Oscillatory Contraction Activity in Physarum

The plasmodial phase of the acellular slime mould Physarum polycephalum represents a multinuclear mass of protoplasm. Migrating plasmodia (Fig. 1a) are differentiated into a front region, i.e. a compact sheet of protoplasm, and in the rear region into protoplasmic ‘veins’ or ‘strands’. Within the veins protoplasm is transported with velocities up to 1·3 mm/s (Kamiya, 1959). Much of the interest in studying the physiology of Physarum lies in the fact that the plasmodia are used to analyse the function of cytoplasmic actomyosin (Komnick et al. 1973; Wohlfarth- Bottermann, 1975b).

The protoplasmic veins (Fig. 1b) are differentiated into an ectoplasmic tube (possessing a complicated system of plasmalemma invaginations with intimately connected cytoplasmic actomyosin structures in the form of sheets and fibrils) and an endoplasmic core (representing the more fluid transported protoplasm (Wohlfarth-Bottermann, 1974; 1975c). The rhythmical ebb and flow of endoplasm has a period within the minute range (shuttle streaming), i.e. protoplasmic streaming oscillates for respect to its direction (Kamiya, 1959; Sachsenmaier & Hansen, 1973; Hülsmann & Wohlfarth-Bottermann, 1978b). A ‘long period oscillation’ (8–10 h) related to synchronous mitoses (Sachsenmaier & Hansen, 1973) will not be discussed in this contribution.

When considering the question of the nature (Hess & Boiteux, 1971) of the physiological oscillator(s) which drives the periodic streaming phenomena, we have to consider the motile mechanisms responsible for endoplasmic flow. The endoplasm flows passively due to hydrostatic pressure differences developed in different areas of the veins (Kamiya, 1959). These pressure differences are brought about by rhythmic contractions of the ectoplasmic tube, i.e. by oscillatory contractions of the cytoplasmic actomyosin within the ectoplasm (Wohlfarth-Bottermann, 1962, 1965; Hatano & Oosawa, 1966; Fleischer & Wohlfarth-Bottermann, 1975; Wohlfarth-Bottermann & Fleischer, 1976). Recordings of contractile phenomena can be made by cinematography and more recently also by tensiometric and other techniques (Kamiya, 1970; Wohlfarth-Bottermann, 1975 a; Samans, Götz von Olenhusen & Wohlfarth-Botter-mann, 1978; Grebecki & Cieslawska, 1978). Tensiometry is a convenient technique in addition to providing a direct measurement of contractile force in millipond (1 kp = 9·80665 Newtons) of single veins. With this equipment it is possible to register automatically

1. longitudinal contractile activity of isolated veins (under isometric and isotonic conditions) (Kamiya, 1970),

2. radial contractile behaviour of veins in situ (Wohlfarth-Bottermann, 1975 a),

3. longitudinal contractions of veins in situ (Hülsmann & Wohlfarth-Bottermann, 1978a),

4. protoplasmic flow rates (in mg/min) of single veins by using the tensiometric electrobalance for weighing inflowing and outflowing endoplasm (Hülsmann & Wohlfarth-Bottermann, 1978b).

What are the time periods of the different oscillating phenomena? The mean average time period for longitudinal contractions of isolated veins under isometric conditions lies in the range of 2·1 min, whereas the corresponding value for radial contractions is 1·3 min (Wohlfarth-Bottermann, 1977). The latter value corresponds with the mean average value of the shuttle streaming period (Hülsmann & Wohlfarth-Bottermann, 1978b).

These observations raise the question of whether there are different contraction frequencies in one vein (i.e. 2·1 min (longitudinal) and 1·3 min (radial and streaming)). If the tensiometric analysis of radial activities and longitudinal activities of veins are performed in situ and both measurements are done simultaneously in one and the same vein by using a twin tension transducer and a special arrangement of the vein (Hülsmann & Wohlfarth-Bottermann, 1978 a, b), it can be shown that there is only one common time period of 1·3 min and that both contractile activities have a close phase identity.

