The species was first documented during biodiversity assessments at shallow rocky reefs of Boca Cangrejo (coordinates: 28.406016 -16.314799, NE Tenerife Island, Fig. 1) in August 2020. This locality is a natural environment, hotspot for diving and recreational activities, located 11 Km south of the island’s main port hub in the capital, Santa Cruz. No individuals had been observed in July of the same year at Boca Cangrejo. A follow-up was performed at this locality during 2020–2021 to monitor its initial spread. A further tally was done at the end of 2024. All assessments were made by deploying 3 to 5 transects (10 × 2 metres) between 2 and 5 m of depth and counting the number of individuals or aggregates found.

In addition, and in the framework of the citizen science platform RedPROMAR (https://redpromar.org), observations of the species have been recorded in recent years at various locations on the island of Tenerife (Fig. 1). REDPROMAR is a network set up by the Canary Islands Government for the monitoring and surveillance of marine life in the Archipelago. Anyone can register and upload sightings, which must be accompanied by a precise location, date, and a picture of the organism. Sightings are validated by a team of specialists in various taxonomic groups.

All specimens examined were sampled at Boca Cangrejo, seven collected in January 2021 and eight in October 2021. Half the specimens were preserved in alcohol for genetic analyses, and the other half in formalin for morphological examination. The latter were dissected and examined under a stereomicroscope, and the cuticular lining of the siphons was excised to examine the siphonal spines with a light microscope.

A small piece (ca. 5–9 mm²) of the muscular mantle tissue of six individuals was extracted with REDExtract-N-Amp Tissue kit (Sigma-Aldrich), following manufacturer’s recommendations. A fragment of ca. 590 bp of the cytochrome oxidase I (COI) mitochondrial gene was amplified with the ascidian-specific primers Tun_forward 5’ TCGACTAATCATAAAGATATTAG 3’, and Tun_reverse2, 5’ AACTTGTATTTAAATTACGATC 3’ (Stefaniak et al. 2009). PCR amplification was done in 20 µL total reaction volume with 10 µL of REDExtract-N-Amp PCR reaction mix (Sigma-Aldrich), 0.8 µL (10 mM) of each primer, 6.4 µL of ultra-pure water (Sigma-Aldrich) and 2 µL of DNA at a concentration of ca. 5 ng/µL. PCR followed an initial denaturation at 94 °C for 5 min, 35 cycles of a denaturation step at 94 °C for 1 min, an annealing step at 50 °C for 1 min, an elongation step at 72 °C for 1 min, and a final elongation step at 72 °C for 7 min. The amplified DNA was purified with Exo-SAP. Sequencing was carried out (both strands) at Macrogen facilities (Netherlands). The resulting sequences were assembled, edited and aligned in BioEdit v.7.2.6 (Hall 1999). The final alignment length was 584 bp.

In order to place the new sequences in a phylogenetic context, we compiled a dataset of StyelidaeCOI sequences available in GenBank. We obtained sequences of all genera available with enough coverage (90% or above) of the sequenced region. We included representative species of these genera and, whenever possible, two sequences per species for a better clade sorting. We also downloaded all sequences from the genus Cnemidocarpa and, in the case of C. verrucosa, we obtained representatives of the two clades (spA and spB) recognized by Ruiz et al. (2020). Three unpublished sequences of Cnemidocarpa in GenBank, two identified as C. areolata (a synonym of C. irene, see Discussion) with accession numbers KX138515 and KX138514, and a Cnemidocarpa sp. (KX650767), were not included in the final dataset. The three of them were > 90% similar, but blast searches of these sequences showed a low similarity with any other ascidian. The best hits were ca. 78% similarity with a Pyuridae (Microcosmus squamiger) and with a Styelidae (Botrylloides sp.). Similarities with all other Styelidae sequences in the dataset had a mean of 72.8%. As there is no publication and we could not verify their identity, we considered these divergent sequences as being misidentified.

The aligned sequences were trimmed to a common length (536 bp) and checked with the modelTest function of the R package phangorn (Schliep 2011) to select the best-fit evolutionary model of nucleotide substitution based on the Akaike Information Criterion (AIC). This model was then selected in a maximum likelihood tree search in Mega X (Kumar et al. 2018) with default options and 100 bootstrap replicates. A sequence of the pyurid Halocynthia papillosa was used as an outgroup. A fasta file with the alignment is provided as Suppl. material 1.

Cnemidocarpa irene is a solitary species of the family Styelidae. However, it is most often present in nature forming aggregates of 2–10 individuals. For simplicity, we will hereafter use “aggregates” as the abundance unit (albeit occasionally what we counted were single individuals). In the monitored point at Punta Cangrejo, the abundance of C. irene increased in little more than one year since first detection (August 2020) to values over 1 aggregate per square meter. Four years after arrival, it reached densities of 1.9 aggregates*m^-2 in December of 2024 (Fig. 2).

