According to the phylogeny presented in Fig. 3, several taxa are united in a monophylum. Systematic relationships of some groups have been discussed for a long time (Acochlidiacea, Rhodopidae), others have been considered to belong to the Cephalaspidea (Philinoglossidae, Runcinidae). Their sister-taxon relationship is not solved yet and the presented cladogram is debatable. Nevertheless they have in common some evolutionary traits, e.g. the complete loss of the shell, their small size compared to other opisthobranchs and their food preference for diatoms and detritus. According to histological results, no particular defensive glands could be detected and defensive strategies probably lay in habits. Many of them burrow in sand and/or are cryptic in colour.

A shell is generally considered to be a protection against predators, such as fish, crabs and other vagile organisms. "If the shell of a whelk (e.g. Buccinum, a "prosobranchiate" caenogastropod, annotation of the authors) is broken away and the soft animal is then offered to a hungry cod, it is eaten readily." (p: 115) 26. Reduction, internalisation or loss of the shell within Opisthobranchia implies other defensive strategies. Shell reduction within molluscs is uncommon, and occurs mainly in the highly mobile cephalopods. In gastropods, shell loss is rare in paraphyletic prosobranchs, and known only from few groups of Pulmonata, e.g., the Gymnomorpha and the stylommatophoran groups Arionidae and Limacidae. However, shell reduction occurred many times within the different subgroups of Opisthobranchia. Here, an internalization or complete loss occurs within the Cephalaspidea s.str, Anaspidea, Sacoglossa, Acochlidiacea and Pleurobranchoidea (Fig. 3). Whereas complete loss of the shell is not known from any member of the small taxon Tylodinoidea (about 15 species), this character state occurs in the stemline of the Nudibranchia and Gymnosomata. When estimating species numbers with no shell or a rather tiny internal shell and comparing this to the number of species with a larger external shell, the former outnumber the latter by far.

Loss of the shell therefore can be assumed to have advantages compared to the presence of a protective but heavy shell. Advantages probably lay in the exploration of new habitats, which are more difficult to reach when being protected by a shell. This can be observed e.g. in a subgroup of the Cladobranchia. The Aeolidoidea are able to graze on fragile hydrozoans (Fig. 8E). This kind of prey is used by few other invertebrates, e.g., Solenogastres, members of the Pycnogonida and of the Amphipoda 272829. Burrowing forms with a shell, e.g., Scaphander or Acteon Montfort, 1810, have an elaborate cephalic shield that partially covers the shell and renders them streamlined. Loss of the shell probably enables slugs to search for food in sandy or muddy habitats more easily. This is the case for members of the Cephalaspidea s.str. and Acochlidiacea.

Basal members of the Sacoglossa have retained a shell, but more derived ones have lost it. Shell loss allowed evolution of a phenomenon that is unique in the animal kingdom. Sacoglossa in general feed on algae by piercing the cells with their tooth and sucking out the contents of the cell. The cytoplasm is digested, but in many species (e.g. Elysia timida (Risso, 1818), Placobranchus ocellatus) the chloroplasts are stored in distinct branches of the digestive gland. Here they are stored for a period of several days to months 30. For this phenomenon, the term "cleptoplasty" is used by several authors 25. The functioning chloroplasts continue with photosynthesis within the slug and provide nutritional metabolites for the metabolism of the gastropod 3133132. Penetration of light into the slug would be hindered by the possession of a shell. A similar system is observed in members of the Nudibranchia, e.g., in Phyllodesmium jakobsenae Burghardt & Wägele, 2004, or Melibe bucephala Bergh, 1902) 3334. Here unicellular algae (zooxanthellae) from the coral food or from the free water column are stored in the digestive system and metabolites of these zooxanthellae are used for the slug's own purposes 353637. According to published phylogenies and to our own results (unpublished data of both authors) on Sacoglossa and Nudibranchia, it can be assumed that uptake of chloroplasts or zooxanthellae first enhanced crypsis (Fig. 1D) 25. The short-term storage allows a continuation of the photosynthetic activity of chloroplasts within the slug. Storage over a longer period allowed the reduction of food uptake with the possibilities to search for new and/or less frequent prey organisms 238. The most effective symbiotic relationships are known for the sacoglossan Elysia chlorotica Gould, 1870, which can survive eight months without food 3, the aeolid Pteraeolidia ianthina (Angas, 1864) and the dendronotoidean Melibe bucephala, both of which survived in our aquaria for 10 months without food (Burghardt & Wägele unpublished data).

Loss of a shell as a protective structure led to an array of different defensive structures. Some of these traits can be observed as a combination in one and the same species.

