Sacoglossans are specialized marine herbivores that tend to have a close evolutionary relationship with their macroalgal hosts, but the widely distributed species Elysia viridis can associate with several algal species. However, most previous investigations on the field abundance and size distribution of E. viridis have focussed on Codium spp. in the British Isles, and algae from this genus are considered superior hosts for E. viridis. In the present study, we investigated the abundance and size distribution of E. viridis on 6 potential host algae with differing morphologies (the septate species Cladophora sericea, Cladophora rupestris, Chaetomorpha melagonium, and Ceramium virgatum, as well as the siphonaceous species Codium fragile and Bryopsis sp.) at 2 sites on the Swedish west coast over the course of a year. In spring, slugs were almost absent from all algal hosts. In summer and autumn, E. viridis consistently occurred on several of the algal species at both sites. The highest number of small E. viridis were found on C. sericea, intermediate numbers of significantly larger E. viridis were found on C. rupestris, while fewer, intermediate sized animals were found on C. fragile. Throughout the study period, only a few E. viridis individuals were found on C. melagonium, Bryopsis sp., and C. virgatum. Our results indicate that E. viridis is an annual species in Sweden, capable of exploiting co-occurring congeneric and intergeneric algal hosts with differing morphologies. These results corroborate previous findings that E. viridis can exploit several different algal species, but does not indicate that C. fragile is a superior host.
Citation: Baumgartner FA, Toth GB (2014) Abundance and Size Distribution of the Sacoglossan Elysia viridis on Co-Occurring Algal Hosts on the Swedish West Coast. PLoS ONE 9(3): e92472. https://doi.org/10.1371/journal.pone.0092472
Editor: Arga Chandrashekar Anil, CSIR-National Institute of Oceanography, India
Received: December 21, 2013; Accepted: February 22, 2014; Published: March 19, 2014
Copyright: © 2014 Baumgartner, Toth. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This work was financially supported by the Swedish Research Council through contract no. 621-2007-5874, the Faculty of Sciences University of Gothenburg through the strategic research platform in Marine Chemical Ecology (MARICE), and via grants from the Colliander foundation. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Sacoglossans are an intriguing clade of predominantly herbivorous opisthobranch molluscs that are considered one of the few examples of specialized herbivores in the marine environment. Each sacoglossan species is generally restricted to a single genus or species of macroalgae and tend to have a close evolutionary relationship with their algal host(s) . A defining characteristic of the group is their possession of a uniseriate radula consisting of intricate modified teeth adapted to pierce different types of algal cells from which they suck cytoplasm . Furthermore, many species are capable of sequestration and/or conversion of algal chemical defence compounds , retention of functional algal chloroplasts from their macroalgal diets, known as kleptoplasty , and demonstrate crypsis in both morphology and colouration .
Despite these adaptive traits having caught the imagination of many researchers, interpretation of their ecological relevance has been limited. This likely stems from a paucity of ecologically relevant field data on host associations in natural sacoglossan populations ,  (but see , ). Observational field data are important in providing an ecological basis to test and/or explain how mechanistic studies fit within the contextual framework of natural systems . Although a substantial amount of data are available on sacoglossan-algal host associations in the laboratory , , with more insight now being gained from the use of molecular techniques to accurately identify sacoglossan diets (e.g. –), quantitative observational field data of abundance and size distribution on algal hosts are still limited for most species (but see , , –).
In the present study, we focus on the widely distributed sacoglossan species Elysia viridis, which can be found on northeast Atlantic coasts from Norway, south, including the British Isles, into the Mediterranean Sea . Unlike many sacoglossan species, E. viridis associates with several genera of macroalgae (e.g. ). However, the most commonly cited food sources for E. viridis are green macroalgae from the siphonaceous genera Bryopsis and Codium, and the filamentous septate genera Cladophora and Chaetomorpha . E. viridis collected on Codium spp. were larger than individuals collected from Bryopsis plumosa, but smaller than animals collected on Chaetomorpha sp. . Furthermore, slugs fed Codium fragile grew to a larger size than those fed Cladophora rupestris, indicating that C. fragile ssp. are a superior food source for E. viridis compared to C. rupestris .
