Olifantiella, RIAUX-GOBIN & COMPERE

OLIFANTIELLA RIAUX-GOBIN & COMPÈRE

Olifantiella cf. rodriguensis Riaux-Gobin ( in Riaux-Gobin & Al-Handal 2012, Riaux-Gobin 2015) was present in Rik16 (east Mangareva), epizoic on Holothuria atra . The latter taxon was also present in Rik10 (same site, intertidal sand and macroalgae). Olifantiella cf. rodriguensis was rare, with no internal SEM view available to check for its buciniportula. A mucus thread was observed through the external opening of the buciniportula, confirming previous remarks on a possible physiological role for this process, in the relationships between the cell and the external environment ( Riaux-Gobin & Al-Handal 2012). A second taxon, O. pilosella Riaux-Gobin ( in Riaux-Gobin & Al-Handal 2012), with all original morphological characters, was found as epizoic in Rik16. A third taxon, Olifantiella muscatinei (Reimer & Lee) Van de Vijver, Ector & Wetzel (2018a , 1), was observed as an epiphyte in Rik4. Note that no Olifantiella specimens were observed in the holothurian scraping Rik5 (west Mangareva).

* From Mangareva (SEM), **,*** type [** Riaux-Gobin et al. 2016, *** Riaux-Gobin et al. 2008)]

Stria densities and axial rows in 10 µm, length & width in µm.

Taxon Length Width L/l RV SV str. SV axial RV str. rows fascia * Cocconeis cf. mascarenica not (n=27) 7.2 4.1 1.76 37.8 40.6 always no *** Cocconeis mascarenica not (n=31) 6.7 3.8 ca.1.7 38 25-35 always no

* Cocconeis sp. (n=8) 7.8 3.8 2.08 38.8 37.2 no no decussate up

* Cocconeis cf. diaphana (n=12) 14.4 6.6 2.20 39.8 52.8 pattern margins half

* Cocconeis cf. molesta (n=5) 9.1 6.6 1.38 46.9 52.8 no valve in zig-

** Cocconeis molesta Kütz. BM ca. 30 zag 22- <half

18381 (n=4) LM 16.4 9.7 1.96>30 23 valve

** Cocconeis diaphana W.Sm. in zig-

var. diaphana BM 23161 zag ca. short,

(n=11) LM 33-42 21-26 1.65 25 26 23 elliptic

** Cocconeis diaphana W.Sm. short, Isolectotype SEM ca. 38 ca. 23 1.65 22 35 unclear elliptic

** Cocconeis dirupta W.Greg. ca. zig- <half BM 1420 (n=16) LM 18-37 17-31 1.12 18.9 16.6 zag valve

EPIZOIC DIATOM ASSEMBLAGES

Rik5 and Rik16 were both cumulative scrapings of several individuals pertaining to a well circumscribed population of Holothuria atra collected in small areas (ca 3 m 2), 1) from Gatavake Bay (west Mangareva, Rik5), 2) from Rikitea nearshore (east Mangareva, Rik16). The two epizoic assemblages were highly dissimilar ( Table 2), with a higher number of taxa in Rik16 ( Table 2). Note that the assemblage in Rik7 (sediments close to the location of the holothurian scraping Rik5) and in Rik10 (sediments close to the holothurian scraping Rik16), are also highly different from what was found in each holothurian’s population. Some taxa present on the holothurians were absent from the nearby environment and vice versa, but the species richness is similar among epizoic and non-epizoic assemblages, being higher for Rik16–Rik10 (Rikitea environment) than for Rik5–Rik7 (Gatavake Bay). There were no diatom taxa unique to a particular epizoic population, or particularly dominant with regards to its abundance in the nearby substrates.

From a shallow and well-oxygenated zone (exposed intertidal zone), Rik9, a scraping of several living specimens of Rochia nilotica (Syn.: Trochus niloticus Linnaeus 1767 ), (ca. 5 cm in length, colonized by diminutive filamentous algae and encrusting red algae, probably pertaining to Corallinaceae, Maggy Nugges comm. pers.), was quantitatively diatom-poor, but with the presence of large diatoms such as Achnanthes cf. brevipes Agardh 1824 and Cocconeis pseudodiruptoides Foged (1975 , 18). Note that the latter taxon was only observed in Rik9 (see supplementary material), possibly linked to the ethology of this mollusk migrating from intertidal to subtidal substrates, whereas the other studied samples (except for the pearl oyster farm samples) were strictly intertidal (< 50 cm deep).