Crypsis can be observed in many groups and is very often achieved by incorporation of the same dyes from the food (Fig. 8C, Phyllidia flava Aradas, 1847). Cryptic appearance also is achieved by mimicking the same patterns or even outline of the substrate. Corambe pacifica MacFarland & O'Donoghue, 1929 perfectly mimics the colour patterns of its prey, the bryozoan Membranipora de Blainville, 1830. Phyllodesmium jakobsenae mimics the feathered polyps of the soft coral Xenia Lamarck, 1816, on which it lives 33, whereas the cerata of P. briareum (Bergh, 1896) are smooth like the tentacles of its prey, the soft coral Briareum Blainville, 1830 (Fig. 8D). Zebra effects are achieved by patterns with blotches like that in Peltodoris atromaculata Bergh, 1880 (Fig. 8A) or by stripes. Looking like unpalatable sponges (Fig. 8B, Jorunna tomentosa (Cuvier, 1804)) is very common in spicule-bearing dorids. According to Gosliner 39, the cryptic species are rather basal taxa, whereas the taxa with aposematic colour patterns are more derived – a hypothesis that has yet to be proven by thorough phylogenetic analyses that include all species of the subgroup in question.

A unique defensive strategy within animals is the storage of cnidocysts ("cleptocnides"), which is typical for nearly all members of the cladobranch Aeolidoidea 4940. This group mainly feeds on cnidarians, with priority on Hydrozoa. The mechanisms of the uptake of cnidocysts, so that explosion is not triggered during consumption, are still not understood. It is assumed that the slug exudes a mucus to hinder explosion 926. Investigated aeolids, like Aeolidia papillosa (Linné, 1761), have a highly glandular oral tube (Fig. 6C) that supports this hypothesis. Another theory implies that there occurs a kind of acclimation process, similar to that discussed between sea anemones and anemone fish 41. According to the investigations of Greenwood and Mariscal 42 only immature cnidocysts are stored in the cnidosac, whereas mature ones are digested. But, histological investigation of many aeolids directly collected from their food have not revealed high numbers of exploded cnidocysts in the stomach (unpublished data of HW). Only Notaeolidia schmekelae Wägele, 1990 from the Antarctic Ocean has been observed to have many exploded cnidocysts in its digestive tract 43.

Presence of spicules in the notum as a defensive strategy was discussed by several authors 1044. Spicules are present in many shell-less Anthobranchia and Acochlidiacea, but also in members of the Pleurobranchoidea, which sometimes have an internalised small shell. Spicules never occur in opisthobranchs with a larger shell. Cattaneo-Vietti et al. investigated the mineral composition of dorid spicules and found calcite (CaCO₃) and brucite (Mg(OH)₂) 45. Smaller spherules are composed only of calcite. Harris described feeding experiments offering various opisthobranchs to specimens of Navanax Pilsbry, 1895 (Cephalaspidea), who is a ferocious predator on opisthobranchs 10. This species rejected all spiculose dorids.

Another evolutionary trait for defence, and discussed as a prerequisite for shell reduction at least in sacoglossans 13, is the uptake or de novo synthesis of secondary metabolites that are toxic to possible predators 546. Uptake by feeding on toxic prey (mainly algae, Porifera, Bryozoa, Tunicata and Cnidaria) is the major source of compounds, whereas de novo synthesis is known only from few taxa 5. When dietary derived, Avila called these cleptochemicals, following the terms cleptoplasts and cleptocnides for incorporation and use of chloroplasts in Sacoglossa and cnidocysts in Aeolidoidea 5. Literature on chemical compounds in opisthobranchs is numerous. Some reviews summarize our knowledge 5746474849. Compounds mainly belong to the terpenoids, especially the insoluble sesquiterpenoids and diterpenoids. Little is known about the function of the biological compounds, although their defensive tasks are very often postulated 56746. Few feeding experiments have been performed in the past, demonstrating a toxic effect on crustaceans and/or fish 2650. Also the translocation from prey into the slug, and the transformation by changing the chemical structures either by degradation through digestion, or by an active mechanism into a more effective chemical, is hardly understood 57. Location of the compounds is investigated only for few species, by analysing certain parts of the body 51, or even by isolating larger organs, like the MDFs 52. Tracing the compounds within the tissue, or even cells, using immunohistochemical methods has never been done. Therefore, it is not possible to correlate chemical bioactivity with certain histological structures, except for the mantle dermal formations in the species Hypselodoris webbi (Chromodorididae) 52.

Inorganic compounds, like sulphuric acid are produced in few groups. Their function and location is better known due to the extensive work of Thompson 535455. He analysed the production of sulphuric acid in different members of Gastropoda, including members of the Pleurobranchoidea, Cephalaspidea and Dorididae. The exudated acid contains inorganic chloride and sulphate anions, and traces of organic substances. He was able to localize the acid by histochemistry within the large vacuoles in the median buccal gland (Fig. 7C) and the subepithelial glands of Pleurobranchoidea (Fig. 7D). There, the acid is held in active form 53. Gillete et al. investigated the role of the central nervous system and peripheral nerves for exudation and showed positive feed back 56.

