According to Professor Boyd Dawkins, the Round Barrow race must have

been Goidels, and not Wends, Finns, or Slaves, because the latter would

not have subsequently retreated eastward ‘against the current of the

Celtic, Belgian, and German invasions’;[1926] while the late Canon

Isaac Taylor[1927] affirmed that the skulls of the well-known ‘Sion

type’, which by some anthropologists are believed to have belonged to

the Celtic Helvetii, resembled those of the round barrows.

Now the view that the tall brachycephalic people of the round barrows

were the Belgae is so utterly absurd that it is difficult to conceive

how writers who posed as authorities on ethnology could ever have

entertained it.[1928] If some benighted classical scholar had ascribed

the Copernican system to Ptolemy, one may imagine how he would have

been derided by scientists; yet such a blunder would not have been

different in degree from that which Thurnam committed and Huxley

approved. For the Belgic invasion began, at the earliest, in the third,

and, as Professor Rhys himself maintains,[1929] in the second century

before the Christian era; and the first invaders of the Round Barrow

race landed in Britain, at the latest, about 1400 B.C.,[1930] and

probably several centuries earlier. The argument which Thurnam bases

upon the alleged similarity between Round Barrow skulls and some which

have been exhumed from French dolmens has no weight. To begin with, the

theory that any Celtic-speaking people invaded Gaul in the Neolithic

Age is contrary to historical and archaeological evidence;[1931] and,

assuming that they did, the resemblance between the skulls to which

Thurnam refers and most of those of the tall Round Barrow skeletons

is purely superficial. Any one may convince himself of this who will

take the trouble to compare the illustrations of Round Barrow skulls

in _Crania Britannica_ with those in _Crania Ethnica_; and Thurnam

himself in more than one passage[1932] admits, indeed emphasizes,

the distinction. Even Broca[1933] denied that there was any physical

affinity between the tall brachycephali of the round barrows and the

[so-called] ‘real Celts of Gaul’; and, as we shall see presently, by

the latter he simply meant the brachycephalic people, descended from

neolithic ancestors, that formed the substratum of the population

whom Caesar called Celtae. Similarly Dr. Beddoe truly says that the

[characteristic] Round Barrow skulls resemble those of Borreby in the

Danish island of Falster, rather than those of Broca’s Celtae.[1934] It

is true indeed, as we have seen, that some of the Round Barrow skulls

resemble some of the neolithic French skulls; but, speaking generally,

the former are far more rugged and in every way more strongly marked

than the latter.[1935]

More striking, however, than the contrast between the skulls of

the characteristic Round Barrow skeletons and those of the French

brachycephalic neolithic race is the discrepancy in stature. The

average height of the former was, as we have seen, on the lowest

computation, 5 feet 8⅖ inches; that of the latter was very little over

5 feet.[1936] Moreover, while the brachycephalic Finns and Danes and

the few modern brachycephalic inhabitants of England are generally tall

or moderately tall and fair, those of France and Central Europe are

generally not only short but dark.[1937]

The argument that since the Long Barrow skulls were pre-Aryan, those

of the round barrows must have been Celtic, begs the question. As we

shall see presently, there are other skulls in museums, which belong to

neither type, and which undoubtedly are Celtic. What reason is there to

deny that the earlier brachycephalic invaders who were buried in round

barrows may, as Mr. C. H. Read[1938] reasonably suggests, have been

pre-Aryan? The British Celts of the later Bronze Age were doubtless

cremated; and therefore their skulls are not forthcoming. And if the

resemblance between the cremation interments of the round barrows and

those described by Caesar proved that the former were all Celtic, it

would also prove that they were Greek![1939]

In answer to Professor Boyd Dawkins it may be said that if the tall

Round Barrow race were not Finns or Slaves, it does not follow that

they were Goidels. And supposing that they were Finns or Slaves, why

should it be necessary to assume that they ‘subsequently retreated

eastward against the current of the Celtic, Belgian, and German

invasions’? Or that they retreated eastward at all? The ‘Iberian’

immigrants certainly did not retreat ‘against the current’ of the Round

Barrow invaders: they retreated, if at all, to the remoter parts of

Britain. The argument that the Round Barrow skulls resemble those of

the Sion type is disposed of by merely comparing the measurements and

the illustrations of the two series. The Sion type, as Rolleston[1940]

says, ‘corresponds to many of our long-barrow skulls,’ and is not

brachycephalic but dolichocephalic or mesaticephalic:[1941] there

is no proof that it was that of the Helvetii;[1942] and, as I have

pointed out elsewhere,[1943] there is strong reason to believe that the

Helvetii did not appear in Switzerland before the Iron Age.

So much for the arguments which have been adduced in favour of the

popular theory. There are facts which absolutely disprove it. First,

there is no evidence that the brachycephalic people who built round

barrows ever reached Ireland, at least in appreciable numbers; for

not a single skull of the characteristic Round Barrow type has ever

been found there, and only four brachycephalic skulls which can be

referred to prehistoric times.[1944] Yet it is needless to say that

since a time long anterior to the Roman invasion of Britain Ireland has

been one of the principal abodes of the Goidelic stock. Secondly, it

is, as we have seen, in the highest degree probable, if not certain,

that the Round Barrow race first invaded Britain in the Neolithic

Age. Let us, however, for the sake of argument, accept Professor Boyd

Dawkins’s assumption that their advent synchronized with the beginning

of the British Bronze Age. Now, according to Professor Montelius, the

Bronze Age in this country began about 2000 B.C.; according to Sir

John Evans,[1945] six centuries later. It is impossible to fix with

certainty the date of the earliest Celtic invasion of Britain; but such

historical evidence as we possess points to the conclusion that it was

not earlier than the seventh century before the Christian era.[1946]

M. Salomon Reinach has argued that a Celtic-speaking people appeared

in North-Western Gaul in the ninth century,--the earliest date which

has ever been proposed by any scholar; but his view is based on the

mere conjecture that κασσίτερος, the Greek word for tin, which occurs

in Homer, is of Celtic derivation.[1947] M. d’Arbois de Jubainville,

indeed, who adopts this conjecture,[1948] supposes that the Celts

actually landed in Britain as early as the ninth century before Christ;

but even if we accept his chronology, we are confronted with the fact

that the very earliest date that has been assigned on historical or

linguistic grounds for the first Celtic invasion[1949] is four or five

centuries later than the latest, ten or eleven centuries later than

the earliest date which has been assigned by archaeologists for the

commencement of the Bronze Age in Britain. Yet anthropologists and

antiquaries will go on repeating the dogma that the builders of the

round barrows, who, at the latest, began to arrive in Britain at the

commencement of the Bronze Age, were Goidelic Celts. The moral is that

anthropologists and antiquaries would not be worse equipped if they

enlarged the sphere of their studies.

Again, the view that a Celtic-speaking people invaded Britain at the

close of the Neolithic or the beginning of the Bronze Age implies that

Celtic and Latin, the nearest of kin in the Aryan family of languages,

had become differentiated long before the Neolithic Age came to its

end. Would any philologist who knew the rudiments of archaeology

sanction a theory so preposterous?[1950]

The foregoing arguments apply equally to the short men whose remains

have been found in the greatest purity in North-Eastern Scotland. The

race to which they belonged began to arrive in Gaul very early in the

Neolithic Age:[1951] they themselves landed in Britain before its

close. Whoever they may have been, they were neither Goidels nor Belgae

nor Brythons of any tribe.

Finally, although I am aware that I am about to tread upon thorny

ground, I affirm that there is not the slightest reason to doubt that

the Celtic invaders of Britain, in so far as they were descended

from the Celtic-speaking people who conquered Gaul, were not a

brachycephalic but a dolichocephalic or mesaticephalic people. I have

already argued in favour of this thesis in a dissertation on ‘the

Ethnology of Gaul’,[1952] and I will now adduce fresh evidence in its

favour. But first let me make my meaning perfectly clear. I do not

mean that the Celtic invaders of Britain were all of the same type.

On the contrary, I assume that the dominant race had intermixed and

intermarried, before they embarked from the Continent, with descendants

of the neolithic stocks. I do not mean that even the invaders who

introduced the Celtic language into Gaul, even those who beat the

Romans on the Allia, were homogeneous. Dr. Beddoe, as I have remarked

elsewhere,[1953] warns us not to believe that there was ever a period

when, for example, all the Caledonians were red-haired. I only mean

that among the Celtic-speaking conquerors of Britain dolichocephaly, as

well as tallness and fairness, was a prevailing characteristic.

Thurnam[1954] asserted that ‘we may ask in vain for a series of ancient

dolichocephalic skulls which, on satisfactory archaeological grounds,

can be assigned to the immediately pre-Roman, and therefore to the

Celtic period, either in England or in France’. Let us consider England

first. Now it happens that the skulls of the ‘Late Celtic’ period, or

Early Iron Age, which have been found in this country are almost all

either dolichocephalic or mesaticephalic.[1955] Canon Greenwell,[1956]

it is true, explains this fact by the assumption that ‘the intruding

round-headed people ... were gradually absorbed by the earlier and

more numerous [Long Barrow] race’. ‘In this way,’ he says, ‘it appears

to me that we may account for the skull type of the Early Iron Age

without the necessity of requiring any immigration into Britain or its

conquest after the time of the presumed occupation by the bronze-using

round-headed people,’ &c. But that necessity is imperative. Had Canon

Greenwell momentarily forgotten his Caesar? The immigration of the

Belgae took place, at the earliest, in the third century B.C., many

centuries after the ‘occupation by the bronze-using round-headed

people’. It is true that some of the British skulls which belong to the

Late Celtic period are of the same type as those of the Long Barrow

race:[1957] but this only proves that the Long Barrow race survived;

and others are of a type which, as Rolleston says, is ‘entirely wanting

... in the series from the long barrows’.[1958] Unfortunately,

however, the Late Celtic skulls which have been found in Britain are

comparatively few;[1959] and hardly any of them can be assigned with

certainty to the Brythonic invaders.

In France, on the other hand, the skulls of the corresponding period

are very numerous; but few of them have been measured. Those few,

however, confirm my argument. They belonged with very few exceptions to

tall mesaticephalic or dolichocephalic men; and two of them may be seen

in Salles IX and X of the Musée de St. Germain, near Paris, the former

having been buried with his war-chariot, iron helmet, and long iron

sword. The mean index of twenty-seven adult male skulls of this type,

found in _tumuli_ of the Early Iron Age in the department of the Marne,

was 78·49; but Broca, who has described them, maintains that the index

of skulls of the purest ‘Kymric’ (or, to use the term which is now in

vogue, ‘Galatic’) type would be considerably lower; for, he argues, as

the Gauls of the Marne lived very near the frontier of the Celtae, they

must have intermarried with the brachycephalic people who formed the

great majority of that group of tribes.[1960]

Again, in a recent article on _tumuli_ of the Early Iron Age in

the department of the Côte-d’Or, Dr. Hamy points to the noteworthy

fact that two brachycephalic skulls, belonging to descendants of an

earlier race, were found ‘among the dolichocephali who predominated

in that population’;[1961] and in a paper which he has just published

on the earliest Gallic invaders of the Iron Age he shows that the

cephalic indices of the available skulls from the Châtillonnais and

the arrondissement of Beaune range between 73·1 and 76·59, while the

average stature was 1 metre 75·7, or just over 5 feet 9⅛ inches.[1962]

The prevalent view in this country is, I am aware, that the Celts were

a brachycephalic people; but it is begotten of sheer confusion of

thought. Professor Ripley[1963] remarks that ‘there is practically

to-day a complete unanimity of opinion among physical anthropologists,

that the term _Celt_, if used at all, belongs to the brachycephalic

darkish population of the Alpine highlands’; and he adds that the only

dissentient is M. G. de Lapouge.[1964] But Dr. Beddoe,[1965] whom he

counts among the professors of the orthodox faith, has emphatically

recorded his opinion that, at the time of the Roman conquest, the

Celtic-speaking people of Southern Britain ‘partook more of the tall

blond stock of Northern Europe than of the thick-set, broad-headed

dark stock which Broca has called Celtic’; and the ‘unanimity’ (which

is far from being ‘complete’) upon which Professor Ripley pins his

faith is due partly to misunderstanding or misinterpretation of Broca’s

famous essay, _Qu’est-ce que les Celtes_, partly to the desire of

establishing a uniform connotation, and partly to the fact that some

physical anthropologists have neglected to supplement their scientific

researches by the study of classical texts. Broca found the term

‘Celt’ used in a multiplicity of senses, and he attempted to put an

end to confusion by attaching to it one limited, conventional, and, as

we shall see, misleading signification. When, in the essay to which

I have just referred, he endeavoured to prove that the Celts were a

dark brachycephalic people, he expressly limited the term ‘Celts’ to

the population of that part of Gaul which, according to Caesar,[1966]

was inhabited by ‘a people who call themselves Celts and whom we

[the Romans] call Gauls’. ‘There is no proof,’ he insists, ‘that the

existence in the British Isles of a people bearing the name of Celts

has ever been authoritatively affirmed’:[1967] according to him, the

invaders of Britain who spoke the so-called Celtic languages were

the Belgae,[1968] for he knew nothing about Goidels or pre-Belgic

Brythons; and, although he allowed himself to be persuaded that the

tall Round Barrow race spoke Celtic, he denied ‘that there is any

other affinity except that of language between the brachycephali of

the round barrows and the real Celts of Gaul’.[1969] When he insisted

that ‘the Celts’ were a dark brachycephalic people, he did not mean

that darkness and brachycephaly were characteristic of the conquerors

who introduced the Celtic language into Gaul: he meant that they were

characteristic of the great mass of the mixed population whom Caesar

called _Celtae_,[1970] who were in the main descended from neolithic

invaders, and whose uppermost stratum, so to speak, consisted of

invaders whom Broca, speaking as a physical anthropologist rather

than a philologist, called ‘Kimris’.[1971] That the name _Celtae_ did

not belong to the people of Gaul until it was introduced by these

Celtic-speaking ‘Kimris’ is evident from the fact that it belongs to

the Celtic tongue:[1972] in other words, the Celts, anthropologically

speaking, were originally identical with the invaders who introduced

the Celtic language first into Germany and then into Gaul.[1973] These

invaders were tall and mesaticephalic or dolichocephalic; and the

Celtic-speaking conquerors of Britain belonged to the same stock.

‘The radical errors in Broca’s definition of the “Celts of history”

[so I wrote some years ago[1974]] are these:--first, he calmly assumes

that no classical writer’s testimony, except Caesar’s, is of any value;

and secondly, he fails to see that Caesar, by saying that the people

who called themselves “Celts” were called by the Romans “Gauls”, makes

it as clear as noon-day that for him and for his countrymen, as for

Polybius and Pausanias, the words “Celt” and “Gaul” were synonymous.

Broca admits that the older population of Gallia Celtica was conquered

by men of the same race as the Gauls or Celts who captured Rome.

Therefore it is absolutely certain that the Celtae of Transalpine Gaul

were called after their conquerors. The truth is that Broca, while

he aimed at putting an end to confusion, only made confusion worse

confounded. Moreover, throughout his discussion, _he simply ignores the

Helvetii, who, according to Caesar, were included among the Celtae_.’

Since the foregoing paragraph was written, I have lighted upon a

passage[1975] in which Broca himself justifies my argument and uses

the word ‘Celt’ in the sense which I attach to it. The Celtae of Gaul,

he remarks, ‘were already mixed before the arrival of the Kimris [or

Gallo-Brythonic invaders], since the name [Celtae] under which they

appeared for the first time in history had been imposed upon them

by the conquering race of the Celts properly so called, which, like

the Kimris and the Germans, came from the east, and, like them, was

dolichocephalic.’[1976]

Professor Ripley appeals to the German ethnologist, Johannes

Ranke,[1977] whose arguments, he insists, are ‘decisive’. But any one

who will take the trouble to read the chapter which Ranke devotes

to the Celts will see that his argument does not support Professor

Ripley’s contention. Virchow, he reminds us, has pointed out that

wherever the Celts are known to have penetrated dark peoples are now

to be found. But, as he fully admits, Virchow himself said, ‘I am not

on that account inclined to assume that the original Celts were ...

dark,’ and reminded his readers that the ancient writers described

the Celts as fair. Ranke points out, further, that wherever the Celts

originally dwelled in Central Europe we now find the people not only

dark but also brachycephalic; but at the same time he warns us to bear

in mind that in certain Celtic districts of Britain dolichocephaly is

unmistakable, and that there is evidence that on the Continent _the

Celtic invaders found a dark brachycephalic people in possession_. In

other words, Ranke does not commit himself to any theory as to the

physical characters of the Celts properly so called,--the invaders who

introduced the Celtic dialects into Germany, Gaul, Britain, and other

countries which they subdued. The reader will also bear in mind that

the writers who identify the tall brachycephalic Round Barrow race with

the Goidelic Celts unanimously maintain that they were fair.

