1. The rapidly exchanging 24Na fraction was found to account for approximately a third of the sodium contained in desheathed terminal abdominal ganglia which had been previously made radioactive by the injection of labelled ions into the haemolymph.

  2. This rapidly exchanging sodium was associated with the extracellular spaces which, with the aid of 14C-inulin, were shown to contain 18·2 % of the ganglion water. Using this figure it was possible to calculate that the extracellular sodium concentration exceeded that in the haemolymph by a factor of 1·8 and was 2·5 times greater than that in the cellular fraction of the ganglion.

  3. Experiments using 22Na, 42K, 45Ca, 36Cl and 3H0H showed that the concentrations of the ions in the rapidly exchanging extracellular fractions of isolated abdominal nerve cords were different from those of the external medium. The three cations were maintained at considerably higher concentrations in the extracellular spaces, chloride ions being present at a much lower concentration than in the external solution.

  4. It was concluded that the ions were distributed between the extracellular spaces and the external solution according to a Donnan equilibrium. The evidence would seem to indicate that a significant proportion of the free anion groups in the extracellular spaces may be those of protein molecules in solution. An appreciable portion of the cations may also, however, be associated with the free anion groups of structural elements such as collagen and mucopolysaccharides.

  5. It is suggested that the relatively rapid depolarization observed in desheathed preparations, as compared with intact nerves, in conditions of high external potassium concentration may result from the effects of the disruption of the Donnan equilibrium as well as to any properties of the nerve sheath as a diffusion barrier.

  6. It was shown that because of their positive charges the diffusion of the cations in the extracellular spaces occurred more slowly than would be expected on the basis of their free-diffusion constants. The efflux of 14C-inulin also occurred more slowly than would be expected from its free diffusion, an effect probably caused by the passage of the polysaccharide molecules through relatively small spaces in the extracellular system.

The experiments of Hoyle (1953) on locust peripheral nerve and of Twarog & Roeder (1956) on the abdominal nerve cord of the cockroach showed that a block to conduction developed rapidly in solutions of high potassium concentration if the nerve sheath was removed or if the solution was injected beneath the sheath. On the basis of these results it was concluded that the nerve sheath acted as a barrier restricting the entry of ions and acetylcholine molecules into the nervous systems of these insects. The experiments of Yamasaki & Narahashi (1959, 1960) on the abdominal nerve cord of the cockroach were also interpreted as demonstrating the presence of an extremely impermeable diffusion barrier in the sheath surrounding the central nervous system.

The insect nervous system, unlike that of vertebrates, is an avascular organ and, apart from the oxygen carried by the tracheae, the only access for ions and molecules is via the fibrous and cellular nerve sheath. Thus despite its impermeability the nerve sheath must nevertheless be specialized, as Wigglesworth (1959, 1960) has pointed out, to allow the necessary exchanges of nutrient and excretory substances to take place. The exchanges and metabolism of 14C-labelled sugars between the haemolymph and the abdominal nerve cord were, in fact, demonstrated to take place relatively rapidly in Periplaneta (Treheme, 1960). A subsequent investigation also showed that sodium and potassium ions exchanged rather rapidly in the abdominal nerve cord of this insect (Treheme, 1961a). In experiments using 24Na it was found possible to demonstrate an active extrusion of sodium ions from intact isolated abdominal nerve cords and it was suggested that this might represent evidence for the part played by the cellular layer of the perilemma in regulating the exchanges of ions between the haemolymph and the central nervous system (Treherne, 1961b). This suggestion was made unlikely by the subsequent discovery that removal of substantial portions of the nerve sheath in an isolated ganglion preparation did not significantly affect the rate of efflux of 24Na from the system (Treherne, 1961c, d). The effluxes measured in these experiments were, in fact, found to be those from the underlying cellular components of the central nervous system (Treherne, 1961 e). The exchanges taking place between the extracellular spaces and the haemolymph occurred so rapidly that they could not be detected in the earlier experiments. These rapid movements of the labelled ions across the perilemma were found to take place passively and appeared to involve only diffusion processes.

It is clear that a paradoxical situation has arisen as a result of these recent investigations on the central nervous system of Periplaneta. It is difficult to reconcile the experiments of Hoyle (1953) and Twarog & Roeder (1956), showing an apparent effective diffusion barrier around the nervous system, with the later results demonstrating relatively rapid passive exchanges of ions taking place across the perilemma between the extracellular spaces and the haemolymph. The experiments to be described in this paper represent an attempt to resolve this apparent paradox.

The experiments to be described in this paper were largely concerned with the efflux of various radio-isotopes from the central nervous system of adult male Periplaneta americana L. The abdominal nerve cords were made radioactive either by the injection of 10–100 μl. of a solution containing the radio-isotopes into the haemolymph or by soaking the isolated organs in a physiological solution containing the radioactive materials. As has been previously described the isolated abdominal nerve cords and terminal abdominal ganglia were ligatured with threads pulled from 15 denier nylon stockings (Treherne, 1961 d, e). In a few experiments isolated terminal abdominal ganglia were completely or partially desheathed using finely ground watchmaker’s forceps as described by Twarog & Roeder (1956).

The physiological solution used in this investigation was based on that of Asperen & Esch (1956) and was similar to that employed in some earlier investigations on the physiology of the cockroach central nervous system (Treherne, 1961a, b). The basic solution had the following ionic composition: Na, 157·0 mw./I.; K, 12·3 mM./l.; Ca, 4·5 mM./l. ; Mg, 4·0 mM./l. ; Cl, 184·1 mM./l. ; HCO3,2·1 mM./l. ; HgPO4, 0·1 mM./l. In some experiments the potassium concentration was altered to 70·0 or 157·0 mM./l., the sodium concentration being proportionally reduced to maintain isotonicity. The solution also contained the following organic compounds: trehalose, 36·9 mM./l.; glucose, 2·2 mM./l.; glutamic acid, 35·0 mM./l.; glutamine, 30·0 mM./l.; glycine, 30·0 mM./l.

The various radio-isotopes used in this investigation are listed in Table 1, together with details of their approximate specific activities and chemical forms. The isotopes with relatively strong β or γ radiations were assayed using a GM tube (Mullard MX 123) linked to a scaler unit (Panax 100c); isotopes with only soft β -radiations (38C1,14C and 3H) were counted using a dioxan liquid phosphor (Naphthalene, D.P.O. and P.O.P.O.P.) in a Panax scintillation counter.

Table 1.

The chemical forms and specific activities of the various radio-isotopes used in these experiments

The chemical forms and specific activities of the various radio-isotopes used in these experiments
The chemical forms and specific activities of the various radio-isotopes used in these experiments

To study the efflux of the labelled ions and molecules from the central nervous system the ligatured abdominal nerve cords or ganglia were carefully dried on filterpaper and then washed for varying periods in successive 0·2 ml. amounts of inactive physiological solution or in isotonic dextrose (0·48 M.). The amounts of radio-isotope remaining in the nerve cords after varying washing times were estimated from the measured activity of the washings.

