Popular Science Monthly/Volume 59/May 1901/Primitive Color Vision

by W. H. R. Rivers

Layout 4

PRIMITIVE COLOR VISION.^[1]

By Dr. W. H. R. RIVERS,

ST. JOHN'S COLLEGE, CAMBRIDGE UNIVERSITY.

THE importance of language as an instrument of anthropological enquiry has been the subject of much difference of opinion. On the one hand, there are those who believe that the relation between language and thought is so close that the former has always been an almost exact mirror of the latter, and that every increase in intellectual development has been accompanied by, if not conditioned by, a corresponding increase in the development of language. On the other hand, the tendency which perhaps now prevails among anthropologists is to attach too little importance to language as an indication of the mental development of a race. The subject of the color sense of primitive races is one which is especially useful in studying how far the capacity for appreciating differences goes with the power of expressing those differences in language. We are able to put to the test how far the ideas of a people may be deduced from their language. We can collect the epithets used for color in various races, both of the present and of the past, and from a study of these epithets we can draw conclusions as to the nature of the color sense in these races. In the case of still existing races, we can then examine the color sense objectively and ascertain how far the conclusions derived from the study of language are verified by the result of the objective examination.

Historically, this is more or less what has been done. In 1858, in his 'Studies on Homer and the Homeric Age,' Gladstone called attention to the great vagueness of the color terminology of Homer; he showed that Homer used terms for color which indicated that his ideas of color must have been different from our own, and he was inclined to go as far as to suppose that Homer had no idea of color as we understand it, but distinguished little beyond differences of lightness and darkness.

Ten years later, Geiger,^[2] from a more extended investigation of ancient writings, also came to the conclusion that the color sense of the ancients must have been very defective. He found, not only in Greek literature, but in the Vedic hymns of India, in the Zendavesta, in the Norse Edda, and in ancient Chinese and Semitic writings that there was evidence of great imperfection, especially in the names for green and blue. In hardly any of these ancient writings is any word used from which we may gather that the people who wrote them had any idea of blue. Geiger advanced the view that there had been an evolution of the color sense in historical times; and he supposed that this evolution had been of such a kind that red had been distinguished first, followed by yellow and green, and that the sense for blue had developed much later than that for the other colors. Magnus^[3] came to the same conclusions on the basis of a still more extended examination of ancient writings, and Gladstone, in 1877, again called the attention of English scholars to the subject in the pages of the 'Nineteenth Century.'^[4]

The subject was taken up both from the literary and scientific points of view. On the literary side it was objected that the special peculiarities of the color terminology of Homer were due to a characteristic of epic style, according to which attention is paid to form rather than to color. It was also pointed out that the language of some modern poets presents the same peculiarities as those of ancient literature. Grant Allen^[5] counted the color-epithets used in Swinburne's 'Poems and Ballads' and found that red occurred much more frequently than blue, and a similar preponderance of red was found to be a feature of Tennyson's 'Princes.' Instances were also given of individual peculiarity in the use of color by many modern poets, one instance being La Fontaine, who, according to Javal,^[6] only used an epithet for blue once in the whole of his poems.

On the scientific side, also, objections were raised. It so happened that about 1877 there were in Germany two parties of Nubians going from town to town in traveling caravans. These Nubians were examined by Virchow and others, and it was found that they exhibited the same peculiarity of color language as ancient writers; they had no word for blue, or, rather, they used the same word for blue as for black or for dark colors generally. On examination, it was found, however, that they were not color blind, and that they sorted colored papers and wools correctly. It was, therefore, concluded that the ideas of Gladstone and Geiger were altogether erroneous, and that there was no necessary connection between color sense and color language.

In this country the views of Gladstone and Geiger were submitted to a comprehensive criticism by Grant Allen in the book, 'The Color Sense,' already cited. The strongest objection raised by this author was based on the existence of a well-developed color sense in many of the lower animals, and it was argued that this sense could not therefore be defective in primitive man. He also brought forward evidence that many existing primitive races made large use of color, and cited the opinions of travelers that savages were able to distinguish colors perfectly. He also pointed out that the decorations of the early Egyptians and of other ancient races showed the existence of a well-developed sense long before the time of Homer.

