Functional elimination of afferent pathways and decreased safety factor during postembryonic development of cockroach giant interneurons

Developmental Brain Research, 8 (1983) 311-320

3 11

Elsevier

Functional Elimination of Afferent Pathways and Decreased Safety Factor During Postembryonic Development of Cockroach Giant Interneurons M. E. SPIRA and Y. YAROM

Department of Neurobiologv, Life Sciences Institute, Hebrew Universi(v(Givat Ram), Jerusalem (Israel) (Accepted December 14th, 1982)

Key words: safety factor - postembryonic development - interneurons - cockroach

The giant interneurons (GIN) from the cockroach CNS undergo two major physiological changes during the postembryonic developmental period: (A) a marked decrease in the number of afferent pathways innervating the GIN at the metathoracic ganglion (Ts); and (B) a gradual decrease in the safety factor for impulse propagation along the intraganglionic segment in T3. In 100% of the experiments (n ~ 100) performed on GIN from early developmental stages, spontaneous postsynaptic potentials (SPSPs) were recorded; in adults, on the other hand, SPSPs have been recorded in only 34% of the experiments (n -- 74), Evoked synaptic potentials can be elicited in nymphal stages by stimulation of 8 nerves ofT3, the contralateral connectives, ipsi- and contralateral nerve roots 2, 3, 5, and by stimulation of adjacent GINs. In adult, PSPs can be evoked by stimulation of adjacent GINs, and contralateral thoracic connectives, but not from nerves 2, 3 and 5. The functional disappearance of synaptic inputs to the GINs does not reflect a general phenomenon of reduction in synaptic transmission efficacy. In previous studies it was demonstrated that high frequency stimulation of adult GIN leads to blockage of impulse propagation in T3. In nymphal stages, the safety factor for propagation of impulses along T3 is higher. The reduction in safety factor appears gradually during the postembryonic developmental period. From analysis of the mechanisms underlying the elimination of functional afferent pathways and the appearance of low safety factor (see consecutive paper by Yarom and Spira43) it is concluded that the functional elimination of afferents is a consequence of decreased transmission efficacy, while the appearance of low safety regions for impulse propagation is a consequence of morphological changes of the GIN segment within ganglion T3. INTRODUCTION

The development of the central and peripheral nervous systems of both vertebrates and invertebrates is not completed during embryogenesis (for review see ref. 2). Changes in geometry of soma, dendrites and a x o n s 1.6.7-16.26,30,3L36-41, a s well as changes in membrane properties and connectivity~s.23,37, have been observed in postnatal stages. Structural modifications of a neuron with or without changes in its membrane properties may alter its integrative and functional properties. As a consequence of such changes, behavioral patterns may be altered. Another common feature in the development of the peripheral and central nervous systems in vertebrates is a reduction in the number of axons that synapse with a given target cell during embryonic and postembryonic development (for review, see Purves and Lichtman32). For exam0165-3806/83/$03.00 (~ 1983 Elsevier Science Publishers B.V.

pie, the number of axons innervating a single muscle cell decrease during the first few weeks after birth of mammals and a one-on-one pattern of innervation is established 32.3-~. A similar phenomenon was documented to occur in autonomic ganglia 2°-22, Purkinje cells ~L and in the visual c o r t e x 1732 34. In the present work, we examine the changes in properties and pattern of innervation of identifiable interneurons during postembryonic development. The ventral giant interneurons (GINs) of the cockroach (Periplaneta americana) were selected for this study, since they can be morphologically identified throughout the various postembryonic developmental stages and are large enough throughout these stages to allow for intracellular recording and stimulation. The giant interneurons of the cockroach Periplaneta americana originate in the last abdomi-

312 nal ganglion (A6) and extend continuously along the ventral nerve cord, at least up to the subesophagial ganglion j3.3~.3'~. In adult cockroaches the giant interneurons receive synaptic inputs from the cercal nerves at A69 and from unidentified pathways at the metathoracic ganglion (T3)4°. The synaptic inputs to the GINs in T3 are located on neurites which extend from the fibers into the neuropile~L In studies from our laboratory j°,:~-29-3~,4°, we have shown that there is a low safety factor for impulse propagation at T3. Thus, stimulation at high frequency causes conduction block at this region. In addition, activation ofsynaptic inputs to T3 produces a local increase in the conductance which can modulate this safety factor4°. In the present article, we describe two major changes that occur during postembryonic development. They are: (a) the number a n d / o r efficacy of afferent pathways terminating on the GINs at T3 is markedly reduced during postembryonic development; (b) the safety factor for impulse propagation is high in the early postembryonic developmental stages and is gradually reduced during development. In the following articlea3 we use morphological and electrophysiological techniques to study the mechanisms that underlie these changes. B

20

15

Et

N

N4 N3

~

.

