An Electronmicroscopic Analysis of the Optic Nerve in the Golden Hamster ROBERT W. RHOADES, LINDA HSU AND GAIL PARFETT Department of Anatomy, College of Medicine and Dentistry of New Jersey, New Jersey School of Osteopathic Medicine, University Heights, P.O. Box 55, Piscataway, New Jersey 08854

ABSTRACT The electronmicroscopic examination of sections taken from the hamster's optic nerve 5 mm behind the globe indicated that the nerve contains 110,165 4,177 (p < 0.05) fibres of which 96.4% are myelinated. The fibre diameter distribution is unimodal with a peak a t 1.2 p m and axon diameters ranging from 0.20 p m to 3.93 pm. Fibres of all sizes are distributed uniformly throughout the cross section of the nerve. The thickness of the myelin sheath surrounding a given axon is highly (0.80) correlated with axonal diameter and the degree of myelination for a fibre of a given size is nearly constant throughout the nerve's cross section. In nerve sections taken just posterior to the globe most (64%)of the fibres counted are unmyelinated and the percentage of unmyelinated axons is highest near the peripheral boundary of the nerve. The process of myelination is essentially complete in sections taken 3.5mm behind the eye. These differences in the myelination of the proximal and distal nerve most probably account for the discrepancy between the results reported here and those provided by a previous study (Tiao and Blakemore, '761 concerned with the structure of the optic nerve in this species.

*

The electronmicroscopic examination of the optic nerve in a number of mammals (Cohen, '67; Forrester and Peters, '67; Potts et al., '72; Treff et al., '72; Hughes and Wassle, '76; Vaney and Hughes, '76; Hughes, '77) has revealed no significant population of unmyelinated fibres. One exception to this generalization is the study of Tiao and Blakemore ('76) which employed the golden hamster. These investigators describe the optic nerve of this species as containing approximately 58%unmyelinated axons. This finding is somewhat surprising in view of the fact that the conduction velocity distributions for the optic nerve fibres which project to the superior colliculus in the hamster (Rhoades and Chalupa, '78, '79) and the rat (Sumitomo e t al., '69; Fukuda et al., '781, a species in which virtually all optic nerve axons are myelinated (Forrester and Peters, '67; Hughes, '771, are very similar. In view of this discrepancy we have, in the present study, reexamined the morphology of the hamster's optic nerve using the electronmicroscope. The objectives of the study were: (1)to provide estimates of the total number of myeliJ. COMP. NEUR. (1979)186: 491-504

nated and unmyelinated fibres in the hamster's optic nerve; and (2) to determine the spectrum of fibre diameters for the optic nerve of this species. The latter point is of some interest since Tiao and Blakemore ('76) have observed t h a t t h e distribution of soma1 diameters for t h e ganglion cells i n t h e hamster's retina reflects two distinct modes and it has also been reported (Rhoades and Chalupa, '78, '79) that the distribution of axonal conduction velocities for those optic nerve fibres which project to the superior colliculus in this species is clearly bimodal. METHODS

Tissue preparation and electronmicroscopy Optic nerves were obtained from five normal adult (5-8 month) hamsters (weight 175210 g m ) . Each animal was deeply anesthetized with 40 mg of sodium pentobarbital (delivered intraperitoneally) and perfused transcardially with 150 ml of 0.9%saline at 37°C. This was followed by 250 ml of a fixative consisting of 1.0% paraformaldehyde, 3.0% glutaraldehyde, 7.0%sucrose, and 0.05%CaCl, in 0.1 N sodium cacodylate buffer (pH 7.31,

491

492

ROBERT W. RHOADES, LINDA HSU AND GAIL PARFETT

also a t 37OC. Immediately after the perfusion the entire length of the optic nerve was dissected free, immersed in the same fixative for 9 to 12 hours a t 4°C and then rinsed in two changes of cacodylate buffer. After postfixation for one to two hours in 2.0% osmium tetroxide (in 0.1 N sodium cacodylate buffer, pH 6.8) the tissues were dehydrated in graded alcohols and propylene oxide, transected into 2 mm segments, and embedded in Epon-Araldite. Thick (1 p m ) or thin (-8OOA) sections were cut with glass or diamond knives on an MT2-B ultramicrotome. The thick sections were mounted on glass slides, stained with toluidine blue and employed for area measurement and the orientation of electronmicrographs. Thin sections of the entire nerve were mounted on copper mesh grids, stained with 4% uranyl acetate and 0.1% lead citrate and examined with a Phillips EM 300 electronmicroscope.