The abdominal nerve cords were weighed on a 5·0 mg. torsion balance with an accuracy of 0·01 mg. Single ganglia were weighed on a quartz fibre balance with an accuracy of approximately 0·002 mg. Haemolymph samples, of between 2·0 and 10·0 μl., were taken from the base of a prothoracic leg with silicone-lined micropipettes. The concentration of sodium in the haemolymph was measured using an EEL flame photometer.

The distribution of labelled sodium and inulin in the terminal abdominal ganglion

Fig. 1 illustrates the rate of loss of sodium ions from a ligatured terminal abdominal ganglion isolated 4–5 hr. after the injection of 24Na into the haemolymph. In each case the ganglia were desheathed prior to the measurements on the rates of efflux in order to eliminate the effects of any surface radioactivity associated with the nerve sheath. As was found in a previous investigation (Treherne, 1961e) the emergence of the labelled ions from the ganglia could be resolved into a two-stage process, an initial rapid component eventually giving way to a second slow phase. In Table 2 the proportions of the radioactivity contained in these two phases are expressed in terms of unit weight of ganglion tissue. These results showed that the ions in the rapidly exchanging fraction accounted for approximately a third of the sodium associated with the tissues of the terminal abdominal ganglion.

Table 2.

The distribution of 24Na in the slowly and rapidly exchanging fractions of desheathed terminal abdominal ganglia

The distribution of 24Na in the slowly and rapidly exchanging fractions of desheathed terminal abdominal ganglia
The distribution of 24Na in the slowly and rapidly exchanging fractions of desheathed terminal abdominal ganglia
Fig. 1.

The escape of 24Na from a ligatured desheathed terminal abdominal ganglion which was previously made radioactive by the injection of the isotope into the haemolymph (closed circles). The fast component of the main curve (open circles) was obtained by subtraction of the straight line extrapolated to xero time.

Fig. 1.

The escape of 24Na from a ligatured desheathed terminal abdominal ganglion which was previously made radioactive by the injection of the isotope into the haemolymph (closed circles). The fast component of the main curve (open circles) was obtained by subtraction of the straight line extrapolated to xero time.

To interpret the data given in Fig. 1 and Table 2 it is essential to know something of the extent of any extracellular spaces in the central nervous system of this insect. To discover this the total water content of the desheathed terminal abdominal ganglion was measured (Table 3) together with the uptake of 14C-labelled inulin by this structure. In these experiments the inulin content of the nervous tissue was measured following the injection of the radioactive polysaccharide into the haemolymph. After 6 hr. the terminal abdominal ganglion was removed, desheathed and the radioactivity extracted from the remaining tissues. Table 4 shows the apparent extent of the tissue water occupied by the 14C-inulin, estimated from a comparison of the radioactivity of the haemolymph and the desheathed ganglion. The mean figure of 18·2% for the inulin space in the terminal ganglion is greater than that of 10·1 % obtained for washed whole abdominal nerve cords in a previous investigation (Treherne, 1961e). As was pointed out in the earlier paper the lower value must be regarded as a minimum estimate of any extracellular space for some radioactivity must have leaked from within the nerve cord during the washing procedure.

Table 3.

The water content of desheathed terminal abdominal ganglia

The water content of desheathed terminal abdominal ganglia
The water content of desheathed terminal abdominal ganglia
Table 4.

The inulin space of the terminal abdominal ganglion

The inulin space of the terminal abdominal ganglion
The inulin space of the terminal abdominal ganglion
In earlier work evidence was advanced for regarding the rapidly exchanging sodium as the ions contained in the extracellular spaces, while the slowly exchanging fraction was shown to be most probably associated with the cellular components of the central nervous system (Treherne, 1961e). With the above information on the content of the two sodium fractions, together with the measured volume of the inulin space, it is possible to estimate the concentrations of the ions in the cells and extracellular spaces of the last abdominal ganglion. Table 5 shows the results obtained from combining the data contained in Tables 2, 3 and 4 compared with the measured sodium concentrations of the haemolymph. These estimated figures are characterized by some individual variation but clearly demonstrate that in all cases the apparent sodium concentration in the extracellular space exceeded that in the haemolymph, the mean value for
formula
The sodium concentration in the extracellular space also greatly exceeded that in the cells of the ganglion, the mean value of
formula
Table 5.

The calculated concentration of sodium ions in the cells and the postulated extracellular spaces of the terminal abdominal ganglion based on the data given in Tables 2–4

The calculated concentration of sodium ions in the cells and the postulated extracellular spaces of the terminal abdominal ganglion based on the data given in Tables 2–4
The calculated concentration of sodium ions in the cells and the postulated extracellular spaces of the terminal abdominal ganglion based on the data given in Tables 2–4

The exchanges and distribution of water and monovalent ions in the whole abdominal nerve cord

The relatively low specific activities of 42K and 36C1 at present available made it impossible to extend the previous studies on the distribution of sodium ions in the terminal abdominal ganglion to these other monovalent ions. To compare the distribution of these ions with those of sodium the effluxes from intact isolated whole abdominal nerve cords were studied.

Fig. 2 shows the rate of loss of 24Na from an abdominal nerve cord which was loaded in vitro by soaking the ligatured preparation in the radioactive physiological solution. As with the single ganglion the efflux could be resolved into a two-stage process with slowly and rapidly exchanging sodium fractions.

Fig. 2.

The escape of 24Na from a whole ligatured abdominal nerve cord (closed circles) which had previously been made radioactive by soaking for 20 min. in a physiological solution containing the radio-isotope. The fast component of the main curve (open circles) obtained by subtraction of the straight line extrapolated to zero time.

Fig. 2.

The escape of 24Na from a whole ligatured abdominal nerve cord (closed circles) which had previously been made radioactive by soaking for 20 min. in a physiological solution containing the radio-isotope. The fast component of the main curve (open circles) obtained by subtraction of the straight line extrapolated to zero time.

In a similar way to the sodium loss the efflux of 42K and 36C1 could be resolved into two-stage processes with equivalent rapid and slowly exchanging fractions (Figs. 3, 4). In both cases, as was shown for the diffusion from the extracellular space in cylindrical muscle (Hill, 1928), the rate of efflux of the rapidly exchanging fractions appeared to be initially complex, eventually assuming a simple exponential form.

Fig. 3.

Results of a typical experiment showing the escape of the slowly (closed circles) and rapidly exchanging fractions (open circles) of 42K from an isolated abdominal nerve cord.

Fig. 3.

Results of a typical experiment showing the escape of the slowly (closed circles) and rapidly exchanging fractions (open circles) of 42K from an isolated abdominal nerve cord.

Fig. 4.

The fast (open circles) and slow (closed circles) components obtained for the efflux of 36CI from an isolated ligatured abdominal nerve cord.

Fig. 4.

The fast (open circles) and slow (closed circles) components obtained for the efflux of 36CI from an isolated ligatured abdominal nerve cord.

