296 



NA TURE 



\yan. 24, 1889 



HUMAN VARIETY} 



IT would have been a pleasure to me in this address, 

 given at the conclusion of my office as your President, 

 to have cast a retrospect over the proceedings of our 

 Institute during the four years that I have had the honour 

 to hold it. But the subjects that have come before us 

 are so varied that it seemed difficult to briefly summarize 

 them in a manner that should not be too desultory. 



On the whole, I thought it might be more useful if I 

 kept to a branch of anthropometry with which many 

 inquiries have made me familiar, and took the opportunity 

 of urging certain views that seem to be worthy the 

 attention of anthropologists. 



Before entering upon these more solid topics, let me 

 mention that the laboratory of which I spoke in my last 

 address has been in work during the past year, and that 

 about 1200 persons have been already measured at it in 

 many ways, some more than once. I lay on the table a 

 duplicate of one of the forms of application to be 

 measured, and of one of the filled-up schedules. It will 

 be observed that I now have the impressions made in 

 printers' ink of the two thumbs of each person who is 

 measured, being desirous of investigating at leisure the 

 possibilities of employing that method for the purpose 

 of identification, not forgetting the success that attended 

 Sir W. Herschel's use of it in India, but conscious at the 

 same time of practical difficulties. There is no doubt 

 that the thumb or finger marks vary so much that a 

 glance suffices to distinguish half a dozen varieties, 

 while a minute investigation shows an extraordinary 

 difference in small, though perfectly distinct, peculiarities. 

 Neither is there any room for doubt that these peculiarities 

 are persistent throughout life ; nor, again, that so satis- 

 factory a method of raising a very strong presumption of 

 identity would be valuable in many cases. It will suffice 

 to quote the following. A newspaper was lately sent me 

 from the distant British settlement of North Borneo, 

 where, owing to the wide and rapid spread of information 

 nowadays, attention had been drawn to an account of a 

 lecture I gave on one of the Friday evenings last spring, 

 at the Royal Institution. It was on " Personal Descrip- 

 tion and Identification," and a writer in the British 

 North Borneo Herald commented upon the remarks 

 there made on finger imprints. He spoke of the great 

 difficulty of identifying coolies either by their photographs 

 or measurements, and that the question how this could 

 best be done would probably become important in the 

 early future of that country. I also am assured that the 

 difficulty of identifying pensioners and annuitants has led 

 to frequent fraud from personation, involving in the 

 aggregate a very large sum of money annually, as there 

 is good reason to believe. If finger imprints could be 

 practically brought into use, such frauds would be 

 extremely difficult. I am still unable to speak positively 

 as to the best way of making them, but the plan adopted 

 at the laboratory is as follows. A copper plate is smoothly 

 covered with a very thin layer of printers' ink, a printers' 

 roller being used, and the plate being cleaned every day. 

 When the layer is thin, no ink penetrates into the delicate 

 furrows of the skin, but the ridges only are inked, and 

 these leave their impression when the inked thumb is 

 pressed on paper. In this way a permanent mark is 

 registered. A little turpentine cleans the fingers effectu- 

 ally afterwards. But for purposes of identification a 

 simpler process is necessary, one by which a person 

 suspected of personation could furnish an imprint for 

 comparison with the registered mark without having 

 recourse to the troublesome paraphernalia of the printer. 

 Such a process is afforded by slightly smoking a piece of 

 sinooth metal or glass over the candle, pressing the 

 finger on it, and then making the imprint on a bit of 



' Address delivered at the anniversary meeting of the Anthropological 

 Institute, on Tuesday, January 22, by Mr. Francis Gallon, F.R.S., President. 



gummed paper that is slightly damped. The impression 

 is a particularly good one, and is sufficiently durable 

 for the purpose. The iron used for the ironing of clethes 

 is excellent for the purpose ; even a smooth penny can be 

 used. As for the gummed paper, luggage labels can be 

 used ; even the fringe to sheets of postage stamps is broad 

 enough to include as much of the impression as is 

 especially wanted — namely, where the whorl of ridges 

 takes its origin. 



I hope at some future time to recur to this subject. 



Correlation. — The measurements made at the laboratory 

 have already afforded data for determining the general 

 form of the relation that connects the measures of the 

 different bodily parts of the same person. We know in a 

 general way that a long arm or a long foot implies on the 

 whole a tall stature — ex pede Herculem ; and conversely 

 that a tall stature implies a long foot. But the question was 

 as to whether that reciprocal relation, or correlation as it 

 is commonly called, admitted of being precisely expressed. 

 Correlation is a very wide subject indeed. It exists wher- 

 ever the variations of two objects are in part due to 

 common causes ; but on this occasion I must only speak of 

 such correlations as have an anthropological interest. 

 The particular problem I first had in view was to ascer- 

 tain the practical limitations of the ingenious method of 

 anthropometric identification due to M. A. Bertillon, and 

 now in habitual use in the criminal administration of 

 France. As the lengths of the various limbs in the same 

 person are to some degree related together, it was of 

 interest to ascertain the extent to which they still admit 

 of being treated as independent. The first results of the 

 inquiry, which is not yet completed, have been to myself 

 a grateful surprise. Not only did it turn out that the 

 expression and the measure of correlation between any 

 two variables are exceedingly simple and definite, but it 

 became evident almost from the first that I had uncon- 

 sciously explored the very same ground before. No 

 sooner had I begun to tabulate the data than I saw that 

 they ran in just the same form as those that referred to 

 family likeness in stature, and which were submitted to 

 you two years ago. A very little reflection made it clear 

 that family likeness was nothing more than a particular 

 case of the wide subject of correlation, and that the whole 

 of the reasoning already bestowed upon the special case 

 of family likeness was equally applicable to correlation in 

 its most general aspect. 



It may be recollected that family likeness in any given 

 degree of kinship— say that between father and son — was 

 expressed by the fact that any peculiarity in the father 

 appears in the son, reduced on the average to just one- 

 third of its amount. Conversely, however paradoxical it 

 might at first sight appear, any peculiarity in a son 

 appears in the father, also reduced on the average to 

 one-third of its amount. The regression, as I called 

 it, from the stature of the known father to the average 

 son, or from the known son to the average father, 

 was here from i to I ; from the known brother to 

 the unknown brother it was 5 ; from uncle to nephew, 

 or from nephew to uncle, it was | ; and in kinship 

 so distant as to have insensible influence, it was from I 

 to o. Whether the peculiarity was large or small, these 

 ratios remained unaltered. The reason of all this was 

 thoroughly explained, and need not be repeated here. 

 Now the relation of head-length to head-breadth, whose 

 variations are on much the same scale, is of the same 

 kind as the above. They are akin to each other in the 

 same sense as kinsmen are. So it would be in the closer 

 relation between the lengths of the corresponding limbs, 

 left arm to right arm, left leg to right leg. The regression 

 would be strictly reciprocal in these cases. When, how- 

 ever, we compare limbs whose variations are on different 

 scales, these differences of scale have to be allowed for 

 before the regression can assume a reciprocal form. The 

 plan of making the requisite allowance is perfectly 



