54 



NATURE 



[NOVEMIJER lO, 1 910 



particularly large quantity, of fish in proportion to the 

 number of hauls ; and they naturally obtained but little 

 information regarding such questions as the depth at which 

 tlic animals live, and their vertical wanderings by night 

 and day. These questions seemed to me to be of special 

 interest at the present juncture, and accordingly an essential 

 part of the work of our expedition was directed towards 

 therr solution. We first constructed some large nets of 

 3*25 metres diameter, partly of coivser silk and partly of 

 prawn-net, arranged to close on the principle of Nansen's 

 closing net (see Figs. 2 and 3). With these we made 

 several successful hauls at various depths, and obtained 

 sufficient catches of the commonest forms to enable us to 

 determine more approximately the actual depth at which 

 they occur. Nevertheless, we soon discovered that even 

 these large nets yielded merely an incomplete collection of 

 the fauna, since many species occur far too sparsely to be 

 caught with vertical hauls. It was therefore found neces- 

 sary to employ large horizontal-fishing appliances and to 

 make hauls of considerable length. 



Such, hauls would, however, take an unduly long time, if 

 they were to be carried out singly at the same station, for 

 hours in succession, at different depths. It was, therefore. 



largest net, in particular, worked splendidly. We hav* 

 thus discovered quite a number of species of pelagic deep- 

 sea fish not previously described. 



As there were so many stations, and we fished in widely 

 differing waters and at all hours of the day and night, a 

 comparison of these catches with each other will afford 

 much information concerning the geographical distribution 

 of the different species, as well as regarding the depth at 

 which they occur by day and by night, and so on. The 

 catches show that the hauls have much in common, and 

 we may accordingly assume that they are in the mam 

 representative of the depth in which the appliances have 

 been towed ; and it is further extremely satisfactory to 

 note that the experiences gained from these hauls and from 

 the vertical closing-net are in close accordance. 



It is too soon yet, and, moreover, would take too long, 

 to describe in full the results of our experiences. I will 

 confine myself, therefore, to mentioning that everywhere 

 in the Atlantic Ocean, from the Wyville Thomson Ridge 

 to the Sargasso Sea, there appears to be, at depths below 

 400 metres, a consistently uniform fauna of small, chiefly 

 black pelagic fish, large red crustaceans, numerous 

 medusae, &c., a fauna which, in any case so far as the 

 fishes are concerned, is probably also shared by other 

 oceans, and which presents the same variety of form that 

 the Valdivia expedition, for instance, has found in the 

 Indian Ocean, and the Challenger in the Pacific. In the 

 upper layers, at depths less than 400 metres, we have 



I 



KiG. 5. 



particularly desirable to drag a number of appliances at 

 several depths simultaneously. The appliances had in this 

 case to be fastened to one or two wire ropes, as one 

 cannot tow many wires at the same time. The technical 

 difficulty now presented itself that long lengths of wire get 

 twisted, when towed, and consequently destroy the appli- 

 ances or displace their position in the water. We solved 

 this by an arrangement, shown in the accompanying figure 

 (Fig. 4), by which a shackle to which the appliance is 

 fastened moves freely round the wire. By this means it 

 became possible to have no fewer than ten appliances out 

 simultaneously from two wires, as shown in the figure 

 (Fig. 5). Here we see a series, consisting partly of nets, 

 partly of Dr. C. G. Joh. Petersen's young-fish trawls, in 

 use at the following lengths of wire : o, 100, 200, 300, 600, 

 1000, J500, 2000, 2500, 3000 metres. The total number of 

 these tbwing stations exceeded thirty. 



The material obtained in this way was very large indeed. 

 From the same station hundreds of pelagic deep-sea fishes 

 an^i litres of large decapods, medusae, &c., were secured. 

 All the same, the hauls showed that the material was not 

 by any means too large, since right up to the very last 

 haul we continued to capture a few species of pelagic fishes 

 that had not occurred in any of the previous hauls. The 



NO. 2 141, VOL. 85] 



discovered numerous younger stages of fish that are not 

 as yet determined, mostly of transparent, colourless form, 

 such as Leptocephali, to take merely one example. 



Trawlings. 

 During previous expeditions in the Atlantic Ocean a 

 great number of hauls have been undertaken either with 

 the dredge or with small trawls. There was, therefore, no 

 pressing necessity for the Michael Sars to investigate the 

 bottom-fauna of the Atlantic, more particularly as hauls of 

 this nature require a considerable expenditure of time, ana 

 could therefore with difficulty be combined with our exact- 

 ing programme of hydrographical and plankton investiga- 

 tions. It was of interest, on the other hand, to try 

 whether a large-sized model of the ordinary otter-trawl 

 (with 50 feet of head-rope) would yield new results. 

 During my previous researches I had succeeded to my 

 satisfaction, and had secured very good catches, by making 

 use of a trawl of this kind at depths down to 1000 fathoms. 

 It was, in my opinion, especially desirable to employ this 

 appliance along the Continental slope from the Wyville 

 Thomson ridge southwards to the tropical coast of Africa, 

 so as to ascertain the composition of the fauna on this 

 long stretch at depths varying from 500 to 1600 fathoms. 



