# Popular Science Monthly/Volume 82/February 1913/The Role of Membranes in Cell-Processes

 THE RÔLE OF MEMBRANES IN CELL-PROCESSES
By Professor RALPH S. LILLIE

UNIVERSITY OF PENNSYLVANIA

THE importance of membranes in vital processes has long been recognized. From the earliest times anatomists have been impressed with the frequency with which thin sheets of solid material occur as elements of structure in organisms. Even elementary methods of analysis show that the materials composing the most various organs often tend to dispose themselves in thin, continuous layers. Thus the entire body is enclosed in an extremely resistant and impermeable layer, the skin. Each of the internal organs has its own characteristic enclosing membrane; the peritoneum lines the body-cavity and invests the intestine and its associated glands, the heart is enclosed in the pericardium, the lungs in the pleura, the central nervous system in the pia mater; the muscles are closely surrounded by thin connective tissue sheaths, or perimysia; the walls of the blood-vessels and of the intestine and other hollow viscera consist of several distinct concentric layers. Various products of animals, like the eggs of birds and reptiles, often show this tendency. Plants also deposit a great part of their structural materials in layers; the wood forms concentric circles; leaves and fruits have thin and often waterproof membranous coverings; the orange is partitioned by a system of membranes and each smaller portion of pulp has a membrane of its own. The instances, in fact, are innumerable. Evidently the tendency to deposit material in thin continuous sheaths is highly characteristic of organisms.

This much was clear at a time when anatomists were limited to direct and unaided vision. When the microscope came into use the existence of a similar tendency soon became evident in the minutest tissue elements. The living substance exhibited itself everywhere as minutely subdivided by innumerable thin partitions, or membranes, giving it a characteristic honeycomb-like or cellular structure. These partitions isolate the enclosed portions of living substance and render them at least mechanically separable. Hence the conception that each of these minute membrane-enclosed masses of gelatinous or viscid "protoplasmic" material is an independently living entity, or elementary physiological unit, gained ground, and, as all know, has been universally adopted in biology. The name "cell," originally applied to the minute spaces themselves, has been transferred to the protoplasmic mass within, by whose activity the enclosing membrane is itself formed.

Thus it was early recognized that cells tend to separate materials from their surfaces and deposit them in the form of definite coherent layers or membranes. Similar membranes may also be formed in the cell-interior. Of these, the best known is the nuclear membrane. Hence, in considering the general organization of the cell, cytoplasm and nucleus are usually described as bounded by definite structurally distinct layers, plasma-membrane and nuclear membrane. Vacuole-membranes, sphere-membranes and plastid-membranes may also exist in certain cells. To all of these structures it has been customary to ascribe a more or less mechanical or simply protective or isolating function. On the other hand, many cells show no optically distinguishable membranes, either at their surfaces or in their interior; certain ameboid cells and the blood-corpuscles of vertebrates are apparently without membranes and are often described as "naked masses of protoplasm." Yet in such cases the nakedness is only apparent, for it can readily be shown that these cells have membranes which are highly definite in character, but whose existence can be demonstrated only by certain forms of physiological experimentation.

The membranes whose physiological rôle forms the subject of this article are not to be identified with those more or less conspicuous layers separated at the surfaces of many animal and plant cells. The cellulose membranes of plant cells and the various cuticular structures of animal cells are dead structures, whose function is typically passive and mechanical. They are to be sharply distinguished from the membranes about to be considered, whose rôle is a characteristically active one, and, as I believe, fundamentally important in the life of all cells. These membranes are present in all living cells without exception, whether a visible external layer is present or not. Thus red blood corpuscles, though typically naked cells, show by their behavior in salt-solutions of varying concentration that they are bounded by a difficultly permeable surface-layer which is different in its physical properties from the internal protoplasm—having in fact the essential properties of a semi-permeable membrane. Plant cells, like those of Spirogyra, also behave in such solutions as if the surface-layer of the protoplasm were semi-permeable; the visible cellulose membrane plays no part whatever in the osmotic process (plasmolysis) observed under such conditions, while the invisible surface-film of the protoplasm is all-important. Hence in the case of plant cells the conceptions of cell-membrane—i. e., the hardened secretion of cellulose—and of plasma-membrane—or semi-permeable surface layer of the living protoplasm—have to be kept sharply distinct. It is the plasma-membrane, the most external layer of the living protoplasm, with which I shall be chiefly concerned in the present article, and I propose to discuss briefly various questions which arise in reference to this structure: what is its physical and chemical nature? what are the conditions of its formation? and how does it influence the characteristic activities of the cell?

