The force-displacement curves of crumbly particulates tested individually or in bulk are irregular and irreproducible. Their degree of jaggedness, however, determined in terms of the apparent fractal dimension, the mean magnitude of their Fourier power spectrum, or the standard deviation of their normalized dimensionless force-displacement curve, is both a consistent and characteristic parameter and a measure of the particles’ crumbliness, brittleness or crunchiness. Similarly, their stiffness can be assessed quantitatively as the fitted force value at a selected deformation level using a polynomial model. In many cases it is more convenient, for technical reasons, to test particulates in a single layer of particles or in bulk rather than individually. As a result of averaging, however, the mechanical signature of the assembly will be considerably smoothed. The jaggedness of the individual particles’ signatures however can still be assessed from that of their single layer assembly by recreating an ’original’ particle signature by multiplying the layer’s signature by the square root of the number of particles in the layer. The procedure is demonstrated with computer-generated random sequences where its applicability was anticipated in light of basic statistical theory, and with the experimentally determined signatures of two crunchy food particulates (cheese balls and croutons). It can be applied to other fragile particles whose strength is too low to be meaningfully determined by regular testing machines. Since the contribution of particles to their layer stiffness is additive, the individual particles’ stiffness could be estimated by simply dividing that of the layer by the number of particles in it. Theoretically, the procedure’s applicability can be impaired if the particles have a wide size range or are tested as a bed deep enough for cushioning to have an effect.