The controversy was carried on for some years, especially in Germany. Magnus^[7] showed that the same defect of terminology for green and blue, which characterizes ancient writings, still exists among many primitive races at the present day, and argued that this wide-spread peculiarity must have had some definite cause, probably of a physiological nature. Nevertheless, the general trend of opinion was strongly against the views of Gladstone and Geiger, and the idea of the evolution of the color sense in man has been almost universally rejected.

As a member of the anthropological expedition which went out from Cambridge to Torres Strait and New Guinea in 1898, under the leadership of Prof. A. C. Haddon, I had an opportunity of re-investigating this question. In addition to a full examination of the color vision of two tribes of Papuans inhabiting one the eastern and the other the western islands of Torres Strait, I was able to make observations on natives of the Island of Kiwai, at the mouth of the Fly River, and on members of several Australian tribes. The languages of these people showed different stages in the evolution of color terminology, which correspond in a striking manner with the course of evolution deduced by Geiger from ancient writings. In an Australian tribe, from the district of Seven Rivers, on the eastern shore of the Gulf of Carpentaria, several natives only used three color epithets; red, purple and orange were called by the same name, 'ǒti'; white, yellow and green were called 'yǒpa,' while black, blue, indigo and violet were all called 'manara.' Other natives from an adjoining tribe used the names 'owang,' wapǒk' and 'unma' in the same sense; some natives applied other names to green and yellow, but those given appeared to be the only terms which were used with any definiteness and constancy.

The next stage in the evolution of a color vocabulary was found in Kiwai. In the language of this island there was a very definite name for red, 'dǒgó-dǒgó,'^[8] and a less definite name for yellow, 'agó-agó agó-agó.' Greens were called by most individuals, 'emasóro' and 'tigiro,' which were also used for white and black respectively, and may probably be translated 'light' and 'dark.' A few used a special word for green, 'poroporona.' Blue and violet were usually called 'wǐbu-wǐbuna,' the word for black, others calling these colors 'tǐgiro' or 'pǒropǒrona.' The brilliant blue of the sky received from these people the same name as the deepest black.

In Murray Island, in the eastern part of Torres Strait, where we stayed for four months, I was able to investigate the language used for color very completely. In this island there was a very definite name for red, 'mamamamam,' while several other names, such as 'kiamikiam,' 'erǒko, mamamamam' and 'somer-mamamamam,' were used for purples and pinks. There were two definite names used for both orange and yellow, 'bambam' and 'siusiu,' and one definite word for green, 'soskěpusoskěp,' while several other words were occasionally used. There was, however, no native name for blue, apart from that used for black, 'golegole.' This word was used by many of the older men, and, as in Kiwai, the brilliant blue of the sky and the deep blue of the sea would often be called by the same name as the darkest black. Many of the natives had, however, adopted the English word, which, by re-duplication and separation of contiguous consonants, had become 'bŭlu-bŭlu,' and many of the younger men believed that this word belonged to their own language.

The color language of the western tribe of Papuans in Torres Strait was fully investigated in the Island of Mabuiag. Here the vocabulary was more definite. There were a limited number of terms which were used by nearly all for the chief colors. Red and yellow were called 'kulkadgamulnga' and 'murdgamulnga' respectively. There was a fairly definite term for green, 'ildagamulnga,' which was, however, sometimes used for blue, and there was a term for blue, 'maludgamulnga,' which was also used not infrequently for green. In addition to these four more or less definite color names, other terms were used for different shades, and a few natives showed extraordinary ingenuity in devising special names, apparently on the spur of the moment, for different shades of color. I have a list of over thirty such names from one individual, all derived from a comparison with natural objects.

In these four languages of Seven Rivers, Kiwai, Murray Island and Mabuiag, we have progressive stages in the evolution of color language; in the lowest there appears only to be a definite term for red apart from white and black; in the next stage there are definite terms for red and yellow, and an indefinite term for green; in the next stage there are definite terms for red, yellow and green, and a term for blue has been borrowed from another language; while in the highest stage there are terms for both green and blue, but these tend to be confused with one another. It is interesting to note that the order in which these four tribes are thus placed, on the ground of the development of their color language, corresponds with the order in which they would be placed on the ground of their general intellectual and cultural development.

