Quantitative effects of methylazoxymethanol acetate on purkinje cell dendritic growth

Brain Research, 143 (1978)499-511 © Elsevier/North-HollandBiomedicalPress

499

QUANTITATIVE EFFECTS OF METHYLAZOXYMETHANOL ACETATE ON PURKINJE CELL DENDRITIC GROWTH

P.

BRADLEY and M. BERRY

Department of Anatomy, Medical School, Bristol University, Bristol and Department of Anatomy, Medical School, Birmingham University, Birmingham (Great Britain)

(Accepted July 7th, 1977)

SUMMARY A quantitative analysis was made of alterations in the dendritic organisation of Purkinje cells in the cerebellum of the rat following the administration of the degranulating agent, methylazoxymethanol acetate (MAM). This drug depleted the granule cell population of the cerebellar cortex and disturbed Purkinje cell alignment such that a number of Purkinje cells became inverted and grew in the white matter. The quantitative changes that occurred in the dendritic trees of these cells (increase in segment length, decrease in segment numbers, trichotomy and branching probability) were similar to those seen following other degranulation procedures. The size of the Purkinje cell dendritic tree was found to be related to the number of parallel fibres present in the molecular layer. These results were discussed in relation to current theories of neuronal development and were shown to lend further support to the filopodial attachment hypothesis of dendritic growth.

INTRODUCTION There have been many qualitative morphological studies of Purkinje cell dendritic growth under conditions of parallel fibre deafferentation, and a wide variety of experimental approaches have been adopted including X-irradiationZ,4,7, s, virus treatmentz°,21, chemical lesioning24,25,40, the use of agranular mutant miceSZ,37 and the development of Purkinje cells in vitroZ0,~1. We have previously reported the quantitative effects of X-irradiation induced granule cell depletion on the growth of the Purkinje cell dendritic tree in the cerebellum of the rat 11. It was found that all dendritic trees in agranular irradiated cortex were markedly reduced in size (as indicated by total dendritic length and total number of segments) although mean path lengths were normal. Segment lengths were normal over proximal branches, but uniformly increased over distal branches.

500 These quantitative changes provided further evidence in support of the filopodial attachment hypothesis of dendritic growth which, though classical in origin 29, has been strengthened by the recent ultrastructural observations of synapse formation between growth cone filopodia and axons22, 23,3s, on the basis of which certain quantitative predictions can be made 9,1°. It was suggested that this hypothesis provided a possible explanation for the genesis of dendritic patterns in terms of an interaction between growing dendrites and their substrate 11, although this interaction may not always be synaptogenic as was originally postulated 9,1°, as has become evident from various studies which demonstrated the formation of axodendritic attachment plates that did not subsequently develop into synapses1,22, 23,33. The possibility that the changes observed in our previous study ~1 occurred as a consequence of the direct effects of irradiation on the Purkinje cell rather than in response to the radiation-induced parallel fibre deficit cannot be completely ruled out, despite reports that low-level irradiation causes no apparent damage to differentiated cells and their processes 6,19,3~,41. A further disadvantage of this previous study was the failure to correlate the dendritic field changes with granule cell counts. Accordingly, it was decided to investigate the growth of Purkinje cell dendritic trees using an alternative method of degranulation, namely the antimitotic agent, methylazoxymethanol acetate (MAM), and to perform a full quantitative analysis of both dendritic field parameters and the densities of granule and Purkinje cells. MATERIALS AND METHODS

Animals Ten litters of Wistar rats were reduced to 5 animals each at birth. One-day-old rats received a single intraperitoneal injection of MAM, 0.5 mg/g body weight delivered in 0.05 ml physiological saline. Eleven animals survived and were sacrificed at 30 days post partum (p.p.), their brains were removed and one-half of the cerebellar vermis was treated by Sholls 36 modification of the Golgi-Cox technique. The remaining half of the vermis was fixed in 10 ~ formol saline, wax embedded, sectioned at 5 #m and stained with cresyl violet.