Before dealing with the case of the living plasma-membrane, it will be necessary briefly to describe some of the methods by which artificial membranes, similar in many of their osmotic and other physical properties to plasma-membranes, may be prepared, and to consider some of the properties of these membranes. The tendency of the living cell (or living system) to surround itself with a membrane will then be seen to be in no sense a distinctively vital peculiarity, but one which it exhibits in common with a great many non-living systems. There is little doubt that the formation of membranes at the surface of small masses of living protoplasm is a particular instance of the general class of phenomena known collectively as "surface-processes"—processes, that is, occuring as manifestations of the special form of energy, surface-energy, which resides at the surface of separation between materials which do not freely intermix. Consider any material system consisting of various kinds of matter in various states of aggregation, i. e., what the physical chemists call a heterogeneous system, or a polyphasic system. Such a system may be analyzed into a certain number of components, each of which is physically and chemically homogeneous. Each such component is a phase. Oil-drops in a permanent emulsion form one phase, the water a second phase, the soap films at the surface of each droplet a third. Living protoplasm is a good instance of such a polyphasic system. It is—at least in certain forms, e. g., the protoplasm of egg-cells—an emulsion-like or foam-like mixture consisting of various fluid droplets or alveoli (which are supposedly droplets of oil or other fluid containing dissolved substances), separated by another fluid which is typically an aqueous colloidal solution of proteins and lipoids with various additional substances—salts, sugars, amino-acids—in solution. Each droplet or alveolus is a phase; so also is each colloidal particle, or each surface-film, or the interstitial suspension-medium or solvent. At the surface of contact between any two phases a certain tension exists, acting tangentially to the surface; this is the "surface-tension" which (if positive in value) tends to minimize the area of the surface. Each surface, or phase-boundary, is thus the seat of a particular form of energy, surface-energy, of which the intensity-factor is the surface-tension (T), the capacity-factor the total area of the surface (A). The total surface-energy (E) resident at any surface thus equals TA. The tension varies according to the nature of both of the contiguous phases: for water in contact with air it is ca. 75 dynes per linear centimeter, i. e., the pull of a ribbon-shaped portion of water-surface one centimeter wide is about one twelfth of a gram; for water in contact with oil it is ca. 23 dynes per centimeter; for oil in contact with air it is ca. 33 dynes. Now the distribution of the substances present in any such system is influenced in a remarkable manner by these surface-energies. Every one is familiar with the fact that oil spreads over the surface of pure water. This is a case in point: why does the oil not simply float in droplets on the surface instead of spreading out in an extremely thin continuous layer? A consideration of the conditions of surface-tension at once explains this (Fig. 1). An oil-drop placed on the surface of water is subjected to the pull of three tensions, viz.: those at its own two surfaces (${\displaystyle t_{1}}$ and ${\displaystyle t_{2}}$), Fig. 1. where it touches air and water, respectively, and which tend to round it off, and that of the pure water at its margin (${\displaystyle t_{3}}$), which tends to spread it. But the tensions ${\displaystyle t_{1}}$ and ${\displaystyle t_{2}}$ are together less than the tension ${\displaystyle t_{3}}$; the oil is thus rapidly drawn out over the surface by the superior pull of the water-air tension at its margin. Hence the water-air surface, that with high tension, disappears and is replaced by a surface with lower tension. The total surface-energy has been diminished, part having been transformed into mechanical energy and heat. If, instead of the case of a floating oil-drop, we take that of some soluble substance which is produced locally within the water near the surface—e. g., a soap or a protein, a solution of which has a lower tension than pure water—we find essentially the same phenomenon; the substance is spread out over the surface, and this effect will continue so long as the addition of further quantities of the substance to the surface-layer continues to lower the surface-tension. The end-effect will be to concentrate the substance at the phase-boundary. This phenomenon is the expression of a general law, the law of Willard Gibbs and J. J. Thomson, which describes the part played by surface-energies in the distribution of soluble substances in a polyphasic system. In the present case, the process of surface-concentration will go on until some equilibrium is reached, e. g., where the loss of substance from the surface by diffusion balances its collection there under the influence of the surface-energy. But in many cases, as with proteins, soaps and certain lipoids, the substance separates at the surface as a continuous solid film before this stage is reached. The formation of solid surface-films is hence highly characteristic of the solutions of such substances. Casein films form on warm milk, soap films form about droplets of rancid oil in the presence of alkali, and protein films about drops of chloroform or oil suspended in protein solutions. Thin solid membranes formed in this manner at phase-boundaries are called "haptogen membranes." In all of these instances we have to do with a surface-condensation, known under certain conditions as "adsorption," of substances which lower the surface-tension at the phase-boundary. Among the colloidal constituents of protoplasm the proteins and the lipoids belong to this class of substances. Hence it is not surprising that isolated portions of living protoplasm should delimit themselves by membranes. The various cell-membranes are to be regarded as essentially surface-films, or haptogen membranes. Not only do such thin films form about the reconstituting nuclei of dividing cells, but they are also deposited about various cell-inclusions, and even about division-spheres, chromatophores and other cell structures under certain conditions. It is well known that portions of protoplasm cut off from living cells—such as egg-cells, protozoa, root-hairs, etc.—exhibit the same osmotic properties as the intact cells, showing that new semi-permeable membranes are quickly formed at the cut-surfaces.

The surface of contact of the living substance with its medium thus becomes the seat of deposition of certain protoplasmic constituents or products which form membranes, often of a high degree of impermeability. This impermeability is a property of fundamental physiological importance. Speculation on the evolutionary origin of living cells usually leads to little result, but we may at least infer that the early protoplasmic systems which survived and became the ancestors of living organisms must have consisted in part of colloids like proteins and lipoids which had the property of forming surface-films sufficiently impermeable to limit or prevent free diffusive interchange with the surroundings. Only systems thus isolated to a sufficient degree from the surroundings could preserve the requisite complexity and constancy of composition, and hence be enabled to develop the properties of so-called living beings—properties which are so widely different from those shown by other natural systems. The surface-films, or plasma-membranes, of living cells at the present time are in fact typically characterized by a remarkably high impermeability to simple crystalloid substances like sugars, neutral salts and amino-acids, all of which are important constituents of protoplasm. Zangger expresses the situation concisely when he says that living cells can contain as permanent constituents only such substances as are not free to diffuse into the surrounding medium. The existence of this diffusion-preventing or insulating surface-film, the plasma-membrane, is thus a necessary condition of the stability of the living system and hence of the continuance of the life-processes. The living condition is in fact incompatible with marked and permanent increase in surface-permeability. During life the semi-permeable condition is retained; on death there is always a marked increase in the permeability of the plasma-membrane; the cell then undergoes a ready and rapid dissolution or cytolysis, and the constituents serve as food to bacteria. It is probable that the various intracellular membranes—nuclear membranes, vacuole-membranes, sphere-membranes, chromatophore-membranes—subserve a similar insulating or differentiating function. Hofmeister has indeed conceived of the protoplasm of living cells as subdivided in this manner into a many-chambered system, which accordingly permits of a high degree of chemical differentiation. A variety of independent processes might coexist side by side in such a system, as appears, for example, to be the case in liver-cells; in this way a "chemical organization," distinct from and yet dependent upon a structural organization, becomes possible.

Haptogen membranes formed thus by deposition of proteins at phase-boundaries may show considerable density and impermeability. The protein in such surface-films may undergo an alteration resembling coagulation, assuming a relatively resistant and insoluble form, Thus Ramsden was able to coagulate protein solutions by prolonged shaking, and Robertson obtained thin films of coagulated casein, gelatine and protamine at the surface of chloroform droplets. Solid films of albumose, saponin, and other substances are formed at the free surfaces of their solutions—the readiness with which such solutions are thrown into foams depends in fact on this condition. The condensed and insoluble protein films formed on chloroform droplets are strikingly similar in many respects to those visible at the surfaces of cells like sea-urchin eggs, and which apparently correspond to the outer layer of the true plasma-membranes.

