Protoplasmic Action and Nervous Action

The addition of traces of CaCL and SrCla to isotonic sugar solutions was found to delay the loss of irritabiHty, and a slight revival was observed when sugar-muscles were returned to weak solutions of Ca, Sr, Ba, or Mg salts in isotonic sugar solution; but no restorative action was observed in pure isotonic solutions of these salts. The results of experiments with nerve-trunks were on the whole similar to those with muscle,^ although on account of the structure of nerve a much longer immersion in sugar solution (in the cold) was necessary in order to abolish irritabiHty completely. The minimal concentration of Na salts for preserving irritability proved to be about the same as in muscle; on nerve as well as on muscle K salts have a characteristic paralyzing action, which is also antagonized by Ca and Sr in the presence of Na salts.

An important section of Overton's work has reference to the influence of salts on the transmission across the myoneural junctions and through reflex arcs. Locke^ (1894) had found that when the frog's sartorius, with nerve attached, was placed in a pure isotonic NaCl solution the tissue within fifteen to twenty minutes lost irritabihty through the nerve for single induction shocks, while it remained directly irritable for some hours; the addition of a little CaCla (0.02 per cent) to the solution restored indirect irritability in a few minutes; a second return to pure NaCl again aboHshed stimulation through the nerve, and the effect could be again reversed by CaCla. Muscles perfused with salt solution show the same phenomenon;^ and other cases of innervation, such as inhibition of the heart through the vagus, are similarly dependent on the salts of the medium.^

Overton^ found that the addition of KCl to the NaCl solution greatly accelerated the junctional paralysis; salts of Rb and Cs acted similarly, also salts of NH3 and organic ammonium compounds, which have long been known to exhibit this curare-like action. When the concentration of NaCl was 0.7 per cent, the addition of .05 per cent KCl was found to shorten the period of irritabihty through the nerve to an eighth or a tenth of its duration in the pure solution; with a lower concentration of NaCl less KCl was required. This blocking effect of potassium is antagonized by calcium; there is,

in fact, enough K in blood plasma to destroy all myoneural transmission, were it not for the Ca also present. Calcium is the only metallic ion in plasma (other than Na) required for myoneural transmission; in mixtures of NaCI and CaCL indirect irritabihty is preserved almost as long as in serum. Sr also antagonizes this action of K, but Ba and Mg do not. A similar blocking of myoneural transmission is caused by BaCl.; this action is also antagonized by Ca. Overton found that in a mixed solution of sodium and calcium chlorides of the same concentration as in serum (0.7 per cent NaCl plus 0.02 to 0.03 CaCla) the addition of 0.05 to 0.06 per cent KCl completely paralyzed the nerve-endings without affecting the direct irritability of the muscle; in order to abolish the latter 0.15 per cent KCl was required; when more CaCL was added to the solution more KCl was required to prevent transmission. In the absence of K a mere trace of Ca is all that is necessary; in a K-free NaCl solution one part of CaClz in 20,000 maintained indirect irritability for twenty-four hours and one part in 50,000 for twelve hours; even one part in 100,000 had a perceptible effect.

Although the nerve-trunk retains its power of conduction for many hours in pure solutions of Na and Li salts, transmission through reflex arcs is quickly prevented by lack of Ca.^ This was shown both in perfusion experiments with intact frogs and in experiments in which the isolated spinal cord with nerves and muscles attached was immersed in the solution. In such a preparation kept at a low temperature in well-oxygenated Ringer's solution reflex activity may continue for days. If the cord and attached sciatic nerves are immersed in isotonic sugar solution, reflexes completely disappear within twentyfour to thirty-six hours; on returning the cord to Ringer's solution they return in a few hours. In perfused frogs, replacement of the blood by sugar solution rapidly causes disappearance of reflexes and later of direct muscular irritability; perfusion with pure NaCl solution then restores direct muscular irritability, but not indirect or reflex, while Ringer's solution restores all three. The quantity of Ca necessary for restoring reflexes (as well as indirect irritability) is very slight.