Tracing of cells 100 #m thick parasagittal sections were cut from the Golgi-Cox impregnated vermis of each of the surviving animals. Fully impregnated Purkinje cells were selected from along the primary fissure if their dendrites exhibited a planar organisation parallel to the plane of the section. One cell was drawn from each animal using a camera lucida method and an oil immersion objective. Cells were classified into two categories, those that were normally oriented with respect to the pial surface (MAM-upright), and those that were inverted and growing in the white matter (MAM-inverted). Six cells of the former type and 5 cells of the latter type were analysed. Ten, 30-day-old, normal Purkinje cells were provided from a previous study 13.

501 Dendritic analysis The basic theory of the network analysis method 9,12,zn and the application of this technique to the study of normal 1° and abnormal T M dendritic fields has been described elsewhere. Dendritic branches were designated using the Strahler ordering method and a topological analysis was undertaken to detect changes in growth. Segment lengths, mean total path lengths, number of segments, frequency of trichotomy, bifurcation ratios and branching probabilities were all estimated by methods described in previous publicationsa-lL In addition the dendritic trees were centrifugally ordered using the method of Coleman and Riesen 17, and the mean segment lengths of the first 6 centrifugal orders were compared. The centrifugal approach was used to obtain a more accurate representation of the proximal dendritic branches than that provided by the Strahler ordering technique. Cell measurements The area and cell density of the various cortical layers were measured from the 5 #m cresyl violet sections. In a total of 5 sections per animal camera lucida tracings were made of the boundaries of the molecular and granular layers; the cross-sectional area of these layers and of the whole cerebellum was estimated by weighing paper cutouts. From a set of standard sampling sites within the cerebellar folia, (lobus centralis, culmen, declive, pyramis and uvula) measurements were made of the width of the internal granular, molecular and Purkinje cell layers and of granule cell density. A count was also made of the total number of Purkinje cells present in each section and their distribution within the folia was plotted. RESULTS Gross cerebellar morphology The size of the MAM treated cerebellum was greatly reduced as seen in a decrease of cross-sectional area when compared to normal 30-day cerebellum (Fig. 1). The complexity of foliation was also reduced. However, there was no apparent loss of Purkinje cells despite the decrease in overall cross-sectional area. The number of Purkinje cells/section in the MAM-treated cerebelli was calculated to be 793 4- 114 as against 829 :k 25 in the normal cerebellum. In the MAM-treated cerebelli, the Purkinje cell somata were aligned normally along the fissures, but there was a clustering of cells at the outer limit of each folium (Fig. 1). The density of cells in the Purkinje cell layer, measured in regions where a monolayer was present, was doubled in MAM-treated cerebelli (2.3/100 #m) compared with the normal distribution (1.1/100

l,m). Cerebellar measurements Table I contains the measurements of vermal cross-sectional area, granule cell density, etc. from sagittal sections of MAM-treated and normal cerebelli. Significant differences at the 1 70 level are marked with an asterisk. It can be seen that although the total number of granule cells (granule cell area × density), was significantly

502

Fig. 1. a and b: cresyl violet stained sagittal sections of cerebellar vermis from control and MAMtreated animals respectively. Scale bar - 1 ram. c and d : Purkinje cell distribution at the tips of cerebellar folia in control and MAM-treated cerebella respectively. Note the apparent immaturity of the unaligned cells clustered at the tip of the MAM-treated folium (d). Scale bar = 50 ltm.

TABLE I

Cerebellar measurements from MAM-treated and 30-day control cerebelli (means 4- S.E.M.)

