Anticlockwise from top: Edible-nest Swiftlet in flight
Photo by Lim Hong Yao (2017)
, perched on nest
Photo by Lim Hong Yao (2018)
, harvested edible nest
Photo by Ediblebirdsnest (2016) on Wikimedia Commons, accessed from https://commons.wikimedia.org/wiki/File:Edible-birds-nest-bowl-shape.png
, prepared bird's nest soup
Photo by Crue (2010) on Blogger, accessed from http://ichikichikehem.blogspot.com/2010/01/thai-hatyai-bird-nest-soup-7ps.html
The Edible-nest Swiftlet has a large range starting from the South of Hainan island spanning southward along the coasts of Vietnam, Myanmar, Thailand and Cambodia, subsequently including Sumatra, Java, Lesser Sundas, Borneo, and West Philippines (Figure 1). Inland Peninsula Malaysia is part of its natural range but Edible-nest Swiftlets from house farms are currently inhabiting the area. 8 current subspecies are recognized. For more details, see subspecies taxonomy below.
Figure 1: Range map of the Edible-nest Swiftlet, adapted from HBW (2018)
Edible-nest Swiftlets are aerial insectivores that catch arthropods on the wing. Diet analyses have been conducted by examining regurgitated food boluses; hymenoptera (bees, wasps and ants), diptera (flies) and ephemeroptera (mayflies) made up majority of the food items, while arachnida (spiders and other arachnids), coleoptera (beetles) and hemiptera (true bugs) made up most of the rest.
The Edble-nest Swiftlets nest in natural caves, and man-made houses. Breeding has been observed to take place year-round but peak in October and February. (Nigel Langham 1979).
As Edible-nest Swiftlets nest in places which are completely dark, they have evolved the ability to echolocate using clicking sounds, along with many other species from the genus Aerodramus. This ability, however, is not diagnostic of the genus as the Pygmy Swiftlet (Collocalia troglodytes) has been proven to possess this ability too.
Sound recording of an Edible-nest Swiftlet's echolocating clicking calls. (Recorded by Lim Hong Yao, 2018)
The Edible-nest Swiftlet was originally described by Carl Peter Thunberg, a well renowned Swedish naturalist from the 19th century. Thunberg collected the specimen in Java and described it as Hirundo fuciphagus in his Kungl. Svenska vetenskapsakademiens handlingar (Remarks about the Swallows that build jelly-like, edible Nests) (1812) (Figure 3), which, with help of Google Translate, roughly states that
Figure 3: Scans of the original swedish description of H. fuciphagus in Kungl. Svenska vetenskapsakademiens handlingar by Thunberg (1812), obtained from Biodiversity Heritage Library.
Thunberg (1812) named the Edible-nest Swiftlets in Java Hirundo fuciphagus, mistakenly grouping them together with the swallows (Hirundo is a genus of swallows, which are passerines). By the 1900s swiftlets were recognized as a separate taxon Collocalini and all its members were lumped in the genus Collocalia, then subsequently divided into 3 genera: Hydrochous (Giant Swiftlets) are sister to Aerodramus (medium sized brownish swiftlets), and Collocalia (small glossy plumaged swiftlets). Multiple recent phylogenetic analyses have been conducted to establish that Collocalia are the basal group that are more related to the swifts, followed by Hydrochous and Aerodramus being sister genera. (Figure 3) ( Cibois et. Al, 2018; Päckert et. al, 2012; Price et. Al, 2005; Rheindt et. al, 2014)
Figure 4: Phylogenetic tree of swiftlets obtained using nuclear and mitochondrial DNA analysis, estimated with Bayesian and Maximum Likelihood methods. Node support is denoted as posterior probabilities/bootstrap values. Adapted from Cibois et. al (2018).
Many studies have attempted to resolve the phylogenetic relationships of the Edible-nest Swiftlet, but A. fuciphagus has often shown up as a paraphyletic clade with one or more individuals of A. fuciphagus are more closely related to A. salangana than other A. fuciphagus (Cibois et. Al, 2018; Price et. Al, 2005; Rheindt et. al, 2014). Even the most recent tree constructed using mitochondrial and nuclear DNA analysis depicts this (Figure 4) (Cibois et. al, 2018). The low genetic divergence between closely related taxa is likely due to occasional hybridisation events, which have been reported in Sabah (Lee, 1996), leading to genetic introgression in the form of mitochondrial DNA (mtDNA) sweeps (Rheindt & Edwards, 2011) that cause some individuals to appear closer to sister species than its own.
Furthermore, some authors disagree with classifying all 8 subspecies under a single species A. fuciphagus. The eight recognised subspecies for A. fuciphagus are as follows (descriptions are relative to nominate race unless stated):
Differences amongst subspecies are often subtle and difficult to distinguish in the field due to variations in lighting as well as difficulty in observing constantly fast-moving subjects.
