Relation of the pelvis to erect posture. An exploratory study

RELATION O F THE PELVIS TO ERECT POSTURE

AN EXPLORATORY STUDY

EDWARD REYNOLDS AND E. A. HOOTON Peabody Mweum, Earvard University

Presumably every scientist has his own private morgue in which repose the desiccated cadavers of his unsuccessful research operations. The results of post-mortem examinations of these defunct enterprises are rarely published, possibly because an inquest presupposes a crime. Such suppressio veri is a manifestly unfair procedure, since it prevents scientific colleagues from profiting by the mistakes of their fellow research students. These mistakes commonly see the light of day only when their perpetrators are unaware of failure and consequently present their abortive products under the erroneous impression that they are fullfledged and creditable off spring.

The authors of this paper labor under no such delusion with respect to the attempt at research here exposed. Humbly and (if the truth must be told) reluctantly, they venture to describe a frustrated investigation upon a subject of basic importance, impelled by moral responsibility and a dim hope that their bitter experiences may be of some value to future students of greater competence.

The variations of the human pelvis with reference to the erect posture are virtually unknown. There are, of course, numerous speculative inferences drawn from dried bones of individuals of unknown posture and from living persons whose skeletal structure cannot be observed. By piecing together these discrete masses of inferential data, certain generaliza- tions concerning the relation of the pelvis to the erect posture

253

AXEBICAN JOURNAL OF PHYSICAL ANTHROPOLOQY, VOL. XXI, NO. 2 AND SUPPLEMENT APRIL-JUNE, 1936

254 EDWARD REYNOLDS AND E. A. HOOTON

have been promulgated and have gained a considerable measure of credence. For example, it is commonly believed that the female has a more pronounced lumbar curve than the male and a pelvic brim more sharply inclined from the horizontal. Neither of these beliefs can be verified until some technique has been devised for ascertaining pelvic and vertebral variations in living, standing, and walking subjects.

Obviously the x-ray offers the only possibly successful approach to this problem. With this unoriginal idea in mind, we some years ago undertook an investigation which has nearly brought down our gray hairs in sorrow to the grave. As first we grappled with the problem enthusiastically and devised a technique which appeared to be promising. We still think that this technique has many merits, as well as a number of grave defects unappreciated until correction was impos- sible. Even now it seems to us probable that the uncertain and unsubstantial nature of our results is due principally to the numerical inadequacy of the series studied, rather than to fundamental errors in method. After crude laboratory observations had been secured, not without difficulty, upon a pitifully small group of subjects, our research assistant, Dorothy M. Johnson, expended an enormous amount of time and skill upon the very arduous and delicate task of recon- structing the corrected drawings of the pelves. Subsequently she also performed the drudgery of statistical reduction of the data. It then devolved upon the authors to attempt to derive some conclusions from the material obtained. We manipulated the data in almost every conceivable way, hoping that by trial and error we might eventually emerge from the morass in which we had become bogged. We have repeatedly laid aside the work in disgust for months at a time, and as often have been flogged back to it by our guilty consciences. The present paper is a desperate attempt to liquidate what- ever of assets this business possesses-probably at the rate of about 5 cents on the dollar, considering the time and worry spent upon it, not to mention a grant of money for its prosecution made by the American Medical Association.

THE PELVIS AND ERECT POSTURE 255

As a result of this prolonged travail we have learned some- thing about the human pelvis in relation to the erect posture, but, frankly, we do not think that any of our results can be regarded as substantial. They are merely more or less inter- esting suggestions, advanced on tenuous evidence. A considerable mass of observations and statistical data have been omitted entirely from this paper on the ground that they are not worth printing. All that we consider of possible value has been summarized with the minimum expenditure of words and figures.

The authors wish to acknowledge their indebtedness to the half hundred of amiable individuals, mostly students of Harvard University and Radcliff e College, who submitted themselves to the ordeal of examination. A good deal of the laboratory observation was carried out by Miss Harriet Hammond and by Mr. Fred S. Hulse, to whom we are very grateful. Dr. Ariel George of Boston most liberally placed at our disposal the facilities of his x-ray laboratory and contributed to our research all the roentgenograms secured. Without his cooperation the work could not have been carried out. Dorothy M. Johnson, as previously stated, did the lion’s share of the analytical work. Her services, contributed for more than a year, were like the gospel, “without money and without price.” If there accrues to this research any modicum of credit, Mrs. Johnson should have it. Finally, we have to thank the American Medical Association, to whom we feel most grateful for its liberal grant in aid, its long enduring patience, and its misplaced confidence in our ability.

METHODS AND TECHNIQUE

Our examination of each subject involved the obtaining of: a. Certain anthropometric data. b. The position of the center of gravity in its relation to

the base of support in the standing position. c. Lateral and anterior photographs. d. Lateral and anterior roentgenograms.


e. The shape and position of the pelvis and its relation to the position of the center of gravity in the standing position of the subject.