MAM-treated 30-day control

Total vermal area (mm 2)

Granular layer Granule cell area (mm 2) density (cells Imm ~)

7.24:1- 1.1" 18.29 ± 1.5

2.67 4- 0.3* 7.26 4- 0.4

Molecular layer Granular layer thickness (l~m) thickness (ibm)

16,758 4- 2975 116.6 4- 11.1' 14,212 4- 1207 145.3 4- 12.7

80.1 ± 16.6 97.3 i 13.1

r e d u c e d b y a d m i n i s t r a t i o n o f M A M , the a c t u a l w i d t h o f the g r a n u l a r layer, a n d t h u s the d e p t h o f cells u n d e r l y i n g each P u r k i n j e cell was n o t altered. H o w e v e r , the d e p t h o f the m o l e c u l a r layer was significantly reduced, suggesting a n effect o f M A M o n e l e m e n t s o t h e r t h a n g r a n u l e cells.

503

Fig. 2. Representativecamera lucida tracings of MAM-upright (A), MAM-inverted (B) and control (C) Purkinje cells.

Dendritic analysis Qualitative observations. Most of the Purkinje cells in the Golgi-Cox stained sections of MAM-treated cerebellum appeared normally oriented with respect to the pial surface, although they seemed somewhat reduced in size. However, there were a few cells which were completely inverted so that their somata were located in the granular layer and whose dendritic trees could be seen arborising in the white matter. These cells had retained their polarity and their entire dendritic tree could be observed within a single plane of focus, thus minimising errors likely to accrue from projection artefact and sectioning loss. Representative drawings of these cells and of a 30-day control cell are shown in Fig. 2. The results for the MAM-upright and MAM-inverted ceils are presented separately since the environmental influences on their developing dendritic trees were totally different. Quantitative data. The mean width of the dendritic field of the normally orientated cells in the MAM-treated cerebelli was 154-4-40 #m, whilst in the normal 30-day cells the width was 253 q- 33 #m. Thus, the width of the dendritic trees of the MAM-upright Purkinje cells was significantly reduced when compared with that of their normal counterparts.

504 •--, 60,

/V~.M - inverted

T

,, - • NP~M - upright

I~

o.....o 30 day control

Z:~

50,

"~

40

I

t

.r~5

30

20,

I

I

Ii

~ - ! ~

t t

............

[ q ,.,"{ "

~.. 10.

"...........

............

.......

oi Strahler order

Fig. 3. Mean segment length (± S.E.M.) plotted against Strahler order for MAM-upright, MAMinverted and 30-day control cells. Both MAM-treated groups have significantly longer segments than the control. N e t w o r k analysis

A plot of mean segment length against Strahler order for normal, M A M upright and MAM-inverted Purkinje cells reveals that the segments of the M A M treated normally orientated cells were longer than those of the control cells but shorter than those of the MAM-inverted cells (Fig. 3). These differences were significant over the first 4 Strahler orders at the 1% level. Significance was calculated using a Student's t-test on segment length data which had undergone a square-root transformation to correct for inhomogeneity of variance. The rest of the data from the network analysis is represented in Table II. It can be seen that there was a decrease in the total number of segments in the dendritic tree both in MAM-upright and -inverted cells, the latter decreasing significantly more than the former, although both decreases were significant at the 1 ~o level. Path lengths were not significantly altered, although there appeared to be a decrease in the M A M inverted cells. These cells did show a significant decrease in the frequency of trichotomous branches when compared with 30-day control rats, and although the upright cells also had decreased trichotomy, the difference was not statistically significant. TABLE II Network analysis data for MAM-upright, MAM-inverted and 30-day control cells (means 4- S.E.M.)

