Abstract The idea of a thin multiwire chamber with thick wires as a possible sampling detector for calorimeters was checked out in practice. Two small test detectors with gas sample thickness of 2 and 3 mm were constructed and tested in the laboratory. Avoiding argon, several pure hydrocarbons or binary mixtures with CO 2 were found to quench photons efficiently enough to make it possible to have charge multiplication factors of the order of 10 7. Efficiency of operation in this probably limited streamer mode for minimum ionizing electrons and with a single avalanche (streamer) response was found to be higher than 98%, with pulse-height resolutions down to 60% fwhm. The effect of gap-length variation is quite small; a gap change of upto 25% results in less than a 20% change in the mean charge released in a streamer.