It is said that there are other races, such as the Todas of Southern India, who have only three definite terms in their color vocabulary, viz., those for red, white and black, while others have also a term for yellow. The absence of a definite term for blue, on the other hand, is very common. The languages which have this characteristic appear to fall into two main classes; those which, as in Kiwai, have the same word for blue and black, and those which have the same word for blue and green. The former class includes the languages of Hovas and Bushmen, as well as of many Australian and Melanesian tribes. The second group comprises a very large number of languages, including one so near home as Welsh, in which there is only one word, 'glas,' for both green and blue.

By many races a word for blue has been borrowed from some other language, as was the case in Murray Island; thus many African races are said to use the term Dru,' obviously a corruption of the English word; in South America the Spanish word 'azul' has been borrowed, and the Battas of Sumatra have borrowed words both from Dutch and Malay. The word used by the Ga people of the Gold Coast for blue and for indigo is said to mean literally 'something that must be learnt,' these people having been taught the use of indigo either by Europeans or by other Africans.^[9]

When in Ceylon I obtained color vocabularies from a number of Singhalese and Tamils, and, though the two languages differed in other respects, both Singhalese and Tamils used the word 'nil' or "nilam' for blue, and this word, which is said to be the same as the name of the river Nile, is found widely distributed among Asiatic languages. The river Nile has another interest in connection with this peculiarity of color language. We are in the habit of speaking of the White Nile and the Blue Nile. The Arabic name for blue is 'azrag,' a word used by the modern Egyptian for blue and for dark colors generally. 'El Bahr azrag' probably originally meant the dark Nile, and, when we speak of the Blue Nile, we are using an expression which is based upon the primitive confusion between blue and black.

Magnus has shown that these defects in color nomenclature cannot be referred to the poverty of language. Some races, such as the Kaffirs and Basutos of South Africa, who have no word for blue, have, nevertheless, a very long vocabulary for the various colors of oxen. Similarly, the Kirghises,^[10] of Central Asia, have many different names for the colors of horses,^[11] though they tend to confuse the designations for green and blue.

I have had an opportunity of examining the color vision of the Eskimos who have lately been in London, and have found that their language presents a very striking contrast to that of the tropical people, to whom my previous work had been limited. The terminology for color appears to be extremely well developed; there are definite names for red, yellow, green and blue, and modifications of these colors are expressed by means of suffixes or by compounding two names; thus, by more than one individual a purple was called 'aupaluktaktungalangaijuk,'^[12] which means bluish-red, while a violet was called 'tungajuktakaupalangaijuk,' which means reddish-blue. This recognition of the fact that violet and purple are mixtures of red and blue shows a high degree of definiteness in the nomenclature of both colors. I have only met with one other individual who behaved in a similar way, viz., a Tamil, who called purple 'sikapu-nilam,' red-blue.

Another peculiarity which appears to characterize a very large number of languages is the absence of a word for brown. In Torres Strait a native would call one brown by a name meaning 'reddish'; another brown by the same name as yellow, while others would be called dark or gray. It was quite clear that they had no generic name for brown. In the Australian and Melanesian languages, I have had the opportunity of studying, as well as in Tamil, Singhalese and Eskimo, I have failed to find any definite term for brown, and the same defect is found in Welsh and in many other languages. The word given for brown in many vocabularies is obviously the same word as that used for red or dark or gray. There is always a danger that one may accept, as a generic name for brown, a word which is only a name for a special brown. This was very well shown in Mabuiag, where brown wools were by some natives called by such names as 'wamauwibadgamulnga' (honeycomb colored), or Vabadgamulnga (Draæena colored), but it was quite certain that these were names used by certain individuals for special browns and were in no sense names for brown in general. Similarly, in other languages in which there is no word for brown there may be names for special browns, such as names for the colors of horses or cattle, but such terms are limited to those objects. We have in our own language similar examples in the words 'bay' and 'dun.'

The idea of brown is so definite with us that it is surprising that a word for brown, and apparently the generic idea of brown, should be absent from so many languages. The fact is perhaps the more strange in that the word 'brown' bears evidence of having arisen in an early stage of our language, and is not, like violet or orange, obviously of recent origin.

