[ 219 ]

GLEES STAINING OF THE MONKEY HYPOTHALAMUS:A CRITICAL APPRAISAL OF NORMAL AND

EXPERIMENTAL MATERIAL

BY W. R. ADEY, ALICE F. RUDOLPH, I. F. HINEAND NANCY J. HARRITT*

Department of Anatomy, University of Melbourne

In a recent paper, Cowan & Powell (1956) have drawn attention to certain appear-ances in the brain of monkey and man when stained with the Glees silver technique.They have found in their material, which they believe to be normal, a 'terminaldegeneration' in septal and hypothalamic areas with a distribution in these regionsresembling that seen after lesions in the olfactory bulb (Le Gros Clark & Meyer, 1947;Meyer & Allison, 1949; Adey, 1953), the amygdaloid nuclei (Adey & Meyer, 1952 a),the frontal lobe (Meyer, 1950; Wall, Glees & Fulton, 1951) and the hippocampus(Simpson, 1952).As we had not encountered this appearance in our own normal material, it seemed

desirable to attempt an assessment of the incidence of such pseudodegeneration innormal preparations stained by the Glees method. A larger number of brains hasbeen examined here than in the series of one human and four monkey brains studiedby Cowan & Powell, particularly as their material included an animal from whichan eye had been recently removed, another which had died of an enteric infection,and a human brain fixed without perfusion. Their published figures do not indicatefrom which of the monkey brains the sections had been prepared.Apart from the question of artefacts, the relative merits of Glees and Nauta

methods of displaying degeneration in the central nervous system have been thesubject of conflicting opinions. Thus Evans & Hamlyn (1956) concluded that thefinest degenerating ramifications, shown by the Glees technique, were not stained inNauta preparations. They considered the Glees method satisfactory for studying theexact site of termination of axonal arborizations in appropriate regions, butthat it was at a disadvantage in the presence of scanty degeneration amongsta mass of normal fibres. By contrast, Blackstad (1956) has maintained that in hisstudies ofrhinencephalic systems, the Glees method often completely failed to displayterminal structures, although pre-terminal degenerating fibres were well showni.The disclosure by Cowan & Powell (1956) of a selectively distributed pseudo-

degeneration might well be a salutary warning of the difficulties which may arise inlstudies of degeneration by silver impregnation techniques, since, if it were found tooccur even occasionally in the hands of earlier workers with the Glees method inrhinencephalic and hypothalamic systems, the results must be seriously questioned.We have unsuccessfully endeavoured to display pseudodegeneration as described

* This study was assisted by grants from the National Health and Medical Research Council ofAustralia and from the Medical Research Committee of the University of Melbourne.

Anat. 9215

W. R. Adey and othersby Cowan & Powell in normal monkey brains, and have therefore re-examinedboth the distribution of hypothalamic degeneration following amygdaloid andfrontal cortical damage and also the effects of a dummy operation.

MATERIAL AND METHODS

A total of thirty-four monkey brains and one baboon brain have been used to datein this study. It is a pleasure to acknowledge the helpful co-operation of theCommonwealth Serum Laboratories in supplying the monkey material. All animalswere immature (weighing 7-10 lb.) and in good health at the time of killing, and hadbeen in the laboratory environment for a minimum period of 8 weeks and hadreceived a diet containing antibiotics.

All animals except one (Rh 8) were perfused with a solution of 10% commercialformalin (non-neutralized) in normal saline. The perfusion was performed throughan incision in the left ventricle which permitted insertion of the cannula into theascending aorta. A drainage incision was made in the right atrium, and thecirculation was first flushed with normal saline until the return through the atrialincision was pale pink. Approximately 400 ml. of formalin solution was theninjected under a pressure varying from 110 to 120 cm. of fluid in the course of theperfusion. After removal, all brains have been fixed in 10% aqueous formalinsolution for periods from 3 weeks to 9 months prior to histological examination.Although sixty-seven animals were perfused under ether anaesthesia immediately

following the aseptic removal of both kidneys for virological investigations, onlytwenty-one of these brains (ten Macacus cynomolgus and eleven M. rhesus) havebeen examined histologically. No attempt at selection has been made in the brainstaken from this series for histological examination. A further series of seven M. rhesusmonkeys were similarly perfused under deep Nembutal anaesthesia and their brainsexamined histologically.

In a further series of six monkeys (all Macacus rhesus) experimental ablations andcontrol procedures were performed. In two, amygdaloid lesions were made, and intwo others frontal lesions were inflicted. A dummy operation, involving removal ofa large fronto-parietal bone flap without opening the dura mater was performed ina fifth animal. The brain of the sixth animal was examined without any prior surgicalinterference, in order to test the remote possibility that the pseudodegenerationmight be correlated with tuberculous infection, since all six animals in this series hadbeen culled from the colony as positive tuberculin reactors. These animals were allperfused under Nembutal anaesthesia. An extensive post-mortem examinationperformed by the staff of the Commonwealth Serum Laboratories failed to disclosetuberculous lesions in any of these animals.One adult female baboon (Papio ursinus) was perfused under deep Nembutal

anaesthesia and the brain examined histologically. Thus, in summary, histologicalinvestigations have been carried out in twenty-eight normal monkey brains and inone normal baboon brain. One monkey brain from a positive tuberculin reactor hasbeen examined. Experimental procedures have been performed in five monkeys,and of these, one amygdaloid and one frontal ablation and a dummy operation haveso far been submitted to histological examination.

220

Glees staining of the monkey hypothalamusHistological procedures. While it is hoped eventually to investigate with the

material currently available such factors as varying impregnation and precipitation,which indubitably beset the Glees method, it was considered paramount to adhereessentially to the method used in previous studies of hypothalamic connexions offrontal, temporal and olfactory structures (Sprague & Meyer, 1950; Adey & Meyer,1952a, b; Adey, 1953; Adey, Merrillees & Sunderland, 1956).

All brains have been sectioned frozen for Glees and Nauta preparations, usinga dural freezing block, in which dry ice is evaporated in 70% alcohol. This methodproduces very even freezing and permits serial section of the whole diencephalon at15/u thickness. Facilities were available for the arrangement of 450 frozen sectionsin sequence.

