■ a a Parasites of the Parasites. 



Sacculina, 50 bore also either one or two specimens of D. curvata, fixed either on the Saccu- 

 lina or on the abdomen of the crab near the point of attachment of the Sacculina. Several 

 crabs have been found with small D. curvata on them which showed no external trace of a 

 Sacculina. but all these crabs on dissection proved to be infected with Sacculina interna: it 

 appears therefore that the larvae of the Liriopsid have some means of finding out crabs that 

 are infected with Sacculina before the latter appear in the external world. It may be that a 

 sense of smell directs them in these cases. 



I may mention here a species of Danalia, which I have named D. ypsilon (see fig. 32 

 Plate 7) and which is found upon Galathea dispersa. This species of Galathea is also fre- 

 quently infected by a Rhizocephalous parasite, Lernaeodiscus galatheae (see fig. 31 Plate 7) 

 but curiously enough the adult forms of Danalia and Lernaeodiscus have never been found by 

 me upon the same individual host, nor have I been able to prove by the presence of roots 

 that the specimens infected with Danalia ypsilon have once carried a Lernaeodiscus or are in- 

 fected with a still internal Lernaeodiscus. It is highly probable therefore that in this case 

 the Danalia has secondarily become a true parasite of the Galathea, instead of the Lernaeodiscus. 



Besides the species mentioned in Bonnier's work, Hansen (6 has described the Crypto- 

 niscid larvae of a number of forms from the Plankton Expedition of 1889. and in his work 

 can be found the most accurate description of the external anatomy of these larvae in exis- 

 tence. The author adopts a different classification to that of Bonnier; in his subfamily 

 Cryptoniscinae he includes all the parasites of the Cirripedia as well as those of the Ostracoda 

 Bonnier's Cyproniscidae), and of the Isopoda and Amphipoda (Bonnier's Cabiropsidae). That 

 this classification may prove correct is highly probable, but since we are only concerned here 

 with the parasites of the Bhizocephala, it is more convenient to adopt the more exclusive 

 classification of Bonnier. 



In comparing the larvae of Cryptoniscinae figured by Hansen (6) with the larvae of 

 Danalia curvata, it will be seen that our larvae (Plate 1 fig. 23) in the characters of the head 

 and of the basal portion of the first antenna correspond most closely to Hansen's Cryptoniscus B, 

 but in our larvae the terminal joint of the sixth pair of thoracic legs is much elongated and 

 not that of the seventh, as in Hansen's type. 



The chief subject of interest with regard to these parasites, besides their association 

 with the Pthizocephala, is their method of reproduction, because, being solitary fixed parasites, 

 some special method of cross-fertilization might be expected. The following paragraphs will 

 elucidate this problem, which has been erroneously conceived hitherto. 



Fraisse (1) has given the only detailed account of the life-history of the Liriopsidae, 

 and although his work is in many respects of value, he made the misleading error of 

 describing the sexual gland of the Cryptoniscus larva as a diverticulum of the gut, while the 

 spermatheca of the adult is considered as a gill. Furthermore an altogether illusory testis is 

 described in the so-called larval males, which does not exist. 



Fraisse conceived that the Cryptoniscus larvae are already at the free-swimming stage 