Fibre counting and measurement The thin sections used for the quantitative analysis of the nerve were all taken 5 m m behind the globe (the length of the optic nerve in normal adult hamsters is 9.2 & 0.34 mm; Rhoades and Chalupa, '78). Additional thick and thin sections were taken from 0-4 mm behind the globe and these were employed to examine the morphology of the nerve just after it leaves the eyeball and passses through the lamina cribosa. The electronmicrographs used for the quantitative analysis were taken a t a linear magnification of 6,550 and printed with no further magnification. They were spaced in a quasirandom manner throughout the nerve. TO avoid an influence of the grid bars on the areas sampled the electronmicrographs employed for the counts and measurements were taken from four serial thin sections; care was taken in the analysis to insure that no group of fibres was counted or measured more than once. A 10 y m X 10 y m square was drawn on each electronmicrograph and all fibres within the square and those touching the upper and lefthand borders were counted. In all, 5,821 axons were counted in 108 electronmicrographs and 2,448 (in 42 electronmicrographs) were measured. Axon diameters were established by taking the average of the long and short axis of each fibre. This method was chosen since i t was employed by Tiao and Blakemore ('76) and thus

would allow the most direct comparison of the data obtained here with those provided by their study. As with the majority of the morphometric studies of t h e optic nerve (Ogden and Miller, '66; Tiao and Blakemore, '76; Hughes and Wassle, '76; Hughes, '771, we included the myelin sheath in our fibre diameter measurements.

Identification of m o n s The criteria employed for the identification of nerve fibre cross sections were identical to those employed in previous studies (Peters, '66; Hughes and Wassle, '76). The differentiation of unmyelinated fibres from axons sectioned a t the node of Ranvier was based on the presence of an electron dense granular substance subjacent to the axolemma or of paranodal axoplasm. Both of these features have been found to be characteristic of axons a t the node of Ranvier (Peters et al., '70). Measurement of myelination In order to correlate fibre size with the thickness of the myelin sheath the internal axon diameter (i.e., diameter excluding the sheath) was established in the manner described above and the number of lamellae counted with the aid of a dissecting microscope for a small number of fibres. A lamella (fig. 1) was identified by the occurrence of a major dense line (Friede and Samorajski, '67; Friede et al., '71). Estimation of compression The compression of the thin section was computed in the manner described by Vaney and Hughes ('76). A photomosaic of one of the thin sections was constructed from a series of low power electronmicrographs and the border of the nerve was interpolated over the grid bars. The dorsoventral axis of the thin section (i.e., the axis perpendicular to the knife edge) was compressed to 93% of the value for the adjacent thick section. There was also a slight (2%) compression across the thin section parallel to the microtome knife. The areal compression was computed by weighing cutouts of tracings of the thick and thin sections on an analytical balance. This procedure yielded a total areal compression for the thin section of 9%. The possibility that the compression varied among the different thin sections employed was assessed by measurements of two large axons in electronmicrographs taken from each section. Since the variability

HAMSTERS OPTIC NERVE

493

Fig. 1 Detail of the myelin sheaths from three axon8 ( X 37,800) showing the dense line (DL) which alternates with the lighter intraperiod line (IL). The number of lamellae surrounding the axon was taken as the number of dense lines counted. The calibration is 0.1 rm.

in these measurements was < 1%,the areal compression computed above was employed for the data obtained from all four thin sections. The total fibre count and density estimates were corrected for the compression artifact. Measurements of axonal diameters were not corrected. RESULTS

General considerations A thick section taken from the optic nerve used for the quantitative analysis is shown in figure 2A. As is readily apparent, the dorsoventral axis is elongated relative to the mediolateral axis. This was a feature common to sections taken from this portion of the nerve. Examination of axon profiles from adjacent thin sections (figs. 2B-D) indicated clearly, however, that the shape of the nerve cross section is not a n artifact of oblique cutting. As can be seen in the electronmicrographs, most of the axon profiles are relatively circular and

there is no evidence of a preponderance of profiles which are elongated in the dorso-ventral plane.

Total fibre count and percentages of myelinated and unmyelinated axons The electronmicrographs presented in figure 2 suggest that myelinated axons comprise the vast majority of the fibre population in the hamster's optic nerve and the quantitative analysis confirmed this impression. Of the 5,821 axons counted 96.4% (N = 5,611) were myelinated. Fibres judged to be sectioned a t a node of Ranvier (N = 64) accounted for 1.1%of this total. This total count for the 11,868 pm2 area (corrected) sampled resulted in an estimate of 106,199 f 4,155 (p c0.05) myelinated and 3,966 f 1,021 (p < 0.05) unmyelinated axons for the nerve analyzed. As illustrated in figure 3, the fibre density was relatively uniform throughout the nerve and equalled 592,000/mm2. This total

494

ROBERT W. RHOADES, LINDA HSU AND GAIL PARFETT

Fig. 2 (A) Thick section taken from the nerve at a point 5 m m behind the globe. The calibration is 0.1 mm. (B), (C) and (D) are electronmicrographs ( X 6,550) taken from an adjacent thin section; they are from the central, midperipheral and peripheral portions of the nerve cross section, respectively. M and U indicate myelinated and unmyelinated axons while N is used to signify an axon sectioned a t a node of Ranvier. In the latter case t h e electron dense granular substance subjacent to the axolemma is clearly visible. The processes of glial cells (G) are also evident. Note particularly t h a t in all micrographs unmyelinated axons are quite rare. The calibration for t h e electronmicrographs is 1pm.