The increase in concentration of the three monovalent ions in the rapidly exchanging fractions of the abdominal nerve cord on soaking in vitro in the physiological solution is illustrated in Fig. 5. These data were derived from experiments on the whole nerve cord in which there was some surface radioactivity associated with the nerve sheath. Thus, for example, the concentration of the rapidly exchanging sodium, which has already been shown to exceed that of the haemolymph, must be regarded as a minimum estimate of the level in the abdominal nerve cord. Fig. 5 shows that the concentration of the sodium quickly rose to a steady level of about 44·9 mM./kg. tissue, while that of the chloride averaged about 16·9 mM./kg. tissue. These facts clearly suggest that the concentration of the rapidly exchanging chloride in the nerve cord was much lower than that of the sodium. The ratio of these ions in the external solution was
formula
The same ratio in the rapidly exchanging fraction of the abdominal nerve cord was
formula
The distribution of the ions in the rapidly exchanging fraction was also determined at different concentrations of sodium and potassium in the external solution. The results from these experiments are summarized in Table 6. The chloride level was relatively low at all three concentrations despite the high external concentration of 184·1 mM./l. The distribution of sodium and potassium in the rapidly exchanging fraction appeared to be directly related to the concentrations in the external solution. The ratios of sodium and potassium to the chloride ions in the rapidly exchanging fraction were of the same order in the three solutions, being 2·8,2·1 and 2·3, respectively.
Fig. 5.

The accumulation of the rapidly exchanging ion fractiona in ligatured isolated abdominal nerve cords. The vertical lines illustrate the extent of twice the standard error of the mean.

Fig. 5.

The accumulation of the rapidly exchanging ion fractiona in ligatured isolated abdominal nerve cords. The vertical lines illustrate the extent of twice the standard error of the mean.

Table 6.

The concentration of sodium, potassium and chloride ions, per unit tissue weight, in the rapidly exchanging fraction of the abdominal nerve cord at different external ionic concentrations

The concentration of sodium, potassium and chloride ions, per unit tissue weight, in the rapidly exchanging fraction of the abdominal nerve cord at different external ionic concentrations
The concentration of sodium, potassium and chloride ions, per unit tissue weight, in the rapidly exchanging fraction of the abdominal nerve cord at different external ionic concentrations

The distribution of water in the abdominal nerve cord was estimated by studying the rates of efflux of tritiated water from this organ. The nerve cords were made radioactive either by injection of 3HOH into the haemolymph or by soaking isolated preparations in the physiological solution containing tritiated water. A typical result obtained in these experiments is illustrated in Fig. 6. As with the monovalent ions the rate of loss of tritiated water from the nerve cord appeared to approximate to a two-stage process. The proportion of tritiated water contained in the fast fraction is summarized in Table 7, where it is expressed as a percentage of the total nerve cord water. The mean figure of 21·6% is reasonably close to that of 18·2% obtained for the inulin space in the desheathed ganglion.

Table 7.

The percentage of tritiated water contained in the rapidly exchanging fraction of the abdominal nerve cord

The percentage of tritiated water contained in the rapidly exchanging fraction of the abdominal nerve cord
The percentage of tritiated water contained in the rapidly exchanging fraction of the abdominal nerve cord
Fig. 6.

The efflux of tritiated water from an isolated abdominal nerve cord 20 min. after the injection of the radioisotope into the haemolymph. The rapidly exchanging fraction (open circles) was calculated by extrapolation of the main curve (closed circles) to zero time.

Fig. 6.

The efflux of tritiated water from an isolated abdominal nerve cord 20 min. after the injection of the radioisotope into the haemolymph. The rapidly exchanging fraction (open circles) was calculated by extrapolation of the main curve (closed circles) to zero time.

Using the figure obtained for the rapidly exchanging water molecules in Table 7 it is possible to convert the data illustrated in Fig. 5 into the appropriate units of concentration for unit volume of water. On the basis of the evidence already presented and of that contained in a previous paper (Treherne, 1961e) the water and ions contained in the rapidly exchanging fraction are identified with the extracellular spaces of the central nervous system. As with experiments on the isolated abdominal ganglia the apparent concentration of the ions in the extracellular spaces differed from those in the haemolymph (Table 8). At the concentrations equivalent to that of the normal haemolymph both sodium and potassium exceeded that of the external medium while the chloride level was considerably below that outside the nerve cord.

Table 8.

The concentration of the monovalent ions in the postulated extracellular spaces of the abdominal nerve cord estimated from the data given in Fig. 5 and Tables 6 and 7

The concentration of the monovalent ions in the postulated extracellular spaces of the abdominal nerve cord estimated from the data given in Fig. 5 and Tables 6 and 7
The concentration of the monovalent ions in the postulated extracellular spaces of the abdominal nerve cord estimated from the data given in Fig. 5 and Tables 6 and 7
The estimated values contained in Table 8 are not inconsistent with the hypothesis that the monovalent ions distribute themselves between the extracellular spaces and the external solution, according to a Donnan equilibrium, where the following relationship should hold :
formula
The actual ratios obtained from these experiments were as follows :
formula
And
formula
The excellent agreement obtained between these two ratios accords with the hypothesis of the existence of a Dorman equilibrium between the haemolymph and the extracellular spaces of the central nervous system of this insect.

The distribution and exchanges of calcium in the abdominal nerve cord

The distribution of the exchangeable calcium in the central nervous system was determined by studying the efflux of 45Ca from abdominal nerve cords which had been soaked in vitro for varying periods in a physiological solution containing radio-calcium. A typical efflux curve for 45Ca is illustrated in Fig. 7 and shows its apparent two-stage process also characteristic of the monovalent ions.

Fig. 7.

The escape of 45Ca from an isolated abdominal nerve cord made radioactive by soaking for 20 min. in a radioactive solution. The open circles show the rapidly exchanging fraction obtained by extrapolation of the main curve to zero time.

Fig. 7.

The escape of 45Ca from an isolated abdominal nerve cord made radioactive by soaking for 20 min. in a radioactive solution. The open circles show the rapidly exchanging fraction obtained by extrapolation of the main curve to zero time.

The rate of accumulation of 45Ca within the abdominal nerve cords is illustrated in Fig. 8. These data, which are expressed in terms of the amount of radio-calcium per unit weight of nerve cord tissue, show the relatively rapid equilibrium achieved by the rapidly exchanging calcium fraction. In Fig. 9 these results have been converted to concentration in terms of the volume of nerve cord water on the assumption of a 21·6% extracellular space postulated earlier. The outstanding feature of these estimated values is the high concentration of calcium in the extracellular space (17·6 mM./l.) relative to that in the outside medium (4·5 mM./l.). The 45Ca in the cellular component of the abdominal nerve cord reached a level of 14·7 mM./l. at the end of the experiment. It is not certain whether this fraction had come into effective equilibrium with the extracellular 45Ca so that this figure must be regarded as a minimum estimate of the content of the exchangeable cellular calcium.

Fig. 8.

The total accumulation of 45Ca (closed circles) compared with rate of appearance of the radio-isotope in the rapidly exchanging fraction (open circles) in isolated abdominal nerve cords. The vertical lines indicate the extent of twice the standard error of the mean.

Fig. 8.

The total accumulation of 45Ca (closed circles) compared with rate of appearance of the radio-isotope in the rapidly exchanging fraction (open circles) in isolated abdominal nerve cords. The vertical lines indicate the extent of twice the standard error of the mean.

Fig. 9.