Despite being considered one of the better-studied sacoglossan species in terms of ecology , , reproductive biology , , and kleptoplasty –, there are surprisingly few rigorous field studies on the abundance and size distribution of E. viridis on different algal host species. Previous observational field abundance and size data have come from the British Isles, often focussing on macroalgae in the genus Codium (e.g. , , ). Trowbridge and Todd  did search for E. viridis on C. rupestris in Scottish slug populations; however, they only found E. viridis on C. rupestris when the introduced host C. fragile ssp. tomentosoides was absent. Furthermore, field populations of E. viridis were reported on algal hosts belonging to different genera and species including Chaetomorpha sp., Cladophora sp., C. rupestris, Bryopsis sp., and Griffithsia sp. . However, the study by Trowbridge and co-workers  was not designed to statistically compare quantitative data on the abundance and size distribution of E. viridis on different co-occurring algal hosts.
On the west coast of Sweden the occurrence of E. viridis has been reported previously , , but no study on its distribution among algal hosts has, to our knowledge, been conducted in Scandinavian waters. Several genera that are potential algal hosts for E. viridis such as Codium, Cladophora, Bryopsis, and Chaetomorpha are fairly common and often co-occur at several sites in this region. In this study, we sampled 6 potential algal hosts during different seasons at 2 sites on the Swedish west coast in order to 1) provide data on the seasonal patterns of abundance and size distribution of E. viridis on algal hosts in Sweden and 2) establish if E. viridis commonly colonise co-occurring intergeneric and congeneric algal hosts.
Materials and Methods
Sampling area and algal hosts
Abundance and size of E. viridis on co-occurring algal hosts was assessed via 4 surveys conducted by snorkelling at 2 sites (<1–5 m depth) in the Koster Fjord, Sweden (Yttre Vattenholmen 58° 52′ 33.5″ N, 11°6′ 22.9″ E and Saltö Lyngnholmen 58° 51′ 45.3″ N, 11° 7′ 52.8″ E) from autumn 2010 to autumn 2011. No specific permissions were required to sample organisms at these locations and the study did not involve any endangered or protected species. Sites were characterized by predominantly submerged (tidal range <0.3 m, ) sloping to vertical rocky outcrops covered by algae.
E. viridis are difficult to locate directly in the field so instead of screening all potential algal hosts present at a site, the surveys focussed predominantly on 6 species. 3 green algal species, the filamentous septate species C. rupestris and Cladophora sp. (likely sericea/albida complex but will be referred to as C. sericea henceforth) and the siphonaceous planar C. fragile, were selected as they commonly co-occur at both sites in reasonable abundance during the summer and autumn months. Furthermore, a combined total of 1187 E. viridis were found on these hosts across both sites during a pilot survey conducted in summer 2010 (late July/early August). In addition, a co-occurring dominant red alga (Ceramium virgatum), which to our knowledge has not previously been reported as an E. viridis host, was selected to assess whether the number of slugs found on different hosts were due to the relative abundance of the algae rather than specific algal-herbivore associations. Finally, individuals of Chaetomorpha melagonium and Bryopsis sp. were collected during some surveys. C. melagonium, whilst present during most surveys, was not particularly abundant with individuals ranging from a single filament to several filaments often <1 g. Bryopsis sp. was only common/visible during the spring survey.
Elysia viridis abundance
E. viridis abundance was assessed by haphazardly sampling an average of 10 samples of each potential algal host from the 2 sites (Yttre Vattenholmen and Saltö Lyngnholmen) in autumn 2010 (late October), spring 2011 (late May), summer 2011 (late July/early August), and autumn 2011 (early October). Algal individuals were carefully removed from the substrate and placed in separate press seal bags with seawater and transported back to the Sven Lovén Centre for Marine Sciences-Tjärnö to be searched for E. viridis. Each algal individual was thoroughly searched for E. viridis, the number of animals recorded, after which the alga was spun in a salad spinner, patted dry with absorbent paper in order to remove excess water, and weighed (±0.1 g). Abundance data were standardized to the number of E. viridis per g algal individual (no. g algal ind−1). Furthermore, we calculated the percentage of thalli that were colonised by E. viridis (% thalli occupied). This was calculated by dividing the number of thalli E. viridis were found on by the total number of thalli searched for each potential host species from each site in each survey and multiplying this value by 100.
