ABSTRACT

Purpose: Previously, measurements of retinal ganglion cellaxon diameter have been used to make inferences aboutthe physiology and clinical pathology of the visual pathway.However, few of these studies were able to unequivocallyrelate axon diameter to retinal ganglion cell type and otherassociated measurements. In this and our previous study wehave examined intraretinal axon diameters to determine ifdifferences in axon diameter may help to explain conduc-tion velocity measurements found previously.

Methods: Individual retinal ganglion cells of a New Worldprimate, the common marmoset (Callithrix jacchus) wereinjected iontophoretically with 2% Lucifer yellow and 4%neurobiotin. Labelled cells were visualized by horseradishperoxidase immunohistochemistry and diaminobenzidineand then retinae were mounted vitreal side up on a glassslide. Cell measurements were made with the aid of acamera lucida attachment and computer-aided morphome-try. Axons were photographed under ×100 oil immersionand measured at a final magnification of ×4600.

Results: A sample of 62 parasol cells, 22 midget cells, 16hedge cells and 11 small bistratified cells were analysed.Dendritic field diameter of the different cell classes showedonly moderate (non-significant) increases with eccentricity.Only the parasol cells demonstrated a significant increase inmean axon diameter with eccentricity. When the parasolclass was examined more closely, it was found that onlyparasol cells of the superior, inferior and temporal retina(SIT group) showed significant positive correlationsbetween different cell parameters (mean axon diameter,soma diameter, dendritic field diameter, eccentricity). Somaand dendritic field diameters of the SIT group were signifi-cantly larger than those of the nasal parasol cells. However,mean axon diameters of the SIT cells were not significantly

different from nasal parasol cells. Axon diameters of nasalparasol cells were very variable and overlapped those ofthe midget and hedge cell classes to a large extent.

Conclusions: The present data show that for marmosetparasol cells there may not be a clearly defined distinctionbetween nasal and superior, inferior and temporal parasolcells on the basis of axon size. Of particular interest in thepresent analysis is the clear separation of superior, inferiorand temporal parasol cells and nasal parasol cells whencomparing soma and dendritic field diameters which is notreflected in the distribution of axon diameters.We suggestthat changes in diameter along the length of an axon,differences between retinal quadrants and the variabilitybetween cells may be related to minimization of spatio-temporal dispersion necessary for accurate perception ofmotion within the visual world.

Key words: axon diameter, biotin, eccentricity, parasol cells,primate, retina.

INTRODUCTION

A number of inferences about the physiology and pathologyof the visual system have been made based on axon morphology and diameter measurements.1–8 Axon measure-ments along the optic pathway show a close correspondenceto size differences between the major retinal ganglion cellclasses and the t1, t2 and t3 conduction velocity groups.6,9–16

Other authors have reported a positive correlationbetween soma and axon diameters and conduction veloc-ity.10,17–20 These correlations and reports demonstrating thatsoma and dendritic field diameters increase with eccentric-ity13,14,21–26 suggest that conduction velocity should alsoincrease. However, other data suggest that conductionvelocity, in both the cat and the Japanese monkey (Macacafuscata), is not graded with respect to eccentricity.10,20,27–29

Furthermore, while numerous authors have shown that

Clinical and Experimental Ophthalmology (2000) 28, 423–430

Original Article

Intraretinal axon diameters of a New World primate,the marmoset (Callithrix jacchus)Natalie Walsh BAppSci(Orth),1,4,5 Krishna K Ghosh PhD1,2 and Thomas FitzGibbon PhD1,3,4

1Institute for Biomedical Research, Departments of 2Physiology and 3Anatomy and Histology, 4Save Sight Institute, Department of ClinicalOphthalmology and 5School of Orthoptics, University of Sydney, Sydney, New South Wales, Australia

■ Correspondence: Dr Thomas FitzGibbon, Department of Anatomy and Histology (F13), Institute for Biomedical Research, University of Sydney, Sydney, NSW

2006, Australia. Email: tomf@cortex.physiol.usyd.edu.au

temporal soma size is on average larger than nasal somasize,2,22,23,30–32 data showing that temporal fibres havehigher conduction velocities are equivocal.7,9,10,29,33 In addi-tion, there is also evidence to suggest that axon diameterand conduction velocity within the visual pathway do notremain constant.2–4,7,31,34–36.