495

HAMSTER’S OPTIC NERVE

h-

c

i*\

608,000

Fig. 3 Drawing of the thick section shown in figure 2A depicting the loci of the points (circles) at which electronmicrographs were taken for the counting of axons and the locations (diamonds) at which fibres were both counted and measured. Average fibre densities for different regions (indicated by boxes) are also illustrated. Local fibre densities ranged from 550,000/mm2to 610,000/ mm2 and, as is evident, were relatively uniform throughout the nerve. The calibration is 0.1 mm.

count and nerve fibre density are both very similar to the values which have been reported for the rat (Hughes, ’77).The percentage of * myelinated axons observed here is, however, much greater than the 42%reported for the hamster by Tiao and Blakemore (‘76). In view of the fact these investigators observed unmyelinated axons predominantly in the periphery of the nerve we tested this possibility in our material. To accomplish this the incidence of myelinated and unmyelinated fibres was compared between eight electronmicrographs taken within 40 p m of the geometric center of the nerve, a second set of eight electronmicrographs having their center within 20 p m of

the edge of the nerve, and a third set of eight centered within 20pm of the midpoint of radii running from the center to the edge of the nerve. The results of this comparison are shown in table 1 and it is readily apparent that there was no significant increase in the relative percentage of nonmyelinated fibres as the periphery of the nerve was approached. The values for the central, midperipheral and peripheral regions were 4.1%,3.7%and 5.1% respectively. Thus, with the exception of the estimate of the total number of axons in the hamster’s optic nerve, the present findings are appreciably different from those provided by Tiao and Blakemore (‘76).One possible explanation for the high percentage of unmyelinated nerve fibres observed by Tiao and Blakemore (‘76) which has been recently put forward by Hughes (‘77) is that the section employed in their study was taken very close to the lamina cribrosa where myelination may not yet be complete (Tansley, ’56). To test this possibility we examined sections of the optic nerve taken from approximately 0.5, 1.5 and 3.5 mm behind the globe. Electronmicrographs from a thin section taken 0.5 mm behind the globe are shown in figure 4. The composition of the nerve a t this point is markedly different from that seen in sections taken closer to the chiasm (see, for example, fig. 2). Of the 1,047 fibres counted within an area of 1,648 pm2 in this distal section 64%(N = 670) were not myelinated. Furthermore, the relative incidence of unmyelinated fibres was clearly greater in the periphery of the nerve. In the five electronmicrographs taken a t the edge of the nerve 303 (81%)of the 374 axon profiles counted in a 491 pm2 area were nonmyelinated. In an area of identical size a t the center of the nerve only 38%(N = 130) of the 342 axons counted were unmyelinated (x’ = 138.21, p < 0.01). Figure 5 shows electronmicrographs obTABLE 1

Myelinated axons

Central nerve Midperipheral nerve Peripheral nerve

405 (95.9%) 393 (96.3%) 470 (94.9%)

Unmyelinated axons

17 (4.1%) 15 (3.7%) 25 (5.191)

Numbers of myelinated and unmyelinated axons counted in areas of 879 pm’ for the central, midperipheral and peripheral portions of the nerve cross section taken 5 mm behind the globe. The percentages of unmyelinated fibres in the different portions of the nerve sampled were not significantly different (xz = 1.14, p > 0.05).

496

ROBERT W . RHOADES, LINDA HSU AND GAIL PARFETT

Fig. 4 Electronmicrographs ( X 6,5501 taken from t h e central (A), midperipheral (B) and peripheral (C) portions of a nerve section 0.5 mm posterior t o t h e globe. The conventions here are t h e same as those in figure 2. Note the increase in the relative percentage of unmyelinated fibres, especially in the electronmicrograph (C) taken near the edge of the nerve. The calibration is 1pm.

Fig. 5 A series of electronmicrographs ( X 6,550) taken from t h e central (AI, midperipheral (B)and peripheral (Cl portions of thin section 1.5 mm posterior to t h e globe from the same nerve used to obtain t h e section illustrated in figure 4. Here again there are a large number of unmyelinated fibres and they are seen predominantly in the peripheral portion of the cross section. The conventions are t h e same as those in figure 2. The calibration is 1pm.

tained from a thin section taken approximately 1.5 mm behind the globe. Here again unmyelinated axons are observed more frequently than in more distal portions of the nerve. In the 1,648 pm2 area in which fibres were counted at this point 35.2%(N = 352) of the 1,001 axons encountered were unmyelinated. As was the case for the section taken closer to the eye, the relative incidence of unmyelinated fibres was greatest in the periphery. In an area of 549 p m 2 counted at the edge of the nerve 249 of the 360 axons (69.1%)encountered were not myelinated. Only 13%(N = 28)

of the 289 fibres sampled in the central area did not have myelin sheaths (x2 = 233.94, p < 0.01). The composition of the sections taken 3.5 mm behind the globe (fig. 6) was very similar to that of the nerve segment from which we collected the vast majority of our quantitative data. At this point 96% (N = 1,060) of the 1,104 axons counted (in an area of 1,648 pm2) were myelinated and there was no difference in the relative incidence of nonmyelinated fibres in the center and periphery of the nerve (x2 < 1.00, p > 0.05).

