Fiber analysis of the pyramidal tract of the laboratory rat

EXPERIMENTAL

NEUROLOGY

87, 503-5 18 (1985)

Fiber Analysis of the Pyramidal Tract of the Laboratory Rat G. W. HARDING

AND A. L. TOWE’

Department of Physiology and Biophysics, University of Washington School of Medicine. Seattle, Washington 98195 Received June 12. 1984; revision received October 29. I984 Light and electron microscopic study of the pyramidal tract of the laboratory rat at a midbulbar level revealed the total number of myelinated fibers on one side to be about 200,000. They ranged from 0.2 pm to more than 5 pm, but clustered strongly in the neighborhood of 1.0 pm (mode of 0.9 pm and mean of I .2 pm), forming the highly skewed fiber spectrum so familiar for mammalian pyramidal tracts and other central fiber pathways. Numerous small clusters of unmyelinated axons were found scattered throughout the tract, adding another 100,000 axons to the estimated number. Not only were the fibers exceedingly small, but also the degree of myelination relative to axon diameter varied widely, suggesting that conduction speed within the tract is not optimal for all fibers. In fact, about half of the fibers in the pyramidal tract would, in theory, conduct faster if they had no myelin wrapping. 0 19X5 Academic Press. Inc.

INTRODUCTION The estimated number of fibers in the pyramidal tract (PT) of the laboratory rat has increased significantly during the last 75 years. In 19 10, Ring (7) identified only 901 PT fibers rostra1 to the decussation of the pyramids, but in 1940, Lassek and Rasmussen (11) increased that number to 73,000. In an allometric sense, this latter number seemed about “right” for a rat-size mammal (19). However, in 1969, Dunkerley and Duncan (5) reported an average of 111,600 fibers per pyramid from six counts made on three rats. Then, in 197 1, Brown (2) estimated from light microscopy Abbreviations: PT-pyramidal tract, LM-light microscopy, EM-electron microscopy. ’ This work was supported by National Institutes of Health grants NS 05 136 and NS 0 179 I. The authors thank Dr. Barbara A. Bohne, Washington University, St. Louis, MO, for the use of her laboratory facilities and for her generous assistance. Excellent technical help was provided by Charles D. Carr, Rosie Saito, and Virginia Schnettgoecke. Mr. Harding’s present address is Department of Neurology, Washington University School of Medicine, St. Louis, MO 63 130. 503 0014-4886/85 $3.00 Copyright 0 1985 by Academic Press, Inc. All rights of reproduction in any form reserved.

504

HARDING

AND

TOWE

that there are about 137,000 PT fibers “at the cervical level,” which suggests that there may be as many as 230,000 PT fibers at a midbulbar level (5, 18). Using electron microscopy, Brown (2) also found numerous unmyelinated fibers interspersed among the myelinated PT fibers, although no count was reported. Finally, in 1982, Leenen, Meek, and Nieuwenhuys (12) estimated that there are only 9 1,000 myelinated, but fully 133,000 unmyelinated, fibers in the PT at a midbulbar level, again using electron microscopy. Clearly, although the estimates should eventually converge in the neighborhood of some particular number, such convergence is not yet evident. The wide differences reported are not readily attributable to intraspecific variation or to age differences; they must be due to differences in technique. There is considerable precedent for such an idea. In the case of man, for example, the averages from various studies have ranged from a low of 496,611 for nine apparently normal adult specimens (20) to highs of 1,087,200 for 2 1 specimens (4) and 1,lO 1,000 for three specimens (10). Such marked differences sometimes can be resolved in a reasonable manner, usually in favor of the higher estimates ( 19). Whether or not such a resolution can be attained for the rat remains an open question. In all species thus far studied, the PT “fiber spectrum” has taken the form of a highly skewed distribution, with a mode near 1.0 pm and with 88 to 98% of the fibers less than 3.0 pm in diameter (1, 5, 6, 12), or even less (12)-much too thin to benefit by the presence of myelin, if optimum conduction speed is to be attained ( 16). The primary problem in defining the shape of the fiber spectrum has been to identify fibers near the limit of resolution of light microscopy. This problem can be overcome by using electron microscopy; it is then possible not only to obtain fairly accurate counts, but also to make more accurate measurements, and thus to define the shape of the fiber spectrum with greater accuracy. Rather than using the standard 1.O- or 0.5~pm bin size, it is possible to use 0. l-pm bins, down to the smallest fibers. In the present study, we describe the pyramidal tract of the laboratory rat using both light and electron microscopy of sections through the pyramids at a midbulbar level. Measurements were made of the circumference of both the fiber and its axis cylinder for a large sample of myelinated PT fibers. Equivalent circular diameters were calculated to determine the PT fiber spectrum. Separate estimates were made of the total number of myelinated and unmyelinated fibers present at that level of the brain stem. METHODS Two adult male Sprague-Dawley rats were fasted 12 h, and then weighed (600 and 476 g) and anesthetized with pentobarbital sodium at 60 mg/kg.