All twenty-eight normal brains have been examined by the Glees and Nautamethods. The Nauta method used was that described by Nauta & Gygax (1954).For the Glees staining, separate batches of sections were stained by two of us(W. R.-A. and A. F. R.). A third series of Glees sections were prepared in seven brains(N.J. H.). Nauta staining was performed by one investigator (A. F. R.). In addition,Bodian (1936, 1937) impregnations were prepared by one investigator (I. F. H.) intwenty-one of the twenty-eight normal brains, including the seven animals perfusedunder Nembutal anaesthesia. The protargol used here was Protargol S (New Com-mission Certified-Winthrop Stearns). Except for the variations mentioned below,the Glees method followed that described by Adey & Meyer (1952b). In all, 397Nauta sections and 677 Glees sections have been prepared in normal and controlmaterial, and an additional 125 Nauta and 70 Glees sections have been stained inexperimental material, a total of 1269 sections.

Modifications tested in the Glees techniqueOnly such minor variations as might contribute to the appearance of pseudo-

degeneration have been examined here.(i) Addition of buffering substances toformalin/tap water mixtures. The local water

supply is of a high order of softness, with low inherent buffering capacity. Thus inthe preparation of formalin/tap water mixtures, the effects of adding citric acid orammonium hydroxide were tested. Commercial formalin neutralized with excesscalcium carbonate was used in these solutions. One to six drops of 1% citric acidwere added to each 10 ml. of formalin/tap water mixture. In the two brains inwhich citric acid was added to the formalin solution (Rh3, Cy5) background stainingwas perceptibly paler and the intensity of impregnation of individual fibres pro-portionately reduced. Due to the reduced intensity of impregnation, fibres appearedbrown in many areas, rather than the usual brownish black.Ammonium hydroxide was used much more widely in the formalin/tap water

mixtures, and in almost every batch of sections one to six drops of ammonia solution(analytical, reagent) were added to each 10 ml. of the formalin/tap water mixture,sometimes in the passage of the early sections to the ammoniated silver bath and par-ticularly in the subsequent final reduction. This variation undoubtedly helped inreducing background staining without perceptible modification of either intensityof impregnation or morphology of the fibre elements in the hypothalamus orelsewhere.

15-2

221

W. R. Adey and others(ii) Variations in methods offixation. Since variations in osmotic interrelationships

and relatively slow fixation might account for the beaded appearances seen inpseudodegeneration, particular note was taken of the adequacy of fixation at thetime of removal of the brain. Persistence of pinkish areas and the presence of bloodin the surface vessels was regarded as evidence of defective fixation. The assess-ment of fixation is presented in tabular form (Table 1).

Quality ofperfusionPoorGoodPoorGoodGoodGoodGoodUnperfusedGoodGoodExcellentGoodGoodGoodPoorFairPoorPoorModerateGoodGoodGoodPoorModerateGoodGoodFairFair

Table 1Quality offixation

PoorGoodMediumGoodGoodGoodExcellentFairExcellentGoodExcellentExcellentGoodGoodInadequateFairInadequateInadequateModerateGoodGoodExcellentPoorModerateExcellentFairPoorPoor

Silver impregnation methodsand staining quality

G+ + N+ + B+ +G+ + N+ + B+ +G+ + + N+ + B+ ++G+ N+ + B+ +G poor N+ B poorG+ + N+ + B+ +G poor N+ B+G++ N++ B+ +G++ N++G++ N++G+++ N++G+++ N++ B+++G++ N++ B+++G++ N++ B+++G+++ N++ B++G++ N++ B++G++ N+ B+++G+ N+ B++G++ N++ B++G+ N++ B++G+++ N++ B++G poor N++ B+ +G++ N++ B++G variable N++ B++G variable N++ B++G++ N++G+ N++G+ N++

An arbitrary assessment of general staining quality from poor to three-plus has been used inthese normal brains in relation to each staining method tested, Glees (G), Nauta (N) and Bodian (B).

To estimate the effects of agonal and autolytic changes, one brain (Rh 8) was leftunperfused. It was removed 2hr. after death and placed in 10% aqueous formalinsolution for 11 days. Subsequent Glees staining of the diencephalon producedadequate impregnation of hypothalamic fibres, although the degree of impregnationvaried more from place to place in a given section than in the perfused material.There was some precipitation along the course of individual fibres, apparently ofadherent metallic particles, but no appearances likely to be construed as pseudo-degeneration could be detected.The effects of storage of cut frozen sections have been examined in one brain

(Rh 7). Separate batches of sections were kept for 3 days after cutting in 10%formalin solution and in 10% formol saline, prior to transfer to ammoniacal alcoholin the Glees procedure. Impregnation after storage in formalin solution was poor,

with heavy precipitation. No beading of fibres was seen. Better impregnationfollowed storage in formol saline solution, but gross beading of fibres occurred

222

BrainRh1Rh2Rh3Rh4Rh5Rh6Rh7Rh8Rh2ORh2lRhl1Rh22Rh23Rh24Rh25Rh26Rh27Rh28Cy 1Cy2Cy3Cy4Cy5Cy6Cy7CyllCy 16Cy 17

AnaestheticEtherEtherEtherEtherEtherEtherEtherEtherEtherEtherEtherNembutalNembutalNembutalNembutalNembutalNembutalNembutalEtherEtherEtherEtherEtherEtherEtherEtherEtherEther

Glees staining of the monkey hypothalamus 223diffusely in many areas of the hypothalamus and elsewhere in the diencephalon.This beading involved large fibres without interruption of continuity and showedno pattern of distribution. The appearance of the fibres, though abnormal, and thediffuse nature of the change, can scarcely lead to confusion with the picture of truedegeneration, particularly as no beading could be seen amongst the terminal plexusesin the hypothalamus or thalamus.A striking difference was seen in the intensity of impregnation of the fine peri-

cellular plexuses in the hypothalamus in the Protargol-stained material whencomparisons were made between animals killed under ether and Nembutal. A muchmore intense impregnation uniformly followed perfusion under Nembutal than withether. These differences could not be discerned in similar comparisons of Glees andNauta preparations from the two groups of brains. However the results with theProtargol method confirm the widely held opinion that the use of volatile anaes-thetics may exert a deleterious effect on subsequent silver impregnation of neuraltissue.