In the observations above most of the fundamental phenomena of salt action are illustrated — the necessity of salts for cellular activity, the toxic action of pure solutions, ion-antagonisms, and especially the remarkable role of calcium in maintaining the normal structural and functional relationships between cells. With regard to this latter phenomenon, Overton cites Herbst's observations on the action of calcium in promoting the coherence of blastomeres,^ and also calls attention to the presence of this element in the middle lamella of plant tissues, where, according to Mangin, it is present as a compound, Ca-pectate (or pectinate), which is necessary for intercellular coherence. When this compound is removed (by weak acid) or when it is substituted by the Na salt (e.g., in pure NaCl solution) the cells fall apart.'' Apparently some Ca compound is necessary for the

' Compare the recent observations of Hansteen on plant tissues; again pure NaCl solutions cause a disintegration due to loss of intercellular coherence: Jahrb. wiss. Bolanik, XLVII (1910), 289; LIII (1914)} 53*^- intimate union between the nerve end-plate and muscle cell, or for the normal properties of the synaptic junctions. Overton suggests that when the tissue is transferred to pure NaCl solution, the Ca in this compound is replaced by Na, producing a compound of greater water-absorbing properties, which swells and interrupts the union. Such an effect is reversed by a return to Ca-containing solutions.

Some of the more general inferences to be drawn from these and recent experiments of a similar kind are as follows: Since irritability disappears in isotonic sugar solutions, which maintain the osmotic balance without apparently changing the crystalloid content of the cell, it seems clear that the action of the salts of the external medium must be superficial; i.e., is a surface-action exerted upon the plasma membrane. This action is probably of a twofold nature : first, the normal composition and physical properties of the colloidal . materials composing the surface-film are preserved only when a certain combination of ions is present in the medium; and, second, the normal state of electrical polarization of the membrane is also dependent on the presence of salts in the medium as well as in the cell-interior; this effect is important because the electrical polarization of the membrane is a factor determining its permeability. These two effects, however, cannot be regarded as independent, since changes in the physical state of the membrane must alter its permeability and (in so doing) its electrical polarization; and, conversely, changing the electrical polarization influences the permeability.

The essential conclusion, however, is that salts may exert physiological action without penetrating into the interior of the cell, solely by altering the state of the plasma membrane, and there is much additional evidence that this is the case. The analytical results of Urano and Fahr^ show that in normal vertebrate muscle Na is almost entirely absent from the cell interior, although, as just shown, it is essential to the maintenance of irritability. Urano kept frogs' sartorii for several hours in isotonic sugar solution, and found that the muscle gave off to the solution relatively much more Na than K. Muscles thoroughly extracted in sugar solution were found to contain only 2 per cent or less of the normal total Na- content, but nearly all of the K and phosphate. The conclusion seems certain that the Na is contained almost entirely in the intercellular spaces of the tissue, and the K and phosphate in the cells. Muscles kept for six hours in sugar solution lose very little ash that may not be accounted for by the salts present in the interstitial lymph. Fahr, in an ash-analysis of the extract of fresh uninjured muscles in isotonic sugar solution, found that only 6 per cent of the original K of the tissue, but 90 per cent of the original Na, was thus recoverable. If four-fifths of the total volume of the muscle be regarded as consisting of muscle cells, and one-fifth of interstitial tissue and lymph spaces, the Na-content of the tissue is completely accounted for by the salts present in the lymph. The latter must therefore exert their physiological influence through their action upon the external protoplasmic layer or plasma membrane.

Hober's experiments on the influence of isotonic solutions of neutral salts on the demarcation-current potential of muscle^ also support Overton's view that the normal uninjured resting muscle is impermeable to sodium salts. Solutions of sodium and lithium salts leave the normal electrical potential of the external muscle-surface unchanged; while salts like those of K, Rb and NH4 (which give other evidence of penetrating the muscle) produce an injury-current or local negativity. Absence of penetration is indicated by failure to change the external potential of the muscle; the alkali earth salts and in part those of caesium are thus indifferent. Nevertheless, they alter the properties of the tissue; thus Mg salts have a strongly anti-stimulating or narcotic action, and Ca and Sr salts in pure isotonic solution similarly render muscle resistant to stimulation; all of these effects are reversible if the exposures are not too prolonged.