That the Celtic-speaking invaders of Gaul and Britain were commonly

dolichocephalic or mesaticephalic is not only attested by the skulls

of warriors of the Iron Age, but is either attested or at least

not disproved by the results of modern observations of existing

Celtic-speaking peoples[1978] and of the country which was formerly

inhabited by the Gallic Belgae.[1979] When Sergi[1980] tells us that

the Gauls who captured Rome were ‘composed of brunet Celts and blond

Teutons’, he makes an assertion which, as it is absolutely unsupported

by any evidence, calls for no refutation; and it would be useless to

ask him who were the ‘blond Teutons’ who were the ancestors of the

red-haired Gauls of the Perthshire Highlands.[1981] As Dr. Beddoe[1982]

puts it, the Gauls of Scotland are probably descended from ‘Iberians’

crossed with ‘a long-faced, harsh-featured, red-haired race, who

contributed the language and much of the character’.[1983]

3. The late Mr. Charles Elton,[1984] referring to Professor A. H.

Sayce’s _Science of Language_,[1985] affirmed that ‘a Finnish idiom has

been traced in several of the British languages’, and inferred that the

tall builders of the round barrows were Finns. The idiom in question

may, for aught that I know, have been traced by some philologist who

had determined to find it, but not by Professor Sayce nor by any one

to whom Professor Sayce refers. Mr. Elton’s argument is as obsolete as

that which Professor Rhys founded upon his imaginary tracing of Basque

in the language of the Picts.

4. Much may be said for the theory of the late Professor Rolleston,

that the tall people of the round barrows came from Denmark or some

of the adjoining islands, if it be duly modified. On the coast near

Flamborough Head are remains of earthworks, which, as has been

demonstrated by General Pitt-Rivers, who excavated them, were erected

by invaders fighting their way inland; and, as he remarks, ‘it is

unlikely that any but Northmen should have landed in this spot.’[1986]

Thurnam himself admits that there is ‘a great resemblance’ between

the characteristic Round Barrow skulls and those from ‘the Giants’

Chamber at Borreby [in the island of Falster], and from other

Scandinavian megalithic tombs’;[1987] and his testimony is confirmed by

Rolleston[1988] and Dr. Beddoe.[1989]

Dr. A. H. Keane[1990] argues, in opposition to Rolleston’s view,

that if any of the Round Barrow invaders had come from Scandinavia,

‘they must have spoken some Low German dialect, of which there are no

clear traces in the tribal and place-names of the Bronze Age.’ The

answer is, first, that, as Mr. C. H. Read[1991] suggests, they may

have spoken not a Low German but a pre-Aryan dialect; and, secondly,

that we know absolutely nothing about either the tribal or the

place-names of Britain in the Bronze Age. Assuming that Low German

tribal or place-names existed in Britain before the Celtic invasion,

they would for the most part have been superseded by Celtic names,

just as the Celtic invaders of Gaul generally substituted their own

tribal and place-names for those of their predecessors, and just as in

certain parts of Scotland Celtic names of rivers gave place to Norse

names.[1992]

5. Messrs. J. Gray and J. F. Tocher infer from their observations of

the physical characteristics of the population of West Aberdeenshire

that ‘a tall, broad-headed, dark-haired, light-eyed people’, whom they

regard as ‘the descendants of the men of the Bronze Age’, formerly

inhabited Aberdeenshire, but were driven inland by later blond

immigrants, who were shorter and had narrower heads, and whom they

identify with North Germans.[1993] The resemblance of the tall dark

people to modern Dalmatians[1994] is, they say, ‘significant when taken

in conjunction with the fact that bronze first came into the British

Isles from South-East Europe.’

‘The fact!’ But is it the fact? Archaeology has certainly shown that

Britain, in the Bronze Age, was commercially connected with Northern

France, which, as Mr. C. H. Read[1995] says, was ‘supplied to a

certain extent from Italy’. But no archaeologist supposes that bronze

was carried all the way from Italy, still less from Dalmatia, into

Britain or even into Northern France by Italians or Dalmatians. It came

through the methods of primitive commerce. Moreover, as we have already

seen,[1996] ‘the men of the Bronze Age,’ by whom Messrs. Gray and

Tocher mean the tall brachycephalic people of the round barrows, were

still in their Stone Age when they began to invade Britain. A direct

immigration from the coasts of the Adriatic into West Aberdeenshire

or even Southern Britain is inconceivable; and if it had taken place

gradually across the Continent, we should find that the immigrants

had left traces of their presence on the way, which is not the case.

Notwithstanding the thoroughness with which Messrs. Gray and Tocher

conducted their investigation, I fear that it throws no new light upon

the ethnology of Ancient Britain. After the successive invasions and

immigrations, the internal migrations, and the intermarriages of 3,000

years, it is utterly impossible to establish by dint of even the most

elaborate census of a living population the fact that the people of

the Bronze Age even in West Aberdeenshire were ‘tall, broad-headed,

dark-haired, and light-eyed’; and if they were, why only in West

Aberdeenshire?

6. The Honourable John Abercromby maintains that the brachycephalic

invaders, or some of them, came at the beginning of the Bronze Age

or in the period of transition between the Neolithic Age and the

Bronze Age from the neighbourhood of the middle Rhine or from some

intermediate district between it and Britain.[1997] Remarking[1998]

that ‘the recorded finds of the last hundred years are sufficient to

establish the fact that the beaker [or drinking-cup] is the oldest form

of fictilia in the Bronze Age of this country’, he argues that the

immigrants who introduced the oldest drinking-cups of the kind which

Thurnam designated as ‘type β’ must have belonged to a tribe who at

one time lived in the valley of the Rhine, because between British and

Rhenish specimens of this type ‘there is a substantial agreement’ both

in form and ornament, which ‘seems too great to be the result of pure

accident’; and he points out[1999] that ‘the type exists not only in

the central Rhine, but also near its mouth’, though the intermediate

stages cannot be traced. The Rhenish cups belong to the Neolithic Age;

and it seems impossible to prove that the earliest British examples

were not made before any objects of bronze were manufactured in or

introduced into Britain:[2000] but Mr. Abercromby has certainly

established a very strong probability in favour of the locality to

which he refers their origin.[2001]

The great mistake that has been made in discussing the question is

the not uncommon assumption that the brachycephalic immigrants who

buried their dead in round barrows arrived in Britain at one time and

came from one place. Some of them certainly appeared before the end

of the Neolithic Age: others may have introduced bronze implements or

ornaments; others doubtless came, in successive hordes, during the

course of the Bronze Age. Some of those who belonged to the Grenelle

race, who certainly came from Eastern Europe and possibly from

Asia,[2002] and whose centre of dispersion was the Alpine region,[2003]

may have started from Gaul;[2004] others could have traced their origin

to some Rhenish tribe; and I am inclined to believe that those who

belonged to the characteristic rugged Round Barrow type crossed over,

for the most part, from Denmark or the outlying islands. That the

first Celtic-speaking invaders landed in Britain before the end of

the Bronze Age I do not deny; and if they came from that part of Gaul

which was inhabited by the Celtae, I have no doubt that many of them

were brachycephalic. But it is nevertheless certain that among these

invaders the dominant element, who were Celtic in blood as well as

in speech, and whose physical type was that described by the ancient

writers, were not brachycephalic but mesaticephalic or dolichocephalic.

And if I am asked where the Celtic skulls of the later British Bronze

Age are to be found, I answer, Nowhere: they were reduced to ashes by

cremation.[2005]

It is interesting to find that, according to Huxley, of the skeletons

that were found in the famous Heathery Burn Cave, near Durham, which

was inhabited in the closing period of the Bronze Age, not one belonged

to either of the brachycephalic types, but all to ‘the same race of

rather small and lightly-made men with prominent superciliary ridges

and projecting nasal bones’[2006] which is represented by the river-bed

skulls of England and Ireland.[2007]

IX. THE CELTS

1. Little can be added to what has been said in the previous section

about the physical characteristics of the Celtic invaders of Britain.

Some Celtic scholars, as we shall presently see,[2008] deny that any

Goidels reached this country before the Roman conquest; but, assuming

that some did so, there is no reason to suppose that they differed

much physically from the Brythons. If Strabo[2009] was right in saying

that the Britons generally were less fair-haired than the Gauls, the

inference would seem to be that the Celtic invaders of Britain had

intermarried more freely than those of Gaul with the descendants of

the aborigines; nor would this inference be weakened by the fact that,

according to the same authority,[2010] they were conspicuously taller

than their Gallic kinsmen.[2011] I believe, however, that Strabo’s

statements were based upon nothing more than his own observation of

the few Britons whom he says that he himself saw in Rome, supplemented

perhaps by hearsay evidence derived from Roman soldiers or traders who

were not trained observers; and that his testimony is worth neither

more nor less than that of Lucan, who speaks of ‘the fair-haired

Britons’.[2012] Dr. Beddoe[2013] has concluded, from his observation

of the modern inhabitants of ‘those parts of Scotland and the north

of England where Kymric blood may well be supposed to remain in large

proportion,’ that the Belgae who invaded Britain as well as those

of Gaul were on the whole somewhat dark: but his arguments, which I

have examined fully elsewhere,[2014] do not prove that the dominant

Celts among the Belgae were dark, but simply that, before they invaded

Britain, they had become largely intermixed with an older dark

population, and that, since they reached this country, they and their

descendants have intermarried with people darker than themselves.[2015]

2. Professor Rhys has more than once changed his opinion about the

Celtic invaders of Britain since he began to handle the subject. In

the second edition of his _Lectures on Welsh Philology_[2016] he

argued that they were not ‘two distinct nationalities, speaking two

distinct languages’; in other words, he maintained that the Goidelic

and Brythonic dialects had been evolved within the British Isles

after the Celts had entered them. In the preface to _Celtic Britain_,

however, which was written in January, 1884, he recanted; and his old

view is now obsolete. For many years past he has maintained that the

earliest invaders were Goidels, or, as he now prefers to call them,

Celticans;[2017] and that the later comers were Brythons. But whereas

until a recent date he held that the only Brythonic invasion was that

of the Belgae, and that Pytheas, who visited Britain towards the end

of the fourth century B.C., ‘is not likely to have found any Brythons

here,’[2018] he now holds, or at all events held a few weeks before

the time when I am writing, that the first Brythonic invaders ‘appear

to have settled here before the middle of the fourth century B.C., for

Pytheas ... gives indirect evidence to their presence’.[2019] To this

view I hope he will firmly adhere. There is, indeed, no direct evidence

that any Brythonic immigrants landed in Britain before the Belgae. But

indirect evidence there is; and that of two kinds. The first has been

already noticed in the section on the Picts. There are good grounds for

believing that the authority whom Diodorus Siculus followed in his

notices of Britain was Pytheas.[2020] Diodorus speaks of the British

Isles as Πρετ(τ)ανικαὶ νῆσοι;[2021] and the _P_ in Πρετ(τ)ανικαί (if

that reading is certain), shows that Pytheas learned the word from lips

which spoke a Brythonic, or Gaulish dialect. M. d’Arbois de Jubainville

asserts that his informants were Gauls:[2022] but that is simply his

opinion; it is open to any one to argue that Pytheas probably learned

the name of the Britons as well as the facts which he reported about

them and their country in Britain, and not in Gaul. Be this, however,

as it may, it is, as we shall presently see, certain that during the

earlier period of the Roman occupation, the greater part of England

and a considerable part of Scotland were inhabited by Brythons; and,

as we shall also see, it is extremely improbable that they were all of

Belgic origin. The question of the chronological order of the various

Celtic invasions is, according to Professor Rhys,[2023] answered by

the present geographical distribution of the Celtic-speaking peoples

of the British Isles: ‘it may be regarded,’ he says, ‘as fairly

certain that those who are found driven furthest to the west were the

earliest comers.’ The argument might be sound enough (though the word

‘driven’ begs the question) if we were considering the British Isles

as a whole, and not merely Britain;[2024] and even those who maintain

that there were no people of Goidelic descent in Britain in the time

of Caesar could hardly answer Professor Rhys unless they assumed that

the Goidelic invaders of Ireland came from Spain, or that they dared

not risk a contest with the Southern Britons; for otherwise it is hard

to believe that they would not have directed their immigration towards

Britain, the nearer country.

Professor Rhys, in his _Celtic Britain_,[2025] endeavours to trace the

distribution of the Brythonic and Goidelic peoples, as he believes

it to have existed at the time of the coming of the Romans; and in

so doing he uses materials on which he founds another argument to

show that there were Goidels in Britain at that time. These materials

are Goidelic inscriptions which have been found in North Wales, in

Cornwall, and in Devonshire:[2026] but not one of them belongs to an

earlier date than the fifth century of our era. With the exception of

the districts in which they occur, of the greater part of Somersetshire

and Dorsetshire, of South Wales and the adjoining parts of England

which lie between the Severn and the Teme, and of Cumberland, part

of Westmorland, the Isle of Man, Dumfries, Kirkcudbright, Wigtown,

Ayrshire, Renfrew, and part of Lanarkshire, the professor regards

the whole of Britain south of the Firths of Clyde and Forth as

Brythonic; and he prints a list of proper names, most of which are

certainly Brythonic, in support of this conclusion.[2027] The northern

part of the island he divides, for reasons which have been already

examined, between Goidels and aboriginal tribes, whom he identifies

with the Picts properly so called.[2028] It will, however, of course

be understood that when he speaks of Goidelic and Brythonic tribes,

he means tribes who spoke the Goidelic and Brythonic dialects. The

former he regards as mingled largely with the aborigines, and the

latter with both Goidels and aborigines. But it is difficult to

understand how he has been able to maintain that the Dumnonii of

Cornwall and Devonshire were Goidels in the face of the fact that most

of the British emigrants who invaded Brittany came from the Cornish

peninsula,[2029] bringing the name _Dumnonii_ with them, and that he

himself formerly insisted that the Dumnonii who inhabited what is now

Renfrew and Ayrshire were Brythons.[2030] I say ‘formerly’, because

this is one of the many opinions which the professor has felt obliged

to discard: ‘the southern portion’ of the Scottish Dumnonii have just

been transformed by a stroke of the pen into ‘Goidels who adopted

Brythonic speech’.[2031] However, as M. d’Arbois de Jubainville says,

referring to the inscriptions upon which Professor Rhys relies, ‘To

conclude from the fact that five Goidels were buried, during the period

which elapsed from about 400 to about 700 A.D., in the territory

of the Dumnonii, that the entire population of that territory was

Goidelic seems extremely rash;’[2032] and, he asks,[2033] ‘if they

were Goidels, how came it that they brought a Brythonic dialect into

Brittany?’ Further, he asks why Professor Rhys maintains that the

Novantae of Galloway were Goidels when he admits that the Trinovantes

of Essex were Brythons;[2034] and the only answer which the professor

vouchsafes to this question is that the name _Novantae_ was ‘given

them probably by Brythons’.[2035] What are the grounds of his opinion,

he does not say. I may add that while he explains[2036] that ‘the

consonantal combination of _cs_ or _x_’ is Gaulish, that is to say,

Gallo-Brythonic, he says[2037] that it is ‘remarkable’ that ‘most of

the early names with _x_ belong to districts which have before been

pointed out as non-Brythonic’. When we look for these districts, we

find[2038] that they were those of the Taexali, the Vacomagi, the

Scottish Dumnonii, the Selgovae, and Cumberland. When we ask on what

grounds the inhabitants of these districts had been ‘pointed out as

non-Brythonic’, we find[2039] that the Taexali and the Vacomagi were

Pictish, that is to say ‘no doubt’ aboriginal; that the Dumnonii,

according to the professor himself,[2040] were ‘undoubtedly Brythons’,

and remained so until, discovering perhaps that he had inadvertently

given his case away, he changed them by his enchanter’s wand into

‘Goidels who adopted Brythonic speech’;[2041] and that the Selgovae are

asserted to have been, like the Novantae, ‘in a great measure ... most

likely a remnant of the aboriginial inhabitants.’[2042] Why? Because

they were afterwards included under the name _Atecotti_, which ‘appears

to have meant old or ancient’, and was ‘possibly given to them by the

Brythons’.[2043] Doubtless they were ‘in a great measure’ aboriginal,

as were doubtless all the British tribes; but seeing that _Uxellon_,

the name of a town in their country, is Gaulish, the natural conclusion

is that their Celtic masters were not Goidels but Brythons.