The accumulation of 45Ca in the cells (open circles) and extracellular fluid (closed circles) calculated from the data illustrated in Fig. 8. The broken line illustrates the calcium concentration in the external solution.

Fig. 9.

The accumulation of 45Ca in the cells (open circles) and extracellular fluid (closed circles) calculated from the data illustrated in Fig. 8. The broken line illustrates the calcium concentration in the external solution.

In order to determine whether the calcium ions are distributed according to a Donnan equilibrium the values obtained in these experiments with 45Ca may be compared with those obtained with labelled sodium ions. The potential difference which would be developed between the extracellular fluid and the external medium can be given for the steady-state distribution of sodium as follows :
formula
where E is the potential difference, R is the gas constant, T is the absolute temperature’ and F is the Faraday. The value of 14·9 mV. for E can be inserted in the following equation for the distribution of calcium ions :
formula
where z is the valency of calcium. The calculated value of 3·5 is in reasonable agreement with the measured value of 3·8 for the ratios of the external and extracellular calcium.

The distribution of sodium and calcium ions in desheathed ganglia loaded in vitro

The distributions of 22Na and 45Ca in ganglia which were desheathed before soaking in radioactive solutions were determined in an attempt to throw some light on the nature of the changes brought about by the desheathing procedure. It was found that the efflux of both these radioactive ions from the desheathed preparations approximated to two-stage processes from which it was possible to estimate the concentrations of the rapidly exchanging fractions (Figs. 10, 11). The concentrations of these ion fractions, expressed in terms of unit tissue weight, are summarized in Table 9 and show, rather unexpectedly, that there was an apparent increase in concentration resulting from the desheathing procedure. The concentration of the rapidly exchanging sodium in the intact ganglia which were desheathed after soaking averaged 34·0 mM./kg. tissue, as compared with the present mean value of 793 mM./kg. tissue for ganglia desheathed before exposure to the radioactive solutions.

Table 9.

The concentration per unit tissue weight, of 22Na and 45Ca in the rapidly exchanging fraction of terminal abdominal ganglia which was desheathed before soaking in the radioactive solutions

The concentration per unit tissue weight, of 22Na and 45Ca in the rapidly exchanging fraction of terminal abdominal ganglia which was desheathed before soaking in the radioactive solutions
The concentration per unit tissue weight, of 22Na and 45Ca in the rapidly exchanging fraction of terminal abdominal ganglia which was desheathed before soaking in the radioactive solutions
Fig. 10.

The escape of 45Na from a terminal abdominal ganglion which was desheathed before soaking for 20 min. in a solution containing the radio-isotope. The fast component (open circles) was obtained by extrapolation of the main curve (closed circles) to zero time.

Fig. 10.

The escape of 45Na from a terminal abdominal ganglion which was desheathed before soaking for 20 min. in a solution containing the radio-isotope. The fast component (open circles) was obtained by extrapolation of the main curve (closed circles) to zero time.

Fig. 11.

The escape of 22Ca from a terminal abdominal ganglion which was desheathed before being made radioactive by soaking for 20 min. in a solution containing the radio-isotope. The fast and slow components are represented by open and closed circles respectively.

Fig. 11.

The escape of 22Ca from a terminal abdominal ganglion which was desheathed before being made radioactive by soaking for 20 min. in a solution containing the radio-isotope. The fast and slow components are represented by open and closed circles respectively.

Measurements were also made of the extent of the extracellular space in desheathed ganglia. Terminal abdominal ganglia were desheathed and were soaked in the physiological solution containing 1·5 % 14C-labelled inulin. From the relative activities in the tissue and the external medium it was found that the extracellular space had greatly increased following desheathing, the mean value being 0·494 l-/kgtissue (Table 10). Insertion of this figure in the calculation of the concentrations of the ions in the extracellular spaces, using the data contained in Table 9, resulted in the figures summarized in Table 11. These calculated values apparently show that the concentrations of both the 22Na and 45Ca had fallen to levels similar to those in the external solutions.

Table 10.

The volume of the extracellular space of desheathed ganglia as measured with 14C-inulin

The volume of the extracellular space of desheathed ganglia as measured with 14C-inulin
The volume of the extracellular space of desheathed ganglia as measured with 14C-inulin
Table 11.

The concentrations of 22Na and 45Ca in the extracellular spaces of ganglia desheathed prior to soaking in radioactive solutions. The calculated values are based on the data given in Tables 9 and 10

The concentrations of 22Na and 45Ca in the extracellular spaces of ganglia desheathed prior to soaking in radioactive solutions. The calculated values are based on the data given in Tables 9 and 10
The concentrations of 22Na and 45Ca in the extracellular spaces of ganglia desheathed prior to soaking in radioactive solutions. The calculated values are based on the data given in Tables 9 and 10

The movements of labelled ions and molecules in the extracellular spaces of the abdominal nerve cord

Besides the information on the distribution of ions and water, the data outlined above also throw some light on the nature of the ionic and molecular movements taking place within the extracellular spaces of the central nervous system of this insect. In addition to the results already described some experiments were also carried out on the efflux of 14C-labelled sucrose molecules from isolated abdominal nerve cords. The nerve cords were made radioactive by soaking them for 20–30 min. in a physiological solution in which the organic compounds were replaced by 136·9 mM./l. 14C-sucrose. A typical rate of loss curve obtained in these experiments is illustrated in Fig. 12. In this experiment the emergence of the labelled molecules assumed an approximately exponential form with a half-time (t0.5) of 280 sec. This value is higher than that of 214 sec. for 14C-inulin obtained in a previous investigation in which the efflux of the labelled polysaccharide molecules was confined to that from the extracellular spaces (Treherne, 1961). It seems reasonable to suppose, therefore, that the slow exponential decline in radioactivity for the 14C-sucrose efflux represents the efflux from some cellular components of the central nervous system. On this assumption it is thus permissible to derive an initial rapid phase of sucrose efflux by extrapolation of the slow phase to zero time. In the data shown in Fig. 12 the fast fraction obtained in this way was shown to have a half-time of 24·0 sec.

Fig. 12.

The escape of 14C-sucrose from an isolated abdominal nerve cord made radioactive by soaking for 20 min. in a solution containing the labelled disaccharide. The exponent portion of the main curve (closed circles) exhibited a slower efflux than that of 14C-inulin from the nerve cord and is identified as that taking place from the cells of the central nervous system. The fast component (open circles) was obtained by extrapolation of the main curve to zero time.

Fig. 12.

The escape of 14C-sucrose from an isolated abdominal nerve cord made radioactive by soaking for 20 min. in a solution containing the labelled disaccharide. The exponent portion of the main curve (closed circles) exhibited a slower efflux than that of 14C-inulin from the nerve cord and is identified as that taking place from the cells of the central nervous system. The fast component (open circles) was obtained by extrapolation of the main curve to zero time.

