|Image analysis techniques: a tool for the identification of bivalve larvae?|Hendriks, I.E.; van Duren, L.A.; Herman, P.M.J. (2005). Image analysis techniques: a tool for the identification of bivalve larvae? J. Sea Res. 54(2): 151-162. dx.doi.org/10.1016/j.seares.2005.03.001
In: Journal of Sea Research. Elsevier/Netherlands Institute for Sea Research: Amsterdam; Den Burg. ISSN 1385-1101, more
Analytical techniques; Identification; Imagery; Larvae; Zooplankton; Bivalvia [WoRMS]; Cerastoderma edule (Linnaeus, 1758) [WoRMS]; Crassostrea gigas (Thunberg, 1793) [WoRMS]; Macoma balthica (Linnaeus, 1758) [WoRMS]; Mytilus edulis Linnaeus, 1758 [WoRMS]; Marine
|Authors|| || Top |
- Hendriks, I.E., more
- van Duren, L.A., more
- Herman, P.M.J., more
Despite the importance of the planktonic larval stage in intertidal bivalves, our understanding of this stage is still insufficient. A major obstacle in the quantification of planktonic larval distributions is the identification of sampled larvae. Identification is difficult due to the uniform morphology of many larval species. We evaluated the morphology of bivalve larvae reared in our laboratory (Crassostrea gigas, Cerastoderma edule, Macoma balthica, Mytilus edulis) and literature data on larvae from the 1960s, using image analysis techniques. We used this dataset to compile species-specific dimensions (length-width of the larval shell) and shape parameters (contour of the larval shell). The first method yielded different slopes when length and width were plotted against each other, but regression lines overlapped, which rendered the technique impractical for field identification. Multidimensional scaling of larval shape within one species showed shape development of the larvae during ontogeny. Linear discriminant analysis did not produce results when the whole data set was used. But discriminant analysis on larger individuals (length > 150 μm) was relatively successful for species of which sufficient individuals were available. The identity of up to 74% of the large larvae could be predicted correctly.