A DEEP-SEA. BACTERIOLOGICAL WATER-BOTTLE. 529 



North Atlantic in 1910. There has been no reason to doubt any of 

 the results obtained with it, and on the single occasion when it failed 

 to close it appears to have been lying on the bottom. It is true that 

 at one station the results are decidedly difficult to explain, but in this 

 case the weather made it necessary to keep the boat moving ahead 

 while the water-bottle was out, and the depth actually reached is 

 problematical; the stray was so great that at the time it was estimated 

 that it might be only half that shown by the amount of wire out. 



Various improvements might be made which would add considerably 

 to the convenience of the water-bottle. In particular, the releasing 

 rods should be arranged centrally, by means of elbows in the case 

 of the longer one. As at present made, the supporting rods are 

 placed asymmetrically with regard to the springs, and there is con- 

 sequently a twisting moment which tends to jamb the guides on 

 the frames ; this can be prevented at present only by a very careful 

 adjustment of the strength of the various springs. 



The larger messenger is also somewhat unsatisfactory; in spite of 

 numerous holes bored in its lower half to allow the water to escape, 

 it takes about half an hour to fall through 800 fathoms, and at the 

 same time it oscillates from side to side so violently that the wire 

 quickly wears out the central hole to a funnel shape. 



