Strictly simultaneous eyelid conditioning in human Ss on the one hand and forward (ISI: 0.45 s), backward (ISI: -0.45 s), and pseudoconditioning on the other hand were compared in two different experimental situations: one-session experiment (100 reinforcements, 20 isolated CS randomly interspersed), and ten-sessions experiment (20 reinforcements, 4 isolated CS in each session), UCS was a light flash (0.5 J, 20 microseconds duration), CS a tone (1000 Hz, 50 ms, 73 dB). In both the experiments, strictly simultaneous conditioning was significantly different from pseudoconditioning, but significantly less effective than forward conditioning. In the ten-sessions experiment, it was also less effective than backward conditioning, while in the one-session experiment, it did not differ from backward conditioning. Statistical mixture decomposition method applied to the one-session- and the ten-sessions- experimental samples of individual acquisition curves resulted into the division of both samples into 4 subgroups according to the trend of acquisition curves. A significant positive association between the conditioning procedure and the distribution of individual curves into the four subgroups was found.