To come now to more directly biological considerations: what is the nature, chemical and physical, of the surface-film of living cells? There are few direct chemical analyses bearing on this question. Liebermann found the vitelline membrane of the hen's egg to consist largely of a keratin-like albuminoid. There is good reason to believe that modified proteins belonging to this class enter very generally into the composition of the surface-films of cells. The tendency to deposit horny or albuminoid material at the cell-surfaces is in fact remarkably widespread in animals. Cuticular and epidermal structures, to which chemical resistance and impermeability are physiologically essential, consist typically of proteins belonging to this class; such proteins have recently been called "scleroproteins" on account of their frequent presence in skeletal or cuticular structures. They are also abundant in the intercellular materials of bone, cartilage and connective tissue. The surface-films of many cells apparently have this composition. Thus in echinoderm eggs the characteristic fertilization-membranes, which Professor Jacques Loeb has shown to arise by separation of a surface-film, consist apparently of modified protein. They are at least non-lipoid in character and are remarkably resistant to reagents, resembling in these respects the protein films formed at the surface of chloroform droplets. The fertilization-membrane, after separation from the cell, proves however to be much more permeable than the true plasma-membrane, or semi-permeable external layer of the unaltered egg, so that it probably corresponds to only a portion—probably the outer layer—of this membrane. The presence of protein in the plasma-membrane of, sea-urchin eggs is also indicated by the fact that the cytolytic action of acids may be lessened or counteracted by neutral salts like sodium or calcium chloride. Such antagonistic actions between acids and salts, while not shown by colloids in general, are peculiarly characteristic of certain proteins. Thus the rate of swelling of gelatine (a typical scleroprotein) in water is greatly increased by the addition of a little acid; this effect is prevented by the addition of neutral salts, and the basis of this form of anti-cytolytic action may possibly lie here—i. e., the disruptive action of the acid on the proteins of the membrane is checked or prevented by the salt.[1] Yet the plasma-membrane undoubtedly contains other constituents, and among these the substances belonging to the group of lipoids appear to be fundamentally important. These substances, fat-like in their solubilities and colloidal in their physico-chemical character, are always present in cells. Much light has been thrown on their physiological significance by the investigations of Overton and his successors, which have shown that ready permeability to lipoid-solvents is highly characteristic of both animal and plant cells. Alcohols, esters, ethers, hydrocarbons and similar compounds, all of which are soluble in lipoids, enter living cells rapidly, in contrast to neutral salts, sugars, amino-acids—the chief crystalloidal constituents of protoplasm—which diffuse into resting cells (with unmodified plasma-membrane) either imperceptibly or with extreme slowness. Overton's results thus indicate that lipoids enter into the composition of the plasma-membrane. This is to be expected. The structure probably consists of a mixture of all those protoplasmic constituents which have marked effect in lowering the surface-tension of the cell-boundary. Lipoids are conspicuous among this group of substances. That they form part of the plasma-membrane is also indicated by the readiness with which the permeability and other properties of this structure may be altered by lipoid-modifying substances. Lipoid-solvents as a class, when present in certain concentrations, have a specific action in increasing, often irreversibly, the permeability of the plasma-membrane. In lower concentrations many appear to decrease this permeability. Their influence on irritability, which is probably a function of the condition of this membrane, also indicates their importance as membrane-constituents. Narcotic action is highly characteristic of lipoid-solvents, and there is good evidence that this action depends on an alteration of the plasma-membrane. I shall refer to this possibility later, in connection with the problem of the relation of membranes to stimulation. All of these facts taken together indicate very clearly that the colloids composing the semi-permeable surface-film of living cells consist of both lipoids and proteins, which are probably intermixed or combined in some characteristic manner and vary in their relative proportions in different cells, according to the specific constitution of the latter.

What are the chief peculiarities in the physical properties of these membranes, on which their physiological importance depends? Two properties appear especially significant. One of these is the semi-permeability which the membranes preserve during life, i. e., the ability to transmit water freely while holding back dissolved substances. The other is their ability to undergo reversible changes in their permeability to such.substances, either in the direction of increase or decrease. These changes of permeability may in some cells be very rapid; and there is evidence that this is especially the case with irritable tissues, and that the power of rapid response to stimuli is directly dependent on this peculiarity. How essential the semi-permeability of the plasma-membranes is to living organisms may be realized with especial clearness in the case of plants. In many of these organisms the rate of growth, the normal form and habit, and the characteristic movements and reactions are intimately dependent on the peculiar condition known as turgor, which is the expression of the outward pressure of the dissolved molecules of the cell-contents against the membranes which enclose them and which they can not pass. The diffusing molecules hence press against these membranes, often with the force of many atmospheres, and keep the cellulose cell-walls stretched and rigid. It is on this condition that the maintenance of the normal form often depends. The entrance of the water into the cell in growth is also largely due to this osmotic pressure. Thus the confinement of the molecules within the cells by membranes impermeable to their outward diffusion is an indispensable condition of the continuance of normal life-processes in these organisms. The same is true of animal cells, although here the condition of turgor is usually unimportant in itself. But, as we have already seen, the preservation of the normal protoplasmic composition in the case of any cell involves the prevention or restriction of any free or unselected diffusive interchange of materials between the cell and its surroundings. The semi-permeability found during life is the expression of the all-importance of this condition. We must therefore ascribe to the insulatory or semi-permeable character of the plasma membrane, not only the existence of conditions like turgor in plants, but even the very possibility of the existence of a stable or permanent chemical organization in any cell.