The special characteristics of primitive color language appear, then, to be the following: The existence of a definite name for red, sometimes with subsidiary names for shades of red; a definite name for orange and yellow; indefinite nomenclature for green; absence of a word for blue, or confusion of the terms for blue and green, and absence of a word for brown, a brown object being called red, yellow or dark, according to its prevailing character.

These features closely resemble those of Homers color terminology, Homer uses several words for red, φοῖνιξ, φοίνιος, μίλτος, ὲρνρὀςθ and πορφύρεος; he has a definite word for yellow, ξανθός, an indefinite word for green, χλωρός; and no word for blue. Two words which later came to mean blue, γλανκός and κυάνεος were used by Homer, but it can not be said that the terms mean more than 'light' and 'dark' respectively. There seems now to be little doubt that κύανος, the substance, was 'lapis lazuli" and also an imitation of this substance made by coloring a glass paste with salts of copper, but κυάνεος is not used by Homer as an epithet for any distinctively blue object (except κύανος,), while it is used for a perfectly black garment.^[13] The substance, κύανος, is also qualified in one place,^[14] as μέλας.

There appears, also, to be no word for brown in Homer, but brown or brownish objects are qualified by the same adjectives which are used for red, thus φοῖνιξ^[15] is used for the color of a horse, and ζαφοινός is applied to jackals^[16] and the skin of a lion.^[17]

The resemblance is so striking that the conclusion seems irresistible that we have to do in Homer with a color vocabulary in the same early stage of development which is found among many primitive races at the present day. Indeed, one might almost go so far as to say that Homer's terminology for color is in a stage of development which is on much the same level as that of Kiwai, and distinctly less developed than those of Murray Island and Mabuiag.

From the nature of the defects of language, it has been concluded that the color sense of both ancient and existing primitive races is in some way defective. The next stage in the inquiry is to investigate the color sense of some existing people, and this I have been able to do most satisfactorily in Mun-ay Island. I tested about 150 natives of this island with Holmgren's wools for color-blindness, and failed to find one case in which there was any confusion between red and green, the common form of the defect in civilized countries. Since red-green blindness exists in about 4 per cent, of the male population of Europe, one may conclude that this form of defective color sense was either absent or was much rarer than with us. I also failed to find a case of color-blindness in Kiwai and Mabuiag and among the Australian aborigines, although I met with three well-marked cases of red-green blindness among eight natives of the Island of Lifu, in the Loyalty Group.

As regards other colors, however, the case was different; blue and green were constantly confused, and also blue and violet. Either owing to lack of interest or to some actual deficiency in color sense, there was a distinct tendency to confuse those colors for which their terminology was deficient. I have also found this tendency to confuse green and blue in several other races.

The behavior of the people in giving names to colors also pointed frequently in the same direction. I have already mentioned that in Mabuiag there was a great tendency to invent names for special colors; on one occasion a man, who seemed to have a special faculty in this direction,-gave me as the name for a bright blue wool 'idiiridgamulnga,' which meant the color of the water in which mangrove shoots had been washed to make 'biiu,' an article of food. In this case there was a deliberate comparison of a bright blue with dirty water, and I frequently came across other instances of the kind, which seemed almost inexplicable, if blue were not to these natives a duller and a darker color than it is to us.

This view was confirmed by quantitative observations, made with Lovibond's Tintometer, which had been kindly lent to the expedition by Mr. Lovibond. When the native looked into this apparatus he sawtwo square patches of light, either of which could be colored in any intensity of red, yellow or blue by means of a delicately-graded series of glasses of those colors. The 'threshold' for each color was then determined by finding the most faintly colored glass which the native could recognize and name correctly. The results showed that the natives recognized a very faint red, a more pronounced yellow, and only recognized blue when of a considerable intensity. Similar observations made on a series of Englishmen showed greatest sensitiveness to yellow and somewhat less to red and blue. The results may be given more definitely in Mr. Lovibond's units of color; in Murray Island, red was perceived on the average at.18 units, yellow at.27 and blue at .60 while the average results for the English observers were .31, .30 and .36 respectively. These figures do not show anything approaching blue blindness, but they do show a relative insensitiveness to blue in the Murray Islander, as compared with the European. The former appears nlso to be relatively more sensitive to red.