HAMSTERS OPTIC NERVE

497

Fig. 6 The series of electronmicrographs ( x 6,550) is taken from a section 3.5 mm posterior to the globe and they are also from the nerve which is depicted in figures 4 and 5. In this instance the vast majority of the fibres in the central (A), midperipheral (B) and peripheral (C) portions of the cross section have myelin sheaths. The conventions are the same as those in figure 2 and the calibration is 1 pm.

While this limited analysis does not permit meaningful quantitative conclusions regarding the percentages of myelinated and nonmyelinated fibres in the portion of the hamster’s optic nerve within and immediately behind the lamina cribrosa it does indicate that the degree of myelination in this portion of the nerve is clearly different from that observed at slightly greater distances from the globe. It also points out that the most likely reason for the differences between the observations provided by Tiao and Blakemore (‘76) and those of the present study result from the fact that the area of the nerve they sampled was proximal (with respect to the retinal ganglion cells) to the point a t which the process of myelination is completed.

Fibre diameter spectrum The overall distributions for the myelinated and unmyelinated axons which we measured are shown in figures 7A and B, respectively. The size distribution for the 2,316 myelinated fibres is very similar to those which have been reported for the pigmented (Hughes, ’77) and albino (Forrester and Peters, ’67) rat. It was positively skewed and unimodal with a single peak a t about l p m . The fibre diameters ranged from 0.37 p m to 3.39 pm. The mean for the distribution was 1.27 p m and the median was 1.19 pm. These values are somewhat greater than those reported by Tiao and Blakemore (‘76) and a possible reason for this difference is discussed below. The distribution

for the small sample of unmyelinated axons was also unimodal with a mean of 0.59pm and a median of 0.63 pm. The diameters of the nonmyelinated fibres which we measured ranged from 0.20 p m t o 1.33 pm. The fact that the fibre diameter distributions for both the myelinated and unmyelinated fibres were unimodal was somewhat surprising in view of the clearly bimodal distributions of retinal ganglion cell diameters (Tiao and Blakemore, ’76)and retinocollicular conduction velocities (Rhoades and Chalupa, ’78 ’79) which have been reported for this species. Hughes and W a d e (‘76) have observed a similarly unimodal distribution of fibre diameters for entire optic nerve of the cat, but at the same time reported data which suggested that the distribution for the peripheral portion of the nerve in this species may have two or three modes. To test whether this might also be the case for the hamster we reanalyzed the data from three sets of eight electronmicrographs which could be clearly identified as being taken from the central, midperipheral or peripheral portions of the nerve (the criteria for inclusion in each of the three groups are detailed above). The histograms constructed for the nerve fibres measured in each of these regions are shown in figure 8. The data for the 1,268myelinated axons are depicted in figures 8A (peripheral nerve), B (midperipheral nerve), and C (central nerve). The mean for the central fibres was 1.29 p m and the median was 1.17 pm. The

498

ROBERT W. RHOADES, LINDA HSU AND GAIL PARFETT

A 2316

I.o

I-

Ly:

3.0

-B

Z W V

2.0

30 -

132

W

a 20

-

10 -

r

I

I

.

DIAMETER (MICRONS) Fig. 7 The distributions of axon diameters for 2,316 myelinated (A) and 132 unmyelinated (B) fibres measured in the sections taken 5mm behind the globe. The median diameter for the myelinated fibres is 1.19pm and that for the unmyelinated axons is 0.63 pm. See text for further details.

respective values for the midperipheral distribution were 1.30 and 1.19p m and those for the peripheral distribution were 1.22 and 1.12 pm. While there is some suggestion of a greater number of large axons in the central and midperipheral nerve than in the periphery a statistical test of this possibility did not yield significant results (F = 1.82, p > 0.05). The fibre diameter distributions for the small number of unmyelinated axons measured a t each eccentricity were also very similar. The mean diameters for the central, midperipheral

and peripheral samples were 0.68 pm, 0.54 pm, and 0.59 p m , respectively. The median value for the central distribution was 0.69pm, that for the midperipheral sample was 0.59 p m and the value for the peripheral sample was also 0.59 pm. The small number of unmyelinated axons encountered precluded any statistical comparison of the samples from the three areas. These findings for the optic nerve of the hamster would then indicate that, in contrast to the case in the cat (Hughes and Wassle, '761, fibres of all sizes are distributed

499

HAMSTER'S OPTIC NERVE

25

393

LL 15

20

E

LL w

'

O

l

L

L

L

L

L

17

l€L 20

0.5

1.0

DIAMETER (MICRONS) Fig. 8 The diameter distributions for the myelinated fibres measured in areas of 786 pm2 in the peripheral (A), midperipheral (B) and central (C) portions of the nerve cross section taken 5 mm behind the globe. The number a t the upper right in each histogram indicates the sample size. The median fibre diameters were 1.12 pm, 1.19 pm and 1.17 pm, respectively and there were no significant differences between the three distributions (F = 1.82, p > 0.05). (D), (El and (F) depict the diameter spectra for the unmyelinated axon8 measured in the peripheral, midperipheral and central portions of the cross section. The respective medians in this case were 0.59 p m , 0.59 pm and 0.69 pm.