In this and our previous reports31,36 we have examinedintraretinal axon diameters in order to determine if differ-ences between and within retinal ganglion cell classes mayhelp to explain previous conduction velocity measure-ments.8,10,17,20,27,35 There are comparatively few studieswhich have correlated axon diameter directly with soma anddendritic field size, eccentricity and retinal location at thesingle cell level.25,27,35 In this study we present data obtainedfrom individually labelled retinal ganglion cells of a NewWorld primate, the common marmoset (Callithrix jacchus).

METHOD

The methodology used to inject and stain retinal ganglioncells has been published previously and will only be outlinedbriefly.23,36,37 All procedures conformed to guidelines set outby the University of Sydney Animal Ethics Committee.

At the completion of physiological experiments, theanimals were overdosed with barbiturate and the eyesremoved. The retina was dissected free from the adjacenttissues, cut into 3–4 pieces, and placed into carboxygenatedAmes medium. The retinal cells were stained with 0.001%acridine orange and then individual cells were penetratedwith borosilicate glass micropipettes filled with a solution of2% Lucifer yellow and 4% neurobiotin in a 0.1 mol/L TrisHCl buffer (pH 7.6). Neurobiotin was injected using0.1–0.3 nA positive current for 20 s to 2 min, depending oncell type and size.

Retinae were fixed in phosphate-buffered 4%paraformaldehyde. To visualize the neurobiotin the retinaewere processed by horseradish peroxidase immunohisto-chemistry and diaminobenzidine, then mounted vitreal sideup on a glass slide in 20% Mowiol in a 1:1 glycerol: watersolution and cover-slipped without counterstaining.Maximal shrinkage was estimated to be 10% linear inMowiol.

Cell measurements (soma diameter, dendritic field diam-eter, eccentricity) and retinal maps were made with the aidof a camera lucida attachment and computer-aided morphom-etry. Only cells which had complete staining of the soma,dendritic field and at least 200 µm of the axon were includedin the study. Retinal pieces were reconstructed to form awhole flatmounted retina in order that eccentricity measure-ments and the distance of the cell body from the optic disccould be calculated (Fig. 1a). The foveal magnificationfactor for the marmoset is 128 µm per degree of visualangle.38 The relationship between visual angle and distancefor the retina at greater eccentricities can be calculated bythe formula y = 0.0538x3 – 0.542x2 + 9.396x – 0.75.38

The retina was then subdivided into nasal, temporal,superior and inferior quadrants. Retinal cuts were made at a

45° angle to the horizontal meridian passing through thefovea and optic disc as shown in Fig. 1a. The retinae weredivided in this manner in order to avoid damaging cells in the foveal and parafoveal regions.23,37 This method ofsubdivision also avoided transecting fibres within thearcuate bundles39,40 allowing a greater length of labelledaxon to be traced across the retina.

Axons were photographed under ×100 oil immersion(Nikon 160/0.17 objective; Nikon, Tokyo, Japan) and thenegatives projected to a final magnification of ×4600.Measurements were then taken at regular intervals along thelength of the labelled axon from the projected negatives.Seven measurements (10 µm apart) per location were thenaveraged to gain a single axon diameter measurement foreach location along the axon. For each cell, a mean axondiameter was calculated by averaging measurements taken atall locations.

424 Walsh et al.

Figure 1. (a) Schematic diagram of the reconstructed retinalwhole-mount illustrating how the retina was subdivided. Theeccentricity (e) of the cell with respect to the fovea (F) was measured as well as the distance (d) of the soma from the optic disc (OD). For each cell, measurements of soma and dendritic fielddiameter were determined by converting the area into an equiva-lent diameter. (b) The number of each cell type measured withinthe nasal, temporal, superior and inferior portions of the retina .