Elysia viridis size distribution
Size of the collected E. viridis individuals was measured during all surveys to assess the variation between algal hosts within sites. All E. viridis individuals found in the autumn collections (2010 and 2011) were patted dry on absorbent paper and weighed (±1 mg). However, in the summer collection of 2011 there were extremely high numbers of E. viridis (>2500 individuals) that were mostly small (<1 mg) preventing accurate assessment of their mass. Instead, E. viridis lengths were measured (±0.083 mm) while slugs crawled across a petri dish, using a microscope equipped with an ocular micrometer. As measuring all individuals was not practical, a subset of animals was measured to provide a relative estimate of E. viridis size on each algal host. A maximum of 20 individual E. viridis were measured from each of the first 5 algal individuals searched of each algal host from each site. This amounted to 77–100 E. viridis individuals measured per algal host per site for C. sericea, C. rupestris and C. fragile.
As all potential algal hosts were not found at each site during each survey, we chose to analyse E. viridis abundance and size data from different sites and surveys in separate statistical analyses. Data from the spring survey were not statistically analysed because too few E. viridis were found (see Results). Data on the mean abundance (no. g algal ind.−1) of E. viridis on different algal hosts were statistically compared for each survey and site by 1-factor analysis of variance (ANOVA) with algal host as a fixed factor. Due to heterogeneous variances around the means, data on E. viridis size (mass, mg) at each site in the autumn surveys were statistically analysed with t-tests using separate variances . Data on the size (length, mm) of E. viridis on different algal hosts at each site in the summer survey were compared using 1-factor ANOVAs with algal host as a fixed factor. When less than 20 E. viridis were found on a host alga at a particular site and survey, data were not included in the analysis in order to reduce strong bias associated with highly uneven sample sizes . Data on algal host colonization (% thalli occupied by E. viridis) were not statistically analysed because only one value was obtained for each algal host species at each site and survey (i.e. n = 1).
Prior to all statistical analyses, data were tested for homogeneity of variances using Levene's test. All variances were heterogeneous apart from data on mean size of E. viridis from the summer survey, which were successfully transformed to log (x). However, since transformations did improve the homogeneity of variances, data on the size of E. viridis from the autumn surveys were also transformed to log (x) and the abundance data from all surveys were transformed to log (x+1). Post hoc comparisons of mean values were made using Dunnett's T3 multiple-comparisons test when variances were heterogeneous , or Tukey's honestly significant different (HSD) multiple-comparisons test when variances were homogeneous.
Elysia viridis abundance
E. viridis were found on several species of co-occurring algal hosts at both sites during 3 of the surveys (autumn 2010, summer 2011, and autumn 2011; Fig. 1 A, C, D). In spring 2011, however, only 2 individuals were found combined across all algal hosts and sites searched (Fig. 1B, data not included in statistical analyses). In general, the abundance of E. viridis was low on the green algal species Bryopsis sp. and C. melagonium, and on the dominant red alga C. virgatum. In contrast, E. viridis numbers were substantially higher on the green algal hosts C. sericea, C. rupestris and C. fragile (Fig. 1). However, the number of E. viridis found on different algal hosts varied substantially between sites and surveys.
Field abundance of Elysia viridis (E. viridis. g algae−1; mean + SE) on different algal hosts at 2 sites (Yttre Vattenholmen and Saltö Lyngnholmen) during 4 surveys: (A) autumn 2010, (B) spring 2011, (C) summer 2011, and (D) autumn 2011. Numbers below bars indicate the number of samples of each algal host searched. ND denotes no data for a particular host, i.e. no samples of the algal host were collected. Algal hosts within a site and survey sharing the same letter do not significantly differ at α = 0.05 (Dunnett's T3 multiple comparisons test).