This being the case, it is not surprising to find that simple modification of permeability may profoundly modify many cell-processes. To take first a relatively simple instance: if the semi-permeability of the plasma-membrane is a necessary condition for continued life in any cell, it ought to be impossible to increase this permeability beyond a certain limited degree for any length of time without inflicting permanent injury on the cell and eventually causing death. Loss of essential cell-constituents through the altered membrane would have this effect. Now there is evidence that a large class of injurious or toxic substances exert their destructive action by altering the surface-films of cells and permanently increasing the permeability. When this occurs in such a cell as a blood-corpuscle or a sea-urchin egg—which is normally in osmotic equilibrium with its medium—the cell first swells (an effect showing loss of osmotic equilibrium) and eventually dissolves or disintegrates, an effect known as cytolysis. Lipoid-solvents, like chloroform or ether, have this effect in concentrations above certain minima: they disrupt the membrane, presumably by altering the condition of the lipoids, and disintegration follows. Many toxic alkaloids and glucosides—like saponin, digitalin, aconitin, etc.—and certain bacterial products—cytolysins and hemolysins—have similar effects. Other substances, as inorganic salts, acids, or alkalis, may cause cytolysis by altering the state of the colloids of the membrane. In certain typical instances there is direct evidence that the toxic action is primarily due to a surface-alteration, and consists in a destruction of the semi-permeable properties of the membrane. Certain fluorescent substances like eosin exert a cytolytic action on many cells in the presence of light, though inactive in the dark (photodynamic action). Harzbecker and Jodlbauer found that blood-corpuscles so treated began to swell before there was any perceptible entrance of the dye into the cell, i. e., the initial stage of cytolysis, involving a loss of osmotic equilibrium, occurred previously to the entrance of the toxic substance. But loss of osmotic equilibrium, unless soon reversed, involves destruction of the cell. The essential or critical toxic action in this case is thus superficial, and what is true of eosin is probably true of many other—possibly most—cytolytic substances.

The peculiar antagonisms existing between the physiological actions of various substances (e. g., muscarin and atropin, toxin and antitoxin, etc.) are probably in many cases to be explained on this basis. The toxic and antitoxic actions of neutral salts form a case in point. Pure solutions of sodium salts, even sodium chloride, are strongly toxic to many cells, particularly those of marine organisms, as the work of J. Loeb and his successors has shown with especial clearness; but if to the pure solution a little calcium salt is added, this toxic action is prevented or greatly diminished; the calcium (or other favorable salt) counteracts the toxic action of the sodium salt—in other words, has an antitoxic action. Now it can readily be shown in certain organisms that the toxic action of the pure sodium salt solution is associated with a strong permeability-increasing action. As test-objects or physiological indicators in the investigation of these effects I have used the pigment-containing eggs of the sea-urchin, Arbacia, and the larvæ of a marine annelid, Arenicola, whose cells contain a water soluble yellow pigment. The eggs or larvae die rapidly in pure isotonic solutions of sodium salts, and this toxic action is associated with a loss of pigment (more or less rapid according to the particular salt employed), i. e., with a marked increase in permeability. But if a calcium or other antitoxic salt is previously added to the solution, both the permeability-increase (as indicated by loss of pigment) and the toxic action are prevented or greatly retarded. Apparently, a pronounced and persistent permeability-increasing action is equivalent to a toxic action; the calcium' prevents or retards this destructive action of the sodium salt on the plasma-membrane, and hence has an anti-cytolytic or antitoxic effect. Professor Osterhout's experiments disclose similar conditions in plant cells; pure solutions of sodium chloride increase permeability—as shown by loss of turgor and increase of electrical conductivity—and have a well-marked toxic action; both of these effects may be prevented by adding a little calcium to the solution. In all of these cases the antitoxic action apparently consists in protecting the surface-film against the permeability-increasing action of the pure sodium salt solution. I have found that not only salts of metals, like calcium and magnesium, but also various lipoid-solvents or anesthetics may prevent the cytolytic action of pure solutions of sodium salts in an essentially similar manner. Evidently certain changes in the state of the lipoids in the membrane render the latter more resistant to the disruptive action of the salt solution. Cytolysis by substances like saponin may also be checked by neutral salts. It seems probable that the relations between bacterial cytolysins and anti-cytolysins are of the same essential nature. The theory of antagonistic salt-actions may thus become of the greatest importance as a guiding principle in practical therapeutics. Such surface-actions as those just described constitute only one form of toxic action, but they are among the most important because of the external position of the plasma-membrane in the cell and its consequent direct accessibility to modification by changes in the surroundings.

The integrity of the plasma-membrane thus appears to be essential to the normal living cell. Injury to this membrane thus means toxic action: prevention of this injury is antitoxic action; restoration of the normal permeability after injury is therapeutic action. But the plasma-membrane does not play only the purely passive role so far indicated. It is intimately concerned in many active cell-processes; and there is evidence that many of the distinctive energy-manifestations of the cell are determined or controlled by changes—largely changes of permeability—which have their seat in this structure. This appears to be true of many forms of cell-movement, of cell-division, and of the stimulation-process in general. Permeability-changes are also concerned in secretion, in the fertilization of the ovum, and probably in the general process of intake of food-materials by cells. The stimulation of irritable tissues is a process which exhibits a peculiarly intimate dependence on the semi-permeable membranes of the irritable elements. Perhaps more is known of the relations of membranes to the stimulation-process than to any other cell-activity, and I shall accordingly consider its conditions in some detail.