30-day control MAM-upright MAM-inverted

Total segments

Path length ( l~m)

Bifurcation ratio

% Trichotomy

986 i 123 298 4- 96 101 4- 15

178.4 ± 25 171.1 ± 22 140.1 4- 15

3.1 ± 0.0 3.24 4- 0.3 3.29 ± 0.43

6.5 ± 1.2 5.3 4- 1.2 3.8 4- 1.7

505 GROUP

%FREQUENCYOFTOPOLOGICALTYPES

,~J5 MAM inverted

22

82 18

MAM upright

86

73 27

30 day control

467

67 33

55 35 10

MAM inverted

20

MAM upright

59

37 42 20

30 day control

321

47 36 ll

MAM inverted

14

MAM upright

36

29 14 36 14 7 0 19 19 17 8 31 6

30day control

238

24

18 19 13 18 8

Fig. 4. Table representing the frequency of topological types over the 4th, 5th and 6th pendant arc

series for MAM-upright, MAM-inverted and control Purkinje cells (see text for full explanation).

Topological analysis The frequency of topological types over the 4th to 6th pendant arc series is shown in Fig. 4 for 30-day control, MAM-upright and MAM-inverted cells. The results of the Chi-squared comparison of control against experimental distributions are given in Table III. It can be seen that over the first two pendant arc series the topological type distribution of MAM-upright Purkinje cells was not significantly different from the control distribution, that is, these trees appeared to have developed by random dichotomous branching at terminals a. The 4th pendant arc series of the MAM-inverted cells did differ significantly from the control distribution and this suggests that these cells grew by means other than random terminal dichotomous branching. TABLE III

The distribution of topological types for 30-day control cells compared against that for MAM-upright and MAM-inverted cells using the Chi-squared statistic A n asterisk denotes a significant difference.

Control vs.

Pendant arc Zz

P

MAM-upright MAM-upright MAM-upright

6 5 4

13.12 3.66 1.33

0.1 * < 10 < 10

MAM-inverted MAM-inverted

6 5 4

31.94 4.27 10.18

< 0.1" < 10 < 0.1 *

MAM-inverted

506 60, ..._= 50,

T

,m,, MAM - inverted

'~

.--e

I"~.

....... 30 day control

,I

= 4o, =

/V~M - upright

"~

30, " ~ - - 4:,,

:~

l

20,

°%

10"

o

T--"

"

%.

°'*%.

.

.

.

.

....................

T

ii

"L

....i.. i

~"

i

Centrifugal order

Fig. 5. Mean segment length plotted against centrifugal order for MAM-upright, MAM-inverted and 30-day control cells. The S.E.M.s for all groups were of similar magnitude and for simplicity are included only with the control data.

Centrifugal ordering The standard errors about the mean segment lengths for the first 6 centrifugal orders (Fig. 5) were so large that statistically significant differences could not be demonstrated. Nevertheless, the pattern of the graph showing these lengths for the three groups seems to mirror that seen in the graph of segment length plotted against Strahler order (Fig. 3) with mean segment length increased in both MAM-treated groups, indicating that the effects of parallel fibre deafferentation are manifested over the proximal as well as the distal dendrites.

Branchingprobability The probability of branching present in the dendritic trees of the control, MAMupright and MAM-inverted Purkinje cells over the period 0-30 days p.p. was calculated from a frequency histogram of non-terminal segment lengths and is tabulated below. The highest branching probability is seen in the 30-day control dendritic trees with the MAM-upright and MAM-inverted trees showing successively lower values (cf. Table IV). These results are inversely correlated with the segment lengths measured for these three groups; the higher the probability of branching, the shorter the dendritic segments that are formed. TABLE IV

Branching probability over the period 0-30 days in 30-day control, M.4M-upright and MAM-inverted Purkinje cells Branching probability 30-day control MAM-upright MAM-inverted