Table 12 summarizes some details of the dimensions and free diffusion constants of the test substances together with the data for the rates of efflux of the various ions and molecules contained in the rapidly exchanging fractions in the abdominal nerve cord. These data include a value for the efflux of 14C-glucose which in a previous investigation (Treherne, 1960) was attributed to molecules associated with the surface of the nerve cord. In the fight of more recent evidence (Treherne, 1961 e) this efflux is identified as that taking place from the extracellular spaces of the abdominal nerve cord. The combined results show no simple correlation between either the size or the diffusion constants of the ions and molecules and their rates of efflux from the abdominal nerve cords. The sucrose molecules, for example, apparently showed a more rapid efflux than the smaller potassium ions which possess a diffusion constant nearly three times greater than that of the non-electrolyte. Chloride ions, on the other hand, which are closely similar to potassium both in size and diffusion constant were also found to have a half-time for efflux which was less than a third of that of the cation.

Table 12.

The rate of efflux (t0.6) of the rapidly exchanging fraction in the abdominal nerve cord together with the diffusion constants and effective radii of the various test substances

The rate of efflux (t0.6) of the rapidly exchanging fraction in the abdominal nerve cord together with the diffusion constants and effective radii of the various test substances
The rate of efflux (t0.6) of the rapidly exchanging fraction in the abdominal nerve cord together with the diffusion constants and effective radii of the various test substances
To interpret the results contained in Table 12 it is relevant to recall the close similarity, which has already been commented upon, between the efflux of these ions and molecules from the central nervous system of this insect and the diffusion from the’ extracellular space in muscle demonstrated by Hill (1928). It was shown that in the case of a cylindrical muscle the diffusion from the extracellular space will be initially complex but will eventually follow a simple exponential with a half-time given by
formula
where r0 is the radius of the muscle and D′ the diffusion constant in the extracellular space. As has already been pointed out (Treherne, 1961e) Hill’s equation can be modified for a complex structure such as the abdominal nerve cord by representing it thus
formula
where a is some constant. The term I/t0.5 is therefore proportional to the diffusion constant in the extracellular spaces of this complex structure. The plot of I/t0.5 against D, the free-diffusion constant, in Fig. 13 illustrates the relation between the movements of the ions and molecules in the extracellular spaces and those occurring in free solution. If the diffusion in the extracellular spaces was unrestricted then it would be expected that there would be a linear relation between D and I/t0.5 for it is reasonable to suppose that a would remain constant under these conditions. The results illustrated in Fig. 13 show that there is not a simple linear relation for all the substances between the diffusion in the extracellular spaces and that in free solution. Clearly the three cations diffused more slowly than would be expected on the basis of their free-diffusion constants. The remaining non-electrolytes, chloride and tritiated water can, however, be related by a line such as the continuous one shown in Fig. 13. The actual theoretical relationship which would be expected if the diffusion in the extracellular spaces was unrestricted cannot be calculated without a knowledge of the factors contained in the constant a. A relation can, however, be calculated if it is assumed that the movement of the tritiated water molecules approximated to that in free solution. This assumption is not necessarily strictly valid but forms the basis for a minimum estimate of the relation to be expected. On this assumption the following equation for freely diffusing molecules and ions should hold
formula
where Dx is the diffusion constant and the half-time for efflux of the test substance. This minimum estimate is shown in Fig. 13 by the broken line which appears to depart from the measured relationship at lower values of D and I/t0.5.
Fig. 13.

The relation between the free diffusion constant (D) and the value 1/t0.5 obtained for the various test substance in their diffusion in the extracellular spaces of the abdominal nerve cord.

Fig. 13.

The relation between the free diffusion constant (D) and the value 1/t0.5 obtained for the various test substance in their diffusion in the extracellular spaces of the abdominal nerve cord.

The concentrations of the rapidly exchanging ion fractions, which have been identified with the extracellular spaces of the central nervous system (Treherne,1961 e), have been shown to be very different from those in the external medium. The cations were estimated to occur at much higher concentrations in the extracellular spaces, the sodium and potassium levels being nearly twice, and that of calcium at least four times, that of the external solution; the chloride ions, on the other hand, were shown to be present at very much lower concentration in the extracellular spaces as compared with that outside the abdominal nerve cord. This differential ionic distribution was shown to be probably the result of some sort of Donnan equilibrium between the haemolymph and the extracellular fluid. The very much higher concentration of the calcium in the extracellular spaces, as compared with that of the monovalent cations, is presumed to have resulted from the enhanced Donnan effect due to the divalent change of the calcium ion. The precise nature of the anion causing the Donnan effect is not altogether clear from the results presented above. Two possibilities seem to exist. The demonstrated ionic distribution could be due to large protein molecules in solution which were effectively confined to the extracellular space as a result of the relative impermeability of some layer of the nerve sheath or to the presence of free anion groups of some structural materials associated with the extracellular spaces. A degree of impermeability of the nerve sheath to large molecules, which the first hypothesis would demand, might perhaps be inferred from the demonstrated restriction to diffusion of the inulin molecules in their efflux from the intact nerve cord. The free anion groups postulated in the second hypothesis might, for example, be those associated with collagen, which according to Tristram (1953) has 77·2 free anion equivalents/106 g. protein. Collagen-like material has been identified in the extracellular spaces of the nervous system of an insect (Gray, 1959). Using the figure of Tristram it can be calculated that if all the anionic groups were available then the collagen required to produce the ionic distribution demonstrated above would represent at least about 5 % of the wet weight of the abdominal nerve cord. The other possible structural element which might contribute to the Donnan effect in the extracellular spaces of the ganglia is the acid mucopolysaccharide demonstrated in the thoracic ganglia of Periplaneta americana by Ashhurst (1961) and Pipa (1961). By analogy with the polysaccharide component of the mucoprotein in cartilage, studied by Kantor & Schubert (1957), it might be supposed that a substantial proportion of anionic groups would be free and associated with inorganic cations in this ground substance in the cockroach central nervous system. No quantitative figures appear to be available for the extent of the free anion equivalents in this material and it is therefore not possible to calculate the quantity which would be required to produce the ionic distribution between the haemolymph and the extracellular spaces of the abdominal nerve cord demonstrated above.

It was hoped that the experiments on the distribution of 22Na and 45Ca in ganglia which were desheathed before being soaked in the radioactive solutions would help to throw some light on the nature of the anion groups which were causing the Donnan distribution in the extracellular spaces. If the anion groups were those of protein molecules in solution then it was presumed that the concentration of the cations in the extracellular fluid would fall to that of the external medium as the large anion molecules were dispersed following the removal of the nerve sheath. On the other hand, if the Donnan effect was due to the anion groups of some structural elements associated with the extracellular spaces, then it was supposed that the concentration of the cations would still be maintained above that of the external solution. These possible effects were, however, obscured by the great enlargement of the extracellular space, equivalent to an approximately sixfold increase, caused by the removal of the nerve sheath. Taking into account this increase in volume, the expected concentration of sodium in the extra-cellular spaces, on the assumption of fixed anion groups, would have been 178 mM./l. in the desheathed preparation as against 287 mM./l. in the intact ganglion. The apparent concentration estimated in the desheathed ganglia was 160·1 ± 10·5 (S.E.) mM./l., which was not significantly different from either the external concentration of 157·0 mM./l. or that of 178·0 mM./l. predicted in the case of the fixed anion groups. On the basis of these experiments, then, it is not possible to differentiate between the two possibilities of the anion groups being those of molecules in solution or those attached to structural elements associated with the extracellular spaces. In the case of the experiments with 45Ca the expected concentration in the extracellular spaces would have been 6·5 mM./l. in the case of fixed anion groups as against the level of 4·5 mM./l. which would have been expected for the case of diffusible anion molecules. The actual estimated concentration in desheathed ganglia loaded in vitro was 4·6 ± 0·5 (S.E.) mM./l. Despite the variability of these results the value of 6·5 mM./l. lies outside the confidence limits of the observed mean (P < 0·02 <0·01) which would suggest that the demonstrated ionic distribution cannot be explained solely in terms of the anion groups attached to structural elements in the extracellular spaces of the abdominal nerve cord. It therefore seems probable that a significant proportion of the cations in the extracellular fluid may be associated with anion groups of large molecules in solution. The possibility still remains, however, that an appreciable fraction may perhaps also be associated with the anion groups of such structural elements as collagen and acid mucopolysaccharides.

