The squares that run diagonally from the top at the left, to the bottom at the right, contain the double events, and it is with these that we are now concerned. Are the entries in those squares larger or not than the randoms, calculated as above, viz. the values of 10 × 19, 68 × 61, 27 × 25, all divided by 105? The calculated Randoms are shown in the first line of Table XXIII., the third line gives the greatest feasible number of correspondences which would occur if the kinship were as close as possible, subject to the reservation explained in p. 127. As there shown, the _lower_ of the A and B values is taken in each case, for Arches, Loops, and Whorls respectively.
+----------------------------------------------+ | | A and B both being | | |----------------------------| | | Arches. | Loops. | Whorls. | |-----------------|---------|--------|---------| | Random | 1·7 | 37·6 | 6·2 | | Observed | 5·0 | 42·0 | 10·0 | | Utmost feasible | 10·0 | 61·0 | 25·0 | +----------------------------------------------+
In every instance, the Observed values are seen to exceed the Random.
Many other cases of this description were calculated, all yielding the same general result, but these results are not as satisfactory as can be wished, owing to their dilution by inappropriate cases, the A. L. W. system being somewhat artificial.
FIG. 24. The "C" set of standard patterns, for prints of the Right Hand.]
With the view of obtaining a more satisfactory result the patterns were subdivided under fifty-three heads, and an experiment was made with the fore, middle, and ring-fingers of 150 fraternal couplets (300 individuals and 900 digits) by Mr. F. Howard Collins, who kindly undertook the considerable labour of indexing and tabulating them.
The provisional list of standard patterns published in the _Phil. Trans._ was not appropriate for this purpose. It related chiefly to thumbs, and consequently omitted the tented arch; it also referred to the left hand, but in the following tabulations the right hand has been used; and its numbering is rather inconvenient. The present set of fifty-three patterns has faults, and cannot be considered in any way as final, but it was suitable for our purposes and may be convenient to others; as Mr. Collins worked wholly by it, it may be distinguished as the "C. set." The banded patterns, 24-31, are very rarely found on the fingers, but being common on the thumb, were retained, on the chance of our requiring the introduction of thumb patterns into the tabulations. The numerals refer to the patterns as seen in impressions of the _right hand_ only. [They would be equally true for the patterns as seen on the _fingers themselves_ of the left hand.] For impressions of the left hand the numerals up to 7 inclusive would be the same, but those of all the rest would be changed. These are arranged in couplets, the one member of the couplet being a reversed picture of the other, those in each couplet being distinguished by severally bearing an odd and an even number. Therefore, in impressions of the left hand, 8 would have to be changed into 9, and 9 into 8; 10 into 11, and 11 into 10; and so on, up to the end, viz. 52 and 53. The numeral 54 was used to express nondescript patterns.