METHODS AKD DATA. 3 



go too deep than too sliallow. At tbe same time, as I have before sliown*, the augle made by the 

 visible wire iu going over is not au acf;urate indication of the lie of the submerged wire : for example, 

 during this cruise, on one occasion (luiuls 35 c and d) my nets^ streamed from the port quarter, fouled a 

 sounding-wire with thermometers let go from the starboard bow ; on discovering the foul, both wires 

 were hauled, and both appeared to be straight up and down for many fathoms, although really foul of 

 one another at 70 fathoms depth under the keel amidships. 



Two attempts to improve the ordinary tow-nets are perhaps worth recording, although both were 

 failures, since they may serve as warnings to other marine naturalists. In the one, a funnel of boulting- 

 cloth hung down loosely into the net from the frame, to serve as a " pocket " which should keep the 

 faster swimmers from escaping ; its effect in practice was to prevent animals from getting into the net by 

 diminishing the opening, rather than to keep them there. Another experiment was a funnel of boulting- 

 cloth, which was fastened to the net just above the collecting-tin, the lower end hanging in the tin, with 

 the idea of saving the organisms which get crushed in the angle between net and tin when the latter is 

 lifted out of water : it was not a success. 



(2) The Mesoplankton net, which was fully described in the Proc. Zool. Soc. for 1898, pp. 568-574, 

 worked fairly well, but with 23 per cent, of failures : these were traced to easily remediable causes, 

 notably to the net-frame being too light for the lock, if lowered fast or used in a heavy swell : these 

 defects have been remedied in nets built subsequently for the Antarctic sliip 'Discovery' and for 

 Dr. Wolfenden. 



(3) The Mesoplankton trawl, having a much heavier net-frame, worked quite satisfactorily as a 

 machine, but caught so little, in comparison with the time occupied on deep hauls, that it was soon 

 abandoned. It had been devised to catch large organisms, fish, and cephalopods, which escape the 

 ordinary net owing to its small size and slow ^peed. I attribute its failure to catch more than it did, to 

 its being still too small and hauled too slowly ; it could only be heaved in at about 2 miles an hour, 

 the maximum rate at which the winch hands could reel up the wire rope as delivered from the deck- 

 engine. When touched, large animals seem to jump clear of a small net : I have therefore designed a 

 17-foot-square trawl to open and shut, which I hope may yield better results, if the winch is geared to 

 the deck-engine. 



The epiplankton nets were out for different periods of time, generally for about an hour. I think 

 that uniform hauls of half an hour would have been better. In an hour more material is, of course, 

 caught ; but it seems to increase the number of specimens rather than of species, with the results that 

 the condition of the material is not so good, and that the labour of sorting the catch, always tedious, 

 becomes doubly so. In consequence, I have taken for identification only about half the material in 

 many of the epiplankton hauls ; from the other half have been taken conspicuous novelties, and 

 organisms of which every specimen was desired for statistical (quantitative) study. 



As regards reagents for dealing «ilh the bulk of the catch, which remains when the larger specimens 

 have been picked out, the chief thing to avoid is mercury bichloride: this foiinsa dirty, sticky, insoluble 

 deposit all over the specimens, apparently consisting of an albuminous conqjound with mercury. My 

 best results, on the whole, were got by adding a little pure formalin to kill, and when the animals had 

 sunk to the bottom of the jar, by preserving in 3 per cent., followed by 5 per cent, solution of formalin 

 in fresh water. 



The details in Table I. are extracted from the ship's log. The force of the wind is 

 expressed ou Beaufort's scale, of which 



1 represents a velocity of 8 miles per hour, 

 3 J, ,) 1<J » „ 



" yy >) ■'■^ >} Ji 



A. 'i'\ 



^ y> J* '*'-' >} 39 



* Troc. Zool. Soc. 189S, p. 5GS. 