PYRAMIDAL

TRACT

FIBERS

505

The medullary pyramids were exposed via the standard ventral approach by removing basioccipital bone and reflecting the dura mater. Each animal was thereafter treated differently, one being used for study of its antidromic response and for light microscopy and the second for both light and electron microscopy. Only the anatomic results will be discussed in this paper. Animal One. The protocol of Lund (13) was followed, the animal being perfused with normal saline, followed by 0.5% glutaraldehyde and 4% paraformaldehyde in Millonig’s phosphate buffer (14). A block of brain stem containing the medullary pyramids was removed and postfixed in 2% osmium tetroxide in Millonig’s buffer. This was followed by alcohol dehydration, Epon embedding, and polymerization. Sections were cut at 2 pm and stained with methylene blue-azure II (15) for light microscopy (LM). Animal Two. Two calibration lesions were produced in the left medulla using a parallel, tandem pair of tungsten electrodes with a tip separation of 233 pm, inserted 1 mm deep to the ventral surface of the medulla in the transverse plane, with the medial electrode about 0.5 mm from the midline. The animal was then perfused at 155 cm Hz0 pressure with a flushing solution of Millonig’s buffer to which 0.05% CaC12 had been added (PH 7.27; osmolarity 293 mOsm). After about 50 ml flushing solution had been perfused through the vascular system, 950 ml 0.5% glutaraldehyde and 4% paraformaldehyde in Millonig’s buffer (Ca added) was perfused for 13 min. Temperature was maintained at 40°C. The animal was then refrigerated 2 h, and then weighed (including tissue removed during surgery). The brain and cervical spinal cord were removed and placed in 50 ml fixative for another 2 h. The tissue was then trimmed to include brain stem, cerebellum, and cervical spinal cord, and was placed overnight in fresh fixative at 4’C. The next morning, the tissue was rinsed in three 15-min changes of Millonig’s buffer and soaked for 1 h in fresh buffer. During the soak, the tissue was trimmed to contain a portion of the pyramidal tract. The size of the final piece was approximately 2 X 2 X 3 mm. It was then postfixed 4 h at 4°C in 2% 0~0~ in Millonig’s buffer, to which 1.5% potassium fenicyanide had been added. The osmium was removed by three 15-min rinses in Millonig’s buffer. After a brief rinse in 35% ethanol and two 5-min rinses in 50% ethanol, the tissue was immersed in 70% ethanol and placed in a refrigerator. The next day the specimen was dehydrated and infiltrated with plastic. During processing, the specimen bottle was continuously agitated on a rotary shaker at 65 rpm. The tissue block was dehydrated in a series of ethanol solutions, 30 min each, of 80%, 95%, lOO%, and 100% alcohol and propylene oxide (two changes). The tissue was then infiltrated with increasing concentrations of Araldite (Durcupan-Fhrda) diluted with propylene oxide