3. Professor Kuno Meyer holds that ‘no Gael ever set his foot on

British soil save on a vessel that had put out from Ireland’;[2044] and

his words are echoed by Dr. Macbain.[2045] Professor Meyer points out

that ‘we have the concurrent testimony of Irish and Welsh tradition

that from the second century of our era till the sixth a series of

partial conquests of Britain took place’.[2046] Dr. Beddoe[2047] has

indeed argued that it is extremely improbable that ‘the Romans would

have allowed the Irish Gael to acquire by violence possession of a

large portion of one of their provinces’; and Professor Meyer, who

admits the difficulty, says that he will not attempt to explain it

away. He might have noted that the author of the panegyric which was

addressed A.D. 296 to Constantius Chlorus[2048] expressly affirms that

such invasions did take place. Professor Meyer also points out that the

Gaelic inscriptions which have been found in Southern Britain belong

almost exclusively to South Wales, the quarter to which the invasions

may be assumed to have been directed, very few having come to light in

North Wales, Devonshire, and Cornwall.[2049] On the other hand, it will

be admitted that the record of these invasions is no proof that Goidels

had not settled in Britain in pre-Roman times.

4. M. d’Arbois de Jubainville holds, as we have already seen,[2050]

that Goidels, or rather a people who spoke ‘the Celtic dialect from

which Goidelic was evolved’,[2051] were masters of the British

Isles in the time of Pytheas, and that between his time and that of

Caesar Britain was conquered by the Cymric Brittones. So far he is

substantially in agreement with the view which, until a recent date,

commended itself to Professor Rhys,[2052] who, as the reader knows,

now believes that there were two successive Brythonic invasions.[2053]

The more important differences between the two scholars lie partly in

their views, which have been already examined, of the Pictish question;

partly in the fact that M. d’Arbois is unable to accept the evidence

which satisfies Professor Rhys that in Caesar’s time and later Goidelic

tribes still remained in Western and Northern Britain. He holds that

many of them had been driven by the Belgae into Ireland, and that in

Britain they only survived as a vanquished people who had been forced

to adopt the language of their Gaulish conquerors.[2054] I am inclined

to believe, from the analogy of Gaul,[2055] that in Caesar’s time

Goidelic was still spoken in remoter parts of the island.

5. Mr. Nicholson has recently attempted to prove that all his

predecessors are entirely mistaken even on the few points on which

they are agreed. According to him, the earliest Celtic invaders of the

British Isles were Brythons, whom, however, he prefers to call Kymri;

after them came a horde of Goidels; in the third century before Christ

the Picts, who were also Goidels, invaded Scotland; and finally came

the Belgae, who were Goidels too! The result was that ‘apparently

the great majority of the tribes inhabiting Roman Britain were

Goidels’,[2056] although ‘of the later Kymric recovery and victory in

Wales and some other parts there is no manner of doubt’.[2057] It will,

at all events, be admitted that a victory, however late, gained by a

small minority, was no mean achievement.

How does Mr. Nicholson set about proving this revolutionary theory?

He tells us that ‘on the map of Roman Britain’ he can only see

one ‘certainly Kymric geographical name’[2058]--Pennocrucium (now

Penkridge) in Staffordshire. The long lists of Cymric names which

have been drawn up by Professor Rhys, Dr. Whitley Stokes, M. d’Arbois

de Jubainville, and Dr. Macbain do not move him at all. When he is

confronted with geographical, tribal, or personal names belonging to

Pictland--names such as _Argentocoxos_, _Epidii_, _Gartnait_, the

_Ochil_ Hills, and the prefixes _aber_ and _pet_--he either ignores

them or, as his opponents would say, explains them away.[2059]

Professor Rhys’s list[2060] is disposed of with the same breezy

self-confidence. _Corstopiton_, _Epeiacon_, (Mons) _Graupius_,

_Leucopibia_, _Maponi_, _Parisi_, _Petuaria_, _Prasutagos_, _Rutupiae_,

_Toliapis_,--these names are either left out of account or explained

as Goidelic by the simple method of affirming or ‘suspecting’ that the

_p_ in each case is ‘Indo-European’.[2061] The reader will form his

own opinion if he can; only he will bear in mind that the weight of

authority is all on one side. When doctors disagree, the patient must

decide for himself which is the quack.

So much for the assertion that the Goidels, who, according to Professor

Kuno Meyer and Dr. Macbain, were non-existent in Britain at the time

when the Roman conquest began, formed then ‘the great majority’ of the

population. What is the evidence for the theory that they came later

than the Brythons?

There is no doubt that the Celts who first entered Gaul were

Goidels[2062] (assuming that Goidelic was then a distinct

dialect[2063]), and that the latest Celtic invaders of Gaul as of

Britain were Belgae.[2064] If the Belgae had been Goidels, we should

then have to admit that Gaul was invaded first by Goidels, then by

‘Cymri’, and finally by Goidels again. Is this likely? And is it not

likely that if Goidels were the first Celts who invaded Gaul, they were

also the first who invaded Britain?

Mr. Nicholson offers the following arguments in favour of his theory.

Remarking that the Menapii were a Belgic tribe, he says[2065] that ‘the

Isle of Man(n) [which Caesar calls _Mona_] is called _Monapia_ by Pliny

(iv, 103)’; and that the Gaelic dialect which is spoken in the island

is evidence that its inhabitants in Pliny’s time were Goidels.

Now I ask, first, is it certain that Pliny’s _Monapia_, rather than

Caesar’s _Mona_, was the name by which the Isle of Man was known to

its own inhabitants? Is it not probable that the name _Monapia,_

which is, at all events presumably, Brythonic, came to Pliny from a

Brythonic source?[2066] Secondly, assuming that the names _Monapia_ and

_Menapii_ are etymologically connected, does it necessarily follow that

_Monapia_ was a name peculiar to the Belgae, seeing that the tribal

name _Ceutrones_ occurs not only in Belgic Gaul but in the Alps?[2067]

Thirdly, is Mr. Nicholson prepared to prove that the Isle of Man was

not colonized by Goidels after it had received the name _Monapia_

from Brythons? Lastly, since Mr. Nicholson himself affirms[2068] that

although the name _Aremorici_ is ‘certainly Kymric’, it nevertheless

‘is no proof that the Aremoricans were Kymric’, why does he insist that

the fact, if it is a fact, that _Monapia_ was Goidelic proves that the

Belgae were Goidels?

Again, he says that the Parisi, who lived near the mouth of the

Humber, were Belgae,[2069] and he believes that ‘their name preserves

Indo-European _p_’.[2070] But Caesar did not include the Gallic Parisii

among the Belgae, and did include them among the Celtae.[2071] Mr.

Nicholson’s belief, that the _p_ in their name is Indo-European, is not

shared by any other Celtic scholar.

Thirdly, he argues that the Atrebates, who were certainly Belgae, were

Goidels; for, he says,[2072] ‘With one exception, no ogam-inscription

has ever been found in these isles outside territory which is known to

have been once in Goidelic occupation. The single exception is that

of the stone found at Calleva Atrebatum (Silchester).’ But, according

to Mr. Nicholson himself, ‘the great majority’ of the British tribes

were Goidelic: yet in only a small minority of their territories are

ogam inscriptions forthcoming; and that minority, with the possible

exception of the Atrebates, is in the west of England. What then is

proved by the solitary inscription at Silchester? The individual who

erected it was doubtless a Goidel:[2073] but if it is to be regarded

as a proof that the Atrebates were Goidels, then the existence of

synagogues in Great Britain proves the truth of that widespread

delusion which Professor Tylor[2074] has described as ‘abject

nonsense’,--the ‘Anglo-Israel theory’.[2075]

Fourthly, Mr. Nicholson remarks[2076] that between the Parisi and

the Iceni, the name of whose king, Prasutagus,[2077] he regards ‘as

containing Ind.-Eur. _p_’, while all other Celtic scholars regard it

as Brythonic, dwelled the Coritani.[2078] ‘From their position on the

coast,’ he says, ‘they should belong to the same Picto-Belgic family,

and I submit that their name is simply Qṛtanoi, Cruitni.’ In other

words, Mr. Nicholson submits that a single tribe, which he assumes

to have been Belgic, called itself by the same name which, on his

own showing,[2079] had been given to the entire population of Great

Britain[2080] long before the Belgae set foot in the land!

6. I have set down the gist of the linguistic evidence which has

been offered in support of the various theories about the Goidels

and the Brythons in order that the reader may be able to form an

independent judgement about its value. It goes without saying that on

any particular question of Celtic etymology no opinion except that

of a competent Celtic scholar is worth listening to: on most of the

questions that concern us competent Celtic scholars differ widely among

themselves: Professor Rhys differs from himself; and Mr. Nicholson,

whose competence I neither affirm nor deny, differs from everybody.

Even the lay reader who has studied the writings of Dr. Windisch,

of Professor Rhys, of Dr. Whitley Stokes, of Dr. Macbain, of Mr.

Nicholson, and of M. d’Arbois de Jubainville, and who has made much

use of Alfred Holder’s _Alt-celtischer Sprachschatz_ cannot but see

how few of the etymologies that relate to ethnology are to be accepted

as certain. It would of course be absurd to sneer at the services

which philology has rendered to ethnology and history; nevertheless

the fact remains that on almost all the fundamental questions of

Celtic ethnology the philologists agree to differ. And, at the risk

of appearing flippant, I cannot help saying that when I read some of

Mr. Nicholson’s pages, when I see how M. Salomon Reinach demonstrates,

with the approval of M. d’Arbois de Jubainville and of Professor Rhys,

who for once find themselves in agreement, that κασσίτερος, the Greek

word for tin, must be of Celtic derivation because the root _cassi-_

is found in numerous Celtic names,[2081] I ask myself whether some

future philologist will not adduce the similarity between _Tamesis_

and _Tamesi_, the name of a Mexican river, as a proof that the Celts

once colonized Central America; whether he will not compare the name of

Admiral Togo with that of the British prince, Togo-dumnos, and prove

that ‘the Japanese Nelson’ was of Celtic extraction.[2082]

7. Caesar, in a familiar passage, states that ‘the maritime districts

[of Britain are inhabited] by people who crossed over from Belgium to

plunder and attack [the aborigines], almost all of them being called

after the tribes from whom the invaders were an offshoot’.[2083] It

is, however, impossible to define the limits of the region which, in

Caesar’s time and during the period that elapsed between the date of

his departure and that of the Claudian conquest, was occupied by the

Belgae. The only tribal names that indicate their presence are those of

the Catuvellauni,[2084] who, about the commencement of the Christian

era, occupied a territory of uncertain area round Verulamium, or St.

Albans, which included Hertfordshire, Middlesex, and probably parts of

Bedfordshire, Buckinghamshire, and Northamptonshire; the Atrebates,

who possessed parts of Hampshire and Berkshire; and the Belgae, whose

chief towns, according to Ptolemy,[2085] were Aquae Calidae, or Bath,

and Venta, or Winchester.[2086] Caesar’s words would certainly lead us

to believe that the Cantii, the Trinovantes, and the Regni were also

Belgic peoples, although their names do not occur in the list of the

Belgic tribes of Gaul.[2087] Professor Rhys indeed affirmed in the

second edition of _Celtic Britain_[2088] that ‘there is no evidence

that the Cantii ... should be considered Belgic’; and this statement

is repeated in the edition which has recently been published: one

feels therefore that the evolution of the professor’s views is quite

normal when one reads in an intermediate volume, published two years

ago,[2089] that the earliest Belgic invaders of Britain were probably

the Brittani,[2090] and that the Brittani were probably the Cantii.

8. Finally, Dr. Macalister regards certain skeletons which have been

found in the War Ditches of Cambridgeshire below layers that contained

traces of late Roman occupation as Anglian[2091]; and it may be that

they testify to a pre-Roman immigration from Northern Germany.

X. CONCLUSION

For the sake of clearness I shall summarize the results which this

inquiry has attained. No human remains, except those of Bury St.

Edmunds and Cattedown, which can be certainly attributed to the

Quaternary Period have been found in Britain; but it is probable that

the earlier inhabitants belonged in part to the Neanderthal stock, and

that towards the close of the Palaeolithic Age they were joined by

immigrants akin to the Chancelade people of the Lozère valley. There

is no conclusive evidence that the earliest neolithic invaders found

this island inhabited; but it has not been demonstrated that even

here there was a ‘hiatus’ between the Old and the New Stone Age. The

source of the first neolithic influx was probably in France, in the

southern parts of which at all events the latest palaeolithic and the

earliest neolithic inhabitants were akin. The neolithic invaders who

built the long barrows of Southern Britain and the chambered cairns of

Scotland, and many of whom built round barrows also, were a branch of

the ‘Mediterranean’ race, and likewise came from France, perhaps in

some cases originally from the Spanish peninsula; but if they are to be

called ‘Iberian’ the term must be regarded as conventional. There is no

evidence that they were related more nearly to the Basques than to some

other branch of the Mediterranean stock.[2092] They certainly spoke

a non-Aryan language; and so probably did the earlier brachycephalic

invaders, of whom the first comers landed in Britain before the end

of the Neolithic Age. These invaders--the principal builders of the

round barrows and the short cists--continued to arrive in successive

hordes during the earlier part of the Bronze Age, some probably from

Gaul, some from the Low Countries and the valley of the Rhine, and

others, who settled in Yorkshire and Northumberland and perhaps in

Derbyshire, from Denmark or Danish islands and possibly also from the

Scandinavian peninsula. The brachycephalic Round Barrow skulls fall

under two different types. Some resemble those of the French Grenelle

race--in other words, the so-called Alpine race of Central Europe--and,

like them, belonged to individuals of low stature; although the

general superiority of the Bronze Age Britons in this respect is so

great as to preclude the supposition that men of the pure Grenelle

type invaded Britain in considerable numbers: others illustrate the

rugged and, in some cases, almost brutal type which Thurnam and

Rolleston have so forcibly described; and some of those of Yorkshire,

especially Rudstone, and Northumberland exhibit these characteristics

in such a degree that they may almost be grouped apart. The majority

would seem to show that people of the two types intermarried, as they

certainly did with the dolichocephalic neolithic population. The first

Celtic invaders were Goidels, who certainly reached Ireland in the

Bronze Age, and who may be supposed to have settled in Britain also

before the time of Pytheas. The first Brythonic immigrants probably

inaugurated the Iron Age in this country, and began to arrive a short

time before the visit of Pytheas. They were succeeded by the Belgae,

who, like them, came in successive hordes, the first probably in the

third century before Christ. The Belgae and the other Brythons spread

over the greater part of Southern Britain and many parts of Scotland.

Both they and the Goidels were doubtless mixed with people of the

‘Iberian’ and Grenelle races with whom they had intermarried before

they left the Continent; but the purer representatives of the two

Celtic stocks--the descendants of the invaders who had introduced the

Celtic languages into Gaul and of their continental kinsmen--belonged

to a type different from both of the Round Barrow types, being not

only tall and generally fair but dolichocephalic or mesaticephalic. A

people characterized by dolichocephaly and low stature, who apparently

were not descended from the Long Barrow race, but whose affinities are

doubtful, were settled in the Early Iron Age in East Yorkshire, and,

it would seem, nowhere else in Britain. The Picts of Romano-British

history were a medley of tribes, among whom Celts were, as everywhere,

predominant, but who probably included a greater proportion of the

descendants of the neolithic and other pre-Aryan peoples than any other

British group. It is possible that in the remoter parts of Pictland a

non-Aryan dialect was still spoken when the Romans invaded Britain;

but the pre-Aryan Picts as a whole had been Celticized, and the Celtic

language had prevailed, although it had been largely modified by the

speech with which it had come in contact. Everywhere in Britain the

pre-Roman stocks have, in greater or less proportions, survived.[2093]

Few Englishmen, Welshmen, or Scotsmen, if their pedigrees could be

traced back far enough, would not be found to count among their

ancestors men of the type who were buried in long barrows, sturdy

warriors of the Bronze Age, and Celts who fought against Caesar or were

subdued by Agricola.

* * * * *

The study of ethnology is as fascinating to its votaries, partly

by reason of its very difficulty, as the attempt to determine the

distances of the less remote stars must have been to Bessel, Henderson,

and Struve; but I can sympathize with those to whom, in both cases,

the quest of knowledge for its own sake appears equally unprofitable.

