In the past, deep-sea water samples for bacteriological purposes have been taken either by means of water-bottles provided with taps, and so made that the sample came in contact with metal, or in evacuated sterilized glass vessels drawn out to a fine point and sealed in the flame. The water-bottle is difficult to sterilize on account of the taps, and the results obtained are vitiated by the bactericidal action of metals. The sealed glass tube is free from these defects, but at even moderate depths it is liable to collapse on the sudden change of pressure when the end of the capillary portion is broken off.