506

HARDING

AND

TOWE

in ratios of 1:3, 1:2, 1: 1, and 2: 1, the specimen being exposed to each of these dilutions for 1 h. For the last 4 h of infiltration, the Araldite was warmed to approximately 4YC by shining a 100-W bulb on the bottle in the rotary shaker. The cork on the bottle was removed during this time. The specimen was then placed in a flat embedding mold in fresh Araldite, and the plastic was polymerized 48 h at 60°C. The polymerized block was trimmed and mounted in a chuck so that the specimen could be sectioned perpendicular to the rostrocaudal axis. Using an LKE! Ultratome III, both l-pm (thick) and thinner (thin) sections were made with 38”-angle knives (solar gray glass, Libby-Owens-Ford). The thick sections were mounted on glass slides, stained with methylene blue-azure II, and viewed by LM. The thin sections were mounted on Formvar-coated, single-slot copper grids or on uncoated mesh grids, stained with 4% aqueous uranyl acetate and 11.6% lead acetate, and viewed with a JEOL 100s electron microscope. Sampling Procedure. Low-power photomicrographs were made of sections from the right pyramid to be measured by LM and adjacent to thin sections to be measured by EM. Camera lucida drawings were made outlining the PT and all blood vessels and transverse fiber bundles within the pyramid. The borders of the PT could be visualized on the basis of fiber sizes and density. The drawings, made at 604X magnification, were then transferred to large sheets of l-mm graph paper and each l-cm square which included PT fibers was numbered. Using the rule that each square to be studied must be entirely within the PT and must contain no large blood vessels or transverse fiber bundles, 36 squares were selected randomly from rat one and 40 squares were selected uniformly from rat two. For rat one, each square was photographed at high magnification and the photographs were enlarged five times to a final magnification of 4469X. A circle equivalent to 7.6 pm radius was then inscribed on each plate, and all fibers within the circle by one-half or more were traced with a sonic graphics tabled on a PDP- 12 computer system (3). All fiber perimeters were traced, and from these tracings the enclosed areas were computed and equivalent circular diameters calculated. The graphics tablet was calibrated before each tracing session. No attempt was made to correct for tissue shrinkage for rat one. For rat 2, the squares selected for study were identified on the adjacent thin section by using measured displacements from clear landmarks on the low-power tracing. Electron micrographs were made of each site and were enlarged three times for a final magnification of 17,100X. Electron microscopy (EM) calibrations were made just before and just after each photographic session. The focus of the enlarger was constant for all EM negatives, and the prints were made on resin-coated paper. Each final plate was 134.67 pm2 in area; all fibers wholly within the plate were traced with an

PYRAMIDAL

TRACT

FIBERS

507

electrostatic graphics tablet on an HP-85 computer system, The outside (fiber) and inside (axon) fiber circumferences on each fiber were traced, and myelin thickness was measured at a representative site on the sheath by tracing a line normal to the myelin sheath from outside to inside. The enclosed fiber and axon areas were computed, and equivalent circular diameters were calculated. The graphics tablet was calibrated before each tracing session. The total circumference of the pyramidal tract was traced from the large drawings for both animals using the HP-85 system. All apparent blood vessels and transverse fiber bundle outlines were also traced, and the enclosed areas were subtracted from the total area of the PT. Dividing this area into the sum of the sample areas yielded the fraction of the PT sampled for each rat: 0.02 1 for rat one and 0.0 17 for rat two. Because both thick and thin sections were taken from the same embedding block, the shrinkages were uniform (5, 9); hence, the calculated fractions of the PT sampled are nearly exact. A few high-power (20,000 to 40,000X) EM pictures were taken near the center of the PT and subsequently enlarged 2.5 to 3.0 times. The thickness of the myelin sheaths was measured and the number of myelin lamellae was counted at a site where the sheath was relatively straight. From this, average lamellar thickness was calculated. The lesions produced in the brain stem by the double electrode were located from 2-pm serial sections through the region. A section containing both lesions was photographed at medium magnification and the distance between the centers of the lesions measured. The double electrode was photographed at the same magnification and the distance between the tips measured. Subtracting the ratio of the two distances from unity yielded an estimate of the tissue shrinkage, which came out to be 0.187. RESULTS Pyramidal Tract Area. Accurate estimation of the cross-sectional area of the PT is important for calculating the total number of fibers in the tract. In some preparations, the border between the PT and medial lemniscus is clearly evident, as with rat one (Fig. 1A), whereas in others it is less clear, as with rat two (Fig. 1B). Care was taken to define a clear border along the tortuous and fuzzy region shown in Fig. lB, including what could be taken only as a small outlying bundle of PT fibers. The criterion for identification of the tract was a high density of fine fibers, with only a few coarser fibers scattered randomly throughout the region. The generally coarser fibers of the medial lemniscus intermingled with these fine fibers along the border, rendering that boarder difficult to specify as a line. On the other hand, the