When radial, longitudinal and streaming phenomena were recorded simultaneously Fig. 2a), all three curves (Fig. 2b) reveal identical frequencies and close phase relationships. In summary, the time period of approximately 13 min seems to tepresent the genuine and characteristic frequency for these oscillatory phenomena in Physarum.

According to Yoshimoto & Kamiya (1978a–d) ‘the strand has no active contractile rhythm in both radial and longitudinal directions as long as it is part of the network or ramified region of the advancing plasmodium’. However, the presence of rhythmicity has been demonstrated by applying infrared (Samans et al. 1978) as well as cinematographic (Grebecki & Cieslawska, 1978) recording techniques which, in contrast to tensiometry, register radial oscillations without touching the veins. Hülsmann & Wohlfarth-Bottermann (1978b) have provided further evidence and arguments against the claims of Yoshimoto & Kamiya (1978d) cited above.

Because the sinusoidal form of oscillation curves often suggest the existence of interference phenomena by two waves with slightly different frequencies (Kamiya, 1959, 1970), it has also to be considered that different oscillators may be operating. However, the observed interference phenomena can also result from small local phase shifts of contraction phenomena working with identical frequencies but having a spatially different arrangement (Hülsmann & Wohlfarth-Bottermann, 1978a, b). Principally, it is undecided, whether

• there are different physiological oscillators interfering with each other, or whether

• there is only one oscillator, and spatially different effector systems interfere with each other.

Contraction–relaxation cycles of strand segments become synchronized under tensiometric conditions (Yoshimoto & Kamiya, 1958a). Grebecki & Cieslawska (1978) concluded that ‘the entire system of major veins forming the network contracts synchronously’. This statement, however, is not supported by data obtained from infrared techniques (Z. Hejnowicz & K. E. Wohlfarth-Bottermann, unpublished).

What is the physiological nature and location of the oscillator and is there some sort of ‘pacemaker’ or ‘trigger’ which governs the contractions? It seems reasonable to discuss the following possibilities concerning the nature and location of the unknown clock:

1. oscillatory behaviour based on the force generating system itself (i.e. the cytoplasmic actin and/or myosin),

2. oscillation phenomena of the enzymes of the energy supply system (glycolysis and/or aerobic respiration),

3. rhythmical translocation of calcium across the internal membrane systems which regulate the free calcium concentration (10⁻⁶ to 10⁻⁷ M) and thereby control contractile activity of cytoplasmic actomyosin,

4. rhythmical ion movement across the plasmalemma which may act as a trigger for the contractile machinery, either directly or indirectly by inducing calcium release from the inner stores.

   One approach is to try and exclude some of these possible mechanisms in order to narrow down the alternatives for future investigations.

The existence of a regulating mechanism synchronizing contraction rhythms throughout a large plasmodium was clearly established by Takeuchi & Yoneda (1977), Yoshimoto & Kamiya (1978a), Krüger & Wohlfarth-Bottermann (1978). Yoshimoto & Kamiya (1978c) conclude from their results that the regulating factor is transmitted by the flowing endoplasm. However, since rhythmicity of the ectoplasmic tube continues when endoplasm is replaced by artificial media (Ueda et al. 1978), an exclusively ‘endoplasmic seat’ seems unlikely.

Let us consider some alternative locations in somewhat more detail.

The example of the oscillatory behaviour of insect flight muscle (Rüegg, 1973) shows that in principle actomyosin can oscillate sui generis. Force output of Physarum veins is increased by stretching the strands (Kamiya, 1970; Wohlfarth-Bottermann, 1975a) and strong stretches induce a phase shift of the oscillation (Krüger & Wohlfarth-Bottermann, 1978) (Fig. 3a, b). These two facts seem to favour the idea that a mechanical component (tension) is involved in the control mechanism. Yoshimoto & Kamiya (1978b) deny the possibility of shifting the phase mechanically and point to the possibility that the clock mechanism ‘advances, even when contraction per se is inhibited’.