There is evidence that a rapid and reversible increase in the general permeability of the plasma membrane is an accompaniment and indeed a primary condition of stimulation in irritable tissues. This evidence comes from many sides and is partly direct and partly indirect. Perhaps the clearest indications of this kind are afforded by the motile mechanisms of certain plants, like the sensitive plant (Mimosa pudica) or the Venus's fly-trap (Dionæa). In Mimosa the characteristic movement, which consists of a dropping of the leaves and a folding together of the leaflets, is due to a collapse of certain turgid cells which form the so-called pulvini, or cushion-like masses of parenchyma at the bases of the leaves and leaflets. A fluid containing dissolved substances rapidly leaves these cells on stimulation; evidently the membranes, semi-permeable during rest, become suddenly permeable to the osmotically active intracellular substances which maintain the turgor. This explanation—first put forward in its essentials by Sachs—is accepted by the majority of plant physiologists, and there is little doubt of its substantial correctness. We have here, therefore, an instance where stimulation depends directly upon a sudden increase in permeability. Now in this case the primary or critical change is apparently the same as in irritable animal tissues; an electrical variation similar to that shown by an active muscle or nerve accompanies the movement, and the conditions which call forth the response are essentially the same in the plant as in the animal. In the case of animals the evidence that increase of permeability is a condition of stimulation is, as a rule, less direct. Yet in certain organisms a sudden increase of permeability may readily be shown to be the equivalent of stimulation. My own observations on the pigmented larvæ of Arenicola illustrate this very clearly. When these organisms are suddenly brought from sea-water into pure isotonic solutions of sodium salts (e. g., ${\displaystyle {\ce {m/2NaCl}}}$) the muscles contract with extreme vigor and persistency, causing the larvae (which are small worm-like trochophores about 0.3 millimeter long) to shorten to half their normal length; at the same time the yellow pigment contained in the cells of the organism diffuses into the solution and colors the latter yellow. The exit of pigment is the expression of a rapid permeability-increasing or cytolytic action; this is equivalent to a strong stimulation. If by the addition of any substance to the solution we check or prevent this permeability-increase, we find that stimulation is checked or prevented at the same time. Thus, if instead of the pure ${\displaystyle {\ce {m/2NaCl}}}$ we use ${\displaystyle {\ce {m/2NaCl}}}$ to which a little calcium or magnesium chloride, or other appropriate salt, has been added, the strong stimulation and loss of pigment are no longer seen—both are simultaneously prevented. The same effect may be produced by various anesthetics; these also protect the cells against the permeability-increasing action of the ${\displaystyle {\ce {m/2NaCl}}}$, and at the same time prevent stimulation. Thus, if Arenicola larvae are exposed for a few minutes to an isotonic solution of a magnesium salt and are then brought into ${\displaystyle {\ce {m/2NaCl}}}$, neither stimulation nor loss of pigment follows. The same is true if they are brought from ether-containing sea-water into ether-containing ${\displaystyle {\ce {m/2NaCl}}}$; and other anesthetics in appropriate concentrations show a similar inhibitory and protective action. These and similar experiments point to the conclusion that a membrane-alteration, in the direction of rapid increase of permeability, is constantly associated with stimulation. It is of course apparent that such increase in permeability must in normal stimulation be perfectly reversible. If the reversibility is incomplete, permanent injury results; and this is in fact the case when Arenicola larva? are stimulated by immersion in pure isotonic sodium salt solutions. We have already seen that this injurious action, as well as the stimulating action, is greatly diminished by the presence of calcium chloride, or some other antitoxic salt. Anesthetics also show an antitoxic as well as an anti-stimulating action.

It is impossible within the limits of this article adequately to discuss the physiology of stimulation. A few of its aspects ought, however, to be touched on here, since otherwise the above relation between permeability-increase and stimulation may appear as a merely empirical or detached observation, without any general or theoretical significance. The most striking physical peculiarity of irritable tissues is their sensitivity to electrical changes in their surroundings. Most persons are accustomed to think of electrical currents as laboratory phenomena par excellence, and as playing little part in nature outside of laboratory walls. Yet living cells are profoundly influenced by such currents. We can in fact imitate the normal conditions more closely by using electrical currents as stimuli, than in any other manner. This preconception is however a completely mistaken one. Not only do irritable tissues respond to electrical currents, but certain electrical changes in the tissues themselves are invariably associated with stimulation, whether normal or artificial, and form perhaps the most constant and essential feature of the stimulation-process. Such a statement may sound like a truism to any one versed even slightly in modern physical chemistry: ions—charged molecules and atoms—are present everywhere in protoplasm, and it would perhaps be surprising if electrical changes did not accompany protoplasmic activities. We have, however, to inquire more particularly into the nature and conditions of the response of irritable tissues to the electrical current, and of the electrical processes originating in the tissues themselves, and to relate these processes, if possible, to the total effects produced by stimulation.

These processes again, like some of those already referred to, appear to be a function of the changing permeability of the plasma-membrane. When we take a tissue consisting of a parallel bundle of cells, like a frog's sartorius muscle, cut it across, place one electrode in contact with the normal uninjured surface of the muscle, and the other with its cut surface, and connect the two with a galvanometer, we find that an electrical current passes—the so-called demarcation-current. The exposed interior (or cut surface) of the cells always shows a lower potential than the exterior; the potential-difference lies usually between a tenth and a twentieth of a volt. This potential-difference depends on the living condition of the cells. It is absent or insignificant in dead muscle. It diminishes when the muscle-surface is treated with cytolytic substances—i. e., with substances which increase the permeability of the plasma-membrane. The evidence, in fact, indicates that the existence of a normal demarcation-current potential is dependent on the semi-permeability of the plasma-membrane. When the permeability is artificially increased, the potential-difference is invariably diminished; its degree thus appears to be dependent on the degree of permeability of the membrane; hence its increase on death or under the influence of membranolytic substances. Now during stimulation the demarcation-current potential always undergoes a marked decrease; this is the change known as the negative variation or action-current, which is an inseparable accompaniment of stimulation. Normally, this change is completely reversible, and when stimulation ceases, the original potential-difference is regained. What is significant from the present point of view is that the direction of the electrical variation accompanying stimulation is the same as in that resulting from death or cytolytic action and associated with an increase of permeability. The phenomenon is thus intelligible on the assumption that during stimulation there is a sudden and marked increase in the permeability of the plasma membrane. This permeability increase, with the accompanying electro-motor variation, differs from that associated with death or cytolysis chiefly in being rapidly and completely reversible. Stimulation may, however, be so excessive under some conditions as to lead to irreversible alterations in the membranes, or even to the death of the cell; i. e., the degree of reversibility is limited, and this consideration explains why excessive stimulation is so injurious—it is in effect equivalent to a cytolytic action or any other action where permeability is irreversibly increased.