After the usual notations of name, age, serial number, nationality and parentage, we measured the height of the subject in centimeters, the height from the floor to the head of the fibula and its horizontal distance from an erect standard to be referred to later, the height and similar horizontal distance of the external malleolus, and the height of the iliac crest. Here we took only one crest, but subsequently regretted not having taken both. We measured the bi-iliac, the head length and width, and took the total weight of the subject on a standardized scale. We next determined the position of the center of gravity in the erect attitude of the subject and in its relation to the base of support. This relation was obtained by a machine which one of us had used with satisfactory results in previous work, and which is illustrated in figure 1, in which the subject is in position for the determination of the antero-posterior position of the perpendicular dropped from the center of gravity to the base.

The subject stands upon a board, one end of which is supported by a knife edge in fixed position upon the base at a known distance from an upright arm and received by a groove upon the under surface of the board. The other end of the board is supported by a knife edge placed upon the under surface of the board itself and resting upon the platform of the scale which previously recorded the total body weight. These knife edges are, for convenience, 50 cm. apart. On the upper surface of the board, most conveniently at one-fourth of the distance between knife edges, there is a strip of wood attached transversely as a heel rest and at right angles to this another, the two forming a letter ‘L’ and serving as heel rests in the two positions of the subject. A sheet of paper is placed in this ‘L’ and the subject stands upon this paper with the heels against the heel rest. The outline of the feet is then traced upon the paper and the antero-posterior relation of the perpendicular dropped from the center of gravity to the


base of support is known by the following principle: The total weight is to the balance weight as the distance between the knife edges is to the position of the center of gravity.

Fig.1 Machine for the determination of the center of gravitp.

This is figured for the individual case by the following formula :

Standing antero-posterior c e d e r of gravity Total weight (W) . . . . lbs. Balance weight (w) . . . . Ibs. L=distance

m=moment of board between knife edges=50 em. (i.e., effect of board alone on the scale).

f.h.= fulcrum to heel.

50w-m Formula - f .h .=X


In this the distance called fulcrum to heel is a constant, being the distance from the knife edge to the heel edge of the heel rest. X then represents the distance from the heel rest to the center of gravity, and its position as calculated is subsequently marked upon the tracing of the feet.

The lateral relation of the center of gravity is then found by placing another sheet of paper in the 'L' of the heel rest, standing the subject upon it with the heels against the other arm of the 'L,' tracing the outline of the feet upon this paper and applying a similar formula.

Standing lateral center of gravity L' = distance

between knife edges=50. f.h.=fulcrum to A.P. heel rest. m=moment of board.

Total weight (W) . . . . lbs. Balance weight (w') . . . . lbs.

50 (w'-m) Formula - f .h. = y em.

This completes the relation of the perpendicular from the center of gravity to the base of support, but for completeness of observation we added a determination of the height of the center of gravity from the base, which was taken by placing the subject supine upon a 6-foot board similarly fitted and supported.

Supine center of gravity Total weight (W) . . . . lbs. Balance weight (w") . . . . lbs. L"=distance

between knife edges. m=moment of board. f.h. =fulcrum to heel.

f.h. (w"-m) Formula = z em.

The determinations so made seem to have been satisfactory and accurate, but certain precautions are necessary in taking them. When the position is first assumed, the subject sways backward and forward quite actively, but after a few moments this sway decreases and observation of the center can be taken, although some slight sway always remains.l As the body sways forward the weight recorded on the dial increases; as the sway is backward it decreases and the mean must be ob-

'The amount of sway can be lessened by making the depth of the forward knife edge such that the board is approximately horizontal when the platform of the scale has been lowered by a subject of average weight.


served. Here is the largest opportunity for error of observation due to personal equation. We found, however, that if two observers, each of whom has been trained with a number of subjects, estimated this mean position independently, they almost always arrived at the same figure, and if there was some slight discrepancy, we assumed the mid-point between their estimates. This mean balance weight having been de- cided upon, the horizontal slides were run forward to touch the subject lightly when at that balance weight, and were fixed in that position by set screws. The subject was then urged to maintain that light contact during the remainder of the observations and one of the observers watched to see that he o r she neither leaned heavily nor swayed away from the rests.

Lateral and antero-posterior photographs of the subject when so fixed in position were taken in each case, and for convenience in using them, a panel marked in gridiron at 10 cm. intervals was used as a background.

The next and h a 1 step in our examination of the subject was to take two roentgenograms at right angles from each other (i.e., lateral and antero-posterior).

