The paper discusses the desirable features to be embodied in the design of apparatus for the detection and measurement of combustible gases and vapours in the atmosphere, and records some of the work carried out during the development of a portable gas detector which was manufactured for use by the Admiralty, Ministry of War Transport and the War Office.The paper describes tests carried out to determine the catalytic reaction of combustible vapours upon platinum and palladium filaments and relates the heat of combustion of a vapour or gas/air mixture to the heat given to electrically heated filaments.It is shown that the combustible properties of a gas can be studied by the use of a suitably designed apparatus employing heated catalytic filaments.The delaying effect of porous enclosures upon the reaction of combustible vapour/air mixtures on a platinum filament such as may be used in automatic methane detectors is studied, and it is shown that the pressure drop set up inside the container is extremely small and is dependent upon the catalytic properties of the filament and also upon the porosity of the enclosure.The difficulty of ensuring that the deposit of dust upon the enclosure will not affect the pressure drop, combined with the fact that the catalytic properties of the filament must be maintained at a constant value when taking comparative measurements, may lead to unreliable results.In the author's opinion the measurement of the catalytic reaction is to be preferred to that of measuring the fall in pressure, which is a secondary effect.Attention is drawn to the need of official adoption of standard methods for the testing and calibration of gas detectors, and the paper describes the methods adopted in the author's laboratory.It is stated that the practice of analysing small samples of gas taken from relatively large spaces is likely to give rise to inaccurate results, and it is claimed that, of known methods, continuously indicating instruments give the most reliable results, particularly when in the hands of unskilled personnel.The permissible inaccuracy of gas detectors is discussed and it is pointed out that detectors are often used for the measurement of concentrations which are below the inflammable range but are in the toxic range. In such cases it is considered desirable that there should be some official lead regarding the limits which are considered safe for normal working operations.Reference is made to research work which has been instituted as the result of the development of convenient methods of producing vapour in a readily measurable concentration