Loss of contractility in non-electrolyte solutions and its prompt return in solutions of sodium salts — especially if some Ca is also present — are characteristic of many varieties of muscle. Among invertebrates the larvae of Arenicola show this phenomenon in a striking manner;^ here also, as with vertebrate muscle, Li and Na salts are alone capable of preserving contractility for prolonged periods, and K salts have a paralyzing action. Mg salts repress contractility very promptly and completely, and the action is readily reversed, especially by solutions of Na salts containing a little Ca.^ The resemblance of Li to Na in its power of maintaining the normal properties of muscle (though less perfectly than Na) seems to be general for both vertebrate and invertebrate muscle; for example. Mines describes this phenomenon in the

molluscan heart (Pecteny and in the heart of elasmobranch fishes. Much further evidence could be cited indicating that changes in the plasma membranes of the living cells, produced by salts or other compounds which show no evidence of penetrating into the cell interior, may profoundly affect the properties of irritable tissues or organisms. An interesting example is seen in the changes produced by isotonic salt solutions in the sensitivity of frog's muscle to various forms of chemical stimuH and to physical agents like heat or contact. A curarized frog's gastrocnemius, placed for a few minutes in a pure isotonic solution of a sodium salt (NaCl, NaBr, NaN03, Nal, NaC103, etc.) and then dipped into a solution containing a stimulating compound (K salt) or a cytolytic agent (chloroform in saturated solution), contracts much more vigorously than a normal muscle which is dipped into the same solution directly from Ringer's solution. The pure Na salt solution increases the responsiveness to the stimulating agent; i.e., induces a sensitization; this effect is readily antagonized by CaCla. The reverse effect (desensitization) is produced by exposure to isotonic CaCla, SrCla, or MgCL; after a brief stay in these solutions the muscle fails to respond to the stimulating solution or responds subnormally. In all such cases normal irritabiUty returns in Ringer's solution.^ Various facts indicate that Ca compounds in the surface layer of the cell are

concerned in normal irritability;^ a striking example is the marked hypersensitivity induced in frog's muscle and nerve by Na salts whose anions precipitate calcium or form Ca salts of limited solubility. Muscles treated for a few minutes with isotonic solutions of these salts (sulphate, phosphate, tartrate, citrate, oxalate, etc.) become highly sensitive to the contact of foreign substances or media, and contract vigorously when exposed to air.^ These phenomena of sensitization are produced so rapidly as to leave little doubt that they depend on alterations in the surface-films of the cells. ^

Direct evidence that changes in the permeability of the plasma membranes play an essential part in processes of stimulation will be cited later. Apparently all substances that alter the physical or chemical state of the plasma membrane (salts, acids, alkalis, narcotic agents) modify the stimulation-process; this influence may be in the direction either of facilitation or of repression, according to conditions."* The precise means by which salts produce these effects has been much debated and is still imperfectly

3 Other effects dependent on surface changes in the cells are related to those just described; e.g., changes in the surface-films of blood corpuscles and bacteria, affecting the critical concentrations of haemolysis or agglutination by H ions, are produced by the addition of proteins; cf. the recent papers of Coulter, Jour. Gen. Physiol., Ill (1921), 309, and IV (1922), 403; Northrop and De Kruif, ibid., IV (1922), 655; also Eggerth and Bellows, ibid., p. 669.

4 Such effects evidently imply an influence on cell metabolism. Warburg's observations on sea-urchin eggs show that the oxygen consumption may be increased several times by the addition of alkali which shows no evidence of penetrating the cell: Z. physiol. Chem., LXVI understood. Perhaps the clearest light on the problem has been afforded by the phenomena of salt-antagonism. These have shown definitely, at least in certain cases, that structural changes in those parts of the cell which are most directly exposed to the action of the solution (plasma membranes and other surface-structures like cilia) are the primary condition of the effects produced. The toxic and other effects are secondary consequences of these structural changes. The physical condition of the structural colloids — state of subdivision, of hydration, of electrical polarization, etc. — is changed by the action of the salt, with corresponding changes in the properties of the cell-structure itself and of the chemical and other activities controlled by it. In many cases the injurious action of the pure salt-solution (NaCl) is referable to a destruction of the nonnal semi-permeability of the plasma membrane; this effect, if not soon reversed, involves chemical and other disorganization followed by death. Any condition preventing or retarding the alterative action on the membrane, such as the presence of an antagonistic salt or non-electrolyte (narcotic), has accordingly a protective or ''anti-toxic" action.^ These general conditions are well illustrated in a simple marine organism much used in experimental work at Woods Hole, the larva of the annelid Arenicola cristata. This is a segmented trochophore larva about one-third of a millimeter in length, having pigmented body cells and swimming by a combination of muscular and ciliary movements. When placed in pure isotonic