As our laboratory has no x-ray plant, we were obliged to take the roentgenograms in each case at another time and place and to accommodate ourselves to conditions there. On this account we supplied the upright and arms which fixed the subject’s position upon the balance machine with millimeter graduations by which we could reestablish the position upoii another stand which we used for the roentgenograms. This accessory stand consisted of a revolving platform fitted with similar upright and arms and with a movable heel rest so that the ascertained center of gravity could be placed directly over the center of rotation of the stand thus corresponding to a vertical wire which was affixed to the buckie and is seen in each plate. The lateral and antero-posterior x-rays were then taken without moving the Crooke’s tube o r buckie. The height of the x-ray tube and plate and the distance from each other were noted for each exposure. I n an anthropometric


THE PELVIS A N D ERECT POSTURE 261

262 EDWARD REYXOLDS A N D E. A. HOOTON

laboratory supplied with an x-ray plant it would be much simpler, and possibly more accurate, to take the roentgenograms while the subject was still upon the center of gravity machine. By moving the x-ray and buckie to certain stops upon the floor this could easily be done without loss of time.

By either method two x-rays should be taken, one exactly from the norma lateralis and the other from a point directly anterior. The fact that the position of the perpendicular dropped from the center of gravity is shown on each plate by

Fig. 3 Apparatus for correcting the distortion of x-rays.

the shadow of the wire attached to the bncltie is a great con- renieme during the subsequeilt work.

The first glance a t a lateral x-ray priiit (fig. 2 ) shows that the two sides of the pelvis as developed by the x-ra~7 are of very different shape and size, and a comparison of this with the antero-posterior print makes it evident that both rcpre- sentations of the pelvis are so greatly distorted that deduc- tions drawn from them as showii would he entirely uni~eliahlt~. In fact, 110 satisfactory stucly of pelvic rclations caii be maclc without correction of these opticaal distortions. Fo r this pur- 1)ose we constructed tlic al)pai.ntus s1ion.n in figure 3. The


table is notched to afford the technician easy access to the two drawings which are to be made. The apparatus consists fundamentally of two rigid arms a t right angles to each other supported at the outer ends by standards which can be moved up, down, or horizontally and which permit free movement of the arms in any direction as well. One of these arms rests at their intersection upon the other arm and the lower one is supported near its inner end by a projecting ledge so attached to a movable upright that it can be set a t any desired height and position. One edge of each arm ends in a sharp point and these two edges are adjusted to cross each other. Outline tracings of the two x-ray prints2 are attached to vertical stands and placed exactly at right angles to each other with the median line shown by the wire on the buckie directly opposite the standards, a t the same distance from the standard as the plate had been from the Crooke’s tube, and with the outer ends of the arms raised up on the standards to the height of the Crooke’s tube in the individual case.3 Three sheets of analytical paper are attached to the table in the positions shown in the figure with the center lines of the middle paper opposite the center lines as shown by the buckie wire. The large brass square shown in the figure and the ‘T’ square lying beside it complete the apparatus. The large square carries, attached to its angle, a graduated upright, each edge of which is exactly opposite a projecting point upon the end of the corresponding arm of the square and each of these points is distant from the corresponding edge of the upright by just the length of the analytical paper.

When the apparatus is so adjusted, the edge of each horizontal arm will show the path taken by the x-ray from the Crooke’s tube to the spot upon the tracing upon which its point is placed. Whenever the points of the two arms are placed upon the same anatomical point in the pelvis, as shown in the tracings, the point at which the edges of the arms cross ‘In practice these were found more convenient than the prints themselves. ‘It is evident that the height of the centers of the plate and tube must vary

up and down with the varying heights of the subjects.

AMERICAN JOURNAL O F PEYSICAL ANTHROPOLOGY. VOL. XXI, NO. 2 AND SUPPLEMENT


will show the true corrected position of that anatomical point. If the large brass square is then placed so that the edge of the graduated upright attached to it is in contact with the angle between the two arms, and the sides of the square are parallel to the lines upon the paper, the points of the projection which correspond with that edge of the upright will be upon those lines of the two outlying analytica.1 papers on which that anatomical point should be represented and the height shown upon the graduated upright will give the position of the anatomical point upon those lines. In determin- ing this some observers will find the small ‘T’ square a convenience.

If this apparatus is used to correct x-rays of a dried pelvis on which successive points along each line which is to be represented on a corrected drawing have been marked in metal, the results obtained are of necessity mathematically correct and can be obtained with ease and rapidity. But in work upon the living subject, such marks cannot be placed upon the pelvis. The process of making corrected drawings is therefore much more tedious and correctness can be obtained only by undergoing the tiresome ordeal of a long series of trials and errors. There are many anatomical landmarks in the pelvis that are not seen by quite the same points upon the bone in plates taken from the two different points of view, but there are fortunately a few which are exactly the same in each tracing, e.g., the middle of the promontory and the points of the ischial spines. Such points can be selected for each line by visual inspection of the dried pelvis from the two points of view. With these as initial data, a correct outline can be obtained.

