Water-Sampling Devices

The types of water-sampling devices that will be described are those that are intended for taking samples at subsurface levels for physical and chemical studies. Samples for the enumeration of phytoplankton and for bacteriological examination may be obtained with these instruments, but for such purposes specially designed samplers are more commonly used. A water sampler for collecting at subsurface levels is so designed that it can be closed watertight at any desired depth, and thus the enclosed sample is not contaminated by water at higher levels or lost by leakage after the bottle is brought on board. Because of the great pressures encountered in deep water the sampling bottles are sent down open and then closed at the required depths by means of messengers or propeller releases. To expedite work at sea, water-sampling devices are used in series—that is, with more than one bottle on the wire rope so that samples can be taken at a number of depths on the same cast. As it is essential that temperatures and water samples be taken at the same depths, the water-sampling bottles are fitted with frames in which one or more reversing thermometers are placed. An exception to this is the insulated Pettersson-Nansen bottle, which can be used for the upper few hundred meters (p. 354). Water-sampling devices must be constructed of noncorrosive materials that will reduce contamination of the water samples to a minimum. The bottles are usually made of brass, plated inside with tin or silver or coated with a special lacquer. For removing the water sample they are fitted with a drain cock and an air vent. To be clearly visible when hauling in, the bottles should be painted white. Many types of sampling bottles have been invented, but the rigorous working conditions and the desirable features that have been listed above have reduced the types in general use to only a few of simple but rugged design.

---

---

Messengers are essential for the operation of many types of oceanographic equipment. Although their size and shape will vary for different types of apparatus, they are essentially weights which are drilled out so that they will slide down the wire rope. In order to remove or attach them they are either hinged or slotted. The speed of travel depends upon the shape and weight of the messenger and upon the wire angle (the angle the wire rope makes with the vertical). With no wire angle the messengers used with the Nansen bottles travel approximately 200 m per minute.

Water-sampling devices are of two general types, depending upon the method of closing, which may be accomplished by means of plug valves or by plates seated in rubber. The Nansen bottle, an example of the first type, is the one most widely used in oceanographic research. The Ekman bottle is of the second type.

The Nansen bottle (fig. 87) is a reversing bottle fitted with two plug valves and holding about 1200 ml. The two valves, one on each end of the brass cylinder, are operated synchronously by means of a connecting rod fastened to the clamp that secures the bottle to the wire rope. When the bottle is lowered, this clamp is at the lower end, and the valves are in the open position so that the water can pass through the bottle. The bottle is held in this position by the release mechanism, which passes around the wire rope, but, when a messenger sent down the rope strikes the release, the bottle falls over and turns through 180 degrees, shuts the valves, which are then held closed by a locking device, and reverses the attached thermometers. After reversing the bottle, the messenger releases another messenger that was attached to the wire clamp before lowering. This second messenger closes the next lower bottle, releasing a third messenger, and so on.

The Ekman bottle, which can also be operated in series, consists of a cylindrical tube and top and bottom plates fitted with rubber gaskets. The moving parts are suspended in a frame attached to the wire rope, and, when the instrument is lowered, the water can pass freely through the cylinder. When struck by a messenger the catch is released and the cylinder turns through 180 degrees, thereby pressing the end plates securely against the cylinder and enclosing the water sample. Reversing thermometers are mounted on the cylinder.

Thermally insulated bottles such as that of Pettersson-Nansen (for illustration, see Murray and Hjort, 1912) consist of several rigidly fixed concentric cylinders each of which is fitted with end plates. When these plates are shut, a series of water samples are isolated one within the other. The outermost cylinder and end plates are constructed of brass and ebonite, the inner ones of brass and celluloid, the cylinders

---

and end plates being mounted on a frame. In open position there is space between the cylinders and the upper and lower end plates, but, when the catch is struck by the messenger, the cylinders and the upper plate slide down, the cylinders are pressed against the lower end plate, and the upper one seals the bottle. Owing to its construction the bottle must be attached to the end of the wire rope. The temperature at the depth of sampling is determined by a thermometer inserted in the innermost cylinder. Corrections must be applied for adiabatic cooling and, under extreme conditions, for heat conduction.

On the Meteor Expedition a glass-lined sampling bottle with a capacity of 4 l was used to collect large samples with a minimum of contamination. This bottle was attached to the end of the wire rope, the closing mechanism being similar to that used for the insulated bottle. A special mounting that reversed when the bottle was closed was provided for thermometers (Wüst, 1932).

