When invasion biology meets taxonomy: Clavelina oblonga (Ascidiacea) is an old invader in the Mediterranean Sea

Taxonomic issues often confound the study of invasive species, which sometimes are unrecognized as introduced in newly colonized areas. Clavelina oblonga Herdman, 1880 is an abundant ascidian species along the southeastern coast of the United States and the Caribbean Sea. It was introduced into the eastern Atlantic and Brazil decades ago. In the Mediterranean Sea, a similar species had been described as C. phlegraea Salfi 1929 and reported from southern Italy and Corsica. In the last few years a species of Clavelina has proliferated in the embayments of the Ebro Delta (NW Mediterranean), a zone of active bivalve culture industry where it has smothered mussel spat, leading to economic loss. We here report the morphological and genetic identity of this species, synonymizing the Atlantic C. oblonga and the Mediterranean C. phlegraea (the latter therefore is a synonym of the former). Thus, C. oblonga has existed in the Mediterranean for over 80 years, but was known under a different name. We also found this species in natural habitats in the Iberian Atlantic coast close to the Strait of Gibraltar, raising concerns about an ongoing expansion. In order to obtain information relevant for management, we monitored growth, reproductive cycles and settlement patterns of this ascidian on bivalve cultures in the Ebro Delta. Its biological cycles were markedly seasonal, with peak abundance and reproduction during the warmest months, followed by regression during the cold season. The settlement period was short, mostly concentrated in a single month each year. Avoidance of mussel and oyster seeding during late summer and early autumn can readily reduce the damage caused by this species.


INTRODUCTION
The universal primers HCO2198 and LCO1490 (Folmer et al. 1994) were used to 142 amplify a fragment of the mitochondrial gene cytochrome c oxidase I (COI). PCR  ropes were taken out of the water for a few minutes, laid on a flat surface, and 163 mussels and ascidians were gently stretched out to avoid overlaps. The ropes then 164 were photographed with a digital camera together with a ruler scale and immediately 165 returned to sea. The perimeter of each colony in each photograph was manually 166 outlined with Photoshop CS4, and colony areas were determined by the Laboratory 167 of Image Analysis of the Scientific and Technological Center at the University of 168 Barcelona. The total area of the colonies on each rope (cm 2 ) was divided by the total 169 rope length (m), to obtain a relative estimate of abundance in cover area/length 170 (cm 2 /m) each month. 171 Second, we deployed plates to study the growth cycle. In a mussel raft located in the 172 center of the bay we placed three ropes in December 2011, each with three PVC 173 plates (20 x 20 cm) at three depths: 20 cm, 1m, and 2m, which were separated by tens of meters. Experiments were monitored monthly until December 2013 (except 175 for November 2012). The PVC plates (both sides) were photographed and processed 176 as above, except that cover was calculated as percent area of the colonies relative to 177 the total surface area of the plate. The colonies could be easily delineated in the 178 photographs, even if made up of a single zooid, as they form whitish masses (Fig. 179 S1). The congeneric species C. lepadiformis (Müller, 1776) was occasionally present 180 on the plates, but could be clearly differentiated by its transparent tunic and white 181 lines in the branchial region.  randomly, with each one from different rafts and preserved in situ in seawater with 10 194 % formaldehyde. Colonies then were dissected under a binocular microscope and 195 ten zooids were randomly selected per colony to determine their reproductive status.

196
Like most colonial ascidians, this clavelinid is a hermaphrodite that broods its 197 offspring. We categorized each zooid as follows: (1) immature, (2) presence of 198 testes, (3) presence of ovary, and (4) presence of brooded larvae. Since stages 2, 3, 199 and 4 are not mutually exclusive, we assigned each zooid to the most advanced 200 stage observed. A maturity index (MI) per month followed López-Legentil et al. 201 (2005), by averaging the category numbers of ten zooids per colony and calculating 202 the mean of five colonies.

203
To assess the recruitment pattern of Clavelina sp., we supplemented the same 204 mussel raft where the ropes with permanent plates were located (see above) with 205 three additonal ropes having PVC plates (20 x 20 cm) at three depths (20 cm, 1 m, 206 and 2 m), as in the previous experiment. We replaced these plates monthly (again except November 2012), took close-up photographs, and counted the number of 208 colonies established on them to estimate recruitment per month and depth. The new 209 colonies could be easily counted in the photos as they were formed by a single or a 210 few zooids.     (Table 1)     Patterns of MI and temperature appear to match, with cross-correlation analysis 330 showing a significant positive correlation at time lag 0, as well as at the two previous 331 months and the following one (lags -1, -2, +1; Fig. 4). Moreover, MI was significantly 332 positively correlated with the present and previous months' salinity (Fig. S4), while 333 levels of O 2 and chlorophyll a were uncorrelated with MI (Fig. S4). On the other hand,

334
MI also was significantly correlated with the coverage of the present, previous, and 335 following months (lags -1, 0, +1, Fig. 4). Some recruitment occurred on the plates during summer 2012 (reaching mean 337 values of ca. 14 recruits per plate in July), peaking markedly in October to more than 338 100 recruits per plate at 20 cm (Fig. 6). In December 2012 only three recruits 339 occurred in total. No recruitment was observed afterwards, until July 2013, when a 340 single recruit was found (at 2 m). In 2013, recruitment was more intense than the

361
We conclude, therefore, that C. oblonga and C. phlegraea are the same species, with 362 the former name having precedence. It is unfortunate that the type specimen of C. 363 phlegraea is unavailable and that no material from the type locality (Fusaro Lagoon, 364 SW Italy) could be found, despite repeated attempts and a thorough survey. This is  side of the river), which has slightly lower summer temperatures.

437
Clavelina oblonga was more abundant at one and two metres in the permanent 438 plates than on the shallower ones. Competition with other species that also grew on 439 the plates near the surface (such as Styela plicata) could explain this pattern.

440
Moreover, coverage on permanent PVC plates was significantly lower in the second 441 monitoring year, when there was an almost two-year-old community on the plates 442 (including other ascidians, mussels, bryozoans, polychaetes, algae, and sponges), 443 and temperature was somewhat different (a cooler spring than the previous year). 444 Although biotic and abiotic effects on C. oblonga remain little studied, previous work caused by the fishermen's attempt to obtain an extra cohort of marketable mussels, 459 involving mussel re-seeding in July, which became completely covered whit mass C.      Fgure S3. Cross-correlation analyses relating the mean monthly coverage (mean percent cover in the permanent PVC plates) of Clavelina sp. with chlorophyll a (µgL -1 ), salinity, and levels of O 2 (%) in the water column. Data series were lagged with respect to one another and the Pearson correlation coefficient computed for each time lag (months). The curved lines represent the threshold for significant (p=0.05) correlation values. Figure S4. Cross-correlation analyses relating the Maturity Index (MI) of Clavelina sp. with chlorophyll a (µgL -1 ), salinity, and levels of O 2 (%) in the water column. Data series were lagged with respect to one another and the Pearson correlation coefficient computed for each time lag (months). The curved lines represent the threshold for significant (p=0.05) correlation values.