0.16 0.10 0.06

507 DISCUSSION The administration of a single dose of MAM did not result in the destruction of the entire granule cell population of the cerebellar cortex. However, the proliferation of these cells was affected, such that their total numbers were greatly reduced by 30 days p.p. Woodward et al. 4° injected MAM into rats on the first 4 postnatal days and studied the recovery of the external germinal layer. This layer was much thinner than normal at 7 days p.p., but by 14 days p.p. became thicker than in controls and persisted for longer than usual to compensate for the early cell loss. Similar destruction of the external germinal layer occurs after the administration of 5-fluorodeoxyuridine but reconstitution subsequently occurs by prolongation of the proliferative phase 34. Irradiation at a level which destroys postmitotic premigratory cells, but leaves intact the germinative precursors also allows recovery of the external germinal layer and can lead to the formation of an ectopic granular layer~, or cause abnormalities of dendritic formT,S,11,1a. However, in none of these situations, despite the compensatory efforts of the external germinal layer, does the granule cell population reach its normal level, and there is always a period early in development during which there are very few parallel fibres formed. The depletion of the granule cell population in MAM-treated cerebellum is accompanied by the failure of Purkinje cells to align in a monolayer and a disorientation of dendritic trees with some cells completely inverted. This phenomenon is seen after all degranulation proceduresll,20,21,24,25,a0,z1,4° and in the Reeler a2 and Weaver as,aT mutant mice. Disorientation may thus be associated with the paucity of granule cells per se or with a decrement in the numbers of Bergmann glia, whose processes could be the main orientating factor in Purkinje cell dendritic organisation. Bergmann glia could be affected by degranulating procedures, since, like granule cells, they are undergoing active division to keep pace with the expanding cerebellum2s. The presence of inverted cells in the MAM-treated cerebellum produced an interesting experimental situation by providing two distinct cell populations. One group of Purkinje cells were normally orientated with respect to the pial surface, but their packing density in the Purkinje cell layer was greatly increased, thus decreasing the area of molecular layer available for each dendritic arborisation, and this was reflected in the decrease in the width of their dendritic trees. The second group of cells were those that became inverted and grew into the white matter. These cells provided an interesting model on which to test the filopodial attachment hypothesis of dendritic growth, since their dendrites were ramifying in an environment specifically lacking the type of axons with which they normally establish contact i.e., the potential for the development of axodendritic interactions was very low, and it would be predicted that these cells should display an increase in dendritic segment length, a decrease in the frequency of trichotomy and a reduced branching probability. These changes were in fact observed in the MAM-inverted cells when they were compared with 30-day control and MAM-upright cells, and provide an indication that the branching pattern of a dendritic tree may depend on

508 the density and location, during its development, of the specific axons which will eventually provide the tree's presynaptic innervation. The filopodial attachment theory would suggest that, in the absence of compatible membrane to which growth cone filopodia would attach, branches would only rarely be established. Hence, segments would grow longer before they branched and, if a branch was established, the likelihood of it being trichotomous would be decreased. In the case of the MAMupright Purkinje cell compatible membrane would be provided by parallel fibres but, in the MAM-treated cells growing inverted in the white matter there would be no compatible membrane to which filopodia could adhere, and branching frequencies would become determined only by the physical constraints of the environment. These results are directly comparable with those from our study of agranular irradiated cortex 11 where both upright and inverted Purkinje cells displayed increases in segment length and decreases in the frequency of trichotomy. This congruity between the results presented here and those obtained with a different method of degranulation would suggest that the results are treatment-independent and are not the consequence of the direct action of MAM or irradiation on the developing dendritic tree per se. It would seem that the decrement in parallel fibre density subsequent to degranulation leads to alterations in the length and order of branching of developing Purkinje cell dendrites. The decrease in parallel fibre density was also related to a reduction in the size of the Purkinje cell dendritic trees growing in the MAM-treated cerebelli. The number of segments in the arborisation of the Purkinje cells growing upright in the molecular layer was significantly reduced to about one-third of the normal value and the number of segments in the inverted cells was reduced to one-ninth of the normal value. This would suggest that there is a correlation between the extent of dendritic growth and the presence of parallel fibres. The fact that granule cell numbers were reduced, whilst the Purkinje cell population remained constant, meant that the ratio of the numbers of granule cells to the numbers of Purkinje cells was greatly reduced. This ratio varies from species to species 2v and is directly related to the complexity of the Purkinje cell dendritic tree. For example, Lange 2v calculated that there were 897 granule cells per Purkinje cell in the rat as against 778 per Purkinje cell in the mouse, a difference which seems to be reflected in the total number of segments in the dendritic tree, i.e. 950 in the rat and only 650 in the mousO 4. Furthermore, although the depth of the granular layer in MAM-treated cerebelli was normal, the width of the molecular layer was significantly reduced. This latter parameter may be determined by three factors, firstly, the number of parallel fibres arising from granule cells directly below the site of measurement, secondly, the ingrowth of parallel fibres from granule cells located at varying distances along the folium ~8, and, thirdly, the relative proportion in the neuropil of glial and vascular elements. Since in the MAM-treated cerebelli the first of these factors was not affected, it is possible that both the growth of parallel fibres along the folium and the proliferation of glia were retarded, thus effectively reducing the number of parallel fibres with which the Purkinje cell dendritic tree could make contact. The idea that there is a direct relationship between the number of parallel fibres