The scraping of empty green-colored tubes of Teredo sp. (probably colonized by unicellular chlorophytes) was almost void of benthic diatoms. This driftwood was subjected to drastic conditions such as high light exposure and intermittent desiccation.

The scrapings of living two-year old Pinctada margaritifera (‘Black-lipped Pearl Oyster’), before grafting (before nucleus transplant) (Rik12–14, Table 2), as well as a one-year old individual (Rik14bis), were surprisingly diatom poor, quantitatively and qualitatively, whereas the macroalgae living on mooring ropes at the same depth (Rik15) were colonized by a diverse array of diatoms, apparently absent from the oyster’s fouling. This low diatom colonization may be related to the fact that the oysters were cleaned every 3 months (via a passage to kärcher, followed by scraping with a knife), which drastically lowers the biofouling potentially detrimental to the oyster growth. This biofouling was composed of macroalgae and small fauna (i.e. bryozoans, ascidians, sponges, see also Lacoste et al. 2021) potentially colonized by diatoms. Note that the subtidal assemblages from the pearl oyster farm (epizoic as well as epiphytic, Table 2) seem to be characterized by the presence of taxa absent in intertidal samples from the same bay, such as Schizostauron citronella (Mann) Górecka, Riaux-Gobin & Witkowski ( in Górecka et al. 2021, 1480), Cocconeis pseudomarginata Gregory (1857 , 492) and Cocconeis peltoides var. archaeana Riaux-Gobin & Compère ( in Riaux-Gobin et al. 2011d, 330).

NMDS ANALYSES

In order to visualize the level of similarity among samples or groups of samples as a function of their diatom assemblages, an NMDS analysis was carried out. The full data are given as supplementary material. Note that the NMDS is testing the grouping of samples in relation to their assemblage compositions, and not whether a certain taxon is a reflection of a particular biotope or ethology. The analysis illustrated the relationships between samples ( Fig. 41 View Fig ).

The epizoic assemblages (blue points in Fig. 41 View Fig , upper diagram) slightly group (but not significantly, see PERMANOVA results below), while the other assemblages (epiphytic and epipsammic taxa) have no particular grouping ( Fig. 11 View Fig , upper diagram). Concurrently, the different locations and bathymetries appear to have a significant influence on the assemblages (see PERMANOVA results below) ( Fig. 41 View Fig , bottom diagram), with Hao atoll and Togegegie motu grouping together negatively, possibly due to the low species richness and particular calcic environment. The subtidal samples group positively on the left of the analysis (violet points, Fig. 41 View Fig bottom diagram) with a particular colonization in terms of species, as noted above in, for example Schizostauron citronella (Mann) Górecka, Riaux-Gobin & Witkowski and also Cocconeis coronatoides Riaux-Gobin & Romero ‘discoid morph’ (acronym cocid2, see supplementary material). The subtidal epizoic assemblage was slightly different from that growing as non-epizoic at the same depth (see supplementary material and violet point), but still grouping with the other subtidal samples. The samples from Gatavake Bay (red points, bottom diagram) group slightly differently from the samples from the Rikitea shore line (blue points).

The PERMANOVA results provide a statistical basis for the above remarks, and indicate that the biotope (upper diagram) did not have a significant effect on the community dissimilarity (Permanova R2 = 0.10, F = 1.14, 2df, p-value = 0.262). In contrast, the effect of the location (bottom diagram) was statistically significant (Permanova R2 = 0.42, F = 3.083, 2df, p-value = 0.001). Furthermore, results from an ad-hoc pairwise comparison among locations (excluding location H - Hao atoll- because of its small sample size), indicated that site S (subtidal samples, pearl oysters) was statistically significant from the rest (p-values <0.01), site R (Rikitea) was statistically different from sites G (Gatavake Bay) and M (Totegegie motu) (p-value <0.01), and sites G and R were also statistically different (p-value <0.01).