508

HARDING

AND TOWE

FIG. I. Photomicrographs of midbulbar pyramidal tracts of rat one (A) and rat two (B). Section from rat two split, but was reassembled for photography. Calibration bar: 0. I mm.

magnitude of this included fiber bundles, two (a more

the error in estimating the area of the PT was less than 6%; the errors in measurement of blood vessels and transverse which made up 2% of the PT in rat one and fully 15% in rat highly vascular region happened to be sampled). The final

PYRAMIDAL

TRACT

FIBERS

509

estimates were 0.293 mm’ for rat one and 0.3 13 mm2 for rat two, with total body weights of 600 and 476 g, respectively. Fiber Count. One of the 34 circular areas selected from rat one for LM study is shown in Fig. 2; the problem of clearly identifying and measuring fibers near the limit of LM resolution can be appreciated from the plate. Many circular images not clearly measurable, but having the character of fibers, were counted separately; they comprised 7.5% of the total count. Thus, of 3,977 PT fibers counted in the 34 circular areas, 3684 were clearly measurable and were used to estimate the fiber spectrum. The remaining 293 fibers are treated in this account as having been between 0.4 and 0.6 pm in diameter. The count of 3977 fibers in the 6183~pm2 sample area implies a total of 188,395 PT fibers at the midbulbar level-higher than most previous estimates. Though an effort was made to identify every fiber, some may have been missed because light microscopy was used. On the

FIG. 2. Electronmicrograph of pyramidal tract of rat one showing one of the circular areas sampled. Calibration bar: 4 pm.

510

HARDING

AND TOWE

other hand, the density of PT fibers per unit area was the same as for rat two. A portion of one of the 40 square areas selected from rat two for EM study is shown in Fig. 3. It is evident that most fibers were round or oval, with little wrinkling of the myelin sheaths and almost no separation of myelin lamellae. The axons completely filled the sheaths and had uniformly distributed organelles. Occasional isolated transverse fibers were present (right of middle in Fig. 3), but their origin and destination remained unknown. Clusters of fine, unmyelinated fibers were scattered throughout the tract (lower left in Fig. 3), their presence being equivocal in a-few instances. Variations in the degree of myelination on axons of the same diameter were striking. The count of 2814 myelinated fibers in the 5387pm* sample area implied a total of 163,530 fibers at the midbulbar level. However, only the fibers contained wholly within each sample area were measured; a better estimate was obtained by reducing each dimension of the sample area by the average fiber diameter of 1.182 brn. This adjustment

FIG. 3. Electronmicrograph of pyramidal tract of rat two showing portion of one square area sampled. Calibration bar: 0.5 pm.

PYRAMIDAL

TRACT

FIBERS

511

yielded an estimate of 206,829 myelinated PT fibers. It also yielded a PT fiber density of 0.661/pm2, compared with 0.643/pm2 for rat one. Thus, if the estimates were much in error, the source of that error lay in calculating the area of the PT in both rats. Fiber Spectrum. It is evident in Figs. 2 and 3 that the PT fibers were very small, though myelinated. The median fiber diameters were 1.054 pm for rat one (no correction for shrinkage) and 1.182 pm for rat two (0.187 correction for shrinkage). The maximum fiber diameters measured were 3.7 and 5.2 pm, respectively. Figure 4 shows the distribution of fiber sizes for the two rats, that for rat one being plotted in 0.2~pm bins and that for rat two in 0.1~pm bins, after correction for shrinkage. A 20% shrinkage for rat one would make the two graphs almost superimposable, the major difference being among the larger fibers. The modal diameter in both distributions was about 0.9 pm, near that for all mammals thus far measured. Graphs of the fiber spectrum for each quadrant of the PT, computed separately, were found to be almost exactly superimposable, thus revealing that the PT fibers are uniformly distributed through the tract with respect to size. The four fiber spectrums are shown in Fig. 5 as cumulative proportion curves. Perusal of these curves revealed that about 6.5% of the myelinated fibers were less than 0.5 pm in diameter and that 50% was less than about 1.2 pm. The inflection on the rising phase of each curve, which shows the mode of the distribution, occurred near 0.9 pm. Only 36% of the fibers were smaller than the mode, but 80% were smaller than twice the mode. The quadrant showing a greater accumulation of large fibers (2.1 to 3.1 pm) was dorsolateral in the tract; it most likely reflected sampling variation. Myelin Lamellae. High-magnification EM pictures were made in order to resolve the lamellar structure. Except for sheaths with only two to three wraps, no interlamellar spaces were evident. A region along the myelin

