Page:Mind (Old Series) Volume 11.djvu/403

 402 L. T. STEVENS : Average of the Standard. total number of Error. reproductions. II. 800 798 - -3 per cent. 515 535 + 3-9 15 506 526 + 4-0 55 500 523 + 4-6 55 471 494 + 4-9 55 III. 960 960 o- 55 933 929 - -4 55 647 681 + 5-3 )5 615 645 + 4-9 )5 588 619 + 5-3 55 493 510 + 3-4 55 IV. 920 908 - 1-3 55 796 862 + 8-3 55 734 783 + 6-7 55 V. 1-115 1-072 - 3-9 5) 733 746 + 1-8 15 VI. 614 620 + i-o 55 VII. 632 657 + 4-0 55 626 635 + 1-4 55 552 591 + 7-0 55 The irregularities consist in reproducing accurately long, in shortening long, and in lengthening short intervals. It must be stated that, in the attempt to get pure results, all those experi- ments were excluded in which fatigue or inattention spon- taneously stated by the subject entered, but no others. The examination of such experiments, however, revealed the fact, that the effect of fatigue is to make the error for short intervals /;///.- instead of minus, and to increase the amount of variation made in the reproduction of long intervals ; and that individuals under experiment are apt, when inattentive, to shorten long and pro- long short intervals. Some of these experiments, at least, are, therefore, regarded by the author as those in which fatigue and diversion of attention were existing but not acknowledged condi- tions. The accurate reproduction of long intervals is regarded as the chance result of the mutual neutralisation of the two opposing factors, the natural tendency to slow, and the effect of inattention. The conclusion drawn in this paper, with reference to the direction of the variation made in estimating long and short inter- vals, is in direct opposition to that of previous investigators. Vierordt (1), receiving time-impressions from the beating of a metronome and after the lapse of a short while reproducing them,