(iii) Variations in the technique of Glees silver impregnation. Since the cut sectionsare washed briefly in water as the initial step in the Glees method, prior to beingtransferred to ammoniacal 50% alcohol for 48 hr., we have tested the effect of aperiod of 24 hr. immersion in water on the sections from one brain (Rh 27). Frozensections were cut on the 35th day after perfusion, which was imperfect in posteriorthalamic areas. Subsequent Glees staining was good in almost all sections, with littleprecipitation and no significant beading in any part of the diencephalon.A similar procedure was tested in sections from one brain (Rh 26) by extending

the period in water to 24 hr. after removal from ammoniacal alcohol and prior toimpregnation with 10% silver nitrate. Good staining followed in these sections. Incomparison with sections from the same brain not so treated in water, but stainedat the same time, the larger fibres in the lateral hypothalamic and thalamic areasshowed a high proportion of 'knobs', darkly stained swellings along the course ofthe fibres which did not appear to be associated with modified structure or stainingof the intervening portions of the axon. This appearance was found diffusely throughthe section. No beading or disruption of finer axonal structures in pericellularplexuses could be detected.Prolonged exposure to 10% silver nitrate solution does not appear to exert a

significant effect on the appearance of the stained fibres. Sections from the brain C 2,which had been subjected to a dummy operation videe infra) were placed in 10%silver nitrate for 63 days and stored in the dark. By this time they had taken ona dark chocolate-brown colour. They were bleached in 1% sodium thiosulphatesolution for 15-30 sec. and washed in five changes of water. They were then returnedto 10% silver nitrate. Batches of sections were stained in the normal way 2j, 5jand 26 hr. later. Particularly satisfactory impregnation occurred in the bed nucleusof the stria terminalis, the hypothalamus, septum and preoptic area without evidenceof pseudodegeneration.We have also examined in two brains (Cy 6 and Cy 7) the effects of an additional

water wash between the conclusion of the first formalin reduction and immersion ofthe section in the ammoniacal silver bath. It appeared that this procedure wasfollowed by unpredictable variability in impregnation and an increase in both

224 W. R. Adey and othersprecipitation and background staining. No appearance suggestive of pseudodegenera-tion was seen in these sections.

It has been found that weaker formalin/tap water mixtures than the customary10% concentration can be used with advantage in displaying the fine fibre elementsin medial hypothalamic areas. The 'development' of the final picture occurs moreslowly, but usually with increased ultimate contrast, in that extremely delicateramifications are well defined against a pale background. Suitable mixtures mayhave a formalin concentration between 0-5 and 3 0 %.

RESULTS

A. Survey offindings in normal materialIt would not seem rewarding to enter into a detailed description of fibre appearancesby the Glees method in the normal diencephalon on the basis of regional differencesin the morphology of fibre terminals or the arrangement of fibre bundles and peri-cellular plexuses. Indeed, the most careful assessment of our material does not suggestthat the method would provide a fruitful basis for such a qualitative differentiation.We may, however, draw attention to certain variations in staining noted in this

study. It appears that certain difficulties attach to the adequate and consistentimpregnation of the region of the median eminence and adjoining medial areas ofthe caudal hypothalamus. Sections which were well stained in other regions fre-quently remained poorly impregnated in these ventral areas, and proper staininghere appeared critically dependent on such factors as the nature and amount of thebuffering substances in the formalin/tap water mixtures, as outlined above.

In such poorly impregnated sections, the fibres took on a pale and ghostlyappearance, and in such cases the range of contrasting densities between glial andneural fibre elements was much reduced, since the neural fibres displayed little ofthe argentophilia characteristic of Glees preparations. Such a confusion has also beennoted in the preparations of the glial and neural elements of the neurohypophysisof the cat and rat (Green & van Breemen, 1955). In such sections, too, tiny scatteredgranular bodies were sometimes discerned at the limits of resolution of the lightmicroscope. Although occasionally arranged in linear aggregates, they could scarcelybe construed as representing the terminals of neural elements by the criteria currentlyemployed, having regard to their small size, their lack of argentophilia and theirdiffuse distribution. These granules showed considerable resemblance to the surfaceprecipitation of impregnating material which commonly accompanied this appear-ance. They showed no similarity to the larger, intensely argentophilic fragmentedterminals seen previously in the dorso-medial and ventro-medial hypothalamicnuclei following amygdaloid lesions (Adey & Meyer, 1952 a).More cogent to any discussion of pseudodegeneration in Glees preparations is the

occasional appearance of scattered fragmented fibres bearing all the characters oftrue degenerating fibres. These have been described with a variety of silver stainingtechniques. In our material they appeared more numerous in Bodian preparationsthan in either Glees or Nauta sections. A very careful search of a Bodian preparationof the whole diencephalon might reveal six to ten such fragmented fibres occurringin a completely random fashion throughout the section. In Glees material from the

Glees staining of the monkey hypothalamuussame brain, a search of several sections was usually necessary to reveal a single suchfibre. Their incidence in the hypothalamus was no higher than in other regions of thediencephalon. It may be suggested that the process of paraffin embedding atrelatively high temperatures may account for the higher incidence of this appearancein Bodian material.

Particular attention has been directed to the appearance of fibres along themargins of sections. Prior to the present study we had observed from time to timetortuous fibres in Glees sections of normal tissues adjacent to the free surfaces of thesection. The fibres so involved might be sharply torn across and separated from anadjoining segment of fibre by a short interval. They might also exhibit beadedswellings on an otherwise normally staining axon. This complicating factor in theinterpretation of degeneration in such bundles as the stria medullaris has beenmentioned elsewhere (Adey et al. 1956). It is our impression that this appearanceoccurs rarely in the freezing method used in the present study, involving the slowinduction of a steady state of freezing with dry ice, as compared with earlier methodsdepending on production of a transient and highly variable frozen state with gaseouscarbon dioxide. The violent rupture of marginal fibres, suggested by the microscopicappearance in such cases, supports the hypothesis that it may result from the rapidfreezing and melting which is an inevitable accompaniment of the gas-freezingmethod. We have not seen any appearance in the present normal Glees material ofperiventricular structures suggesting pseudodegeneration which might conceivablyhave resulted from such an insult to the tissue.We have encountered no difficulty in securing adequate and consistent impreg-

nation of the septum, the bed nucleus of the stria terminalis, the supraoptic andpreoptic areas, and have not seen any evidence of pseudodegeneration amongsteither the larger fibres or fine pericellular plexuses in Glees preparations of theseregions.