There is a striking similarity between the results for the cockroach abdominal nerve cord and those obtained for the distribution of sodium and potassium ions in the sciatic nerve of the cat (Kenjević, 1955). In the case of the mammalian peripheral nerve both of the cations were present in the extracellular space at concentrations which were about 1·6 times greater than that of the plasma. Kenjević considered the possibility of the existence of a simple Donnan equilibrium between the plasma and the extracellular fluid, but favoured the view that the differential ion distribution was perhaps the result of some active mechanism for the removal of water through the nerve sheath. Unfortunately Kenjević made no direct measurements on the chloride level of the extracellular space of the sciatic nerve using 38C1 so that it is not possible rigorously to eliminate the possibility of a simple Donnan effect even in this mammalian nerve.

One consequence of the Donnan equilibrium demonstrated in the present investigation is an excess of osmotic pressure within the extracellular spaces of the central nervous system of Periplaneta. For the ions studied in these experiments the concentration in the extracellular fluid exceeded that in the haemolymph by about 67·0 mM./l. It seems reasonable to assume that the excess of internal osmotic pressure might be opposed by the inextensibility of the outer fibrous neural lamella. The almost explosive bursting-out of nerve substance through small holes made in the nerve sheath described by Twarog & Roeder (1956) is convincing evidence of the mechanical resistance which this membrane offers to swelling.

It now seems possible to attempt to resolve the apparent paradoxical situation which has arisen with regard to the function of the sheath surrounding the insect nervous system. The work of Hoyle (1953) on locust peripheral nerve and of Twarog & Roeder (1956) on the cockroach central nervous system both showed that the presence of the sheath greatly reduced the rate of depolarization of the underlying nervous elements in solutions containing high concentrations of potassium ions. It was concluded on the basis of these results that the perilemma was an extremely impermeable diffusion barrier which prevented the ready access of ions from the haemolymph to the underlying nervous tissues. The postulated impermeability of this membrane has also been invoked to explain the ability of insect axons to function in those species which possess a haemolymph containing high concentrations of potassium ions. More recent work on the cockroach central nervous system has, however, shown that exchanges of both sodium and potassium ions can occur relatively rapidly between the haemolymph and the abdominal nerve cord (Treherne, 1961a). It was also subsequently shown that these exchanges were taking place passively between the haemolymph and an extracellular space (Treherne, 1961e). The present experiments have shown in addition that removal of the nerve sheath results in very dramatic changes in the composition of the extracellular fluid surrounding the neurones and glial cells of the abdominal nerve cord as a result of the disruption of the demonstrated Donnan equilibrium with the ions in the haemolymph. The changes occurring in the extracellular fluid resulting from the desheathing procedure may be summarized as follows:

  • (1) A sixfold increase in the volume of the extracellular water.

  • (2) A 34 % fall in the concentration of potassium ions.

  • (3) A 47 % fall in the concentration of sodium ions.

  • (4) A 73 % increase in the concentration of chloride ions.

  • (5) A fourfold decrease in the concentration of calcium ions.

  • (6) A decrease in osmotic pressure.

  • (7) A decrease in hydrostatic pressure.

It is now relevant to consider some of the effects which such changes have on excitable tissues, especially in the conditions of the high external potassium concentration employed in the experiments of Twarog & Roeder (1956). The fall in the concentration of potassium ions in the extracellular fluid, which at equilibrium would be even higher than that in the external solution, would by itself be expected to reduce the rate of depolarization of the conducting elements in the central nervous system. The fall in the level of sodium ions in the extracellular spaces might, however, be expected greatly to increase the rate of depolarization at high potassium concentrations. In frog nerve fibres, for example, it has been shown that the block of conduction by high potassium is greatly delayed by excess sodium (Lundberg, 1951). It may be significant that the twofold increase in sodium concentration required to produce this effect is approximately equivalent to the concentration difference maintained between the extracellular space and the haemolymph of the cockroach abdominal nerve cord. Lundberg quotes the example of nerve fibres treated with twice normal sodium in which the total block to conduction developed after 40 min. in 35·0 mM./l. KC1, whereas at normal sodium level the same concentration of potassium gave a total conduction block after only 3 min. The marked increase in concentration of chloride ions resulting from the removal of the nerve sheath may be presumed to produce little polarization change, for substitution of this ion with glucose or sucrose produced only effects similar to those caused by removal of sodium alone in experiments with squid giant axons (Hodgkin & Katz, 1949). The change in concentration of calcium on desheathing, on the other hand, might be expected to produce changes in the electrical properties of the central nervous tissues, for this ion is well known as a stabilizing substance (cf. Shanes, 1958 a, b). In frog nerve fibres, for example, lowering of the calcium level of the bathing solution results in an accentuation of potassium depolarization (Stkmpfli & Nishie, 1956). The reduction in the osmotic pressure of the extracellular fluid could also lead to an accelerated potassium depolarization in desheathed preparations, for small nerve bundles have been shown to be depolarized more slowly in Ringer made hypertonic with sucrose and more rapidly in hypotonic Ringer (Stampfli & Nishie, 1955). By analogy with the vertebrate condition it might be supposed, however, that the effects of osmotic pressure on the rate of potassium depolarization would be smaller than those induced by decreased sodium concentration (Lundberg, 1951).

The above examples have served to show that the complex changes produced in the extracellular fluid by the desheathing procedure may contribute to the accelerated rate of potassium depolarization demonstrated in desheathed preparations by Twarog & Roeder (1956). In particular the reduction of the sodium and calcium levels resulting from the disruption of the Donnan equilibrium are likely to have profound effects on the blocking times measured by these authors. The possibility must not be overlooked that some other unspecified essential factors may also be released following the desheathing procedure. At all events the relatively rapid exchanges of ions between the extracellular space and the haemolymph apparently show that the resistance to diffusion offered by the perilemma is likely to account for only a fraction of the time taken to produce the conduction blocks demonstrated by Twarog & Roeder in solutions of high potassium concentration.

