212 DISCOVERY REPORTS 



Water Bottles and Thermometers. Water bottles were supplied by Dr Martin 

 Knudsen from the Laboratoire Hydrographique at Copenhagen and are of three kinds : 

 the Nansen-Pettersson insulating water bottle, the Ekman reversing water bottle and 

 the Knudsen full speed water bottle. 



In the Nansen-Pettersson instrument the water sample is insulated by a number of 

 concentric jackets and the temperature is read by a thermometer of normal type in- 

 serted in the top of the water-chamber. This pattern is made for attachment to the end 

 of the wire and in consequence only one instrument can be used at a time; it can, how- 

 ever, be shot and hauled very rapidly since the thermometer will take up the tempera- 

 ture while being hauled to the surface. If used in deep water the temperature is liable 

 to rise, owing to the heat given off by the expansion of the water as the instrument is 

 hauled to a higher level; it was, however, never used at greater depths than 400 m., 

 the error arising from this cause being then negligible. 



The water bottle is of the type described and figured by Knudsen in 1923.^ The 

 release mechanism in this pattern is effected by a messenger with a conical excavation 

 in the bottom, which impinges on a small striking plate set in the conical headpiece 

 of the bottle. This mechanism was not considered entirely satisfactory and a different 

 form, in which a flat-bottomed messenger strikes a horizontal lever, was substituted. 

 It is believed that a similar arrangement has been adopted in the latest models. The 

 Nansen-Pettersson bottle proved most reliable on both vessels, thousands of tempera- 

 tures and water-samples being taken without the smallest hitch. 



For deeper water the Ekman reversing bottle was used, the model also being that de- 

 scribed by Knudsen in 1923.^ In this pattern the cylinder in which the water is collected 

 capsizes when the messenger reaches the instrument, and the thermometers, two in 

 number and carried in brass tubes attached to the cylinder, are capsized also. The 

 thermometers are of a special type in which the mercury column is cut off at a particular 

 point when they are reversed, and the only alteration in the reading while the instru- 

 ment is being hauled to the surface is that due to the contraction or expansion of the 

 cut-off column of mercury. In order to correct this error, which is always small, a 

 secondary thermometer is sealed up in the same tube as the main thermometer, the 

 correction being calculated from the two readings by means of a simple formula. 



A single bottle of this type cannot be operated so quickly as one of the insulating 

 pattern, for it must remain at the desired level until the thermometer has taken up the 

 temperature. The bottles, however, are made so that they can be clamped on the wire 

 at any point, and each is fitted with a device by which a messenger can be released on 

 the wire below it. In consequence a series of bottles can be used at a single operation, 

 and in deep-water work a great saving of time is effected. 



The Ekman reversing bottle is a most reliable instrument. The only defect noted is 

 that the cylinder is if anything too evenly balanced on its pivots, with the result that 

 there were occasions when it failed to engage in the clip on the lower part of the frame 



1 Knudsen, Pub. Circ, Cons. Explor. Mer, 77, pp. i-i6, 1923. 