B. Appearances of degeneration in experimental materialIn view of an inability to mimic the appearances of significant degeneration in

normal material by the Glees method, we have turned to an examination of thediencephalon in a series of operated brains. The findings following an amygdaloidlesion (Am 1) and a frontal ablation (Fr 1) will be presented here, since they contrastsharply with the negative findings in the large amount of normal material examined.Two other brains, with amygdaloid and frontal lesions respectively, were preparedcontemporaneously with Am 1 and Fr 1, but will be examined later as part of acontinuing study.

(i) Control brain (C 1). Since all animals in this group were positive tuberculinreactors, this subject was used as a control to test the remote possibility thatpseudodegeneration might be connected with the tuberculous infection. This animalwas caged with Am 1 andFr 1 preoperatively and during their post-operative survivalperiod. Histologically, its brain showed no significant abnormality with the Gleesand Nauta methods.

(ii) Effects of dummy operation (C 2). In this animal a large fronto-parietal boneflap was removed under Nembutal anaesthesia. The dura mater was not opened.Five days post-operatively perfusion was performed under Nembutal anaesthesia.

225

W. R. Adey and othersNo evidence of degeneration was found in the hypothalamus, septum or preoptic area.The effects of very long exposure to silver nitrate on sections from this brain havebeen described above.

(iii) Distribution of degeneration following an amygdaloid lesion. Since previousstudies using both Marchi (Fox, 1940, 1943) and Glees (Adey & Meyer, 1952 a)methods have indicated extensive amygdalo-hypothalamic pathways running inpart at least through the stria terminalis, we have examined the diencephalon in thebrain of an animal (Am 1) killed 6 days after an amygdaloid lesion.The extent of the amygdaloid involvement (Text-fig. 1 A-C) resembles that of

another brain (M 85) described previously (Adey & Meyer, 1952 a). Anteriorly, thelesion has destroyed the cortex of the inferior part of the tip of the temporal lobeand on the lateral aspect of the hemisphere extends for a short distance into thelower part of the superior temporal gyrus and the middle temporal gyrus. On themedial side of the temporal lobe the primary olfactory cortex in the temporalprepyriform area has been extensively resected. Further posteriorly, at the level ofthe optic chiasm, the lesion has removed the anterior portion of the entorhinalcortex. In the overlying amygdala, the medial and most of the lateral parts of thebasal nucleus, the cortico-amygdaloid transition area and a portion of the lateralnucleus have been removed. Only the medial nucleus, parts of the lateral nucleus,and the dorsal divisions of the basal nucleus appear to have been spared at this level.Caudally the lesion is sharply circumscribed within the anterior entorhinal cortexand overlying amygdaloid nuclei, including the entire basal complex, but sparingthe central nucleus, the medial nucleus and a narrow strip of the lateral nucleus.The anterior end of the hippocampus is not involved.Not only has the examination of this brain confirmed the general distribution of

hypothalamic degeneration described in M 85 by Adey & Meyer following an amyg-daloid lesion, but new data has also been gleaned concerning certain additionalpathways from the temporal lobe to the hypothalamus. Examination of both Gleesand Nauta sections at each level has served to emphasize the merits of each methodin different situations.Nauta sections at the level of the anterior commissure showed considerable

numbers of degenerating fibres traversing the commissure (Text-fig. 1 E). Thesefibres were widely dispersed through all dorso-ventral levels of the commissure(PI. 1, fig. 3). They could be traced across the mid-line to the contralateral limits ofthe commissure, and a few appeared to leave it dorsally and enter the lateral partof the base of the septum on both sides of the mid-line. In Glees sections at the samelevel, the degenerating fibres tended to be obscured by the enormous numbers ofheavily stained normal fibres.

Coronal sections of the septum in a plane slightly anterior to the median portionof the anterior commissure (Text-fig. 1 D) showed degenerating fibres passingdownwards through the septum in a paramedian plane bilaterally (PI. 1, figs. 1, 2).They appeared first near the upper border of the septum and extended inferiorly tothe level of the ventral border of the anterior commissure. It was our impressionthat these fibres were more numerous on the operated side. Scattered degeneratingterminals were also detected along the course of these fibres which were morestrikingly displayed in Nauta than Glees preparations.

226

Glees staining of the monkey hypothalamus 227Further posteriorly, coronal sections through the middle of the anterior commis-

sure again showed degenerating fibres coursing ventrally through the septum onboth sides of the mid-line (Text-fig. 1 E). These fibres were more profuse than atthe level of Text-fig. 1 D, and they appeared to arise from the fornix bundles visiblein the upper part of the section. Many degenerating terminals are visible in the

Rhesus Am IL. amygdaloid lesion

A(

R I

B

R

C

R

D

L

E

R

L

- -VM

Text-fig. 1. Extent of temporal lesion (A, B, C) is indicated in solid black. Secondary degeneration asseen in Nauta and Glees preparations is seen in D-G. Degenerating fibres are indicated byarrows and degeneration in terminal ramifications by dots. Abbreviations: A.C., anteriorcommissure; C.N., caudate nucleus; D.M., dorso-medial hypothalamic nucleus; F., fornix;G.P., globus pallidus; I.C., internal capsule; N.ACC., nucleus accumbens; OPT.CH., opticchasm; SEPT., septum; STRIA TERM. (HYPOTHAL.), hypothalamic bundle of striaterminalis; STRIA TERM. (PREOPT.), preoptic bundle of stria terminalis; STRIA TERM.(SUPRACOMM.), supracommissural bundle of the stria terminalis; V.M., ventro-medialhypothalamic nucleus; 1, cortico-amygdaloid transition area; 2, cortical amygdaloid nucleus;3, medial amygdaloid nucleus; 4, accessory basal amygdaloid nucleus; 5, medial part of basalamygdaloid nucleus; 6, lateral part of basal amygdaloid nucleus; 7, lateral amygdaloidnucleus; 8, central amygdaloid nucleus; OLF. TUB., olfactory tubercle.