509 present and the size of the Purkinje cell dendritic tree is supported by other quantitative observations on partially and completely agranular cerebelli 11,13. The work of Privat and Drian 31 and of Calvet et al. 15 with isolated explants of cerebellum cultured in the presence of MAM has shown that, even in the complete absence of climbing fibres and cerebellar interneurons, a rudimentary Purkinje cell dendritic tree develops and is covered with unattached spines. The factors responsible for this initial dendritic outgrowth are not known, but it seems that the presence of parallel fibres is necessary for the induction of a dendritic arborisation beyond this initial stage, the extent of arborisation being related to the total number of parallel fibres present. The value of this relationship in terms of the economy of the animal's resources is obvious, ensuring, as it does, that dendritic growth does not fall short of or exceed the axonal demands in any particular location of the CNS, yet, at the same time, allowing maximal use to be made of the afferents that are present; and any growth mechanism that subtends such a relationship will have strong selective pressures in its favour. Furthermore, such a relationship meets the requirements for gene-saving mechanisms in the specification of neuronal networks18, 89 since a single mechanism could potentially control the size of all dendritic arrays. Thus, if dendrites were to grow according to precepts laid down by the filopodial attachment hypothesis, they would rapidly arborise in dense arrays of synaptically compatible axons and would grow straight through regions of low axon density until their growth was arrested by intrinsic or extrinsic factors. An example of this is provided by those cells growing in the white matter in MAM-treated cerebellum, in which very few dendritic segments were formed. Nevertheless, those segments that were formed appeared mature and the dendritic trees of the MAM-inverted cells had clearly passed through their early developmental phases and had an adult, though sparse, appearance. This is evidenced by the fact that the path lengths of these cells were not significantly different from those of control cells, a situation similar to that of Purkinje cells in agranular irradiated cortex 11. These dendrites had grown as far from the soma as they would normally, but had branched less in doing so. ACKNOWLEDGEMENTS We are grateful fo Dr. R. Flinn for his assistance with the computer analysis, and to W. McKenna and B. Bhogal for the histology. This work was supported by a grant from the S.R.C.

REFERENCES 1 Altman, J., Coated vesicles and synaptogenesis. A developmental study in the cerebellar cortex of the rat, Brain Research, 30 (1971) 311-322. 2 Altman, J., Experimental reorganisation of the cerebellar cortex. III. Regeneration of the external germinal layer and granule cell ectopia, J. comp. NeuroL, 149 (1973) 153-180. 3 Altman, J., Experimental reorganisation of the cerebellar cortex. V. Effects of early X-irradiation schedules that allow or present the acquisition of basket cells, J. comp. NeuroL, 165 (1976) 31-48. 4 Altman, J. and Anderson, W. J., Experimental reorganisation of the cerebellar cortex. I. Morpho-