In autumn 2010, there was a statistically significant difference in the abundance of E. viridis on different algal hosts at both Yttre Vattenholmen (ANOVA, F3,40 = 13.520, p<0.001; Fig. 1A) and Saltö Lyngnholmen (ANOVA, F4,49 = 12.077, p<0.001; Fig. 1A). When means were compared using Dunnett's T3 multiple-comparisons test, we found that E. viridis abundance was significantly (p<0.05) higher on C. rupestris than on C. fragile, C. virgatum, and C. melagonium at Yttre Vattenholmen (Fig. 1A). C. sericea and Bryopsis sp. were not found at this site during the survey. In contrast, the abundance of E. viridis was significantly (p<0.05) higher on C. sericea, and lower on C. fragile, C. virgatum and C. melagonium, compared to C. rupestris at Saltö Lyngnholmen (Fig. 1A). No specimens of Bryopsis sp. were found at this site during autumn 2010.
In summer 2011, we found a statistically significant difference in the abundance of E. viridis on different algal hosts at both Yttre Vattenholmen (ANOVA, F3,36 = 70.613, p<0.001; Fig. 1C) and Saltö Lyngnholmen (ANOVA, F3,36 = 48.568, p<0.001; Fig. 1C). Post hoc comparisons (Dunnett's T3 multiple-comparisons test), revealed that E. viridis abundance was significantly (p<0.05) higher on C. sericea than on C. rupestris, C. fragile and C. virgatum at Yttre Vattenholmen (Fig. 1C). Furthermore, at Saltö Lyngnholmen, E. viridis were significantly (p<0.05) more abundant on the 2 Cladophora spp. than on C. fragile and C. virgatum (Fig. 1C). C. melagonium and Bryopsis sp. were not collected/found at either site during the summer 2011 survey.
Very few slugs were found at Yttre Vattenholmen in autumn 2011 (a total of 45 individuals on C. rupestris and 2 on C. fragile), and therefore data on abundance of E. viridis between different algal hosts were not statistically compared (Fig. 1D). C. sericea and Bryopsis sp. were not present at this site during the survey. However, there was a statistically significant difference in the abundance of E. viridis between different algal hosts at Saltö Lyngnholmen (ANOVA, F2,30 = 22.741, p<0.001; Fig. 1D). Dunnett's T3 multiple-comparisons test showed that E. viridis were significantly (p<0.05) more abundant on C. rupestris than on C. fragile and C. virgatum (Fig. 1D). C. sericea, C. melagonium, and Bryopsis sp. were not found at this site during the survey.
Patterns of host colonisation (i.e. the percentage of algal thalli occupied by E. viridis) demonstrated consistently high use of the 2 Cladophora hosts by E. viridis in summer and both autumn surveys (between 65–100% thalli occupied; Fig. 2A, C, D) compared with C. fragile where colonisation was more variable (between 10–100% thalli occupied; Fig. 2A, C, D). Few algal thalli were colonised by E. viridis in spring 2011 with only 2 individuals found, one on Bryopsis sp. and the other on C. virgatum (Fig. 2B).
Thalli occupied by Elysia viridis (%) on different algal hosts at 2 sites (Yttre Vattenholmen and Saltö Lyngnholmen) during 4 surveys: (A) autumn 2010, (B) spring 2011, (C) summer 2011, and (D) autumn 2011. Numbers below bars indicate the number of samples of each algal host searched. ND denotes no data for a particular host, i.e. no samples of the algal host were collected.
Elysia viridis size distribution
In autumn 2010, there was a statistically significant difference in the size of E. viridis collected from different hosts at Saltö Lyngnholmen (t-test, t1,51 = 7.30, p<0.001; Fig. 3A). E. viridis collected from C. rupestris were on average 8.6 times larger than those from C. sericea. Size data for E. viridis collected from C. fragile were not statistically analysed due to the low number of animals, but their mean mass lay in between animals collected from the 2 Cladophora spp. (Fig. 3A). At Yttre Vattenholmen, C. sericea was not present in autumn 2010 and this species was therefore not included in the statistical analysis. However, there was a statistically significant difference in size between E. viridis individuals collected from C. rupestris compared to those collected from C. fragile (t-test, t1,173 = 3.16, p = 0.003; Fig. 3A) with E. viridis from C. rupestris on average 1.4 times larger than slugs collected from C. fragile (Fig. 3A). Only 2 E. viridis individuals were found in spring 2011, 1 on Bryopsis sp. and the other on C. virgatum (Fig. 3B).
