May io, 1894] 



NATURE 



41 



land 2 of Table I., are entered in Fig. I. Thus 13S milli- 

 metres is the just-perceptible interval of the median man, and 

 1 1 8 that of the median female. In one sense, but only in an 

 imperfect one, the relative sensitivity of the two sexes is given 

 by these figures as being about 7 to 6. Much more has, how- 



observed values could be obtained by closer attention to the 

 second decimals, but such minuteness is uncalled for in a case 

 like this. It will be seen from columns B, C, and b, c, in 

 Table II., that the sums of the ol)served and calculated deciles 

 closely accord, and that the differences between the several 



Table I. 



Summary of Observations ^932 .Vales and 377 Females, showing the number in whom a just-perceptible feeling of doubleness was 

 given by the pressure of two points across the nape of the neck, and separated by the various intervals, as below. 



* These figures are protracted in both cases as go '5, inasmuch as the accordance of the two preceding and of the four subsequent entries make the 

 correction reasonable as well as convenient. 



ever, to be specified before the relation can be adequately ex- 

 pressed, because it is obvious from the diagram, that what is true 

 'for persons having medium sensitivity, is not true for those having 

 ihigh, and still less for those having low, sensitivity. We are, how- 



pairs of them, headed B — C and /' — c, are as nearly alike as we 

 have a right to expeC. The calculated deciles, and the curves 

 drawn through them, in Fig. 2, may therefore be accepted as a 

 just rendering of what is more roughly indicated by the obser- 

 vations in Table I. and Fig. I. In the following remarks 

 reference will be made almost exclusively to the calculated 

 values, but the results can and will usually be checked by 

 reference to the observed ones, with which they tally sufficiently 

 well. 



Table II. 



FE.MALES. 



100° 



jFlG. I. — Traces and deciles from observations. The dots refer to the ob- 

 served values as given in the 3rd and 6th lines of the table. They are 

 connected by straight lines. The figures are the values of the corre- 

 sponding deciles — that is, of the ordinates to the traces erected at each 

 successive tenth part of the base. 



iver, able to specify what is wanted very compendiously, because 

 30th of the traces conform fairly well to the law of frequency of 

 ;rror, at least between the limits of the ist and the gth decile. In 

 he case of nial -s, the median is taken at y^o millimetres, and 



Fig. 2. — Deciles and curves by calculation. 

 Data for males : median = i3'5o, // = 3"25 mm. 

 Dat.i for females : median = i2'oo, ^ ^ 3'7o mm. 



he (/ (= half the difference between the two quartiles, a v.ilue 

 vhich is identical with that of the ± probable error of a single 

 bservation) at 3'25 ; in females, the corresponding values are 

 2"00 and 370. A somewhat nearer approximation to the 



NO. 1280, VOL. 50] 



The average ratio between the sensitivities of the females and 

 males is the same as that between the sums (or means) of columns 

 C and c in Table II., namely as 125'5 Io loSi ; or to speak 

 more modestly, as no trust can be reposed on the minute pre- 