~

N

~]

~1

t

Naz~'~a-51 Aa-4 ] An-3 I Aa- 2 ~ L ~

Ad--i ;

n¢

'I insto,

Fig. 1. Definition of the postembryonic development stages. A: The body length of 222 cockroaches collected randomly from our laboratory culture was measured. Note the 7 distinct peaks; the last peak (dashed line) corresponds to the adult stages (AD), the other 6 peaks correspond to the nymphal stages ( A d - 6 - A d - l ) . B: schematic drawing of the metathoracic ganglion, the peripheral and central nerves. Thoracic connectives N t, abdominal connectives N'7, and 5 nerve roots N2-N 6. Microelectrodes for voltage recording and intracellular stimulation (R, St) were inserted into giant interneurons at the caudal and rostral edges of the ganglion.

MA I'ERIALS AND MEFHODS

Animals The cockroaches (Periplaneta americana) were cultured in the laboratory at a temperature of 23 30°C, The culture contained nymphs in various stages of development and adult specimens. The developmental stages of the individuals collected from our culture were determined by body measurement (from the first thoracic to the last abdominal segment). Examination of body length frequency histogram of 222 individuals collected randomly from the culture (Fig. IA) showed 7 fairly distinct peaks at 5.11, 14, 17, 21, 25 and 30 ram. The number of peaks served as an estimate of the number of nymphal stages during development under our culture conditions. The last peak corresponded to the adult stages (Ad) as marked by the appearance of the wing buds. In this paper, the nymphal stages were defined by the number of molts prior to reaching adulthood. For example, the last nymphal stage is referred to as adult minus one ( A d 1) and nymphs of ! 0 mm are referred to as A d - 5 (adult minus 5).

Preparation The ventral nerve cords of adult and larval cockroaches were isolated, as described by Spira et al. 3~. The nerve roots to the metathoracic ganglia (T3) were separated from the coxal muscle to allow electrical stimulation by suction electrodes. The nerve cord of the larval preparation was secured to the bottom o f a Sylgard chamber by the application of a small drop of 2 5% Agar solution (Bacto-agar, Difco) to the nerve cord close to ganglion T3. This method does not cause any obvious damage to the nervous system and allows stable intracellular recording. All preparations, including the adult, were secured to the bottom of the chamber, ventral side up, to facilitate recording from the largest ventral GINs 3~.

Solutions Experiments were performed in continuous flow (1 2 ml/min) of a physiological solution composed of 214 mM NaC1, 3.1 mM KC1, 7 mM

313 stages, spontaneous postsynaptic potentials (SPSPs) were recorded (more than 100 experiments), while in adults, SPSPs were detected in only 34% of the experiments (74 experiments to increase the resolution for detection of SPSPs records were made at a higher gain and after hyperpolarization of the membrane). The SPSPs recorded from adult and nymphal stages differed in several aspects: in the nymphal stages the SPSP mean frequency was 24/s (ranging between 18 and 35/s, 10 experiments, Ad-4), while the mean frequency in adults was 9/s (ranging between 3 and 16/s, 7 experiments). The range of SPSP amplitude distribution in the nymphal stages (Fig. 2E, F) was always larger than in the adult preparations (Fig. 2G, H). Finally, in most of the experiments two peaks were resolved in the amplitude histogram of the nymphal stages (2F), while a single peak, with distribution close to normal, was seen for the adult SPSPs (Fig. 2G, H). Although the SPSP frequency, amplitude and amplitude distribution in nymphal GIN differ significantly from the adult GIN, their rise times are similar. The mean SPSP rise time is 2.8 ms in nymphal stages and 3 ms in adults (ranging in both between 2 and 4 ms). An example of re-

CaCI2, and I mM Tris. The pH was adjusted to 7.4-7.6. Bath temperature was 20-25 °C.