In connection with wire sounding and bottom sampling, it is often desirable to obtain the temperature and the water samples close to the bottom and, as an added check, to obtain the depth by means of an unprotected thermometer. In order to avoid waiting for a messenger to travel all the way to the bottom, possibly one half hour or more, special sampling devices activated by propeller releases are used (Soule, 1932; Parker, 1932). These are usually reversing bottles in which the release pin is attached to a propeller. When the apparatus is being lowered, the propeller holds the pin in place, but, as soon as hauling in is commenced, the propeller rotates and withdraws the pin.

Spilhaus (1940) has devised a multiple sampler consisting of six small valve-closing bottles that can be closed individually at predetermined depths by releases which are activated by the hydrostatic pressure. The instrument was designed to be used in conjunction with the bathythermograph (p. 352) from a vessel while under way.

The general procedure for taking water samples and temperatures by means of Nansen bottles or other serial sampling devices is as follows. In order to keep the wire rope taut and to reduce the wire angle, a stray weight is attached to the end of the wire rope, usually of 50 to 100 lb, depending upon the size of the rope, the depths at which observations are to be taken, and the general working conditions. A certain amount of the wire rope (25 to 50 m) is then payed out so that the weight will not strike the ship when the bottom bottle is being attached or detached, and also to reduce the possibility of damaging the bottle if the weight strikes bottom.

The depth at which samples are to be collected should be planned before the cast is begun. The first bottle, adjusted to the set position, is then attached to the wire rope, the thermometers are checked, and the meter wheel is set at zero. When the wire has been lowered the required

---

amount, the second bottle is attached. To operate the bottles, messengers must be attached to the messenger releases on the second bottle and on all those above. The number of bottles attached on a single cast is determined by the strength of the wire and by the operating conditions. Up to twelve or more are sometimes used at one time. After all the bottles have been attached at appropriate intervals, they are lowered to the required depths. Since the meter wheel was set at zero when the first bottle was attached, the entire cast must be lowered by an amount equal to the distance between the sea surface and the height at which the bottles were put on. After the bottles are lowered to the required depths, they are allowed to remain there for about 10 minutes to permit the thermometers to reach the temperatures of their surroundings, and then the messenger is dropped. If the bottles are located at depths of less than about 500 m, their proper functioning may be checked by touching the wire, as it is usually possible to feel the jerks when the messengers strike the bottles. When the samples are taken at great depths or when the wire angle is large, it is impossible to feel the jerks, and therefore sufficient time must be allowed for the messengers to travel to the deepest bottle before hauling in. When the wire angle is large, the slower rate of travel of the messenger must be taken into consideration. The wire is then hauled in and the bottles are removed and placed in their rack in the deck laboratory, care being taken to avoid turning the bottles from the reversed position, since the reversing thermometers may set themselves again.

Just before the messenger is released the wire angle should be estimated or measured. Knowledge of the wire angle is useful when determining the depth of sampling, as will be shown later.

Certain accessories are necessary when handling water-sampling bottles. As a safety measure a light line with a harness snap should be attached to the rail. This line is snapped on the bottle before it is handed to the operator on the working platform and is not removed until the bottle is firmly clamped on the wire rope. It is also attached when removing the bottle. To hold the wire rope steady and close to the platform a short line with a large hook should be attached to the platform. The hook is placed on the wire rope when instruments are being attached to it or removed from it. When the vessel is drifting with wind or surface currents, the wire will not hang vertically and will sometimes trail so far away that the wire angle may be as much as 50 or 60 degrees. Under these circumstances the wire rope must be pulled in by means of a boat hook or block and tackle in order to attach the hook.

Because of the larger vertical gradients in the distribution of properties in the upper layers, observations are taken at relatively close intervals near the surface and at increasingly larger intervals at greater depths. The International Association of Physical Oceanography,

---

in 1936, proposed the following standard depths at which observations should be obtained directly or by interpolation from the distribution at other levels. The lower limit is determined by the depth of water or the plan of operations. The standard depths, in meters, are: surface, 10, 20, 30, 50, 75, 100, 150, 200, (250), 300, 400, 500, 600, (700, 800, 1000, 1200, 1500, 2000, 2500, 3000, 4000, and thereafter by 1000-m intervals to the bottom. The depths in parentheses are optional. In addition to observations at standard depths, it is often desirable to obtain temperatures and water samples close to the bottom. In deep water this usually means within about 50 m of the bottom, but in moderate and shallow depths it may be much less, depending upon the bottom topography and the working conditions.