experiments. The sensory attention, in the course of comparison-experiments, adapts itself (stellt sich ein) to the normal time; deviations from the normal call up surprise or expectation. If the normal time be given to the reagent in varying order, the secondary impressions occur only with comparatively large time-differences; sensible discrimination decreases. These facts hold alike for the cases in which the compared intervals succeed one another immediately, and in which they are separated by a pause. They have been tested for intervals up to 2". For intervals under 0.3" the conditions of the sensory attention are different; the basis of comparison, however, is the same. §2. The rapidity of a rotating drum, on which syllables are printed at equal intervals, is estimated by the strain of the attention necessary for the cognition and pronunciation of the syllables. §3. If lines be drawn on the rotating drum, in place of the syllables, the rapidity of the movement of these lines is estimated by the strain of attention necessary to follow them all for a moment with the eyes. But this is, of course, not the sole basis of such estimation. §4. The psychophysic of attention is very incomplete. It seems certain that a tendency to automatic activity at definite intervals can be impressed on the sensory centres; and this helps to explain attentional adaptation. Perhaps expectation-strain implies the existence of an internally aroused strain-feeling. (2) The results of previous investigation. §5. Mach and Vierordt. Valuable in Mach's investigation are: the determination of the maximum of relative sensible discrimination (at 0.3"-0.4"); the emphasis laid on adaptation; the calling in of secondary sensations to explain the time-sense. Vierordt employed a reproduction-method; imagining that the mean variable error measured sensible discrimination, and that the reproduced time gave the "corresponding temporal sensation-content." Both assumptions are untenable. §6. The Leipzig investigations. That of Glass alone deserves notice. He found that large times are over-estimated (cf. Vierordt); and that there is a periodicity of the constant error. This last fact (which is quite clear only in the final series of experiments) is perhaps to be explained by the influence of respiration. §7. Münsterberg. He also emphasizes strain-sensations of expectation, and (for large times) respiration. But he recognizes the periodicity-law of Estel, Mehner, and Glass. Firstly, this law is not in the experimental results; secondly, what of periodicity there is, is not explicable by respiration-influence; thirdly, the mean variable error of Glass does not exhibit the periodicity of the constant error. And the contrast-phenomena, to which he appeals, he has described incorrectly. His employment of Mehner's remark, that an interval is shortened by weakness of the limiting stimulus, is invalid, because the fact is wrongly stated. §8. Lesser contributions. Exner; Buccola; Hall and Jastrow (whether
