J. Neurol. Neurosurg. Psychiat., 1954, 17, 204.

THE EFFECT OF CARDIAZOL CONVULSIONS ON THEDISTRIBUTION AND ACTIVITY OF SOME PHOSPHATASES IN THE

AREA POSTREMA OF THE RATBY

JORGE A. INSUA* and D. NAIDOOFrom the Department of Neuropathology, Maudsley Hospital, University ofLondon

During previous studies on the localization ofphosphatases in the central nervous system of therat it was observed that the area postrema showedintense enzyme activity to a number of importantphosphoric acid esters in the acid and alkalineranges. This structure therefore provided a con-venient point upon which to test the effects in situon some brain enzymes of substances used inneuropsychiatric therapy.

Since anatomical information on the area post-rema of the rat was limited, a preparatory study ofthe medulla was undertaken to establish the normalhistology. Cardiazol was the test substance ofchoice because accurate doses could be administeredintravenously and because the subsequent convul-sions could be assessed readily according to lengthand severity.

Material and MethodsPreparatory anatomical and histological work was

carried out on 22 adult albino rats weighing between 200and 250 g. The methods applied were the ordinaryhistological staining procedures including various silverimpregnation techniques.A histochemical study of the area postrema was made

on the frozen-dried brains of 45 adult albino rats. Thefemoral veins of anaesthetized ratswere exposed and 05 g.of "cardiazol" in a 10% solution in double-distilledwater was injected. The convulsions began immediatelyand the animal was transferred to a chamber into whicha small constant flow of oxygen was directed to avoidanoxia. The rats were in convulsions between a half andfour hours. Control animals were injected with anequal volume of distilled water and placed in the sameatmosphere. An experimental animal and its controlwere anaesthetized together and both brains wereremoved rapidly. The cerebellum was then cut throughapproximately between its anterior two-thirds andposterior third. The caudal block containing the areapostrema was immediately dropped into isopentanecooled to below -65° C. in a bath of acetone and CO2ice.

* British Council scholar.

After about two to five minutes the material wastransferred to a drying tube made of pyrex glass 15 cm. x3 cm. provided with a B34 ground socket. The tube waspreviously filled to the level of about 2.5 cm. withparaffin wax (m.p. 520 C.) and cooled to -78° C. in theacetone and CO, ice bath. The material was placed onthis wax bed. The tube was then affixed to the vacuumpump (2.S.50 Edwards) by means of a glass wide-boreadaptor. Between adaptor and pump was interposed anisolation valve and a P205 trap. The tube was thensurrounded by a bath (a 2-litre Dewar flask) of phenyl-ethyl-ether cooled to about -75° C. with CO2 ice. Thepump was started and the process of dehydration began.The phenyl-ethyl-ether bath was allowed to rise toabout -30° C. at which temperature it was maintainedby the addition of CO2 ice. The pressure was frequentlymeasured by a hot-wire Pirani gauge incorporated withthe adaptor connecting tissue tube to the vacuum pump.The working pressure was about 0-02 to 0-002 mm. Hgand the time required for adequate drying was about 72hours. When this period of time had elapsed the tem-perature was allowed to rise to room temperature byremoval of the refrigerant. After 30 minutes at roomtemperature the pressure in the tube containing tissuewas brought up to between 2 and 10 mm. Hg by theintroduction of dry air. Having done this, the paraffinbed in the tube was slowly melted by surrounding thetube with a warm water bath. In this manner vacuumembedding in dry air was obtained. The temperature ofthe wax was not allowed to rise beyond about 56° C.After wax penetration for 20 minutes the tube was re-moved and the blocks embedded in paraffin of the samemelting point as that which made up the wax bed.These blocks were stored at 0° C. in a dessicator. Thetissue blocks were cut within one month of embedding.Sections 15 ,u thick were mounted on albuminized 3 xI in. slides without floating on water. The sections wereplaced in a 37° C. incubator for 60 minutes. De-waxingwas carried out by successive immersions for less thantwo minutes altogether in light petroleum (b.p. 100 to1200), petroleum (b.p. 60 to 800), and isopentane (b.p.27 to 310 C.). The last solvent evaporated immediatelyafter the section was removed from the de-waxing jar.The sections were placed at once in the prepared sub-

204

Protected by copyright.

on May 1, 2020 by guest.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.17.3.204 on 1 August 1954. D

ownloaded from

EFFECT OF CARDIAZOL CONVULSIONS ON PHOSPHATASES IN RATS 205

strate media contained in coplin jars. The media weremade up as the following formulae indicate.Adenosine Triphosphate pH 6-5

0-002 M stock solution ATP... 75 ml.0-1 M sodium succinate buffer .. . 15-0 ml.0-04 M lead acetate (neutral)... 0-75 ml.