Size of Elysia viridis (mg or mm, mean + SE) collected from different algal hosts at 2 sites (Yttre Vattenholmen and Saltö Lyngnholmen) during 4 surveys: (A) autumn 2010, (B) spring 2011, (C) summer 2011, and (D) autumn 2011. Numbers below bars denote the total number of E. viridis measured from each algal host. Data were included in the statistical analysis when at least 20 E. viridis were collected from a particular host, site and survey. Algal hosts within a site and survey sharing the same letter do not significantly differ at α = 0.05 (Tukey's honestly significant difference, HSD, multiple comparisons test).
In summer 2011, there was a statistically significant difference in size between E. viridis collected from different algal hosts at both Yttre Vattenholmen (ANOVA, F2,291 = 7.774, p = 0.001; Fig. 3C) and Saltö Lyngnholmen (ANOVA, F2,262 = 13.904, p<0.001; Fig. 3C). However, a large proportion of the E. viridis individuals were in the range of 1–4 mm, and the differences in size of E. viridis on different hosts were small (<1 mm; Fig. 3C). E. viridis collected from C. rupestris were longer compared to individuals collected from C. sericea and C. fragile at both sites (Fig. 3C). Size data for E. viridis on C. virgatum were not included in the analysis, as only 6 individuals in total were found on this algal host.
In autumn 2011, E. viridis individuals collected from C. rupestris were significantly larger than those collected from C. fragile at Saltö Lyngnholmen (t-test, t1,111 = −2.58, p = 0.011; Fig. 3D). A statistical comparison could not be made between E. viridis collected from different hosts at Yttre Vattenholmen as only 2 E. viridis were found on C. fragile in contrast to the 45 E. viridis found on C. rupestris. However, mean E. viridis size on C. rupestris and C. fragile hosts were similar at both sites (Fig. 3D).
Sacoglossans are a group of herbivorous marine molluscs that commonly have specialized adaptations to their algal hosts and therefore also tend to have restricted taxonomic host associations . However, some sacoglossans, e.g. several species in the genus Elysia, have less restricted host use, associating with/feeding on algae from several different families (e.g. , , , , ). We surveyed the abundance and size of E. viridis on 6 different potential host algae (the green algal species C. sericea, C. rupestris, C. fragile, C. melagonium, Bryopsis sp., and the red alga C. virgatum) at 2 sites on the Swedish west coast. To our knowledge, this is the first investigation that has rigorously tested the abundance and size distribution of E. viridis on several co-occurring macroalgal hosts through statistical comparisons of data acquired through structurally designed collections. In spring, slugs were almost absent from all algal hosts. In summer and autumn, E. viridis consistently occurred on several of the algal species at both sites. The highest number of small E. viridis were found on C. sericea, intermediate numbers of significantly larger E. viridis were found on C. rupestris, while fewer, intermediate sized animals were found on C. fragile. Throughout the study period, few E. viridis individuals were found on C. melagonium, Bryopsis sp., and C. virgatum.
Previous investigations of E. viridis abundance and size distribution on different algal hosts have been conducted in the British Isles (e.g. , , , ). These studies were primarily designed to study the interaction between E. viridis and the introduced species C. fragile spp. tomentosoides and did not sample several co-occurring potential algal host species simultaneously (but see  where the authors searched for E. viridis on C. fragile and C. rupestris simultaneously) nor did they report quantitative, statistically analysed abundance and size data of E. viridis on the other algal host species. However, E. viridis were found to occur on Codium spp., Cladophora spp., and Chaetomorpha sp. within a 1 km2 Irish lough, as well as on Codium spp., Chaetomorpha sp., and Bryopsis sp. at other sites in the British Isles , , indicating a broad taxonomic host use. Scottish populations of E. viridis, on the other hand, demonstrated a more selective pattern of host use in which sacoglossans only occurred on C. rupestris when C. fragile spp. tomentosoides was absent . In our study, few E. viridis were found on C. melagonium and Bryopsis sp., probably because the biomass of these species was low (C. melagonium often grew as single filaments and Bryopsis sp. was only present in spring when E. viridis was mostly absent). Furthermore, E. viridis colonised C. rupestris hosts at percentages that were either the same or higher compared to C. fragile, and E. viridis were larger and more abundant on C. rupestris, indicating that C. rupestris is a better host than C. fragile on the Swedish west coast.
