Fig. 1. Ochre stars Pisaster ochraceus sheltering in a crevice on the west coast of Vancouver Island, British Columbia during low tide
Sea star diversity on the Pacific coast of North America is the highest of anywhere in the world. On a single SCUBA dive one can expect to see a dozen or more species. Diversity of intertidal forms is less, but number of individuals, for example, ochre stars Pisaster ochraceus (Fig. 1) may be great.
NOTE “seastar” or “starfish”? According to the Oxford English Dictionary, the first literature reference to asteroids was in 1538 to a type of “sterrefyshe”, and the designation “sea star” did not appear until 1672. Contemporary invertebrate textbooks refer to asteroids as “sea stars” or "seastars", and both terms will be used in the ODYSSEY. The first part of the name reminds us that echinoderms live only in the sea. As for the pentaradiate symmetry implied by the second part, consider this. Of 37 species of asteroids in British Columbia, 30 have 5 arms (a classic "star"-shape), and 7 have more than 5, so the “star” designation seems mostly applicable. The sunflower star Pycnopodia helianthoides has up to 40 arms and, at sizes up to 1.3m across, is the largest-diameter asteroid in the world. Larger in mass than Pycnopodia, however, is the pink seastar Pisaster brevispinus, and this species is the largest-mass 5-armed sea star in the world (60cm dia)
NOTE the idea that P. ochraceus exists in 2 subspecies, ochraceus and segnis, appears to have been put to rest from results of a genetics investigation by a consortium of Mexican researchers. The researchers analyse tube-feet tissue of specimens from 17 localities spanning British Columbia to Baja California and find no significant genetic differentiation. Pisaster ochraceus apparently exists as an homogeneous species. Frontana-Uribe et al., 2010 p. 187 In, Echinoderms: Durham (Harris et al., eds.) Taylor & Francis Group, London
Frontana-Uribe et al. 2010 p. 187 In, Echinoderms (Harris et al., eds.) Taylor & Francis Group, London
Fig. 1. Possible echinderm phylogenies
Past morphological and recent molecular phylogenetic studies have created two hypotheses for echinoderm evolution named Asterozoa (G. “star” “animal”) and Cryptosyringida (Fig. 1). At the morphological level the basis of the asterozoan hypothesis is the similarity of the five-rayed body plan of the ophiuroids and asteroids, placing them in their sister clade. The second hypothesis is supported mainly from characters relating to comparative anatomy and embryology, most notably an apparent homology of the ophiopluteus and echinopluteus larvae, and similarities in certain aspects of embryonic development, placing the ophiuroids, echinoids, and holothuroids in their own sister clade. Past molecular phylogenetic analyses have bounced back and forth between the two hypotheses, with some supporting asterozoa and some cryptosyringida, but other researchers have suggested that the specific phylogenies derived in these studies may be sensitive to the particular analytical method being used. Most recent phylogenetic analyses, however, have favoured the asterozoa plan and, with strong supporting data from a recent phylogenomicanalysis done by a consortium of British, French, and American workers, the final nail may indeed have been driven into the Cryptosyringidacoffin.
NOTE the name cryptosyringida (G. “hidden” “pipe”) refers to the fact that in holothuroids, echinoids, and ophiuroids the radial water canals and radial nerves become covered during development, and these three classes are combined into the Super-class Cryptosyringida by the author who created the concept (A.B. Smith 1984 Palaeontology 27: 431). However, as noted above, the idea has fallen into disfavour
NOTE the analysis involves cDNA library construction using embryonic stages and some adult material from freshly collected specimens, with supporting nucleotide sequences obtained from online databases
Telford et al. 2014 Proc R Soc B 281: 2014479
Research Study 2
Suppose your research project involves photographic identification of miscellaneous benthic invertebrates obtained as bycatch in thousands of trawl sets with the aim, say, of determining appropriate siting of a Marine Protected Area1? You could hire a large cadre of people, train them extensively in invertebrate identification, then engage them in the tedious task of sorting and most likely error-prone identifications...or, you could apply a machine-learning methodology2 developed by a team of French scientists to automate the job. The methology requires first photographing each of thousands of species caught in a few representative bycatches, then incorporating them into a complex deep-learning computerised programme. With this as a start, each of your thousands of photos of trawl sets noted above could be automatically analysed for species content and outputted to your data bank. The authors show that their programme, based on the Faster Region based Convolutional Neural Network (RCNN) works, but at this early stage with only limited level of precision. Results depend upon the level of physical and colour distinction of each invertebrate. Thus, echinoderms with their relatively distinct images are identified to about 70% precision3, while arthropods and annelids with their variable morphology score at best only 50%, and sponges and cnidaria with their soft morphology score 20% or zero. Also scoring zero were molluscs but, although not specified, these may have been gastropods with foot extended or octoposes; certainly, squids and nautiluses would have higher scores. Still, the process holds promise, and with further fine-tuning is sure to improve. The project should attract great interest from fisheries management and related personnel, and the authors deserve kudos for their "thinking-outside-the-box" contribution.
NOTE1 this is just an idea, and in truth the primary aim of the researchers was to development computerised deep-learning methodology for invertebrates, but seemingly not for any specific application
NOTE2 machine learning or, in this case, deep learning, is an artificial intelligence approach that has been used forpattern recognition in many applications. The study is done on species collected in the Kerguelen archipelago area of south Indian Ocean (50o South Lat., 70o East Long.)
NOTE3 sea stars with their distinctive morphology and body colour scored highest, at a level explaining the inclusion of this technical report in this part of the ODYSSEY
Fig. 1. Typical trawl bycatch from the Kerguelen area, containing various species of echinoderms, ascidians, sponge and corals. The authors used over 92,000 individual images from nearly 2,000 bycatches as data-mining resources for the study
Fig. 2. Example of fairly precise detection and classification of several sea stars identified by computer with no human involvement
Fig. 3. Example of comparatively imprecise detection and classification of a mix of crinoids, seaweeds and a single gastropod (lower left) identified by computer with no human involvement
Select a TOPIC from the menu at the top of the screen
OR: play the animation to the left
OR: follow the snail's ODYSSEY by CLICKING on any X-marked invertebrate on the map above
Phylum Echinodermata (lit. “spiny skin” G.) including sea lilies, sea stars, sea urchins, sea cucumbers, and brittle stars
Class Asteroidea (lit. “like a star” G.), including sea stars
NOTE some recent name changes in west-coast asteroids include: bat stars changed from Patiria miniata to Asterinaminiata and then back again. Blood stars Henricia spp. remain a “puzzling complex”; however, whatever species name used by an author will be the one used in the ODYSSEY
Mah 2007 In, The Light and Smith Manual Intertidal invertebrates from central California to Oregon (Carlton, ed.) U Cal Press, Berkeley
Echinodermata
