On 1 August 1889, the journal Nature carried a letter submitted by an undergraduate mathematician from Trinity College, Cambridge. The letter criticised a paper delivered by Francis Galton to the Anthropological Institute, reprinted in Nature a year earlier, in which he had made a series of observations based on cranial measurements produced by the Cambridge Philosophical Society. Signing his letter with the initials ‘F.M.T’, the correspondent noted inconsistencies in the figures produced during repeated measurements of two of his friends:
The products are seen to vary in the first case from 236.7 to 250.1, nearly 6 per cent. So far are the figures capable of affording good evidence of head growth of either individual, they are so inaccurate as not to make certain whether X. or Y. had the bigger head.
The exchange between Galton and ‘F.M.T’, a student of medicine named Frederick Meadows Turner who had recently taken a B.Sc. from London University, continued through subsequent editions of the journal. Whilst Galton argued that measurement errors could be accounted for within his statistical analysis, questions about the value of inexact measures continued throughout the Cambridge Philosophical Society’s anthropometric work. Such discussions highlight the difficulties presented to quantitative methodologies by the human body, and the doubts surrounding the value of anthropometry in late 19th-century Britain.
The Limits of Measurement
The designs of anthropometric instruments illustrate the challenges presented to measurement practices by the human body. Such apparatus were described in the professional publications and catalogues produced by instrument makers and designers. In 1887, Francis Galton described his instruments to the Anthropological Institute of Great Britain and Ireland. His talk was immediately followed by a paper from Horace Darwin, instrument designer for the Cambridge Scientific Instrument Company. The two men corresponded on the design and construction of instruments, and both contributed to the framing of the Cambridge Philosophical Society’s anthropometric investigation. In the same year, the Company produced a descriptive list of their apparatus, with detailed explanations of their designs. Finally, in 1899 the Instrument Company produced a catalogue of their physiological instruments, including a section on anthropometry. These sources, and the contingencies built into the objects they describe, testify to the difficulty involved in subjecting the human body to accurate measurement.
Anthropometric Data Cards CPS/12/1/10, reproduced by permission of the Cambridge Philosophical Society (C)
Human bodies present a dynamic and changing subject, undergoing cycles of replication and regeneration at both cellular and systemic scales. The shape and size of bodily features vary both relatively and absolutely, these variations occurring at unpredictable time scales. Further, the body is composed of numerous tissues, combining different material states and offering few consistent properties. Perhaps most challenging, however, are the philosophical question posed by particular aspects of human physicality. The measurement of physiological responses such as reflex action or sensory perception raises questions about the relationship between the mind and the body.
Disciplining the Skull
Cranial measurements, such as those identified and criticised by Frederick Meadows Turner, provide illustrations of the difficulties facing anthropometric practitioners. The design and contingencies of the Horizontal Head Spanner produced by the Cambridge Scientific Instrument Company offers an insight into the problems posed by a living subject. The irregular surface of the head could easily influence results. To mitigate this, the point of contact between the spanner and a subject’s head was composed of pairs of steel rods: ‘The reason why the rods are in pairs and not single, is to give an extension of surface, the rods passing through the hair like the teeth of a comb.’ However, the instrument provided only linear measurements, restricted to a horizontal plane. In order to estimate cranial volume, a measure of height was necessary. This required the defining of conventions for a vertical dimension, and the designing of a suitable apparatus:
The third dimension required is the maximum height of the upper surface of the head above the plane passing through the holes in the ears and through the lowest part of the orbital cavity of the eyes. The Company hope soon to issue a satisfactory instrument after making some further experiments.
Once measurements for length, width, and vertical height, were obtained, the resulting figures could be applied to a formula for estimating volume. However, it was appreciated that the proportions of various parts of the skull were different for particular groups, and the accuracy of calculating cubic capacity was observed to vary accordingly. A paper presented to the Royal Anthropological Institute in 1910 by John Anderson, an anatomist working in Melbourne, explored the differences between estimates of cubic capacity and those produced through displacement techniques. Of the various formulas for estimating capacity from linear measurements, Anderson identified two credible methods. The first of these had been established by the French anthropologist and Madeleine Pelletier and the second by the biometric technician Alice Lee. Anderson noted: ‘I have ascertained that Pelletier’s formula gives the more accurate result in 37.5 per cent. of cases, and Lee’s formula in the remaining 62.5 per cent.’ However, both methods were compromised by the unpredictable shape and thickness of the skull, and the impossibility of defining standard cranial measurements.
Even the bodies of dead subjects resisted accurate measurement. One method, proposed by the surgeon and palaeontologist George Busk involved filling the skull with materials such as sand or seeds and then measuring the material. This method presented its own problems: the point at which a skull might be considered full was difficult to judge, the texture of the skull’s internal surface affected the accommodation of material, and the properties of the aggregate were difficult to control: ‘As the quantity of sand that can be introduced into a skull differs considerably according to the way in which it is introduced, it is very necessary that this should be done in a uniform manner.’ As with the measurement of living subjects, Busk’s work relied on improvised definitions of the extent and limits of the skull.