relatively uniformly throughout the cross section of nerve. Quantitative analysis of nerve fibre myelination It has been noted in several fibre pathways (Friede and Samorajski, '67; Friede et al., '71; Hokoc and Oswaldo-Cruz, '78) that the thick-

ness of the myelin sheath can be predicted well by internal fibre diameter (the diameter of the axon excluding the myelin sheath). In order to determine whether or not this was the case for the optic nerve of the hamster the number of lamellae in the myelin sheath and the internal axon diameter (METHODS) were correlated for 150 fibres. Fifty of the axons

ROBERT W. RHOADES, LINDA HSU AND GAIL PARFETT

3 40

2c

AXON DIAMETER (MICRONS)

NUMBER OF LAMELLAE

.." .... ..........-. . . m

.

lot 5

w

15

;.I 5

z

5

I

.

. .. . .. ... .-- ." .. .. "..... : I

I I

e

.

I' .. . . 4 ......:.... . ..

15

0

...m

I

w

n

0

.

.

.

"

AXON DIAMETER (MICRONS) Fig. 9 The frequency distribution for the number of lamellae surrounding 150 myelinated axons from the sections taken 5 mm posterior to t h e eye. The mean for t h e distribution is 8.8 3.9. (B) The distribution of internal axon diameters for the fibres employed to obtain the data shown in (A); the shape is very similar to that for the larger sample where the external diameter of the fibre was measured (fig. 7A). The mean for the distribution is 1.14p m and the median is 0.96 Fm. (0,(D)and (E)depict the correlation diagrams for the 50 fibres analyzed in the peripheral (C), midperipheral (D) and central (El portions of the nerve c r o ~ ssection. There are no significant differences in the slopes or y-intercepts of the regression lines for each of the three scatterplots. See text for further details.

*

50 1

HAMSTERS OPTIC NERVE

used in the analysis were taken from the central portion of the nerve (as defined above), 50 from the midperipheral portion and 50 from the periphery. The results of the analysis are depicted in figure 9. The minimum number of lamellae observed in a given myelin sheath was three and the smallest myelinated fibre included in the sample was 0.45 pm. Two axons were encountered which had sheaths consisting of 20 lamellae and the axon diameters in these cases were 2.06pm and 1.91p m . As can be seen in figure 9B, the shape of the distribution of axon diameters measured in this sample matches fairly closely with that for the larger sample of fibres for which the external diameter was measured. Thus it is reasonable to expect that the data from the small sample taken here reflect relatively accurately the relationship between fibre size and myelination for the whole cross section of the nerve. Figures 9C, D and E depict the correlation diagrams for internal fibre diameter and the number of lamellae in the myelin sheath for the peripheral, midperipheral and central samples, respectively. Two conclusions are readily drawn from the scatterplots. First, at all eccentricities, the degree of myelination is predicted well by axon diameter. For the central fibres the correlation between these two variables was 0.80, for the midperipheral fibres i t was 0.83 and for the axons sampled in the periphery it was 0.82. The second point which is apparent from the data is that the degree of myelination for a fibre of a given size is virtually identical in the three areas of the nerve sampled. A regression line was fitted to each of the data sets using the method of least squares and the slopes for the central midperipheral and peripheral samples were 8.1 lamellae/pm, 7.8 lamellaelpm and 7.7 lamellae/pm, respectively. The results of a test for differences between the distributions produced no significant results (F < 1.00, p > 0.05). Thus, even though fibres in the central portion of the nerve are myelinated proximally to those in the periphery (see above), a t a point roughly midway between the globe and the chiasm fibres of a given caliber in all portions of the nerve tend to have equally thick myelin sheaths. Another discrepancy between our data and those of Tiao and Blakemore (‘76) which was mentioned above concerns the distribution of fibre diameters for those axons possessing myelin sheaths. The mean and median diam-

.. . . . .. . ’. . ... . .

W

a -I

-1

8

0

W

1

a

I

t

.

.

**

*’.’

-*

-1

LL

0

B

E W

m

z

3 16 Z

. . ... . ....... .. --..... .. . . -

8-

0

0

. . I

i I

U .

1

I .o

2.0

3.0

AXON DIAMETER (MICRONS) Fig. 10 (A) The correlation diagram relating internal axon diameter and number of lamellae for 50 fibres from the center of a nerve cross section taken 1.5 mm behind the globe. The slope for the regression line fitted to these data was 5.03 lamellae/pm and the y-intercept was -0.70. (B)A similar scatterplot for 50 central fibres from the same nerve in a section taken 3.5 mm posterior to the eye. The slope of the regression line in this case was 6.43 lamellae/pm and the y-intercept is +0.67. The difference between the two distributions was significant (t = 4.69, p < 0.01). See text for further details.

