At this point we have completed the survey in which, with the help of the combined disciplines of osteology, archaeology, history, and linguistic science, we have attempted to trace the development of racial entities in the territory occupied by the white race, from the earliest human times to the Middle Ages, the threshold of the modern period. We are now faced with the problem of working with a different body of material - that furnished by the anthropometry of living peoples. We must further attempt to fit this material into the frame furnished us by our study of the dead, so that from the combination of the two a complete and orderly reconstruction will result.

While we were dealing with the data gleaned from the measurement and observation of bones, the chief difficulty which faced us was the lack of adequate samples in most of the periods, regions, and cultural units under consideration. On the other hand, while metrical accuracy was by no means to be assumed, yet the measurements on the dry skulls and long bones were for the most part comparable, and technical difficulty was subordinate to the paucity of documents. In dealing with the living material, however, we have vastly larger samples. In some countries, as in Norway, Sweden, and Poland, these comprise the entire military age group of the nation, and thus cease to be samples in the strict sense, and assume the character of total populations. In relatively few regions is it necessary to use samples of less than one hundred individuals.

Our authority has, therefore, increased immensely. We may speak with some confidence of the superficial physical composition of most European nations. But, at the same time, what we have gained in volume, we have to a certain extent lost in accuracy, for the present state of anthropometry is partly one of confusion and mistrust in regard to technical methods. Despite various attempts in the past and in the present to establish a standard corpus of technique,¹ different schools have arisen in different countries. What discrepancies may exist between the work of members of each school can usually be determined and allowed for; but this is not the root of the trouble. The chief difficulty is that much measuring has been done not by professional anthropometrists but by amateurs, while some with professional status have not been properly trained. Therefore we cannot be sure that such men belong to any school, nor that they follow any standard other than their own. The accuracy of existing documents on the living is far less than that of skeletal data, and it is not always possible to know what techniques have been used. This lack of consistency is often an obstacle to mathematical comparison, but not enough of an obstacle to render many series wholly useless. We still have a better tool for the study of race in the living than we had in the documents of the dead.

Let us review the more important measurements in which technical difficulties most commonly arise. Stature, unfortunately, heads the list. One would suppose that the maximum height of the body while standing would be a constant dimension and one easy to measure, but neither assumption is true. Some investigators allow the subject to be measured in his shoes, and then attempt to make a standard subtraction for the heel. This is seldom if ever satisfactory. On the other hand barefooted negroes with horny soles are raised up several millimeters by their callouses, when compared to thin-soled white men standing with their shoes removed. Differences in posture, and in degree of conscious stretching, may attain the dimensions of centimeters.

Furthermore, it has been established² that the human body, except in senility, shrinks as much as 2.5 cm. during a daytime spent either afoot or in a chair, the amount depending partly on the degree of and nature of the day's activity. It makes some differences, therefore, what time of day the investigator habitually chooses for his work. At the same time the state of nutrition and of health makes some difference, and one must beware of series measured entirely in hospitals.

For the reasons above outlined, and without doubt for others as well, we must not, in studying stature as a statistical criterion of racial value, even if our samples are equivalent in age, expect to find accuracy down to the millimeter. Therefore the common statistical devices used to check the validity of the series on the basis of the sampling process are set at too fine an adjustment in view of the coarseness of the measurement itself, and in view of the great variability caused by factors other than sampling or racial attributes. What applies to stature applies in varying degree to measurements of its segments and of other bodily dimensions; the breadths of the shoulder and hips, and the diameters of the chest, are dependent in some degree on the highly variable amounts of sinew, muscle, and fat present at the points of measurement.

In the dimensions of the head and face, most of the difficulties found in stature and bodily measurements cease to exist. On the whole, a much greater accuracy is not only possible but has been attained. There are but two important matters in which serious inaccuracies arise with any frequency; these are the measurement of auricular head height and the location of nasion.

The first of these, the measurement of the height of the cranial vault, is without doubt the least satisfactory of all common anthropometric techniques. Although technique #15 of Martin³ is considered standard, not all use it, and few do it in the same way. Some investigators use special metal head-spanners which measure the height of the vault from the middle of the ear hole, others measure from the top of the ear hole; still others, following Martin, from tragion. There is also a dispute as to whether the height taken should be to the vertex, as stated by Martin, or to a point exactly above the ear hole when the head is held in an approximation to the eye-ear plane.

As a result of these technical difficulties in taking head height on the living, differences of from ten to fifteen millimeters exist between the results of different investigators working on identical populations, and reports embodying these discrepancies are published without comment. Since the difference between techniques is as great as the difference between extremely disparate racial groups of mankind, head height on the living is a useless criterion when employed uncritically. Unless the compiler knows the technical peculiarities and personal equation of each investigator whose work he uses, he should leave this material alone. In the present work, this ruling immediately excludes from consideration the majority of published data on head height.