' Cf. the discussion in my paper on antagonism between salts and anaesthetics, A^nerican Journal of Physiology, XXXI (1913), 255; cf. pp. 275 fif. NaCl (or similar unbalanced solution of alkali salt) the larvae contract strongly and the yellow pigment begins to diffuse from the cells; by degrees the cilia cease movement and undergo a visible breakdown or disintegration (suggesting liquefaction or absorption of water). In a solution of NaCl containing a little CaClz (95 volume m/2 NaCl plus 5 vols, m/2 CaCL) all of these immediate effects are prevented, and the larvae retain their normal appearance and behavior for some time and die much more gradually.^

In this case it is clear that the immediate action of the pure salt-solution is upon the surface structures — cilia and plasma membranes — which quickly lose their normal structural coherence and continuity; the result of this change in the cilia is physical breakdown, and in the plasma membrane a marked increase of permeability, hence the diffusion of soluble cell-constituents to the exterior and the progressive disorganization. In many other organisms and cells it can also be shown that an abnormal increase of permeability is produced by pure solutions of Na salts and prevented by the addition of CaCla or other antagonistic salt.^ The experiments of Osterhout on the electrical conductivity of Laminaria and other plant tissues afford perhaps the clearest evidence that the toxic action of pure NaCl solutions is the result of a destruction of semi-permeabihty and that antagonistic salts (CaCla and others) produce their anti-

^ In an. earlier paper {Biological Bulletin, XVII [1909J, 188) I have given a summary of observations showing the correlation between permeability-increasing action and toxicity for compounds other than salts. toxic effects by counteracting the structural change in the membrane.^ The question of just how, in the physico-chemical sense this result is accomplished is a fundamental one; salt antagonism is shown by all groups of organisms from bacteria^ to vertebrata, and is apparently a universal phenomenon in living protoplasm. It should be noted that other effects produced by the pure Na salt solution are also antagonized by Ca and other salts; e.g., its stimulating action, shown in the production of twitches in vertebrate and other muscle,^ the sensitizing action on muscle just described, and the activation of unfertilized eggs."* The fact that prevention (by Ca and narcotics) of increase of permeability in various irritable tissues and organisms is associated with prevention of stimulation has a general interest as evidence of the essential part played by membranes in stimulation, and will be considered more fully later.

The problem of the physico-chemical basis of salt antagonism has been approached in various ways and ^ Osterhout, loc. cit. Related observations are those of Hamburger on the action of calcium in preventing the increase in permeability produced in frog's kidneys perfused by NaCl solutions containing dextrose (cf. Hamburger, Biochem. Zeitschrift, LXXXVIII [1918], 97). In 1904 J. B. Mac Galium had observed the antidiuretic action of calcium salts and had related it to the action in decreasing permeability (Journal of Experimental Zoology, I [1904J, 179; University of California Publications, Physiol., II [1905], 93). It is well known that calcium antagonizes cytolysis by saponin and other permeability-increasing compounds. Chiari's observations on the formation of exudates and many other observations are related; cf. Hober, op. cit., pp. 544 ff.

^ For the case of bacteria cf. Shearer, Journal of Hygiene, XVIII on the whole most effectively by comparing the action of salts on colloidal solutions and emulsions with their action on living cells.' The physiological effects produced by salts and salt combinations vary with the nature of the ions in a manner which in many cases shows a close parallelism with the physical effects produced by the same salts in simple colloidal systems. For example, in the counteraction of the toxic effects of pure Na salt solutions the results all indicate that the cation of the antagonistic salt is the effective agent; and in this case certain relations highly characteristic of the action of ions on colloidal systems are shown clearly; thus with the cilia of Arenicola all salts of bivalent heavy metals (Co, Ni, Cd, Zn, Mn, Pb, Fe++, Cu) were found to produce their maximal antitoxic effects in concentrations of the order m/400 to m/i6oo; while with trivalent metals (Al, Cr, Fe^'O the physiologically corresponding concentrations were from 50 to 100 times less.^ This increase in effectiveness with increase in valence is characteristic of the action of salts on negatively charged suspensoid systems (rule of Schulze and Hardy); and apparently the observations above indicate that the cations produce their effects in the living system by influencing the state of subdivision of the suspensoid colloids forming part of the protoplasmic structure. Other cases of rapid increase in physiological effectiveness with increase of valence are well known.