Broad histological investigations of the Opisthobranchia show that many species are characterized by a large array of glandular structures 811575859. These comprise single glandular cells lying in the outer epithelia, or subepithelially. Glandular follicles composed of several cells usually lie subepidermally and open via a duct to the outside. Larger organs are the MDFs, or the glandular tubules of the median buccal gland in the Pleurobranchoidea. Some of these structures have been known for a long time and their defensive tasks were discussed in more detail by Hoffmann 8. Well known are the ink gland (Blochmann's glands) and the opaline gland (Bohadsch gland) in the Anaspidea. Both glands exude substances that have been shown to be toxic to cnidarians 1. Probably these substances also caused severe damage of the liver of a 40-year-old man, who ate Aplysia kurodai Baba, 1937 60. By experimental studies it was shown that the repellent substance in the ink gland is a monomethyl ester of phycoerythrobilin and is derived from phycoerythrin from the consumed red algae 61. The role of the opaline gland is less known. According to Carté, the prosteroglandine with the highest known biological activity is Dolastatin 10, a natural product extracted from the anaspidean Dolabella auricularia (Lightfoot, 1786) 62. This large species of more than 10 cm lives on the intertidal flats in the tropical Indo-Pacific, where it would represent an ideal food for birds and fishes, if not for that highly toxic chemical. This substance is already applied in medical treatments (see http://www.clinicaltrials.gov), and seems to be one of the most potent anticancer agents.

Information on other glandular structures are rare, and nearly nothing is known about their contents and their functions. At the moment we are not able to trace the different substances in these glands to find out whether there are any constraints concerning structure (and therefore function) and the stored chemicals.

Only few hypotheses are formulated concerning acquisition of toxicity and loss of the shell. Faulkner & Ghiselin assumed that chemical defence based on metabolites derived from food preceded the reduction of the shell and that chemical defence has been a driving force behind the evolution of Opisthobranchia 134663. Cimino et al., by analysing the different compounds and their origin, came to the conclusion that evolution within the Sacoglossa started with the uptake and storage of sesquiterpenoids from algae in species still having a shell 7. Within the shell-less members of the family Elysiidae, diterpenoids from the algae were stored, whereas in highly evolved forms, like Elysia timida, the slugs switched to a de novo synthesis of polyproprionates. Cimino & Ghiselin also mentioned that handling and utilization of a particular kind of defensive metabolite allowed the switch to food with similar compounds quite easily, and therefore has driven adaptive radiation 46. As an example they named the dorids and in particular the family Chromodorididae, which show a large array of usage of biochemicals from different sponges. Again, the bio-synthesis of compounds, as observed in Dendrodoris Ehrenberg, 1831, is considered to be the most derived form of defence within the Anthobranchia.

Information on defensive strategies, as listed above, is available now for several groups of the Opisthobranchia. More and more reliable phylogenies are becoming available, which allow the identification of well-supported branches and stemlines. Combining this knowledge, it becomes evident that several defensive systems evolved before the loss of the shell (several glandular structures, e. g., the hypobranchial gland, mantle rim glands, Bohadsch gland). Here we would like to extend the hypothesis of Cimino et al. by addressing the problem of excretion 7. It can not be ruled out that certain glandular structures evolved as a kind of excretory system to get rid of ingested toxic substances. Therefore it is not storage in special organs that preceded the use of toxic substances, but the necessity to expel them. By analysing phylogeny, it is evident that many defensive structures evolved after the internalisation or loss of shells (e.g. acid glands in the notum, cleptocnides, MDFs).

But we still have to identify the location of the compounds for a better understanding of the evolutionary history concerning the acquisition of toxicity, which certainly was a driving force in the evolution of these fascinating opisthobranchs. New techniques, e.g., the oligonucleotide aptameres, could help to solve this question 64. We also have to keep in mind that chemical substances might not only play a role in defence (allomones), but also in reproduction and development (pheromones).

About 300 different species of Opisthobranchia have been collected and investigated by the authors in the past 20 years. Collecting was performed from the intertidal (e.g. Australia, Helgoland), to the sublitoral zones (e.g. Mediterranean Sea, North Atlantic, tropical waters in the Red Sea, Australia and others) down to depth of 1000 m (Antarctica). Collecting techniques comprised hand collecting in the intertidal and while diving, or using trawls, like the Agassiz trawl in Polar Seas. Specimens were preserved in 4 – 10 % formaldehyde/seawater for histological investigations, or 96% ethanol for molecular investigations.

Investigation of morphology and anatomy was performed by macroscopical and histological techniques. For histological investigations, entire animals (when small in size) or parts of the animals were embedded in hydroxyethylmethacrylate for serial sectioning (2 μm). Sections were stained with toluidine blue and investigated by light-microscopy. Toluidine blue stains acid mucopolysaccharides in various shades of red to violet, whereas neutral mucopolysaccharides are staining in blue colours. Observation of living animals aided the understanding of external characters and life strategies. Data used for the phylogeny comprise 79 taxa and 110 characters based on morphology and histology. Polarity of characters was obtained by outgroup comparison with Littorina littorea. Due to some trivial characters, an all-zero outgroup was chosen. The characters are explained in full detail by Wägele & Klussmann-Kolb (in prep). Analyses were performed by PAUP 4.0 beta 3a (Swofford, 1999) 65. Parameters of maximum parsimony analyses were: ACCTRAN, all characters unordered and unweighted; heuristic search options: stepwise addition = random, branch-swapping option = TBR.