510

5

6 7

8

9 i0 11 12

13 14 15 16 17 18 19

20

21

22 23 24 25 26

27 28

logical effects of elimination of the microneurons with prolonged X-irradiation started at birth, J. comp. Neurol., 146 (1972) 355-406. Altman, J. and Anderson, W. J., Experimental reorganisation of the cerebellar cortex. 1I. Effects, of elimination of most microneurons with prolonged X-irradiation started at four days, J. comp. Neurol., 149 (1972) 123-152. Altman, J., Anderson, W. J. and Wright, K. A., Selective destruction of the cerebellar cortex with fractionated low-dose x-rays, Exp. Neurol., 17 (1967) 481497. Anderson, W. J. and Stromberg, M.W., Effects of low-level x-irradiation on cat cerebella at different post-natal intervals. 1. Quantitative evaluation of morphological changes, J. comp. Neurol., 171 (1977) 17-38. Anderson, W. J. and Stromberg, M. W., Effects of low-level x-irradiation on cat cerebella at different postnatal intervals. II. Changes in Purkinje cell morphology, J. comp. Neurol., 171 (1977) 39-50. Berry, M. and Bradley, P. M., The application of network analysis to the study of branching patterns of large dendritic fields, Brain Research, 109 (1976) 111-132. Berry, M. and Bradley, P. M., The growth of the dendritic trees of Purkinje cells in the cerebellum of the rat, Brain Research, 112 (1976) 1-35. Berry, M. and Bradley, P. M., The growth of the dendritic trees of Purkinje cells in irradiated agranular cerebellar cortex, Brain Research, 116 (1976c) 361-387. Berry, M., Hollingworth, J., Anderson, E. M. and Flinn, R. M., Application of network analysis to the study of the branching patterns of dendritic fields. In G. W. Kreutzberg (Ed.), Advances in Neurology, Vol. 12, Raven Press, New York, pp. 217-245. Bradley, P. M. and Berry, M., The effects of reduced climbing and parallel fibre input on Purkinje cell dendritic growth, Brain Research, 109 (1976) 133-151. Bradley, P. M. and Berry, M., The growth of dendritic trees in mutant mouse cerebellum. A quantitative Golgi study of weaver and staggerer mice, in preparation. Calvet, M., Lepault, A. and Calvet, J., A Procion yellow study of cultured Purkinje cells, Brain Research, 111 (1976) 399-406. Changeux, J. P. and Danchin, A., Selective stabilisation of developing synapses as a mechanism for the specification of neuronal networks, Nature (Lond.), 264 (1976) 705-712. Coleman, P. D. and Riesen, A. H., Environmental effects on cortical dendritic fields. I. Rearing in the dark, J. Anat. (Lond.), 102 (1968) 363-374. Haddara, M. A. and Nooreddin, M. A., A quantitative study on the postnatal development of the cerebellar vermis of the mouse, J. eomp. Neurol., 128 (1966) 245-254. H~mori, J., Development of synaptic organisation in the partially agranular and in the transneuronally atrophied cerebellar cortex. In R. Llin~is (Ed.), Neurobiology of Cerebellar Evolution and Development, Vol. 3, Amer. Med. Ass., Chicago (1969) pp. 845-858. Herndon, R. M., Margolis, G. and Kilham, L., The synaptic organisation of the malformed cerebellum induced by perinatal infection with feline panleukopenia virus (PLU) (I). The Purkinje cell and its efferents, J. Neuropath. exp. NeuroL, 30 (1971) 557-569. Herndon, R. M., Margolis, G. and Kilham, L., The synaptic organisation of the malformed cerebellum induced by perinatal infection with the feline panleukopenia virus (PLU) (II). The Purkinje cell and its afferents, J. Neuropath. exp. Neurol., 30 (1971) 557-569. Hinds, J. W. and Hinds, P. L., Synapse formation in the mouse olfactory bulb. II. Morphogenesis, J. eomp. Neurol., 169 (1976) 15~,0. Hinds, J. W. and Hinds, P. L., Synapse formation in the mouse olfactory bulb. I. Quantitative studies, J. comp. NeuroL, 169 (1976) 41-62. Hirano, A. and Jones, M., Fine structure of cycasin induced cerebetlar alteration, Fed. Proc., 31 (1972) 1517-1518. Hirano, A., Dembitzer, H. M. and Jones, M., An electron microscopic study of cycasin induced cerebellar alteration, J. Neuropath. exp. NeuroL, 31 (1972) 113-125. Hollingworth, T. and Berry, M., Network analysis of dendritic fields of pyramidal cells in the neocortex and Purkinje cells in the cerebellum of the rat, Phil. Trans. roy. Soe. B, 270 (1975) 227262. Lange, W., Cell number and cell density in the cerebellar cortex of man and some other mammals, Cell Tiss. Res., 157 (1975) 115-124. Lewis, P. D., Ffil6p, Z., Haj6s, F., Balazs, R. and Woodhams, P. L., Neurologia in the internal granular layer of the developing rat cerebellar cortex, Neuropath. appl. Neurobiol., 3 (1977) 183-190.