Glass-distilled water to .. . 30 0 ml.Adenosine Monophosphate pH 6-5

0-1 M sodium succinate buffer .. . 150 ml.0-04 Mlead acetate... 1-5 ml.0-01 M adenosine monophosphate. 3-0 ml.

Glass-distilled water to.30 0 ml.Aneurin Pyrophosphate pH 6-9

0-1 M sodium maleate buffer .15-0 ml.0-005 M aneurin pyrophosphate 3-0 ml.0-1 M magnesium chloride. 6-0 ml.0-04 M lead acetate. 0-75 ml.

Glass-distilled water to.30 0 ml.Glcverophosphate pH 5-3

0-1 Msodium acetate buffer .15-0 ml.0 04 M lead acetate (neutral). 15 ml.0-1 M stock solution glycerophosphate 3-0 ml.0.1 M magnesiumchloride. 1-0 ml.

Glass-distilled water to.30 0 ml.Glycerophosphate pH 9-1

0-1 M sodium barbiturate buffer. 75 ml.1-0 M calcium chloride. 30 ml.0-1 M sodium glycerophosphate 6-0 ml.

Glass-distilled water to.300 ml.Aneurin 7Pyrophosphate pH 9-1

0-1 M sodium barbiturate buffer. 75 ml.10- M calcium chloride. 30 ml.0-005 M aneurin pyrophosphate . 60 ml.

Glass-distilled water to.30-0 ml.

The preparation of adenosine triphosphate, adenosinemonophosphate, and of the buffer solutions followed themethods advised by Naidoo and Pratt (1951 and 1952aor b).The substrate medium was degassed immediately

before the introduction of the sections. In order toensure further that no gas bubbles would adhere to thesurface of the tissue the slides were flicked into themedium with a sharp movement of the fingers so thatthey went in very rapidly. Incubation proceeded forintervals of 3, 10, 20, 40, 80, and 120 minutes. Thesections were then removed, washed in four changes ofice-cold, glass-distilled water, and then placed for twominutes in 2% (NH4)2S solution.A comparison between the experimental and control

series was made by cutting ribbons of sections on arotatory microtome from single blocks of each series, andone anatomically similar section each from two com-parable ribbons from separate blocks were incubatedtogether on the same slide at the same level on the slide.Three separate criteria were employed in assessinghistological enzymatic activity. First, a histologicalcomparison of both sections incubated for the sametime, secondly the difference in incubation time requiredto produce an equal depth of staining (Naidoo andPratt, 1953), and thirdly a comparison of the response toinhibitors used with the normal control and the experi-mental material (Naidoo and Pratt, 1953).

ResultsAnatomical and Histological.-The ependymal

conus (area postrema) of the rat starts caudally at a

point some 650 u below its cephalic margin at themidpoint of the dorsal median raphe behind thecommisural nucleus of Cajal. Vessels from the

posterior median fissure are directed towards thissite which is made up of mesodermal and glialfibres. About 150 ,u rostral to this point the conus islarger, round, and its ventral limits are clearlydemarcated by a dense glial wall whereas laterallythe borders are indistinct. Further forward theabsence of a definite lateral boundary gives theimpression that the structure of the conus becomesintimately associated with that of the nucleussolitarius. At this point the central canal becomesT-shaped in cross section and the conus morewedge-shaped. There is a thick glial fibre bundleconnecting the central part of the ventral surfaceof the conus to the central canal. In between thesefibres there are some astrocytes and ependymalcells. As the restiform bodies diverge and thefasciculus solitarius with them, the conus is stretchedout into a fan-shaped structure whose handle liesin the roof of the central canal. Further cephaladthe stretched conus forms the roof of the centralcanal but the anterior boundary is reached soonafter the conus becomes an intraventricular structure.At each lateral end of the anterior boundary, whichmay be visualized as the base of a triangle, a short,thick prolongation of the conus is seen.The conus, then, is dorsally related to the central