eters for such fibres, calculated from their grouped data, were 0.99 p m and 0.95 pm, respectively. These are somewhat smaller than the values which we obtained (1.27 p m and 1.19 pm). It is possible that this discrepancy can be explained by sampling differences since the number of myelinated axons which they measured was relatively small. It is also possible that the discrepancy may have resulted

502

ROBERT W. RHOADES, LINDA HSU AND GAIL PARFETT

from a difference in either axonal diameter or myelin sheath thickness in the portions of the nerve sampled in the two studies. Our qualitative impressions suggested that those axons which were myelinated in the distal portion of the nerve had thinner sheaths than those from the sections farther behind the globe. We tested this possibility in one nerve by measuring internal axon diameters and counting the number of lamellae surrounding 50 myelinated axons in the central portion of the thin section taken 1.5 mm behind the eye and comparing the data with those obtained for 50 central axons from the section 3.5 mm posterior to the eye. The results of this analysis are shown in figure 10. In the proximal segment the average internal diameter for the fibres we measured was 1.21 pm, a value which was not significantly different from that (1.19 pm) obtained for the central fibres measured in the segment 3.5 mm behind the globe (t < 1.00, p > 0.05). In the proximal segment, however, the average number of lamellae surrounding a l p m fibre was only 5.0 and this was significantly fewer than the average (8.3) obtained from the section taken farther from the globe (t = 4.69, p < 0.01). Thus, it would appear that a t least a portion of the difference between the fibre diameter spectrums provided by the two studies may be the result of a systematic difference in the thickness of the myelin sheaths included in the measurements. DISCUSSION

The anatomical organization of the hamster's optic nerve is very similar to that of the rat (Forrester and Peters, '67; Treff et al., '72; Hughes, '77). It contains approximately 110,000 fibres which are distributed with a fairly even density (592,000/mm2)throughout the nerve. The vast majority (96.4%)of these axons are myelinated and, as is the case in the rat, the spectrum of fibre diameters is positively skewed and unimodal with a fairly extensive "tail." The thickness of the myelin sheath surrounding a given axon is proportional to the axonal diameter and is also relatively constant throughout the nerve's cross section. Our estimate of the total number of optic nerve fibres agrees well with that reported by Tiao and Blakemore ('76) and also with their estimate of the number of ganglion cells in the retina of this species. With these exceptions the results of the present study are clearly dif-

ferent from those of Tiao and Blakemore. The discrepancy in the estimates of the percentage of myelinated fibres from the two studies and the probable cause of this difference has been discussed in detail above. Another difference between our results and those of Tiao and Blakemore, which is less easily explained, concerns the fibre density in the hamster's optic nerve. As was noted above we found the axon density to be relatively uniform throughout the cross section of the nerve. Tiao and Blakemore, on the other hand, reported that the packing density varied from 370,000/mm2 a t the core to approximately 2,350,000/mm2 a t the edge of the nerve. This latter value is over three times that which has been reported for a given portion of the optic nerve in any mammal. It is also much higher than the densities which we observed a t the edge of the nerve in sections taken either 0.5 mm (1,127,000/mm)2 or 1.5 mm (680,000/mm2)behind the globe. Tiao and Blakemore ('76) also suggested that the mean fibre size for both myelinated and unmyelinated axons increased slightly in the periphery of the nerve. This was not the case for the sample of axons which we measured. Such an increase in fibre size as the edge of the nerve is approached would be predicted on the basis of the fact that the relative incidence of large ganglion cells increases in the peripheral retina (Tiao and Blakemore, '76) and the additional assumptions that fibre size is positively correlated with ganglion cell diameter and that the topography of the retina is maintained in the optic nerve. This would appear to be the case in the cat where the relative percentage of large ganglion cells is greatest in the retinal periphery (Stone, '65, '78; Fukuda and Stone, '74) and the relative incidence of large axons is also greatest near the edge of the optic nerve (Hughes and Wassle, '76). A close topographic relationship between retina and optic nerve may also be the case in the monkey. In this species both Van Buren ('63) and Bunt et al. ('75) have reported that small ganglion cells dominate the central retina and in the optic nerve there is a corresponding high density core of relatively small axons (Potts et al., '72). In both the rabbit (Vaney and Hughes, '76) and rat (Fukuda, '77) the relative incidence of large ganglion cells is also greatest in the periphery of the retina. In both of these species, however, there is no increase in the relative percentage of large axons near the boundary of the optic nerve (Vaney and Hughes, '76; Hughes, '77). The

