THE REARING OF SEA-FISH LARVJ?. 79 



effect on the fishes was noticeable, but the tank was appreciably 

 darkened. Next day five of the larvae had completely disappeared, 

 and no trace of their remains could be discovered in the gravel or 

 elsewhere. The circulation was all right, and escape by the siphon 

 overflow (a broad funnel imbedded deeply in the gravel) was quite 

 impossible. On further examination it became practically certain that 

 the larvae had been devoured by Harpacticid Copepods (almost ex- 

 clusively Idya fuTcata) which had entered their tank through the 

 siphon from the supply tank above. These Copepods were present in 

 the fishes' tank in myriads — on the slate sides, on the glass, in the 

 gravel, and freely swimming ; and I subsequently found, from an 

 examination of the supply tanks, that they abounded in many of these 

 tanks, especially on the glass fronts, from which, in this case, they had 

 clearly been drawn through the siphon supplying the portable tank. 

 The siphon had inadvertently been displaced, so that its mouth rested 

 against the glass front of the supply tank. 



The rapacity of these Copepods was easily tested upon dead or 

 moribund larva3 derived from other experiments. The Idya appears to 

 fix itself to its prey by the second antennae in front, and the second 

 maxillipeds behind. The mandibles, or neighbouring appendages (the 

 precise organs were not determined), then make a rapid series of bites 

 from before backwards, accompanied by an incessant scratching move- 

 ment carried on by the setae of the mouth parts. The fin membrane of 

 a larva was quickly gnawed away in semicircular patches, like the holes 

 on the edge of a leaf by caterpillars. To test the rate of dismember- 

 ment, three entire corpses of Blenny larvae were placed at 4 p.m. in a 

 small dish containing sixty-eight Copepods, all of which, as determined 

 after the experiment, were Idya furcata. Next morning the corpses 

 had entirely disappeared except a fragment of one larva, and the 

 faintest possible traces on the sites of the other two. 



In order to see whether the Copepods would also attack living prey, 

 I placed four healthy young Gobius minuius, each about one inch long, 

 in the tank. On examining one, which was near the glass front, some 

 time afterwards with a lens, I could see two of the Copepods on the 

 base of the tail fin, busily engaged eating the fin membrane. The Goby 

 appeared to feel no serious discomfort from their proceedings, thougli it 

 occasionally made restless movements as though slightly irritated. Its 

 movements, which were repeated three or four times while I watched, 

 did not dislodge the Copepods. Next morning one of the Gobies was 

 dead, and had been browsed upon by one at least of its companions, as 

 well as by the Copepods, Its death, however, seemed attributable to 

 the latter. 



On the other hand, the Copepods appeared to prefer dead to living 