According to the subspecies range, Edible-nest Swiftlets observed in Singapore should be A. f. germani which extends into the Malay Peninsula, but specimens collected appeared identical to the nominate race A. f. fuciphagus, likely because colonies in Malaysia and Singapore are of the house farmed variety (see House Farming below), which is suspected to be of Javan origin (ssp. fuciphagus). To complicate matters, their feeding ranges are likely overlapping.
Several authors believe that this species should be split into two or three. Table 1 on the below presents a summary of the 3 different treatments.
Table 1: Summary of the authors' different species treatments of the Edible-nest Swiiftlet complex.
|HBW & Stresemann (1931)||A. fuciphagus||all 8 listed above|
|Clement's checklist||A. germani|
fuciphagus, vestitus, inexpectatus, perplexus, dammermani, micans
|Cranbrook et al. (2013)||A. fuciphagus|
fuciphagus, vestitus, dammermani, micans
inexpectatus, germani, perplexus
The Clement's checklist treatment appears to be based on morphology and original descriptions of the subspecies, but does not appear to explicitly explain the treatment in any publication or platform. Despite this, many authors have adopted this treatment.
Stresemann (1931) postulated that populations with a paler rump from A. f. germani and those of a darker rump from A. f. vestitus and germani formed a transition zone over Peninsula Malaysia where intergrades of rump colour can be observed. Thus, it was proposed that the populations were interbreeding, and thus the taxa were regarded as subspecies of A. fuciphagus. Medway (1966) also arrived at a similar conclusion.
Cranbrook et. al (2013) on the other hand, re-examined the specimens used in Stresemann (1931) and postulated that there was not a cline in morphology, but rather an overlap of feeding range in Peninsula Malaysia. It was observed that the specimens could be grouped into two main groups: grey-rumped and brown rumped (Figure 5). Given that the rump colour was maintained as a character amongst these two groups with no gradation, it was concluded that these two should be grouped into two species, while amechanus was considered an endemic given its unusual glossy colouration and variable rump band. Grey-rumped swiftlets were grouped under A. inexpectatus while brown-rumped swiftlets remained as A. fuciphagus. Mitochondrial DNA analysis was also conducted in the study with both Maximum Parsimony and Neighbour Joining methods using cyt-b haplotypes but the nodes were all poorly supported and the tree appeared inconclusive regarding the phylogenetic relationships between populations (Figure 6).
However, upon examining the plates, it is apparent that rump colouration can vary significantly within a subspecies; A. inexpectatus germani of plate 3A (Figure 5) clearly shows an individual with a brown rump instead of a whitish rump as described for germani and shown in plate 1A. This raises concerns about using rump colour as a diagnostic trait to treat the species complex. Even if the morphological breaks separating the populations in different geographical ranges are real, in applying the Biological Species Concept, there is insufficient evidence to establish reproductive isolation given that the colonies breed in allopatry. Additionally, Cranbrook et. al (2013) suggested that the house farmed swiftlets could be a hybrid population of fuciphagus and inexpectatus, or fuciphagus and germani, indicating that these taxa may not withstand the test of sympatry when brought together in nest houses. Therefore, A. fuciphagus is currently treated as a single species (encompassing all 8 subspecies) on this page and by several other authors.
In considering the Phylogenetic species concept sensu Wheeler & Platnick (2000), most of these subspecies are likely to be elevated to the rank of species given that the different populations appear to have a unique set of character states in terms of size and plumage differences. However, more work needs to be done as well to adequately sample the different populations to establish the existence of these different character states.
Future taxonomic work is required to concretely establish the relationships between these taxa, both in terms of confirming morphological differences as well as investigating molecular evidence. Goh et. al (2018) recently investigated more house farmed Edible-nest Swiftlets from Peninsula Malaysia and found that they appeared to be closest to A. f. vestitus rather than A. f. fuciphagus (which was first to be recorded in the Malay Peninsula), but the node was not very well supported (bootstrap = 78) and no conclusive statements about their origin could be made. Given the many unsuccessful attempts with mtDNA thus far, it is likely that mtDNA is unsuitable as a marker for intraspecific studies as it is only maternally inherited. Coupled with the issues of mtDNA sweeps, it is evident that mtDNA is not an effective taxonomic indicator for the Edible-nest Swiftlets, and adopting genomic methods using Next Generation Sequencing is likely the way forward to unravel the true relationships between these populations.
Figure 5: Photo plates by Cranbrook et. al (2013) depicting the grey-rumped and brown rumped Edible-nest Swiftlet specimens examined.
Figure 6: Maximum Parsimony tree created using cyt-b haplotypes from Edible-nest Swiftlet samples by Cranbrook et. al (2013). Numbers at nodes are the Neighbour Joining/Maximum Parsimony bootstrap values (tree topology was identical for both methods).