On inheritance of hair and eye colour / by John Brownlee.

  • Brownlee, John, 1868-1927.
Date:
[1913]
Catalogue details

Licence: In copyright

Credit: On inheritance of hair and eye colour / by John Brownlee. Source: Wellcome Collection.

Provider: This material has been provided by The University of Glasgow Library. The original may be consulted at The University of Glasgow Library.

    29/38 (page 11)
    Table VI.—Scarlet Fever, Belvidere Iloftpital. Number of Deaths. Theoretical. Actual Difference. Difference Squared Divided by Theoretical Numbers. 0 10-4 10 •4 •00 1 23-4 25 1-6 •11 2 268 23 2-8 •30 3 18-7 22 3-3 •59 4 98 9 •8 •07 6 4-0 2 2'0 1-00 6 7 1-3) 1 c •4; 1-3 I'OO 93-8 94 3 07 These figures show a very fair correspondence between fact and theory. Possibly the fit might be better or might be worse with larger numbers, for in the figures as given there is a correlation between high numbers of deaths or low numhers of deaths in the corresponding fifties or hundreds to be expected as the fevers vary somewhat in severity from period to period, but the numbers are not sufficient to determine the amount definitely. 20. We are now in a position to explain the proof of the chief theorem in probable error as applied to vital statistics. The problem is : if we have a population of AMndividuals consisting of s groups y-,, to find the standard deviation of the group yp. The chance of one individual being drawn from this group is evidently and likewise the chance of his not being drawn is 1 N If then m individuals have been selected by chance the proportional distributions will be represented, as has been seen before, by the terms of the expansion of ] \ ~ W/ 1 standard deviation of which is Now we do not know the ratio of y^ to N. All that is known is the ratio which the samples of these quantities bear to each other. We may, however, assume, subject to subsequent investigation, that these ratios are for practical purposes identical, keeping in mind that this at present is only an assumption. If then y'p denotes the actual number of yp found, we have the standard devi- ation of the error of yp since the total number of observations is m, repre- sented by m \ m or suppressing tlie accents by m