canal and connected to its ependymal lining by aventral glio-ependymal prolongation from it; it isrelated, by making contact with it, to the nucleussolitarius by a prolongation running laterally belowthe posterior funiculi of the medulla. The nucleussolitarius lies between this prolongation and thedorsal nucleus of the vagus. The intermediate zonebetween the pars magnocellularis of the nucleussolitarius and the ventrolateral border of theependymal conus is occupied by what has beencalled the subnucleus tracti solitarii parvo cellularisby Meesen and Olszewski (1949).The boundaries of the conus are definite except

antero-laterally and at the ventral midpoint whereprolongations exist. It is a highly vascular andcellular area. The vascular arrangement is that of asinusoidal bed in which are set, among the sinusoids,astroglia, microglia, fibroblasts, and a type of cellcontaining a large oval clear nucleus with often aneccentric nucleolus and a small amount of chromatin.The larger of the cells lie at the periphery. Biel-schowsky's silver technique reveals a very smallnumber of protoplasmic processes. The astrocyticprocesses are not traceable readily to the walls ofthe sinusoids. Large amounts of collagen are seenaround the vessels and the glial wall delimiting theconus is seen to be built up by processes of fibrousastroglia. Extensive search after the use of varioussilver techniques failed to demonstrate any structure

Protected by copyright.

on May 1, 2020 by guest.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.17.3.204 on 1 August 1954. D

ownloaded from

JORGE A. INSUA AND D. NAIDOO

FIG. 1 FIG. 4

_'

.j.464,,

FIG. 2

FIG. 5

FIG. 1.-Area postrema showing relation to medullary structures.x 26. Heidenhain stain.

FIG. 2.-Anterior end of area postrema showing bridge overcentral canal at apex of the fourth ventricle. x 14. Nisslstain.

FIG. 3.-Adenosine-monophosphatase pH 6-5: relatively no

activity. x 90.

FIG. 4.-Glycerophosphatase pH 91: intense activity of areapostrema. x 12.

FIG. 5.-Aneurinpyrophosphatase pH 6-9 showing strong activityin parenchymal cells. x 350.

FIG. 3

206

Protected by copyright.

on May 1, 2020 by guest.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.17.3.204 on 1 August 1954. D

ownloaded from

EFFECT OF CARDIAZOL CONVULSIONS ON PHOSPHATASES IN RATS

FIG. 6 FIG. 8

FIG. 9

FIG. 7

FIG. 6.-Glycerophosphatase pH 5 3 showing strongly active areapostrema and intense activity in the intermediate zone. x 140.

FIG. 7.-Adenosinetriphosphatase pH 6-5: intense activity of nucleiof all cells and of nuclei of capillary cells. x 125.

FIG. 8.-Glycerophosphatase pH 5 3 (after cardiazol convulsions):activity much reduced. x 160.

FIG. 9.-Aneurinpyrophosphatase pH 6-9 (after cardiazol convul-sions) : activity much reduced. x 350.

FIG. 10.-Aneurinpyrophosphatase pH 9.1 (after cardiazol convul-sions) : activity increased, activity in capillaries greatly in-creased x 320.

207

FIG. 1 0

Protected by copyright.

on May 1, 2020 by guest.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.17.3.204 on 1 August 1954. D

ownloaded from

JORGE A. INSUA AND D. NAIDOO

which could be indubitably described as an axon.

Reticulin methods show abundant reticular fibresaround the sinusoids and a thick layer in the outersurface of the conus. The larger cells with ovalnuclei, which form the main cell of the parenchymaof the area postrema, appear to be modifiedcpendymal cells.

Adenosine-monophosphatase: Normal.-The hy-drolysis of adenosine monophosphate is found tobe situated in the fibres of the medulla but by sharpcontrast the area postrema stands out as a clear,unstained triangular area (Fig. 3). The outer borderof the conus, however, appears a little more deeplystained than the remaining fibres of the medulla.Within the conus itself it is difficult to discern any

staining of separate cellular elements. Here andthere a short, thin fibre can be seen to be stained.At the longer incubation periods the cells, particu-larly those described above as parenchymal cells,begin to show slight staining. Nucleolar staining isinconsistent and appeared only in a few cells. Themain impression is the absence of hydrolytic activityto adenosine monophosphate in the area postrema.