For this purpose the two best established points upon the given line are selected and marked upon the papers. Then each arm is directed to a point midway between them on each print and the result jotted down in pencil; then to the quarter points, et cetera. If the points of both arms are at any of these spots placed exactly upon the same points upon the bone, the new dot upon the paper will be correct and a succes- sion of such dots will make a fair curve. But if any one point


as measured upon one tracing is not precisely the same as that measured upon the other, as will often happen from the differ- ing distortions, the resulting dot will be more or less out to one side and one of the other arms must be moved slightly until the dot comes into place for a true curve. At first the whole process seems not only endless but hopeless. One feels as if any one of many curves could be so obtained and a t first it may be necessary to adjust many of the dots again and again. But with practice the technician soon learns to know which arm should be moved in each case and about how much. Our results and conclusions have been drawn from tracings so obtained.

It is to be noted that if from any cause the subject shifts his attitude between the times of taking the two x-rays, the machine at once shows this by the impossibility of obtaining a true curve and, as happened in several of our series, these prints must be thrown out of consideration.'

The center of the base of support was determined as follows. The outline of the feet fled with the examination sheet was pricked out on a piece of Bristol board, together with the position of the center of gravity. The Bristol board was then cut along the outer edges of the foot print and the straight lines tangent to the heels and the most anterior projections of the toes. This roughly trapezoidal piece of Bristol board represents the area of the base of support. The Bristol board was then suspended loosely from a vertical plane at two successive points by means of thumb tacks from which hung a plumb line. The intersection of the two lines drawn on the Bristol board marking the respective positions of the plumb lines is the center of the base of support.

'It first seemed to us that the stereoscopic x-ray plates which came iuto use after our work had been completed might have been much more convenient in the subsequent work. However, on further reflection, it seems evident that although the mathematical formula by which the true length of a diameter can be obtained from them no doubt furnishes a sufficiently correct result for many single lines, the construction of the true shape of the whole pelvis by this method would be but little, if any, easier or more rapid, and that, since the limit of error from views taken at a small angle from each other is much greater than from a pair at 90" separation, the result would probably be much less correct.


When all these data had been obtained, we were in a position to a.ttempt their interpretation.

ANALYSIS OF METRIC DATA

Our total utilizable series consisted of nineteen adult males and twenty-f our adult females. A considerably larger number of each sex was observed in the laboratory, but in a fairly large percentage the data were incomplete, either because the x-ray films of the pelves were unsatisfactory, or because the subjects did not fulfill their appointments a t the x-ray laboratory.

Table 1 presents the means and other statistical constants of those measurements upon the living and of the measurements and indices taken from the reconstructed pelvic drawings which seem to be worth recording. A brief commentary upon the implications of these data follows.

A g e , weight, height

In general the series represents young adult Americans of college age, although the mean age of the males is elevated by the inclusion of one 50-year-old person. Height and weight conform to expectation for a sample of the college population.

M a x i m u m transverse diameter between. superior borders of Acetabula

2.04 mm.) is only 2.16 times its probable error and hence statistically unreliable. Nevertheless it may be a true sex difference re- lated to the greater depth of the acetabula and the larger diameter of the femoral heads in males.

The difference in favor of the males (4.41 mm.

Minimum diameter betzcreerc femoral heads

1.71 mm. The difference is 3.80 times its probable error and hence significant.

The females exceed the males by 6.50 mm.

TABLE 1 Zeans of measurements on the subjects and, pelvic reconstrudwns of the subjects

Weight

Height

Maximum transverse diameter vs. superior borders of acetabulum

Minimum diameter vs. femoral heads

Maximum bi-ischiatic diameter

Maximum bi-trochant,eric diametei

Maximum transverse of sacrum

Maximum transverse diameter oi brim

External bi-iliac d i e t e r

Bi-iliac diameter from x-ray reeon structions

Angle of inclination of lumbai curve with vertical center 01

gravity (5th lumbar)

Angle of inclination of lumbz curve with vertical center 01

gravity (4th lumbar)

Angle of inclination of lumba curve with brim inclination (5tl lumbar)

Angle of inclination of lumba curve with brim inclination (4tl lumbar )

Index between lumbar curve ant perpendicular to its axis (5tl lumbar)

Index between lumbar curve ani perpendicular to its axis (4 t lumbar)

Angle of brim

mMBEB - 19 22

19 24

19 24

19 24

19 24

16 24

1 2 20

18 24

19 24

16 24

16 24

15 24

8 12

15 24

8 12

15 24

8 12

19 24

-- BaNQB

21- 50 21- 30

115-175 98-147

166-184 153-174

188-225 183-220

108-139 116-149

129-178 143-181

271-327 264-310

96-131 103-126

106-138 113-144

260-325 248-305

240-284 238-297

36- 59 33- 70

34- 50 23- 56

78-109 71-113

88-120 84-119

11- 22 13- 22

18- 24 16- 2E

43- 65 40- 71

MEAN

25.47 f 1.07 23.41 f 0.39

47.95 f 2.69 26.25 f 1.81

74.90 f 0.78 63.29 f 0.64

04.74 f 1.82 00.33 f 1.30

23.58 f 1.39 30.08 f 1.00

.50.44 f 1.97 60.62 f 1.22

196.42 f 3.01 187.90 f 1.67

13.17 f 1.32 .14.12 f 0.71

.21.79 f 1.40

.28.29 f 0.92

!82.88 2 2.79 !76.62 f 2.08

!63.62 f 2.21 162.38 f 2.02

51.00 f 1.12 47.96 f 1.26

40.75 f 1.17 38.67 2 1.81

92.87 f 1.39 95.00 f 1.37

104.00 f 2.39 101.00 2 1.76

16.13 f 0.5: 16.71 2 0.32

20.75 f 0.45 20.92 f 0.54

55.90 f 1.04 53.12 2 1.01

S.D.

