334 University of California Publications in Zoology. [Vol.8 



(see Tanner, 1894) seems to be subject to increasing error as the 

 depth is increased. Repeated instances occur in practice (see 

 Fowler, 1898) in which bottom has been struck by apparatus 

 when the dredging quadrant indicated that it should be some dis- 

 tance above the bottom. Giesbrecht (1893), as a result of experi- 

 ments in aquaria with miniature apparatus, has arrived at the 

 conclusion that much of the curvature, at least under certain 

 conditions of strain, may occur in the upper part of the cable and 

 the lower end be more and more approximated to the perpen- 

 dicular. 



In any event it is possible with the net here described, sus- 

 pended by a %-inch plow-steel cable, to a depth of 550 fathoms, 

 to tow sloMdy with the visible part of the cable in an almost 

 vertical position. Even with greater speed and some obliquity of 

 the cable it is possible to reduce the vertical component of the 

 path of the net far below that which is unavoidable in all vertical 

 closing nets and perhaps, because of its greater weight, to reduce 

 that hitherto obtainable in other types of horizontal nets. 



"When the towing is completed and the cable is again perpen- 

 dicular the second messenger is sent down and releases the upper 

 jaw which falls to the lower position, is held firmly against the 

 lower jaw by the spring, and thus closes the net (pi. 23, fig. 3). 

 In the original design spring-catches were used to hold the jaws 

 firmly shut after closure. It has been found in practice that the 

 closing secured by springs is adequate and constant. The cable 

 is then reeled in, the spring in the supporting apparatus per- 

 mitting the bucket to drop downward and thus tending to con- 

 centrate the catch towards the lower end of the net during ascent. 

 Upon arrival on deck the rim of the net is released by unscrewing 

 the thumb-nuts on the doors to chambers containing the hubs and 

 dropping the doors downward (pi. 23, fig. 3). The set screws at 

 the rear of the hubs are released and the net is lifted from the 

 frame, its sides rinsed down to concentrate all the catch in the 

 bucket, which is then removed. The net may be quickly returned 

 to position for a second descent, by replacing the hubs, closing 

 the gates, suspending the bucket and setting the spring. 



This net has been used extensively in the form of the older 

 model since the summer of 1908 (see Esterly, 1911a, b, and 