Adenosine-monophosphatase after Cardiazol.-Thehistochemical observations on the area postrema ofanimals convulsed with " cardiazol " show that thesites of hydrolysis of adenosine monophosphate inthe area postrema are unaltered after convulsions.The area remains clearly demarcated as an unstainedwedge. The rate of staining in the fibres of themedulla shows no change.

Glycerophosphatase pH 91 and 5-3: Normal.When sections which have been incubated withglycerophosphate atpH 9 1 are examined, the conus

appears after 40 minutes' incubation as a blacktriangle set in a grey mottled background of medulla(Fig. 4). The conus thus shows intense enzymeactivity with this substrate. The activity is so greatthat a positive result was obtained after an incu-bation period of two minutes. At such shortincubation periods activity of the capillary endo-thelium can be distinctly demonstrated. Longerincubation periods increase the staining of thecapillaries and the intersinusoidal parenchyma alsobegins to turn grey. The greyness does not start inthe cells or collagen but is general and evenlydistributed over the whole area. The area postremaand the choroid plexuses are the parts of the brainwhich stain most deeply after incubation withglycerophosphate at pH 9 1. At pH 5 3, however,the localization of the sites of enzyme activity differfrom those found in the alkaline pH. In the area

postrema the parenchymal cells are stained but not

the sinusoids or glial cells. The parenchymal cellsthemselves appear to vary in their reaction toglycerophosphate. Some stain much more deeplythan others over the same incubation period. Thatsegment of tissue described as the intermediate zone(subnucleus tracti solitarii parvo-cellularis) betweenthe conus on the one hand and the vagal and hypo-glossal nuclei on the other, also shows strong enzymeactivity (Fig. 6). The nuclei of the vagus and hypo-glossal nerves show neuronal staining and in thelonger incubation periods some parenchymalstaining as well. The neurons show deep cytoplasmicstaining, a clear nucleus, and a stained nucleolus.At the level of the fibroglial boundary there is aclear demarcation between the homogeneouslystained intermediate zone and the conus in whichonly the parenchymal cells are stained.

Glycerophosphatase after Cardiazol.-In the sec-tions incubated in glycerophosphate at pH 9 1 thereaction in the tissue from convulsed animals isconsiderably accelerated. After an incubationperiod of two minutes the capillaries in the experi-mental tissue are much darker than those of thecontrol tissue. The whole area also appears darker:no definite cytological elements can be seen in theintercapillary spaces but the parenchyma generallytakes part in the increased rate of staining. Evenagainst this darker background the staining of thecapillary walls is proportionately greater. Afterlong incubation periods of one hour or more thedifferences become less discernible and by 90minutes both normal and experimental areaepostremae are equally black. At pH 5.3 the areapostrema of the experimental animal stains con-siderably less than that of the control. In thoseanimals in which convulsions were of relativelyshort duration-about half-an-hour-before theremoval of the brain, the intermediate zone (sub-nucleus tracti solitarii pars parvocellularis) stainsdeeply but when the convulsions are prolonged up toabout four hours the intermediate zone as well asthe conus fails to stain. The longer the animalremains in a state of convulsion the more slowly doesthe intermediate zone hydrolyse glycerophosphateat pH 5 3. In the normal the neurons of the cranialnerve nuclei only showed staining of the cytoplasmof the nucleolus and of some karyosomes, whereasin the experimental animal the processes from thecells were also stained for a short distance from thecell body.

Vitamin B1 Diphosphatase pH 6-9 and 91:Normal.-When the sections are incubated withaneurin pyrophosphate (vitamin B1 diphosphate)

208

Protected by copyright.

on May 1, 2020 by guest.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.17.3.204 on 1 August 1954. D

ownloaded from

EFFECT OF CARDIAZOL CONVULSIONS ON PHOSPHATASES IN RATS 209

they present quite a different picture. The areapostrema and both parts of the nucleus solitarius aredeeply stained whereas the nuclei of the vagus andthe hypoglossus are only lightly stained. In the areaitself only a few of the parenchymal cells showactivity (Fig. 5). Those parenchymal cells whichdemonstrated staining show that their cytoplasm aswell as their nuclei and karyosomes stained.Between these cells of the conus, the vessels stand outclearly because of the deep staining of the endo-thelium. The vessels appear at moderate and longincubation periods. The glial cells stain lightly. Inthe alkaline range the vessels are distinctly stainedbut in the acid pH no vessels at all are seen. Thecranial nerve nuclei stained are similar to thosefound in the acid range. The details of neuronalstaining are clearly seen in the large neurons of thecranial nerve nuclei. The cytoplasm and the nucleo-lus are most deeply stained and the nucleoplasm isconsistently a little more lightly stained. Scatteredthroughout the nucleoplasm is a number of discretegranules of deeply staining material. In the cyto-plasm the staining appeared to be deepest in someaggregates of amorphous material. It is difficult tosay whether these bodies represented Nissl substance.

