Our results for the infection of Japanese eels from SW Taiwan are in good agreement with previous findings of Münderle et al.  on eels from the same area: the only swimbladder nematode found was Anguillicoloides crassus. Anguillicola globiceps was not found in hosts from this area. Furthermore infection parameters with adult A. crassus for both wild eels and eels from aquaculture are in agreement with these previous findings. In contrast the larval prevalence as well as the larval mean abundance and larval mean intensity reported here for wild eels are surprisingly low compared to the Münderle et al. data, a fact that could possibly accord to seasonal fluctuations in the prevalence of A. crassus (also noticed by Münderle et al.).
We were able to show, that the infective pressure of A. crassus was extremely high in eels from aquaculture. This fact was indicated by slightly elevated levels of infections with adults but more importantly by strongly elevated numbers of larvae found in the swimbladder wall of cultured eels compared to wild hosts. Characteristically for macro-parasite infections , the frequency distribution of adult and larval A. crassus approaches negative binomial distribution with high degree of over-dispersion both in the wild and in cultured eels.
Regarding the higher infection parameters in aquaculture some inferences can be drawn to general aspects of host-parasite interaction in Japanese eels. Obviously, there is a sufficient number of intermediate hosts available for the larval development in aquaculture. The higher infective pressure therefore seems to be a by-product of the elevated host density. It would also seem possible, that moderately elevated levels of infection with adults increase the infective pressure on hosts, because of a relationship between adult infection and L2 larvae produced in the next generation. This second hypothesis is not supported by our data, showing a comparable percentage of hosts in which L2 larvae are present and therefore reproduction of parasites is successful in wild and cultured eels at a comparable level. This indicates a mechanism preventing A. crassus from fast and elevated reproduction in Japanese eels.
A high mortality of A. crassus larvae has been reported before in the swimbladder wall of Japanese eels . We were able to show for the first time systematically, that this mortality is even more pronounced in the intestinal wall. This suggests a regulation of A. crassus populations by the immune-system of their natural host in east Asia (concomitant immunity), while A. crassus populations seem to be much more regulated by intraspecific processes in European eels [15, 18]. This raises the question of how immunocompetent Japanese eels respond to infection under high infective pressure. Our data on encapsulated larvae allows some hypotheses to be formulated on the mechanisms and succession of destruction of larvae. We could clearly show that the establishment of encapsulated larvae inside the intestinal wall is related to killing of larvae in the swimbladder wall: significant numbers of encapsulated larvae in the intestinal wall were not observed when capsules in the swimbladder-wall were absent. Furthermore, no capsules in the intestinal wall could be found in single, non-repeated experimental infections of Japanese eels, while larvae are killed in the swimbladder wall. This observation shows that larvae are first encapsulated in the swimbladder wall and encapsulation inside the intestinal wall follows only repeated heavy infections. These features suggest a major role of acquired or infection induced immunity in the formation of capsules.
In the antibody response against A. crassus, differences in European and Japanese eels have been found. There is an earlier onset of antibody production in Japanese eels, possibly directed against larvae [19, 20]. Induction of resistance was shown in immunization experiments using irradiated L3 larvae in the Japanese eel but not in the European congeners .
Regarding the cellular immune response against A. crassus, epithelioid macrophages and granulocytes have been found aggregating around larvae of European eels and forming parasitic nodules in the walls of swimbladder and intestine . In these studies capsules in the swimbladder wall were more common than in the intestinal wall. For European eels conclusions from this observations, especially whether these cells are initiating the formation of connective tissue capsules  or just absorbing damaged tissue  remain debatable. Encapsulated "objects" have been noted from the intestinal wall of Japanese eels in studies comparing the immune response of Japanese and European eels [14, 24].
It is advised to take into account the stomach- and intestinal wall as a second location of direct host-parasite interaction in infection experiments. In experimental studies, where L3 larvae are expelled from the copepod intermediate hosts and eels are infected through a stomach tube, capsules should be expected in the stomach wall. This seems to be an artificial situation, as the wild larvae ingested inside their copepod intermediate host are found in the intestinal wall, especially in the more posterior region. The high numbers of A. crassus L4 larvae and adults alive in the swimbladder of the Japanese eels experimentally double-infected with high doses of larvae demand further investigation. Possibly it is a result of the concentrated application during infection, compared to a situation in the wild or in aquaculture, where although under high infection pressure the host is gradually exposed to the larvae.
Studies on free living hosts should consider the intestinal wall as a location where dead larvae can be found. Our PCR based approach allows an identification of degraded larvae inside the swimbladder wall. This is desirable as in European eels the absence of swimbladder-capsules in the wild is assumed (mainly because these capsules are absent in experimental infections of European eels) leaving some sparsely found "objects" in the swimbladder-wall unidentified .
In the present study we established that it is possible to use a molecular approach to identify A. crassus, even if encapsulated and degraded within intestinal tissue of its eel host. We suggest a simple PCR based method providing a reliable alternative to morphological identification of A. crassus. We show that if swimbladder nematodes inside capsules containing degraded tissue are investigated, a PCR approach with subsequent analysis on agarose gels yields sufficient diagnostics, as the PCR product of the nematode LSU-gene can be distinguished from the chimeric amplification product of the the eel-host's LSU on an agarose gel. Subsequent sequencing of the amplified marker gene furthermore allows reliable identification of the exact parasite species, even in the absences of adult worms. To guarantee correct amplification of the nematode LSU, the collagen sheath of the capsule must be disrupted in order to allocate the nematode DNA. The inability to amplify the LSU fragment from some of the smallest capsules, may result from the already advanced degradation of the nematode DNA making it impossible to amplify long target fragments. This represents a obvious limitation of our method when applied to the most disintegrated specimen. A further development of comparable approaches using shorter target fragments for amplification can therefore be advised for authors specially interested in these issues. Nevertheless our method provides a means for identification of encapsulated objects superior to visual inspection.