Amongst the many dangers to which ships crossing the Atlantic are exposed is that of collision with icebergs. These are large masses of ice which have become detached from the mighty ice-fields of the north, and which travel slowly and majestically southwards, growing smaller and smaller as they pass into warmer seas. Icebergs give no warning of their coming, and in foggy weather, which is very prevalent in the regions where they are encountered, they are extremely difficult to see until they are at dangerously close quarters. Attempts have been made to detect the proximity of icebergs by noting the variations in the temperature of the water. We naturally should expect the temperature of the water to become lower as we approach a large berg, and this is usually the case. On the other hand, it has been found that in many instances the temperature near an iceberg is quite as high as, and sometimes higher than the average temperature of the ocean. For this reason the temperature test, taken by itself, is not at all reliable. A much more certain test is that of the salinity or saltness of the water. Icebergs are formed from fresh water, and as they gradually melt during their southward journey the fresh water mixes with the sea water. Consequently the water around an iceberg is less salt than the water of the open ocean. The saltness of water may be determined by taking its specific gravity, or by various chemical processes; but while these tests are quite satisfactory when performed under laboratory conditions, they cannot be carried out at sea with any approach to accuracy. There is however an electrical test which can be applied accurately and continuously. The electrical conducting power of water varies greatly with the proportion of salt present. If the conductivity of normal Atlantic water be taken as 1000, then the conductivity of Thames water is 8, and that of distilled water about 1/22. The difference in conductivity between normal ocean water and water in the vicinity of an iceberg is therefore very great. [Illustration: _By permission of_] [_Dr. Myer Coplans._ FIG. 43.--Diagram of Heat-compensated Salinometer.]