They may well ask the ethnologist why he does not proceed to deduce

from what he knows conclusions that would interest all students of

history and of human nature. ‘There are few fields,’ says Professor

Bury,[2094] ‘where more work is to be done or where labourers are more

needed than the Celtic civilisations of Western Europe. In tracing

from its origins the course of western history in the Middle Ages, we

are pulled up on the threshold by the uncertainties and obscurities

which brood over the Celtic world. And for the purpose of prosecuting

that most difficult of all inquiries, the ethnical problem, the part

played by race in the developement of peoples and the effects of race

blendings, it must be remembered that the Celtic world commands one of

the chief portals of ingress into that mysterious prae-Aryan foreworld,

from which it may well be that we modern Europeans have inherited

far more than we dream.’ But when we have entered the ‘prae-Aryan

foreworld’, how shall we map out its various provinces, and what clue

shall we have gained to the solution of ‘the ethnical problem’? That is

as complex as the problem, which theoretically may not be insoluble,

of forecasting remote meteorological as accurately as astronomical

phenomena; and its solution is more hopeless still. We want to know

what contributions the various British races which we have identified

made to the formation of the British character, which is so obviously

different from that of any other nation, and which is, so to speak, the

generalized manifestation of the characters of the English, Scottish,

and Welsh peoples, and, descending the scale, of the characters of

the inhabitants of every district, and finally of every man.[2095]

Again, we want to trace the manifold sources from which the ‘Celtic’

character, with the idiosyncrasies of which we are all more or less

familiar, is derived. But the Celtic character is not everywhere the

same. Study it in Wales, in Man, in the Scottish Highlands, in Ireland,

in Cornwall, in France, and you will find that while it is Celtic

everywhere, everywhere it is different;[2096] that everywhere it has

become what it is because it is compounded, in different degrees,

not only of Celtic, not only of pre-Celtic and pre-Aryan, but also

of post-Celtic elements. And all these elements have been modified

and moulded by different geographical and climatic influences and

by adventitious circumstances too numerous to be particularized and

too elusive to be estimated.[2097] Those who know Ireland well have

observed that the character of Anglo-Irishmen, whose blood is neither

more nor less Celtic than that of many Englishmen, has acquired a

quasi-Irish tinge, which is discernible in their children even when

they have been born and bred in England; and this sets us thinking,

though we think in vain. We all know the passage in which Mommsen

compares the Gauls to the modern Irish: the ethnologist knows enough

to see that it is as misleading as it is brilliant; but he knows too

little to attempt to rewrite it. Anthropologically speaking, the Gauls

(I use the word in its most comprehensive sense) were very different

from the modern and indeed from the ancient Irish; and if Mommsen’s

analogy were more than superficial, we should be forced to conclude

that the character of the Gauls, as it is revealed in ancient writings,

was that of the dominant Celts, perhaps mostly Gallo-Brythonic, alone;

and that the character of the Irish is simply that of Celts, mostly

Goidelic, who were once but have long ceased to be dominant. Who will

attempt to differentiate the respective shares of the pre-Aryan Long

Barrow race, of the few representatives of the pre-Aryan Grenelle race

who settled in this land, of the tall harsh-featured Round Barrow

people, of the Goidels, and finally of the Brythons in building up the

character which was to be further modified by the Roman, the Saxon, the

Dane, the Norman, the Fleming, and aliens of every nationality, who

each and all contributed something to a result which, influenced by

the Continent, by the Far West, and now by the Far East, is still in

process of evolution?

THE NAMES ΠΡΕΤΑΝΙΚΑΙ ΝΗΣΟΙ, _BRITANNI_ AND _BRITANNIA_

Πρεττανοί, which (written with a single τ) is supposed to represent the

Brythonic or the Gaulish equivalent of a Goidelic word _Qrtanoi_--the

assumed progenitor of the Irish _Cruthni_[2098]--is found in certain

manuscripts of Strabo[2099] instead of the more usual Βρεττανοί:

Diodorus Siculus[2100] (who derived part of his information about the

British Isles indirectly from Pytheas[2101]) Strabo, Ptolemy,[2102]

and Marcian,[2103] appear to have described them as Πρετανικαὶ νῆσοι,

for perhaps they were not responsible for the ττ which appears in

manuscripts; and Stephanus of Byzantium speaks of Πρετανίας and

Πρετανίδες.[2104] According to Professor Rhys[2105] and M. d’Arbois

de Jubainville,[2106] the form Βρεττανικαὶ (νῆσοι), which occurs in

most of the manuscripts, is to be accounted for by the fact that

_Brittani_, the Goidelic name of the Brythonic invaders of Britain,

which had no connexion with Πρετανοί, was eventually confounded with

it: ‘the confusion,’ says Professor Rhys, ‘is to be detected in the ττ

of Πρεττανική;’[2107] and he attributes it to scribes. The questions

connected with all these names are very difficult. The first puzzle

is this:--if, as the professor says,[2108] Πρετανικαὶ νῆσοι, ‘under

the influence of the name of the Brythons, Βρεττανοί, became at last

Βρεττανικαὶ νῆσοι, that is to say “Brythonic isles”,’ why did Diodorus

Siculus, Ptolemy, and Marcian persist in calling them Πρετανικαὶ νῆσοι?

Again, the professor’s views about the Brittani, who, he tells us,

called themselves Brittones, have lately undergone a sweeping change.

In 1902[2109] he regarded ‘the first of the Belgic peoples to cross

over to this country’ as an offshoot of ‘the Brittani or Brittones

whom Pliny seems to have found so called in the valley of the Somme’

(and whom, by the way, Pliny,[2110] whatever he may have found them

called, called neither Brittani nor Brittones, but Britanni); and he

considered that their name, ‘from being exclusively that of the first

settlers, came to be extended to the successive hordes, so that at the

last it actually denoted all the settlers here of Belgic descent.’

But the Britanni who are mentioned by Pliny were a Belgic tribe of

such small importance that Caesar either ignored or had never heard of

them; and, granting that some of them invaded Britain, of which there

is no evidence, it is to the last degree improbable that they rather

than the Belgae, of whom they would only have formed an item, should

have imposed their name upon the people of the whole island. In 1902

Professor Rhys maintained that the Brythonic invaders of Britain were

all Belgae. In 1904[2111] he distinguished the Belgae from the other

Brythons, maintained that they were the second group of Brythonic

invaders, and gave the name _Brittones_ not only to them but also to

their predecessors. Whether or not he still holds that these two groups

of Brittones derived their name from Pliny’s Britanni, I cannot say. If

so, it is somewhat puzzling that one tribe of the so-called Brittones

were the Belgae, who presumably called themselves after the Gallic

group of tribes which, as a whole, was designated by that name: the

professor’s theory would lead to the startling conclusion that while

a single horde of the second group of Brythonic invaders were called

_Belgae_ after the entire nation of which they were an offshoot, the

two Brythonic groups of invaders, Belgic and non-Belgic alike, were

conjointly called after the most obscure tribe of the second group! He

explains the name _Brittones_ as connected with the Welsh _breithyn_,

cloth, and concludes that ‘the word Brython and its congeners meant

a clothed or cloth-clad people’, and that ‘the race with which the

Brythons contrasted themselves to their own satisfaction, when they

began to give themselves that name, was probably some of the aboriginal

tribes whose home they invaded on the Continent’.[2112] But if so,

it seems wellnigh inexplicable that none of the Continental Belgae,

none of the other Gauls, were called either Brittones or Brittani, and

that only one petty Belgic tribe, which was unknown to or unnoticed by

Caesar, was even called Britanni. As Windisch says,[2113] Professor

Rhys’s etymology has to contend with serious difficulties; and it must,

I think, be admitted that if the Brythons were called either Brittani

or Brittones, the mention by Pliny of the Belgic Britanni throws no

light upon the origin of the name. It is perhaps conceivable that, as

Dr. Macbain[2114] has suggested, ‘the tribe on the Somme were some

returned emigrants from Britain.’

Britain, says Professor Rhys, is traceable to _Britannia_, and

_Britannia_ to _Britanni_,--‘the Latin name of the people’. He observes

that the Greek form of _Britanni_ was Βρεττανοί, and he adds that ‘the

practical identity between the Latin and Greek forms makes it probable

that it was from or through the Greeks of Marseilles that the Romans

first heard of these islands. This,’ he continues, ‘is not all, for the

Latin _Brittanni_, and especially the Greek Βρεττανοί, have their exact

counterpart in the Medieval Irish plurals Bretain, genitive Bretan,

which had at times to function as the name both of the Brythons and of

the island. It is to be noticed that neither Βρεττανοί or Britanni, nor

the Irish Bretain has anything corresponding to it in the dialects of

the Brythons themselves. From whom, then, did the Greeks hear the word

which served as the basis of their names for Britain and its people? It

cannot have been from the Brythonic peoples of the south-east of the

island, or any, perhaps, of the Gauls of the Continent: it was probably

from the natives of the south-west who brought their tin to market, and

in whose country the only Celtic speech in use was as yet Goidelic.

When, however, the Romans came to Britain they learnt the name which

the Brythons gave themselves in the south-east of the island, and this

was not Britanni, or Brettani, but Brĭttŏnes.’[2115] On the other

hand, Dr. Macbain[2116] suggests that the ‘Greek form Prettania [or

rather _Pretania_, the form which is assumed to have been derived from

_Priten_ or its older equivalent, and indirectly from _Qrtanoi_] gave

rise to the name Britain,--a bad Latin pronunciation’. Mr. Nicholson

objects that ‘in neither Greek nor Latin is _p_ known to pass into

_b_’.[2117] Is it possible that the Latin pronunciation, if it was bad,

may have been traceable to a bad Greek pronunciation, which gave rise

to Polybius’s[2118] Βρεττανικαί (νῆσοι), and which was itself due to a

defect not in pronunciation but in hearing?

THE BIRTHDAY OF RELIGION

Those who, like Professor Tylor, reject the theory that certain savage

tribes have no religious belief would probably accept the evidence

which Lord Avebury[2119] adduces in its favour: only they attach to

the word Religion a meaning different from his. Indeed he himself, in

one passage,[2120] uses the word in Professor Tylor’s sense; for he

remarks that ‘one of the lowest forms of religion is that presented

by the Australians, which consists of a mere unreasoning belief in

the existence of mysterious beings’; and he admits that religion,

in this sense, ‘is general to the human race.’[2121] Dr. Frazer,

however, would apparently refuse to make even this concession. He is,

or was, inclined to believe that ‘faith in magic is probably older

than a belief in spirits’;[2122] for ‘magic is nothing but a mistaken

application of the very simplest ... processes of the mind, namely,

the association of ideas by virtue of resemblance or contiguity’,

while ‘religion assumes the operation of conscious or personal agents,

superior to man, behind the visible screen of nature. Obviously,’ he

continues, ‘the conception of personal agents is more complex than a

simple recognition of the similarity or contiguity of ideas.’[2123] I

can only say that to me this is not obvious; the fancy of a primitive

savage that fire, running water--everything that moves--is alive, is

doubtless a less rudimentary mental act than the fear of a horse that

a traction-engine is a formidable monster, but the difference is only

one of degree.[2124] And Dr. Frazer’s definition of magic is singularly

narrow: magic and religion were rooted in the same soil; and their

branches intertwined.[2125]

To M. Salomon Reinach also ‘it appears evident that the true primitive

savage ... does not believe himself to be surrounded by spirits; he is

in the state which Herbert Spencer calls _passive atheism_.... The most

backward primitive savages whom we know are in the neolithic age....

The superstition (δεισιδαιμονία, dread of demons) which dominates their

whole existence ... is ... the outcome of a long evolution.’[2126] But

did not the process begin when the primitive savage, conscious of life,

fancied that sun and stars, flood and fire were also alive? And how can

M. Reinach make it ‘evident’ that there ever was a savage so primitive

that he had no such fancy? It is not true that the most backward

savages whom we know, or at least have known, are in the Neolithic Age.

The Tasmanians, a hundred years ago, were in their Palaeolithic Age,

but they believed themselves to be surrounded by spirits.[2127] Lord

Avebury indeed affirms that ‘some races entirely disbelieve in the

survival of the soul after the death of the body’;[2128] nevertheless,

if they believe in spiritual beings, they have the germ of religion.

For M. Reinach[2129] religion was born at the moment when man, finding

himself constrained to do what he feared might offend malignant

spirits, began to devise means of conciliating them. But may it not be

said with equal truth that the birthday of religion was when man began

to form the conception, on which religion, in the ordinary sense, is

based, that spiritual beings exist?

M. Reinach has recently pronounced that ‘fire-worship preceded the use

of fire, just as the worship of cereals preceded and prepared the way

for their cultivation’.[2130] One must infer that the ‘true primitive

savage’, who, according to M. Reinach, was in a state of ‘passive

atheism’, and therefore had not begun to worship fire, had not found

out how to produce it. If M. Reinach is right, the ‘passive atheist’

must have been primitive indeed.

Professor Robertson Smith held that ‘religion in the only true sense of

the word’ began ‘not with a vague fear of unknown powers, but with a

loving reverence for known gods who are knit to their worshippers by

strong bonds of kinship’.[2131] But it was in the ‘vague fear’ that the

‘loving reverence’ had its germ.

Dr. J. G. Frazer, in a recent article[2132], argues that the Australian

aborigines have no religion: but by religion he means ‘a propitiation

or conciliation of the higher powers’;[2133] and he admits that

some Australian tribes ‘have a notion of spiritual beings who can

help or injure them’.[2134] In other words, their belief fulfils

Professor Tylor’s ‘minimum definition of religion’; and Professor A.

C. Haddon justly remarks that ‘it is doubtful whether more than a few

anthropologists of repute would deny the term religion to the beliefs

and practices of the Arunta’ of Central Australia.[2135]

DUMBUCK, LANGBANK, DUNBUIE

I have said nothing in the first part of this book about the famous

‘crannogs’, or pile-dwellings (so called), which were discovered a few

years ago at Dumbuck and Langbank in the estuary of the Clyde, the

hill-fort of Dunbuie by Dumbarton Castle, and the remarkable objects

which they contained, because it is admitted that they belong to a

period several centuries later than the Roman conquest of Britain;

but, for a reason which will presently be apparent, they must not

be ignored. Everything worth reading that has been written upon the

subject is included in two recent books--_Archaeology and False

Antiquities_, in which Dr. Robert Munro contends that the disputed

objects are spurious, and the _Clyde Mystery_, in which Mr. Andrew

Lang endeavours to show that the difficulty of regarding them as

forgeries is at least as great as the difficulty of maintaining their

authenticity, and that, if they are genuine, they prove the survival of

ritual and magical ideas that must have belonged to the Stone Age.

It may be premised that Professor Boyd Dawkins,[2136] after a careful

examination of certain engraved oyster shells, which were a part of

the finds at Dunbuie, reported that he ‘had satisfied himself that

two of the shells were American blue points’, and, as he somewhat

superfluously added, ‘consequently of very modern date.’ Mr. Lang,

admitting this, suggests that, as Dunbuie was left unguarded for

several months, the shells were introduced by some local wag.[2137] At

the same time he argues that if the disputed objects were not genuine,

either the forger must have been a man of extraordinary erudition,

who had studied the archaeology of England, America, France, Germany,

Italy, Spain, Portugal, and Australia, or, by a coincidence which is

incredible, he produced objects which are found in all those countries.

I would suggest, however, that he may himself have been a person of

quite ordinary education, who was either employed by an archaeologist

with a peculiar sense of humour or learned what was necessary for his

purpose from some one more erudite than himself. Mr. Lang reminds us,

further, that if the disputed objects have been found in Britain only

in the basin of the Clyde, certain painted pebbles, similar to those of

Mas d’Azil, which have been found in Scottish brochs, are also unique

in Britain, and yet are disputed by nobody; and he might have said the

same of the strange objects of Mycenaean type which were found in a

barrow on Folkton Wold.[2138]

Dr. Munro, on the other hand, can see no resemblance whatever between

the disputed objects and the genuine productions of Australia or

certain other questionable ‘antiquities’ that recently startled the

explorers of a Portuguese dolmen.[2139] The reader, as Mr. Lang says,

must decide for himself; and I doubt whether he will see eye to eye

with Dr. Munro.

The doctor also insists that if the Scottish objects are survivals, ‘we

ought to find, at least somewhere in Britain, decided and undisputed

evidence of the existence of a phase of culture in the Stone Age in

which the prototypes ... would be the prevailing forms in general use.

But,’ he adds, ‘of such archaic remains there is not a vestige.’[2140]

No; but the earth has not yet given up all the vestiges of the Stone

Age: the first discovery of a Scottish interment of the Early Iron Age

has been made within the present century,[2141] and the doctor will

admit that it is probably not unique; besides, do not the brochs and

Folkton Wold suggest an answer to his argument?