228 W. R. Adey and othersmedial septal nucleus dorsal to the anterior commissure. Further laterally at thislevel, scattered degenerating fibres are present in the supracommissural bundle ofthe stria terminalis on the operated side, but no degeneration could be detectedin the commissural bundle of the stria nor could any degeneration be detected inthe preoptic region ventral to the anterior commissure. The stria terminalis bundlescontralateral to the lesion were normal at this level.At the posterior border of the anterior commissure (Text-fig. 1 F) the hypothala-

mic bundle of the stria terminalis clearly displayed numerous degenerating fibres.These could be traced dorsally in the section along the medial border of the internalcapsule to the ventral part of the head of the caudate nucleus, following the courseof the bed nucleus of the stria terminalis (P1. 1, figs. 7-9). At this level the fornixcolumns lie close to the upper border of the anterior commissure. There werescattered degenerating fibres in the fornix bilaterally (P1. 1, figs. 4, 5). In sectionsslightly further forward, these fibres ran in considerable numbers at the medialborder of the fornix and appeared to terminate in the medial septal nucleus. Ventralto the anterior commissure at the level of Text-fig. 1 F, the degeneration rapidlydecreased, with only scattered fibres and terminals degenerating in the paraven-tricular nucleus. No degeneration was seen in the suprachiasmatic or supraoptic nuclei.

Further caudally (Text-fig. 1 G), examination of Nauta preparations revealeda wide distribution of degenerating fibres. On the operated side fine fascicles ofdegenerating fibres were found to sweep medially into the thalamus ventral to thecaudate nucleus. Within the thalamus they were found mainly in the reticularnucleus adjoining the internal capsule, in the mid-line nuclei and in the intralaminarnuclei (P1. 1, figs. 10, 11). A similar but less extensive distribution of degeneratingfibres occurred in the contralateral thalamus. The involvement of the mid-line nucleiwas heavy bilaterally. No degeneration was found in the caudate nuclei (apart fromthe fascicles at its ventral border) nor in the internal capsules. Ventrally in thehypothalamus degenerating fibres and terminals occurred bilaterally in the dorso-medial and ventro-medial hypothalamic nuclei (P1. 1, fig. 6), but gradually decreasedin amount at the more ventral zones. The lateral hypothalamus showed no degenera-tion on either side, nor were degenerating fibres found in the fornix on either sideat this level. In the globus pallidus and lentiform nucleus on the operated sideobliquely directed fascicles of fine degenerating fibres and degenerating terminalswere seen. Less numerous degenerating fibres and terminals were seen in the contra-lateral globus pallidus (P1. 2, figs. 12, 13).While the Nauta preparations allowed an excellent appraisal of the general

distribution of the degeneration in a large volume of tissue, it was our impressionthat the method might not impregnate the finest elements of a degenerating peri-cellular plexus. This impression was supported by our observation that degenerationin the dorso-medial and ventro-medial hypothalamic nuclei in this brain was bestseen in its finest pericellular ramifications in Glees preparations (P1. 1, fig. 6),although the search for these appearances amongst many normal fibres requiredconsiderably greater care and patience.

(iv) Hypothalamic appearances following a frontal lobe lesion. A right frontalcortical resection was performed under Nembutal anaesthesia in an immaturemacaque (Fr 1). The animal was perfused 6 days later under Nembutal anaesthesia.

Glees staining of the monkey hypothalamusThe lesion (Text-fig. 2) mainly involved cortex on the medial side of the hemispherein areas 8B and 9, extending inferiorly slightly into the anterior end of area 24. Thelesion did not involve the frontal pole anteriorly, nor did it appear to extend sig-nificantly into area 6 posteriorly.

8-8 46

24

Text-fig. 2. Extent of frontal lesion indicated in solid black. Distribution of secondary degenerationin the hypothalamus is indicated by arrows (fibres) and dots (terminals). Abbreviations:C.N., caudate nucleus; D.M., dorso-medial hypothalamic nucleus; F., fornix; G.P., globuspallidus; V.M., ventro-medial hypothalamic nucleus.

We have confined our histological examination of this brain to Glees and Nautasections at the level of the dorso-medial hypothalamic nuclei. Glees preparationsshowed degenerating fibres entering the dorso-medial nucleus from its dorsal aspecton the operated side (P1. 2, fig. 14). There were numerous fragmented and swollenterminals in more ventral parts of the ipsilateral dorso-medial nucleus (P1. 2, fig. 15).Similar fragmented terminals were present in the contralateral dorso-medial nucleuswith a similar density of distribution (P1. 2, fig. 16), but no degenerating fibres couldbe seen entering the dorsal aspect of this nucleus. In the ventro-medial nuclei manymore fragmented terminals were seen on the operated than the unoperated side.Careful examination of Nauta preparations at the same level has disclosed very littlefibre or terminal degeneration on either side of the hypothalamus.

229

W. R. Adey and others

DISCUSSION

While it can be unequivocally stated that this study of the normal monkey dien-cephalon has failed to disclose a pseudodegeneration of the type described by Cowan& Powell (1956) in preparations by either the Glees or the Nauta methods, it isdisappointing that the variations in the Glees technique tried here have added littleto our knowledge of how this appearance might arise. It may be conceded that suchunusual procedures as absence or inadequacy of perfusion, lengthy maceration of thetissues in water at various stages of the staining, or marginal damage during thefreezing process may alter the intensity of impregnation and even occasionally thefibre morphology in an unpredictable fashion, but it has not been our experiencethat these changes mimic in any way the appearances of true degeneration. Nor dothe changes in impregnation in Bodian sections seen here after killing under etheranaesthesia produce a picture of pseudodegeneration.

Nevertheless, there is a considerable body of work published prior to the obser-vations of Cowan & Powell (1956) supporting the presence of beaded and fragmentedfibres in the infundibular region as an apparently normal phenomenon. Hagen (1952)drew attention to the factors of age and disease in initiating morphological changesin the fibre and cellular elements of the diencephalon. In an extensive investigationof the human diencephalon and hypophysis obtained from accident victims and fixednot more than 12 hr. after death, Hagen noted in Bielschowsky-Gros and Bodianpreparations neurofibrillar swellings along the course of the finest axis cylinders inthe hypophyseal stalk and neurohypophysis, and further, that these knotted thicken-ings appeared to lose continuity with nerve fibre elements, their linear arrangementthen indicating the original course of the nerve fibre. Hagen drew attention to thepossible confusion of these elements with degenerating nerve fibres, and consideredthat they represented a normal physiological process.