Why should a change in the permeability of the membrane produce electrical effects of this kind? The phenomenon becomes intelligible when we remember that membranes act by limiting or preventing diffusion, and that they may limit the diffusion of ions—the mobile, electrically charged atoms and atomic groups present in salt solutions—just as they do that of uncharged molecules. The ions formed by the dissociation of any electrolyte have as a rule unequal diffusion-velocities, and presumably unequal solubilities and other physical properties, in correspondence with their chemical differences; and hence we may infer that they possess unequal abilities to pass through membranes. If this is so, a membrane separating two electrolyte-solutions becomes the seat of a potential-difference; i. e., a potential-difference, which may be considerable, will exist between its opposite faces. This suggestion, first made by Ostwald in 1890, has formed the basis of the chief prevailing view—the so-called "membrane-theory" of the nature of the bioelectric processes. Ostwald's suggestion, modified to suit the conditions Fig. 2. Illustrating the supposed conditions of polarization of the plasma membrane. The electrolytes are lactic and carbonic acids; the membrane is supposed to be permeable only to H-ions. in cells, was essentially as follows. Imagine the cell enclosed in a plasma-membrane freely permeable to the cations (positive ions, e. g., hydrogen ions or potassium ions) and impermeable to the anions (negative ions) of a certain electrolyte (which we may suppose to be lactic or carbonic acid) contained in the protoplasm (Fig. 2). The cations then pass outward, carrying their positive charges, while the anions remain behind; this will proceed until the potential-difference thus arising is sufficient to compensate the diffusion-tendency (equivalent to the osmotic pressure) of the cations. A condition of equilibrium with outer surface positive and inner negative thus results. The membrane becomes the seat of an electrical polarization (normal or physiological polarization) which is dependent on its impermeability to anions. If the permeability of such a membrane were to increase sufficiently to transmit the anions, a fall of the potential-difference between the exterior and the interior of the cell would at once follow. An effect of just this kind is seen in muscle and nerve during stimulation, and is attributed by Bernstein and other upholders of the membrane-theory to the changing ionic permeability of the membrane. The selective permeability to ions of different sign, on which the potential-difference between exterior and interior depends, disappears along with the general increase in permeability accompanying stimulation: hence a negative electrical variation is always associated with this process. The precise arrangement imagined by Ostwald has not yet been satisfactory realized, although, according to Brünings, precipitation-membranes of copper ferrocyanide show sorre of the properties required by this theory. But certain natural membranes present a much closer approach to the theoretical requirements; thus the surface-membranes of apples, which Beutner and Loeb have recently studied, behave as if decidedly more freely permeable to cations as a class than to anions, and it is possible that this condition is typical for the plasma-membranes of cells. The membranes of irritable tissues, however, may belong to another type; certain membranes (consisting of thin films of glass) whose electrical polarization depends on the relative hydrogenion concentrations in the solutions which they separate, have recently been investigated by Haber; and in some respects the phenomena presented by these membranes appear to correspond more closely to the conditions in irritable tissues. Hydrogen-ions would be the polarizing cations in the case of these membranes; and in fact irritable tissues are as a rule remarkably sensitive to changes in the H-ion concentration of their medium. We are not yet in a position to decide between such alternatives. But for the present purpose it is sufficient to recognize that a membrane which interferes unequally with ionic diffusion may become the seat of a potential-difference when it separates two solutions; and the evidence that plasma-membranes and other cellmembranes are of this kind appears very strong, even at the present time. In general, phase-boundaries are the seat of electrical energies, and these largely depend on the ionic content of the adjoining media. Membrane-polarization is a special instance of this general class of phenomena. The precise conditions of the normal physiological polarization in irritable tissues have to be determined by future investigation.

Membranes in their electrochemical aspect are to be regarded, on the present theory, as ion-transmitting surfaces, just as the metallic plates in ordinary electric batteries are ion-forming or ion-combining surfaces. The electrical properties exhibited by all of these surfaces are conditioned in essentially the same manner, and Nernst's theory applies to all. A system composed of solutions separated by membranes may thus, under the proper conditions, show the same essential properties as a system of batteries connected in series. The potentialdifferences of the individual elements may be summed by appropriate arrangement so that the electric tension between the terminals may be very large. In the electrical organs of Gymnotus and other fish, systems of this kind have actually been realized in nature, and have been applied to defensive or other purposes.

Let us now consider in a little more detail the conditions of stimulation of an irritable tissue by an external electric current. The surface-film of the muscle-cell or the nerve fiber is to be regarded as electrically polarized in the sense already indicated. Why does the tissue respond in its characteristic manner to the electric current? The first fundamental suggestion as to the mode of action of the current was made by Nernst in 1899. He pointed out that the current in passing through a living tissue—a system equivalent to a solution containing electrolytes and subdivided by semi-permeable membranes—can produce decided changes of condition only at the semi-permeable surfaces, where the movement of ions is blocked; changes of electrical polarization would be produced at such surfaces; ions of a given sign would be carried against one face of the membrane by the current and would concentrate there until the back-diffusion equalled the current-transport; the same effect, with the signs of the ions changed, would result at the other face (Fig. 3). He conceived that in Fig. 3. Illustrating the polarization of the current on a membrane difficultly permeable to ions. The anions and cations of the electrolyte, ${\displaystyle {\ce {NaCl}}}$, move in the direction indicated by the arrows. The current, passing from left to right, carries cations toward and anions away from the left face of the membrane; at its right face the conditions are reversed. The membrane thus becomes electrically polarized, with its left face at the higher potential. electrical stimulation something of the kind occurs. The essential or critical change occurs at the semi-permeable membrane, and consists in carrying to this membrane sufficient ions to produce a given ionic concentration-difference corresponding to a given electrical polarization. This is the determining condition of stimulation. A certain time will be required for the process, depending on the strength of the current, and on the specific diffusion-rate of the ions. Nernst estimated that on this hypothesis the stimulating action (S) of a given current ought to vary directly with its strength (i), and with the square root of its duration (t) (${\displaystyle S=Ki{\sqrt {t}},}$ K being a constant characteristic of the tissue). The experimental data show that a more intense current requires for stimulation a shorter time than a weaker current, and in approximately this proportion. The more recent work of Lapicque, Lucas and Hill has confirmed and amplified Nernst's theory. There is therefore strong evidence that a current stimulates by producing an electrical polarization at the membranes.