Mr. Lang, concluding that at present the only position which the

impartial _savant_ can reasonably assume is a seat upon the proverbial

fence, admits that ‘the very strong point against authenticity is this:

_numbers_ of the disputed objects were found in sites of the early

_Iron Age_. Now,’ he continues, ‘such objects, save for a few examples,

are only known--and that in non-British lands--in _Neolithic_ sites.

The theory of survival may be thought not to cover the _number_ of the

disputed objects.’[2142] May it not also be said that as an ignorant or

sportive forger undoubtedly carved the oyster shells, so the disputed

objects may have been smuggled into the sites by a forger who was well

informed?[2143]

INHUMATION AND CREMATION

Dr. R. Munro[2144] says, on what authority I do not know, that the

object of cremation was ‘to liberate the spirit more quickly’. Is it

then to be concluded that in cases where inhumation and cremation were

practised simultaneously in the same barrow,[2145] it was intended that

certain spirits should be liberated quickly and others slowly?

Mr. W. C. Borlase[2146] remarks that ‘the transformation which would

have taken place when incineration was introduced ... would ... have

... been from a cult which was probably filthy and material to one

which was pure and spiritual’. We have seen that the ‘probably filthy’

and the ‘spiritual’ cult were practised simultaneously by the same

people; are we to assume that when inhumation was reintroduced in the

Early Iron Age filth and materialism were revived?

Professor Boyd Dawkins[2147] insists that cremation was introduced

into Britain by ‘the bronze-using Celtic tribes’; and Dr. Munro[2148]

apparently agrees with him. Putting aside the fact that most of the

tribes to which the professor refers were not Celtic,[2149] there is no

evidence that cremation was first introduced by bronze-using tribes:

if it was, the long barrows in which primary cremation interments have

been found must have been erected in the British Bronze Age! It may

or may not be true that, as Canon Greenwell suggests,[2150] some of

the Yorkshire round barrows were erected in the Stone Age; but at all

events they were later than the long barrows of the same county. Those

long barrows, according to Dr. Munro and Professor Boyd Dawkins, must

have been erected after a bronze-using people had introduced cremation

into Britain. How then would the professor and the doctor explain the

fact that in the round barrows of the Yorkshire Wolds there was a

reaction in favour of inhumation, seeing that Canon Greenwell[2151]

found in them 301 interments of unburnt and only 78 of burnt bones?

Dr. Munro[2152] remarks further that, ‘so far as available evidence

has been adduced, it would appear that the only sepulchral remains,

proved to have been older than the custom of cremation, are the

chambered cairns in the south-west of England. When, however, the

analogous cairns of Argyllshire, Caithness, and the Orkney Islands were

constructed, the religious wave had already enveloped Northern Britain.

Hence, though generally destitute of bronze relics, these structures

were generally contemporary with the Bronze Age burials elsewhere in

Britain.... The explanation ... is that in out-of-the-way localities

... the Stone Age civilisation lingered longer than in those on the

main routes of commercial intercourse.’ Certainly; but no sepulchral

remains in Britain are ‘_proved_ to have been older than the custom of

cremation’. Inhumation preceded cremation in Cornwall;[2153] but there

is no evidence that when inhumation was first practised there cremation

was not practised in other parts of Britain. Though cremation was very

rare in the chambered long barrows of Gloucestershire and Wiltshire, it

was not unknown:[2154] it was almost universal in the unchambered long

barrows of Yorkshire; and it cannot be proved that they were later than

the chambered long barrows of Wiltshire and Gloucestershire.[2155] The

chambered cairns of Scotland were not only ‘generally’ but absolutely

‘destitute of bronze relics’. Very likely some of them may have been

erected after the Bronze Age had commenced in Southern Britain; but

even this can hardly be proved. What has been proved is that even in

the Palaeolithic Age in the caves near Mentone cremation was already

practised side by side with inhumation.[2156]

[Since the rough draft of this note was written Professor Boyd

Dawkins[2157] has asserted that the long chambered barrow of Stoney

Littleton belonged to the Bronze Age, while he admits, apparently

because it did not contain cremated interments, that the long chambered

barrow of Rodmarton was neolithic.[2158]]

Professor Ridgeway’s views, which are expounded in his well-known

chapter, ‘Cremation, Inhumation, and the Soul,’ have been noticed in

the first part of this book. In regard to Western usage he blunders

in a way which makes me hesitate to accept his statements about

archaeological details that I have not myself studied. He says that

‘in Dorsetshire ... the extended position seems to be the prevalent

one’,[2159] a remark which I have already noted[2160] as an instance of

the danger of relying upon second-hand evidence; he implies that the

invaders who ‘conquered Dorset, Wiltshire, and Cornwall’ in the Bronze

Age were Belgae;[2161] and he states that ‘in France inhumation was

universal before the age of metal’,[2162] which, as I have shown,[2163]

is contrary to fact.

SEPULCHRAL POTTERY

Some antiquaries have maintained that drinking-cups, food-vessels,

incense-cups, and urns were not specially made for sepulchral purposes,

but were merely ordinary domestic vessels.[2164] On the other hand,

it has been urged that most of them were too fragile to stand rough

usage; that many are so contracted at the bottom that they would

have been ill adapted to serve as table or culinary ware; that the

food-vessels and the drinking-cups were too porous to hold fluid long,

while the shape of most of them would have made them inconvenient for

any ordinary purpose; and that all are wholly unlike the domestic

pottery which has actually been found in hut-circles, forts, barrows,

and the Heathery Burn Cave.[2165] Mr. J. R. Mortimer[2166] replies

that drinking-cups and food-vessels were quite strong enough for

domestic use; that ‘the form of the typical drinking-cup is well chosen

for the purpose its name implies, and most of the food-vessels are

the prototypes of our ... porringers, jars’, &c. It may be admitted

that some few food-vessels, for instance the one figured by Thurnam

in _Archaeologia_, xliii, 381, are, apart from their decoration,

not unlike domestic bowls; but what about incense-cups? The truth

perhaps lies between the opposing views; for drinking-cups and

food-vessels have been exhumed from pit-dwellings near Taplow;[2167]

and Pitt-Rivers,[2168] speaking of an urn which was found on the bottom

of the ditch of the camp in South Lodge Park in his estate, observes

that ‘it is more probable that the urn would be found in the ditch

thrown away as refuse if it was in ordinary use, than if it were only

fabricated for ceremonial purposes’. He remarks further that ‘the

large quantities of pottery of the same quality ... afterwards found

in different parts of the Camp, confirms this opinion [that sepulchral

pottery was used for domestic purposes], as it could not all have

been used for funeral urns’.[2169] Moreover, fragments of ornamental

pottery of the drinking-cup type were found by Pitt-Rivers in a pit

in Martin Down Camp.[2170] Still, the fact remains that only a very

small proportion of the pottery which is commonly called sepulchral has

been found outside sepulchres; and even it may have been intended for

sepulchral use.

STONEHENGE

Stonehenge has exercised the minds of many generations of antiquaries.

An exhaustive bibliography, filling 169 pages and containing the

titles of 947 books and articles, was published in the _Wiltshire

Archaeological and Natural History Magazine_ for 1901: but nearly all

the works therein enumerated are obsolete; and any one who wishes to

form an independent judgement will find all the necessary materials in

the volumes which will be referred to in this article.

I. Modern opinion has for some time been tending to the conclusion

that Stonehenge was erected, or at least began to be erected, in the

Bronze Age. Excavation has proved that it did not exist before the

use of copper or bronze, however uncommon it may have been, was known

in this country;[2171] and the arguments of Rickman,[2172] James

Fergusson,[2173] and others who contend for a Roman or post-Roman date

have been or can be demolished. To refute them in the text of this

article would be useless; for no competent archaeologist now regards

them as worth discussion.[2174]

Dr. Arthur Evans maintains that Stonehenge was built in the earlier

half of the third century before Christ, although some parts of it may

be of later date; that ‘sun worship was at most a secondary object in

its structure’; and that it was ‘one of a large series of primitive

religious monuments that grew out of purely sepulchral architecture’.

Let us first consider the question of date.

Dr. Evans has no difficulty in establishing, what has already been

demonstrated in Part I of this book,[2175] that ‘Stonehenge was

at least begun before the close of the Wiltshire “Round-Barrow”

Period’.[2176] At the same time he holds that ‘its foundation

belongs to the conclusion of this period’. He points out that ‘of 36

disk-shaped barrows [in the neighbourhood of Stonehenge] 35 contained

cremation interments.... The number of glass beads contained in these

barrows is also,’ he continues, ‘evidence of their comparatively late

date[2177].... The general inference which we draw from the intimate

structural connexion between Stonehenge and these disk-shaped barrows

is, that the great stone circles themselves were erected towards

the close of the Round-Barrow Period. The proportionately frequent

occurrence of gold relics in barrows in the immediate neighbourhood

of Stonehenge, 4 out of 5 such discoveries having been made within

half-a-mile of this monument, points in the same direction.’ And

if it should be argued that the barrows may have been built after

the erection of Stonehenge, his answer would be that ‘the barrows

themselves, with the exception of the two within its own area, are

disposed without any reference to Stonehenge, and do not in any way

cluster about it, as we might reasonably have expected them to do had

the bulk of them been reared after the Stone Circle’. Dr. Evans then

observes that an amber collar ‘found in one of the Lake barrows about

two miles from Stonehenge ... is of a form and arrangement identical

with the amber necklaces found in the great cemetery at Hallstatt,

and from the similar character of the boring of the beads must in all

probability have come from the same centre of manufacture’; and he

endeavours to show that we may infer from recent discoveries that ‘a

large proportion of the Hallstatt remains reach down to the period

between the approximate dates of 450-300 B.C.’[2178] On the other hand,

Late Celtic antiquities, which began to appear in Britain ‘at least

as early as the second century B.C.’, are absent from the barrows of

Wiltshire; and the latest date which can be assigned to these barrows

is about 250 B.C. Dr. Evans concludes that ‘we may approximately refer

the foundation of Stonehenge to the end of the fourth or beginning of

the third century’.[2179]

Some of these arguments do not appear to have much weight. Dr. Evans

himself admits that there is a great structural distinction between

Stonehenge and the disk-shaped barrows:[2180] in the latter the

surrounding ditch is _inside_ the bank; in the Stonehenge vallum it

is _outside_. Those barrows at all events in which chippings of the

stones were found were later than Stonehenge;[2181] and whatever

conclusion may be drawn from the arrangement of the barrows, their

number is so great as to suggest the inference that many of them were

erected there because Stonehenge was regarded as a holy place. In a

recent article,[2182] however, Dr. Evans has reinforced his argument,

pointing out that the circle called the Rollright Stones, which stands

on a hill overlooking the valley of the Warwickshire Stour, also stands

‘in immediate relation to a large group of [disk-shaped] sepulchral

barrows’, and giving additional and conclusive evidence as to the late

date of these particular monuments. He would not, however, I believe,

now assign to Stonehenge quite so early a date as the ‘beginning of the

third century’; for the commencement of our Early Iron Age is commonly

referred to about 400 B.C.[2183] Of the two barrows in which Hoare

found chippings of the Stonehenge stones one was bell-shaped,[2184] the

other belonged to the kind which he called ‘flat’,[2185] but which, as

Thurnam points out,[2186] is simply a variety of the bowl-barrow; but

Thurnam has given reasons for believing that many barrows of this form

may not have been earlier than disk-shaped mounds.[2187]

In 1901 excavations were made at Stonehenge, but only in ‘a fraction

of the whole site’, under the superintendence of Professor W.

Gowland.[2188] The principal objects discovered were chippings from the

‘sarsens’ and ‘blue-stones’; more than one hundred stone implements,

many of which were of flint, and had evidently been used for dressing

the softer stones of the monument, while others consisted of ‘the hard

quartzite variety of sarsen’; bones of domestic animals; ‘splinters of

antlers of deer’; ‘a portion of a large antler with its lowest tine

worn away,’ apparently from its having been used as a pick; and Roman

coins, which, however, were only found in the superficial layers.[2189]

‘The layers of the excavations,’ says Professor Gowland, ‘in which

the flint and stone tools were found was absolutely undisturbed

ground’;[2190] and the chippings were found as far down as the surface

of the bed rock. Only one trace of copper or bronze was visible, namely

a stain, described by the formula CuCO₃, on a sarsen block, seven feet

below the surface. The work of trimming the stones appears to have been

done with stone implements only. The copper stain, however, proves that

copper or bronze must have been in use at the time when the builders of

Stonehenge were at work. Professor Gowland[2191] affirms that the stain

‘can only have been produced by prolonged contact with some very small

object of copper or bronze or some material containing copper.... It

may perhaps have been an ornament, but cannot possibly have been an

implement.’ He argues, further, that ‘even if metal tools were of no

use for this particular work, it is difficult to believe that, if the

monument were of the Bronze Age, no bronze implement would have been

lost in the course of its erection’.[2192] I, on the other hand, would

remind him that no bronze has been found in the hut-circles of Dartmoor

or in various cemeteries which undoubtedly belonged to the Bronze

Age,[2193] and I suggest that if the workmen who built Stonehenge had

no use for bronze tools when they were building it, they were not more

likely to lose them on the site than the masons who built St. Paul’s

Cathedral to drop their table knives within the area of the churchyard.

However, Professor Gowland does not pin his faith upon this argument.

He points out[2194] that many of the flint implements which he

discovered at Stonehenge closely resemble those which were discovered

by Canon Greenwell at Grime’s Graves, and which were attributed by

him to the close of the Neolithic Age, or, at the latest, to a period

when bronze had not come into general use.[2195] But nobody who has

learned the ABC of archaeology needs to be told that stone implements

were used long after the introduction of bronze;[2196] and no expert

sees anything improbable in the theory that such tools were used in

constructing Stonehenge towards the end of the Bronze Age. Besides, as

Professor Gowland admits, Dr. Maskelyne has pointed out that ‘bronze

tools would not work sarsens’;[2197] assuming, then, that Stonehenge

was erected in the Bronze Age, how could the sarsens have been dressed

except with implements of stone? But the discovery upon which Professor

Gowland lays the most stress is that of the deer-horn pick. Similar

picks were found in large numbers at Grime’s Graves. The one which

Professor Gowland found, if it really was a pick, must have been used

for excavating the pits in which the stones of Stonehenge were erected;

and Professor Gowland argues that if bronze tools had been in use at

the time, ‘it would seem not unreasonable to assume that they would

have been employed, as they would have been so much more effective for

such work than the picks of deer’s horn.’[2198] But no bronze pick has

ever been found in this country; and deer-horn picks have been found

in interments of the Bronze Age,[2199] and even in a Romano-British

deposit in the village of Woodyates on Cranborne Chase.[2200] Professor

Gowland provisionally assigns the date ‘about 2000-1800 B.C.’ for

the erection of Stonehenge; and he adds that Sir Norman Lockyer’s

astronomical calculation ‘gives an approximate date ... of 1680 B.C.,

with a margin of error ± 200 years’.

More than one attempt has been made to determine the date of Stonehenge

from the orientation of its axis; and these attempts have been founded

upon the assumption that one, at all events, of the objects which the

builders had in view was the worship of the sun. ‘The chief evidence,’

as Sir Norman Lockyer and the late Mr. F. C. Penrose have observed,

‘lies in the fact that an “avenue” ... formed by two ancient earthen

banks, extends for a considerable distance from the structure, in

the general direction of the sunrise at the summer solstice.’[2201]

On the avenue, 100 feet from the so-called Slaughter Stone, stands a

large monolith, called the ‘Friar’s Heel’, or the ‘Heel Stone’. At one

time it was generally assumed that on Midsummer Day, at the time when

Stonehenge was built, an observer, standing on or behind the ‘Altar

Stone’, could see the sun rising above the tip of the Heel Stone. At

the present time, however, as Mr. Arthur Hinks points out, ‘the sun

rises further south than it has done for the last ten thousand years’;

and yet, from the point of view of an observer standing behind the

Altar Stone, ‘it still rises north of the stone.’[2202] In fact ‘it is

some seven days before or after midsummer day when it rises directly

over the stone’.[2203] Moreover, as Professor Flinders Petrie[2204]

says, the ‘skew position’ of the Altar Stone would seem to show that

it is not now in its original position. Accordingly Sir Norman Lockyer

felt obliged to leave the Friar’s Heel out of his calculations, and

to confine himself to attempting to determine the orientation of the

avenue.[2205] The method which he and his colleague adopted was to peg

out as accurately as possible ‘the central line between the low and

often mutilated banks’ of the avenue, and then to measure ‘the bearings

of two sections of this line near the beginning and the end’.[2206]

‘The resulting observations,’ he tells us, ‘gave for the axis of the

avenue nearest the commencement an azimuth of 49° 38′ 48″, and for

that of the more distant 49° 32′ 54″.’[2207] But neither of these

measurements was adopted by Sir Norman. He found, or thought that he

found, that the mean between the two values which he had obtained,

namely, 49° 35′ 51″, was ‘confirmed by the information, supplied by

the Ordnance Survey, that from the centre of the temple [Stonehenge]

the bearing of the principal bench mark on the ancient fortified hill,

about eight miles distant, a well-known British encampment named....

