The NIOSH P-and-CAM 318 method for analysis of formaldehyde (50000) collected on a solid sorbent (charcoal) was studied in order to discover the cause of erroneously low analytical results obtained with this method. A total of 40 formaldehyde samples were collected from a test atmosphere and divided into eight groups which were analyzed by different methods. One group was collected and analyzed according to P-and-CAM 354 and used as a reference to which all other results were compared. All the other groups were collected by P-and-CAM 318 and analyzed as follows: two groups immediately, one in the authors' laboratory and one in a contact laboratory; two groups after storage at either ambient or elevated (40 degrees-C) temperature; and two groups after one had been stored in the laboratory while the other was mailed to and returned from the contact laboratory, in order to determine any possible effects of shipping. The results obtained after storage at ambient temperature or 40 degrees-C were 25 and 75 percent lower, respectively, than those obtained after immediate analysis. There were no significant differences between the results from the two laboratories, and the samples which were shipped and then returned to the authors' laboratory gave higher results than those which had not been shipped. Analysis of a further 16 samples collected according to P- and-CAM 318, showed no effects of light, reduced pressure, or humidity, but reduced recovery resulted from storing samples for 1 month and from not keeping samples refrigerated. The authors conclude that samples collected according to P-and-CAM 318 are unstable over time, and that stability can be increased by refrigeration. They recommend the use of P-and-CAM 354 for collection and analysis of formaldehyde.