Subsequently Knocke (1953) examined Bielschowsky-Gros preparations of thehypophysis and diencephalon, mainly in young dogs whose brains were perfusedimmediately after death with 10 or 20% formalin solution. Knocke found nodularfibres in the tuber cinereum, the supraoptic and paraventricular nuclei, and in theneurohypophysis. At times the nodular fibres appeared as discontinuous finegranules, and in the peripheral parts of the infundibulum these deeply staining fibreswere condensed into tracts arborising in the infundibular region and around the bloodvessels of the neurohypophysis.While fixation of the brain was not initiated in either of the foregoing investiga-

tions until after death, our results in one unperfused brain do not suggest that thisis necessarily a cause ofpseudodegeneration. However, neither of those investigationsdisclose the widespread distribution of pseudodegeneration reported by Cowan &Powell, involving the lateral septal nucleus, nucleus accumbens, bed nucleus of thestria terminalis, anterior hypothalamic areas and the region of the tuberal nuclei. Amore localised distribution of this appearance in the vicinity of the tuber cinereumand adjacent supraoptic and paraventricular nuclei is suggested by both Hagen andKnocke.

It is a matter of some interest that Cowan & Powell should include as a controlexhibiting typical pseudodegeneration the brain of an animal with an eye recently

230

Glees staining of the monkey hypothalamusremoved, since previous work from the same laboratory has shown that in the ferret,at least, the normal gonadal response to visual stimulation may involve opticpathways to the subthalamus or the ventral nucleus of the lateral geniculate body(Le Gros Clark, McKeown & Zuckerman, 1938). Gonadal activation invariablyfollowed retinal stimulation even after essentially complete bilateral section of theoptic tracts at the ventral border of the dorsal nucleus of the lateral geniculate body.Other work from the same laboratory has shown that in the ferret fibres from theoptic chiasm can be traced into the hypothalamus in close relation to the wall of thethird ventricle (Jefferson, 1940), although it is claimed that these are aberrant fibreswhich eventually rejoin the optic tract. However, the Marchi technique used in thatstudy would not disclose the presence of unmyelinated collaterals terminating in thehypothalamus.A neurosecretory function has been ascribed to these beaded hypothalamic fibres

(Hagen, 1952; Mazzi, 1954; Cowan & Powell, 1956). However, the wide distributionof the beading seen by Cowan & Powell in Glees preparations extends appreciablybeyond the regions to which neuroendocrine functions are normally ascribed, andincludes the septum and the bed nucleus of the stria terminalis. A more cautiousattitude is taken by Green & van Breemen (1955) who noted in Gomori preparationsof the rat a gradual appearance of Gomori substance in the cells of the neural lobein the first two weeks after birth. However, the hypophyseal stalk and medianeminence remained clear of Gomori material in all these preparations, a surprisingfinding if it is assumed that the Gomori material is being produced in hypothalamicnuclei and is passing through nerve fibres of the neural stalk to be stored in theneural lobe. Moreover, Green & van Breemen were unable to see an accumulationof Gomori material in the tractus hypophysius during recovery from dehydration orfrom saline administration, as might be expected on the basis of a neural secretionhypothesis.

Furthermore, Green & van Breemen concluded that, if the granule masses seen inthe neural lobe by electron microscopy correspond to the Gomori substance, and ifthe nerve fibres resemble those seen in silver impregnations, then the granule massesare so large that they cannot possibly lie within nerve fibres. Yet the granule massesare enclosed in membranes suggesting that they lie in distorted fibres or separatedparts of the parent cell. These membranes may arise from nerve fibres, pituicytes ormodified ependymal cells, but at this stage their glial or neural origin cannot besettled. It is interesting that Green & van Breemen found granules in the medianeminence in an intra-axonal position but these granules were only one-third thediameter of those in the neural lobe and in the Herring bodies.Turning briefly to our findings in experimental material, which contrast so sharply

with our negative results in the substantial volume of normal and control material,the brain Am 1 with an amygdaloid lesion provides interesting evidence of directseptal connexions. Such a pathway had been suggested from electrophysiologicalstudies (Gloor, 1955), which had revealed that the shortest latency responses toamygdaloid stimulation in the cat are recorded in the base of the septum, the nucleusaccumbens, the preoptic area and the anterior hypothalamus. It is of particularinterest that Gloor's studies disclose projections to the bed nucleus of the striaterminalis and to the ventro-medial hypothalamic nucleus from the amygdala, thus

231

232 W. R. Adey and otherssupporting both the present histologic results and those described elsewhere (Adey& Meyer, 1952 a).Fox (1943) considered that fibres of the precommissural fornix of the cat termi-

nating in the lateral septal nucleus arise in the anterior end of the hippocampusadjacent to the amygdala. Since the lesion in Am 1 clearly spares the anterior endof the hippocampus, it would appear that these septal projections through the fornixmay in fact arise in the amygdala in the monkey.The lesion in our brain had also involved the tip of the temporal pole and adjoining

cortex on the lateral surface of the temporal lobe. Electrophysiological evidence hasindicated additional reciprocal projections between the temporal pole and the septalregion (Stoll, Ajmone-Marsan & Jasper, 1951). Stimulation of the septal nucleiproduced bilateral responses in the temporal poles, with a latency only slightlylonger in the contralateral than the ipsilateral temporal tip. Similar bilateralresponses in the temporal poles followed stimulation of the nucleus lateralis posteriorof the thalamus. Stoll et al. present evidence that the connexions with the septumrun in part through the fornix system. These findings are supported by the bilateraldegeneration we have found so clearly displayed in Nauta sections, and affords someexplanation for the difficulty experienced by Cowan & Powell (1956) in acceptingsuch bilaterality as evidence of valid degeneration in the septum and hypothalamus.McLardy (1955 a, b) has described bilateral degenerative changes in the lateral septalnucleus of the macaque following both fornix and temporal pole lesions and hasproduced evidence of a partial interfornix decussation in at least the septopetalcomponent of this two-way temporoseptal system.