In spite of many experiments, we have been unable to restore oscillatory activity in a glycerinated or detergent-treated strand (i.e. in a ‘cell free’ system) in a significant and reproducible manner. Thus, the question of whether the oscillator is influenced by tension is still undecided. Oscillatory behaviour of the actomyosin itself remains to be established, but it should be mentioned that the contraction-relaxation cycle needs a periodic actin transformation (Isenberg & Wohlfarth-Bottermann, 1976). If this actin cycle is inhibited, e.g. by the injection of Phalloidin (Fig. 4d), which shifts the equilibrium irreversibly towards the F-form of actin all oscillations cease as soon as the drug reaches the contractile elements (Götz von Olenhusen & Wohlfarth-Bottermann, 1979a).

The regulatory mechanisms which control this actin transformation are thus potential candidates for the oscillator. Conceivably, changes in tension may influence not only the actin equilibrium, but also calcium homeostasis by triggering mechanosensitive calcium channels (e.g. in the plasmalemma or in the membranes enclosing internal calcium stores).

It is known that the glycolytic system of yeast cells can oscillate (Hess & Boiteux, 1971; Richter, Betz & Giersch, 1975). Sachsenmaier & Hansen (1973) have speculated that the rhythmical contractions of Physarum might be driven by such a glycolytic oscillator. Daniel (1970) has observed that the reduced pyridin nucleotide level in Physarum seems to oscillate in phase with contractile activity. However, it is difficult to monitor such biochemical changes using conventional fluorometric techniques because such ‘optical oscillations’ may result mainly from rhythmic alterations of vein geometry (Sachsenmaier & Hansen, 1973). Another approach to this problem is to study the response of longitudinal contractions to anaerobic conditions and to inhibitors of glycolysis (Fig. 3c). In contrast to earlier observations about the effects of these conditions upon protoplasmic streaming (Kamiya, 1959; Sachsenmaier & Hansen, 1973), there was a decline in the longitudinal force output recorded tensiometrically during oxygen deprivation. After reaeration of the outer medium, contractile amplitude recovered. Application of 5 × 10⁻⁵ M iodoacetate (IAA) also reduced contractile amplitude. Complete inhibition of oscillatory activity occurred only after the simultaneous application of IAA and the removal of oxygen, i.e. a complete energy blockade (Fig. 3c). The energy source seems to be variable, dependent upon the availability of oxygen. In the experiments described in Fig. 3(c), variations in frequency were not significant, irrespective of whether aerobic or anaerobic cell respiration was available. This seems not to be in favour of an assumption that there is a metabolic oscillator which functions to create regular oscillations via the level of ATP.

The cytoplasmic actomyosin of Physarum is sensitive to changes in the level of free calcium (Hatano, 1970). Contractile activity can be controlled by injecting varying calcium concentrations (Ueda, Götz von Olenhusen & Wohlfarth-Bottermann, 1978). Histochemical and biochemical techniques have revealed the existence of internal stores which can accumulate calcium against concentration gradients by means of energy-requiring calcium pumps (Braatz & Komnick, 1973; Kato & Tonomura, 1977). However, there is no unequivocal evidence to suggest that a rhythmical release and accumulation of calcium by these stores is the basis of the oscillator. Ridgeway & Durham (1976) used the photoprotein aequorin in an attempt to detect fluctuations in intracellular calcium. Unfortunately, the oscillations in light output which they described may be explained by rhythmic radial volume changes of the veins and may not necessarily reflect rhythmical changes in the level of calcium.

The use of streaming direction by Ridgeway & Durham (1976) as a criterion for contractile phases is of limited value, since streaming and contraction activity are not always in identical phases. However, the thesis that there is a rhythmical translocation of calcium across the membranes of the inner calcium stores is nevertheless a very reasonable hypothesis which must be followed up in future experiments.

