The food of the Scyphozoa varies a good deal. _Charybdea_ and _Periphylla_, and probably many others with large mouths, will capture and ingest relatively large fish and Crustacea; but _Chrysaora isosceles_[342] apparently makes no attempt to capture either Copepoda or small fish, but preys voraciously upon Anthomedusae, Leptomedusae, Siphonophora, Ctenophora, and pelagic worms. Very little is known about the food of the Rhizostomata, but the small size of the mouths of these forms suggests that their food must also be of minute size. The frequent association of small fish with the larger jelly-fish is a matter of some interest that requires further investigation. In the North Sea young whiting are the constant guests of _Cyanaea capillata_.[343] Over a {312}hundred young horse-mackerel (_Caranx trachurus_) may be found sheltering under the umbrella of _Rhizostoma pulmo_. As the animal floats through the water the little fishes hover round the margin, but on the slightest alarm dart into the sub-umbrella cavity, and ultimately seek shelter in the sub-genital pits.[344] Two species of fish accompany the American Medusa _Dactylometra lactea_, one a Clupeoid, the other the young of the Butter-fish (_Stromateus triacanthus_). According to Agassiz and Mayer[345] this is not an ordinary case of mutualism, as the fish will tear off and devour fragments of the tentacles and fringe of the Medusa, whilst the Medusa will in its turn occasionally capture and devour one of the fish. A great many of the Scyphozoa, particularly the larger kinds, have the reputation of being able to sting the human skin, and in consequence the name Acalephae[346] was formerly used to designate the order. Of the British species _Aurelia aurita_ is almost harmless, and so is the rarer _Rhizostoma pulmo_; but the nematocysts on the tentacles of _Cyanaea_, _Chrysaora_, and _Pelagia_ can inflict stings on the more delicate parts of the skin which are very painful for several hours, although the pain has been undoubtedly greatly exaggerated in many popular works. The soft structure of the Medusae does not favour their preservation in the rocks, but the impressions left by several genera, all belonging apparently to the Rhizostomata, have been found in Cambrian, Liassic, and Cretaceous deposits. There is reason to believe that many Scyphozoa exhibit a considerable range of variation in the symmetry of the most important organs of the body. Very little information is, however, at hand concerning the variation of any species except _Aurelia aurita_, which has been the subject of several investigations. Browne[347] has found that in a local race of this species about 20 per cent exhibit variations from the normal in the number of the statorhabs, and about 2 per cent in the number of gastric pouches.