June 8, 1888.] 



SCIENTIFIC NEAVS. 



545 



ot pajpers!, Hectiire^, etc* 



ROYAL INSTITUTION. 

 The Friday evening discourse on the 25th ult. was deli- 

 vered by Mr. Francis Gallon, who lectured on personal 

 identification and description. He said he greatly felt 

 the inadequacy of language to express from, when he was 

 dealing with hereditary likenesses, with grouping human 

 features into classes, and with many other topics. The 

 present lecture was the result of the considerable pains 

 that he had taken to overcome this difficulty. There 

 proved to be many alternative ways of doing it, but as 

 he had not yet satisfied himself as to the best, he would 

 avoid details, and confine himself to general remarks. It 

 was perfectly possible to define and to measure resem- 

 blance, taking the least discernible difference as the unit 

 for each degree of unlikeness. This simple principle 

 could be easily applied to outlines such as those of sil- 

 houettes. It was better at first to go even further in the 

 direction of simplicity, and to consider only that part of 

 the outline of the face which lay between the brow and 

 the parting of the lips. The least discernible difference 

 between two silhouettes might be taken as equal to the 

 one-hundredth part of an inch. Though an exceed- 

 ingly large number of profiles might be drawn that 

 differed by this small amount in some part of their 

 outline, the number that differed by ten times that 

 amount would be comparatively very small. We might 

 aim at producing a collection of standard portraits, drawn 

 with coarse outlines of one-tenth of an inch in width 

 and then reduced to a small scale, to serve as standards 

 of reference. No profile that fell wholly within the 

 two edges of the coarse outline could differ from its 

 centre by more than five grades of unlikeness. This 

 would be a useful and a first degree of approximation ; it 

 would not be a likeness in any other sense than that of 

 excluding the very large proportion of profiles that were 

 still more unlike than those that fell within the specified 

 limits. A mechanical apparatus, described further on, 

 afforded the means of rapidly finding the standard or 

 standards to which any given profile conformed. Indi- 

 viduals differed in a measurable manner in so many 

 respects that a person might be identified with consider- 

 able precision by a statement of his measures. The 

 lecturer explained the measurements that were most 

 useful for this purpose, and how registers of them could 

 most easily be searched in order to find whether those of 

 a particular person were contained in them or not. Dif- 

 ferences that might be measured, or otherwise clearly de- 

 fined, existed on a small as well as on a large scale. The 

 curious variety of imprints made by the inked finger- 

 tips admitted of being classed and catalogued. They 

 seemed to be singularly persistent, judging from four 

 specimens that were exhibited of the digit marks of Sir 

 W. Herschel, m.ade in the years i860, 1874, 1885, and 

 1888 respectively. Though there was a difference of 

 twenty-eight years between the dates of the first and 

 the last, no difference could be perceived between the 

 impressions. The forms of the spirlas remained the 

 same, not only in general character, but in minute and 

 measurable details, as in the distance from the centre of 

 the spiral and in the direction at which each new ridge 

 took its rise. Sir W. Herschel had made great use of 

 digit marks for purposes of legal attestation among 

 natives of India. Prisoners were now identified in 



France by measurements of their heads and limbs accord- 

 ing to the ingenious method of M. Alphonse Bertillon. 

 The measures of each prisoner were all entered on the 

 same cards, and the cards were classified according to 

 the successive measures they contained, just as words 

 were arranged in a dictionary, according to their succes- 

 sive letters. The classification did not take more note of 

 the measures than by placing each in the category of 

 large, medium, or small, as the case might be. Thus 

 one measurement gave rise to three possible groups, two 

 to nine, three to twenty-seven, and so on. The lecturer 

 exhibited the rough working model of an apparatus he 

 had contrived that could select by a single movement 

 those cards out of many hundreds whose measures cor- 

 responded within any desired limit with those of any 

 given person. It was free from the objections inherent 

 in all methods of hard-and-fast classification, such as that 

 of M. Bertillon, and could act on a large scale, and with 

 great rapidity and simplicity. The apparatus consisted 

 of a large number of strips of card, or of metal, pivotted 

 through one of their ends by a common axis, while their 

 other ends rested on a frame that turned about the same 

 axis. When the frame was raised all the cards were 

 lifted by it, when it was lowered the cards dropped each 

 independently of the rest by its own weight. The lower 

 edge of each card was variously notched to indicate the 

 measures of the person to whom it referred. The key 

 that made the selection was a board some 30 in. of clear 

 length, with wires stretched lengthways and supported 

 by several bridges. The positions of the wires were ad- 

 justed by the same scale as that by which the notches 

 were cut, and the wires were set to represent the 

 measures of the given person. When the board was set 

 crossways under the cards, and these were allowed to 

 fall, they were all checked in their descent by the wires, 

 except those few whose notches.corresponded with them. 

 The notches were made wide enough not only to admit 

 the wires but also to allow for some inaccuracy of 

 measurement and misfit. A key of the above length 

 would test 500 cards at once, and might be used in 

 quick succession with any number of sets of cards. The 

 profile of a face, he continued, could be measured with 

 as much precision on the sharp outline of a photograph. 

 The observed differences in feature were severally small, 

 but they were numerous and more independent of one 

 another than the lengths of the various limbs. The best 

 base from which to measure horizontally seemed to be 

 the tangent line that touched the convexity of the chin, 

 and the concavity between the brow and the nose. When 

 the chin was bearded, the position of the concavity must 

 be guessed. A good unit of horizontal scale was given 

 by the distance between the line just mentioned and one 

 drawn parallel to it that just touched the nose. It was 

 better to keep the unit of vertical scale separate, and to 

 use for it the distance between the pupil of the eye and the 

 parting of the lips, measured parallel to the above lines. 

 One of the objects the lecturer had in view in the pre- 

 sent inquiry was to discover measurable and independent 

 peculiarities that would assist in hereditary investiga- 

 tions. He had some hope that by noting many of these 

 it might be possible to trace in every person clear evi- 

 dence of his parentage and near kinship. 



ROYAL SOCIETY OF EDINBURGH. 

 On May 21st Dr. Archibald Geikieread a paper on "The 

 History of Volcanic Action During Tertiary Time in the 



