6. WATER SAMPLING AND THERMOMETERS 



H. F. P. Herdman 



1. Sampling 



The collection of subsurface water samples and the determination of an 

 accurate temperature at the depth where the sample is taken are still important 

 factors in oceanography, and the method of taking routine serial temperatures 

 and samples remains essentially as described by Sverdrup, Johnson and 

 Fleming (1942). These authors also mention two types of subsurface water- 

 sampler commonly used (Ekman and Nansen) which are still standard equip- 

 ment, but do not mention a further type which is in common use in European 

 waters. This is the Knudsen reversing water bottle (1929), which is similar to 

 the Nansen bottle in that it is frameless, being clamped to the wire rope at the 

 bottom and held to the wire at the top by a catch. It differs from the Nansen 

 bottle, however, in that it is sealed by independent hinged lids at either end. 

 When lowering the sampler these lids are held open, against an internal spring, 

 by a simple mechanism connected to the release catch. When the messenger 

 hits this catch, the lids are immediately closed and the bottle falls away 

 through 180°, pivoting on the bottom clamp. This action also reverses the 

 thermometers, which are housed in a frame independently attached to the 

 bottle. 



None of these samplers is entirely satisfactory. In the Nansen type, for in- 

 stance, constant attention is required to prevent leakage at the valves, and 

 distortion of the cam-operated push rods and end plates of the Ekman bottle 

 is often the cause of a poor seal in this type. The Knudsen bottle is more 

 positive in its sealing action owing to the narrow openings and the small 

 hinged lids but it is difficult to clean the inside of the tube. With valve-type 

 bottles it is virtually impossible either to sight or to clean the interior. In 

 recent years it has been found that this inability to clean the inside of a water 

 bottle has led to deterioration of any metallic plating (commonly silver or 

 nickel) used as a lining. This can be a serious defect when samples are being- 

 collected for the determination of dissolved oxygen, as the brass tubing normally 

 used for the bottle is then exposed and will in time occlude oxygen. If the 

 sample is only in the bottle for a few minutes, this effect is not appreciable, but 

 it can be substantial for samples from the greater depths, which may be held 

 in the water bottle for an hour or more. To overcome this difficulty, the Munro- 

 Ekman water bottle, used by the National Institute of Oceanography in Great 

 Britain, was first of all lined with a silver tube spun into the bottle, then with 

 a proprietary brand of a phenol-formaldehyde plastic and, latterly, with nylon. 

 All these linings are inert in sea-water, but nylon has been preferred since it is 

 flexible and so not easily cracked. 



More recently, Fjarlie (1953) has described a simplified water-sampling 

 bottle with better flushing arrangements. In this type only the thermometer 

 frame reverses, the bottle remaining clamped to the wire rope. Closure of the 

 tube is effected by flexibly mounted lids with "mousetrap" type springs on 



[MS received Jvly, I960} 1 24 