, Parasol; h, midget; , hedge; j, small bistratified cells.

RESULTS

Cells were classified as either parasol, midget, small bistrat-ified (SBS) or hedge cells, or were unclassified as outlined inprevious reports.23,37 The numbers of each cell type andtheir retinal locations are summarized in Figure 1b. The

number of cells in each group in no way implies relativedensity of the separate cell populations. Note that a pre-ponderance of parasol cells were sampled compared to themore numerous midget cell population. This occurredbecause of inherent difficulties in obtaining well-filled cellsand axons at eccentricities less than 2 mm due to the smallsoma diameters in the region of highest cell density near thefovea and the thickness of the nerve fibre layer in theparafoveal region. Also, the resolution of foveal axons whichcan be smaller than about 0.4 µm becomes unreliable at thelight microscopic level. Examples of labelled cells and axonsare shown in Fig. 2 and in our previous reports.23,37

Comparisons between cell classes

A total of 111 retinal ganglion cells from four retinae weresuitable for measurement. Histograms summarizing data forsoma diameter, mean axon diameter and dendritic fielddiameter measurements are shown in Fig. 3. Note the lack ofthe smallest midget and non-parasol cell somata (> 10 µm)due to undersampling.23,37,38,49 The lack of foveal cells in the sample is also reflected in the mean axon diameter for all axons (1.09 µm, SD = 0.3) compared to measurementsmade with the electron microscope (0.5 µm, SD = 0.2,n = 6668; FitzGibbon, unpublished results). Also note thedegree of overlap between cell classes for all measurements.In order to examine any changes associated with retinalposition the data in Fig. 3 were replotted as a function ofeccentricity (Fig. 4).

Previously we have shown that, although the trends werepositive, only midget cells showed a positive correlationbetween soma size and mean axon diameter (r = 0.787,P < 0.01).36 Dendritic field diameter and mean axon diame-ter of both parasol and midget cells showed significant cor-relations (r = 0.357, P < 0.05 and r = 0.734, P < 0.01,respectively) although this trend was not significant forhedge and SBS cells. In general, dendritic field diametershowed only moderate (non-significant) positive changesover the eccentricities sampled in this study (Fig. 4a). Note,however, that the SBS cells were obtained over a limitedrange of eccentricities. The cell classes were more clearlydefined in these graphs although there was some overlap atall eccentricities. As shown in Fig. 4, mean axon diameterwithout reference to cell class (grouped data) was positivelycorrelated with changes in eccentricity (P < 0.01); however,of the four cell groups only parasol cell axon diameter was significantly correlated with eccentricity (r = 0.277,slope = 4E-05, P < 0.05). From Fig.4b it is clear that ateccentricities greater than approximately 3.5 mm, thelargest axons found were most likely to be parasol axons,although some overlap between parasol axon diameters andother cell classes occurs at all eccentricities. Although thehedge cells show considerable variation in dendritic fielddiameter, axon diameters appear to fall within a relativelynarrow range and are generally the smallest axons seen at alleccentricities. Comparison between the trend lines (F-test ofthe z-values) showed that there were no significant differ-ences between cell classes.