Unfortunately, the preparation of membrane fractions from Physarum plasmodia in our experience is much more difficult as compared with other cells because the outer slime layer causes different cell components to stick together after homogenization. This hinders their separation during differential centrifugation. A clear-cut demonstration of rhythmical activities of inner Ca²⁺ stores and their functional analysis has to be attempted on purified ultracentrifugal fractions (Kato & Tonomura, 1977).

Similar automatic contraction events as in Physarum are the slow rhythms (minute rhythms) of some smooth muscle (Golenhofen, see Bülbring & Shuba, 1976) and myogenic activity of heart muscle. These are supposed to be triggered by a periodic influx of Ca²⁺ across the plasmalemma which may then trigger a further release of calcium from inner Ca²⁺ stores through a mechanism of calcium-dependent calcium release (Fabiato & Fabiato, 1975; Endo, 1977; Lüttgau, 1977). Therefore the existence of a similar trigger mechanism must be taken into account for Physarum, and it is important to know whether rhythmic changes in ionic permeabilities or electrogenic pump activity across the plasmalemma may play a role as a pacemaker mechanism.

Electrophysiological recordings on Physarum reveal that the membrane potential fluctuates by 54 mV (Meyer & Stockem, 1979). The period of the bioelectrical phenomena responsible for this oscillatory behaviour has a time period in the range of streaming-and contractile-activity (Kishimoto, 1958a,b; Kamiya, 1959; Rhea, 1966). This observation suggests that changes in ionic permeabilities and/or electrogenic ion pumps may trigger contractile activity. The plasmalemma should have an important function at least in the modulation of contractile activity (Hato et al. 1976; Ueda et al. 1975, 1976; Ludlow & Durham, 1977), because external chemotactic signals must cross this structure before they lead to appropriate alterations of the oscillating motor. Unfortunately, however, electrophysiological measurements of membrane potentials are difficult to perform because Physarum has a strong tendency to separate off microelectrodes by forming new plasmalemma around them (Rhea, 1966). Furthermore, ion flux measurements (Ludlow & Durham, 1977) cannot give reliable information because there is an extended extracellular space within the veins in the form of highly branched pockets of deeply penetrating plasmalemma invaginations (Fig. 1 (b), PI). Thus, flux measurements and electrophysiological methods are of restricted applicability and possible physiological events occurring in the plasmalemma have to be deduced indirectly from tensiometric recordings of responses to externally and internally applied test solutions. Physarum offers several advantages for this type of experimental approach. It is possible to study:

1. The effects of external application on persistence of longitudinal contractile activity of veins under isometric conditions of measurement (Wohlfarth-Bottermann, 1975a) (Fig. 4a, c).

2. The effects of injected test solutions on persistence of radial contractile activity of veins in situ (in addition, internal and external solutions can be applied simultaneously) (Fig. 4b, d)(Ueda et al. 1978).

3. The effects of externally applied test solutions on de novo generated contractile activity of protoplasmic drops (Fig. 6) (Wohlfarth-Bottermann & Götz von Olenhusen, 1977; Götz von Olenhusen & Wohlfarth-Bottermann, 1979b).

These modes of application as well as the different preparations (veins or drops at different age stages) have various suitabilities for answering certain questions. Advantages and shortcomings cannot be discussed here in detail (see Götz von Olenhusen & Wohlfarth-Bottermann, 1979b).

The importance of calcium for the contractile rhythm has been investigated in several different ways. Studies with radioactive calcium seem to show a rapid and continuous ⁴⁵Ca²⁺ uptake into migrating plasmodia. This influx may represent either an active uptake process or a passive inflow down the large concentration gradient which exists between the external (10⁻³ M) and internal compartment (10⁻⁷ M). This large influx of calcium is balanced by a corresponding efflux which may be associated with slime secretion. The uptake of ⁴⁵Ca²⁺ into migrating plasmodia is strongly reduced by 5 mm-La³⁺ without obviously decreasing the viability of the plasmodium over a period of many hours (Dierkes, unpublished).