Sidbury, is 49° 34′ 18″; and that the same line continued through

Stonehenge to the south-west strikes another ancient fortification,

namely, Grovely Castle, about six miles distant, and at practically the

same azimuth, viz., 49° 35′ 51″. For the above reasons,’ he says, ‘49°

34′ 18″ has been adopted for the azimuth of the avenue.’[2208] Having

regard to the rate of change in the obliquity of the ecliptic,[2209]

he concluded that the date of the foundation of Stonehenge was 1680

B.C.; but he admits that this date ‘may possibly be in error by ±200

years’.[2210]

It would appear then that, if Sir Norman Lockyer’s calculations are

well founded, Stonehenge was erected at some time between 1880 and 1480

B.C. Certainly the conclusion does not err on the side of excessive

precision. But the foundation upon which the calculations rest has

been shown by Mr. Hinks to be rotten. To begin with, the assumption

that Sidbury Hill was connected with the erection of Stonehenge is

absurd. Does Sir Norman Lockyer mean to suggest that the bench mark was

prehistoric? ‘In our climate,’ says Mr. Hinks, ‘Sidbury is probably

not visible from Stonehenge at sunrise once in twenty years.’[2211]

In point of fact it is never so visible: only the trees on the top of

the hill are to be seen. Furthermore, as Mr. Hinks points out, Sir

Norman Lockyer has assumed that ‘for [the temple of] Karnak the moment

of sunset was when the sun’s centre had just reached the horizon;

for Stonehenge sunrise was the moment when the first tip of the sun

appeared above the hill. It was necessary to adopt these precise yet

different phases for the two cases, because any other assumptions

would have led to results obviously absurd.’[2212] Finally, Sir

Norman Lockyer is obliged to assume that the builders of Stonehenge

could tell the exact day on which the midsummer solstice occurred.

The utter improbability of this assumption must be apparent to any

one who remembers that the astronomer who constructed the Julian

calendar miscalculated the dates both of the summer and of the winter

solstice.[2213]

Mr. E. J. Webb, whose brilliant article in the _Edinburgh Review_

of October, 1894, demolished Sir Norman Lockyer’s theory as to the

orientation of the Egyptian temples,[2214] has written me a letter in

which the futility of attempting to determine the date of Stonehenge

by astronomical reasoning is explained with a clearness which leaves

nothing to be desired. ‘As,’ he writes, ‘the sun in our latitudes

does not rise at right angles to the horizon, but with a considerable

slant, it follows that the place where his upper rim begins to appear

is appreciably further towards the north than the place where his

centre appears, and this again than the place where he is first seen

fully risen,--that is, where his lower edge touches the horizon.’ Now I

think myself that, even if we could credit the builders with complete

accuracy, attempts to get the date of the building astronomically would

be vain, because (1) we do not know the exact place (if such there

was) at which the observer’s eye was supposed to be placed. (Flinders

Petrie does to some extent get over this difficulty by supposing that

the observer took up a position from which the point of the Heel Stone

appears exactly level with the horizon. I doubt, however, whether we

have a right to be sure that the point is exactly where it was at

first. Some of the stones have leaned over considerably, and why not

this? But the difficulty is much greater for Lockyer, who takes no

account of the Heel Stone.) (2) We do not know whether the ancients

would have understood by the moment of sunrise the moment when the

sun’s upper rim appears (_A_), or the moment when his centre appears

(_B_), or the moment when his lower rim appears (_C_). (3) Even if we

did know this, yet, as every one who has watched the sun rise must

admit, it is practically quite impossible to be certain when any one of

these moments occurs. Lockyer tacitly admits this when he arbitrarily

takes as the moment of first appearance the time when 2′ (about 1/16)

of the sun’s disc are risen.

‘It is clear that (_A_ being assumed) when Stonehenge was built, an

observer looking along Flinders Petrie’s line of sight would see

the Friar’s Heel considerably to the south of the place of sunrise,

inasmuch as, though that place has ever since been moving southwards,

we see it slightly to the south even now. Lockyer therefore puts the

Friar’s Heel out of his theory altogether, in the belief that a stone

which did not exactly mark the place of the most northerly sunrise

could be of no use. I think, on the other hand, that a stone placed a

little too far south would probably suit what is likely to have been

the purpose in view even better than one which exactly marked the

solstitial sunrise. For if, instead of asserting that Stonehenge was

roofed over, and a beam of light admitted at the moment of sunrise

to its darkened sanctuary[2215]--all of which is pure guesswork--we

suppose that the builders were contemplating merely such bonfires and

rejoicings as, by Lockyer’s own admission, certainly have taken place

in various parts of the world, may we not ask how people in those

ancient days knew when these festivities were to be held. For the later

days, of which we have knowledge, the answer is easy enough: then

people had the Julian calendar, according to which St. John’s Day, or

whatever day was selected, always recurred at the same place in the

solar year, whether at, or before, or after the solstice. But we do not

know whether in pre-Roman times the inhabitants of Wiltshire had any

settled calendar at all; and if they had, it is probable that, as in

almost all ancient calendars, the days of the month, and therefore most

likely the festivals, were reckoned by the moon.[2216] The fifteenth,

let us say, of a particular month meant the day when a particular moon

was fifteen days old; and if this day should coincide in one year with

the solstice, it would not coincide with it the next year, and could

not have coincided with it the year before. How then could people

tell when the Midsummer festival ought to be held? I answer that they

might have very easily done so some time beforehand by the aid of a

stone set up so as to mark, not the solstitial sunrise itself, but an

earlier--and therefore of course also a later--one. If the Friar’s

Heel stood, as on Lockyer’s theory it did, some little way to the

south of the place of the midsummer sunrise, then the sun must have

risen over it twice--first towards the end of his journey north, just

before the solstice, and secondly on his return southward, just after

the solstice. Now if the Stonehenge people looked out for the morning

on which the sun first seemed to rise over the stone, and counted the

days to a morning when he seemed to rise there again on his return

journey, they could, by halving this number, obtain the time of the

solstice with as much accuracy as they could have required. After doing

this once, they could in following years always know, by watching the

sun’s first approach to the Friar’s Heel, for what day to appoint the

midsummer rejoicings. That these rejoicings took place at sunrise I do

not assert. Bonfires, at least in these times, usually take place at

night.... I do not think we have any right to say with certainty that

any solstitial festival ever took place at Stonehenge or near it. For

even granting that, as seems not unlikely, Stonehenge was orientated

more or less closely to the solstitial sunrise, and that the Friar’s

Heel was really used for the observation of the sun, it does not follow

that Stonehenge was a “solar temple” any more than Milan Cathedral,

which is orientated more or less closely to the equinoctial sunset, and

has had a meridian line traced upon its pavement. And even if we knew

that it was a solar temple, we should have no right to infer what kind

of worship went on there.’[2217]

Although Dr. Evans’s arguments are not all equally strong, there

can be little doubt that his view as to the date of the erection of

Stonehenge is approximately correct. The stones were certainly not

standing when round barrows were first erected on Salisbury Plain;

for one is contained within the _vallum_, which, moreover, encroaches

upon another.[2218] Mr. F. R. Coles has shown that ‘so far as direct

evidence has been obtained by rightly conducted excavations, the

outstanding feature of all the Scottish circles that have been

investigated is the presence within them of interments of the Bronze

Age’.[2219] That Stonehenge was erected before the close of this

period, or at all events before the dawn of the Iron Age in Wiltshire,

is certain; and, as it was the most elaborate and highly finished of

all the stone circles of Great Britain, we may fairly infer that it was

one of the latest of them all.

II. The most interesting pages of Dr. Evans’s article are those in

which he attempts to trace the pedigree, so to speak, of Stonehenge,

and to divine the purpose of its builders. He cites instances to show

that ‘wherever the meaning of these great stone monuments has been

clearly revealed to us, we find them connected either directly or

indirectly with sepulchral usage’.[2220] He contends that in the most

characteristic examples ‘the Circle is an enlarged version of the ring

of stones placed round the grave-mound; the Dolmen represents the cist

within it; the Avenue is merely the continuation of the underground

gallery, which in our earliest barrows leads to the sepulchral

chamber’.[2221] But is there any evidence that interments ever did take

place within the precincts of Stonehenge? General Pitt-Rivers remarked

that the question could be definitely settled by excavation;[2222] but

scientific excavation, as we have seen, has hitherto been confined

within a small area. The evidence amounts to this:--a vessel, which

Dr. Evans calls an incense-cup, was discovered by Inigo Jones,[2223]

and incense-cups have never been found except in association with

interments;[2224] while the numerous bones of domestic animals which

have been exhumed, along with charcoal and fragments of pottery, from

the interior circle,[2225] point to the conclusion that Stonehenge was

the scene of sepulchral rites such as we know to have been performed

in barrows.[2226] Furthermore, the older monument of Avebury contains

two smaller stone circles, within each of which are the remains of a

stone chamber, which, Dr. Evans argues, ‘there can be little doubt once

contained interments.’[2227] But Dr. Evans is at no great pains to

argue that Stonehenge was itself a cemetery: it is on its connexion,

close or distant, with sepulchral usage that he lays stress. While he

points out that ‘in the case of the Chambered Barrows the [surrounding]

stones may be said still to fulfil an original structural function’,

he holds that ‘in the case of the Circles they bear a more purely

ritual signification. In some cases,’ he adds, ‘we find transitional

examples in which the stone circle is actually seen in the act as it

were of separating itself from the earth barrows. Thus in the great

monument of New Grange [in Ireland] the stone circle is separated by

an interval of some twenty feet from the central mound.’ Then, going

to the Far East for an illustration, he tells us that, while the stone

circles and dolmens which are still erected by the Khasis of Assam

‘are in themselves non-sepulchral’, they ‘are reared as a propitiation

either to the departed Spirits of their own ancestors or to any other

Spirit’.[2228] But Dr. Evans does not deny that Stonehenge was also a

solar temple: he admits, indeed, that its orientation ‘certainly seems

to associate the Sun in the religion of the spot’.[2229] This theory is

supported by observations made by Professor Gowland in Japan. ‘There,’

he tells us, ‘on the seashore at Futa-mi-gaura ... the orientation of

the shrine of adoration is given by two gigantic rocks which rise from

the sea as natural pillars. The sun, as it rises over the mountains of

the distant shore, is observed between them, and the customary prayers

and adorations made ... the point from which the sun is revered is

marked by a structure of the form of a trilithon,’[2230] &c.

But although some evidence has been collected in support of the theory

that certain stone circles in the British Isles and elsewhere were

orientated more or less closely to the Midsummer sunrise, it does

not necessarily follow that they were solar temples;[2231] and a

scientifically conducted examination of the circles of Kincardineshire

and Aberdeenshire has shown that their main diameters ‘are in scarcely

any instance oriented (_sic_) to any point of the compass as we

understand the term’;[2232] while Mr. W. C. Lukis, pointing out that

on Dartmoor and in Cornwall circles are to be found in clusters, and

that there are three circles quite close to one another at Stanton

Drew, asks, ‘if they were temples, why should the worshippers have been

gathered into separate congregations?’[2233] The only answer which I

can suggest is that while each of these circles was probably erected

either for sepulchral purposes or in honour of a dead ancestor, the

rites which were from time to time solemnized within them may have been

connected with the worship of the sun. There is no reason to believe

that in any megalithic circle in the British Isles solar worship was

more than incidental.

About forty years ago the late distinguished archaeologist, Professor

Nilsson, wrote an article,[2234] the main object of which was to

prove that Stonehenge was a temple of Phoenician origin, consecrated

to the worship of Baal; but the evidence upon which he relied was so

unsubstantial that no useful purpose would be served by summarizing his

arguments, which, indeed, are virtually obsolete.

Professor Flinders Petrie[2235] argues that certain parts of Stonehenge

are much later than others; and Dr. Evans, who agrees with him, remarks

that ‘this is strongly shown by the fact that each of the Stone Circles

as well as the Earth Circle has a different centre’.[2236] Dr. Evans

also points out that, in the case of the circles which are still

erected in the East, ‘the huge blocks are not all put up at one time

but in batches of an equal number of stones at intervals of time.’

Professor Gowland has shown that the sarsen stones in the outer circle

must have been erected before the trilithons, and the trilithons

before the blue-stones.[2237] ‘That the stones,’ he remarks, ‘of

the central trilithon were erected from the inside of the circle

has been conclusively demonstrated by the excavations; hence the

“blue-stones” in front cannot have been erected before them. Moreover,

the “bluestone”, No. 68, the base of which was laid bare in Excavation

V, was found to be set in the rubble which had been used to fill up

the foundation of No. 56, and further, in a lower layer than its base,

there were two ... blocks of sarsen with tooled surfaces.... If [the

outer sarsens were set up] from the inside [of the circle], their

erection must have preceded that of the trilithons and hence of the

“bluestones”. On the other hand, should the outer sarsens have been

reared from the outside, it would not be possible for the “bluestones”

to have been placed in position before them, as they would then have

seriously interfered with, if not altogether prevented the erecting

operations.’ Mr. William Cunnington, however, observes that ‘the fact

that specimens of all the varieties of rocks which constitute the inner

circle of Stonehenge have been found in the mixed substance at the base

of ... [the stump of one of the blue-stones] proves that they were all

on the spot when the inner ellipse was erected’;[2238] and Professor

Gowland, who confirms this view, concludes that ‘no long interval

of time separated the erection of the sarsen and the “bluestone”

monoliths, although the work must have occupied a considerable period’.

III. Unwarned by the _Edinburgh Review_ and Mr. Hinks, Sir Norman

Lockyer published in _Nature_[2239] a series of ‘Notes on Stonehenge’,

which might be safely ignored if his authority had not made converts,

even among archaeologists and men of science who happen to be ignorant

of certain essential facts. He now maintains that the sarsens

‘and above all the trilithons of the magnificent naos represent a

re-dedication and a re-construction of a much older temple’; and,

further, that ‘the older temple dealt, primarily but not exclusively,

with the May year’, while ‘the newer temple represented a change

of cult, and was dedicated primarily to the solstitial year’. It

is unnecessary to examine in detail the process by which he has

endeavoured to establish these conclusions; but I shall give a few

specimens of his work.

‘Acting,’ says Sir Norman, ‘on a very old tradition, the people from

Salisbury and other surrounding places go to observe the sunrise

on the longest day of the year at Stonehenge. We therefore,’ he

concludes, ‘are perfectly justified in assuming that it was a solar

temple.’[2240] Not improbably it was--from one point of view; but how

old is the tradition? The earliest extant mention of Stonehenge is in

the _Historia Anglorum_[2241] of Henry of Huntingdon, who lived in the

twelfth century, but who does not refer to the tradition. Stonehenge,

according to Sir Norman Lockyer, was rebuilt in 1680 B.C. It is

therefore impossible to prove that the tradition originated even as

early as two thousand nine hundred years after the alleged date of the

alleged second dedication of Stonehenge. Tentatively I would suggest

that it may have arisen after 1771, when the astronomical theory was

anticipated by a Dr. John Smith.[2242]

Among the ‘considerations’ to which Sir Norman would ‘direct

attention’ in support of his theory the fifth[2243] runs as

follows:--‘It is quite possible that the rebuilding of the temple

in 1680 B.C. was part of a very large general plan which could only

have been undertaken by a large, powerful and comparatively civilized

tribe or people under strict government, commanding the services of

skilled mathematicians, for Stonehenge, Old Sarum, and Grovely Castle

occupy the points of an equilateral triangle of _exactly_ six miles in

the sides, and the three sides are continuations of the entrances at

Stonehenge and Old Sarum and of a ditch running through the centre of

Grovely Castle, and the line Stonehenge--Old Sarum passes _exactly_

through Salisbury Spire, which again is exactly two miles from Sarum.

We ought to restore the old name, Solisbury.’