During life, however, the membranes are apparently already the seat of a pre-existent polarization, as we have seen. The polarization produced by the external current must, therefore, modify this. Now it appears that in most, if not all irritable tissues, stimulation results when the physiological polarization is diminished suddenly, but not when it is increased. This is the simple inference from the law of polar stimulation. When a current is passed through a tissue the external positivity of the irritable elements is lowered on the side directed toward the cathode and increased on the side directed toward the anode, as may be seen by reference to Fig. 3. Now it has long been known that the stimulus originates on the cathodal side of an irritable tissue when the current is made, and on the anodal side when the current is broken; i. e., we obtain stimulation when the preexisting polarization of the irritable elements is rapidly diminished—in other words, when there is a depolarization. We may formulate the essential relations thus: stimulation is equivalent to depolarization, i. e., to a rapid decrease of the already existing or physiological polarization of the plasma-membranes.

Stimulation, however, is also connected with a change in the permeability of the membrane, as we have seen. We must therefore conclude—since a sudden change of polarization stimulates—that simple alteration of the electrical polarization alters the permeability of the membrane. Decrease of the potential-difference between the opposite faces of the membrane—i. e., depolarization—apparently increases permeability, and often to a remarkable degree. Irritability seems, in fact, to be an expression of this peculiar relation. The electric current thus alters the polarization of the semi-permeable membranes of the irritable tissue, and in so doing alters the permeability. This change becomes the condition of the characteristic electrical variation of the tissue; the latter is self-propagating, and thus the effects of the local stimulus are transmitted to other regions of the cell. These appear to be the essential changes in the stimulation-process as such.

According to this point of view we must conceive of the plasma-membrane of an irritable element as possessing during rest a characteristic impermeability or semi-permeability to which corresponds a definite polarization, or potential-difference between its outer and inner surfaces, of the value of (e. g.) one tenth volt. Now the permeability of the membrane is determined by a number of conditions, some of which are, its chemical composition, the temperature, the chemical changes in the protoplasm and the surroundings, and probably the state of-mechanical tension of the membrane. Another factor is, however, of fundamental importance: this is the existing state of electrical polarization of the membrane. Alteration of this polarization alters the permeability; if we decrease it we increase the permeability and stimulation may follow; if we increase it we presumably alter the permeability in the inverse direction—hence in all probability the lowered irritability at the anode (anelectrotonus) during the passage of a constant current through a muscle or nerve. Such a view ascribes peculiar properties to the plasma-membrane, but the facts lead directly to this interpretation. Girard has shown experimentally that changing the electrical polarization of a membrane of bladder or parchment alters the permeability to neutral salts. The electrical state of a membrane may thus determine its permeability. The plasma-membrane of irritable tissues has apparently acquired extreme sensitiveness to changes in its electrical polarization, such that slight electrical disturbances in the surroundings may lead to a large increase of permeability, and hence to marked stimulation.[2]

On this hypothesis we can also understand why the state of excitation is transmitted from one region of the irritable element to another. It is highly probable that the effect of a local stimulation is propagated over the surface of the muscle-cell or nerve-fiber because of the electrical variation which the permeability-change at the excited region itself produces. This electrical variation affects the adjacent regions of the membrane, and alters their permeability, with corresponding electrical effects, and so the effect spreads. The explanation of the conduction-process in a nerve or other irritable tissue is on this view identical with that of the stimulation-process. There is, in fact, good evidence that the region in a state of excitation simply excites the adjoining regions electrically by means of its action-current, and that the effect is transmitted in essentially this manner.

It is possible to change the polarization of the membrane, and hence its permeability, in other ways than by passing an electrical current. Or we may alter the permeability directly, by acting on the cell by chemical substances, or by suddenly changing the temperature, or by mechanical action. When such treatment produces a sufficient increase of permeability, we may suppose that all ions become free to pass the membrane, and that a polarization-change then occurs, with consequent stimulation which, like other forms of stimulation, is self-propagating. On such a view the ordinary forms of mechanical and chemical stimuli are at bottom electrical in their nature. Such stimuli act by directly altering the permeability of the membrane and hence its electrical polarization.

On the other hand, the properties of the membrane may be so modified under certain conditions that it fails to respond to changes of polarization by changes in its permeability. This occurs, for instance, in narcosis. I have found that narcotics, in the concentrations at which they anesthetize the musculature of Arenicola larvæ, also check or prevent the permeability-increasing action of isotonic sodium chloride solution on the pigment-containing cells of this organism; at the same time they decrease or prevent the stimulating action of this solution. They also protect the organism against its toxic action, as we have already seen. An anesthetic action is thus the equivalent of both an anti-stimulating and an anti-cytolytic action. Both effects depend upon a modification of the plasma-membrane; under the influence of the anesthetic this structure becomes more resistant than normally to conditions that otherwise increase its permeability. We may infer in general that the degree of responsiveness of an irritable tissue is dependent on the state of its plasma-membranes; and that anesthesia corresponds to a condition of decreased susceptibility, and hyper-irritability to one of increased susceptibility, to the action of permeability-increasing agencies. Sensitization and desensitization, on this view, are primarily surface effects, dependent on alteration of the limiting membranes.

The polarization-changes accompanying stimulation may be extremely rapid in some cases. During the contraction of a man's voluntary muscle under the influence of the will, the existence of a rhythmical electrical variation with an average rhythm of about fifty vibrations per second has recently been demonstrated by the thread-galvanometer. The negative variation accompanying a single muscular twitch occupies from one hundredth to one two-hundredths of a second in a frog's voluntary muscle at ordinary temperatures; that accompanying a single nerve impulse lasts about one thousandth of a second; while more slowly reacting tissues, like heart-muscle or smooth muscle, show correspondingly slower electrical variations. On the membrane-theory the corresponding permeability-changes in the membrane must occupy similar times; and this consideration indicates the extreme delicacy of the adjustment between permeability and electrical polarization that must exist in the membranes of highly irritable tissues.

