of the lake imply that there is little, if any, demographic exchange between dolphins in Songkhla Lake and those in the Gulf of Thailand ? almost certainly fewer than one successful migrant per year (the criterion for defining a subpopulation according to the IUCN Red List Categories and Criteria Version 3.1). Although we have no information on the geomorphic history of the lake, the dynamic sedimentation and erosion processes typical of marine appended lakes would suggest that there were previous primary (i.e., non-anthropogenic) isolation events. These may have allowed population-level genetic differences to develop between the dolphins in Songkhla Lake and those in coastal waters of the Gulf of Thailand, although no data are available to investigate this possibility. A precautionary approach is to consider the dolphins in Songkhla Lake geographically isolated, pending future evidence to suggest otherwise.

There are no quantitative data for estimating population trends. However, circumstantial evidence indicates declining numbers. Compared to the 2001?2002 surveys reported in Beasley et al. (2002), the 2003 survey used three rather than two observers and search effort was conducted from a raised platform about two meters higher above the water level. This should have improved searching efficiency. Even with no change in efficiency, the 2003 survey should have produced at least one or two sightings. It is possible that the difference in results simply reflects random variation in sighting biases, but a precautionary interpretation would be that the very small population that existed in 2001 and 2002 has been reduced even further. This interpretation is reinforced by the high mortality experienced by the population (as evidenced by the large number of recorded deaths; see above) in relation to its extremely low (although precisely unknown) population size.

Decreasing