6.89 2 0.75 2.72 f 0.28

17.41 f 1.90 13.12 f 1.28

5.02 2 0.55 L.66 f 0.45

1.78 f 1.29 3.48 2 0.92

9.87 f 0.98 7.27 f 0.71

1.70 f 1.40 5.86 f 0.86

5.45 f 2.13 1.05 f 1.18

8.29 f 0.93 5.13 f 0.50

9.03 f 0.99 6.68 f 0.65

6.53 f 1.97 5.07 f 1.41

3.10 2 1.56 4.68 f 1.43

6.41 f 0.75 9.12 f 0.81

4.89 f 0.8: 9.28 f 1.21

7.99 f 0.91 9.92 f 0.9:

0.01 f 1.6! 9.06 f 1.2!

3.18 f 0.3! 2.39 f 0.2:

1.98 f 0.3: 2.75 f 0.3;

6.70 f 0.7, 7.90 f 0.7

V.

!7.05 f 2.96 .1.62 f 1.18

l.77 f 1.29 .0.39 f 1.01

2.87 f 0.31 2.85 f 0.28

5.75 f 0.63 4.73 f 0.46

7.26 f 0.79 5.59 f 0.54

7.78 f 0.93 5.52 f 0.54

5.21 2 0.72 3.84 f 0.41

7.32 f 0.82 4.50 f 0.44

7.41 f 0.81 5.21 f 0.51

5.84 f 0.70 5.45 f 0.53

4.97 f 0.59 5.59 f 0.54

12.57 f 1.55 19.02 f 1.85

12.00 -t 2.02 24.00 2 3.30

8.60 f 1.06 10.44 2 1.02

9.62 -t 1.62 8.97 f 1.24

19.72 f 2.43 14.30 f 1.39

9.54 f 1.61 13.14 f 1.81

11.99 f 1.31 14.87 f 1.45

26 7


Maximum bi-ischiatic diameter

The difference in favor of females (10.18 mm. k 2.32 mm.) is 4.38 times its probable error and hence may be accounted real.

Bi-trochant eric diameter

This measurement was not taken externally because it is inaccurate. On the pelvis the mean diameter of males exceeds that of females by 8.52 -t- 3.44 mm. This difference is not certainly significant since it is only 2.44 times its probable error. Comparative data in Martin relative to external bi- trochanteric diameters of German women indicate that the latter have much broader hips than the American college girls. We should have to allow 17 mm. for thickness of the soft parts on each side to bring our females up to Martin’s German averages and 32.5 mm. on each side to make them equal the means of Bach, quoted by Martin5

Maximum tramverse o f sacrum The difference in favor of females is less than 1 mm. and

consequently insignificant. Comparative data taken by direct measurement upon dried bones indicate that there are no consistent sex differences in this measurement and that. our means derived from pelvic reconstructions are 1 to 3 mm. below averages for whites.

Maximum tramverse diameter of brim

The female mean of the transverse diameter of the pelvic brim exceeds that of the males by 6.50 mm. k 1.67 mm. This difference equals 3.89 times its probable error and is significant. Comparative measurements on dried pelves again indicate a slight inferiority of the dimensions derived from our pelvic reconstructions.

6Martin, R., Lehrbueh der Anthropologie, 1928, p. 358.


Bi-iliac diameter (exterrcal and x-ray)

Our series consisted fo r the most part of slender individuals who were rather tall. The external bi-iliac diameters are small when compared with those of many other observers. This fact conforms with the small diameters of the reconstructed pelves. The difference in external bi-iliac breadth is 6.26mm. 3.48 mm. in favor of the males. It is not statistically reliable. The males exceed the females in x-ray bi-iliac diameter by only 1.24 mm. which is less than its own probable error.

The external bi-iliac diameter exceeds the x-ray diameter by 19.26 mm. in males and by 14.24 mm. in females. These excesses represent the thickness of the soft parts if measurements and reconstructions are correct. The difference of these differences between males and females is 5.02 mm. or 1.09 times its probable error and hence unreliable.

The coefficients of correlation f o r external and x-ray bi-iliac diameters are 0.87 -e 0.03 for twenty-four females and 0.76 f 0.07 for sixteen males.