Apart from a broad taxonomic host use, we found an annual variation in the abundance and size of E. viridis on the Swedish west coast. E. viridis were mostly absent in the spring survey, while high numbers of small slugs were found in summer. In autumn, E. viridis were larger but lower in number. In contrast,  consistently observed large E. viridis throughout the winter, spring, and summer in Scotland. Results from the present study indicate that E. viridis is an annual species in Sweden, and that the larvae that settle in spring or early summer may arise from external populations. E. viridis larvae have a large potential for dispersal (E. viridis larvae are planktonic for ca. 30 days, ), and it might be hypothesized that long distance transport, and hence increasing age prior to settlement, make the larvae “desperate” (i.e. settling on any host), thereby reducing host selectivity and resulting in a broader host-use pattern. However, E. viridis did not demonstrate a pattern suggesting desperate larvae evidenced by their consistently low abundance/absence on C. virgatum compared to the less common green algal hosts in summer and autumn surveys.
Host selection in natural populations of E. viridis may be controlled by physical and/or biological constraints (e.g. removal via wave motion, and/or predation), although their impacts on E. viridis distributions are not yet well understood . Increased intensity of wave exposure has been linked to changes in size distribution and decreased abundance of the sacoglossan Placida dendritica , , and algal host complexity has been suggested to increase the abiotic refuge value for amphipods . Physical disturbance through wave motion may explain the general decreases in E. viridis abundance observed from summer to autumn surveys in the present study, as strong autumn storms are common in the Koster Fjord and may facilitate variable removal of E. viridis from different hosts. Furthermore, although we are not aware of any published reports of predators that use E. viridis as prey, several potential predators (e.g. different species of crabs and fish) co-occur with E. viridis in the study area (G. B. Toth, pers. obs.). Therefore, predation could also be a potential, but not mutually exclusive, explanation for the seasonal decreases in E. viridis abundance and variation in size structure between hosts from summer to autumn. Algal host morphology and colour could contribute to the probability of E. viridis survival. Small E. viridis were often visible on the light green C. sericea whilst snorkelling (F. Baumgartner, pers. obs.) and although anecdotal, this evidence might suggest detection by visual predators could be substantial, explaining the lack of any large E. viridis on this alga. Furthermore, the morphology of C. fragile and C. melagonium likely provides little refuge through crypsis when E. viridis are large.
The broad taxonomic host-use pattern observed in our study area and in the British Isles could be due to ontogenetic or adult host switching by E. viridis, or to differential host exploitation by different E. viridis individuals within populations. E. viridis demonstrates variation in feeding behaviour at several ontogenetic stages (e.g. , , ), but initial genetic studies have revealed no apparent differences between E. viridis associated with different algal hosts . Together, these results indicate that E. viridis is capable of host switching in some areas. Host selection, intra- and inter- specific host specialization, and host switching are considered important mechanisms for sacoglossan speciation . For example, speciation within the genus Elysia has been linked to algal host switches with sympatric congeners commonly exploiting different algal diets . Furthermore intergeneric and congeneric algal host use by sympatric conspecific sacoglossans has been observed (: Aplysiopsis enteromorphae, : Placida dendritica, : Elysia clarki, : E. viridis) and, although considered rare in this clade, intraspecific variation and/or specialization in host use are hypothesized to be important sources of polymorphism and processes for speciation in many other groups of animals .
In conclusion, host selection in small, sedentary herbivores like E. viridis, which use algal individuals both as food and shelter, most likely depends on a combination of several intrinsic algal and herbivore traits, as well as extrinsic physical and biological constraints. From our data it is clear that E. viridis colonise co-occurring congeneric and intergeneric green algal hosts on the west coast of Sweden with C. rupestris seemingly providing the most stable host. This quantitative observational data will be vital in providing a framework to which mechanistic studies on intraspecific variation in host use can be related. Furthermore, it will provide a stronger ecological basis to address questions regarding both physical and biological constraints on E. viridis abundance and size distributions. The fact that so few studies have focussed on the distribution of E. viridis in natural field populations is surprising considering that several aspects of E. viridis biology and behaviour have been well-studied in the laboratory (e.g. , –). To be able to draw ecologically relevant conclusions about the natural history, biology and/or ecology of a species, results from laboratory experiments must be compared with observational field data.