Recording and stimulation The abdominal and thoracic connectives were stimulated by bipolar silver chloride hook electrodes. For stimulation of the ipsi- and contralateral roots N2, N3, N4, and N 6 (nomenclature after Guthrie and TindalU 5, Fig. IB), suction electrodes were used. The responses were recorded by glass microelectrodes filled with 3 M KC1 (10-15 M£) inserted into a GIN at either the caudal or rostral edge of ganglion T3. The same microelectrodes served for both voltage recordings and current injection (WPI-microprobe system). RESULTS

Synaptic activity The spontaneous and evoked synaptic activity from various nymphal stages and adult GINs were compared by intracellular recordings at the caudal base of ganglion T3.

Spontaneous synaptic potentials In 100% of the experiments on the nymphal A

E

C

G

40 f

2 F 40v B

D

uoom~ 20

m

8ms

40

0

~

0

2

4

L_ 4

H

40I

2

Iil !

m m

i'

6 2 4 Amplitude of PSP (mV)

Fig. 2. Spontaneous synaptic potential in adult and nymphs. A-D: intracellular recording made at the caudal base of the metathoracic ganglion, from nymphal stage A d - 4 (A, B) and adult cockroach (C, D). B, D: superimposed sweeps of a simultaneous recording from the rostral and caudal edges of ganglion T3. B, nymph; D, adult. Note the similarity in rise times and amplitudes of the spontaneous PSPs in the caudal and rostral electrodes (upper and lower beam). E, H: amplitude distribution of spontaneous synaptic potentials in two nymphal preparations (Ad-6, A d - 5 ) (E, F), and 2 adults (G, H). Note that the amplitude distribution of adult is close to normal while in nymphal stages at least 2 peaks can be seen. These differences indicate that more presynaptic pathways terminate on nymphal GINs.

314 cords from which the rise times were measured is shown in Fig. 2B and D in which simultaneous recordings from both sides of the ganglion T3 were made. In addition, these records demonstrate that the rise time and amplitude of a given SPSP recorded simultaneously from both sides of the ganglion T3 are similar. Since subthreshold potentials are attenuated to 70% along the axonal segment in T3, it is reasonable to assume that these PSPs originate at T3 and not in neighboring ganglia. The decline in number of preparations in which SPSPs were detected, from 100% in nymphs to 34% in adults, occurred during the last two nymphal stages.

Evoked synaptic activity To characterize the pathways that terminate on the GINs at T3, we electrically stimulated the contralateral thoracic connectives between ganglia T2 and T3 (Nl), the ipsi- and contralateral nerve roots N2 N6 and the contralateral abdominal connective between ganglia A~ and A2 (NT)

A

4

B

I0

i ms

Fig. 3. PSPs evoked by stimulation of Nt, adult GIN. (Nl) was recorded at the caudal base ofTY A - C increase in stimulus intensity.

A

B

G

H

IO0%-~lOmV

IO0%q IOmV~IIT

50°/o'

50% I t t ..,..~ r~

5mY

Ad-5Ad-4Ad~2/~:1-1Ad

30ms

5mY II~,i] ~ t _F] II.,LIIJ.II,LIIJ~=,~ Ad-SAd-4Ad-2Ad-IA¢

Fig. 4. PSPs evoked by stimulation of Nj in nymph stage Ad-5. The responses to contralateral TI T2 connective (N1) stimulation were recorded at the caudal base of ganglion T3. The stimulus intensity was increased from A to D, Variability in the PSPs amplitude and shape is revealed in D and E in which the stimulus intensity was kept constan t. The PSPs are composed of early and delayed peaks. The delayed peak can be evoked independently of the early one by reversing the polarity of stimulation (F). The mean maximal amplitude (empty bars in G and H) and percentage of preparations in which PSPs could be evoked (filled bars), is gradually reduced during the course of development. G: early peak. H: delayed peak.