TOO 2

600

2

500

z v0 I g

400

2

100

300 200

0 0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

Fiber

4.0

Diameter

0

0.5

1.0

1.5

2.0

2.5

3.0

3.5

4.0

(pm)

FIG. 4. Fiber spectrum of rat one (A) and rat two (B). Hatched portion in 0.5rrn bin in A shows the 293 countable but unmeasureable fibers from rat one.

512

HARDING AND TOWE

FIBER

DIAMETER

FIG. 5. Cumulative proportion curves of fiber size distributions for the four quadrants of the pyramidal tract of rat two. Pluses mark positions of mode (0.9 pm), median (1.2 pm), and 97th percentile (3.0 pm).

sheath where the lamellae could be clearly resolved was selected for study in 153 PT fibers. Average lamellar thickness was obtained for each fiber by dividing the myelin thickness at the site counted by the number of lamellae. Though the relation was not strong (r = 0.55), average lamellar thickness decreased as the number of wraps increased, from 112 A at four wraps to 90 8, at 30 or more wraps. Except for somewhat thinner innermost and outermost wraps, lamellar thickness was uniform across the sheath (Fig. 6). Thus, the observed relationship did not result from averaging-in successively thinner lamellae as myelin thickness increased. Another interesting feature, illustrated in Fig. 6, was that adjacent fibers often had the same number of wraps, independently on their axon diameters. This is the inverse of the observation stated earlier that axons of the same diameter often had markedly different myelin thicknesses. A few EM pictures were made of the material from rat one in order to obtain an idea of what was being measured by LM. As is evident in Fig. 7A, a large amount of myelin sheath wrinkling and lamellar separation was present. This contrasted sharply with the condition in rat two (Fig. 7B), in which potassium ferricyanide was used. Here, the sheaths were relatively

PYRAMIDAL

TRACT

FIBERS

513

FIG. 6. Electronmicrograph of PT fibers illustrating myelin lamellae and unmyelinated fibers. Calibration bar: 0.1 pm.

smooth and no interlamellar spaces were evident. It therefore seems likely that the outer circumferences measured in rat one overestimated the actual values, which partially offset the lack of correction for shrinkage.

514

HARDING

AND TOWE

FIG. 7. Electronmicrographs from rat one (A) and rat two (B) illustrating differences in the two types of preparation. Calibration bar: 0.5 pm.

Unmyelinated Fibers. The total number of fibers with no apparent myelin was estimated from the EM pictures used for measuring myelinated fibers. The high-magnification micrograph of Fig. 8 shows at least 12 unmyelinated

PYRAMIDAL

FIG. 8. Electronmicrograph 0.1 pm.

TRACT

FIBERS

515

of rat two showing unmyelinated fibers in tract. Calibration bar:

fibers within a 1.2~pm* area. Several such clusters are seen in Figs. 3, 6, and 7B, and some are evident even in the poorer material of Fig. 7A. The magnification of Fig. 8 shows these fibers to be mainly 0.1 to 0.3 pm in diameter. The lowest of three separate counts made on the 40 sample areas of rat two yielded 1370 unmyelinated fibers, implying that about one-third of the fibers within the medullary pyramids of adult laboratory rats are unmyelinated. If the cell bodies of these fibers were in the cerebral cortex and if they were a part of the pyramidal system, then the midbulbar PT of rat two contained about 30 1,82 1 PT fibers. DISCUSSION Both the LM and EM fiber counts were higher than in most previous studies; the reasons for this difference are unclear. The strategy in any fiber count (unless the entire tract is counted) is to calculate the area to be sampled, find the fiber density within the sample area, and multiply that density by the total area of the tract. Thus, there are three sources for error: (i) in total area, (ii) in area sampled, and (iii) in fiber count. The first source