Such cases of antagonism are consistent with the hypothesis that the normal state of the living protoplasmic structure (e.g., a cilium or plasma membrane) requires a certain state of subdivision of the chief structural colloids. Apparently in the normal medium, e.g., sea water, the influence of the various ions is so balanced as to preserve this state; the collective effect of the negative charges carried by the anions (CI, SO4, HCO3, etc.) is just balanced by the collective effect of the positive charges carried by the cations (Na, K, Ca, Mg). In the pure solution, however, of NaCl, in which no bivalent cations are present, the influence of the anions is insufficiently compensated, and the colloids are altered in a manner injurious to cell structure: i.e., a preponderant influence of anions is the essential toxic factor in such cases. The addition of bivalent or trivalent' cations of any kind has under these conditions a compensating and hence antagonistic or antitoxic influence; trivalent ions exercise this effect in much lower dilution than bivalent ions, and bivalent than monovalent ions. At an appropriate concentration, the balance is restored, protoplasmic structure remains normal, and life continues. The reverse condition, in which the toxic action of the pure solution results from a prepotency of cation action, and in which accordingly the addition of small quantities of salts with powerfully acting anions has an antitoxic effect, is apparently also realized in some cases; e.g., the ciliated epithelium of the molluscan gill {Mitylus) in solutions of SrCla; here various Na salts (NaOH, NaBr, NaT, NaCNS, NaaSOj show well-marked antagonistic action.^ Such purely physical explanations, while apparently partly applicable in some cases, prove inadequate in

many others, where the foregoing relation of antagonism to valence does not appear to hold, and especially in those numerous cases where different cations similar in valence have widely different antagonistic or other physiological actions.^ In such cases the special chemical properties of the compounds formed between the ions and the structural colloids of the cell must apparently be taken into consideration. Thus various facts indicate that calcium proteinates or calcium soaps (or both) are of special importance as constituents of cell structures; and these compounds cannot be replaced satisfactorily by the corresponding compounds of other bivalent elements.^ Such a conception explains why strontium, the metal most closely resembling calcium in general chemical properties, usually comes nearest to calcium in antagonistic effectiveness; the other alkali earth cations come next; while the bivalent heavy metals are effective in only a few special cases (Mn. Co, Fe++), the great majority of these metals being too strongly toxic to act as efhcient antagonists.^

^ Cf. Hober's discussion in his recent papers on the physiological action of calcium; Arch. ges. Physiol., CLXVI (191 7), 531, and 2 The cases are numerous where calcium has a specific action which cannot be replaced by that of other cations, even Sr; instances are the specificity of calciimi in phagocytosis (Hamburger), in the inhibition of swimming plates, and in the preservation "of contractility in heartstrips. On the other hand, in its simple stabilizing or protective action, (e.g., in cytolysis, etc.) Ca can often be replaced by other cations (cf. Hober's papers, loc. cit.; cf. also Wiechmann, Arch. ges. Physiol., CXCV

3 Nevertheless even the most toxic cations, like Ag and Hg, may show definite antitoxic action in certain cases; e.g., the beat of the ctenophore swimming plate in pure isotonic NaCl solution is prolonged several times by AgNOj and HgCL in concentrations of m/ 100,000 to m/ 200,000 (cf. American Journal of Physiology, XVI [1906], 117). The evidence, taken as a whole, indicates that the salts act chiefly by altering the physical state of the structural colloids of the cell; and it is to be presumed that general physical factors (of the kind regarded as acting in all colloidal phenomena) and special chemical factors specific for each form of protoplasm are both concerned in producing the total effect. In the directly toxic or injurious action of salt solutions, a frequent, if not invariable, factor is a destruction of the semipermeable properties of the protoplasmic partitions, primarily of the plasma membranes. In physiological salt-actions of other kinds, e.g., stimulation, sensitization, inhibition, it is to be presumed that the physical properties of the plasma membranes undergo special modifications of a corresponding kind, but that the eft'ects do not exceed a certain range and are therefore reversible. The general fact that a certain combination of salts, usually of Na, Ca, and K, is required in the external medium for most forms of normal protoplasmic action indicates the fundamental importance of the influence of salts on the structural colloids of protoplasm. Apparently the structural conditions required for the continuity and closeness of texture necessary in semi-permeable membranes depend on the maintenance of a definite equilibrium between the ions in the medium and those associated (chemically or otherwise) with the structural coUoids. Shght changes in the salt-content imply corresponding structural changes, with dependent physiological effects.

