58o 



SCIENTIFIC NE^A^S. 



[June 22, 1888. 



lines of carriages, and a footway. After having com- 

 pleted one round of the tower the vertical ascent is equal 

 to about 14 feet. The clear opening of the swing bridge 

 is 91 feet, and the distance between the underside of 

 this bridge and the water is about 49 feet. It is pro- 

 posed to fit up the inside of the towers as offices or shops. 



WHAT IS A "GENERALISED" 

 PORTRAIT ? 



THIS question pertains to one of the most interesting 

 of the many important applications of photo- 

 graphy — an application which will doubtless lead to 

 results scarcely as yet anticipated. If we consider a 

 portrait, whether photographic or of other kind, if only 

 it be accurate, we must admit that it blends a great 

 variety of impressions. The photograph of John Styles 

 exhibits, in the first place, characteristics peculiar to 

 himself alone, and which serve to distinguish him from 

 all other human beings. But it further reveals features 

 or expressions common to John and to his brothers, his 

 father, his uncles, and, in short, to all members of the 

 Styles family, and which at once distinguish them as a 

 class from any member of the Nokes family over the 

 way. The ordinary phrase, " family likeness," recog- 

 nises the existence of such traits. 



Going a step further, we find that the portrait in ques- 

 tion shows points common to all the Styles, to all the 

 Nokes, and, generally speaking, to all Englishmen ; and 

 which differentiate them as a body from all Frenchmen, 

 Spaniards, or Italians. Here, again, we have common 

 observation with us, since we not unfrequently hear it 

 said that some person has a genuine English face, or, on 

 the contrary, that he has a foreign look. But the por- 

 trait may have still more to reveal. Suppose John 

 Styles aforesaid has for year's followed some occupation 

 or expression, associating chiefly with certain classes of 

 persons and being conversant mainly with some set of 

 ideas. Say, to " fix our ideas," that he is a member of 

 the Stock Exchange. Is it not possible that we may 

 find in him something common to stock or share brokers 

 in general, and not common to men engaged in other 

 callings ? 



Now, the idea occurred some years ago to Mr. Francis 

 Galton, F.R.S., a cousin of Charles Darwin, that if we 

 could throw together in some manner the portraits of all 

 the Stokes family, those features which are peculiar to 

 John Stokes would relatively disappear, and those which 

 are the common attributes of the Stokes would come 

 into prominence, thus yielding a generalised portrait of 

 the "Stokes" family. This suggestion has been carried 

 into successful operation on several occasions, and that 

 with the most remarkable results. Two very instructive 

 " generalised " portraits were produced at the joint 

 meeting of the British and the American Associations 

 for the Advancement of Science. As a great number of 

 scientific men were present, it was proposed to take 

 generalised portraits respectively of the physicists and 

 the naturalists, or, to speak more accurately, the 

 biologists. Each group included men of different nation- 

 alities — English, Irish, Scottish, Americans, Germans, 

 and Canadian French. But when the two portraits were 

 completed it was found that in each there was a charac- 

 teristic something not capable of being rendered in 

 words, which differentiated the one group from the other. 



In like manner it will be possible to determine the 

 type of different races and tribes whom explorers 



encounter in their travels, as a generalised portrait will 

 depict not the mere individual, but the race to which he 

 belongs. 



A very curious application of the same idea is the 

 production of what may be called — contradictory as the 

 words may sound — a generalised portrait of an indi- 

 vidual. It is well known that one and the same person, 

 especially if he has what is commonly called an expres- 

 sive countenance, may appear very different at difterent 

 times, according to his frame of mind, his state of 

 health, and other varying conditions. If, then, we have 

 a number of portraits of such person, each accurate tor 

 the moment when it was taken, we may obtain from 

 them a truly typical portrait. In this manner Mr. Galton 

 has photographed six antique medals of Alexander the 

 Great and five of Cleopatra, and has obtained images of 

 each finer and more characteristic than any of the 

 medals taken separately. 



But now comes the question. How are such generalised 

 photographs to be obtained ? To this end a variety of 

 methods have been proposed. The first step was the 

 use of the stereoscope. But in this manner only two 

 images were superimposed, which was quite insufficient 

 for producing a general type. 



Mr. Galton and Mr. Herbert Spencer had the idea of 

 reproducing portraits of the same size upon very thin 

 paper, or upon films, and then laying them upon each 

 other and examining them by transmitted light. Here, 

 also, only a very small number of proofs could be em- 

 ployed, and the image obtained was very confused, the 

 general traits being obscured as well as the more peculiar 

 features. On the contrary, what is wanted is that those 

 traits which give, so to speak, the key-note of the gene- 

 rality should be brought out the most strongly, whilst 

 the others should be less and less accentuated in propor- 

 tion as they belong more and more to the individual, and 

 should at last disappear completely when they are only 

 the transitory and accidental modification of the physiog- 

 nomy. 



This end may be reached by a photographic procedure 

 recently devised. If a portrait of an in dividual be 

 taken with an exposure which is quite insufficient, the 

 impression received on the sensitive plate is, of course, 

 so faint that it cannot be developed. If a second portrait 

 be then taken under exactly the same conditions, then a 

 third, a fourth, and so on, continuing the operation as 

 many times as necessary, the totality of the exposures 

 makes up the time necessary for a portrait, and this will 

 be the general result of all the superpositions. 



Let us suppose that the complete time for an exposure 

 is sixty seconds, and that we have to combine twenty 

 models, whether they are living persons, portraits pre- 

 viously taken, medals, etc. On exposing each for three 

 seconds, no photographic image is obtained, but by the 

 succession of the twenty objects we have the sixty 

 seconds necessary for the production of a proof to which 

 each will have contributed a twentieth share. Then all 

 the characteristic general traits, being repeated by the 

 twenty objects, are impressed upon the sensitive surface 

 at the same points, and come out with the greatest in- 

 tensity. In proportion as the traits are more peculiar 

 to single individuals, they come out less and less strongly, 

 and any individual deformity disappears altogether. 



A curious result is that the generalised portrait is 

 always more beautiful than any of the individuals who 

 have served for its production. There takes place a 

 kind of idealisation of the expression and the features. 