Reducing external calcium to 10⁻⁵ to 10⁻⁸ M in the external medium likewise had little effect in that the automaticity of longitudinal contractions persists for 1 to 2 h. After this time a declining longitudinal contraction force may be traced back to an increasing loss of internal Ca²⁺, which can be prevented by the addition of 5 mm-La³⁺ to the 10⁻⁸ M-Ca²⁺ EGTA buffer (Wohlfarth-Bottermann & Götz von Olenhusen, 1977)-

Verapamil (0·5 mm), a calcium antagonistic drug, abolishes longitudinal force generation within a few minutes after external application (Fig. 4a). However, since the action of this drug is inhibited by 5 mm-Ca²⁺, La³⁺ or Mn²⁺ (not by Mg²⁺ or Sr²⁺) it is difficult to believe that this drug acts to inhibit calcium movement across the plasmalemma as it does in smooth muscle: the outer application of 5 mm-La³⁺ probably is not compatible with a persisting rhythmical Ca²⁺ transport across the plasmalemma. Dörrscheidt-Käfer (1977) points to a possible action of the Ca²⁺ antagonist upon inner stores of the skeletal muscle.

After the introduction of a new technique for replacing of endoplasm by artificial media, it was possible to compare the influence of drugs when applied to either the inside or to the outside of the ectoplasmic tube (Ueda et al. 1978). It became apparent that theophylline and caffeine (Fig. 4b) have a strong effect when applied externally, but not when injected internally. The strong relaxing effect of 15 mm external caffeine was inhibited when 5 mm-La³⁺ was added to the drug solution. This external sensitivity to caffeine (Hatano & Nakajima, 1961; Hatano, 1970) on the outside and the insensitivity of strands after injection of the drug to the inside may indicate that the drug influences contractile activity via the plasmalemma (blockade of phosphodiesterase activity?).

A corresponding differential sensitivity was found when heavy water (D₂O) was applied either externally or internally (Fig. 4c, d). Longitudinal contractile activity stopped immediately when 50% D₂O was added to the bathing solution, but the injection of 99·7% D₂O did not lead to a cessation of contractile rhythmicity. The addition of Phalloidin to the heavy-water proved that the injected medium actually reached the contractile elements (compare Götz von Olenhusen & Wohlfarth-Bottermann, 1978, 1979a). The addition of 10% polyethyleneglycol (PEG) to the external D₂O did not slow down the fast action of the heavy water. These observations on the affects of both caffeine and D₂O on veins initially suggest that both substances act primarily on some processes on the plasmalemma which may be important for the continuation of contractile automaticity. One has, however, to consider (i) that caffeine and D₂O do rapidly penetrate into the cell, and (ii) the effective concentration of substances within the ectoplasmic tube may be low after injection in comparison to external application. Therefore, D₂O and caffeine could still act within the cell, e.g. by altering calcium movement across internal stores (Syson & Huddart, 1976; Kaminer, 1977).

When discussing the effects of certain chemicals applied externally upon veins, we have to consider 6 principal compartments: (1) the extracellular medium, (2) the slime layer, (3) the cell membrane, (4) plasmalemma invaginations with slime content, (5) the cytoplasm, (6) different internal Ca²⁺ stores.

The use of another preparation, i.e. protoplasmic drops (Wohlfarth-Bottermann, 1963; Götz von Olenhusen, Jücker & Wohlfarth-Bottermann, 1979; Götz von Olenhusen & Wohlfarth-Bottermann, 1979b), has four essential advantages in comparison to the use of veins:

• Instead of the persistence of contractile activities, the de novo generation of rhythmicity is used as a criterion for the effects of externally applied substances.

