PULSE DURING MENTAL AND PHYSICAL WORK. 229 



Probably the most insistent impression from a casual inspection of 

 table 41 will be the enormous variability of the pulse in the same indi- 

 vidual under what might appear to be identical conditions. In the 

 case of Subject II, for example, the normal rest-pulse has an average 

 difference in one case (December 5) of +34 and in the other case 

 (March 17) of 245. No changes due to the effect of alcohol exceed 

 this change on different normal days. It might seem that out of such 

 chaotic data nothing could be learned. But the data are not so chaotic 

 as they might seem at the first uncritical glance. If one refers to the 

 other average normal differences for Subject II (December 5 and 

 March 17) which are given in table 41, it will be noted that the rest- 

 pulse differences of +34 and 245 belong to quite different series of 

 experiments. The former corresponds to an experimental series which 

 included the relatively vigorous muscular activities which are involved 

 in rising from the steamer-chair, standing, and the double genuflections. 

 The latter corresponds to an experimental series in which there was a 

 minimum of physical activity. We would not deny that there is large, 

 even gross, variability in the pulse differences under what were intended 

 to be similar circumstances. It was something of a revelation to us 

 that apparently similar conditions could be so different. But the 

 accidental variations are only a small fraction of the apparent variations 

 which are really due to the differences in the experimental series. 



In view of these differences, it may be questioned if we are not com- 

 mitting a gross statistical blunder by combining into a single table 

 results which developed under such various conditions. In answer, 

 let us insist that, except in rare instances, normal and alcohol data both 

 appear for each set of conditions. Since data from each set of condi- 

 tions are obviously directly comparable with respect to the effect of 

 alcohol, when the different sets of conditions are added together the 

 alcohol differences will not thereby disappear or be quantitatively 

 changed. One will merely average the changes due to alcohol which 

 occurred in the rest-pulse of all the experimental series. We have a 

 right to assume that the effects, which result from differences in the 

 experimental series, will balance. Changes which are due to accidental 

 disturbances would also tend to balance the more completely the greater 

 the number of instances. The changes which represent the real ten- 

 dency of alcohol should therefore most clearly appear in the total 

 averages of all the results which were obtained under all the different 

 sets of homologous conditions. 



An inspection of the general averages of the differences given in the 

 extreme right-hand column of table 42 shows that in this group of 

 experimental circumstances, just as in the association experiments, 

 there was a gradual retardation of the pulse during the 3-hour session. 

 On normal days this retardation averaged greatest in the finger-move- 

 ment experiment, and least in the more violent muscular activities. 



