In a series of experiments, we looked at a "cutoff" effect for the detection of eye irritation from neat vapors of homologous n-alkylbenzenes and 2-ketones. Stimuli comprised pentyl, hexyl, and heptyl benzene, 2-dodecanone, and 2-tridecanone, presented to each eye at 4 and 8 l/min for 6 sec, using a three-alternative forced-choice procedure against blanks. Detection probability corrected for chance (i.e., detectability) decreased with carbon chain length such that heptyl benzene and 2-tridecanone were virtually undetectable, irrespective of flow rate to the eye. Heating both stimuli sources to 37 degrees C (body temperature) from 23 degrees C (room temperature) increased vapor concentration by 5.0 and 6.9 times, respectively, for heptyl benzene and 2-tridecanone. Still, both chemicals failed to show increased detection for 13 of the 21 participants. In addition, plots of experimentally measured and calculated eye irritation thresholds as a function of carbon chain length for each series indicated that, based on the trend, the concentration of the two cutoff homologs at 37 degrees C should have been high enough to allow detection. Taken together, the results suggest that these cutoffs rest on limitations related to the dimension of the molecules rather than on limitations related to their vapor concentration. For example, the stimulus molecule could exceed the size that allows it to fit into the receptor pocket of a receptive protein. Plots of calculated molecular dimensions across homologous alkylbenzenes, from ethyl to dodecylbenzene, and across 2-ketones, from 2-octanone to 2-octadecanone, provided additional support to the above conclusion.