Another method of investigating the subject quantitatively which I employed was to determine the distance at which small spots of different colors could be recognized. I found in Murray Island that natives could see a red spot 2 mm. square at over 20 meters, while a blue spot of the same size was confused with black at even 2 or 3 meters. Europeans, however, also recognized red at a much greater distance than blue, and I have not at present sufficient comparative data to enable me to say that there is any marked difference between the Murray Islander and the European in this respect.

These results do not show that these islanders are blue blind, but they do show fairly conclusively that they have a certain degree of insensitiveness to this color, as compared with a European. We have, in fact, a case in which deficiency in color language is associated with a corresponding defect in color sense.

On the question of the cause of this insensitiveness there is room for differences of opinion. It is, of course, possible that the insensitiveness may be apparent only and may be merely due to lack of interest, but there is, I think, little doubt that it depends on physiological conditions of some kind.

The Murray Islander differs from the Englishman in two important respects; he is more primitive and he is more pigmented, and his insensitiveness to blue may either be a function of his primitiveness or of his pigmentation. In other words, it is possible that his insensitiveness may depend on the lack of development of some physiological substance or mechanism, which acts as the basis of the sensation blue in ourselves, or it may only depend on the fact that the retina of the Papuan is more strongly pigmented than that of the European. There is some reason to think that this latter factor is the more important. We know that the macula lutea in the retina, which contains the region of most distinct vision, is pigmented, and that as a consequence of the reddish-yellow color of its pigment, blue and green rays are more strongly absorbed than red and yellow; we have reason to believe further that the macula of dark races is more pigmented than that of ourselves.

The consequence would be that, in dark races, blue and green would be more strongly absorbed, and consequently there would be a certain degree of insensitiveness to these colors, as compared with red and yellow. In the observations made with the tintometer, the patches of color were so small that only the macula would have been stimulated. The probability that the insensitiveness to blue depended, at any rate partially, on the pigmentation of the macula lutea is increased by the fact that the natives were able to recognize blue readily on the peripheral retina.

It would, of course, be wrong to make any wide generalization on the basis of these observations. One would not be justified in directly applying the conclusions arrived at in the case of Murray Island to all existing races whose color nomenclature is defective, and still less so in applying them directly to the races of the past. Nevertheless, the fact remains that, in the only race which has been investigated with any degree of completeness,^[18] the characteristic defect in color language has been found to be associated with a corresponding defect in color sense.

There are other sources from which evidence on the evolution of the color sense in man may be derived. It has already been mentioned that at an early stage in the controversy the evidence of ancient monuments was brought forward against the views of Gladstone and Geiger. It was pointed out that, long before the time of Homer, green and blue pigments were used in Egyptian sculpture and decorations. In the Berlin Museum there is a palette with seven depressions, which appear to have been used for seven colors, white, black, red, yellow, green, a bright blue and a dark color which may have been either blue or brown. Indeed, blue appears to have been the predominant color of Egyptian pottery, and blue and green beads have been found in the graves of the prehistoric Egyptian race. Green and blue appear also to have been used in the decoration of the ancient Assyrians and Chaldeans. Greek architecture has also been found in Thera, Tiryns and Mycenæ of a date earlier than that of Homer, in which the colors used include blue.

Mr. Bénaky, of Smyrna, has recently collected^[19] the evidence derived from the coloration of ancient monuments, and believes that it decisively disproves the existence of any defect of color vision of the ancient Egyptians and Greeks. Two considerations must, however, be borne in mind when dealing with evidence of this kind. In the first place, it may be conceded that the monuments of the Egyptians show that these people had a perfectly developed color sense, and yet the color sense of the Greeks one thousand years or more later may have been defective. Just as we find different races at the present day in different stages of evolution as regards color, so it may have been three or four thousand years ago. The state of the color sense of the Egyptians has no direct bearing on that of the Greeks. It is a point of interest that the high development of the color sense in the ancient Egyptians, as shown by their decorations, appears to have been accompanied by a corresponding development of language, for it is stated^[20] that in the ancient Egyptian language there were two words for green and one for blue.