The authors thank S. Enge for her assistance in the field and J. Pavia for his assistance in the lab.
Conceived and designed the experiments: FAB GBT. Performed the experiments: FAB. Analyzed the data: FAB GBT. Wrote the paper: FAB GBT.
- 1. Krug P (2011) Patterns of speciation in marine gastropods: a review of phylogenetic evidence for localized radiations in the sea. Am Malacol Bull 29: 169–186.
- 2. Jensen KR (1997) Evolution of the Sacoglossa (Mollusca, Opisthobranchia) and the ecological associations with their food plants. Evol Ecol 11: 301–335.
- 3. Marín A, Ros JD (2004) Chemical defenses in sacoglossan opisthobranchs: taxonomic trends and evolutive implications. Sci Mar 68: 227–241.
- 4. Pierce SK, Curtis NE (2012) Cell biology of the chloroplast symbiosis in sacoglossan sea slugs. Int Rev Cell Mol Biol 293: 123–148.
- 5. Clark KB, DeFreese D (1987) Population ecology of Caribbean Ascoglossa (Mollusca, Opisthobranchia) a study of specialized algal herbivores. Am Malacol Bull 5: 259–280.
- 6. Williams SI, Walker DI (1999) Mesoherbivore-macroalgal interactions: feeding ecology of sacoglossan sea slugs (Mollusca, Opisthobranchia) and their effects on their food algae. Oceanogr Mar Biol Annu Rev 37: 87–128.
- 7. Marín A, Ros JD (1992) Dynamics of a peculiar plant-herbivore relationship: the photosynthetic ascoglossan Elysia timida and the chlorophycean Acetabularia acetabulum. Mar Biol 112: 677–682.
- 8. Trowbridge CD (1993) Feeding ecology of the ascoglossan opisthobranch Aplysiopsis enteromorphae (Cockerell & Eliot): patterns of distribution and impact on tidepool-dwelling green algae. J Exp Mar Biol Ecol 169: 233–257.
- 9. Underwood AJ, Chapman MG, Connell SD (2000) Observations in ecology: you can't make progress on processes without understanding the patterns. J Exp Mar Biol Ecol 250: 97–115.
- 10. Händeler K, Wägele H (2007) Preliminary study on molecular phylogeny of Sacoglossa and a compilation of their food organisms. Bonn Zool Beit 55: 231–254.
- 11. Curtis NE, Massey SE, Pierce SK (2006) The symbiotic chloroplasts in the sacoglossan Elysia clarki are from several algal species. Invertebr Biol 125: 336–345.
- 12. Curtis NE, Pierce SK, Massey SE, Schwartz JA, Maugel TK (2007) Newly metamorphosed Elysia clarki juveniles feed on and sequester chloroplasts from algal species different from those utilized by adult slugs. Mar Biol 150: 797–806.
- 13. Händeler K, Wägele H, Wahrmund U, Rüdinger M, Knoop V (2010) Slugs' last meals: molecular identification of sequestered chloroplasts from different algal origins in Sacoglossa (Opisthobranchia, Gastropoda). Mol Ecol Resour 10: 968–978.
- 14. Maeda T, Hirose E, Chikaraishi Y, Kawato M, Takishita K, et al. (2012) Algivore or phototroph? Plakobranchus ocellatus (Gastropoda) continuously acquires kleptoplasts and nutrition from multiple algal species in nature. PLoS ONE 7: e42042
- 15. Christa G, Wescott L, Schäberle TF, König GM, Wägele H (2013) What remains after 2 months of starvation? Analysis of sequestered algae in a photosynthetic slug, Plakobranchus ocellatus (Sacoglossa, Opisthobranchia), by barcoding. Planta 237: 559–572.
- 16. Brandley BK (1984) Aspects of the ecology and physiology of Elysia cf. furvacauda (Mollusca: Sacoglossa). Bull Mar Sci 34: 207–219.
- 17. Trowbridge CD (1991) Diet specialization limits herbivorous sea slug's capacity to switch among food species. Ecology 72: 1880–1888.
- 18. Trowbridge CD (2004) Emerging associations on marine rocky shores: specialist herbivores on introduced macroalgae. J Anim Ecol 73: 294–308.