(Fig. 1B). The responses to these stimuli were intracellularly recorded at the caudal base ofT3. Synaptic potentials could be evoked by stimulation of Nt in only 24% of the experiments on adult cockroaches (74 experiments). The amplitude of the evoked PSP increased with the increase in stimulus intensity, indicating convergence of terminals on the GINs in T3. The maximal PSP amplitude was 2 mV with a rise time of 2 4 ms (Fig. 3). The responses to stimulation of N~ in early nymphal stages differed from those of adults. As shown in Fig. 4, a gradual increase in stimulus intensity of N~ (Fig. 4A-E) evoked a compound PSP with 2 distinct peaks. The early PSP was sustained even at high frequencies (20 Hz) for about 60 s. The second peak, which appeared after a delay of 20~ 30 ms, was labile and disappeared within 10 s at a stimulation rate of 2 5 Hz. The delayed responses could be evoked independently of the early one (Fig. 4F), indicating that the 2 components were activated by 2 independent pathways. The long delay and the susceptibility to stimulation at high frequency indicates that delayed response is mediated through a polysynaptic pathway: even the short-

315 er latency responses are still long enough to include one interneuron, thus these too are most likely polysynaptic afferent pathways. The decline with age in percent of preparations revealing the early and delayed responses is shown in Fig. 4G, H (black bars). The reduction in the average maximal PSP amplitude (evoked by supermaximal stimulation) is indicated by the empty bars. The early component of the evoked PSP was detected throughout nymphal stages A d - 5 , Ad-1. The amplitude of this response was reduced from 10 mV in stages Ad-5, Ad--2 to 6 mV in stage A d - 1 . In adults, when the response was detected (only 25% of the preparation), the maximal response was 2.5 mV (Fig. 4). The delayed component of the PSP was observed in only 60% of the experiments of stage Ad-5. The percentage of preparations in which this response was detected and its average amplitude gradually declined during the course of development (Fig. 4H). Stimulation applied to the ipsi- and contralateral roots N2, N3, N4, and N5 failed to produce any response in adult GINs. To facilitate the probability of detecting a synaptic input, we hyperpolarized the membrane to -120 mV and, even under these conditions, could detect no evoked PSPs. On the other hand, synaptic potentials could be evoked by stimulation of N 2, A

C

Ioo%i

I

[]

Ad-5 Ad-4Ad-2Ad-I Ad B

13

IO0 °/oi

50ms

•

Ad-5 Ad-4Ad-2 Ad-I Ad

Fig. 5. Synaptic potentials generated by stimulation of the ipsi-(A) and contralateral (B) N 2, in nymphal stage A d - 5 . The mean maximal PSP amplitude (empty bars) and percentage of preparations in which the responses could be evoked (filled bars) is reduced during the postembryonic development. Note that in adult, stimulation of N2 failed to evoke PSPs.

A

m

B

I m

lOms Fig. 6. Reciprocal synaptic interactions between GINs within the same connective. The records are from experiments on nymph stage A d - 5 . Two GINs were impaled by microelectrodes. One at the caudal edge ofT3 and the second at its rostral edge. A: intracellular stimulation of one GIN produced a PSP in the other. B: stimulation of the second GIN produced a PSP in the first one. This type of interaction persists throughout the development. Voltage calibration: upper trace, 60 mV; lower trace, 6 inV. N 3, and N 5 in the nymphal stages. An example of these responses is illustrated in Fig. 5. Ipsi-and contralateral stimulation of N 2 evoked a compound PSP after a delay of 25-30 ms (Fig. 5A, B). Both ipsi- and contralateral responses were observed in all the experiments conducted on nymphal stages A d - 5 and Ad-4. The number of preparations which revealed these inputs was reduced during stages A d - 5 , A d - 1 (Fig. 5C, D) and disappeared completely in the adult stage. The responses evoked by stimulation of N3 and N5 behaved in the same way as that described above for nerves 1 and 2, i.e. the average PSP amplitude and frequency of appearance declined during stages A d - 5 to A d - 1 and disap-

316 peared completely in the adult. Unlike the evoked synaptic potential described above, synaptic interactions between adjacent GINs persisted throughout the developmental stages and in the adults. These synaptic interactions in adult and nymphal preparations were described by us earlier4°-42. Fig. 6 shows an experiment in which 2 GINs within the same connective were impaled by microelectrodes. Intracellular stimulation of one GIN was followed by a PSP in the other GIN. The connection between the GINs was always reciprocal. These PSPs were characterized by their fast rise time (1 1.5 ms) and slow decay (60 100 ms). The frequency of appearance and average amplitude of the PSP were not altered during development.

