A comparison of Elysia (=Tridachia) crispata (Mörch, 1863) from the Virgin Islands with elysiid slugs from the mangrove swamps and canals in the Florida Keys that have previously been identified as E. crispata reveals many differences in habitat, gross and microscopic anatomy, food preferences of juveniles, sources of symbiotic chloroplasts and their localization within the digestive tubules, radular morphology, and nucleic acid sequences of two genes. The differences between the two groups of slugs are such that the Florida Keys animals are considered to represent a new species, Elysia clarki. Elysia clarki feeds on siphonaceous algae, and intracellularly sequesters the chloroplasts, which actively photosynthesize for up to 4 months. We have determined the algal source of the chloroplasts in adult E. clarki from 2 populations in the Florida Keys, using molecular techniques, feeding experiments, and electron microscopy. Our results clearly demonstrate that adult E. clark i sequester chloroplasts from 7 different species of algae, representing two genera, of which 5 were identified; Penicillus lamourouxii, P. capitatus, Halimeda incrassata, H. monile, and Bryopsis pennata. In addition, chloroplasts from more than 1 species of algae are sequestered in the same digestive cell simultaneously. Phylogenetic analysis of rbcL sequences from the order Bryopsidales showed that E. clarki feeding was restricted to calcareous members of the family Udoteaceae and the family Bryopsidaceae. Feeding experiments were conducted, using individuals raised in the laboratory from egg masses laid by E. clarki adults which had been collected from Grassy Key, Florida, USA, and 29 species of macroalgae. For the first 14 d post-metamorphosis, juveniles ate only the thin filamentous coenocytes, Bryopsis plumosa or Derbesia tenuissima. Electron microscopy showed that the chloroplasts from both algae were sequestered intracellularly in juvenile slugs. Individuals offered any other macroalga, including the four calcareous species fed on by adults, did not feed on or incorporate any chloroplasts, and soon died. Juveniles switched from B. plumosa to P. capitatus at a length of ~ 1.0 cm, and fixed for microscopy 14 days later had intact intracellular chloroplasts from both algae.