- 19. Baumgartner FA, Motti CA, de Nys R, Paul NA (2009) Feeding preferences and host associations of specialist marine herbivores align with quantitative variation in seaweed secondary metabolites. Mar Ecol Prog Ser 396: 1–12.
- 20. Middlebrooks ML (2012) Consequences of kleptoplasty on the distribution, ecology, and behavior of the sacoglossan sea slug, Elysia clarki. PhD dissertation, University of South Florida, Tampa, FL. Available: http://scholarcommons.usf.edu/etd/4162. Accessed 2014 Feb 25.
- 21. Trowbridge CD, Todd CD (2001) Host-plant change in marine specialist herbivores: ascoglossan sea slugs on introduced macroalgae. Ecol Monogr 71: 219–243.
- 22. Trowbridge CD, Little C, Stirling P, Farnham WF (2008) Sacoglossan gastropods on native and introduced hosts in Lough Hyne, Ireland: larval retention and population asynchrony? J Mar Biol Assoc UK 88: 277–286.
- 23. Trowbridge CD (2000) The missing links: larval and post-larval development of the ascoglossan opisthobranch Elysia viridis. J Mar Biol Assoc UK 80: 1087–1094.
- 24. Hinde R, Smith DC (1975) The role of photosynthesis in the nutrition of the mollusc Elysia viridis. Biol J Linnean Soc 7: 161–171.
- 25. Vieira S, Calado R, Coelho H, Serôdio J (2009) Effects of light exposure on the retention of kleptoplastic photosynthetic activity in the sacoglossan mollusc Elysia viridis. Mar Biol 156: 1007–1020.
- 26. Evertsen J, Johnsen G (2009) In vivo and in vitro differences in chloroplast functionality in the two north Atlantic sacoglossans (Gastropoda, Opisthobranchia) Placida dendritica and Elysia viridis. Mar Biol 156: 847–859.
- 27. Thomasson MA, Tunberg BG (2005) Composition and vertical distribution of the motile epifauna on a vertical rock wall in Gullmarsfjorden, western Sweden, using an alternative sampling approach. Mar Biol Res 1: 107–117.
- 28. Nohrén E, Pihl L, Wennhage H (2009) Spatial patterns in community structure of motile epibenthic fauna in coastal habitats along the Skagerrak - Baltic salinity gradient. Estuar Coast Shelf Sci 84: 1–10.
- 29. Johannesson K (1989) The bare zone of the Swedish rocky shores: why is it there? Oikos 54: 77–86.
- 30. Quinn GP, Keough MJ (2002) Experimental design and data analysis for biologists. Cambrige: Cambridge University Press. 537 p.
- 31. Klochkova TA, Han JW, Chah K-H, Kim RW, Kim J-H, et al. (2012) Morphology, molecular phylogeny and photosynthetic activity of the sacoglossan mollusc, Elysia nigrocapitata, from Korea. Mar Biol 160: 155–168.
- 32. Trowbridge CD (2002) Local elimination of Codium fragile ssp. tomentosoides: indirect evidence of sacoglossan herbivory? J Mar Biol Assoc UK 82: 1029–1030.
- 33. Sotka EE (2005) Local adaptation in host use among marine invertebrates. Ecol Lett 8: 448–459.
- 34. Trowbridge CD (1992) Mesoherbivory: the ascoglossan sea slug Placida dendritica may contribute to the restricted distribution of its algal host. Mar Ecol Prog Ser 83: 207–220.
- 35. Trowbridge CD (1992) Phenology and demography of a marine specialist herbivore: Placida dendritica (Gastropoda: Opisthobranchia) on the central coast of Oregon. Mar Biol 114: 443–452.
- 36. Sotka EE (2007) Restricted host use by the herbivorous amphipod Peramphithoe tea is motivated by food quality and abiotic refuge. Mar Biol 151: 1831–1838.
- 37. Jensen KR (1989) Learning as a factor in diet selection by Elysia viridis (Montagu) (Opisthobranchia). J Mollusc Stud 55: 79–88.
- 38. Bolnick DI, Svanbäck R, Fordyce JA, Yang LH, Davis JM, et al. (2003) The ecology of individuals: incidence and implications of individual specialization. Am Nat 161: 1–28.