Safetyfactor for impulse propagation across 7"3 In previous papers from our laboratories ~°.28-29,38,4°,we demonstrated the phenomenon of impulse propagation blockade at T3 during high-frequency stimulation. We have shown that in adult cockroaches, stimulation of the

GINs at frequencies above 50 Hz induces conduction block at T3. While passing through the ganglion the GIN diameter is reduced and several neurites course off from the main axon into the neuropile. This special structure creates an area of low safety factor for impulse propagation. We have suggested that during high frequency stimulation, potassium ions accttmulate around the axon causing depolarization and conductance increases. As a result, the action potential is reduced in amplitude and fails to propagate across the metathoracic ganglion. In the present report, the safety factor of the GIN at T3 in the nymphal and adult GINs was characterized by measuring 2 parameters. namely the duration and frequency of stimuli which bring about the conduction block, and the minimal action potential amplitude which can propagate across the ganglion. A low amplitude would represent a high safety factor, In order to measure the minimum amplitude of a spike, which is still capable of propagating through the gangliom we impaled a fiber at the caudal base A ~ I00

B

~ ~

I ' ~

o. 50

O I

A d -5 Ad-4 A d - 2 A d - I

Ad-5 Ad"4 Ad-2 Ad-I Ad

Ad

C

D

~lO0 rY

5ms Fig. 7. The minimal action potential amplitude which propagates across the GlN in the metathoracic ganglion of adult and larval preparations. Recording was made from the caudal edge ofT3 stimuli applied to the abdominal connectives between A3 and A4. A, B: adult preparation. A: control. B: after 5 s of stimulation at 20 Hz, C, D: nymphal preparation (Ad-5). C: control. D: after 3 rain of stimulation at 40 Hz. The axon was depolarized by about 20 mV. The back reflection of the action potential in the rostral edge of the ganglion is marked by an arrow. Note that the reflection potential appeared in the adult when the action potential amplitude was reduced by about 20 mV, while in the nymphal preparation the reflection appeared when the action potential was attenuated by almost 50 inV. (For further details, see text.)

30

} so

20 I0 Ad-5 Ad-4 M - E Ad'l

Ad

Ad~5 Ad-4 Ad- 2 Ad- I Ad

Fig. 8. Membrane properties and safety factor for propagation of a train of impulses at various developmental stages. A: action potential amplitude. B: resting potential, C: threshold for action potential initiation. D: the minimal amplitude of ascending action potential that propagate through T3. Resting potential, action potential, amplitude and threshold for spike initiation are not altered during the development. The safety factor for impulse propagation gradually decreases during the postembryonic development: each bar represents the standard error of 8 12 experiments.

317 of T3 and stimulated the abdominal connective. When the action potential in the anterior wide region of the axon is delayed (as a consequence of travelling from a narrow to a wide region), a decremental potential is seen on the falling phase of the spike recorded at the caudal edge of the ganglion (see Figs. 5 and 6 of Spira et al.4°, and Parnas et al. 29, respectively). With further delay of the anterior spike, the decrementaUy reflected potential is further delayed and, finally, when the spike fails to propagate into the anterior part of the fiber, the reflected potential disappears. Fig. 7 shows the results of such experiments. In these experiments the abdominal connective was stimulated by repetitive short trains. The intratrain interval was 3 ms and the trains were repeated at 20 Hz and 40 Hz in adult and immature animals, respectively. The records were made at high (upper trace) and low gain to allow clear recording of the reflected potential, as well as the spike amplitude. Failure of action potential propagation across T3 in the adult GIN was evoked after 5-10 s of stimulation, as indicated by the disappearance of the reflected potential (arrow in Fig. 7B, upper trace). At this time the spike amplitude was reduced from 115 mV to 85 mV. A much longer train of stimuli (3 min) had to be given to the nymphal GIN in order to produce the blockade; at that time the axon was depolarized by about 20 mV and the amplitude of the ascending action potential was reduced to 45 mV (Fig. 7C, D). As seen in Fig. 8D, the safety factor for impulse propagation across ganglion T3 is gradually reduced during the postembryonic developmental stages. The safety factor for impulse propagation along an axon can be theoretically influenced by the action potential amplitude, resting potential, threshold for spike initiation, and the geometry of the fiber. As shown in Fig. 8, the action potential (Fig. 8A), resting potential (Fig. 8B), and threshold (Fig. 8C) are not changed during postembryonic development. Therefore, it is most likely that the gradual decrease in safety factor for propagation of a train of impulses is due to changes in the geometry of the fiber at T3.