503

HAMSTER'S OPTIC NERVE

finding in t h e latter species lends support to Lashley's ('34) earlier observation that there is little retinotopic order in the optic nerve of the rat. If it is assumed that the relationship between conduction velocity and fibre diameter is linear for myelinated axons (Ogden and Miller, '66; Waxman and Bennett, '72; Waxman and Swadlow, '77) one would also expect a close correspondence between the distributions of conduction velocities and axon diameters in the optic nerve. Clearly this has not been t h e case. Studies in t h e hamster (Rhoades and Chalupa, '78, '79; the present study), rat (Forrester and Peters, '67; Sumitom0 et al., '69; Hughes, '77; Fukuda, '77; Fukuda et al., '781, opposum (Rowe et al., '76; Freeman and Watson, '781, and cat (Stone and Freeman, '71; Stone and Fukuda, '74; Kirk et al., '75; Hughes and Wassle, '76; Rowe and Stone, '76) have all failed to obtain clear correlations between the conduction velocity distribution and the fibre diameter spectrum for the optic nerve. A similar discrepancy exists in those instances where t h e attempt has been made to relate t h e compound action potential recorded from the optic nerve with the distribution of fibre diameters (Bishop et al., '53; Ogden and Miller, '66; Forrester and Peters, '67; Vaney and Hughes, '76; Freeman and Watson, '78). This lack of correlation has prompted several investigators (Landau et al., '68; O'Flaherty, '71; Vaney and Hughes, '76; Freeman and Watson, '78) to use empirical methods in reconstructing the optic nerve compound action potential from the frequency distribution of fibre diameters. In most of those instances in which physiological data have been related to anatomical findings obtained with the electronmicroscope (Vaney and Hughes, '76; Freeman and Watson, '78) these methods have also been unsuccessful. One possible reason for this lack of correlation, at least in the hamster, may stem from the fact t h a t many optic nerve fibres which will eventually become myelinated have no myelin sheath for approximately 20% of their travel in the optic nerve. It has been noted in several species (Wendell-Smith e t al., '66; Stone and Hollander, '71; Freeman, '78; Bussow, '78) that ganglion cell axons traveling in the nerve fibre layer of the retina are unmyelinated (a partial exception to this rule is the rabbit; see Blunt et al., '65) and that the intraretinal portion of the ganglion cell axon conducts more slowly than the medullated seg-

ment of t h a t axon in the optic nerve (Stone and Freeman, '71; Kirk e t al., '75; Schiller and Malpeli, '77; Caldwell and Daw, '78). The present work points out that for a t least one species the process of myelination is incomplete for a significant portion of the length of the optic nerve. Thus if this portion of the nerve is included in the segment over which conduction velocity is measured the values obtained for many axons will be the average of the conduction velocity in the myelinated and nonmyelinated portions of a single fibre. ACKNOWLEDGMENTS

This work was supported in part by NIMH Grant 32897 to R.W.R. and by N.J.O.E.F. Grants 39-2510 to R.W.R. and 39-2501 to LH. Thanks to Eric Proshanksy for help with the light photomicroscopy and to Ms Sharon Mayer Scoles for typing the manuscript. LITERATURE CITED Bishop, P. O., D. Jeremy and J. W. Lance 1953 The optic nerve. Properties of a central tract. J. Physiol. (London), 121: 415-432. Blunt, M. J., C. P. Wendell-Smith and F. Baldwin 1965 Glia-nerve fibre relationships in the mammalian optic nerve. J. Anat., 99: 1-11. Bunt, A. H., A. E. Hendrickson, J. S. Lund, R. D. Lund and A. F. Fuchs 1975 Monkey retinal ganglion cells: morphometric analysis and tracing of axonal projections with a consideration of the peroxidase technique. J. Camp. Neur., 164: 265-286. Bdssow, H. 1978 Schwann cell myelin ensheathing C.N.S. axons in the nerve fibre layer of the cat retina. J. Neurocytol., 7: 207-214. Caldwell, J. H., and N. W. Daw 1978 New properties of rabbit retinal ganglion cells. J. Physiol. (London), 276: 257-276. Cohen, A. 1967 Ultrastructural aspects of the human optic nerve. Invest. Ophthalmol., 6: 294-308. Forrester, J., and A. Peters 1967 Nerve fibres in the optic nerve of the rat. Nature, 214: 245-247. Freeman, B. 1978 The retinal origins of the optic nerve conduction latency groups in the brush-tailed possum, Trichosurus uulpecula. J. Camp. Neur., 179: 753-760. Freeman, B., and C. R. R. Watson 1978 The optic nerve of the brush-tailed possum, Trichosurus uulpecula: fibre diameter spectrum and conduction latency groups. J. Camp. Neur., 179: 739-752. Friede, R. L., T. Miyaghishi and K. H. Hu 1971 Axon calibre, neurofilaments, microtubules, sheath thickness and cholesterol in cat optic nerve fibres. J.Anat., 108: 365-373. Friede, R. L., and T. Samorajski 1967 Relation between the number of myelin lamellae and axon circumference in fibers of vagus and sciatic nerve nerve of mice. J. Camp. Neur., 130: 223-232. Fukuda, Y. 1977 A three group classification of rat retinal ganglion cells: histological and physiological studies. Brain Res., 119: 327-344. Fukuda, Y., and J. Stone 1976 Retinal distribution of Y-, Xand W-cells of the cat's retina. J. Neurophysiol., 37: 749-772.