Intraretinal axon diameter 425

Figure 2. Photomicrographs of cell somata and axons which have been filled with neurobiotin and visualized with diamino-benzidine. (a) Parasol cell and axon (solid arrowheads; soma diameter = 18.2 µm, dendritic field diameter = 109 µm, mean axondiameter = 1.3 µm, eccentricity = 3.5 mm). The axon indicated bythe full arrow is connected to a more peripheral parasol cell. Thecell body of the very fine axon indicated by the open arrowheadswas not identified but on the basis of its fine gauge it is presumedto be that of a peripheral C-cell. (b) Midget cell from the superiorretina (soma diameter = 20.1 µm, dendritic field diameter = 82 µm,mean axon diameter = 0.9 µm, eccentricity = 5.9 mm). (c) Parasolcell from nasal retina (soma diameter = 17.0 µm, dendritic fielddiameter = 116 µm, mean axon diameter = 1.3 µm, eccentric-ity = 3.5 mm). (d) Hedge cell from superior retina (somadiameter = 14.1 µm, dendritic field diameter = 250 µm, mean axondiameter = 0.8 µm, eccentricity = 6.2 mm). Note that this cell hasa large dendritic field and is more eccentric than the other cellsshown, however the axon is clearly finer than parasol or midgetaxons. (Scale bar = 40 µm.)

Comparison of nasal and non-nasal parasol cells

Previously we showed that cells in the temporal, inferior andsuperior quadrants show the same increase in dendritic fieldsize in relation to eccentricity. In contrast, cells of the nasalretina have a much lower rate of increase in dendritic fieldsize with eccentricity and do not increase appreciably ateccentricities greater than 5 mm. In addition, somata ofnasal cells are on average smaller at equivalent eccentrici-ties.23,37 Therefore, as in our previous studies, superior, infer-ior and temporal parasol cells (SIT) were grouped together(n = 36) and compared to nasal cells (n = 26) as shown inFig. 5. Due to low cell numbers of non-parasol cells in eachquadrant these were not analysed further.

All the correlations shown for the SIT parasol cells inFig. 5 were significant (P < 0.05–0.01). By contrast, nasalcells only showed a significant positive correlation betweendendritic field diameter and soma diameter (P < 0.05;Fig. 5b) and a significant negative correlation between somasize and eccentricity (Fig. 5d). Note that nasal cells gener-ally do not have soma diameters greater than 25 µm.

It is apparent from Fig. 5a–c that nasal parasol cellstended to have larger axons than SIT cells within a similarsoma and dendritic size range. It should be noted that whileparasol cell soma and dendritic field diameters of the nasalretina were on average smaller than SIT parasol cells, therange of axon diameters between nasal and SIT cells wasvery similar. Comparison of the trend lines (F-test of the z-values) showed that in all three comparisons there was asignificant difference in the slopes of the trend lines indicat-ing that nasal and SIT populations are behaving in differentways.

The differences between nasal and SIT cells were readilyapparent when the data were plotted as a function of eccen-tricity (Fig. 5d–f). Although both SIT and nasal somatashowed a significant correlation with eccentricity (Fig. 5d)comparison between the trend lines did not reach signifi-cance (P > 0.07) and appears to be due to the greater variability of the nasal soma diameters at smaller eccentrici-ties. In contrast, only SIT dendritic field diameters were significantly correlated to eccentricity and there was a significant difference between the two trend lines (Fig. 5e).

426 Walsh et al.

Figure 3. Distribution of soma diameter, dendritic field diameter and axon diameter for sample of cells analysed. (a) Parasol, n = 62; (b) midget, n = 22; (c) hedge, n = 16; (d) small bistratified cells, n = 11. Mean ± SD.

These results indicate that dendritic field and soma dia-meters plateau or decrease with increasing eccentricity.Although only SIT axons showed a significant correlation

with eccentricity, comparison of the trend lines showedthere were no significant differences suggesting that bothgroups tend to increase with eccentricity. Differences inmean axon diameter between these groups are not asmarked as that seen between somata and dendritic fields andthere is considerable overlap between the nasal and SITmean axon diameter at all eccentricities (Fig. 5f). There is adegree of scatter in the mean axon diameter of nasal cells ateccentricities greater than 7.5 mm which is not reflected inthe distribution of soma and dendritic field diameter at thesame eccentricity.