• The slime layer of protoplasmic drops is much thinner than that of veins, i.e. there is faster and more intimate contact of test solutions with the external plasmalemma.

• The extracellular space within the plasmalemma invagination system is much smaller in normal protoplasmic drops as compared with the extended invagination system in veins. Therefore bathing solutions are more likely to reach the innermost pockets of the invaginations.

• In caffeine-treated drops (Fig. 5), plasmalemma invaginations are absent thus removing all possible doubt that solutions can reach all parts of the plasmalemma. In spite of their inability to form plasmalemma invaginations, the caffeine-treated drops (as normal drops) undergo a fibrillogenesis of cytoplasmic actomyosin (drop age 5–15 min) and subsequently a de novo generation of contractile activity (drop age 20 min).

  Fig. 5.

  

  View largeDownload slide

  Drawings of semithin sections of normal protoplasmic drops (control) and of 5 mm caffeine-treated drops demonstrating the differences in interior membrane areas (plasmalemma invaginations and vacuoles). The cytoplasmic actomyosin fibrils are not depicted. Magnification 90 × (Götz von Olenhusen, Jücker & Wohlfarth-Bottermann, 1979).

  Fig. 5.

  

  View largeDownload slide

  Drawings of semithin sections of normal protoplasmic drops (control) and of 5 mm caffeine-treated drops demonstrating the differences in interior membrane areas (plasmalemma invaginations and vacuoles). The cytoplasmic actomyosin fibrils are not depicted. Magnification 90 × (Götz von Olenhusen, Jücker & Wohlfarth-Bottermann, 1979).

Fig. 6(a) shows that contractile activity of normal protoplasmic drops (control in Fig. 5) can be generated de novo if the outer concentration of Ca²⁺ ions is as low as 10⁻⁸ M, and even in the presence of the calcium antagonistic drug D 600 (0·5 mm). Caffeine (5 mm) does not hinder the initiation of activity (Fig. 6b), but 5% D₂O inhibits the newly generated activity (Fig. 6c). The results characterized in Fig. 6(a–c) were obtained in normal drops. In order to be completely sure that the total area of the plasmalemma is reached by the experimental solutions, we tested the influence of different chemicals on caffeine-treated drops, i.e. on drops without plasmalemma invaginations (see Fig. 5).

The relative amount of interior membrane in normal drops compared to caffeine-treated drops at 10 and 20 min was estimated by a morphometric analysis of the semithin sections shown in Fig. 5. The interior membrane area (invaginations and vacuoles) amounts to 265% (10 min) and 365% (20 min) of the exterior plasmalemma area of normal drops, whereas the corresponding values in the caffeine-treated drops are only 18 and 65% respectively (representing mainly vacuoles).

The large amount of interior plasmalemma and vacuoles in normal drops may indicate physiologically important structures delimiting at least partly an extracellular space in the drops’ interior which perhaps is not readily accessible to externally applied test solutions. Thus, the caffeine-induced plasmodial stage which lacks plasmalemma invaginations is completely accessible to experimentally applied solutions. Fig. 6(d) demonstrates that such drops equilibrated with 5 mm caffeine generate contractile activity even in the presence of 5 mm La^s+ and 50 mm outer K⁺. Even 100 mm-K⁺ does not hinder completely the de novo generation of an oscillation in caffeine-treated drops. Oscillatory phenomena cannot be registered in the combined presence of 5 mm caffeine and 70 mm-K⁺. Likewise no oscillation appears when high K⁺ is replaced by high Na⁺ (Fig. 6d).

Evaluating these data one can conclude that oscillation seems to be widely independent of a contraction triggering plasmalemmal Ca²⁺ pump (Ludlow & Durham, 1977). There are no data favouring the view that a rhythmical Ca²⁺ transport across the plasmalemma triggers the oscillation.

