lengths of the segments of the spinal cord. The measurements of these lengths recorded by STILLING (3, p. 619), are insufficient, having been made between the uppermost and lowest fila of each nerve-thus omitting some of the cord, where, as is conspicuously the case in the thoracic region, the line of roots is not continuous. A number of fresh observations were therefore made. The measurements to determine the lengths of the segments in the adult human cord were made on one specimen (W)-preserved in normal size in 10 per cent. formalin; -on the two careful delineations (X, Y) published by Kadyi (4), and on the photolithographic chart (Z) of RÜDINGER (5). In the three plates the cord is depicted in natural size. The cords in the order named, are designated W, X, Y, Z.

In the cord W, the condition of the specimen did not permit the measurement of the first four cervical segments. In cords X and Y, the first cervical segment could not be measured on the dorsal aspect nor on the coccygeal segment at all. In cord Z, measurements on the dorsal aspect alone could be made, and even these could not be extended below the level of the 12th thoracic segment.

To determine the length of a segment, the distance between the uppermost fila of successive nerves was found, beginning with the uppermost filum of the first cervical nerve.1 This was done both on the dorsal and ventral aspects of the cord. In making the measurements, the distance was marked off with a pair of spring compasses, and then this distance was measured on a metal scale to the nearest tenth of a millimeter. Each measurement has been separately entered in Table I.

1 This method of measurement credits to any segment the entire "intersegmental space" which lies caudad to it. The method of LÜDERITZ (6) was to credit to any segment one-half the length of the intersegmental spaces lying on either side of it. The difference in the results would be the amount by which, in any instance, the intersegmental space caudad to the part of the segment to which the roots were attached differed in length from the sum of half the intersegmental spaces above and below the same parts. This difference in general would be small.

TABLE I.

Showing in millimeters the lengths of the segments in the adult

The last column in the Table (I), gives the average of all the measurements for each segment, and these average lengths were those used for the divisions of the abscissa in the accompanying chart I.

(a) Total Length of the Spinal Cord. The lengths of all the segments taken together, should, of course, equal the length of the entire cord. The average length obtained for the segments will therefore depend on the lengths of the cords. But before presenting the lengths of the cords which were here ex

amined, it will be best to state what is already known concerning the length of the human spinal cord.

The correlation in development between the medulla spinalis and the columna vertebralis indicates that the longer columna vertebralis would contain the longer medulla spinalis.

The observations of RAVENEL (7) on the length of the adult human spinal cord, show in II adult males, a range in length of 390-480 mm., with an average of 448 mm. In 11 adult females, the range is from 370-460 mm., with an average of 413 These results plainly exhibit the greater average length. of the medulla spinalis in the male. The measurements were made on fresh material from the level of the upper edge of the atlas to the lowest filum of the coccygeal nerve on the conus medullaris.

mm.

The cords examined for the length of the segments in the present investigation give the following lengths:

It is thus seen that this average length lies between the average length for the males and that for the females, though rather nearer the former, as determined by RAVENEL (7). It concerns us here, however, merely to show that the average length obtained is a medium one—-differences according to sex being disregarded. The average lengths of the segments determined from the several cords (W, X, Y, Z) are presented in Table I. When these average lengths are summed, they give 441.6 mm., for the length of the cord, against 440.5 mm., as determined by direct measurements. The former number is the one employed in the construction of the chart.

LÜDERITZ (6) is the only investigator who has made a detailed study of the length of the segments of the spinal cord.

On comparing the results just given (Table I), with those

obtained by LÜDERITZ (6), it appears that the averages of the sums of the segments in his two men-33 and 37 years of age -show the total to be 450.5 mm., or about 2% more than the total sum in our case. This is a difference which is well within normal limits, since the measurements of RAvenel (7) show for the male cord a range in length from 390-480 mm. At the same time, the lengths of the segments as determined by LüDERITZ, agree substantially with our own, when the differences in the total length of the cord and the great individual variations in the lengths of the segments are both taken into ac

count.

This can be shown by the accompanying Table (III), in which the percentage values of the lengths of the different regions of the cord as determined from our own Table (I), and from the observations of LÜDERITZ on two male cords, are compared with the range in the lengths in these regions, as found by RAVENEL in II male cords.

TABLE III.

To show the percentage value of the lengths of the spinal cord within the regions named in the table, as compared with the range in these percentages as determined by RAVENEL.

For these reasons we are justified in employing our own data in marking off the abscissa for the curve showing areas, since, in so doing, we shall give an approximately true picture of the relations in the human cord of medium size.

For the abscissa, it was decided therefore to take 441.6 mm. and this base line was divided into lengths equal to the average lengths of the segments as recorded in Table I.

II. Areas of Cross-Sections.

In the first instance charts were drawn life size, so to speak, with the abscissa 441.6 mm. long. The ordinates were drawn

on a scale of one linear millimeter for each square millimeter in the area of the transverse section of the cord. The original charts thus made, were too large to publish, and have therefore been reduced photographically to exactly one-third their linear dimensions. As a consequence of this reduction, multiplying the length of the ordinates in the accompanying chart by three, will give a line as many millimeters long as the section has square millimeters of area, and multiplying the length of any segment in the chart by three will give the length of the segment as represented in the column of averages in Table 1.

The data for the areas of the cross sections represented by the ordinates were taken from STILLING (3) and comprise his measurements on four adult cords.

TABLE IV.

Giving the pages in STILLING'S "Neue Untersuchungen über den Bau des Rückenmarks"-1859, where the records are to be found.

The sections from each segment of these cords are also depicted in Table XXVIII of STILLING'S Atlas (1859).

This is the place to call attention to the condition of the cords measured by STILLING. All of the measurements of areas used in this present article-including those on the immature cords to be mentioned later on-were made on material hardened in chromic acid and preserved in 97% alcohol. STILLING makes a statement of his method on pp. 1032 and 1033, but it appears to be erroneous in that it calls for so large an amount of chromic acid, practically an 8% solution. Preliminary observations made on the spinal cord of the white rat, hardened in chromic acid 0.6%, followed by 97% alcohol, indicate a slight increase in the volume of the cord after this treatment. This suggests that these measurements by STILL