Vitamin B1 Diphosphatase after Cardiazol.-Thehydrolysis of vitamin B1 pyrophosphate is less thanthe normal control after " cardiazol " convulsions.The difference in staining is seen most clearly at incu-bation periods of 20 to 30 minutes (Fig. 9). At pH9 1, however, a further difierence is seen in thevessels, which, in tissue from animals injected with" cardiazol ", stain much more deeply than thedefinite but lightly stained vessels of the con'rolanimal (Fig. 10).

Adenosinetriphosphatase: Normal.-In the materialincubated with adenosinetriphosphate, incubationperiods of 20 minutes are needed to show even verylight staining. The nuclei of the nerve cells and theglial cells of the medulla appear at this length ofincubation (Fig. 7). No vessels or other elementsare seen to be stained. In the area postrema thenuclei and nucleoli of cells irrespective of typeappear.

Adenosinetriphosphatase after Cardiazol.-Noalteration in the rate or sites of adenosinetriphospha-tase activity was detected in the area postrema afterconvulsions. Longer incubation periods increasedthe staining of the nuclei but no new structurescould be seen. The parenchymal cells of the conusshowed distinct, large, light- to deep-staining nucleiwith deeply stained nucleoli. After a 40-minuteincubation the nuclei and nucleoli of the cranial

nerve neurons are clearly stained but the cytoplasmof the cells whether neuronal or ependymal do nottake part in the hydrolysis of adenosinetriphosphate.

DiscussionHistochemically the area shows a consistent and

definite pattern of enzyme activity. Adenosinemonophosphate is generally hydrolysed in the nervefibres. It has been shown previously that the cellbody is free from any hydrolytic properties for thissubstrate. Nucleolar staining should, at present, beregarded as a possible artefact of the method. Thenucleolus may form a focus of crystallization orhave a special affinity for the phosphate freed fromthe phosphoric acid ester. The area postremaappears to have no nerve fibres and consequentlyno enzyme activity for adenosine monophosphate.

It is possible that the occasional short fibre,rarely more than one or two in an entire section ofthe area, may be a nerve fibre with its cell bodyelsewhere. The hydrolysis of glycerophosphate atpH 9-1 shows that this alkaline phosphatase ispractically confined to the sinusoidal wall. Muchhas been written about the role of alkaline phos-phatase in glycolysis (Richter, 1951) and in theprocesses of contractility carried out by proteins(Fell and Danielli, 1943). Its bilaminar arrangementin the intestinal mucosa (Johnson and Kugler,1953) has given weight to hypotheses concerningthe phosphorylation and transport of intestinalglucose. The possibility that the sinusoids of thearea postrema may be especially sensitive to meta-bolites in the blood on the one hand, or to changesin the closely applied parenchymal cells on the other,arises. The direction of the processes assisted byenzyme action in the sinusoidal wall could con-ceivably be bidirectional: from the blood into theparenchyma as well as from the parenchyma intothe circulation.

Examination of the cerebral cortex and otherareas showed in earlier work that acid phosphataseat pH 5-3 with glycerophosphate as substrate isactive strongly in the cytoplasm of neurons. Bodianand Mellors (1944) have shown that the activity ofthis enzyme increases during the degeneration andregeneration of Nissl substance. The parenchymalcells are not all active in the sections and there is asuggestion that those closest to the sinusoids are.most prominent in their staining. As the sinusoidalbed is very dense, this point could not be clarifiedfurther. More interesting, however, is the stainingof a zone of brain tissue representing the small-celledpart of the nucleus solitarius lying between theconus and the vagal and hypoglossal nuclei, calledfor convenience the intermediate zone. Previous

Protected by copyright.