DISCUSSION The results demonstrate that the number of pathways from which synaptic potentials can be evoked onto the giant interneuron is markedly reduced during postembryonic development. In the early nymphal stages (Ad-5 and A d - 4 ) the GINs receive synaptic inputs from 8 sources: the contralateral thoracic connective (N1) ipsi- and contralateral nerve roots 2, 3, 5 and from GINs within the same connective (N7). The ipsi- and contralateral input from nerves 2, 3 and 5 could not be detected in the adults. In addition, the safety factor for propagation of trains of impulses gradually decreased during development. It has been well established that invertebrate and vertebrate neurons continue to grow during the postembryonic stages. This growth involves "elongation and arborization of dendrites, as well as changes in the diameters of different segments of the neuron 1.6,~7.Such spatial changes in the structure of a neuron may result in an increase in the electrotonic distance between a given synaptic input and other segments of the neuron. Therefore, one possible mechanism that could explain the disappearance of the synaptic inputs from nerves 2, 3, and 5 in adult cockroaches would be elongation of the giant axon within T3 or elongation of the neurites on which these presynaptic pathways terminate. Such an increase in the electrotonic distance could also be the result of a reduction in the specific membrane resistance of the neurites or the axon. In order to determine an increase in the electrotonic distance during postembryonic development, thorough characterization of the morphology and passive membrane properties of the GIN within T3 is required. However, some observations indicate that this is not the mechanism which underlies the apparent elimination of synaptic inputs: (a) during postembryonic development from nymphal stage A d - 4 to Ad-1, the length of the metathoracic ganglion and segment of the GIN within it is increased by a factor of 1.7 (from 0.72 _ 0.03, 1.13 ± 0.1; Yarom and Spira43). On the other hand, in cases where spontaneous or evoked synaptic potentials had been recorded

318 from adult preparations, the PSP rise time was simil'a~f°tothat recorded from immature animals, i:e. functional synaptic terminals in adults and immature animals are located at the same electrotonic distance from the recording site: (b) the observation that the reciprocal synaptic connections between adjacent GINs are not altered during postembryonic development indicate that the relative electrotonic location of this synapse does not change; (c) even though the major changes in synaptic inputs take place during metamorphosis (stage Ad--1 to Ad), no significant change in the dimensions of the ganglion is noticeable during this period. However, we cannot rule out the possibility that specific and restricted changes in membrane properties or morphological alterations of certain neurites cause an increase in the electrotonic distance between some synapses and the recording site. For this reason, we studied the morphology and membrane properties of the GINs within ganglion T3 and compared them in adult and immature preparations. The results obtained from this study 43 revealed that neither morphological changes nor changes in membrane properties can account for the functional disappearance of the synaptic inputs. An alternative mechanism that could account for the 'disappearance' of the synaptic inputs during the transition between the last nymphal stage and adult would be a decrease in the efficacy of transmission along the afferent pathways. Such decrease could be due to several mechanisms: (a) decrease in synaptic transmisson efficacy, due to a decrease in postsynaptic sensitivity or decreased quantal content of neurons comprising the pathway14.24.25.44: (b) development of inhibitory connections onto the neurons within the afferent pathway; (c) elimination of synaptic terminals by regression or degeneration3-5.s~lc2°.21-27.35~ (d) failure of propagation along afferent pathways due to structural or physiological changes of interneurons within the afferent pathway.

Safeo; .[actor ,~or impuLs'e propagation a/on~ the GINs in T3 The safety factor along the metathoraeic segment of the adult GINs is sufficient for the propagation of a single impulse but insufficient to allow propagation of impulses delivered at high frequencyJ0,2~.3~.4~. In the present investigation we found that in the nymphal stages the safety factor for spike propagation along T3 is high for both single spikes and trains of impulses. In many of our experiments on nymphal stages A d - 5 , A d - 4 , we were unable to produce conduction block even by stimulation at high frequencies for long periods of time (100 Hz, 3 rain). In cases where conduction block was observed, the propagating spike had to be reduced by 45% for conduction block to occur. The threshold for spike initiation in the GINs of nymphs and adults is identical (Fig. 8). This suggests that most of the difference in the safety factor of adult and larval GINs is probably due to a difference in the geometry. Nevertheless, it is possible that different membrane properties are responsible for a change in safety factor. The results described in this article show that there are clear and defined differences between GINs from adult and immature animals. The functional disappearance of afferent pathways to the GIN in T3 does not reflect a general phenomenon of reduction in synaptic transmission efficacy in the cockroach CNS. (For example, the cercal nerve GIN transmission is functional and effective in both nymphs and adults.) Thus, the observed changes point out that the GINs and afferent neurons to them undergo specific changes which express a certain development process. The mechanisms underlying these changes are treated in the following artiCI~ 3 .

