Lambir halfbeak project

From BioDivBorneo2010

Jump to: navigation, search

Halfbeaks of the Lambir Hills: At the Interface of Morphology and Stream Environment

Sachi Oshima and Alex Kim



Using 9 continuous measurements, we analyzed morphometric variation in a single species of halfbeak (cf. Dermogenys pusillus: Hemiramphidae) in Lambir Hills National Park (Sarawak, Borneo, Malaysia). In addition, environmental parameters were measured at the four sites from which the fish were collected. Analysis of similarities (ANOSIM) and clustering dendrograms generated using the two data sets failed to support the “monophyly” of the sites' fish. Possible explanations for these results are discussed.


The streams of the Lambir Hills are sluggish to swift-flowing, turbid to clear, and vary both in depth and substrate, providing a broad range of habitats for freshwater organisms. Freshwater fish are integral components of most tropical streams, occupying a number of trophic categories and exhibiting considerable variability in body form. Many of the fish species present in the Lambir Hills belong to exclusively freshwater families; however, there are also groups like the Hemiramphidae, or halfbeaks, which have both marine and freshwater representatives (Inger & Kong, 2002).

Morphometric analysis can be used to assess micro-evolutionary as well as macro-evolutionary relationships amongst organisms. Since selective pressures and environmental factors which affect development can differ between microhabitats, there may be a correspondence between physical form and site of origin, even within a continuously distributed species.

We chose to investigate how morphometric variability in the halfbeak (cf. Dermogenys pusillus) relates to physical and chemical qualities of the streams they inhabit in Lambir Hills National Park (Sarawak, Borneo, Malaysia). Studying this question may allow us to view the mechanisms of evolution on a fine scale.

Figure 1: Freshly-netted halfbeak


How does morphometric variability in the halfbeak (cf. Dermogenys pusillus) relate to physical and chemical qualities of the streams they inhabit in Lambir Hills National Park (Sarawak, Borneo, Malaysia)?

Investigation of this issue may allow us to view the mechanisms of evolution on a fine scale.


Field methods

Seven halfbeaks were collected from each of four sites (defined as a 10m stretch) along the length of the stream that flows from Latak Waterfall. We recorded the following environmental parameters at each site: temperature, total dissolved solids, conductivity, salinity, average stream depth, average stream width, and flow rate.

Chemical measurements were taken using an ExTech ExStik II. Width and depth (at the midpoint of the stream) were recorded using a tape measure at three points: the upstream end, downstream end, and midpoint of each site. Flow was recorded by placing a surface-floating leaf at the upstream end of the site and recording the time (in seconds) necessary for it to traverse the entire 10m length.

Lab methods

Using electronic calipers, we measured 9 traits adapted from Pethiyagoda et al. (2008) and Chakrabarty et al. (2008): total length, beak length, head width (at beak base), body depth (at operculum), orbit diameter, caudal penducle height, caudal fin length, and underbelly length (defined as the distance between the junction of the gills and the anal fin) of each fish, to the nearest 0.01 mm.


We formatted the environmental and morphometric data into two matrices, correcting the halfbeak measurements by dividing other eight measurements with total length, which we discarded. Next, we used the vegan package in R (v. 1.17-2 and v. 2.11.0, respectively) to generate dendrograms of Euclidean distance between specimens and, in a separate tree, between sites. Finally, by applying the “daisy” function to the morphometric data, a distance matrix was generated, and an analysis of similarities (ANOSIM) chart created.

We hypothesized that the halfbeak specimens would form four “monophyletic” clusters corresponding to their site of origin. These clusters would be topologically analogous (having the same hierarchical relationships) as the dendrogram derived from environmental data for the sites themselves. With respect to ANOSIM, our null hypothesis was that dissimilarity within each site would be equal to that amongst sites.


Out of our four sites, sites 1 and 2 were the most environmentally similar, while site 4 was the most dissimilar (Figure 1). In the dendrogram for relationships amongst individual halfbeaks (Figure 2), the four sites were not recovered as monophyletic units. Some pairs of individuals from the same site (e.g., 2C & 2F) were sister to one another, but no “clade” of more than 3 individuals was comprised entirely of representatives from a single site.

Furthermore, ANOSIM revealed that median amongst-site variance was greater than that between sites for all locations except site 2; the P-value, 0.257, was not statistically significant, so we failed to reject the null hypothesis of equal morphometric dissimilarity for fish amongst and between sites.

Figure 2: Cluster dendrogram of relationships amongst the sites (numbered from upstream to downstream), based on environmental data
Figure 3: Cluster dendrogram of relationships amongst halfbeak individuals, based on morphometric data
Figure 4: Analysis of similarities (ANOSIM) chart based on matrix of morphometric distances amongst halfbeak individuals


Analysis did not reveal a conclusive relationship between halfbeak morphology and the physical and chemical properties of the sites.

Our results may have been affected by our inability to sample sites that varied significantly from one another. Sites 1 and 2 were the most similar, which may be related to their physical proximity. Site 4 was the most dissimilar, which corresponds to the fact that it was located in a smaller stream, prior to a merge with the larger stream in which the 3 other sites were located.

There are many reasons that could explain the results seen in Figures 2 and 3. Compared to those from sites 3 and 4, specimens from sites 1 and 2 were more closely associated with one another. An explanation for this can be seen in Figure 3, which shows that the within-site variability for sites 3 and 4 was greater than that for the other two.

However, it may well be that population connectivity between the four sites is extensive enough (or environmental differences negligible enough) that their halfbeaks have not experienced significant differentiation.

Possible improvements for this study include: sampling a greater number of halfbeaks over a greater geographic area (multiple drainages) and considering more morphological measurements (e.g., meristic values like fin ray counts). Alternatively, genetic relationships could be assessed; the mitochondrial D-loop or microsatellites are likely to provide good resolution.


  • Inger, R.F. & Kong, C.P. (2002). The Fresh-Water Fishes of Borneo. Natural History Publications (Borneo) Sdn. Bhd., Kota Kinabalu, Sabah, Malaysia.
  • Pethiyagoda, R., Kottelat, M., Silva, A., Maduwage, K., & Meegaskumbura, M. (2008). A review of the genus Laubuca in Sri Lanka, with description of three new species (Teleostei: Cyprinidae). Ichthyol. Explor. Freshwaters, 19(1): 7-26.
  • Chakrabarty, P., Amarasinghe, T., & Sparks, J.S. (2008). Redescription of Ponyfishes (Telostei: Leiognathidae) of Sri Lanka and the Status of Aurigequula Fowler 1918. Cey. J. Sci. (Bio. Sci.), 37(2): 143-161.