DISCUSSION

Our data support and extend previous reports in which therelationships between retinal axon diameter and other cellparameters have been examined.3,25,31,36 Rodieck et al.reported that there was a significant positive correlationbetween axon diameter and eccentricity for human parasoland midget cell axons.25 The present results indicate that, ingeneral, for the marmoset there is a modest increase in axondiameter with respect to eccentricity for all cells but whichis more apparent for the temporal parasol cells (Figs 4,5).Previously, we have shown that dendritic field diameter andmean axon diameter of both parasol and midget cells weresignificantly correlated; however, while SBS and hedge cellsshow similar trends these were not significant.36 Although inthe present study only the parasol cell class was examined in relation to nasal and temporal differences, the results ofGhosh et al. indicate that all cell classes within the nasalretina show only modest changes in soma or dendritic fielddiameter with respect to eccentricity (i.e. at eccentricities of2–10 mm), compared to SIT retinal quadrants.23 However,the relationship of axon diameter to cell class and eccentric-ity may be masked by size differences between on- and off-centre cells wherein on-centre cells are approximately30–50% larger than their off-centre counterparts.21,23,41

Of particular interest in the present analysis is the clearseparation of SIT and nasal parasol cells when comparingsoma and dendritic field diameters which is not reflected inthe distribution of axon diameters. The graphs comparingsoma and dendritic field size with respect to eccentricityshow a distinct clustering of peripheral nasal parasol cells (ateccentricities greater than 7 mm). In contrast, axon diame-ters for this same group of cells are very variable and overlapthose of the midget and hedge cell classes to a large extent.In a number of species, including primates, there is evidenceto suggest that somata of temporal cells and axons are onaverage larger than nasal cells at equivalent eccentrici-ties.2,3,14,21–24,31,33,42,43 The present data show that for mar-moset parasol cells at least, there may not be a clearlydefined distinction between nasal and SIT axon size. Whileonly parasol axons were analysed in this respect, previousanalysis of soma and dendritic field diameters of other cellclasses suggests the same trends will be apparent.23,37 In thisrespect, the significant variables affecting axon diameter arecell class and eccentricity for both nasal and SIT cells rather

Intraretinal axon diameter 427

Figure 4. (a) Dendritic field diameter as a function of eccentric-ity. (b) Mean axon diameter as a function of eccentricity. d, Hedge; h, parasol; +, midget; n, small bistratified cells.

than retinal quadrant. These results and the overlap in axonsize between cell classes also have implications for currenttheories related to glaucomatous damage.1,40,44–46

Implications for conduction velocity

The present results indicate that axon diameters tend to belarger as the parent soma is found at greater eccentricities.Furthermore, axon measurements in other studies suggestthat axon diameter does not remain constant along thevisual pathway. Wässle et al. reported that there was a goodcorrelation between soma size and conduction velocity,17

and on this basis it is suggested that conduction velocity

would also show an increase with eccentricity. However,data obtained from both the cat and the Japanese monkey(Macaca fuscata), suggest that this is not the case.7,9,10,20,27,29,33

In contrast to some of this earlier work, the present resultssuggest that in the marmoset nasal and SIT parasol conduc-tion velocities may be very similar but gradually increasewith eccentricity, albeit with a large amount of variability.Furthermore, the axon volume/diameter changes along theintraretinal axon that we have demonstrated previously31,36

and within the optic nerve and tract4,7,8,31,34 imply that conduction velocity within the visual pathway will not beconstant. Indeed, the variation in axon size demonstrated for some parasol cells suggests that conduction velocity