The other consideration, to which, in my opinion, sufficient attention has not been paid, is whether the various colors are used appropriately. To prove the existence of a well-developed color sense from monuments, it is not enough to show that certain colors were used; it must be shown that they were used appropriately. Even in the case of Egyptian art, one reads of statues of human figures with blue hair, and in the case of early Greek art, the inappropriate use of one color, blue appears to have been very common.

I am informed by Mr. E. E. Sikes, to whom I am glad of this opportunity of expressing my thanks for much kind advice, that in the Acropolis at Athens, such examples of coloration are to be seen as a blue bull, a blue horse, a man with blue hair, beard and mustache, these probably dating from 600 B. C, certainly later than the time of Homer.^[21] When such examples of eccentric coloration in blue are found associated with the defect in nomenclature for the same color, it is difficult to believe that the sense for this color can have been as highly developed in those times as it is among civilized races at the present day. The whole subject of the use of color in ancient monuments, in its bearing on color vision, requires a more thorough investigation than it has hitherto received.

Another line of objection to the views of Gladstone and Geiger, which has already been mentioned as having been taken up by Grant Allen, is derived from the high degree of development of the color sense in many of the lower animals, and especially in insects and birds. To many of those who have taken part in the controversy, this objection appears to have been regarded as conclusive. A well-developed color sense in any one branch of the animal kingdom does not, however, necessarily imply the existence of the same in other, even if higher, forms. We have many instances of the independent development of closely similar mechanisms in very widely separated branches of the animal kingdom, and there is nothing improbable in the view that this may have been so in the case of the color sense. If the color sense were found to be highly developed in mammals, the fact would naturally have a closer bearing on the color sense of man than has the presence of a similar development in birds. The evidence, however, of such development in mammals is very defective. Graber,^[22] who has carried out the most comprehensive investigations of the color sense in different branches of the animal kingdom, obtained much less definite evidence from mammals than from other animals, and altogether failed to obtain evidence in the case of some species. Again, if the anthropoid apes were found to have a well-developed color sense, the fact would have a still closer bearing on the condition of primitive man, but here again the scanty evidence is negative. The only experimental investigation with which I am acquainted is that made by Romanes^[23] on the chimpanzee 'Sally' in the Zoological Gardens. After having successfully taught this animal to recognize numbers, Romanes proceeded to apply a similar method to teach her colors, but wholly without success, and he was obliged to conclude that the animal was probably color-blind. It may be objected that the brilliant coloration of the mandrill and other species points to the existence of a color sense in the primates, but little weight can be attached to such indirect evidence in the absence of experimental investigation.

Another subject which has some bearing on the question is that of the color sense of the human child. It is now a more or less accepted principle in biology that the history of the individual presents the same stages of development as have occurred in the history of the race. Darwin^[24] was the first to point out that the power of distinguishing colors is a very late accomplishment in childhood; he found that his children were unable to name colors correctly at an age in which they knew the names of all familiar objects. This subject has since been the subject of much investigation, the most important work having been done by the late Professor Preyer^[25] and by Garbini.^[26] Preyer made a very large number of investigations on one child, while Garbini has based his results upon the observations of no less than 600 children. Both agree in the conclusion that the child is unable to distinguish colors at all till towards the end of the second year, and they also agree that red is distinguished and named correctly at an earlier age than blue, although there is some difference of opinion as to the exact order of development of other colors. Garbini points out further that the power of distinguishing colors develops earlier than the power of naming colors, language appearing to lag behind sense. If any importance is to be attached to the bearing of the history of the child on the history of the race, the evidence from childhood is in favor of the view that the color sense of man is a comparatively recent acquirement.

Whatever room for difference of opinion there may be on the question of the evolution of the color sense, there can be no doubt that there has been an evolution of color language. The possibility that the course of this evolution has been determined by physiological conditions has been considered, but there can, I think, be little doubt that these have not been the only factors upon which the characteristic defects of language have depended. The deficiency in the sense for blue, which I found in Torres Strait, is only partial and can not wholly account for the absence of a word for that color. Even to those with normal sensitiveness to blue, I think, there is no doubt that there is a closer resemblance between blue and black and between green and black than between red and black, and this difference in the degree of similarity between the different sensations of color and that of blackness may account in some measure for the difference in the definiteness of nomenclature.