It is obvious that this study has not extended into many important facets of theproblem of silver impregnation of neural tissue. It has been our prime purpose totest the extent to which pseudodegeneration might be seen in the indubitably normalprimate diencephalon when stained by us with the Glees method, having regard tothe previous use of the method by one of us in studies of the hypothalamic con-nexions of the temporal lobe (Adey & Meyer, 1952 a). The more common difficultieswhich beset the use of silver impregnation techniques in the display of degeneratingneural elements have been extensively discussed by Glees & Nauta (1955) and needno elaboration here. Our failure to find pseudodegeneration with the Glees method,even in such a relatively large volume of normal material, and despite a rigoroussearch of those hypothalamic areas to which unusual staining reactions and morpho-logical appearances have been attributed, would seem in itself a relatively fruitlessand negative conclusion to such a study. We may perhaps take heart that inexperimental material our results following an anterior temporal lesion confirm inboth Glees and Nauta preparations the general arrangement of hypothalamicdegeneration previously reported (Adey & Meyer, 1952 a). Moreover, the directcomparison possible here of Glees and Nauta preparations from the same region hasled to the conclusion that each may have its place in the scheme of investigation ofneural degeneration. As already suggested by Evans & Hamlyn (1956), the Gleesmethod would seem to achieve a more extensive impregnation of the finest terminalramifications than can be secured with the Nauta technique. On the other hand,the clarity of Nauta preparations allows immediate and easy assessment oflimited degeneration in such bundles as the anterior commissure and fornix, where

Glees staining of the monkey hypothalamusheavily stained normal fibres in Glees sections readily obscure degeneratingelements.

SUMMARY

1. The appearance of the diencephalon has been examined with a variety of silverstaining methods in thirty-four monkey brains and one baboon brain. All animalswere in good health at the time of killing. Glees, Nauta and Bodian techniques havebeen used in this study. A total of 1269 sections stained by the Glees and Nautamethods have been examined.

2. In unoperated and control material, comprising 397 Nauta sections and 677Glees sections, we have been unable to discern pseudodegeneration of the typedescribed by Cowan & Powell (1956). A particular search for this appearance hasbeen made in the hypothalamus, bed nucleus of the stria terminalis and septalnuclei. The control material included an animal subjected to a dummy operationinvolving removal of a large fronto-parietal bone flap but leaving the dura materintact.

3. Various modifications of the Glees staining procedure have been tried inunsuccessful endeavours to produce an appearance of pseudodegeneration. Etheranaesthesia prior to perfusion impaired impregnation in Bodian preparations of thehypothalamus in comparison with similar preparations from brains perfused underNembutal anaesthesia.

4. In a brain with a unilateral amygdaloid and temporopolar lesion, we haveconfirmed the findings of Adey & Meyer (1952 a) of bilateral terminal degenerationin the dorso-medial and ventro-medial hypothalamic nuclei, in both Glees and Nautapreparations. We have confirmed the role of the stria terminalis as a pathway to thehypothalamus on the side of the lesion. Nauta preparations in this brain disclosebilateral degeneration in septal nuclei, with bilateral degeneration in fibres of thefornix in its course adjoining the septum. Extensive bilateral terminal degenerationwas seen in the dorso-medial hypothalamic nuclei following a unilateral frontal lesion.It would appear that Nauta preparations permit ready assessment of scattereddegenerating fibres in such dense bundles as the fornix and anterior commissure,whereas Glees preparations display more suitably degenerating terminal ramifica-tions in such regions as the dorso-medial and ventro-medial hypothalamic nuclei.

It is difficult to acknowledge adequately the opportunity afforded one of us(W. R.A.) under a Royal Society and Nuffield Foundation Commonwealth Bursaryto discuss matters relevant to this study with other workers in this field in GreatBritain.

It is a pleasure to acknowledge the assistance of Dr P. L. Bazeley, Director of theCommonwealth Serum Laboratories and members of his staff in the provision of themonkey material, and without whose assistance this study would not have beenpossible.We wish to express our grateful appreciation to Prof. S. Sunderland for his

continuing help and encouragement.

233

234 W. R. Adey and others

REFERENCESADEY, W. R. (1953). An experimental study of the central olfactory connexions in a Marsupial

(Triclosurus vulpecula). Brain, 76, 311-336.ADEY, W. R., MERRILLEES, N. C. R. & SUNDERLAND, S. (1956). The entorhinal area; behavioural,

evoked potential and histological studies of its interrelationships with brainstem regions.Brain, 79, 414-439.

ADEY, W. R. & MEYER, M. (1952a). Hippocampal and hypothalamic connexions of the temporallobe in the monkey. Brain, 75, 368-384.

ADEY, W. R. & MEYER, M. (1952b). An experimental study of hippocampal afferent pathwaysfrom prefrontal and cingulate areas in the monkey. J. Anat., Lond., 86, 58-74.

BLACKSTAD, T. W. (1956). Commissural connexions of the hippocampal region in the rat, withspecial reference to their mode of termination. J. comp. Neurol. 105, 417-538.

BODIAN, D. (1936). A new method for staining nerve fibres and nerve endings in mounted paraffinsections. Anat. Rec. 65, 89-97.

BODIAN, D. (1937). The staining of paraffin sections of nervous tissue with activated protargol.The role of fixatives. Anat. Rec. 69, 153-162.

COWAN, W. M. & POWELL, T. P. S. (1956). A note on terminal degeneration in the hypothalamus.J. Anat., Lond., 90, 188-192.

EvANs, D. H. L. & HAMLYN, L. H. (1956). A study of the silver degeneration methods in thecentral nervous system. J. Anat., Lond., 90, 193-203.

Fox, C. A. (1940). Certain basal telencephalic centers in the cat. J. comp. Neurol. 72, 1-65.Fox, C. A. (1943). The stria terminalis, longitudinal association bundle and precommissural fornix

fibres in the cat. J. comp. Neurol. 79, 277-295.GLEES, P. & NAUTA, W. J. H. (1955). A critical review of studies on axonal and terminal degenera-

tion. Mschr. Psychol. Neurol. 129, 74-91.GLOOR, P. (1955). Electrophysiological studies on the connections of the amygdaloid nucleus in the

cat. I. The neuronal organization of the amygdaloid projection system. EEG clin. Neuro-physiol. 7, 223-242.

GREEN, J. D. & vAN BREEMEN, V. L. (1955). Electron microscopy of the pituitary and observationson neurosecretion. Amer. J. Anat. 97, 177-227.

HAGEN, E. (1952). tber die feinere Histologie einiger Abschnitte des Zwischenhirns und derNeurohypophyse des Menschen. Acta Anat. 16, 367-415.

JEFFERSON, J. M. (1940). A study of the subcortical connexions of the optic tract system of theferret, with special reference to gonadal activation by retinal stimulation. J. Anat., Lond., 75,106-134.

KNoCKE, H. (1953). tber das Vorkommen eigenartiger Nervenfasern (Nodulus-fasern) in Hypo-physe und Zwischenhirn von Hund und Mensch. Acta Anat. 18, 208-223.