Axis of the lumbar curve and indimation of the sacrum

We were unable to measure the lumbar curve, because the x-ray films were not large enough to include the entire pelvis and lumbar column. Several measurements of the lumbar- sacral axis were devised and executed, but none was satisfactory. These measurements involved the drawing of an axis from the most anterior superior point of the fifth (or fourth) lumbar vertebra,, to the inferior anterior point on the sacrum intersected by the ilia (approximately the posterior inferior iliac spines). The angle of this axis with the vertical representing the axis of the center of gravity was then laid off. Evidently it would increase with greater anterior projection of the fifth lumbar vertebra and with more pronounced backward tilting of the inferior part of the sacrum. Whether this axis was laid off from the fourth or fifth lumbar, it nevertheless resulted in insignificant superiorities of the male


means over the female. If these results were valid, they would indicate a sharper lumbar convexity and a more pronounced sacral tilt in males than in females. Inspection of lateral photographs of the subjects, and consideration of brim inclination indicate that this statistically unsound result may be valid, so far as our small series are concerned.

This lumbar-sacral axis was also measured in its angular relation to the slope of the pelvic brim. The value in males was insignificantly smaller than in females in the case of axes measured from the fifth lumbar, and slightly larger for axes measured from the fourth lumbar. Thus it is hinted that the male lumbar column at its most anterior convexity juts more abruptly forward toward the center of gravity in males than in females, and that the fZth lumbar-sacral axis ap- proaches close to the axis of pelvic brim inclination. The reverse result in the case of the fourth lumbar-sacral axis would indicate that the convexity of the lumbar column is shorter and sharper in males than in females.

The authors now think that a much better method would have been to measure separately the inclination of the sacrum, and the axes of the anterior faces of the lumbar vertebrae. This procedure would have made possible the determination of the sacro-vertebral angle. The axes employed are com- posite, and ambiguous in their mechanical and morphological significance.

We did indeed attempt an estimate of the prominence of the sacro-vertebral angle by taking the ratio of the perpendicular from the sacral promontory to the lumbar-sacral axes previously described. No significant sex differences were found by this method.

Angle of brim imlircation The angle of inclination of the pelvic brim cannot be de-

termined externally with any satisfactory degree of accuracy. In the present investigation the axis of the ilio-pectineal line from the pubic symphysis to its juncture with the sacrum was determined. The inclination of that axis from the horizontal


was then measured. The angle thus obtained is somewhat lower than that of the conjugata Vera. The results show a mean angle for females which is 2.78" f 1.51" less than the corresponding mean for males. However this difference is not statistically significant.

Correlation of angle of brim i!aclimtion with aagle of femoral shaf t axis

An attempt was made to correlate the angle of inclination of the brim from the horizontal with the angle of deviation of the femoral shafts from the vertical. The larger the angle of brim inclination the more steeply inclined is the pelvis; the smaller the angle of femoral shaft axis the more markedly the femora are inclined to slope from above downward and forward. The femoral axis could be determined in only thirteen males and twenty-four females. The coefficient of correlation for the former was found to be - 0.40 f: 0.16 and for the latter -0.27 0.13. These results are hardly ponderable.

Taken at their face value they might suggest that persons with steeply tilted pelves tend to stand w3h the femora less fully extended than do those whose pelvic brims approach more closely to the horizontal.

Cowelatiow of angle of brim imlinatiow with the awtero- posterior position of the axis o f the center of gravity

The axis of the center of gravity is represented by a straight vertical line in the tracings of both lateral and antero-posterior roentgenograms. Its anterior or posterior position with reference to the middle points of the upper rim of the acetabula was measured in the lateral pelvic reconstructions. When the axis passes in front of the middle of the acetabula a positive value is recorded in millimeters. When the axis passes behind the acetabula the value is negative. These measurements were then correlated with the angle of inclination of the pelvic brim. The coefficients of correlation were:


Twenty-four females, - 0.55 -c 0.10; sixteen males, - 0.41 -r- 0.14. These correlations are statistically of probable significance, and express the extent to which increasing degrees of forward and downward tilt, of the pelvis are accompanied by shifting forward of the axis of the center of gravity.

OBSERVATIONS AND ASSOCIATIONS

A considerable number of points of interest lend themselves to categorized observations, but not to measurements. In such cases, the authors, having agreed upon technique, proceeded, each separately, to make the entire list of observations upon each subject. The results were then compared and equated. These observations have been combined in association tables and the nature of the relationships has been sub- jected to analysis. Most of these elaborate tables yielded no results which would justify the expense of printing them. Only one of them is reproduced here, and the trends of the rest are summarized in the text.