‘Skilled mathematicians’ on Salisbury Plain in 1680 B.C., a thousand

years before the dawn of mathematics in Greece,[2244] busily engaged

in forming, for some recondite religious purpose, gigantic equilateral

triangles! Sir Norman italicized the word ‘exactly’. Evidently then

he wished to impress upon us, in proof of the mathematicians’ skill,

not only that they made their triangle equilateral, but that each

side measured six miles,--no more and no less. Is it not a remarkable

coincidence that the unit of measurement in the British Bronze Age

was the English statute mile? I confess that I cannot grasp the

significance of the prolongation of ‘the line Stonehenge--Old Sarum’

to Salisbury Spire, or of the fact that this additional section was

‘exactly two miles long’, unless the builders of Stonehenge were

Christians as well as mathematicians and Salisbury Spire was standing

in 1680 B.C. Nor indeed, it should seem, can Sir Norman himself: at

all events in _Stonehenge and other British Monuments Astronomically

considered_--a book which is, in the main, a reproduction of his

‘Notes’--the passage which I have quoted disappears: equilateral

triangle and skilled mathematicians are left to the kindly obscurity

of _Nature_. But if ‘the line Stonehenge--Old Sarum’ and the line

Stonehenge--Grovely Castle have lost all significance, why persist

in staking a hopeless case upon the imaginary importance of the line

Stonehenge--Sidbury Hill?

Sir Norman Lockyer has not restricted his researches to Stonehenge,

sun-worship, and the solstitial year. He has discovered instances in

which stone circles have been used for the observation not of the sun

but of the stars, and in which, ‘on account of the change in a star’s

place due to precession,’ ‘the sight line has been changed in the

Egyptian manner.’[2245] Among these astral temples were ‘the three

circles of the Hurlers, near Liskeard’ and ‘the circles at Stanton

Drew’. After an interesting calculation he announces that ‘we have the

following declinations approximately:--

The Hurlers. Lat. 50° 31′ Stanton Drew. Lat. 51° 10′

Dec. N. 38½° Dec. N. 37°

” 38° ” 36½°

” 37°

Here then,’ he observes, ‘we have declinations to work on, but

declinations of what star? Vega is ruled out as its declination is

too high.’ He concludes that the star which ‘the astronomer-priests’

observed was Arcturus, and that ‘the approximate dates of the use of

the three circles at the Hurlers’ are 1600 B.C. for the southern, 1500

for the central, and 1300 for the northern circle; and at Stanton Drew

1260 B.C. for the great circle and 1075 B.C. for the south-western

circle.[2246]

Once more I am puzzled. Sir Norman remarks that all these circles are

considerably older than Stonehenge.[2247] Stonehenge, he says, was in

use as a solar temple in 1680 B.C. and a good deal earlier: none of

the older circles began to be used as an astral temple until 1600 B.C.

Why? Surely not because Arcturus, Capella, and Vega all refuse to fit

in with ‘the sight lines’ which Sir Norman has discovered except at

inconveniently late dates? Again, ‘Vega is ruled out as its declination

is too high.’ But the present declination of Vega happens to be exactly

38½°. ‘In other words,’ as Mr. Webb writes to me, ‘there exists

between the circle and one of the brightest stars in the sky a perfect

correspondence, which is nevertheless, beyond all possibility of doubt,

wholly accidental.’ Why did Sir Norman omit to mention this significant

fact?

But second thoughts or kind friends have once more come to Sir Norman’s

rescue. In his book ‘Vega is ruled out as its declination _was_ too

high’[2248] (the italics are mine). ‘He had become aware,’ remarks the

lynx-eyed Mr. Webb, ‘of the damaging fact that the present declination

of Vega actually _is_ 38½° N., in other words that, on his own

principles, we can prove that the Hurlers were set up to-day.’

THE CASSITERIDES, ICTIS, AND THE BRITISH TRADE IN TIN

I. THE CASSITERIDES

I. The identity of ‘the tin-islands’, which ancient writers called

the Cassiterides, is still a matter of dispute. Professor Haverfield,

indeed, has affirmed that ‘the recent researches of Usener [for which

read Unger], Rhys, and others, have made it almost certain that the

Cassiterides were off N.W. Spain’.[2249] Professor Rhys shall speak

for himself. ‘M. Reinach,’ he says,[2250] ‘argues, convincingly as it

seems to me, that the Cassiterides meant the Celtic islands, or, as I

may call them, the British Isles.’ And, if anything relating to this

question is certain, it is that the islands off North-Western Spain,

which are supposed to have been the tin-islands, have never produced

any tin at all.[2251]

One group of scholars insists that all the ancient writers who

mentioned the Cassiterides associated them with Spain. But what if the

ancient writers were misinformed, or misunderstood their informants?

Another group insists that for the metallurgists of ancient Europe the

sole source of tin was the British Isles; and with this pronouncement

they would apply the closure to the debate. But the British Isles were

not the sole source;[2252] and the debaters persist in wrangling. If

the only question were, From what parts of Europe did the Greeks and

Romans derive tin, it could be answered in a sentence:--from Galicia

in Spain, Cornwall, and, possibly, the Scilly Islands.[2253] But this

is not the only question. What we want to know is, Were the ancient

writers misled into believing that the Cassiterides were islands? If

they were not misled, were they all thinking of the same islands?

Or did they attempt to indicate the position of the Cassiterides by

simply guessing? If they were misled, was the district to which their

informants alluded Galicia or Cornwall, or did they refer to both? Did

the ancient writers fancy that islands used as depots for tin were the

places in which the mines were situated? Did those who professed to

inform them intentionally mislead them?

The theories which have recently held or still hold the field

are, first, that the Cassiterides were a group of islets off the

north-western coast of Spain; secondly, that they were headlands of

the same coast; thirdly, that they were the Scilly Islands; fourthly,

that they were Cornwall, which is supposed by some writers to have been

regarded either as an island or as a group of islands, separated by

estuaries, which were erroneously believed to be channels; and, lastly,

that they were the British Isles.[2254]

II. Diodorus Siculus,[2255] after stating that tin was produced in

Britain and in many parts of Iberia, goes on to say that there are

many tin mines in the islands called Cassiterides, which are situated

in the ocean, off the coast of Iberia and above the country of the

Lusitani.[2256] Strabo mentions the Cassiterides four times. In the

first passage[2257] he says that the extremity of the Pyrenees is

opposite the western parts of Britain, and that the Cassiterides, which

are situated in the same latitude as Britain, are in the open sea,

opposite to and north of the Artabri. In the second[2258] he mentions

both the Cassiterides and the British Isles, clearly distinguishing the

two groups. In the third[2259] he says that, according to Posidonius,

tin was produced in the country beyond [that is to say, north of] the

Lusitani, and also in the Cassiterides; that tin was conveyed from the

British Isles to Massilia, and that, according to the same authority,

tin, silver, and gold were produced in the country of the Artabri,

the most remote tribe of Lusitania, who face the north-west. In the

fourth[2260] he says that the Cassiterides are ten in number and lie

close together in the open sea, north of the harbour of the Artabri;

that one of them is uninhabited; and that the inhabitants of the rest

wear black robes reaching down to their feet, and walk about with

staves in their hands, ‘like the Furies in tragedy.’ They are, he says,

nomadic, and live upon flesh meat; and they barter tin and hides with

merchants for pottery, salt, and articles of bronze. Formerly, he adds,

the Phoenicians monopolized the trade from Gades, or Cadiz, with the

islanders; and they kept the route a close secret, which, however, the

Romans, after numerous attempts, succeeded in discovering. Finally,

Publius Crassus sailed across (διαβάς) to the islands, ascertained that

the tin lay near the surface, and indicated the route for the benefit

of traders, ‘although the passage was longer than that [from the

continent] to Britain.’ In another passage[2261] Strabo says that the

Artabri dwell in the neighbourhood of the north-western promontory of

Iberia, which he identifies with the Nerian promontory.[2262] Pomponius

Mela,[2263] himself a Spaniard, immediately after speaking of Baetica

and Lusitania, and immediately before mentioning the island of Sena,

which was off the coast of Brittany, states that the Cassiterides

are situated _in Celticis_. Pliny[2264] says that the Cassiterides,

so called from the abundance of tin which they produce, are situated

over against Celtiberia, and that opposite the promontory of the

Arrotrebae are ‘the six islands of the Gods’, which some call ‘the

Fortunate Isles’. In another passage[2265] he says that tin was first

fetched from ‘the island Cassiteris’ (or from ‘the tin island’) by

Midacritus, whom M. d’Arbois de Jubainville,[2266] wrongly, according

to M. Salomon Reinach,[2267] identifies with ‘Melkarth, who personified

the Phoenician race’. In a third passage[2268] Pliny says that tin has

been fabulously reported to have been obtained from islands in the

Atlantic. ‘Now,’ he continues, ‘it is certainly known to be produced in

Lusitania and Galicia.’ By ‘islands in the Atlantic’ Pliny certainly

did not, as Professor Ridgeway supposes,[2269] consciously mean the

British Isles; for in his geographical system the northern limit of the

Atlantic was marked by the north-western promontory of Spain.[2270]

Ptolemy[2271] says that the Cassiterides are ten in number, and are

situated in the western Ocean. Finally, Dionysius Periegetes[2272] says

that ‘the western isles’, which produce tin, and are situated below the

Sacred Promontory, or Cape St. Vincent, are inhabited by Iberians.

From a comparison of these statements it is clear, first, that the

ancient geographers who mentioned the Cassiterides regarded them as

distinct from the British Isles; secondly, that they believed them

to be situated somewhere off the coast of Spain (although, as we

shall presently see, the words of Strabo are not inconsistent with

the theory that he identified them with the Scilly Islands, or even,

unconsciously, with the British Isles); thirdly, that of those who

attempted to define their position one associated them with the

south-western, the others with the north-western coast; fourthly,

that one writer mentioned _an_ island, Cassiteris, from which tin was

fetched; and, lastly, that this same writer, having affirmed that the

tin islands were opposite Celtiberia, nevertheless denied that any

islands in the Ocean which extended as far north as the north-western

promontory of Spain produced tin.

III. Assume, for the sake of argument, that the Cassiterides were off

the north-western coast of Spain. What, then, are the islands with

which they are to be identified?

1. Unger[2273] remarks that it may be inferred from Ptolemy’s statement

that they were south of the Nerian promontory and off the western

coast of Galicia. Strabo, it is true, places them northward of the

northern coast: but, says Unger,[2274] Strabo is wrong; for on the

northern coast there are no islands distant more than one German mile

[or between four and five English miles] from the shore, whereas Strabo

himself says that the Cassiterides were further from the continent

than Britain. Dionysius Periegetes was, by common consent, mistaken.

Let us see then what islands we can find off North-Western Spain.

The coast between Cape Ortegal and the mouth of the Douro is broken

by several inlets or fiords, which are called Rias. East of Cape

Finisterre, in the Ria de Corcubion, are three very small islands;

and off the south-eastern entrance of this Ria, by Cape Minarzo, are

six tiny islets. About 20 miles south-east of Cape Finisterre is the

Ria of Muros and Noya: on the west of Mount Louro, which dominates the

entrance of this fiord on its northern side, are the four small islands

of Bruyos; and there are others within the fiord itself and south

of it. Unger[2275] remarks that the small size of all these islands

harmonizes with the word νησῖδες, or ‘islets’, which Diodorus applies

to the Cassiterides; and that the smallest of them may have been left

out of account when the number was given as ten. Off Pontevedra Bay,

which is north of Vigo Bay, are the islands of Ons and Orcela.

Strabo says that the harbour of the Artabri, north of which he places

the Cassiterides, was formed by a gulf on which were situated numerous

cities.[2276] Mela[2277] describes a gulf in the country of the Artabri

as having a narrow entrance and a wide circuit, and adds that four

streams flowed into it; and Ptolemy,[2278] having first mentioned a

harbour of the Artabri, immediately south of the Nerian promontory,

speaks, in the next section but one, of ‘the Great Harbour’, on the

shore of which he places Brigantium. The gulf mentioned by Mela

and Ptolemy’s ‘Great Harbour’ correspond with the Ria of Betanzos

and Ferrol, which is between Cape Finisterre and Cape Ortegal; but

there are no islands north of this harbour. The identification of

the harbour of the Artabri which Ptolemy places immediately south of

the Nerian promontory depends of course upon the identification of

the promontory itself. The latter is generally identified with Cape

Finisterre; and if this view is correct, the harbour must have been

the Ria de Corcubion. Unger, however, identifies the Nerian promontory

with the bluff of land, between Cape Finisterre and Corunna, from

which project the headlands of Punta del Roncudo, Punta de Nariga,

and Cape de S. Adrian;[2279] and if he is right, Ptolemy’s harbour

was the Ria of Corme and Lagos. But, as Unger points out, there are

no islands off this harbour or north of it.[2280] It is clear, then,

that if the Cassiterides really lay in Spanish waters and on the

north of a harbour of the Artabri, that harbour must be looked for

further south. Now Unger[2281] observes that Posidonius, as quoted by

Strabo,[2282] makes the territory of the Artabri extend southward as

far as the river Douro; for he says that their territory produced gold;

and, says Unger, in the country of the Artabri, in the narrower sense,

there are no auriferous streams. Accordingly, Unger identifies the

harbour for which he has been searching with the Puerto de Bayona,--the

southernmost inlet north of the Douro; and from this harbour he

maintains that Crassus sailed to the Cassiterides, which he identifies

with the islands of Bruyos. He points out that the distance from the

northern entrance of the harbour, opposite the island of Bayona, to the

islands of Bruyos is eight German miles. This distance exceeds that of

the shortest passage between Britain and the continent; and accordingly

Unger insists that it harmonizes with the statement of Strabo. But the

point which he most strenuously labours is that ‘in this case only is

explicable the circumstance, strikingly calculated to cause such a

mistake as Strabo made, that Crassus sailed northward from a harbour on

the west coast [of Galicia], and yet sailed in the open sea’.[2283]

Mr. Cecil Torr,[2284] on the other hand, insists that ‘unless it can

be shown that there were tin mines on the islands [near Vigo], the

story [of the voyage of Crassus] cannot be used to show that Crassus

visited those islands’. Strabo, he adds, states precisely that the

Cassiterides ‘lay to the north of Ἀρτάβρων λιμήν’, which ‘is obviously

the gulf that now holds Ferrol and Corunna’. Here, as we have seen,

there are, _in Spanish waters_, no islands; and Mr. Torr argues that

‘Strabo is so very accurate in his description of this part of Spain

that his account of the Cassiterides cannot be explained away as an

inaccurate description of the islands at Vigo.... It must be a bit

of downright fiction repeated in good faith.’[2285] To this latter

argument it might be replied, first, that just as Strabo was mistaken

in supposing that the direction of the Pyrenees was from north to

south, so he may have been mistaken in supposing that the Cassiterides

were on the north of Spain; and, secondly, that the other writers

whose testimony has been quoted place them off the west coast. No

other reply, indeed, could be made by those who hold, like Unger, that

the Cassiterides were in Spanish waters. But, for reasons which shall

presently be given, I agree with Mr. Torr that Ἀρτάβρων λιμήν must

have been ‘the gulf that now holds Ferrol and Corunna’, and also that

Strabo’s ‘account of the Cassiterides cannot be explained away as an

inaccurate description of the islands at [or rather near] Vigo’. Only

I believe that he is wrong in regarding that account as ‘fiction’.

Strabo’s Cassiterides _were not in Spanish waters at all: they were,

as he says, in the open sea, and far to the north of Corunna_. Mr.

Torr’s other objection rests upon the fact that tin was never produced

in any island off the coast of Spain, except, possibly, in Ons,[2286]

which, for reasons obvious to any one who consults the map of Spain,

Unger does not include among the Cassiterides. The only possible

answer to this objection has been already suggested in this article:

it is that the islands may have served as depots to which the tin

was conveyed from the mainland opposite, and that they may have been

confounded with the districts in which the tin was actually produced.