The properties of the nerve sheath as a significant diffusion barrier have also been inferred from observations on the effects of acetylcholine on the insect central nervous system (Twarog & Roeder, 1956). It was shown that, while the extremely high concentration of acetylcholine of 10-2M. was without effect in intact ganglia, synaptic function was rapidly modified in desheathed preparations. The lowest effective concentration causing synaptic block in desheathed preparations was between a half and a third of that in the intact ganglia (3–5 ×10-3M.). A similar effect of desheathing was noted in the investigation of Yamasaki & Narahashi (1960) on the synaptic transmission of the last abdominal ganglion of the cockroach. In the fight of the evidence presented above, the function of the perilemma as a structure rigidly excluding acetylcholine molecules from the central nervous system seems unlikely. An explanation can, however, be sought from among the very dramatic changes produced in the extracellular fluid of the central nervous system resulting from the disruption of the Dorman equilibrium by the desheathing procedure. It is well known that the transynaptic excitation is dependent on the chemical environment of cells and terminations in synaptic regions (cf. Brink, 1954). In particular the calcium ion has been shown to be of extreme importance in transynaptic excitation processes in sympathetic ganglia. The excitability of postsynaptic cells is enhanced by the perfusion of acetylcholine in the absence of calcium ions; correspondingly the excitability is reduced during perfusion with solutions containing higher than normal calcium chloride (Brink, Bronk & Larrabee, 1946; Bronk, 1939; Harvey & Macintosh, 1940). It might thus be only necessary to suppose that the fall in the calcium concentration in the extracellular fluid, from 17·1 to 4·6 mM./l., resulting from the desheathing procedure changed the level of excitability of the synaptic membranes to acetylcholine in the experiments of Twarog & Roeder and Yamasaki & Narahashi.

Recent work on the permeability relations of the terminal abdominal ganglion of the cockroach nerve cord has shown that molecules of the dye trypan blue did not penetrate readily beyond the outer fibrous sheath, or neural lamella (Wigglesworth, 1960). It was concluded on the basis of these results that the inner cellular layer, the perineurium, was responsible for restricting the entry of the dye molecules into the deeper layers of the ganglion. In view of the demonstrated accessibility of the abdominal nerve cord to the relatively large inulin molecules it is perhaps surprising that trypan blue, with a molecular weight of 868, should apparently be excluded from the extracellular spaces of the ganglion. The ganglia of the cockroach do, however, contain appreciable amounts of a mucopolysaccharide substance (Ashhurst, 1961; Pipa, 1961). Similar substances in vertebrate tissues have been shown to possess numerous free anionic groups and it has already been suggested that these groups might contribute to the Donnan equilibrium described above for this insect. The trypan blue molecule possesses four negative charges and in experiments with cartilage ground substance, which is formed of mucopolysaccharide in association with protein, this dye has been shown to be completely excluded from the tissue because of the presence of the free anionic groups (Kantor & Schuber, 1957). Large positively charged dye molecules were, however, shown to penetrate into the ground substance relatively rapidly. It can thus be postulated that the exclusion of trypan blue from the ganglion depended not only upon the properties of the cellular layer of the nerve sheath but also upon the presence of an anionic ground substance demonstrated by Ashhurst.

In a previous paper it was shown that the ratio of the rates of efflux of 14C-labelled inulin molecules and 24Na ions was similar to that for the free diffusion constants of these substances (Treherne, 1961e). These results were interpreted as showing that the movements of the polysaccharide molecules were relatively unrestricted in the extracellular spaces of the central nervous system. The present data have shown that this interpretation is not valid, for the movements of the cations in the extracellular spaces were apparently much slower than would be expected on the basis of their free diffusion constants or their effective ionic radii. Thus the similarity of the ratio for the diffusion of the sodium ions and of the polysaccharide molecules in the extracellular spaces to the ratio of their free diffusion constants was, in fact, due to the slowing down of the movements of the positively charged ions in the extracellular spaces. The present results have clearly shown that the relatively large inulin molecules were in fact diffusing in these spaces much more slowly than would be expected on the basis of their free diffusion constant. This effect is presumed to result from the restricted diffusion caused by the passage of the polysaccharide molecules through relatively small spaces at some point in the extracellular system of the abdominal nerve cord. The factors causing the slowing down of the movements of the positively charged ions cannot be precisely identified at the moment. It is possible, for example, that such an effect could result from a peripherally situated cation barrier or from the attraction of the free anion groups encountered in the passage of these ions through the extracellular spaces of the abdominal nerve cord.

The apparent diffusion constant in the extracellular spaces, D, calculated for sodium in the previous study was 5·77 ×10-7 cm2, sec.-1 which was approximately onethirtieth of that for free diffusion of sodium ions (Treherne, 1961 e). Following Harris & Burn (1949) this reduction of the diffusion coefficient was attributed to the increase in the effective path length of the ions diffusing between a complex collection of cellular structures. From the present experiments it is now known that in addition to the effect of the increased path length there is also a reduction in the rate of diffusion of these ions due to their positive charges. By comparing the figure for D′ with chloride ions, which have been shown to be not restricted in this way, it can be calculated that the diffusion constant in the extracellular spaces would be very approximately one-eighth of that in free solution. In the apparent absence of significant restriction to diffusion of these small ions it seems reasonable to attribute this effect to the increased effective path length traversed in diffusing through the extracellular spaces.

Finally, it is of interest to consider the effects which the composition of the extracellular fluid might have on the electrical behaviour of the axons in the central nervous system of this insect. The resting potential of cockroach giant axons at low external concentrations of potassium has been shown to be about 70–77 mV. (Yamasaki & Narahashi, 1959; Narahashi & Yamasaki, 1960). According to the diagram given by Yamasaki & Narahashi the resting potential at the concentration of the extracellular fluid (17·1 mM./l. K) was approximately 54 mV., as against that of 58 mV. at the level of the haemolymph potassium (12·3 mM./l.). At the relatively high sodium concentration, of 283·6 mM./l., in the extracellular fluid the action potential measured by these authors was about 26 mV. The resting potential was shown not to be significantly altered by sodium solutions of the order of concentration found in the extracellular fluid. The concentration of calcium ions found in the extracellular fluid (14·7 mM./l.) was much higher than the normal value of 1·8 mM./l. used by Narahashi & Yamasaki (1960) in their experiments. This high level of calcium would, according to the results given by these authors, only result in a very small degree of hyperpolarization for the resting potential. The magnitude and shape of the action potential in a calciumrich solution, of 18·0 mM./l., were, however, demonstrated to undergo some changes as compared with those in the normal solution. Thus it would be expected that the action potential quoted above would be slightly increased in height and also that the onset of the falling phase would be delayed in the calcium-rich conditions of the extracellular fluid. The post-spike phases were also affected by solutions of high calcium concentration, the positive phase being depressed and the negative after-potential being augmented.

I am indebted to Dr R. H. Adrian and Prof. J. W. S. Pringle, F.R.S., for some helpful discussions during the course of this work.

