case for enquiry, but no more.

We may fairly ask why negroes in Hayti speak French, and why Frenchmen speak a modification of Latin. We may also fairly ask why some of the languages commonly called Aryan differ from one another according to certain general rules; why two or more languages agree with one another in certain sounds, and yet differ from the rest of the Aryan family. Why do the Roman and the Gael use the sound k where the German uses the sound f, and where the Greek and the Welshman agree in using the sound p? Why do the Roman, the German, and the Gael agree in using the sound s where the Welshman and the Greek agree in using the sound h?

Yet

It would, I think, be less absurd to found an ethnological argument on these definite philological rules than to assume that every linguistic connexion, however remote, implies blood-relationship. this assumption is made by almost every philologist who touches upon ethnology, no less by those who declare most European nations to be of the Aryan 'stock' than by those who lose themselves in a search for the lost tribes of Israel.

6

But it has been my aim to avoid all assumptions, and to draw all my conclusions from evidence. I should be inconsistent with myself were I to assume that any philological laws are identical with any

if

any,

ethnological laws. But, precisely as I enquire what, is the ethnological significance of the fact that the English speak a language of which the grammar is to a great extent Teutonic, I enquire what, if any, is the ethnological significance of certain other philological facts and rules pointed out in the second chapter. The importance of these facts is in each case determined by independent evidence.

I may be permitted further to point out that, if the whole of the philological discussion were omitted from this essay, I should still have a mass of evidence hardly the less powerful in favour of my conclusions; and, yet further, should all the evidence be rejected by which I have attempted to arrive at a positive and definite result, there would still be enough to demonstrate that we Englishmen are not what we are commonly supposed to be.

In my anxiety to escape from a particular danger I have said more on this subject than I ought perhaps to have said in a preface; especially as I have made similar remarks elsewhere. I shall endeavour to atone for this offence by saying here as little as possible on the other branches of the enquiry, the treatment of which must speak for itself.

In dealing with the higher mental characteristics of the different peoples compared, and in attempting to classify them, I have been met by this very great

difficulty: it is nearly, if not quite, impossible for any one to make himself a competent judge of all intellectual manifestations; and I fear that anyone may be accused of presumption who enters upon the task which I have undertaken. I can only say that I am quite sensible of my own deficiencies; that I have done my best to draw a sketch which, though incomplete, I believe to be correct as far as it goes; and that in the fine arts I have trusted authors of different schools who agree on those points to which my attention has been directed, and have carefully avoided giving any criticisms of my own.

I have to express my obligations to Dr. Beddoe, of Clifton, who has most kindly allowed me to quote two unpublished papers which were of considerable service to me in the investigation of physical characteristics; to Dr. Barnard Davis, who took the trouble to send me the measurements of some skulls in his collection; to Mr. Newton, and to Mr. Vaux, of the British Museum, who, in their respective departments, gave me every possible information, and who made some very valuable suggestions; to Professor Longmore, and to Dr. de Chaumont, of the Army Medical School at Netley, whose assistance enabled me to measure some skulls in the Medical Staff Museum in the least possible time; and to Mr. C. Carter Blake, from

whom I have more than once obtained some very useful anatomical information.

I believe I have in all cases acknowledged my obligations to workers in the same field with myself; but in order that I may escape the charge of even an unintentional plagiarism, I append a list of works consulted. Some of these have been studied for years; others have only been looked into in search of evidence bearing upon particular points. But my itself can best tell what use I have made of my predecessors.

essay

I am responsible for the translations of quoted passages except in those instances in which the translator's name is mentioned.

I have for two reasons retained the old-fashioned spelling of some names found in classical authors and elsewhere. In the first place (to give an instance) the general reader is, even now, more accustomed to Celt than to Kelt, and likes it better; in the second place, I think he has good grounds for his liking. There are only two excuses for writing Kelt: one of which is that the Greek form of the word is in that way preserved, the other that Kelt suggests the truc pronunciation. But I cannot see why we need discard an old custom in order to adopt the Greek spelling of a word for which we are indebted at least as much to Latin as to Greek authors. And if

pronunciation is to be our guide, there is an end of all our old orthography. It is as easy to give a hard sound to the first letter of the word Celt as to give the sound off to the letters ough in the word cough.

A phonetic system of spelling may be desirable; but be that as it may, I doubt whether such a system is possible. By what system of spelling, for instance, could an Englishman unacquainted with any language but his own be taught to pronounce the German word ich, or the French word rue? And if he cannot be taught the true pronunciation of a French or German word, what guarantee have we that he can be taught the true pronunciation of a word which appears as a foreign name in ancient Greek or Latin authors?

To discard an old and adopt a new system of spelling on pretence of securing accurate pronunciation, is simply to mislead. No one who reads English history supposes Lewis to be an accurate representation of the French Louis, or would care to see it printed Looee. Nor does anyone suppose that an Englishman pronounces the words Cincius and Cicero exactly as a Roman pronounced them. But if we are not to translate French vowels into English vowels, why need we pretend to give a faithful rendering of Latin or of Greek consonants? And if we are to affect accuracy in the consonants, why not in the vowels