428 Walsh et al.

Figure 5. Parasol cells from the superior, inferior and temporal (SIT) quadrants (r) grouped together and compared to parasol cellssampled in the nasal quadrant (✳). (a)–(c) Comparisons between the various cell parameters: soma diameter, dendritic field diameter andmean axon diameter; (d)–(f) cell soma diameter, dendritic field diameter and mean axon diameter as functions of eccentricity. (a) SIT:slope = 0.0314; r = 0.548**; nasal: slope = 0.0001, r = 0.001; F(z) = 0.99; P = 0.01*. (b) SIT: slope = 9.6352, r = 0.776**; nasal:slope = 4.0002, r = 0.436*; F(z) = 0.98; P = 0.02*. (c) SIT: slope = 0.003, r = 0.655**; nasal: slope = 0.0024, r = 0.297; F(z) = 0.96; P = 0.03.(d) SIT: slope = 0.019, r = 0.766**; nasal: slope = –0.0009, r = 0.551**; F(z) = 0.93; P = 0.07. (e) SIT: slope = 0.0198, r = 0.651**; nasal:slope = –0.0026, r = 0.179; F(z) = 0.99; P = 0.01*. (f) SIT: slope = 5E–05, r = 0.372*; nasal: slope = 2E–05, r = 0.136; F(z) = 0.8; P = 0.2.**Significance at the 0.01 level; *significance at the 0.05 level.

could change by approximately 0.2–0.3 m/s (velocity =1.8 √D – 0.24) over the first 2 mm of intraretinal length.27,47

Stanford, using single cell recordings from the X-cellclass of the cat retina, demonstrated that although there wasconsiderable variation in the conduction velocity betweenX-cells, there was a strong negative correlation betweenintra- and extraretinal conduction times.35 The effect of thisrelationship was an almost constant transmission time fromsoma to central targets within the brain for the X-cell classregardless of changes in eccentricity of the soma. Stanfordpostulated that a constant conduction time could thus main-tain the spatiotemporal central representation of the image;however, he did not elaborate on how this constant conduc-tion time might be achieved. We have argued in our previ-ous report that changes in diameter along the length of anaxon and the variability between cells may be a necessity tominimize spatiotemporal dispersion for accurate perceptionof motion within the visual world. Of particular interest tothese theories are those patients with intraretinal myelina-tion in which conduction velocity variations might be exac-erbated.4,48

ACKNOWLEDGEMENTS

This research was supported by grants from the SydneyFoundation for Medical Research and NHMRC (0931011).Author K Ghosh was a University of Sydney PostgraduateMedical Foundation Research Scholar and is currently a vonHumbolt Postdoctoral Fellow.

REFERENCES

1. Quigley HA, Sanchez RM, Dunkelberger GR, L’Hernault NL,Baginski TA. Chronic glaucoma selectively damages largeoptic nerve fibres. Invest. Ophthalmol. Vis. Sci. 1987; 28: 913–20.

2. FitzGibbon T, Funke K. Retinal ganglion cell axon diameterspectrum of the cat: mean axon diameter varies according toretinal position. Vis. Neurosci. 1994; 11: 425–39.

3. FitzGibbon T, Wingate RJ, Thompson ID. Soma and axondiameter distributions and central projections of ferret retinalganglion cells. Vis. Neurosci. 1996; 13: 773–86.

4. FitzGibbon T, Nestorovski Z. Morphological consequences ofmyelination in the human retina. Exp. Eye Res. 1997; 65:809–19.

5. Williams RW, Chalupa LM. An analysis of axon caliber withinthe optic nerve of the cat: evidence of size groupings andregional organisation. J. Neurosci. 1983; 3: 1554–64.

6. Bishop GH, Clare MH, Landau WM. Further analysis of fibergroups in the optic tract of the cat. Exp. Neurol. 1969; 24:386–99.

7. Freeman B. Myelin sheath thickness and conduction latencygroups in the cat optic nerve. J. Comp. Neurol. 1978; 181:183–96.

8. Carras PL, Coleman PA, Miller RF. Site of action potential initiation in amphibian retinal ganglion cells. J. Neurophysiol.1992; 67: 292–304.

9. Bishop PO, Jeremy D, Lance JW. The optic nerve. Propertiesof central tract. J. Physiol. 1953; 121: 415–32.

10. Rowe MH, Stone J. Conduction velocity groupings amongaxons of cat retinal ganglion cells, and their relationship toretinal topography. Exp. Brain Res. 1976; 25: 339–57.