The electrical phenomena of stimulation are, however, relatively inconspicuous—if we except the case of the electric eel or torpedo. The characteristic and biologically important "response" of the tissue varies with its special nature. A muscle contracts, for instance; a gland secretes. The relation between the rapid change of polarization, which is the primary event in stimulation, and the resulting mechanical and chemical effects remains to be inquired into. The problem is a difficult one, and insufficiently investigated. The energy of muscular contraction is derived from the oxidation of energy-yielding compounds, like sugar. We must conclude that the polarization-changes at the cell-surface influence the chemical processes in the muscle-cell. Stimulation is known to increase many times the rate of oxidation in muscle-cells. I have lately attempted to modify the rate of formation of indophenol (a deeply colored organic oxidation-product) in the blood corpuscles of the frog by passing induction-shocks; and I find that the rate of formation of this compound through intracellular oxidation can be greatly accelerated by this means, especially in leucocytes, where the oxidation-rate is relatively rapid. I am inclined, therefore, to attribute to the variations in the electrical polarization of the membranes an important general role in varying the rate and possibly the character of the energy-yielding intracellular oxidations. On this view, intracellular metabolism would be largely controlled by membrane-processes. How this is possible may be illustrated by the case of anesthesia just discussed. The ether-impregnated plasma-membrane is relatively unaffected, as compared with the normal membrane, by isotonic sodium chloride solution; and consequently the stimulation, with its resultant increase in oxidation, is prevented by thus altering the membrane. The precise nature of the conditions in these and similar phenomena can be elucidated only by further study.

I had hoped to discuss the rôle of membrane-processes in other cellactivities, such as fertilization, cell-division and development, but the space at my disposal is insufficient. Before closing, however, I wish to refer briefly to the large class of physiological processes in which a regular rhythmical repetition of the same change, e. g., contraction, is the essential characteristic. Such processes include ciliary activity, the action of contractile vacuoles, the action of the heart and of nerve-cells like those of the respiratory center or the heart-ganglia of certain animals. In the division of cells during early development, a definite though slower rhythm is also seen. Now an electrical rhythm accompanies the physiological rhythm in muscle and nerve cells, probably in cilia, and almost certainly in dividing cells, as indicated by the experiments of Miss Hyde on dividing fish-eggs. The existence of a chemical rhythm—of carbon dioxide production—has been demonstrated in dividing cells (sea-urchin eggs) by Dr. Lyon, and we may infer its presence in the other rhythmical processes. The electrical rhythm indicates a rhythm of changing permeability, and of this there is some direct evidence in dividing egg cells. In all of these cases we have to do with automatic processes whose rhythm proceeds of its own accord, provided the external conditions remain normal. Each cycle in the rhythm furnishes itself the conditions for its own recurrence. The question arises: from what physico-chemical point of view is it best to regard this class of phenomena? In the case of a rhythmical contractile tissue three interdependent and synchronous rhythms may be distinguished—a chemical, a mechanical (presumably the expression of surface-tension changes), and an electrical. An elementary model of these phenomena is, I believe, furnished by the experiments of Bredig and his pupils on the rhythmical catalytic decomposition of hydrogen peroxide in contact with metallic mercury. When a ten per cent, solution of hydrogen peroxide is poured over the surface of pure mercury, a film of peroxidate at once forms over the surface of the metal. Its formation alters the surface-tension of the mercury by changing the potential-difference between the metal and the solution. Consequently, the form of the mercury-surface changes. Under appropriate conditions this deformation causes a mechanical rupture of the film at some portion of its surface; there follows on this an electrolytic decomposition of the peroxidate at the margin of the fissure, an effect which spreads over the whole surface and involves the dissolution of the film, and its reduction to metallic mercury, together with the liberation of oxygen. The film then reforms, and the process is repeated. Thus a regular rhythm, involving a form-change, a chemical decomposition, and a change of electrical polarization, is started and continues automatically. The rate of rhythm may be altered, just as in organic processes, by altering the chemical character of the medium, e. g., by changing its alkalinity, or by the addition of various other chemical substances. The velocity with which the film is laid down and dissolved may thus be influenced, and the whole rhythm correspondingly affected. Graphic records showing the variation in the rate of oxygen-liberation present a marked resemblance to the records of rhythmical organic processes like the heart-beat. Now the general conditions determining the rhythm in this phenomenon are strikingly like those which, on the foregoing theory of stimulation, determine the physiological rhythms. The surfacefilm of peroxidate may be compared to the plasma-membrane. Its rupture is equivalent to a local increase of permeability. This change is the direct condition both of the chemical change and of the electro-motor change, on which last depends the variation of surface-tension conditioning the form-change. While the living system is indefinitely more complex than the mercury-peroxide system, yet in its rhythmical character and in the essential nature of the controlling conditions this automatic rhythmical catalysis bears an undeniable and striking resemblance to the action of living tissues like the heart, in which a rhythmical autostimulation is the distinguishing characteristic. In both cases an alteration of a surface-film is the critical change; and the rate of this change determines the rate of the other rhythmical events of the cycle. We may infer that if we could control the condition of the plasma-membranes of cells we could control the entire range of cell-processes. But I do not wish to prejudge these questions; I make the above comparison chiefly in order to suggest possibilities, and to indicate the desirability of devoting more careful study to the surface-films of cells. Investigation of the conditions of their formation, their permeability and their physical and chemical nature is certain to lead to results of far-reaching importance for biology.

1. This consideration, however, is not demonstrative. The precipitation of lecithin by acid can be prevented by salts in concentrations which in themselves do not precipitate, as Handowsky and Wagner have recently shown. Lecithin, which seems always to be present in cells, probably forms an important part of the plasma membrane, in which case changes in its physical condition would influence the properties of the latter.
2. The assumption of a permeability-increase at the time of stimulation 1 is the only hypothesis, so far as I know, that accounts at once for the two characteristic and invariable accompaniments of stimulation, (1) the negative electrical variation, and (2) the temporary loss of irritability (refractory period) during the electrical variation. The time-relations of these two outwardly diverse phenomena coincide, as Tait has shown, and both are to be regarded as expressions or consequences of the same change, namely, a temporary increase in the permeability of the limiting membranes. This increase involves a temporary loss of the semi-permeability which is essential to the maintenance of the normal polarization of the membrane, and also—according to Nernst's theory—essential to electrical stimulation. I therefore regard the existence of a refractory period as furnishing strong support to the general theory of stimulation and conduction outlined above.