504

ROBERT W. RHOADES, LINDA HSU AND GAIL PARFETT

Fukuda, Y., D. A. Suzuki and K. Iwama 1978 A four group classification of t h e rat superior collicular cells responding to optic nerve stimulation. Jap. J. Physiol., 28: 367-384. Hokoc, J. N., and E. Oswaldo-Cruz 1978 Quantitative analysis of the opposum’s optic nerve: an electronmicroscope study. J. Comp. Neur., 178: 773-782. Hughes, A. 1977 The pigmented r a t optic nerve: fibre count and fibre diameter spectrum. J. Comp. Neur., 176: 263-267. Hughes, A,, and H. Wassle 1976 The cat optic nerve: total count and diameter spectrum. J. a m p . Neur., 169: 171-184. Kirk, D. L., B. G. Cleland, H. Wassle and W. R. Levick 1975 Axonal conduction latencies of cat retinal ganglion cells in t h e central and peripheral retina. Ex.Brain Res., 23: 85-90. Lashley, K. S. 1934 The mechanism of vision. VII. The projection of the retina uDon the Drimarv oDtic centers in the .at J Comp Neur , i9 341-i73 Landau, W M , M H Clare and G H Bishop 1968 Recon struction of myelinated nerve tract action potentials: an arithmetic method. Exp. Neurol., 22: 480-490. Ogden, T. E., and R. F. Miller 1966 Studies of the optic nerve of th e rhesus monkey: nerve fiber spectrum and physiological properties. Vision Res., 6: 485-506. OFlaherty, J. J. 1971 The optic nerve of the mallard duck: fiber diameter frequency distribution and physiological properties. J. Comp. Neur., 143: 17-24. Peters, A. 1966 The node of Ranvier in t he central nervous system. Quart. J. Exp. Physiol., 51: 229-236. Peters, A,, S. Palay and H. F. Webster 1970 The fine structure of the nervous system: t he cells and their processes. New York, Harper and Row. Potts, A. M., D. Hcdges, C. B. Shelman, K. J. Fritz, N. S. Levy and Y. Magnall 1972 Morphology of the primate optic nerve. 11. Total fibre size distribution and fibre density distribution. Invest. Ophthalmol., 1 I: 989-1003. Rhoades, R. W., and L. M. Chalupa 1978 Conduction velocity distribution of the retinocollicular pathway in the golden hamster. Brain Res., 159: 306-401. 1979 Conduction velocity distribution of the retinal input to the hamster’s superior colliculus and a correlation with receptive field properties. J. Comp. Neur., 184: 243-264. Rowe, M. H., and J. Stone 1976 Conduction velocity groupings among axons of cat retinal ganglion cells and their relationship to retinal topography. Exp. Brain Res., 25: 339-357. “

I

Rowe, M. H., E. Tancred, B. Freeman and J. Stone 1976 Properties of ganglion cells in the retina of the brush-tailed possum, Ifi.ichosurus uulpecula. Neurosci. Absts., 11; 1089. Schiller, P. H., and J. G. Malpeli 1977 Properties and tectal projections of monkey retinal ganglion cells. J. Neurophysiol., 40: 428-445. Stone, J. 1965 A quantitative analysis of the distribution of ganglion cells in the cat’s retina. J. Comp. Neur., 124: 337-352. 1978 The number and distribution of ganglion cells in the cat’s retina. J. Comp. Neur., 180: 753-772. Stone, J., and R. B. Freeman 1971 Conduction velocity groups in the cat’s optic nerve classified according to their retinal origin. Exp. Brain. Res., 13: 489-497. Stone, J., and Y. Fukuda 1974 Properties of cat retinal ganglion cells: a comparison of W-cells with X- and Ycells. J. Neurophysiol., 37: 772-748. Stone, J., and H. Hollhder 1971 Optic nerve axons measured in the cat retina: some functional considerations. Exp. Brain Res., 23: 498-503. Sumitomo, I., K. Ide, K. Iwama and T. Arikumi 1969 Conduction velocity of optic nerve fibers innervating the lateral geniculate body and superior colliculus in the rat. Exp. Neurol., 25: 378-392. Tansley, K. 1956 Comparison of the lamina cribrosa in mammalian species with good and indifferent vision. Br. J. Ophthalmol., 40: 178-182. Tiao, Y-C., and C. Blakemore 1976 Regional specialization in the golden hamster’s retina. J. Comp. Neur., 168: 439-458. Treff, W. M., E. Meyer-Koning and W. Schlote 1972 Morphometric analysis of a fibre system in the central nervous system. J. Microsc., 95: 337-343. Van Buren, J. M. 1963 The retinal ganglion cell layer. Charles C Thomas, Springfield. Vaney, D. I., and A. Hughes 1976 The rabbit optic nerve: fiber diameter spectrum, fiber count, and comparison with retinal cell count. J. Comp. Neur., 170; 241-252. Waxman, S. G., andM. V. L. Bennett 1972 Relative conduction velocities of small myelinated and non-myelinated fibres in the central nervous system. Nature New Biol., 238: 217-219. Waxman, S. G., and H. A. Swadlow 1977 The conduction properties of axons in central white matter. Progress in Neurobiol., 8: 297-324. Wendell-Smith, C. P., M. J. Blunt and F. Baldwin 1966 The ultrastructural characterization of macroglial cell types. J. Comp. Neur., 127; 219-240.