11. Burke W, Burne JA, Martin P. Selective block of Y optic nervefibres in the cat and the occurrence of inhibition in the lateralgeniculate nucleus. J. Physiol. 1985; 364: 81–92.

12. Ogden TE, Miller RF. Studies of the optic nerve of the rhesusmonkey: nerve fiber spectrum and physiological properties.Vision Res. 1966; 6: 485–506.

13. Boycott BB, Wässle H. The morphological types of ganglioncells of the domestic cat’s retina. J. Physiol. 1974; 240:397–419.

14. Hughes A. Population magnitudes and distribution of themajor modal classes of cat retinal ganglion cells as estimatedfrom HRP filling and a systematic survey of the soma diameterspectra for classical neurones. J. Comp. Neurol. 1981; 197:303–39.

15. Kolb H, Nelson R, Mariani A. Amacrine cells, bipolar cells andganglion cells of the cat retina: a Golgi study. Vision Res. 1981;21: 1081–114.

16. Fukuda Y, Stone J. Retinal distribution and central projectionsof Y-, X- and W-cells of the cat’s retina. J. Neurophysiol. 1974;37: 749–72.

17. Wässle H, Levick WR, Kirk DL, Cleland BG. Axonal conduc-tion velocity and perikaryal size. Exp. Neurol. 1975; 49:246–51.

18. Sakai H, Woody CD. Relationships between axonal diameter,soma size, and axonal conduction velocity of HRP-filled, pyra-midal tract cells of awake cats. Brain Res. 1988; 460: 1–7.

19. Fukuda Y, Hsiao CF, Watanabe M, Ito H. Morphological cor-relates of physiologically identified Y-, X-, and W-cells in catretina. J. Neurophysiol. 1984; 52: 999–1013.

20. Hsiao CF, Watanabe M, Fukuda Y. The relation between axondiameter and axonal conduction velocity of Y, X and W cells.Brain Res. 1984; 309: 357–61.

21. Dacey DM, Petersen MR. Dendritic field size and morphol-ogy of midget and parasol ganglion cells of the human retina.Proc. Natl Acad. Sci. USA 1992; 89: 9666–70.

22. Stone J, Johnstone E. The topography of primate retina: astudy of the human, bushbaby, and new- and old-worldmonkeys. J. Comp. Neurol. 1981; 196: 205–23.

23. Ghosh KK, Goodchild AK, Sefton AE, Martin PR.Morphology of retinal ganglion cells in a New World monkey:the marmoset Callithrix jacchus. J. Comp. Neurol. 1996; 366:76–92.

24. Watanabe M, Rodieck RW. Parasol and midget ganglion cellsof the primate retina. J. Comp. Neurol. 1989; 289: 434–54.

25. Rodieck RW, Binmoeller KF, Dineen J. Parasol and midget ganglion cells of the human retina. J. Comp. Neurol. 1985; 233:115–32.

26. Rodieck RW, Watanabe M. Survey of the morphology ofmacaque retinal ganglion cells that project to the pretectum,superior colliculus, and parvicellular laminae of the lateralgeniculate nucleus. J. Comp. Neurol. 1993; 338: 289–303.

27. Fukuda Y, Watanabe M, Wakauwa K, Sawai H, Morigiwa K.Intraretinal axons of ganglion cells in the Japanese monkey(Macaca fuscata): conduction velocity and diameter distribu-tion. Neurosci. Res. 1988; 6: 53–71.

28. Stanford LR. W-cells in the cat retina: correlated morphologi-cal and physiological evidence for two distinct classes. J.Neurophysiol. 1987; 57: 218–44.

Intraretinal axon diameter 429

29. Kirk DI, Cleland BG, Levick WR. Axonal conduction latenciesof cat retinal ganglion cells. J. Neurophysiol. 1975; 38: 1395–402.