It should be added that other general physical conditions, also dependent on the presence of salts, are of importance in the normal activity of protoplasm. such as the electrical conductivity of protoplasm and medium and the electrical polarization of the plasma membranes. These conditions will be considered later under the subject of stimulation. The special relations of the three chief cations of the protoplasmic media, Na, K, and Ca, to protoplasmic activity appear to depend on chemical conditions of a kind still imperfectly understood. The differences between the physiological actions of Na (and Li) and of K (and Rb and Cs) in their relation to vertebrate muscle and nerve cannot be satisfactorily explained on the basis of the physico-chemical constants of these ions. Na and K, though chemically closely similar, appear in many forms of protoplasm, e.g., vertebrate muscle, to act as physiological antagonists; in others their physiological differences are relatively slight; e.g., in the case of fish eggs (Fundulus) and sea-urchin eggs (Arbacia) isotonic KCl solution is even less toxic than NaCl solution; both solutions are antagonized by Ca.

It is remarkable that the striated muscle cells of vertebrata appear to be readily permeable to some K salts (KCl, KBr, KI, KNO3) but not to others (K2SO4, K- tartrate, K-phosphate) ;^ in this respect K salts exhibit a striking contrast to Na salts. The special permeabihty to KCl is shown in frog's muscle by rapid increase in the weight of the tissue when it is immersed in isotonic solutions of this salt; similar conditions have been found by Seebeck^ in the frog's kidney, which, like muscle, shows normal semi-permeability toward Na and Li salts. The special physiological action of K is prob-

ably related to this peculiarity, but the nature of the relationship is unknown. Recently Zwaardemaker has attributed importance to the slight radioactivity of K, on the ground of certain striking parallels between the action of K salts and uranium salts on the frog's heart ;^ but other evidence from experiments with marine organisms fails to support this view.^ Recently Meigs^ has investigated the osmotic behavior of smooth muscle in solutions of salts and non-electrolytes, and finds a number of remarkable differences between this tissue and striated muscle. For example, the stomach muscle of the frog gains weight when immersed in isotonic solutions of sugar, alanin, or NaCl, and even in Ringer's solution. He concludes that smooth muscle cells differ fundamentally in their structure and mode of action from striated muscle cells, and that they are not surrounded by semi-permeable membranes. In the adductor muscle of the marine clam {Venus mercenaria), sl smooth muscle which reacts slowly and maintains its state of tension for a long time, closely similar conditions were found. Pieces of muscle immersed in sea water take up chloride from the latter and increase in weight; they even fail to lose weight in lo per cent NaCl or in sea water of twice the normal concentration. "*

^ R. F. Loeb, Journal of General Physiology, III (1920), 229. A. J. Clark fails to confirm Zwaardemaker's claim that uranium can act as a substitute for K in restoring the heartbeat (Journal of Pharmacology and Experimental Therapeutics, XVIII [1922], 423), In explanation of these apparent discrepancies it has been suggested that smooth muscle cells are easily injured and that the normal semi-permeability had disappeared at the time when the tissue was examined, but this possibility is rejected by Meigs. It is difficult to escape the conclusion that a continuous semi-permeable membrane does not invest the whole cellular element in this tissue. Possibly the contractile part of the entire structure has a relation to the protoplasmic part similar to that which the fibers in connective tissue have to the cells by which they are formed. The sluggishness of the movement and its slow reversibility suggest a fundamentally different type of organization from that of striated muscle cells. Semi-permeability appears to be universal in indifferentiated cellular elements and in the majority of specialized cells; but a differentiation in which part of the total structure acquires non-cellular properties is not infrequent in organisms, as illustrated in connective tissues and skeletal structures; and the above-cited types of contractile tissue may exemplify this general condition.

The penetration of salts through the egg-membrane of the sea-minnow, Fundulus, as shown in various toxic effects and antagonisms, illustrates many conditions of great interest, and especially indicates the importance of the purely chemical factors in permeability. This membrane is a dead structure, or chorion, external to the living protoplasm of the egg, and apparently consisting chiefly of a keratin-like protein. In its normal state, it is almost completely impermeable to water or the salts of the medium; hence the eggs will develop either in distilled water or in concentrated sea water.