Loose agreement of a radio position of low accuracy with that of some object listed in the NGC is not sufficient to provide the identification of a radio source. Even satisfactory coincidence of a precise position with that of an astronomical object requires supporting evidence. Agreement of the size of the source with that of the visible object, at least in order of magnitude, is an important argument in favour of an identification; exact agreement of sizes can be expected only where radio and optical emission are physically connected. The radio spectrum, the optical spectrum, and the physical characteristics of the visual object also have to be taken into account. Observations of the radio spectrum should be particularly useful to support the identification of sources with H 11 regions which can be recognized from their thermal emission even if they are obscured and optically inaccessible. If all data are available, satisfactory agreement exists between optical and radio observations. The best example of this kind at the moment is perhaps NGC 2237, the Rosette nebula, reported as a source by Ko and Krauss (1955)  and also observed by Mills, Little and Sheridan (1956 ; see also paper 18).