on May 1, 2020 by guest.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.17.3.204 on 1 August 1954. D

ownloaded from

JORGE A. INSUA AND D. NAIDOO

studies (Naidoo and Pratt, 1951) show acid phos-phatase (glycerophosphatase) to be a cytoplasmicenzyme, but this small zone shows staining of cellsand tissue between the cells as well. The stainingextends from the border of the conus to the vagaland hypoglossal nuclei. The neurons of the vagal andhypoglossal nuclei themselves show the typicalcytoplasmic staining observed in earlier studies.Banga, Ochoa, and Peters (1939) found aneurin tobe more effective in passing the permeability barriersthan the diphosphate, and it is possible that in thisform it enters the cells, takes part in the oxidativedecarboxylation of pyruvic acid, so that metabolismpasses beyond the pyruvate stage. The dephos-phorylation of vitimin B1 diphosphate occurs in thearea postrema in both the acid and alkaline ranges.At the alkaline pH the vessels as well as the paren-chymal cells take part in the dephosphorylatingprocess. The separate identity of glycerophosphatasehas already been established. The enzyme hydro-lysing glycerophosphate at pH 9 1 was shown to bechemically and histologically distinct from theenzyme hydrolysing aneurin pyrophosphate at thesame pH (Naidoo and Pratt, 1953). Adenosinetriphosphate is an energy-rich phosphate. Therupture of the pyrophosphate bonds is now believedto yield considerable quantities of energy. It waspreviously found that in all regions of the brain thenuclei of cells of all types showed activity in thissubstrate. The findings for the area postrema donot differ.

After " cardiazol" the cleavage ofglycerophosphatein the alkaline range is considerably acceleratedwhereas the hydrolysis of the same substrate in theacid range is retarded. The increase in alkalinephosphatase activity is restricted mainly to thecapillary endothelium where activity normally takesplace, whereas the function of acid phosphatase issuccessively reduced in the area postrema and theintermediate zone depending on the length ofconvulsions. Aneurin pyrophosphate is brokendown more slowly after convulsions than by normaltissue in the acid range, but in the alkaline range anadditional factor is that the sinusoids stain moreprominently than in the control. It is possible

therefore that non-specific alkaline phosphataseactivated by "cardiazol" convulsions acts uponaneurin pyrophosphate, breaking it down in thecapillaries where the enzyme is situated.The rest of the medulla seen in each preparation

of the area postrema was also studied to exclude thepossibility that the differences seen were generaldifferences applicable to the medulla as a whole.Changes in the cells of the reticular formation werenot sought. The changes described were limited tothe area postrema except for the increased activityof alkaline phosphatase which appeared in the sinu-soids of the area postrema as well as in the capillariesof the medulla.

SummaryThe effects of " cardiazol " convulsions upon the

distribution and activity of enzymes hydrolysingadenosine triphosphate, adenosine monophosphate,aneurin pyrophosphate, and glycerophosphate inthe area postrema of the rat are described.There are clear differences between the cytological

enzyme activity observed in the normal animal andin the animal injected with " cardiazol ".The significance of the normal distribution of

enzyme activity and of the changes after " cardiazol"convulsions is discussed.

We wish to thank Professor A. Meyer for his supportof and interest in this work, and the Bethlem Royal andMaudsley Hospital Research Fund for financial support.We wish to thank also Dr. 0. E. Pratt for his collabora-tion in the purification of some of the substrate materialsused.

REFERENCESBanga, I., Ochoa, S., and Peters, R. A. (1939). Biochem. J., 33, 1109.Bodian, D., and Mellors, R. C. (1944). Proc. Soc. exp. Biol., N. Y.,

55, 243.Fell, H. B., and Danielli, J. F. (1943). Brit. J. Exp. Path., 24, 196.Johnson, F. R., and Kugler, J. H. (1953). J. Anat., 87, 247.Meesen, H., and Olszewski, J. (1949). A Cytoarchitectonic Atlas of

the Rhombencephalon of the Rabbit. Karger, Basel.Naidoo, D., and Pratt, 0. E. (1951). Journal of Neurology, Neuro-

surgery and Psychiatry, 14, 287.-(1952a). Ibid., 15, 164.

(1952b). First Congress of the Histopathology of the NervousSystem, Rome.

(1953). Enzymologia, Amst., 16, 91.Richter, D. (1951). In Recent Progress in Psychiatry, 2nd ed.

Churchill, London.

210

Protected by copyright.

on May 1, 2020 by guest.

http://jnnp.bmj.com

/J N

eurol Neurosurg P

sychiatry: first published as 10.1136/jnnp.17.3.204 on 1 August 1954. D

ownloaded from