166 



SCIENCE 



[Vol. YI., Ko. 334. 



obtained were practically alike. In the case 

 of the national academicians, the gentleman 

 referred to as representing the type was not 

 one of the sitters at all. The explanation 

 of this peculiarit}^ (excluding all possibilit}^ of 

 mechanical error in photographing) seems to 

 be possible in two ways. It ma}^ be regarded 

 as an example of what has been spoken of as 

 ' prepotency,' which means that one set of 

 featm-es was so powerful and characteristic 

 as to outweigh the effect of all others not in 

 harmony with it. On the other hand, it may 

 be said that the peculiar face is reall}^ yqyj 

 nearly the arithmetical mean in point of ap- 

 pearance of its class. For example, if we 

 desired the average height of twenty-one men, 

 five of whom were five feet ten inches high; 

 five, five feet nine inches high ; one, five feet 

 eight inches ; five, five feet seven inches, and 

 five, five feet six inches, each, we might take 

 the man of five feet eight inches as the average 

 of the lot. This individual would then be the 

 mean of his class, just as the one face was 

 the mean of its class. If, now, we omit the 

 man of just five feet eight inches in height, the 

 average of the group is not altered in the least. 

 So, also, in omitting the one photograph, as 

 Mr. Galton observes, the type is not altered. 

 This leads probably to the latter explanation. 



The problem is reduced to its simplest terms 

 in the combination of but two faces. The 

 resultant here ought to be the exact mean be- 

 tween the two originals. If the composite 

 resembles one more than the other, we must 

 regard that one as the more powerful charac- 

 teristic face ; if it resembles each of the ori- 

 ginals, the}' probably resemble each other ; if 

 it diflfers from both, the}^ differ from each other. 

 That is, the mean of three inches and five 

 inches does not diflfer much from either, nor do 

 the}' differ much from each other : the mean of 

 one inch and seven inches is also four inches, 

 but the differences between them and the mean 

 are large. 



I have attempted to experiment with ordi- 

 nary photographs, and without the elaborate 

 apparatus of composite photography, and give 

 my results, in the hope of thus placing the 

 process within the reach of ever}- one. 



The most natural method of combining two 

 pictures is b}' means of the well-known Brewster 

 stereoscope. I tried this, and was surprised at 

 the splendid result. I went to work with all the 

 fervor of an original discoverer, but afterwards 

 found, on looking at Mr. Galton's article, that 

 he had done the same. I mention this, because 

 Mr. Galton himself, in the same place, ^ admits 



1 Inquiries into human faculty. 



an independent suggestion, both of this and of 

 composite photograph}', in the shape of a letter 

 from Mr. Austin of New Zealand, to Charles 

 Darwin. 



It was easy to arrange a device by which 

 one of the two photographs could be raised or 

 lowered, moved laterally, and also swung 

 around its centre so as to bring its image ex- 

 actly in correspondence with that of the other 

 photograph. The only requisites are that the 

 two pictures be approximately of the same size 

 and position. AVhen these requisites were not 

 satisfied, I found other methods of combining 

 them, as will appear below. With an ordinary 

 stereoscope and a family album, any one has 

 the means of an amusing and instructive study. 

 A few of the results gotten in this way may be 

 worth recording. 



As was said before, in uniting two pictures, 

 say of two sisters that are commonly consid- 

 ered to resemble each other, the composite is 

 very much like either. Some will call it more 

 like the one, some more like the other. By 

 alternately closing each eye, and then opening 

 both, the observation becomes more striking. 

 If the pictures represent persons who are total 

 strangers to each other, the result is often an 

 entirely different face from either. The effect 

 is peculiar of combining the photographs of 

 two persons of opposite sex. The male face 

 seems to predominate ; but this is probably 

 due to the influence of the beard, mustache, 

 etc. ; for in combining children of opposite sex, 

 or using photographs of adults with smooth 

 faces, the predominance disappears. Like the 

 beard, so, too, the hair, costume, etc., are apt 

 to lend an undesirable pecuharity to the com- 

 posite. This can be partly avoided by cutting 

 out the shape of the face proper in white paper, 

 and attaching it with a rubber band to the pho- 

 tograph, thus combining the features alone. 

 On the same principle, one can combine the 

 upper part of one face with the lower part of 

 another by covering with paper the suitable 

 parts of each. 



One can combine persons of different ages 

 with good result. A young lady of twenty, com- 

 bined with her mother of sixty, gives a lady of 

 about forty years of age. A still more striking 

 case is, when a girl and her grandmother give 

 as a composite a middle-aged woman much 

 more like her mother than hke the girl herself, 

 or her grandmother, although a family resem- 

 blance runs through the group. 



As Mr. Galton says, the effect is often to 

 idealize the faces. A composite of two photo- 

 graphs of the same person gives a composite 

 far better than any photograph actually taken 



