June 26, 1890] 



NATURE 



lOI 



through the lamp so that the shielded or inclosed leg is 

 now the negative one, and the galvanometer is able to 

 detect no current, whether connected in one way or the 

 other. We establish, therefore, the conclusion that it is 

 the negative leg of the carbon loop which is the active 

 agent in the production of this " Edison effect," and that 

 if it is inclosed in a tube of either glass or metal, no cur- 

 rent is found flowing in a galvanometer connected between 

 the positive terminal of the lamp and this middle collect- 

 ing plate. 



Another experiment which confirms this view is as 

 follows .-—This lamp (see Fig. 7) has a middle plate, 

 which is provided with a little mica flap or shutter on one 



^ ^ 



Fig. 7.— Glow-lamp having mica shield s interposable between middle 

 plate M and negative leg of carbon, thereby diminishing: the "Edison 

 effect." 



side of it. When the lamp is held upright the mica shield 

 falls over and covers one side of the plate, but when it is 

 held in a horizontal position the mica shield falls away 

 from the front of the plate and exposes it. Using this lamp 

 as before, we find that when the positive leg of the carbon 

 loop is opposite to the shielded face of the plate, we get 

 the " Edison effect " as before in any position of the lamp. 

 Reversing the lamp current and making the same leg the 

 negative one, we find that when the lamp is so held the 

 metal plate is shielded by the interposition of the mica, 

 and the galvanometer current is very much less than 

 when the shield is shaken on one side and the plate 

 exposed fully to the negative leg. 



{To be continued.) 



SOME EXPERIMENTS ON FEEDING FISHES 

 WITH NUDIBRANCHS. 



VX/'ITH the view of testing the theory that the remark- 

 * '' able shapes and colours of N udibranchs are either 

 of a protective or of a warning nature,^ and are definitely 

 related to the edibility or the reverse of the animals, I 

 have been offering lately various kinds of N udibranchs 

 to the fishes in the aquarium of the Liverpool Free Public 

 Museum,- and have carefully noted the result of each 

 trial. 



Although these experiments will have to be repeated, 

 and additional evidence accumulated, still it may be in- 

 teresting to other biologists working on similar lines to 

 have this account of the inquiry, in its present stage, laid 

 before them, and I need scarcely say that I would be 

 glad of any suggestions which would be useful in future 

 investigations. 



Most of the experiments were made in three large fish- 

 tanks, which may be called A, B, and C. A and B are 



' Quart. Joiim. Microsc. Set'., vol. xxxi. p. 41. 



- With the kind permission and- assistance of Mr. T. J. Moore, the 

 curator, and his assistants, who were present at all the experiments. 



NO. 1078, VOL. 42] 



rectangular slate and plate-glass wall-tanks, lit from the 

 top, measuring 7^ feet long, s\ feet wide, and 3 j feet 

 high. A has a gravel bottom, and contains about 20 

 very healthy and active shannies {Blennius pholis) ob- 

 tained from the Menai Straits ; while B has a sandy 

 floor, and is devoted to flat-fish. It contains a consider- 

 able number of soles and plaice, a few small thornback 

 rays, turbot, and brill, and on one of the occasions had 

 some young cod. The average size of these flat-fish is 

 6 inches in length, and there are over 60 of them in the 

 tank. Both A and B have some rock-work. C is an 

 octagonal centre-tank with a sandy bottom, measuring 

 4i feet in diameter, and 17 inches in depth. It contains 

 various small fishes, viz. bullhead, goldsinny, pogge, 

 gemmeous dragonet, five bearded rockling, viper-weever, 

 and young cod. 



AH these fishes are apparently in a healthy con- 

 dition, and some of them have been living undis- 

 turbed in their tanks for periods varying up to four 

 years. They are usually fed upon mussels, cockles, and 

 occasionally worms, which are thrown in at the top of 

 the tank and allowed to sink through the water. Such 

 food-matters are usually seen at once, and eagerly pounced 

 upon and eaten during their descent. I adopted the same 

 plan in putting most of the Nudibranchs into the tanks, 

 and as the fishes were not fed on the days I intended to 

 experiment with them, and had usually been fasting for 

 24 hours when I began, they may be regarded as being 

 unusually eager to seize any object dropped into the water. 

 At the beginning and again at the end of each day's ex- 

 periments, we threw a couple of cockles or mussels into 

 the tanks, and found that they were at once caught and 

 bolted in the usual manner. 



I. October 29, 1889. [A supply of Doris^ bilamellata 

 was obtained from the rocks at New Brighton.] 



Tank A. — Doris. 



(1) Seized, when falling, by a shanny, and taken at once 

 to dark corner. 



(2) Seized and at once rejected, seized by another 

 shanny and at once rejected, seized by a third and re- 

 jected, then allowed to lie on Ijottom of tank. 



(3) Seized and rejected by two fish in rapid succession, 

 then seized by third and taken to dark corner. 



(4) Seized and rejected by first fish, taken to dark 

 corner by second. 



(5) Seized and rejected by three fish in rapid succession, 

 and then left. 



Tank ^.— Doris. 

 (i) Seized and rejected in rapid succession by a turbot, 

 a sole, another sole, and a plaice, and then left lying on 

 the sand. 



Tank C— Doris. 



(i) Seized and rejected by a goldsinny, tried again by 

 same and again rejected, then left. 



(2) Seized and rejected by a bullhead and by a 

 dragonet in rapid succession, and then left. 



Finally, another Doris was dropped gently into a fourth 

 tank containing a conger eel, so as to fall in front of its 

 nose ; but although the fish passed close to the Nudi- 

 branch several times while under observation, it appar- 

 ently took no notice of it. 



From these nine experiments it seems probable that 

 Doris bilamellata is distasteful to at least most of these 

 eight kinds of fishes tried. This was an unexpected 

 result, as Doris has no stinging apparatus, and certainly 

 seems to be protectively coloured. The distastefulness 

 may be due to the spicules in the skin or to the abundant 

 mucus covering the body. 



' I use throughout this article the old well-known generic names Doris and 

 Eolis. instead of the modern genera, only known to specialists, in which the 

 species I am dealins: with have been placed. No possible confusion can 

 arise from doing so. 