It is a characteristic of the language of primitive races to have special names for every natural object, and often for very many individual parts of a natural object. If the savage has one name for one blue flower and another name for another, and so on, he will not require a name for the abstract quality of blueness. It is possible that he only begins to require names for colors when he begins to use pigments. If this be the case, it may help to explain the earlier development of names for red and yellow, for in many parts of the world pigments of these colors are by far the most common. In Torres Strait there were both red and yellow pigments, but no green pigment, and the nearest approach to a blue pigment was a slate-colored shale, and there appear to be many parts of the world where a blue pigment is wholly absent. Probably the most widely distributed blue pigment is indigo, and I have endeavored to ascertain whether those races which are familiar with indigo have a word for blue, but the evidence I have at present is too scanty to allow me to express an opinion on this point. It is probable, however, that the distribution of pigments has helped to determine the characteristic features of primitive color nomenclature, the greater frequency of red and yellow pigments being probably one of the factors which account for the more definite nomenclature for those colors.

Another factor, which may have been of importance, is the absence in the savage of an aesthetic interest in nature. The blue of the sky, the green and blue of the sea and the general green color of vegetation do not appear to interest him. It is, however, possible that the sky and sea do not interest the savage, or interest him less than the civilized man, because their colors are less brilliant than they are to us, and consequently this factor is not one on which much stress can be laid.

The widespread defect in the nomenclature for blue is rendered more striking by the fact that a name for red is universally present in primitive languages, while in many languages, as in that of Murray Island, various shades of red are not only discriminated, but also receive special names. In the experiments made in Torres Strait it seemed to me that this definiteness in the nomenclature for red was associated with a high degree of sensitiveness to this color, apparently greater than that of the average European. I also found, both by observation of their clothing and by direct questioning, that red was the favorite color of these people. In reading accounts of primitive man, one can not help being struck by the great predominance of red in the decoration of their houses, weapons and implements. This predominance may partly be due to the striking nature of the color and also to the prevalence of red pigments, but it seems possible that it may also be connected with the fact that red is the color of blood. Many savage races appear to be in a state of constant warfare, and in the religious rites and ceremonies of nearly all primitive races blood pays a great part.

The suggestion may even be hazarded that the earliest use to which red pigments were put was to smear the body in the war-dance, to imitate the blood-stained victor,^[27] or to replace blood in the various ceremonies of which it so often forms an essential feature. In his 'Legend of Perseus'^[28] Mr. Hartland has collected a number of instances in which it is perfectly obvious that vermilion or other red pigment has been used in the place of blood. Both in Murray Island and Mabuiag the chief words for red were derived from the name for blood, and this derivation is found in many languages, including our own. To whatever cause it may be due, there is no doubt that red is the most important color in the life of the savage, and it is natural that the predominant color should also be that which has the most definite name.^[29]

The main conclusions may be summed up as follows: The language used for color in ancient writings shows a characteristic defect, from which it has been concluded that the color sense of ancient races was also defective.

Existing primitive races agree in showing the same defect of color language as is found in ancient writings, and, in at least one such race, there has been found to be a corresponding defect in color sense, consisting in a certain degree of insensitiveness to those colors for which the nomenclature is defective.

Evidence, derived from ancient monuments and from the color vision of animals, which has been held to disprove the existence of any defect in the color senses of the ancients, appears to be inconclusive, and might, indeed, be held to support the views which have been derived from the study of language.

The observations made on the color vision of childhood may be regarded as indirect evidence that color vision has been a comparatively recent acquirement of the human race.

In addition to possible physiological conditions there are certain other factors which may have taken part in the production of the characteristic features of primitive color language.