This suggestion, however, leaves unexplained the definite statement of

Strabo, that Crassus sailed across to the islands and found that the

islanders worked the tin easily because it lay near the surface. There

remain three other objections, which, unless Strabo’s authority is to

be absolutely discarded, appear insuperable. First, Strabo, I repeat,

distinctly states that the islands were ‘in the open sea’ (πελαγιαι);

and of the islands which Unger identifies with the Cassiterides not one

is more than four statute miles from the mainland, while the nearest

is not more than two. Secondly, as they are all within sight of land,

their situation could never have been kept secret. Lastly, since it

was unnecessary for those who desired to reach them to sail from the

harbour of Vigo, and easy to sail across from the neighbouring Ria

of Muros and Noya, it is difficult to understand why Strabo should

have said that the distance which separated them from the mainland

was greater than the distance from Gaul to Britain. It is true that

the islanders, according to Strabo, dressed in black, and that,

according to the same authority,[2287] the inhabitants of Lusitania did

likewise; but any one who regards this as an argument for identifying

the Cassiterides with the islands near Vigo must make up his mind to

reject nearly all the details which are given by Strabo, and to pin his

faith to the undoubted fact that the Cassiterides are placed by most

of the ancient authorities off the coast of Spain. M. Salomon Reinach,

however, with whom I agree, argues that ‘the fact that numerous

[ancient] writers place the Cassiterides in geographical connexion

with Spain only proves--what we knew before--that Phoenician Spain had

commercial relations with those islands’.[2288]

2. M. Hans Hildebrand[2289] thinks that the Cassiterides were headlands

of the Galician coast. He argues that if they are to be located ‘in

England’, the name Cassiterides must be applied to headlands in

Cornwall; accordingly, he says, ‘je demande la même concession pour

ma théorie espagnole, savoir que ce nom désigne des caps.’[2290]

But Cornwall is part of an island which is itself one of a group

of islands: Spain is not an island at all. M. Hildebrand’s theory

can by no ingenuity be defended except on the assumption that the

ancient writers were misled by the fact that in the language of the

Phoenicians, from whom the earliest notions about the Cassiterides

may be supposed to have been derived, there was no word which

specially denoted islands; and if it is accepted, not only must all

the statements of those writers which relate to the situation of the

islands, their number, their inhabitants, the mode in which the tin

was extracted, and the voyage of Crassus, be rejected as absolutely

fictitious, but it is utterly impossible to conceive how they should

have originated.[2291]

3. The old-fashioned view, which identified the Cassiterides with

the Scilly Islands, has even of late years had adherents of high

reputation, such as Dr. von Gutschmid,[2292] Emil Hübner,[2293] and

Mommsen.[2294] Although the bulk of the tin which supplied the wants

of ancient Europe came from Cornwall and Spain, it is nevertheless not

improbable that some came from the Scilly Islands.[2295] If so, the

real Cassiterides were the Scilly Islands and ‘the adjacent island’ of

Great Britain. But of the ancient writers there were only two of whom

it can be maintained that when they referred to the Cassiterides they

were thinking either of the Scilly Islands or of Cornwall,--Festus

Avienus and Strabo.

Festus Avienus was a writer of the fourth century, whose _Ora maritima_

was based either upon a Greek version of the Carthaginian account of

the voyage of Himilco,[2296] or, as seems more probable, upon a Greek

poem, which had itself been compiled from two distinct Greek narratives

of different dates, the latter being assignable to the period between

240 and 150 B.C.[2297] After describing the rocky peninsula,

Oestrymnis, he says that in the gulf formed by it lie the islands

called Oestrymnides, which are widely scattered and rich in tin.[2298]

He does not mention the Cassiterides at all. The Oestrymnides, however,

are generally, and, if Festus was right in saying that they produced

tin, necessarily identified with the Cassiterides: the peninsula is

rightly identified with Brittany,[2299] or, more strictly speaking,

with the promontory formed by Finistère; and therefore the gulf is

either the Bay of Biscay, or the gulf in which lie the Channel Islands.

After describing the Oestrymnides, Festus goes on to say that ‘from

here it is two days’ sail to the Sacred Island’, that is to say,

Ireland (_hinc duobus in sacram sic insulam Dixere prisci solibus

cursus rati est_[2300]); and then, remarking that ‘the island of the

Albiones’, or Britain, is near, he says that the Tartesii used to

resort for trade to the Oestrymnides, and that the Carthaginians also

used to sail ‘these seas’.[2301]

The question of the identity of the gulf is discussed by Friedrich Marx

in an article on the _Ora maritima_,[2302] of which a summary has been

given by Mr. W. H. Stevenson.[2303] Marx, says Mr. Stevenson, ‘explains

the _sub vertice_ [of the promontory or peninsula, Oestrymnis] of

Avienus[2304] as referring to the maps of antiquity, and as having

the sense of “northwards of”, so that the Tin Islands are conceived

of as north of the promontory of Finistère.... The Tin Islands must

therefore be the mainland of Britain and the Isle of Wight (which Marx

considers to be included among the _laxe iacentes insulae_ of Avienus),

and cannot be explained as the Scilly Islands, which have nothing

beyond their insular nature to favour the identification.’ But why

‘therefore’? What has the Isle of Wight beyond _its_ ‘insular nature’

to ‘favour the identification’? If Marx is right in his interpretation

of _sub vertice_, the gulf (_sinus_) lay north of ‘the promontory of

Finistère’; and since it can hardly be maintained that this gulf

was the English Channel, it must have been the gulf in which lie the

Channel Islands. If it was the English Channel, Marx can hardly venture

to argue that ‘the mainland of Britain’ is _in_ the Channel. If any

conclusion can be drawn from the words of Festus, it must be either

that the _laxe iacentes insulae_ were the Channel Islands and the

_sinus_ the gulf in which they are situated, or that the _sinus_ was

the Bay of Biscay and the _insulae_ Ushant and the adjacent islets.

But I agree with Mr. Stevenson that ‘the Tin Islands [of Avienus] ...

cannot be explained as the Scilly Islands’, unless the indications

which Avienus gives of their situation are utterly misleading.

Strabo, as we have seen, says that the Cassiterides were in the

open sea northward from the harbour of the Artabri; and Mr. H. F.

Tozer[2305] argues that, according to Strabo’s ‘idea of the relative

position of these countries [Spain and Britain] this would place them

a great distance to the west of the Scilly Islands’. This objection,

however, assumes that Strabo was aware that the Scilly Islands were

comparatively close to the Land’s End. Strabo imagined that the

direction of the Pyrenees was from north to south; that the coast of

Gaul extended in a straight line from the northern extremity of the

Pyrenees to the mouth of the Rhine; and that the southern coast of

Britain extended from a point nearly opposite and close to the northern

extremity of the Pyrenees, parallel with the coast of Gaul.[2306] He

expressly states that the Cassiterides were in the same latitude as

Britain; and therefore, if he had intended to identify the Cassiterides

with the Scilly Islands, it would have been quite natural for him

to say that they lay north of the harbour of the Artabri.[2307]

Müllenhoff[2308] indeed dismisses the claims of the Scilly Islands with

contempt; but all that he has to say against them is that they never

produced tin, and that they are small. The former objection is, as we

have seen, unfounded; the latter is irrelevant, for small islands may

contain mines, and the islands off the coast of Spain are smaller still.

But I am not concerned to argue that the sailors from whom the ancient

writers, directly or indirectly, derived their information intended to

convey that the Cassiterides were the Scilly Islands and the Scilly

Islands alone; for, although the Scilly Islands did produce some tin,

by far the greater part of the British supply of that metal doubtless

came from Cornwall. Professor von Gutschmid indeed explains that ‘the

tin was supposed [by the ancient writers] to be produced where it was

exchanged,--a very common case’;[2309] and although the place where the

Cornish tin was exchanged by the merchants who used the overland route

was Ictis, or St. Michael’s Mount,[2310] the Phoenicians may possibly

have found it convenient to occupy one of the Scilly Islands.[2311] But

it seems to me safer to conclude that the Scilly Islands may have been

originally included with Britain under the designation, _Cassiterides_.

The late distinguished geographer, H. Kiepert, maintained that although

the name, _Cassiterides_, had been originally used by the Greeks to

denote the tin-producing districts of Britain, it was erroneously

applied by Strabo to the Scilly Islands. ‘Only to this group’, he

insists, ‘can Strabo’s account of the discovery of the ten small

_Cassiterides_-islands on the north of Hispania by ... Publius Crassus

refer, as there are no other islands in this part of the ocean.’[2312]

4. George Smith, for whom the Cassiterides represented simply the

Cornish peninsula,[2313] observed, anticipating a similar argument

of Mr. Cecil Torr,[2314] that ‘the Hebrew, Phoenician, and cognate

languages had no terms which distinctly specified islands, peninsulas,

&c.; one word being used to signify islands, sea-coasts, and even

remote countries. In these languages the whole coast of Cornwall

and Devonshire might be termed island or islands.’ It may be

objected that the very same argument might be used to show that the

name, _Cassiterides_, really denoted the headlands on the coast of

Galicia.[2315] But it is easier to conceive how the misconception

should have arisen in the case of Cornwall, part of a remote island

in the northern ocean and close to the Scilly Islands, than in the

case of Galicia; and, moreover, the Galician theory leaves the story

of Crassus’s voyage unexplained. But the problem of the Cassiterides

cannot be satisfactorily solved by the simple statement that they were

Cornwall.

5. Müllenhoff,[2316] M. Salomon Reinach, and various other writers

identify the Cassiterides with the British Isles. According to M.

Reinach,[2317] ‘the whole question resolves itself into this:--what

islands in western Europe produce tin? The British Isles alone fulfil

this condition; therefore we must recognize in them the archipelago of

the Cassiterides.’ ‘If,’ he adds, ‘Strabo does not identify them with

the British Isles, though he mentions both the one group and the other,

this is because in the different chapters [of his work] he follows

different authorities, some of whom allude to the Cassiterides from

hearsay evidence collected in Spain, while the others describe the

British Isles from experience derived on the spot.’ Then, remarking

that the alleged derivation of κασσίτερος (the Greek word meaning

‘tin’) from a Sumerian word and from an Assyrian word have been proved

to be fanciful, he argues that κασσίτερος did not, as most ancient and

modern writers have supposed, give its name to the Cassiterides, but

on the contrary derived its name from theirs. Similarly, he points

out, at least four names of metals have been derived from the names

of places which produced them, namely, copper from Cyprus; silver (in

Gothic _silubr_) from the town of Salybe in Pontus;[2318] bronze from

Brundisium; and _Kalay_, the Turkish word for tin, from Kalah in the

peninsula of Malacca. M. Reinach goes on to argue that as the Greeks

derived their knowledge of the Cassiterides from the Phoenicians, the

termination ιδες must have been added by them. There remains therefore

_cassiteros_, of which the first part is found in numerous Celtic

words, for example, _Cassi_, _Cassi_-vellaunus, Velio-_casses_, &c.

M. Reinach gives reasons, which appear to me unsatisfactory, for the

conjecture that _Cassiterides_ means the same as _insulae extimae_

(‘the remotest isles’); and he holds that the name was given to the

British Isles by the Celts of Western Gaul.[2319]

Whatever M. Reinach’s argument may be worth, he and Müllenhoff are

unquestionably right in one sense: the British Isles, taken as a

whole, were the only islands from which the ancients derived tin. But

this truism did not require demonstration. The question is, whether

the identification of the Cassiterides with the British Isles can be

reconciled with what was written about them by the ancient geographers.

IV. The story which Strabo tells about Publius Crassus presents

some difficulty. As we have seen, he says that after the Romans had

discovered the route to the Cassiterides in spite of the efforts which

the Phoenicians made to conceal it, Crassus sailed across to the

islands, ascertained that the tin lay near the surface, and indicated

the new route for the benefit of traders. The first question is,

who was Crassus? Unger[2320] maintains that he was the consul of 95

B.C. who conquered the Lusitanians. If so, he must have sailed from

the mainland to the islands near Vigo which Unger identifies with

the Cassiterides. But, as I have already pointed out, these islands

are quite close to the coast: their distance from the mainland is

not greater, but many times less than the distance of Britain from

the Continent; their whereabouts could never have been kept secret;

and they have never produced tin. Therefore, if Publius Crassus was

the consul of 95 B.C., Strabo’s story is utterly untrustworthy.

Mommsen[2321] holds that Crassus was Caesar’s lieutenant of that name,

and that he sailed from Gaul to the Scilly Islands before Caesar’s

first invasion of Britain.[2322] How then are we to account for the

ignorance of Caesar, who tells us that tin was produced ‘in the

midlands’ (_in mediterraneis regionibus_[2323]) of Britain? Professor

Ridgeway, who believes that the Cassiterides were the islands near

Vigo, also identifies Crassus with Caesar’s lieutenant, who, as he

reminds us, invaded Aquitania--the south-western division of Gaul--in

56 B.C. ‘He is all the more likely,’ writes Professor Ridgeway,[2324]

‘to have passed into Northern Spain, inasmuch as the people of that

region had given great assistance to the Aquitani ... (_B. G._, iii,

23). Without doubt he was fully aware of the mineral wealth of that

country, as is shown by Caesar’s remark (iii, 21) on their skill in

defending cities, in consequence of their having numerous copper mines

and other works in that region. As is plain from Strabo’s words, the

Romans already knew how to reach the tin islands by sea, coasting

round from the Mediterranean and up from Gades on the old Phoenician

track. Crassus, then, by opening up a far shorter route, that of a

short sea voyage from the Cassiterides to the coast of Gaul (possibly

to the Garonne), at once developed this trade. The ore lay near

the surface. The distance by sea was greater than that across the

English Channel, but the readiness with which the tin was obtained,

combined with the shorter land transit, more than compensated this.

Strabo is evidently contrasting the rival tin-producing regions when

he introduces the allusion to Britain.... From this achievement of

Crassus and its results we can now understand in its proper light

the famous expression of Pytheas, that “the northern parts of Iberia

are more accessible towards Keltiké than for those who sail by the

ocean”.... He found, as Publius Crassus found three centuries later,

that the mineral regions and islands of North-Western Spain were far

more accessible for the Massaliotes by a land journey across Gaul

and a short sea voyage than by the long and perilous route round by

Gibraltar.’ But Professor Ridgeway mistranslates ‘the famous expression

of Pytheas’,--τὰ προσαρκτικὰ μέρη τῆς Ἰβηρίας εὐπαροδώτερα εἶναι

[τοῖς] πρὸς τὴν Κελτικὴν ἢ κατὰ τὸν ὠκεανὸν πλέουσι.[2325] He fails

to see that the word πλέουσι refers to πρὸς τὴν Κελτικήν as well as

to κατὰ τὸν ὠκεανόν. The passage simply means that it is easier to

sail along the northern coast of Iberia (Spain) from west to east in

the direction of Keltiké (Gaul) than to sail along the southern coast

from east to west in the direction of the Atlantic.[2326] This, as

Müllenhoff[2327] observes, is perfectly true, owing to the set of the

current and the prevalence of westerly winds. Moreover, Professor

Ridgeway does not seem to be aware that there are no ‘tin islands’ off

the coast of Spain: he does not explain how Crassus could have found

time in 56 B.C. to make the ‘short sea voyage’ of five hundred miles or

more from the mouth of the Garonne to the neighbourhood of Vigo, when

he was campaigning in Aquitania until the approach of winter;[2328]

nor, finally, does he explain how the Massaliotes would have gained by

conveying tin five hundred miles from the neighbourhood of Vigo to the

mouth of the Garonne, and then considerably more than three hundred

miles across Gaul to Massilia, instead of overland across Spain. Mr.

Tozer[2329] disposes of the difficulty by simply discrediting Strabo’s

account. ‘There is no reason,’ he says, ‘to doubt that Crassus made

such an expedition; but whatever the place was to which he went, his

account is quite untrustworthy, because he represents the Cassiterides

as producing tin, whereas that metal is not found in any of the groups

of islands which lie off the coasts of Gaul, or Britain, or Spain. The

explicit character of his statements, however, seems to have deceived

his contemporaries, and Strabo among them.’ But what theory can Mr.

Tozer frame to account for the gratuitous mendacity which he imputes

to Crassus, who, by the way, was not Strabo’s contemporary?[2330]

Strabo’s story is, in any case, obviously inaccurate:[2331] but I agree

with Mr. Tozer that it contains a kernel of truth; and I can only

suppose that Crassus, when he was in Brittany in 57-56 B.C.,[2332] was

directed by Caesar to visit and report upon the tin-producing districts

of the British Isles.[2333] And if I am asked how I account for the

mistake which Caesar made when he said that tin was produced in the

interior of Britain, I offer the following suggestion. Crassus may have

contented himself with landing on the coast, perhaps at or near St.

Michael’s Mount, where the tin was delivered to the merchants:[2334]

if so, he was doubtless informed that the tin was actually won in the

interior, as, in literal truth, it of course was;[2335] and Caesar may

have hastily concluded from his report that the tin mines were far from

the coast. As to the details with which Strabo embellished his story,

it would be idle to conjecture from what source they were obtained. We

may be sure that he did not invent them; but he may have confused items

of information furnished by different authorities.[2336]