The second major difficulty, the location of nasion in the living, while not quite as inaccurate, is even more serious, since three important vertical diameters of the face, morphological face height, morphological upper face height, and nose height, are theoretically limited, at their upper boundary, by this landmark; and nasion is an extremely hard point to determine. Ashley-Montagu, however, has recently devised a method which promises to overcome this difficulty in most cases.⁴ On adult male whites, luckily, there is usually enough ruggedness of facial relief to make this difficulty less serious than with mongoloids or negroids. Still technical differences of from five to ten millimeters render the works of different investigators incomparable, and one must again be sure of the individual equation of each investigator, or of the school in which he was trained. Since the facial and nasal indices depend upon vertical as well as lateral diameters, and hence upon nasion, these important racial criteria must be taken with great reserve, for the constancy of the lateral diameters serves only to exaggerate, in the indices, the differences between the vertical dimensions.

So much for the most serious metrical difficulties. In measurements on the living we see a more bountiful but less accurate counterpart of the criteria already familiar to the craniologist. There is another large body of data, however, unique in living material; the observations on the soft parts, including such features as hair form, hair texture; skin, hair, and eye color; the shape of the various component segments of the nose, the lips, and the external eye. These are important diagnostic racial characters and deserve as careful study as do measurements and indices. But, unfortunately, accurate comparisons between the work of different investigators is even less possible here than with metrical data, since observation is a matter of judgment, and no two men's judgments are the same.

The use of standard pigment scales in determining hair, skin, and eye color has helped enormously, but has not entirely eliminated the diffi-culties in the pigmentation field. There is no really adequate eye-color scale on the market, although Martin's series of sixteen glass eyes is far better than nothing. Von Luschan's skin-color scale does not always approximate human shades, and this is especially true with whites. The Sailer-Fischer hair-color scale, made from actual human hair, is excellent, in most respects, but has not yet come into common use; the earlier Fischer scale, made of bleached and dyed vicufia hair, is also good.

Unfortunately, however, the majority of our observational data has been collected without reference to scales, and published without accurate definitions, and it is impossible to tell, in many instances, what color or what degree of blondism or pigmentation is implied by a given term. Then too, environment and age make great differences in pigmentation; the degree of tanning or of uncleanliness in regard to the skin color is seldom indicated; eyes often grow lighter with age, and the deposit of fat in the cornea, called arcus senills, which gives a grayish-blue tone to the peripheral zone of the iris, is often mistaken for eye blondism. Hair color is notoriously transitory, changing, in all but pure brunets and extreme blonds, continuously from birth to grayness, baldness, or death.

Most observations, other than those referring to pigmentation and the morphology of the pilous system, are divided into the following categories: absent, sub-medium, medium, pronounced. These are frequently expressed by the symbols, abs., sm., +, + +. Often ssm. and + + + are added for greater refinement. In general, the standard for the + or medium category is a roughly estimated and ideal mean or intermediate white or European male condition, Thus in nasal tip thickness almost all negroes would be + + or + + +; in beard development almost all Eskimos would be abs., ssm., or sm. There is a tendency for the observer to make the mean condition of the people he is studying + or medium, or to be unconsciously influenced by his own facial form.

Various attempts have been made to standardize these quantitative observations, and the most promising is perhaps that of the Moscow school, where a series of plaster casts has been made to show standard stages of sm., +, and + + in each of the more commonly studied criteria. Still, whatever standards are used, the location of the borderline between categories must always be a matter of individual judgment.

Our first difficulty with the study of race from existing data on living populations, whether these data be metrical or observational, is therefore one of technical inaccuracy and inconsistency. But it is not the greatest difficulty which will be encountered, and it is not insuperable. The careful compiler can usually discover what are the technical idiosyncracies of a given investigator, and if he is familiar with the material as a whole, he can usually sense improbable divergences from standard technique. The comparison of different samples selected from the same population by different investigators often makes a standard adjustment possible.

Technical inconsistencies and inaccuracies render the study of race on the living something less than an exact science, but it remains something more than a plaything. The manipulation of metrical data requires experience and judgment, and the uncritical use of existing materials on a purely statistical basis, no matter how erudite in the mathematical sense, can never be more than a sterile exercise. Those who employ experience and judgment, and who make a discreet use of the simpler statistical methods, may learn much from the handling of the immense body of anthropometric data.

Notes:

1 Cf. The Geneva agreement of 1912; the standards established by R. Martin in his Lehrbuch der Anthropologie; the present laudable attempt of Miss Miriam Tildesley to bring about unification.

2 Backman, G., FUL, N. F. vol. 29, 1923-24, pp. 255-282.

3 Martin, R., Lehrbuch der Anthropologie, vol. 1, pp. 185-186.

4 Ashley-Montagu, M. F., AJPA, vol. 20, 1935, pp. 81-93; vol. 22, 1937, #3, Suppl. p.6.