On the more special question of the color sense of Homer, I believe that Gladstone and Geiger went too far. The evidence seems to me to suggest one of two possibilities. It is possible that to the Greeks of the time of Homer green and blue were less definite, possibly duller and darker colors than they are to us. It is, however, possible that the language used by Homer was only a relic of an earlier defect of this kind, the defect of nomenclature persisting after the color sense had become completely developed, language lagging behind sense in the race, as it appears to do in the child. According to the latter view, the defective terminology of Homer would be a phenomenon of the same order as the absence of a word for blue in such languages as Welsh, Chinese and Hebrew at the present day. It would not necessarily show the actual existence of a defective color sense, but would suggest that at some earlier stage of culture there had been defective sensitiveness for certain colors.

The evidence derived from poetry and art must always be in some degree unsatisfactory, owing to the great part which convention plays in these productions of the human mind. Still, every convention must have had a starting point, and though, in some cases, it is possible that considerations of technique^[30] may have originated the conventional use of color, it seems more probable that the predominance of red and deficiency of blue, both in the color language and in the decoration of the ancient Greeks, however conventional they may have become, nevertheless owe their origin to the special nature of the development of the color sense.

The subject of the evolution of the color sense in man is one which can only be settled by the convergence to one point of lines of investigation which are usually widely separated. The sciences of archæology, philology, psychology and physiology must all be called upon to contribute to the elucidation of this problem. I do not wish to do more than reopen the subject, and shall be contented if I have shown that the views of Gladstone and Geiger cannot be contemptuously dismissed as they were twenty years ago.

1. A lecture delivered at the Royal Institution, January 25, 1900.
2. 'Contributions to the History of the Development of the Human Race,' p. 48.
3. 'Die geschichtliche Entwickelung des Farbensinnes,' Leipzig, 1877.
4. t Vol. II., p. 366; 1877.
5. 'The Color Sense.' London, 1879: p. 264.
6. 'Bull. de la Soc. d'Anthropologie de Paris.' T. XII., p. 480; 1877.
7. 'Untersuchungen über den Farbensinn der Naturvölker, Jena, 1880,' and 'Ueber ethnologische Untersuchungen des Farbensinnes,' Breslau, 1883.
8. 'Ó' stands for the sound of 'aw' in the word 'law.'
9. For many other instances of defective color terminology among savage and semi-civilized races, see the papers of Magnus already quoted, and also Andree, 'Zeitsch f. Ethnologie,' Bd. X., p. 328; 1878.
10. Radloff, 'Ztschr. f. Ethnol.' Bd. III., p. 285; 1871.
11. Strictly speaking, these names and those of the Kaffirs are not names of colors, but rather names for distribution of color and marking; thus among the Kirghises a brown horse with a black mane and tail would have one name, and a brown horse with a white mane and tail, another name.
12. 'Au' stands for the sound of 'ou' in 'house,' and 'ai' for that of 'i' in 'ice.'
13. Il., XXIV., 93.
14. Il., XL, 24.
15. Il., XXIII., 454.
16. Il., XI., 474.
17. Il., X., 23.
18. A full account of these investigations will be given in the reports of the Cambridge Expedition.
19. 'Du sens chromatique dans l'antiquité.' Paris, 1897.
20. See 'Kosmos,' Bd. I., s. 430; 1877.
21. See also Gardner, 'Handbook of Greek Sculpture,' p. 28.
22. 'Grundlinien zur Erforschung des Helligkeits ixnd Farbensinnes der Tiere.' Prag., 1884.
23. Proc. Zoolog. Soc.,' 1889; p. 316.
24. 'Kosmos,' Bd. I., s. 376; 1877.
25. 'Die Seele des Kindes.' Leipzig, 1884.
26. 'Arch, per l'Anthropologia e la Ethnologia,' Vol. XXIV., pp. 1l and 193; 1894.
27. This custom appears to have persisted down to Roman times, in which the conqueror painted his body red when taking part in the triumphal procession.
28. Vol. II., pp. 242, 264; notes, 342, 354-5.
29. Since the above was written, the question of the predominance of red haa been fully discussed by Havelock Ellis. (Popular Science Monthly, August and September, 1900.)
30. As an instance of the origin of a convention in technique, Mr. Sikes has suggested to me the red figures on Greek vases. Early Greek vases were made of a reddish material, on which the figures were designed in black. At a later time, the vessels were black and the figures red, the conventional persistence of red decoration in this case having had its starting-point in the special nature of the material originally used in the manufacture of the vases.