ACKNOWLEDGEMENTS

This work was supported by a grant from the United States-Israel Binational Science Foundation (BSF) Jerusalem, Israel.

319 REFERENCES 1 Altman, J. S. and Tyrer, N. M., Insect flight as a system for the study of development of neuronal connections. In L. Barton Browe (Ed.), Experimental analysis of insect behavior, Springer, New York, 1974. 2 Anderson, H., Edwards, J. S. and Palka, J., Developmental neurobiology of invertebrates, Ann. Rev. Neurosci., 3 (1980) 97-139. 3 Aguayo, A. J., Peyronnard, J. M., Terry, L. C., Romine, J. S. and Bray, G. M., Neonatal neuronal loss in rat superior cervical ganglia: retrograde effects on developing preganglionic axons and Schwann cells, J. Neuroeytol., 5 (1976) 137 155. 4 Bennett, M. R. and Pettigrew, A. G.. The formation of synapses in striated muscles during development, J. Physiol. (Lond.), 241 (1974)515-541. 5 Bennett, M. R. and Pettigrew, A. G.. The formation of neuromuscular synapses, Cold. Spring Harb. Syrup. quant. Biol.. 40 (1976) 409-424. 6 Bently, D. R., Postembryonic development of insect motor system. In D. Young (Ed.), Developmental Neurobiologv ofA rthropods. Cambridge University Press, 1973. 7 Berry, M. and Hollingworth. T., Growth of dendrites of the Purkinje cells in the rat cerebellum, J. Anat., 111 (1972) 491-492. 8 Brown, M. D., Jansen, J. K. S. and Van Essen, D., Polyneuronal innervation of skeletal muscle in newborn rats and its elimination during maturation, J. Phvsiol. (l, ond.). 261 (1976) 387-422. 9 Callec, J. J., Synaptic transmission in the central nervous system of insects. In J. E. Treherne (Ed.), Insect Neurobiology, North-Holland, Amsterdam, 1974, pp. 120-185. 10 Castel, M., Spira, M. E., Parnas, I. and Yarom, Y., UItrastructure of the region of low safety factor in an inhomogeneous giant axon of the cockroach, J. Neurophysiol., 39 (1976) 90(~908. I 1 Crepel, F., Marianai, J. and Delyhaye-Bouchaud, N., Evidence for a multiple innervation of Purkinje cells by climbing fibres in the immature rat cerebellum, J. Neurobiol.. 7 (1976) 567-578. 12 Dennis, M. J. and Yip, J. W.. Formation and elimination of foreign synapses on adult salamander muscle, J. Phvsiol. (Lond.). 274 (1978) 299-310. 13 Farley, R. D. and Milburn. N. W., Structure and function of the giant fiber system in the cockroach Periplaneta americana, J. Insect Physiol.. 15 (1969) 457-476. 14 Grinnel, A. D., Rheuben, M. B. and Letinsky, M. S., Mutual repression of synaptic efficacy by pairs of foreign nerves innervating frog skeletal muscle, Nature (Lond.), 265 (1977) 368 370. 15 Guthrie, D. M. and Tindall, A. R., The Biology of the Cockroach, Edward Arnold, London, 1968. 16 Gymar, A. and Edwards, J. S., The development of the insect nervous system I. An analysis of postembryonic growth in the terminal ganglion ofAcheta domesticus, J. Morph.. 123 (1976) 191- 197. 17 Hubel, D. H. and Wiesel, T. N., Receptive fields of cells in striate cortex of very young, visually inexperienced kittens, J. Neurophvsiol.. 26 (1963) 994~ 1017. 18 Huiser, P., Kuno, M. and Miyate, Y., Differentiation of motoneurons and skeletal muscles in kittens, J. Physiol. ¢Lond.), 252(1975)465 479. 19 Jacobson, M., Developmental Neurobiologv, Plenum