The Colour and the Shape
Anthropometric scrutiny was applied not only to size and shape but also to qualitative aspects of the human body. The colour of a person’s hair and eyes were recorded, these characteristics presenting further obstacles to standardised measurement. The data recorded by the Cambridge Philosophical Society reflects the difficulty in categorising colours, and duplicate records for particular individuals often assign different values to the same person’s features. In the anthropometric records of the Philosophical Society, eye colour is recorded in categories ranging between light and dark. These categories were embodied in an instrument produced by the Cambridge Scientific Instrument Company, that reflected the experience and advice of Francis Galton: ‘At the Health Exhibition Mr Francis Galton used three artificial glass eyes as standards for comparison. The colours were Blue, Brown, and an intermediate colour Grey-Brown.’ Similarly, in its 1899 catalogue of physiological apparatus, the Instrument Company advertised a hair colour scale using real samples:
Five samples of Human Hair are mounted behind a sheet of glass, and a cover is arranged by means of which the light can be excluded. The samples match those used by Mr Galton. Numbers are engraved over each thus: Red, No. 1; Flaxen, No. 2; Light Brown, No. 3; Brown, No. 4; Black, No. 5.
Despite the production of elaborate scales for hair and eye colour, the measurement of these characteristics was impossible to standardise. Such information relied on the perceptual faculties of an observer, and was subject to the variability of human judgement. The same problem had compromised the measurement programmes of natural philosophers throughout the 19th Century. For these reasons the evaluation and recording of features such as skin type, represented on the Philosophical Society’s data cards by predefined categories such as ‘pale’, ‘ruddy’, ‘freckled’, and ‘dark’, was of limited value.
Dynamic and Sensory Measurements
The measurement of dynamic characteristics such as strength or physiological response presented further challenges to anthropometric practitioners and instrument designers. Apparatus for measuring the strength of a subject’s grip could not control for hand size and shape, muscular distribution, and psychophysical factors. Galton’s investigations at the Health Exhibition had made use of a ‘hand dynamometer’, but the data it produced displayed a serious flaw: ‘in the results given by the original pattern of instrument, the size of the hand was an important factor as well as the strength.’ John Venn, a Cambridge philosopher closely involved with the anthropometric programme, noted: ‘The instrument cannot be regarded as very accurate, since the effective range through which the muscles of the hand can thus work is but small.’
Francis Galton’s Second Anthropometric Laboratory, Science Museum, South Kensington, 1885-1890. Karl Pearson, The Life, Labours, and Letters of Francis Galton, Vol. 2 (Cambridge, 1914), p. 378. Reproduced with kind permission from UCL Special Collections. Galton/Pearson Papers, UCL Library Services, Special Collections (C).
Tests for the acuteness of sensory perception were compromised by both practical and theoretical challenges. The faculties of sense and perception could easily be conflated. For example, the Instrument Company designed an apparatus for testing the appreciation of musical pitch, in which the difference between two musical notes was reduced until a subject could not differentiate between them. The apparatus was composed of an organ pipe of variable length, producing two tones separated by a predefined period of time. However, this instrument could not differentiate between physiological and perceptual factors in hearing. Further, the question of pitch sensitivity was revealed to be more complex than previously imagined, as the appreciation of pitch varied within individuals at different frequency ranges. The use of the equipment also required a control of environmental factors that was difficult to achieve: ‘Care should be taken to keep the air pressure constant during each experiment.’
The difficulty of separating physiological and perceptual factors also applied to the evaluation of a subject’s eyesight. Tests involving the appreciation of colour and length could not avoid conflating vision and perception. The Instrument Company’s 1899 catalogue featured an apparatus for measuring the maximum distance at which a subject could read printed letters in small type: ‘A carriage containing the card of test type, with a candle for illuminating the card, slides along the graduated bar.’ Whilst this apparatus controlled for the ambient light in which the test took place, it could not control the lighting conditions to which the subject had previously been exposed, or factors such as fatigue. Previous instruments had been undermined by the influence of schematic factors in visual perception:
At the Health Exhibition a piece of an easily procured edition of the New Testament in small type was adopted; but Mr F. Galton thought that a series of figures set at random would give a more accurate test, as the effect of memory would be thus eliminated and the instrument could be used by persons not knowing English.
Whilst the integrity of biometrical data was the product of considerable labour on the part of instrument designers and practitioners, the most significant challenge to the anthropometric programme was the difficulty of establishing significant correlations. To give meaning and significance to the mass of anthropometric data, a demonstrable connection with other measurable data was needed. Galton, for example, wanted to explore variation in anthropometric results between generations. Whilst this did not require investigators to identify relationships with factors external to the data, it was undermined by the time scale such an investigation required: the instrumentation and methods of measurement changed more rapidly than the bodies of human subjects.
Later researchers such as Karl Pearson attempted to illustrate a correlation between the physical and mental characteristics of his subjects. However, the measurement of intelligence presented insuperable obstacles, and such investigations produced disappointing results. The data produced by the Cambridge Philosophical Society was compared to the corresponding degree classifications that its subjects achieved. On 24 April 1888, John Venn gave a talk to the Anthropological Institute, conceding:
We find then that, in regard to all the ordinary elements of health and strength, there does not seem to be the slightest difference between one class of our students and another: that is, they are equally tall, they possess the same weight, the same muscular strength of hand, and the same breathing capacity.