516

HARDING

AND TOWE

may account for much of the variation among the different studies, because identification of a border separating the PT from adjacent tissue depends on the criterion of fiber size and dispersion of size. Perusal of a well stained section through the medulla leaves the clear impression that a border exists within a rather narrow band of error, and this impression is reinforced by comparison of a completely degenerated pyramid with the normal side. The error in this measurement should be less than 10% along about half of the border of the pyramidal tract. Subtraction of the area occupied by blood vessels and transverse fibers improves the accuracy of the estimate. If shrinkage corrections are properly applied, they should have no effect on the results. Measurement of the sample areas is a matter of proper calculation of EM magnification, which can readily be checked by the use of reference scales. This error should not exceed 2%. The remaining source of error is in the identification of fibers when counting. It is likely that rather large errors have crept into several studies, as may be surmised from the specific shapes of the fiber spectra. However, not enough information is available to allow adjusting for such differences. The estimates of the total number of myelinated fibers in the present study, though larger than usual (5, 11, 12), were about the same using both LM and EM techniques. The clarity of the PT border in rat one suggests that the LM estimate may not be much in error; only the smallest fibers were missed. On the other hand, placement of a border in rat two was more difficult, and could have erred enough to raise or lower the estimate by about 15,000 fibers. In either case, the number is about twice the “proper” number for a half-kilogram mammal, if the allometric relation calculated from a broad range of mammals is taken as descriptive (19). That relation was calculated on LM data, and it was stated that “There is little doubt that many of the counts are deficient in the smallest fibers; the work should be done with electron-microscopic techniques” (19). In doing so, and in being careful in the LM counts, it appears that the estimated number of PT fibers may double (or triple, if the unmyelinated fibers belong to the pyramidal system). A recent EM fiber count in the domestic cat yielded an average of about 356,000 myelinated and 60,000 unmyelinated fibers at the midbulbar level (1). Again, this is about twice the “proper” number for a 2.5-kg mammal (19). If this increasing estimate accompanies all EM studies, then one might expect to find about two million fibers in each PT of man. On the other hand, more fiber counts have been carried out on man (and rhesus monkey) than on any other mammal, and it is likely that those counts have been pushed as high as the histologic material would allow. Nonetheless, the human material is usually much less than ideal, rendering many fibers obscure, and thereby decreasing the estimated total number. Man weighs about 140 times as much as the laboratory rat,

PYRAMIDAL

TRACT

FIBERS

517

but has only five to six times as many PT fibers; why the rat has so many, or man so few, remains to be discovered. Conduction distances are considerably greater in man than in the rat, yet the fiber spectra are similar between the two species. As for mammals in general ( 18), the modal diameter is close to 1.O pm and more than 90% of the fibers are less than 3.0 pm in diameter. Measurement of human PT fibers by Lankamp (9) put 70% of them less than 1.0 pm and 84% less than 2.0 pm; the comparable values for rat one, by LM, were 49% and 91%, respectively. The only real difference is that about 7% of the PT fibers in man are larger than any found in the rat. If the same rules concerning conduction speed apply to both species, then the rat has considerable temporal advantage over man. However, conduction speed may not be the issue, because more than half of the myelinated PT fibers are too small to gain any speed advantage from the myelin; in fact, they should conduct slower than if the myelin were not present (16). Why they have myelin remains to be discovered. Unmyelinated Fibers. Although the first EM study of a rodent (6) failed to reveal unmyelinated fibers in the PT, subsequent studies have found them in seeming abundance. Leenan and co-workers (12) calculated the average number for five rats at 9 1,000 myelinated and 133,000 unmyelinated PT fibers, which may be misleading in that the rats were young (190 to 2 10 g body weight) and may not have had fully myelinated tracts. On the other hand, several estimates could have been low because of the difficulty in identifying the very small fibers. Samorajski and Friede (17) found the average unmyelinated fiber to be less than 0.5 pm, Leenen et al. (12) placed it at 0.16 pm (no correction for shrinkage), and Brown (2) stated that “numerous non-myelinated axons 0.1-0.2 pm in diameter may be found interspersed among the large axons,” We likewise found many exceedingly small unmyelinated axons; whether they are a part of the pyramidal system is not known. It is clearly evident that a great many unmyelinated axons exist in clumps among the larger, but still quite small, myelinated PT fibers, not only in the medulla, but also in the spinal cord (8). There is an unmistakable overlap in fiber sizes between the unmyelinated and the myelinated fibers, suggesting that they may comprise functionally distinct groups rather than reflecting those fibers that have and those that have not attained the size threshold that might trigger a myelination process. It is quite obvious that the pyramidal system is not a primarily high-speed system. REFERENCES 1. BIEDENBACH, M. A., J. L. DEVITO, AND A. C. BROWN. axon size and morphology, submitted.

