Who can forget, even if he has only once seen it, this phantasmagoria of customs and physiognomies which developes itself before our eyes at every moment; here a gigantic Circassian, there a smaller sized Copt, with an arched nose; a Nubian, with his “violet ebony” colour, but with a pleasing figure, nose straight and small, thin lips, well arranged teeth; a Turk, with as white and transparent a skin as a man of the north; a Negro, with crisped hair, flat nose, prominent cheek-bones, thick lips, large and projecting teeth; a Fellah, with olive complexion; a Bedouin, almost as black as a Nubian, but tall, with aquiline nose, thin lips, and kingly bearing. We must not seek for a pure population in the streets of Paris, London, Marseilles, Trieste, or Constantinople: we only find in these capitals isolated facts, good specimens, perhaps, of different species, but lost in the multitude of hybrids. We can only study in these places _individuals_, not _species_. In those parts alone which we must make centres of observation, can we see the same man indefinitely multiplied among really primitive people, still free from intermixture, or with the least possible taint of the same. Then we must hasten to seize his general characteristics, and take both his physical and moral portrait. The physical portrait in particular comprises two series of _data_, features and colour. As to feature, photography is an unequalled resource, but it belongs to anthropological study to settle its application in a clear manner: we must always choose some individual presenting the usual type of the population in the midst of which he is found, rather than among the chiefs or nobles of the land. We must select this type in the prime of life, when the animal œconomy has arrived at its perfect development, and has not yet commenced to decay, and still shines in all the splendour of its reproductive force; this would be, for man, from twenty-two to twenty-seven years of age. For photographic portraits to be of real utility to anthropologists, they ought to represent the individual completely full face, or in profile; thus only can they be of use in measurements. For it is important not to confound anthropology with ethnology, as is done every day. They are two things entirely different. _Dressed-up_ portraits are the domain of the latter, the _natural_ history of man demands _always_ absolutely nude representations, and the best will be those which show us the individual with untouched beard and hair.