568 DE. G. H. rowLEE ON THE [June 21, 



point, I endeavoured to introduce into it such improvements as 

 would obviate what appeared to me to be its weaknesses, namely : 

 (1) The position of the vsires when the net had shut, Mhich 

 necessitate the mouth being alwaj's slightly open ; (2) the lack of 

 power to keep the net-mouth shut in a roll of the ship or a check 

 on the Hue, as the attachmeuts of the wires by which it then hangs 

 are so close together; (o) the speed at which the whole structure 

 must be towed in order that the screw-propeller, and the rod to 

 which it is fixed, may overcome the frictional resistance offered by 

 the rings on which the weight of the net is hanging. 



I decided to construct a net for vertical and not for horizontal 

 use, because it seems to me, on the basis of my small experience, 

 impossible to be certain of the depth at which a net is being towed 

 horizoutally. The usual method for this is to lower the net 

 vei'tically, and to begin towiug with the rope straight up and down ; 

 then to observe the angle made by the rope with the horizon by 

 means of a quadrant, and to calculate the vertical depth of the net 

 by traverse tables on the assumption that the towing-line is the 

 hypotenuse of a right-angled triangle. Unfortunately for this 

 method, however, the towing-rope is not a hypotenuse, but 

 forms an uuknown catenary, which varies with the weight of the 

 net, its resistance to the water, and the pace of towing ; this forms 

 an increasing source of error, the greater is the length of towing- 

 warp out. As an example of the uncertainty of this method, — I 

 struck bottom at 398 fathoms in the Faeroe Channel, when by 

 quadrant and traverse tables the net should have been at 300 

 fathoms with 450 fathoms of rope out. There are so many forces 

 at work as to make it impossible for any but a highly skilled mathe- 

 matician to calculate the probable position of the net, and this 

 only after tedious experiment. 



Description of the Apparatus. 



This consists of the net, the net-frame and chains, and the 

 locking-gear. As the first of these were used both in lS9t5 and 

 1897, they will be described in detail ; the locking-gear of the 



1896 pattern will only be sufficiently sketched to enable future 

 workers in this field to profit by my experience of failures ; the 



1897 pattern will be fully described. 



The net is made of Swiss Silk Boulting Cloth, by far the best 

 material known to zoologists for every form of tow-net ; it was 

 supplied by Messrs. Staniar of the Manchester Wii-e Works ; this 

 material will stand almost any fair pull, but, as it is very liable to 

 be cut by anything sharp, when coming inboard, the actual net 

 is surrounded by a loose case of common mosquito-netting. A net 

 with a twenty-inch square mouth, tapered to a four-iuch diameter 

 cod-end, and six feet in length, was found to be a good working 

 size. It should be sewn throughout by hand, not by machine ; 

 and with strong sewing-silk, not thread. If washed nightly in 

 fresh water and dried in the air, a net of this sort w ill last for a 

 very long time. 