30. Murakami D, Sesma MA, Rowe MH. Characteristics of nasaland temporal retina in Siamese and normally pigmented cats:ganglion cell composition, axon trajectory and laterality ofprojection. Brain Behav. Evol. 1982; 21: 67–113.

31. FitzGibbon T, Funke K, Eysel UT. Anatomical correlationsbetween soma size, axon diameter and intraretinal length forthe alpha ganglion cells of the cat retina. Vis. Neurosci. 1991; 6:159–74.

32. Curcio CA, Allen KA. Topography of ganglion cells in thehuman retina. J. Comp. Neurol. 1990; 300: 5–25.

33. Stone J, Leventhal A, Watson CR, Keens J, Clarke R.Gradients between nasal and temporal areas of the cat retinain the properties of retinal ganglion cells. J. Comp. Neurol. 1980;192: 219–33.

34. Baker GE, Stryker MP. Retinofugal fibres change conductionvelocity and diameter between the optic nerve and tract inferrets. Nature 1990; 344: 342–5.

35. Stanford LR. Conduction velocity variations minimize con-duction time differences among retinal ganglion cell axons.Science 1987; 238: 358–60.

36. Walsh N, FitzGibbon T, Ghosh KK. Intraretinal axon dia-meter: a single cell analysis in the marmoset (Callithrix jacchus).J. Neurocytol. 2000; (in press).

37. Ghosh KK, Martin PR, Grünert U. Morphological analysis ofthe blue cone pathway in the retina of a New World monkey,the marmoset Callithrix jacchus. J. Comp. Neurol. 1997; 379:211–25.

38. Wilder HD, Grünert U, Lee BB, Martin PR. Topography ofganglion cells and photoreceptors in the retina of a NewWorld Monkey: the marmoset Callithrix jacchus. Vis. Neurosci.1996; 13: 335–52.

39. FitzGibbon T. The human fetal retinal nerve fiber layer andoptic nerve head: A DiI and DiA tracing study. Vis. Neurosci.1997; 14: 433–47.

40. FitzGibbon T, Taylor SF. Retinotopy of the human retinalnerve fibre layer and optic nerve head. J. Comp. Neurol. 1996;375: 238–51.

41. Wässle H, Peichl L, Boycott BB. Morphology and topographyof on- and off-alpha cells in cat retina. Proc. R. Soc. Lond. B Biol.Sci. 1981; 212: 157–75.

42. Perry VH, Oehler R, Cowey A. Retinal ganglion cells thatproject to the dorsal lateral geniculate nucleus in the macaquemonkey. Neuroscience 1984; 12: 1101–23.

43. Yamada ES, Silveira LCL, Perry H. Morphology, dendriticfield size, density, and coverage of M and P retinal ganglioncells of dichromatic Cebus monkeys. Vis. Neurosci. 1996; 13:1011–29.

44. Quigley HA, Flower RW, Addicks EM, McLeod DS. Themechanism of optic nerve damage in experimental acuteintraocular pressure elevation. Invest. Ophthalmol. Vis. Sci. 1980;19: 505–17.

45. Morgan JE. Selective cell death in glaucoma: does it reallyoccur? Br. J. Ophthalmol. 1994; 78: 875–80.

46. Morgan JE, Uchida H, Caprioli J. Retinal ganglion cell deathin experimental glaucoma. Br. J. Ophthalmol. 2000; 84: 303–10.

47. Rushton WA. A theory of the effects of fibre size in medullatednerve. J. Physiol. 1951; 115: 101–22.

48. Straatsma BR, Foos RY, Heckenlively JR, Taylor GN.Myelinated retinal nerve fibers. Am. J. Ophthalmol. 1981; 91:25–38.

49. Goodchild AK, Ghosh KK, Martin PR. A comparison ofphotoreceptor spatial density and ganglion cell morphology inthe retina of human, macaque monkey, cat and the marmosetCallithrix jacchus. J. Comp. Neurol. 1996; 366: 55–75.

430 Walsh et al.