Ashhurst
,
D. E.
(
1961
).
An acid mucopolysaccharide in cockroach ganglia
.
Nature, Lond
.,
191
,
1224
5
.
Asperhn
,
K. Van
&
Esch
,
I. Van
(
1956
).
The chemical composition of the haemolymph in Periplaneta americana
.
Arch nierl. Zool
.
11
,
342
60
.
Brink
,
F.
(
1954
).
The role of calcium ions in neural processes
.
Pharm. Rev
.
6
,
243
98
.
Brink
,
F.
,
Bronk
,
D. W.
&
Larrabee
,
M. G.
(
1946
).
Chemical excitation of nerve
.
Ann. N.Y. Acad. Sei
.
47
,
457
85
.
Bronk
,
D. W.
(
1939
).
Synaptic mechanisms in sympathetic ganglia
.
J. Neurophysiol
.
2
,
380401
.
Gray
,
E. G.
(
1959
).
Electron microscopy of collagen-like connective tissue fibrils of an insect
.
Proc. Roy. Soc. B
,
150
,
233
9
.
Harris
,
E. J.
&
Burn
,
G. P.
(
1959
).
The transfer of sodium ions between muscle and the surrounding medium
.
Trans. Faraday Sac
.
45
,
508
28
.
Harvey
,
A. M.
&
Macintosh
,
F. C.
(
1940
).
Calcium and synaptic transmission in a sympathetic ganglion
.
J. Physiol
,
408
16
.
Hill
,
A. V.
(
1928
).
The diffusion of lactic acid through tissues
.
Proc. Roy. Soc. B
,
104
,
39
96
.
Hodgkin
,
A. L.
&
Katz
,
B.
(
1949
).
The effect of sodium ions on the electrical activity of the giant axon of the squid
.
J. Physiol
.
108
,
37
77
.
Hober
,
R.
(
1945
).
Physical Chemistry of Cells and Tissues
.
Philadelphia
:
Blakiston
.
Hoyle
,
G.
(
1953
).
Potassium ions and insect nerve muscle
.
J. Exp. Biol
.
30
,
121
35
.
International Critical Tables (1929)
. Ed.
Washbum
,
E. W.
New York
:
McGraw-Hill
.
Kantor
,
T. G.
&
Schubert
,
M.
(
1957
).
The difference in permeability of cartilage to cationic and anionic dyes
.
J. Histochem. Cytom
.
5
,
28
32
.
Kortüm
,
G.
&
Bockris
,
J. O. M.
(
1951
).
Textbook of Electrochemistry. II
.
Amsterdam
:
Elsevier
.
Krnjevió
,
K.
(
1955
).
The distribution of Na and K in cat nerves
.
J. Physiol
.
128
,
473
88
.
Lundberg
,
A.
(
1951
).
On the ability of some cations to inhibit the potassium depolarization of frog nerve fibres
.
Acta Physiol. Scand
.
22
,
365
75
.
Narahashi
,
T.
&
Yamasaki
,
T.
(
1960
).
Mechanism of the after-potential production in the giant axons of the cockroach
.
J. Physiol
.
151
,
75
88
.
Pappenheimer
,
J. R.
(
1953
).
Passage of molecules through capillary walls
.
Physiol. Rev
.
33
,
387
423
.
Pipa
,
R. L.
(
1961
).
Studies on the Hexapod nervous system. III. Histology and histochemistry of cockroach neuroglia
.
J. Cell. Comp. Physiol
.
116
,
15
26
.
Renkin
,
E. M.
(
1954
).
Filtration, diffusion and molecular sieve through porous cellulose membranes
.
J. Gen. Physiol
.
38
,
225
43
.
Shanes
,
A. M.
(
1958a
).
Electrochemical aspects of physiological and pharmacological action in excitable cells. I. The resting cell and its alteration by extrinsic factors
.
Pharm. Rev
.
10
,
59
164
.
Shanes
,
A. M.
(
1958b
).
Electrochemical aspects of physiological and pharmacological action in excitable tissues. II. The action potential and excitation
.
Pharm. Rev
.
10
,
165
273
.
Stxmpfli
,
R.
&
Nishie
,
K.
(
1955
).
Effects of aniso-osmotic solutions on resting and action potential of myelinated nerve
.
Helv. physiol, acta
.
13
,
33
34
.
Stampfli
,
R.
&
Nishib
,
K.
(
1956
).
Effects of calcium-free solutions on membrane-potential of myelinated nerve fibres of the Brazilian frog Leptodactylus ocellatus
.
Helv. Physiol. Pharm. Acta
,
14
,
93104
.
Treherne
,
J. E.
(
1960
).
The nutrition of the central nervous system in the cockroach, Periplaneta americana L. The exchange and metabolism of sugars
.
J. Exp. Biol
.
37
,
513
33
.
Treherne
,
J. E.
(
1961a
).
Sodium and potassium fluxes in the abdominal nerve cord of the cockroach, Periplaneta americana L
.
J. Exp. Biol
.
38
,
315
22
.
Treherne
,
J. E.
(
1961b
).
The movements of sodium ions in the isolated abdominal nerve cord of the cockroach, Periplaneta americana L
.
J. Exp. Biol
.
38
,
629
36
.
Treherne
,
J. E.
(
1961c
).
Exchanges of sodium ions in the central nervous system of an insect (Periplaneta americana L
.).
Nature, Lond
.,
191
,
1223
4
.
Treherne
,
J. E.
(
1961d
).
The efflux of sodium ions from the last abdominal ganglion of the cockroach, Periplaneta americana L
.
J. Exp. Biol
.
38
,
729
36
.
Treherne
,
J. E.
(
1961e
).
The kinetics of sodium transfer in the central nervous system of the cockroach (Periplaneta americana L
.).
J. Exp. Biol
.
38
,
737
46
.
Tristram
,
G. R.
(
1953
).
The amino-acid composition of proteins
.
In The Proteins
,
1
, p.
224
, ed.
Neurath
,
H.
&
Bailey
,
K.
New York
:
Academic Press
.
Twarog
,
B. M.
&
Roeder
,
K. D.
(
1956
).
Properties of the connective tissue sheath of the cockroach abdominal nerve cord
.
Biol. Bull., Woods Hole
,
111
,
278
86
.
Wang
,
J. H.
,
Robinson
,
C. V.
&
Edelman
,
I. S.
(
1953
).
Self-diffusion and structure of liquid water. III. Measurements of the self-diffusion of liquid water with H2, H3 and O13 as tracers
.
J. Amer. Chem. Soc
.
751
,
466
70
.
Wigglesworth
,
V. B.
(
1959
).
The histology of the nervous system of an insect, Rhodmus prolixus (Hemiptera). II. The central ganglia
.
Quart. J. Mier. Sei
.
100
,
299
313
.
Wigglesworth
,
V. B.
(
1960
).
The nutrition of the central nervous system in the cockroach, Periplaneta americana L. The role of perineurium and glial cells in the mobilization of reserves
.
J. Exp. Biol
.
37
,
500
12
.
Yamasaki
,
T.
&
Narahashi
,
T.
(
1959
).
The effects of potassium and sodium ions on the resting and action potentials of the cockroach giant axon
.
J. Im. Physiol
.
3
,
146
58
.
Yamasaki
,
T.
&
Narahashi
,
T.
(
1960
).
Synaptic transmission in the last abdominal ganglion of the cockroach
.
J. Ini. Physiol
.
4
,
1
13
.