After identification of the species in 2021, we could in retrospect identify some observations previously registered at the RedPROMAR citizen science platform (RedPROMAR 2025), but as yet unassigned. The identification was based on pictures of individuals matching the general appearance of the species and in which the siphons are open and their typical aspect (see below) could be ascertained. This pushed back C. irene’s arrival in Tenerife to 2018 (Fig. 1). Since that year, in the Canary Islands, this species has only been observed on Tenerife. The first records in the platform (2018–2019) were located in the northeastern part of the island. Subsequently (2020–2022), additional sightings were recorded at other locations along the eastern coast. In recent years (2023–2024), sightings have also been reported along the southern and western coasts. In total, 74 sightings have been documented on the RedPROMAR platform during this period (Fig. 1).

The synonymy of the species is complicated (see Discussion), and detailed morphological descriptions are necessary to verify the taxonomic status of any new finding. Thus, we aim here to provide a comprehensive account of the external and internal morphology of the specimens collected.

The individuals measure up to 5 cm in maximal dimension. The tunic is reddish to orange in live specimens (turning brown in fixative) (Fig. 3A–E). The siphons are distinctive, with the inner surface covered by a velum extending almost to the siphon rim. This gives the inner siphons a marbled appearance, predominantly whitish, yellowish or greenish in tone. In some specimens, the inner surface of the siphons is darker, brown to reddish, featuring yellow-green patches (Fig. 3F). Scale-like spines (20 µm in width) are present in the upper inner tunic of the siphons, but they are few and hard to observe. The tunic surface ranges from near-smooth to rough with tubercles of several sizes. The ascidians are covered by other epibionts, mainly algae species. However, when individuals are found in areas with sea urchins they are cleaner than otherwise, likely due to the effect of sea urchin grazing upon epibionts. The ascidians are fixed ventrally and posteriorly, and the intersiphonal distance is ca. half the total tunic length. It is not uncommon to see individuals on the carapace of sponge crabs (Dromia spp., Fig. 3E).

The tunic is consistent but not coriaceous, it has an inner whitish layer and the mantle is easily separated from it. The mantle is dark-brown, weakly muscular, and allows the gonads to be seen from outside (Fig. 4A, B).

The oral siphon is lined by a variable number (20–30) of simple tentacles (Fig. 4C) arising from a short membrane with a denticulate margin. The pericoronal area is narrow, and the aperture of the neural gland is U-shaped with horns more or less rolled (Fig. 4E, F). The branchial sac has four folds on each side that are well separated and do not overlap. The dorsal lamina is simple and low. The branchial formula varies among individuals, but for a 4.5 cm specimen at the posterior part of the branchial sac it was:

Right: Dorsal lamina-5-(14)-7-(18)-5-(13)-5-(14)-5-Endostyle

Left: Dorsal lamina-7-(12)-5-(14)-5-(13)-5-(14)-5-Endostyle

There are up to 7–10 stigmata in a branchial mesh (Fig. 4G), and parastigmatic vessels are sometimes present, particularly near the dorsal lamina.

The digestive system forms a closed loop, with the ascending and descending branches of the gut close together (Fig. 4C). The secondary loop is wide open. The digestive system does not reach the posterior end of the body, leaving a free space. The stomach is sac-like with a smooth outer surface, although internally there are up to 20 laminar plications (Fig. 4D). The anus is multi-lobed and opens near the atrial siphon (Fig. 4H). In this same area open the gonoducts, and a ring of filiform tentacles is present at the basis of the atrial velum (Fig. 4H).

The gonads are elongated and tubular, in variable numbers. We have found generally four (sometimes three) on the right side, but often some of them are bifurcated (one gonad divides distally in two branches) or two gonads are fused distally (e.g., Fig. 4C). On the left side there are generally two (sometimes three) gonads, anterior to the digestive loop. The gonads have the male follicles in the lower (mantle-wise) part, while the upper (branchial-wise) portion contains the oocytes. In all cases the gonads were ripe (samples examined were from January and October). The sperm ducts from the male follicles surround the oocyte mass and join in the upper part of the gonad into a common sperm duct that runs the length of the gonad until the distal male papilla situated just above the female opening (Fig. 4I).

There are a few endocarps interspersed with the gonads on the right side of the body, but they are more frequent both inside the gut loop and following the outer margin of the intestine curve (Fig. 4C).

We obtained 4 different haplotypes from the 6 individuals sequenced (H1, H2, H3: one individual each; H4: three individuals). The sequences have been deposited in GenBank (accession numbers: PV849603 to PV849606). Sequence identities between them ranged from 0.917 to 0.984, being H2 the most divergent with respect to the other three haplotypes. They were added to a database of 48 sequences of Styelidae encompassing 14 genera (Suppl. material 1). The modelTest function revealed that the best-fit model of nucleotide selection for this dataset was the General Time Reversible model with a gamma distributed rate variation among sites and a proportion of invariable sites (GTR+G+I). This model was input in the ML algorithm of Mega and the corresponding phylogenetic tree obtained is depicted in Fig. 5 (G parameter = 0.731, I parameter = 43.76%).

It can be seen that the four haplotypes obtained formed a monophyletic clade with high bootstrap support. In turn, this clade joins the one formed by 4 sequences of Cnemidocarpa verrucosa and 2 of Asterocarpa humilis (each species forming a separate clade). The sister clade to this group includes the five Styela species in the dataset. It is noteworthy that the only other Cnemidocarpa sequence included, C. finmarkiensis, was rather divergent from the ones in the Cnemidocarpa + Asterocarpa clade, forming part of an unresolved polytomy.