Relation of antero-posterior position of asis of cemter of gravity t o variations in relative posterior projection

t o shoulders a d buttocks Table 2 records the associations between three categorized

positions of the axis of the center of gravity (anterior to acetabula, through acetabula, back of acetabula) and three variations of the shoulder-buttocks relation (shoulders back of buttocks, buttocks back of shoulders, even plane) as judged from lateral photographs of subjects standing in front of a grille. Analysis of the interrelationships is carried out by comparing the percentage frequencies of each individual cell in the tables with the total percentages of rows and columns. The following facts may be observed from a study of the table: 1) In both sexes and in the combined series there is a positive association of an anteriorly placed axis of the center of gravity with the buttocks back of the shoulders; 2) conversely, in both sexes there is a positive association of


YAm

Number I %

YALE AND FEMALE

Shoulders back 2 Even plane 3 Buttocks back 7

Total 12

66.67 I 1 25.00 3 42.86 60.00 0 0 3 42.86 70.00 9 50.00 16 57.14

66.67 10 41.67 22 52.38

Shoulders back 0 0 Even plane 1 20.00 Buttocks back 3 30.00

Total 4 22.22

Relation of the antero-posterior position of the axis of the center of gravity to amount of lumbar curve

The lumbar curve of each individual was classified by observation of his lateral photograph. The criteria employed were: Depth of the curve as judged from the shoulder buttock tangent, sharpness of the curve as estimated from the

2 50.00 2 28.57 2 100.00 3 42.86 7 38.89 10 35.71

11 45.83 15 35.71

Shoulders back Even plane Buttocks back

Total

1 33.33 1 25.00 2 28.57 1 20.00 0 0 1 14.29 0 0 2 11.11 2 7.14

2 11.11 3 12.50 5 11.90

Shoulders back Even plane Buttocks back

Total

3 16.67 4 16.67 7 16.67 5 27.78 2 8.33 7 16.67 10 55.56 18 75.00 28 66.67

18 100.00 24 100.00 42 100.01

a posteriorly placed axis of the center of gravity with the shoulders back position ; 3) the buttocks back of the shoulders position predominates in both sexes, but especially in females ; 4) a position of the axis of the center of gravity anterior to the acetabulum predominates in both sexes, and is commoner in males.

TABLE 2

Relation of antero-posterior center of gravity ia pelvis to shoulder-buttock plane

274 EDWARD REYNOLDS A N D E. A. HOOTON

lumbar-sacral angle. The categories of the curve recognized were small, medium and pronounced. The association table (not published here) shows that, on the whole, the farther forward the axis of the center of gravity falls, the more pronounced is the lumbar curve. As the lumbar curve becomes feeble, the axis of the center of gravity shifts backward.

Relation of antero-posterior position of the mis of the celzter of gravity to degree of pelvic tilt

This association table merely confirmed our previous find- ing that the more sharply tilted the pelvis, the further anterior to the centers of the acetabula is the axis of the center of gravity.

Relation of lateral displacements of the axis of the center o f gravity t o variations of the relative height

of the pelvic sides

Another of our association tables shows that the left side of the pelvis is usually higher in both sexes (72.09 % of the combined series) and that pelves in which both sides rise to the same level, and those in which the right side is higher, occur in approximately equal proportions of both sexes. There is no consistent association between the direction of lateral displacement of the axis of the center of gravity and superiority in height of one or other side of the pelvis.

R e h t i o n of the antero-posterior position of the axis of c e d e r of gravity in the pelvis to amtero-posterior position

of center of gravity i m the base of support

In another table the antero-posterior position of the center of gravity in the base of support was calculated as a percentage of the antero-posterior diameter of the base of support. The positions were then divided into three categories: Posterior, median and anterior. These categories were then associated with the categories of antero-posterior position of the axis of center of gravity in the pelvis. No relationship


between the various positions of the center of gravity in the pelvis and in the base of support was revealed by this association ta.ble.

Relation. of pelvic rotatiom t o long leg The antero-posterior pelvic reconstructions show which

side of the pelvis is higher and which leg is longer. The lateral reconstructions show which side of the pelvis is twisted forward (from the relative position of the acetabula). Long left leg occurs in the majority of both sexes (63.16 % of males; 73.91 % of females). Legs of substantially equal length occur in approximately the same proportion of each sex (21 % ) . In both sexes the right acetabulum is generally twisted forward (57.89 % of males; 69.57 % of females). About one- third of each sex shows no pelvic rotation. There is no dis- cernible relation between pelvic rotation and long leg.6

Relatiort of lateral displacememt of ce.lzter of gravi ty im base to lolzg leg

About one-third of the subjects of each sex have the center of gravity in the middle of the base, a third to the right of the middle, and the other third to the left. These variations in position seem unrelated to long leg.

Relatiort of la terd pelvic thrust t o long leg In both sexes the axis of center of gravity in the pelvis

is more frequently displaced to the left than to the right. This displacement is apparently independent of the long leg.

Classification of sub jecEs ( f r o m photographs) imto postural types

The authors spent a great deal of time in attempting to classify the subjects into postural types, for the purpose of investigating metric data concerning the center of gravity

OThe possible relation of these variations to the very common lateral distortions of the spinal column was not investigated.


and pelvic relations to such types. In the case of the females the following classifications were adopted and carried through by the use of the frontal and lateral photographs of each subject. Z. Feminine type. Characterized by buttocks much poste-

rior to shoulders, concavity of back mainly or exclusively in lumbar region, knees strained backward in lateral photograph, short chest and long lower abdomen distinguishable in both views, but especially in the antero-posterior view. 2. Sway-back type. Shoulders on a plane with buttocks, o r

behind them; back hollow with concavity extending up into dorsal region, knees without back strain ; thoraco-abdominal proportions varying.

