To understand ecological and behavioral aspects of inshore migration of tropical eels, Anguilla spp., we collected 32 836 glass eels in the Poigar River estuary on north Sulawesi Island, Indonesia, in monthly collections from 1997 to 1999. The specimens were identified using morphological characters, and 3 species, A. celebesensis, A. marmorata, and A. bicolor pacifica, were found each year in fluctuating abundances. A. celebesensis was the most abundant species and comprised 73.5, 79.5, and 81.9% of all glass eels recruiting to the estuary of the Poigar River in 1997, 1998, and 1999, respectively. This species was relatively abundant in all 3 years with peaks during June in 1997 and 1998, and during September in 1999. A. marmorata was the second most abundant species and comprised 23.8, 18.8, and 17.7% of the yearly catches, respectively, and reached peaks in abundance during June in 1997 and 1998, and during January in 1999. A. bicolor pacifica comprised only 2.7, 1.7, and 0.3% of the yearly catches, respectively, with peak catches in June in 1997, in January in 1998, and in January and February in 1999. A. celebesensis and A. marmorata were collected almost throughout the year in 1997, 1998, and 1999, suggesting that in contrast to the temperate eels that recruit during half the year from winter to spring, these tropical eel species recruit to some degree throughout the year. Glass eels clearly were more abundant during new moon sampling events, with a total of 3075, 183, 3, and 90 glass eels collected in 10 sampling periods during new moon, waxing moon, full moon, and waning moon, respectively. All glass eels were collected at night (18:00 to 05:00 h) when light intensity ranged from 0 to 1.86 lux, and no glass eels were collected during the day (06:00 to 17:00 h) when light intensity ranged from 500 to 150 000 lux. During new moon, more glass eels were collected during flood tide than during ebb tide. These results suggest that a greater number of glass eels recruited to the Poigar River estuary at night, during new moon and during flood tide.

Environmental DNA (eDNA) has the potential to provide more comprehensive biodiversity assessments, particularly for vertebrates in species-rich regions. However, this method requires the completeness of a reference database (i.e. a list of DNA sequences attached to each species), which is not currently achieved for many taxa and ecosystems. As an alternative, a range of operational taxonomic units (OTUs) can be extracted from eDNA metabarcoding. However, the extent to which the diversity of OTUs provided by a limited eDNA sampling effort can predict regional species diversity is unknown. Here, by modelling OTU accumulation curves of eDNA seawater samples across the Coral Triangle, we obtained an asymptote reaching 1531 fish OTUs, while 1611 fish species are recorded in the region. We also accurately predict ( R ² = 0.92) the distribution of species richness among fish families from OTU-based asymptotes. Thus, the multi-model framework of OTU accumulation curves extends the use of eDNA metabarcoding in ecology, biogeography and conservation.

DNA barcoding opens new perspectives on the way we document biodiversity. Initially proposed to circumvent the limits of morphological characters to assign unknown individuals to known species, DNA barcoding has been used in a wide array of studies where collecting species identity constitutes a crucial step. The assignment of unknowns to knowns assumes that species are already well identified and delineated, making the assignment performed reliable. Here, we used DNA‐based species delimitation and specimen assignment methods iteratively to tackle the inventory of the Indo‐Australian Archipelago grey mullets, a notorious case of taxonomic complexity that requires DNA‐based identification methods considering that traditional morphological identifications are usually not repeatable and sequence mislabeling is common in international sequence repositories. We first revisited a DNA barcode reference library available at the global scale for Mugilidae through different DNA‐based species delimitation methods to produce a robust consensus scheme of species delineation. We then used this curated library to assign unknown specimens collected throughout the Indo‐Australian Archipelago to known species. A second iteration of OTU delimitation and specimen assignment was then performed. We show the benefits of using species delimitation and specimen assignment methods iteratively to improve the accuracy of specimen identification and propose a workflow to do so.

Remarkably little is known about the life histories of the many tropical anguillid eels distributed across the Indo-Pacific region, and since the Danish expedition to study eels in the region in 1928 and 1929, research on these eels has only begun again in recent years. Sampling for anguillid leptocephali in the Indonesian Seas has been carried out recently to learn about the spawning ecology and larval distributions of tropical eels there. The leptocephali of Anguilla marmorata, Anguilla bicolor pacifica, Anguilla borneensis, Anguilla interioris, and Anguilla celebesensis were collected around Sulawesi Island both in May 2001 and October of 2002. The development of genetic identification techniques has enabled these leptocephali to be identified to species, and their distributions and sizes during different seasons indicated that there are differing life history patterns among sympatric species in the region. A. celebesensis was found to have been spawning in Tomini Bay of northeastern Sulawesi Island in March and April 2001, but apparently, no spawning had occurred in the late summer and fall of 2002. Studies on anguillid glass eels have suggested that tropical anguillids may spawn throughout much of year, but our research on leptocephali in Tomini Bay and data on the downstream migration of tropical anguillids in the major tributary to Tomini Bay indicate that A. celebesensis may have a distinct seasonal pattern of spawning possibly related to the regional monsoon cycles. This is the first evidence of seasonality of spawning in tropical anguillid eels whose life histories are only just beginning to be revealed.

Environmental DNA (eDNA) has the potential to provide more comprehensive biodiversity assessments particularly for vertebrates in species-rich regions. Yet, this method requires the completeness of a reference database, i.e. a list of DNA sequences attached to each species, which is never met. As an alternative, a diversity of Operational Taxonomic Units (OTUs) can be extracted from eDNA metabarcoding. However, the extent to which the diversity of OTUs provided by a limited eDNA sampling effort can predict regional species diversity is unknown. Here, by modelling OTU accumulation curves of eDNA seawater samples across the Coral Triangle, we obtained an asymptote reaching 1,531 fish OTUs while 1,611 fish species are recorded in the region. Besides, we also accurately predict (R² = 0.92) the distribution of species richness among fish families from OTUbased asymptotes. Thus, the multi-model framework of OTU accumulation curves extends the use of eDNA metabarcoding in ecology, biogeography and conservation.