1985. Pyramidal

tract of the cat:

518 2.

3. 4. 5. 6. 7. 8. 9. IO. 1I. 12. 13. 14. 15. 16. 17.

18. 19. 20.

HARDING

AND

TOWE

BROWN, L. T., JR. 197 1. Projections and termination of the corticospinal tract in rodents. Exp. Brain Res. 13: 432-450. COWAN, W. M., AND D. F. WANG. 1973. A computer system for the measurement of cell and nuclear sizes. J. Microsc. 99: 33 l-348. DEMYER, W. 1959. Number of axons and myelin sheaths in adult human medullary pyramids. Neurology 9: 42-47. DUNKERLEY, G. B., AND D. DUNCAN. 1969. A light and electron microscopic study of the normal and the degenerating corticospinal tract in the rat. J. Comp. Neural. 137: 155184. GOLDBY, F., AND G. N. KACKER. 1963. A survey of the pyramidal system in the coypu rat, Myocastor coypu. J. Anat. (London) 97: 517-53 I. KING, J. L. 1910. The corticospinal tract of the rat. Anat. Rec. 4: 245-252. LANGFORD, L. A., AND R. E. COGGESHALL. 1981. Unmyelinated axons in the posterior funiculi. Science 211: 176-l 77. LANKAMP, D. J. 1967. The Fiber Composition of the Pedunculus Cerebri (Crus Cerebri) in Man. Thesis, Leiden, Luctor et Emergo. LASSEK, A. M., AND G. L. RASMUSSEN. 1939. The human pyramidal tract. A fiber and numerical analysis. Arch. Neurol. Psychiatry 42: 872-876. LASSEK, A. M., AND G. L. RASMUSSEN. 1940. A comparative fiber and numerical analysis of the pyramidal tract. J. Comp. Neural. 72: 417-428. LEENEN, L., J. MEEK, AND R. NIEUWENHUYS. 1982. Unmyelinated fibers in the pyramidal tract of the rat: a new view. Brain Rex 246: 297-301. LUND, R. D. 1972. Synaptic patterns in the superficial layers of the superior colliculus of the monkey, Macaca mulatta. Exp. Brain Rex 15: 194-21 I. MILLONIG, G. 1961. Advantages of a phosphate buffer for OsO., solutions in fixation. J. Appl. Physiol. 321 1637. RICHARDSON,K. C., L. JARRETT, AND E. H. FINKE. 1960. Embedding in epoxy resins for ultrathin sectioning in electron microscopy. Stain Technol. 35: 3 13-323. RUSHTON, W. A. H. 1951. A theory of the effects of fiber size in medullated nerve. J. Physiol. (London) 115: 101-122. SAMORAJSKI, T., AND R. L. FRIEDE. 1968. A quantitative electron microscopic study of myelination in the pyramidal tract of rat. J. Comp. Neural. 134: 323-338. TOWE, A. L. 1973. Motor cortex and the pyramidal system, pp. 67-97. In J. D. MASER, Ed., Efirent organization and the integration ofbehavior. Academic Press, New York. TOWE, A. L. 1973. Relative numbers of pyramidal tract neurons in mammals of different sizes. Brain Behav. Evol. 7: l-1 7. VERHAART, W. J. C. 1950. Hypertrophy of pes pedunculi and pyramid as a result of degeneration of contralateral corticofugal fiber tracts. J. Comp. Neural. 92: l-16.