LE GROS CLARK, W. E., McKEoWN, T. & ZUCKERMAN, S. (1938). Visual pathways concerned ingonadal stimulation in ferrets. Proc. Roy. Soc., B, 126, 449-468.

LE GROS CLARK, W. E. & MEYER, M. (1947). The terminal connexions of the olfactory tract inthe rabbit. Brain, 70, 304-328.

McLARDY, T. (1955 a). Observations on the fornix of the monkey. I. Cell studies. J. comp. Neurol.103, 305-326.

McLARDY, T. (1955b). Observations on the fornix of the monkey. II. Fiber studies. J. comp.Neural. 103, 327-344.

MAzzI, V. (1954). Recent advances in neurosecretion. Scient. Med. Ital. 3, 41-59.MEYER, M. (1950). Vth International Anatomical Congress, Oxford, pp. 128-129. Cambridge

University Press.MEYER, M. & ATLLSON, A. C. (1949). Experimental investigation of the connexions of the olfactory

tracts in the monkey. J. Neurol. Psychiat. 12, 274-286.NAUTA, W. J. H. & GYGAX, P. A. (1954). Silver impregnation of degenerating axons in the central

nervous system: a modified technique. Stain Tech. 29, 91-93.SIMPSON, D. (1952). The efferent fibres of the hippocampus in the monkey. J. Neurol. Psychiat.

15, 79-92.SPRAGUE, J. M. & MEYER, M. (1950). An experimental study of the fornix in the rabbit. J. Anat.,

LOnd., 84, 354-368.

Journal of A nato my, Vol. 92, Part 2P

%

4' 1

AI Is

I,

/

I

b*a

2

9II.

As.

. '

.,. Sv a

'1 1

.1

5;.,(..,, ..5 {,.

Is.F

t ,,

d> ies

't1C.}

.

.e'

III

*t t

~~~~~i 'I.-V~~~~~~~~~~~~~~~~~~~~~~~~

a I

r

' f

ftAI

y

* ^.1

0

i .I

a~~~~~~~~

il & {~~~~~~~

* 4. 08 i-

i

"I

V

J

A

Isll 1

f

tN

4

'i

I

'* IB

i.'9*I

a

f

#..I

'Ii

*

k . 9

ADEY AND OTHERS GLEES STAINING OF THlE MONKEY IIYPOTIIALAMLS(Facing p. 234)

"r i, 1.*<

Iw

I

%,

'Vl.S

I.4

AVt1 f

I.1~ f

4$I

\ 0.

w {.

r!.

I

..1

fIV&

I

t.I

plextc I

..t

N

jt 41

w

i' .%.::

I

0

4.

it

Journal of Anatomy, Vol. 92, Part 2

. "~~~~~~~~~~~~~~~~~~~~.,l ....... ;e r0

_~~~~~~ov I|i0.4-

.o417"-00-i1

.4' bi

10 _10

4_"p.!,.RI.R:~~~~~t

rw

*

1.~~~~~~~~~~~~~1/~ ,* m + 12

a I~~~~~~~~~~~~~~~

_a,A f w _ 5 *e x~*S N_,.12~~~~~~~~~13_ s

.

.I"'*.

* rS

_-IP% 0

A_*, 40 W.-

, _% ,_*..

;'* til *V 4, o,

VA,

ftoqji .Ia t_- 41

S. -V :e.-

:

a.

/4 1, \.15.

ADEY AND OTHERS GLEE.S STAINING OF THE MONKEY HIYPOTHIALAMUS

Plate 2

C

I4

....

.*

qA~a

ft

r *

..

11

r _ "" -

JopAip

..111~

:.

N.~ ~ ~ ~ w: ¢ _'.

14

AS

'A40" 1/x v-

.I

0 I Ws'I16

Glees staining of the monkey hypothalamus 235STOLL, J., AJMONE-MARSAN, C. & JASPER, H. H. (1951). Neurophysiological studies of subcortical

connexions of anterior temporal region in cat. J. Neurophysiol. 14, 305-316.WALL, P. D., GLEES, P. & FULTON, J. F. (1951). Corticofugal connexions of posterior orbital

surface in Rhesus monkey. Brain, 74, 66-71.

EXPLANATION OF PLATESPLATE 1

All figures in this plate are from Rhesus Am 1, with an anterior temporal lesion.Fig. 1. Degenerating fibres at base of septum ipsilateral to lesion. Nauta method. x 800.Fig. 2. Degenerating fibres at base of septum contralateral to lesion. Nauta method. x 800.Fig. 3. Degenerating fibres in the anterior commissure contralateral to the temporal lesion. Nauta

method. x 800.Fig. 4. Degenerating fibres in the ipsilateral fornix. Nauta method. x 1400.Fig. 5. Degenerating fibres in the contralateral fornix. Nauta method. x 1400.Fig. 6. Droplet terminals in the contralateral ventro-medial hypothalamic nucleus. Glees method.

x 1800.Fig. 7. Degenerating fibres in the ipsilateral bed nucleus of the stria terminalis. Nauta method.

x 800.Fig. 8. Ipsilateral hypothalamic bundle of stria terminals. Nauta method. x 600.Fig. 9. Ipsilateral hypothalamic bundle of stria terminalis. Nauta method. x 800.

PLATE 2

Fig. 10. Rhesus Am 1. Contralateral intralaminar thalamic nuclei, showing degenerating terminal.Nauta method. x 1400.

Fig. 11. Rhesus Am 1. Ipsilateral paramedian thalamic nucleus, with degenerating terminal.Nauta method. x 1400.

Fig. 12. Rhesus Am 1. Ipsilateral globus pallidus with numerous swollen terminals. Nauta method.x 1200.

Fig. 13. Rhesus Am 1. Ipsilateral globus pallidus, showing degenerating fasciculi traversing thenucleus. Nauta method. x 1050.

Fig. 14. Rhesus Fr 1. Degenerating fibre entering ipsilateral dorso-medial hypothalamic nucleus.Glees method. x 800.

Fig. 15. Rhesus Fr 1. Degenerating terminals in ipsilateral dorso-medial hypothalamic nucleus.Glees method. x 800.

Fig. 16. Rhesus Fr 1. Degenerating terminals in contralateral dorso-medial hypothalamic nucleus.Glees method. x 800.

Anat. 9216