The same can be concluded with respect to the involvement of a hypothetical Na⁺-K⁺ pump. The ineffectiveness of ouabain on persistence and de novo generation of activity and on ATPase activity of Physarum homogenates (Achenbach & Achenbach, 1979) ^rules out any role for the Na+-K+ pump in triggering oscillations., Caffeine obviously prevents the formation of the plasmalemma invagination. perhaps by interfering with Ca²⁺-sensitive membrane sites which may influence the Anchoring of actin molecules to the plasmalemma and/or blocking plasmalemmal phosphodiesterase activities. Only speculations are possible at the present time about the molecular action of heavy water (conformational change of membrane proteins and interaction with structured water which surrounds membranes). In muscle, D₂O uncouples excitation from contraction by affecting inner calcium stores (Kaminer, 1977).

The possible existence of a proton pump should be mentioned. However, according to pilot experiments, 10 mm KSCN and 10⁻⁶ M valinomycin did not inhibit the initiation of contractile activity in invagination-free drops (Gotz von Olenhusen & Wohlfarth-Bottermann, 1979(b) and unpubl. results). Experiments depicted in Fig. 5 (d) also argue against an oscillation triggering role for some hypothetical proton pump (H+-HCO₃⁻ or H⁺-K⁺) (Daniel & Eustace, 1972; Matveeva et al. 1978).

In summary, we can probably exclude calcium and Na⁺-K⁺ ion pumps as trigger mechanism on the plasmalemma. This does not preclude the existence of some other unknown mechanism located in the plasmalemma or within the peripheral cell cortex. The role of external Ca²⁺ as a force modulator in the sequence of chemotactic responses remains to be analysed. The bioelectrical phenomena observed to date (Meyer & Stockem, 1979) do not prove rhythmic plasmalemma ion pumps as a trigger, since the phenomena may either reflect intracellular ion fluxes or represent a consequence of contraction activity (Kishimoto, 1958a, b): this activity continues also when streaming activity is restrained (Wohlfarth-Bottermann, 1975 a) and thus can also be present under conditions when there is no spontaneous streaming. The strict dependence of oscillatory behaviour upon bioelectric trigger phenomena in the plasmalemma could only be shown by demonstrating that intrinsic contractile activity ceased when these electrical signals in the plasmalemma were blocked experimentally.

To further analyse the role of the surface membrane it would be very important to determine whether oscillations would occur in strands whose plasmalemma is completely destroyed by detergents. Teplov et al. (1978) recently described that oscillation could be revived in Tween-treated strands. When trying to reproduce these interesting results in our laboratory, we confirmed that the Tween-treated veins behaved as described by Teplov et al. (1978). However, when we put the Tween-treated ‘models’ on to nutrient agar, at the end of the experiment, small parts of the tensiometrically measured veins developed into a new microplasmodium 12 h later. This means that the revival of oscillations supposed by Teplov et al. (1978) was perhaps not performed with a real ‘cell free’ system, but that the revival of contractile activity may be traced back to small areas of the veins possibly remaining unaffected by the detergent. In any case, in our laboratory (Achenbach, unpublished), it was not possible to revive oscillations in those veins which afterwards did not grow to new microplasmodia, i.e. in veins which were in fact ‘dead’ and thus in fact a cell free system.

The caffeine-treated protoplasmic drop, which lack plasmalemma invaginations thus allowing free access to external solutions, may prove a useful experimental model for future measurements of ion fluxes and electrical phenomena. It may be easier to identify the precise role of the plasmalemma in generating oscillatory activity using this simpler plasmodial stage. A recent observation in our laboratory (F. Achenbach, Unpublished) that invaginations reappear after removal of caffeine underlines the vitality of this stage and presents the opportunity of studying the contribution of these invaginations to oscillatory activity.

I wish to thank Prof. H. C. Lüttgau (Bochum) for helpful discussions and Drs R. L. Snipes (Giessen) and M. J. Berridge (Cambridge) for reading the manuscript.