3. Straight or columnar type. Figure columnar through- out; no knee strain, concavity of back slight and general, buttocks little, if at all, posterior to shoulders; thoraco- abdominal proportions variable.

4. Nofindescript. Generally intermediate between feminine and straight.

The males were classified, probably unsatisfactorily, in two groups: a ) with dorsal concavity chiefly or wholly confined to the lumbar region; b) with the concavity of the back extending into the dorsal region.

Table 3 presents the meager results of this classification. The figures for the females suggest that the straight or columnar type of female figure has a center of gravity further back in the base of support (nearer the heels) and a lesser degree of pelvic inclination than the other imperfectly distinguished types. The so-called sway-back type seems to present a more anterior position of the axis of center of gravity with reference to the acetabula, and a possibly more pronounced pelvic inclination than the other types. But these inferences are dubious, in view of the inadequate numbers of individuals represented in the several types. The figures for males are not instructive, except in so fa r as they show the much more anterior position of the pelvic center of gravity (with relation to the acetabula) than is found in the females.


C. of G. in relation to 5th

(a) Antero-posterior length of

(b) Center of gravity to heel (b)/(a) Percentage of foot

C. of G. in relation to

metatarsal

foot

length

acetabula Indination of pelvis Weight

One of us believes, as the result of long continued clinical observation, that examples of the feminine type are far more common in the community at large than in our series. It should be noted too that our males were also all college students,' and, as it happened, exclusively of the intellectual type. We do not think that our series embodied a satisfactory cross section of the community as a whole.

-~ Feminine

( 5 )

+ 2.0 23.8 10.2

42.94

+ 1.5 54" 120.5

TABLE 3

Center of gravity relations by postural types

I

+ 2.9 24.4 10.3

42.04

+ 1.9 51" 132.5

F E M A b E ( t Y E A N B

+ 1.5 24.1 9.9

41.01

2.0 53' 125.3

.-

Swapback (6 )

+ 1.3 23.9 10.1

42.38

+ 3.4 55" 126.3

Straight ( 8 )

+ 0.6 24.2 9.2

38.26

+ 1.5 51" 123.9

I MAL-MEANS

- C. of G. in relation to 5th

(a) Antero-posterior length of

(b) Center of gravity to heel (b)/(a) Percentage of foot

C. of G. in relation to

Inclination of pelvis Weight

metatarsal

foot

length

acetabula

Lumbar ( 14)

+ 2.4 26.1 10.7

40.94

+ 28 56" 153.2

Dorsdumbar ( 9 )

+ 2.2 27.2 9.9

36.35

+ 33 55" 153.0

Total (23)

+ 2.3 26.6 10.4

39.14

+ 30 56" 149.2

SUMMARY

The authors have attempted to investigate the pelvis in the living in its relation to the erect posture. A small series of subjects was measured, photographed, and the antero- ' The one exception had an unsatisfactory x-ray.


posterior and lateral positions of .the center of gravity were determined. X-ray photographs of the pelvis were taken with reference to the position of the center of gravity. A n apparatus was devised whereby the distorted x-ray views of the pelvis were corrected so that accurate reconstructions of the pelvis in antero-posterior and lateral aspects were obtained.

The metric data resulting from measurements on the living and on the pelvic reconstruct,ions were analyzed. Few results of certain validity were obtained; probably because of the inadequate number of subjects studied. Apparently, most of the ordinary sexual differences found in dried pelves were demonstrable upon the pelvic reconstructions. Our small series seemed to suggest, however, that pelvic inclination is perhaps slightly greater in males than in females, and lumbar convexity shorter and sharper. Such a &ding might be re- versed if a more adequate series were studied. The more horizontal the pelvic brim, the more fully extended are the femora, and vice versa. Again, the more inclined or tilted the pelvis is, the further forward of the acetabulum is the axis of the center of gravity.

The buttocks project back of the shoulders in a plurality of both sexes, and particularly in females. The further anterior the axis of the center of gravity, the more common is the projection of the buttocks back of the shoulders in the erect standing position. Similarly, the further anterior the axis of the center of gravity, the sharper the lumbar curve and the more pronounced the pelvic tilt. Asymmetries of the pelvis and unequal length of legs are apparently unrelated to lateral displacements of the center of gravity in the pelvis o r in the base of support. No association between pelvic rotation and long leg could be determined.

The authors regard their findings as merely suggestive, and by no means conclusive. They strongly recommend that this problem of the relation of the pelvis to the erect posture be re-investigated by other students who will avoid their errors in method and technique and who will utilize a statistically adequate sample of subjects.

Documents

Relation of the pelvis to erect posture. An exploratory study