Press, NY, 1978. 20 Lichtman, J. W., The reorganization of synaptic connexions in the rat submandibular ganglion during post-natal development, J. Physiol. (Lond.), 273 (1977) 155~ 177. 21 Lichtman, J. W., On the predominantly single innervation of submandibular ganglion cells in the rat, J. Physiol. (Lond.), 302 (1980) 121-130. 22 Lichtman, J. W. and Purves, D., The elimination of redundant preganglionic innervation to hamster sympathetic ganglion cells in early post-natal life, J. Physiol. (Lond.), 301 (1980)213-228. 23 Llinas, R. and Sugimori, M., Calcium conductances in Purkinje cell dendrites: their role in development and integration. In M. Cuenod, G. W. Kreutzberg and F. E. Bloom (Eds.), Development and Chemical Specificity of Neurons, Progress in Brain Research, Vol. 51, Elsevier, Amsterdam, 1979, pp. 323-334. 24 Marotte, L. R. and Mark, R. F., The mechanism of selective re-innervation of fish eye muscles. I. Evidence from muscle function during recovery, Brain Res., 19 (1970) 41-62. 25 Mark, R. F., Marotte, L. R. and Mart, P. E., The mechanism of selective re-innervation offish eye muscles. IV. Identification of repressed synapses, Brain Res., 46 (1972) 149- 157. 26 Murphy, R. K. and Levine, R. B., Mechanisms responsible for changes observed in response properties of partially deafferented insect interneurons, J. NeurophysioL, 43 (1980) 367-382. 27 Oppenheim, R. W. and Majors-Willard, C. C., Neuronal cell death in the brachial spinal cord is unrelated to the loss of polyneuronal innervation in wing muscle, Brain Res., 154(1978) 148 152. 28 Parnas, I., Spira, M. E., Werman, R. and Bergman, F., Non-homogeneous conduction in giant axons of the nerve cord of Periplaneta americana, J. exp. Biol., 50 (1969) 615-627. 29 Parnas, I., Hochstein, S. and Parnas, H., Theoretical analysis of parameters leading to frequency modulation along an inhomogeneous axon, J, Neurophysiol., 39 (1976) 909-923. 30 Parnavelas, J. E., Mounty, E. Z., Bradford, R. and Liberman, A. R., The postnatal development of neurons in the dorsal lateral geniculate nucleus of the rat: a Golgi study, J. comp. Neurol., 171 (1977)481-499. 31 Purpura, D. P., Comparative physiology of dendrites. In G. C. Quarton, T. Melnechik and F. O. Schmitt, (Eds.), The Neurosciences, The Rockefeller University Press, 1967 372-392. 32 Purves, D. and Lichtman, J. W., Elimination of synapses in the developing nervous system, Science, 210 (1980) 153 157. 33 Rakic, P., Prenatal genesis of connections subserving ocular dominance in the rhesus monkey, Nature (Lond.), 261 (1976) 467- 471. 34 Rakic, P., Prenatal development of the visual system in the rhesus monkey, PhiL Trans., 13 278 (1977) 245-260. 35 Redfern, P. A., Neuromuscular transmission in newborn rats, J. Physiol. (Lond.), 209 (1970) 701-709. 36 Sbrenna, G., Postembryonic growth of the ventral nerve cord in Shistocerca gregaria forsk, Boll Zool., 38 (1971) 49- 74. 37 Spitzer, N. C., Ion channels in development, Ann. Rev. Neurosci., 2 (1979) 363-397.

320 38 Spira, M. E., Parnas, 1. and Bergman, F., Organization ~H" the giant axons of the cockroach Periplaneta americana, J. exp. Biol,. 50(1969)615 627. 39 Spira, M. E.. Parnas, I. and Bergman, F., Histological and electrophysiological studies on the giant axons of the cockroach Perlplaneta americana, J. exp. BioL. 50 (1969) 629 634. 40 Spira, M. E., Yarom. Y. and Parnas. I., Modulation of spike frequency by regions of special axonal geomet U and by synaptic inputs, J. Neurophvsiol., 239 (1976) 882 899. 41 Truman, J. W. and Reiss, S. E., Dendritic reorganization of an identified motoneuron during metamorphosis of

the tobacco hornworm moil1, ,S'