511 29 Morest, D. K., The growth of dendrites in the mammalian brain, Z. Anat. Entwickl.-Gesch., 128 (1969) 290-331. 30 Privat, A., Dendritic growth in vitro. In G. W. Kreutzberg (Ed.), Advances in Neurology, Vol. 12, Raven Press, New York, pp. 201-216. 31 Privat, A. and Drian, M. T., Postnatal maturation of the rat. Purkinje cells cultivated in the absence of two afferent systems: an uttrastructural study, J. comp. Neurol., 166 (1976) 201-244. 32 Rakic, P. and Sidman, R. L., Synaptic organisation of displaced and disoriented cerebellar neurones in Reeler mice, J. Neuropath. exp. Neurol., 31 (1972) 192. 33 Rakic, P. and Sidman, R. L., Organisation of cerebellar cortex secondary to deficit of granule cells in Weaver mutant mice, J. comp. Neurol., 152 (1973) 133-162. 34 Shimada, M. and Langman, J., Repair of the external granular layer after postnatal treatment with 5-fluorodeoxyuridine, Amer. J. Anat., 129 (1970) 247-260. 35 Shofer, R. J., Pappas, G. D. and Purpura, D. P., Radiation induced changes in morphological and physiological properties of immature cerebellar cortex. In J. J. Haley and R. S. Snider (Eds.), Response of the Nervous System to Ionizing Radiation, Little-Brown, New York, 1964, pp. 476-508. 36 Sholl, D. A., The organisation of the visual cortex of the cat, J. Anat. (Lond.), 89 (1955) 33-46. 37 Sotelo, C., Dendritic abnormalities of Purkinje cells in the cerebellum of neurologic mutant mice (Weaver and Staggerer). In G. W. Kreutzberg (Ed.), Advances in Neurology, Vol. 12, Raven Press, New York, pp. 335-351. 38 Vaughn, J. E., Henrikson, C. R. and Grieshaber, J A., A quantitative study of synapses in motor neuron dendrite growth cones in developing spinal cord, J. Cell Biol., 60 (1974) 664-672. 39 Wolpert, L. and Lewis, J. H., Towards a theory of development, Fed. Proc., 34 (1975) 14-20. 40 Woodward, D. J., Bickett, D. and Chanda, R., Purkinje cell dendritic alteration after transient developmental injury of the external germinal layer, Brain Research, 97 (1975) 195-214. 41 Yamazaki, J. M., Bennett, L. R. and Clements, C. D., Behavioural and histological effects of head irradiation in newborn rats. In. T. G. Haley and R. S. Snider (Eds.), Response of the Nervous System to Ionizing Radiation, Little-Brown, Boston, Mass., 1964, pp. 224-269.